CHAPTER TWO

Safe Spaces

image

The Pruitt-Igoe housing project, a complex of thirty-three eleven-story buildings that opened in North St. Louis between 1954 and 1956, is perhaps the most famous failure in the history of American public housing.1 The initial design, by the highly regarded architect Minoru Yamasaki (who went on to design the World Trade Center), called for a mixed-rise set of towers connected by walking paths and “a river of trees,” but the Public Housing Administration forced a revision to bring down costs. The housing agency standardized the plan and substituted cheaper construction materials than those Yamasaki wanted. Ultimately, each of the thirty-three identical buildings featured an open first floor for community activity; a corridor with common rooms, laundry facilities, and a garbage room on several “anchor floors”; and elevators and stairwells shared by all of the roughly one hundred families residing in the tower. Families rushed into the new apartments, which were in great demand. In 1957, more than 90 percent of all units were occupied.2

A few architecture critics initially praised the high modernist project for its spatial efficiency and abundance of green space. But major problems soon emerged. In the 1960s, Oscar Newman, a young professor of architecture and urban planning from Washington University in St. Louis, initiated a study to determine what was wrong. Newman had read about the conditions at Pruitt-Igoe, but that hadn’t prepared him for what he saw in person. Vandals had destroyed the laundry and garbage facilities. Graffiti covered the common areas wherever he went. There was garbage all over the public areas, both inside and outside the buildings, and the sidewalks were coated with broken glass. Residents complained about the spike in prostitution, drug dealing, burglary, and violent crime. Families with children had grown fearful and began hunkering down in their private residences. “The corridors, lobbies, elevators, and stairs were dangerous places to walk,” Newman writes.3 “Women had to get together in groups to take their children to school and to go shopping.” People began moving out, and soon there was an exodus. By 1971, about half the units in the complex had been shuttered and the occupancy rate was 35 percent.4

Newman acknowledges that his first reaction was to blame the residents for ruining what could have been a beautiful, modern housing development. “Walking through Pruitt-Igoe in its heyday of pervasive crime and vandalism, one could only ask: What kind of people live here?” he says.5 That’s not surprising, since prevailing theories about the causes of crime have long focused on the characteristics of people who break the law. When criminologists examine the factors that lead to high levels of lawlessness, they typically consider “background” factors like a person’s race, gender, income, family situation, and education level, or, in some cases, their moral values and capacity for self-regulation. Initially, Newman did the same, and he observed what most government officials did: a community with a high concentration of poor, black, single mothers and their children. It appeared as if no one was able to exert control.

But as Newman spent more time at Pruitt-Igoe, he realized that the situation was more complex. Most residents, for instance, kept their own apartments “neat and well maintained—modestly furnished perhaps, but with great pride.”6 Moreover, Newman visited the small balconies shared by two families, and found them to be “pockets that were clean, safe, and well-tended.” He wanted to know what accounted for the differences between the project’s private and public spaces, and he began searching for answers.

The well-kept apartments and semiprivate landings offered Newman one clue; but what he observed in the housing project adjacent to Pruitt-Igoe taught him even more. “Across the street from Pruitt-Igoe was an older, smaller, row-house complex, Carr Square Village, occupied by an identical population,” he writes.7 “It had remained fully occupied and trouble-free throughout the construction, occupancy, and decline of Pruitt-Igoe.” People made good use of their semiprivate gardens and public areas, which meant that there were plenty of what Jane Jacobs, whose ideas greatly influenced Newman, called informal surveillance through “eyes on street.”8 Families felt safe and relatively comfortable, and with good reason: the crime level was three times lower than it was in Pruitt-Igoe. Newman was fascinated. “With social variables constant in the two developments,” he explains, the underlying reasons “that enabled one to survive while the other was destroyed” had to involve “physical differences,” not the characteristics of the residents.9

Newman compared the physical features of the two projects, and the differences were stark. In Carr Square Village, each row house contained just a few families, all of whom could identify one another as neighbors, if not friends. They shared a compact entryway and a semiprivate outdoor area, and it was relatively easy to establish standards for using and maintaining the space since only a few neighbors would be sharing it. In Pruitt-Igoe, however, the design and management of the buildings made it impossible for residents to regulate behavior in and around them. So many people shared the same public areas that no one person could manage or maintain them, and, unlike high-rise buildings in middle-class areas, there were no doormen or resident superintendents who were paid to do that job. What’s more, the population of the thirty-three-building complex was so large that it was “impossible to tell resident from intruder.” For Newman, the lesson was straightforward. The dire situation at Pruitt-Igoe was not due to the characteristics of the people who lived there, but to the project’s physical infrastructure. Building and landscape design “play an important role in reducing crime and in assisting residents in controlling behavior in their housing environments,” he concludes. “Residents maintained and controlled those areas that were clearly defined as their own,” but the larger, shared public spaces “evoked no feelings of identity or control [and] … made it impossible for even neighboring residents to develop an accord about acceptable behavior.”10

Newman included an account of the physical factors that made Pruitt-Igoe so much more dangerous than Carr Square Village in his 1972 report, Defensible Space.11 That same year, the government began to demolish the Pruitt-Igoe complex, a process that would go on until 1976, when the last doomed building came down. Newman’s theory of defensible space gained considerable influence among urban planners and criminologists and helped shape the design of new housing projects throughout the world. (Newman even became a character in the 2015 HBO miniseries Show Me a Hero.) Subsequent research has questioned whether all of Newman’s findings are generalizable. In some cities, we know, poor neighborhoods with high-rise public housing are safer than poor neighborhoods without it, and Newman failed to identify which conditions make some projects more successful than others. He was wrong to conclude that, in the case of crime and housing, “the apartment tower itself … is the real and final villain.”12 But Newman’s main point, that the built environment helps determine local crime levels, is widely accepted. In fact, the evidence for it is now stronger than ever.

