A Definition, and a Little History
Understanding racial profiling requires a good working definition. Though some prefer other names for the tactic—for example, biased policing or bias-based policing—racial profiling remains the most common term. For purposes of this discussion, I define racial profiling as the law enforcement practice of using race, ethnicity, national origin, or religious appearance as one factor, among others, when police decide which people are suspicious enough to warrant police stops, questioning, frisks, searches, and other routine police practices. Notice that this definition does not require that racial or ethnic appearance acts as the sole factor motivating what an officer does; such a narrow conception would define racial profiling out of existence because few if any law enforcement encounters occur based on a single factor. Note also that using a reasonably detailed description that includes the race of a suspect who has been observed is not racial profiling. If a witness describes the person seen running from the convenience store that just experienced a robbery as male, black, five foot ten inches tall, 20 to 24 years old, wearing a green sweatshirt and black Nike sneakers, and having a mustache and a goatee, a police bulletin using this description does not constitute racial profiling. Rather, it constitutes good police work. Race, in the context of a reasonably detailed description, is a better descriptor than any clothing worn or facial hair; a suspect can change both of those with ease. Racial appearance, in contrast, does not change, and sticks in the mind.
Historically, racial targeting by police did not start in the late twentieth century. It has constituted a fact of life for African Americans as long as there have been organized police forces in the United States—indeed, even before that, with the slave patrols of the American Antebellum South. But what we think of as racial profiling, in a somewhat systematic modern form, really took shape in the last two decades of the twentieth century, beginning in Florida. (See, e.g., David A. Harris, Profiles in Injustice: Why Racial Profiling Cannot Work (2002).) In the 1980s, a Florida state trooper named Bob Vogel, engaged in drug interdiction along the state’s highways, began to put together a list of factors that, he said, kept coming up in all of his biggest and most important drug busts. Never mind that the lists were by nature incomplete and selective; they only included instances in which the trooper had stopped drivers and found drugs, not ones in which the effort had been unsuccessful. The tactic made Vogel somewhat of a celebrity; he was interviewed on network television news shows, and he won election as Sheriff of Volusia County, Florida. The DEA took notice and created its own system of factors, systematizing them into profiles. This all resulted in an effort that brought the federal government into the effort in a big way: Operation Pipeline. This program, which began in 1986, used millions of dollars in federal funding to train police from all over the country in the fine points of profiling; those officers would then return to their own departments, ready to train others and set up interdiction units. (See David Kocieniewski, New Jersey Argues That the U.S. Wrote the Book on Racial Profiling, N.Y. Times (Nov. 29, 2000).) The DEA denies that it made racial or ethnic appearance a factor in any of its training, but the evidence says otherwise. And the proof, as they say, was in the pudding: Once Pipeline tactics made it into the training and tactics of police forces around the country, police targeting of black drivers became systematic and common. And nowhere did this show up more clearly than in New Jersey and Maryland.
In the early 1990s, state troopers in both New Jersey and Maryland appeared to have taken Operation Pipeline’s instruction and encouragement into use, putting large-scale drug interdiction into effect on some of their most traveled interstate highways. In New Jersey, the most intense efforts came on the New Jersey Turnpike; in Maryland, the enforcement took place along I-95. Even in the early 1990s, the Pipeline-based tactics of both police departments had drawn lawsuits, challenging both the legality of the practice on Fourth Amendment grounds and the racial discrimination intertwined with these vehicle stops. Interdiction and enforcement, it seems, had a distinctly racial cast, and a wildly disparate impact on African Americans. In New Jersey, the Soto case put these matters into focus. (State v. Pedro Soto, 734 A.2d 350 (N.J. Super. Ct. Law Div. 1996.) New Jersey’s attorney general, representing the state police, bitterly contested every allegation of wrongdoing in the case, especially charges of racially biased enforcement. But the existence of the legal action created, for the first time, the opportunity to measure whether the state police were, in fact, targeting black drivers. Using a study protocol designed by Dr. John