Reinstatement of HUD's Discriminatory Effects Standard

AGENCY:

Office of the Assistant Secretary for Fair Housing and Equal Opportunity, U.S. Department of Housing and Urban Development (HUD).

ACTION:

Final rule.

SUMMARY:

The Fair Housing Act prohibits discrimination in the sale, rental, or financing of dwellings and in other housing-related activities. This prohibition extends to practices with an unjustified discriminatory effect, regardless of whether there was an intent to discriminate. In 2013, HUD published a rule which formalized a burden-shifting test for determining whether a given practice has an unjustified discriminatory effect. In 2020, HUD published a rule that would have altered the standards set forth in the 2013 rule. However, a preliminary injunction prevented the 2020 rule from ever going into effect. On June 25, 2021, HUD published a proposed rule to recodify the 2013 rule. After considering public comments, HUD in this final rule reinstates and maintains the 2013 rule and rescinds the 2020 rule.

DATES:

Effective: May 1, 2023.

FOR FURTHER INFORMATION CONTACT:

Jeanine Worden, Associate General Counsel for Fair Housing, Office of General Counsel, U.S. Department of Housing and Urban Development, 451 7th Street SW, Washington, DC 20410–0500, or telephone number 202–402–3330 (this is not a toll-free number). HUD welcomes and is prepared to receive calls from individuals who are deaf or hard of hearing, as well as individuals with speech or communication disabilities. To learn more about how to make an accessible telephone call, please visit: https://www.fcc.gov/consumers/guides/telecommunications-relay-service-trs.

SUPPLEMENTARY INFORMATION:

I. Background

The Fair Housing Act and Its Goals

Title VIII of the Civil Rights Act of 1968, as amended (“Fair Housing Act” or “Act”), prohibits discrimination in the sale, rental, or financing of dwellings and in other housing-related activities because of race, color, religion, sex (including sexual orientation and gender identity), disability, familial status, or national origin. Through the Act, Congress expressed its intent to eradicate discrimination and proclaimed that “[i]t is the policy of the United States to provide, within constitutional limitations, for fair housing throughout the United States.” The Act's protections are meant to be “broad and inclusive.” Congress passed the Act in the wake of the assassination of Dr. Martin Luther King, Jr., recognizing that “residential segregation and unequal housing and economic conditions in the inner cities” were “significant, underlying causes of the social unrest” and that both open and covert race discrimination were preventing integrated communities. As the Supreme Court reiterated more recently, the Act's expansive purpose is to “eradicate discriminatory practices within a sector of the Nation's economy” and to combat and prevent segregation and discrimination in housing. Congress considered the realization of this policy “to be of the highest priority.”

The Act gives HUD the authority and responsibility for administering and enforcing the Act, including the authority to conduct formal adjudications of complaints and to promulgate rules to interpret and carry out the Act. Through that authority, HUD promulgates this rule.

Discriminatory Effects Law Under the Fair Housing Act Prior to HUD's 2013 Rule

HUD's 2013 rule, titled “Implementation of the Fair Housing Act's Discriminatory Effects Standard” (“2013 Rule”), broke no new ground, but instead largely codified longstanding judicial and agency consensus regarding discriminatory effects law. Courts had long found that discrimination under the Act may be established through evidence of discriminatory effects, i.e., facially neutral practices with an unjustified discriminatory effect. Indeed, before HUD's issuance of the 2013 Rule, all federal courts of appeals to have addressed the question had held that liability under the Act could be established by a showing that a neutral policy or practice either has a disparate impact on a protected group or creates, perpetuates, or increases segregation, even if such a policy or practice was not adopted for a discriminatory purpose. As the Sixth Circuit explained, the Act “proscribes not only overt discrimination but also practices that are fair in form, but discriminatory in operation.”

Consistent with this judicial consensus, HUD has for decades concluded that facially neutral practices that have an unjustified discriminatory effect on the basis of a protected characteristic, regardless of intent, violate the Act. For example, in 1994, HUD, along with nine other agencies and the Department of Justice, issued a joint policy statement that recognized disparate impact liability under the Act. Although there had been some minor variation in the application of the discriminatory effects framework prior to the 2013 Rule, HUD and the federal appellate courts were largely in agreement. HUD has always used a three-step burden-shifting approach, as did many federal courts of appeals prior to the 2013 Rule.

HUD's 2013 Discriminatory Effects Rule

In February 2013, after notice and public comment, and considering decades of case law, HUD published the 2013 Final Rule. The 2013 Rule “formalize[d] [HUD's] long-held recognition of discriminatory effects liability under the Act and, for purposes of providing consistency nationwide, formalize[d] a burden-shifting test for determining whether a given practice has an unjustified discriminatory effect, leading to liability under the Act.” In promulgating the 2013 Rule, HUD noted the Act's “broad remedial intent;” HUD's prior positions, including that discriminatory effects liability was “imperative to the success of civil rights law enforcement;” and the consistent application of discriminatory effects liability in the four previous decades (with minor variations) by HUD, the Department of Justice, nine other federal agencies, and federal courts.

Among other things, the 2013 Rule codified a three-part burden-shifting framework consistent with frameworks on which HUD and courts had long relied: (1) The plaintiff or charging party is first required to prove as part of the prima facie showing that a challenged practice caused or predictably will cause a discriminatory effect; (2) if the plaintiff or charging party makes this prima facie showing, the defendant or respondent must then prove that the challenged practice is necessary to achieve one or more substantial, legitimate, nondiscriminatory interests of the defendant or respondent; and (3) if the defendant or respondent meets its burden at step two, the plaintiff or charging party may still prevail by proving that the substantial, legitimate, nondiscriminatory interests supporting the challenged practice could be served by another practice that has a less discriminatory effect.

The 2015 Inclusive Communities Supreme Court Decision

In 2015, the Supreme Court confirmed that the Act provides for discriminatory effects liability in Texas Department of Housing and Community Affairs v. Inclusive Communities Project, Inc. The State of Texas presented two questions to the Court (1) Whether disparate-impact claims are cognizable under the Act, and (2) if they are, what standards and burdens of proof should apply, but the Court declined to consider the second question. On the first question, the Court found that disparate-impact claims are cognizable, concluding that Congress's use of the phrase “otherwise make unavailable” in Section 804(a) of the Act and the term “discriminate” in Section 805(a) are each parallel to language that the Court had previously held to provide for discriminatory effects liability under other civil rights statutes.

In reaching this holding, the Court explained that from its first decision to recognize disparate impact liability, in Griggs v. Duke Power Co., it “put important limits” on the scope of liability. For example, with respect to employment discrimination claims under Title VII of the Civil Rights Act of 1964, Griggs explained that an employer can justify a practice that has a disparate impact with a “business necessity” defense, such that Title VII “does not prohibit hiring criteria with a `manifest relationship' to job performance.” Similarly, after holding that the Act provided for disparate impact liability, the Inclusive Communities Court noted that, under the Act, “disparate-impact liability has always been properly limited in key respects . . .' ” Quoting Griggs, the Court explained that it has always been true that disparate impact liability under the Act “mandates the `removal of artificial, arbitrary, and unnecessary barriers,' not the displacement of valid governmental policies.”

The Court then sketched out some of these long-standing limitations on the scope of disparate-impact liability, including: (1) The requirement that “housing authorities and private developers [have] leeway to state and explain the valid interest served by their policies . . . analogous to the business necessity standard under Title VII;” and (2) the requirement that a “claim that relies on a statistical disparity must fail if the plaintiff cannot point to a defendant's policy or policies causing that disparity.”

HUD's 2016 Notice: Application of the Fair Housing Act's Discriminatory Effects Standard to Insurance

In 2016, HUD published a document (“2016 Notice”) supplementing its response to certain comments concerning homeowners' insurance received during rulemaking for the 2013 Rule in accordance with the district court's decision in Property Casualty Insurers Association of America (PCIAA) v. Donovan. In that Notice, HUD stated, among other things, that “[a]fter careful reconsideration of the insurance industry comments in accordance with the court's decision . . . HUD has determined that categorical exemptions or safe harbors for insurance practices are unworkable and inconsistent with the broad fair housing objectives and obligations embodied in the Act” and that “commenters' concerns regarding application of the discriminatory effects standard to insurance practices can and should be addressed on a case-by-case basis.”

HUD's 2020 Disparate Impact Rule

On June 20, 2018, HUD published an Advance Notice of Proposed Rulemaking (“ANPRM”), inviting public comment on “what changes, if any” to the 2013 Rule were necessary as a result of Inclusive Communities. HUD then published a Notice of Proposed Rulemaking on August 19, 2019 (“2019 Proposed Rule”) proposing to change the 2013 Rule.

In response to the 2019 Proposed Rule, HUD received approximately 45,000 comments, most of which opposed the proposed changes and many of which raised significant legal and policy concerns with the 2019 Proposed Rule. Commenters objected that the proposed changes did not align with case law, created problematic defenses and made discriminatory effects claims effectively impossible to plead and prove in many instances, thus contravening the core holding of Inclusive Communities. On September 24, 2020, HUD published a final rule titled “HUD's implementation of the Fair Housing Act's Disparate Impact Standard” (“2020 Rule”), which, among other things removed the definition of discriminatory effect, added demanding pleading elements that made it far more difficult to initiate a case, altered the burden-shifting framework, created new defenses, and limited available remedies in disparate impact claims.

Massachusetts Fair Housing Ctr. v. HUD Order Staying Implementation of the 2020 Rule

Following publication of the 2020 Rule, HUD was sued in three separate federal courts—: Massachusetts Fair Housing Ctr., et al. v. HUD, No. 3:20–cv–11765 (D. Mass.); Nat'l Fair Hous. All., et al. v. HUD, No. 3:20–cv–07388 (N.D. Cal.); Open Cmtys., et al. v. HUD, No. 3:20–cv–01587 (D. Conn.). The plaintiffs in each case contended that the 2020 Rule was invalid because it was inconsistent with the Act and its promulgation violated the Administrative Procedure Act (“APA”). Prior to the effective date of the 2020 Rule, the U.S. District Court for the District of Massachusetts in Massachusetts Fair Housing Ctr. v. HUD issued a preliminary injunction staying the implementation and postponing the effective date of the 2020 Rule. Because of this preliminary injunction, the 2020 Rule never took effect, and the 2013 Rule remained in effect.

In its order, the district court preliminarily found that many significant changes made by the 2020 Rule were likely not supported by Inclusive Communities or other case law. Similarly, the court concluded that the 2020 Rule did not appear to bring the clarity to the discriminatory effects framework that it was intended to foster, but rather introduced new concepts that had never been part of disparate impact case law without fully explaining their meaning. In support of its conclusions, the court identified numerous provisions in the 2020 Rule as problematic, including § 100.500(b) (“requiring at `the pleadings stage,' among other things, that plaintiffs `sufficiently plead facts to support' . . . `[t]hat the challenged policy or practice is arbitrary, artificial, and unnecessary to achieve a valid interest or legitimate objective such as a practical business, profit, policy consideration, or requirement of law' ”); § 100.500(c)(2) (permitting defendants to “ `rebut a plaintiff's allegation under (b)(1) . . . that the challenged policy or practice is arbitrary, artificial, and unnecessary by producing evidence showing that the challenged policy or practice' merely ` advances a valid interest ' ”) (emphasis in original); § 100.500(c)(3) (requiring “at the third step of the burden-shifting framework that the plaintiff prove `a less discriminatory policy or practice exists that would serve the defendant's identified interest (or interests) in an equally effective manner without imposing materially greater costs on, or creating other material burdens for, the defendant' ” (emphasis in original)); § 100.500(d)(1) and (d)(2)(iii) (“conflating of a plaintiff's prima facie burden and pleading burden”); and § 100.500(d)(2)(i) (the outcome prediction defense).

The district court found that the “practical business, profit, policy consideration” language, the “outcome prediction” defense, changes to the third element of the burden-shifting framework, and the conflating of a plaintiff's prima facie burden and pleading burden, ran the risk of “effectively neutering” discriminatory effects liability under the Act, and were all likely unsupported by Inclusive Communities or other judicial decisions. The district court also stated that the 2020 Rule's use of “new and undefined terminology altered the burden-shifting framework, and perplexing defenses” accomplished “the opposite of clarity” and were likely “arbitrary and capricious.” The court stated that “[t]here can be no doubt that the 2020 Rule weakens, for housing discrimination victims and fair housing organizations, disparate impact liability under the Fair Housing Act. . . . In addition, the 2020 Rule arms defendants with broad new defenses which appear to make it easier for offending defendants to dodge liability and more difficult for plaintiffs to succeed. In short, these changes constitute a massive overhaul of HUD's disparate impact standards, to the benefit of putative defendants and to the detriment of putative plaintiffs.”

HUD's Reconsideration of the 2020 Rule and the 2021 Notice of Proposed Rulemaking

On January 26, 2021, President Biden issued a Memorandum ordering the Department to “take all steps necessary to examine the effects of the [2020 Rule], including the effect that amending the [2013 Rule] has had on HUD's statutory duty to ensure compliance with the Fair Housing Act” and “take any necessary steps . . . to implement the Fair Housing Act's requirements that HUD administer its programs in a manner that . . . furthers . . . HUD's overall duty to administer the Act [ ] including by preventing practices with an unjustified discriminatory effect.”

Consistent with the President's Memorandum, HUD began a process to reconsider the 2020 Rule. On June 25, 2021, after reviewing prior public comments on the previous rulemakings described above, HUD's responses to those comments, HUD's 2016 supplemental explanation regarding the 2013 Rule's applicability to the insurance industry, legal precedent including Inclusive Communities, the Massachusetts Fair Housing Center court's order, and HUD's own experience with discriminatory effects cases over 40 years, HUD promulgated a proposed rule titled “Reinstatement of HUD's Discriminatory Effects Standard” (“proposed rule”) that proposed to recodify the 2013 Rule. The proposed rule advocated returning to the 2013 Rule because HUD believed that the 2013 Rule established a workable framework that was more consistent with existing case law and the purpose of the Act than the 2020 Rule.

As HUD described in the proposed rule, in HUD's experience, the 2013 Rule set a more appropriately balanced standard for pleading, proving, and defending a fair housing case alleging that a policy or practice has a discriminatory effect. HUD believed that the 2013 Rule provided greater clarity about what each party must show by relying on concepts that have a long history in judicial and agency precedent and that it appropriately balanced the need to ensure that frivolous claims do not go forward with a realistic understanding of the practical challenges to litigating these claims. With regard to the 2020 Rule, HUD's experience investigating and prosecuting discriminatory effects cases informed its views that many of the points made by commenters and the District Court in Massachusetts Fair Housing Center were, in HUD's opinion, correct. In particular, the changes the 2020 Rule made, such as amending pleading standards, changing the burden shifting framework, and adding defenses, all operated to tip the scales in favor of respondents, introduced unnecessary confusion, may have precluded otherwise valid claims, and, at worst would have made discriminatory effects liability a practical nullity.

HUD further stated its belief that the 2013 Rule was more consistent with the Act's purpose; prior case law under the Act, including Inclusive Communities; other civil rights authorities, including the Equal Credit Opportunity Act and Title VII; and HUD's prior interpretations of the Act. In its 2020 Rule, HUD noted that the rule was intended to better reflect Inclusive Communities, but HUD now believes that the 2020 Rule was itself inconsistent with the holding of Inclusive Communities, which maintained the fundamentals of long-established disparate-impact precedent rather than changing them. Moreover, based on HUD's experience investigating and litigating discriminatory effects cases, HUD believed that the practical effect of the 2020 Rule's amendments was to severely limit HUD's and plaintiffs' use of the discriminatory effects framework in ways that would substantially diminish that frameworks' effectiveness in accomplishing the purposes that Inclusive Communities articulated.

By comparison, in HUD's experience, the 2013 Rule provided a workable and balanced framework for investigating and litigating discriminatory effects claims that is consistent with the Act, HUD's own guidance, Inclusive Communities, and other jurisprudence.

HUD noted that Inclusive Communities heavily relied on Griggs, which is the foundation of Title VII disparate impact jurisprudence, to illustrate the well-settled principles of disparate impact under the Act, and HUD believed Inclusive Communities to be fully supportive of the 2013 Rule. Inclusive Communities explained that in Griggs, “[w]hat is required by Congress [in Title VII cases] is the removal of artificial, arbitrary, and unnecessary barriers to employment when the barriers operate invidiously to discriminate on the basis of racial or other impermissible classification.” Quoting from its foundational decision in Griggs, the Supreme Court in Inclusive Communities observed that “[d]isparate impact liability mandates the `removal of artificial, arbitrary, and unnecessary barriers,' not the displacement of valid governmental policies.” HUD proposed that this quotation from a seminal decision of longstanding disparate impact doctrine is properly read as maintaining existing law, not changing it. HUD highlighted that Inclusive Communities explicitly stated, “disparate-impact liability has always been properly limited in key respects” (emphasis added), making clear that the Court was not adding additional pleading or proof requirements or calling for a significant departure from pre-existing precedent under the Act and Title VII. Furthermore, HUD stated that reading Inclusive Communities to support a heightened pleading standard is contradicted by the fact that the “heartland” cases cited by the Court would not have survived a motion to dismiss under that standard because plaintiffs in those cases did not have specific facts to plausibly allege that a policy or practice was arbitrary, artificial, or unnecessary until after discovery. Finally, HUD explained that because Inclusive Communities considered a judgment reached after discovery and bench trial, the Court had no occasion or opportunity to consider the proper pleading standards for cases brought under the Act. The parties did not brief or argue such questions to the Court, making it particularly unlikely that the Court intended to reach them.

For these reasons and others, HUD proposed that Inclusive Communities' quotation of Griggs' decades-old “artificial, arbitrary, and unnecessary” formulation would be best construed as maintaining continuity with longstanding disparate-impact jurisprudence, as reflected in the 2013 Rule. HUD stated in the proposed rule its belief that other changes the 2020 Rule made would create problems that could be cured by a return to the 2013 Rule. For example, the 2020 Rule eliminated the 2013 Rule's definition of “discriminatory effect,” stating that the definition was unnecessary because it “simply reiterated the elements of a disparate impact claim.” In eliminating this definition, the 2020 Rule erased “perpetuation of segregation” as a recognized type of discriminatory effect distinct from disparate impact, which was contrary to well established precedent. HUD proposed to reaffirm that perpetuation of segregation remains, as it always had been, a basis for contending that a policy has an unlawful discriminatory effect.

HUD described how the 2020 Rule also eliminated from the Act's prohibitions policies or practices that could “predictably result[] in a disparate impact on a group of persons,” i.e., those for which the disparate impact has not yet manifested but will predictably do so. HUD noted, as it stated in 2013, that the Act prohibits discrimination that is predictable because it defines an “aggrieved person” as any person who “believes that such person will be injured by a discriminatory housing practice that is about to occur.” HUD noted that courts have found that predictable discriminatory effects may violate the Act: “[t]o establish a prima facie case of racial discrimination, the plaintiff need prove no more than that the conduct of the defendant actually or predictably results in racial discrimination; in other words, that it has a discriminatory effect.” HUD stated in the proposed rule that the 2020 Rule did not adequately explain how the Act and case law construing it can be read to require waiting until harm is inflicted before an action with predictable discriminatory effects can be challenged, nor did HUD perceive that any such explanation would be availing, given the plain language of the Act and the case law interpreting it.

In addition, in the 2021 proposed rule, HUD recognized and agreed with concerns that the 2020 Rule created new and confusing defenses at both the pleading and post-pleading stage, including the new defense allowing a defendant to show that the challenged policy or practice is “reasonably necessary to comply with a third-party requirement.” The 2020 Rule's preamble stated that this defense would not require a showing that the challenged policy is the only way to comply with such a requirement, only that the policy serves that purpose. In the 2021 proposed rule, HUD stated that this new defense was inconsistent with the Act, which specifies that state and local laws requiring or permitting discriminatory housing practices are invalid. HUD expressed its concern that the defense would preclude many otherwise proper discriminatory effects claims, because, for example, a plaintiff may not have any practical means of knowing whether some other party's policies also contributed to the defendant's practice. HUD reasoned that nothing in Inclusive Communities suggests this defense is required, let alone reasonable, for the agency to create.

HUD noted further in the proposed rule that the 2020 Rule also created a new “outcome prediction” defense which HUD believed would in practice exempt most insurance industry practices (and many other housing-related practices that rely on outcome predictions, such as lending practices) from liability under a disparate impact standard. In the proposed rule, HUD stated that it considered this defense to be inconsistent with HUD's repeated finding, including in the 2020 Rule, that “a general waiver of disparate impact law for the insurance industry would be inappropriate.” HUD reconsidered the defense and explained in the proposed rule that it believed the defense was unclear and would suggest that comparators be used, which were, in HUD's experience, inappropriate. HUD stated that at the very least, the defense would introduce unnecessary confusion into the doctrine.

