National Housing Law Project Housing Law Bulletin

As part of the current drive to increase the number of employed families living in public housing, many public housing authorities are deciding to take applicants who are employed ahead of applicants who are not employed. HUD authorized such preferential treatment by regulation in August of 1994,¹ and, with Congress' suspension of the federal preferences since January 1996,² more and more PHAs have been giving preferential consideration to employed applicants. In 1996, the New York City Housing Authority (NYCHA) proposed to adopt such a preference, as well as to expand its site-based waiting list policies. The legality of those changes recently arose in a motion for a preliminary injunction in a previously settled case involving racial discrimination by the NYCHA. Davis v. New York City Housing Authority, 1997 WL 407250 (S.D.N.Y. July 18, 1997). The court issued a limited injunction against the employment preference, but did not find the expansion of the site-based waiting list policy to be unlawful.

Background

The Davis case was filed in 1990 by a class of African American and Latino public housing applicants to challenge the PHA's discriminatory tenant selection and assignment practices. A companion suit was later filed by the United States.³ The parties settled the matters by a consent decree in 1992 and, in the course of the settlement proceedings, the PHA in effect admitted that it had engaged in racial steering of applicants to create and maintain certain projects that could be racially identified as "white" projects. The consent decree required the PHA not to discriminate, and to adopt and follow a new tenant selection and assignment plan designed to eliminate its discriminatory practices and to desegregate the remaining 11 white projects.⁴

In the summer of 1996, NYCHA proposed to change the tenant selection and assignment program, allegedly to increase the number of employed families in its buildings. Under the plan established by the consent decree, all applicants were required to submit written applications. In the first stage of the selection process, the staff reviewed the applications to assign the applicants' proper priority status. Applicants who qualified for the now-suspended federal preferences were assigned priority status based upon those preferences. Applicants who qualified for the PHA's local preferences were assigned priority status based on those preferences. About 15 percent of the applicants had high enough priorities to be called in for an interview.

In the second stage of the process, the applicants called in for interviews were assigned to waiting lists. At the interview they were asked to select the borough in which they wished to live, and, if they needed two bedrooms or fewer and were not emergency cases, they were asked to select the particular project in which they wished to live. Applicants for three or more bedrooms were not allowed to select their project.

In stage three of the process, as vacancies arose, people on the borough or project waiting lists for appropriately sized units were offered the vacant units. The applicants on each waiting list were divided into three income tiers: Tier I for those with the lowest incomes, Tier II for the middle group, and Tier III for those with the highest incomes. The vacancies were offered on a rotating basis through the tiers, the first vacancy to the applicant from Tier III who applied first, the next to the Tier II applicant who applied first and the third to the Tier I applicant with the earliest application date. Once 25 percent of the anticipated annual vacancies were filled with Tier III (high-income) applicants, the PHA stopped offering vacancies to the Tier III applicants and split the rest of the vacancies between the Tier II and Tier I applicants. The goal was to rent 25 percent of all vacancies to Tier III applicants, and 37.5 percent to each of the other two tiers.

The Proposed Changes and Their Impact

The PHA's proposed changes fell into two general categories. First, with regard to preferences, the PHA proposed to reduce the proportion of units going to federal preference holders to 50 percent and to assign the rest of the vacancies to local preference holders. Among the federal preference holders, those who were employed or disabled would be offered units before any non-working federal preference holders. Among the local preference holders, Tier III (high-income) applicants got the highest priority, Tier II applicants got the next priority and Tier I applicants who were employed or disabled got the last priority. Non-working Tier I applicants would get no local priority at all. The purpose was to rent more apartments to higher income applicants and to employed applicants.

The second proposed change was to allow non-emergency applicants for three- and four-bedroom apartments to select the projects in which they wished to live, in addition to such applicants for one- and two-bedroom apartments. Again, the idea was that more employed and high-income people would apply for public housing if they could select the project in which they would live.

The Davis plaintiffs challenged both the proposed employment preferences and the proposed expansion of the site-based waiting lists. Their claims were that the changes would mean fewer class members and more white applicants would be admitted generally and that the desegregation of the racially identifiable white projects would be slowed down or halted.

The statistical analysis indicated that the percentage of newly admitted applicants who were white would rise from 4.2 percent to 9.9 percent under the new system. The percentage of admitted applicants who were African American would drop from 45.9 percent to 44.1 percent. For Puerto Rican families, the drop would be from 30 percent to 22.7 percent. Although the new percentages represented significant drops for people of color, they would be more consistent with the white and Puerto Rican groups' presence in the applicant pool. Whites would be 9.9 percent of the new admissions and were 10 percent of the applicant pool. Puerto Ricans would be 22.7 percent of the new admissions and were 22.3 percent of the applicant pool. African Americans, however, would be 44.1 percent of the new admissions but were only 35.6 percent of the applicants. In contrast, non-Puerto Rican Hispanics were 18.4 percent of the applicant pool but would receive only 14.2 percent of the vacant units.

