Subscribe
Memorial murals of George Floyd and Breonna Taylor beside the Allegheny River in Pittsburgh.

Memorial murals of George Floyd and Breonna Taylor beside the Allegheny River in Pittsburgh. (Jeff Swensen/for The Washington Post)

In 1988, as Joe Collum drove the New Jersey Turnpike to his new job at a local TV news station, he noticed a recurring scene on the side of the highway: White state troopers rifling through the belongings of Black and Latino motorists.

Collum, an investigative reporter, scoured arrest records in dozens of municipalities. Along the turnpike, he found, Black and Latino drivers accounted for the vast majority — 80% — of all state police arrests on the turnpike. “Without Just Cause,” his investigative report on WWOR-TV Seacaucus, introduced the world to the term “racial profiling.”

“There was this big initiative to stop drugs coming in and out of New York City, and they were transparent about that,” Collum said, “but the troopers individually and collectively, to some degree, decided that the best way to catch people with drugs was to target dark-skinned people.”

Collum’s reporting exposed how troopers’ biases and assumptions about people of color had infected policing along New Jersey’s main artery. State police denied they were targeting minorities. But a group of attorneys, motivated by Collum’s reports, sued troopers, and a state court affirmed for the first time the existence of racial profiling by law enforcement. The Justice Department ordered an end to the practice.

But nearly 30 years later, the most recent audit revealed that Black drivers were still being subjected more often to searches, arrests and uses of force after traffic stops by state police.

The attempted reforms after Collum’s revelation of racial profiling have led to yet another chapter in the long line of attempts to cleanse policing in the United States of its persistent afflictions — an ongoing exercise in reform that never ends.

For decades, police misconduct and the use of controversial tactics have fueled cycles of outrage that have been followed by commissions, studies and orders or promises to reform. In 1929, the federal Wickersham Commission produced 14 volumes of reports documenting widespread police corruption, including the use of the “third degree” to extract confessions. Eighty-six years later, after the killing of a Black man, Michael Brown, by a white officer in Ferguson, Mo., the Obama administration convened the Task Force on 21st Century Policing, which produced 116 pages of recommended changes in U.S. law enforcement.

Last year, the deaths of George Floyd and Breonna Taylor at the hands of police renewed the calls for change.

Since then, more than 2,000 policing-related bills have been introduced nationwide, according to the National Conference of State Legislatures. In April, Maryland became the first state to repeal its powerful Law Enforcement Officers’ Bill of Rights, a set of legislative protections that included scrubbing any record of complaints against an officer after a period of time. And in Congress, the Democrat-led House in March passed a bill named for Floyd that bars chokeholds and no-knock warrants; it is stalled in the Senate.

The Washington Post examined three historic firsts in policing reforms: the effort to stop racial profiling by troopers in New Jersey, the deployment of early warning technology to identify troubled deputies in Los Angeles and the use of federal intervention to force change on police in Pittsburgh.

The legacies of these firsts reveal the difficulty of remaking law enforcement. At each agency, the attempts have been stifled by entrenched cultures, systemic dysfunction, shifts in leadership and swings in public mood. Outrage at officers’ conduct eventually gives way to demands for aggressive enforcement when crime flares, and the cycle continues.

“Eventually, too many people get murdered, too many people get raped, and people start saying, ‘Come on. Beef up the police,’ “ said former Los Angeles County sheriff’s assistant chief Neal Tyler, who helped develop the first computerized early warning system. “And we do that, and then slowly, eventually, we lose our resolve and the pendulum swings back the other direction, and there’s a police scandal of epic proportions.

“Like clockwork.”

Sheldon Williams is a former Pittsburgh police officer and now a pastor and member of the Citizen Police Review Board.