In the proposed rule, HUD explained that the 2020 Rule inappropriately limited remedies in discriminatory effects cases in three respects. It specified that “remedies should be concentrated on eliminating or reforming the discriminatory practice so as to eliminate disparities between persons in a particular protected class and other persons.” It prohibited HUD in administrative proceedings from pursuing anything but “equitable remedies” except that “where pecuniary damage is proved, HUD will seek compensatory damages or restitution.” And it restricted HUD from seeking civil penalties in discriminatory effects cases unless the respondent had been adjudged within the last 5 years to have committed intentional unlawful housing discrimination under the Act. In the proposed rule, HUD proposed that these limitations have no basis in law and run contrary to public interest and the purpose of the Act. While the 2020 Rule cited Inclusive Communities as supporting these limitations, HUD noted that no part of Inclusive Communities suggested such limitations. Moreover, HUD viewed these limitations as in conflict with the plain language of the Act, which provides in all cases for a wide variety of remedies, including injunctive relief, actual damages, punitive damages, and civil penalties. HUD clarified that whereas Congress explicitly has limited the remedies available in disparate impact cases under Title VII, it has chosen not to do so in cases brought under the Act.

In sum, HUD stated in the proposed rule that it believed that the 2013 Rule would be preferable to the 2020 Rule. It believed the 2013 Rule would be more consistent with judicial precedent construing the Fair Housing Act, including Inclusive Communities, as well as the Act's broad remedial purpose. Based on its experience interpreting and enforcing the Act, HUD also believed the 2020 Rule, if put into effect, threatened to limit the effectiveness of the Act's discriminatory effects doctrine in ways that are inconsistent with the doctrine continuing to play its critical role in “moving the Nation toward a more integrated society.” Furthermore, HUD stated that it believed that the 2013 Rule provided clarity, consistency, and a workable, balanced framework, recognized by the Supreme Court, under which to analyze discriminatory effects claims, and under which HUD could better ensure it has the tools to further its “duty to administer the Act [ ] including by preventing practices with an unjustified discriminatory effect.”

II. This Final Rule

HUD received 10,113 comments in response to the proposed rule. HUD reviewed and carefully considered these comments and, as explained in the responses to the comments below, HUD has decided to recodify the 2013 Rule. HUD has confirmed that the concerns it expressed in the proposed rule are consistent with the public comments received in response to the proposed rule, HUD's previous rulemakings and notices, and relevant discriminatory effects case law under the Act, including cases using the 2013 Rule and the 2020 Rule.

HUD continues to believe that, as compared to the 2020 Rule, the 2013 Rule more accurately describes discriminatory effects law in a manner that is consistent with both the Act and the Supreme Court's ruling in Inclusive Communities. As in the 2013 Rule, this final rule does not impose any new liability, but merely provides a consistent, nationwide framework for determining whether a given practice has an unjustified discriminatory effect, leading to liability under the Act. HUD believes the 2013 Rule best aligns with Fair Housing Act jurisprudence and is most consistent with the Act's remedial purposes. As described in greater detail below, HUD believes that the 2013 standard is consistent with and was implicitly endorsed by Inclusive Communities.

Moreover, even if the 2020 Rule were a permissible approach to discriminatory effects law and HUD had no doubts about the legality or appropriateness of the 2020 Rule under the Act, HUD would recodify the 2013 Rule as an exercise of the discretion Congress gave HUD to make rules under the Act. The 2013 Rule's framework is practical and, in contrast to the novel and complicated 2020 Rule, has worked well in discriminatory effects cases. The 2013 Rule's framework adequately balances the interests of plaintiffs and defendants and encourages the latter to seek a less discriminatory alternative when a policy or practice causes a discriminatory effect, without imposing an excessive burden on their substantial, legitimate, non-discriminatory interests. As described in greater detail below, HUD declines to create any exemptions or safe harbors in this rule or to proscribe specific conduct that per se has an unjustified discriminatory effect. As Inclusive Communities recognized in affirming that discriminatory effects claims are cognizable under the Act, “the [Fair Housing Act] must play an important part in avoiding the Kerner Commission's grim prophecy that `[o]ur Nation is moving toward two societies, one black, one white—separate and unequal.' ” For the reasons discussed in HUD's 2013 Rule, in the proposed rule, and below in response to the public comments, HUD rescinds the 2020 Rule and recodifies the 2013 Rule.

HUD adopts one amendment made by the 2020 Rule to HUD's general fair housing regulations at § 100.70(d)(5). This amendment provides additional illustrations of prohibited activities under the Fair Housing Act generally, though it is not specific to discriminatory effects cases. HUD proposed keeping these additional examples in the proposed rule and received no public comments specifically opposing these additions. In this final rule's amendatory instructions, HUD includes instructions to “republish” § 100.70(d)(5) without change from the 2020 Rule to clearly show that HUD is adopting this language in this final rule.

III. Public Comments

General Comments in Support

Commenters generally supported the proposed rule, which would reinstate the 2013 Rule. Commenters stated that the proposed rule is consistent with President Biden's memorandum directing agencies to redress America's history of housing discrimination and the 1994 interagency fair lending guidance under the Act and the Equal Credit Opportunity Act. Commenters also stated that the proposed rule is an important and appropriate exercise of HUD's rulemaking authority.

Among the supportive comments were those stating that the proposed rule: is appropriately broad, inclusive, and will be instrumental in ensuring optimal compliance with the Act and in challenging covert or latent discrimination that can be intentionally or unintentionally embedded in facially neutral policies and practices; is critical for ensuring equal opportunity under the Act; would help secure equal opportunity in a wide variety of housing areas, including in land use and zoning, affordable and public housing, environmental permitting, air quality, and utility burdens; would be effective in protecting against housing discrimination based on all of the Act's protected characteristics, as well as related groups such as persons without English language proficiency or who are survivors of domestic violence or sexual assault; would advance sustainable homeownership and affordable housing programs; would benefit both real estate professionals and consumers; may prevent segregated housing patterns that might otherwise result from covert and illicit stereotyping; is essential to challenging blanket refusals to accept Housing Choice Vouchers, which are disproportionately used by people of color, households with children, and persons with disabilities; and would address de facto and de jure discrimination in housing policies, construction, and tenancy.

Commenters noted that the proposed rule's burden-shifting framework is consistent with long-standing case law, including Inclusive Communities, and well-established agency practice. Commenters explained that the proposed rule contains the traditional burden shifting framework for disparate impact claims, which was endorsed by the Supreme Court in Inclusive Communities and is consistent with the framework for disparate impact claims under Title VII and the Equal Credit Opportunity Act.

Commenters stated that out of more than 40 federal appellate and district court decisions in disparate-impact fair housing cases following Inclusive Communities, very few, other than Inclusive Communities Project v. Lincoln Prop. Co., found any inconsistency between the 2013 Rule and the Supreme Court's Inclusive Communities decision. Commenters pointed to Avenue 6E Investments, LLC v. City of Yuma , which cited the 2013 Rule as authority for the proper burden-shifting framework without noting any inconsistencies between that rule and Inclusive Communities, and Mhany Mgmt., Inc. v. Cnty. of Nassau , which found that the Supreme Court implicitly endorsed the 2013 Rule's framework in Inclusive Communities. Commenters also noted that the court in Mhany Mgmt., Inc. v. Cnty. of Nassau, as well as numerous other cases successfully utilized the 2013 Rule's burden shifting framework to reach decisions.

Commenters supporting the proposed rule stated that it provides a clear, simple, and effective standard that would promote consistency between judicial and administrative venues and throughout the housing industry. Commenters explained that this standard would maintain continuity for regulated entities and enable them to better comply with the Act, since this regulatory framework has been in place since 2013. Commenters described the framework as pragmatic, fostering fair and sound business practices and finding the appropriate balance between fair housing concerns and business necessities.

Commenters expressed support for the burden-shifting framework, describing it as clear, easy to follow, practical, and striking the appropriate balance between competing interests. Commenters stated that the 2013 Rule settled the law on several important issues, including whether the burden-shifting framework is appropriate and which party bears the burden of demonstrating the business necessity for a particular policy and the existence of a less discriminatory alternative. A commenter noted that the 2013 Rule is a fair and accurate codification of longstanding jurisprudence of discriminatory effects liability under the Act and posed no significant departure from previous HUD interpretation or the weight of judicial authority. Commenters noted that plaintiff's burden under the proposed rule is not easy to meet, which eliminates the danger of an onslaught of groundless litigation. A commenter described the proposed rule as balancing the need to prevent frivolous claims from moving forward with a process that allows potentially meritorious claims to be substantiated or disproved. A commenter compared the proposed rule's three-tiered framework to the 2020 Rule's five-tiered test, noting that the former provides a clear way to challenge policies that may unnecessarily restrict housing, while the latter is vague and allows discrimination to continue unchallenged. Comments also stated that the 2020 Rule conflicted with decades of legal precedent, including the Supreme Court's decision in Inclusive Communities and that discriminatory effects claims that sought to challenge neutral policies that actually caused discrimination would not survive under the test contained in the 2020 Rule.

General Comments in Opposition

Other commenters generally opposed the proposed rule, suggesting that HUD withdraw it and retain the 2020 Rule. A commenter stated that the 2020 Rule thoroughly explained its reasoning and was consistent with Inclusive Communities. Another commenter described the proposed rule as unclear and overly burdensome. Commenters also suggested that the proposed rule lacks limitations on how and where it applies, thus adding a new layer of complexity and uncertainty to discriminatory effects law. A commenter stated that the proposed rule would harm the people it purports to benefit by applying a complex, court-created legal framework to a public policy issue and requiring all issues to be resolved in expensive litigation in federal court. Another commenter stated that the proposed rule will not create a uniform mechanism to resolve discriminatory effects disputes but will instead encourage courts to develop alternative approaches to handling such cases. A commenter stated that HUD and others have used the 2013 Rule to bully housing providers into expanding access to housing even if landlords cite legitimate business reasons for restricting housing based on certain admission or occupancy policies.

HUD Response: HUD disagrees with the commenters who opposed the proposed rule. As discussed in the preamble to the proposed rule and elsewhere in this preamble, HUD believes that this final rule establishes the appropriate, balanced framework for assessing claims of discriminatory effects and is entirely consistent with Inclusive Communities and long-standing judicial precedent. In contrast, HUD finds that the 2020 rule, if retained, would limit liability in a manner inconsistent with the Act's purpose and judicial precedent. HUD further believes that some of the standards announced in the 2020 rule might lead some courts to develop alternative approaches to assessing discriminatory effects claims that are inconsistent with the text and broad remedial purposes of the Act. HUD believes that the framework in the proposed rule sets out a consistent nationwide approach to evaluating discriminatory effects claims and adopts the majority view of judicial opinions interpreting the Act. As a result, this final rule affords housing providers the opportunity to maintain policies and practices so long as they do not have an unjustified discriminatory effect because of a protected characteristic. And it does not require allegations of discriminatory effects to be resolved in federal court. Rather, housing providers may avoid potential litigation and liability by reviewing their policies and practices to ensure that they do not have an unjustified discriminatory effect. The discriminatory effects framework is not intended to force housing providers to take any particular course of action but rather to ensure that an important goal of the Act—to safeguard fair housing throughout the country—is accomplished.

General Comments Concerning Clarity

Issue: Commenters disagreed about the clarity that would result from setting aside the 2020 Rule. A commenter stated that the 2020 Rule should be retracted because it created a legal landscape in which HUD, other federal regulators, and courts would have different standards for analyzing discriminatory effects claims, and because it created confusion that would disadvantage housing discrimination victims. However, other commenters asked HUD to retain the 2020 Rule so as to avoid confusion and uncertainty because different forms of the rule have been promulgated and retracted over the last several years. A commenter stated that HUD should recognize the practical implications of repeatedly and drastically changing policies and justification for those policies and requested that HUD solidify clear and consistent long-term standards in order to minimize confusion and uncertainty for federal funding recipients. The commenter said it makes little sense to change procedures with each new administration and that reinstating the 2013 Rule will provoke litigation and disputes between courts rather than provide clarity. Another commenter noted a particular concern about confusion for businesses and damage to their ability to know and comply with the law since litigation concerning the 2020 Rule is pending.

HUD Response: HUD agrees with the commenters who stated that the 2020 Rule introduced a new standard that is incompatible with the standards used by courts and other federal regulators, creating confusion and uncertainty. In contrast, this final rule will provide clarity consistent with well-established judicial and agency interpretations of the Act by eliminating the novel and undefined standards introduced by the 2020 Rule. HUD also notes that the 2020 Rule never went into effect and has never been enforced by HUD. HUD has considered potential reliance interests and believes that no significant reliance was created by the 2020 rule, because unlike a regulation that even briefly governed conduct or supplied benefits, the 2020 Rule never did so. While HUD proposed revising the rule in 2019 and subsequently issued a final rule in 2020, the 2013 Rule, which is recodified in this final rule, is and has been the only promulgated rule governing the standard for discriminatory effects liability that has ever taken effect since the Act became law in 1968. HUD agrees that the 2020 Rule introduced a new standard that is incompatible with the Act and with the standards used by courts and other federal regulators. Had HUD used the 2020 Rule, while other federal agencies and courts used rules analogous to the 2013 Rule or created their own rules in response to Inclusive Communities, there would be substantial confusion in discriminatory effects jurisprudence. HUD believes that it is important that those affected by or accused of discrimination know what standard governs their housing related activities and that that standard does not unnecessarily vary depending on the forum in which a case is decided. Having differing standards would increase litigation costs for the parties and likely result in the dismissal of claims in some forums that are upheld in others. Restoring the 2013 Rule will help ensure the consistency of federal discriminatory effects law and will avoid the confusion caused by the 2020 Rule.

This final rule sets out a usable and uniform framework that is fully consistent with the requirements established by courts, as well as the text and purpose of the Act.

Comments Concerning Harmony Between Other State and Federal Civil Rights Statutes

Issue: A commenter noted that the Rule will bring HUD's regulations back into conformity with state civil rights laws.

HUD Response: HUD acknowledges that many state courts and agencies that interpret and enforce civil rights laws utilize a burden-shifting framework that is similar to this final rule and that HUD's 2020 Rule created confusion and conflicting standards. HUD believes that it is important for plaintiffs to have access to consistent relief in state and federal jurisdictions.

Issue: Commenters applauded the rule for being consistent with other civil rights laws and their discriminatory effects liability frameworks, including Title VII and ECOA. A commenter also noted that courts, including the Supreme Court in Inclusive Communities, have often drawn on Title VII's jurisprudence when interpreting the Act and vice versa because of the similarities between the statutes' texts, structures, purposes, and dates of enactment. The commenter expressed support for the rule because it aligns with judicial precedent that interprets the Act and Title VII similarly. The commenter also stated that the proposed rule furthers the principle that language that is similar across statutes should be given similar meaning.

HUD Response: HUD agrees that the rule is consistent with other civil rights laws and their discriminatory effects liability frameworks, including Title VII of the Civil Rights Act of 1964, as amended (Title VII), and the Equal Credit Opportunity Act (ECOA). HUD acknowledges that courts have generally interpreted these statutes consistently and agrees that HUD should do the same to promote consistency and clarity, particularly for entities whose actions must be compliant with both ECOA and the Act.

HUD notes that the preamble to the 2013 Rule explained in great detail how its framework operates harmoniously with other civil rights laws, including Title VII and ECOA, and best effectuated the important goals of the Fair Housing Act. As HUD noted in the 2013 Rule, the discriminatory effects framework borrowed from Title VII and Griggs is the fairest and most reasonable approach for resolving disparate impact claims, in part because it does not require either party to prove a negative, and it provides the parties the opportunity to obtain adequate information in discovery to meet their burdens.

Comments Concerning Massachusetts Fair Housing Center

Issue: Commenters stated that although the district court in Massachusetts Fair Housing Center stayed implementation of the 2020 Rule, it did not require HUD to totally abandon the 2020 Rule. The commenters stated that the decision primarily addressed three elements of the 2020 Rule—the outcome prediction defense, the requirement that plaintiffs present an equally effective alternative, and the conflation of the plaintiff's prima facie burden and their pleading burden. The commenters also stated that the court acknowledged the requirement that a plaintiff must plead that a challenged policy is “arbitrary, artificial, and unnecessary to achieve a valid interest or legitimate objective,” may have some grounding in case law. The commenters also stated that the court did not address the 2020 Rule's recognition that the Act does not and cannot supplant state laws concerning insurance, or its codification of Inclusive Communities' guidance on remedies.

Other commenters stated that Massachusetts Fair Housing Center criticized the 2020 Rule for introducing onerous pleading standards, defenses that lacked precedent in case law, for conflicting with the remedial purpose of the Act, and for likely being arbitrary and capricious.

HUD Response: While the Massachusetts Fair Housing Center court enjoined HUD from implementing or enforcing the 2020 Rule in any manner and ordered HUD to “preserve the status quo pursuant to the regulations in effect as of the date of this Order,” HUD is not basing its decision to abandon the 2020 Rule and recodify the 2013 Rule on the Massachusetts Fair Housing Center order. Rather, HUD declines to retain any part of the 2020 Rule's substantive disparate impact language based on its own interpretation of and decades of experience in implementing the Act. HUD also finds other aspects of the 2020 Rule that the court left unaddressed or uncriticized to be equally troublesome.

Comments Concerning Inclusive Communities

Issue: Commenters supported reinstatement of the 2013 Rule because it is consistent with Inclusive Communities. Commenters stated that the Court cited the 2013 Rule with approval, noting each step in the 2013 Rule's burden-shifting framework without critique. Commenters also noted that multiple courts since Inclusive Communities, including courts of appeals, have read Inclusive Communities as affirming or implicitly adopting the 2013 Rule's burden-shifting test and have applied the 2013 Rule's framework. A commenter pointed out that the district court in Inclusive Communities stated on remand that, “[a]s a result of the Fifth Circuit's decision adopting the HUD regulations, and the Supreme Court's affirmance (without altering the burden-shifting approach), the following proof regimen now applies to ICP's disparate impact claim under the [Act].” A commenter also cited multiple district court decisions that have incorporated the language of Inclusive Communities when applying the 2013 Rule's framework. Another commenter noted that Inclusive Communities endorsed “heartland” cases, all of which used burden shifting frameworks consistent with the proposed rule. Commenters also stated that the 2020 Rule did not meaningfully address MHANY Management, Inc., de Reyes v. Waples Mobile Home Park Limited Partnership, or Avenue 6E Investments, LLC v. City of Yuma, which found that the 2013 Rule remained valid after Inclusive Communities. A commenter added that in Property Casualty Insurance Association of America v. Carson, a lawsuit directly challenging the validity of the 2013 Rule, the district court held that Inclusive Communities affirmed HUD's burden-shifting approach and did not identify any aspect of the approach that required correction.

Other commenters opposed the proposed rule, stating that it is inconsistent with Inclusive Communities. In support of this, commenters noted that the 2013 Rule preceded Inclusive Communities and stated that the 2013 Rule does not adequately incorporate the holdings of that case. Commenters requested that HUD retain the 2020 Rule or incorporate additional language from the Inclusive Communities decision into this final rule. Commenters stated that although Inclusive Communities mentioned the 2013 Rule, it did not endorse the rule. Others stated that the 2013 Rule does not align with the Supreme Court's caution against injecting racial considerations into every housing decision and perpetuating race-based considerations rather than moving beyond them. A commenter said that compliance with the rule, as opposed to Inclusive Communities, will lead to costly litigation. Commenters noted that the Supreme Court specifically limited the scope of Inclusive Communities to the first question presented (whether disparate impact claims were cognizable under the Act) so references to the 2013 Rule cannot be viewed as approving the 2013 framework. Commenters further stated that the Court in Inclusive Communities did not state that the 2013 Rule incorporates the appropriate limits of disparate impact liability.

Another commenter stated that courts, such as the court in Woda Cooper Dev., Inc. v. City of Warner Robins, Civ. No. 5:20–CV–159 (MTT), 2021 WL 1093630, *1, at *7 (M.D. Ga. Mar. 22, 2021), have struggled to apply the 2013 Rule's framework in the wake of Inclusive Communities, with some choosing to ignore the rule entirely. The commenter stated that Inclusive Communities identified a number of safeguards to prevent abusive disparate impact cases but did not provide detailed explanations of those safeguards or guidance on how courts should apply those safeguards. The commenter urged HUD to elaborate on those safeguards in the final rule.

