On July 1, 1992, plaintiffs, Latino and African American individuals residing in or eligible for New York City Housing Authority ("NYCHA") housing ("the Davis plaintiffs"), the United States, and NYCHA signed a Consent Decree to resolve two actions alleging that NYCHA had engaged in discrimination in the selection and assignment of public housing tenants in violation of various federal statutes.1 The Decree provides, inter alia, for: (1) injunctive relief barring future housing discrimination on the basis of race, color or national origin; (2) the implementation of a new Tenant Selection and Assignment Plan ("TSAP") that substantially revises NYCHA's tenant selection and assignment systems and which prohibits further discrimination; (3) remedial relief for approximately 2,000 Latinos and African Americans who were denied public housing between 1985 and 1990 based on race; and (4) significant record keeping and reporting by NYCHA concerning tenant selection and assignment practices.2

The TSAP created under the Decree and currently in place incorporates a three-stage process for selecting applicants for vacancies in NYCHA housing. During the first stage, applicants are selected for eligibility interviews from an applicant pool. NYCHA assigns priority codes to each application, based on various federal priorities3 and local preferences. Applicants with a certain priority code are then interviewed, at which time they are asked to state their preference as to a borough or project. Typically, approximately 15% of applicants reach the interview stage.

The second stage of the process entails placing interviewed individuals on waiting lists for vacancies.

During the final stage, applicants on waiting lists are chosen to fill vacancies as they arise. In order to achieve an income mix, the primary consideration in selecting tenants is the applicant's "income tier." Tier I families have the lowest incomes, while Tier III families have the highest incomes. Under the current TSAP, NYCHA attempts to rent 25% of all vacancies to Tier III families, 37.5% to Tier II families and 37.5% to Tier I families.

The selection protocol currently used to achieve this income mix is as follows: the first vacancy is offered to the Tier III applicant with the highest priority (e.g. local and federal priorities) and oldest date of certification; the second vacancy is offered to the Tier II applicant with the highest priority and oldest date of certification; the third vacancy is offered to the Tier I applicant with the highest priority and oldest date of certification; and the sequence is repeated until 25% of the expected annual vacancies are filled with Tier III families, then the remaining vacancies are available only for Tier I and II applicants.

Ten years ago, nearly half of all families living in NYCHA developments were working families, whereas today fewer than one third are employed. It is undisputed that unless a higher proportion of applicants with higher incomes receive rentals, the stability of the projects will be jeopardized. In 1995 NYCHA proposed to the United States Department of Housing and Urban Development ("HUD") two changes to the TSAP: (1) implementation of a preference for working families and (2) extension to larger families a limited choice of the project as opposed to the borough in which they wish to reside. Only the first change, the "working family preference," is at issue in these proceedings.

NYCHA proposed implementing the working family preference through the TSAP as follows: first, NYCHA would use all of its local preferences (up to 50% of all rentals) to give first priority to Tier III families, second priority to Tier II families, and last priority to Tier I working or disabled families. Second, for those who already qualify for federal preferences, NYCHA would provide another priority to working families.

On November 15, 1996, the Davis plaintiffs filed a motion in district court to enjoin NYCHA from implementing the proposed changes to the TSAP. Over the next six months, Judge Sweet accepted numerous written submissions from the parties and, on one occasion, heard oral argument. By Opinion and Order dated July 17, 1997, Judge Sweet permitted NYCHA to implement the "project choice" modification to the TSAP, but enjoined NYCHA from implementing the working family preference. Davis v. New York City Housing Authority, 1997 WL 407250 (S.D.N.Y. July 18, 1997) (Sweet, J.). Judge Sweet rejected plaintiffs' argument that the working family preference as proposed by defendant would have a disparate impact on plaintiffs, but found that it was likely that plaintiffs could show that the plan would perpetuate segregation at 11 developments, and therefore enjoined implementation of the preference at all 322 NYCHA developments. Id. at * 12.

NYCHA filed a notice of appeal and moved for reconsideration in district court. On November 13, 1997, Judge Sweet denied NYCHA's motion on grounds that the arguments were not properly advanced on reconsideration, but narrowed the prior ruling to enjoin the working family preference at only 21 NYCHA developments, at which white families currently constitute more than 30% of the tenant population. Davis v. New York City Housing Authority, 1997 WL 711360 (S.D.N.Y. November 13, 1997) (Sweet, J.). NYCHA filed another notice of appeal.

On appeal, NYCHA seeks to overturn the preliminary injunction on various grounds, including that the court's findings on the issue of perpetuation of segregation were vague, conclusory and unsupported.

We review for abuse of discretion a district court's grant of a preliminary injunction. Suffolk Parents of Handicapped Adults v. Wingate, 101 F.3d 818, 822 (2d Cir.1996), cert. denied, 520 U.S. 1239, 117 S. Ct. 1843, 137 L. Ed. 2d 1047 (1997). However, we turn first to NYCHA's contention that the preliminary injunction is procedurally flawed, that is, that the finding regarding perpetuation of segregation lacked sufficient detail to serve as a basis for reasoned appellate review.

