Consider the following case study in the complex interaction of race and law enforcement. An officer from the Drug Enforcement Administration stops and questions a young man who has just stepped off a flight to Kansas City from Los Angeles. The officer has focused on this man for several reasons. Intelligence reports indicate that black gangs in Los Angeles are flooding the Kansas City area with illegal drugs, and the man in question was on a flight originating in Los Angeles. Young, toughly dressed, and appearing very nervous, he paid for his ticket in cash, checked no luggage, brought two carry-on bags, and made a beeline for a taxi upon deplaning. Oh, and one other thing: the officer also took into account the fact that the young man was black. When asked to explain himself, the officer declares that he considered the individual's race, along with other factors, because doing so helps him efficiently allocate the limited time and other resources at his disposal.
How should we evaluate the officer's conduct? Should we applaud it? Permit it? Prohibit it? As you think through this example, be aware that it is not a hypothetical one. Encounters like this take place every day, all over the country, as police attempt to battle street crime, drug trafficking, and illegal immigration. And this particular case study happens to be the fact pattern presented in a federal lawsuit of the early '90s, United States v. Weaver, in which the U.S. Court of Appeals for the Eighth Circuit upheld the constitutionality of the officer's action.
"Large groups of our citizens," the court declared, "should not be regarded by law enforcement officers as presumptively criminal based upon their race." The court went on to say, however, that "facts are not to be ignored simply because they may be unpleasant." According to the court, the circumstances were such that it made sense for the officer to regard blackness, when considered in conjunction with the other factors, as a signal that could be legitimately relied upon in the decision to approach and ultimately detain the suspect. "We wish it were otherwise," the court maintained, "but we take the facts as they are presented to us, not as we would like them to be." Other courts have agreed with the Eighth Circuit that the Constitution does not prohibit police from routinely taking race into account when they decide whom to stop and question, as long as they do so for purposes of bona fide law enforcement (not racial harassment) and as long as race is one of several factors that they consider.
These judicial decisions have been welcome news to the many police officers and other law enforcement officials who consider the racial selectivity of the sort deployed by the DEA agent an essential weapon in the war on crime. Such defenders of what has come to be known as racial profiling maintain that, in areas where young African American males commit a disproportionate number of the street crimes, the cops are justified in scrutinizing that sector of the population more closely than others--just as they are generally justified in scrutinizing men more closely than women. As Bernard Parks, chief of the Los Angeles Police Department, explained to Jeffrey Goldberg of The New York Times Magazine: "We have an issue of violent crime against jewelry salespeople.... The predominant suspects are Colombians. We don't find Mexican-Americans, or blacks, or other immigrants. It's a collection of several hundred Colombians who commit this crime. If you see six in a car in front of the Jewelry Mart, and they're waiting and watching people with briefcases, should we play the percentages and follow them? It's common sense."
For cops like Parks, racial profiling is a sensible, statistically based tool that enables them to focus their energies efficiently for the purpose of providing protection against crime to law-abiding folk. To borrow a concept from economics, it lowers the cost of obtaining and processing information, which in turn lowers the overall cost of doing the business of policing.
Moreover, the very fact that a number of cops who support racial profiling are black, like Parks, buttresses their claims that the practice isn't motivated by bigotry. Indeed, these police officers note that racial profiling is race-neutral in that various forms of it can be applied to persons of all races, depending on the circumstances. In predominantly black neighborhoods and other places in which white people stick out in a suspiciously anomalous fashion (as potential drug customers or racist hooligans, for example), whiteness can become part of a profile. In the southwestern United States, where Latinos often traffic in illegal immigrants, apparent Latin American ancestry can become part of a profile. In a Chinatown where Chinese gangs appear to dominate certain criminal rackets, apparent Chinese ancestry can become part of a profile. Racial profiling, then, according to many cops, is good police work: a race-neutral, empirically based, and, above all, effective tool in fighting crime.