HUD Response: HUD agrees with the commenters who stated that the 2013 Rule is consistent with the Inclusive Communities holding. The Court in Inclusive Communities did not call into question the 2013 Rule's framework for analyzing discriminatory effects claims, nor did it suggest that HUD should make any modifications to that framework. To the contrary, the Court cited HUD's 2013 Rule several times with approval. For instance, the Court noted that the burden-shifting framework of Griggs and its progeny, adopted by HUD in the 2013 Rule and retained in this final rule, adequately balanced the interests of plaintiffs and defendants by giving housing providers the ability “to state and explain the valid interest served by their policies.” The Court also discussed the history of HUD's promulgation of the 2013 Rule, noted that lower courts had relied on it, and repeatedly cited its three-part burden shifting test. Notably, other courts have recognized these findings and relied on the 2013 Rule's burden shifting framework without difficulty since Inclusive Communities was decided. Moreover, HUD agrees that Inclusive Communities' discussion approving the holdings of the “heartland cases” supports reinstating the 2013 Rule. HUD also agrees that the 2020 Rule did not adequately address the well-considered and thorough reasoning of MHANY Mgmt., de Reyes, and Avenue 6E Investments, LLC, each of which found that the 2013 Rule remained valid after Inclusive Communities.

HUD disagrees with the commenters who stated that the 2020 Rule should be retained because it is consistent with and incorporates the “safeguards” described in Inclusive Communities. As discussed above, in Inclusive Communities, the Court did not express any disapproval of the 2013 Rule's framework or specify that it lacked any safeguards. Rather, the Court observed that “disparate-impact liability has always been properly limited in key respects,” making clear that it was not calling for any significant departure from pre-existing precedent under the Act or the 2013 Rule. HUD believes that had the Court intended to overhaul disparate impact jurisprudence, the Court would have done so expressly, rather than citing the 2013 Rule favorably. Moreover, HUD notes that the Court declined to accept certiorari on the proper standard for assessing disparate impact cases. And, as noted above, multiple courts have since read Inclusive Communities as affirming or endorsing the 2013 Rule's burden- shifting framework. Even if the Court did not endorse the 2013 Rule in Inclusive Communities, it did not discard or significantly alter preexisting disparate impact jurisprudence. The 2013 Rule adopts the majority view of preexisting law. HUD believes that to the extent that some courts have attempted to impose limitations greater than those described in the 2013 Rule, they have misread Inclusive Communities. Moreover, the 2013 Rule did not inject racial considerations into housing decisions, and nothing in Inclusive Communities indicates that the Court believed the Rule improperly did so. Accordingly, HUD continues to believe that the burden-shifting test articulated in the 2013 Rule is the most appropriate framework for litigating discriminatory effects claims consistent with the Act and Inclusive Communities.

Issue: Commenters cited Lincoln Property, Oviedo, River Cross Land Co., County of Cook, Ill. v. Wells Fargo & Co, and Nat'l Fair Hous. All. v. Travelers Indem. Co. as evidence that several courts have held that the 2013 Rule was inconsistent with Inclusive Communities. By contrast, other commenters stated that out of more than 40 federal appellate and district court decisions in disparate impact cases following Inclusive Communities, only Lincoln Property, an appellate decision, and district courts bound by Lincoln Property, found any inconsistency between the 2013 Rule and Inclusive Communities.

HUD Response: HUD disagrees that the cases the commenters cited compel the conclusion that this rule is inconsistent with Inclusive Communities. As HUD has previously stated on many occasions, including in the preamble to the 2020 Rule, the 2013 Rule is consistent with Inclusive Communities. The vast majority of courts to consider this issue subsequent to Inclusive Communities, including at least three federal appellate courts, have agreed. Multiple courts have specifically read Inclusive Communities to have affirmed or endorsed the 2013 Rule's burden-shifting framework. For example, in River Cross, one of the decisions commenters characterized as demonstrating incompatibility between the 2013 Rule and Inclusive Communities, the court in fact recognized that Inclusive Communities approvingly cited the 2013 Rule, applied the 2013 Rule, and found it to be easily reconciled with Inclusive Communities. HUD has determined that the small number of courts that reached contrary conclusions misinterpreted the scope of the Inclusive Communities holding, and HUD declines to adopt the minority views of these courts.

In light of the views of a majority of courts and HUD's experience applying the Act, HUD finds that the Fifth Circuit's conclusions in Lincoln Property do not require it to change course. In that case, the majority of a divided panel acknowledged that Inclusive Communities reviewed and affirmed the Fifth Circuit's earlier judgment in that case, remanding to the trial court to apply the 2013 Rule's burden-shifting framework, and that the Court did not explicitly call into question the 2013 Rule's requirements. Nonetheless, the Lincoln Property court found that because the Supreme Court in Inclusive Communities had not explicitly stated that it was adopting the 2013 Rule's framework, whether the Court accepted the framework or modified it remained unresolved. The court construed language from Inclusive Communities as calling for courts to make it more difficult to plead a discriminatory effects claim in some fashion, but acknowledged that Inclusive Communities provided no clear direction as to how it was thus changing the law. While acknowledging that other appellate courts had interpreted Inclusive Communities to have “implicitly adopted the 2013 framework,” the panel's review of certain passages from Inclusive Communities and of subsequent decisions from the Fourth, Eighth, and Eleventh Circuits led the panel to conclude simply that Inclusive Communities “announce[d] a more demanding test than that set forth in the HUD regulation” but “did not clearly delineate its meaning or requirements.” Finding no consensus even among those who believed Inclusive Communities made some change, it concluded that the claim at issue in that case was not properly pleaded under any of several possible standards it could apply, making it unnecessary to state with more specificity how, in its view, Inclusive Communities had changed the law.

HUD believes Lincoln Property's language concerning a more demanding standard is not a reason to change the standard it promulgated in 2013. As stated earlier, HUD disagrees that anything in Inclusive Communities is inconsistent with the 2013 Rule's requirements for discriminatory effects claims. Rather, HUD agrees with the Fourth Circuit that the 2013 Rule “is similar to the framework the Supreme Court ultimately adopted in Inclusive Communities,” and with its observation that “some courts believe the Supreme Court implicitly adopted the HUD framework altogether.” But even if the Fifth Circuit were correct in identifying inconsistencies between the 2013 Rule and Inclusive Communities, Lincoln Property does not provide persuasive reasoning for HUD to modify the 2013 Rule, because the court only found ambiguity in the law after Inclusive Communities rather than specifying the way in which HUD needed to change course. Additionally, the other circuit courts that have analyzed the robust causation discussion in Inclusive Communities have either defined it in a way that is consistent with this final rule or were similarly non-specific in explaining robust causality's meaning.

HUD notes that, while acknowledging that other appellate courts had interpreted Inclusive Communities to have “implicitly adopted the 2013 framework,” the Fifth Circuit panel's review of certain passages from Inclusive Communities as well as subsequent decisions from the Fourth, Eighth, and Eleventh Circuits, led the panel to conclude that Inclusive Communities “undoubtedly announce[d] a more demanding test than that set forth in the HUD regulation.” HUD believes that in two of these decisions, the courts gave more deference to the 2013 Rule than the commenters recognized. Additionally, in the district court decisions cited by the commenters, and in Lincoln Property' s progeny, HUD believes that the courts misread Inclusive Communities as creating heightened pleading standards. Even Lincoln Property only requires a plaintiff to plausibly demonstrate a robust causal connection between a discriminatory practice and an alleged disparate impact. HUD adopts the view of courts that found Inclusive Communities endorsed the 2013 Rule's framework.

HUD also notes that Lincoln Property —a suit between private parties—was decided without the benefit of input from HUD on what effect, if any, Inclusive Communities had on Fair Housing Act disparate impact claims. As the agency to which Congress has delegated the responsibility to interpret and enforce the Fair Housing Act, HUD believes that its reasonable reading of any ambiguities in the meaning of the Act following Inclusive Communities is entitled to deference. Thus, to the extent Lincoln Property identified such an ambiguity and came to conclusions that conflict with those HUD has reached, HUD declines to adopt the court's conclusions. Any risk that litigants in the Fifth Circuit would be subject to a different standard than litigants elsewhere is created by the Lincoln Property decision, not by HUD's promulgation of this rule.

In short, HUD does not believe that the cases cited by the commenters support revisions to the rule.

Issue: Commenters stated that the proposed rule conflicts with what they characterized as Inclusive Communities' holding that a “robust causality requirement . . . protects defendants from being held liable for racial disparities they did not create.” Some commenters asked HUD to expressly add a robust causality requirement to the final rule, while others asked HUD to retain the 2020 Rule, stating that it appropriately reflected Inclusive Communities' robust causality requirement.

Some commenters urged HUD to adopt the view that, in stating that disparate impact claims may not be established simply by demonstrating a “statistical disparity” in outcomes, Inclusive Communities held that such claims must meet a higher causation standard than in the proposed rule. Other commenters stated that the proposed rule does not require proximate cause or a direct link between the policy and the discriminatory effect, which, they said, Inclusive Communities requires. Commenters said that if plaintiffs are not required to establish “robust causality” or “direct proximate cause,” defendants would be liable in cases where discrimination does not actually exist. Commenters also stated that without an explicit robust causality requirement, race will be used in a pervasive way, leading to the use of numerical quotas and raising constitutional questions. Commenters stated that the requirement is necessary so that regulated entities can make practical business choices and profit-related decisions. A commenter suggested revising the proposed rule to provide that to establish robust causality, the plaintiffs have the burden of proving that a challenged practice is the sole and proximate cause, or reasonably predicted cause, of a discriminatory effect.

Commenters who supported the proposed rule said that it incorporates Inclusive Communities' protections for defendants who may fear liability for disparities their policies did not create. Commenters noted that the proposed rule does not permit liability based on statistical disparities alone.

HUD Response: The 2013 Rule and this final rule contain a robust causality requirement by requiring the plaintiff to prove at the first step of the framework that a challenged practice caused or predictably will cause a discriminatory effect. As discussed above, in HUD's view, the framework in this final rule, which includes the requirement that the challenged practice causes a discriminatory effect, is consistent with Inclusive Communities. The Inclusive Communities Court did not announce a heightened causality requirement for disparate impact liability, a requirement which would find no support in the statutory text or case law. Rather, in considering a district court opinion where the trial court had found a violation of the Act without ever requiring the plaintiff to identify a causal link between a specific policy and the challenged disparate impact, the Court merely reiterated that plaintiffs must identify a causal link between the challenged practice and the alleged disparate impact that is sufficiently robust to permit that connection to be scrutinized at each stage of the case. The 2013 Rule, and this final rule require exactly that. The 2013 Rule and this final rule do not use the precise words “robust causality” and (as explained elsewhere in this preamble) nothing in Inclusive Communities requires these words. What Inclusive Communities requires is that a court's examination of causality be robust. Both the 2013 Rule and this final rule implicitly incorporate this requirement by requiring a plaintiff to link a specific practice to a current or predictable disparity. Ultimately, the error identified both by the Fifth Circuit and then by the Supreme Court in Inclusive Communities came from the district court's failure to fully apply the 2013 Rule's framework, not the 2013 Rule's framework itself. Through its framework this rule ensures that, as required by Inclusive Communities, defendants are not held liable for racial disparities they did not create. The rule thus already requires a showing of causation, not just correlation, between the policy or practice and the disparate impact, and so is fully consistent with Inclusive Communities.

HUD also believes that the rule's burden-shifting framework does not preclude businesses from making business and profit-motivated choices, even if they cause a discriminatory effect, so long as they do not create an unjustified discriminatory effect. Once a plaintiff meets its burden of proving that a policy causes a disparate impact because of a protected characteristic, the burden then shifts to the defendant to prove that the policy is necessary to serve the defendant's substantial, legitimate, nondiscriminatory interest. This safeguard allows housing providers and others to make practical business choices and profit-related decisions. The third step of the framework then shifts the burden back to the plaintiff to prove that an alternative policy would have a less discriminatory effect than the challenged policy. This rule balances the interests of the parties by allowing defendants to implement policies that meet their needs, as long as there is no unjustified discriminatory effect, while providing plaintiffs the opportunity to identify policies that serve those needs with less discriminatory effects based on protected characteristics.

HUD notes further that although the 2013 Rule has been in effect for ten years—with similar judicial precedent effective even longer, it is unaware of any case applying the 2013 Rule in a manner that would impose quotas.

Issue: Commenters requested that HUD include in the final rule a requirement that plaintiffs plead that the challenged policy is “artificial, arbitrary, and unnecessary” in addition to the traditional elements of a disparate impact claim, as the 2020 Rule did. Commenters stated that Inclusive Communities required this additional element when the Court stated that “[d]isparate-impact liability mandates the `removal of artificial, arbitrary, and unnecessary barriers' ” to “avoid the serious constitutional questions that might arise under the Act, for instance, if such liability were imposed based solely on a showing of a statistical disparity.” Another commenter explained that the district court in Massachusetts Fair Housing Center did not invalidate the “arbitrary, artificial, and unnecessary” language in the 2020 Rule, but rather noted that it came from Inclusive Communities and other case law, like Ellis v. City of Minneapolis, 860 F.3d 1106, 1112 (8th Cir. 2017).

Other commenters disagreed, stating that if such a requirement were added to the rule, it would be impossible to challenge discriminatory policies absent facts showing discriminatory intent, thus negating Inclusive Communities' holding that violations of the Act may be established through proof of disparate impact. The commenters explained that pleading that a policy is “artificial” is essentially pleading that a policy is pretextual—a showing required in cases alleging intentional discrimination, not discriminatory effects. Commenters also noted that the phrase “artificial, arbitrary, and unnecessary” originated in Griggs and pointed out that in applying this phrase in Fair Housing Act cases, courts have applied it consistent with the 2013 Rule's burden shifting framework, essentially using it as short-hand for the three-step framework, not as a separate, independent element. As examples, these commenters cited City of Black Jack, which Inclusive Communities describes as a heartland case, as well as Graoch Assocs. #33, L.P. v. Louisville/Jefferson Cty. Metro Human Relations Comm'n. A commenter stated that the three-step burden-shifting framework, and especially the defense at the second step—that the policy was necessary to achieve a legitimate interest—already ensures that as the Inclusive Communities Court described, “disparate-impact liability mandates the `removal of artificial, arbitrary, and unnecessary barriers,' not the displacement of valid governmental policies.”

HUD Response: HUD declines to add an “artificial, arbitrary, and unnecessary” pleading standard or substantive element to this final rule. As previously explained, HUD does not construe Inclusive Communities to require the agency to add specific elements or pleading standards for disparate impact cases that go beyond what “has always” been required. Rather, when the Inclusive Communities Court quoted Griggs' decades-old formulation that disparate impact claims require the removal of artificial, arbitrary, and unnecessary barriers, it did so as part of restating the safeguards and requirements that it found (and HUD agrees) have always been a part of disparate impact jurisprudence. In this context, the Court quoted Griggs' short-hand formulation for the type of policy that traditionally has been held to create an unjustified discriminatory effect at the end of the burden shifting analysis. HUD believes that Inclusive Communities, following Griggs as well as earlier Fair Housing Act cases, went on to describe policies invalidated by longstanding precedent as either “arbitrary” or “artificial” as a shorthand for those found to violate the Fair Housing Act under traditional jurisprudence. HUD does not believe this language, when read in context, is best read to require the agency to impose a requirement for plaintiffs and the charging party to plead and prove, in addition to the traditional elements, that policies are artificial and arbitrary and unnecessary. HUD notes, moreover, that the source of this language is Griggs, a decades-old case at the bedrock of disparate impact jurisprudence, and notes that Griggs did not require plaintiffs to establish that the practice at issue met each of these three descriptors, let alone that such evidence be pleaded in a complaint. In addition, HUD believes that reading Inclusive Communities or other cases to support a heightened pleading standard for plaintiffs, such as in the 2020 Rule, is contradicted by the fact that the “heartland” cases cited favorably by the Court would not have survived a motion to dismiss under that standard because plaintiffs in those cases did not allege facts that would plausibly support a claim that a policy or practice was arbitrary, artificial, and unnecessary to the extent those terms are construed as requiring more than satisfaction of the traditional elements. Simply put, in HUD's experience implementing the Fair Housing Act, plaintiffs likely would not have had access to such facts until after discovery. HUD further believes that adding such a standard would also conflict with the text and broad remedial purpose of the Act which provides “within constitutional limitations, for fair housing throughout the United States.” HUD thus concludes that a heightened pleading and proof standard would frustrate the clearly expressed intent to use the maximum allowable power under the law to secure equal housing opportunity. Finally, HUD observes that Inclusive Communities did not specify how courts and agencies should apply a new pleading and proof standard, nor did it come close to clearly stating that it intended to create new elements. To the extent that leaves ambiguity in the law, as a matter of policy, HUD believes it is preferable to retain existing standards that have decades of case law and administrative actions specifying their content rather than impose ones that are undefined and untested.

Comments on Bank of America

Issue: Commenters stated that the proposed rule is inconsistent with Bank of America Corp. v. City of Miami, a 2017 Supreme Court case which held that “proximate cause under the [Act] requires some direct relation between the injury asserted and the injurious conduct alleged.” A commenter suggested that HUD add the phrase “some direct relation” to the proposed rule's burden of proof standard. Another commenter suggested revising the proposed rule to provide that in order to establish a “robust causal link,” the plaintiffs have the burden of proving that a challenged practice is the sole and proximate cause, or reasonably predicted cause, of a discriminatory effect.” Another commenter suggested that HUD state that the causation analysis must consider whether a practice is too remote to give rise to liability.

HUD Response: HUD believes that it is not required to add language to this rule to ensure consistency with Bank of America. In that case, which involved a municipality suing a lender on the theory that predatory lending practices had caused foreclosures which in turn eventually led to damages to the municipality such as reduced tax revenues, the Supreme Court held that, because actions for damages under the Act are akin to tort actions, such suits are “subject to the common-law requirement that loss is attributable to the proximate cause, and not to any remote cause.” The Court declined to further explain the proximate cause requirement as applied to Fair Housing Act claims and did not suggest that such a requirement would otherwise alter analyses under the Act. For example, HUD believes that Bank of America has no impact on the ability of organizational plaintiffs to prove standing by tracing their injuries to the challenged policy.

HUD believes, although the Bank of America decision was in the context of a disparate impact claim, it is not inherently specific to and does not create an additional burden for disparate impact claims. To the contrary, HUD believes that the proximate cause requirement Bank of America described for standing applies to all Fair Housing Act cases, not just disparate-impact claims, and so HUD does not believe it is appropriate to add a proximate-cause requirement to the regulatory requirements that are specific to disparate-impact claims. More broadly, this rule does not purport to address the requirements for Fair Housing Act standing, and neither Bank of America nor any other case requires HUD to add such considerations to this rule. Accordingly, HUD believes that adding the suggested language to this final rule, which purports only to set out the framework for analyzing the merits of disparate impact claims, is unnecessary. Nothing in this rule creates a conflict with Bank of America or bars a court from applying its requirements. This rule simply does not touch on that subject matter.

HUD additionally observes that, in its view, Bank of America applies to claims such as the one in that case that involve unusual claims in which the policy challenged has an unusually attenuated connection to the alleged harm to the plaintiff. HUD does not construe Bank of America as having a larger impact on longstanding principles of Fair Housing Act standing.

Discriminatory Effects as Applied to Insurance ¹¹³

Issue: Commenters asked HUD to exempt homeowners insurance—in whole or in part, as well as risk-based pricing and underwriting in particular—from liability for any unjustified discriminatory effects, advancing a number of reasons. Among other things, commenters stated that the fundamental nature of insurance does not allow discriminatory effects liability; such claims cannot succeed as a matter of law; and the McCarran-Ferguson Act bars claims. A commenter said that applying the rule to insurers is unnecessary because there have been no allegations or findings of unlawful discriminatory effects against an insurer prior to or since 2013. Other commenters disagreed, stating that HUD should not create exceptions for any industry, including insurance, because such categorical exemptions are unworkable and inconsistent with the Act's purpose, which is broad and inclusive. Commenters also stated that exemptions would allow some discriminatory practices to go uncorrected.

