In America, homelessness, defined as living in public spaces or in facilities that accommodate people who otherwise would be living in public spaces, continues to be a large and intractable problem. The scholarly literature includes a study that estimates the number of people nationwide who, over a five-year period, have experienced a spell of homelessness at nearly six million.1 A recent survey conducted by the federal Department of Housing and Urban Development (HUD) found that nationally, on a single night in January 2013, there were 591,768 homeless persons and that about 1.42 million people spent at least one night during that year in an emergency shelter or a transitional housing program.2 The federal government spent $4.042 billion on projects targeted to the homeless in fiscal year 2012.3 This is, of course, not very much in an over $3 trillion federal budget. However, there is a lot of federal spending that is not directly targeted to homeless people but nevertheless serves them. For example, many homeless people are served by the Medicaid program, but we do not have data about how many such people there are. How much in total is spent by all state and local governments on homelessness is also not known. The National Alliance to End Homelessness estimates that, at a minimum, each federal dollar spent on the homeless is matched at the state or local level.4
Even so, it cannot be said that America spends a great deal, relatively speaking, on the homeless. However, homelessness has a greater salience than budget figures imply. To be without shelter is to be without a basic necessity of life. The conditions under which many homeless people live can be very bad. The quality of life for even the sheltered homeless can also be poor. Further, homeless people often suffer from other debilitating conditions. Thus, according to a 2010 HUD report, 34.7 percent of sheltered homeless people were substance abusers, 26.2 percent were severely mentally ill, 12.3 percent were victims of domestic violence, and 3.9 percent were persons with HIV/AIDS.5 It is not surprising, then, that a March 2013 Gallup Poll found that 75 percent of the American public said they worry either a “great deal” or a “fair amount” about homelessness and hunger.6 In short, homelessness is widely recognized as an important social problem.
The ability of the American political system to deal effectively with pressing social problems has often been questioned. Much political science and public administration literature has claimed that change, when it happens in policymaking and implementation, occurs slowly and incrementally.7 Claims that national policymaking has ground to a halt are frequent today as a Republican Congress faces off against a Democratic president.8 But concern over gridlock and drift goes back to at least the early 1960s and has continued through the following decades and into the twenty-first century.9 As one introduction to the American policymaking process notes, “One of the most enduring criticisms of American government has been its seeming inability to respond quickly and responsibly to important policy issues of the day.”10
How effective, then, has our political system been at dealing with the pressing problem of homelessness? For various reasons it is best to address this question by looking not at the federal level but at lower levels of government. To begin with, probably the majority of the money spent on homelessness is spent at the state and local levels.11 Almost all shelters and other services delivered to the homeless are operated by nongovernmental organizations under contract to local governments, or by local governments themselves. And the key experience that results in much of the public concern over homelessness is that of seeing and confronting a homeless person, which is, in its nature, an event that takes place at the local level.
The jurisdiction that has developed the most ambitious policy to address the problem of homelessness is New York City. In April 2015, New York City sheltered 59,285 homeless people (including both single individuals and members of families), and an estimate based on a street survey done in February of that year indicates that there were an additional 3,182 persons living in public spaces.12 In fiscal year 2011 the city’s Department of Homeless Services spent $1.47 billion. No other American city spends nearly as much on the homeless as New York or has close to as large a shelter system. The poor quality of life for at least some of the city’s homeless has received wide attention. New York City’s infamous welfare hotels were icons of the suffering of the urban poor.13 Just as disturbing are reports of a small population of homeless people who live in the city’s tunnels and other underground spaces.14
New York City is also the jurisdiction with the longest history of coping with homelessness. The plight of the so-called disaffiliated alcoholics of the Bowery was documented in the early 1960s and had been dealt with by the city in various ways for decades before then. The 1960s also saw the development of “hotel families,” that is, families that had been burned out of or otherwise lost their housing and were put up in hotels at the city’s expense.15 These episodes belong to what might be called the prehistory of homelessness policy in New York.