In the 11 projects that were predominantly white, the new system would substantially slow down or halt desegregation. In four of the projects, desegregation would have been reversed. In another four it would have halted. In the remaining three, it would have slowed down significantly.

The Legal Analysis

To determine whether to grant the requested preliminary injunction, the court had to decide whether the plaintiffs were likely to succeed on their claim that the new system violated the Fair Housing Act, which had been incorporated into the consent decree. The court concluded that the plaintiffs were likely to succeed on the claim that the preference for working applicants would perpetuate past segregation in the 11 projects that remained as racially identifiable. It thus enjoined implementation of that preference until the PHA could show that a modified version would not adversely affect the desegregation. On the other hand, the court concluded that plaintiffs were not likely to succeed on their more general attack on the employment preference or on their challenge to the expansion of the site-based waiting lists.

Desegregation. To analyze plaintiffs' claim that the employment preference would perpetuate past discrimination, the court began with a review of the record in the case in which there were substantial allegations and evidence that NYCHA had steered applicants to maintain its predominantly white projects as predominantly white. Although there had been no trial and no adjudication of that fact, the court considered the consent decree, NYCHA's failure to rebut the evidence of discrimination, the evidence submitted at the fairness hearing, and NYCHA's admissions about its past practices at the fairness hearing as sufficient to establish past discrimination and segregation. In the process, the court rejected NYCHA's argument that it had expressly refused to admit liability in the consent decree and thus could not be considered to have engaged in past discrimination.

The court reviewed the plaintiffs' evidence and concluded that the employment preference would effectively freeze the desegregation process in the racially identifiable projects, because it would more than double the number of white applicants admitted and most of those applicants would go to the "white" projects. The court also rejected NYCHA's claim that freezing the desegregation of those projects was insignificant because the vast majority of its projects were predominantly minority and because its new system would still not admit white applicants at a rate higher than their presence in the applicant pool. The court correctly perceived that the Fair Housing Act violation was the segregation of these particular projects, not the state of the PHA's entire inventory.

Having concluded that plaintiffs had established a likelihood of presenting a prima facie case on the claim that the employment preference would perpetuate segregation in the white projects, the court moved on to consider the PHA's argument that the preference was the least discriminatory way to further its goals of producing income integration and increasing the number of working families in public housing.

The court concluded that the goals of the employment preference were legitimate, were being pursued in good faith and would be furthered by the preference. The issue was whether they were the least discriminatory means of achieving those goals. The plaintiffs' claim was that income integration and an increase in the number of working families could be achieved by granting the proposed preferences to the upper tier applicants, without also granting a preference for working applicants in the lowest tier. That special preference for working applicants in Tier I increased the chances of white applicants being admitted, because the proportion of white applicants among the employed Tier I applicants was almost twice as high as their representation among Tier I non-working applicants.

In analyzing this point, the court rejected the PHA's position that abandoning the employment preference within Tier I would frustrate the goal of increasing the number of employed residents. In doing so, the court made two points. One was that the burden was on the PHA to prove that an employment preference limited to the Tier II and III applicants would not increase the number of employed tenants. The second was that it was plausible that the applicants who were admitted under the Tier II and Tier III preferences would be employed families because they had higher incomes. The court also concluded that, with the PHA's wide discretion to develop local preferences, the burden was on the PHA to demonstrate that there were no other measures it could adopt to increase rentals to working families without perpetuating segregation.

On the basis of that analysis, the court enjoined the PHA from implementing the employment preference as proposed. However, it also made clear that it would modify the injunction if the PHA presented an alternative employment preference scheme that did not perpetuate segregation.

The More General Attack on the Employment Preference. Plaintiffs' success on their claim that the employment preference would perpetuate segregation in the 11 predominantly white projects was not equaled by their claim that the employment preference would have an adverse impact upon minority applicants. The claim was based on the fact that the employment preference would more than double the percentage of admitted applicants who were white, while imposing offsetting decreases in people of color. The difficulty with the claim was the fact that the increase in the acceptance rate among whites just brought them up to the rate at which they appeared in the applicant pool, i.e., 10 percent (actually 9.9 percent, with a 10-percent presence in the pool). Under the existing system, white applicants were admitted at a rate much lower than their presence in the applicant pool: 4.2 percent admitted compared to 10 percent in the pool.

As a result of those statistics, plaintiffs were unable to make the more ordinary claim that the employment preference caused minority applicants to be selected at a rate lower than their representation in the overall applicant pool. Instead, they focused on the fact that the employment preference would increase the number of white applicants who were admitted and reduce the number of non-white admissions.

