When David Dinkins ran for mayor in 1989, he set forth an ambitious plan to change the way the city dealt with homeless families. Homelessness, he argued, was basically a housing problem, one that could be solved by rapidly relocating families from shelters and welfare hotels into permanent housing.
The blueprint for the new administration’s homelessness policy had been prepared in March 1987 by a task force that Dinkins, then Manhattan borough president, appointed. Its report, entitled A Shelter Is Not a Home, defined family homelessness in economic terms and argued that the solution was to provide more permanent housing.1 Since the report’s authors thought the private market incapable of supplying affordable housing, the city itself would have to do the job.
In fact, the city was already providing homeless families with permanent housing, placing them in apartments rehabilitated by the Department of Housing Preservation and Development (HPD). These were the so-called in rem apartments, which the city came to own as a result of tax delinquency by their owners. The number of HPD units distributed to homeless families had been rising steadily during the latter years of the Koch administration. In 1985, 2,401 such units were provided to homeless families; in 1987, 3,635. In 1988, the administration planned to increase the number to 4,042. The Dinkins task force called on the city to increase its production of rehabilitated apartments for homeless families to 8,000 per year.
The Dinkins proposal assumed that doubling the supply of city-provided permanent housing would not stimulate demand for temporary shelter. In a section of its report entitled “Myths and Facts about Homeless Families,” the task force criticized what it called “Myth 7: Families with other housing options voluntarily enter the shelter system in order to upgrade their housing at the public’s expense and get priority for rehabilitated permanent housing. Fact: There is little evidence that families forsake stable housing arrangements hoping to get an apartment through the city.”
The report pointed out that although the number of families sheltered by the city had grown steadily during the mid-1980s (from 1,153 in January 1983 to 4,365 in September 1986), the number of families requesting shelter each month did not rise. “There is no indication that the availability of emergency housing has produced an increase in demand for shelter or an increase in homelessness,” the report stated. The task force concluded that the growth of family homelessness resulted from too few shelter-system families moving out to permanent housing, not from too many marginally housed families moving into the shelter system.
The task force further contended that homeless families did not suffer unusually high rates of personal or behavioral problems that would impair their ability to maintain stable housing arrangements. The report pointed out that, at the time, no scientific studies had demonstrated that such dysfunctions were disproportionately common among shelter families:
Myth 3: Homeless families are the “worst” of New York’s welfare families. Family homelessness is the result of drug addiction, mental illness, or behavioral dysfunction. Fact: No facts support the belief that addiction or behavioral problems occur with more frequency in the homeless family population than in a similar socioeconomic population. Homeless families are not demographically different from other public assistance families when they enter the shelter system. . . . Family homelessness is typically a housing and income problem: the unavailability of affordable housing and the inadequacy of public assistance income.
The claim was, in effect, that since scientific research had not proven that the homeless were much worse in these terms than other poor people, behavioral problems were not the primary causes of homelessness, which were therefore largely structural and economic. It followed that there was no reason to expect that formerly homeless families, once they were provided with permanent housing, would not thrive and escape the shelter system permanently. “In most cases homeless families are able to manage households and can be fully integrated into ordinary permanent housing,” the task force declared.
When Dinkins took office in 1990, his administration put its ideas to the test. The mayor appointed Nancy Wackstein, an advocate with the Citizens’ Committee for Children who had served on the staff of his task force, to head the Mayor’s Office on Homelessness and SRO Housing. Under her direction, the city vastly increased the supply of permanent housing for homeless families and continued a Koch administration initiative to give homeless families priority for vacant public housing units.
A program was already in place in 1990 to provide homeless families with permanent housing. In August 1988, in response to threats by the federal government to stop funding welfare hotels, the Koch administration had stepped up the supply of permanent housing and stopped assigning new families to the hotels. Instead, they were placed in barracks-style shelters (Tier I) or shelters with private rooms, congregate dining facilities, and stricter supervision than at hotels (Tier II). Families that had been in any part of the system for twelve months (later reduced to nine months) became eligible for permanent housing. Human Resources Administration (HRA) reports at the time showed that most homeless families greatly preferred the relative autonomy of hotel accommodations to the closely supervised Tier II shelters.2 Thus, the Koch administration was consciously making the conditions under which shelter-system families would await apartments less attractive. As Koch himself put the matter, “we are going to, whenever we can, put people into congregate housing like the Roberto Clemente shelter—which is not something people might rush into, as opposed to seeking to go into a hotel.”3 As a result, the number of families entering the shelter system fell during the later Koch years, even though the number of permanent housing placements increased.
Dinkins, too, wanted to end the use of welfare hotels and supply permanent housing to shelter-system families. But he pursued these goals by means quite different from those of Mayor Koch. For one thing, the Dinkins administration could draw from a greater supply of permanent housing, made available by Koch’s programs that were beginning to produce results. The number of units provided by the Department of Housing Preservation and Development to homeless families, which had dropped from 4,042 in 1988 to 3,201 in 1989, rose to its former level. Under a 1988 agreement between the city and the New York City Housing Authority (NYCHA), an increasing number of homeless families were moved into public housing. And other homeless families were relocated to privately owned apartments under the Emergency Assistance Rehousing Program (EARP), which paid landlords rent and gave them other substantial incentives to accept formerly homeless tenants.
The Dinkins administration hoped to take advantage of the increased supply of permanent housing to put an immediate end to the use of welfare hotels, which were already being phased out. But unlike Koch, Dinkins was unable or unwilling to move hotel families into less desirable accommodations to await permanent housing; further, in May 1990, the City Council had enacted legislation to phase out the use of Tier I shelters by June 1992. At the same time, Dinkins had to contend with the ongoing enforcement of McCain.
Judge Freedman’s handling of McCain from the mid-1980s until the case was settled in 2008 has been controversial. For example, Demers argues that “[i]n several instances . . . the court essentially substituted its views of the appropriate standards to be applied for those of the executive agency officials whose responsibility it was to develop standards.”4 The conservative City Journal claimed that “[i]n scores of rulings since 1985, Freedman has stymied New York’s efforts to take control of its policy for homeless families.”5 Eventual city corporation counsel Michael Cardozo would claim that “the city . . . over the next twenty-five years litigated an unending barrage of motions in which the plaintiffs, relying on that initial preliminary injunction, challenged the city’s evolving efforts to address homelessness. The litigation resulted in over fifty court orders touching on and regulating virtually every aspect of the city’s shelter program.”6 Even the New York Times, which was generally supportive of Freedman’s role, could take note of the “more than 40 court orders that prescribed everything from where and when homeless families could sleep to whether bottle warmers and infant formula must be available at city intake offices.”7 We can get a sense of how fair these charges are by looking at some of the court’s most noted decisions.
On June 1, 1990, Justice Freedman issued a stipulation and order that required, among other things, that after July 31, 1990, the city would stop placing homeless families in housing that was not in compliance with a state regulation mentioned earlier, 18 NYCRR 352.3. The homeless families and their lawyers claimed that the city was placing families in noncompliant welfare hotels. The city argued that it was in compliance with some of the requirements of the June 1, 1990, court order and asked to be relieved from others because an unexpected increase in applicants made compliance impossible. New York State was also a defendant in the case, as we shall see. The case involved a complex web of claims and counterclaims. Many homeless families provided affidavits that the hotels did not provide basic furniture or safety features as state regulations required. City employees stationed at the hotels provided affidavits claiming that clients’ basic needs were being met. Justice Freedman cited with approval the statements of the families but found the employees’ statements “unconvincing” because many of the families in the hotels claimed they never received help from, or even saw, these city workers.8 It would seem that it was necessary at this point to go beyond city and client anecdotes and make a systematic, large-scale evaluation of hotel conditions. Justice Freedman implicitly acknowledged that such a study was necessary when she declined the city’s invitation to visit the hotels because “[i]t would be necessary to visit every room at every hotel and interview a large number of families in order to establish the level of compliance.” Courts typically don’t have the bureaucratic capacity to undertake such evaluations, which is one reason courts are ill suited to oversee complex organizations.
Nonetheless, Justice Freedman went forward. Many of the hotel rooms in which the city placed clients lacked private kitchen facilities. The city argued that such placements were in compliance with 18 NYCRR 352.3, part (e) of which read that “[h]otel/motel accommodations without cooking facilities shall be utilized only when accommodations with such facilities are not available.”9 Part (h) of the regulation, which required that the city determine whether a hotel is compliant with relevant state and local laws, referred, according to the State Department of Social Services, to fire and safety codes only. Part (g) of the regulation, which asserted, among other things, that “[p]rimary consideration shall be given to the needs of children,” did not seem to be directly on point. Since interpretations of a regulation by the agency that issues them are binding, unless that interpretation is “irrational or unreasonable,” it would seem that the kitchenless hotel units were compliant with the regulations.
Justice Freedman disagreed, writing,
While the Court would ordinarily defer to an agency’s own interpretation of its regulation, the City’s and State’s interpretation of 18 NYCRR 352.3 (h) to apply to safety and fire codes only, is not logical in view of the salutary purpose of regulations set forth in 18 NYCRR 352.2. Thus, the combined effect of 18 NYCRR 352.3 (e) . . . of 18 NYCRR 352.3 (h) . . . and 18 NYCRR352.3 (g) . . . lead[s] this court to conclude that cooking facilities are mandatory.10
Freedman simply isn’t convincing on this matter, especially since the Appellate Division, in Lamboy v. Gross, had made such a point of deferring to an agency’s interpretation of its own regulations.