Some of that evidence comes from a school of crime prevention that emerged around the same time as Newman’s defensible space theory. Crime Prevention Through Environmental Design (CPTED) begins with the insight that a person who is likely to commit a crime in a certain environment would never consider doing so in another. As C. Ray Jeffery, the criminologist who conceived CPTED, puts it, “There are no criminals, only environmental circumstances which result in criminal behavior. Given the proper environmental structure, anyone will be a criminal or a noncriminal.” It follows, then, that crime control measures are unlikely to work if they are designed to target individual offenders. Instead, crime is best managed “through the manipulation of the environment where crimes occur.”13

To this day, however, most policies that aim to reduce crime focus on punishing people rather than improving places. The president has called for a national “stop and frisk” police program; the attorney general wants more severe sentencing; advocates of “law and order” are resurgent. We invest little in housing and far less in safe sidewalks and neighborhood amenities like libraries, senior centers, and community gardens, which draw people into the public realm and put more eyes on the street. We spend even less to address criminal “hot spots”: specific places, such as empty lots, abandoned buildings, and liquor stores, that are known to foster illegal activities. Funds for improving community and street-level conditions in the places most likely to suffer from crime and violence are in short supply.

Government officials may have political reasons to crack down on potentially dangerous people rather than on demonstrably dangerous places, but the scientific grounds are shaky.

We have other more effective and less expensive ways to reduce crime today. And a growing body of scientific research shows that some of the best options involve investing in social infrastructure.

dot

In the nineteenth century, British researchers began studying the variation in crime rates between and within cities. While some of this work offered relatively simple accounts in which concentrated poverty led to higher crime, others pushed to another level, asking why there was so much variation in crime rates among poor neighborhoods. “Most of this work was descriptive and offered theories,” writes the University of Pennsylvania criminologist John MacDonald, “but did not attempt to provide guidance on how to curb crime.”14 He compares this tradition, unfavorably, with the work of British health scholars, most notably John Snow, whose research on cholera “noted the importance of the spatial environment in shaping human health and suggested the separation of sewers and drinking water wells to prevent water-borne diseases.” Reducing crime is more difficult than preventing cholera, but MacDonald, who’s done pioneering experimental research on how places influence crime rates, is one of many contemporary environmental criminologists with something new and significant to offer.

Social scientists have long played a major role in shaping crime policies. Consider the “broken windows” theory, which the Harvard political scientist James Q. Wilson and the Rutgers criminologist George Kelling introduced in the Atlantic in 1982. According to Wilson and Kelling, criminals perceive broken windows and other forms of neighborhood disorder as signals of weak social control and, in turn, as evidence that crimes committed there are unlikely to be checked. “Though it is not inevitable,” they argue, “it is more likely that here, rather than in places where people are confident they can regulate public behavior by informal controls, drugs will change hands, prostitutes will solicit, and cars will be stripped. That the drunks will be robbed by boys who do it as a lark, and the prostitutes’ customers will be robbed by men who do it purposefully and perhaps violently. That muggings will occur.”15

“Broken Windows” is not only one of the most cited articles in the history of criminology, it’s also one of the most influential works of public policy research, sometimes referred to as “the bible of policing” and “the blueprint for community policing.”16 Since the 1980s, cities throughout the world have used Wilson and Kelling’s ideas as motivation for “zero tolerance” policing, wherein officers carefully monitor petty crimes such as graffiti, loitering, public intoxication, even panhandling, and courts severely punish those convicted of committing them. “If you take care of the little things, then you can prevent a lot of the big things,” said the former Los Angeles and New York City police chief William J. Bratton, who used broken windows theory not only as a guide in both places but also in his global consulting work.17 In practice, this meant stopping, frisking, and arresting more people, particularly those who live in high-crime areas. It also meant a spike in reports that police were unfairly targeting minorities, particularly black men.

Despite some relatively recent experimental evidence supporting elements of the theory, broken windows always worked better as an idea than as a work of empirical science.18 As the Columbia University law professor Bernard Harcourt writes, “the famous broken windows theory has never been verified,” and “the existing social-scientific data suggest that the theory is probably not right.”19 The problems, which include the fact that perceptions of disorder generally have more to do with the racial composition of a neighborhood than with the amount of broken windows or graffiti, are numerous and well documented. A veritable A-list of renowned scholars, including the Harvard sociologist Robert Sampson, the University of Chicago sociologist Stephen Raudenbush, and the Berkeley criminologist Franklin Zimring, have identified flaws at all levels of the broken windows argument, and with the policies it inspired.20

For present purposes, though, I’m less interested in the validity of the broken windows theory than in the way it was framed and interpreted. The authors, Wilson and Kelling, encouraged policy makers to crack down on the petty crimes that lead to things like broken windows, which meant more aggressive street-level policing. Had they been more interested in the influence of social infrastructure, however, they might have taken another tack.

