PUBLIC HOUSING BREEDS CRIME
Fritz Umbach and Alexander Gerould
“Everyone knows how quickly . . . housing projects in big cities turn into dangerous, demoralized slums.”
—Howard Husock, 20031
Howard Husock, Harvard University’s director of public policy case studies, was so confident that public housing’s dangerous nature was common knowledge that he neglected even to provide citations for the sweeping generalization that opens his 2003 book, America’s Trillion-Dollar Housing Mistake: The Failure of American Housing Policy. But a decade and half of robust scholarship has made clear that what “everyone knows” about crime and public housing is almost certainly inaccurate. At times, some public housing has undoubtedly proved to be more dangerous than other shelter for the poor. But that has not been the case consistently, and when it has been true, the origins of the disorder turn out to be very different from what is commonly believed.
Popular and academic discussions of public housing crime have focused on three broad questions. Does something about public housing generate more crime than private housing? Does public housing make surrounding neighborhoods more dangerous? And does dispersing public housing residents simply displace disorder? Answering these questions requires understanding that although criminal offenses may have precise legal definitions, the collection, analysis, and presentation of crime statistics is a social undertaking. Making sense of the numbers often means asking why, how, and for whom they were assembled.
Sorting out something approaching the truth about crime and public housing developments (PHDs) has been neither easy for analysts nor, as it turns out, much in demand. Empirical precision has often seemed beside the point to an American public that has woven crime, PHDs, and racial minorities into an “unholy trinity.” As criminologist Garth Davies observed in his 2006 book on the topic, “that public housing projects are rife with serious crime is . . . accepted as fact despite a discernible paucity of evidence.”2 So powerfully, indeed, has popular belief fused crime with PHDs that even the nationwide demolition of high-rise projects failed to quiet long-standing fears. In 2008, Atlantic magazine writer Hanna Rosin picked up where Husock left off, laying blame for spiking suburban violence on displaced residents of razed public housing. That Rosin neither offered much in the way of proof nor considered alternative explanations did not prevent her “American Murder Mystery” from being included in Best American Crime Reporting 2009.3
Figure 3.1 “Eliminate crime in the slums through housing.” New York: Federal Art Project, 1936. Courtesy of the Library of Congress.
Portrayals of public housing complexes as “criminal paradises” stretch back to the late 1950s, when it became obvious that reformers’ utopian visions of redeeming blighted urban tenements through modern architecture and state management had failed to deliver (Figure 3.1).4 But this unsavory image emerged more from journalistic exposés than academic research, which was itself often sparse and incomplete. Only recently have scholars taken up the topic of crime and public housing with anything approaching statistical thoroughness.
Until the late 1990s, however, efforts to grasp the extent of public housing’s influence on crime generally stumbled over two research challenges, reflecting both the limits of crime data and shortsighted understandings of “the projects.” First, scholars hoping to establish a statistical relationship between PHDs and crime found there were no reliable numbers to crunch. Most police departments simply did not keep separate blotters for public housing (except in New York City, as discussed below). As late as 1998, federal Housing and Urban Development (HUD) researcher Harold Holzman was able to observe matter-of-factly that “valid statistics on the level of crime in public housing do not exist.” Second, scholars tended to see the distinction between public housing and other forms of housing as one primarily of physical form. As a result, these studies attempted to correlate PHDs’ designs with their crime rates or, more simplistically, to compare crime rates inside and outside public housing.5 Such approaches, however, overlooked that PHDs are not simply brick-and-mortar constructions but also sociopolitical ones. As such, they came to be shunted into the most troubled neighborhoods; sheltered a demographic that victimization research identified as most likely to fall prey to crime; and, as communities, often suffered from more than their fair share of municipal neglect.6 Ignoring such contexts meant that research often left open two questions: First, was the disorder criminologists found in PHDs actually generated there, or did the developments merely (and unwillingly) play host to criminal activity from elsewhere? And, second, if public housing did, in fact, elevate the risk of violence and other crimes, what share of the blame lay with their physical design, what share with their social makeup, and what share with managerial incompetence?
The importance of questions regarding PHDs’ neighborhoods becomes more apparent in light of public housing’s location nationally. Some 60 percent of units are to be found in the country’s big cities. Establishing what impact the densely populated, highly segregated, and overwhelmingly poor neighborhoods surrounding most projects have on crime requires pinpointing where crime occurs. But most police departments tally up crime using large geographic units, such as precincts or districts. Better answers to some questions had to wait until the early 2000s, when affordable geographic information systems (GIS) technology became available, allowing crime mapping on small budgets.7 By that time, however, the nation’s public housing was deep into a life stage that scholars have called “housing of last resort” or “welfare housing,” when many PHDs had descended into “severe distress”—to use policymakers’ peculiar euphemism.8 The sometimes-bleak conclusions about crime drawn from recent GIS data are often snapshots of particular complexes at arguably their most difficult points. And so, what follows proceeds chronologically, to highlight the importance of changing social, political, and technological contexts.
Oscar Newman and the Making of a High-Rise Myth
How is it that, decades before the arrival of real statistical knowledge about crime and public housing, the “projects” had already become synonymous with violence, not just among urbanites but also among professional urbanists? The first widely-read academic account of crime and public housing—and one that unintentionally fortified notions of such communities as deviant—appeared in Lee Rainwater’s 1970 Behind Ghetto Walls: Black Families in a Federal Slum, a study of St. Louis’s soon-to-be-razed Pruitt-Igoe. Rainwater numbered among a small army of mostly white, progressive social scientists who poured into poor, minority neighborhoods in the 1960s and observed residents’ “adaptations” to poverty and racism. These idealists aimed to counter conservative talk of a “culture of poverty” by presenting ghetto norms as a coping strategy in response to deprivation rather than an inherited way of life that created it. But the practical impact of such work at the time, explained historian Robin Kelley, was that it “reinforced monolithic interpretations of black urban culture.” Quickly forgotten, for example, was that Rainwater had researched not just any public housing development but perhaps the nation’s most troubled. The National Institute of Mental Health (NIMH) had, in fact, hired him to evaluate last-ditch efforts to halt the uniquely massive decline of the 57-acre complex. But the crime-plagued Pruitt-Igoe of Rainwater’s portrayal became a stand-in for all PHDs and a parable of failed antipoverty programs.9
Within weeks of the publication of Behind Ghetto Walls, St. Louis declared it would tear down Pruitt-Igoe. And just months after that too-oft-commented-upon demolition, architect Oscar Newman published Defensible Space: Crime Prevention through Urban Design, sealing public housing’s criminal reputation.10 After Newman, the high-rise towers of many cities’ PHDs would loom as large in the public’s mental landscape as they did in the urban landscape. Newman’s work also marked the first sustained effort to use crime statistics to explore public housing’s influence on crime. That Newman had numbers at all reflected the unusual fact that between 1952 and 1995, New York had not one but three police departments—one for its streets, one for its subways, and one for its public housing—and each one reported crime data separately.11 But the existence of such separate statistics for Newman to use was not the same thing as a guarantee that he would do so with skill.
