Langlois v. Abington Housing Authority

234 F. Supp. 2d 33 (2002)

Kelley LANGLOIS, et al., Plaintiffs,
v.
ABINGTON HOUSING AUTHORITY, et al., Defendants.

No. CIV.A. 98-12336-NG.

United States District Court, D. Massachusetts.

November 27, 2002.

Christopher Whittingham, Southeastern Massachusetts Legal Assistance Corporation, New Bedford, MA, Amy Copperman, Judith Liben, Massachusetts Law Reform Institute, Boston, MA, for plaintiffs.

Timothy H. White, White & White, Weymouth, MA, for defendants.

MEMORANDUM AND ORDER RE: SUMMARY JUDGMENT

GERTNER, District Judge.

TABLE OF CONTENTS

I. FACTUAL BACKGROUND ........................................................... 38
A. Facts ..................................................................... 38
1. The Section 8 Program .................................................. 38
a. 1980-1995: Federal Preferences ...................................... 39
b. 1995-1998: Administrative Suspension of the Mandatory
Preferences .......................................................... 39
c. OHWRA: Permanent Repeal of Mandatory Preferences
Coupled with Targeting of Extremely Low Income Families ............ 39
(1) 75% Rule ........................................................ 41
(2) Nondiscrimination ............................................... 41
(3) Affirmative Furtherance of Fair Housing ......................... 42
2. The Defendant PHAs' Policies ........................................... 42
a. The Application Process ............................................. 42
b. Formulation of the Waiting Lists: Residency Preferences ............. 43
3. The Plaintiffs' Interaction with the PHAs .............................. 43
B. Procedural History ........................................................ 44
II. LEGAL ANALYSIS ............................................................... 46
A. Summary Judgment Standard ................................................. 46
B. Causes of Action .......................................................... 46
1. The Blessing Test: § 1983, Statutes, and "Rights" ................. 47
2. Enforceability of Administrative Regulations and Executive Orders
under § 1983 ........................................................ 49
C. Abington and Rockland: Compliance with the 75% Rule ....................... 54
1. The 75% Rule: Does It Meet the Blessing Standard? ...................... 54
2. The 75% Rule: Does the Rockland Plan Comply? ........................... 54
D. Disparate Impact (Title VIII/Fair Housing Act) ............................ 55
1. The Prima Facie Case for Disparate Impact .............................. 56
a. The Preliminary Injunction Tests .................................... 57
(1) The "Four-fifths" Rule .......................................... 57
(2) The Comer Test .................................................. 58
b. The Summary Judgment Stage .......................................... 58
(1) The Track Records of Pre-1998 Residency Preferences ............. 59
(2) The Effect of Pre-existing Residency Preferences on
Current Participation ............................................. 59
(3) The Effect of the Preliminary Injunction ........................ 60
(4) The Measure of Delay ............................................ 61
c. Conclusions ......................................................... 62
2. The Prima Facie Case for Disparate Impact: Application
Procedures ............................................................... 64
3. The "Simple Justification Test" ........................................ 66
a. Statutory Purpose and Intent ........................................ 67
b. The Broader Statutory Context: Local Needs Harmonized with
Civil Rights Enforcement ................................................ 68
c. Proving More than Just "Residential Preferences Are Allowed" ........ 69
4. Less Discriminatory Alternatives ....................................... 70
E. Duty to Affirmatively Further Fair Housing ................................ 70
1. Does the "Affirmatively Further" Standard Comply with Blessing
and Other Procedural Prerequisites? ........................................ 71
a. 42 U.S.C. § 1437c-1(d)(15) (QHWRA) ................................71
b. 42 U.S.C. § 3608(e)(5) (Title VIII) ..............................71

(1) The Statute ................................................ 71
(2) The Regulations ............................................ 75
c. Executive Orders 11063 & 12892 ................................. 76
2. Did the PHAs' Plans Meet the Affirmative Furtherance
Requirement ......................................................... 77
F. Miscellaneous Provisions ............................................. 78
1. 42 U.S.C. § 3604(c): Advertising ............................. 78
2. PHAs' Compliance with Section 8 Plans ............................. 79
III. CONCLUSION ............................................................ 79

This case challenges the administration of Section 8 rental assistance programs by eight Massachusetts suburban public housing authorities ("PHAs").^[1] The plaintiffs are four racial minority, lower-income women residing in Randolph and Brockton, the class of similarly situated individuals that they represent,^[2] and the Massachusetts Coalition for the Homeless ("MCH").^[3] The communities in which the PHAs are situated are predominantly white, with low percentages of racial minority residents and a low overall rate of poverty.

The plaintiffs' principal challenge is to the use of a preference for local residents in determining a Section 8 applicant's position on the waiting list, as well as the logistics of the application processes. Suing under 42 U.S.C. § 1983, they maintain that the PHAs' policies effectively discriminated against minorities by favoring local, predominantly white applicants and violated the PHAs' duties to target housing to extremely low-income families, and to "affirmatively further" fair housing.

Defendants counter on a number of fronts, both procedural and substantive. They claim that recent decisional law casts doubt on the use of § 1983 to enforce the very statutes, regulations, and executive orders plaintiffs cite. Moreover, even if § 1983 is available to enforce these provisions, defendants deny that they have violated any of their legal obligations under them. Rather, they allege that the 1998 Quality Housing and Work Responsibility Act ("QHWRA") allows them to administer Section 8 housing as they have done, namely with preferences for local residents.

The following issues, then, are joined:

(1) Can § 1983 be used to enforce the particular statutes, regulations, and orders at issue here? I conclude that § 1983 enforcement is proper here, based on the language of the statutes and their implementing regulations, on the one hand, and § 1983's unique history and expansive language, on the other. Of all the cases in which § 1983 has been used to enforce state statutes and regulations, few come closer to the statute's purpose and history since its enactment post Civil War than cases like the instant one—cases that seek to enforce antidiscrimination laws in general, and fair housing provisions in particular.

(2) Have the plaintiffs met the standard for summary judgment on the question of whether the law has been violated, particularly given the 1998 Act's provisions on local housing preferences? I conclude that—with certain limitations described below—the plaintiffs have proven such violations, notwithstanding the language of the QHWRA. The Act's reference to "local needs and priorities" does not give the PHAs carte blanche to effect preferences for local residents. Rather, it was intended to encourage local authorities to determine the kind of housing mix that suits local needs, within the framework of the applicable laws.

For the following reasons, Langlois et al.'s motion for summary judgment [docket entry # 76] is GRANTED IN PART AND DENIED IN PART, and Abington Housing Authority et al.'s motion for summary judgment [docket entry # 86] is GRANTED IN PART AND DENIED IN PART.

I. FACTUAL BACKGROUND

A. Facts

1. The Section 8 Program

In 1974, Congress established the Section 8 program as a vehicle for providing an adequate supply of housing for low-income families. Housing and Community Development Act of 1974, Pub.L. No. 93-383, tit. II, § 8(a), 88 Stat. 633, 662 (codified at 42 U.S.C. § 1437f).^[4] The component of the Section 8 program at issue here has been known variously as the Section 8 Existing Housing Program, the Certificate/Voucher^[5] Program, or the Section 8 "Tenant-Based" Program. See 42 U.S.C. § 1437f(o) (2002); Comer v. Cisneros, 37 F.3d 775, 781 (2d Cir.1994). The Program provides subsidies to private landlords. In Massachusetts, about 44,000 Section 8 tenant-based subsidies are administered by about 130 local housing authorities ("PHAs") and the Massachusetts Department of Housing and Community Development.

Under the tenant-based provisions of the Section 8 Program, the United States Department of Housing and Urban Development ("HUD") enters into annual contracts with PHAs to fund Section 8 vouchers. See 42 U.S.C. § 1437f(o). In connection with this contract, each PHA must submit a plan to HUD describing the administrative details of the Section 8 program and its compliance with federal and state equal housing requirements. 42 U.S.C. §§ 1437c-1(b) (annual plan requirement), 1437c-1(d)(15) (2002) (civil rights certification requirement); 24 C.F.R. §§ 982.53, 982.54; see also 60 Fed. Reg. 34660, 34661 (July 3, 1995).^[6]

To participate in a Section 8 program, an eligible family^[7] may apply to any state PHA. Each PHA maintains a waiting list and then awards subsidies—in the form of vouchers, as they become available—to eligible participants on its list. 42 U.S.C. § 1437f(o)(13)(J). Generally, participants may use their vouchers to rent eligible dwelling units anywhere in the state with rents meeting the fair market rent standards set by HUD. 42 U.S.C. § 1437f(r)(1). A Section 8 family conveys its voucher together with the family's previously determined rent contribution—generally about 30% of its income—to the landlord, who in turn forwards the voucher to the PHA for the remainder of the rent. 42 U.S.C. §§ 1437a(a), 1437f(o)(2)(A)(i). HUD then reimburses the local PHA. 24 C.F.R. § 982.151 (annual contributions contract).

The federal administrative requirements for the PHAs have shifted over the past several years. It is the latest change in the federal requirements that the PHAs identify as the basis for the disputed changes they have made in their programs. It is useful to outline that history:

a. 1980-1995: Federal Preferences

From 1980 until 1995, Congress required the PHAs to rank eligible families on their waiting lists according to three statutorily mandated selection preferences ("federal preferences"): (1) families who occupy substandard housing (including the homeless); (2) families who are involuntarily displaced; and (3) families who are paying more than 50% of their family income to rent. 42 U.S.C. §§ 1437d(c)(4)(A)(i), 1437f(o)(3)(B) (repealed October 21, 1998).^[8]

b. 1995-1998: Administrative Suspension of the Mandatory Preferences

These mandatory "federal preferences" were suspended from 1995 through 1998 in various HUD Appropriations Acts in which Congress gave PHAs some discretion to set their own admissions preferences. See Pub.L. No. 104-99, § 402(d), 110 Stat. 26, 41 (Jan. 26, 1996); Pub.L. No. 104-204, § 201(c)(2), 110 Stat. 2874, 2893 (Sept. 26, 1996); Pub.L. No. 105-65, § 201(d)(2), 111 Stat. 1344, 1364 (Oct. 27, 1997).

c. QHWRA: Permanent Repeal of Mandatory Preferences Coupled with Targeting of Extremely Low Income Families

On October 21, 1998, President Clinton signed the Quality Housing and Work Responsibility Act of 1998 ("QHWRA"), which permanently repealed the Section 8 federal preferences, giving PHAs discretion within certain limits to establish local preferences in their agency plans. Pub.L. No. 105-276, § 514(b), 112 Stat. 2461, 2547-48 (amending 42 U.S.C. § 1437f(d)(1)(A)) ("[F]or the purpose of selecting families to be assisted, the public housing agency may establish local preferences, consistent with the public housing agency plan submitted under section 5A by the public housing agency ...."); see also 42 U.S.C. § 1437f(o)(6)(A)(i) (2002) ("Each public housing agency may establish a system for making tenant-based assistance under this subsection available on behalf of eligible families that provides preference for such assistance to eligible families having certain characteristics ....").^[9] The question raised by this litigation is the scope of that discretion.

Both the statutory text and the legislative history reflect Congress' concern that federal over-regulation was impeding the PHAs' ability to function effectively. Specifically, Congress was concerned that federal preferences did not afford PHAs sufficient flexibility to channel assistance where the local need was greatest. See, e.g., Pub.L. No. 105-276, § 514(a)(1), 112 Stat. at 2547 (amending 42 U.S.C. § 1437d(c)(4) to require that PHAs develop systems of preferences "based upon local housing needs and priorities"). Congress also felt that removing the federal preferences and allowing preferences based on local needs would ultimately lead to diminished residential segregation by income. See, e.g., Pub.L. No. 105-276, § 502(a)(3), 112 Stat. at 2520 (noting Congress' finding that "the public housing system is plagued by a series of problems, including the concentration of very poor people in very poor neighborhoods"); id. § 502(b)(3), 112 Stat. at 2521 (articulating a statutory purpose of, inter alia, "facilitating mixed income communities and decreasing concentrations of poverty in public housing").

But the PHA's flexibility was not without limits. Senator Kerry described the rationale for repealing the previous federal preferences while still retaining the framework for the protection of the most needy applicants:

[T]he reform measure permanently repeals Federal preferences, which had the unintended consequence of concentrating poverty in public housing developments. The bill allows PHAs to develop their own preferences, including a preference for working families, but requires that at least 40 percent of all public housing units and 75 percent of all section 8 units that become available each year be provided to people making below 30 percent of area median income. These protections ... will benefit residents at all income levels by facilitating the creation of mixed income developments.

The value of mixed income developments cannot be overstated. Working families stabilize communities by offering hope and opportunity in environments of despair. In recognition of this important principle, the reform bill will require housing authorities to develop plans for the economic desegregation of their distressed communities. Each PHA must develop their plan in consultation with its residents, and all plans will be submitted to HUD for approval.

144 Cong. Rec. S11833-02, at S11839 (daily ed. Oct. 8, 1998).

It is worth noting that nowhere in the legislative history or in the statute itself is there reflected any concern that the previous system disadvantaged local residents, or any desire that the new system should afford them greater priority. In other words, while the statute allows PHAs to develop local preferences, nothing in the statute or legislative history shows any desire of Congress to elevate preferences for local residents over fair housing concerns.

There are several significant restrictions on the PHAs' flexibility: the 75% rule, the guarantee of nondiscrimination and the PHAs' duty to affirmatively further fair housing.

(1) 75% Rule

First, in order to maintain protection for the most needy of applicants, the QHWRA includes an admissions-targeting requirement for extremely low-income families (the "75% Rule"):

Of the families initially provided tenantbased assistance under section 1437f of this title by a public housing agency in any fiscal year, not less than 75 percent shall be families whose incomes do not exceed 30 percent of the area median income, as determined by the Secretary with adjustments for smaller and larger families; except that the Secretary may establish income ceilings higher or lower than 30 percent of the area median income on the basis of the Secretary's findings that such variations are necessary because of unusually high or low family incomes.

42 U.S.C. § 1437n(b)(1) (codifying Pub.L. No. 105-276, § 513(a), 112 Stat. at 2545-46 (1998)); 24 C.F.R. § 982.201(b)(2).^[10]

(2) Nondiscrimination

A second set of restrictions ensures that local preferences comport with principles of nondiscrimination. PHAs must certify that their plans comply with civil rights laws, including Title VI of the Civil Rights Act of 1964 and Title VIII of the Civil Rights Act of 1968 (the "Fair Housing Act"), 42 U.S.C. § 3601 et seq. 42 U.S.C. § 1437c-1(d)(15).

Within this framework, the QHWRA and its accompanying regulations authorized the PHAs, when compiling their waiting lists for vouchers, to develop their own "system[s] of local preferences for selection of families admitted to the program." 24 C.F.R. § 982.207(a)(1) (2002). Such systems "must be based on local housing needs and priorities, as determined by the PHA." Id. § 982.207(a)(2) (2002).

At the same time, PHAs are explicitly forbidden to adopt residency requirements, and although the regulations allow them to adopt residency preferences, they may do so "only ... in accordance with non-discrimination and equal opportunity requirements listed at 5.105(a) of this title." Id. § 982.207(b)(1)(i) (emphasis added). Significantly, the PHAs are required to specify in their administrative plans that any residency preferences they adopt "will not have the purpose or effect of delaying or otherwise denying admission to the program based on the race, color, ethnic origin, gender, religion, disability, or age of any member of an applicant family." Id. § 982.207(b)(1)(iii).

(3) Affirmative Furtherance of Fair Housing

The actions of the PHAs must also take place within the framework of the duty to affirmatively further fair housing, as stated in 42 U.S.C. § 1437c-1(d)(15) (United States Housing Act) and its implementing regulation, 24 C.F.R. § 903.7(o); 42 U.S.C. § 3608(e)(5) (Fair Housing Act), and its implementing regulation, 24 C.F.R. § 982.53(b)-(c); Executive Order 11063, 27 Fed.Reg. 11527 (Nov. 20, 1962) and its implementing regulations, 24 C.F.R. §§ 5.105(a) and 107.21; and Executive Order 12892, 59 Fed.Reg. 2939 (Jan. 17, 1994).

