Worse yet, the lines that government draws, putting people into groups, necessarily are often imperfect. Surely there are young adults seventeen years and nine months old who could vote as sensibly as eighteen-year-olds (and plenty of eighteen-year-olds who are not the most discriminating of voters). Why isn’t the speed limit fifty-three, or sixty for that matter? Any one of these choices might be as rational or sensible as the next. If every line could be challenged in court and struck down simply because there was another equally plausible (or imperfect) line, there would be no more laws left. Laws often are the product of interest-group bargaining, and not all the necessary compromises to get a law passed are going to make the most sensible of lines. That’s just unavoidable.20

  For these reasons, the ordinary response to government line drawing and group forming is a certain amount of deference. As long as the lines are sensible enough, the groups roughly the right ones, we give the government the benefit of the doubt. Courts often say they won’t “second-guess” government choices.21

  Under some circumstances, though—such as when the government draws lines among people based on race, religion, ethnicity, gender and sexual orientation, etc.—we look more closely. We call these “suspect classifications.” We believe that these categories rarely are or should be relevant to most things government does. Their very use is “suspect”: we worry that when government discriminates against someone in one of these groups, it is because of a stereotype or simple dislike, rather than a sound rationale.22

  This closer look at suspect classifications has a name—“strict scrutiny.” Strict scrutiny asks the same two questions: What is government trying to accomplish? And does it need to treat this group differently to do so? But under strict scrutiny the courts are more demanding about the answers. If, say, the government wants to treat one racial or religious group differently than others, it needs a super-strong reason, and the classification must be absolutely necessary to achieve the government’s goal. If the government can’t meet that test, it must find some other, nondiscriminatory way of meeting its goals. In examining this issue we ask for proof: Is the problem we are concerned about much more prevalent in the disfavored group than any other? And is the incidence of the problem high enough to justify imposing on everyone in the group?23

  Which brings us back to Oneonta, and the federal court’s claim that relying on race in a victim description simply does not implicate the Equal Protection Clause of the Constitution.

  CAN THE POLICE USE A RACIAL DESCRIPTION TO FIND A SUSPECT?

  The right question in a case like Oneonta is not whether the Equal Protection Clause applies—of course it does—but whether the use of a racial description of a suspect meets the test of strict scrutiny. Courts have offered a number of reasons for why policing based on a victim or witness racial description should get a free pass from any analysis under the Constitution’s equality guarantee, but none of them are very persuasive. The best one is the argument that the government didn’t dream up the racial classification itself, it is just what a witness or victim told the police. Fair enough, but still it is the government that’s using race as a basis for choosing who to stop and search. Surely government can’t evade the requirements of the Constitution because someone else gave the police the idea of doing it. And it is not like government doesn’t make decisions about what exactly to do once it gets this information: witness the dragnet that occurred in Oneonta.24

  Still, it seems obvious, does it not, that in most cases a witness’s description of a suspect that includes race will easily justify police following that description. It passes the Equal Protection Clause’s strict scrutiny test. If a crime has been committed, the government’s interest in catching the perpetrator is obviously high—“compelling” in the lingo of the Equal Protection Clause. And if the witness says the perpetrator is a person of a particular race dressed in a green coat who limps, that is obviously who the police should be looking for. Looking for a person of a different race would make no more sense than seeking a marathon runner in red.

  On the other hand, a suspect description that relies only on race will almost always be impermissible. That is one of the problems the Equal Protection Clause is concerned about: blaming a group for the possible actions of some of its members (or worse yet, stereotypes about those members), when most of them are innocent. That was the real problem in Oneonta. The suspect was probably a youngish black man, and maybe he cut his hand, but uncertainty on this score is the only possible explanation (other than out-and-out harassment) why the police were stopping women and older black men as well. In essence, all the police could be sure of was one fact: the culprit was black.

  The only thing that made relying on race in Oneonta remotely plausible was the relatively small number of black people who lived there, but still this seems hard to swallow. If a crime was committed somewhere, and all we knew about the suspect was that he was one of a thousand or more white men in the vicinity, would we suppose the police could stop all of them? In the absence of any other information, we would be extremely reluctant to make suspects of all these people. So why were things different in Oneonta, where there were several hundred black men, all of whom were apparently quite innocent?

  While the judges upheld what happened in Oneonta, lots of other people thought the police got it wrong. One kid at the college who was on the black list explained, “[T]he only list I ever wanted to be on was the Dean’s List.” In time, many expressed regret about the indignities in Oneonta. The state police apologized. The college did the same; the administrator who turned over the list was demoted. The office of New York Attorney General Eliot Spitzer defended state officials in the case. After they’d prevailed, Spitzer himself read the court opinion and told the press, “We won the case, but it makes your skin crawl.”25

  RACIAL PROFILING

  Agreed, then: if a suspect is described as a person of a particular race wearing a green coat, it makes no sense to search for someone of a different race wearing a red coat. But what if the characteristics that define the suspect are part of a profile, not a witness description? Are these the same thing? This was just the sort of argument people were making after 9/11: Why are we searching grandmothers from Topeka, when we should be searching Muslim men or men of some particular heritage or appearance? Should the rules be the same when there is a profile rather than a witness description?