Newman spun for his readers, as he phrased it, “a tale of two projects.” In Brooklyn, he had found the ideal setting for a natural experiment: two projects across the street from each other and practically identical in all regards. They housed a similar number of residents at roughly the same density and were under the same management. All that differed was their physical design—and crime rates. The low-slung Brownsville Houses experienced so few crimes that mothers left their doors open so as to keep an eye on children playing in hallways. But at the high-rise Van Dyke Houses, mothers kept their youngsters behind locked doors, fearful of the chaos that, Newman held, often erupted in the project’s halls and stairwells. From this contrast, Newman drew lessons about physical environment and crime that seemed to him commonsensical. The tower-in-the-park design of many PHDs created public areas in and around the buildings where neither tenants nor management exerted much control, inviting stranger-on-stranger crimes in spaces with weak community surveillance. “High Rise = High Crime” read the headline to the New York Times article covering Newman’s assault on an earlier generation’s doctrinaire modernism, complete with a photograph of the young, rakishly bearded architect. Newman’s thinking—which emphasized the role of informal community controls on crime—was more complex than the headline’s distillation, but Newman did little to prevent such generic characterizations.12 Ultimately, his take on public housing was decidedly that of an architect: people, policies, and history receded into the background, leaving only physical structures in view. As Newman concluded, “It is the apartment tower itself which is the real and final villain.”13
Defensible Space reflected not simply Newman’s professional training but also his institutional funding. In a sign of the changing research economy of the 1970s, Newman ignored the large-scale social forces—from rising costs to vanishing jobs—that were troubling public housing nationwide and turned instead to small-bore design tweaks focused on crime. As architectural historian Joy Knoblauch has noted of these years, “funding sources shifted from large federal programs to promote the health, education, and welfare of the population to franchise-state efforts to prevent crime, punish offenders, and manage the population through incentive structures.”14 President Lyndon Johnson’s sensitivity to the political potency of the public’s growing crime fears led to the Law Enforcement Assistance Administration (LEAA) and opened for the first time the spigot of federal dollars to local agencies for crime initiatives. In Defensible Space and other works, Newman—lacking a permanent institutional home—followed the new money, creating putative knowledge about public housing’s criminogenic nature and occasionally peddling Band-Aids for the problem. In contrast, a call for a renewed and expensive War on Poverty would have won Newman, as a crime-policy entrepreneur, few grants. But federal funding of social science research into crime in general—and in public housing in particular—only deepened the public’s exposure to both notions. And the simplicity of Defensible Space’s environmental determinism guaranteed that the book’s assertions of public housing’s crime-breeding qualities reached large audiences. Defensible Space ended up outselling all other architecture books that decade. But for the many urban public housing authorities whose brick-and-glass empires consisted largely of high-rises, Newman’s choice of villain was particularly damaging.
Although Newman labored to give his conclusions the aura of slide-rule precision, Defensible Space suffered from a simple but profound defect. Newman’s evidence pointed in one direction, but his claims pointed in another. The Brownsville and Van Dyke Houses were not as comparable as Newman suggested. The high-rise Van Dyke Houses were not simply higher, they also sheltered significantly larger families. And, apparently, this aspect of public housing life mattered immensely. The regression tables tucked away in Defensible Space’s final pages quietly contradict the book’s conclusions by demonstrating that crime in the complexes was more closely correlated with family size (a rough proxy for child-to-adult ratios) than building height. Less adult supervision can elevate crime rates, as historian D. Bradford Hunt has documented using evidence from Chicago’s public housing data.15 Had Newman’s been an “academic study,” observed Knoblauch, his “methods of proof through data would have been inadequate and he would have received little attention.” But Defensible Space was a hybrid of journalistic narrative and social science analysis, taking advantage of the former’s lower methodological bar while simultaneously benefitting from the latter’s greater authority.16
Defensible Space’s tale of architecture as destiny had the additional advantage of sidestepping thorny national debates about poverty’s causes and society’s responsibility. If bad design were the real cause for disorder, then chicken-or-egg debates over ghetto norms and state solutions were irrelevant. Instead, cash-strapped cities should simply “design out crime” by hiring better architects. This perspective however neglected the increasingly troubled tenancy of many projects. The combination of loosened resident screening, growing white flight, and vanishing blue-collar jobs meant that by the late 1960s public housing sheltered a population far more desperate than it had a decade before. But Newman’s popular writings—ignoring policy, poverty, and history—reinforced the notion that crime was a permanent fixture of public housing’s iconic high-rises. Newman had access to statistics demonstrating that New York’s public housing developments historically had lower rates of crime than did their surrounding neighborhoods, but he clung to the towers’ design as his tale’s culprit.17 This insistence had real consequences. Newman’s subsequent consultancy with the New York City Housing Authority (NYCHA) led to expensive and largely ineffective structural design changes to individual projects. Although the growing poverty of public housing’s residents was perhaps beyond their control, policymakers and administrators ignored it in favor of an environmental determinism that distracted from other potentially effective crime control strategies—strategies grounded in policies that public housing authorities might have been able to control, at least in tight real estate markets. Recent scholarship on New York suggests that tighter screening policies, stronger eviction practices, and community policing might have made a meaningful difference. When even a small number of bad apples could duck eviction (at least until legal changes in the 1990s), it had an outsized impact on project disorder.18
Residents played such bit parts in Newman’s drama of high-rise pathology that he neglected even to notify them when his firm— employing Defensible Space’s theories—imposed major changes on their homes.19 It is worth noting, however, that the author’s sidestepping of human factors in thinking about project crime was an elite dance. Significant numbers of PHD residents nationwide consistently reported in surveys (and often confirmed through their own activism) that they believed it was public housing’s changing tenancy that had made project life more dangerous. And few reported believing questions of design mattered at all.20
Newman’s verdict on high-rise towers looks even less reliable in light of more recent research. Aiming to tease out the relative impact of physical and social factors in PHD crime, Tamara Dumanovsky used the same data as Newman—NYCHA tenant demographics and Housing Police crime reports—but subjected them to the careful statistical modeling the architect had avoided. Reviewing more than a decade of data (1985–1996) from the nation’s largest public housing system, Dumanovsky concluded in her 1999 study that the height of a complex’s buildings had a very modest impact on violent crime. The poverty of a complex’s tenants was by far the best predictor of how much violence occurred there.21 Similarly, in 1996, Holzman sought to answer the same question nationwide. His large, stratified sampling found crime correlated more closely with the number of buildings and residents in a development than it did with building height.22 The irony, then, of Newman’s legacy is that New York—the one city that might have yielded a rich statistical understanding of public housing crime—actually ended up giving rise to perhaps the most pervasive and pernicious myth about it. Newman himself would eventually bring more nuance to his arguments, but the damage to public housing’s image had been done.23
Newman’s work remains popular in contemporary criminology—albeit more because of his theoretical contributions than his empirical ones. Much of Defensible Space can be seen as a direct precursor to an alphabet soup of acronyms in environmental criminology and police science: SCP (situational crime prevention), CPTED (crime prevention through environmental design) and POP (problem-oriented policing). Likewise, Newman’s attention to a community’s ability to exert informal social control is compatible with theories about neighborhood conditions and criminal activity that are popular across the political spectrum, from broken windows (favored by criminologists on the right) to collective efficacy (favored by those on the left).24
HUD Surveys and Unexpected Safety from Violence
At the time of its publication, Defensible Space captured the imagination not just of the public but also of public housing officials, and so launched a decade of research into PHD crime focused largely on physical design and funded mostly by HUD. Understandably, HUD officials hoped to rescue the most chaotic projects. Research money flowed to measure crime in the most troubled complexes in order to guide federal (and ultimately failed) rehabilitation efforts.25 The unintended result, however, was that HUD paid for and distributed a picture of public housing lawlessness drawn not from representative projects but instead from distressed outliers—repeating the impact of Rainwater’s Pruitt-Igoe study a decade earlier.