Sheldon Williams is a former Pittsburgh police officer and now a pastor and member of the Citizen Police Review Board. (Jeff Swensen/for The Washington Post)

In 1995 in the suburbs of Pittsburgh, an officer pulled over Jonny Gammage, a Black businessman who was at the wheel of a luxury sports car. He had been stopped for erratic driving. Police wrestled Gammage to the ground, putting pressure on his chest and neck until he died of asphyxia. His last words: “I’m only 31.”

Gammage’s death, along with fatal shootings by the Pittsburgh police and other allegations of officer misconduct, pushed the American Civil Liberties Union to sue the department. The ACLU had been investigating the police for three years, gathering evidence of alleged civil rights violations, including by jump-out squads, free-roaming patrol units that stopped people on the street.

After filing suit, the ACLU asked the Justice Department to take on the case.

“A short time later, they had three attorneys in Pittsburgh hungrily downloading everything we’d learned,” said Witold “Vic” Walczak, the ACLU of Pennsylvania’s legal director, who handled the case.

The Justice Department investigated and for the first time invoked a civil power tucked into the Violent Crime Control and Law Enforcement Act of 1994, filing its own lawsuit against Pittsburgh to address a “pattern or practice” of civil rights abuses by the city’s police. In 1997, the city became the first to enter into a federal consent decree, agreeing to changes that would be monitored by the government.

In 83 paragraphs, Justice laid out a plan for the department of 900 officers, including mandates to implement diversity training, better document traffic stops, reduce the use of strip searches and track complaints filed against officers.

But when the federal monitors moved on nearly a decade later, the pressure for reform had dissipated.

Sheldon Williams, who joined the department as an officer in 1997, experienced it firsthand.

Before Williams became a Pittsburgh officer, his encounters with police had included running from them as a child in Wilkinsburg, a mostly Black borough in the Pittsburgh metropolitan area. In high school, he said, his family moved to a white part of town, where Williams said most residents did not fear officers. “It’s hard to make people see the unpredictability Black Americans face when dealing with the police,” he said.

Williams, who signed up as the department pushed to diversify, took a job in a new public integrity unit, a team created to look for internal corruption. He said he thought he would be applauded for working to restore public confidence. Instead, he was shunned.

Commanders embraced the reforms, he said, but the rank and file often did not. “As I came in, I started seeing the shifts and the changes right away,” he said. “But I also heard the kicks and scrapes and the complaints about it.”

Under the consent decree, officers were taught how to navigate their own biases and understand why Black people might fear them. The decree also gave police nonlethal tools, including pepper spray, he said. “We used to get issued blackjacks to control suspects,” said Williams, referring to the leather-wrapped piece of metal with a wrist strap.

Officers had to document more precisely why they stopped and questioned people and provide more justification for the use of force.

And the department agreed to track officers accused of misconduct.

“There was a sense from members of the community that the department was doing things that increased the professionalism of the officers,” said Williams, who retired in 2011 and has since become a pastor and member of the Citizen Police Review Board.

In 2002, federal monitors pronounced the Pittsburgh Bureau of Police rehabilitated, ending the consent decree except for a provision to monitor the internal affairs department as it investigated a backlog of misconduct cases. That ended in 2005.

The decree worked initially, said David Harris, a professor at the University of Pittsburgh who wrote a book about the intervention, “A City Divided.”

“It changed some of the most dysfunctional parts of the way the department was operating,” Harris said. But once the federal pressure ended, reforms slipped, he said. “When the public pressure is off, departments tend to backslide.”

In 2006, there was a newly elected mayor, and Police Chief Bob McNeilly lost his job. Old problems reemerged.

“For reform to work, you have to have a chief that can lead it, a city government that supports it and officers and unions who go along,” said McNeilly, who was chief of the department from 1996 through 2006. “But politics change, and that’s usually what ends a reform. They hire supporters instead of leaders.”

McNeilly noted that the chief who replaced him held the job just eight months, and the next lasted eight years before going to federal prison for stealing more than $30,000 of the department’s money.