Consider the famous scenario in which the authors propose how spirals of disorder and decay get started. “A piece of property is abandoned, weeds grow up, a window is smashed,” they write. “Adults stop scolding rowdy children; the children, emboldened, become more rowdy. Families move out, unattached adults move in. Teenagers gather in front of the corner store. The merchant asks them to move; they refuse. Fights occur. Litter accumulates. People start drinking in front of the grocery; in time, an inebriate slumps to the sidewalk and is allowed to sleep it off. Pedestrians are approached by panhandlers.”21 From there it gets worse.

What’s curious, I think, is that the first two steps of this vicious cycle—“A piece of property is abandoned, weeds grow up”—have disappeared in the public debate about why some neighborhoods have such high crime rates. The third step—“a window is smashed”—inspired the article’s catchy title and took center stage. In academic circles, scholars had long been concerned about the security risks posed by abandoned properties and empty lots. But the popular and policy debates about Wilson and Kelling’s theory ignored the two problems at the root of their story and jumped straight to the criminal behavior. We get “Broken Windows,” not “Abandoned Property,” and a very different policy response ensues.

Imagine what might have happened if Wilson and Kelling had pushed their readers—many of whom were mayors and police chiefs—to think more carefully about the social infrastructure. What if vacant buildings and empty lots had gotten the attention that was showered on petty criminals instead? What if neighborhood crime prevention efforts focused on inspecting and remediating dangerous properties rather than stopping and frisking suspected people?

This is not a rhetorical question; it’s a puzzle that drove John MacDonald, the Penn criminologist, and his collaborator, the epidemiologist Charles Branas, to start one of the most exciting research experiments in contemporary social science.

dot

Branas, who’s now the chair of epidemiology at Columbia University, is a leading scholar of gun violence, a topic he got to know viscerally while working as a paramedic. Branas often found himself responding to calls for victims of gunshot wounds. On a typical day, more than ninety Americans die from gun violence, and two hundred more are injured.22 And although gun violence has decreased dramatically in recent decades, it remains far higher in the United States than in other affluent nations—the US gun homicide rate is about twenty-five times greater than the average among other high-income societies. It is stubbornly high in particular cities, some midsize (New Orleans, St. Louis, Buffalo) and some large (Detroit, Philadelphia, Chicago); and, within these cities, it’s especially high in very specific places—not just neighborhoods, but blocks. Branas and MacDonald wanted to understand why certain places were so dangerous.

It’s unusual for epidemiologists and criminologists to collaborate, but Branas and MacDonald met in the 2000s when they were working at the University of Pennsylvania, first in a seminar on gun violence at the medical school’s Trauma Center, and then, once the conversation got started, on the campus grounds. Both were frustrated by the science that linked crime to neighborhood disorder. “A lot of it, from ‘Broken Windows’ on, was just descriptive,” Branas told me. “You didn’t know exactly what counted as disorder. And it wasn’t actionable. Outside of policing, which was obviously complicated, there wasn’t much you could do about it.” One day, Branas read an article in Science based on experimental research done by Kees Keizer in the Netherlands.23 “These researchers somehow got permission to create disorder in certain neighborhoods. They actually added graffiti and litter to some places and not others, and they could measure the effects.” The paper showed a strong relationship between disorder and incivility—some, though by no means full, validation for the broken windows theory. But Branas was less interested in the findings than in the method. They knew that no institutional review board or community organization would ever let them add graffiti and litter to an American neighborhood, but perhaps they could design another experiment to test place-based crime reduction techniques.

While they were brainstorming, Branas was invited to discuss his research on guns and alcohol at a conference sponsored by the Federal Reserve Bank in Philadelphia. During his presentation, he briefly mentioned his interest in running an experiment on the physical factors related to gun violence. “When I finished, someone from the Pennsylvania Horticultural Society [PHS] approached me,” Branas recalled. They were convinced that vacant properties were driving up violent crime in poor neighborhoods. Philadelphia had passed an ordinance that required owners of abandoned buildings to install working doors and windows, but there were thousands of buildings out of compliance, and tens of thousands of empty lots that were violating the anti-blight ordinance too. The horticultural society had incredible data, and they offered to help.

Branas and MacDonald were excited about the idea. There is, after all, an established literature showing a relationship between abandoned properties and crime. In 1993, the criminologist William Spelman published a paper showing that, in Austin, Texas, “crime rates on blocks with open abandoned buildings were twice as high as rates on matched blocks without open buildings.” And, in 2005, the sociologist Lance Hannon showed that the number of abandoned houses in a given census tract was correlated with homicide levels in New York City’s high-poverty and extreme-poverty areas.24 Branas, who had taken the lead on the project, invited more scientists to join them: a health economist, a professor from the Department of Emergency Medicine, and a medical anthropologist. There was an enormous amount of data to collect and analyze.