Even more misleading, because most cities lacked New York City’s separate PHD crime statistics, HUD’s researchers resorted instead to a substitute: victimization surveys. Such surveys were, in fact, the new criminological tool of the 1970s and an effort to compensate for underreporting of crime.26 But surveying public housing residents (or a sample of them) only enabled researchers to compare crime rates in a particular project and its host city or the nation as a whole. Local officials typically saw to it that public housing projects were built in cities’ most disadvantaged and crime-ridden neighborhoods. Did higher PHD crime rates reveal that something about public housing increased crime, or did those rates simply announce the not particularly surprising news that projects were located in more dangerous neighborhoods?
Surveys, in short, did not make it possible to consider public housing disorder within its larger “ecology”—criminologists’ term for the constellation of social, physical, and spatial factors surrounding a research site. And so what passed in the 1970s for broad knowledge about public housing crime was really a collective portrait of some of the nation’s worst projects presented without the necessary context. And that picture was certainly not pretty. The risk, for example, of getting robbed in Washington, D.C.’s Capper Dwellings was twice that of the Capitol generally and five times the national average. Matters were even worse in Baltimore. Residents of the Murphy Homes, for instance, experienced three times as many robberies as did other poor residents of the city, and sixteen times as many as Americans generally. And burglars targeted the Murphy Homes four and a half times more frequently than they did Baltimore’s private residences.27 But even within this sketch of crime in some of America’s most difficult projects, there were important details that belied public housing’s reputation and likely would have surprised readers of Rainwater and Newman.
Consider Baltimore’s Murphy Homes. Sliced off from the city by a highway and encircled, thanks to urban renewal, by weed-filled lots during the 1970s, this modernist stew of low-rises and fourteen-story towers photogenically fit public housing’s lurid image. But in 1976, when HUD researchers surveyed the residents, they discovered that the Murphy Homes were not, in fact, a particularly violent place for the poor to live as compared to Baltimore as a whole. Residents were actually no more likely to be assaulted than were other poor citizens. Nor were the Murphy Homes unusual. Tenants of DC’s Capper Dwellings, for example, were actually somewhat less likely to be assaulted than the poor living in private housing. More remarkably, given the public housing of popular imagination, Capper Dwellings residents were twice as safe from assault than were poor Americans generally. Indeed, this pattern of PHD residents suffering from higher property crime rates but benefiting from equivalent or even lower violent crime rates (including robbery) than those experienced by the host city’s poor held true across nearly all of HUD’s research sites (and, in New York City, where HUD had not dispatched a team, both violent and property crime rates were lower in public housing). But no best-selling author trumpeted these unexpected findings, which emerged from a more rigorous statistical methodology than had underwritten Defensible Space. The HUD researchers themselves, acknowledging that their discoveries did not exactly square with conventional wisdom, insisted that what really mattered was public housing residents feared crime more than did other Americans.28
Curiously, this pattern of greater property crime but equal or lower violent crime was reversed in the 1980s and 1990s, when public housing residents started to experience more physical violence than property loss. Researchers in those decades often concluded that higher rates of violent crimes than property crimes merely reflected fixed economic and spatial facts of project life. Public housing residents, they asserted, simply possessed less that might tempt burglars, and superblock architecture, lacking retail stores and parked cars, occasioned fewer opportunities for theft. Such conclusions seemingly forgot the findings of the 1970s. Moreover, scholars associated with the “routine activity” school of criminology have recently challenged these economic and spatial interpretations of public housing’s asymmetrical property and violent crime rates (see below).29
Scarce Funding, Conflicting Findings, and Methodological Challenges
Between the absolute failure of HUD’s rehabilitation efforts in the 1970s and President Ronald Reagan’s disinterest in urban affairs, federal funds for public housing crime research essentially vanished in the 1980s. As urban housing specialist Langley Keyes noted of the priority shift, any “serious investigation of crime reduction strategies was over. Research was out.”30 Academics tried to fill the void, but shoestring budgets meant their efforts lacked HUD’s multicity scope. However, once freed from the necessity to provide practical guidance for troubled projects, scholars were able to choose more representative research sites. In 1982, for example, John Farley compared crime rates for ten St. Louis PHDs with their locations’ expected crime rate “based on five predictors: distance from downtown, percentage black, percentage of population consisting of children under 18, median rent, and percentage of households with female head.”31 Farley, like other researchers outside of New York, did not have separate crime statistics for public housing and so had to impute them by using city blocks that were primarily filled with public housing. To Farley’s surprise, St. Louis’s projects experienced crime rates that were generally no higher than the expected crime rate for their neighborhood and sometimes significantly lower, leading him to wonder in his article’s title, “Has Public Housing Gotten a Bum Rap?”