In 2010, officers in a jump-out squad - similar to the ones cited in the original ACLU complaint — arrested and beat 18-year-old Jordan Miles, a violinist from a prestigious art school, as he walked to his grandmother’s house.

Miles said he thought he was being robbed and ran; police said they thought he had a gun, though one was never recovered. Images of the teen’s swollen face and wounds in his scalp where police had torn out his dreadlocks outraged residents. Police later settled the Miles case for $125,000.

“The Miles case was an indication to me that whatever good the consent decree had done had worn off,” Harris said.

The department was mired in one controversy after another.

In 2012, officers shot and paralyzed Leon Ford during a traffic stop; police said they thought he was reaching for a gun. Ford sued, and the city settled the case for $5.5 million.

In 2014, an investigation of theft and corruption within the Pittsburgh department led newly elected Mayor Bill Peduto to declare in an interview with PublicSource: “[M]any of the officers who get concerned when I talk about this don’t realize how close we are to going into another consent decree.”

The next year, the police settled another ACLU lawsuit by agreeing to remedy allegations that they discriminated against Black applicants; Blacks accounted for 26% of the city’s population but only 4% of new officers.

In 2017, a former police officer was sentenced to prison after he was captured on video beating a drunken teenager at a high school football game. The officer said he thought the teen was about to attack him.

Last year, Pittsburgh police drew widespread criticism and another class-action lawsuit, this one alleging that officers used excessive force on people protesting Floyd’s death.

Christy E. Lopez, who conducted civil rights investigations for the Justice Department from 2010 to 2017, said Pittsburgh showed that consent decrees alone cannot fix broken departments. Since the agreement with Pittsburgh, Justice has used federal consent decrees, or settlements, at least 40 times.

She said federal officials have come to understand that troubled departments need to focus on underlying issues and disincentivize aggressive policing — for example, ending rewards to officers for large numbers of arrests. “We realized there were more structural issues at play, and we could not focus only on use-of-force reporting,” Lopez said.

Last June, Peduto, the mayor, vowed to reform the Pittsburgh department again, creating a task force to “to lay out a blueprint ... that will make Pittsburgh a safe and healthy place for all of our citizens, especially for members of our Black community.”

In October, the task force recommended changes including more transparency about the use of force and better tracking and disciplining of officers who break the rules.

Many of the recommendations closely resembled the changes outlined in the consent decree 23 years earlier.

Joe Collum, seen at home in Plantation, Fla., in April 2021, introduced the term "racial profiling" in reporting he did for a New Jersey TV station in 1989.

Joe Collum, seen at home in Plantation, Fla., in April 2021, introduced the term "racial profiling" in reporting he did for a New Jersey TV station in 1989. (Zack Wittman/for The Washington Post)

Pittsburgh police Sgt. Tiffany Costa of the Community Engagement Office plays chess with civilians in Market Square.

Pittsburgh police Sgt. Tiffany Costa of the Community Engagement Office plays chess with civilians in Market Square. (Jeff Swensen/for The Washington Post)

The Justice Department found in 2015 that Lancaster Station sheriff's deputies had harassed and racially profiled Black and Latino residents of federally subsidized Village Pointe, shown June 2.

The Justice Department found in 2015 that Lancaster Station sheriff's deputies had harassed and racially profiled Black and Latino residents of federally subsidized Village Pointe, shown June 2. (Kyle Grillot/for The Washington Post)

In 1992, the Los Angeles County Sheriff’s Department tapped then-Lt. Neal Tyler to lead a computer-based initiative: an early-intervention system that would collate complaints against officers and their use of force. The data would be used to determine whether an officer was a danger to the community and needed counseling.

The department hired a handful of tech-savvy civilians to convert more than 7,000 deputies’ personnel files into a digital database. It was the first such system in the country, according to Merrick Bobb, a policing expert who monitored the department in the 1990s and 2000s.

But the system eventually fell into disuse, according to interviews and reports.