Lamberth of Temple University, the expert retrained by the Soto team, hard numbers based on in-person observation on the New Jersey Turnpike—the location of the Pipeline-inspired interdiction efforts—proved that the race of the drivers played the dominant role in determining which drivers police pulled over and which vehicles and people were searched by obtaining so-called voluntary consent. Lamberth first confirmed the conventional wisdom among police officers who spent their time enforcing traffic laws: Police could stop any driver they chose for a traffic infraction because all drivers committed some traffic infraction over a minimal time and distance. Second, and much more importantly, Lamberth showed that while all drivers could be stopped, it was African Americans drivers who police stopped at a vastly higher rate than one would expect, given the percentage of African American drivers on the Turnpike. With blacks only 13.5 percent of those on the road, blacks made up approximately 35 percent of all of those stopped by police—19.45 standard deviations. (Statisticians calculate standard deviations to determine whether the difference between two numbers is real—i.e., statistically significant—or the result of pure chance. Generally, the difference between two numbers is considered statistically significant at about two standard deviations.) The chances of such a difference between those driving and those stopped occurring accidentally, Lamberth said, was “substantially less than one in one billion.” The data collected in his observational study allowed Lamberth to conclude: “Absent some other explanation for the dramatically disproportionate number of stops of blacks, it would appear that the race of the occupants and/or drivers of the cars is a decisive factor or a factor with great explanatory power. I can say to a reasonable degree of statistical probability that the disparity outlined here is strongly consistent with the existence of a discriminatory policy, official or de facto, of targeting blacks for investigation. . . .” (Harris, Profiles in Injustice, supra at 53–56.) Despite contrary conclusions from the state’s expert, the judge in the case concluded that the state police had indeed been targeting African Americans on the Turnpike for years. Two more years passed after the judge’s 1996 decision before the state finally conceded what the numbers and other evidence had proven: Racial profiling of African Americans on the Turnpike was “real—not imagined.”
A similar scenario played out in Maryland. When the Maryland state police stopped and searched a car driven by Robert Wilkins, in which Wilkins and his family members were returning to the Washington, DC, area from a funeral out of state, the police made a number of mistakes. They refused to explain satisfactorily why they had decided to stop Wilkins for a petty infraction, and they refused to abide by his statement to them that he did not consent to a search of the vehicle when asked if he would do so. Instead, Wilkins and his family members were forced to stand by the roadside as the police searched the car, even using a drug-sniffing dog, treating them like criminal drug-smuggling suspects. But perhaps the biggest mistake the police made was who they treated this way. Mr. Wilkins was not an average young black man driving a rental car with several family members; he was an attorney, steeped deeply in the law of search and seizures as part of his job at Washington’s prestigious Public Defender Service. When they refused to listen to Mr. Wilkins tell them what the law did and did not allow them to do, they bought themselves a federal civil suit. (Wilkins v. Maryland State Police, Civil No. MJG-93-468 (D. Md., 1993).) That lawsuit featured another observational study by Dr. John Lamberth, which found that while 17 percent of the driving population on the interstate highway in Maryland were black, 72 percent of those stopped and searched were black—a difference that Lamberth described in his report as “literally off the charts.” In language similar to that he used in New Jersey, Lamberth said that “[w]hile no one can know the motivation of each individual trooper in conducting a traffic stop, the statistics presented herein . . . show without question a racially discriminatory impact on blacks” that is “sufficiently great that, taken as a whole, they are consistent and strongly support the assertion that the state police targeted the community of black motorists for stop, detention and investigation within the Interstate 95 corridor.” The lawsuit also uncovered something else: a Maryland State Police internal “Criminal Intelligence Report” dated just days before the stop of Mr. Wilkins, containing an explicit profile targeting African Americans. The Wilkins case settled soon after. (Harris, Profiles in Injustice, supra at 60–62.) (Robert Wilkins was later appointed a federal district judge; he now sits on the US Court of Appeals for the DC Circuit.)
Is Racial Profiling Constitutional?