While Farley was interpreting research from St. Louis, Dennis Roncek was examining statistics from Cleveland. Roncek attempted to measure “spillover” effects of crime from Cleveland’s eighteen public housing complexes into neighboring city blocks from 1971 to 1977, asking, in essence, Did public housing generate more crime than it could consume locally, and so endanger nearby neighborhoods? The question had particular import in the wake of growing resistance to public housing construction nationwide. With GIS technology still on the horizon, the use of city blocks—the smallest geographic unit for which reliable data could be tabulated—made sense as a starting point for finding an answer to such a difficult question. Roncek found that there were no statistically significant PHD spillover effects for property crime. Violent crime did spill over, but the effect shrank once the socioeconomic conditions of the surrounding blocks were considered. In fact, after Roncek controlled for such factors, even the presence of a one-thousand-unit project within a mile probably added no more than one-tenth of one violent crime annually. Roncek did find, in contrast to Farley, that blocks made up of public housing had higher rates of violent and property crimes than did nearby blocks without public housing.32
Roncek had hit upon an important methodological issue in assessing PHD crime—an issue that was often ignored by studies before his and has bedeviled similar research since. Statistical arguments frequently hinge on a study’s use of a “control group.” Ideally, the control group is just like the population being studied, except they have not been exposed to the experiment’s conditions. By comparing outcomes between the control groups and “treatment groups,” one can tease out the impact of a particular factor, from the efficacy of cancer treatments to the effects of welfare payments. Of course, the real world rarely gives social scientists the gift of experimental conditions, and they have learned to make do with control groups that share many, if not all, demographic variables with the treatment group, while adjusting for the rest. Public housing, however, may offer a case so extreme as to render it impervious to researchers’ algorithmic maneuvering.
Appropriate control groups for public housing may be impossible to identify because public housing is not randomly distributed across space—rather it is concentrated in areas that often suffer from disorder and deterioration unrelated to whatever effects public housing may bring. The reasons are many. Municipal officials, eager to preserve or extend patterns of segregation, diverted public housing construction to some of urban America’s most downtrodden districts. Likewise, wealthier neighborhoods were often able to resist such incursions. At the same time, it was typically only on the cities’ fraying edges, or dying industrial pockets, that even well-meaning planners found the large swaths of real estate necessary for public housing’s huge towers. Nor were public housing residents themselves randomly selected; whether because of deeper pockets or lighter skin, those with more options increasingly took them. And so the nation’s most vulnerable people were often housed in its most vulnerable places, frequently isolated from even the declining employment and social services offered by the deindustrializing city.33 According to criminologist Garth Davies, “The physical and social features of public housing [neighborhoods] differ so from those found in typical residential neighborhoods that contrasting areas with and without public housing . . . may be entirely inappropriate.”34
It is not obvious how all of the jigsaw pieces of evidence from the 1970s might fit together to form a coherent picture of PHD crime. It is no surprise, however, that studies conducted at different times and places generated findings that varied in their details: crime fluctuated significantly across the decade, and PHD management practices varied greatly from city to city.35 But like Sherlock Holmes’s dog that did not bark in the night, what the researchers in the 1970s and early 1980s did not find may be the real story. Investigators did not discover that public housing was predictably a more violent place for the poor than was private housing. Nor did they find that public housing complexes particularly menaced surrounding neighborhoods. But the last research on PHD crime during this period appeared in 1982, and the next round of studies would not be published for more than a decade. The intervening years were hard ones for public housing. Deindustrializing urban economies created fewer opportunities for residents, and a flourishing new (and violent) informal economy centered on drugs emerged. Some projects faced both physical deterioration and depopulation as housing authorities, ravaged by funding cuts, sought to rid themselves of their most troubled buildings by withholding maintenance. Such neglect proved especially hospitable to crime—particularly narcotics trafficking, which brought money. With the money came guns and a violent scramble for markets.36 When researchers did turn their attention again to public housing, the accumulated consequences of these changes made the projects look very different than they once had.
Federal Funding Returns and a Shifting Crime Mix
Federal funding for public housing crime research returned in the late 1980s amid the highly publicized arrival of crack cocaine and its devastating impact on urban communities.37 For good reasons and bad, drugs occupied center stage in this era’s PHD crime studies. In what became a creation myth of sorts for the new wave of research, Jack Kemp, HUD secretary during the elder George Bush’s administration, reportedly witnessed drug deals firsthand while touring Baltimore’s public housing projects in 1989. So “shocked” was Kemp that he ordered the nation’s public housing directors “to report to him on the volume of drug trafficking occurring in the developments under their jurisdiction.”38 HUD money flowed to researchers to document drug dealing in public housing. And discover it they did—in spades.
There was just one catch. Drug deals are not like other crime—at least from the perspectives of sellers and buyers. The business is, generally, consensual and threatens no one’s rights when money and product change hands. Moreover, everyone in the game has a keen interest in concealing their activities. In short, a drug sale frequently requires a cop’s presence for it to exist as an offense “known to the police”—unlike, say, homicide, which is nearly always reported. And so researchers hoping to measure drug crime often have only the decidedly imperfect proxy index of drug arrests. Various studies in this period reported that public housing residents—from New Orleans to Los Angeles—were arrested for drug offenses far in excess of their share of the population. Had these studies uncovered extraordinarily elevated drug use in public housing?39 Or had they just documented the consequence of more police in public housing? More arrests could indicate more cops just as easily as they revealed more drugs. Few doubted—and journalistic and ethnographic research confirmed—that trafficking had taken hold in public housing in a major fashion and often upended tenants’ lives.40 But whether or not drug distribution and use was more of a problem in PHDs than would have been true in a poor neighborhood without public housing was not empirically knowable. At a time when the Public Housing Drug Elimination Program was providing hundreds of millions in grants to cash-strapped housing authorities to combat drugs, few wanted to question the accuracy of figures upon which such money depended. Indeed, a disproportionate slice of antidrug money went to the projects. In 1992, the $1.5 billion federal dollars for all drug-related activities (enforcement, education, treatment, and prevention) amounted to an expenditure of about $7 per citizen, but totaled more than $47 per public housing resident with the additional investments.41 As numerous scholars have pointed out, narcotics is one of the few places where the federal government, and politicians eager for votes, can inject themselves into America’s otherwise highly local criminal justice system.42 And so research, and the production of knowledge, followed federal drug money.
Moreover, some of the drug research suggested that crime in general was higher in public housing. One 1994 study commissioned by the National Institute of Justice (NIJ) compared violent and property offenses in PHDs in Los Angeles, Phoenix, and Washington, D.C., with nearby neighborhoods that were not home to projects. But as the researchers conceded, these neighborhoods were merely “comparison” areas rather than the “control groups” that statisticians might insist upon. Although the paired neighborhoods suffered from problems such as poverty and unemployment higher than citywide rates, such problems were “less grave in these nearby neighborhoods than in public housing.” While the NIJ researchers did not control for these differences, their findings revealed levels of violent crime in LA and DC’s public housing that were twice that of neighboring areas—a disparity so large that it probably would not have melted away entirely even after proper statistical controls.43 The NIJ’s raw figures for Phoenix, however, recorded that public housing complexes there suffered from just 20 percent more violent crimes than their “comparison” neighborhoods. This suggests that had the authors used multiple regressions, the difference between PHDs and nearby areas in Phoenix might well have been quite modest. (Such studies also make clear that public housing crime was far from monolithic; there were certainly high crime projects, but there were also low crime projects.)