2. The Defendant PHAs' Policies^[11]

The defendants in this case are eight public housing authorities located in the suburban South Shore area of Massachusetts. The communities in which the PHAs are situated are predominantly white, with relatively low percentages of racial minority residents (ranging from 1.4% to 8.5% minority) and a low overall rate of poverty (ranging from 3.8% to 6.8% of the population in poverty).^[12]Langlois, 1998 WL 1029208, at *4 n. 13. By way of comparison, Brockton, Massachusetts, a large neighboring city where named plaintiffs Fabian, Rivera, and Stewart reside, is 22.5% minority, with 13.6% of its population living in poverty. Id.

In 1998, after determining that their current Section 8 waiting lists would soon be exhausted, the eight defendant PHAs consulted with each other and decided jointly to open up their waiting lists to new applicants via a new round of lotteries. Id. In so doing, they adopted, inter alia, the following policies and procedures, in which local residents had a priority position.

a. The Application Process

The eight PHAs, in an effort to generate a high rate of participation in the application process while placing some limit on their respective administrative burdens, all decided that they would require prospective Section 8 applicants to request their applications during the same two-day period. In October 1998, the PHAs issued a public notice^[13] stating that applications could be requested from each individual PHA, either in person or over the phone, between 9:30 a.m. and 2:30 p.m. on October 29 and 30, 1998. Id. at *5. The applications were to be returned by hand or postmarked no later than 12:00 noon on November 17, 1998. Each PHA would then hold its separate lottery on December 1, 1998, at 1:00 p.m.

In addition to the procedures set forth in the public notice, social service agencies and housing advocates were permitted either to telephone each PHA as often as they could and request five applications with each telephone call, or to send an employee to each PHA to pick up as many applications as they needed. Id. at *5. The success of these support organizations varied, depending on their access to manpower, transportation, time, and the happenstance of getting through on the telephone. Id. Moreover, some of the PHAs apparently did permit organizations to fax or mail in a list of applicants' names, rather than requiring the workers to submit the lists in person. Id. at *5 n. 29.

b. Formulation of the Waiting Lists: Residency Preferences

The PHAs also adopted a procedure for determining which applicants would make the waiting lists and in what order. The first stage of the process consisted of an individual randomized lottery for each PHA that would determine which applicants would make the respective waiting lists for the available Section 8 vouchers. Id. at *4-5. Once that lottery was complete, the PHAs would then place the selected applicants in order on their Section 8 waiting lists. Significantly, the PHAs gave priority to applicants currently living or working in the community where the PHA was located. Id. at *5. In other words, once it was established who made the waiting list and who did not, local residents on the waiting list moved to the front of the line for receipt of a Section 8 voucher.

For some of the PHAs, this residency preference was initiated in 1998; for others, the preference had been in place before, in some cases as early as the 1970s.

3. The Plaintiffs' Interaction with the PHAs

The named plaintiffs in this case—Kelley Langlois, Yasmine Rivera, Lissett Fabian, and Annette Stewart—are all extremely low-income women of color. In 1998, they were either homeless or had severe housing problems. Id. at *6. Langlois currently resides in Randolph, MA; Rivera, Fabian, and Stewart reside in Brockton, MA. These four women, along with other members of the class they represent —i.e., extremely low-income individuals residing in the Brockton and Boston metropolitan areas—attempted to participate in the defendant PHAs' lottery systems in October 1998 and encountered various logistical and substantive problems with the PHAs' procedures. Id. at *6 n. 25.

For example, both applicants and housing advocates reported serious difficulties obtaining lottery applications. Relatively few of the applicants own cars—advocates estimate that between 0 and 10% of their clients have cars—and it would have been extremely difficult, if not impossible, to visit all eight PHAs by public transportation within the two-day window allotted for picking up applications. Some housing advocates and social workers with access to cars were able to pick up applications, but not all of them had the resources to drive to all eight PHAs. The telephone option did not altogether alleviate the situation: both applicants and advocates reported that it was often difficult to reach some of the PHAs by phone during the two-day window. Moreover, some advocates claimed that, when they did get through on the phone, they were allowed to request only a certain number of applications during that single telephone call and were directed to call back if they wanted more.

In addition, even when those seeking applications were able to get them, the application forms each contained notice of the PHA's intention to give waiting list priority to those applicants living in its own community. None of the plaintiffs reside in any of the defendant PHAs' communities, and none of them would have qualified for the residency preference. Notice of the PHAs' residency preferences discouraged many potential applicants from bothering to apply at all.

On October 30, 1998, the plaintiffs' counsel wrote to the Executive Directors of the eight PHAs to express their concerns about the application procedures and to suggest that the procedures might violate state and federal fair housing and antidiscrimination laws. This lawsuit followed.

B. Procedural History

The case at bar was filed on November 16, 1998. The complaint, as later amended on December 1, 1998, originally listed five counts for relief.^[14] However, at the present time, only three counts remain: Count I (violation of federal Fair Housing Act, 42 U.S.C. § 3608(e)(5); QHWRA § 511(a); Executive Order 11063, 27 Fed.Reg. 11527 (Nov. 20, 1962); and Executive Order 12892, 59 Fed.Reg. 2939 (Jan. 20, 1994)), Count II (violation of various provisions of QHWRA), and Count IV (42 U.S.C. § 1983).

Three days after filing suit, on November 19, 1998, the plaintiffs moved for class certification^[15] and for a temporary restraining order ("TRO") allowing the defendant PHAs to proceed with the randomized housing lotteries scheduled for December 1, but prohibiting any other action, including ordering the waiting list according to the residency preference or distributing any available Section 8 certificates. This Court issued the requested TRO on November 30, 1998, pending resolution of the plaintiffs' motion for preliminary injunction.

After examining the data on the outcome of the December 1 lotteries, I issued findings of fact on December 16. The facts suggested to me that a preliminary injunction was warranted, based on the likelihood that the procedures of Abington and Rockland PHAs, if fully implemented, would violate the 75% Rule and that the proposed residency preference in Avon, Holbrook and Middleborough, if implemented, would work a discriminatory disparate impact on minority applicants.^[16] A preliminary injunction issued on December 30, 1998, by which, inter alia,

(1) the Abington and Rockland PHAs were enjoined from distributing any of their available Section 8 certificates to applicants on their December 1 waiting lists according to their residency preference policies until they presented this Court with a specific plan to ensure compliance with the 75% Rule^[17];

(2) the Avon, Holbrook, and Middleborough PHAs were enjoined from distributing any of their available Section 8 certificates to applicants on their December 1 waiting lists according to their residency preference policies;

(3) the Abington, Avon, Holbrook, Middleborough, and Rockland PHAs were ordered to begin distribution of their Section 8 certificates, as they became available, to applicants on their December 1 waiting lists according to their randomized ranking scheme, i.e., without residential preferences;

(4) all of the defendant PHAs were permitted to begin immediate distribution of available Section 8 certificates to the households on their pre-existing waiting lists; and

(5) the Bridgewater, Halifax, and Pembroke PHAs were ordered to report to the Court after they exhausted their pre-existing waiting lists, and prior to any distribution of certificates to applicants on their December 1 waiting lists.

The defendants appealed the entry of the preliminary injunction to the First Circuit on February 10, 1999. On March 27, 2000, the First Circuit affirmed the preliminary injunction enjoining Abington and Rockland from distributing Section 8 vouchers to their December 1 lists because of lack of compliance with the 75% Rule. Langlois v. Abington Housing Auth., 207 F.3d 43, 52 (1st Cir.2000). The First Circuit also affirmed the injunction against the Avon, Holbrook, and Middleborough PHAs, which had been based on a finding that there was a likelihood of success on the merits of plaintiffs' disparate-impact claim, although the First Circuit parted company with this Court on its legal analysis. Id.

On the disparate impact issue, the First Circuit held that the Fair Housing Act prohibits "actions that have an unjustified disparate racial impact." Id. at 49. However, while the First Circuit left undisturbed the trial court's conclusion that the defendants' plans had a disparate impact on the plaintiffs, that finding served only to create a prima facie case of a Fair Housing Act violation. The next stage of the analysis dealt with whether the demonstrated disparate impact was "justified by a legitimate and substantial goal of the measure in question." Id. at 50-51. Insofar as the trial court's rationale for granting the preliminary injunction appeared to rest on "balancing" the merits of that goal against the magnitude of the disparate impact, the First Circuit rejected that approach, observing that a "balancing" role for the district court would violate separation of powers vis-à-vis Congress and is "in tension with the course taken by the Supreme Court and Congress under Title VII where a standard of justification is constructed and applied." Id. at 51.

Nevertheless, the First Circuit ordered that within 90 days the injunction was either to be vacated or "reinstituted based upon the grounds not yet considered by the district court." Id. at 52.^[18] Those grounds consisted of alleged violations of the duty to "affirmatively further fair housing," under 42 U.S.C. § 1437c-1(d)(15); 24 C.F.R. § 982.53(b)-(c); 24 C.F.R. § 903.7(o); and Executive Orders 11063 and 12892, which had not been addressed in the preliminary injunction proceedings.

The plaintiffs filed their motion for summary judgment on September 8, 2000; the defendants cross-moved for summary judgment on October 24, 2000. The first motion hearing was held on November 28, 2000. On June 4, 2001, the Court requested supplemental briefing from the parties, in light of recent developments in the law on § 1983 enforcement, and a second motion hearing was held on June 7, 2002.

II. LEGAL ANALYSIS

A. Summary Judgment Standard

Summary judgment is warranted only "if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c). In assessing a motion for summary judgment, I am required to view the facts in the record in the light most favorable to the nonmoving party, drawing all reasonable inferences in that party's favor. Thomas v. Eastman Kodak Co., 183 F.3d 38, 42 (1st Cir.1999); see also Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S. Ct. 2548, 91 L. Ed. 2d 265 (1986).

B. Causes of Action

The plaintiffs' legal claims stem from two alleged "wrongs": the PHAs' proposed implementation of residency preferences and the logistics of the lottery application process. At this stage of the litigation, all of the plaintiffs' remaining claims are brought under the jurisdictional umbrella of 42 U.S.C. § 1983. In other words, the plaintiffs allege no implied right of action for each of their legal claims; rather, they sue under § 1983, claiming deprivations of rights secured by the respective federal laws allegedly violated by defendants.

Those rights fall into five basic categories:

(1) With regard to the Abington and Rockland PHAs, violation of the 75% Rule, as stated in 42 U.S.C. § 1437n(b)(1) (QHWRA) and 24 C.F.R. § 982.201(b)(2);

(2) With regard to Avon, Holbrook and Middleborough, violations of the duty to affirmatively further fair housing, as stated in 42 U.S.C. § 1437c-1(d)(15) (United States Housing Act) and its implementing regulation, 24 C.F.R. § 903.7(o); 42 U.S.C. § 3608(e)(5) (Fair Housing Act) and its implementing regulation, 24 C.F.R. § 982.53(b)-(c); Executive Order 11063, 27 Fed.Reg. 11527 (Nov. 20, 1962), and its implementing regulations, 24 C.F.R. §§ 5.105(a) and 107.21; and Executive Order 12892, 59 Fed.Reg. 2939 (Jan. 17, 1994);

(3) With regard to Avon, Holbrook and Middleborough, utilization of policies that have a discriminatory disparate impact on minority applicants and/or violate other civil rights laws, in violation of 42 U.S.C. § 3604(a)-(b) (Fair Housing Act) and 24 C.F.R. § 982.207(b);

(4) With regard to all of the PHAs, discriminatory advertising with respect to the sale or rental of a dwelling, as prohibited in 42 U.S.C. § 3604(c) (Fair Housing Act); and

(5) With regard to all of the PHAs, violation of HUD regulations at 24 C.F.R. §§ 982.54, 982.204(a)-(b), for the PHAs' failures to comply with their own Section 8 administrative plans and to maintain waiting lists including racial and other data.

Before launching into the merits of each of these claims, defendants raise an important threshold question, namely, which of the federal laws cited by the plaintiffs may support claims for relief under § 1983.

In this regard, defendants suggest that the ground may have shifted under the plaintiffs since they initially brought their claim. Recent case law, they argue, has substantially narrowed the scope of § 1983, particularly for statutory and administrative claims. I will proceed as follows: First, I will set forth the standards for assessing the viability of a cause of action under § 1983. Then I will address each category of the plaintiffs' claims in turn, first whether the statutory, regulatory, and other provisions at issue meet the standard for creating rights cognizable under § 1983, and if so, the merits of those claims.

1. The Blessing Test: § 1983, Statutes, and "Rights"

In Blessing v. Freestone, 520 U.S. 329, 117 S. Ct. 1353, 137 L. Ed. 2d 569 (1997), the Supreme Court set forth its interpretation of the scope of § 1983. It reaffirmed the fact that § 1983 "safeguards certain rights conferred by federal statutes," and accordingly, that § 1983 suits may be premised on deprivations of such rights. Id. at 340, 117 S. Ct. 1353 (citing Maine v. Thiboutot, 448 U.S. 1, 100 S. Ct. 2502, 65 L. Ed. 2d 555 (1980)).

At the same time, the Blessing court stressed that the § 1983 remedy is available for violations of federal rights, not just of federal statutes per se. Id. at 340, 117 S. Ct. 1353. Rights enforceable under § 1983 center on the needs of particular people, thereby reflecting "individual entitlements" as opposed to yardsticks for measuring "systemwide performance." Id. at 343, 117 S. Ct. 1353 (emphases in original); see also Gonzaga Univ. v. Doe, 536 U.S. 273, 122 S. Ct. 2268, 2274, 153 L. Ed. 2d 309 (2002). The Blessing court, extrapolating from its earlier case law, outlined a three-factor test for determining when a federal statute gives rise to such a right:

First, Congress must have intended that the provision in question benefit the plaintiff. Second, the plaintiff must demonstrate that the right assertedly protected by the statute is not so "vague and amorphous" that its enforcement would strain judicial competence. Third, the statute must unambiguously impose a binding obligation on the States. In other words, the provision giving rise to the asserted right must be couched in mandatory, rather than precatory, terms.

Id. at 340-41, 117 S. Ct. 1353 (citations omitted); see also, e.g., Visiting Nurse Ass'n of North Shore, Inc. v. Bullen, 93 F.3d 997, 1002-03 (1st Cir.1996).

Significantly, the inquiry was whether the statute created a federal right. It was not whether Congress intended to create a right enforceable through § 1983. Section 1983 by itself announced Congress' intention to provide a remedy for state violations of rights. As this Court noted in Furtick v. Medford Housing Auth., 963 F. Supp. 64 (D.Mass. 1997): "Congress is presumed to legislate against the background of section 1983 and thus to contemplate private enforcement of the relevant statute against state and municipal actors absent fairly discernible congressional intent to the contrary." Id. at 71 n. 18; see also Henry Paul Monaghan, Federal Statutory Review Under Section 1983 and the APA, 91 Colum. L.Rev. 233, 249 (1991).

A plaintiff's successful showing that a federal statute creates a cognizable individual right raises "only a rebuttable presumption that the right is enforceable under § 1983." Id. at 341, 117 S. Ct. 1353. At that point, the burden of proof shifts to the defendant to show that Congress "specifically foreclosed a remedy under § 1983." Id. at 341, 117 S. Ct. 1353 (quoting Smith v. Robinson, 468 U.S. 992, 1005 n. 9, 104 S. Ct. 3457, 82 L. Ed. 2d 746 (1984)) (internal quotation marks omitted). Congressional foreclosure may be either explicit—written into the statute itself—or implicit, when the reviewing court identifies a "comprehensive enforcement scheme that is incompatible with individual enforcement under § 1983." Id. (citing Livadas v. Bradshaw, 512 U.S. 107, 133, 114 S. Ct. 2068, 129 L. Ed. 2d 93 (1994)).

The burden of showing an implicit rejection of § 1983 remedies by dint of a "comprehensive enforcement scheme" is a substantial one. As one scholar notes, the "[i]mplied preemption of a section 1983 remedy on the basis of the assertedly comprehensive nature of the remedial scheme created by the federal legislation is not favored." Monaghan, supra, at 247. The mere availability of some administrative mechanism to protect the plaintiff's interest is not enough to foreclose recourse to § 1983. E.g., id. at 347, 117 S. Ct. 1353 (citing Golden State Transit Corp. v. Los Angeles, 493 U.S. 103, 106, 110 S. Ct. 444, 107 L. Ed. 2d 420 (1989)); Wright v. City of Roanoke Redevelopment & Housing Auth., 479 U.S. 418, 427-28, 107 S. Ct. 766, 93 L. Ed. 2d 781 (1987) ("[T]he existence of a state administrative remedy does not ordinarily foreclose resort to § 1983."). Rather, the administrative scheme must be both federal and "sufficiently comprehensive ... to demonstrate congressional intent to preclude the remedy of suits under § 1983." Id. at 346, 117 S. Ct. 1353 (quoting Wilder v. Virginia Hosp. Ass'n, 496 U.S. 498, 521, 110 S. Ct. 2510, 110 L. Ed. 2d 455 (1990)) (alterations in original) (internal quotation marks omitted).