The team’s first major research project involved assessing the impact of two natural experiments in Philadelphia.25 In one, the Penn group examined violent crime around 2,356 abandoned buildings that had been in violation of Philadelphia’s anti-blight ordinance. A set of 676 buildings had been remediated by the owners, which meant they had been “treated” with replacement doors and windows; the rest had not. Every month, for a three-year period between 2010 and 2013, the researchers compared violent crime levels around the treated buildings with violent crime levels around a randomly selected, geographically matched group of buildings that remained in disrepair.

The team did something similar in the second project, which compared violent crime around vacant lots. There were 49,690 “abandoned parcels of open land with no buildings on them” in Philadelphia, the researchers report. PHS had remediated 4,436 of them, which meant they had cleared trash and debris, graded the land, planted grass and trees to create a parklike setting, and installed low wooden post and rail fences with walk-in openings to facilitate recreational use and deter illegal dumping. Again, Branas and his colleagues compared the treated sites with a set of randomly selected, geographically matched properties. In this study, they measured crime annually, over a full decade, from 1999 to 2008.

On a warm and windy day in September, I visited Philadelphia to observe the sites in their experiments, including empty lots and abandoned houses that PHS had remediated and the untreated properties that they used as controls. Keith Green, a bald, large-set African American man with a salt-and-pepper beard, picked me up in his blue Ford pickup truck, and told me that we’d begin by driving to West Philadelphia, where PHS maintains 2.3 million square feet of vacant land. Green, who grew up in a part of Philadelphia that’s so gray they called it “the concrete city,” started working at PHS twenty-one years ago, first as an intern and then on community garden projects. “I never thought I’d be doing this for so long,” he told me, “but I found my niche when we started fixing up abandoned property. I remember one of the first jobs: The city asked us to clean up a two-block area in North Philadelphia where there was a flea infestation. We got there and it was like the entire area had turned into a jungle. Weeds, tall grass, messed-up trees. People were using it as a dumping ground. There were junk piles. Old cars. Broken bottles. Mattresses. Just a mess. We wound up treating 125 empty lots on four city blocks. I’m not kidding: 125! It was a horrible job, but when we finished you could tell that the neighborhood was going to be different. And people were so happy. I’d have kids running up to my truck yelling, ‘Mr. Keith! Mr. Keith! Can you come back tomorrow?’ They treated me like I was Mister Softee!” He laughed at the memory. “I kept thinking, ‘You know, this is something we have to do all over Philadelphia.’ And that’s basically what we did.”

Green drove slowly up Fortieth Street in West Philly. “You’re gonna want to keep your eyes open,” he instructed. “You won’t believe what you’re gonna see.” In fact, the area looked a lot like the depleted neighborhoods I’d studied in Chicago, like Englewood and North Lawndale, places where row houses and apartment buildings, some empty, some well kept, sat adjacent to large open lots that were thick with weeds, debris, and six-foot-high grass. “See that?” He pulled over at a corner lot with a low-lying wooden fence, two benches, trimmed trees, and a neatly cut lawn. “That’s one of our treated sites. You can tell because it’s so well maintained.” We got out, walked through the pocket park and over to a vacant house and large lot a few steps away. There, the grass had grown both high and wide, so that it now came through the sidewalk and out to the curb. “Now this, this is a disaster,” Green said. “It’s probably got an owner who wouldn’t let us work here, or someone we couldn’t track down. If you live here, you’ve got to deal with all the problems this attracts into the neighborhood: Pests. Insects. Garbage. Crime. And you know it’s gonna make it hard for new development to work here. People see that and they want to run.”

We crossed the narrow street to look at another property, and when we got there Loretta, an African American woman in her late twenties who was out for some exercise, was walking briskly in our direction. I paused and asked her if she lived here. “No,” she replied. “But I walk around this neighborhood all the time.”

“Have you noticed all the little parks with small fences?” I inquired.

“Not really.” She looked around, took them in. “They’re nice, though.”

“What about the abandoned lots with all the weeds and garbage?”

“Um, yeah,” Loretta answered, cracking a little smile. “Why do you think I’m walking on the other side of the street!” She paused for a beat, then looked over at the lot and started explaining. “Those places are scary. You don’t know what’s going on in that mess, who’s around. There’s a lot of places like that around here, and I just try to keep away.”

Loretta returned to her exercise and Green and I headed up the road again before turning onto Westminster Street. He pointed to a large remediated lot that residents had converted into a community park, with picnic tables and a small garden. “A guy who owns a store a few blocks away helped fix up this block,” Green explained. “He just wanted the neighborhood to look nice, to get more people out on the sidewalks and gardens. We see a lot of that. If we maintain things, residents go a little further, and put in what they like.”

We crossed over to a set of three row houses that had pocket parks, one large, one small, on either side. Micky, who had gray hair, sunglasses, a wooden cane, and a few missing teeth, was sitting on a picnic table and talking on a flip phone. He stood up and nodded when he saw me approach. I asked if he came to the park often. “I sure do,” he responded. “It’s a popular place. The Muslims had an event this past weekend. There was a health fair not long ago too. I’m seventy-three years old and I walk around here every day. This is a nice spot to stop.”