  That brings us to some recent history about racial profiling in this country.

  In 1999 the then–Attorney General of New Jersey did the unthinkable: he released a report condemning his own state police for rampant racial profiling. (Prosecutors don’t succeed at their jobs by pointing fingers at law enforcement.) Troopers on the New Jersey Turnpike had been accused of stopping many more minority motorists, and searching them more often. The problem of “disparate treatment” of minorities, the AG concluded, “is real not imagined.”26

  It’s no wonder the New Jersey AG did this, though: by the time he released his report, his hand had been forced by a wealth of evidence of racial profiling by the New Jersey highway patrol, including a quite remarkable study performed in a case called State v. Soto. In Soto the State of New Jersey had argued that more minorities were stopped either because there were more of them on the highway or because they broke the traffic laws more often. But Dr. John Lamberth, an expert in statistical methods and social psychology, conducted a rigorous study to prove this was not the case. To eliminate these possibilities, Lamberth put spotters by the side of the Turnpike with binoculars to count the racial makeup of drivers on the highway. Then, he utilized rolling survey vehicles to see who was violating the traffic laws. Dr. Lamberth’s study established that while some 13 percent of the drivers on the Turnpike were African American, and while some 15 percent of the traffic violators were African American, African Americans constituted from 35.6 percent to 46.2 percent of the stops. Relying on this data, and on the testimony of police officials, the judge in State v. Soto found that there was a “de facto policy” on the part of the State Police “of targeting blacks for investigation and arrest.”27

  As it happened, matters in New Jersey were much, much worse than the Soto judge thought. Troopers would park their vehicles perpendicular to roadways, rendering their radar guns ineffective but making it easier to see the race of the driver. Some of the troopers pretty much arrested only minorities. Governor Christine Todd Whitman fired the superintendent of the state police, Colonel Carl Williams, for telling the press, “The drug problem is mostly cocaine and marijuana. It is most likely a minority group that’s involved with that.” State officials hid data from the Soto judge and from the U.S. Department of Justice, which conducted its own investigation, ultimately concluding that New Jersey had engaged in a consistent pattern of racial profiling.28

  The Attorney General’s report documented that racial discrimination was particularly prevalent in the use of so-called consent searches. When troopers had no basis to search a vehicle—and traffic stops typically provide none, for what would be the hidden evidence of speeding or having a taillight out?—they frequently would ask the motorist for permission to search. The AG’s report found that in some instances, 80 percent of the consent search requests were directed at minorities. The more discretion a trooper had, the more likely there would be disparate treatment of minorities. In one locale, troopers operating radar guns (which can’t tell blacks from whites) were issuing tickets to about 18 percent of African Americans, while troopers with the most discretion were issuing tickets to over 34 percent of blacks. “[O]fficers who had more time to devote to drug interdiction,” said the AG, “may have been more likely to rely upon racial stereotypes.”29

  New Jersey was hardly unique when it came to racial profiling. In another thorough study, Dr. Lamberth found that in Maryland 17 percent of the drivers were African American and yet a whopping 72 percent of those being stopped and searched also were, a disparity Lamberth called “literally off the charts” of statistical significance. Researchers in Michigan decided to get a peek behind the decision to stop, by examining queries to the mobile data terminals cops have inside their cars. Not only were queries of supposedly suspicious drivers made more often of African Americans, but the number of such queries increased significantly the further blacks were seen driving into whiter neighborhoods. In North Carolina, blacks were 68 percent more likely to be stopped by the North Carolina State Patrol than whites. Colorado officials paid more than $800,000 in damages to motorists stopped on I-70 based on a drug courier profile that targeted minorities, none of whom were ticketed or arrested.30

  This sort of racial profiling is pervasive in society. A report by Professor Ian Ayres of Yale Law School concerning LAPD stops in 2003–2004 found, after controlling for crime rates and other variables, that “African Americans and Hispanics are over-stopped, over-frisked, over-searched, and over-arrested.” Per 10,000 residents, there were 4,500 stops of African Americans, but only 1,750 for nonminority residents. Indeed, in some districts there were more stops in the reporting period than there were residents. In New York, more than 87 percent of the 700,000 people stopped in 2011 by the NYPD were either black or Latino. From 2007 to 2010, in Boston, there were more than 200,000 “Field Interrogation/Observation/Frisk and/or Search Incidents.” More than 60 percent of that number were black even though Boston has fewer than 25 percent African Americans.31

