Police initiatives to use data to tackle high-crime areas have shown mixed results.

By Ryan Lessard, Granite State News Collaborative

Editor’s note: This article is another installment of “Invisible Walls,” an ongoing joint project of the Granite State News Collaborative, NH Business Review, Business NH Magazine. The series explores how exclusionary zoning laws have reinforced areas of persistent poverty, impacting many aspects of community life, including crime, public health, affordable housing and access to economic opportunity in Manchester. The team used Manchester as a case study, but the same sorts of exclusionary zoning practices present in Manchester are common across the state, and likely have had similarly broad effects.

Joseph Lascaze grew up in Manchester with relatives who emigrated from Haiti. He remembers dribbling soccer balls with his cousins outside his uncle’s house on Laurel Street, playing basketball with friends at Pulaski Park and visiting a Union Street store for its Latin soft drinks after church. Nearly everywhere he went, he would see police scouring the center city.

“They were either monitoring someone or they were arresting someone,” Lascaze said. “We would either run from them or get away on our bikes if were doing stupid stuff when we were kids, or we wouldn’t talk to them.”

By 2005, when he was 17, Lascaze was arrested and charged as an adult for burglary, theft of firearms and home invasion. He served 13 months at Valley Street Jail. 

“I developed this ‘us versus them’ mentality real young, so I started making childish decisions,” he said.

From there, Lascaze was trapped in the revolving door of the justice system; it wasn’t until he served 13 years in the New Hampshire State Prison for Men in Concord that he escaped it. He’s now a statehouse lobbyist for the American Civil Liberties Union of New Hampshire, working to reduce the state’s incarceration rate and combat racial disparities in the justice system. 

Lascaze said police clearly prioritized patrols in the center city neighborhoods where he grew up. 

“There was a heavy concentration of a police presence in the neighborhoods where you had low income housing,” Lascaze said. “Based on actions, I would say there was definitely a bias towards those neighborhoods (and) the people in those neighborhoods. … The North End is not treated like Central or Union or Laurel Street.”

Lt. Matthew Barter of the Manchester Police Department has been thinking about these issues. For the past eight years, he has developed four different versions of the department’s hot spot policing programs.

The most intensive iteration of the program was in place between 2016 and 2018, when the department was using a predictive policing algorithm to direct so-called “hot spot patrols” to areas where trouble was expected. The software predicted where crime was likely to occur based on patterns identified in the crime data. 

During that time, a few neighborhoods were consistently flagged, Barter said. Those included pockets of the West Side and the center city area that includes census tracts 13, 15 and 16, a swath that stretches eastward roughly from Veteran’s Memorial Park and the SNHU Arena to Beech Street.

It didn’t take Barter long to realize why those areas kept getting flagged.

“These are the lousiest socioeconomic parts of the city,” he said.

They also happened to be the neighborhoods with the highest percentage of minority residents, according to census data, which would later complicate police strategy.

Retired Manchester Detective Bob Freitas observed that the center city has long had a number of large tenement buildings, a variety of low-income housing and absentee landlords. These factors, rather than the simple presence of high-density housing, were contributing to crime rates.

According to an analysis by the Granite State News Collaborative, these issues are localized today due to a long history of exclusionary zoning, a mapping process by local municipalities that restricts the development of certain building types — like large apartment buildings — to some parts of the city at the exclusion of others.

UNH Law Professor and Portsmouth Police Commissioner Buzz Scherr said it’s no accident the crime map mirrors the zoning.

“It’s hard to divorce the economic profile of those hotspots and what created that economic profile,” Scherr said, not speaking on behalf of the Portsmouth Police Commission. “It’s a conscious decision by those controlling zoning to create these kinds of areas where there’s a lot of people, and tending to be a lot of low income people, packed into a space.”

Hot spot policing

Over the years, police in the Queen City have done their best to proactively deter or catch criminals. And they’ve long known where the problem areas are.

Former Manchester Police Chief David Mara said the center city was always a known locus of crime in the Queen City, in addition to a few spots to the east and south of the epicenter and on the west side across the Granite Street bridge. 

“We have always had areas in the city that have been higher crime areas, but the focus isn’t just on those areas. It’s wherever the crime is being committed,” Mara said. 

An April 2020 report by Manchester police and scholars at Southern New Hampshire University, summarizes the work done by the Project Safe Neighborhoods (PSN) Initiative, which began in August 2018. The authors alo describe earlier work by the department to implement “hot spot policing.”

Criminology literature has found that a majority of crime tends to occur in only 3 to 6% of an urban space. Those regular problem areas on the map are called hot spots. 