Those who do not practice criminal law regularly may find it surprising to learn that this way of policing does not violate the Fourth Amendment to the US Constitution, which governs search and seizure practices. But in fact, using traffic enforcement as a pretext to investigate other crimes for which police have no evidence falls within the bounds of permitted search and seizure practice. In 1996, the US. Supreme Court decided Whren v. United States., 517 U.S. 806 (1996), a case in which the Court had before it the well-known police practice of using traffic enforcement as a way to stop drivers to investigate the possibility of other crimes, primarily drug trafficking. The facts of Whren illustrated the practice perfectly. Police officers tasked with drug enforcement—not traffic patrol duties—saw a late-model sport utility vehicle driven by a young African American man, with another African American man in the passenger seat. The officers had no reason to suspect these men or the vehicle they drove of involvement in drug trafficking or any other crime, but they followed until the driver committed a minor traffic offense. The officers pulled the vehicle over, and when they came to the windows and looked in, they observed what they immediately knew to be cocaine. Both driver and passenger were arrested, and later convicted on drug charges. Before the Supreme Court, the defendants argued that even if the police had probable cause to believe that the driver had violated the traffic code in some small way, the police had used that violation as a pretext for a stop—a Fourth Amendment seizure—for another offense (drug trafficking), for which they had no evidence. The stop and the subsequent seizure, they argued, violated the Fourth Amendment. But the Supreme Court disagreed, siding with most of the US courts of appeal that had decided cases on the issue before. As long as the police had observed an actual traffic offense that violated the jurisdiction’s criminal code, they had full probable cause to stop the vehicle and the driver. It mattered not at all that the officer’s real motivation had nothing to do with traffic enforcement, and everything to do with drug enforcement, for which they had no evidence. The actual motivations of the officer did not matter.
That much would have been enough to give the use of pretext-based enforcement the Court’s blessing. But what about the presence of race as a factor in deciding who to stop and search? Would that make for a constitutional violation? In their briefs, the defendants confronted the Court with the statistical evidence from the New Jersey and Maryland cases, showing that this kind of pretext-based enforcement had a clear racial skew: The police used it to target African Americans and Latinos. The Court acknowledged the statistical evidence but said that kind of statistical picture had no bearing on the Fourth Amendment questions of search and seizure at issue. A claim of racially discriminatory enforcement, the justices said, should come under the Equal Protection Clause of the Fourteenth Amendment, in a civil lawsuit. Courts would not suppress any evidence seized under the Fourth Amendment; defendants or other aggrieved parties not arrested, like Robert Wilkins, would have to bring a lawsuit, an often difficult, long, and expensive endeavor with no guarantee of success.
Thus, for Fourth Amendment purposes, the Whren decision gave a green light to the practice of racial profiling through traffic stops. Whren joined a line of earlier Supreme Court cases that allowed officers to ask for “voluntary” consent to search without any evidence of criminal behavior, as long as they did not use coercion to obtain consent (Schneckloth v. Bustamonte, 412 U.S. 218 (1973)), which officers used to turn traffic stops into searches; to question motorists during traffic stops without administering Miranda warnings (Berkemer v. McCarty, 468 U.S. 420 (1984)); to order drivers to get out of their vehicles during any traffic stop, for any reason or no reason at all (Pennsylvania v. Mimms, 434 U.S. 106 (1977)); and to look for evidence in “plain view” by looking inside the passenger compartment after stopping the vehicle (Coolidge v. New Hamphire, 403 U.S. 443 (1971)). Combined, these cases—and others still to come—allowed the full transfiguration of any traffic stop into a drug investigation or other criminal probe on the side of any highway or street, at any officer’s almost unfettered discretion. And, as in Maryland and New Jersey, the great bulk of this activity focused on people of color: African Americans and Latinos engaged in that most universal of American activity: driving. This level of legal police discretion to stop and search—virtually without limits, for Fourth Amendment purposes—is what allowed the practice to take root and grow.
The legal challenges to racial profiling have been numerous, but the successes have been few, given the ways in which the Supreme Court has cleared the way for police to operate. The legal theory of the Soto case in New Jersey hung not on the Fourth Amendment, but on New Jersey’s own law, at the state level, that gave citizens more protection from discriminatory practices than the US Supreme Court afforded them. And the Wilkins case, in Maryland, was a civil rights lawsuit—the type of suit that the Court in Whren said should find its foundation in the Equal Protection Clause. Success in these suits is almost always difficult; most fail—sometimes because plaintiffs have criminal records and would make bad witnesses; sometimes because they lack the resources or the patience to carry a multiyear legal battle through to a conclusion; sometimes because the evidence is equivocal. Robert Wilkins was as good a plaintiff and witness as a lawyer could hope for: a Harvard Law–educated criminal defense attorney who had asserted his rights at every step in unfailingly respectful tones, and who had not a blemish on his background; a man who had a deep familiarity with the legal system, and what it took to sustain a lawsuit over the long haul. And the evidence could not have been stronger: There were the statistics, and then that internal memo written by and for the state police.
Did Racial Profiling Help Police Catch More Bad Guys and Make Us Safer?