Federal Rule of Civil Procedure 52(a) requires that in granting or refusing interlocutory injunctions, the court shall set forth the findings of fact and conclusions of law which constitute the grounds of its action. Fed.R.Civ.P. 52(a); Inverness Corp. v. Whitehall Laboratories, 819 F.2d 48, 49 (2d Cir.1987). The rule serves several purposes. First, it aids the appellate court in understanding the ground or basis for the trial court's decision. Id. (citations omitted). "It is of the highest importance to a proper review of the action of a court in granting or refusing a preliminary injunction that there should be fair compliance with Rule 52(a) of the Rules of Civil Procedure." Mayo v. Lakeland Highlands Canning Co., 309 U.S. 310, 316, 60 S. Ct. 517, 520, 84 L. Ed. 774 (1940).

Second, the rule encourages the trial judge to ascertain the facts with due care and render a decision in accord with the evidence and the law. Inverness, 819 F.2d at 50. What is adequate to serve these two purposes depends on the importance of an issue, its complexity, the depth and nature of the evidence presented, and similar elements that may vary from case to case. See Kelley v. Everglades Drainage Dist., 319 U.S. 415, 420-22, 63 S. Ct. 1141, 1144-45, 87 L. Ed. 1485 (1943); Knapp Shoes, Inc.v. Sylvania Shoe Manufacturing Corp., 15 F.3d 1222, 1228-29 (1st Cir.1994); 9A Charles A. Wright and Arthur R. Miller, Federal Practice and Procedure § 2579 (2d. ed.1995).

In finding that it was likely that plaintiffs would be able to show that the proposed TSAP would have a discriminatory effect by perpetuating segregation in NYCHA developments, Judge Sweet first outlined the record evidence of a history of segregation. Davis v. NYCHA, 1997 WL 407250 at * 11. He then defined segregation in public housing as existing "if housing projects are significantly out of racial balance," or "predominantly white," and noted that under the current TSAP, projects are considered racially imbalanced if "more than 30% white." Id. at * 12. He then found that eleven NYCHA projects remained over 50% white and were out of racial balance due to past segregation. Id. Finally, he found that the proposed TSAP would perpetuate segregation at certain NYCHA developments. Id. It is this ultimate finding that defendant attacks as insufficient.

In the challenged portion of his Opinion, Judge Sweet found that

assuming that existing turnover and move-in rates persist and that for each vacancy set aside for claimants under the Davis decree, a family of color would replace a white family, implementation of the Working Family Preference would effectively "freeze" the desegregation process at the racially identifiable projects, in violation of Title VIII. Specifically, the Working Family Preference would reverse the desegregation process at four of the eleven majority-white projects, there would be no further desegregation at another four, and desegregation would be slowed significantly at three other projects.... [It] will have this effect because it will more than double white admission rates and because existing trends demonstrate that many of these additional white families will be concentrated in predominantly white developments.

Id. On this basis, he enjoined use of the proposed TSAP at all 322 NYCHA developments. On reconsideration, he relied on this same finding to modify his original order to apply to 21 developments where whites currently constitute more than 30% of the tenant population, including the eleven developments covered by the original injunction and 10 additional developments about which plaintiffs' expert had offered no opinion. Davis, 1997 WL 711360 at * 4.

In view of the importance of the issue and the importance of NYCHA's interest in the viability of its system, we consider Judge Sweet's findings on the issue of perpetuation of segregation insufficient under Rule 52(a). While not entirely devoid of detail, Judge Sweet failed to adequately explain the subsidiary facts and methodology underlying the ultimate finding. For instance, Judge Sweet refers to "existing trends demonstrat[ing] that many of the[ ] additional white families [admitted under the proposed modification] will be concentrated in predominantly white developments," but does not discuss or attempt to explain these trends or the data reflecting them. He also fails to address the time period during which the purported impact of the proposed TSAP was assessed and/or how future "trends" might affect application of the TSAP to developments that are not currently, but subsequently become, predominantly white. Further, while he focuses on the racial imbalance the proposed TSAP will cause at specific developments within the NYCHA system,4 he does not identify them by name5 or state the number, fraction or percentage of additional white families who will be admitted to each of the 21 developments as a result of the proposed TSAP.

Of course, we may proceed with appellate review despite inadequate findings if we are able to discern enough solid facts from the record to permit us to render a decision. Tekkno Laboratories, Inc. v. Perales, 933 F.2d 1093, 1097 (2d Cir.1991).