HUD Response: HUD declines to provide an exemption for the insurance industry in whole or in part. HUD responds below to the specific reasons commenters advanced for exempting homeowners insurance. However, as a threshold matter, HUD lacks the authority to create exemptions that are not in the text of the Act. When Congress passed the Act in 1968 and amended it in 1988, it established exemptions for certain practices but not for insurance. Furthermore, courts have routinely applied the Act to insurers and have found that discriminatory effects liability applies to insurers under the Act. Moreover, nothing in this rule precludes insurers from raising a defense based on the McCarran-Ferguson Act or from arguing that claims cannot succeed as a matter of law in particular cases. What HUD is declining to do, and what it believes it has no authority to do, is provide a single industry or a set of specific practices a blanket exemption from liability from all claims regardless of whether those claims otherwise would satisfy the rule's (and the Act's) requirements.

As further explained above and below, the Fair Housing Act was intended to have a very broad impact on housing and communities across the country. The plain text, purpose, and structure purpose, structure, and plain language of the Act make clear that the Act was intended to apply to all sectors of the housing industry so that each would have common duties under the Act. For example, the plain text of the Act does not refer to an actor, but rather a prohibited action, meaning that all actors in all sectors of the housing industry are subject to the Act. With regard to purpose, the Act was enacted to replace segregated neighborhoods with “truly integrated and balanced living patterns.” It was structured to address discriminatory housing practices that affect “the whole community” as well as particular segments of the community, with the goal of advancing equal opportunity in housing, and to “achieve racial integration for the benefit of all people in the United States.”

The Supreme Court in Inclusive Communities similarly noted that the Act “was enacted to eradicate discriminatory practices within a sector of our Nation's economy” and discussed that the viability of disparate impact claims is “consistent” with the Act's “central purpose.” In order to “eradicate” discriminatory practices within the housing sector, as the Court acknowledged was the purpose of the Act, it would logically flow that the Act was intended to apply to all sectors of the housing industry. Notably, the court used strong language, saying the purpose was to “eradicate,” rather than weaker language like “reduce”, making clear that the Act was meant to reach all sectors, otherwise eradication would not be possible. Nor did the Court suggest that any portion of the housing sector was not reached by the Act.

In HUD's experience, insurance plays a significant role in the housing industry and in securing equal opportunity in housing in communities nationwide. Home seekers must be able to access mortgage insurance and homeowners insurance in order to become home owners. Multifamily housing owners and managers must be able to obtain property and hazard insurance in order to obtain financing and manage the risks of their operations. These examples show how different sectors of the housing economy interact, and how the exclusion of one sector of the housing economy from the Act's coverage would pose a barrier to equal opportunity in housing. In its fair housing investigations, HUD has encountered housing providers who will not rent to individuals with disabilities because of insurance-related concerns. HUD is also aware that multifamily housing providers face barriers obtaining insurance when they attempt to lease to low-income families, including people of color and individuals with disabilities who use voucher programs to pay rent. Because of the pivotal role insurance plays in all types of housing, an exemption or safe harbor would undermine and be contrary to the Act's broad purposes.

Even if HUD had authority to exempt insurance categorically, HUD finds that such an exemption for a single industry would neither be workable nor consistent with the purpose of the Act. HUD makes this determination for the reasons it stated in its 2016 Supplemental Notice regarding this subject, some of which is reiterated here, as well as for the following additional reasons. Congress has stated that the Act is intended to provide for fair housing throughout the United States, and the Supreme Court has recognized the Act's broad remedial purpose. The Act's prohibitions on discrimination in housing are intended to eliminate segregated living patterns and move the nation toward a more integrated society. Among other things, the Act requires HUD to affirmatively further fair housing in all of its housing-related programs and activities, one of which is the administration and enforcement of the Act. HUD finds that wholesale exemptions for insurance practices would contravene the text and purposes of the Act, and, as explained further below, would also likely be overbroad in most if not all instances, as such an exemption would allow some practices with unjustified discriminatory effects to go uncorrected. HUD also finds that wholesale exemptions also would be likely to immunize potential intentional discrimination in the insurance market, because as the court in Inclusive Communities stated, “disparate-impact liability under the [Fair Housing Act] also plays a role in uncovering discriminatory intent.” As the Court found in that case, the availability of disparate-impact claims, “permits plaintiffs to counteract unconscious prejudices and disguised animus that escape easy classification as disparate treatment.”

HUD notes that multiple court decisions have long found discriminatory effects claims against insurance practices to be actionable. And even if the commenters were correct that the industry's practices generally will not give rise to discriminatory effects liability, that fact does not provide a sufficient justification for exempting the entire industry from liability in all circumstances, even where there is a practice with an unjustified discriminatory effect. Especially in light of the broad remedial purposes of the Act, HUD finds that the final rule strikes the appropriate balance for insurance industry practices. Furthermore, HUD notes that some types of discrimination are more difficult than others to prove, and this is particularly true when individuals who are denied a service or quoted a particular price for a service in a residential real estate-related transaction would typically have no way of knowing the specific reasons for a denial or pricing decision. Simply because claims are difficult to prove and may not end up in litigation does not mean that the underlying conduct can or should be exempted from regulation in all instances.

HUD finds that the concerns raised by the insurance industry do not outweigh these fundamental considerations. This rule sets out a framework by which liability under the Act may be determined; liability arises only for those insurance practices that actually or predictably result in a discriminatory effect and lack a legally sufficient justification. The framework takes into account any defendant's legitimate interest in the challenged practice—including an insurance defendant. As discussed below, HUD finds that any conflict with a specific state insurance law can and should be addressed on a case-by-case basis in the context of that state law.

In sum, the case-by-case approach set out in this final rule appropriately weighs the relevant factors, which include HUD's obligation to enforce the Act, the diversity of potential discriminatory effects claims, the variety of insurer business practices, and the differing insurance laws of the states, as they currently exist or may exist in the future. Given these considerations, HUD believes that it would be impossible for the agency to define the scope of insurance practices covered by an exemption with enough precision to avoid case-by-case disputes over its application. Accordingly, HUD has determined that categorical exemptions or safe harbors for insurance practices are unworkable and inconsistent with HUD's statutory mandate.

Issue: Commenters stated that if HUD does not provide an exemption for insurance practices, insurers would be forced to evaluate whether their practices lead to segregation and to learn what statistical disparities are permissible.

HUD Response: HUD disagrees. Any obligation to evaluate practices comes from the language of the Act itself, not this final rule. As explained above, this final rule does not impose any new liability upon insurers, so it will not require insurers to start new reviews of their practices. Any such obligation to review their practices arose long before the 2013 Rule was promulgated and originates from the statutory language. Judicial precedent applying disparate impact analysis to insurance companies long predates the 2013 Rule, let alone this rule. Any costs entities may now choose to incur will not be due to any new requirement, and in any case will simply be the ordinary costs of complying with any preexisting statute, administrative practice, and case law governing nondiscrimination in housing and housing-related practices. In any event, evaluating and re-evaluating current practices are not unreasonably burdensome activities for a business or industry to undertake. As explained elsewhere, many other industries, such as lending, engage in risk-based practices and show that it is possible to consistently evaluate and re-evaluate their policies and practices to endeavor to avoid those that may cause unjustified discriminatory effects. Yet those industries have not suffered the dire consequences that the insurance industry claims it will suffer. HUD does not believe the insurance industry stands on different footing from other industries in that respect such as to warrant differential treatment.

Issue: Commenters, citing NAACP v. Am. Family Mut. Ins. Co., asked HUD to exempt all homeowners insurance practices from liability for unjustified discriminatory effects, stating that the Act covers only insurance practices that make housing unavailable, thus effectively precluding homeownership. Homeowners insurance practices, they stated, do not make housing unavailable. In addition, citing Southend Neighborhood Improvement Assoc. v. St. Clair, commenters stated that section 804(b)'s prohibition against discrimination in the provision of services in connection with the sale or rental of a dwelling applies only to services generally provided by governmental units, such as police and fire protection or garbage collection, not insurance.

HUD Response: HUD declines to exempt homeowners insurance from liability for the reasons stated previously and explained more fully below. Neither NAACP nor Southend Neighborhood Improvement Ass'n support such an exemption. The commenters are incorrect in stating that insurance practices cannot make housing unavailable or that the Act only covers insurance practices that make housing unavailable. A discriminatory practice that precludes a person from obtaining homeowners or renters insurance may indeed make housing unavailable to that person, as insurance is usually required as a condition for obtaining a mortgage or a lease. Moreover, while section 804(a) prohibits discrimination that “make[s] unavailable” a dwelling, other provisions in the Act may prohibit insurance practices, including pricing, regardless of whether they make housing unavailable. For example, section 805(a) prohibits discrimination in the “terms or conditions” of “residential real estate-related transactions,” and section 804(b) prohibits discrimination in the “terms, conditions or privileges of sale or rental of a dwelling or in the provision of services . . . in connection therewith.” Indeed, since 1989, HUD's fair housing regulations have specifically prohibited “[r]efusing to provide . . . property or hazard insurance for dwellings or providing such . . . insurance differently” because of a protected characteristic. Courts have applied the Act's provisions to various insurance practices, including insurance pricing, marketing and claims processing, irrespective of whether the discriminatory conduct occurred when the unit became available or in conjunction with or subsequent to the acquisition of a dwelling.

In addition, HUD finds that the commenters have misconstrued the referenced cases. HUD notes, for example, that NAACP did not hold that the Act only prohibits insurance practices that effectively preclude homeownership; rather, the court, in considering whether the Act prohibited intentional insurance redlining practices, concluded that it did, and affirmed HUD regulations which “include, among the conduct prohibited by section 3604: `Refusing to provide . . . property or hazard insurance for dwellings or providing such . . . insurance differently because of race.' ” In that case, the plaintiff brought suit under both section 804(a), asserting that the insurer made housing unavailable, and section 804(b), asserting that the insurer discriminated in the provision of services in connection with the sale or rental of a dwelling. The Seventh Circuit, in discussing the viability of plaintiff's claims, stated that § 804 “applies to discriminatory denials of insurance, and discriminatory pricing, that effectively preclude ownership of housing because of the race of the applicant.” The court could not read section 804(b) as requiring a showing that housing was otherwise made unavailable as that language is not present in section 804(b); rather it is in section 804(a). Accordingly, the court's quote cannot be read as applying to the section 804(b) claim especially because it was talking about the plaintiff's claims generally, including its section 804(a) claim, which has the “make unavailable” language. Thus, NAACP cannot be fairly read to hold that the Act only applies when insurance practices make housing unavailable.

Furthermore in NAACP, the Seventh Circuit also clarified its earlier statement regarding governmental services in Southend Neighborhood Improvement Ass'n. In NAACP, the court stated, “[w]e once suggested in passing, [in Southend ] that `service' in section 3604 means `services generally provided by governmental units,' but the subject was not before us—and the suggestion that section [804] is limited to governments is hard to reconcile with another plain-statement principle requiring Congress to be especially clear if it wants to regulate the conduct of state and local governments. . .So it is hard to understand section [804] as restricted to garbage collection and like services.”

Issue: Commenters stated that an exemption for insurance practices is warranted because the judicial and legislative branches have not specifically authorized HUD to become involved in insurance.

HUD Response: Congress authorized HUD to interpret and enforce the Act, and as discussed above, provided no exemption for insurance practices. As also discussed above, courts have routinely applied the Act to insurance practices and have found that, as with other housing-related practices, insurers may be liable for practices that create discriminatory effects under the Act. In promulgating this final rule, HUD is exercising the authority Congress gave it. Any liability originates from the Act itself, not HUD or the rule.

Fundamental Nature of Insurance

Issue: Commenters requested an exemption for insurance practices because of the fundamental nature of the industry, alleging that the proposed rule would fundamentally and problematically alter insurance practices. Commenters said that the foundation of the business of insurance is the ability to classify insurance policyholders by risk and that insurers make decisions based on actuarial and business principles that group policyholders for the purpose of treating those with similar risk profiles similarly. Commenters stated that the industry is predicated on setting rates and making underwriting decisions based on relevant, mathematical, and objective risk factors that accurately predict loss. Commenters said that risk-based pricing has been a bedrock principle of state insurance regulation for more than 150 years, acting as a primary tool for ensuring rates are adequate, not excessive, not unfairly discriminatory, accurately predictive of risk, and protective of the solvency of insurers. Commenters stated that the insurance market functions best when each insured pays a rate that accurately reflects the cost of providing insurance to similarly-situated policy holders. Commenters stated that although professional underwriters routinely avoid or exclude risks for which they lack expertise, underwriting judgment, or actuarial data, they still are required to consider similar factors bearing on risk of loss and do not consider protected traits.

Commenters noted that risk-based pricing is the primary tool to ensure that rates are not unfairly discriminatory, as defined by state insurance codes. Commenters stated that in the context of insurance, unfair discrimination means treating similar risks in a dissimilar manner, which is different from discrimination under the Act. They stated that a rate is unfairly discriminatory if the premium differences do not correspond to expected losses and average expenses.

Commenters stated that the proposed rule would force insurers to eliminate actuarially sound risk-based practices, which is central to the effective determination of insurance premiums, in favor of substitutes that are less effective at furthering an insurer's legitimate, nondiscriminatory interests. A commenter stated that the proposed rule would penalize insurers for relying on sound risk factors that disproportionately affect a protected class, because they would be held liable for disparities they did not create. A commenter stated that the rule will require uniform rates, regardless of risk. Commenters disagreed with the proposed framework's case-by-case analysis. For example, commenters stated that insurers implement polices accounting for risk factors through actuarially sound methodologies, and that it would be impossible for a plaintiff to identify a less discriminatory alternative because any alternative would necessarily correspond to a different risk than the factor at issue, identified through actuarially sound methodology. As a result, if the plaintiff's alternative was adopted, the risk challenged in the lawsuit would no longer be reflected in the price of insurance, resulting in overcharging low-risk customers and likely driving them from the markets.

Other commenters disagreed, stating that the proposed rule appropriately applies to insurance. A commenter stated that application of the 2013 Rule and 2016 Supplement to insurance is consistent with sound actuarial practices because it accommodates underwriting decisions that satisfy the shifting burden framework. Commenters explained that ratemaking, though largely actuarially based, can incorporate elements of non-actuarially based subjective judgments. Commenters cited ratemaking, price optimization, and credit scoring as examples of insurance practices that are not entirely risk-based. Commenters further noted that consideration of these non-purely risk-based factors had not led to the demise of the industry. A commenter indicated that over the past few decades, the insurance industry has removed barriers that restrict homeowners insurers from writing policies in communities of color and, in response to disparate-impact challenges, some insurers have refined underwriting and pricing systems to eliminate arbitrary barriers to the availability of adequate homeowners coverage, resulting in business growth. Commenters concluded that subjecting insurers to disparate impact liability does not “threaten the fundamental nature of the insurance industry.” Commenters noted that other risk-based industries, such as mortgage lending, are subject to liability for unjustified discriminatory effects under the Act and have not had to forego risk-based analysis to avoid liability under the Act.

HUD Response: HUD disagrees that the fundamental nature of insurance warrants the exemptions requested by some commenters, whose comments were premised upon the faulty assumption that the proposed rule generally prohibits risk-based practices. It does not. This final rule does not declare any activity per se unlawful. It merely provides a framework for determining if a particular policy or practice causes an unjustified and unlawful discriminatory effect. HUD recognizes that risk-based decision making is an important aspect of sound insurance practices, and nothing in this final rule prohibits insurers from making decisions that are in fact risk-based. Under the framework established by this rule, practices that actually are risk-based, and for which no less discriminatory alternative exists, will not give rise to discriminatory effects liability. The rule simply requires that if an insurer's practices are having a discriminatory effect and “an adjustment . . . can still be made that will allow both [parties'] interests to be satisfied,” the insurer must make that change.

Risk-based decision making is not unique to insurance, and discriminatory effects liability has proven workable in other contexts involving complex risk-based decisions, such as mortgage lending, without the need for exemptions or safe harbors. Indeed, all businesses covered by the Act make risk-based decisions. For example, landlords assess risk when they select tenants, set rental rates, and decide whether to require deposits. The Act requires that such risk-based determinations not be based on protected characteristics, in whole or in part. Moreover, some states specifically provide for discriminatory effects liability against insurers under state laws, further undermining the claim that providing for such liability as a matter of federal law threatens the fundamental nature of the industry.

Unfortunately, the history of discrimination in the homeowners insurance industry is long and well documented, beginning with insurers overtly relying on race to deny insurance to persons of color and evolving into more covert forms of discrimination. For example, minorities were denied access to insurance through property-location and property-age restrictions, even when data demonstrated that such restrictions were not justified by risk of loss. This history of discrimination led to persons of color being unjustifiably denied insurance policies or paying higher premiums. As described more fully in other responses, HUD believes that discriminatory effects liability continues to play an important role in preventing unjustifiable discrimination, including in the insurance industry.

Furthermore, HUD's long experience in administering the Act counsels that discriminatory effects liability does not threaten the fundamental nature of the insurance industry. Putting aside the length of time insurers have been subject to discriminatory effects liability under the statute itself, the industry has been subject to the 2013 Rule for ten years and the calamitous results commenters claimed would come to pass have not occurred. HUD's position that discriminatory effects liability applies to insurance dates back more than three decades, as does the industry's concern that such liability makes it “near impossible for an insurer to successfully defend himself.” HUD has maintained for decades that remedying discrimination in insurance, including in cases involving discriminatory effects claims, requires examination of each allegedly discriminatory insurance practice on a case-by-case basis, and HUD sees no reason to deviate now from this longstanding approach.

Based on its experience in administering and enforcing the Fair Housing Act, HUD believes that a broad exemption would immunize a host of potentially discriminatory insurance practices that do not involve actuarial or risk-based calculations, such as marketing, claims processing, and payment. In addition, a discriminatory effects claim can challenge an insurer's underwriting policies as “not purely risk-based” without infringing on the insurer's “right to evaluate homeowners insurance risks fairly and objectively.” For example, plaintiffs have challenged insurer policies that deny insurance to landlords because they rent to Section 8 voucher holders. Even practices such as ratemaking that are largely actuarially-based can incorporate an element of non-actuarially-based subjective judgment or discretion under state law. Indeed, many of the state statutes referenced by commenters that mandate that rates be reasonable, not excessive, not inadequate, or unfairly discriminatory, permit insurers, in the very same section of the insurance code, to rely on “judgment factors” in ratemaking. The example of price optimization practices, which some states have started regulating, illustrates how non-actuarial factors, such as price elasticity of market demand, can impact insurance pricing in a manner similar to the pricing of products in non-actuarial industries. The term “price optimization” can refer to “the process of maximizing or minimizing a business metric using sophisticated tools and models to quantify business considerations,” such as “marketing goals, profitability and policyholder retention.” The term “price elasticity of demand” refers to “the rate of response of quantity demanded due to a price change. Price elasticity is used to see how sensitive the demand for a good is to a price change.” Therefore, by using these practices, insurers are already using factors unrelated to risk to help determine price. Relying on factors unrelated to risk, therefore, has not doomed their business model.

HUD likewise declines to craft a safe harbor for any specific risk-based factor because it would be overbroad, foreclosing claims where the plaintiff could prove the existence of a less discriminatory alternative, such as an alternative risk-based practice.

For HUD to select a few factors for per se exemption as a matter of law based on commenters' bare assertions about their actuarial relevance, without data and without a full survey of all factors utilized by the homeowners insurance industry, would also be arbitrary. Even if such data were available and a full survey performed, safe harbors for specific factors would still be overbroad because the actuarial relevance of a given factor can vary by context. In addition, while use of a particular risk factor may be generally correlated with probability of loss, the ways in which an insurer uses that factor may not be. Furthermore, the actuarial relevance of any given factor may change over time as societal behaviors evolve, new technologies develop, and analytical capabilities improve.

The Act's broad remedial purpose is “to provide . . . for fair housing throughout the United States.” Thus, the Act plays a “continuing role in moving the Nation toward a more integrated society.” Ensuring that members of all protected classes can access insurance free from discrimination is necessary to achieve the Act's objective because obtaining a mortgage for housing typically requires obtaining insurance. Likewise, obtaining insurance may be a precondition to securing a home in the rental market. Insurance is also critical to maintaining housing because fire, storms, theft, and other perils frequently result in property damage or loss that would be too costly to repair or replace without insurance coverage.

In light of the long, documented history of discrimination in the homeowners insurance industry, including the use of “risk factors” by insurers and regulators that were subsequently banned as discriminatory and the non-actuarial or hybrid nature of many insurance practices, HUD considers it inappropriate to craft any exemptions or safe harbors for insurance practices. HUD's longstanding case-by-case approach can adequately address any concerns and better serves the Act's broad remedial purpose and HUD's statutory obligation to affirmatively further fair housing, including by supporting fair housing efforts undertaken by states.