Another famous episode in the McCain litigation came to a head on November 20, 1992, when Justice Freedman found the city in contempt for violating state laws and court orders and, among other penalties, required four officials of the Dinkins administration to spend a night in the EAUs. The origins of this development date back to a stipulation and order issued by Justice Freedman on June 1, 1990. That document was essentially a contract, supervised by the court, between the plaintiff homeless families and the city. Prior to the agreement, the city had often placed families in shelters or hotels that did not comply with state regulations or earlier court decisions. The agreement was complex, but essentially the city pledged to implement a detailed plan that committed them to placing families only in accommodations that complied with state regulations, including 18 NYCRR 352.3, by June 30, 1990. The city also agreed to stop placing families in barracks-style shelters for more than one night. As their part of the agreement, the plaintiffs agreed to refrain from taking action to enforce prior orders of the court while the city implemented the plan.
By January 1991 the city was failing to live up to the plan. The city continued to place families in noncompliant hotels. The city brought its failure to live up to this aspect of the June 1 stipulation and order to the attention of the court in September and October of 1990, and Justice Freedman therefore decided not to hold the city in contempt on this matter. The city invoked the perversity argument and claimed that by transferring families from emergency shelter to permanent housing, it had unintentionally created an incentive for doubled-up families to move into the system. The city shared with the judge its plans for the Alternative Pathways Program (APP), which would offer permanent apartments to doubled-up families and provide various other sorts of assistance to prevent families from moving into the shelter system. In any case, the city was continuing to place families overnight in EAUs, and welfare offices continued to place families with pregnant women and young infants in barracks-style shelters. Justice Freedman wrote that therefore “[t]he court is left with little choice but to issue contempt findings.”11 The city was found in civil contempt, and sixty-two homeless families that were housed in violation of court orders received a total of $16,500.12
On September 10, 1991, the parties were back before the court. The plaintiffs sought civil and criminal contempt charges against the city for 1,325 violations of court orders that they had documented since January.13 The violations included sending pregnant women to barracks-style shelters and placing families in rooms with folding cots or other inappropriate furniture. The city admitted to the violations, and Freedman concluded that “a contempt finding is inevitable” but called for further hearings to decide which specific city officials would be found in contempt and what penalties would be imposed.
Soon thereafter Justice Freedman ordered that the city would have to stop placing families in hotel rooms without kitchenettes by September 30. Further, Local Law 18 required that the city stop using barracks-style shelters entirely by the same date. The city argued that it simply could not meet the deadline. “We can’t comply with the law,” First Deputy Mayor Norman Steisel confessed after a City Hall news conference. “We cannot meet the timetable. We’ve got to figure out what remedies are available to us while still basically fulfilling the overall intent of the law.”14 The city was in the midst of struggling to implement a new plan to deal with the influx of homeless families. This was the Alternative Pathways Program (APP), implemented in October 1990. Under APP, some of the HPD apartments that would otherwise go to shelter-system families were instead made available to families doubled up in private apartments. Much of the plan for the APP that the city had presented to the court back on January 8, 1991, had not been realized. The distribution of apartments to doubled-up families, later studies would show, had little impact on the number of families that entered the shelter system. A component of the program that provided legal services to poor families facing eviction had been delayed, and a $12.4 million program to pay rent for some thirty-two thousand families that were in arrears had not been approved by the state.
It is possible to conclude from all this delay that from 1990 through 1992 the city was simply making no progress on improving its provision for the homeless. Certainly the record that was being presented to the court was unimpressive. Steven Banks’s assessment was that “the record of the past year and a half is replete with failure to stop placing families in welfare hotels, failure to comply with court orders to provide lawful shelter and failure to implement the city’s own homelessness prevention program.”15 It is worth remembering, however, that between September 1991 and May 1992, the Cuomo Commission—led by Andrew Cuomo, whose task was to rethink all facets of the city’s homeless policy—was being formed, was deliberating, and was ultimately embraced by the Dinkins administration. The administration was going through the enormously complicated task of rethinking homeless policy, changing direction, and developing a new organizational structure. All this, however, was happening outside the court system, which concluded that the city was simply spinning its wheels, or worse, not acting in good faith.
By November 1992, Justice Freedman had had enough. On November 13, 1992, she again found the city defendants in civil contempt, and on November 20 she issued contempt findings against four individual city officials, Barbara J. Sabol, commissioner of the Human Resources Administration; Kenneth Murphy, deputy commissioner for Crisis Intervention Services at HRA; First Deputy Mayor Norman Steisel; and Jeffrey L. Carples, assistant to Deputy Mayor Cesar A. Perales. (Sanctions against Carples would be dropped when a reassignment left him no longer responsible for EAU operations and were directed instead against his boss, Perales.) The court ordered that on a given evening, these officials would each have to stay overnight in separate EAUs and remain there until all families applying for shelter had been placed. The court also levied fines against the city that were to be paid to applicant families that had not been placed in compliant shelter between September 20, 1992, and November 20, 1992.16 Justice Freedman commented, “I have been ordering and ordering and ordering cures of these violations for two, three, six and seven years. I just want you and everyone to know that my orders have not been effective.”17 A sense of the tension for all concerned can be gotten from a press account of the proceedings: “‘What do you want us to do?’ the City Corporation Counsel, O. Peter Sherwood, asked plaintively. The judge [Freedman] replied, enunciating each word like an aggravated wife instructing her dimwitted spouse: ‘Do not keep them all night in welfare offices.’”18
In December 1992 the Dinkins administration decided to appeal the court’s contempt findings.19 On July 29, 1993, the Appellate Division issued its decision and showed little sympathy with the city: “[I]t is no defense that the municipal defendants were attempting to comply or acting in good faith. . . . The City’s contention that compliance is impossible is belied by the fact that the City was able to do so in the recent past.” Why it necessarily follows that what was possible in the recent past must therefore be possible in the present is not clear. The Appellate Division upheld the fines that Justice Freedman had imposed on the city. However, the order of the four officials to spend a night in EAUs was deleted because “the only purpose served by such a penalty is to punish these individuals.” The Appellate Division did find that the individual officials could be fined or imprisoned and called on the lower court to “impose an appropriate sanction.”20 On December 8, 1993, Justice Freedman imposed a fine of seventy-five hundred dollars on each of the individual officials and also on Chip Raymond, then commissioner of the newly established Department of Homeless Services.
First Deputy Mayor Norman Steisel described his reaction to being held in contempt:
Obviously, I was pissed off. I guess there are two points of view. For me, it was not a conscious effort to be in contempt. For me, it was not knowingly withholding resources or . . . in a calculated fashion preventing people from doing their job, which was to me contempt. On the other hand, to the extent that we didn’t meet our plans because the bureaucratic hurdles were more formidable . . . then I guess we were in contempt. . . . I was really pissed because I felt I was trying to move the thing forward and I thought to some extent we were being undone by the very same people [who] contributed to the problems in the first place. We had this whole thing about negotiating improved agreements or adequate base agreements to get rooms in SROs and making certain of the hotels that had not been used as SROs . . . available. Obviously, that’s a complicated negotiation because there was such a horrible history. If you were a hotel operator . . . you might have been economically desperate for more money, [but] you had to be careful of who comes in to these facilities given what you perceive [as] the problems and baggage they bring with them. Remember, HRA was always making commitments and saying they [had] agreements lined up and then the agreements would vaporize. At the same time, when communities got wind of the fact that we were talking to some apartment owners or some hotel, they would be distressed. They would let you know about it and my recollection is HRA either persuaded the local property owner or HRA [had] to back off. That was an enormous sense of frustration. So were we in contempt? From the point of view of the right people in the city, the city didn’t do what it said it was going to do.21
Ironically, Deputy Mayor Cesar A. Perales was being held in contempt partly for being unable to comply with Administrative Directive 83 ADM-47, which he had promulgated and interpreted when he was New York State social services commissioner. He also commented on his experience before the court:
I knew Judge Helen Freedman. . . . I’m a civil rights lawyer. I came out of that system. I was a neighborhood legal services lawyer in my youth. And so I know these people, the Legal Aid Society. I was seen as one of them. I was seen as a reformer as [New York State] Social Services Commissioner. . . . My regulations were seen as putting enormous pressure on government to do something about homeless people. So I was, in the eyes of many people, a good guy.
And so it was strange to find myself in court, being told that I was not doing enough or doing it quickly enough, and those were, in essence, the reasons that were periodically brought before the court. This was a question of whether or not we could house people more quickly and meet their needs. And for me it was strange because I had always been seen as being somebody on the other side, fighting for poor people, and here I was trying to run, not run, but oversee—because I was a deputy mayor, right?—HRA. And so that was a very difficult position.