Green asked if the park made the neighborhood better. “Oh, you know it does,” Micky replied. He pointed to the front porch of the row house next door, where Joyce, in sandals and a white T-shirt, was relaxing on a rocking chair. “Ask her, she knows.”

Joyce was nodding. “I’ve been staying here ten, twelve years now. Those lots were bad when I first got here. Drugs and all that. Kids up to no good. People would let their dogs run all around them too, and oh, did it smell!” She grimaced and shivered a little from the memory. “But they fixed it up pretty soon after I got here. Put them tables in, big umbrellas too. Kids started coming around. We got the garden going. Before, everybody would avoid this block. It was ugly, and dangerous,’cause you didn’t know who was gonna jump out of those bushes. Now it’s a lot better. We’ve got the park, we’ve got shade. It’s a pretty good place to be.”

Green and his colleagues at the horticultural society suspected that fixing up the empty lots and buildings was improving Philadelphia’s poor neighborhoods, but they weren’t certain. Branas and MacDonald had a more specific hypothesis: they thought that remediation would reduce gun-related crime around them. “It’s not simply that they are signs of disorder,” Branas told me. “It’s that the places themselves create opportunities for gun violence; they take what might just be a poor neighborhood and make it poor and dangerous.” The reasons are straightforward. Abandoned houses are good places for people involved in criminal activities to hide when they are on the run. They’re also good places to store firearms. Untended empty lots are notoriously good places for drug dealing—in part because most law-abiding residents avoid them, and in part because dealers can hide their products in the weeds and tall grass if the police drive by.26 For communities, and for the police, they are hard places to monitor and control.

But compelling theories, as critics of broken windows know all too well, are often betrayed by evidence that discredits them. That’s why Branas was so surprised by the dramatic findings from their first experimental study on blighted environments and violence: There was a 39 percent reduction in gun violence in and around abandoned buildings that had been remediated. There was a smaller but still meaningful 5 percent reduction in gun violence in and around vacant lots that had been remediated as well. These are extraordinary numbers, at a level of impact one rarely sees in a social science experiment. But they weren’t the only ones that impressed Branas and his team. Equally powerful, he said, is that there was no evidence that the violence had simply shifted to nearby places: the declines were real. Moreover, these reductions lasted from one to nearly four years, making the benefit far more sustainable than those in other crime reduction programs.27 “Honestly,” he confessed, “it was a bigger effect than we’d expected to find.”

According to Branas, “the main goal of the two projects was to find a way of making inexpensive, place-based changes that can be scaled to entire cities and that are relatively straightforward to sustain.” He’d failed to do this earlier in his career, he acknowledged, when he was doing more conventional antiviolence research and advocacy, the kind that focused on the people who are most likely to commit crimes. “When I started at Penn, we had been working hard to reduce gun crime in Philadelphia. We had the interrupters, the social workers, the community leaders. Some of them were amazing, and we had some successes. But they were always short-lived. They lasted only as long as we could keep the people there in the hood. We spent a considerable amount of money on that project, and in the end it wound up helping only, I don’t know, about fifty kids, just the ones who were there at the time. We wanted to have a bigger impact, one that would continue even after we left.”

The Philadelphia studies suggest that place-based interventions are far more likely to succeed than people-based projects. “Tens of millions of vacant and abandoned properties exist in the United States,” write Branas and his team. Remediation programs “make structural improvements to the very context within which city residents are exposed on a daily basis.”28 They are simple, cheap, and easily reproducible, so they can be implemented on a large scale. What’s more, they impose few demands on local residents, and the programs appear to pay for themselves. “Simple treatments of abandoned buildings and vacant lots returned conservative estimates of between $5.00 and $26.00 in net benefits to taxpayers and between $79.00 and $333.00 to society at large, for every dollar invested,” their paper in the American Journal of Public Health reports.29 It’s not only more dangerous to leave the properties untended; it’s also more expensive.

That’s why, once Branas began publishing and presenting his findings on the impact of fixing up abandoned properties, cities and universities throughout the United States were eager to enroll in the experiment. “In the last few years we’ve had people here from so many cities,” Keith Green told me. “Detroit, Chicago, Trenton, and Seoul. When the guy from Chicago was here, he kept saying, ‘This is incredible! This is incredible! We need this in Chicago. I don’t know why we don’t do this there!’”

By 2016, the team had raised millions of dollars in federal grants and launched blight remediation projects in New Orleans, Louisiana; Newark and Camden, New Jersey; Flint, Michigan; and Youngstown, Ohio. Each experiment included local academic partners and, at Branas’s insistence, the frontline researchers were community residents who were trained and paid for their contributions. “We’re proud that we’ve been able to employ people in these neighborhoods,” Branas said. “But the bigger, more sustainable effect will come from fixing places.”

dot

Policing will always be a key component of crime reduction, and there’s no question that some policing strategies are better and more humane than others. But in recent years, small-scale, place-based policies in some of the world’s most dangerous cities—including Rio de Janeiro, São Paulo, and Johannesburg—have yielded substantial declines in violent crime. In some cases, they’ve improved the quality of everyday life too.