But the defenders of racial profiling are wrong. This, in itself, is not a particularly original claim. Indeed, ever since the Black and Latino Caucus of the New Jersey State Legislature held sensational hearings a few months ago, complete with testimony from victims of the New Jersey State Police force's allegedly overly aggressive racial profiling, the air has been thick with public denunciations of the practice. In June, at a forum organized by the Justice Department on racial problems in law enforcement, President Clinton condemned racial profiling as a "morally indefensible, deeply corrosive practice." Vice President Al Gore has promised that, if elected president, he would see to it that the first civil rights act of the new century would end racial profiling. His rival for the Democratic nomination, Bill Bradley, has countered that Gore should prepare an executive order and ask the president to sign it now.
Unfortunately, though, many who condemn racial profiling do so without really thinking the issue through. One common complaint about racial profiling is that using race (say, blackness) as one of several factors in selecting targets of surveillance is fundamentally and necessarily racist. But racial selectivity of this sort can be defended on nonracist grounds and is, in fact, embraced by people who are by no means anti-black bigots and are not even cops. Even Jesse Jackson once revealed himself to be an amateur racial profiler. "There is nothing more painful to me at this stage in my life," he said in 1993, "than to walk down the street and hear footsteps and start to think about robbery and then look around and see somebody white and feel relieved." The reason Jackson felt relief was not that he dislikes black people. He felt relief because he estimated, probably correctly, that he stood a somewhat greater risk of being robbed by a black person than by a white person.
A second standard criticism of racial profiling involves a blanket denial of the central empirical claim upon which the practice rests: that in certain jurisdictions individuals associated with particular racial groups commit a disproportionate number of the crimes. But there's no use pretending that blacks and whites commit crimes (or are victims of crime) in exact proportion to their respective shares of the population. Statistics abundantly confirm that African Americans--and particularly young black men--commit a dramatically disproportionate share of street crime in the United States. This is a sociological fact, not a figment of the media's (or the police's) racist imagination. In recent years, for example, victims of crime report blacks as the perpetrators in around 25 percent of the violent crimes suffered, although blacks constitute only about twelve percent of the nation's population.
So, if racial profiling isn't necessarily bigoted, and if the empirical claim upon which the practice rests is sound, why is it wrong?
The argument begins with an insistence upon the special significance of racial distinctions in American life and law. Racial distinctions are and should be different from other lines of social stratification. That is why, since the civil rights revolution of the 1960s, courts have typically ruled--pursuant to the Fourteenth Amendment's equal protection clause--that mere reasonableness is an insufficient justification for officials to discriminate on racial grounds. In such cases, courts have generally insisted on applying "strict scrutiny"--the most intense level of judicial review--to the government's actions. Under this tough standard, the use of race in governmental decision-making may be upheld only if it serves a compelling government objective and only if it is "narrowly tailored" to advance that objective. Strict scrutiny embodies the recognition, forged in the difficult crucible of American history, that the presence of a racial factor in governmental decisionmaking gives rise to the presumption that officials may be acting in violation of someone's civil rights.
A disturbing feature of the debate over racial profiling is that many people, including judges, are suggesting that decisions distinguishing between persons on a racial basis do not constitute unlawful racial discrimination when race is not the sole consideration prompting disparate treatment. The court that upheld the DEA agent's detainment of the young black man at the Kansas City airport declined to describe the agent's action as racially discriminatory and thus evaded the requirement of subjecting the government's action to strict scrutiny. More recently, as Goldberg showed in his New York Times Magazine article, New Jersey Governor Christine Todd Whitman has been willing to denounce as wrongful racial discrimination only racial profiling in which "race is the only factor." For Whitman, when race is just one of a number of factors, the profiling ceases to be "racial" and becomes instead a defensible technique in which a police officer merely uses "cumulative knowledge and training to identify certain indicators of possible criminal activity." This dilution of the meaning of discrimination is troubling not only because it permits racial profiling to continue without adequate scrutiny. Even worse, this confusion will likely seep into other areas of racial controversy, causing mischief along the way.
