Court: Court of Appeals for the Second Circuit; January 2, 2002; Federal Appellate Court
The case involves the New York City Housing Authority (NYCHA) and its compliance with a 1992 consent decree aimed at addressing racial discrimination in public housing admissions. Following earlier appellate remands in 1999 and 2000, the district court found that NYCHA's proposed introduction of a working-family preference would perpetuate past segregation in 20 low-income housing projects. Consequently, the court permanently enjoined NYCHA from implementing this preference at those sites. NYCHA's appeal focused on two main issues: its interpretation of segregation and the assessment that segregation would be significantly perpetuated. The appellate court affirmed the injunction regarding 14 projects but reversed it for six others. The consent decree stemmed from allegations of discrimination against Latino and African American applicants, with NYCHA acknowledging past racial steering practices and agreeing to the decree as a means of remedying these issues.
Following a fairness hearing, the Consent Decree was approved by then-District Judge Pierre N. Leval, establishing that the Housing Authority engaged in unlawful discrimination against Black and Hispanic applicants during specific periods. Key discriminatory practices included: using codes that restricted assignments to white families, implementing zip code and geographic restrictions, setting racial goals for project occupancy, and assigning families to projects without anticipated vacancies. The Consent Decree permanently enjoins the Housing Authority (NYCHA) from any actions that deny equal housing access based on race, color, or national origin, prohibiting preferences in tenant applications and transfer policies based on these factors. Other injunctions prevent racial considerations in transfer denials and false representations about project vacancies linked to race.
While some aspects of the Decree are set to dissolve after eight years, the permanent injunctive provisions remain in effect. The Decree mandates the implementation of a new three-stage tenant selection and assignment plan (TSAP) designed to eliminate unlawful discrimination, aligning with Title VI, the Fair Housing Act, and HUD regulations. HUD approved the TSAP for five years, which acknowledges federal and local preferences for certain applicants. The process includes three stages: initial eligibility interviews, assignment to waiting lists categorized into three income tiers, and vacancy allocation, aiming for 25% of vacancies for the highest income tier (Tier III) and the remaining 75% split between the lower tiers, with priority based on housing need.
The Decree mandates that no applicant will be disqualified from projects under the Tenant Selection and Allocation Plan (TSAP) due to minimum income requirements that discriminate based on race, color, or national origin, in violation of Title VI or the Fair Housing Act. However, the Housing Authority may aim for a diverse tenant body representing various income levels, in line with HUD standards. If NYCHA introduces new minimum income requirements for existing projects, plaintiffs have the right to challenge these changes within five years post-TSAP implementation. NYCHA must fully implement the TSAP within a year of the Decree's entry and provide 60 days' notice to plaintiffs for any proposed modifications during the five-year period, allowing plaintiffs to seek court intervention. Additionally, for three years following this period, plaintiffs can request court modifications of the TSAP if it violates fair housing laws.
In 1995, NYCHA proposed a Working-Family Preference (WFP) that would prioritize working or disabled federal preference holders over others, altering the local preference hierarchy. This proposal would shift focus from housing need to the ability to pay, aiming to increase the presence of working families and enhance income integration in public housing. Plaintiffs opposed the WFP, arguing it would favor white families and perpetuate racial segregation, suggesting an alternative that would maintain Tier I families' lower priority but offer them equal preference among themselves. NYCHA acknowledged the potential for increased income integration from this alternative but rejected it, claiming it would undermine their goal of increasing rentals to working families.
The district court recognized NYCHA’s need to address concerns about tenant income distribution to maintain project stability. Historically, low-income applicants made up about one-third of NYCHA rentals, but by 1995, they represented 77.6% of new admissions due to increased homelessness. NYCHA argued, without dispute from plaintiffs, that a higher proportion of higher-income applicants was essential for stability. The court noted NYCHA's submission of its proposed Working Family Preference (WFP) and "Project Choice" to HUD. While HUD approved Project Choice, it did not endorse local preferences but allowed NYCHA to create them with notice and comment, cautioning that they must comply with the Davis consent decree and civil rights laws.
The court analyzed the WFP’s potential impact on NYCHA’s obligations. It found that only 15% of applicants reached the interview stage under the existing system, with a mere 7.2% of the top applicants being white. Implementation of the WFP would likely increase the percentage of white families among the top applicants, a point not contested by NYCHA. Dr. Cupingood's analysis indicated that the WFP could hinder desegregation at several projects, potentially slowing or reversing progress at some locations.