It should be noted that the Blessing inquiry is a different "rights-creating" inquiry than the one undertaken by the Supreme Court in Alexander v. Sandoval, 532 U.S. 275, 121 S. Ct. 1511, 149 L. Ed. 2d 517 (2001). In Sandoval, the "rights-creating language" that the Court found absent from § 602 of Title VI referred to a private right of action flowing directly from the statute. Id. at 288-89, 121 S. Ct. 1511. If courts imply a right of action from a statute, when Congress itself provided no explicit remedies, separation of powers concerns are arguably raised.

Court implementation of § 1983, in contrast, does not trigger the same level of concern for the separation of powers. Section 1983 is a statute of considerable breadth, passed by Congress after the Civil War to provide judicial remedies for state violations of federal law and, in particular, to provide federal remedies for discrimination against racial minorities:

It is plain that Congress intended to create new enforcement mechanisms to redress state violations of federal law. The experience following the Civil War called for a dramatic expansion in the role of the national government, including the federal courts, in regulating the conduct of the states. It is no doubt true that Congress was primarily concerned with providing a remedy for constitutional violations and unlawful invasions of rights protected by civil rights laws. But it is consistent with the historical evidence to understand the underlying purposes as more general than that, reaching all violations of federal law.

Cass R. Sunstein, Section 1983 and the Private Enforcement of Federal Law, 49 U. Chi. L.Rev. 394, 408-09 (1982).

As I describe below, the statutes at issue in this case, especially those concerned with housing discrimination, fit squarely within the language of § 1983, its purpose, and its historical mission.

2. Enforceability of Administrative Regulations and Executive Orders under § 1983

Enforcement of administrative regulations and executive orders directly under § 1983, the defendants suggest, does raise separation of powers concerns not unlike those raised in Sandoval. While there is some ferment in this area, existing Supreme Court law suggests a broad rule for § 1983 enforceability of regulations: Where an administrative regulation is validly issued pursuant to a congressional delegation of authority and carries the force of law, it may serve as the premise for a § 1983 suit so long as the regulation itself meets the Blessing criteria. Other case law suggests a narrower view: Only statutes confer rights under § 1983; regulations can provide an interpretation of the statute, to which deference is due. See Chevron U.S.A., Inc. v. Natural Resources Defense Council, 467 U.S. 837, 844, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984). While the difference between the two is less than clear,^[19] as I describe below, plaintiffs succeed under either test.

The broad view is suggested in Wright v. City of Roanoke Redevelopment & Housing Auth., 479 U.S. 418, 107 S. Ct. 766, 93 L. Ed. 2d 781 (1987). Wright involved a challenge by tenants of low income housing projects who claimed that utility overcharges violated the Brooke Amendment to the United States Housing Act of 1937, 42 U.S.C. § 1437(a). The Brooke Amendment imposed a ceiling for rents charged to tenants of low income housing projects; their rent had to be a specified percentage of their income. Id. at 420, 107 S. Ct. 766. In the Act's implementing regulations, the Department of Housing and Urban Development ("HUD") defined the statutory term "rent" as including sums for reasonable utility costs. Id. at 420-21 & nn. 3-4, 107 S. Ct. 766 (quoting 24 C.F.R. § 860.403 (1982) and citing 24 C.F.R. § 865.470 (1983)). Arguing from both the statute and the regulations, the tenants claimed that the housing authority had exceeded the statutory ceiling by overbilling them for utilities.

The Court agreed. Id. at 430, 107 S. Ct. 766. The Court found that the statute met the first and third prongs of Blessing, that the Brooke Amendment imposed a rent ceiling intended to benefit the plaintiffs, and that it was a binding obligation on the states. Further, using language found in the narrow test, the court concluded that HUD's regulations comprised an interpretation of the term "rent" that was entitled to deference. Id. With respect to the second prong of the Blessing test, however— whether the regulations were too amorphous and thus, beyond the competence of the Court to enforce—the Court's language was broad. It suggested that the regulations may be an independent source of rights under § 1983 so long as they have the "force of law." Id. at 431-32, 107 S. Ct. 766.

Justice O'Connor dissented, arguing for the narrow approach—that Congress had conferred § 1983 rights only under the statute and not under the regulations. While deference was due to the agency's interpretation of those rights through its regulations, the regulations were not independently actionable. The broader view— that "any regulation adopted within the purview of the statute creates rights enforceable in federal court," she suggested, was "troubling indeed." Id. at 437-38, 107 S. Ct. 766 (O'Connor, J., dissenting) (emphasis in original).^[20]

Three years later, in Wilder v. Va. Hospital Ass'n, 496 U.S. 498, 110 S. Ct. 2510, 110 L. Ed. 2d 455 (1990), the Wright holding was reaffirmed. Without much analysis, however, the Court characterized its earlier holding in Wright as suggesting that "the [statute] and its implementing regulations did create rights enforceable under § 1983." Id. at 511, 110 S. Ct. 2510. Likewise, ten years later, in his dissent in Alexander v. Sandoval, Justice Stevens characterized Sandoval's denial of relief on respondents' direct cause of action under Title VI as "something of a sport" because "[l]itigants who in the future wish to enforce the Title VI regulations against state actors in all likelihood must only reference § 1983 to obtain relief." 532 U.S. at 299-300, 121 S. Ct. 1511 (Stevens, J., dissenting); see also Guardians Ass'n v. Civil Serv. Comm'n of New York, 463 U.S. 582, 638, 103 S. Ct. 3221, 77 L. Ed. 2d 866 (1983) (Stevens, J., dissenting) ("[I]t is clear that the § 1983 remedy is intended to redress the deprivation of rights secured by all valid federal laws, including statutes and regulations having the force of law.").

The lower federal courts have split on the meaning of this case law. The Sixth Circuit, and the Ninth Circuit, together with district courts in Michigan, California, and Florida, have followed the broader rule. See, e.g., Loschiavo v. City of Dearborn, 33 F.3d 548, 551 (6th Cir.1994) ("As federal regulations have the force of law, they likewise may create enforceable rights"); see also Buckley v. City of Redding, California, 66 F.3d 188, 192 (9th Cir.1995),^[21]Hill v. San Francisco Housing Auth., 207 F. Supp. 2d 1021, 1027 (N.D.Cal. 2002),^[22]White v. Engler, 188 F. Supp. 2d 730, 745 (E.D.Mich.2001).

Other courts (this circuit and this Court not among them) have taken the view suggested by Justice O'Connor's dissent in Wright—that regulations may help define the scope of the statutory right created by Congress, but they do not independently create rights. See S. Camden Citizens in Action v. New Jersey Dep't of Envtl. Protection, 274 F.3d 771, 790-91 (3d Cir.2001) (finding that environmental disparate-impact regulations adopted pursuant to Title VI of Civil Rights Act of 1964 § 602, 42 U.S.C. § 2000d-1, do not create rights enforceable under § 1983 because the right only appears in the regulation, not in the statute). See also Harris v. James, 127 F.3d 993, 1009 (11th Cir.1997); Smith v. Kirk, 821 F.2d 980, 984 (4th Cir.1987); Ceaser v. Pataki, No. 98 CIV.8532(LMM), 2002 WL 472271, at *3 (S.D.N.Y. Mar.26, 2002).

I find the broader view—also endorsed by the dissenting judge in South Camden, 274 F.3d at 797-98 (McKee, J., dissenting) —to be more consonant with Supreme Court doctrine, the academic commentary and significantly, the fundamental purpose of § 1983.^[23] On its face, § 1983's broad language encompasses regulations. Section 1983 provides:

Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory of the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in any action at law, suit in equity, or other proper proceeding for redress ....

42 U.S.C. § 1983 (emphasis added). While the plain meaning of "laws" is broad,^[24] the statutory text, as the Blessing court noted, incorporates the limitation that recourse to § 1983 is available only to redress the "deprivation of rights, privileges, or immunities" secured by the "laws" in question.

The sticking point appears to be the source of the right in question. The Third, Fourth, and Eleventh Circuits, as well as the dissenters in Wright, are troubled by the idea that a § 1983 suit could vindicate "rights" stemming not directly from Congress but from an agency's regulation. These courts have to be saying more than the truism that no action will lie to assert a right conferred by an agency acting outside its statutory authority. Rather, they must be suggesting that regulations that derive from broad delegations of congressional authority rather than narrow grants of authority are somehow too far removed from congressional action to meaningfully permit § 1983 enforcement.

But the lodestar under Blessing is congressional intent to confer rights. To say that such rights, contained in valid regulations issued pursuant to narrow congressional delegations, are enforceable, but that the same rights contained in valid regulations enacted pursuant to broad delegations are not, poses a distinction that exists nowhere in the language, the doctrine, the rationale or the history of § 1983. Moreover, the effect of such a distinction is quite paradoxical: The more latitude Congress expressly intended to allow the agency, the fewer § 1983 "rights" that agency may validly create through its regulations.

In any event, the distinction between the broad and narrow tests is beside the point here. The statutes at issue already contain the requisite "rights-creating" language. They meet the Blessing test, both in letter and spirit. Under those circumstances, it should make little difference whether Congress specifies the right even further in the statutory text or charges the agency to do so via regulations.

Moreover, two of the three claims involve statutory provisions concerned with discrimination—redressing illegal discrimination on the one hand, affirmatively furthering fair housing on the other. They are precisely the kinds of rights with which § 1983, passed immediately after the Civil War, was uniquely concerned. They are phrased in mandatory not precatory terms. And antidiscrimination provisions are hardly beyond the competence of courts to administer.^[25] Courts have been the place where such rights have been enforced for decades.

To the extent the fair housing statutes delegate details of enforcement to regulation, the regulations implementing these statutes are not materially different from court decisions implementing like statutes in other areas concerned with discrimination.^[26] Thus, however one characterizes the test—broadly as the Wright majority did, or more narrowly, as in the Wright dissent, the result is the same. The regulations^[27] here are either independent sources of rights because they have the force of law, or they represent HUD's authoritative interpretations of the statutory antidiscrimination rights, or they flowed from a statutory scheme that itself met the Blessing standard.^[28]

C. Abington and Rockland: Compliance with the 75% Rule

1. The 75% Rule: Does It Meet the Blessing Standard?

The QHWRA's "75% Rule" resides in the statute, 42 U.S.C. § 1437n(b)(1), which provides:

Of the families initially provided tenant-based assistance under section 1437f of this title by a public housing agency in any fiscal year, not less than 75 percent shall be families whose incomes do not exceed 30 percent of the area median income, as determined by the Secretary with adjustments for smaller and larger families ....

A straightforward application of Blessing to this provision recognizes that a private action under § 1983 to enforce the 75% Rule is appropriate.

The statute is crafted to ensure that a designated class of families—those described in HUD regulations as "extremely low income families"—comprises at least three quarters of the total number of families awarded assistance by a given PHA. The named plaintiffs all fit this classification, and the plaintiff class is precisely tailored to include only those individuals "with incomes at or below 30% of area median income." There is no doubt that Congress intended the 75% Rule to benefit the plaintiffs; the 75% Rule therefore passes the first component of the Blessing test.

Nor is the right so "vague and amorphous" that a court cannot enforce it. The statute gives an explicit directive to PHAs: make sure that 75% of the families awarded aid are at a certain level of income. The preliminary injunction issued against the Abington and Rockland PHAs in this case is demonstrative of this provision's enforceability in the courts. The available data showed that the wait-lists, as configured by Abington and Rockland, threatened to violate the 75% Rule. Langlois, 1998 WL 1029207, at *3-4. An order followed to ensure their compliance.

Finally, Congress took pains to phrase the 75% Rule "in mandatory, rather than precatory, terms." 42 U.S.C. § 1437n(b)(1) ("[N]ot less than 75 percent shall be families whose income exceeds 30 percent ...." (emphasis added)).

2. The 75% Rule: Does the Rockland Plan Comply?

As mentioned above, on April 4, 1999, the defendants submitted to this Court the "specific plan" that the Rockland and Abington Housing Authorities proposed to ensure compliance with the 75% Rule. The proposed amendment to both PHAs' administrative plans is as follows:

Of the families initially provided tenant based assistance under section 8 by the [Rockland/Abington] Housing Authority in any fiscal year, not less than 75% shall be families whose incomes do not exceed 30% of the area median income as determined by the Secretary. The local preference policy is subordinate to this provision, and admissions to the program will be monitored to assure compliance.

Abington has chosen to cease administration of Section 8. This analysis applies to Rockland only.

The plaintiffs respond that the proposed amendment is not at all "specific"; rather, it is just a statement of intended compliance without any details as to how compliance will take place. The First Circuit, in affirming the preliminary injunction against the Abington and Rockland PHAs, left it to this Court to resolve whether the amendment constitutes compliance with the 75% Rule. Langlois, 207 F.3d at 48-49.

I agree with the plaintiffs that the proposed amendment is essentially more of a promise of compliance than a plan for compliance. The amendment is not at all specific as to what statistical and procedural mechanisms the Abington and Rockland PHAs would use to ensure compliance.

At the same time, however, a given PHA's compliance with the 75% Rule does not strike me as requiring a particularly complex assessment. As I observed in my Memorandum and Order of December 30, 1998, the plaintiffs in this case were easily able to assess projected compliance by (1) counting the number of extremely low-income households due to receive Section 8 certificates and (2) assessing what percentage of those households would in fact receive certificates under the system in place at a given PHA. Langlois, 1998 WL 1029207, at *3-4. What the Rockland PHA's amendment appears to mean is that the 75% Rule will function as a backstop in formulation of the waiting lists, and that it will take precedence over any system of local preferences—as the law in fact requires.

At this point, however, I am willing to give the Rockland PHA the benefit of the doubt. Rockland is now on more than adequate notice as to what compliance means; there is little to gain from hamstringing it further at this point. Should it turn out that, despite the amendment to its administrative plan, the Rockland PHA still fails to comply with the 75% Rule, the plaintiffs are free to come back to court and seek an injunction requiring them to do so.

D. Disparate Impact (Title VIII/Fair Housing Act)

In relevant part, 42 U.S.C. § 3604 (Title VIII/Fair Housing Act) provides:

As made applicable by section 3603 of this title and except as exempted by sections 3603(b) and 3607 of this title, it shall be unlawful—

(a) To refuse to sell or rent after the making of a bona fide offer, or to refuse to negotiate for the sale or rental of, or otherwise make unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status, or national origin.

(b) To discriminate against any person in the terms, conditions, or privileges of sale or rental of a dwelling, or in the provision of services or facilities in connection therewith, because of race, color, religion, sex, familial status, or national origin.

There is no question that the Fair Housing Act meets the Blessing framework. As I have explained, § 1983 was passed after the Civil War as part of Congress's effort to give racial minorities ammunition to fight state-sponsored and private discrimination. Application of Blessing confirms what the history of § 1983 suggests: (1) the Fair Housing Act was enacted to protect racial minorities, (2) courts are well-positioned to enforce norms against discrimination and have done so for decades, and (3) Title VIII's prohibition of discrimination—and the PHAs' obligation to comply with them—are nondiscretionary.^[29]

Moreover, there is no question that Fair Housing Act claims may rest on a disparate impact theory of discrimination.^[30]Langlois, 207 F.3d at 49; see also cases cited supra note 25. The First Circuit held that once a plaintiff establishes a prima facie case of disparate impact, the question then becomes only whether the disparate impact is "justified by a legitimate and substantial goal of the measure in question"—the "simple justification test," id. at 51—and whether there is no less discriminatory means by which the defendant can meet that objective, Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 936 (2d Cir.), aff'd, 488 U.S. 15, 109 S. Ct. 276, 102 L. Ed. 2d 180 (1988); see also Langlois, 207 F.3d at 50 n. 6 (inviting the parties to review the question of less discriminatory alternatives on remand).