With the statistics from New Jersey and Maryland offering undeniable proof that the tactic of racial profiling did not spring from the collective imagination of people stopped by police, but that it was real, the argument from police and their supporters shifted. Denial no longer seemed tenable (though many persisted in that direction), so many supporters began to argue that police used racial profiling as a reasonable and rational measure. Profiling was a good tool for catching bad guys because it helped police focus on the right people and put their resources where they would be most effective. Marshall Frank, a former police captain from Miami, put it this way in an opinion piece in the Miami Herald on October 19, 1999: “Label me a racist if you wish, but the cold fact is that African Americans comprise [sic] 12 percent of the nation’s population, but occupy nearly half the state and federal prison cells. African Americans account for 2,165 inmates per 100,000 population, versus 307 for non-Hispanic whites and 823 for Hispanics.” This meant, Frank argued, that African Americans commit most of the serious crimes, and until that changed, police should continue to use profiling. It was the smart thing to do.
Thus, effectiveness became the question. Did focusing on the right people by using race or ethnic appearance help the police get more bad guys? Did it help them take more guns or drugs off the street?
With the statistical studies in Maryland and New Jersey, and data from elsewhere, it became possible for the first time to answer this question. In my own work, I called this question the “hit rate” issue: In jurisdictions where the data showed that the police did focus their efforts on blacks and Latinos, we could compare results of stops and searches of blacks and Latinos, where race or ethnicity came into the mix of factors police used, and stops and searches of whites, where those factors played no role. How did the rate at which police “hit”—found drugs, found guns, made arrests—compare, when stops and searches of blacks and Latinos were put side-by-side with stops and searches of whites?
The results surprised many. In jurisdictions using racial targeting in their stops and searches, police hit more often when they stopped and searched whites than when they searched either blacks or Latinos. The evidence squarely refuted the profiling hypothesis of Marshall Frank and so many others—that police would do better when they focused on those right people who were racial minorities. In fact, racial targeting did not improve police accuracy; it did not even result in a rate of hits the same as the rate of hits of whites. It made police accuracy—hit rates—worse. And these hit rate results held across different police departments, in different geographic areas with different missions. (Harris, Profiles in Injustice, supra at 79–84.)
All in all, the statistics produced in the Soto case in New Jersey and the Wilkins case in Maryland, combined with the statistics and research on hit rates and the early efforts in Congress on the Traffic Stops Statistics Act, seemed to create a public consensus. By 1999, polling showed that almost 60 percent of all Americans understood what racial profiling was and believed it to be a widely used police tactic. Even more surprising, 81 percent of Americans said that, no matter how widespread it was, they disapproved of racial profiling and wanted it stopped. In a country long divided by race, an 81 percent consensus represented a startling development. (Frank Newport, Racial Profiling Seen as Widespread, Particularly Among Young Black Men, Gallup Poll, Dec. 9, 1999.)
But the terrorist attacks of September 11, 2001, altered this understanding.
New Waves of Profiling
If the wave of police profiling of African Americans and Latinos as drug suspects had met significant public opposition, new waves of profiling lay ahead. With the terrorist attack on the United States, a new public consensus began to form. While Americans may have believed that racial profiling of blacks and Latinos in drug interdiction should end, they felt differently about profiling when the tactic focused on those viewed as potential terrorist threats. By a nearly two-to-one margin, Americans believed that the government should subject people of “Arab descent” to profiling in airports. (Jeffrey M. Jones, Americans Felt Uneasy Toward Arabs, Gallup Poll Monthly, Sept. 28, 2001.) Stanley Crouch, a well-known African American academic, author, and social critic, put it this way in his column in the New York Daily News on March 14, 2002: “[A]ll those who denounce so-called Arab profiling . . . need to put their faces in a bowl of cold water for a few seconds and wake up. . . . [I]f pressure has to be kept on innocent Arabs until those Arabs who are intent on committing mass murder are flushed out, that is the unfortunate cost they must pay to reside in this nation.”