I think the courts were productive. I do think that without court intervention government would have moved more slowly and would have been less responsive to the problem.22
Justice Freedman would eventually try other strategies to achieve city compliance but continued to impose fines up through 2006. The New York Law Journal reported in 2003 that the city had been fined $8 million by that point.23
How effective were the fines and what does this whole episode say about the effectiveness of institutional reform litigation as a mechanism for improving bureaucratic performance? Concerning the effectiveness of the fines, we have to ask what the purpose of the fines was. Steven Banks has made the fair point that the civil contempt charges against the city resulted in compensation for families that had been denied mandated services. Insofar as the fines aimed at compensation, they were effective. But the fines levied against individual administrators and the requirement that these administrators be confined, one evening, to the EAU until all applicants received shelter do not involve compensation. Instead, they aimed at getting the shelter bureaucracy to comply with court orders. Further, Judge Freedman’s famous complaint when she ordered the fines—“I have been ordering and ordering and ordering cures of these violations for two, three, six and seven years . . . [and] my orders have not been effective”—suggests that the purpose of the fines was to get the court orders complied with. The bare facts of the case do not show that city administrators sprang into compliance once they faced contempt charges. The advocate-plaintiffs continued to be unhappy with city performance after the fines were levied and continued their litigation against the city for a long time. It is true that over the long term the pressure of litigation and enforcement of court decrees resulted in a vastly improved shelter system. But the present question is whether Judge Freedman’s findings of contempt and the fines that went with them had a shorter-term positive impact on the system.
An early article on implementation problems in institutional reform litigation held that “defendant administrators do not always have the ability to make their institutions comply. In such cases the contempt sanction is useless and unfair; contempt should be imposed only for failings within the defendant’s control.”24 But what strikes one now about the situation in which the city found itself in 1992–1993 was the complexity of the phenomena the city was struggling with and the sheer stubbornness of the problem they were attempting to solve. It is not at all clear that in this case the city administrators had control over their ability to comply with the courts’ rulings. It is true that the city frustratingly presented one compliance plan after another to the court and failed to follow through on many of them; some of them that they did follow up on, like the APP, proved to be fruitless. This string of abandoned plans apparently gave the courts the impression that the city simply lacked the will to implement plans the city itself had said would work. A more plausible interpretation of this history is that as the various plans turned out to be infeasible or ineffective, the city dropped them and in the end ended up simply not knowing what to do to bring themselves into compliance.
One alternative enforcement mechanism that was not tried by the court was the use of positive incentives. Fines and contempt charges are essentially negative incentives, and psychologists have generally found positive incentives to be more effective in changing behavior. It has been suggested that “[o]ne alternative to the contempt sanction . . . is the use of positive incentives. . . . One possibility might be to have administrators award organizational prerequisites and promotions partly on the basis of compliance with the decree by unit managers within the organization.”25 Interestingly, when the city faced the problem of getting private shelter providers to comply with performance contract specifications, it relied on positive incentives in the form of bonuses for compliance. Judge Freedman never tried this positive approach and so left one alternative to contempt unused.
Under Dinkins the direct transfer of hotel and other shelter families to permanent housing began apace. One reason for this development was that in May the City Council passed a law requiring that the city stop the use of barracks-style (Tier I) shelters by June 1992, so homeless families needed to be moved elsewhere quickly. The Dinkins administration assigned large numbers of apartments to homeless families during the spring and summer of 1990. Between March and August, an average of 561 shelter families per month, or 55 percent of the shelter population, were referred to city-provided permanent housing; the average for the previous six months had been only 262, or 36 percent.26
By some indications, the new strategy seemed to be working. As of July 1990, only 147 families remained in the infamous welfare hotels, compared with 3,306 families when Koch committed to ending the use of the hotels in 1988.27
But for whatever reason, the number of entries into the system went up in short order. Immediately after the Dinkins administration began increasing the provision of permanent housing in spring 1990, new entries to the shelter system jumped from about 530 per month to about 650 per month by July. Researchers from the Bureau of Management Information Systems, a division of the city’s Human Resources Administration (HRA), concluded that the number of entries into the shelter system “appears to be highly responsive to changes in housing and shelter policy. . . . It appears that ‘word gets out’ to the clients and community within months after a significant policy shift occurs, and that the perceived implications of the policy change affect the decision to enter the system or not.”28 Dinkins’s 1987 report had, of course, argued that no such effect would occur. When entries to the shelters suddenly started going up, Wackstein recalls, “this was something pretty wild, something we hadn’t contemplated.”29
The trend created a crisis for the city, which was legally bound by the court order issued in the case of McCain v. Koch, to supply every applicant with shelter. “We became simply obsessed with numbers,” Wackstein says. “Compliance was based on providing so many beds, so many units. There was no time for long-range planning at all.”30
By the early fall of 1990, according to some participants, Wackstein and deputy HRA commissioners Jeffrey Carples and Kenneth Murphy and First Deputy Mayor Norman Steisel were convinced that the greater supply of permanent housing had stimulated the increased demand for shelter. Deputy Mayor Cesar A. Perales agreed. Dinkins himself came to believe this. According to Steisel, “It took a long while for him [Dinkins] to accept that [a perverse incentive] was going on.”31
But in the end even Dinkins concluded at the time that “[u]nintentionally, an incentive for people to use welfare hotels has been created. . . . Families are frequently entering our city shelter system in order to find a permanent apartment. We do not want doubled-up families to make themselves homeless in order to find city housing.”32 In his memoirs, published twenty-three years later, Dinkins would be somewhat more equivocal but would strongly imply that he believed a perverse incentive had been created:
When it became clear that the homeless were indeed receiving priority and that the supply of higher-quality housing was increasing, the number of people entering the shelter system began to rise dramatically. Those who had been concerned about such an eventuality were proved correct. Were there people on the edge who decided to become homeless in order to gain a better deal? Did an increased availability of housing stock, the belief that one could obtain this housing by declaring homelessness, and a national economic downturn that threw people out of their jobs and homes all conspire to create a homeless epidemic? It is entirely possible.33
But HRA commissioner Barbara Sabol and private homeless advocates continued to insist that family homelessness was primarily caused by a shortage of affordable housing. According to Wackstein, those officials who argued that something had to be done to discourage inappropriate use of the shelter system were “roundly denounced” by advocates for the homeless.34
But had the Dinkins policy really created a perverse incentive? This episode, called the “Dinkins deluge” by Cragg and O’Flaherty, is one of the most studied sets of events in the history of New York City’s homelessness policy. The most extensive treatment is my own “Hard Lessons about Homelessness: The Education of David Dinkins,” in the City Journal in summer 1993. Another detailed account is “The Failure of Liberal Homeless Policy in the Koch and Dinkins Administrations,” by J. Phillip Thompson. The matter is also dealt with by Gordon Berlin and William McAllister in various forms and by Christopher Jencks.35
As some of their titles suggest, these authors all came to the conclusion that the policy of moving families rapidly out of shelters and hotels and into refurbished apartments or NYCHA housing created an incentive for poorly housed families to become homeless and so increased entries into the system. Cragg and O’Flaherty strikingly disagreed on the basis of an entirely different way of analyzing the situation. Dinkins administration officials who came to believe that a perverse incentive had been created based their conclusion on the historical record, personal experience, and gut instinct. As Norman Steisel put the matter, “I think there was evidence of it [a perverse incentive]. I wouldn’t say it was entirely hard fact-based evidence but there was enough of it to feel that that is what happened.”36 And Cesar A. Perales based his conviction on his experience with shelter clients:
I began, as I spent more and more time visiting Emergency Assistance Units, . . . to feel in having conversations with people that the way that people could improve their housing situation was to declare themselves homeless, go to the Emergency Assistance Unit and . . . to, in essence, skip to the front of the line and get into public housing or get some other form of housing for instance. And that began to trouble me because, obviously, that had not been my intent.37
Such was what Cragg and O’Flaherty described as the “conventional wisdom,” and they stated that “[w]e test the conventional wisdom and reject it. Better prospects of subsidized housing increase flows into the shelter system, but this effect is not nearly large enough to offset the first-order shelter effect—taking families out of the shelters reduces the number of families in them.”38 In support of this conclusion, the authors offered a sophisticated statistical model of the market for family shelter during the Dinkins administration.
A few qualifications are in order. First, we should note that Cragg and O’Flaherty found that perverse incentives had been created. The problem was not that such incentives didn’t exist; they just weren’t nearly as strong as had been thought, with seven placements into permanent housing being required to entice one family into the system. Cragg and O’Flaherty acknowledged that “the lure of subsidized housing does draw a substantial number of families into shelters—not as many as the journalistic accounts suggest, but still a substantial number.”39 Further, the authors found that one policy decision did have a strong perverse incentive effect. This was increasing the percentage of families that were lodged in Tier II shelters as opposed to welfare hotels. (Cragg and O’Flaherty defined Tier II shelters as follows: “Tier II shelters, generally operated by nonprofit organizations, offer more counseling and some other amenities than hotels do, but they significantly restrict residents’ freedom.”)40 Both Koch and Dinkins increased the use of Tier II shelters to reduce the number of families staying in the infamous welfare hotels. Cragg and O’Flaherty found that had the percentage of families in Tier II shelters not increased, the shelter population by December 1993 would have been only half as big as it was.41 And finally, the authors noted that their analysis did not imply that perverse incentives may not be more important in other situations.
The Dinkins deluge theory is a classic example of a staple of conservative rhetoric that A. O. Hirschman calls the perversity thesis, which is that “the attempt to push society in a certain direction will result in its moving all right, but in the opposite direction.” Anyone who follows political polemics is familiar with this move: the minimum wage increases unemployment; rent control reduces the supply of affordable housing; welfare breeds poverty. What then does the Dinkins deluge episode have to tell us about the use of the perversity thesis in policy discourse? The perversity thesis ought to be applied with great caution. It is neither mere common sense, as public officials tend to believe, nor a vicious lie, as advocates usually claim. It is most useful as a springboard to a discussion of the total set of incentives that policy creates. It is positively harmful when used as a quick put-down of reformers, or as an unanswerable charge against bureaucrats. We ought to be skeptical about casual use of this thesis. The perversity argument is often an effective rhetorical strategy for defenders of the status quo as it allows them to move off the defensive and criticize reformers or advocates. But precisely for this reason it often provokes a heated reaction and beclouds rather than encourages careful thinking. But neither should we simply reject the thesis out of hand.