1. The Prima Facie Case for Disparate Impact: Residency Preferences

As the First Circuit observed in the appeal of the preliminary injunction in this case, the "Supreme Court has said that no single test controls in measuring disparate impact." Langlois, 207 F.3d at 50 (citing Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 995 n. 3, 108 S. Ct. 2777, 101 L. Ed. 2d 827 (1988)). It is my task, then, to review each PHA's residency preference against the several tests that are before me.

It is important, however, to put this analysis in context. As a simple matter of logic, the residency preference of every one of the defendant PHAs raises concerns. All of the communities in this case have significantly lower percentages of minority residents than their urban neighbors; all of them have fewer minority residents per capita than the state average. It follows logically, then, that any policy that facially favors the residents of these communities will disproportionately favor whites over minorities in the long run.

But the law requires proof of a substantial disparate impact. Fudge v. City of Providence Fire Dep't, 766 F.2d 650, 657-58 (1st Cir.1985). The various tests for disparate impact that the plaintiffs propose (the defendants themselves propose none) yield a variety of conclusions; occasionally a PHA's residency preference passes a particular test, sometimes there is insufficient data to apply a given test to a PHA. No single methodology is airtight, as the defendants hasten to point out, and I will address the imperfections of each test below. Nevertheless, I want to keep my eye on the context: Predominantly white suburban communities have enacted a Section 8 preference to bring their own residents to the front of the line for housing vouchers. That there will be a disparate impact over time seems clear. The question is the extent of that impact and whether it meets the applicable legal standards.

a. The Preliminary Injunction Tests

In this Court's Memorandum and Order of December 30, 1998, I applied, at the plaintiffs' urging, two different standards for proving disparate impact on minority applicants: the Equal Employment Opportunity Commission's "four-fifths rule" and the Comer standard. Langlois, 1998 WL 1029207, at *5-6. My focus at that time was on the 1998 wait lists and minority prospects for selection in the coming year.

(1) The "Four-fifths" Rule

The four-fifths rule provides that "a selection rate for any race, sex, or ethnic group which is less than four-fifths (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the Federal enforcement agencies as evidence of adverse impact ...." 29 C.F.R. § 1607.4(D). A typical application of the EEOC test would find fault if, say, 13% of white applicants were given employment as against 4% of black applicants (4% being less than four fifths of 13%). See Fudge, 766 F.2d at 658 n. 10 (applying the test in this manner, but noting that small sample size makes the comparison of percentages less reliable).

The EEOC test is particularly effective in employment cases where an applicant pool is ascertainable. In our case, however, it is not. The waiting lists generated by the 1998 lotteries do not necessarily reflect all qualified applicants. There is evidence that notice of the residency preference discouraged members of the plaintiff class from applying for the defendant PHAs' vouchers. The plaintiffs point to additional barriers to application—for example, the time constraints, the PHAs' requirements that applications be retrieved in person or by phone. Obviously, it is not possible to account for the number of thwarted or deterred potential applicants with any statistical certainty. Cf. U.S. v. City of Warren, Michigan, 138 F.3d 1083, 1092 (6th Cir.1998) (holding that the plaintiff's "inability to isolate the specific reason for the dearth of black applicants was not fatal to its claim" that recruiting practices worked a disparate impact against qualified minority applicants for employment). In addition, at the preliminary injunction stage, the variation of the four-fifths rule that I applied did not compare the selection rate of minorities to the selection rate of whites, as the EEOC test suggests. Instead I compared the percentage of minority applicants selected with the residency preference scheme to the percentage of minorities selected without the preference in place. Thus, if a particular PHA's selection rate for minority applicants with the residency preference in place was less than four fifths of its selection rate for minorities without the preference, the PHA would fail the EEOC test. Langlois, 1998 WL 1029207, at *6.

At that time, I found that the Avon (25% selection rate with the preference versus 38% without), Holbrook (0% with the preference versus 33% without), and Middleborough (40% with the preference versus 57% without) PHAs failed the four-fifths rule.

The plaintiffs submitted further evidence suggesting a disparate impact for Rockland (47% with the preference versus 73% without), but because I had already identified the QHWRA's 75% Rule as a sufficient ground to suspend Rockland's preference, id. at *4, *5 n. 21, I reserved ruling against Rockland on the disparate impact issue. Abington passed the four-fifths test (54% with the preference versus 63% without). The Bridgewater, Halifax, and Pembroke PHAs had not exhausted their existing wait lists and would not reach the lists compiled under the new regime in 1999. Due to the lack of evidence of any projected impact in the coming year, disparate or not, the preliminary injunction's disparate impact analysis did not take account of these three PHAs' programs.

(2) The Comer Test

The second test I applied at the preliminary injunction stage was borrowed from Comer v. Cisneros, 37 F.3d 775 (2d Cir. 1994). Comer involved challenges by minority residents of Buffalo to, inter alia, a suburban PHA's residency preference in distribution of Section 8 subsidies and the Buffalo PHA's practice of keeping innercity plaintiffs from using their subsidies in suburban communities.

In the Comer case the Second Circuit assessed disparate impact by comparing the percentage of minority families holding Section 8 certificates from the suburban housing authority with the percentage of minorities on the PHA's waiting list for housing. Id. at 793. At the preliminary injunction stage I was concerned with the immediate effect of the residency preferences, so I adjusted the Comer test to focus on the PHAs' projected selections for the coming year. A PHA's residency preference would pass muster if the percentage of minority representation in its projected 1999 selections was equal to or greater than minority representation on the entire wait list. My charge at this stage, however, is to determine the impact of residency preferences on minorities in future years.^[31]

At the preliminary injunction stage, Avon failed the Comer test because only 25% of the applicants to be admitted to the program in 1999 were minorities, as against a wait list that was 55% minority. Abington (minorities comprised 54% of projected admissions on a wait list that was 55% minority), Holbrook (0% share of admissions from a wait list that was 51% minority), and Rockland (47% share of admissions from a wait list that was 56% minority) also failed the test. Again, no data was available for Bridgewater, Halifax and Pembroke, as they had not yet exhausted their wait lists.

On appeal, the First Circuit found no reason to disturb these findings. Langlois, 207 F.3d at 50. It did, however, make brief comment on the methodology, and the EEOC rule in particular. The panel took note of the defendants' objection that the four-fifths rule—as with most statistical formulas—is "less reliable when the numbers are small; and the numbers in each of the PHAs in this case are fairly small." Id. (citation omitted). Ultimately, however, this alleged deficiency was not enough to overturn this Court's finding that the plaintiffs were likely to succeed on the merits of the disparate impact issue, particularly given the prospect of expert testimony on the subject. Id.

b. The Summary Judgment Stage

Taking the First Circuit's cue, the plaintiffs have come forward with the opinions and analysis of an expert witness, Professor Chris Tilly. Dr. Tilly has reviewed the available data as to each PHA, and the plaintiffs now propose a battery of four entirely new tests for disparate impact. The defendants have offered no expert.

(1) The Track Records of Pre-1998 Residency Preferences

First, the plaintiffs offer evidence of the effect of prior (namely pre-1998) residency preferences upon Section 8 selection. In effect, with this approach, the plaintiffs seek to develop a track record of past practices, implying that such past practices are a reasonable prediction about the impact of post-1998 residency preferences. This is a viable methodology only to the extent that the PHAs maintained such preferences and there is data on race and residency available to assess their impact. Only Abington, Pembroke, and Rockland are reviewable here, for two reasons: (1) only these PHAs maintained residency preferences prior to 1998 and (2) only these PHAs kept sufficient data on the race, residency, and disposition of their applicants. Avon and Middleborough had no preferences in place before 1998, and Halifax, Bridgewater, and Holbrook had preferences but did not keep data sufficient to show the racial impact of their residency preferences upon the selection process.

The story that the plaintiffs tell about these PHAs is compelling. Looking to Abington's most recent reorientation of its wait list, in 1996, the plaintiffs find that implementation of the residency preference pushed back the distribution of vouchers to nonresidents by fourteen months. All six of the Abington residents moved to the top of the list were white. Seven minorities ultimately received vouchers, albeit over a year later, as opposed to the eleven who would have without the residency preference in place. Similarly, Pembroke gave preference to twelve residents in 1996. Nine of the preferred residents were white; thus, while 54% of Pembroke's wait list were minorities, the group brought to the top of the list was 75% white. It took Pembroke thirteen months to reach the nonresidents on its list. Ultimately, because 54% of the wait list was minority, Pembroke should have given vouchers to thirteen minority candidates, but because of the residency preferences, only seven received vouchers. Rockland, the plaintiffs suggest, demonstrated no disparate impact in this analysis.

(2) The Effect of Pre-existing Residency Preferences on Current Participation

The plaintiffs' second test also examines the past performance of the defendant PHAs' pre-1998 residency preferences, as presently reflected in each PHA's roster of individuals who actually hold vouchers. This test compensates for the limitations of the first. Whereas the first test was able to address only those towns that had a track record of selection and data about the racial makeup and residency of its applicants, this test records the minority presence among current voucher holders (all of whom were selected under preexisting residency preferences), and compares it to the percentage of minorities on the 1998 wait list.^[32] This test's focus on the body of current voucher holders and not past selection rates enables an assessment of all of the six PHAs' pre-1998 residency preferences, not just those with extant past data about the race, residency, and disposition of applicants.

The downside of this approach is that emphasizing voucher holding over voucher selection does not take account of the applicants who were selected but refused vouchers or who accepted them, but are no longer in the program. Voucher holding and selection are different. There is no one-to-one correlation between selection at some time in the past and current participation in the voucher program.

To verify the statistical significance of the data,^[33] the plaintiffs' expert applied two standards. The first calculated the standard deviations between the minority participation share and application share. If there were more than two standard deviations between the percentage of voucher holders who are minorities and the percentage of applicants who are minorities, the PHA would fail. The second standard applies a four-fifths rule: if the percentage of current Section 8 voucher holders who are minorities is less than four fifths of the percentage of applicants who are minorities, the PHA would fail.

For example, Abington showed that 25% of its voucher holders were minorities, though in 1998 55% of applicants were minorities. There was a difference of 4.6 standard deviations between these two figures, and 25% is less than four fifths of 55%. Abington therefore failed in this analysis by both standards. Bridgewater (15% minority voucher holders, 52% minority applicants, 5.8 standard deviations), Holbrook (20% minority voucher holders, 54% minority applicants, 5.3 standard deviations), Pembroke (24% minority voucher holders, 60% minority applicants, 6.1 standard deviations), and Rockland (18% minority voucher holders, 60% minority applicants, 8.8 standard deviations) also failed by both standards. The numbers for Halifax (27% minority voucher holders, 55% minority applicants, 1.8 standard deviations) are passable by the standard-deviation standard, but not under the four-fifths rule. Avon and Middleborough, again, had no residency preference in place prior to 1998 and could not be studied.

(3) The Effect of the Preliminary Injunction

The plaintiffs' third proposed approach evaluates the effect of this Court's preliminary injunction on the roster of each PHA's voucher recipients. The preliminary injunction ordered the parties to proceed with the distribution of the vouchers, but without the residency preferences in effect. Thus, the PHAs with unexhausted wait lists would exhaust them, then dip into the lotteried pool of 1998 applicants, and those with entirely new wait lists would make selections according to the lottery results, without any favoring of residents.

The plaintiffs pay particular attention to the percentage of present minority voucher holders who were taken from the 1998 waiting list. These percentages suggest that the preliminary injunction did wonders for minority participation in the Abington (87%), Middleborough (62%), and Rockland (74%) Section 8 programs. The numbers for Avon (13%), Bridgewater (10%), Halifax (0%), Holbrook (0%), and Pembroke (6%) are substantially more equivocal. The plaintiffs suggest that the data "raises a strong inference that the preferences would have retarded the admission of minorities had they not been enjoined."

The statistics here, in my estimation, are telling as to some of the PHAs, but not all. First, while data oriented around the current rosters of voucher holders is helpful, the connection is not airtight because of the variables that I describe above—selected candidates rejecting vouchers or moving quickly through the program. Second, though the data shows that the preliminary injunction's suspension of residency preferences dramatically altered Abington, Middleborough, and Rockland's rosters of voucher holders, the injunction had a negligible impact at best on the remaining five PHAs. Moreover, Middleborough had no residency preference in place prior to 1998; implementation of the preliminary injunction merely preserved the status quo.

This data suggests that some factor other than suspension of a residency preference may also be at work. This does not suggest that the residency preference did not have a substantial disparate impact on the minority participation within the meaning of the law. Rather, it is not at all clear that suspension of the residential preference was alone responsible for Middleborough's striking 62% result under this analysis.

(4) The Measure of Delay

Fourth and finally, the plaintiffs deploy statistical analysis of the extent of delay nonresidents would experience as a result of the residency preferences before winning selection to the PHAs' Section 8 programs. To perform this analysis, Dr. Tilly extrapolated from each PHA's current pace of voucher issuance and found that Abington (1.2 month delay), Avon (2.7 months), Bridgewater (2.6 months), Middleborough (1.4 months), Pembroke (3 months), and Rockland (1.4 months) promised considerable delays to nonresidents. Halifax and Holbrook's residency preferences would result in delays of less than one month. The plaintiffs propose that this test would find a disparate impact for each PHA that promises a delay of more than one month. It follows, then, that Halifax and Holbrook pass scrutiny.

The problem with this analysis, at first glance, is that it quantifies the delays to nonresidents, but not to minorities, even though the percentage of minorities in the nonresident group is high. The quintessence of a residency preference is that it impacts nonresidents by delaying their selection to the Section 8 program. Dr. Tilly's analysis only reinforces with numbers what anyone could deduce from the face of the scheme. More to the point is the extent to which minorities as a group would suffer delay in receiving their vouchers as a result of the preference. While it seems eminently logical to assume that minorities will bear the brunt of the delay to nonresidents (because of the demographic breakdown of the defendant communities), Dr. Tilly's statistics do not specifically address that question.

At the same time, Dr. Tilly's data appears to understate the amount of delay that might well befall nonresident applicants in the future. The test assumes a constant rate of selection based on each PHA's pace of voucher issuance in the months following the preliminary injunction. As I described above, when Abington launched its Section 8 program in 1996, its residency preference resulted in a fourteen-month delay to nonresidents, and when Pembroke initiated its distribution of vouchers that same year, thirteen months passed before nonresidents were selected. The First Circuit itself observed that "given the limits on funding, the delays apparently stretch into years, so delay may for practical purposes be the equivalent of denial." Langlois, 207 F.3d at 48.

c. Conclusions

As both the plaintiffs and defendants recognize, no single test leaps off the page as ideal. The EEOC test I applied in my preliminary injunction decision may be inappropriate due to the small sample size. Langlois, 207 F.3d at 50. The Comer test at the preliminary injunction stage dealt with projected selections from existing wait lists, but only for the upcoming calendar year. Data insufficiencies stymie application of the first summary judgment test proposed by the plaintiffs to five of the PHAs. The second, and third tests, relying as they do on current rosters of voucher holders, do not reflect selection with 100% accuracy because they do not account for candidates who are offered vouchers and refuse them or who have left the program. The delay test describes the impact on nonresidents and not minorities.

The plaintiffs urge me to view their four tests as an integrated whole and not to consider them individually. Dr. Tilly, for example, acknowledges that the data on delay "standing alone may not be conclusive, [but] when reviewed in conjunction with the other information ..., it adds to the conclusion that the preference will negatively impact minority applicants." Accordingly, the plaintiffs propose that I find a prima facie case of disparate impact as to any PHA that fails more than one of the tests. Abington fails all four tests, Bridgewater, Pembroke, and Rockland fail three of the four, and Avon and Middleborough fail two. Halifax and Holbrook, in the plaintiffs' view, are to be exonerated for failing only one of the proposed tests on summary judgment.