The court suggested that NYCHA could adjust the WFP to mitigate discriminatory outcomes, allowing the preference if white applicants are not placed in over-represented projects. The court accepted Dr. Cupingood’s projections and found that plaintiffs had a strong case against the WFP concerning predominantly white housing projects. Consequently, the court issued a preliminary injunction against implementing the WFP at all 322 NYCHA projects but encouraged NYCHA to propose modifications to limit the injunction to disproportionately white projects.
The parties agreed that the Working Family Preference (WFP) evaluated in Davis II would not promote past discrimination at New York City Housing Authority (NYCHA) projects, except for 21 projects where over 30% of apartments are occupied by white families, referred to as the "Disproportionate Projects." The court in Davis III limited the injunction against the WFP to these 21 projects following NYCHA's appeal. However, the appellate court in Davis IV found that the district court had not sufficiently justified its findings that the WFP would likely perpetuate segregation at these projects, citing a lack of detailed explanation and data to support the conclusions drawn by Judge Sweet. The appellate court noted that while some trends were mentioned, they were not adequately discussed, particularly the impacts of the proposed Tenant Selection and Admission Plan (TSAP) on future demographics of the developments. The court emphasized the need for specific numerical data regarding the additional white families that would be admitted to each development.
Following this, the district court held a remand hearing, considering new evidence and arguments about whether segregation existed and if the WFP would significantly perpetuate it. In its August 11, 1999 decision in Davis V, the court affirmed both the existence of segregation and the conclusion that the WFP would perpetuate it, leading to a permanent injunction preventing NYCHA from applying the WFP in 20 housing projects. The court defined a project as segregated if more than 30% of its apartments are occupied by white families, consistent with previous rulings in the litigation.
The court affirmed that the 30% white benchmark in the TSAP, incorporated into the Consent Decree, is not arbitrary but a negotiated figure indicating that projects with over 30% white families are considered disproportionately white and thus affected by past segregation. The TSAP includes a "Borrowing Provision" allowing housing projects with excess vacancies to borrow applications from other projects, but it prohibits borrowing if both projects have white tenant bodies exceeding 30%. The court emphasized that remedies must relate to the goal of eliminating discrimination without infringing on third parties' interests. It highlighted that without evidence of past segregation in projects exceeding the 30% white threshold, the Borrowing Provision would be unlawful. The court also noted findings from Dr. Cupingood regarding the impact of the WFP on the racial demographics of public housing, using data from 1991-1994 to predict future move-ins without distortion from NYCHA's prior discrimination. Dr. Cupingood's analysis indicated that the WFP would significantly impact disproportionately white projects, either by increasing their white populations or affecting their desegregation process.
The Davis V opinion presents various tables reflecting the impact of the WFP (White Family Preference) on white occupancy rates in disproportionately white housing projects over five and ten years, considering historical turnover by race. Key findings include:
- At Middletown Plaza, the percentage of white families was projected to increase from 51.85% to 60.1% after five years under the WFP, rather than decreasing to 49.4%.
- In other projects like Cassidy Lafayette, the decline in white occupancy would slow, from 53.6% to 50.4% instead of dropping to 42.4%.
- Table 5 illustrates these changes in white occupancy percentages across various projects.
- After ten years, similar trends persisted, with the WFP leading to a net increase of 722 white families in disproportionately occupied projects, contrary to the scenario without the WFP.
- NYCHA contested Dr. Cupingood’s analysis by arguing that the predicted increase in white family admissions would be lower than stated. However, the district court identified significant methodological flaws in Dr. Peterson's arguments, including an underrepresentation of families moving into NYCHA housing and the influence of the "Pipeline Effect."
- Even with adjustments for Dr. Peterson's undercount, a corrected white admission rate in 1998 was found to be 8.53%, indicating a higher percentage than NYCHA claimed.
The findings were documented in specific tables referenced in the opinion, demonstrating the quantitative impacts of the WFP on housing demographics.
The court dismissed Dr. Peterson's analysis in favor of Dr. Cupingood's findings. It rejected NYCHA's argument that a decline in white occupancy rates under the WFP does not equate to perpetuating segregation. The court clarified that "perpetuate" means to extend over time, which the WFP would do. In assessing housing discrimination, after remedies for segregation are ordered, the focus shifts from whether minority applicants are underrepresented to whether proposed changes will significantly perpetuate segregation.
