Depolicing

During the 1990s, police officers in the United States came under attack for the alleged practice of racial profiling. Activists accused law enforcement of singling out African-American men for stops and searches on the basis of race. The term racial profiling was used by civil liberties groups to describe this allegation. The decade also witnessed several high-profile police shootings of civilians, primarily minorities, which were blamed on officer bias as well. As a result of the intense criticism that accompanied these incidents, many officers backed off of assertive policing—a phenomenon known as depolicing. In many jurisdictions where the charge of racial profiling was particularly vitriolic, the result was a dramatic decrease in stops and arrests with a corresponding dramatic increase in crime.

The 1990s will be remembered in law enforcement circles as the decade of racial profiling, a term used by the American Civil Liberties Union and others to refer to the alleged police practice of stopping civilians according to race.

For the media, racial profiling became the very hallmark of policing, even though sound statistical evidence for it is nonexistent.

Profiling studies typically compare police enforcement data, such as the number of stops, searches, or arrests, to census data. If the rates of stops or searches for minorities are disproportionate to their share of a population, critics conclude that the police use skin color, not behavior, in deciding whom to stop. For many in the law enforcement community and social sciences, such an analysis is flawed for two reasons: First, it assumes that law breaking is spread evenly across the population, and second, it omits every factor that, unlike race or census data, actually does determine whom the police approach.

The most important benchmark for aggregate enforcement data is crime rates. Police departments target their resources where crime is highest, and that is usually in inner-city minority neighborhoods. In Chicago, for example, homicide rates from 1991 to 2001 were 11 times higher in black beats than in white beats. Such a disparity means that antiviolence initiatives will take place overwhelmingly in black neighborhoods. Once deployed in those areas, police officers look for behavioral cues of criminal-ity—nervousness, furtive gestures, resemblance to a suspect, or hitching up a waistband as if concealing a gun, among other possible factors. Racial profiling would be irrelevant, because the vast majority of residents are of the same race.

As long as crime rates remain racially disproportionate, so will the enforcement data that police critics seize to show racial bias. In 1998, for example, black males between the ages of 14 and [Page 134]24 committed 30% of all homicides, even though they made up only slightly more than 1% of the population. Only by ignoring law breaking could the police produce enforcement data that would match the nation's demographic profile—the litmus test for avoiding racial profiling accusations.

But although crime rates are the necessary precondition for sound racial profiling analysis, they are not sufficient. Also essential are the demographics of highways and city streets, which change hourly and daily; the distribution of high-profile crimes that would lead to high-intensity policing; age demographics of various populations; and the prevalence of parolees and warrant absconders in different racial groups, to name just a few of the factors that bear on enforcement data. To benchmark highway stops and searches, the analyst needs to know, at a bare minimum, rates of speeding and equipment violations, as well as of drug trafficking.

...