I agree with the plaintiffs that it is critical to see the case as a whole. Indeed, given the context, the plaintiffs paint a more compelling picture than they recognize. Though the defendant PHAs heavily critique the approach the plaintiffs have taken, they propose no better methodologies and offer no expert testimony to rebut Dr. Tilly's. See Boston Police Superior Officers Federation v. City of Boston, 147 F.3d 13, 22 (1st Cir.1998) (emphasizing the value of expert testimony in statistical proof of discrimination and taking notice of the plaintiffs' failure to provide the rebuttal testimony of an expert). Perhaps most importantly, they give me no reason to question the overarching intuitive principle that compromises their case: where a community has a smaller proportion of minority residents than does the larger geographical area from which it draws applicants to its Section 8 program, a selection process that favors its residents cannot but work a disparate impact on minorities.^[34]

Significantly, the defendants give me no reason to believe that the faults they identify in the plaintiffs' individual tests necessarily skew the results of those tests in the plaintiffs' favor. For example, the second and third tests, which key on the current roster of voucher holders rather than statistics about selection, are troubling because selected parties often refuse vouchers, or they may have been cycled in and out of the program already. But no evidence is before me to suggest that minority applicants are more likely than white applicants to refuse vouchers or that they spend a shorter time participating in Section 8 programs.^[35] Accordingly, I have no reason to believe that the data pertaining to current voucher holders understates the actual admission of minorities to the program.

Similarly, the defendants point out that a number of the PHAs maintained first-come first-serve open lists prior to 1998. Federal preferences were also in place in a number of PHAs prior to the enactment of the QHWRA. Accordingly, the defendants argue, the first and second tests' inquiry into the effect of the residency preference under these earlier schemes is not necessarily instructive of how the preference will play in the future under a lottery scheme without federal preferences.^[36] But again, the defendants offer no evidence to suggest that these differences materially hurt their position.

The defendants fault the first test because data is available to examine the track record of only three of the six PHAs' preexisting residency preferences. Bridgewater, Halifax, and Holbrook PHAs did not keep data on the race, residency, and selection order of applicants that would be sufficient to permit an informed assessment of the impact of preexisting residency preferences on minority selection.

As I describe below, the PHAs' failure to keep such statistics is itself in derogation of their duty to affirmatively further fair housing. It was incumbent upon the PHAs to determine the extent to which their Section 8 policies affected fair housing principles, and that included gathering and reviewing data on exactly the question of the impact of their residency preferences on the availability of vouchers for minorities. Though the plaintiffs appear to concede the point, I find it inappropriate to treat Bridgewater, Halifax and Holbrook as "passing" plaintiffs' first test. At best these PHAs deserve a grade of "Incomplete." The defendants summon no evidence and offer no argument to suggest that the absent data would have supported a finding of no impact—and it was their duty under the law to keep that data.

Though the defendants' critiques of the plaintiffs' methodologies do not explain why the flaws they cite are material, the plaintiffs offer ample evidence to suggest that the very fact of the residency preferences resulted in fewer application submissions from nonresident minorities. Notice that the process favored residents from the PHA communities deterred nonresidents from applying. More nonresident minority applicants would mean that a greater portion of the wait list was comprised of minorities who would receive no advancement as a result of the residency preference.

For that matter, the plaintiffs provide census statistics demonstrating that nonresidents who would have applied are disproportionately minorities. More than this, the plaintiffs' evidence that a lower percentage of minorities own cars and phones than do whites—even within the plaintiff class—supports an inference that the logistics of the process made it more difficult for minority nonresidents as a group to submit Section 8 applications than for white nonresidents. Though the number of minorities and whites who would have applied in the absence of these logistical barriers is not quantifiable, it may be inferred from this undisputed evidence that the tests that focus on the composition of the PHAs' wait lists actually understate the residency preferences' impact on minorities. This warrants a further push toward finding a prima facie case of disparate impact. Though the defendants have plenty of reasons why the tests might not describe the impact of the preferences with perfect accuracy, they offer no argument or evidence that would require a push in the opposite direction.

I therefore find that the plaintiffs have established a prima facie case that the residency preferences of Abington, Avon, Bridgewater, Middleborough, Pembroke, and Rockland work a disparate impact on racial minorities.

2. The Prima Facie Case for Disparate Impact: Application Procedures

The plaintiffs mount a specific challenge to the application procedure, which they seek to enjoin as creating a disparate impact on minorities. They claim that notice of the residency preference and the logistics of the application process kept minority candidates for Section 8 vouchers from even applying. The plaintiffs offer evidence that suggests it was next to impossible for potential applicants in neighboring communities to retrieve applications from the PHAs without cars and phones. Public transportation, for example, was the only alternative for applicants without access to cars and phones Brockton or Boston residents, and the distance between the PHAs and the nearest train stations would make trips to the PHAs hours long. Section 8 candidates only had two days to retrieve applications from the defendant PHAs.

The plaintiffs offer census data that demonstrates that greater portions of minorities lack access to cars and phones than do whites. The conclusion to infer from this is that it was substantially more difficult for minorities, as a group, to obtain and submit applications than it was for similarly-located whites. And, the plaintiffs argue (and as I discuss above), the notice given to potential Section 8 applicants that the defendant PHAs had implemented residency preferences "prescreened" nonresident candidates who found it not worthwhile even to apply. Again, deterrence of applications from nonresidents is alleged to disproportionately impact minorities, because of the demographics of the PHAs' communities vis-à-vis their neighbors'.

Finally, the plaintiffs offer anecdotal evidence describing the effects of the PHAs' chosen application processes and the residency preference on qualified candidates' ability—and decisions—to apply to the PHAs for vouchers. At least one affidavit describes a nonresident's decision not to bother applying, because of the residency preference. One housing advocate described heroic expenditures of resources trying to pick up applications for clients; another described clients' lack of access to phones and cars and her own difficulty reaching the PHAs by phone—the lines were always busy. Candidates with phones who tried to call for themselves had similar luck.

The plaintiffs urge that the combination of the announcement of a residency preference and the logistical barriers to applications works a disparate impact against minorities in violation of the Fair Housing Act. They cite United States v. City of Warren, Michigan, 138 F.3d 1083 (6th Cir. 1998), an employment case, which found disparate impact in the combination of the city's announced residency preference for applicants and its failure to advertise municipal job openings beyond a circumscribed, predominately white local area. Id. at 1094.

In City of Warren, however, the United States proceeded on much more solid statistical ground than the plaintiffs do here. The city reversed its recruitment practices for policemen and firefighters midstream, advertising job openings more broadly. A statistically significant increase in minority applicants followed. On this evidence the district court found that the city's recruitment practices for policemen and firefighters worked a disparate impact on minorities. The city did not change its recruiting practices for other municipal jobs, and the United States pressed the court for a similar finding against the city's practices for non-police, non-firefighting positions. The Sixth Circuit found just such a disparate impact, notwithstanding the absence of statistical evidence relating to the impact of the limited recruitment on non-police, non-firefighting positions. Id. at 1092-93.

The same kind of smoking-gun evidence is not present here. The plaintiffs cannot point to dramatic increases in minority applications given an application process more accommodating to out-of-towners. Through circumstances not of their creation, they can only speculate about the number of qualified minorities who would have applied absent the difficult conditions that they challenge. The plaintiffs must rely instead on inferences drawn from evidence that eligible minorities have less access than do eligible whites to the cars and phones that would have facilitated their retrieval of Section 8 applications from the PHAs.

Dr. Tilly offers census data showing that 37.7% of black urban households in the United States do not have access to cars, compared to 14.1% of their white urban counterparts. This is a difference of 80.4 standard deviations, but there is no ready data that breaks down these numbers by income group. The defendants have a strong argument that the differential in car access between whites and minorities is significantly reduced at the pertinent income level: that is, at the level of income where Section 8 eligibility triggers.

Dr. Tilly has also found in Massachusetts that more whites than blacks own phones—but by a statistically insignificant differential of 1.9 standard deviations. A significant differential of 3.6 standard deviation separates whites and Hispanics. Again, however, this data does not control for income level. Households nationwide earning less than $20,000 nationwide suggest that more whites than blacks own phones at a clip of 5.5 standard deviations, and that more whites than Hispanics have phone service, by a gap of 4.3 standard deviations. But these numbers reflect access to phone service nationwide, and the disparities might well vary significantly by reason.

Based on the evidence of car and phone ownership, as well as the established disproportionately high numbers of qualified minority nonresidents, I have no doubt that the application procedures worked a disparate impact on racial minorities. While poor white families may not have had access to phones and cars, if they lived in the relevant towns, they did not need it.

Given the current record, however, while I can find a disparate impact, I cannot determine how disparate the impact was.^[37] Under the circumstances, I decline to take the position that the PHAs' application procedures violated the "because of race" provision in the Fair Housing Act.

The point is moot. As I describe above, I used this data to buttress the conclusion that the residency preferences had a substantial disparate impact on minorities. And, as I will make clear below, the PHAs' failure to consider the possible effect of its application procedures on the ability and inclination of qualified minority residents to apply for their Section 8 vouchers is a clear violation of their duty to affirmatively further fair housing. I will therefore proceed with a remedy on that basis.

3. The "Simple Justification Test"

In the present case, the First Circuit suggested that the defendant PHAs had passed the simple justification test for their residency preferences, pointing to the "considerable history" of preferences for local residents, and moreover to "the 1998 amendment [in which] Congress itself endorsed the use of locally determined preferences in distributing section 8 vouchers." Langlois, 207 F.3d at 51 (referring to 42 U.S.C. § 1437f(o)(6), as well as earlier regulatory provisions expressly allowing residency preferences). The First Circuit concluded that, absent a showing of a different theory of intentional discrimination, the residency preferences at issue here "[do] not violate the ``because of race' provision of the Fair Housing Act standing alone." Id.

What the Court was saying, in effect, was that if Congress has announced that residency preferences are a legitimate goal of the QHWRA, then that is the end of the analysis. It does not matter if these preferences also have a substantial disparate impact on minorities. However, a closer analysis of the legislative history of the Act than was possible during the preliminary injunction litigation suggests the following: First, as I indicated above, Section I.A.1.(c), supra, and summarize below, neither the statutory text nor the legislative history of the QHWRA suggests that Congress wished to encourage implementation of preferences for local residents by PHAs. Rather, the statute suggests that local needs and preferences should enter into the mix of housing promoted by Section 8 vouchers. Second, the QHWRA plainly had multiple goals. Even if the QHWRA implied an endorsement of preferences for local residents, it cannot be read to trump other civil rights statutes that preceded it, and which are also part of the statutory and regulatory framework.

Let me make this clear: It is not an issue of balancing one goal against another, "with individual judges deciding which seem to them more worthwhile," which the First Circuit disapproved. Langlois, 207 F.3d at 50. Rather, it is an attempt to discern what Congress had in mind by listing multiple objectives which may only be harmonized as follows—residential preferences are permissible but only so long as other civil rights laws are complied with.

a. Statutory Purpose and Intent

The First Circuit accurately observed that in creating and passing QHWRA, Congress gave voice to its concern that federal over-regulation was impeding the PHAs' ability to function effectively. As I have discussed, Congress concluded that the pre-QHWRA federal Section 8 preferences regime did not afford PHAs sufficient flexibility to respond to local needs. Indeed, Congress' hope was that by enhancing local preferences the ultimate effect would be to promote mixed-income housing and diminish residential segregation by income. See, e.g., Pub.L. No. 105-276, § 502(a)(3), 112 Stat. at 2520 (reporting a congressional finding that "the public housing system is plagued by a series of problems, including the concentration of very poor people in very poor neighborhoods"); id. § 502(b)(3), 112 Stat. at 2521 (articulating a statutory purpose of, inter alia, "facilitating mixed income communities and decreasing concentrations of poverty in public housing"); 144 Cong. Rec. S11833-02, S11839 (daily ed. Oct. 8, 1998) (remarks of Sen. Kerry).

What is conspicuously absent from both QHWRA and its legislative history is any indication of concern that somehow local residents were getting short shrift in their own PHAs' housing lotteries. It may be the case, as the First Circuit observed, that "Congress itself endorsed the use of locally determined preferences in distributing section 8 vouchers," Langlois, 207 F.3d at 51 (emphasis added), but that is not the same thing as preferences for local residents. From Congress' desire to promote flexibility and local autonomy for PHAs in the administration of their fair housing programs it does not follow that Congress intended specifically to endorse preferences for local residents—particularly in situations in which such preferences threaten to conflict with other congressional mandates.

Indeed, the Secretary's own interpretation of the PHAs' obligations under QHWRA—which is itself entitled to deference under Chevron—suggests precisely the opposite. The regulations demonstrate an explicit concern about the racial impact of the PHAs' implementation of residency preferences. 24 C.F.R. § 982.207(b) provides, in relevant part:

(1) Residency requirements or preferences.

(i) Residency requirements are prohibited. Although a PHA is not prohibited from adopting a residency preference, the PHA may only adopt or implement residency preferences in accordance with non-discrimination and equal opportunity requirements listed at § 5.105(a) of this title.

(ii) A residency preference is a preference for admission of persons who reside in a specified geographic area ("residency preference area"). A county or municipality may be used as a residency preference area. An area smaller than a county or municipality may not be used as a residency preference area.

(iii) Any PHA residency preferences must be included in the statement of PHA policies that govern eligibility, selection and admission to the program, which is included in the PHA annual plan (or supporting documents) pursuant to part 903 of this title. Such policies must specify that use of a residency preference will not have the purpose or effect of delaying or otherwise denying admission to the program based on the race, color, ethnic origin, gender, religion, disability, or age of any member of an applicant family.

(Emphasis added.)

This regulation can hardly be said to constitute an unlimited endorsement of local residency preferences without regard to their impact on racial minorities. Rather, while not prohibiting PHAs outright from adopting residency preferences, the regulation cautions that such preferences may only be adopted or implemented in accordance with an extensive list of nondiscrimination and equal opportunity requirements contained in 24 C.F.R. § 5.105(a). That list includes, in relevant part: The Fair Housing Act and its implementing regulations; Executive Order 11063, as amended by Executive Order 12259, and its implementing regulations; and Title VI of the Civil Rights Act of 1964 and its implementing regulations.

If anything, the regulation suggests a hierarchy of congressional mandates, with civil rights laws taking precedence. Only if they are met are residency preferences permissible.^[38] Indeed, to ignore the civil rights impact of residential preferences is to do precisely what the First Circuit cautioned against—namely, to give effect to an individual judge's view that one goal of the statute, concern for local needs, is somehow more important than another, concern for fair housing.

b. The Broader Statutory Context: Local Needs Harmonized with Civil Rights Enforcement

QHWRA was enacted within a framework of civil rights statutes and regulations—one of which Congress and the Secretary were both well aware. In the face of the civil rights laws already discussed so far, it is simply not enough to point to Congress' authorization of local preferences and proclaim that any PHA's local preference adopted thereunder trumps other statutes because Congress authorized it. See, e.g., Comer, 37 F.3d at 795 ("Although the U.S. Housing Act, by its terms, does permit a local preference, such preference is subject to various limitations including that its administration must be consistent with the Constitution and civil rights laws.... Therefore, the application of the local preference is null and void to the extent that the local preference is inconsistent with the Constitution ... or any of these acts.").

To put it another way, even taking as a given the First Circuit's articulation of the "simple justification" test—that the justification has to be a "legitimate and substantial goal of the measure in question," Langlois, 207 F.3d at 51. Part of what goes into determining the legitimacy of that goal is how it fits into the statutory and regulatory landscape. If the justification is in tension with other laws, it is difficult to see how it can qualify as "legitimate" in order to justify a measure with a racially disparate impact.

The answer to the question—what is the legitimate and substantial goal of the practice that justifies its substantial disparate impact on minorities—cannot be "preferences for local residents." It is not enough. To say that it is elevates some congressional objectives over others, when legislative history and authoritative administrative regulations say otherwise.

c. Proving More than Just "Residential Preferences Are Allowed"

Since Congress did not endorse preferences for local residents without limitation, in order to evaluate the legitimate and substantial justification for the disparate impact deriving from their use, I must go more deeply into the purpose of the challenged practice in these communities at this time. The defendants must set forth the reasons why they want the preferences in place. And it is the reasons that must be legitimate in light of the statute's hybrid goals.^[39]

The reasons the defendants have offered—that they want to protect their administrative fees, that they want to make it easier for their residents to keep living in their communities, that it is important for community morale to know that the PHAs are working for the town's own residents—all collapse into the very definition of residency preferences. If I accepted these as legitimate justifications, residency preferences in and of themselves would forever justify the disparate impacts that they cause.^[40]

The "more" that is required, then, is the following: Defendant PHAs must offer a record of local conditions and needs that suggests why the residency preferences are necessary.^[41] Nowhere do they demonstrate that a residency preference would correct housing problems of the kind described in the QHWRA's legislative history. And, as I describe below, nowhere do the defendants demonstrate that they even considered the impact of their residency preferences on minorities and on their compliance with other civil rights obligations. Likewise, the record before me is devoid of any evidence suggesting that the PHAs met their duty to affirmatively further fair housing, which included an obligation to investigate the potential effects of their proposed residency preferences before their implementation. 24 C.F.R. § 903.7(o). Accordingly, I find that the justifications offered by the defendants do not overcome the prima facie case of disparate impact the plaintiffs have established.