Issue: Commenters opposed the rule or requested an exemption because they believe the rule would force insurers to consider protected traits that are prohibited in the rating and underwriting process and are not risk predictive, contrary to Inclusive Communities' caution against injecting race into housing decisions. Commenters wrote that insurance works best when it is blind to protected traits, as they have no relationship to ratemaking or underwriting and that state insurance law prohibits them from using such data to make decisions concerning eligibility, underwriting, and pricing. Commenters also stated that the rule will require insurers to charge different rates for members of different protected classes but similar risk profiles, violating state insurance laws and regulations and compromising insurers' ability to set fair, accurate, and non-discriminatory rates and reliably predict the probable financial consequences of risk. Commenters stated that an insurer could be liable for considering a protected trait or not considering the trait.

HUD Response: HUD disagrees that this final rule will force insurers to consider protected traits of individuals in the rating and underwriting process. Instead, to ensure compliance, a regulated entity may wish to examine whether a facially neutral policy or practice causes an unjustified discriminatory effect, as defined by the regulation. This is no different from the analysis that any other entity regulated by the Fair Housing Act, such as mortgage lenders and housing providers, might want to perform to ensure compliance. Inclusive Communities rejected the argument that such an analysis would raise equal-protection concerns, reasoning that “awareness of race” can help “local housing authorities [that] choose to foster diversity and combat racial isolation with race-neutral tools.” Such awareness of the impact of facially neutral actions can also benefit other housing providers and entities covered by the Act, including insurers, to achieve the goals that many commenters stated they share, i.e., achieving a more equitable and just society. This sort of awareness of race (and other protected classes), combined with an understanding of how its own policies, practices, and assessment tools impact those protected classes, can inform the covered entity on whether its approach actually or predictably results in a discriminatory effect. HUD notes that awareness of protected traits and the impact of policies based on protected traits is different from considering or making decisions based upon a protected trait, which would constitute discriminatory treatment. Commenters pointed to no state law, and HUD knows of no state law, that prohibits insurers from examining their own underwriting factors and practices to determine whether these factors and practices unjustifiably cause a disparate impact on protected classes or otherwise serve as a proxy for race. This kind of self-examination is encouraged, generally, by this final rule, is consistent with Inclusive Communities and the Act, and is intended not to lead to liability under the Act but rather to protect entities from liability. Indeed, lenders and others covered by the Act regularly engage in such self-examination without threat to their business models. In sum, the industry has been subject to the 2013 Rule for ten years, and iterations of the same burden-shifting framework as imposed by courts for even longer, and none of these dire outcomes predicted by the industry have come to pass.

Issue: Commenters stated that prohibiting risk-based pricing and underwriting, and forcing insurers to consider protected traits, would lead to negative consequences. Commenters stated that the proposed rule could lead to serious and damaging unintended consequences for the industry including, interfering with underwriting; destabilizing insurance coverage; threatening insurer solvency; distorting the market; collapsing the industry; and increasing insurance costs and premium rates, having a negative impact on policyholders and small businesses. As another example, commenters stated that the inability to rate risks will make it prohibitively expensive to insure high-risk properties so insurers will withdraw specific lines of business or insure only low-risk properties. Commenters stated, citing to NAACP, that charging the same rates to individuals posing different levels of risk results in lower-risk individuals subsidizing higher risks, eliminating incentives for insureds to mitigate risk, forcing low-risk consumers out of the market and diminishing insurers' ability to broadly spread risk.

HUD Response: HUD disagrees with the commenters' views on the final rule's impact on the fundamental nature of insurance and that such negative consequences will come to pass. Each example is premised upon the faulty assumption that the rule prohibits risk-based practices or would require insurers to use protected traits. As explained in further detail above, it does not. The rule merely provides a framework for determining if a particular policy or practice causes an unlawful discriminatory effect. Furthermore, as noted above, insurers have been subject to discriminatory effects liability since well before the 2013 Rule and have been subject to the 2013 Rule for ten years, yet to HUD's knowledge the commenters' fears have not come to pass. Certainly, no commenter has provided any evidence that such fears have materialized.

Whether Inclusive Communities Supports an Insurance Exemption

Issue: Commenters cited Inclusive Communities in support of their request for an exemption for risk-based insurance practices. Commenters stated that applying the rule to insurance would run afoul of the limitations on disparate impact liability articulated in Inclusive Communities, and affect their ability to accurately price for risk, making risk assessment more expensive, penalizing consumers, and adversely impacting the insurance market. Some commenters, citing Inclusive Communities' discussion of “legitimate business practices,” asserted that risk-based insurance practices are examples of legitimate business practices and merit an exemption. Commenters stated that restricting insurers' use of objective risk-based factors would run afoul of Inclusive Communities because it would undermine the Act's purpose and the free-market system by making insurers fearful of liability, restrict innovation, and hold insurers liable for disparities they did not create, irreparably distorting the market.

Other commenters opposed an exemption for insurers, with a commenter specifically noting that Inclusive Communities did not discuss exemptions from liability. One commenter noted in Nat'l Fair Hou. All. v. Travelers Indemnity Co., the court rejected defendants' argument that Inclusive Communities introduced new standards such that insurers could not be held liable, stating that the refusal to provide insurance to Section 8 voucher holders remained the “type of clear, non-speculative, connection . . . that Inclusive Communities requires to make out a prima facie claim of disparate impact.”

HUD Response: HUD finds no support in Inclusive Communities for exempting the insurance industry from discriminatory effects liability. As discussed above, Inclusive Communities did not introduce any new limitations to discriminatory effects law, did not address the application of the 2013 Rule or disparate impact principles to risk-based homeowners insurance practices, and did not discuss or suggest exemptions to liability for insurers or anyone else. Inclusive Communities discusses “business necessity,” and “legitimate needs” in the context of the Title VII disparate impact framework, which, like this rule, provides that a practice that is deemed a “business necessity” may still violate the statute if the plaintiff proves there is a less discriminatory alternative. Rather than support an exemption for risk-based insurance practices, this language supports the framework of this final rule. The Court in Inclusive Communities also stated that governmental entities “must not be prevented from achieving legitimate objectives.” This requirement is consistent with the final rule which, at the second step allows the defendant to show that a challenged practice serves a substantial, legitimate, non-discriminatory interest, so as to defeat a disparate impact claim unless the plaintiff can prove there is a less discriminatory alternative that serves that substantial, legitimate, nondiscriminatory interest.

Issue: Commenters stated that because the facts in Inclusive Communities involve decisions on the location of housing, which are distinguishable from the facts and decisions in insurance cases, the principles of Inclusive Communities are inapplicable to the insurance industry. This distinction, they said, supports an exemption for insurance.

HUD Response: HUD agrees that Inclusive Communities had different facts than a case involving insurance. That does not mean that Inclusive Communities supports an exemption or safe harbor for insurance. Inclusive Communities did not limit the use of discriminatory effects claims to any particular industry and provides no support for exempting insurance practices. The Court's holding that discriminatory effects claims are cognizable under the Act applies to all such claims under the Act, and does not exclude practices particular to any industry, including insurance. HUD notes that the potential application of disparate-impact analysis to the insurance industry long predated Inclusive Communities, which generally reaffirmed disparate-impact doctrine.

Whether Other Supreme Court Precedent Supports an Exemption

Issue: Commenters stated that Wards Cove and Watson require an exemption for insurance because they set a higher burden of proof for plaintiffs than the proposed rule does.

HUD Response: HUD disagrees with the commenters. Neither Wards Cove nor Watson provide a basis for an exemption for insurance practices. Both cases, which involve Title VII claims, were decided prior to the Supreme Court's controlling precedent in Inclusive Communities, with which the final rule is consistent. And as explained more fully below, neither case necessitates a revision to plaintiff's burden of proof in Fair Housing Act cases. Simply stated, they provide no basis to exempt insurance practices.

Whether Claims Against Insurers Will Fail as a Matter of Law

Issue: Commenters stated that insurance practices should be exempt because challenges to risk-based pricing and underwriting will fail as a matter of law under Inclusive Communities and Graoch. They stated that insurance claims will fail as a matter of law because Inclusive Communities mandates the removal only of “artificial, arbitrary, or unnecessary barriers” and risk-based pricing does not create such barriers and because plaintiffs would be unable to identify less-discriminatory practices that will allow the insurer to pursue their valid interest. According to the commenters, this is because it is grounded in mathematics, is objective and fair, and advances substantial, legitimate, nondiscriminatory interests. Other commenters stated that making sure that insurance rates accurately reflect the risk of future loss is a valid interest and that Inclusive Communities requires that businesses have “leeway to state and explain the valid interest served by their policies.” In addition, commenters said that the Graoch court held that categorical bars are justified when plaintiffs have no chance of success, a holding that commenters argued the proposed rule ignores.

Commenters further stated that all insurance claims will fail as a matter of law because there can never be a robust causal link between legitimate risk factors and any disparate impact. According to them, risk-based factors do not consider protected characteristics, and they are mandated or approved by state law, limiting insurer discretion. These commenters stated that any disparate impact caused by socioeconomic factors is beyond the control of insurers. Moreover, they stated that because state laws limit insurer discretion, these laws make it impossible to ascribe any discriminatory effects in underwriting and pricing to an insurer's own choices.

A commenter suggested that if HUD does not exempt or provide a defense for insurers, HUD should state in the final rule that disparate impact claims against risk-based pricing and underwriting practices cannot succeed. Commenters also asked HUD to commit not to bring disparate-impact challenges to risk-based insurance practices.

HUD Response: HUD disagrees with the commenters who claimed that lawsuits against insurers based on a discriminatory effects theory will necessarily fail as a matter of law and that therefore insurers are entitled to an exemption. As discussed in detail above, courts have found that insurers are subject to discriminatory effects liability under the Act. HUD also declines to commit not to bring discriminatory effects challenges against insurers or to specify that any claims based on insurance practices will necessarily fail. As discussed at length, insurance practices may be subject to disparate impact liability and insurers may be proper defendants in lawsuits alleging disparate impact under the Act. Indeed, the Act requires HUD to file charges of discrimination if reasonable cause exists to believe discrimination occurred.

Graoch provides no basis for such an exemption. First, the Graoch court stated that “we cannot create categorical exemptions from [the Act] without a statutory basis” and “[n]othing in the text of the [Act] instructs us to create practice-specific exceptions. Absent such instruction, we lack the authority to evaluate the pros and cons of allowing disparate-impact claims challenging a particular housing practice and to prohibit claims that we believe to be unwise as a matter of social policy.” While the Graoch court said that “categorical bars are justified when . . . plaintiffs have no chance of success,” it did not find such a situation and in fact noted the possibility of success on a claim against a landlord seeking to withdraw from a Section 8 program. It made no finding that challenges against insurance practices—which were not the subject of the lawsuit—were impossible under the Act. To the extent that Graoch is relevant, it establishes a high bar—the literal impossibility of making out a particular type of claim—that would have to be established before a categorical bar would be appropriate. And in HUD's belief, it is, in fact, possible for a claim against an insurer to succeed, as demonstrated by several court opinions, so the standard set out by Graoch is not met.

Some comments are premised on the faulty assumption that Inclusive Communities introduced different standards for discriminatory effects claims. As explained above, Inclusive Communities described and endorsed the same disparate impact framework that this rule sets out. In that case, the Supreme Court explained, that policies and practices that are artificial, arbitrary, and unnecessary are invalid under the Act when the longstanding disparate impact elements as set forth in this rule are satisfied. However, the Court did not require plaintiffs to show that a policy or practice is artificial, arbitrary, and unnecessary in addition to proving an unjustified discriminatory effect. Rather, the Court, in quoting “artificial, arbitrary, and unnecessary” from the decades-old case Griggs, was describing the types of policies that will fail under the rule's traditional shifting burden framework, which is consistent with this final rule. In other words, if a practice with a discriminatory effect is not necessary to achieve a substantial and legitimate interest, or when an alternative, less discriminatory practice exists, the challenged practice is invalid under the Act because it is artificial, arbitrary, or unnecessary. Insurance practices, like other practices related to housing, may sometimes create artificial, arbitrary, and unnecessary barriers. Further, as part of the disparate impact framework set forth in this rule, insurers, like all defendants, are provided the opportunity to show a valid interest supporting any practice challenged under the Act. Therefore, a specific exemption for insurers is unwarranted.

HUD finds that claims against insurers will not fail categorically as a matter of law. HUD believes, contrary to commenters' assertions, it is possible for plaintiffs to establish a causal connection between an insurance practice and a discriminatory effect. HUD also believes that it is possible for plaintiffs to prove a less discriminatory alternative. HUD notes that the fact that risk-based pricing does not facially consider protected characteristics provides no support for the contention that plaintiffs cannot—or should be precluded from the opportunity to—prove that a particular policy that defendants claim is risk-based causes an unjustified discriminatory effect. A violation of the Act based on a discriminatory effects claim requires proof of an unjustified discriminatory effect because of a trait protected by the Act, not proof of intentional discrimination. The fact that state laws mandate that rates be actuarially sound, or risk-based, does not necessarily negate causation because insurers may not in fact be using risk factors that are actuarially sound or the least discriminatory set of risk factors that would achieve that end. Specifically, an actuarially sound practice may nonetheless cause an unjustified discriminatory effect if a less discriminatory alternative is available that also is actuarially sound and otherwise complies with state law. As other examples, commenters referenced ratemaking, price optimization, and credit scoring as examples of largely actuarially based practices that can incorporate elements of non-actuarially based subjective judgment or discretion, and thus cause an unjustified discriminatory effect. HUD finds this comment persuasive. HUD acknowledges that there may be scenarios where plaintiffs will be unable to show causation or demonstrate the existence of a less discriminatory alternative, but it is incorrect to say that all claims will fail as a matter of law. Thus, HUD declines to grant a categorical exemption on this basis.

State Regulation

Issue: Commenters stated that insurance practices should be exempted from discriminatory effects liability, with some advocating for retention of the 2020 Rule, because, according to the commenters, states are better at regulating insurance and should be the primary or sole regulators, and federal regulation creates a patchwork of rules, leading to higher costs. Commenters stated that the state regulatory system is comprehensive; protects consumers; effectively and efficiently regulates the insurance industry; ensures that premiums are actuarially justified and not excessive, inadequate, or unfairly discriminatory; and has increased affordability and availability over the past 150 years. Commenters stated that one of the primary aims of state regulation is to protect insurer solvency by ensuring that insurance providers charge premiums that adequately cover current and future claims and provide adequate surplus for capitalization, asset and reinsurance purchases and liquidity. Commenters also stated that state regulations already preclude the type of discrimination they believe the rule addresses, though others noted that unfair discrimination under insurance laws is not the same as discrimination under the Act. Commenters said that state regulators understand the unique conditions in their state affecting market and consumer needs; are structured so as to promote consistency and sufficiently flexible to promote innovation; and have always set the right regulations for local conditions. Commenters said that interfering with a system that works well will have negative effects, undermining state insurance regulations and consumer protection laws and upending the commonsense structure of state regulation. Commenters stated that federal regulation would subvert the role of state regulators and undermine the accuracy of risk-based pricing, leading to premium increases.

Commenters, citing to Cole v. State Farm Insurance Co. (Alaska 2006) and Cain v. Fortis Insurance Co. (S.D. 2005), stated that courts have recognized that state laws ensure that the insurance market functions fairly. A commenter stated that every state has effective civil and criminal insurance anti-discrimination laws, regulations, and enforcement divisions.

Other commenters, however, warned that a broad exemption for the homeowners insurance industry could go beyond underwriting practices to exclude unregulated practices like marketing, claims processing, and claims payment from disparate impact liability.

HUD Response: HUD disagrees with commenters who say that this final rule will upend the state regulatory system or create insurer insolvency. The rule recodifies the rule that has been in effect since 2013—and that itself codified jurisprudence which has included application to insurers for decades—during which time no such upending has occurred. The rule makes no change to the status quo, and so there is no basis for claims that it will upend anything. As discussed above, the rule does not prohibit risk-based pricing or modify the ability of states to regulate insurers as they have done for decades. And Congress has delegated authority to HUD to regulate under the Act.

State regulators may effectively and efficiently ensure that premiums are actuarially justified and not excessive, inadequate, or unfairly discriminatory as defined by state insurance codes. However, as commenters arguing for the exemption themselves recognize, “unfairly discriminatory” as defined by insurance codes, is related to treating similar risks differently, which is wholly distinct from housing practices that are unlawful because they discriminate because of the protected characteristics under the Act. State insurance codes generally require only that policies and practices are aimed at a legitimate objective without regard to whether that objective discriminates because of a protected characteristic or whether a less discriminatory alternative exists to achieve that objective. As an example, many state statutes mandating reasonable rates that are not excessive, inadequate, or unfairly discriminatory, permit insurers, via the very same section of the insurance code, to rely on discretionary “judgment factors” in ratemaking. These judgment factors, although permissible under the insurance code, may result in unlawful discrimination under the Act. Moreover, it is the responsibility of HUD—not of state regulators—to promulgate regulations related to compliance with the Act.

McCarran-Ferguson Act

Issue: Commenters stated that HUD should exempt all insurance practices, or at least risk-based pricing and underwriting, because imposing the rule on insurers would violate the McCarran-Ferguson Act. Commenters stated that the McCarran-Ferguson Act established the states as the primary regulator of insurance and that state insurance laws preempt federal laws, such as the Act, when (1) the federal law does not expressly relate to the business of insurance; (2) the state law is enacted for the purpose of regulating insurance; and (3) the application of federal law might “invalidate, impair, or supersede” state laws regulating insurance. Commenters stated that the Act is not expressly related to the business of insurance and therefore its application to insurers would be inconsistent with the McCarran-Ferguson Act.

Commenters, citing Humana Inc. v. Forsyth stated that the rule contravenes the McCarran-Ferguson Act because it could invalidate or conflict with risk-based insurance pricing or underwriting policies that are permitted or required under state law. Commenters stated that state insurance laws permit or require risk-based pricing and underwriting, so any claim under the Act will always be preempted. Commenters said insurers would be caught between conflicting state and federal law and forced to either comply with state approved rates based on objective risk factors permitted or required by state law or comply with the Act by considering protected traits. Commenters stated that under state laws, insurers make underwriting decisions based on actuarial risk factors, and that risk-based differences in charges could affect demographic groups differently. Commenters also stated that the majority of states require insurers to set rates based on neutral actuarial factors, requiring insurers to take risk into account to remain solvent. Commenters said that permitting the showing of a less discriminatory alternative at step three of the burden shifting framework requires insurers to adopt alternate risk-based practices that are less effective and will result in less accurate pricing, in violation of state law. Commenters stated that the rule violates McCarran-Ferguson because a federal court may be called upon to enjoin the insurer's state-approved risk-based practice in favor of an alternative that may not be equally effective at predicting loss. Commenters stated that this violates state laws prohibiting inadequate rates and unfair discrimination between individuals with comparable risk profiles and would force insurers to use factors that are prohibited in the underwriting process.

Commenters, citing Humana, stated that federal law must not be read to authorize regulations that, if applied, would “frustrate any declared state policy or interfere with a State's administrative regime” concerning insurance. Commenters further stated that the rule impermissibly interferes with the state regulatory system for various reasons. The proposed rule, they stated, would interfere with a state's administrative regime by substituting the judgment of a federal court for state regulators. They also cited Saunders II, in support of the assertion that it is improper to empower federal courts to reject rates that were reviewed and approved by state regulators under state law. Commenters stated that even if a federal court does not reject the rate, allowing such a claim to proceed in federal court would render insufficient the assurance of lawfulness that the state approval provides. Commenters noted that the Saunders II court stated that “HUD has never applied a disparate-impact analysis to insurers” and expressed doubt that it could. Commenters also cited Ojo v. Farmers Insurance Company, which found that the McCarran-Ferguson Act barred a disparate impact claim against Farmers because it would frustrate Texas's regulatory policy, which does not prohibit an insurer from using race neutral factors in credit scoring to price insurance, even if it creates a disparate impact. A commenter pointed to Dehoyos v. Allstate Corp, which stated that “a disparate impact claim goes to the heart of the risk adjustment that underlies the insurance business” to show that the proposed rule would interfere with a state's administrative regime. Commenters stated that because unfair discrimination as defined by state insurance laws is different than discrimination prohibited by the Act, the rule disrupts states' regulatory regimes.