Judge Sweet states that he credited the opinion of plaintiffs' expert, Dr. Leonard Cupingood. Indeed, he adopted Paragraph 13 of Dr. Cupingood's Second Affidavit almost verbatim. In that paragraph, Dr. Cupingood stated that

I was also asked to calculate the effect that the proposed changes to the TSAP would have on slowing down the desegregation in projects which were more than 50 percent white as of June 1996. Information supplied to me indicated that the following 11 projects were more than 50 percent white as of June 1996--Berry, Cassidy-Lafayette, Forest Hills, Independence, Middletown Plaza, New Lane, Robbins Plaza, South Beach, Taylor-Wythe, Todt-Hill, and Williams. Some of these projects are affected by the Davis Consent Decree. For these "Davis" projects, I first conservatively assumed that all units set aside for Davis tenants would represent replacements for white tenants. Then, assuming the turnover rates and move-in rates from Appendix E of [NYCHA expert] Dr. Peterson's affidavit continued and that the proposed changes to the TSAP were adopted, I calculated that a) there would be no further desegregation at four of the projects, b) the desegregation would be reversed at four other projects, and c) the desegregation would be slowed significantly at three projects. Thus, adoption of the proposed changes to the TSAP would adversely affect the desegregation at these majority white projects.6

This paragraph raises more questions than it answers. As an initial matter, it is unclear to which proposed change--project choice or working family preference, or both--plaintiffs' expert attributes the perpetuation of segregation. Plaintiffs' expert begins by stating that he was asked to calculate the effect "the proposed changes" would have in slowing down desegregation in developments with white populations exceeding 50%.7 The conclusion portion, however, fails to link each proposed modification to a purported effect.

The vagueness of Paragraph 13 on the issue of the segregative effect of the working family preference is further highlighted by the specificity of the expert's opinion concerning the impact of project choice. While Paragraph 13 of the Second Affidavit only vaguely discusses the effect of the working family preference, Paragraph 14 specifies the exact effect project choice will have on NYCHA developments. Correspondingly, Judge Sweet's findings on project choice are much more detailed. Compare Davis, 1997 WL 407250 at * 178 with Davis, 1997 WL 407250 at * 12, cited in pertinent part hereinabove.

Another difficulty with plaintiffs' expert's opinion is that he identifies his source of statistics generally as Appendix E of NYCHA's expert's report, which consists of data compiled by NYCHA, but does not cite to any precise numerical data underlying his opinion. His only other reference to data in Paragraph 13 is to "[i]nformation supplied to me."9 Plaintiffs' counsel has attempted, both in the district court and on appeal, to explain the subsidiary facts and methodology relied on by plaintiffs' expert in reaching the ultimate opinion. Among other things, plaintiffs' counsel has argued that it is a matter of "simple arithmetic and common sense."10 Neither common sense nor anything plaintiffs' counsel has offered to date, however, elucidates Paragraph 13 enough to enable us to conduct a meaningful review. In any event, plaintiffs' expert, not plaintiffs' counsel, is the proper person to explain his methodology.

We are persuaded that plaintiffs' expert's opinion was so vague and conclusory as to preclude defendant from effectively challenging the accuracy of the opinion and to preclude Judge Sweet from being able to specify the subsidiary facts underlying the ultimate finding of perpetuation of segregation. Because we are unable to discern the relevant matters from the record, we are unable to make a reasoned determination of whether Judge Sweet's finding of perpetuation of segregation in the system was "clearly erroneous." We therefore remand for proceedings consistent with this opinion.

Defendant has not appealed on grounds that Judge Sweet failed to conduct an evidentiary hearing,11 but it appears from the record that, in making his findings, Judge Sweet relied solely on the experts' affidavits. We note for purposes of remand that while affidavits may be considered on a preliminary injunction motion, motions for preliminary injunction should not be resolved on the basis of affidavits that evince disputed issues of fact. Forts v. Ward, 566 F.2d 849, 851 (2d Cir.1977). When a factual issue is disputed, oral testimony is preferable to affidavits. Id.; Schulz v. Williams, 38 F.3d 657, 658 (2d Cir.1994).

NYCHA also contends that Judge Sweet abused his discretion when, in ruling on NYCHA's motion for reconsideration, he rejected NYCHA's argument that any effect the proposed TSAP had on desegregation was de minimis. Considering that NYCHA had filed a notice of appeal prior to seeking reconsideration, it appears questionable whether Judge Sweet had jurisdiction to consider the motion. In any event, in view of our decision to remand for further proceedings, we need not address this issue other than to state that we consider it in the interest of justice that the court thoroughly address any arguments NYCHA offers to demonstrate that the proposed working family preference would, of itself, have only a de minimis segregative effect on the NYCHA system. The proper standard to be applied on remand is whether the proposed use of the working family preference will significantly perpetuate segregation at the relevant NYCHA developments. Gilmore v. City of Montgomery, 417 U.S. 556, 568, 94 S. Ct. 2416, 41 L. Ed. 2d 304 (1974); Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 938 (2d Cir.), aff'd, 488 U.S. 15, 109 S. Ct. 276, 102 L. Ed. 2d 180 (1988).

We leave the preliminary injunction intact until the district court has had the opportunity to address these matters on remand. Inverness, 819 F.2d at 51.