4. Less Discriminatory Alternatives

Even if I were to find that the defendants rebutted the prima facie case merely by pointing to language in the regulations permitting residency preferences, the defendants would still have to show that no less discriminatory alternative is available to meet its justified ends. Huntington Branch, NAACP, 844 F.2d at 939.

The plaintiffs describe at least two initiatives that would enable the defendants to recover their administrative fees. First, they might work harder to recruit Section 8 voucher recipients to take up residence in their communities—a fee-preserving measure with the incidental benefit of fostering community integrations. Second, the PHAs could endeavor to contract for the administration of their Section 8 vouchers in other communities, thereby recovering their fees.

The plaintiffs also take note of the possibility of qualifying the residency preferences rather than applying them strictly. They propose an innovative preference system by which the defendant PHAs would develop "reciprocal regional residency preference" programs in partnership with the PHAs of nearby urban communities with much larger minority populations. For example, a PHA could align with the city of Brockton to allow residents of both communities preferential status in both communities' Section 8 selection processes. This tempered approach would still help support local residents in their efforts to maintain their residencies in the defendant communities, while at the same time keeping the strategy from running afoul of the fair housing requirement of no disparate impact.

The defendants, for their part, do not address these proposals or even argue that no less discriminatory means is available to achieve the ends they describe.

E. Duty to Affirmatively Further Fair Housing

Plaintiffs argue that two principal statutory provisions support the PHAs' duty to affirmatively further fair housing: 42 U.S.C. § 1437c-1(d)(15), by which PHAs must submit annual plans to the Secretary of HUD that include a certification that the PHAs will affirmatively further fair housing; and 42 U.S.C. § 3608(e)(5), which requires the Secretary of HUD to "administer the programs and activities relating to housing and urban development in a manner affirmatively to further the policies of this subchapter." The plaintiffs also identify two executive orders in support of their claims: Executive Order 11063, 27 Fed.Reg. 11527 (Nov. 20, 1962), and Executive Order 12892, 59 Fed.Reg. 2939 (Jan. 20, 1994). I will consider each statute or executive order and its accompanying regulations in turn before proceeding to the merits of the plaintiffs' affirmative furtherance claim.

1. Does the "Affirmatively Further" Standard Comply with Blessing and Other Procedural Prerequisites?

a. 42 U.S.C. § 1437c-1(d)(15) (QHWRA)

An annual public housing agency plan ... for a public housing agency shall contain ... [a] certification by the public housing agency that the public housing agency will carry out the public housing agency plan in conformity with title VI of the Civil Rights Act of 1964, the Fair Housing Act, ... and will affirmatively further fair housing.

In its decision affirming the preliminary injunction in this case, the First Circuit first raised a question of timeliness. The court observed that there was an open question as to whether this provision of QHWRA applied to the defendant PHAs' 1998 lotteries. Langlois, 207 F.3d at 52 & n. 9. QHWRA § 511(e) provides that "[t]his section [i.e., § 511] shall take effect, and the amendments made by this section are made on, and shall apply beginning upon, the date of the enactment of this Act"—i.e., October 21, 1998. Pub.L. No. 105-276, § 511(e), 112 Stat. at 2539. 42 U.S.C. § 1437c-1(d)(15) is among the statutory provisions added by § 511 of QHWRA. However, §§ 1437c-1(a), which governs PHAs' five-year plans, and (b), which applies to PHAs' annual plans, both specify an effective date of October 1, 1999 for the requirement that a PHA submit a plan "under this subsection." Accordingly, Congress has specified that, notwithstanding the general effective date for the QHWRA, the effective date for the requirements set forth in § 1437c-1—including this certification—is October 1, 1999.

The PHAs' lotteries at issue here all occurred on December 1, 1998. Because of this Court's injunction, no other lotteries or application procedures have taken place since then that the plaintiffs argue violate § 1437c-1(d)(15). Accordingly, I find that the plaintiffs in this case do not have a cause of action under § 1437c-1(d)(15), because this subsection of the statute was not in effect at the time the PHAs' contested conduct occurred.

b. 42 U.S.C. § 3608(e)(5) (Title VIII)

(1) The Statute

The Secretary of Housing and Urban Development shall ... administer the programs and activities relating to housing and urban development in a manner affirmatively to further the policies of this subchapter[.]

"This subchapter" refers to ch. 45, subchapter I, of title 42 of the United States Code. The declaration of policy for subchapter I provides, "[i]t is the policy of the United States to provide, within constitutional limitations, for fair housing throughout the United States." 42 U.S.C. § 3601.

Both the statutory text and the case law construing it make clear that § 3608(e)(5) is fundamentally concerned with, and aimed at, affirmatively promoting and protecting the right to fair housing. In addressing the significance of this very statute, the First Circuit observed that there was a substantial difference between a statute that merely exhorts officials not to discriminate in effect (a negative obligation) and one that exhorts them to take steps to promote fair housing (an affirmative obligation):

[A] statute that instructs HUD to administer its grant programs so as ``affirmatively to further' the Act's fair housing policy requires something more of HUD than simply to refrain from discriminating itself or purposely aiding the discrimination of others.... If one assumes that many private persons and local governments have practiced discrimination for many years and that at least some of them might be tempted to continue to discriminate even though forbidden to do so by law, it is difficult to see how HUD's own nondiscrimination by itself could significantly "further" the ending of such discrimination by others.

N.A.A.C.P. v. Sec'y of Housing & Urban Development, 817 F.2d 149, 154 (1st Cir. 1987); see also id. at 155 (finding the legislative history to reflect "the desire to have HUD use its grant programs to assist in ending discrimination and segregation, to the point where the supply of genuinely open housing increases"); id. (surveying the case law and noting that "every court that has considered the question has held or stated that Title VIII imposes upon HUD an obligation to do more than simply refrain from discriminating (and from purposely aiding discrimination by others)").

Section 3608(e)(5) also meets the Blessing criteria. It could not be clearer from the statute, the legislative history, and the case law construing it that this provision was intended to benefit the plaintiffs here: people in desperate need of access to fair housing, minorities and the poor.

Nor is the duty to affirmatively further too vague and amorphous for the courts to enforce. Essentially, as the Court in Blessing noted, this is an issue of judicial competence. As noted above, it is an anti-discrimination statute and few discrimination statutes are more specific. See, e.g., 42 U.S.C. § 2000e-2(a)(1) ("It shall be an unlawful employment practice for an employer to fail to or refuse to hire any individual, or otherwise so discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual's race, color, religion, sex, or national origin ...."). Judicial decisions and administrative regulations have given content to these provisions over the past twenty years.

The First Circuit in N.A.A.C.P. v. HUD held that the obligation to "affirmatively further" fair housing gave rise to "a plausible claim of a Title VIII violation," id. at 157, and cited numerous other cases in which other courts recognized the obligation as enforceable under § 1983, id. at 155. HUD's regulations give content to the provision, spelling out specific ways in which a PHA is supposed to affirmatively further fair housing. For the most part, these regulations create logical obligations. They describe how the PHA might monitor the impact of proposed programs on the groups the provision was intended to serve:

A PHA shall be considered in compliance with the certification requirement to affirmatively further fair housing if the PHA fulfills the requirements of § 903.2(b) and:

(i) Examines its programs or proposed programs;

(ii) Identifies any impediments to fair housing choice within those programs;

(iii) Addresses those impediments in a reasonable fashion in view of the resources available;

(iv) Works with local jurisdictions to implement any of the jurisdiction's initiatives to affirmatively further fair housing that require the PHA's involvement; and

(v) Maintains records reflecting these analyses and actions.

24 C.F.R. § 903.7(o)(3).

The final Blessing criterion—whether the statutory language at issue imposes an unambiguous and binding obligation on the state—is also met here. The language of the statute is mandatory: it provides that the Secretary "shall" administer the relevant programs in a manner affirmatively to further the policies of the statutory subchapter.

The remaining question is whether the mandatory obligation imposed by § 3608(e)(5) on the Secretary of HUD can be enforced against the PHAs. When viewed in the larger context of Title VIII, the legislative history, and the case law, there is no way—at least, none that makes sense—to construe the boundary of the duty to affirmatively further fair housing as ending with the Secretary.

That is precisely what the Second Circuit held in Otero v. New York City Housing Authority, 484 F.2d 1122 (2d Cir.1973), a seminal case in this area. The court observed, in its analysis of what is now § 3608(e)(5):

[W]e are satisfied that the affirmative duty placed on the Secretary of HUD by § 3608[ (e)](5) and through him on other agencies administering federally-assisted housing programs also requires that consideration be given to the impact of proposed public housing programs on the racial concentration in the area in which the proposed housing is to be built. Action must be taken to fulfill, as much as possible, the goal of open, integrated residential housing patterns and to prevent the increase of segregation, in ghettos, of racial groups whose lack of opportunities the Act was designed to combat.... Senator Mondale, who drafted § 810(a) of the Act, 42 U.S.C. § 3610, pointed out that the proposed law was designed to replace the ghettos "by truly integrated and balanced living patterns." 114 Cong. Rec. 3422.

Id. at 1133-34 (emphasis added) (citations omitted). Indeed, for the past thirty years, claims arising under Title VIII and § 3608(e)(5) have been enforced consistently with Otero—recognizing an affirmative duty imposed upon the Secretary of HUD and through him on entities like the PHAs.^[42]

Precedent interpreting different statutory and regulatory schemes is simply inapposite here. In Stowell v. Ives, 976 F.2d 65 (1st Cir.1992), the First Circuit ruled that various provisions of the Social Security Act relating to the Aid to Families with Dependent Children (AFDC) program, 42 U.S.C. §§ 601-15 and 1396a(c)(1), could not sustain a cause of action under § 1983 because they obliged the federal government, and not the states, to take action. Stowell, 976 F.2d at 69; see also id. (holding that "section 1396a(c)(1) provides incentives—not commands—to the States"). Accordingly, the First Circuit held, "when a provision in a statute fails to impose a direct obligation on the States, instead placing the onus of compliance with the statute's substantive provisions on the federal government, no cause of action cognizable under section 1983 can flourish." Id. at 70 (citing Suter v. Artist M, 503 U.S. 347, 356-57, 112 S. Ct. 1360, 118 L. Ed. 2d 1 (1992)).

This case is multiply distinguishable from both Stowell and Suter, on which it rests. First, and significantly, the Suter Court's analysis of the scope of the obligation conferred by the statute in question did not end with the statutory language itself. Rather, the Court also examined the agency's implementing regulations and found that "the regulations are not specific and do not provide notice to the States that failure to do anything other than submit a plan with the requisite features, to be approved by the Secretary, is a further condition on the receipt of funds from the Federal Government." Suter, 503 U.S. at 362, 112 S. Ct. 1360.

This is simply not the case here. 42 U.S.C. § 3608(e)(5) places an affirmative obligation on the Secretary to ensure that the PHAs comply with civil rights laws. This obligation is embodied in the statutory provisions that require the PHA to certify its compliance, 42 U.S.C. 1437 c-1(d)(15). 24 C.F.R. § 982.53(c) confers on the PHAs an independent obligation to affirmatively further fair housing themselves. Moreover, 24 C.F.R. § 903.7(o)(3), incorporated by reference into § 982.53(c), explicitly lists precisely what steps PHAs have to take in order to be considered in compliance with their obligation. Thus, insofar as the holdings of Suter and Stowell both rested on the respective courts' conclusions that neither the statute in question nor the agency's own interpretation thereof placed any specific obligation on the states, Title VIII's provisions warrant different treatment.

Moreover, both Suter and Stowell addressed statutes very different from the civil rights provisions at issue in this case. Both the Adoption Assistance and Child Welfare Act provision at issue in Suter and the AFDC provision in Stowell involved conditional funding grants or "incentives" to the states that hinged on the states' assumption of some systemic obligation that they were not otherwise required to meet. Suter, 503 U.S. at 350-51, 112 S. Ct. 1360 (reading the statute at issue to provide "that States will be reimbursed for a percentage of foster care and adoption assistance payments when the State satisfies the requirements of the Act"); Stowell, 976 F.2d at 69 ("States are not obliged by federal law to sponsor medical assistance plans or to accept federal funds for this purpose.").

Title VIII, however, is different. It is a civil rights statute focused on ensuring individual rights to fair housing. Its obligations are not ancillary to a federal-state spending contract. Rather, its provisions operate directly on local PHAs, spelling out their obligations with precision. Gonzaga Univ. v. Doe, 536 U.S. 273, 122 S. Ct. 2268, 2277, 153 L. Ed. 2d 309 (2002) (finding that, in the Spending Clause context, "[u]nlike the individually focused terminology of Titles VI and IX (``no person shall be subjected to discrimination'), FERPA's provisions speak only to the Secretary of Education, directing that ``no funds shall be made available' to any ``educational agency or institution' which has a prohibited ``policy or practice.'"); see also id. at 2282 (Stevens, J., dissenting) (observing that the majority "contrasts FERPA's ``no funds shall be made available' language with ``individually focused terminology' characteristic of federal antidiscrimination statutes").

This interpretation is not only consistent with the statute, its legislative history, and the overall context of civil rights law, as I have described. Significantly, this interpretation comports with HUD's own understanding of Title VIII; reflected in HUD's implementing regulations—which, as the plaintiffs point out, are entitled to substantial deference under Chevron U.S.A. v. Natural Resources Defense Council, 467 U.S. 837, 843-44, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984) ("If Congress has explicitly left a gap for the agency to fill, there is an express delegation of authority to the agency to elucidate a specific provision of the statute by regulation. Such legislative regulations are given controlling weight unless they are arbitrary, capricious, or manifestly contrary to the statute.").

(2) The Regulations

24 C.F.R. § 982.53(b)-(c) provides as follows:

(b) Civil rights certification. The PHA must submit a signed certification to HUD that:

(1) The PHA will administer the program in conformity with the Fair Housing Act, Title VI of the Civil Rights Act of 1964, section 504 of the Rehabilitation Act of 1973, and Title II of the Americans with Disabilities Act.

(2) The PHA will affirmatively further fair housing in the administration of the program.

(c) Obligation to affirmatively further fair housing. The PHA shall affirmatively further fair housing as required by § 903.7(o) of this title.

24 C.F.R. § 903.7(o) provides, in relevant part:

(o) Civil rights certification.

(1) The PHA must certify that it will carry out its plan in conformity with title VI of the Civil Rights Act of 1964 (42 U.S.C.2000d-2000d-4), the Fair Housing Act (42 U.S.C. 3601-19), section 504 of the Rehabilitation Act of 1973 (29 U.S.C. 794), and title II of the Americans with Disabilities Act of 1990 (42 U.S.C. 12101 et seq.). The PHA also must certify that it will affirmatively further fair housing.

* * * * * *

(3) A PHA shall be considered in compliance with the certification requirement to affirmatively further fair housing if the PHA fulfills the requirements of § 903.2(b) and:

(i) Examines its programs or proposed programs;

(ii) Identifies any impediments to fair housing choice within those programs;

(iii) Addresses those impediments in a reasonable fashion in view of the resources available;

(iv) Works with local jurisdictions to implement any of the jurisdiction's initiatives to affirmatively further fair housing that require the PHA's involvement; and

(v) Maintains records reflecting these analyses and actions.

These regulations unambiguously impose mandatory requirements on the PHAs not only to certify their compliance with the fair housing laws, but actually to comply. Section 982.53(c) provides that the PHAs have an independent obligation to affirmatively further fair housing, and § 903.7(o)(3) sets out with precision just what PHAs must do in order to be deemed in compliance.