Thus, a new reason for racial (or ethnic or religious) profiling came along, to put the old wine of the tactic in new bottles. But there was no reason to believe that using profiling based in part on racial, ethnic, or even religious appearance would do any better job of helping police—or, in this iteration of the tactic, aviation security or antiterrorism forces—find the bad actors. And those with the greatest degree of experience in the sphere of tracking terrorism recognized this almost immediately, despite the public’s eagerness to bring profiling to the “war on terror.” Just weeks after the September 11 attacks, an internal government memorandum surfaced, written by senior US intelligence officials, urging those on the front lines to avoid at all costs using profiling in pursuit of possible terrorist suspects. The document, reported in the Boston Globe, urges those on the front lines to avoid the mistake of using ethnic or religious appearance as a proxy for dangerousness. They did not voice concerns about political correctness, or bad publicity, or lawsuits. Rather, their reasons were simpler and more straightforward: Profiling would fail. The only way to ferret out potential terrorists was through intelligence gathering and through careful observation of suspicious behavior. As one of the officials told the Globe, “believing that you can achieve safety by looking at characteristics instead of behavior is silly. If your goal is to prevent attacks . . . you want your eyes and ears looking for pre-attack behaviors, not characteristics.” (Bill Dedman, Airport Security: Memo Warns Against Use of Profiling as Defense, Bos. Globe, Oct. 12, 2001.)
There were additional reasons to see danger in profiling as an anti-terrorism tool, even if the public accepted it. First, allowing or even encouraging front-line antiterror officers to use profiling would make our security system easier to defeat. Terrorist groups would simply shift to using agents with a less Middle Eastern or Muslim appearance, or to agents who had Western identities, such as the British citizen who attempted to ignite a bomb hidden in his shoe on a Detroit-bound US airliner on Christmas Day of 2009. Second, using racial or ethnic appearance—even as just one factor in addition to others—would produce great numbers of false positives, all of which security personnel must sort through and investigate, and would drain significant amounts of resources away from real threats. Third, investigating the many false positives, based on Middle Eastern or Muslim appearance, could not help but alienate these populations within the United States. If there was real danger in the U.S. of so-called sleeper cells of terrorists, as the government repeatedly claimed, police and security services would need help in spotting these individuals. Simply put, Middle Eastern and Muslim communities were the ones who could best do this. They knew the languages, the cultures, and habits of those who lived among them; they would have access to situations in which there might be real dangers of extremist threats evolving. In short, Middle Eastern and Muslim communities in the United States were the allies we needed most. And this obvious truth proved out in the earliest case of government disruption of a terrorist cell on US soil. In Lackawanna, New York, the government apprehended, charged, and imprisoned six young Middle Eastern men who had formed a cell for possible terrorist action. The authorities were able to accomplish this only because the Yemeni community in Lackawanna came forward with this information, bringing it to a community policing officer from Lackawanna’s own police force—someone they knew and trusted. (Dina Temple-Rastin, The Jihad Next Door: The Lackawanna Six and Rough Justice in the Age of Terror (2006).) Wholesale targeting of these same communities, through widespread appearance-based profiling, would make cooperation with law enforcement and intelligence operations much more difficult. The trust necessary for cooperation would be driven out by fear and resentment.
The next wave of racial profiling—the third—arrived in the mid to late 2000s, with new political interest in suppressing illegal immigration to the United States. Ten years earlier, in 1996, Congress had passed the Illegal Immigration Reform and Immigrant Responsibility Act. Part of that law, section 287(g), offered state and local police forces the opportunity to agree, voluntarily, to become involved in immigration enforcement through agreements with the federal government; the local department, or a contingent of the department, would be effectively deputized to assist federal authorities with immigration enforcement, usually task forces and other kinds of specialized efforts. Nearly all US police departments either ignored the invitation or spoke up against it because they believed it would turn their immigrant communities against them because their officers would end up targeting people by ethnic (usually Latino) appearance. Local police knew that their immigrant communities contained both documented and undocumented individuals, often mixed together within families. Turning local police into adjunct immigration agents would cause people living in immigrant communities, whether documented or not, to fear calling the police. They would cease reporting crime and cooperating with local officers. Thus, efforts to recruit state and local police into immigration enforcement section 287(g) largely failed; even today, only 80 police departments out of approximately 18,000 in the United States operate under section 287(g), with the vast majority operating only inside jails and focusing on those already arrested. (Delegation of Immigration Authority Section 287(g) Immigration and Nationality Act, U.S. Immigration & Customs Enf’t.)