In São Paulo, for instance, the local government has supplemented policing programs with policies that reduce opportunities for crime in specific “hot spots,” while incentivizing young people to spend more time in safe places, particularly schools. Social scientists have long observed that the presence of bars and liquor stores tends to increase local levels of violence, especially in high-crime areas. History shows that prohibition is not a wise response to this problem, not only because it generates more crime but also because it fails to reduce overall consumption and often shifts consumption to more dangerous substances. But there’s good evidence that reducing the hours when bars and liquor stores operate helps. Between 2001 and 2004, for instance, sixteen of the thirty-nine São Paulo municipalities implemented a “dry law,” which forced bars and liquor stores to close between 11 p.m. and 6 a.m. When the Brazilian economist Ciro Biderman examined the impact on violent behavior, he discovered that the dry laws caused “a 10% reduction in homicides” and had similar effects “on battery and deaths by car accidents.”30 Although the restrictions were suspended after the crime rate began to drop, they showed that subtly tweaking the social environment can dramatically increase safety.

While one of São Paulo’s crime-fighting techniques involved reducing access to dangerous places, another aimed to increase access to a cherished safe spot that generates multiple benefits for children and communities: schools. The government had good reason to focus on the one place where young people have regular interactions with responsible adults. According to a World Bank report on crime prevention in Brazil: “In São Paulo, among those crimes for which the age of the suspected offender is known, between 20 and 25 percent of robberies, thefts and motor vehicle crimes are committed by individuals under the age of 18.”31 In 2003, São Paulo initiated the Family School policy, which opened 5,306 public schools on the weekends and provided extensive programming for children. Drawing on models from Mexico and Colombia, the government also started CCT Bolsa Família, a policy that offered modest but meaningful cash transfers to poor families, on the condition that their children complete secondary schools. Both programs, the World Bank claims, made important contributions to São Paulo’s great crime drop.32

Not all environmental designs to reduce crime produce widespread benefits. Consider the gated community, an architectural form that has emerged throughout the world to assuage anxieties about crime and violence, and which some critics see as an extreme expression of CPTED’s underlying principles. After all, gated communities facilitate local control of shared territory and establish clear lines between public and private space. They sponsor and, through signage, call attention to active surveillance by security guards and cameras. They often encourage the kind of informal social interaction among residents that allows for eyes on the street. They maintain all public areas, which sends a clear signal that residents monitor and care about the physical environment. Wherever possible, they reduce opportunities for crime.

In City of Walls, the Berkeley anthropologist and urban planning scholar Teresa Caldeira documents the fortification of urban space in São Paulo during the 1980s and 1990s. “In the last two decades, in cities as distinct as São Paulo, Los Angeles, Johannesburg, Buenos Aires, Budapest, Mexico City, and Miami, different social groups, especially from the upper classes, have used the fear of violence and crime to justify new techniques of exclusion and their withdrawal from traditional quarters of the cities,” she writes. Caldeira attributes the rise of gated communities and private security in Brazil not only to high crime but also to anxieties about the end of military governance and emerging democratic politics in an unequal society. She calls attention to the pervasive “talk of crime,” which “simplistically divides the world into good and evil and criminalizes certain social categories,” while also legitimating the use of gates and private security “to ensure isolation, enclosure, and distancing from those considered dangerous.”33 Caldeira grants that these systems protect those who can afford them, but worries, quite rightly, that they weaken democracy, deepen social divisions, and endanger as well as infuriate the people whom they exclude.

Urban critics raised similar objections to the rise of gated communities in postapartheid South Africa. Some, including the architect Karina Landman, blamed advocates of CPTED for not recognizing how their ideas could be used to justify new forms of social exclusion. In the late 1990s and early 2000s, Landman observed two distinct kinds of gated communities in South Africa: “Enclosed Neighborhoods,” where white residents of formerly open residential communities put up gates, fences, and booms to reduce and control access points, and “Security Villages,” which were larger places with mixed-use facilities, including office parks, retail outlets, and luxurious residential developments.34

Landman reports that both types of gated community successfully maintained low crime levels inside their protected environments. But the segregated social infrastructures they supported created countless other problems and directly undermined the urgent project of rebuilding democratic order in postapartheid South Africa. “Many people object to a restriction of access to public roads.35 It is also prohibited by the South African Constitution,” she observes. Moreover, “the question of access often leads to resentment,” particularly among city dwellers who had previously been able to visit or travel through the enclosed neighborhoods without scrutiny. Landman does not explicitly condemn gated communities, but there’s no question about the message written between the lines. She’s essentially saying that in South Africa, fear of crime is doing the work that fear of blacks did during apartheid: justifying the creation of an unequal society sustained by segregated social infrastructure, in which one group can take extralegal measures to protect itself and the others can only fume.

Intriguingly, there is at least one well-documented case of an impoverished, crime-ridden community taking control of their own territory and using unauthorized security devices to monitor violence perpetrated against them by police. But their low status and lack of political power made the project short-lived. In 2001, while the University of Texas anthropologist João Costa Vargas was conducting fieldwork in Rio de Janeiro, a neighborhood association in the city’s second largest favela, Jacarezinho, installed large gates and cameras at several access points. On a popular political website, leaders of the association demanded “their right” to “more and better social programs focusing on health, education, and job training, and public transportation. In short, they demanded full citizenship,” Vargas writes.36 Gating themselves into a protected community was a political tactic that local leaders used to publicize both their social problems and their struggles with violently abusive police.