The basic strategy is based on the “routine activities theory” from the late 1970s, which argues crime is primarily caused by the convergence of three things: a motivated offender, a suitable target and a lack of capable guardianship. Police seek to correct for the third item on the list by installing a police presence in the so-called hot spots, hoping that will be enough to deter crime.

In the wake of this theory and subsequent supporting studies, departments across the country have adopted this simple doctrine: More cops in the right place means less crime. Barter said this mindset continued with the use of data-driven hot spot policing, which told them where the right places were more reliably.

Manchester police had every reason to believe it would work. A 2008 study of policing in Lowell, Massachusetts concluded that those strategies resulted in statistically significant reductions in calls for service, robberies, non-domestic assaults and burglaries within Lowell’s hot spots.

Prior to Manchester’s data-driven efforts, the department partnered with community organizers and the health department to bring resources to disadvantaged neighborhoods. For decades it increased police presence in areas where criminal behavior appeared most concentrated. In its attempts to explain why some areas frequently required increased patrols, the department examined other issues related to unemployment, stress and substance use disorder, but one thing the department hasn't looked at in recent memory was the impact of historical zoning, according to Barter.

Whether police were responding to anecdotal reports of issues in an area or patterns identified by sophisticated software, they generally sent uniformed police patrols there in the hopes their presence would deter further crime. 

Measuring success beyond arrests

Mara, who retired in 2015, said it was always his goal as chief to focus more on individual offenders than whole neighborhoods, and to get the community involved in whatever the police department was doing. The department also participated in federally-funded social service efforts in the early 2000s called Weed and Seed, which targeted disadvantaged neighborhoods in the inner city with resources.

But before he was chief, Mara doesn’t recall many programs that targeted specific parts of the city as a matter of overarching strategy, though the notion of increasing police presence to deter crime was no less popular then. He remembers the department obtained Community Oriented Policing Services (COPS) grants in the 1990s, an initiative during the Clinton administration, which effectively resulted in departments hiring more police officers.

Today, outcomes are measured by crime statistics like a decrease in burglaries from year to year. But in the ‘80s and ‘90s, Mara said officers were more incentivized to make arrests and department metrics leaned on arrest figures to demonstrate the successful use of city resources. 

“Back then… patrol officers were judged by their production,” Mara said. 

But Manchester police began to move away from that model, Mara said, acknowledging that sometimes a significant focus on enforcement of the law can exacerbate social pathologies on aggregate and cause widening divisions between law enforcement and minority communities, which can be disproportionately targeted. Conversely, Mara said community policing emphasized building constructive relationships with residents and sometimes prevented an escalation in crime.

Making arrests is still an important function of law enforcement, he said, but officer performance should be defined and evaluated as much by how well they interact with the community in a positive way. 

A data-driven approach

Using data and computers was a seachange in policing, Barter said. He and like minded folks in other departments across the country have been pushing for this sort of thing for years. But leaders in staid police departments have often been resistant to listen to their more bookish counterparts.

“We call ourselves the nerd herd of cops,” Barter said. “There's a lot of stuff that we’ve done in policing [that]… has zero evidence base to it, can be harmful to communities, and we don’t want to do those. We want to stay away from those.”

In 2014, while Mara was still chief, Barter first proposed using crime statistics to create a hot spot policing program in Manchester. The program has gone through a number of changes over the years but was most consistently used between 2016 and 2018 in tandem with a special, custom-built algorithm that reviewed 12-weeks of crime data daily to predict where criminal activity was likely to occur. This form of regression analysis was called “predictive” policing. 

“Without going somewhat overboard, I agree that [Barter] is somewhat of a visionary,” Mara said. “I’ve watched with pride as what they are doing has evolved.”

The idea behind using software was to target police presence more strategically by basing it on some form of evidence. 

“We want to reduce crime in the neighborhoods that need it most,” Barter said.

Barter said the specific deterrence strategy the department took was based on the Koper Curve principle. The idea is based on an experiment in Minneapolis published in 1995, and establishes that a 15 -minute patrol presence in an area every two hours optimizes crime deterrence.

The predictive policing program that started in 2016 focused on three types of crimes: robberies, burglaries and thefts from motor vehicles. Barter said the department was dedicating more resources to these crimes because they were a bigger problem then, but they were also occurring at substantial enough numbers that the algorithm could make more reliable predictions based off of them.

At first, it wasn’t easy getting officers used to the new system. But eventually, it became part of the normal routine. 

“This was a paradigm shift in how we police,” Barter said.

Did it work?