With section 287(g) attracting little law enforcement support, and with efforts to update and repair the broken immigration system failing, a few state governments began to fill the vacuum. Arizona took the lead. The most well-known state effort—to some, the most notorious—was Arizona’s S.B. 1070, the “show me your papers” law. (S.B. 1070, 49th Leg., 2d Sess. (Ariz. 2010).) The law had multiple provisions, but the one most pertinent to this discussion required that any police officer conducting a detention of any kind—for example, a traffic stop or a stop and frisk—must inquire about the suspect’s immigration status any time the officer had any suspicion of an immigration issue. Failure by any law enforcement agency to carry out this legislative directive could subject the agency or even an individual officer to a lawsuit. Police officers in Arizona rightly saw the law as trapping them between the proverbial rock and a hard place. On the one hand, if they failed to ask about immigration status, they could face legal liability. On the other, they would be forced, in practice, to make this judgment based on “looking Mexican” and perhaps on the person’s accent. Because Mexican appearance was common among Arizona residents—Arizona, after all, had been part of Mexico well into the 19th century, and was not an American state until the 20th century—and because Spanish-accented English remained common, officers would practically be forced to engage in ethnic profiling. Arizona—specifically, Maricopa County, the state’s most populous county—was also home to former Sheriff Joe Arpaio, who built his well-worked media brand—he called himself “America’s toughest sheriff”—partly on cracking down on Latino immigrants. He was eventually found liable for unconstitutional racial profiling in federal court, and then held in criminal contempt for (among other things) disobeying the court’s order to stop the practice. (The president eventually pardoned Arpaio on the criminal contempt charge.) All in all, the use of profiling against undocumented immigrants from Mexico and Central America has not let up in the years since; with the advent of the Trump administration’s dramatic change in immigration policies and actions at the border with Mexico, this third wave of profiling, in immigration, remains very much with us.
Where Are We Now?
Back in 2002, in my book Profiles in Injustice: Why Racial Profiling Cannot Work, I argued that we had the statistics and the other evidence needed to come to some conclusions. First, no one could doubt that racial profiling—the targeting of African Americans and Latinos by police, using their racial or ethnic appearance as one factor among others in deciding who police should stop, question, search, or frisk—was real. This did not happen in every police department and was not a tactic used by every police officer. But in some police agencies—the New Jersey and Maryland state police, the New York Police Department, and the US Customs Service, just to name a few—the evidence was beyond question in 2002. And there was every reason to believe that many other police departments used racial profiling as well, especially given the conscious government effort to spread the practice nationwide through Operation Pipeline. Second, the hit rate data proved just as solid. In every police agency in which measurement had occurred, the use of race-based stops and searches rendered police efforts less effective than policing that did not reference race. In departments using profiling, hits among blacks and Latinos were not higher than hits among whites; they were not the same as hits among whites. They were, in fact, lower than the rate of hits among similarly situated whites. And the practice of racial profiling was not only ineffective and inefficient law enforcement; it damaged and eroded the important connections and trust police needed to have with communities they served.
Fast forward almost two decades later, and we have a much more detailed, data-heavy, and even nuanced picture. In the late 1990s and early 2000s, a number of states passed laws mandating collection of racial profiling data on traffic stops and searches. Most of these laws stayed in effect for just two or three years; they were temporary, a way to conduct something like pilot studies of racial profiling. But a few of these laws did not sunset; they remained effective and required ongoing data collection and even reporting by state agencies. For example, Missouri’s law, passed in 2000, requires that each police department in the state submit traffic stop data to the attorney general of the state every six months; the attorney general must then report to the public on the data. (Mo. Rev. Stat. § 590.650 (2000).) Another of those states was North Carolina. (N.C. Gen. Stat. § 143B-903 (1999).) It required the ongoing collection of data every year, like Missouri’s law. In 2018, Frank Baumgartner, Derek A. Epp, and Kelsey Shoub published Suspect Citizens: What 20 Million Traffic Stops Tell Us About Policing and Race. Baumgartner and his colleagues obtained the massive amount of data North Carolina had collected over almost 20 years and did a rigorous job of analyzing it. The mountain of information they had on police traffic stops far outstripped anything available to me or other researchers at the turn of the 21st century, when I wrote my book; the results would say much about whether or not we were right when we looked at the considerably more sparse data on racial profiling almost two decades before. In these lines of their opening chapter, Baumgartner, Epps, and Shoub summarize their findings.
We conduct the most comprehensive analysis to date of traffic stops in a single state, North Carolina. . . . [P]owerful disparities exist in how police interact with drivers depending on their outward identities: race, gender and age, in particular. . . . First, there are stark differences. Second, young men of color are clearly targeted for more aggressive treatment. Third, these differences are not fully justified by differences in criminality. Fourth, the aggressive use of traffic stops as a tool to investigate possible criminal behavior, though justified as part of the war on crime, is surprisingly inefficient, rarely leading to arrests for contraband. When we contrast the costs of targeted and aggressive policing with the benefits of it, we find that the social and community costs are high (in terms of reduced trust in and cooperation with the police), but that the number of crimes solved by traffic stop-related investigations is minimal.