When a court-ordered plan is in place to eliminate past segregation, any changes proposed must be compared to the original plan to determine if they contribute to or delay desegregation. The court concluded that the WFP would significantly impede desegregation in NYCHA's 20 disproportionately white projects, as it would more than double white admission rates and likely lead to increased concentrations of white families in predominantly white developments.
The court evaluated both "statistical significance," which assesses whether the observed effects are likely due to chance, and "legal significance," which determines the legal implications of those effects. It referenced a standard two-standard deviation test as a method for assessing statistical significance, indicating that results deviating significantly from expected norms are unlikely to be random. Courts typically consider deviations of two to three standard deviations as statistically significant.
A statistical analysis indicated that if the difference between expected values and observed numbers exceeds two or three standard deviations, it raises doubts about the hypothesis of non-discriminatory hiring practices. In the context of the WFP's impact on white move-ins at disproportionately white projects, the court found significant deviations. Over five years, the WFP would lead to approximately 1,139 white move-ins compared to 625 without the WFP, a difference of 514 move-ins or 15.39 standard deviations. Over ten years, this would increase to 2,240 white move-ins, reflecting a 21.92 standard deviation difference. Statistically significant differentials were noted in 18 out of 21 projects after five years and in 20 out of 21 projects after ten years. For example, at Holmes Towers, the WFP would increase white move-ins from 34 to 75 (4.3 standard deviations), while at Pomonok, the figures would rise from 138 to 287 (8.21 standard deviations).
The court emphasized the legal significance of even small percentage-point increases in white occupancy, stating that at Pomonok, a 3.2 percentage point increase after five years could mean 66 fewer apartments for non-white families, worsening segregation. The WFP's benefits were acknowledged, but the adverse effects on desegregation rates and compliance with the Consent Decree were deemed more impactful. Consequently, the court permanently enjoined NYCHA from implementing the WFP at the 20 relevant developments, and upon appeal, the case was remanded for further information without resolving the merits.
The WFP, implemented in 1998 across 302 projects, was under scrutiny regarding its impact on tenant demographics at 20 projects covered by an injunction. Key inquiries included: (1) the correlation between actual and projected tenant move-outs and their influence on expert conclusions; (2) the expected duration to achieve a white family occupancy rate below 30% at the 20 projects with and without the WFP; and (3) whether the white admissions rate should be revised from Dr. Cupingood’s estimate of 9.9% to either 8.53% or 8.28%.
In June 2000, the district court addressed these inquiries, concluding that actual move-out rates for 1998 would not enhance the accuracy of white admissions rate projections. It determined that the reference to an 8.53% rate was a clerical error, intended to reflect 8.28%. Despite this, the court found Dr. Cupingood's rationale for using 9.9% more compelling than Dr. Peterson's arguments for a lower rate, and concluded that using the 8.28% rate would not significantly alter the findings regarding the WFP's role in prolonging segregation at the 20 projects.
In response to the second question, the court provided comparative estimates of desegregation timelines under the TSAP with and without the WFP. Following these findings, NYCHA’s appeal was reinstated. NYCHA contended that the district court misapplied the definition of segregation and erred in its analysis of the WFP’s effects. The appellate review upheld the district court's factual findings as not clearly erroneous, except for its conclusion regarding the legal significance of the WFP's effects at six projects.
The phrase "clearly erroneous" refers to a standard used by appellate courts to review findings made by district courts. A finding is deemed "clearly erroneous" when, despite supporting evidence, the appellate court is left with a firm conviction that a mistake has been made. This standard does not allow appellate courts to overturn findings merely because they would have reached a different conclusion. Appellate courts must not replicate the role of the district court and should not decide factual issues anew. When the district court's interpretation of the evidence is plausible based on the entire record, the appellate court cannot reverse it, even if it would have weighed the evidence differently. If there are two reasonable interpretations of the evidence, the factfinder's choice is not considered clearly erroneous.
The standards apply regardless of whether the district court's findings are based on oral testimony or documentary evidence. Findings related to racial discrimination are also subject to this clearly erroneous standard. Additionally, the appellate review is limited to determining if the trial judge's conclusions are clearly erroneous, while conclusions regarding legal questions or mixed questions of fact and law are reviewed de novo.
A determination of the legal significance of effects identified by the court is a conclusion of law, and the interpretation of a consent decree is also a legal issue subject to appellate review. Consent decrees function as both contracts and enforceable orders, with the appropriateness of permanent injunctions reviewed for abuse of discretion. While courts cannot arbitrarily alter consent decree terms, they possess broad discretion in enforcement to ensure compliance.
