hill country observerThe independent newspaper of eastern New York, southwestern Vermont and the Berkshires


News & Issues February-March 2017


Unequal justice?

Bennington, Rutland rank high in study of police bias


rutland policeBy TRACY FRISCH
Contributing writer


Black drivers who are pulled over by city police in Rutland are at least six times more likely than white drivers to wind up being searched.

But police searches of black drivers in Rutland and elsewhere across Vermont are less likely than searches of white drivers to turn up drugs or other contraband.

These are among the findings contained in a new study that offers the most comprehensive analysis to date of racial disparities in policing in Vermont. The study, released in January, examined data from 29 law enforcement agencies across the state, including the Vermont State Police and municipal police departments in Rutland, Brandon, Bennington and Manchester.
The report found that in Bennington and Manchester, black drivers stopped by police in 2015 were five times more likely than white drivers to be subjected to searches.

The statistical study, entitled “Driving While Black and Brown in Vermont,” used data collected under a new state law that required police agencies to keep track of the race of drivers in all traffic stops, beginning in 2015.

The study’s authors – economics professor Stephanie Seguino of the University of Vermont and urban planning professor Nancy Brooks of Cornell University -- analyzed race and gender data for traffic stops from 2015 as well as data from previous years in jurisdictions where it was available.

State legislators crafted the new data-reporting law in 2014 after a series of cases around the state in which police agencies were accused of racial profiling. Advocates who pushed for the reporting law say the study’s findings confirm their longstanding concerns about racial bias in Vermont’s police agencies.

“A steady stream of data bear out what people of color have been saying for decades in Vermont and across the country,” said Jay Diaz, a staff lawyer at the state chapter of the American Civil Liberties Union. “People of color are stopped much more frequently. And they’re not being stopped due to actual infractions or infractions that are policed equally, but based on the color of the driver or another person in the car.”

Bennington Police: Tony Israel photo
Overwhelmingly white
The issue of racial bias in policing may seem an unlikely controversy in Vermont, which remains one of the two whitest states in the nation. Only 1.6 percent of the state’s population is black, according to census data used in compiling the new report.

But civil rights advocates say Vermont’s incarceration rate alone raises questions about whether the state’s laws are being evenly enforced: Fully 10 percent of the state’s prison population is black.

Robert Appel, a Burlington lawyer who served as chairman of the Vermont Human Rights Commission for more than a decade, said he suspects the relative rarity of people of color in Vermont, especially in rural areas, contributes to racial bias.

He recalled the experience of a former neighbor, now deceased, who was born in 1946 and had never seen a black person until he entered the Army. Absent any first-hand experience, the neighbor took his ideas about black people from television, which in those days portrayed them mainly in crime dramas. Although Vermont has become slightly more diverse since then, many areas remain overwhelmingly white, and most police officers are recruited from within the state, Appel said.

Rutland Police StationAcross Vermont, the new study found black and Hispanic drivers were more likely than white or Asian drivers to be pulled over by police in discretionary traffic stops. In Bennington, the black share of traffic stops was nearly 2.5 times higher than the black share of the county population.
Statewide, the rate at which black drivers were pulled over ranged from 161 percent to 193 percent more than their share of the driving or community population respectively. For Hispanic drivers, the stop rate was 179 percent greater. Conversely, white and Asian drivers were stopped less frequently than would be expected based on their population share.

The discrepancy cannot be explained as a statistical anomaly related to the state’s small black population. The study points out, for example, that Asians represent a similarly small share of the state population – 1.8 percent – but the rate at which black drivers were searched was seven times the rate for Asian drivers.

Overall, black drivers who are pulled over by police across Vermont are four times more likely than whites to be searched. For Hispanic drivers, the likelihood of a police search is three times greater than for whites. All of these statistics are for discretionary police stops – in other words, in cases where there is no warrant for a person’s arrest.

Though they were searched far more frequently when stopped, the study found black and Hispanic drivers were less likely than white and Asian drivers to be found with contraband. Seguino, the study’s co-author, explained that this finding suggests searches of black and Hispanic drivers were subject to a lower threshold of evidence.

“The search disparities we find are very consistent with findings in other states,” Seguino said.
But she added that the disparity in search rates was even wider in Vermont than in some other states where racial bias in policing has become a bigger issue, such as in Missouri and North Carolina.

Although black and Hispanic drivers were more likely than white drivers to face tickets or citations after Vermont traffic stops, those who were searched were far less likely to be arrested. Citation and arrest data provide strong evidence that police often were not finding anything serious enough to warrant taking these drivers into custody, Seguino explained.