When, in a policy dispute that we will consider in another chapter, the Bloomberg administration appealed to the perversity argument, advocates were quick to reject the very idea as absurd; Coalition for the Homeless claimed that Bloomberg was saying “Up Is Down, Night Is Day, 2+2=5, Etc. Etc.”42 This polarizing rhetoric was unnecessary, especially given that in this particular case the plain facts told against the Bloomberg administration, as we shall see. We have seen that city bureaucrats take the reality of perverse incentives very seriously, and it is never a good idea to just dismiss the experience and wisdom of public officials, even when they seem unscientific. And there is evidence that perverse incentives might in some situations be very real. For example, while there is no evidence that those housing subsidies such as Section 8 vouchers lure homeless families into the shelter system, there is evidence that under certain conditions vouchers do have perverse incentive effects. Economists Brian A. Jacob and Jens Ludwig have found that receipt of Section 8 vouchers discourages work. Section 8 families had 6 percent higher unemployment and 15 percent higher welfare receipt than similar families that did not receive the subsidy.43
We ought to encourage the development of what Hirschman calls a “democracy friendly” kind of discourse, one that realizes “there are dangers and risks in both action and inaction. The risks of both should be canvassed, assessed, and guarded against to the extent possible.” More often than conservatives realize, the perversity argument undermines the development of such a dialogue.
In any case, by the winter of 1990 much of the Dinkins administration concluded that they were facing perverse incentives of such a magnitude that they could not be ignored. Since the problem had supposedly been caused by an increase in the supply of permanent housing and a lessening of reliance on shelters, these trends, it was thought, had to be reversed. But the tactic the Koch administration had used—deterring demand for shelter by consciously making the system less attractive—was unpalatable to the Dinkins administration.
The result was the Alternative Pathways Program (APP), implemented in October 1990. Under APP, some of the HPD apartments that would otherwise go to shelter-system families were instead made available to families doubled up in private apartments. The city explained that the program would give doubled-up families an incentive to stay out of the shelters by convincing them that they had a chance of getting their own apartment. It was doubtful that APP was effective as a reward for doubled-up families to stay put. Because families enrolled in APP were assigned apartments by lottery, they could not be certain of getting their own apartment. By contrast, all families entering the shelter system were assured of an eventual transfer to permanent housing.
APP did, however, cut the number of HPD-produced units allotted to shelter-system families by 961 in 1991, thereby reducing the yearly supply of units available to those families to about its 1989 level. By allowing the city to cut back on the supply of permanent housing to shelter-system families, the program increased the amount of time such families would have to wait before being transferred to permanent housing. And family entries into the system did fall shortly after the implementation of APP.44 This led some observers, including the present author, to conclude that the reduction in transfers to permanent housing caused the drop in family entries. However, Cragg and O’Flaherty convincingly contest this aspect of the Dinkins deluge theory and find that APP, in their models, had no impact on family entries into the system. APP is interesting primarily as evidence that the Dinkins administration was hitting various policy buttons in an effort to adjust incentives for entry.
In September 1991, a discouraged Wackstein resigned. In a July 1992 interview with the New York Times, the former point person for homelessness policy in the Dinkins administration discussed her disillusionment with the efforts to move homeless families into permanent housing. “That was our big mission,” she said. “It was the mayor’s campaign issue. Boy, I believed it. I approached it with energy and passion.” Yet her efforts “didn’t seem to be making a difference.”45 She would later say,
I wanted to make the world better for poor people. I wanted to give them more options, more access to services, and housing and everything else and here I was . . . having meetings with HRA officials and OMB officials about what we’re going to do to keep them out. . . . I know this is not what I want to be doing. And that’s really why I left.
She decided she could do more for the homeless working in the private sector:
I was just perfectly delighted when I left, I was perfectly happy to go. I went to run a settlement house . . . Lenox Hill, and I was just thrilled to be dealing with . . . three hundred homeless people rather than thirty thousand homeless people because it felt like I could make more of a difference in running a direct service agency and doing decent programs and that these people didn’t try to deal with this unreal policy thing.46
The city’s policy of placing homeless families in public housing was originally a Koch initiative, but its results would challenge the Dinkins administration’s assumption that homelessness is primarily a problem of economics, rather than personal behavior. In 1988, as part of its response to the federal government’s threat to cut off funding for welfare hotels, Mayor Koch persuaded NYCHA to commit to a four-year plan offering two thousand apartments a year—one in four of the authority’s vacancies—to homeless families. NYCHA, which initially resisted Koch’s policy, had to reorganize itself to fulfill its new obligation.
In June 1988, the authority established the Special Rental Team (SRT), whose mission was to expedite the process of determining whether homeless families were eligible to live in public housing. Those who had untreated drug, psychological, or medical problems or were suspected of criminal or disruptive behavior were supposed to be screened out.
Eligibility determinations for nonhomeless families were usually conducted in Housing Authority offices just one day a week. The SRT, by contrast, conducted daily hearings at separate offices more conveniently located for homeless families, and in some cases would send staffers out to shelters to conduct interviews. Homeless families, therefore, got through the eligibility process in an average of two to three months, with perhaps one in three declared eligible immediately after the interview. Meanwhile, nonhomeless families faced an average six-month wait and were almost always subjected to further investigation before their eligibility was determined. Because the screening process for homeless families was thus less rigorous than for nonhomeless ones, some Housing Authority officials suspected that SRT was interpreting eligibility criteria rather loosely. SRT staff, however, hotly denied that they were failing to weed out dysfunctional homeless families.
In 1990, the authority further stepped up its efforts to place homeless families by establishing the Command Center, a centralized unit that aimed to speed up the process of matching eligible homeless families with vacant apartments. Normally, when a housing project had a vacancy, it would be reported up through the regular chain of command: the local project manager would inform the Housing Authority’s district office, and the district office would inform the central office, whose application department maintained the authority’s waiting list. Under the new regime, vacancies were first reported to the Command Center, which would immediately match a new vacancy with an eligible homeless family. Early on, the Command Center interpreted its mission as ensuring that the Housing Authority filled its yearly quota of apartments for the homeless before any vacant units were made available to nonhomeless families on the waiting list. Accordingly, the Command Center would “warehouse” vacancies; that is, if a homeless family were not immediately available for a given apartment, the Command Center waited until one became available, rather than release the apartment to a waiting nonhomeless family. Only if a homeless family declined a particular vacancy, which was its prerogative, would an apartment be offered to a waiting-list family. With the establishment of the Command Center, placement of homeless families picked up sharply. In all of fiscal 1990, 2,418 homeless families received apartments, but the majority of them—1,318, or 55 percent—were placed in the final quarter of that year alone (April through June 1990), just after the Command Center had been established.47
The mission of accommodating homeless families, however, conflicted with the Housing Authority’s institutional interests, as well as the interests of its tenants. The city’s public housing managers had held “an article of faith for many years that the New York City Housing Authority maintains its reputation as the very best of the large public housing authorities in the nation because of its deliberate policy of an ‘economic mix,’” as Sally Hernandez-Piñero, who became the authority’s chair in February 1992, explained.48 In order to maintain social stability, the authority tried to balance the population in its roughly 179,000 apartments at about one-third elderly, one-third working poor, and one-third very low income (families on public assistance). But between 1985 and 1991, the mix of entrants into the authority changed drastically. Very low-income families increased from 32 to 54 percent of total yearly entrants, while the elderly decreased from 49 to 35 percent and the working poor dropped from 19 to 12 percent. Housing Authority officials came to see the homeless referrals as a threat to their longstanding policy of “economic mix.”49
Public-housing tenants resented the preferential treatment accorded to homeless families. In 1990, Housing Authority projects were home to some one hundred thousand members of families doubled up in Housing Authority units, in addition to their 470,000 authorized residents. These doubled-up residents were eager to apply for public housing vacancies but found that none would be available until SRT had filled its quota for homeless referrals. Housing Authority tenants did not take kindly to seeing their own “internal homeless” pushed to the back of the line.
By 1990, tenants had become so frustrated that they held public demonstrations protesting the expedited placement of homeless families in public housing. City Hall dispatched First Deputy Mayor Norman Steisel and HRA commissioner Barbara Sabol to soothe tempers at a public meeting with Housing Authority tenants. Steisel later reflected that
[t]here was even a point in time where Barbara Sabol advocated very strongly to . . . put people into the Housing Authority. It creates all sorts of problems. The Housing Authority, for whatever they are, they have their own sets of problems. They are pretty stable communities and it was clear that there was information available that happened to show that the homeless people got other kinds of support. It made problems in the community, both real and perceived, so that many of the Housing Authority [families] . . . resisted. . . . There was high resistance to it and some with good cause.50
Steisel and Sabol promised to transfer some one million dollars from HRA to the authority in order to cover costs associated with integrating the formerly homeless families. The situation was helped not at all when these promised funds were eventually cut from the HRA budget and therefore were never transferred to the Housing Authority.
