Does CompStat Reduce Crime?

CASE STUDY
CompStat (short for COMPuter STATistics or
COMParative STATistics) originated in the
New York City Police Department (NYPD)
in 1994 when William Bratton was police
commissioner. CompStat is a comprehensive,
city-wide database that records all reported crimes or
complaints, arrests, and summonses issued in each
of the city’s 76 precincts. City officials had previously
believed that crime could not be prevented by better
information and analytical tools but instead by using
more foot patrols in neighborhoods along with the
concept of “community policing” in which efforts
were made to strengthen the involvement of community groups. In contrast, Bratton and Rudy Giuliani,
then the mayor of New York City, believed that
police could be more effective in reducing crime if
operational decisions took place at the precinct level
and if decision makers had better information.
Precinct commanders were in a better position than
police headquarters to understand the specific needs
of the communities they served and to direct the
work of the 200 to 400 police officers they managed.
CompStat gave precinct commanders more authority
and responsibility, but also more accountability.
At weekly meetings, representatives from each
of the NYPD’s precincts, service areas, and transit
districts are put on the “hot seat” at police
headquarters and required to provide a statistical
summary of the week’s crime complaint, arrest,
and summons activity, as well as significant cases,
crime patterns, and police activities. Commanders
must explain what has been done to reduce crime
in the districts under their command, and if crime
has gone up, they must explain why. Commanders
are held directly accountable for reducing crime in
their area of command. In the past, they were
evaluated primarily on the basis of their administrative skills, such as staying within budget and
deploying resources efficiently.
The data these commanders provide, including
specific times and locations of crimes and
enforcement activities, are forwarded to the NYPD’s
CompStat Unit where they are loaded into a
city-wide database. The system analyzes the data
and produces a weekly CompStat report on crime
complaint and arrest activity at the precinct, patrol
borough, and city wide levels. The data are summarized by week, prior 30 days, and year-to-date for
comparison with the previous year’s activity and
for establishing trends. The CompStat Unit also
issues weekly commander profile reports to
measure the performance of precinct commanders.
The weekly commander profile reports include
information on the commander’s date of appointment, years in rank, education and specialized
training, most recent performance evaluation rating, the units that person previously commanded,
the amount of overtime generated by police under
that commander, absence rates, community demographics, and civilian complaints.
Using MapInfo geographic information system
(GIS) software, the CompStat data can be displayed
on maps showing crime and arrest locations, crime
“hot spots,” and other relevant information.
Comparative charts, tables, and graphs can also be
projected simultaneously. These visual presentations help precinct commanders and members of
the NYPD’s executive staff to quickly identify
patterns and trends. Depending on the intelligence
gleaned from the system, police chiefs and captains
develop a targeted strategy for fighting crime, such
as dispatching more foot patrols to high-crime neighborhoods, or issuing warnings to the public when a
particular model of vehicle is susceptible to theft.
During Bratton’s 27-month tenure, serious crime
in New York dropped by 25 percent and homicides
went down by 44 percent. Crime in New York City
has dropped by 69 percent in the last 12 years.
Skeptics do not believe that CompStat was responsible for these results. They point to the decline in
the number of young, poor men, an improved
economy, programs that reduced welfare rolls
while giving poor people access to better housing,
increasing the size of the NYC police force, and giving precinct commanders more decision-making
responsibility and accountability.
Nevertheless, Bratton, convinced that CompStat
was the catalyst for New York’s drop in crime,
implemented the system in Los Angeles to further
prove its worth. Since the introduction of
CompStat, combined violent and property crimes
in Los Angeles dropped for six consecutive years.
Yet the ratio of police officers to residents is only
half that of New York and Chicago. CompStat hasalso been adopted in Philadelphia, Austin, San
Francisco, Baltimore, and Vancouver, British
Columbia.
Skeptics point out that crime has fallen in all
urban areas in the United States since 1990
regardless of whether the cities used CompStat. In
fact, a critical study of CompStat by the Police
Foundation found that CompStat encouraged police
to be only reactive rather than pro-active in fighting
crime. Sending police to where crime has become a
problem is, in other words, too late. CompStat
encouraged what the Police Foundation called
“whack-a-mole” theory of policing, similar to the
game played in amusement parks. Rather than
change police departments into nimble crime
fighters, the Foundation found that a database had
been attached to traditional organizations, which
themselves remained unchanged.
Because of the emphasis placed on reducing
crime and because of the newfound importance of
crime statistics to officers’ careers, CompStat has
created pressure on some precinct commanders to
manipulate crime statistics to produce favorable
results. Officers must continue to improve their
crime statistics, despite shrinking budgets and
dwindling numbers of officers. A study conducted
in 2009 via a questionnaire given to 1,200 retired
police captains and more senior officers concluded
that nearly a third of respondents were aware of
unethical manipulation of crime data.
More than 100 survey respondents said that
intense pressure to produce annual crime reductions led some supervisors and precinct
commanders to manipulate crime statistics. For
example, officers were known to check catalogs,
eBay, and other sites for items similar to those
reported stolen, looking for lower prices they could
use to reduce the values of the stolen goods for
record-keeping purposes. Grand larceny, a felony,
is considered to be theft of goods valued at $1,000
or more, whereas theft of goods valued at less than
$1,000 is only a misdemeanor. Using this method,
precincts could reduce the number of felony thefts,
considered an “index crime” and tracked by
CompStat. Surveys and anecdotal evidence also
indicated a lack of receptiveness on the part of
police in some areas, possibly motivated by a
desire to reduce the number of crime incidents
reported.
Some survey respondents stated that precinct
commanders or aides dispatched to crime scenes
sometimes tried to persuade victims not to file
complaints or urged them to change their accounts
of what happened in ways that could downgrade
offenses to lesser crimes.
Previous studies of CompStat encountered an
unwillingness by the NYPD to disclose their data
reporting methods. A professor performing a study
that ultimately praised CompStat’s influence on
crime in New York City was given full access to
NYPD crime data, but the NYPD did not cooperate
with the Commission to Combat Police Corruption
(CCPC), an independent board that monitors police
corruption. The commission sought subpoena power
to demand the NYPD turn over its data and data collection procedures to uncover potential wrong doing
by the police. Unfortunately, the commission was
denied access to this data after strong police department opposition.
On the other hand, versions of CompStat have
been adopted by hundreds of other police departments across the United States, and the CompStat
approach has been credited with improving police
work in many cities. In New York City itself, much
of the public believes that crime is down, and that
the city has become a safer and more pleasant place
to live.
Sources: William K. Rashbaum, “Retired Officers Raise Questions
on Crime Data,” The New York Times, February 6, 2010; A.G.
Sulzberger and Karen Zraick, “Forget Police Data, New Yorkers
Rely on Own Eyes, The New York Times, February 7, 2010; Luis
Garicano, “How Does Information Technology Help Police
Reduce Crime?” TNIT Newsletter 3 (December, 2009); and New
York City Police Department, “COMPSTAT Process,”
www.nyc.gov/html/nypd/html, accessed October 9, 2006.
CASE STUDY QUESTIONS
1. What management, organization, and technology
factors make CompStat effective?
2. Can police departments effectively combat crime
without the CompStat system? Is community policing incompatible with CompStat? Explain your
answer.
3. Why would officers misreport certain data to
CompStat? What should be done about the misreporting of data? How can it be detected?