Studies, however, continued to point in contradictory directions. Casting different methodological nets into different public housing developments, criminologists brought up different findings. The importance of controlling for socioeconomic conditions and historical forces, for example, was made clear in a study of Washington, DC., and Cleveland conducted by the Urban Institute the same year as the NIJ’s study of Los Angeles, Phoenix, and DC. The researchers found that after considering rates of physical decay, social disorganization, and economic deprivation in comparison neighborhood and PHDs, public housing did not consistently elevate risks of violent crime.44
Despite the various studies’ contradictory claims, it is still possible to divine broad trends. If the range of findings from the 1970s and early 1980s suggested public housing was either less violent or no more violent than similarly situated private housing, the findings from the 1990s suggested public housing was either about as or more dangerous. Moreover, it was not merely the level of crime that was changing in PHDs, it was the mix of crimes as well. For all of the methodological nitpicking that the NIJ’s 1994 DC-Phoenix-LA study invites, the researchers were the first to observe the early signs of changes in project life that would become a near constant in later research. In contrast to what researchers had found in PHDs over the previous two decades, the NIJ team discovered that DC and LA’s PHDs had lower rates of property crimes than either the city at large or their immediate neighborhoods—even as they suffered from more violent crime. (In Phoenix, public housing suffered from property crime at a rate four times that of the city at large, a figure more in keeping with HUD’s 1970s research.)45
The Current Research Paradigm
Out of the inconsistencies and ambiguities of research, two significant steps forward came in the late 1990s, compliments of HUD money and changing technology. In 1994, HUD—eager to evaluate the impact of the hundreds of millions spent on the Public Housing Drug Elimination Program—contracted with the Research Triangle Institute (RTI) to “retrofit” the National Crime Victimization Survey for use in public housing. Public housing, HUD believed, posed distinct problems for such surveys. In response, the RTI researchers designed a methodology adapted to public housing’s distinctive world. RTI produced a particularly powerful toolkit; not surprisingly, it was too expensive for use by individual public housing authorities. But before HUD decided that they had bought more than they wanted, the survey allowed Holzman to conclude (as noted earlier) that, contrary to Oscar Newman’s assertions, public housing crime correlated far more closely with the number of residents in a complex than with building height.
While surveys proved to be a methodological cul-de-sac, the development of affordable GIS technology in the mid-1990s (combined with significant HUD money) did lead to both an explosion of PHD crime data and greater confidence in the findings. The emerging technology has also helped answer long-standing questions about the causal role public housing residents play—as both perpetrators and victims—in crime. Scholars have wondered, for example, whether public housing itself elevates crime, or if it simply stacks together individuals whose everyday lives, by dint of poverty and segregation, leave them more vulnerable to victimization regardless of where they live. The new GIS research suggests that the latter is probably more accurate than the former. HUD researchers, using crime data from two cities (identified for anonymity as City X and City Y), discovered that black women living in public housing suffered aggravated assaults more frequently than did black women living in private housing. Even for criminologists who traffic in such risk differentials, the heightened dangers confronting public housing residents were startling. After controlling for a host of demographic factors, black women living in City X’s public housing were thirty percent more likely to fall prey to an aggravated assault than were those who did not; worse yet, black women living in City Y’s public housing suffered assaults twice as often as did those in private housing. And while those numbers alone would have been significant and troubling, the real surprise in the study’s findings—made possible by GIS—was not in the assaults’ frequency but in their location. Indeed, when considered alongside other research, the geography of assaults in these two cities may point to new and important ways to think about PHD crime.46
In City X, seven out of ten of women living in PHDs who were assaulted suffered their ordeal on project grounds. But in City Y’s public housing, where assault rates on females were even higher, the grounds of the projects themselves actually proved to “present little risk of serious personal violence for residents and visitors.” Indeed, only 8 percent of assaults on City Y’s female PHDs residents happened on project grounds; the vast majority took place somewhere else.47 In short, although City Y’s public housing was home to women who were particularly likely to suffer an aggravated assault, that violence rarely transpired in a project. In contrast, despite the fact that City X’s PHDs also concentrated women who were more likely to be victims, when such women were assaulted, it happened in the PHDs themselves. The relative size of the projects in cities X and Y, the researchers speculated, may explain these differences. Their thinking suggests that the small scale of City Y’s public housing complexes—with many fewer units than City X’s developments—allowed the project communities to subject residents and visitors alike to greater scrutiny, discouraging potential attacks in ways that City X’s larger PHDs could not. Regardless of the deterrent value of particular PHD designs, however, the high assault rates for female residents in cities X and Y, coupled with the very different locations of those assaults, underscore that projects often concentrate a population that is already at particular risk of victimization outside of whatever ill effects public housing itself might contribute —whether such putative dangers spring from flawed architecture, incompetent management, or political neglect.
Researchers employing the same GIS technology also found that by the late 1990s, violent crime was significantly higher in public housing complexes than in their immediate neighborhoods. In a three-city analysis, Holzman concluded that (1) PHDs more dangerous than areas within 300 meters of projects, and (2) even those 300-meter buffer zones were, in turn, more dangerous than the city at large. Violent crime, it seemed, clustered in and around public housing—confirming long-standing stereotypes. Property crimes in public housing, however, told a different story. Holzman’s research confirmed the curious mix of crimes identified by both the NIJ and Dumanovsky: property crime was lower in public housing than in adjacent neighborhoods and, perhaps even more surprising, was often lower than in the host city itself. Some analysts have concluded that property crime in PHDs is lower simply because public housing residents make unattractive targets for thieves. Such thinking has a certain logic: having less to steal lowers one’s odds of being stolen from. Holzman, however, proposes an alternative explanation, arguing that criminology’s “routine activity theory” helps us to see this mix of higher violent and lower property crime rates in public housing as two sides of the same coin. Moreover, combining Holzman’s hypothesis with the historical shifts in public housing crime data may help shed significant light on the changing experience of public housing tenants during the past half century.