It was only a matter of time before Housing Authority officials would side with their tenants, partially as a means of strengthening their own hand in the inter-bureaucratic struggles over responsibility for the homeless. “I can’t have any moral authority if I have to say to our people that they have to do something that I know is undermining our housing,” explained NYCHA chair Hernandez-Piñero. “That just goes to the core of the relationship I’m trying to develop with the tenants. The stories of some of the people on our waiting lists would break your heart too.”51
The resistance to homeless families among public-housing tenants probably came as a surprise to the Dinkins administration. A Shelter Is Not a Home had attacked “Myth 13: No community wants housing for the homeless in its own backyard. Fact: . . . Communities have accepted housing for the homeless when the facilities are small-scale and well-supervised by an accountable non-profit organization, priority is given to homeless people from the same area, and advance planning includes local input.”52 Public-housing tenants certainly would have been more receptive to the policy if it had given priority to “homeless people from the same area”—that is, to doubled-up families in public housing. But this is as much as to say that the Housing Authority should not have accepted outside homeless families in the first place, since its own internal homeless could have filled most, if not all, of the available vacancies.
Even if the homeless families referred to public housing had possessed the social skills of diplomats, regular tenants would probably have resented a policy that put them at a disadvantage in getting vacant apartments. But were the tenants’ doubts about the functioning capabilities of the homeless justified, or were they, as Dinkins’s task force had argued in 1987, the result of “stereotypes of homeless people which have aroused unfounded fear and prejudice”?53 In September 1989, New York University’s Health Research Program published a scientific study that could be interpreted as supporting the tenants’ view.54 The control-group study compared 704 homeless families requesting shelter at one of the city’s Emergency Assistance Units with 542 nonhomeless families on public assistance interviewed at welfare offices.
The findings of the study contradicted several of the assertions made in Dinkins’s 1987 report. For one thing, homeless families were in fact demographically different from other public-assistance clients. They were more likely to be black, less likely to be Hispanic, and much younger on average than nonhomeless poor families. Further, homeless families have consistently higher rates of behavioral problems than their housed counterparts. “On almost every measure of ‘well-offness’ or ability to function independently,” the authors remarked, “homeless families are in somewhat worse shape than the housed public assistance population.”55
It is true that reported substance-abuse and mental health problems involved only a minority of homeless families. One should not conclude, however, that the higher reported rates of behavioral problems among homeless families are too small for such problems to play a role in causing homelessness. A combination of such “small” differentials can have a powerful impact. “Although the differences in the two groups are not dramatic for many of the individual measures, the cumulative effects of the higher incidence of numerous problems may be the major cause of many families’ homelessness,” the study noted.56
However, follow-up research to the initial Knickman and Weitzman study suggested that the differences between homeless and nonhomeless poor families are less dramatic than their initial work indicated. From 1990 to1996, the National Institute of Mental Health supported a long-term follow-up study, in which 70 percent of those families in the initial study who had not used shelter prior to the 1988 interview were re-interviewed. One study based on these new interviews found that the problems of homeless families did not prevent them from achieving housing stability once they received some form of housing subsidy.57 Another follow-up study found that, five years after entering a shelter, formerly homeless children and continuously housed children exhibited few differences, although such differences as there were favored the continuously housed children.58 It has also been found that there are few differences in the social networks of homeless and housed families.59 This follow-up research suggests that however risky homeless families might have seemed at first, they would have successfully stabilized themselves once they got access to NYCHA apartments. But in 1990 Housing Authority families and staff seemed to have reason to worry, and they resisted accordingly.
Therefore the Housing Authority reduced its efforts to place homeless families. In the summer of 1991, the Command Center abandoned its practice of “warehousing” vacant apartments for the homeless. In June 1992, Hernandez-Piñero abolished the SRT. Homeless families applying for vacancies in public housing were once again subject to the same eligibility review as everyone else. And Hernandez-Piñero renegotiated the 1988 agreement, so that the authority was obliged to accept only about fourteen hundred homeless families a year.
Even before he became mayor, Dinkins’s interest in the homeless had especially focused on homeless families. Thus the document that would become the guide to administration homeless policy, A Shelter Is Not a Home, focused exclusively on homeless families. Looking back from 1992, the Cuomo Commission would hold that homeless single adults had been treated like “a forgotten population.”60 Wackstein confirmed this judgment by noting that “in our administration, like in many others, subsequent and before, the family issue tended to overwhelm everything else.”61
But, as mayor, Dinkins would have to address that ongoing problem as well. For singles, his administration had relied, as Koch’s had, on former armories that had been converted to mass barracks-style shelters, with perhaps hundreds of beds filling up the buildings’ large open spaces.
Converted armories provided protection from the elements but were otherwise not satisfactory. One researcher provided a vivid picture of the huge Fort Washington Armory as it was in the early nineties:
The low-ceilinged donut-shaped first-floor corridor that ran the length of the block was crowded with African American and Latino men in every state of dress and undress. . . . In the distance rats scurried under ancient standing metal radiators in the empty eastern end of the long hallway.
A dark room off the hall revealed dozens of men dozing under the faint blue light radiating from an aging nineteen-inch color television mounted high up in a corner. Past the TV room, the hallway curved around to the other side of the building where ropes marked off the cafeteria line and its conflicts over stale bologna sandwiches and mealy apples. Even the air was different in “the Fort.” Industrial disinfectant mingled with the smell of human sweat, cafeteria food and residual radiator moisture. Not unbearable, the smell was unpleasant and just nauseating enough to put one a little on edge. . . .
Ascending the broad open staircase at the end of the first floor corridor, the third-world bazaar of used small electronics, hair cutting, drug deals, and other petty commerce gave way to the vast drill floor where nearly seven hundred beds stretched out in long rows spaced two feet apart. The wooden floor was larger than a football field, and was enclosed high above by a cavernous sloping, domed ceiling. This main room was like something between an indoor sports arena and one of the orphans’ work houses in Oliver Twist. Residents were forbidden to go up into the bleachers that ringed the room about twenty-five feet above the floor, where the director of the shelter had his office.
At one end of the drill floor was a giant steel shutter door that was raised electronically every morning, including in winter, “in order to air this smelly place out,” as shelter workers put it. This was also seen as a way of getting residents out of bed and making it too cold to comfortably stay there during the day. . . .
The men near the center of the drill floor, who continued to sleep through the noise, cold air, and movement of a late winter mid-morning, lay on metal cots with spindly iron legs stuck into tennis shoes, oxfords, boots, and all variety of other footwear. . . . Despite the common precaution of putting the legs of one’s bed into shoes and tying the shoes to bed frames with tight knots, some shelter residents told me that they had actually lost shoes to thieves who lifted up their bed while they slept.62
A description from the Times in 1992 confirms the above account of the Fort Washington Armory shelter:
After the lights go down at night at the Fort Washington Armory in upper Manhattan, paranoid schizophrenics lie nervously next to ex-convicts they rightfully fear. Noises arise in the darkness: the moans of men having sex with men, the cries of the helpless being robbed, the hacking coughs of the sick, the pounding of feet running through a maze of 700 cots packed into one vast room.63
Wackstein’s deputy director at the Mayor’s Office was Anne Teicher, who was in charge of policy for singles. Teicher’s focus became eliminating the armory shelters and replacing them with many smaller shelters dispersed throughout the city. She described the administration’s singles policy as follows:
So actually the policy was pretty progressive. It just crashed and burned. The concept was to create smaller [shelters], to get away from these huge armory-style shelters. At that time there was still the Fort Washington Armory and these huge shelters. . . . I think there were at that time still hundreds of beds. But to create these smaller shelters around the city that would be more like transitional housing. Because at that time it was still like three hots and a cot theory. There weren’t a lot of . . . on-site services to really help them while they were in the shelter. So the concept was to create program-rich, smaller shelters. That was the intention.64
A City Council law required the administration to come up with a five-year plan for sheltering the homeless and so the focus of that plan became developing a blueprint for building a network of smaller, program-providing shelters that would allow the large armory shelters to be closed. By late 1990, work on the scheme had begun, which would be released in October 1991 as the New York City Five-Year Plan for Housing and Assisting Homeless Adults, and which addressed policy for single persons exclusively.
The plan had two main thrusts. The first, “primary,” goal was “to transform the service system for homeless adults from one that merely provides shelter to one that provides homeless people with the training, rehabilitation and other services that would enable them to participate as fully and independently as possible in the life of the city.”65 The idea that homeless individuals needed “rehabilitation” was an important development and reflected the belief that Wackstein and others in the administration had come to that homelessness frequently was the result of some disability or “underlying cause” that policy had to address. Wackstein described the crystallization of this controversial idea as follows:
There was a belief that I had and that others had that the people who ended up in the shelter system needed things other than shelter. And that . . . engendered some controversy with groups like Housing Works, which was an advocacy group for people with AIDS, and the Coalition for the Homeless, because there was a real ideological rift, saying that homeless people, all they needed was a home, all they needed was a place to stay. And I did not agree and I still don’t agree. . . . People, whether it’s families or individuals, they don’t end up in the shelter system, generally, unless there are other needs that need to be addressed.66
Wackstein’s belief that the homeless had other problems beside a lack of housing proved so unpopular with Housing Works that the group pasted posters on Wackstein’s block that showed her photo with an “X” through it. This notion that homelessness was usually rooted in some underlying disability was the beginning of the paternalistic understanding of homelessness, and would be picked up and developed in detail by the Cuomo Commission, as we shall see.