In rejecting NY-CHA's argument against using a 30% white family population level as a standard for segregation in public housing projects, the district court justified its decision based on the Consent Decree and related litigation. The 30% threshold was explicitly referenced in the Decree and the TSAP's Borrowing Provision, which restricts borrowing applications from similarly populated projects. The Decree also identified 31 “Affected Developments,” requiring priority for families adversely affected by discrimination since 1985. At the time of the decree, a significant majority of these developments had white family populations exceeding 30%, validating the use of this figure as indicative of projects affected by past segregation. NYCHA's claim that compliance with the Decree negates segregation status contradicts the Decree's provisions.
The Decree permits plaintiffs to seek an injunction against modifications to the Tenant Selection and Assignment Plan (TSAP) within the first five years if such changes conflict with the Decree or the Fair Housing Act (FHA). After this period, plaintiffs may request modifications based on violations of the FHA in the TSAP's implementation. The assessment of segregation in New York City public housing is to be measured by FHA standards rather than solely on individual victim relief. Data from NYCHA indicated that white families comprised only 7% of the tenant population, leading the court to define a project as segregated if the white population was 23 percentage points above the system-wide average or more than four times that average. This approach aligns with precedents such as United States v. Yonkers Board of Education, which recognized significant demographic disparities as indicative of segregation. NYCHA's argument that the litigation was resolved without a trial and that they did not concede liability is deemed unconvincing; a violation of the FHA can be established without proving discriminatory intent, merely by demonstrating a discriminatory effect. The FHA empowers courts to mandate affirmative actions to rectify past segregation. NYCHA's records confirm ongoing segregative housing patterns across its projects.
In the context of the Consent Decree, 62.66% of white families living in New York City Housing Authority (NYCHA) projects were concentrated in 31 designated Affected Developments, with an additional four projects (Haber, Independence, Taylor-Wythe, and Williams) housing a significant percentage of white tenants, despite not being classified as Affected Developments. By the end of 1992, these four projects housed 1,298 white families, contributing to over 71% of the white public-housing population residing in 35 out of roughly 320 NYCHA projects. The district court found that NYCHA had engaged in practices resulting in unlawful racial discrimination against Black and Hispanic applicants, which NYCHA acknowledged in support of the Consent Decree, stating that historical policies favored white applicants and involved racial steering. As of the Decree's entry, only 7% of families in NYCHA were white, while projects with over 30% white occupancy were deemed segregated. The court's methodology, including the 30% threshold for segregation, was upheld as reasonable, and it rejected NYCHA’s arguments against the findings regarding the impact of the tenant selection assistance program (TSAP) on tenant demographics, affirming the credibility of the district court's analysis and conclusions.
Decisions regarding which witness to credit and which inferences to draw are the exclusive responsibility of the district court as the factfinder. The court determined that Dr. Cupingood's data extrapolation was based on a reasonable, undistorted period, and his assumptions, which utilized historical averages to mitigate uncertainties, were deemed reasonable. Both sides' experts agreed that using actual turnover data would not alter their opinions on the anticipated effects of the WFP moving forward. While predictive analyses are inherently more speculative than statistical analyses of existing data, the necessity for predictions in this case was unavoidable. The district court found Dr. Peterson's analysis flawed, but this does not affect the permissibility of his views. The acceptance of Dr. Cupingood’s analysis over Dr. Peterson’s was not clearly erroneous. NYCHA's argument that Dr. Cupingood's analysis would automatically yield statistical significance over time was rejected, as the inquiry here does not require traditional statistical analysis. The court adequately analyzed the WFP's effects at both the project and aggregate levels, and the inclusion of multiple tables in prior rulings confirmed that individual projects were considered. Lastly, the court's comparison of anticipated desegregation effects under the WFP versus those under the original TSAP was appropriate, as the TSAP remedies aimed to address previous racial discrimination.
The district court's analysis, aligned with Supreme Court precedents, deemed it reasonable to assess the impact of NYCHA’s proposed changes against the expected outcomes under the original TSAP. The legal significance of the differences in desegregation rates between the original and modified TSAP is a question of law subject to de novo review. The court emphasized the importance of understanding how actions that slow desegregation could perpetuate past segregation, referencing cases such as Arthur v. Nyquist and Sarabia v. Toledo Police Patrolman’s Ass’n.