Few racially discriminatory decisions are animated by only one motivation; they typically stem from mixed motives. For example, an employer who prefers white candidates to black candidates--except black candidates with clearly superior experience and test scores--is engaging in racial discrimination, even though race is not the only factor he considers (since he is willing to select black superstars). There are, of course, different degrees of discrimination. In some cases, race is a marginal factor; in others it is the only factor. The distinction may have a bearing on the moral or logical justification for the discrimination. But it cannot logically negate the existence of racial discrimination. Taking race into account at all means engaging in racial discrimination.
Because racial discrimination is discouraged by both law and morality, proponents of racial profiling should bear the burden of persuading the public that such discrimination is justifiable. Instead, defenders of racial profiling frequently neglect the costs of the practice. They unduly minimize (or ignore altogether) the large extent to which racial profiling constantly adds to the sense of resentment felt by blacks of every social stratum toward the law enforcement establishment. Ironically, this is a cost of racial profiling that may well hamper law enforcement. In the immediate aftermath of O.J. Simpson's acquittal, when blacks' accumulated anger at and distrust of the criminal justice system became frighteningly clear, there existed a widespread recognition of the danger that threatens all Americans when cynicism and rage suffuse a substantial sector of the country. Alienation of that sort gives rise to witnesses who fail to cooperate with the police, citizens who view prosecutors as "the enemy," lawyers who disdain the rules they have sworn to uphold, and jurors who yearn to "get even" with a system that has, in their eyes, consistently mistreated them. For the sake of better law enforcement, we need to be mindful of the deep reservoir of anger toward the police that now exists within many racial minority neighborhoods. Racial profiling is a big part of what keeps this pool of accumulated rage filled to the brim.
Yet the courts have not been sufficiently mindful of this risk. In the course of rejecting a 1976 constitutional challenge to actions by officers of the U.S. Border Patrol who selected cars for inspection in Southern California partly on the basis of some drivers' apparent Mexican ancestry, the Supreme Court pointed to what it viewed as positive results. The Court noted that, of the motorists passing the checkpoint involved, fewer than one percent were stopped for questioning. It also noted that, of the 820 vehicles inspected pursuant to the profiling during the eight days surrounding the challenged arrests, roughly 20 percent contained illegal aliens.
As Justice William J. Brennan noted in dissent, however, the Court provided no indication of the ancestral makeup of all of the persons stopped in conformity with the Border Patrol profile. It is likely that a large percentage of the innocent people who were stopped and questioned were persons of apparent Mexican ancestry who found themselves in the position of having to prove their obedience to the law simply because others of their national origin have engaged in misconduct.
The burden placed on innocent people stopped by law enforcement officers because of racial profiling is typically underestimated. In the case of the Border Patrol, the Supreme Court maintained that the agents' intrusion on those selected for questioning is "quite limited," involving "only a brief detention of travelers during which all that is required ... is a response to a brief question or two and possibly the production of a document."
There is reason, however, to be skeptical of this upbeat portrait of quick and courteous police intervention. The justices seemed to forget that people who look Mexican and live in border regions are what game theorists call "repeat players." Their national origin, actual or apparent, remains the same long after the first time they get pulled over. Unlike Anglos, Mexicans and Mexican-Americans must contemplate not just the possibility of one or two stops in their lifetimes, but many.
Moreover, everyone involved in such an encounter knows that race played a role in the officer's decision to stop the car, which sets up a downward spiral in relations between the Border Patrol and Latinos. Officers who start out doing their duty courteously will encounter people who resent having been stopped in part because of their racial or national heritage. The people stopped will vent their resentment. The officer will respond in kind, which will provoke the person in the car further. Next thing you know, there's a violent incident. And don't forget that the cops are repeat players in this game, too. How courteous--and how sincerely nonracist--can we expect them to be after a few months of such hassles?
My case against racial profiling concludes on a frankly ideological note. Racial profiling undercuts a good idea that needs more support from both society and the law: that individuals should be judged by public authority on the basis of their own conduct and not on the basis--not even partly on the basis--of racial generalization. Race-dependent policing retards the development and spread of such thinking; indeed, it encourages the opposite tendency.