The five-year plan described its other thrust, which would have more infelicitous consequences, as follows: “The cornerstone of this Plan is a five-year, $200 million development program to create 2,500 new beds in model transitional housing facilities to replace an equal number of beds in existing shelters.”67 To accomplish this goal, the city committed itself to a process of building about twenty-four shelters throughout the city. The process of finding appropriate sites, according to the plan, was to involve “Early and Meaningful Community Involvement,” “Fair Distribution of New Facilities around the City,” and “Undertaking a Comprehensive Approach to Siting” (italics in the original).68 For the first time the Fair Share provision of the city’s new charter, which called for a fair distribution of “burdens and benefits associated with city facilities,” would be implemented to ensure the distributive justice of the process. The New York Times described the planning process:
Meet the 1990’s incarnation of the political machine: an I.B.M. 3090 computer. It was enlisted by the city to help formulate its proposed dispersal of the homeless from barracks-style dormitories to two dozen smaller shelters scattered in every borough. . . .
The computer spewed out 490 locations after exempting 20 community districts deemed to be oversaturated. Its human handlers ruled out 252 and visited the remaining 238, eventually culling 35—a third of them in commercial or industrial settings—from which 24 are to be selected.
From the beginning, the Dinkins administration sought to enlist other elected officials in the process and is doing so again. Except for Ruth W. Messinger, the Manhattan Borough President, many preferred the political cover of the computer.69
In short, the plan was to be everything such processes are supposed to be: rational, comprehensive, transparent, consultative, and fair. The administration was optimistic; A Shelter Is Not a Home had rejected the “myth” that “[n]o community wants housing for the homeless in its own backyard” and asserted that “[c]ommunities have accepted housing for the homeless when the facilities are small-scale and well-supervised by an accountable non-profit organization, priority is given to homeless people from the same area and advance planning includes local input.”70
But when the five-year plan was released and a ninety-day public comment period began, things didn’t go as planned. Things got off on the wrong foot when a copy of the plan was leaked to a Staten Island newspaper, which reported that shelters were to be located in that borough. Anne Teicher reported that “[w]hen we went to a meeting in Staten Island I was afraid. They were horrible. I was afraid I was going to come out and find flat tires.” Things got no better in the rest of the city, Teicher found:
After the plan came out we went to a lot of community meetings. . . . They were brutal. They were really brutal. It was really painful. Because of the way things came out. It just hit the fan. We were going to put these shelters all over the city. They were going to be horrible; you know what shelters are like, blah, blah, blah. And we don’t want these people. And NIMBY. It just activated the NIMBY response all over. . . .
We knew that that was a major issue. There’s always opposition to siting anything. I think we might have been a little naïve. We were trying to do it comprehensively. Instead of dribbling out these sites over a period of time, we would identify all the sites at one time and we could then show the Fair Share aspects of it. And it was very rational. But it was a little naïve that you could do this and not have this uprising of NIMBYism.71
Wackstein had the same experience:
I probably still have the scars on my back from going to Community Board meetings. They didn’t want homeless family facilities but even more they didn’t want facilities for single adults. Because there was this very widespread perception that they were either addicts, crazy, or like people with AIDS. And remember this was back in the late eighties, early nineties, everybody was up in arms about people with AIDS in their communities. So there was that sense that these were people who were going to bring disruption to neighborhoods. So it was very hard to get anything sited.72
Almost no one outside the administration stepped forward to defend the siting process. City council speaker Peter F. Vallone summed up the general public reaction when he said, “You’re unnecessarily frightening all the people of the city, and it doesn’t make any sense. . . . It’s spreading homelessness and hopelessness throughout. It’s almost as if you’re saying we have a serious disease and we’ll spread it so everybody will suffer from it.”73
Clearly, the high-profile planning process had served not to defuse opposition with its obvious fairness but to offer unpersuaded community leaders a target for their fire. A crucial mistake had been the decision to have the proposed shelters, which were to be managed by nonprofits, actually built by the city. With the city in charge of the building process, a very visible, centralized siting plan seemed to the Dinkins administration to be the only way to avoid the “unfairness” of the Koch administration’s nonconsultative practices. But visibility resulted in vulnerability without viability. Much more workable would have been a plan to have nonprofits not only manage but build the shelters. Then individual nonprofits could have, piecemeal, entered into individual negotiations with given neighborhoods, which would have lowered the profile of the process and avoided the concentrated fire of the whole city at once. This was the approach that was in fact ultimately followed at the advice of the outside advisor, Andrew Cuomo, whom the Dinkins administration brought in.
Early on, key members of the Dinkins administration concluded that their approach to homelessness was not working. As a response, in September 1991, Mayor Dinkins farmed out the job of rethinking homeless policy to a Commission on the Homeless, headed by Andrew Cuomo, the governor’s son and eventually governor himself. In February 1992, the commission released its report, The Way Home: A New Direction in Social Policy, which challenged the assumptions that had guided the early Dinkins policy.74
The commission endorsed the Dinkins deluge hypothesis and argued that placing thousands of homeless families into permanent housing contributed to an enormous surge of families entering the shelter system:
After inheriting a system with approximately 1,500 families in hotels at the end of 1989, Mayor Dinkins was able to dramatically reduce the number of families in hotels to about 150 in August 1990. Placing thousands of homeless families, many of whom had only recently entered the shelter system, into permanent housing appears to have contributed to an enormous surge of families entering the system in the later part of 1990. By 1991, the City was forced to return, to a lesser but still substantial degree, to hotels. Today, there are almost 1,000 homeless families. It is unfortunate that hotels are once again a significant source of emergency shelter for the City’s homeless.75
The Way Home also claimed that behavioral problems were an important cause of homelessness. It conducted its own extensive survey of homeless individuals and families in the city:
The results were enlightening. The commission found that homelessness is frequently a symptom of some underlying problem, such as lack of job skills or education, a substance-abuse problem, or mental illness. It results when one or more of these problems interact with a number of social and economic factors, including a shortage of affordable housing.
The commission’s findings were striking: 42 percent of the families surveyed seemed to have either a mental health or drug problem, and no less than 29 percent of adults in shelter families tested positive for illegal drugs.76
The commission went even further, arguing that the sociological reality of homelessness in New York was incompatible with a policy of unconditional right to shelter. The advocacy groups that had won recognition in various ways of a right to shelter pursued such an approach in the belief, shared by Dinkins in his 1987 report, that homelessness was primarily an economic problem. The “basic thrust” of this strategy, according to the Community Service Society, was “to shift the focus from rehabilitation, the attempt to return the disabled to normalcy, to welfare.” Hard experience forced the Cuomo Commission to insist again that shelter regimes had to emphasize rehabilitation at least as much as welfare or shelter:
The disabilities suffered by homeless single adults have been ignored for too long by administrations as well as some advocates. It is not enough to provide only shelter to these vulnerable and troubled individuals. Many of them need significant social services and medical attention. . . . Society’s honest acknowledgment of the problem is the first step toward a solution.77
The commission articulated what might be called the “underlying problem” approach to homelessness, which focused on determining and addressing the reasons why a particular person or family is homeless: “Homelessness is often a symptom of one or more underlying problems for which the answer is not simply access to housing, but also access to social and related services, including job training and assistance, independent living programs, substance abuse treatment, and mental health care.”78
Another crucial conclusion of the commission was the need for what might be called the “mutual obligation” approach to providing shelter. Potential clients had some kind of an underlying problem that contributed to their homelessness. Therefore they had to address that problem as a condition of receiving aid. This was especially important because not every request for aid could be answered. “Government should not, and cannot, be expected to provide housing to everyone who asks for it,” the commission stated. “A determination of need must be made so that government’s limited resources can be targeted to the most needy.”79 Those in need had to address the reason why they were in need. Thus the homeless had a responsibility as well as a right. The Way Home put the matter as follows: “The emergency shelter system must incorporate a balance of rights and responsibilities. A social contract and a mutuality of obligation must exist between those receiving help and society-at-large.”80
Eventually the Giuliani administration would give administrative substance to this call for “mutuality of obligation” by developing a shelter system that, insofar as was consistent with the city’s legal obligations, would be structured to strongly nudge clients to enroll in rehabilitative services in return for being provided with shelter. The general call for mutual obligation was thus the germ of what would be an enforced quid pro quo: shelter in return for compliance.
In other words, the Cuomo Commission articulated for the first time in detail the paternalistic paradigm of homeless policy. The two key assumptions of this philosophy would be, first, the belief that there was “something wrong” with homeless people, an “underlying cause” that was crucial to their being homeless. The second assumption was that this underlying cause must be addressed through a quid pro quo of shelter for self-help in the form of enrollment in some form of rehabilitation. Other policy recommendations of the Cuomo Commission included the creation of a new “small, entrepreneurially styled” city agency that would coordinate efforts to deal with homelessness; provision of rent vouchers to homeless families; increased reliance on nonprofit organizations, rather than city agencies, to provide services to the homeless; and establishment of requirements that homeless families undergo screening and treatment before being assigned to permanent housing.
The Dinkins administration’s response was ambivalent. When the report was released in February 1992, the mayor would only say, “I’m fully committed to examining whether we can go forward.”81 Perales and Sabol criticized the proposal to develop a new agency.82 As the Times noted, “The controversy left the Mayor in the awkward position of deciding whether to take the advice of his staff or the commission he appointed in September.”83 Dinkins quickly created another panel to explore the practicality and political viability of the commission’s proposals.