The study also found a significant gender disparity by race. While 62 percent of white drivers in traffic stops statewide were male, 88 percent of the black drivers stopped were male. The study says this suggests that “black and Hispanic males, in particular, are targets for heightened police scrutiny.”


Targeting the heroin trade
Locally, police in Bennington and Rutland both have faced accusations of racial profiling in recent years, and both agencies ranked high on some measures of racial disparity in the new study.
Police in both communities have been struggling in recent years to curb the flow of heroin that has been feeding Vermont’s growing problem with opioid addiction.

Bennington Police Chief Paul Doucette did not respond to multiple requests for an interview for this report.

Seguino suggested black and Hispanic drivers are experiencing heightened scrutiny because of the opioid crisis in Vermont. She stressed, however, that despite the stereotype that blacks and Hispanics are disproportionately involved in drug trafficking, “our data doesn’t support this assumption.”

Rutland Police Chief Brian Kilcullen said his department is continuing to work to reduce bias in its policing.

“Going forward, we’ll be looking at this as baseline,” Kilcullen said of the figures used in the new study, which included five years’ worth of data from Rutland covering 2011-15.

The police chief said he’ll focus on improving data quality and identifying teachable moments for his officers. Although nearly 3 percent of traffic stops in the city’s dataset lacked information on race, for example, he said this problem decreased over the course of the five-year period.
The data showed 5.5 percent of black drivers stopped in Rutland were subjected to searches, while just 0.9 percent of white drivers were searched.

Kilcullen responded to this by pointing out that only 21 black drivers were searched across the five-year period.

Even so, when compared with the proportion of white drivers who were subjected to searches, these 21 drivers represented a much larger share – 625 percent larger – of the pool of black drivers in traffic stops. And the number of black drivers stopped was already larger than would be expected if drivers of every race and ethnicity were stopped at the same frequency.

In addition, of the 21 black drivers who were searched, not a single one wound up being arrested, whereas 13 of the 108 white drivers searched during the same period were arrested.
When tickets or citations were factored in, Rutland police found contraband resulting in a ticket or arrest in 72 percent of the searches of white drivers but in only 32 percent of searches of black drivers. This makes the “hit rate” for blacks searched less than half of the rate for white drivers who were searched.

“That would suggest black drivers are being over-searched, and that perhaps officers have a lower threshold of evidence when deciding to search a black driver,” Seguino said. “That is the real story here.”


Data challenge assumptions
Kilcullen said he believes heroin trafficking accounts for the racial disparities seen in traffic stops by Rutland police. But his theory on the role played by drug crimes is quite different from Seguino’s. He said his agency could extract information for each of the incidents to get a fuller picture.

“We need to look at context,” Kilcullen said. “Drugs are trafficked into Vermont by train and by car. What we’ve seen are African-American males coming to Vermont from New York City. Some of them develop a relationship with particular Vermont females. A white female passenger may have drugs.”

Told of the police chief’s comments, Seguino said his view isn’t supported by the data she reviewed.

“His stereotype about blacks is precisely what the data dispute,” Seguino said. “Blacks are less likely to be carrying contraband than white or Asian drivers. The police are casting too wide a net. The great benefit of data is that we can test hypotheses such as his, and his hypothesis is simply refuted by the data.”

Both Kilcullen and Seguino agreed that additional data would be illuminating. The data on traffic stops, for example, focuses on drivers and does not include information about passengers, their races and whether they were searched.

Kilcullen took over as police chief in November 2015, so nearly all of the figures examined in the new report predate his tenure. But Rutland police have struggled over the past decade with scandal, poor morale and incidents of officer misconduct – including allegations of racial bias. Kilcullen’s predecessor, James Baker, was brought in to clean up the department in 2012.
Many details of the Rutland force’s problems came to light two years ago in a discrimination lawsuit brought by Andrew Todd, who for many years was Rutland’s only black officer. Todd alleged a supervisor and co-worker created a hostile work environment, routinely using racial slurs and engaging in racial profiling of black citizens. Among other examples, the lawsuit claimed city police observed the arrival of the nightly Amtrak train from New York City and detained and even strip-searched black passengers without probable cause, while white travelers were not searched.

Todd resigned in 2012 and later became a state trooper; Rutland hired another black officer in 2015. The city ultimately settled Todd’s lawsuit for $975,000 -- a sum Appel characterized as one of the largest settlements, if not the largest ever, for an employment discrimination claim in Vermont.