What about the fact that in some jurisdictions it is demonstrable that people associated with a given racial group commit a disproportionately large number of the crimes? Our commitment to a just social order should prompt us to end racial profiling even if the generalizations on which the technique is based are buttressed by empirical evidence. This is not as unusual as it may sound. There are actually many contexts in which the law properly enjoins us to forswear the playing of racial odds even when doing so would advance certain legitimate goals.
For example, public opinion surveys have established that blacks tend to be more distrustful than whites of law enforcement. Thus, for purposes of convicting certain defendants, it would be rational--and not necessarily racist--for a prosecutor to use race as a factor in seeking to exclude black potential jurors. Fortunately, the Supreme Court has outlawed racial discrimination of this sort. Similarly, it is a demographic fact that whites tend to live longer than blacks. Therefore, it would be perfectly rational for insurers to charge blacks higher life-insurance premiums than whites. Fortunately, though, the law forbids that, too. And, given that, statistically, whites tend to be better educated than blacks, it might make business sense for an employer to give a racial edge to white applicants. But a battery of laws proscribes racial discrimination in the workplace, even under circumstances in which it would strengthen a business's bottom line.
The point here is that racial equality, like all good things in life, costs something; it does not come for free. Politicians often speak as if all that Americans need to do in order to attain racial justice is forswear bigotry. They must do that. But they must do more as well. They must be willing to demand equal treatment before the law even under circumstances in which unequal treatment is plausibly defensible in the name of nonracist goals. They must even be willing to do so when their effort will be costly.
Since abandoning racial profiling would undeniably raise the information costs of policing to some extent, with some attendant potential loss in effective crime control, those of us who would do away with it must advocate a responsible alternative. Mine is simply to spend more on other means of enforcement--and then spread the cost on some nonracial basis. This is hardly infeasible. One possibility is hiring more police officers. Another is subjecting everyone to closer surveillance. A benefit of the second option would be to acquaint more whites with the burden of police intrusion, the knowledge of which might prompt more whites to insist upon reining the police in. As it stands now, this burden falls with unfair severity upon minorities--imposing on Mexican-Americans, blacks, and others a special kind of tax for the war against illegal immigration, drugs, and other forms of criminality. The racial character of that tax should be repealed. I am not saying that police should never be able to refer to race. If a young white man with blue hair robs me, the police should certainly be able to use the description of the perpetrator's race in efforts to apprehend the felon. In this situation, though, whiteness is a trait linked to a particular person with respect to a particular incident. It is not a free-floating proxy for risk that hovers over young white men practically all the time--which is the predicament in which young black men currently find themselves. Nor am I saying absolutely that race could never be legitimately relied upon as a signal of increased danger. In an extraordinary circumstance in which plausible alternatives appear to be absent, officials might appropriately feel bound to resort to racial profiling. This would be right, however, only in a rare instance in which a strong presumption against racial profiling has been overcome by evidence of compelling circumstances. This is a far cry from the situation today, in which racial profiling is routine and is subjected to far less scrutiny than it warrants.
Now that racial profiling is a hot issue, the prospects for policy change have improved. In June, President Clinton directed federal law enforcement agencies to collect and report information aimed at determining the extent to which their officers in the field stop or search individuals on the basis of race. The Customs Service has indicated that it is rethinking its practice of using apparent ethnicity or nationality as a basis for selecting certain subjects to investigate more closely than others. The Federal Aviation Administration has been reevaluating the security procedures it recommends for the nation's airlines; it wants the airlines to combat terrorism with computer profiling, a system that is purportedly less race-based than checks by airport personnel.
Unfortunately, though, a minefield of complexity occupies the terrain beneath many of these actions--which, after all, have been spurred by politicians' simplistic rhetoric. Unless these complexities are understood, this fleeting opportunity to address racial profiling properly will be wasted. Take the basic issue of defining racial profiling. As we have seen, many police officials, politicians, and community activists say that racial profiling occurs when a police officer stops, questions, or arrests someone solely on the basis of his or her skin color. They propose that definition for different reasons. Community activists like it because it helps them fire up the emotions of their followers by evoking a scene that is unequivocally evil: the crudely bigoted officer, to gratify his or her racist itch, harassing and humiliating black men for no legitimate purpose. Police officials and politicians like this definition, too, because it enables them to condemn resoundingly something that is already thoroughly discredited.