A whole new policy framework was created by the signing of a consent decree in the case of Callahan v. Carey on August 26, 1981. As a result of this and other litigation by advocates for the homeless, the city is one of the few local governments with a court-recognized and enforceable policy of providing shelter to anyone who requests it. New York City is therefore the main stage on which the pressing national problem of homelessness has been addressed.
But just as there is a literature on the shortcomings of the national policy process, there is a similar literature on the limits of urban politics, especially New York City politics. Urban policymaking especially is held to be slow and incremental. A standard textbook on urban politics sums up the consensus of the literature on the nature of urban politics as follows: “Urban political structures change slowly in an incremental, evolutionary fashion.”16 Classic statements of pluralist and corporatist theories of urban politics predict that urban politics, in particular New York City politics, will tend towards “stasis” as proposed changes are fought by the many interest groups with stakes in the status quo to defend.17 Later, structuralist accounts of urban politics have argued that various sorts of economic, constitutional, and social structures drastically limit the possibilities for change.18
Urban politics generally, New York City politics particularly, and New York City public bureaucracies especially are held to be resistant to change. In her analysis of the city’s agencies, Blanche Blank begins with the observation that “[i]t may be the very essence of any bureaucracy to move cautiously, even ponderously,” and concludes that “[c]hange is always incremental as it is too in New York City government.”19 And in their updating of Wallace Sayre and Herbert Kaufman’s work, the authors of Power Failure: New York City Politics and Policy since 1960 hold that “[t]here is much inertia within the system, which makes change difficult. . . . The major difficulty is in putting forward changes, regardless of the direction of the change.”20 And finally, of all bureaucracies, street-level bureaucracies—that is, public service agencies that employ a significant number of front-line workers who interact with the public with substantial discretion—are held to be very resistant to change. Thus in his classic account of street-level bureaucracies, Michael Lipsky writes, “I have argued that the determinants of street-level practice are deeply rooted in the structure of the work. . . . These observations contribute to our understanding of the stability of the institutions and their unlikely response to significant reform activities.”21
Some recent literature has challenged this picture of slow change and has argued that more dramatic change is possible. The most comprehensive statement of this challenge is The New Politics of Public Policy, edited by Marc Landy and Martin Levin (1995). The contributors to this volume argue that the combination of the professionalization of reform, a fragmented institutional framework, divided government, a highly competitive political environment, and the rights revolution have encouraged various types of policy entrepreneurs to compete with each other to have the best claim to popular political ideas. The result has been that “[a] shift has taken place from a ‘politics of interests’ (which tend to be fixed and thus to change slowly) to a ‘politics of values and ideas’ (which tend to be more open, fluid, and responsive to change and reason).”22 However, the contributors to New Politics are concerned with national political developments and have little to say about urban politics. The question is, Can this new paradigm of change, which we can call the ideational paradigm, be usefully applied to the supposedly very static forum of city politics?
This book is part of an ongoing effort to test the validity of the received wisdom on urban political change, and the usefulness of the ideational paradigm, by examining in detail various efforts that have been made by successive mayoral administrations to introduce major changes in New York City’s public welfare policy and bureaucracy. I have elsewhere argued that throughout the last thirty years, homelessness and welfare policy in New York City have developed in a series of quantum or nonincremental jumps with dramatic, short-term changes in funding, administration, and policy “philosophy.”23 Put another way, nonincremental change is the product of the “politics of values and ideas,” with its key features of expertise, policy entrepreneurs, institutional fragmentation, competitive political environments, and, especially, public ideas. My purpose here is to develop this account in more detail and at greater length in the context of the city’s efforts to deal with the challenging social problem of homelessness.
Another theme broached by this case history is that of what might be called the perversity thesis, which is the idea that addressing a social problem can actually make it worse. An obvious problem with the jerry-built system as it developed in the 1980s was the poor quality of the shelter it provided. Many clients, both single individuals and families, were quartered in large barracks-style shelters. Welfare hotels for families were also of low quality. One solution to these problems was to move families to in rem apartments that the city had come to own as a result of tax delinquency and abandonment by their former owners and that had been refurbished. This solution was pursued vigorously in the early days of the Dinkins administration.