Neighboring states lag behind
Seguino said 31 states collect some demographic data on traffic policing, but very few do an adequate job of it. New Hampshire’s law, for example, pertains only to state police, and its data categories are more limited than Vermont’s.

Other states don’t make the data publicly available or lack transparency in other ways. In Maine, for example, police agencies submit traffic stop data only on paper, which makes the information nearly impossible for researchers to access and analyze.

New York is among the states that do not collect any data on the race of drivers in traffic stops. Race must be recorded only when a person is arrested. State legislators have sponsored bills on racial profiling and data collection in traffic stops since at least 2003, but only the Assembly has passed such legislation. But some individual jurisdictions, including the city of Albany, have responded to public concerns by opting to collect race data for all traffic stops.

Massachusetts in 2000 became one of the first states to pass legislation aimed at eliminating racial and gender profiling in traffic stops, but time has revealed major deficiencies in the law. Officers only collect demographic information for traffic stops if they issue warnings or tickets or makes arrests. In those situations, they must indicate if a search is conducted. The law also lacks any provision for enforcement if police agencies don’t comply with the reporting requirements.

In 2004, the Northeastern University Institute for Race and Justice analyzed 27 months of data from more than 300 law enforcement agencies in Massachusetts. The analysis found significant racial disparities in more than 80 percent of these police forces. Under the 2000 law, that finding allowed the state to require those police agencies to collect more comprehensive data for one year, but it appears the state never exercised this option. Nor does it appear that any subsequent data analysis was conducted.

The same day the final Northeastern report was released, the Massachusetts Chiefs of Police Association issued a 32-page position paper attacking the data collection and analysis. The police chiefs contended that the overall racial disparity in traffic stops was “so low” that no police departments should be compelled to collect additional data.


Officer-by-officer breakdown
As part of her study in Vermont, Seguino examined data for individual police officers in departments that provided data. Only officers that made at least 50 traffic stops were included in this analysis. Seguino calculated the percentage of officers in each agency that stopped black drivers at a rate 50 percent or more above their share of the population.

The top three police agencies on this measure were all located in Rutland and Bennington counties. In Brandon, 67 percent of officers stopped black drivers at a rate above the threshold, while 63 percent of officers did so in Manchester, and 60 percent did in Bennington.

So in Bennington, where 1 percent of the overall population is black, black drivers accounted for 2 percent or more of the traffic stops recorded by 13 of the town’s 24 police officers. And with four officers, black drivers were behind the wheel in more than 6 percent of the vehicles stopped.
And in Manchester, where slightly more than 1 percent of the population is black, seven of the town’s eight officers reported that black drivers accounted for 2.5 percent or more of their traffic stops.

Individual officers are not identified by name in the data.
State Rep. Bill Lippert, D-Hinesburg, who shepherded the new data-reporting requirement through the Statehouse, said the law is the result of nearly a decade of effort to collect more information about racial profiling and police bias. Initially, he said, lawmakers held back while advocates and police agencies tried to create a voluntary reporting system. Although some, including Rutland’s police force, took part, most police departments did not.

Even with mandatory reporting now in place, some data gaps remain.
“Some departments did a stellar job; others did not comply,” said state Rep. Kiah Morris, D-Bennington, who has focused on fair-policing issues and, as a black woman, said she has at times been treated with suspicion by police because of the color of her skin.

In southwestern Vermont, the Bennington County Sheriff’s Department has not provided data under the new law. Only three of the state’s 11 county sheriff’s departments did so, and another 20 municipal police agencies, mainly in smaller towns, also failed to comply with the mandate. But the 24 municipal departments that did supply data accounted for 78 percent of the Vermont population in communities with local police forces.

Last year, Vermont Public Radio requested public records for traffic-stop demographic data from about 20 police agencies around the state. In many jurisdictions, race and gender data were spotty or nonexistent -- or were not made accessible. VPR reported that the Bennington County sheriff said it would cost $750 to see the traffic-stop records, even though the records are supposed to be readily available under state law.

“Many of the smaller police agencies don’t have the resources or skill sets to submit the data electronically,” Seguino said. “They probably need the support of the state.”
In her own review of the data, Seguino said she found a number of gaps. The Rutland County sheriff, for instance, provided no data on search outcomes or subsequent police actions, and other agencies had many instances in which information about race was missing in traffic-stop records.