To fulminate against police officers who engage in such practices, however, requires no real confrontation with the status quo, because hardly anyone of substance or standing defends police surveillance triggered solely by race. The legal system stands ready to punish the officer who uses race alone as a signal of suspicion. Attacking this version of racial profiling is like attacking the racially motivated burning of churches. It is fine as long as one acknowledges that the target of the condemnation is, in the overall scheme of things, a rather marginal phenomenon. It is not so fine insofar as such rhetoric deepens public confusion. Much of the talk about police "targeting" of suspects on the basis of race is, in this sense, misguided and harmful. It diverts attention from the major part of the story to a side issue.
The better definition of racial profiling embraces a much more widespread police practice: using race as a factor in deciding whom to place under suspicion and/or surveillance. Relatively few police officers detain a person solely on the basis of race. Not even Mark Fuhrman was known to detain elderly women who happened to be black. Much more typical is questioning or detaining a person because of the confluence of a variety of factors--age (young), dress (hooded sweatshirt, baggy pants, etc.), time (late evening), geography (the person is walking through the "wrong" neighborhood)--that include race (black). It is the confluence of such factors that constitutes the profile that triggers police attention. Properly understood, then, racial profiling occurs whenever police routinely use race as a negative signal that, along with an accumulation of other signals, causes an officer to react with suspicion.
Thus, we must be skeptical and realistic as we watch and participate in the growing debate over racial profiling. One danger that I have already mentioned is the threat of demagoguery through oversimplification. When politicians discuss "racial profiling," we must insist that they define precisely what they mean and ask in particular whether they merely reject police surveillance based solely on race or whether they also condemn police surveillance that is at all triggered by a racial factor (except in the exceptional circumstances I have noted). Another danger is evasion--putting off making hard decisions in the guise of needing more information. This motivation, I fear, is behind Clinton's directive ordering federal agencies to gather statistics on the racial demographics of stops, questionings, and arrests. There is no mystery about whether federal and state police agencies use race routinely in making determinations of suspicion. Officials have said repeatedly that they do. The issue is whether the Clinton administration is willing to go beyond easy rhetoric and actually direct federal law enforcement personnel to cease their routine use of race as a factor in selecting subjects to scrutinize or question. On this point, Bradley was right when he criticized Gore for promising to end racial profiling only after he's elected president. The Clinton-Gore administration does not need to wait for court rulings or legislation or more data. It can and should act now. True, the federal government's law enforcement agencies are smaller than the combined forces of local and state authorities. But the DEA, the FBI, the Border Patrol, and other federal agencies are still important. And state and local cops often emulate their policies.
Finally, it must be soberly acknowledged that, even if this or any other administration does prohibit routine racial profiling, the practice will not cease quickly. After all, it will be difficult to prove that an officer made a given decision on a partly racial basis if he or she keeps silent regarding the racial aspect of his or her decisions and mentions, when pressed, only the nonracial cues. Moreover, supervisors and judges would probably be loath to reject an officer's proffered nonracial explanation. For one thing, doing so would typically lead not only to concluding that an officer acted on an unlawful basis but also, generally, to ruling that the officer lied about his or her conduct. Generalized avoidance of such confrontations is a fact of social psychology--albeit one that enables a considerable amount of illegal racial discrimination to pass undisturbed.
Even so, it would be helpful for Clinton to promulgate a strict anti-discrimination directive now. That would at least send a signal to conscientious, law-abiding officers that there are certain criteria they ought not to use. If nothing else, it might force racial profiling into the realm of conduct that officers may practice on the sly but don't dare admit to. To be sure, it's not optimal to create a norm that can't be fully enforced. But even if it can't be, that should just encourage us all to work that much harder to close the lingering gap between our laws and the actual conduct of those in positions of public authority. Even if a new rule against racial profiling would, to some degree, be made to be broken, it would still be worth having--for it would at least help set a new standard for legitimate government.
Randall Kennedy is a professor at Harvard Law School and the author of Race, Crime, and the Law.