Rapid rehousing to renovated apartments seemed to have created a perverse incentive to become homeless in order to qualify for housing. Many people in the Dinkins administration came to believe this as the number of families entering the system rose for a time. A blue ribbon advisory commission established by Dinkins and headed by Andrew Cuomo later endorsed this claim, as did the present author.
Later scholarship, at the least, cast serious doubt on the reality of perverse incentives in the city’s shelter system.24 Michael Cragg and Brendan O’Flaherty have convincingly argued that the immediate impact of rapidly rehousing a family is to reduce the shelter population by one family and that whatever perverse incentive was created was not strong enough to pull more than one new family into the shelter system. Nonetheless, city administrators and some observers continued to be concerned. Near the end of his administration, Mayor Bloomberg himself suggested that his administration took the possibility of perverse incentives very seriously.25 New York City’s homelessness shows us how ideas, in this case what A. O. Hirschman calls the “perversity thesis,” can influence policy.26
Let me anticipate some of my findings here by noting that in New York City, the process of establishing shelter as a right has gone through three distinct stages or moments. These are described in the following sections.
Simply establishing that there indeed is a right to shelter and then delivering on that entitlement is one of the central challenges to policy. The courts and various advocacy groups such as Coalition for the Homeless are primarily concerned with this aspect of homelessness policy. These interests push policy in the direction of developing a shelter system that is large, court supervised, and primarily concerned with service delivery. Establishing and implementing a right to shelter is one major theme in New York City’s policy, a theme that was especially prominent in the early days—that is, through the eighties to the early nineties—of modern homelessness policy.
The right to shelter completely transformed the city’s homelessness system. The system grew tremendously in the early eighties. While 7,584 individuals were sheltered in 1982, 21,154 were sheltered in 1985. Spending grew from $6.8 million in 1978, just before the litigation to establish a right to shelter began, to $100 million in 1985.27 To cope with the rapidly expanding demand, the city rushed to open large, barracks-style shelters where hundreds of clients would sleep in cots laid out in open spaces. During these years the city also relied on commercial welfare hotels to shelter homeless families at the cost of $72 million in 1986. The shelter system during these years was satisfactory neither from a conservative nor from a liberal point of view. The right to shelter was absolute, and unbalanced by any requirements to work, participate in rehabilitation, or seek permanent housing. Moreover, shelter quality was often very poor, and few services were offered to clients. The city had created a system that guaranteed the right to free, low-quality shelter.
Entitlement is one axis around which New York City homelessness policy has spun. As time went by, however, the limits of this purely entitlement-based, emergency-oriented system showed themselves.
The unconditional right to shelter proved to be problematic in various ways. Behavioral problems—such as substance abuse, nonwork, and criminal activity—of some of the homeless required that the entitlement to shelter be conditioned on proper behavior, including participation in work and treatment programs. Strong conceptions of the rights of the mentally ill sometimes had to be limited in order to provide necessary protection and therapy. This set of challenges is of particular concern to mayors and administrators who, unlike the courts or advocates, are responsible for the actual operation of the shelter system. These bureaucracy-based actors therefore push policy in a paternalistic direction, one in which rights are conditioned on good behavior and on participation in programs such as drug treatment, work, and activities designed to move clients out of the shelters as soon as possible.
During the Giuliani administration, the shelter system was much changed from what it had been in the eighties and early nineties, mostly in a paternalistic direction. While in the 1980s most shelters were government run, the system was privatized or, more accurately, not-for-profitized. That change improved shelter quality. Not-for-profitization has also made it possible for the system to impose work or rehabilitation requirements on clients. The city still provided shelter to everyone who asked for it. But not-for-profit shelters can require their clients to work, or participate in rehabilitation, in order to stay in that particular shelter. (Clients who decline to participate are sent back to a city-run, general-intake shelter.) In other words, privatization made paternalism possible.28
Family homelessness policy has also over the years shifted away from entitlement and towards paternalism. Throughout the nineties and the early twenty-first century, one of the worst aspects of the shelter system was the intake point for homeless families known as the Emergency Assistance Unit (EAU).29 During that period, the EAU was typically crowded with families seeking shelter. Processing their requests took many hours and families often slept overnight on the office floor waiting for a determination. One of the main problems was what to do with families that had been found ineligible and who frequently reapplied almost immediately, thus blocking up the process. One knowledgeable observer of city homelessness policy said of the EAU at that time that “[i]t is probably one of the most disturbing places on Earth.”30
However, since 2003 things are much changed. A court decision that held that DHS may deny shelter to anyone who reapplies within ninety days after his or her application is rejected, unless there has been a change in his or her circumstances, helped end the bottleneck and improve conditions. This was a significant relaxation of the right to shelter, one that allowed DHS to redefine the task performed at the intake point from one of sheltering and determination of eligibility to one of diversion and prevention. Chapter 4, which covers homelessness policy during the Bloomberg administration, describes in detail this process of restructuring intake procedures to the city’s family shelter system.