Thus, more than 20 years after the first statistical studies of profiling by John Lamberth in New Jersey and Maryland, and almost 20 years after the publication of my own book on the subject, it’s all there. Baumgartner et al. find that the targeting of people of color is certainly real; this racial targeting also comes with a more aggressive style of policing. They sustain the early findings on hit rates, concluding that the data show racially targeted policing is highly inefficient and rarely uncovers contraband or solves cases. And it tears at the all-important fabric of trust that the police and the public depend upon for cooperation with each other. One can only hope that every law enforcement officer—not just in North Carolina but across the country—will take the time to absorb the book’s lessons. And those lessons are echoed in a large-scale analysis by the Stanford Computational Policy Lab, which analyzed nearly 100 million state and municipal police patrol stops, finding racial bias in both who police stopped and who they searched after they stopped. (Sharad Goel et al., A Large Scale Analysis of Racial Disparities in Police Stops Across the United States (Mar. 13, 2019) (Stanford Computational Pol’y Lab).)
Profiling by Proxy
We cannot leave the subject of racial profiling without noticing a newly observed phenomenon, a new type of racial profiling that should extend the ways in which we have come to view the practice. It is important to say, again, that this is not a new experience for people of color; it is simply newly noticed by the mainstream, predominantly white culture. This is due in no small part to the ubiquity of the cell phone video camera system. I call this new phenomenon profiling by proxy: White civilians, seeing African Americans in places in which the white people don’t feel the African Americans belong, summon police as their proxies, to investigate the behavior of the (to them) out-of-place and therefore suspicious blacks. Some of these examples will be familiar; others may not be.
• On April 12, 2018, two African American men in a Starbucks Coffee shop in Philadelphia waited for a business associate to arrive for a meeting. The men asked to use the bathroom but were refused permission because they had not purchased anything. When the men still did not buy anything as they waited, the manager asked them to leave. When they did not, the manager called police, who arrested, handcuffed, and charged the men. They were later released and charges were dropped, and the men reached a civil settlement with the city. (Scott Neuman, Men Arrested in Philadelphia Starbucks Reach Settlements, Nat’l Pub. Radio (May 3, 2018).)
• On April 29, 2018, a white woman called the police in Oakland, California, because she observed a black family group in a public park getting ready to have a barbeque with (she said) the wrong type of grill. The incident took place in a “designated barbeque zone” in the well-used park in which barbequing is a regular activity. Police responded, detaining and questioning the family for an hour. They made no arrests. (Carla Herreria, Woman Calls Police on a Black Family for BBQing at a Lake in Oakland, Huff. Post (May 11, 2018).)
• In May 2018, as four people—three of them black—checked out of an Airbnb rental in Rialto, California, and loaded their luggage in a car, preparing to leave, seven police cars swarmed the area and stopped them. A police sergeant approached, ordered them to put their hands in the air, and told them they were also being tracked by a police helicopter. The police had a report by a (white) neighbor that she did not recognize them, and they might be burglarizing the house. The guests showed the sergeant their paperwork (he said he did not know what Airbnb was) and contacted the owner of the property to confirm. They were held for 45 minutes and then released. (Three Black People Checked out of Their Airbnb Rental. Then Someone Called the Police on Them, CNN (May 10, 2018).)
• On May 8, 2018, a black graduate student at Yale University decided to take a nap in her dorm’s common area during a long night of paper writing. A white graduate student entered the room, flipped on the lights, and told the black student she had no right to sleep there; she then called the campus police. When police responded, the black graduate student had to spend almost an hour with them to prove that she had a right to be present. (Elliot Hannon, A Black Graduate Student at Yale Says a White Student Called Police on Her for Napping in a Common Room During an All-Nighter, Slate (May 9, 2018).)
And these are far from the only incidents; new examples appear frequently. Profiling by proxy represents a new development in racial profiling: racial and ethnic minorities suffering extra police attention while not engaged in criminal behavior, but in this version, at the instigation of white civilians.
The Bottom Line
Racial profiling has a long history in the United States. It is not new; more to the point, it is not gone, even 25 years after the first litigated cases and statistical studies of the tactic. It corrodes relationships between police and communities, leaving people of color knowing that they have been treated unfairly and understandably resentful of police; in the bargain, it does nothing to fight crime or make us safer. This, of course, helps no one.