The court argued that failing to consider the implications of proposed actions on desegregation could lead to an indefinite delay in achieving it. However, NYCHA's proposed Working Families Preference (WFP) serves a legitimate countervailing interest—enhancing the number of working families in public housing to promote income integration and stability. This aligns with federal mandates prohibiting the concentration of low-income families in public housing and requiring agencies to implement policies aimed at income deconcentration. The statute mandates that agencies must develop admissions policies that facilitate a mix of income levels without enforcing specific income or racial quotas.
Public housing agencies are authorized to implement income-mix criteria for resident selection in public housing, as outlined in 42 U.S.C. 1437n(a)(1). HUD regulations allow for preferences towards “working families” to promote diversity and role models within housing projects (24 C.F.R. 960.205(a), 960.206(b)(2)). Although HUD supports this flexibility, it did not endorse the New York City Housing Authority’s (NYCHA) proposed Working Family Preference (WFP), nor did it criticize it despite its duty under Title VI of the Civil Rights Act to ensure compliance with civil rights laws in federally funded housing. HUD reminded NYCHA to notify about the WFP proposal and adhere to the Davis consent decree's injunctive relief and civil rights obligations.
The assessment of the WFP's legal significance requires balancing the interests of addressing past segregation effects with achieving financial stability in public housing. A district court recognized the growing percentage of applications from homeless or low-income families, which rose from 33% to 77.6%, indicating that without a higher proportion of higher-income applicants, NYCHA’s projects face stability risks. However, it is emphasized that public housing agencies must not implement financial deconcentration plans that violate civil rights laws. The concentration of white families in a small percentage of NYCHA projects resulted from the agency's prior discriminatory practices, including racial steering and misrepresentations, which had led to over 71% of white families being placed in only 35 out of 320 projects at the time the consent decree was entered.
As of December 31, 1992, over 51% of white families in NYCHA projects (7,929 out of 15,530) resided in the 20 projects relevant to this appeal, with minimal decline in concentration post-TSAP implementation. By January 1, 1999, 12,185 white families lived in NYCHA projects, with 5,771 in the same 20 projects, representing 47.36% of the white families across NYCHA's 322 housing projects. The district court concluded that the desegregation delays from the WFP's implementation were legally significant, considering both percentage impacts and real-time delays. Distinctions were identified among the projects: some were effectively desegregated pre-injunction, while others faced varying delays, from minimal (two to twelve months) to substantial. Notably, Holmes Towers and Straus had white populations below 30% at the time of the injunction, and there was no prediction of their exceeding this threshold due to the WFP. The court deemed that these two projects should not be considered segregated, thus suggesting that the WFP's potential delays in these cases were not legally significant. In four projects, delays ranged from two to twelve months, with the shorter delays deemed de minimis and the longer delays not considered excessive, as significant progress had been made towards desegregation since the Consent Decree.
On December 31, 1992, Isaacs and Sheepshead Bay had white tenant populations of 52.9% and 54.1%, respectively, which dropped to 36.3% and 35.0% by January 1, 1999. The document concludes that the 11 and 12 month delays in these two projects are not legally significant given their historical context and the pursuit of financial stability. Therefore, the implementation of the WFP (Workforce Program) at Pelham Parkway, Bay View, Isaacs, and Sheepshead Bay should not be enjoined.
In contrast, the WFP’s effects on 14 other projects, which have delayed their desegregation significantly, are deemed legally significant. Despite the law permitting some reasonable delay, almost a decade has passed without achieving a 30% desegregation level. These projects had white tenant populations above 30% in 1992 and still maintain similar levels in 1999, with nine exceeding 50% white tenants. The original TSAP (Tenants' Strategy Action Plan) allowed for some delay, but NYCHA's proposed delays are viewed as excessive, leading to a permanent injunction against the WFP at these remaining projects.
Specific projections suggest that some projects like Independence, Williams, and Taylor/Wythe would not achieve desegregation for over 50 years. The WFP would further extend the time to desegregation by an additional 23 to 56 years. In six projects, while the WFP does not delay reaching the 30% level by more than 12 months, its effects are not legally significant. However, in the other 14 projects, the delays are significant enough to justify the injunction.
The district court retains the discretion to modify its injunction based on changed circumstances. NYCHA may seek to modify the injunction if the white tenant population in any project decreases to the 30% level. The court's order is partially reversed, allowing the WFP to be implemented at specific projects but affirming the injunction for the others.