Lippert said the data made available through the law is essential. But he also said the effort is insufficient “unless you use the data to correct the disparities that are found in roadside stops.”
And what occurs in traffic stops “doesn’t begin to address all the issues in the criminal justice system,” he added. “There are much broader issues in terms of implicit bias.”

To tackle the broader issues, Morris said she and other lawmakers are proposing to establish a racial justice oversight commission.

But trying to bring about change on race and policing at the Statehouse is not something that legislators can do in a vacuum, she noted. Taking up the issue brings out “multiple stakeholder groups from Border Patrol to immigrant activist groups and everything in between,” Morris said.


Shedding light on bias
As an economist, Seguino studies racial and gender disparities and their effects on employment and income. As a member of the Burlington school board, she said she’s become aware of the racial disparities in student suspensions.

“You learn that the issues are quite similar,” she said.
Seguino said she undertook her analysis of the statewide traffic stop data as a voluntary commitment. She had worked on a similar project as faculty adviser to an earlier study of data voluntarily collected by city police in Burlington. She was not paid for this work, though she said she was able to obtain funding from the University of Vermont for an assistant to “clean up the data.”

Before Seguino took on the project, Northeastern University had analyzed five years of race and gender data collected by the Vermont State Police.

“A number of people felt that Northeastern’s analysis wasn’t adequate,” Lippert said, because it “lumped Asian-American drivers with other nonwhites.”

Seguino agreed that Northeastern’s approach was problematic, because “Asians get fairly equitable treatment compared to blacks and Hispanics.

“Putting all three groups together obscures the disparities,” she explained.
Lippert said the purpose of collecting the data is to illuminate the differences in how people are treated, not to minimize them.

In framing the new analysis, Seguino said she set out to clear up other shortcomings in the previous effort.

“We have been interested in a straightforward methodology that can be replicated by others,” Seguino said. “We know the Vermont context and have done ride-alongs with police here, giving us a sense of the underlying issues.”


Reducing risks
Some of those working to curb police bias point out that routine traffic stops can easily escalate into situations with dire consequences. If the stop wasn’t justified or an officer’s handling of it is unfair, that unfairness can quickly be magnified: Traffic stops can turn violent or deadly, and confrontations may be more apt to escalate when drivers feel they are being targeted or scrutinized unfairly.

“Disproportionate contact with law enforcement leads to disproportionate arrests,” Appel said.
Undocumented immigrants also face greater risk from traffic stops, as getting pulled over may lead to their being detained by Immigration and Customs Enforcement officers and potentially deported, separated from their families and livelihoods.

Lippert said Migrant Justice, an organization that represents “a lot of undocumented folks who are keeping our farm economy going,” mobilized members and offered testimony at the Statehouse in favor of impartial policing.

Lippert and Morris promoted a requirement that all police agencies adopt a policy for fair and impartial policing. The model policy created by Vermont State Police and the state attorney general contains a provision directing police to refrain from automatically contacting Immigration and Customs Enforcement when they encounter someone who may be undocumented. Although this provision reduces the risk of deportation, it still can occur, Lippert noted.

Advocates point out that when people have experienced racial profiling or fear an encounter with police might lead to deportation, they are less likely to seek out or cooperate with law enforcement in addressing crime.

“The distrust between the community and law enforcement just gets amplified,” Morris said.

State Police embrace change
In the effort to curb racial bias in policing, the State Police have taken the lead in Vermont. The agency has set an example for diversity in recruitment, community outreach, and training aimed at countering implicit bias.

“We have been working proactively toward bias-free policing since 2006 or 2007,” explained Capt. Ingrid Jonas, who directs the agency’s Fair and Impartial Policing program.

Early on, the Vermont State Police set up a Committee on Fair and Impartial Policing that has counted both Lippert and Appel among its members. The initiative has drawn praise from groups like the ACLU.

“We do applaud the State Police for their efforts,” Diaz said. “They have made some important strides and are taking it very seriously.”

Jonas said the agency has been working on “training, supervision and hiring the right people” – and that fair policing now is part of every review for promotion.

The agency screens against explicit bias, or overt bigotry, in hiring, but it also understands that everyone has implicit bias and that it must be addressed, she said.

Jonas defined implicit bias as “an unconscious mindset we might not be aware of.” Dealing with implicit bias requires “self-reflection, realizing that we have the potential to make unconscious judgments about people, places and things,” she said.

“We have to override that tendency to serve the public,” Jonas said.