Holzman points out that PHD residents are more homebound than the typical city dweller. And that difference, Holzman argues, helps shape project crime. The more frequent presence of PHDs residents in their apartments makes them, on the one hand, “capable guardians” of their possessions (deterring theft); on the other hand, it means that they are more likely to be assaulted there (increasing violent crime rates). As Holzman notes, only a third of PHD households have wage income, while the remaining survive on public assistance, social security, pensions, and (at times) informal or illicit markets. Nearly half of all PHD households have children, and 30 percent are headed by the elderly. Collectively, then, higher rates of unemployment, infirmity, retirement, and child-rearing mean that comparatively more PHD residents are home at any given hour. And so, Holzman argues, “there is no shortage of potential guardians in public housing developments to discourage property crime.” 48
There is, however, the coin’s other side. These “capable guardians” also face a greater chance of violent assault. Statistically speaking, to be a young African American woman in a low-income, urban household is to live with a greater risk of violence at the hands of a husband or boyfriend. And this demographic group disproportionately calls public housing home. According to Holzman, the same capable guardian “who, by her very presence, protects her home against thieves” is vulnerable “to physical assault by persons known to her” in her apartment.49 In short, PHDs now gather a comparatively homebound population both more vulnerable to criminal violence and less likely to experience property crime.
Holzman’s hypothesis goes a long way toward explaining the somewhat unexpected shift in public housing’s crime mix over the past four decades. Recall that researchers in the 1970s found that public housing residents were more at risk of property crime rather than violent crime. But that pattern of offenses had reversed itself by the time criminologists finally returned to PHDs in the 1990s. As researchers at the NIJ discovered, by the middle of that decade, PHD residents faced more violent crime than property crime. Changing tenancy within public housing may well have caused this seismic shift in the mix of project crimes. As late as the 1970s, public housing nationally still contained a good many working adults whose jobs took them away from their apartments and whose paychecks allowed some off-project amusements. Those absences exposed homes to theft and placed residents in a variety of alternative environments where they might or might not suffer personal violence. But as wage earners increasingly abandoned public housing, their places were taken by those whose sources of income and affordable pleasures did not tug them from their project. Daily life unfolded closer to project grounds than had been true for earlier generations—cutting property crimes but also increasing PHD violence. And so the transformation of public housing’s crime mix may well reflect the larger transformation of public housing’s tenancy.
Does Public Housing Make Neighborhoods Dangerous?
Researchers have also turned the klieg light of GIS onto perhaps the longest-standing fear about public housing—that projects make neighborhoods dangerous. Such worries take several forms. Does public housing radiate violence outward? Or does it draw violence inward from surrounding areas?50 As with other research questions, scholars studying different cities have produced findings that point in multiple directions. Davies, taking advantage of New York’s uniquely rich PHD crime statistics, argues for what he calls “two-way diffusion” effects. With few exceptions, he observed that crime in Gotham’s boroughs clusters in “hot spots.” PHDs, however, seem to foster such hot spots. Hot spots appear near PHDs at twice their expected rate, and “extreme hot spots” three times so. But crime seeps in both directions: poor neighborhood conditions increase the number of assaults inside public housing, while drug markets centered in the developments push up homicide rates in surrounding areas.51
Research from other cities suggests that how, how much, and where project crime spills over into surrounding neighborhoods depends on local conditions and history. Sociologists Thomas McNulty and Steven Holloway assert that the impact of Atlanta’s public housing on neighborhood crime rates is inextricably linked to the city’s history of segregation. One consequence of that near-apartheid experience, they argue, has been that public housing is a “feature of the institutional fabric of Black Atlanta, from which White Atlanta is socially and physically insulated.” And so when Atlanta’s projects rub up against white neighborhoods, that proximity “has little effect on violent crime rates.” But in nonwhite neighborhoods, nearness to public housing “is closely associated with violent crime.” Segregation’s legacy of social isolation shapes public housing’s criminogenic impact across space.52 Researchers in Los Angeles have found similar patterns. Elizabeth Griffiths and George Tita discovered that although LA’s public housing is a disturbingly lethal place, that deadly violence sticks to where residents spend their time—on project grounds—and seldom spills out. Griffiths and Tita argue that residents’ tightly circumscribed worlds help explain this geography of crime. The homicide statistics tell the story: murderers who do not live in public housing are three times more likely to take the life of someone living in a different neighborhood than are killers living in public housing. The researchers conclude that PHD offenders strike “close to home; they are not responsible for escalating violence in the larger community.”53
Public Housing Crime in Historical Perspective
Many of these recent studies would seem to confirm stereotypes that have followed public housing residents for a half century. And in a sense, they do. But in other ways, the research can mislead. While the findings are genuinely disquieting, this recent scholarship is best understood within a larger historical context. These studies document public housing long after the program had decayed, in many places, into a shelter of last resort. Consequently, the research does a better job of capturing the accumulated costs of decades of neglect than it does in establishing an immutable relationship between public housing and crime. Moreover, because these studies do not always address how some projects sunk as low as they have, the research risks implying that some innate quality of public housing (or its residents) creates disorder. The problem stems partially from research conditions. In the absence of historical crime data, criminologists have tended to compare public and private housing areas at a single (and recent) point in time. Longitudinal studies of public housing violence have been rare. And so, in contrast to much criminological scholarship, where crime rates rise and fall (and structural causes can be pinpointed), public housing crime in the literature often just is.
New York City, however, provides an exception. There, the Housing Authority Police Department (1952–1995) kept separate, and reasonably accurate, crime statistics. These fortuitous circumstances allowed a charting of crime in NYCHA—the nation’s largest public housing program—across time.54 NYCHA’s violent crime figures reveal that through the mid-1970s, New York’s public housing was, on average, 60 percent safer than the city at large. Although violent crime rates rose in both NYCHA and the city during those decades, NYCHA’s started lower and grew more slowly. Between 1967 and 1977, violent crime rates in NYCHA complexes tracked but did not intensify movements in New York’s violent crime rate. Indeed, what might be dubbed the “safety dividend” of living in public housing grew in these years. In 1967, NYCHA’s complexes were 42 percent safer than New York’s streets; in 1974 they were 78 percent safer. Little wonder that NYCHA’s projects enjoyed a reputation among the poor at the time as a refuge from Gotham’s escalating violence. After 1977, however, the trend underwent a dizzying reversal, as every bump in the city’s crime rates was accompanied by a greater jump in NYCHA’s. When, for example, New York’s violent crime rate floated up 13 percent between 1986 and 1988, NYCHA’s leaped 40 percent. By the late 1980s, NYCHA’s developments were more dangerous than their surrounding neighborhoods and the city at large (and remain so as of this writing). Changing tenant demographics, shifting policing and management practices, declining funding, rising informal markets, and the arrival of crack cocaine all help explain these patterns. But it is clear from NYCHA’s experience that, contrary to popular belief, public housing can provide safe and stable neighborhoods—and did so for decades.55
Nor is NYCHA’s history unique in the way it challenges the narratives of criminality that encrust public housing and, increasingly, shape public policy. No doubt, such images have suited some big-city projects in recent years. But as housing historian Edward Goetz points out, “the overwhelming majority” of public housing developments are not marked by criminality and “provide a better alternative than private-sector housing in poor neighborhoods.”56 Even in larger cities with high crime rates, the connection to public housing is not always clear. Sociologist Zaire Dinzey-Flores, for example, has observed that during the 1990s, police sectors in San Juan, Puerto Rico, containing public housing enjoyed lower crime rates than those without. According to Dinzey-Flores, these numbers undercut politicians’ rationale for that city’s highly publicized practice of militarized, predawn housing project sweeps as part of Puerto Rico’s policy of mano dura contra el crimen (“strike hard against crime”).57
Does Dismantling Public Housing Simply Displace Crime?