Beginning in the late Dinkins administration and continuing through the Giuliani administration and much of the Bloomberg administration, city homelessness policy developed in a paternalistic direction, one that emphasized the importance of getting homeless people who are able to do so to take responsibility for their housing situation. The drive to develop such a paternalistic policy has required the city to get itself out from under the constraints of the many lawsuits that drive the city’s homelessness policy. The city has had to “reinvent” its Department of Homeless Services as a more decentralized and flexible system. In short, New York’s homelessness system has evolved from its beginning as a centralized, highly constrained, and entitlement-based system to one that is much more decentralized and privatized and that emphasizes clients’ responsibilities as well as their rights.
But paternalism turned out to have its limits, just as entitlement did. Paternalism greatly improved management of homeless services and responded to political demands for more responsibility on the part of recipients. What paternalism did not do was offer much hope of eventually “solving” the problem of homelessness. Despite efforts to diagnose and then treat the “underlying causes” of homelessness, the number of people on the street and of families entering shelters remained frustratingly high. The overall shelter census continued to go up, as did the budget for services for the homeless. The paternalistic reforms, promoted under the Dinkins administration by a special commission led by Andrew Cuomo, and implemented with much fanfare during the Giuliani years, seemed not to be making a dent in these two fundamental measures of success. Paternalism had done a better job at managing homelessness but had failed as a strategy for solving homelessness.
The next moment in New York City homelessness policy had its origins in efforts to come up with a strategy that would “solve” homelessness. A crucial part of that effort was what amounted to a redefinition of the homelessness problem by the well-known researcher Dennis Culhane. In the late nineties, only 10 percent of the single homeless persons in New York—who were the most disabled and whom Culhane identified as the “chronic” homeless—accounted for almost half of the shelter days provided by the city. This discovery allowed the homelessness problem to be redefined in such a way that a “solution” seemed within reach: Focus on the relatively small chronic population, house them, thus making a disproportionate impact on reducing shelter use, and declare victory.
The question then became where to house the chronically homeless. The answer was provided by an approach that began to be developed by housing activists in the mid-1980s and that would be expanded between 1990 and 2005 in a series of agreements between New York City and New York State: supportive housing. Supportive housing is subsidized housing for people with various sorts of disabilities who are provided medical and social services on-site so that they can live independently off the streets. Supportive housing was an appropriate place to send the worst-off street dwellers and would receive its largest commitment of resources under the New York/New York III agreement signed during Bloomberg’s first term.
In its first iterations, supportive housing projects generally embraced the paternalistic approach and demanded that homeless people sober up, take their medications, and demonstrate that they were “housing ready” before they would be moved into an apartment. The problem was that many street dwellers declined to accept this deal, which made achieving dramatic reductions in street homelessness difficult. Sam Tsemberis, a psychologist experienced in outreach work to the street homeless and founder of the innovative service organization Pathways to Housing, came up with a response. His “Housing First” approach to outreach involved breaking with the paternalist quid pro quo and providing street dwellers with housing before asking them, or perhaps without asking them, for compliance with rehabilitation. Many single homeless people, it turned out, who had previously declined shelter on paternalistic terms were willing to take this deal.