There’s little question as to the accuracy of Manchester’s crime statistics and where crime is happening, but there is much debate over how effectively it was put to use.

For instance: did increasing police presence deter crime?

Former drug dealer Carl Connor, 40, who served 11 years in federal prison for a narcotics charge, said police presence never made a difference to him when he was committing crimes.

“I didn’t care if I saw five cops or zero,” Connor said.

In the end, it is unclear if the overall program was successful, because there is a lack of data linking it to outcomes.

Barter said he set up the predictive policing program without using internal control groups — parts of the city they could measure against — to gauge the impact of the program.

“At this time, I wasn’t very knowledgeable at running experiments. So we just kind of rolled out and did it,” he said.

When pressed to measure success, he compared Manchester’s burglary, robbery and theft from motor vehicle rates during the time of the program to national rates in similar-sized cities with populations of 100,000 to 249,000. Using this non-scientific comparison, there was some indication that Manchester’s property crime rates dropped at a faster rate than those at cities of comparable size, Barter said. But he acknowledges there are too many factors at play to determine why with any certainty.

Aside from that, Barter points to a number of anecdotal examples of hot spot policing working as it was intended, such as when the program predicted copper thefts from a building and police officers caught a man in the act due to a targeted patrol.

However, overall crime hasn’t gone down over time. In fact, calls for service have risen fairly consistently over the past decade until 2020, when they declined, likely due to pandemic lockdowns and people staying at home. Gun crimes have become an increasingly frequent occurrence, and increased 27 percent between 2015 and 2016, according to police. 

“I think it’s gotten worse,” Connor said of crime in the city. “My job was partly created because there are so many shootings.”

Connor now works as a street outreach worker for MY TURN in Manchester, a nonprofit that provides mentorship and other services to disadvantaged teens and young adults.

He is no stranger to the consequences of crime. Connor’s father died of a heroin overdose in a Manchester Street home, and his own 17-year-old son Jaden Connor was shot and killed while fleeing from an attempted robbery in 2020. 

When he was young, Connor said he had negative experiences with police. In one instance, he recalls when an officer pointed his firearm at another child because they were playing with a BB gun. 

Today, Connor said police are doing better at providing resources for mentorship programs at the Manchester Police Athletic League and MYTURN. But he doesn’t think hot spot policing has turned the tide. Connor works as an Uber driver in the city, and said he commonly witnesses people high on methamphetamine and other drugs.

The program evolves

The predictive policing program was phased out with the introduction of a new data storage system in Manchester in 2018. Since then, the use of crime stats and mapping data evolved, while the belief that hot spot policing should work in Manchester continued. 

The next iteration was in the 2018 to 2020 program known as Project Safe Neighborhood (PSN). The PSN initiative did away with the focus on robberies, burglaries and thefts from motor vehicles, and instead focused on gun crimes

It too ended with mixed results. According to Manchester Police’s final report on its latest hot spot program that ended in 2020, a lack of detailed field reports from patrol officers and partnering agencies left a lot of questions unanswered.

The report’s authors analyzed crime statistics in three hot spots that had the highest number of gun crimes, looking at the period spanning the 14 months before police interventions and 14 months during the PSN program.

“It is unclear how many community interactions occurred in each of the three hot spot areas and to what extent those interactions may have had deterrent effects or contributed to reporting biases,” the report states.

Still, the program appeared to have some positive effects. Though gun crimes increased city-wide by 3% during the program, two of the three hot spot areas saw reductions, according to the final report. Overall violent crime decreased by 2.39% city-wide. The PSN report also claimed the program had a deterrent effect predicted by program coordinators which resulted in reduced felony gun crime arrests. There were 15 such arrests in the 14 months prior to the program period, compared to 10 arrests during the initiative. 

Another goal of the program was to prioritize gun crime investigations, but the report noted the clearance rate actually got worse during the intervention period; from about 47% of cases cleared in the prior 14 months to about 38% during the program.

Criticisms

Not everyone believes data-driven hot spot policing is as groundbreaking as it may seem. UNH Law Professor Buzz Scherr said he thinks hot spot policing is the same sort of “back end” work that police usually do; reactively dedicating resources when they learn about a problem area.

“That kind of policing has been done forever. They’ve dressed it up with statistics, but it’s… what police have done for decades, without having algorithms and things like that,” said Scherr, who is also the Portsmouth Police Commissioner, but was not commenting in that role.

But sending police to an already economically depressed area with a lot of minorities is more likely to make the underlying problems, such as district and economic factors, worse, Scherr believes.