There are many police departments that have made efforts to curb these practices. They explicitly recognize the practice and prohibit it, by policy, training, and practice. Some—a smaller number—collect data on all traffic stops, stops and frisks, and other routine police practices in an effort to track the practice and correct their officers. (For a list of the 50 or so largest police departments in the United States that collected some data on traffic stops as of 2012, see David A. Harris, Across the Hudson: Taking the Stop and Frisk Debate Beyond New York City, 16 N.Y.U. J. Leg. & Public Policy 853 (2013).)
What we know is that the practice must be curtailed, even if we cannot seem to end it. We’ve long had information proving that the practice does not help police and does injure communities; it is not a matter of not enough data, but if it ever was, the work of Baumgartner and his colleagues and others has filled any remaining gap. That information can become part of police training. But two other things must happen in order to curtail this pernicious practice.
First, the idea that people consent to searches voluntarily—the cornerstone of the US Supreme Court’s 1973 Schneckloth decision—when faced with an armed police officer who has the power to hand out a traffic ticket or make an arrest—or to do neither—is absurd on its face. As shown in a path-breaking study by Ilya Lichtenberg, fewer than 10 percent of everyone given this “choice” actually asserts and stands by his or her right to refuse a search. (See, e.g., Illya D. Lichtenberg, Voluntary Consent or Obedience to Authority: An Inquiry into the Consensual Police-Citizen Encounter (1999) (unpublished Ph.D. dissertation, Rutgers University).) Lichtenberg’s work showed that the heart of consent searches is not actually consent to search by the civilian in any true sense of the word; rather, it is a direct result of the power and authority of the officer; the possible consequences the officer may choose to visit upon the civilian, within the officer’s discretion; and the very human and overwhelmingly common desire to conform to the wishes of any authority figure. Lichtenberg’s findings showed that we could only view consent searches as based on actual consent only in the narrowest, most legalistic terms. And yet, consent searches have always served as one of the key components of traffic enforcement–based racial profiling; they are the tool that converts a routine traffic stop, which (without probable cause to suspect some other crime) would never allow a search of the vehicle, to become an investigation for drug offenses or other crimes for which police have no evidence. In that situation—a traffic stop in which police have no probable cause or even reasonable suspicion to suspect drug trafficking or other crimes—police could not legally insist on a search of the vehicle, its drivers, or passengers. The simple expedient of so-called voluntary consent allows searches to take place anyway, without evidence, 90 percent of the time because the vast majority of people will not defy the authority of the requesting officer or are afraid that refusing the request will make them “look guilty.” The racially skewed statistics on who ends up being searched pursuant to consent, in study after study, show just how important consent searches have become to this pernicious practice. Taking away the ability to search, absent probable cause or reasonable suspicion to believe an offense has been committed, would go a long way toward curtailing these practices. The US Supreme Court has made clear that it has no interest in reexamining the constitutionality of so-called consent searches. (See, e.g., Ohio v. Robinette, 519 U.S. 33 (1996), reaff’g Schneckloth v. Bustamonte, 412 U.S. 218 (1973).) But eliminating this tool, except in the presence of evidence that would support a finding of probable cause, would not require a change in constitutional law. State lawmakers can make this happen with the passage of statutes that prohibit police searches or requests to search without probable cause; law enforcement leaders can do it themselves, by creating police department rules and regulations that do the same thing. We need not wait for the US Supreme Court to act.
Second, every police department must keep data on all traffic stops, including not just who is stopped and appropriate demographic data on each person but also on post-stop activity such as searches and recovery of contraband. The same must be done for all other types of routine enforcement that results in searches, such as stop and frisk activity. As discussed earlier, Missouri and North Carolina have such statutes; in Illinois, which had a law passed 15 years ago (sponsored by then state senator Barak Obama) that would expire June 30, 2019, the legislature passed a new law that will make data collection an ongoing obligation. Illinois House Bill 1613 passed both houses of the state legislature on May 21, 2019. (At this writing, the governor has not yet signed the bill.) Every state must pass such a law. And with those data, supervisors and all police commanders must analyze them to see whether racially skewed patterns of enforcement, with no explanation other than racial discrimination, emerge. If and when those patterns become evident, the responsible officers, units, precinct commanders—everyone—must be held accountable.