Just as criminologists started using GIS to answer timeworn questions about crime and public housing, the frequent target of this new statistical firepower—the large, inner-city project—was vanishing from view. Between 1990 and 2010, housing authorities sold or demolished more than 220,000 such units. As Goetz asked in a 2011 article title, “Where Have all the Towers Gone?”58 In a reversal of public housing policy, nearly seven billion federal dollars have been spent on what is arguably the nation’s most ambitious urban renewal scheme—HOPE VI. Aiming to “deconcentrate” poverty, the program tears down PHDs in varying states of distress—and occupied exclusively by the poor—and erects in their stead mixed-income developments with fewer impoverished residents. HOPE VI’s environmental premise was as simple as its policy promise was grand. Once the towers disappeared, so too would the ills of the urban poor. Those former tenants excluded by the new developments’ much higher rents or newly exacting screening standards, would (in theory) be given housing vouchers that would (again, in theory) allow them to fend for themselves on the private market, where (theoretically) individual landlords would shelter them better and at a lower cost than the state had done.59
Public housing’s hulking, inner-city towers came down first, but soon even smaller cities with low-rise developments got in on the act. No doubt, HOPE VI’s many critics are correct in fingering real estate interests hungry for profits and neoliberal politicians eager to remake the urban order as driving forces behind the program’s wide embrace. But criminological thinking also played its role. Support for HOPE VI often came from a circle of increasingly pessimistic housing specialists. Many had been chastened by Chicago’s failure in the 1990s to turn around its lawless complexes, even after spending hundreds of millions of dollars on elaborate community crime prevention efforts and aggressive policing. What this group of scholars came to share, observed Susan Popkin, was the “sense that a new, radical approach was the only way to improve the life chances for residents.” Likewise, as Goetz documents, even in midsized cities like Tucson, Arizona, crime rates correlated closely with how frequently wrecking balls tore into a city’s public housing. HOPE VI’s results have been mixed at best, but the near extinction of the “projects”—those “towers in the park,” so long and widely maligned—has changed the criminological research agenda.60
Two new questions—one posed by housing scholars and one forced upon them—have emerged. First, if to live in public housing is now frequently to live in a “bad” neighborhood that is truly injurious to children and families, would moving to better neighborhoods make for better lives (including decreased exposure to and participation in crime)? Answering this question was the explicit goal of the Moving to Opportunity experiment (MTO), a social experiment of unprecedented size and scope: $80 million spent, more than five thousand participants followed, and a decade’s worth of longitudinal data collected from five cities. The second question—did HOPE VI relocatees raise crime rates when they moved to new their communities?—was the focus of a smaller but well-designed study.
MTO aimed to answer the first question through a study conducted under conditions approaching “experimental”—that elusive grail of social scientists. In doing so, MTO researchers hoped to avoid a methodological predicament that vexed previous studies of the court-ordered Gautreaux Project, which had assisted tenants of Chicago’s hypersegregated projects move to suburbia. Although the life circumstances of Gautreaux participants seemed to improve in the suburbs, evaluations of the program could not answer a crucial policy-relevant question: If the PHD residents who opted to move out differed in meaningful ways—more ambitious, say—from those who stayed behind, might comparing the two groups tell us more about the power of individual traits than it did about the impact of one environment or the other? To avoid this problem, MTO used a lottery to randomly assign families (almost all female-headed and minority) to different research groups; each group would receive different types of help in moving, or not, from high poverty neighborhoods to lower poverty ones. The use of random assignments, the researchers believed, would generate comparable groups and so uncover the potential effects of neighborhood on how well families fared.61
MTO’s results were more mixed than anyone had expected—and no more so than in the impact of environment on former public housing residents’ experience of crime. That the program’s benefits on that score were so ambiguous has a particular poignancy, since four out of five MTO participants identified escaping the lawlessness of their PHD as their overriding goal in moving to a new neighborhood. MTO’s mixed findings spring from a puzzling gender split revealed by the experiment. Moving to lower poverty neighborhoods benefited adolescent girls in regard to a host of crime and risk-related aspects of their lives, but the move seemed only to make matters worse for boys. By leaving public housing for less poor neighborhoods, girls experienced a “profound drop in sexual pressures, predation, and related types of risk.” Such gains in safety and sense of security had profound benefits for the mental health of relocating girls. Their levels of depression dropped by a fifth—a reduction researchers noted was “comparable to that achieved by some of the most successful drug treatments.” But for boys, researchers concluded, “the picture is a sad one.” Compared to those who remained in public housing, moving to a lower poverty neighborhood did not reduce their chance of being arrested for violent crime—and actually increased property crimes arrest rates. The researchers considered a number of explanations: increased police surveillance in the new neighborhoods, more (and more expensive) goods available for thieving, and differing parental gender expectations. The researchers tentatively speculated that complex interactions between parenting and neighborhood best explained these gender differences (and rejected increased police attention as a likely causal factor).62 But given other studies reviewed here, the boys’ increased arrest rates may simply reflect the fact that their new neighborhoods lacked the “capable guardians” of PHDs’ comparatively more homebound population.
The second question was propelled not by academics but by Hanna Rosin’s widely read and provocative Atlantic article, discussed briefly at the beginning of this chapter. Rosin painted a vivid picture of HOPE VI as a “grand anti-poverty experiment” that was backfiring. In Rosin’s telling, former project tenants with a housing voucher in one hand and a Saturday Night Special in the other were “migrating” to the suburbs, bringing a crime spike in their wake. Despite holes in the article’s arguments, its popularity obliged criminologists to explore this second question more explicitly. As displaced public housing residents moved to private market housing, did their problems—including violent crime—accompany them?