“It doesn’t improve the problem, it exacerbates it,” Scherr said. “It’s a complex problem that is really resistant to saying ‘oh we’ve flooded this area with cops and it’s reduced crime.’”

Scherr thinks police presence isn’t enough to make a dent in crime, and believes more resources should be spent on investigations.

The growing unpopularity of hot spot policing in the national discourse was one of the reasons Manchester Police decided to phase out their predictive analytics program, according to Barter. 

In 2016, the American Civil Liberties Union along with 16 other civil rights organizations came out against predictive policing in the United States, generally, pointing to examples of racial bias and a lack of transparency in some programs.

One of the common criticisms levied against hot spot policing is the tendency of some departments to use data points generated by police officers themselves, such as arrests, complaints filed by cops or gang surveillance reports.

“That’s where you get the feedback loops,” Barter said.

To avoid this, Manchester Police used data exclusively from calls for service from the general public. They also hired a third-party consultant to make their own algorithm to do the regression analysis, so they know exactly how it reached its conclusions.

Lessons learned

The legacy of the department’s previous hot spot policing programs live on in newer iterations that learned from the lessons of the past and are aimed at solving societal problems with a more holistic and cooperative approach with partner agencies and organizations.

Today, the department’s hot spot program’s focus remains on gun crimes. Hot spot data is used to target those areas where gun crimes occur, but instead of enhanced patrols, a more hybrid approach is being taken, with a new emphasis on investigating the small group of individual offenders, and with regular beat cops walking through the center city and other areas in the hopes they would have positive interactions with the residents there.

But ‘positive interactions’ are often in the eye of the beholder, especially when it comes to law enforcement.

Lascaze, who is Black, thinks there is value in finding ways to build trusting relationships between cops and residents. Yet he is against predictive policing strategies in general, because he said the approach has a history of disproportionately targeting minorities. He’s skeptical of the present-day iteration of Manchester’s hot spot policing program. 

“(Police) being seen in the same area every day is going to do one of two things with people,” he said. 

If they have a positive association with police, they will feel safe when they see them, Lascaze said. But there are many others, particularly minorities, who have had negative and traumatizing experiences with law enforcement whether it was in Manchester, another state, or another country. 

“It’s going to make them feel targeted, unsafe, paranoid,” Lascaze said.

Ophelia Burnett grew up in a high-density area on Orange Street, slightly north of the center city. Burnett, who is Black, said police presence sometimes created problems where none would exist, simply because of mutually heightened tensions and misread body language.

“The cops would come by, they’d ask us what’s going on,” Burnett said. And people would get nervous.

Barter said it’s his hope the department will earn the trust of communities of color over time. But not everyone is so optimistic.

“That is a really difficult complicating factor,” Scherr said. “What is the reaction when they see police presence? Is the reaction that people who would otherwise engage in those robberies and burglaries and breaking into cars? Are they less inclined to engage in that? Or is it more complicated? Does it build distrust in those neighborhoods with increased police presence?” 

The department is not blind to the effects of housing and poverty on certain neighborhoods today, according to Barter. Insofar as the socioeconomic issues of a neighborhood are to blame for the crime rates, Barter hopes new efforts with the health department and public works will focus on cleaning up neighborhoods and providing resources targeting health issues and young people. 

Increasing police presence is no longer seen as the silver bullet it was once hoped to be, according to Barter. But hot spot mapping can still help, he said.

“Essentially, the concept with the program was that [police] have a lot of data on violent crime. We leverage that data to help inform policing strategy, but it is also valuable to help inform non-law enforcement responses,” Barter said.

One of the ways they’re doing this is by working with nonprofit organizations like MY TURN in the hopes that outreach workers like Connor will be able to intervene to deescalate flare-ups of violence. The department funded a program called Project Connect with an initial grant of $33,000, and it provides the organization with useful intel on young violent offenders.

There are also ways the city can improve areas with a multi-pronged approach. A couple years ago, Barter said the department was drawn by hot spot data to an alleyway off Auburn Street where a lot of shootings and drug deals were taking place. It was a regular hangout for bad actors, he said. But police presence alone didn’t solve the problem. As soon as cops showed up, the people there would scurry and hide. 

Barter said the solution was more complex and specific: they had to team up with public works to install better lighting, remove a wooden guardrail that was used as a bench, clean up the tree brush and tow abandoned cars. Police arrested 11 people in the alley after an investigation in October 2020. After that, it was no longer a suitable spot for drug deals and violence.

“Since that day, there’s not been an issue in that alley,” Barter said.

These articles are being shared by partners in The Granite State News Collaborative as part of our race and equity project. For more information visit collaborativenh.org.