Several factors may explain the initial apparent lack of official enthusiasm for the commission’s recommendations. In objecting to the proposed new agency, for example, Sabol may simply have been trying to protect her own bureaucratic turf. The same may have been true of HPD commissioner Felice Michetti, who questioned the use of rent subsidies. It is also possible that some of the commission’s positions—such as the claims that perverse incentives to enter the system had been created, that dysfunctionality was widespread among the homeless, and that homelessness was more than a housing problem—may have been difficult for liberals in the administration to absorb.
Nonetheless, by September 1992, Dinkins embraced the commission’s recommendations. Besides possible ideological problems, the administration’s initial coolness towards implementing the Cuomo Commission’s recommendations may have reflected an appreciation of the practical problems that remained to be resolved. That task was thrust upon Charles, or Chip, Raymond, who would eventually become the first Department of Homeless Services commissioner, but who was before that appointed to a small entity within the mayor’s office to study the commission’s recommendations and create a detailed implementation plan. That office was called the Mayor’s Office of Homeless Facilities and Service Development. This represented, at last, the clear decision to go ahead with the new agency. On December 2, 1992, Raymond became director of that office.84 One of his first moves was to bring in as his assistant chief of staff Muzzy Rosenblatt from the Mayor’s Office of Construction. Rosenblatt remembered some of the issues that remained to be worked out:
Does a new municipal agency need to have, because of its new mandate, new functions, that didn’t exist as part of HRA? For example, its procurement activity as designed was going to be and became much more robust than the procurement activity that was going on at HRA, [which] wasn’t in the business of outsource. So how do we . . . embrace the notion of privatization, [of] outsourcing everyone? How do we do it? The notion was going to be we were going to take [a] municipally run system with municipal employees and contract out to private entities, not for profits. What happens to the workers? Now the Cuomo Commission didn’t answer that question. It didn’t ask that question. And so that question needed to be asked and needed to be answered and done in a way that was thoughtful. . . .
So, that was a major issue. Another issue was the budget. Was this agency going to wind up spending more . . . without getting anything more? A big mantra at the time . . . that the editorial boards began to embrace was cost mentality. Can’t cost any more or less. Now when [you’re] decentralized, you create certain diseconomies. You may get greater effectiveness, you may get greater results, but the cost of achieving those results may also be greater even though the cost benefit may be positive. . . . But the City Council said, no more—we cannot cost more. . . . So it was much easier said than done to create a new agency. In the end we did it, but those are some of the major issues at the time.85
Among Raymond’s first tasks was drafting legislation to create his department and then guiding the bill through the City Council. But rather than the streamlined, entrepreneurial agency suggested by the Cuomo report, the proposed Department of Homeless Services had a staff of nearly three thousand, most of whom transferred from HRA. Legislative approval was necessary to establish the new department, but by mid-May 1993, the proposal was stymied in the City Council. Speaker Peter Vallone, who had enthusiastically endorsed the Cuomo Commission recommendations when they were released, objected to the size of the proposed agency and argued that the mayor had not specified how the revamped shelter system would deal with such matters as drug treatment, mental health care, and job training. Raymond testified before the council that the administration had been so occupied with the logistics of establishing the agency that it had not had time to “present that kind of gigantic plan.”86 The City Council would finally pass the legislation creating DHS on July 1, 1993. However, in order to retain City Council power over the department, City Council speaker Peter Vallone insisted on having the power to vote it out of existence after two years.87 The legislation also stipulated that the department would have to cease the use of welfare hotels in order to be continued.
Raymond was in some ways an unusual choice to be the first commissioner of the new Department of Homeless Services (DHS). Most observers were struck, not always positively, by the fact that during the 1980s he had managed ballet companies, including the New York City Ballet and the Joffrey Ballet. Asked how his experience with managing ballet companies qualified him to run the shelter system, Raymond replied that “he wasn’t responsible for what went on the stage, but that the lights went on. . . . What we don’t have now is somebody selling the vision of what goes on the stage.”88 But Raymond also had extensive managerial experience that was less unusual. He had worked in the Lindsay administration for four years, first in the Human Resources Administration, then as assistant commissioner of the Department of Sanitation, and later for a task force on SRO housing. He had also worked as deputy commissioner of the Department of Mental Health, Retardation, and Alcoholism, and during the Koch administration, as deputy commissioner for housing, he had been in charge of the city’s extensive holdings of in rem apartments. He had even been a part, early in his career, of a consulting company that had been responsible for implementing various reforms at the infamous Willowbrook institution. In that effort Raymond worked with Norman Steisel, who would become Dinkins’s first deputy mayor and was a strong supporter of the Cuomo Commission’s recommendations.
Raymond described his career as follows:
In my career it’s very easy to pick out a few things that are disconnected, but the thing I always try to do is carry, you know, a kit bag of very good management tools. I rely very heavily on other people. I don’t have to be expert in any field because I’d feel that you could find people who are experts. And learning about, you know, homelessness, and learning about in rem housing, and learning about mental health, that’s the easy part. The tough part is managing a system and that’s what I felt I was pretty good at.89
In short, Raymond was a manager, not an advocate. That was his appeal. His appointment represented a victory for Steisel, who was skeptical of the call of the advocates for more resources, and who felt that better management could make a real difference. Steisel had clashed with Sabol at HRA who, like the housing and homeless advocates, called for an increase in the housing allowance of the state welfare grant and who was skeptical of the value of a new agency. According to Steisel,
HRA . . . didn’t have the managerial talent or focus to drive through some of these questions about expanding or improving the facilities. . . . So when we created the agency [DHS], it’s the reason we hired Chip. He wasn’t out of that world. He wasn’t a social service advocate or activist. . . . He was a real generalist. . . . He was a very skillful manager and the idea was we wanted to give the impression to the world as well as to our agencies that we were going to work with them, we now had someone in there who was a manager. . . . So that was the shift in emphasis we were trying to demonstrate by bringing him in.90
The eventual embrace of the Cuomo Commission’s recommendations, the creation of DHS, and the appointment of Steisel’s “protégé,”91 Raymond, all represented the at least temporary victory in New York City of what may be called the “managerial” as opposed to the “structural” approach to the homelessness problem. In his classic article entitled “The Quandaries of Shelter Reform: An Appraisal of Efforts to ‘Manage’ Homelessness,” which was contemporaneous with The Way Home, Dennis P. Culhane argued that
the ability of shelters to serve as homeless management agencies is constrained by the structural causes of the homeless problem. Without renewed commitments from public agencies responsible for the treatment of mental illness and substance abuse, for reintegrating recently paroled persons in the community and for promoting child welfare, affordable housing, and income maintenance, the shelter system will remain overburdened and unmanageable.92
New York City was in effect embracing the idea that shelters could be managed effectively, and make a substantial contribution to solving the homelessness problem, even in a situation that did not promise much immediate improvement in the social welfare system. It was an approach that was to dominate city policy through the Giuliani and Bloomberg years.
By his own account, Raymond was never very successful at what he identified as “the department’s biggest problem . . . through-put, getting people in and out the back door.” Throughout his tenure, the EAU continued to be terribly overcrowded, with so many families sleeping in chairs and on the floors that “there were periods of time when you could hardly get in the door.” “I think that my biggest disappointment was that we couldn’t . . . solve the problem and the problem to me was always bringing people in and getting people out and at that we were never terribly successful.”93 Neither did Judge Freedman find Raymond’s efforts satisfactory. On December 8, 1993, two days before his resignation, Freedman imposed contempt findings on Raymond because “he was the official whose decisions about emergency shelter placements from EAU’s most flagrantly violated Court orders.”94
To be fair to Raymond, he had, from the very first day of his appointment, expressed doubt that he could relieve crowding at the EAU. “I’m not sure that in reality we will be able to reach the point where we won’t have people sleeping in EAU’s,” he said on being appointed. “I’m not sure we will. I just don’t think the system can deal with it.”95 And indeed, the city’s first efforts to manage homelessness were not successful.