Commenters also cited to Mutual of Omaha, stating that the proposed rule contravenes the McCarran-Ferguson Act because it allows courts—rather than states—to determine if rates are actuarially sound. Commenters stated that in Mutual of Omaha, the Seventh Circuit held that the McCarran-Ferguson Act preempted application of the Americans with Disabilities Act because it would require insurers to litigate whether the challenged insurance practices were actuarially sound, thus stepping on the toes of state regulators. Specifically, commenters stated that steps two and three of the burden shifting framework would force federal courts to second guess the actuarial soundness of state-regulated insurance. Commenters stated that under Mutual of Omaha, a case-by-case approach to whether McCarran-Ferguson preempts the Act's application is inappropriate because of the uniformity in state laws permitting or requiring the use of risk factors and because second guessing state regulators itself is improper, regardless of outcome.

Commenters stated that state anti-discrimination laws are irrelevant to whether the McCarran-Ferguson Act preempts a case under the Act, because McCarran-Ferguson asks only whether the application of federal law would invalidate, impair, or supersede state laws enacted for the purpose of regulating the business of insurance, and state antidiscrimination laws are not enacted for such purpose. Commenters said that even if a state's fair-housing law were identical to the Act and would permit a disparate-impact challenge to risk-based practices in state court, any federal litigation under the rule would still require federal courts to second-guess the actuarial soundness of insurance practices regulated by state law—- contrary to the express holding of Mutual of Omaha.

Other commenters stated that the proposed rule does not undermine the state regulation of insurance and thus presents no conflict with McCarran-Ferguson. They stated that state authority to regulate insurance does not, on its own, create a conflict with federal law; rather this is a fact-specific determination that depends on the relevant state law, the conflict claimed and other case-specific variables. Commenters stated that many states have regulations that complement disparate-impact liability under the Act and, even if they do not, that does not necessarily mean there is a conflict with state law. Commenters cited Dehoyos, Humana, and Wai to show that the need for a fact-specific inquiry depends on the relevant state law, the conflict claimed, and other case-specific variables. Commenters stated that the District of Columbia, California, and North Carolina, for example, expressly provide by statute for disparate impact claims. Commenters said that given the variation in state insurance laws, an exemption for insurers is inappropriate, and a case-by-case evaluation is the better approach.

HUD Response: HUD believes that the McCarran-Ferguson Act neither creates nor justifies a wholesale exemption for insurers from liability for policies and practices that have an unjustified discriminatory effect. Some discriminatory effects claims against insurers will be preempted under McCarran-Ferguson but others will not, depending on a host of case-specific variables, so wholesale exemptions would be overbroad. The McCarran-Ferguson Act provides that “[n]o Act of Congress shall be construed to invalidate, impair, or supersede any law enacted by any State for the purpose of regulating the business of insurance . . . unless such Act specifically relates to the business of insurance.” As interpreted by the Supreme Court in Humana, McCarran-Ferguson applies to preempt federal law only when a particular application of that law directly conflicts with a specific state insurance regulation, frustrates a declared state policy, or interferes with a State's administrative regime. That is, McCarran-Ferguson preemption is assessed on an application-by-application basis and does not operate at the wholesale level commenters sought here. Accordingly, the mere fact that a state has the authority to regulate insurance or has adopted ratemaking regulations does not on its own create the kind of conflict, frustration of purpose, or interference that triggers preemption under McCarran-Ferguson. Rather, the inquiry required by Humana depends on the relevant state law and other case-specific variables.

For example, in Dehoyos v. Allstate, the Fifth Circuit rejected a McCarran-Ferguson defense to a disparate impact claim where the insurer did not identify a specific state law that was impaired. The Fifth Circuit reasoned that the Seventh Circuit's holding in Doe v. Mutual of Omaha that McCarran-Ferguson barred a particular claim of discrimination under the Americans with Disabilities Act did not foreclose all discriminatory effects claims against insurers. Instead, the Fifth Circuit distinguished Doe, by explaining that “[i]n Doe, there was an actual state insurance law which purportedly conflicted with the application of the ADA to the particular insurance question at issue.” Thus, where no state law is impaired, McCarran-Ferguson will not require preemption of a discriminatory effects claim against an insurer.

HUD finds that whether in fact a particular policy or practice would create a conflict so as to preempt the Act is highly fact specific and depends on the particular state law and fair housing allegations in question. Accordingly, HUD has determined that a case-by-case approach is necessary and justified. McCarran-Ferguson, by its nature, requires such case-by-case analyses and contains no requirement that HUD provide categorical exemptions. McCarran-Ferguson requires a fact-intensive inquiry that will vary state by state and by claim. Even those cases in which an impermissible impairment under McCarran-Ferguson was found support the case-by-case approach herein adopted by HUD rather than the wholesale exemption sought by some commenters. For example, in Saunders v. Farmers Insurance Exchange, prior to ruling that McCarran-Ferguson barred a discriminatory effects claim under the Act, the Eighth Circuit remanded the case for further inquiry into the facts and Missouri law.

Precedent also demonstrates that, in some instances, state law may not preempt discriminatory effects claims against insurers even when an insurer points to a specific state law and alleges that it is impaired. Although the commenters provided examples of cases in which state laws were found to preempt particular discriminatory effects claims, other cases provide examples of state laws that were not. For instance, in Lumpkin v. Farmers Group (Lumpkin II), the court rejected a McCarran-Ferguson defense to a disparate impact challenge to credit scoring in insurance pricing, holding that disparate impact liability in that context did not impair the state's law mandating that “insurance rates cannot be `unfairly discriminatory.' ” In so ruling, the court held it erroneous to read a state law prohibiting “unfairly discriminatory” rates “too broadly” and rejected the insurer's argument that such state laws require that practices with an unjustified discriminatory effect must be permitted “as long as the rates are actuarially sound.” The court then cited other provisions of the state's insurance code specifically dealing with credit scoring, concluding that they too were not impaired.

The many ways in which one state's insurance laws can differ from another's, as well as the ways in which a single state's insurance laws can change over time, mean that even an exemption for specific insurance practices would be overbroad and quickly outdated. For example, variations in state insurance laws have resulted in discriminatory effects challenges to similar insurance practices surviving a McCarran-Ferguson defense in some states but not in others. Precedent also demonstrates that the insurance laws of each state can change over time in significant ways, and state insurance regulators respond to new practices as they become common and their effects become clear. Given the variation in state insurance laws across more than 50 jurisdictions and over time, HUD declines to fashion a one-size-fits-all exemption that would be overbroad, quickly outdated, and inevitably insulate insurers engaged in otherwise unlawful discriminatory practices from liability under the Act that would not be precluded by McCarran-Ferguson.

A one-size-fits-all exemption is also inappropriate because insurance practices are not governed solely by “hermetically sealed” state insurance codes, but are also governed by a range of other state laws, including state fair housing laws. Many state fair housing laws track the Act's applicability to insurance and provision of effects liability, indicating that those states do not consider disparate impact liability to conflict with the nature of insurance. Categorical exemptions or safe harbors of the types requested by some commenters would deprive all states of this federal support in addressing discriminatory insurance practices—even those states that welcome or depend on such support. This outcome would be at odds with the purpose of McCarran-Ferguson to support the autonomy and sovereignty of each individual state in the field of insurance. Connecticut's Discriminatory Housing Practices Act, for example, “provides similar (albeit broader) protection against housing discrimination as the [Act], which is [a] strong indication that application of the federal antidiscrimination law will not impair Connecticut's regulation of the insurance industry, but rather is complementary with Connecticut's overall regulatory scheme.” Similarly, a state court found that “the disparate-impact approach does not conflict with Ohio Insurance law” and thus allowed a disparate impact claim against an insurer to proceed under the state's fair housing law. In another case where the court rejected a McCarran-Ferguson defense to a discriminatory effects claim against an insurer, the court explained that it was “not persuaded that California law would allow [the challenged] practice” and therefore ” the [] Act complements California law in this regard.” Furthermore, the allocation of authority to enforce a state's protections against discrimination in insurance can impact whether McCarran-Ferguson is a viable defense to a discriminatory effects claim in a given state. The case-by-case approach thus affirms state autonomy and furthers the Act's broad remedial goals by ensuring that HUD is not hindered in fulfilling its statutory charge to support and encourage state efforts to protect fair housing rights.

Furthermore, HUD finds that comments claiming there is necessarily always a conflict with state laws in violation of the McCarran-Ferguson Act rest on the false presumption that this final rule prohibits the use of risk-based pricing or would require insurers to consider protected traits of individual insureds in making decisions. As described in greater detail above, it does not. HUD also disagrees with commenters who stated that even if a state fair housing law prohibits practices having an unjustified discriminatory effect, the rule contravenes McCarran-Ferguson. As courts have found, and HUD agrees, in such circumstances there would be no conflict between the federal and state law at issue. Step three of the burden-shifting framework, allowing plaintiffs to prove a less discriminatory alternative, also does not necessarily interfere with the state regulation of insurance. All the rule requires is that if an insurer's practices have a discriminatory effect and “an adjustment . . . can still be made that will allow both [parties'] interests to be satisfied,” the insurer must make that change. It does not require insurers to violate state laws.

HUD disagrees with Mutual of Omaha to the extent it implied that any claim requiring a court to assess the actuarial soundness of a policy and/or whether a policy or practice is consistent with state law necessarily interferes with a state administrative regime. HUD notes that in promulgating a rule of nationwide effect it is not bound to follow the decision of a single appellate court, but may reasonably conclude that the Act allows for a different result. HUD believes courts should continue to decide through a case-by-case assessment whether requiring a court to assess actuarial soundness or consistency with state law necessarily interferes with an administrative regime, as this is an underdeveloped area of the law and case law could evolve differently in the circuits.

In any event, disparate impact claims challenging insurance practices do not necessarily require courts to ascertain whether a practice complies with state law or is actuarially sound. Therefore, not all claims even implicate the reasoning of Mutual of Omaha. As the Court in PCI explained, “[w]hile some states require insurers to use risk-based pricing, other states merely permit risk-based pricing.” Accordingly, Mutual of Omaha does not necessarily preclude claims that challenge practices that rest on subjective business judgments, rather than actuarially sound principles or state law requirements because in adjudicating such claims, the court would not necessarily need to ascertain whether the insurer's practices are actuarially sound and/or consistent with state law. For example, a plaintiff may not dispute that an insurer's practice complies with state law, but rather may show that there are alternative practices that also comply with state law that do not cause a discriminatory effect. Such a claim would not require the court to evaluate whether the challenged practice complies with state law or is actuarially sound and thus would not run afoul of Mutual of Omaha. The analysis of the challenged practice instead focuses on whether or not it produces a discriminatory effect and, if the insurer states a legitimate interest justifying the practice, the plaintiff may show that there is a less discriminatory alternative that would serve defendant's substantial, legitimate, nondiscriminatory interest. While another risk-based practice could be a possible alternative at step three, it is not necessarily the only alternative. And, even if the alternative is a risk-based practice, a court may not need to assess the actuarial soundness of the alternative practice. For example, the evidence might show that an insurer opted for one of two risk-based practices which were both equally actuarially sound and compliant with state law, even though one produced a greater discriminatory effect. In such a case, the actuarial soundness of the alternative risk-based practice and its compliance with state law would already have been determined by the insurer itself. The court's analysis, therefore, would be limited to assessing the efficacy of the alternative risk-based practice in serving the insurer's business interests—the type of “unremarkable task” regularly undertaken by courts. As the court in PCI stated, Mutual of Omaha called into question the viability of some disparate impact claims. HUD agrees, but notes that while Mutual of Omaha may prevent some claims from going forward due to the McCarran Ferguson Act in Seventh Circuit district courts, it does not necessarily preclude all claims. The Act's purpose is broad and inclusive, and because Mutual of Omaha would not prevent all claims against insurers from proceeding even in its own circuit, HUD believes it is important not to foreclose meritorious claims by creating a wholesale exemption; indeed, doing so would run counter to the Act's purposes. HUD believes that case-by-case adjudication is appropriate to balance the purpose of the Act and to account for any differences that emerge in the circuits.

Finally, HUD disagrees that the rule will lead to a deluge of lawsuits. The industry has been subject to the 2013 Rule for ten years and a HUD regulation on insurance for well over 30 years and commenters have provided no evidence of an uptick in lawsuits. And as explained above, the insurance industry was subject to disparate impact liability long before the 2013 Rule, with many courts using a framework similar to the rule. Therefore, because the statute itself is the source of liability and the rule merely provides a framework for assessing the evidence, the rule cannot be the cause of any increase in lawsuits going forward.

Issue: Commenters stated that applying the rule to insurance is contrary to Congressional intent because of the McCarran-Ferguson Act. A commenter noted that the McCarran-Ferguson Act specifically exempts the Sherman, Clayton, and Federal Trade Commission Acts but not the Fair Housing Act, so under the statutory canon of construction expressio unius est exclusio alterius, Congressional intent was to exclude only the specified statutes from pre-emption. Therefore, commenters stated, an exemption for the insurance industry from the rule is justified.

HUD Response: HUD disagrees. Even assuming that at least some Fair Housing Act claims are pre-empted by McCarran-Ferguson, that does not mean that all disparate impact claims under the Act are categorically preempted. The fact that a statute is not specifically exempted from application of the McCarran-Ferguson Act simply means that the McCarran-Ferguson analysis may be applied on a case-by-case basis to claims brought under that non-exempt statute; it does not mean that McCarran-Ferguson categorically bars all such claims. The Supreme Court explained this in Humana, when it held that the McCarran Ferguson Act did not create a field exemption and that claims under Racketeer Influenced Corrupt Organizations Act (“RICO”), which like the Act is not explicitly listed as exempt from preemption, were not barred by McCarran-Ferguson. Thus, HUD has determined that the arguments put forth by commenters regarding congressional intent and expressio unius est exclusio alterius to justify an exemption from discriminatory effects liability are unpersuasive.

Issue: Commenters made various comments concerning the impact of Inclusive Communities on the McCarran Ferguson Act, including that Inclusive Communities did not invalidate McCarran-Ferguson or expand disparate impact liability to insurance; that applying the rule to insurance would bring about an undesirable “specter” of litigation in conflict with Inclusive Communities; and that the rule would increase the likelihood of a conflict between the Act and state laws regulating insurance because the rule does not conform to Inclusive Communities.

HUD Response: As discussed above, HUD believes that Inclusive Communities had no impact on the application of this final rule to insurance practices. Inclusive Communities also had no impact related to the application of the McCarran-Ferguson Act. HUD agrees that Inclusive Communities did not invalidate the McCarran-Ferguson Act, as the Court did not address insurance or McCarran-Ferguson. Nor did Inclusive Communities expand liability for unjustified discriminatory effects to insurers, who were subject to such liability long before the decision. Moreover, since there is no conflict between this rule and Inclusive Communities, as discussed above, there is no likelihood of the rule leading to litigation in conflict with that precedent or with state laws.

Issue: Commenters requested that HUD retain the provision in the 2020 Rule that recognized the McCarran-Ferguson Act. Another commenter disagreed, stating that the 2020 Rule attempted to undermine the nuanced position HUD took in 2016, when it stated that there is a circuit split as to whether McCarran-Ferguson applies at all to “subsequently enacted civil rights legislation” This commenter also stated that HUD had no authority to interpret McCarran-Ferguson in the 2020 Rule.

HUD Response: HUD declines to retain the portion of the 2020 Rule that references the McCarran-Ferguson Act because it was confusing. While the 2020 Rule did not mention the McCarran-Ferguson Act in its regulatory text, it borrowed from some of the statute's language, stating that “[n]othing in this section is intended to invalidate, impair, or supersede any law enacted by any state for the purpose of regulating the business of insurance.” HUD expressed in its 2019 Proposed Rule that this language was meant to “codify the general applicability of the `reverse preemption' provisions of the McCarran-Ferguson Act as it applies to the Fair Housing Act” and “clarify that the Fair Housing Act does not `specifically relate to the business of insurance.” In comments to that proposed rule, commenters stated that this language would create an exemption for insurance practices or preempt all such possible claims. HUD responded in the 2020 Rule that it was “neutral” as to McCarran-Ferguson's application in specific cases and pointed to cases in which the Act had been not preempted and cases in which it had been preempted. HUD repeated that it was not exempting the insurance industry and was “only clarifying that its disparate impact rule is not specifically related to the business of insurance.”

HUD believes that some commenters appear to have misread the 2020 Rule to provide a complete exemption from disparate impact liability for insurers. It is plain from reading the 2020 Rule that it neither provided an exemption nor specified that McCarran-Ferguson reverse preemption always applies to insurance practices. It simply stated that the Fair Housing Act was not intended to “invalidate, impair, or supersede any law enacted by any state for the purpose of regulating the business of insurance.” HUD has reconsidered the 2020 Rule and concludes that this provision does not clarify how the Act and the McCarran-Ferguson Act interact. Nothing in the McCarran-Ferguson Act requires HUD to make this statement and it is not HUD's responsibility to interpret the McCarran-Ferguson Act. HUD has decided this statement is unnecessary and confusing, as evidenced by commenters' misreading of the provision, and declines to retain it.

As HUD stated in 2016, the agency has adopted a case-by-case approach on McCarran-Ferguson reverse preemption, as that law requires. This approach is appropriate given the variations in jurisprudence across circuits that currently exist and may continue to evolve over time. HUD continues to believe that a case-by-case approach is appropriate. It therefore declines to incorporate the 2020 Rule's language into this final rule. HUD leaves it to the courts to decide, as they encounter individual cases, whether the McCarran-Ferguson Act preempts application of the Act in each case.

Filed-Rate Doctrine

Issue: Commenters stated that insurance practices merit an exemption because the proposed rule would violate the filed-rate doctrine, which prohibits federal courts from reexamining rates filed by a regulated entity and subject to the review and approval of a regulatory agency, as these rates are “presumed reasonable and unassailable in judicial proceedings brought by ratepayers.” Commenters stated that the Eighth Circuit decision in Saunders v. Farmers Ins. Exch., on which HUD relied in the 2016 Supplement, is inconsistent with the weight of case law holding that the filed-rate doctrine bars challenges under federal laws to rates filed with state agencies. Commenters stated that the proposed rule would upend the protections afforded the filed-rate doctrine, threatening the health, solvency, and competitiveness of the insurance market.

HUD Response: HUD disagrees that this final rule conflicts with the filed-rate doctrine. The doctrine primarily serves two purposes: preventing litigants from securing more favorable rates than their non-litigant competitors, and preserving for agencies rather than courts the role of ratemaking. HUD is not aware of any case, and no commenter cited one, in which a court has applied the filed-rate doctrine to defeat a claim under the Act, although several courts have rejected such attempts, including for discriminatory effects claims. For example, Wegoland Ltd. held that “[t]he [filed-rate] doctrine bars suits against regulated utilities grounded on the allegation that the rates charged by the utility are unreasonable.” (emphasis added). Whether a rate causes an unjustified discriminatory effect is a different issue than whether it is reasonable; discriminatory effects claims do not challenge the reasonableness of insurance rates but rather their discriminatory effects.

Multiple courts examining the filed-rate doctrine in the context of Fair Housing Act claims have found the doctrine inapplicable, noting that the Supremacy Clause, rather than the filed-rate doctrine, applies. Unlike filed-rate doctrine cases involving a conflict between federal ratemaking and a federal statute, applying the filed-rate doctrine to prioritize state ratemaking over a federal statute “would seem to stand the Supremacy Clause on its head.” Moreover, the filed-rate doctrine “does not preclude injunctive relief or prohibit the Government from seeking civil or criminal redress,” which are types of relief often obtained for violations of the Act. A filed-rate doctrine defense requires an examination of the facts in context of the laws and ratemaking structure at issue. The case-by-case approach best accommodates these variations.

As discussed above, HUD disagrees that the rule would threaten the health, solvency, and competitiveness of the market because no conflict exists with the filed-rate doctrine. Furthermore, insurers have been subject to the 2013 Rule for ten years, and disparate impact liability generally even longer, and the market effects alleged by commenters have not come to pass.

Case-by-Case Adjudication Cost for Insurers

Issue: Commenters opposed the rule's application to the insurance industry because case-by-case litigation in federal court is costly and, they contended, these costs outweigh the benefits. A commenter stated that even if a case is resolved in favor of the insurer, another suit with slightly altered facts may quickly follow. Commenters asserted that insurers would have to defend various risk factors on a regional basis, with courts possibly reaching inconsistent judgments. Commenters stated case-by-base adjudication would be a waste of judicial resources. Commenters also stated that requiring insurers to defend risk-based practices in court will make insurance less affordable. According to commenters, the costs are unjustified because rates are risk-based as required by state insurance law and have been approved by state regulators, and plaintiffs may bring claims that are hypothetical and speculative. Commenters stated that the vagueness and uncertainty of the rule threatens insurer insolvency.