The scholarly response to Rosin’s article was certain and swift. Within months, Shelterforce, a prominent affordable housing journal, published a rebuttal by Peter Dreier and Xavier de Souza Briggs and endorsed by two dozen housing experts. Dreier and Briggs asserted that Rosin had simply gotten her facts wrong. Few of the “migratory” poor arriving in suburbs did so with government help; the MTO experiment maligned in the article actually played no role in Memphis (the focus of her piece); and most former PHD residents moved to places that were not only near their original homes but also “already on the decline, with rising crime rates.” HOPE VI could hardly have brought violence to once-bucolic suburbs. Rosin, likewise, failed to consider alternative explanations for suburban crime, leaving readers “inappropriately confident in her ghetto-migration hypothesis.”63
In particular, Rosin overlooked that HOPE VI was only one of several forces remaking suburbia. Many of the poor did leave the worst ghettos in the 1990s, but it was market forces—gentrifying urban neighborhoods and robust job markets—that largely provided the push-and-pull dynamics. Government intervention played, at most, a small role. Additionally, most of the poor relocated to “first-ring suburbs”—older neighborhoods circling cities’ ragged edges—that were often just barely better off than the areas they left. Indeed, the share of the poor living in suburbs quadrupled between 1970 and 2000. HOPE VI relocatees were a small rivulet in a larger demographic flow. Finally, Rosin offered no evidence that HOPE VI’s beneficiaries contributed meaningfully to growing poverty in their new neighborhoods, let alone committed more than their share of crime there. But Dreier and Briggs themselves had no evidence that the relocatees were not increasing disorder. The answer to this question, however, would come shortly.64
In 2011, Susan Popkin, Michael Rich, Leah Hendey, Chris Hayes, and Joe Parilla undertook the most methodologically sophisticated study to date of relocated residents’ impact on local crime. Their research tapped the rich veins of statistical evidence offered by the nation’s two largest HOPE VI sites, Chicago (1999–2008) and Atlanta (2002–2009). The good crime news of recent years, however, complicated matters, making it difficult to determine whether or not PHD transplants elevated crime when crime itself was tumbling nearly everywhere. Rather than simply look at crime trajectories, the researchers constructed yardsticks of predicted crime declines against which to compare actual drops after HOPE VI relocations. The results suggested residents’ fears (amply revealed in focus groups) of HOPE VI were largely, but not entirely, misplaced. In both cities, violent and property crime declined dramatically in neighborhoods where PHDs were demolished. The effects in neighborhoods receiving HOPE VI residents were a little more complex. As long as the number of former project residents in a new neighborhood remained low, there was no real increase in crime. But “once the number of relocated households reached a certain threshold, crime rates, on average, decreased less than they would have if there had been no former public housing in-movers.” When relocated households made up more than six per one thousand households, the violent crime rate was 11 percent higher on average in Atlanta, and 13 percent in Chicago. When such neighborhood density exceeded fourteen, the average crime rate was 21 percent higher in both cities. So, although crime still dropped in such neighborhoods, it dropped less than if the neighborhood had not received HOPE VI residents. Although these modest increases belie Rosin’s alarmist account, it was also clear from the data that HOPE VI voucher-assisted households who had previously lived in public housing brought more troubles with them than conventional Section 8 holders who had not lived in public housing. As Popkin et al. observed, “Traditional voucher holders have much smaller effects on crime rates than relocated [HOPE VI] households, and it takes a much higher density of traditional voucher holders before we see any effect at all.”65
Such studies run the risk of casting public housing residents as innately criminogenic elements that, like radioactive ash, must be held below a certain threshold in the environment. To their credit, Popkin et al. use their findings to bolster an argument for “responsible relocation,” with a particular emphasis on supportive services. But the political and social choices that created the conditions for former public housing residents’ criminal participation can easily vanish from view given the framing of such research.
Conclusion
If questions about the relationship between public housing developments and crime rates have been of such concern to researchers, why were good answers so long in coming? And why did those answers often seem to be frustratingly mixed, if not outright contradictory, when they did arrive? No doubt, a public that had already made up its mind about public housing—whether that opinion was favorable, as during the New Deal, or hostile, as during the Reagan Revolution— did not clamor for complete and complex answers. And even specialists can overlook methodological lapses when findings align with useful paradigms. The 1930s experts who championed the crime-fighting potential of “hygienic” structures and open spaces were as imprecise as those who, four decades later, linked modernist architecture to criminal invasion.66 As Michel Foucault reminds us, historical moments generate the authoritative specialists and putative knowledge broader social agendas require.67 But even doggedly independent researchers with large budgets and open eyes might have found it difficult to assess the moving target that public housing presented. As a program, public housing has long been both varied and changing. What was true for St. Louis was not true for Baltimore, and what was true for New York in 1965 was not true there in 1980. Different methodologies applied in different cities at different times generated—unsurprisingly—different findings. But the simple and frequently invoked idea that “public housing increases crime” is demonstrably false as both an analytic conclusion and a causal model, even as it has proved remarkably powerful as a cultural narrative. The variations in public housing’s context, physical form, and management policies require answers with more explanatory suppleness. Nearly two generations of research on public housing and crime leaves us with this fairly unsatisfying generalization: some public housing seems to have increased crime rates under certain conditions, while other projects, under different conditions, are actually safer than private housing.
Taking a historical approach to research about public housing crime highlights the importance of context for understanding both the nature of public housing itself and how knowledge about it gets produced. Simplistic comparisons of crime rates between PHDs and their surrounding neighborhoods generated research questions that ignored the larger social and historical forces that located public housing in the most vulnerable neighborhoods. This made it nearly impossible for researchers to identify the “comparison areas” required by statistical analysis. Likewise, the research environment of the 1970s narrowed scholars’ analytic vision from larger social forces to the criminogenic impact of architectural details. The lure of federal dollars a decade later pushed discussions (and finger-pointing) away from design and toward drugs, despite a dearth of reliable crime numbers.
The emergence over time of more discerning analytical tools has made it possible to disaggregate the data and so achieve greater specificity about crime at the neighborhood level. Pinpointing crime with quantitative data, however, underscores the importance of attention to the larger social context. Public housing’s shifting tenancy changed both how much crime occurred and the relative mix of property and violent offenses. The architectural determinism of “high rise = high crime” proved overly simplistic under more rigorous analysis, as the number and demographic characteristics of residents proved to be much more important for crime than the type of structure in which they lived. As more and more deeply disadvantaged people found themselves concentrated in public housing and left without meaningful support, the chances that their PHD would be marred by violence grew as well.
But it was not just the research questions about public housing crime that proved vexingly complex; the large-scale policy solutions of recent years have not unfolded as originally envisioned. Despite social scientists’ straightforward predictions and politicians’ easy platitudes, the safety benefits of breaking up the concentrated poverty of PHDs have proved to be highly contingent and sometimes contradictory. In practice, deconcentration has benefited girls more than boys; worked less well with PHD residents than Section 8 holders; and occasionally elevated crime. As this and so many other examples confirm, questions about crime and public housing offer no easy answers for stakeholders—researchers, policymakers, and public housing residents alike.