Previous early-warning systems in New York and Miami had been maintained on paper. As early as 1980, a U.S. Civil Rights Commission report, titled “Who’s Guarding the Guardians,” had recommended that all police departments adopt such systems.

The department’s move to track officer conduct came after the Los Angeles Times and the San Fernando Valley Daily News reported on settlements paid by the county, revealing which deputies had cost taxpayers the most money, Tyler said.

“It’s kind of a simplistic view, but 90% of your problems are from 10% of your people, right? Well, we didn’t know who the 10% were,” he said.

The process of data collection and entry alone took about four years, Tyler said. The next step was to write guidelines for supervising officers on how to intervene and mentor troubled officers.

That is where Tyler’s team experienced the most resistance. “It forced [supervisors] into a situation where they weren’t just parental figures to their subordinates, that they really were there to examine on behalf of the department whether or not the use of force was valid,” said Mike Graham, then-assistant chief and Tyler’s boss.

Tyler ran into another unanticipated problem: The system seemed to suggest the department should re-examine the work of some of what were considered to be its best officers.

“Among the people who were flagged by the [early-intervention system] were what we call hard-charging, you can say, aggressive, assertive, initiative-taking officers who didn’t just wait around for something to happen, but were actively seeking out criminal activity,” Tyler said.

When New Orleans police set up their system in the late 1990s, officers flagged for intervention were sent to a class for additional training. Officers came to refer to the class as “Bad Boy School” and it became a “badge of honor” in some circles, according to a later Justice investigation.

The mixed-messaging on aggressive policing also surfaced in a 2003 Justice Department-funded study by researchers Geoffrey Alpert and Sam Walker. “One disturbing finding was a slight tendency of early-warning officers to be promoted at higher rates than control officers,” they wrote.

In Los Angeles, the system was credited with dropping the rate of force-related lawsuits and payouts. But Bobb, who monitored the department for the county’s board of supervisors, believed it was capable of much more.

In 2003, Bobb reported that the system was slipping into disrepair. Information sat for six months before it was logged in the database by the Risk Management Bureau, which was understaffed. And many managers had little understanding of how the database worked, he wrote.

“LASD has the Rolls Royce of risk management software and procedures,” Bobb wrote in his 117-page semiannual report. “And indeed it is the Rolls. But when the odometer is checked, it is apparent that it has hardly ever been taken out of the garage.”

Settlement costs began rising steadily for the sheriff’s department in the 2010s, from $5.6 million in 2011 to $50.9 million in 2016, according to public records.

In 2011 in Antelope Valley, Black and Latino residents of federally subsidized housing alleged that sheriff’s deputies were harassing them. The allegations, including warrantless searches of homes and surveillance, triggered a federal civil rights investigation.

The Justice Department concluded in 2015 that deputies had abused and profiled residents in Section 8 housing, noting the failure of the department’s early warning system in its two Antelope Valley stations.

“LASD’s early warning system does not adequately identify or effectively respond to Antelope Valley deputies with repeated complaints or other histories indicating a need for intervention,” the Justice Department’s report said.

The ensuing consent decree focused on the Antelope Valley stations but called for department-wide changes, including a revamped early warning system to help eradicate racial profiling by flagging “deputies with the highest stops of people of color.”

Last year, the federal monitors reported that the sheriff’s department had finally “started the process for this type of station-level review.”

The county’s inspector general, Max Huntsman, who since 2013 has been tasked with monitoring the sheriff’s department, said that nothing as comprehensive as the system Tyler implemented in the 1990s exists today.

“The Los Angeles Sheriff’s Department is an illuminating example of why any reform effort can be destroyed if it’s not part of a universal reform effort with real teeth,” Huntsman said.

The sheriff’s department did not respond to requests to interview Sheriff Alex Villanueva or answer questions about the status of its early intervention system.