Other commenters stated that the 2013 Rule, 2016 Supplement, and proposed rule appropriately state that a case-by-case analysis is the correct approach for assessing whether discriminatory effects are unjustified for all industries, including insurers. Commenters explained that to create an exemption, HUD would need to outline highly specific standardized rules, which would not be possible as actuarial practices are constantly changing and evolving.

HUD Response: As demonstrated by the relatively few cases filed against insurance companies in the decades-long history of disparate impact liability and in the ten years since the 2013 Rule was promulgated, there is no reason to believe that a continued case-by-case approach will lead to increased litigation, increased expenses in defending against claims of unjustified discriminatory effects, insurer insolvency, or increased premiums for customers. Nor did commenters provide support for these assertions.

HUD also disagrees with comments predicting that the proposed rule would create increased compliance costs. As discussed above, insurers have been subject to discriminatory effects liability since well before the 2013 Rule and have been subject to the 2013 Rule for ten years, yet commenters have not demonstrated that the 2013 Rule has led to significantly higher compliance costs. Prior to the 2013 Rule, in adjudications, HUD always used a three-step burden-shifting approach, as did many federal courts of appeals, but one federal court of appeals applied a multi-factor balancing test, other courts of appeals applied a hybrid between the two, and one court of appeals applied a different test for public and private defendants. By formalizing the three-part burden-shifting test for proving such liability under the Act, the 2013 Rule provided for consistent and predictable application of the test on a national basis. Reduced compliance costs would be expected to result because housing providers could look to a uniform standard at HUD and in the various courts across the country. It also offered clarity to persons seeking housing and persons engaged in housing transactions as to how to assess potential claims involving discriminatory effects. HUD now recodifies the burden shifting framework of the 2013 Rule, continuing the clarity, consistency, and predictability that accompanied that rule.

Issue: Commenters stated that HUD has provided no basis for the statement that it would cost as much for an insurer to demonstrate eligibility for an exemption for risk-based practices as it would to litigate the actuarial soundness of a challenged practice on a case-by-case basis in multiple jurisdictions at different points in time.

HUD Response: It appears that commenters may be referencing HUD's discussion from its 2016 Supplement of granting safe harbors for specific risk-based factors. In 2016, HUD did not discuss the cost to insurers of demonstrating eligibility for a general exemption for “risk-based practices.” Rather, HUD discussed how the arguments and evidence that insurers would need to demonstrate to show they qualified for an exemption would be the same as the arguments and evidence that they would need to meet their burden at step two. As HUD explained, if HUD were to provide a safe harbor for the use of any factor that an insurer could prove is purely risk-based, entitlement to the safe harbor would inevitably necessitate the insurer to establish it qualifies for the defense, i.e., that the use of the factor is, in fact, risk-based. If an insurance practice is provably risk-based, and a plaintiff cannot establish that a less discriminatory alternative exists, the insurer will have a legally sufficient justification under this final rule. The arguments and evidence that would be necessary to establish whether a practice qualifies for the requested exemption would effectively be the same as the arguments and evidence necessary for establishing a legally sufficient justification. Consequently, on the one hand an exemption for all provably risk-based factors would offer little added value for insurers, in terms of avoiding litigation costs. On the other hand, an exemption would foreclose potentially meritorious claims in contravention of the Act's broad remedial goals and HUD's obligation to affirmatively further fair housing.

Other Comments Related to Insurance

Issue: Commenters urged HUD to retain the 2020 Rule for numerous reasons. Commenters said that different forms of this rule have been enacted and retracted over the past few years, leading to confusion and that reinstating the 2013 Rule would be a step backwards. A commenter stated that in 2013, HUD expanded the scope of the Act to cover the insurance industry. Commenters stated that the 2020 Rule did not apply to insurance, so this rule should not create liability for homeowners insurers. Commenters noted that retracting the 2020 Rule so soon after it was promulgated was problematic for policy holders and the insurance industry, as risk-based pricing should not be subject to fleeting changes in policy.

Other commenters stated that it makes practical sense for insurers to be covered by the proposed rule given a long and well documented history of discrimination in the insurance industry. Commenters noted that the insurance industry has been subject to discriminatory effects liability for several decades. A commenter noted that in the more than twenty years since the Act was amended, courts that have considered the issue have consistently held that the Act prohibits acts of discrimination by homeowners insurers.

HUD Response: HUD declines to retain the 2020 Rule and notes that the 2020 Rule also did not exempt insurers. Commenters appear to misunderstand HUD's prior rules. Insurance practices have long been subject to liability under a disparate impact theory; that liability did not begin with the 2013 Rule and did not end with 2020 Rule, which contained no exception for such practices. Indeed, since 1989, HUD's fair housing regulations have explicitly prohibited “[r]efusing to provide . . . property or hazard insurance for dwellings or providing such . . . insurance differently” because of a protected characteristic. And the 2020 Rule explicitly stated that it “does not establish an insurance industry exemption.” Moreover, since the 2020 Rule never went into effect, there have been no changes in policy. In promulgating this final rule, HUD is recodifying a standard that has been in effect for ten years, has proven workable, and is supported by decades of caselaw both before and following its enactment.

Issue: A commenter requested that the rule include a specific defense for risk-based ratemaking, as provided in the 2020 Rule. Other commenters stated that HUD should add a substantive defense for risk-based practices whereby if a defendant can show it relied on risk-based practices at step two of the burden-shifting framework, the plaintiff should not have the opportunity to rebut the defense at step three.

HUD Response: HUD notes that the 2020 Rule did not in fact provide defenses specific to risk-based ratemaking and it declines to add such a defense now. Step two of the burden-shifting framework already provides a defense for substantial, legitimate nondiscriminatory interests, which will allow a defendant to prevail absent the plaintiff's ability to show a less discriminatory alternative. Eliminating the third step would remove the requirement for insurers to adopt the least discriminatory alternative that serves their substantial, legitimate, nondiscriminatory interest, undermining the purpose of the Act. In sum, by suggesting that the third step be eliminated, the commenter is asking for an exemption from liability for policies and practices having a discriminatory effect, which may have a legally sufficient justification, but for which a less discriminatory alternative may exist, which as explained above, HUD declines to do.

Issue: Commenters noted that in 2017, the U.S. Department of Treasury recommended that HUD reconsider whether its 2013 Rule is consistent with the McCarran-Ferguson Act, whether the disparate impact rule would have a disruptive effect on the availability of insurance, and whether the rule is reconcilable with actuarially sound principles.

HUD Response: As discussed above in greater detail, HUD has considered these issues and finds that the 2013 Rule and its framework, as adopted in this rule, is consistent with the McCarran Ferguson Act, is reconcilable with actuarially sound principles, and would not have a disruptive effect on the availability of insurance. Treasury believes that HUD, in its reconsideration of the 2013 Rule, has addressed the concerns Treasury noted in the 2017 report regarding the Rule's application to the insurance industry. Treasury no longer has the concerns expressed in that report.

Issue: Commenters stated that the 2013 Rule and 2016 supplement adequately considered the issue of application to insurance and adequately addressed the industry's concerns. Other commenters stated that the 2016 Supplemental Explanation failed to adequately explain why the filed-rate doctrine would not bar challenges to insurance rates under the Act.

HUD Response: While these comments are outside the scope of this final rule, since HUD is now re-finalizing the 2013 Rule and responding to the current comments received in response to its 2021 Notice of Proposed Rulemaking, HUD agrees with the commenters who stated that the 2013 Rule and 2016 Supplement adequately considered the 2013 Rule's application to insurance and adequately addressed the industry's concerns. And this final rule thoroughly responds to comments from the insurance industry, including those concerning the filed-rate doctrine.

Section 100.5(d): Data Collection

Issue: Commenters disagreed about whether to include the 2020 Rule's language that nothing in HUD's fair housing regulations requires or encourages the collection of data relevant to characteristics protected by the Act. Some commenters opposed including such a provision, stating that: its inclusion was unnecessary and unwise; data collection can be used to identify policies and practices that may have a discriminatory effect; and discouraging data collection would have a grave effect on discriminatory effects litigation.

Other commenters asked HUD to include such a provision, stating that otherwise, the rule's burden shifting framework necessitates data collection, which will create unnecessary costs and be especially burdensome and expensive for insurers because they do not already collect this data. Commenters stated that without the provision, the rule would expose businesses to liability risks by requiring them to obtain and store personal and potentially sensitive information about an individual's protected characteristics. Another commenter stated that requiring the collection of data would inappropriately shift the burden of proof from a plaintiff to a defendant.

Commenters also expressed concern that the only way for insurers to collect data regarding protected characteristics would be through self-reporting, which may result in incomplete or erroneous data, making compliance with the rule difficult. A commenter stated that disparate-impact challenges to risk-based practices in insurance would improperly inject race into the business of insurance by incentivizing or compelling insurers to collect and analyze data on protected characteristics to be able to mount a defense in the event of a disparate impact challenge. Commenters stated that insurers may be prohibited under state law from collecting protected trait data. Commenters added that: insurance company employees will be uncomfortable asking current or potential policy holders for information about their membership in a protected class; collecting demographic data regarding protected traits would invade customer's privacy; and asking about protected class characteristics could discourage applicants for insurance from seeking quotes.

HUD Response: HUD believes that this final rule need not include data collection language. HUD agrees that data collection can play an important role in assessing whether a policy or practice may have an unjustified discriminatory effect. HUD also agrees with the Court in Inclusive Communities, when it acknowledged that “awareness of race” can help industries “[that] choose to foster diversity and combat racial isolation with race-neutral tools.” This supports the idea that this Rule should not discourage the collection of this information. But HUD is also not requiring data collection. The purpose of this final rule is to recodify a long-recognized legal framework, not to describe how data and statistics may be collected, obtained, or used in the application of the framework.

HUD notes further that while data collection can be a means to identify practices that have or predictably will have a discriminatory effect, there are other ways of identifying such practices that do not require examining a business' own client pool. For example, businesses can look to publicly available datasets or studies related to their practices to see if their practices cause or predictably will cause a discriminatory effect. As HUD explained regarding the use of criminal records, “[a]cross the United States, African Americans and Hispanics are arrested, convicted and incarcerated at rates disproportionate to their share of the general population. Consequently, criminal records-based barriers to housing are likely to have a disproportionate impact on minority home seekers.” A business need not collect data from its own clients to ascertain that relying on criminal records in its policies or practices likely has a discriminatory effect on certain populations. In addition, independent data gathering is not necessary to defend a lawsuit alleging discriminatory effects. Plaintiffs must meet their initial burden at step one to show a disparate impact. Defendants need not present their own statistics in response to this step one evidence, but may defend in numerous other ways, including by showing that the data put forward by plaintiff is incorrect or wrongly analyzed, or by showing at step two that the challenged practice is necessary to achieve a substantial, legitimate, nondiscriminatory interest. This is true for all defending parties, including insurers, who bear no increased burden.

Moreover, concerns about how the rule would change industry practices—in particular what commenters say is the insurance industry practice of not collecting demographic data—do not square with the fact that current industry practice is based on a rule that has been in place, uninterrupted, since 2013, and based on the underlying law that has been in place for decades prior. Businesses that have not collected data over the past several decades will not be facing any change in the laws regulating their practices with HUD's recodification of the 2013 Rule.

Issue: Commenters requested that HUD clarify expectations and provide protections for lenders that collect demographic data for use in fair lending self-testing.

HUD Response: As discussed above, HUD believes that demographic data can be helpful in assessing whether a policy has an unjustified discriminatory effect. HUD notes further that lenders routinely collect data on protected characteristics as part of their Home Mortgage Disclosure Act reporting obligations. However, HUD is not requiring either collection of demographic data or self-testing. It is unclear what “protections” commenters meant for HUD to provide to lenders who collect demographic data and use that data to engage in self-testing. HUD notes that the self-testing privilege as described in 42 U.S.C. 3614–1 already applies to lenders. This self-testing privilege will not provide a lender (or any other entity) with an exemption from liability under the Act, but if a complaint is made to HUD against a lender alleging practices that have an unjustified discriminatory effect, HUD is prohibited from obtaining self-testing results covered by this self-testing privilege to investigate a lender's compliance with the Act. Of note, HUD will not absolve a lender of potential liability merely because the lender collects demographic data and does self-testing. Doing so would abdicate HUD's basic obligation to enforce the Act by ceding substantive compliance authority from HUD to private lenders.

Section 100.500: The Discriminatory Effects Rule

Section 100.500(a): Removing “Predictably” From the Definition of Discriminatory Effect

Issue: Commenters asked HUD to remove the word “predictably” from the proposed rule's definition of discriminatory effects in § 100.500(a), asserting that it violates the Act. A commenter stated that the plain language of section 804(b) does not include practices that might result in a discriminatory effect. Other commenters asserted that the “predictably results” language violates Inclusive Communities' “robust causality” requirement. According to one commenter, this “new” robust causality standard requires a plaintiff to prove that a practice already caused the discriminatory effect, not just that a practice will predictably do so. Commenters similarly suggested that Inclusive Communities' bar on claims that are based solely on statistical evidence rules out claims based on predictable or hypothetical impacts.

Other commenters wrote in favor of retaining the “predictably” language in the rule. Commenters pointed out that courts, including Inclusive Communities, have interpreted the “predictably” language in the proposed rule to contain a “robust causality” requirement, including a bar on claims that are based solely on statistical evidence of discriminatory effects. One commenter noted that the robust causality requirement that Inclusive Communities discusses is simply the 30-year-old requirement that a plaintiff, to prevail in a disparate impact challenge, must show that the disparate impact is causally related to, not merely correlated with, the identified practices of the defendant. Another commenter noted that in the very first case in which an appeals court recognized discriminatory effects liability, the 8th Circuit required the plaintiff to bear the burden of showing that defendants' conduct actually or predictably resulted in a discriminatory effect. One commenter noted that HUD in 2013 explained how the “predictably” language was supported by the plain language of the Act and case law, and HUD ignored this justification when it attempted to remove the language in the 2020 Rule.

One commenter acknowledged that “predictability” is a necessary element to assess the disparate impact of a policy and an issue that Inclusive Communities did not address. Other commenters noted multiple cases in which courts have utilized the proposed rule's predictably standard in practical, effective ways, such as in Georgia Conference of the NAACP v. City of LaGrange, and Fortune Society v. Sandcastle.

A commenter noted that caselaw and practical common-sense support that one need not wait until actual harm is inflicted before an action can be challenged. Another commenter noted that removing the predictably standard would unnecessarily increase the risk of harm to communities by taking away the ability to make claims for reasonable, foreseeable harm.

HUD Response: HUD declines to remove “predictably” from this final rule's definition of discriminatory effects. As explained in the 2013 Rule, the plain language of the Act supports the inclusion of this language. The Act defines an “aggrieved person” as anyone who, among other things, “believes that such person will be injured by a discriminatory housing practice that is about to occur. ” Furthermore, the Act explicitly authorizes HUD to take enforcement action and Administrative Law Judges (ALJs) and courts to order relief with respect to discrimination that “ is about to occur. ” In addition, courts interpreting the Act have agreed that predictable discriminatory effects may violate the Act. HUD further believes it would be contrary to HUD's duty to affirmatively further fair housing if it could not take action to prevent the harm of a predictable discriminatory effect and instead had to first allow individuals to be subjected to discrimination before any enforcement action could be taken. As explained above, the Court in Inclusive Communities did not announce a new “robust causality” requirement. Nor did it indicate any intention to exclude from liability cases that allege predictable discriminatory effects. Rather, the Court simply described the longstanding requirement that a plaintiff must establish a causal connection between the policy or practice and the discriminatory effect. Inclusive Communities explained that a plaintiff raising a “disparate-impact claim relying on a statistical disparity” must “point to a defendant's policy or policies causing that disparity.” Consistent with Inclusive Communities, this final rule requires—whether for a disparity that has already occurred or one that will occur—that the plaintiff point to a defendant's policy or policies that cause the disparity, and not rely on a statistical disparity alone.

Issue: Commenters also objected to the “predictably results” language in proposed § 100.500(a) (and the “predictably will cause” language at § 100.500(c)(1)) saying that it is inconsistent with case law under Title VII and the Age Discrimination in Employment Act (ADEA). They stated that the Title VII cases Wards Cove Packing Co. v. Atonio and Watson v. Fort Worth Bank & Trust preclude discriminatory effects claims based on policies which predictably, rather than actually, cause a discriminatory effect. Commenters stated that in Wards Cove, the Supreme Court stated that “[a] plaintiff must demonstrate that it is the application of a specific or particular . . . practice that has created the disparate impact under attack. Such a showing is an integral part of the plaintiff's prima facie case.” Commenters quoted Watson, which said that “the plaintiff must offer statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their membership in a protected group.” Similarly, a commenter cited Meacham v. Knolls Atomic Power Lab'y for the proposition that plaintiffs in cases brought under the ADEA must prove an existing disparate impact—not a future one.

HUD Response: HUD believes that the commenters' reliance on these Title VII and ADEA cases is misplaced because these cases only considered the question of whether certain policies already had had a disparate impact, not whether they would “predictably” have one in the future. Furthermore, the Act explicitly defines an aggrieved person as including “any person who believes that such person will be injured by a discriminatory housing practice that is about to occur” Finally, courts interpreting the Act have agreed that predictable discriminatory effects may violate the Act.

Section 100.500(a): Perpetuation of Segregation in the Definition of Discriminatory Effect

Issue: Commenters disagreed about the proposed rule's inclusion of perpetuation of segregation as a type of unlawful discriminatory effect. Some commenters stated that including liability for practices that perpetuate segregation is too broad and may have a chilling effect on the development of affordable housing. One commenter said that prohibiting practices that perpetuate, create, increase, or reinforce segregated housing patterns based on protected classes was a more stringent standard than Inclusive Communities announced. Another commenter stated that this language would expand liability to cover any action or any absence of action that reinforces or perpetuates segregated housing patterns, which is inconsistent with Inclusive Communities' requirement that plaintiffs demonstrate that the challenged practice is a direct cause of the disparate impact.

In contrast, other commenters stated that including the perpetuation of segregation provision is crucial to combatting segregation (including segregation based on disability and race), which is still a major problem today and can have devastating impacts on communities. Commenters said if HUD did not include this language, it would mean that HUD had adopted the view that perpetuation of segregation was not a central or relevant concern of disparate impact, that perpetuation of segregation was no longer a basis for liability under the Act, and/or that perpetuation of segregation liability would be collapsed into disparate impact liability, and only be evidence of a disparate impact claim, rather than an independent means of establishing a violation in and of itself. Commenters noted that reinstating the perpetuation of segregation language was important to eliminate the confusion that the 2020 Rule had caused through its removal, and to clarify that perpetuation of segregation is a distinct type of discriminatory effect under the Act. Commenters gave examples of activities which may unlawfully perpetuate segregation, including facially neutral zoning decisions whose real but disguised purpose is to exclude people of color, and the demolition or displacement of affordable housing leading to severely limited opportunities for people of color. Commenters said that removing the perpetuation of segregation provision would conflict with Inclusive Communities. One commenter stated that federal appellate courts have long recognized perpetuation of segregation as a distinct basis for discriminatory effects liability. Commenters stated that the 2020 Rule, which eliminated perpetuation of segregation, conflicted with HUD's duty to affirmatively further fair housing, which is a central goal of the Act.

HUD Response: HUD agrees with the latter commenters that perpetuation of segregation is prohibited by the Act and, as such, should be included in the definition of discriminatory effects in this rule. The elimination of segregation is a central goal of the Act, one that was highlighted by Inclusive Communities and has long been recognized by other courts. Inclusive Communities also recognized that practices that perpetuate segregation independently violate the Act. HUD also notes that every federal court of appeals to have addressed the issue has agreed with HUD's interpretation in the 2013 Rule. HUD finds that the rule is consistent with Inclusive Communities' causation requirement because it plainly requires that a practice “causes or will cause” a discriminatory effect. While Inclusive Communities did not directly address a claim brought under a “perpetuation of segregation” theory, it discusses disparate impact's long-standing limits, including its causation requirement, as in harmony with its aim to prohibit “perpetuating segregation.” HUD believes that eliminating the perpetuation of segregation language will cause inconsistency between HUD's rule and judicial precedent and create the mistaken impression that HUD believes that practices that perpetuate segregation are not practices which create discriminatory effects.