Housing First was developed as an outreach strategy directed to street dwellers but also had an impact on policy toward homeless families. From the eighties to the mid-1990s, it was thought by some observers—including the present author—that homeless families were much more troubled than similar, nonhomeless poor families with problems such as drug use, mental illness, criminal activity, and “underclass” pathologies. Here again, the thought was that there was an underlying cause of the homelessness of many families. By the mid-1990s, research indicated that homeless families, though they suffered higher rate of such problems than similar poor families, were not as dramatically worse as had been thought. In any case, research also showed that whatever their problems, homeless families could generally stay stably placed in permanent housing even if they did not receive any rehabilitative services. The key to rapidly rehousing them was not services, but subsidies. Homeless families, whatever their troubles, could usually live outside the shelter system if they received access to public housing or Section 8 vouchers and other forms of rental subsidy. Thus, under the influence of the Housing First strategy for singles, policy for families began to move away from diagnosing underlying causes and providing appropriate services to planning for rapid rehousing of shelter families, with some form of subsidies being a prominent part of that plan.
The post-paternalistic features of the city’s homelessness policy were broached during the early Bloomberg years. It was under Bloomberg that, with much publicity and acclaim, a five-year plan was introduced, the expressed purpose of which was to “overcome” or end homelessness. Ending homelessness really meant having a disproportionate impact on the use of shelters and services by focusing on the chronically homeless, as Culhane had suggested, sending them to supportive housing, and doing so without demanding “good behavior” first, in keeping with the Housing First policy. Implementation of Housing First strategies proceeded apace under Bloomberg, as did the analogous family policy of rapid rehousing, which, in Bloomberg’s first term, involved a reliance on various sorts of housing subsidies.
The results of post-paternalism have been mixed, perhaps because this policy philosophy has been incompletely implemented. The Housing First strategy for single homeless people has been effective in considerably reducing the city’s population of street dwellers, by about 24 percent between 2005 and 2014.31 The situation with the shelter population was much different. The census in the shelter system rose throughout the Bloomberg years and was at an all-time high at the end of his final term.
This may be the case because the Housing First strategy was never fully implemented for families. Rapid rehousing consisted mostly in planning to move families out of the shelter almost as soon as they entered, rather than waiting for various sorts of rehabilitative programs to take effect. But a signature Bloomberg policy for dealing with homeless families was “delinking,” that is, ending priority access of homeless families to Section 8 vouchers and vacant public housing units. Such delinking was supposed to put an end to the “perverse incentive” of receiving subsidies upon becoming homeless, and was therefore expected to abate the flow of families into the shelter system. Also under Bloomberg, an important rent subsidy for homeless families, the Advantage program, came to an end under complicated circumstances. The delinking strategy and the end of rent subsidies were out of keeping with post-paternalism, which, when applied to families, implied reliance on rent subsidies to achieve rapid access to permanent housing.
We have, then, three stages in the development of homeless policy in New York City: entitlement, paternalism, and post-paternalism. Actually, these stages are more like facets or aspects. Paternalism did not end entitlement; paternalism assumed the homeless had a right to shelter but located the cause of homelessness in the homeless person and demanded that he or she “give something back” in return for shelter and services. Post-paternalism would have undermined paternalism, but has been incompletely implemented. The result is that paternalism has been imposed on top of entitlement, and post-paternalism on top of paternalism. The city’s homeless policy is therefore quite complex, and is driven by three distinct “philosophies.”
Finally, it is important to note what this book does not do. It does not offer policy prescriptions for how to end homelessness. Such work has been done very ably by Dennis Culhane in his analysis of the problem of chronic homelessness and by Sam Tsemberis and other researchers examining the Housing First approach to sheltering street dwellers. Both these approaches are discussed in the course of the case history told here. This book also does not offer a theory of homelessness. As I have noted elsewhere, the dominant theory of homelessness among academics is a structural theory, that is, one in which longstanding economic and social trends that individuals can have little immediate influence over are held to be the main causes of homelessness while personal characteristics of homeless individuals are held to be of secondary importance.32 One specific form this structural approach has often taken is to argue that since the late seventies, a decrease in inexpensive housing and an increase in the number of poor people has turned the affordable housing market into a game of “musical chairs,” in which some people must end up on the street due to a sheer lack of appropriate units. In this situation it is the weakest—the mentally ill, the addicted, or those who are otherwise incapable of coping—who end up displaced. So, structural features of the housing market and the economy determine the essential fact that there will be a certain amount of homelessness, while individual characteristics explain precisely who will make up the homeless. This book perhaps could, therefore, begin by elaborating that structural theory in detail as a backdrop against which policy history must be understood.