Tyler retired in 2018 when Villanueva became sheriff. These days, Tyler, 69, teaches a class at the Los Angeles Museum of Tolerance to incoming sheriff’s department cadets. He sees early idealism curdle into cynicism in follow-up visits.

“You start with a group of young, fresh-faced guys who know what the right thing to do is. But after the rigors of this high-stress, high-pressure job, they lose sight of that somehow, and they start dealing in different shades of gray,” he said.

Nationwide, the use of early warning systems appears to have declined. In a 2007 federal survey of nearly 1,000 departments, 39% reported some sort of early intervention system. In 2016, when the survey was updated across 2,746 agencies, the proportion using such systems had dropped to 25%.

Last June, Alpert, the researcher, testified before mostly Southern law enforcement leaders on President Donald Trump’s Commission on Law Enforcement and the Administration of Justice. He mentioned early warning systems. Some had no idea what he was talking about.

“They thought that was a new concept,” Alpert said.

Joe Collum’s multipart report in 1989 on racial profiling by New Jersey State Police troopers drew condemnation from the police and some members of the public: It happened to be broadcast in the same week that a state trooper was shot and wounded during a traffic stop.

“There were people who said, ‘Hey, they’re making arrests. What’s the problem?’ Our switchboard was inundated,” Collum said. “But we got very little pushback from Blacks or Hispanics on what we reported.”

State police insisted they pulled people over at rates in line with the driver demographics on the turnpike. Only arrests were disparate and troopers made them without racial bias, they said. (In the 1990 Census, 79% of New Jersey residents identified as white.)

Two of the troopers involved in training for the stops were Brian Caffrey and Andy Mastella. Caffrey won Trooper of the Year in 1982 after making five drug seizures that year worth more than $1.2 million, according to the award citation. Mastella won his second Trooper of the Year award in 1984 after arresting 35 people and seizing more than 100 pounds of cocaine and a half-ton of marijuana that year, according to the agency.

The Justice Department paid for Caffrey and Mastella, who are both white, to travel the country teaching “highway drug interdiction” to police departments in more than 40 states, Caffrey and Mastella told The Washington Post.

Together, they developed a line of questioning that they believed uncovered liars and another script for compelling consent for a search, Mastella said.

Caffrey and Mastella, interviewed at their homes in Florida, where each has retired, acknowledged the existence of racial profiling by some of their colleagues and what they said was a minuscule number of police officers.

“I’d guess a fraction of 1% of officers racially profiled,” Caffrey said. “I didn’t. I don’t see race.”

Mastella said their training was based on common sense. “If you see two people speeding,” Mastella said, “and one is a grandma and one is a car full of young men, which one are you going to pull over? It’s not about race.”

After Collum’s investigative report, a handful of public defenders took up the cause of 17 Black and Latino people charged with drug crimes based on turnpike stops and seizures. The attorneys argued that state police had used selective enforcement, violating the equal protection and due process clauses of the 14th Amendment.

The plaintiffs analyzed available police data and collected statistics on 42,706 turnpike drivers by standing on a roadside and using binoculars to count Black and white drivers. The attorneys said Black motorists were more than four times as likely to be stopped as whites between turnpike exits 1 and 3.

“We felt that doing it on the turnpike was sort of an easier thing than trying to make judgments in cases about profiling, let’s say, in a town or city, because the turnpike was a closed system with its own population,” Jeffrey Wintner, one of the public defenders, told The Washington Post.

In court, state police generally argued that it was usually impossible for them to tell whether drivers were Black or white before they were stopped.

Caffrey was called to testify as a teacher of interdiction techniques. He said that “ethnicity is something to keep in mind,” though not a tipoff, according to court documents.

In 1996, the New Jersey Superior Court ruled in favor of the plaintiffs, becoming the first state court to affirm the existence of racial profiling. The ruling suppressed evidence in hundreds of arrests and led to the dismissal of more than 150 drug cases. “The utter failure of the State Police hierarchy to monitor and control a crackdown program operated by the State Police or investigate the many claims of institutional discrimination manifests its indifference if not acceptance,” the court wrote in its decision.