  WHY RACIAL PROFILING HAPPENS

  Even if one chooses to ignore the stunning human toll of racial profiling, it is unquestionably ineffective policy. Study after study shows minorities are not carrying or using drugs at higher rates than Caucasians; indeed, just the opposite is true. When police search, the hit rates—the number of times drugs or other contraband are found—are consistently higher for whites than for minorities. In North Carolina, where the State Patrol was stopping far more blacks than whites, the “hit rate” for whites who were searched was 33 percent, while it was 26 percent for blacks. In Illinois, even after years of working on the racial profiling problem, an annual report in 2014 concluded that minority drivers are still about twice as likely to be the subject of a vehicle consent search than other drivers, relative to how frequently they are stopped, yet police officers conducting consent searches are 50 percent more likely to find contraband in a vehicle driven by a white driver than one driven by a minority driver. Ayres’s report on the LAPD found that “frisks and searches are systematically less productive when conducted on blacks and Hispanics than when conducted on whites.” As the New Jersey AG’s report pointed out, “many of the stereotypes about drug use are simply wrong”; among high school students, “white students are actually more likely than black or Hispanic students to report having ever used” drugs or alcohol.32

  If profiling doesn’t work, why is it so pervasive? One possible answer is widespread intentional racism. But that answer is both too simple and too disheartening to accept without considering alternatives. And it turns out that at least one alternative is readily at hand and instructive. It is an answer grounded in the idea of unconscious racial bias.33

  A chapter of the New Jersey AG’s report, titled “The Circular Illogic of Race-Based Profiles,” describes a “self-fulfilling prophecy” that explains why racial profiling occurs. Think about it this way. You like to fish, so you ask people what’s a good spot. They say Trout Pond. You go there, and sure enough you catch some fish. Occasionally you go somewhere else, and you catch some fish there, too, but you’ve been told Trout Pond is a surefire bet, and you’ve seen some evidence of that, so you keep coming back to it. Now, if you’d done a careful study, you’d have learned that Trout Pond was no better than any other spot, and in fact might have been less good. There were more fish in other places, like Town Wharf or Towd Point. But having been told of Trout Pond and had your information confirmed, that is where you went. The AG’s report described the hits officers got by stopping minorities—and then told one another about—as the “statistics … used to grease the wheels of a vicious cycle.”

  Consistent with our human nature, we in law enforcement proudly display seized drug shipments or ‘hits’ as a kind of trophy, but pay scant attention to far more frequent ‘misses,’ that is, those instances where stops and searches failed to discover contraband … Logically, of course, one cannot hope to judge the overall effectiveness of any practice or program by looking solely at its successes, any more than by looking only at is failures.34

  By the late 1990s all this was well-enough known that a national consensus against racial profiling had developed. President George W. Bush had condemned the practice, as did many, many others. Congress was considering legislation that would have required state and local police forces to keep statistics that would reveal profiling where it occurred. A whopping 81 percent of respondents told the Gallup poll they disapproved of the practice.35

  Then came September 11. Radical Islamists attacked the United States and the consensus against racial profiling quickly melted away. On the day of the attack, New Jersey’s new Attorney General, John Farmer, Jr., was in Atlantic City at a conference talking about the progress the state was making in addressing the profiling problem. But within two weeks of 9/11 he published an article in The Star-Ledger titled “Rethinking Racial Profiling.” In it Farmer called the fact that before 9/11 “we were able to condemn universally the practice of racial profiling” a “luxury.” “More than 6,000 people are dead, some would argue, because of insufficient attention to racial or ethnic profiles at our airports … How can law enforcement not consider ethnicity in investigating these crimes when that identifier is an essential characteristic of the hijackers and their supposed confederates and supporters?”36

  So, what’s right? What we said before 9/11, or after it?

  WHEN IS RACIAL PROFILING OKAY?

  The second incorrect thing courts—and others—say about race and policing is that profiling is okay so long as race is not the “sole” item in the profile. It’s okay to look at race, they say, so long as there are other factors in play as well. But this, too, deviates from the way we usually deal with questions of racial discrimination when the issue is not policing.

  The standard rule under the Equal Protection Clause is that if race is a “motivating” factor in a government decision, strict scrutiny must be applied, even if it is not the “sole” factor. Strict scrutiny requires us to compare the incidence of the problem that the government is concerned about in both the favored and disfavored (here, searched) groups. Although in very rare occasions race may be a relevant factor in a profile, it’s unlikely to be in most instances.37

  To see why race or other suspect classifications rarely will be appropriate in a profile, it helps to look at a familiar example: airport drug interdiction cases. Typical was United States v. Weaver, yet another case involving our old friend from Chapter 6, DEA Agent Carl Hicks. In this case the police spotted Arthur Weaver getting off a plane in Kansas City, coming from Los Angeles. Hicks explained Weaver was suspect because he was “a ‘roughly dressed’ young black male who was carrying two bags and walking rapidly, almost running, down the concourse toward a door leading to a taxi stand.” Hicks also testified that he “was aware that a number of young roughly dressed black males from street gangs in Los Angeles frequently brought cocaine into the Kansas City area, and “walking quickly towards a taxicab was a common characteristic of narcotics couriers at an airport.” So Hicks stopped Weaver. Weaver declined to allow Hicks to search his bags, and told him to get a warrant. Hicks persisted in grabbing at Weaver after he got in a cab, so Hicks and Weaver ended up tussling over Weaver’s bags. After Weaver hit Hicks’s hand in order to pry it off his bag, Hicks arrested Weaver, patted him down, and—having obtained a warrant—searched the bags and found drugs.38