The Brazilian media was fascinated by the story. But, as Vargas reports, the stigma around Jacarezinho and the low status of its poor, dark-skinned community made journalists skeptical of the neighborhood association’s claims that police, not residents, were the main perpetrators of violence. The news media consistently denied the legitimacy of the residents’ demands and, in turn, of the gates. Officials insisted that drug dealers were using gates and cameras to create a zone of lawlessness where police couldn’t operate. Whereas the white neighborhoods in South Africa had managed to defend their neighborhood enclaves, the poor, mostly black slum dwellers of Rio did not have that privilege. Their gates came down quickly, and the militarized police force reasserted control.

dot

Battles to control the social infrastructure are usually not so dramatic. In the United States and Europe, disputes over safety and the social infrastructure are more likely to unfold slowly and quietly, and they’re often centered in gentrifying neighborhoods where new commercial and residential developments change street-level conditions in ways that can either protect or—if they lead to displacement or heightened racial discrimination—imperil local residents.37 Gentrification is one of the most controversial issues in affluent urban areas, so it’s no surprise that there’s a burgeoning literature on how it affects crime rates. Unfortunately, there is not yet a clear answer. Some studies show an increase, likely because new commercial development creates new targets and opportunities for crime, while others show a decrease, often attributed to more eyes on the street. The variation suggests that gentrification plays out differently depending on the local context, and also that groups are likely to be affected by gentrification in divergent ways.

In all neighborhoods, though, commercial establishments are important parts of the social infrastructure. As Jane Jacobs and Ray Oldenburg famously argued, grocery stores, diners, cafés, bookstores, and barbershops draw people out of their homes and into the streets and sidewalks, where they create cultural vitality and contribute to the passive surveillance of shared public space. When I did research in Chicago, I discovered that poor neighborhoods with active retail corridors were surprisingly resilient during the devastating heat wave, because people who lived in them could easily go out and get air-conditioning or support from neighbors. Remarkably, the social activity generated by street-level commerce protected both the majority and the minority populations in Chicago neighborhoods: it wasn’t the products that made residents safer, it was the people who came out to shop and socialize. Yet Chicagoans who lived in depopulated and commercially depleted neighborhoods didn’t have the same opportunities for casual interaction, and as a result they were more likely to remain in their broiling homes.38

But, as the story of São Paulo’s successful dry laws illustrates, some commercial outlets are more likely to foster crime than to inhibit it. Bars and liquor stores are obvious examples; banks, currency exchanges, and automatic teller machines can have similarly deleterious effects on vulnerable neighborhoods, for the simple reason that they create new targets for robbery and assault. In the 1960s, the urban planning scholar Shlomo Angel examined patterns of illicit behavior in Oakland, which, like many cities at the time, was experiencing a worrisome spike in street crime. Angel found that retail corridors were hot spots, particularly after hours, when most consumers were at home and informal surveillance was weak. Although he shared Jane Jacobs’s enthusiasm for the protective value of eyes on the street, Angel warned that commercial outlets were the wrong way to attract them, and urged cities to restrict their development.39

Recently, though, social scientists have taken a closer look at these patterns, and they’ve discovered that most retail outlets and commercial corridors are more protective than Angel realized.40 This is especially true in gentrifying neighborhoods, where new retail businesses, such as coffee shops and restaurants, symbolize an “invasion” of white, affluent residents and often spark public debates about displacement. In one intriguing paper, “More Coffee, Less Crime?,” a team of sociologists led by the Yale professor Andrew Papachristos examined the relationship between gentrification and crime rates, using the annual number of neighborhood coffee shops that opened as a proxy measure for new local retail development. The article is agnostic on the general question of whether gentrification is good or bad for residents and cities more broadly. But it makes a compelling case that retail outlets like coffee shops help keep a neighborhood safe.

After analyzing data from the Chicago Police Department, the US Census, and the Chicago business directories, Papachristos and his team make some striking observations. First, even after controlling for other factors, an increase in the number of neighborhood coffee shops is associated with a decrease in the number of murders. They report that this pattern holds regardless of whether the neighborhood is primarily white, black, or Latino. Second, and not surprisingly, not every group benefits equally. Street robberies, for instance, tend to go down when new coffee shops open in gentrifying neighborhoods that are primarily white or Latino. But they tend to go up in gentrifying neighborhoods that are primarily black, quite likely for the reasons that Angel proposed: they have fewer commercial outlets to attract a steady stream of consumers, and consequently they lack enough informal surveillance to deter crime.41