Mastella remains unmoved.

“All that racial profiling stuff started because some lawyers couldn’t get their guilty clients off, because all the evidence was against them, so they needed to cook up some reason the cops were wrong,” Mastella said.

Collum said that years later, the Oxford English Dictionary informed him that his reporting contained the first use of the term “racial profiling.” He later wrote a book titled “The Black Dragon: Racial Profiling Exposed.”

“When I think back to then,” Wintner said, “there was something of a movement in the direction of stopping it, but not like a tidal wave that could actually do it. white people were never as concerned as minorities about it. Many of them aren’t at all concerned.”

Despite the court’s ruling, allegations of racial profiling on the New Jersey Turnpike persisted.

In 1997, the Newark-based Star-Ledger reported that minorities represented 75% of all turnpike arrests. Several high-profile lawsuits over the stops — including one in 1998 in which troopers fired into a van and injured three unarmed young men of color on the turnpike — prompted a federal response. In 1999, the Justice Department put the force of more than 2,500 state troopers under a federal consent decree, ordering the department to track racial disparities in turnpike enforcement and eliminate discrimination.

That year, for the first time, the New Jersey attorney general publicly acknowledged racial profiling by troopers, calling it “real, not imagined.”

Then, in 2003, 12 motorists who sued the state police arguing that they had been racially profiled won more than $775,000.

Finally, in 2010, a U.S. District judge ended the decree, indicating that the state police had satisfied the requirements. The state police superintendent, Col. Rick Fuentes, called it a “watershed moment for the division,” adding that the reforms had “instilled public confidence and trust.”

But by 2016, disparity in traffic stop outcomes had returned: Black drivers accounted for 20% of traffic stops by troopers in the first half of the year, but were 39% of the drivers subjected to additional actions after the stop, a review by the state attorney general’s office found.

For example, Black drivers represented the largest proportion of motorists subjected to probable cause searches after traffic stops — 40%, compared with 38% for white drivers, who account for 59% of all stops.

“As noted throughout this report, the proportions of Black drivers involved in specific post-stop activities were high. ... Black drivers made up the majority of stops involving vehicle exits, probable cause searches, use of force, and charges filed in the current reporting period,” according to the 2018 traffic enforcement review.

The report said Black drivers were more likely to have outstanding warrants, representing 52% of all wanted people. “Thus, the trooper has no discretion and must arrest the individual, which may be related to the larger proportion of Black individuals arrested or Black drivers who were arrested and searched,” the report noted.

Following the 1999 consent decree, state law required monitoring reports twice a year, though the 2018 report is the most recent. State police and the attorney general’s office declined to share more current statistics. The office, however, said it is working on an online dashboard with the latest traffic enforcement data.

In May 2020, the New Jersey comptroller’s office, also tasked with monitoring the Justice reforms, released a report indicating that state police methods for tracking racial discrimination in traffic stops were inadequate because troopers’ stops were being compared to one another, not to the overall population of the turnpike. If officers were biased across the board, the comptroller’s office argued, the state police would have no way of knowing.

Col. Patrick J. Callahan, superintendent of the New Jersey State Police since 2017, told The Post the department is doing all it can to train officers to police equitably and is tracking officer behavior to prevent profiling. Callahan, however, said his department “should be doing a better job” of educating people about its efforts to end racial profiling.

“With all the mitigation efforts we have in place, we haven’t uncovered [racial profiling],” Callahan said. “The last thing people want to hear is, ‘It is what it is,’ but . . . I would hold us up against any agency in the country as far as the scrutiny we put ourselves under.“

Sign Up for Daily Headlines

Sign up to receive a daily email of today's top military news stories from Stars and Stripes and top news outlets from around the world.

Sign Up Now