Plainly, these regulations, enacted pursuant to a rights-creating statute, confer a right cognizable under § 1983. They could scarcely have any other purpose than to benefit specific individuals in need of fair housing. The agency was very specific about both the obligation to affirmatively further fair housing and the content of that obligation, demonstrating that, far from vague and amorphous, the right admits of a clear plan of enforcement. And the language speaks solely in terms of what the PHAs, not the agency, "must" and "shall" do. Compliance is not optional.

c. Executive Orders 11063 & 12892

Plaintiffs are entitled to enforce the executive orders, Orders 11063 and 12892, through § 1983. First, they have the force and effect of law. They were enacted "pursuant to a statutory mandate or delegation of authority from Congress." Second, quite apart from whether they are independently enforceable, both executive orders differ from those in cases that have denied § 1983 enforcement. While not necessarily promulgated pursuant to a federal statute, they were subsequently incorporated explicitly into a federal statute. 42 U.S.C. § 3608 sets out the obligations of the Secretary of Housing and Urban Development in administering the Fair Housing Act. Section 3608(e)(6) requires the Secretary to make annual reports to Congress and the public regarding the data underlying various HUD-administered programs that falls under the coverage of a list of specified laws. That list of laws is set forth in § 3608(f), and both Executive Orders 11063 and 12892 are on it. 42 U.S.C. § 3608(f)(11).

In other words, Congress has now explicitly made the Secretary responsible for keeping track of whether the programs he administers—which surely include the Section 8 program—comply with Executive Orders 11063 and 12892. Accordingly, regardless of whether the Executive Orders would have had the force of law standing alone, I find that they do so now, as incorporated by Congress into the Fair Housing Act.

Indeed, precisely because I treat these orders as part of the statutory scheme, I must decide if they confer rights enforceable through § 1983 under Blessing. I conclude that they do. Executive Order 11063, titled "Equal Opportunity in Housing," is deeply and fundamentally concerned with the right to fair housing. President Kennedy directed "all departments and agencies in the executive branch of the Federal Government, insofar as their functions relate to the provision, rehabilitation, or operation of housing and related facilities, to take all action necessary and appropriate to prevent discrimination because of race, color, creed, or national origin ...." Exec. Order 11063, § 101, 27 Fed.Reg. 11527 (Nov. 20, 1962). The mandate here could not be clearer or less ambiguous. The language is plainly mandatory, not precatory.

The same is true of the regulations adopted by the Secretary pursuant to Executive Order 11063: 24 C.F.R. §§ 5.105(a) and 107.21. 24 C.F.R. § 5.105, in relevant part, provides:

The following Federal requirements apply as noted in the respective program regulations:

(a) Nondiscrimination and equal opportunity. The Fair Housing Act (42 U.S.C. 3601-19) and implementing regulations at 24 CFR part 100 et seq.; Executive Order 11063, as amended by Executive Order 12259 (3 CFR, 1959-1963 Comp., p. 652 and 3 CFR, 1980 Comp., p. 307) (Equal Opportunity in Housing Programs) and implementing regulations at 24 CFR part 107; title VI of the Civil Rights Act of 1964 (42 U.S.C. 2000d-2000d-4) (Nondiscrimination in Federally Assisted Programs) and implementing regulations at 24 CFR part 1 ....

24 C.F.R. § 107.21 provides: "All persons receiving assistance from, or participating in any program or activity of the Department involving housing and related facilities shall take all action necessary and proper to prevent discrimination on the basis of race, color, religion (creed), sex or national origin."

Likewise, in Executive Order 12892, President Clinton directs the Secretary of HUD to ensure that affirmative furtherance of fair housing occurs and requires the heads of other agencies that administer relevant programs to cooperate and work in conjunction with HUD. Exec. Order 12892, § 2, 59 Fed.Reg. 2939 (Jan. 17, 1994).

2. Did the PHAs' Plans Meet the Affirmative Furtherance Requirement?

The parties agree that this Court owes no deference to the defendant PHAs' certifications of compliance with the obligation to affirmatively further fair housing. The local PHAs have no demonstrated expertise in the pursuit of fair housing by which this Court might owe them deference as to their compliance with federal law.

Indeed, here again, the fair housing provisions are different from other provisions in which deference to the PHAs may be due. For example, while HUD regulations specify that courts are to defer to the decisions of PHAs regarding utility allowances, "unless found to be arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with the law." 24 C.F.R. § 965.502(e), the implementation of antidiscrimination provisions rests on the courts. Congress has historically legislated to address problems, like civil rights enforcement, left unresolved by states and localities and which courts are uniquely suited to review. Notwithstanding the certifications of the individual PHAs, then, it is the task of this Court to review the defendants' compliance with the affirmative furtherance requirement.

The First Circuit teaches that the Fair Housing Act's affirmative furtherance obligation "requires something more ... than simply to refrain from discriminating itself or purposely aiding the discrimination of others." N.A.A.C.P., 817 F.2d at 154 (noting further that otherwise the statute would afford protections no greater than are available under the Equal Protection Clause). As an example of the kind of affirmative duty imposed by the statute, the N.A.A.C.P. court cited the Second Circuit's finding that "a Housing Authority might even ignore its own regulations giving assignment priority in urban redevelopment projects to former site residents if it is convinced that doing so is the only way to prevent ghettoization of the area." Id. at 155 (citing Otero, 484 F.2d at 1134). It would seem to follow, a fortiori, that Section 8 residency preferences violate the affirmative furtherance principle when their institution would fortify predominantly white communities against minority entry.

To give more specific content to the affirmative furtherance requirement, it is appropriate to look to HUD's view on the matter as Congress' delegate. HUD regulations for the administration of PHA plans specify that a PHA complies with the Fair Housing Act when it

(i) Examines its programs or proposed programs;

(ii) Identifies any impediments to fair housing choice within those programs;

(iii) Addresses those impediments in a reasonable fashion in view of the resources available;

(iv) Works with local jurisdictions to implement any of the jurisdiction's initiatives to affirmatively further fair housing that require the PHA's involvement; and

(v) Maintains records reflecting these analyses and actions.

24 C.F.R. § 903.7(o)(3). The defendants caution against a reading of this regulation that would hold all Section 8 selection systems to an unrealistic standard requiring precise racial correlation between Section 8 applicant pools and selection rates.^[43] But this Court need not go so far to accept the plaintiffs' view that the defendant PHAs did not remotely satisfy their affirmative furtherance obligations under the Fair Housing Act.

Whatever "affirmative furtherance" may mean in other settings, in this setting it is clear. It should have occurred to the PHAs, prior to their adoption of the 1998 plans, to, at the very least, investigate the potential implications for fair housing of the proposed residency preferences and application processes. They did not. Indeed, they could not. They did not bother to keep the kinds of records that would enable them to determine the impact of their new processes. They did not bother to identify potential impediments to fair housing that their application procedures might present. As such, they could not even begin to monitor their compliance with the civil rights laws. The absence of any record of fair housing considerations in the PHAs' development of its selection systems flatly runs afoul of the recording requirement in 24 C.F.R. § 903.7(o)(3)(v).

The difficulties ultimately cited by the plaintiffs were not so unforeseeable that the requirements of 24 C.F.R. § 903.7(o)(3) were not triggered. The deposition testimony cited by the plaintiffs, in which housing authority directors could not even articulate precisely how their organizations affirmatively furthered fair housing, supports an implication of "naked certification."

The law is clear that the mere recital of racial neutrality is not enough. N.A.A.C.P., 817 F.2d at 154. The filings before me give no indication that the defendant PHAs did any more than this before proceeding with their programs. Accordingly, the plaintiffs' motion for summary judgment for their claims under the affirmative furtherance provision of the Fair Housing Act is GRANTED.

F. Miscellaneous Provisions

1. 42 U.S.C. § 3604(c): Advertising

As made applicable by section 3603 of this title and except as exempted by sections 3603(b) and 3607 of this title, it shall be unlawful—

(c) To make, print, or publish, or cause to be made, printed, or published any notice, statement, or advertisement, with respect to the sale or rental of a dwelling that indicates any preference, limitation, or discrimination based on race, color, religion, sex, handicap, familial status, or national origin, or an intention to make any such preference, limitation, or discrimination.

The plaintiffs' argument here is that the PHAs' announcements and advertisements in connection with the December 1998 lottery—including the application form itself—communicated a discriminatory racial preference by including notice of the local residency preference as to the PHAs' predominantly-white communities.

Although the First Circuit has not yet addressed the issue, other courts have construed the requirement in § 3604(c) as follows: the standard for a § 3604(c) violation is whether "an ad for housing suggests to an ordinary reader that a particular race is preferred or dispreferred for the housing in question." Ragin v. New York Times Co., 923 F.2d 995, 999 (2d Cir.1991) (citing United States v. Hunter, 459 F.2d 205, 215 (4th Cir.1972) and Spann v. Colonial Village, Inc., 899 F.2d 24 (D.C.Cir.1990)); see also Jancik v. Dep't of Housing & Urban Devel., 44 F.3d 553, 556 (7th Cir.1995) ("Significantly, no showing of a subjective intent to discriminate is ... necessary to establish a violation of [§ 3604(c)]."); see also Ragin, 923 F.2d at 1000.

The plaintiffs then point to the affidavits of various applicants and advocates who indicated that the local residency preference directly dissuaded them from applying to the PHAs in question.^[44] However, even taking these affidavits at face value and accepting that notice of the residency preference in fact dissuaded some applicants of color from applying to the various PHAs because of the residency preference, I cannot agree that the advertisements and announcements at issue here, simply by notifying applicants of the residency preference, communicated a race-based preference. That they communicated a preference is beyond question, but that preference, as is by now crystal clear, is for local residents of the PHAs' communities.

Accordingly, summary judgment is hereby GRANTED to the defendants on this claim.

2. PHAs' Compliance with Section 8 Plans

Finally, the plaintiffs seek to enforce, via § 1983, several regulatory provisions relating to the PHAs' alleged failure to comply with HUD's requirements for the contents and maintenance of administrative plans. For example, 24 C.F.R. § 982.54 requires, inter alia, that each PHA adopt a written administrative plan that establishes local policies for administration of its programs in accordance with HUD requirements, and that PHAs administer such programs in accordance with their own administrative plans. 24 C.F.R. §§ 982.204(a) and (b) pertain to Section 8 waiting lists and require PHAs (a) to select participants from the waiting list in accordance with the admission policies in their administrative plans, and (b) to organize and maintain their waiting lists based on specific information as to each applicant, including racial and ethnic information.

These provisions in and of themselves fail to meet the criteria set forth in Blessing for a cause of action under § 1983. However, they are relevant to the "affirmative furtherance" issues described above, to the extent that they require that the PHAs maintain data that relates to civil rights enforcement. Accordingly, I find that the plaintiffs do not have a § 1983 cause of action against the PHAs on these grounds alone. Summary judgment on this issue is GRANTED to the defendants.

III. CONCLUSION

Accordingly, for the foregoing reasons, Langlois et al.'s motion for summary judgment [docket entry # 76] is GRANTED IN PART AND DENIED IN PART, and Abington Housing Authority et al.'s motion for summary judgment [docket entry # 86] is GRANTED IN PART AND DENIED IN PART.

SO ORDERED.

NOTES

[1] The defendant PHAs are located in Abington, Avon, Bridgewater, Halifax, Holbrook, Middleborough, Pembroke, and Rockland, Massachusetts. However, defense counsel notified this Court in an affidavit dated April 30, 2002 that three of the defendants — the Abington, Halifax, and Holbrook PHAs — no longer administer their Section 8 programs.

[2] The plaintiff class consists of "all extremely low-income individuals with incomes at or below 30% of area median income who reside in the Brockton and Boston Primary Metropolitan Statistical Areas, but are not residents of [any of the towns of the defendant PHAs]." Approximately 34% of the class are racial minorities.

[3] Plaintiff Massachusetts Coalition for the Homeless is a non-profit corporation — comprised of homeless and formerly homeless persons, organizations, and agencies — that works to provide immediate shelter and transitional and permanent housing to homeless persons and families.

[4] As discussed below, the program was substantially modified with the passage of the Quality Housing and Work Responsibility Act ("QHWRA") on October 21, 1998. Pub.L. No. 105-276, tit. V, § 545, 112 Stat. 2461, 2596-2604 (codified as amended in scattered sections of 42 U.S.C.).

[5] Prior to 1998, there were separate Section 8 programs for "vouchers" and "certificates." However, following the enactment of the QHWRA and promulgation of its attendant regulations, these programs merged in October 1999, and only vouchers are now awarded. 42 U.S.C. § 1437f(o) (2002); 24 C.F.R. § 982.502 (2002).

[6] As part of its Administrative Plan, a PHA must describe the geographic area that its program serves and its policies and procedures for (1) outreach to eligible families, (2) selecting families for participation without discrimination because of race, and (3) assisting housing voucher holders who allege that illegal discrimination is preventing them from leasing suitable units. See 24 C.F.R. §§ 982.53, 982.54; 60 Fed.Reg. 34660, 34661.

[7] Section 8 units rent only to families who qualify as low-income at the time of their initial occupancy of the units. See 42 U.S.C. § 1437a(a). "Low-income families" are defined as those whose incomes do not exceed 80% of the median family income for the area as determined by the Secretary. See id. § 1437a(b)(2). "Very low-income families" are defined as low-income families whose incomes do not exceed 50% of the median family income for the area. See id. The HUD regulations recognize a third "extremely low-income" group, which includes families whose incomes are between 0 and 30% of the median income of the area. See 24 C.F.R. § 91.5.

[8] Even under the federal preference system, HUD regulations made provisions for "approved" residency preferences, by which local residents who also qualified for federal preferences would have waiting-list priority over non-local residents. See Comer, 37 F.3d at 782 (citing 24 C.F.R. §§ 882.219(b)(2)(iii)(A); 887.157(b)(2)(iii)(A)—now reserved).

[9] 42 U.S.C. § 1437d(c)(4), the federal preference statute, was amended to read as follows:

[T]he public housing agency shall comply with such procedures and requirements as the Secretary may prescribe to assure that sound management practices will be followed in the operation of the project, including requirements pertaining to ... (A) making dwelling units in public housing available for occupancy, which shall provide that the public housing agency may establish a system for making dwelling units available that provides preference for such occupancy to families having certain characteristics; each system of preferences established pursuant to this subparagraph shall be based upon local housing needs and priorities, as determined by the public housing agency using generally accepted data sources, including any information obtained pursuant to an opportunity for public comment as provided under section 5A(f) and under the requirements applicable to the comprehensive housing affordability strategy for the relevant jurisdiction.

Pub.L. No. 105-276, § 514(a) (1998), 112 Stat. at 2547 (emphasis added).

[10] Both the 75% rule and the repeal of the federal preferences became effective on the date of the statute's enactment: October 21, 1998. Pub.L. No. 105-276, §§ 513(b), 514(g), 112 Stat. at 2547, 2549.

[11] This section is based on the parties' agreed statement of undisputed facts and the findings of fact in support of the preliminary injunction issued in this case. Langlois v. Abington Housing Auth., No. C.A. 98-12336-NG, 1998 WL 1029208 (D.Mass. Dec. 16, 1998).

[12] All statistics are based on the 1990 Census data.

[13] The PHAs had issued another public notice earlier in the month announcing their decision to eliminate federal preferences and to install a local residency preference in its place. This first notice sought public comment by October 28, 1998.

[14] The original counts were as follows: violations of the federal Fair Housing Act (Title VIII of the Civil Rights Act of 1968), Title VI of the Civil Rights Act of 1964, and the Equal Protection Clause of the U.S. Constitution (Count I); violations of federal housing law (Count II); violation of the Due Process Clause of the Fourteenth Amendment of the Constitution (Count III); violation of civil rights (Count IV); and violation of the Massachusetts public accommodations statute (Count V). In November 2000, the plaintiffs moved voluntarily to dismiss Count III, Count V, and the equal protection claim in Count I, a motion that was granted.

Later, on June 7, 2002, the plaintiffs moved to dismiss the claim in Count I arising from Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq., and its implementing regulations at 24 C.F.R. § 1.4; this motion was also granted. Based on counsel's representations at oral argument on June 7, 2002, I am proceeding on the assumption that the § 1983 claim premised on the alleged § 2000d violation was also dropped. If this assumption is incorrect, counsel are invited to move for clarification of the issue and I will address it at that time.

[15] The motion for class certification was ultimately granted on June 7, 2000.

[16] For a more detailed description of this analysis and the standards used, see this Court's Memorandum and Order of December 30, 1998. Langlois v. Abington Housing Auth., No. C.A. 98-12336-NG, 1998 WL 1029207 (D.Mass.1998), aff'd, 207 F.3d 43 (1st Cir.2000).