Third, and importantly, business owners are less likely to open coffee shops in gentrifying black neighborhoods than in those that are white or Latino. When they do, it’s often a sign that developers are planning a large-scale transformation of the area, one that will likely result in significant displacement of established African American residents. In 2017, this threat helped fuel a small-scale uprising in Denver’s Five Points neighborhood, one of countless gentrifying urban areas around the United States. The flame came from Keith Herbert, the proprietor of a local Denver coffee chain, ink!, who allowed his staff to post signs saying “Nothing says gentrification like being able to order a cortado” and “Happily gentrifying the neighborhood since 2014” outside the company’s new shop in Five Points. Herbert thought the signs would be funny, but residents and community organizations concerned about their fate found them offensive. Some responded by vandalizing the establishment; others organized protests against both ink! and the gentrification process it represented. “I am embarrassed to say that I did not fully appreciate the very real and troubling issue of gentrification, and I want to sincerely apologize to those who understand firsthand the hardship and cultural consequences that gentrification has caused,” said Herbert in a Facebook post after the protests began. But this hardly satisfied his critics, who have continued pressing for their neighborhood to preserve places where everyone, not only affluent young professionals, feels at home.42

Residents of the Cabrini-Green housing projects in Chicago organized a similar protest movement in the late 1990s, when the city announced that it would soon demolish, privatize, and revitalize the public housing stock, and civic groups accused officials of pushing poor African Americans off what had become valuable urban real estate. As protesters marched in front of City Hall, developers built a shopping center with a Starbucks across the street from Cabrini. The coffee shop, Papachristos and his team write, became a driver and a symbol of neighborhood transition, one that would “ensure this area will soon experience rapid gentrification.” The housing project did experience a decline in homicide, but it remained far more dangerous than the areas around it. “Here, we might see a ‘positive’ effect of gentrification in the form of long-term neighborhood crime reduction,” Papachristos and his co-authors write, “but at the severe expense of the displacement of Cabrini residents.”43 For this reason, they are adamant that although local retail outlets may well make neighborhoods safer, gentrification imposes formidable social costs to impoverished and vulnerable people, and is by no means the ideal way to reduce crime.

dot

As it happens, research conducted in another public housing complex in Chicago points to a more desirable way to reduce crime with social infrastructure. In a fascinating series of natural experiments, the landscape architect William Sullivan and the environmental scientist Frances Kuo, both professors at the University of Illinois, discovered the extraordinary power of vegetation to reduce crime in a high-poverty residential development.44

Neither Sullivan nor Kuo knew what they would find when they began their two-year study of crime patterns across ninety-eight apartment buildings in the Ida B. Wells project on Chicago’s South Side. They were advocates of greening urban neighborhoods. But it was the 1990s, and at the time urban planners who had studied places like Pruitt-Igoe were convinced that trees and grass in poor neighborhoods created opportunities for criminal behavior, because the green canopy made spaces beneath hard to monitor and control. Some older research supported this view, but Kuo and Sullivan noted that several recent studies reported lower levels of fear, less aggressive and violent behavior, and more civility in green areas. Their project, the first to use police crime data to measure the impact of green ecologies on public housing, would help establish a better answer.

Fortunately, the Ida B. Wells project was an ideal place to conduct a natural experiment. The residents of the buildings were demographically similar: nearly all were African American and poor enough to qualify for public housing, and the buildings contained a mix of family types. Kuo and Sullivan measured the amount of tree and grass cover outside the ninety-eight buildings, sorted them into categories, and analyzed which areas had higher or lower levels of crime. The findings were unambiguous: the greener the immediate surroundings of a building, the lower the rate of total crime. This pattern held for both violent crimes and property crimes, and it remained even after controlling for the number of apartments per building, building height, and vacancy rates.45

In Ida B. Wells and another massive housing project on the South Side, the Robert Taylor Homes, Kuo and Sullivan probed to learn more about how green space helped reduce crime and violence. Part of the answer was predictable: when the green spaces were well maintained, residents used them often, and this meant more passive surveillance from eyes on the street as well as greater feelings of ownership and control.46 But they also discovered surprises. Through interviews and observations by public housing residents whom they trained to do research on their team, they found that people whose buildings were surrounded by vegetation felt less aggression and mental fatigue than those who lived amid concrete. Although the levels of aggression public housing residents reported were significantly higher than those in the general US population, people who lived in a green environment said they were involved in less aggression and violence against their partner and less aggression against their child than those whose environment was primarily gray.

The research by Sullivan and Kuo made a major impact on the way designers thought about public housing. Unfortunately, by the 1990s there was little funding available to build new public housing projects, and in some American metropolitan areas political leaders condemned those that existed as failures. Soon after Sullivan and Kuo released their research findings, Chicago and the federal government announced the “Plan for Transformation,” which would demolish seventeen thousand public housing apartments throughout the city, including every building in the Ida B. Wells and Robert Taylor developments. There was a severe shortage of affordable housing in Chicago, and tens of thousands of families were stuck on multiyear waiting lists for public units. No matter. As in St. Louis during the 1960s and 1970s, local and national political officials believed that Chicago’s hulking public housing projects were beyond repair.

The shortage of affordable housing is a national crisis today, and violent crime is once again on the rise in some cities. Public investment in these matters is inevitable, because the problems they generate are too difficult for citizens to endure and too serious for political officials to ignore. For decades, building prisons for the poor has been our main crime reduction policy, and the social costs have been as great as the economic expense. If we want a better, more equitable, and sustainable solution for the challenges facing our cities and suburbs, we’d be better off building social infrastructure instead.