Contemporaneous with the publication of The Way Home was a development that would eventually, under the Bloomberg administration, usher in a new era in homelessness policy, one that would challenge the paternalist paradigm that was first completely articulated by the Cuomo Commission. This was the founding by Sam Tsemberis in 1992 of Pathways to Housing, an organization that would revolutionize outreach to street dwellers by implementing the Housing First approach. Pathways to Housing and Housing First grew out of Tsemberis’s increasing frustration at the city’s Health and Hospitals Corporation, where he ran the outreach program Project HELP, infamous for its ultimately frustrated effort to involuntarily transfer Joyce Brown, aka Billie Boggs, and other mentally ill homeless people, off the streets and into the hospital psychiatric wards. Tsemberis described what he saw as the limited results of this approach:
We had realized at that time—this is like four years into the Project Help experience—that taking people to the hospital, even if they may have been a danger to themselves or others, was not resulting in any good outcomes, other than to get them out of the immediate danger for a short time. They kept returning to the street, a vicious cycle of making this terrible anguished decision about do we need to hospitalize this person, should we take them now or not. And then after a lot of debate we would take them. Three weeks later or four they were back out on the street. . . . I was desperate to try something new.96
The approach Tsemberis and his collaborators—who included psychologists William Anthony, David Shern, and Mikal Cohen—decided to apply was psychiatric rehabilitation, or psych rehab, whose “major focus . . . is on achievement of individually defined goals.”97 William Anthony, who had developed the psych rehab approach at Boston University, described it as “an intervention that essentially . . . helped people figure out what they wanted, then helped them figure out what they needed to have or do to get what they wanted, and then teach them the skills and provide them the supports to get what they wanted.”98 In the context of Project HELP’s street outreach work, implementing the psychiatric rehabilitation model implied that instead of being told that they would be transported to psychiatric services, street dwellers would be asked what they wanted and the outreach team’s job would be to accomplish that request. To implement and evaluate this approach among Project HELP clients, Tsemberis and Shern received a grant in 1990 from the National Institutes of Mental Health entitled “Taking Psych Rehab to the Streets.” The insistence on client choice in the psych rehab model was so radical that Tsemberis himself had trouble sticking to it consistently, although in the end that commitment prevailed. Tsemberis recalled,
We had a big debate when we got that psych rehab grant. Bill Anthony and Mikal Cohen were like, no you have to ask people. I thought like, we just take people off the street and bring them into housing. That was the whole point of doing something different. I sort of knew they wanted housing. And [Anthony and Cohen] said, no, you have to ask them what they want. And if it’s not housing, then you’ll give them what they want. If it’s haircuts, that’s what you’re going to be doing. [I said] I can’t believe that we went through all this trouble and we’re giving people haircuts! And I had to appreciate the importance of the principle of what choice really means. Choice really means you’re offering support for whatever it is the client needs.99
The results of the “Taking Psych Rehab to the Streets” experiment were striking. When Tsemberis and company compared the outcomes for clients who received the experimental, psychiatric rehabilitation outreach to the outcomes of clients who received the standard Project HELP approach, they found that
[c]ompared with persons in standard treatment (n=77), members of the experimental group were more likely to attend a day program (53% vs 27%), had less difficulty in meeting their basic needs, spent less time on the streets (55% vs 28% reduction) and spent more time in community housing [as opposed to shelters and the streets]. They showed greater improvement in life satisfaction and experienced a greater reduction in psychiatric symptoms.100
But what did street dwellers want when they were asked? According to Tsemberis,
People were quite articulate. . . . People knew exactly what they wanted and the terms that they wanted it on. Everybody wanted a place to live first and then maybe they would consider treatment. Even though we were finding that using psych rehab we could engage people much more quickly, we had a much more therapeutic alliance, we were able to meet their demands in the sequence they wanted them, like food or shelter or like that, but they, what they wanted was a place to live and the program was really an outreach program, not a housing program, so we couldn’t deliver on what they most urgently needed. And I could not persuade the Health and Hospitals Corporation to start a housing program. . . . out of Bellevue. . . . They were like, we’re a hospital, we don’t do housing, it’s not our mission, it’s not our mandate. So I left the Health and Hospitals Corporation in 1992 and I started . . . Pathways to Housing.101
According to the homeless mentally ill themselves, the key to getting them off the streets was to offer them housing first, and then, possibly, treatment for their illness, rather than forcing them into treatment first with the expectation that treatment of the underlying cause of their homelessness would eventually make them fit for housing. What was lacking in the psych rehab model as Tsemberis was implementing it at Project HELP was the offer of permanent, independent housing that did not require sobriety or compliance with treatment. Such then were the prerequisites for moving from a shelter or similar community residence to supported housing, that is, a permanent apartment where psychiatric and other services were provided. At the inauguration of his not-for-profit Pathways to Housing, Tsemberis received a five hundred thousand dollar grant from the New York State Office of Mental Health to add this housing component to his outreach efforts. Tsemberis described the results:
[Other] contracts for supported housing were already in place, we didn’t invent that. All we had to do was serve fifty people and provide services for them. Everybody else in the city was basically doing this kind of a program but they were doing it as a graduation out of community residences.
So, if you lived in a community residence for a long time, were compliant, and were clean and sober, you’d be put into an apartment usually shared by one or two other people. We took that same contract and gave an independent apartment to a person, no treatment, no sobriety. Just like they wanted. . . . and they wanted the housing first, that’s what people wanted. And then we did the case management support after they were housed.
. . . We started around November ‘92 and so at the end of ‘93 we reported to the Board of Directors at Pathways. . . . So, our report after that first year was 84 percent of the people we had taken from the streets—these Project HELP people—and brought in to apartments, delusions, addiction, and all, were still housed after that first year.
And we thought, “Oh, my God, we’re on to something new.” Who knew that people were so capable? Who knew that these people who seemed so disorganized, actually, when you think about it, they had such incredible skills on the street. They knew where to eat, they knew where it’s safe to sleep, they knew where soup kitchens are open or not, they knew how to not get arrested, and they knew who to trust with their stuff. All that stuff. All these skills, not really visible, initially, clearly manifest once the person was housed. Living a life in an apartment was a piece of cake compared to living a life on the street. They did spectacularly well and the program has been growing ever since.102
Thus was born Housing First, which in time would develop into a paradigm that would begin to displace the paternalistic philosophy developed in The Way Home. But first the immediate future would see the city develop the paternalistic shelter system that the Cuomo Commission was calling for.
Dinkins’s slow grasp of the Cuomo Commission’s agenda created an opening for Rudolph Giuliani in the mayoral election of 1993. Owing to some complicated sparring over exactly who should be included in the mayoral debates of 1993, it ended up that the only forum in which Dinkins and Giuliani confronted each other directly was one sponsored by the Citizens Housing and Planning Council and devoted to housing issues. Giuliani thus had maximum opportunity to step into the breach. The New York Times reported that
Mr. Giuliani, who had avoided in-depth discussions of these issues until last night, mocked the Democratic Mayor for naming a panel to study the study of his own commission on homelessness, rather than moving immediately to carry out its recommendations. “Give me a break—make a decision,” said Mr. Giuliani, the Republican-Liberal candidate for mayor.103
Giuliani, of course, won that election. How important his campaign promises to reshape homelessness policy were in his victory is not known. Some of his specific proposals—such as the unrealistic suggestion of imposing a ninety-day limit on the length of shelter stays—may even have hurt him, although probably more by mobilizing the homelessness-policy community against him than by their lack of popular appeal. Giuliani appointed Joan Malin, a former Dinkins administration official, to head the new Department of Homeless Services. How Giuliani reshaped city homelessness policy is the subject of the next chapter.
Mayor Dinkins and some of his top administrators deserved considerable credit for facing up to the truth that they had pursued a wrongheaded policy. They accepted what became the conventional wisdom apparently based on recent experience that perverse incentives and behavioral problems are central to the problem of family homelessness, and the Cuomo Commission has developed a detailed plan for reform. But the administration was not vigorous enough in translating its own hard-won insights into policy. Whether because of ideological reservations or because of bureaucratic turf wars, the administration was slow to embrace unequivocally, much less to implement effectively, the recommendations of the Cuomo Commission report. Implementing the plan would become a priority for the incoming Giuliani administration.
During the Dinkins administration, the first moment in the development of shelter as a right came to a close; this was the moment of entitlement. Throughout the Koch and most of the Dinkins administrations, policy spun around, first, establishing that there was indeed a right to shelter and, then, making good on that right by in fact providing shelter largely on demand. Two characteristics of this stage of the development of shelter as a right stand out. First, court intervention was a necessary, but not sufficient, condition of establishing the right to shelter. It is conceivable that a right to shelter could have been established by legislative means, but in fact the movement to establish that right focused on a court-oriented strategy from the beginning. Legislation by the City Council came after the right had been established and focused on regulating the provision of shelter.
However, it is important to understand exactly how the courts helped establish a right to shelter. There was not a single landmark decision that created the right. It is an error to think that the courts simply discovered a right to shelter in New York State’s constitution and issued a Brown-like decision that so found. The fact that article 17 of the state constitution famously mandated the “aid, care and support of the needy” undoubtedly created a legal environment in which litigation demanding a right to shelter was at least colorable. But the constitution left the implementation of the mandate to aid the needy to the discretion of the legislature and was never interpreted by the courts as clearly establishing a right to shelter. The courts’ role in helping to establish a right to shelter lay in creating a complex web of findings and decisions in the Callahan, Lamboy, McCain, and other cases that relied on a various legislative and constitutional provisions.
Second, other features and actors of state and city politics were necessary to consolidate the right. As we saw, through the Koch and Dinkins administrations, the courts never established a right to shelter in so many words. Courts held that if the city undertook to provide shelter, it would have to do so on certain terms and provide housing of a certain quality. The fact that the city was already providing shelter—albeit in the form of the old Men’s Shelter at 8 East Third Street and the infamous welfare hotels—was crucial. The city itself was thus an important actor in establishing the right to shelter because the court mandates all assumed that the city was, as a matter of contingent fact, providing shelter. The city was also a crucial actor insofar as it voluntarily decided to sign the Callahan consent decree. This is true despite the fact that the city was undoubtedly under strong legal pressure without which it would not have established a right to shelter on its own. For as we saw, in the end the city thought that, given the circumstances, signing the consent decree was the right thing to do, admittedly in part because they underestimated the size of the commitment they were taking on.
During the Dinkins administration, the limitations, or perceived limitations, of the entitlement stage of homeless policy became apparent. A response to those limitations, that of paternalism, was also articulated under Dinkins. Implementing that response would be the challenge of the Giuliani administration. The policy learning process continued.