[17] Once Rockland submitted such a plan, that plan was to be reviewed by this Court for possible violation of the Fair Housing Act prior to any implementation of the residency preference policy. Plans as to both Abington and Rockland were submitted to the Court on April 7, 1999; the next day, in a motion for further preliminary relief, the plaintiffs argued that these one-paragraph submissions were insufficiently specific to pass as compliant with the Court's injunction, and they requested that the injunctions remain in place until more specific details on compliance were provided. The injunctions remained in place; the issue of the sufficiency of the plans, however, was never ruled upon and remains a live issue here.

[18] The parties agreed to the continuation of the injunction pending resolution of the remaining issues in the case.

[19] In fact, there are more than two views here:

1) Administrative regulations can create rights under § 1983, when they have the force of law and meet the Blessing criteria. The statute pursuant to which they have been enacted may or may not confer rights under Blessing. The phrase "laws" in § 1983 refers to any valid source of federal law that creates rights.

2) Administrative regulations can create rights under § 1983 when they are enacted pursuant to a statute that meets the Blessing criteria. If the statute confers rights and the regulations are enacted pursuant to that statute, then the regulations are also enforceable under § 1983.

3) Administrative regulations can never create rights; only statutes can. Administrative regulations may be used to interpret the scope of the statutory provisions, but only statutes may be the source of § 1983 rights.

As I note above, the statutes and regulations at issue here meet all three tests.

[20] While the majority's language is not clear, Justice O'Connor's account may be overstating the "broad view." Viewed in context, one could interpret the majority not as taking a position that any old valid regulation would be enforceable under § 1983, as Justice O'Connor interpreted it. Rather, in Wright, the regulation met the Blessing criteria, and it was enacted under a statute that also met the Blessing standard.

[21] While the court in Buckley referred to enforcing statutory rights under § 1983, in fact, the statute at issue, Federal Aid in Sport Fish Restoration Act ("the Act"), 16 U.S.C. §§ 777-777k, did not deal with the plaintiff's issue, the regulations governing personal watercraft. It was, in effect, a suit seeking § 1983 enforcement of a regulation.

[22] Hill confronted a set of statutory provisions and regulations different from those at issue in this case: 42 U.S.C. § 1437d(f)(2) (providing for establishment of housing quality standards) and implementing regulations 24 C.F.R. §§ 200.925a, 965.805, and 941.203. The plaintiffs in Hill were tenants in a public housing development who sued the city housing authority for violations of the U.S. Housing Act and the Fourteenth Amendment. The Hill court found that the statute and regulations in question did not pass the Blessing test because, inter alia, they were not intended to confer a federal right on the plaintiffs. Hill, 207 F.Supp.2d at 1027-30.

[23] The South Camden and Harris decisions have been critiqued in a number of respects. First, it is suggested that the courts effectively asked the wrong question—not whether the statute created federal rights, the inquiry in Blessing, but rather, whether Congress intended the statute to create federal rights enforceable by § 1983. Brendan Cody, South Camden Citizens in Action: Siting Decisions, Disparate Impact Discrimination, and Section 1983, 29 Ecology L.Q. 231, 244, 257 (2002); Recent Case, Harris v. James, 127 F.3d 993 (11th Cir.1997), 111 Harv. L.Rev. 2444, 2447 (1998).

Moreover, the South Camden decision has been criticized for adopting dicta in Sandoval concerning the scope of section 602 of Title VI, without analysis. The Third Circuit took the position that the statute by its terms only created a right to sue under a disparate treatment theory, not a disparate impact theory, citing to the Sandoval decision. South Camden, 274 F.3d at 788-89. It then concluded that regulations embodying the latter theory under section 602 were not independently actionable under § 1983. But Sandoval did not address the issue raised in South Camden. Sandoval concerned whether there is a private cause of action to enforce regulations precluding disparate impact discrimination. South Camden concerned whether § 1983 provides an independent avenue to enforce disparate impact regulations promulgated under § 602. See South Camden, 274 F.3d at 796 (McKee, J., dissenting).

[24] The most recent edition of Black's Law Dictionary defines "law" as follows, in relevant part:

law. 1. The regime that orders human activities and relations through systematic application of the force of politically organized society, or through social pressure, backed by force, in such a society; the legal system . 2. The aggregate of legislation, judicial precedents, and accepted legal principles; the body of authoritative grounds of judicial and administrative action . 3. The set of rules or principles dealing with a specific area of a legal system . 4. The judicial and administrative process; legal action and proceedings . 5. A statute ....

Black's Law Dictionary 889 (7th ed.1999).

[25] It would indeed be difficult to suggest that enforcement of the ban on housing discrimination "strains judicial competence," as courts so regularly undertake this task and have done so for at least twenty-five years. See, e.g., Comer v. Cisneros, 37 F.3d 775 (2d Cir.1994) (assessing a disparate impact theory of discrimination under the FHA); Edwards v. Johnston Cty. Health Dep't, 885 F.2d 1215 (4th Cir.1989) (same); Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926 (2d Cir.1988) (same); Arthur v. City of Toledo, 782 F.2d 565 (6th Cir.1986) (same); Metropolitan Housing Development Corp. v. Village of Arlington Heights, 558 F.2d 1283 (7th Cir. 1977) (same); see also Nat'l Fair Housing Alliance, Inc. v. Prudential Ins. Co., 208 F. Supp. 2d 46, 58 (D.D.C.2002) ("[E]very Circuit Court except the District of Columbia Circuit has held that disparate impact claims are cognizable under the FHA. These courts have recognized that the premise of a disparate impact claim is that housing practices may operate in a manner that is functionally equivalent to intentional discrimination." (citing, inter alia, Casa Marie, Inc. v. Sup.Ct. of Puerto Rico, 988 F.2d 252, 269 n. 10 (1st Cir.1993))).

[26] To be sure, in South Camden and Ceaser, courts denied § 1983 enforcement to regulations even when the statute and the regulations each independently met the Blessing test. The problem in both cases was the courts' view that the statute, Title VI, said one thing—prohibiting intentional discrimination —and the regulations said another—prohibiting discrimination that had a disparate impact on racial minorities. Thus, the Third Circuit in South Camden concluded that "inasmuch as the disparate impact regulations go far beyond the intentional discrimination interdiction in section 601 of [Title VI], the district court's reliance on Wright is misplaced." Id. at 783. Likewise in Ceaser, the court suggested that the regulations were not "linked to Title VI in the same way that the regulation was linked to the housing statute in Wright." Ceaser, 2002 WL 472271, at *3.

While I do not agree with the interpretation of Title VI by the courts in South Camden and Ceaser, for the reasons described above, the statute and regulations at issue in this case are distinguishable, and the relationship between them more like the relationship the Supreme Court identified in Wright. The statute in Wright imposed a rent ceiling; the implementing regulations interpreted a reasonable rent to include a reasonable amount for the use of utilities.

Here the statute and its regulations have been interpreted in the same way—to permit a disparate impact theory. See cases cited supra note 25. Indeed, in Huntington Branch, NAACP, the Second Circuit observed that a strict requirement of discriminatory intent "would strip the statute of all impact on de facto segregation." 844 F.2d at 934. The Huntington Court adopted the analysis of Resident Advisory Board v. Rizzo, 564 F.2d 126 (3d Cir.1977), where the Third Circuit emphasized the Senate's rejection of an amendment that would require proof of discriminatory intent to establish claims under Title VIII. Id. (citing Rizzo, 564 F.2d at 147). In fact, Title VIII's language in the Fair Housing Act parallels the language of Title VII. Compare 42 U.S.C. § 3604(a) ("[I]t shall be unlawful to refuse to sell or rent ... or otherwise make unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status, or national origin.") and 3604(b) ("[I]t shall be unlawful to discriminate against any person in the terms, conditions, or privileges of sale or rental of a dwelling, or in the provision of services or facilities in connection therewith, because of race, color, religion, sex, familial status, or national origin."), with 42 U.S.C. § 2000e-2(a) ("It shall be an unlawful employment practice to fail or refuse to hire any individual, or otherwise to discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual's race, color, religion, sex, or national origin."). Just as that language in Title VII gave rise to two theories of discrimination —disparate treatment and disparate impact —so did the language in Title VIII.

Likewise, the duty to affirmatively further fair housing is articulated in the statutory scheme, 42 U.S.C. § 3608(e)(5); the regulations interpret it precisely as the regulations did in Wright.

[27] The analysis is the same for executive orders.

[28] Establishing that the laws at issue in this case confer rights upon individuals essentially ends the Blessing analysis. There is no comprehensive scheme in place to supplant § 1983 enforcement of the 75% Rule, the Fair Housing Act's antidiscrimination mandate, or its requirement that HUD and state and local PHAs affirmatively further fair housing. It is clear that Congress enacted against a backdrop of longstanding § 1983 enforcement of the fair housing laws, and absent any affirmative evidence that it preferred some other means of enforcement, § 1983 must provide the mechanism by which the plaintiffs here can enforce their rights. In fact, regarding the requirement that PHAs certify to HUD their plans' compliance with the Fair Housing Act and their duty to affirmatively further fair housing, Congress expressly stated that HUD's determinations would not foreclose recourse to federal courts under § 1983. 42 U.S.C. § 1437c-1(4)(B).

[29] For that matter, § 1437c-1 expressly contemplates judicial review of a PHA's compliance with that section's stated requirements for the submission of annual and five-year plans to HUD. Id. § 1437c-1(4)(B) (noting that a HUD finding of compliance "shall not preclude ... an action regarding such compliance under section 1979 of the Revised Statutes of the United States (42 U.S.C. 1983)"). One of those requirements is to certify compliance with the Fair Housing Act. 42 U.S.C. § 1437c-1(15).

[30] The lottery-and-preference process adopted by the defendant PHAs is precisely the sort of nondiscretionary selection procedure that is susceptible to disparate impact review. A disparate impact theory of discrimination is generally available "to challenge selection procedures based on facially neutral selection criteria which allow no exercise of discretion." David C. Baldus & James W.L. Cole, Statistical Proof of Discrimination § 1.2, at 46.

[31] See Complaint ¶¶ 79, 81 & Part VII.d (alleging that the residency preferences discriminate against minorities and asking that they be permanently enjoined).

[32] This approach is similar in principle to the Comer analysis performed at the preliminary injunction stage, but here the plaintiffs analyze the universe of current participants in a PHA's program against the composition of the wait lists formulated in the 1998 lottery. The earlier analysis compares the projected selections for 1999 to the composition of the 1998 wait lists.

[33] Dr. Tilly's dual standards here seek to immunize his analysis from the challenge that the defendants made under Fudge to my application of the four-fifths test at the preliminary injunction stage. In Fudge, an employment discrimination case, the First Circuit found application of the EEOC test inappropriate where sample sizes were so small that a four-fifths disparate impact might be the result of chance. Fudge, 766 F.2d at 658. The First Circuit held that "in cases involving a narrow data base, the better approach is for the courts to require a showing that the disparity is statistically significant, or unlikely to have occurred by chance, applying basic statistical tests as the method of proof." Id. Dr. Tilly has applied a standard-deviation analysis to affirm the statistical significance of the identified disparities. Only Halifax has a statistically insignificant disparity; thus, if the sample size were small enough to warrant abandonment of the four-fifths standard, Fudge vindicates only Halifax under this test.

[34] The defendants instead expend considerable effort disputing the implications of the logic. For example, they object that this logic leads ultimately to a ban on all residency preferences unless a community is "politically perfect," that is, unless the racial breakdown of the community's residents is statistically similar to that of nonresidents.

Defendants overstate the problem. The standard is not just disparate impact, but substantial disparate impact; a "politically perfect" community is not required.

In any case, though HUD does expressly permit residency preferences, 24 C.F.R. § 982.207(b)(1), it does not declare that communities are entitled to institute them without regard to their substantial impact on minorities. In fact, as I will show below, it is clear from the regulations that residency preferences are allowable only insofar as they do not conflict with fair housing principles.

[35] Defendants have offered evidence that when a Section 8 family "leases up" in a community other than the one that provides its voucher, the community where the family leases up will occasionally "absorb" the voucher holder into its system. Thus, a family might lease up in Boston with a voucher from, say, Rockland, with the result that Boston elects to provide the voucher to the family instead, as it is handling much of the local administration of the subsidy in any event. Because nonresidents are more likely to be absorbed away from the rosters of a given PHA, and because, in this case, nonresidents are disproportionately minority, minorities are more likely to be absorbed into other voucher programs. That phenomenon would result in the defendant PHAs' current rosters understating their selection.

Though well taken, this narrower point depends on the same logic that fundamentally undermines the defendants at every turn: Because nonresident Section 8 candidates are disproportionately minority, anything that affects nonresidents affects minorities more than whites. Though accepting the phenomenon of absorption suggests resolving in the defendants' favor some of the ambiguity left open by these tests, the fact that nonresidents are disproportionately minority works against the defendants too often to save them. As I note below, evidence is also in the record that notice of the residency preference discouraged nonresident applications, that the two-day window for retrieving applications frustrated nonresident applications, and that the residency preference devalued nonresident applications. These effects on nonresidents greatly favor the plaintiffs' case and the counterweight of nonresident absorption is not enough to overcome them.

[36] Some account must be taken, as well, of the difficulty of the task before the plaintiffs. They must prove, prior to the implementation of the PHA's new residency preference regimes, that these regimes are likely to have a substantial disparate impact on minority applicants. Entry of the preliminary injunction in their favor has complicated matters still more: as long as the programs' residency preferences are suspended, assessment of their impact upon minorities under specific, controlled conditions is not possible.

[37] The plaintiffs might have constructed a workable record with survey evidence. A survey of Section 8 candidates in neighboring areas, for example, could have developed data on how many nonresidents (and how many minorities) were interested in applying, attempted to apply, or elected not to apply to the PHAs programs and why. The plaintiffs might have performed identical surveys within the PHAs' communities and compared the results. Or the plaintiffs could have compiled data about the Section 8 application processes of other white, suburban PHAs in the state. A showing of a statistically significant disparity between minority application totals to the defendants' programs and those with more accommodating application processes would be a major stop to a prima facie case against the processes.

[38] Indeed, it is significant that authorization for residency preferences is found in the regulations and not the statute. 24 C.F.R. § 982.207(b)(1)(i). The statute refers to "local housing needs and priorities." 42 U.S.C. § 1437d(c)(4)(A).

[39] Any other view would make this analysis entirely circular: Plaintiffs show a disparate impact occasioned by the preferences for local residents; defendants respond that the impact is justified by the fact that there are preferences for local residents. That analysis would make no sense.

Title VII suits, for example, require a showing that the challenged practice is "job related for the position in question and consistent with business necessity." 42 U.S.C. § 2000e-2(k)(1)(A)(i). A private employer, put to it to justify a practice of residency preferences that works a disparate impact on minorities, cannot prevail by showing merely that he is entitled to hire predominantly from within the community. To overcome a prima facie case of disparate impact, some showing must be made that the preference bears on the ability of the employees to do their work.

[40] The defendants in the City of Warren, Michigan case, for example, could not justify the adverse impact of their residency preferences for municipal jobs with claims that the preferences were necessary to improve the unemployment rate in Warren and to nourish city income tax revenues.

[41] A PHA might install residency preferences if a fire in the community has left an abnormally high number of residents homeless. Economic factors that hit the PHA's community especially hard—a plant closing, for example—might provide a justification for a residency preference to overcome a prima facie case of disparate impact. No such rationale is before me in this case.

[42] The most recent case endorsing this view is Reese v. Miami-Dade County, 210 F. Supp. 2d 1324 (S.D.Fla.2002). The court held explicitly that the provision requiring governmental agencies to "affirmatively further fair housing" meets all of the tests for § 1983 enforcement. Id. at 1329.

[43] Though the defendants' point here is not pertinent to my finding that the PHAs violated the affirmative furtherance requirement of the Fair Housing Act, I find that it does bear on the more difficult question of crafting an appropriate injunctive remedy for the violation.

[44] Note, however, that the applicants simply said they were dissuaded. They did not say they were dissuaded on the basis of race, but rather that they felt they had little chance at success because the local residents would get priority over them. See, e.g., Munera Affidavit, ¶ 4.