Plaintiffs had to get around the fact that, even though the employment preference decreased the rate of minority admissions, it did not bring their admission rate lower than their overall presence in the applicant pool. They sought to do so by taking the position that the relevant comparison was between minorities' presence in the pool of people likely to be admitted before and after implementation of the employment preference. The court rejected that analysis for two interrelated reasons. First, even though the pool of applicants likely to be admitted under the employment preference was more white than the pool without the preference, in the court's view that fact would make a prima facie case only if the white applicants' presence in that pool exceeded their presence in the overall applicant pool. It was not enough that the employment preference would increase the proportion of white applicants in the pool of those likely to be admitted. Second, plaintiffs had not presented any evidence to show that white applicants' presence in the pool of people likely to be admitted would exceed their presence in the overall applicant pool.

In explaining why decreasing minority applicants' chances of being admitted, in itself, did not establish a prima facie case, the court distinguished a series of cases plaintiffs had relied upon, primarily because in each of the other cases the net effect of the defendant's practices was to reduce the selection of minority applicants beneath the acceptance rate one would expect in the absence of any discriminatory practices.⁵ After distinguishing those cases, the court also posited a policy reason for rejecting plaintiffs' position. In the court's view, plaintiffs were arguing that any change that reduced minority applicants' chances of being admitted from the status quo would present a prima facie violation of the Fair Housing Act. The court reasoned that the effect of that position would be that a PHA could never alter a selection policy if the change would reduce minority applicants' chances of being selected. As a result, the court concluded, a PHA would be deterred from establishing policies that favored minority applicants by the fear of never being able to change those policies.

Expansion of the Site-Based Waiting Lists. The other part of the proposed change to the tenant selection system that the plaintiffs challenged was the plan to allow applicants for three- and four-bedroom units to specify the development to which they wished to move. Under the existing system that had been developed as part of the consent decree, only applicants for one- and two-bedroom units were allowed to apply for specific developments and thus did not have to accept a unit offered elsewhere in their borough.

Historically, after assignment of tenants to particular developments on the basis of race was made illegal, some housing authorities were still able to maintain segregated projects by keeping separate waiting lists for each project and either steering applicants to the different lists based upon their race, or allowing them to self-select. In response, HUD's predecessor required PHAs to establish centralized waiting lists and to send applicants to the bottom of the list if they rejected offers of more than two units from that list.⁶

Plaintiffs challenged the proposal to allow applicants for three- and four-bedroom units to select the development they wished to wait for on the grounds that doing so would perpetuate the segregation of the 11 predominantly white lists. Because HUD had approved the proposed change, the court concluded that its role was analogous to judicial review of informal agency action under the Administrative Procedure Act, i.e., its role was to determine whether HUD had acted arbitrarily or capriciously. The court concluded HUD had not acted arbitrarily, because of the facts of this particular case. Plaintiffs were able to show that the change would make only 17 additional white families eligible to select their development, and only eight of those families would move into projects in which the white occupancy rate was above 10 percent. Those facts convinced the court that allowing applicants to select the developments they would wait for would not perpetuate any segregation that existed in New York City public housing.

This decision and the issues it covers should be of continuing significance as more and more housing authorities adopt employment preferences and other tenant selection devices for public housing and certificates and vouchers that favor applicants with higher incomes. In cases where such devices impede actions that are being or should be taken to desegregate predominantly white developments, the courts are likely to step in and prevent implementation of the new schemes. The courts are also likely to step in where the statistics show that the new devices would decrease the rate at which people of color or people from particular protected classes will be admitted below their presence in the eligible population. The courts may also step in if the rate of admissions for protected classes drops below the level of their presence in the applicant pool. Neither of those latter two situations were presented in this case.

The more difficult cases to decide will be ones like this one, where the changes from the status quo clearly work to the disadvantage of protected classes, but not to the extent of reducing their rate of admission below their presence in the applicant pool. This court was not willing to treat such an adverse impact as presenting a prima facie case under the Fair Housing Act, but there does appear to be something invidious about policies that increase the amount of time that non-white applicants have to wait for housing and decrease the waiting time for white applicants. That is especially true when the changes are applied to applicants who are already on the waiting list and are close to the time when they would be receiving a unit. No doubt over the next few years, these questions and others like them will be examined often by the courts.

1. 24 C.F.R. § 960.205(a) (1997).
2. Pub. L. No. 104-99, § 402(d), 110 Stat. 26 (Jan. 26, 1996).
3. United States v. New York City Hous. Auth., No. 92 Civ. 4873 (S.D.N.Y. complaint filed 1992).
4. See Davis v. New York City Hous. Auth., 1992 WL 420923 (S.D.N.Y. Dec. 31, 1992), subsequent opinion, 839 F. Supp. 215 (S.D.N.Y. 1993), subsequent opinion, 940 F. Supp. 80 (S.D.N.Y. 1996).
5. See Connecticut v. Teal, 457 U.S. 440 (1982); Waisome v. Port Auth., 948 F.2d 1370 (2d Cir. 1991); Giles v. Ireland, 742 F.2d 1366 (11th Cir. 1984).
6. See Notice on Site-Based Waiting Lists, 62 Fed. Reg. 1,026 (Jan. 7, 1997).