There are several reasons why I have not begun by laying out this theory. First, developing a theory of homelessness in New York City is beyond the scope of this book, which has the already ambitious objective of chronicling the policy and politics of homelessness. Also, despite excellent work that had been done in applying the structural theory to social reality by authors such as Martha Burt and Brendhan O’Flaherty, I remain somewhat skeptical of the structuralist account and feel that for all its explanatory strength there remains something mysterious about the phenomenon of mass homelessness in New York City. For New York City’s population of about sixty thousand (sheltered and unsheltered) homeless people is “small” compared to the total housing stock of the city, which totals 3,084,861 units, including 1,533,107 rentals.33 It is certainly true that most of these units are too expensive or otherwise not available to the homeless. But more than three million housing units makes for a great deal of physical space with very many unused rooms, attics, basements, garages, and other potential accommodations. Many homeless people have at least some income, if only from welfare and disability payments, and it is reported that a fair number work.34 Why don’t the owners of all that unused space rent it out at a pittance and make at least some money from an otherwise unprofitable asset? In a classic article, Milton Friedman and George Stigler demonstrated that due to such renting out of previously unused space, there was no housing shortage after the San Francisco earthquake of 1906 even though the city’s housing stock was reduced by more than half.35 In New York City today, unused space is often converted—frequently illegally—in this fashion, creating a “housing underground” that has been estimated to include many more than 114,000 apartments.36 An unknown number of low-income immigrants live in this housing underground, and a recent exploration of this practice among Chinese immigrants found that “[t]he existence of unauthorized apartments provides immigrants, low-wage workers and students a roof over their heads and prevents some of them from becoming homeless.”37 Why doesn’t this housing underground expand, perhaps considerably, and take up at least more of the city’s “small” homeless population? Working out and testing my skepticism of the dominant structural model of homelessness is a project for another book.
But let’s assume this skepticism is unwarranted—that in fact there is no mystery over why people are homeless and that an overview of the structural causes of homelessness could perhaps serve as an introduction to this book. There is another, more important reason for not taking this approach. This project is to a considerable extent historical: it covers a period of more than thirty-five years. The early chapters, at least, are concerned with understanding the actions, decisions, and political judgments of various actors in the past. To understand what people did in the past, what is known now may or may not be relevant. But what was known at the time the actors made their decisions is crucial. Whether or not homelessness should be a mystery to us now, many policy actors of the seventies, eighties, and nineties at least thought homelessness was a mystery. As the interviews conducted for this book will bear out, those actors were confused about the roles deinstitutionalization, personal disability, cultural factors, family demographics, policy history, as well as income and housing trends played in creating the homelessness crisis. And they were legitimately confused, however certain we may be of things now. Kicking off a history with the very latest account of what happened and then judging past actors according to how closely they approximated that understanding is anachronistic. This is not to say that some mysteries didn’t clear up over time. In particular, we have learned much more over time about the phenomenon of perverse incentives as it applies in homeless policy, and about how to improve outcomes in housing mentally ill street dwellers. In these matters, I feel it is fair to say that later actors should have known more than was known at the beginning, and to judge their decisions accordingly. But earlier actors, including the present author, who played a minor role in some of the developments chronicled here, had a right to be unsure.
The following chapters tell the story of New York City’s homelessness policy from 1979, when the first work that would lead to the Callahan v. Carey consent decree was begun, to the early days of the de Blasio administration up to 2015. The administrations of Koch, Dinkins, Giuliani, Bloomberg, and de Blasio’s early years each receives its own chapter. A final chapter draws conclusions about the themes broached in this introduction and about the nature of urban politics.