HUD also disagrees that the final rule would chill the development of affordable and fair housing, including in predominantly minority neighborhoods. Commenters did not provide, and HUD is not aware of, any support for the proposition that this rule would have such an effect. Instead, this rule provides a framework for plaintiffs to challenge discriminatory housing decisions. And Inclusive Communities specifically noted that HUD's discriminatory effects rule recognized that disparate impact liability does not mandate that affordable housing be located in neighborhoods with any particular characteristic. Eliminating the provision on perpetuation of segregation would also be inconsistent with HUD's duty to affirmatively further fair housing, which applies, inter alia, to HUD's program of administering, implementing, and enforcing the Fair Housing Act.

In sum, HUD declines to eliminate the provision on perpetuation of segregation because doing so would lead to uncertainty over the state of the law, the provision is consistent with Inclusive Communities and well established caselaw, and doing so would undermine one of the core goals of the Act, i.e., ending the perpetuation of segregation.

Section 100.500: Racial Quotas or Unfair Advantages to Plaintiffs

Issue: Commenters expressed concern that the proposed rule's framework would cause them to adopt quotas to avoid unlawful disparities. One commenter stated that this is because the rule does not require any causal connection between the policy and any disparity and would therefore pose the risk that financial services and businesses would adopt a quota-based approach to avoid disparities. Another commenter similarly suggested that the proposed rule does not contain a robust causality requirement, stating that Inclusive Communities emphasized a robust causality requirement to prevent housing providers and businesses from resorting to racial quotas. Another commenter asserted that to align the rule with Inclusive Communities' robust causality requirement and therefore reduce the incentive for housing providers to use racial quotas, while still maintaining the essence of the 2013 Rule, HUD should modify the final rule to say that “discrimination on a group of persons is predictable through a robust causal link by the challenged policy or practice.”

In contrast, a commenter stated that the proposed rule would not require businesses to consider race or quotas. Other commenters stated that by requiring that a plaintiff prove that the challenged practice caused or predictably will cause a disparate impact rather than imposing liability based on statistical disparities alone or general societal discrimination, this rule addresses any concerns that disparate impact liability would cause defendants to resort to quotas.

HUD Response: HUD disagrees that this final rule will incentivize quotas. While the Court expressed concern in Inclusive Communities that “without adequate safeguards at the prima facie stage,” disparate-impact liability might lead to the use of “numerical” or “racial quotas, this rule already contains these “adequate safeguards.” In particular, the rule requires plaintiffs to demonstrate that “a challenged practice caused or predictably will cause” (emphasis added) a discriminatory effect. Furthermore, it defines “a practice that has a discriminatory effect” as one where the practice “ actually or predictably results in a disparate impact” (emphasis added) or in segregation cases, where the practice “creates, increases, reinforces or perpetuates segregated housing patterns.” As explained previously in this preamble, this connection between the challenged practice and the discriminatory effect is the causality that Inclusive Communities spoke of when discussing how safeguards would prevent the use of racial quotas. And it was in the context of the Inclusive Communities district court's failure to require this connection (by finding the defendant liable based solely on discrepancies in outcomes, without requiring the plaintiff to show that a particular practice caused those outcomes) that the Fifth Circuit remanded the matter with instructions to follow the 2013 Rule, a judgment the Supreme Court ultimately affirmed. Crucially, therefore, far from invalidating the 2013 Rule for failing to require this connection, the Fifth Circuit and Supreme Court decisions both support HUD's position that that applying the 2013 Rule's framework is the correct method of ensuring that disparate impact liability does not improperly require the use of racial quotas.

Further, HUD's discussion above regarding the insurance underwriting processes explains the difference between being aware of protected traits to avoid discrimination (consistent with this final rule, Inclusive Communities, and the Act) and violating the Act by making decisions based upon a protected trait.

In addition, it is unclear how the commenter's proposed alternative language, that “discrimination on a group of persons is predictable through a robust causal link by the challenged policy or practice” would improve the rule or disincentivize quotas. On the contrary, HUD believes modifying the rule to incorporate this language would create confusion about the causal link between the policy and the effect discussed by Inclusive Communities. For example, it is unclear what “by” means in the proposed sentence, and the sentence does not make clear that the practice must cause (predictably or actually) the discriminatory effect. HUD further believes incorporating the “robust causal link” language is unnecessary and could confuse people about a heightened standard that Inclusive Communities did not create, as detailed elsewhere in this preamble.

Issue: Commenters stated that the proposed rule would create an uneven playing field in favor of plaintiffs through the requirements and burdens placed on defendants, as compared to plaintiffs. Some commenters stated that the proposed rule requires defendants to show that their policy will not cause a disparate impact on a protected group. Other commenters said the proposed rule allows plaintiffs to use hypothetical or speculative evidence, or no evidence at all, to show that a practice causes a discriminatory effect, while at the same time requiring defendants to meet their burden at step two with evidence that is not hypothetical or speculative, thus placing the entire burden of proof on defendants. A commenter said the rule allows plaintiffs to raise hypothetical or speculative impacts at step one (because of the “predictably results” language), while barring defendants from raising hypothetical or speculative defenses at step two.

On the other hand, commenters supported HUD's continuation of the 2013 Rule's framework, stating that the 2020 Rule unjustifiably favors defendants because plaintiffs must meet a preponderance of the evidence standard to prove discrimination, but defendants are only required to show that a policy advances a legitimate interest. The commenters stated that this conflicts with well-established case law placing the burden on the defendant to prove that the challenged practice is necessary to achieve one or more substantial, legitimate, nondiscriminatory interests.

HUD Response: HUD believes that the burdens and requirements in the rule are appropriately balanced, and that the concerns that the rule is tipped in favor of plaintiffs are based on misunderstandings of the rule.

First, the rule does not require a defendant to show that its policy or practice does not cause a disparate impact. In fact, this rule does not require any party to prove a negative. While a defendant may choose to present evidence that the defendant's policy does not cause a discriminatory effect to rebut the plaintiff's evidence that it does, the plaintiff has the ultimate burden of proving that a defendant's policy caused (or predictably will cause) a discriminatory effect.

Nor does this rule place a greater evidentiary burden on defendants than on plaintiffs or otherwise shift the burden of proof entirely onto defendants. Under the rule, the plaintiff must prove through evidence (not speculation) that a challenged practice caused or predictably will cause a discriminatory effect (step one). Assuming the plaintiff meets this burden, the defendant must prove through evidence (not speculation) that a challenged practice is necessary to achieve one or more of its substantial, legitimate, nondiscriminatory interests (step two). If the plaintiff fails to meet its step one burden, defendant prevails, and if the defendant fails to meet its step two burden, the plaintiff prevails. It is the plaintiff—not the defendant—who carries the burden at two of the three steps in the burden shifting framework, including the final one. As HUD said in the 2013 Rule: “Requiring the respondent or defendant to introduce evidence (instead of speculation) proving that a challenged practice is necessary to achieve one or more substantial, legitimate, nondiscriminatory interests in order to benefit from the defense to liability is not different in kind from requiring the plaintiff to introduce evidence (not speculation) proving that a challenged practice caused or will predictably cause a discriminatory effect. As discussed in this preamble, the language of the Act makes clear that it is intended to address discrimination that has occurred or is about to occur, and not hypothetical or speculative discrimination.”

Although commenters specifically called out evidence that could support a complaint concerning a “predictable” disparate impact as “hypothetical” or “speculative” under the rule, this is incorrect. In the final rule's framework, neither the plaintiffs nor defendant may rely on hypothetical or speculative evidence. All parties must rely on evidence that is sufficiently rigorous and not speculative, and there is no requirement that either side rely solely on evidence of the already existing effects of defendants' adopted policy. For example, lenders routinely assess proposed policy changes using current data to determine whether, if adopted, the policy would have a disparate impact in the future. Data analysis like this—of the effects that a policy will have, rather than the effects a policy already has had—is neither “hypothetical” nor “speculative” and could be used by either a plaintiff or a defendant to support or rebut a predictable effects claim at step one. And just as a plaintiff can rely on evidence that the defendant's policy will predictably have certain effects, a defendant can rely on evidence that a proffered less discriminatory alternative to its policy will not work.

Moreover, characterizing the impact of a “predictable effects” showing at step one as “hypothetical or speculative” is incorrect. Even if the impact has not yet occurred, this final rule still requires that plaintiffs prove that it predictably will occur. If plaintiffs show only that the discriminatory impact is “hypothetical,” or “speculative,” they will not prevail. Defendants may prove that a policy or practice with a discriminatory effect was necessary to meet a substantial, legitimate, interest. Hypothetical or speculative defenses articulated in support of a policy or practice will not be sufficient, because defendants know the actual reason for the policy or practice at issue. Allowing defendants to present different reasons than their actual reasons for implementing policies with discriminatory effects would allow pretextual reasons to justify discriminatory policies, thus defeating the important role of discriminatory effects liability in uncovering discriminatory intent, and would permit, rather than remove, arbitrary and artificial barriers to housing.

Finally, HUD agrees with commenters who noted that the burden shifting framework in this rule strikes the appropriate balance between the interests of plaintiffs and defendants, and that the 2020 Rule upset this balance. For example, it required defendants to identify only a legitimate interest rather than an interest that is also substantial and nondiscriminatory. It removed the requirement that the defendant's challenged practice be necessary to achieving that legitimate interest. Additionally, the defendants' burden was reduced from one of proof to one of production. HUD notes that these changes were neither consistent with nor justified by the text of the Act or case law interpreting it. And to the extent the Act and case law provide discretion, HUD exercises its policy judgment to maintain the 2013 Rule's burden shifting framework for the reasons stated above.

Section 100.500(a) and (c)(1): Clarifying Causation

Issue: Commenters suggested HUD provide guidance to help clarify causation in the final rule or modify the causation standard in the rule to make it more detailed or specific. Some commenters asked HUD to clarify that a challenged practice may be too remote from the alleged discriminatory effect to give rise to liability. Other commenters criticized the proposed rule for not making clear that a plaintiff must identify a specific policy or practice that caused the alleged disparate impact (as opposed to challenging a more general array of practices), with some saying that Wards Cove requires this. Other commenters suggested that the rule specify that the discriminatory impact be “significant” because Wards Cove and Inclusive Communities require it. According to the commenters, the latter's warning that race should not be used in a pervasive way or injected into every housing decision necessitates a “significant” discriminatory impact. Others suggested the rule needs to be clearer on what evidence is required to show causation and should establish statistical standards, with one commenter stating that the rule should require some threshold of showing credible, localized, statistical proof that a challenged practice has a discriminatory effect. A commenter said clarification is needed because HUD's 2016 Guidance on criminal records, by pointing to historic nationwide incarceration rates, shows that HUD has interpreted the proposed rule to allow the plaintiff to meet the initial burden using sweeping generalizations about statistics and impact with little or no showing of statistically valid discriminatory impact. The commenter further stated that courts have interpreted the initial burden under Title VII, including in Wards Cove, as much higher than the burden articulated in the rule, including requiring that the practice has an adverse impact on a specific protected class that is qualitatively different from other classes and that can be demonstrated to have a materially different impact based on statistics for the relevant geographic area.

In contrast, other commenters stated that the proposed rule already contains a sufficiently clear causation requirement. A commenter wrote that the 2013 Rule and federal jurisprudence have appropriately rejected any potential single test to define “discriminatory effect” through evaluating statistical evidence of causation, citing Mt. Holly Gardens Citizens in Action, Inc. v. Twp. of Mount Holly, Bonasera v. City of Norcross, and Langlois v. Abington Hous. Auth. and noted that further defining “discriminatory effect” (including that a disparate impact is “significant”) is inappropriate because of the wide variety of policies and practices challenged.

HUD Response: HUD believes that revising the causation requirement in this final rule is inappropriate. The final rule already requires plaintiffs to show a causal link between the challenged practice and the alleged discriminatory result. That requirement, in turn, necessitates consideration of whether a challenged practice is too remote from the alleged discriminatory effect for liability to arise under the Act.

In HUD's experience, identifying the specific practice that caused the alleged discriminatory effect will depend on the facts of a particular situation and therefore must be determined on a case-by-case basis. As has been recognized in the employment context under Title VII after Wards Cove, the elements of a decision-making process may not be capable of separation for analysis, in which case it may be appropriate to challenge the decision-making process as a whole. For example, in a reverse redlining case, there may be multiple acts or policies that together result in a discriminatory effect. Finally, in some instances, the absence of a policy may amount to a practice. And while Wards Cove limited plaintiffs' ability in an employment matter to aggregate multiple practices in showing that a practice or practices cause a disparate impact until Congress amended Title VII, Inclusive Communities did not endorse a wholesale application of Wards Cove to disparate impact cases under the Act. Indeed, the Court only cited Wards Cove for an uncontroversial and undisturbed portion of its holding, i.e., that simply pointing to racial imbalances within a company is insufficient to show that a policy caused a disparate impact. And Inclusive Communities explicitly noted when it cited Wards Cove that the “robust causality” requirement it attributed to Wards Cove did not incorporate any part of the opinion that was superseded by Title VII's statutory amendments:

A robust causality requirement ensures that `[r]acial imbalance . . . does not, without more, establish a prima facie case of disparate impact” and thus protects defendants from being held liable for racial disparities they did not create.' Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 653 (1989), superseded by statute on other grounds, 42 U.S.C. 2000e–2(k) (emphasis added).

HUD further declines to set statistical standards, including statistical thresholds, to require localized statistics, or note a “significance” requirement. HUD continues to believe, as it did in 2013, consistent with courts, that analyzing causation in these matters on a case-by-case basis is the best approach, especially given the wide variety of policies, practices, and discriminatory effects at issue in these types of cases. Courts have recognized a variety of circumstances—both under the Act and Title VII—in which using national statistics, rather than local statistics, is appropriate. HUD's 2016 Guidance recognizes this, while also noting that “state or local statistics should be presented where available and appropriate based on a housing provider's market area or other facts particular to a given case.” The Supreme Court has recognized that a case-by-case approach is appropriate in the Title VII context when it comes to statistical thresholds and requirements and levels of significance. And, as HUD noted in 2013, the decision not to codify a significance requirement is consistent with the 1994 Joint Policy Statement on Discrimination in Lending, the statutory codification of the disparate impact standard under Title VII, and the Consumer Financial Protection Bureau's interpretation of the disparate impact standard under ECOA.

Section 100.500(c)(1): When Multiple Factors Produce Discriminatory Effects

Issue: Commenters stated that the “actually or predictably results” language in step one ignores situations in which multiple factors may produce discriminatory effects.

HUD Response: This rule requires plaintiffs to prove that the challenged policy caused or predictably will cause the alleged discriminatory effect. Therefore, plaintiffs are required to show that the policy they challenge is a cause of a discriminatory effect. The rule does not require the challenged policy to be the sole factor that causes or predictably will cause the discriminatory effect. Such an approach is consistent with HUD's position that in disparate treatment cases, the Fair Housing Act is violated even if discriminatory animus was only one of the factors motivating the defendant's actions.

Section 100.500(c)(2): Proving That the Challenged Practice Is Necessary To Achieve One or More Substantial, Legitimate, Non-Discriminatory Interests

Issue: A commenter characterized the proposed rule as requiring defendants to show “hefty” evidence at step two that the policy or practice is necessary to achieve a substantial, legitimate, nondiscriminatory interest, placing an almost insurmountable burden on defendants.

HUD Response: HUD disagrees that the rule places an insurmountable or unreasonable burden on defendants. Whether a defendant's own policy or practice is necessary to achieve one or more substantial, legitimate, nondiscriminatory interests of the defendant is well within the knowledge of that defendant, who is uniquely able to meet this burden. Furthermore, the rule does not specify what evidence is necessary to meet this burden; it merely states that a legally sufficient justification must be supported “by evidence.”

As HUD explained in 2013, the requirement that a defendant prove with evidence the substantial, legitimate, nondiscriminatory interest supporting the challenged practice and the necessity of the challenged practice to achieve that interest is consistent with HUD's longstanding application of an effects framework under the Act, and is similar to the approach taken by other federal regulatory and enforcement agencies under ECOA and Title VII. This requirement is furthermore consistent with most federal courts' interpretations of the Act after Inclusive Communities. Nowhere has HUD seen this approach present an insurmountable burden on defendants, except where appropriate: when defendants do not have a legally sufficient justification.

Issue: Commenters disagreed about whether the defendant's burden at step two in § 100.500(c)(2) should be a burden of proof or production based on Inclusive Communities and Wards Cove. Some commenters stated that the proposed rule places a more onerous burden on defendants than what Inclusive Communities requires and that defendants should have only a burden of production at step two. Commenters said that although in Inclusive Communities, the Court did not address defendants' burden, it analogized it to the business necessity defense of Title VII under Wards Cove, which is one of production. They stated that the Court also made clear in Wards Cove that the defendant's obligation was only a burden of production and that this “conforms to the usual method for allocating persuasion and production burdens.” They said that the Inclusive Communities Court instructed that disparate impact claims must be limited to give insurers latitude to consider market factors. Another commenter focused on the Inclusive Communities statement that “housing authorities and private developers [are provided] leeway to state and explain the valid interest served by their policies” and concluded that a defendant need only explain how its policy interests are reasonably served by the particular practice, rather than prove it.

Other commenters supported the proposed burden of proof on defendants. These commenters noted that Wards Cove is no longer good law because the Civil Rights Act of 1991 specifically placed the step two burden of proof on defendants under Title VII. Commenters stated that the proposed rule contains the necessary protections for defendants, allowing them “leeway to state and explain the valid interest served,” consistent with Inclusive Communities. Commenters pointed out that Inclusive Communities specifically described defendants' burden as a burden of proof rather than production, and as “important and appropriate.”

HUD Response: As HUD noted in 2016, in over 25 years of case law since Wards Cove, no circuit court of appeals had ever applied the Wards Cove burden-shifting framework to the Act. Since then, only one circuit court of appeals has applied Wards Cove' s holding that step two requires a defendant to produce, rather than prove, its interest. HUD believes that the court's explanation in that case of why it applied Wards Cove to the Fair Housing Act case before it is unpersuasive and notes that it conflicts with the other circuits. Moreover, as explained above, Inclusive Communities' sole reference to Wards Cove was limited to a discussion that was not overruled by statute, and had nothing to do with the burden under step two; it instead related to the causation analysis required as part of a plaintiff's prima facie case under § 100.500(c)(1). Far from endorsing a burden limited to production, Inclusive Communities explicitly noted and approved of the requirement that defendants “prove” the necessity of their policies. And contrary to what some commenters wrote, Inclusive Communities did not, analogize the business necessity defense to Wards Cove's Title VII standard (which is a burden of production); instead, Inclusive Communities analogized the business necessity defense to Title VII's modern standard (which is a burden of proof). Further, Inclusive Communities specifically and favorably cited HUD's 2013 Rule as “properly limit[ing] disparate impact liability . . . to give housing authorities and private developers leeway to state and explain the valid interest served by their policies.”

Issue: Commenters suggested revising the requirement that defendants show a policy is “necessary” in step two to something less burdensome. One commenter suggested HUD should require a defendant to show only that the challenged policy is rationally-related to a legitimate, nondiscriminatory interest of the defendant. Another urged HUD to require that a defendant show its practice simply serves a valid interest of the defendant. Other commenters stated that in Wards Cove, the Supreme Court expressly rejected a “necessity” requirement, concluding that such a requirement would impose a degree of scrutiny impossible to meet.

Other commenters disagreed, stating that the proposed rule is consistent with Inclusive Communities, which requires defendants to prove that the challenged practice is necessary to achieve a valid interest. They cited the Court's statement that a housing provider should be allowed to maintain a policy if it is “necessary to achieve a valid interest” and noted a lower standard would conflict with well-established disparate impact jurisprudence, including under Title VII. A commenter also noted that in 2013 HUD specifically rejected a suggestion to remove “necessary” from the rule, because “necessary” is clear, uniform, in compliance with the 1994 interagency guidance, and effectuated the Act's broad remedial goal.

HUD Response: HUD declines to change the defendant's burden in step two because doing so would be inconsistent with longstanding judicial and agency interpretations and because HUD believes that the defendant's burden in step two best effectuates the broad, remedial goals of the Act. Moreover, the Court in Ward's Cove never expressly rejected a “necessity” requirement, but rather rejected a standard which required a showing that the challenged practice be essential or indispensable to the employer's business. The proposed rule does not require a defendant to show that a challenged practice is essential or indispensable, but only that it is necessary to achieve a substantial, legitimate, nondiscriminatory interest of that business.

Furthermore, the Court in Inclusive Communities specifically cited to the 2013 Rule's explanation of step two of the burden shifting approach as being analogous to the business necessity standard of Title VII when explaining that “this step of the analysis” of disparate impact liability is “an important and appropriate means of ensuring that disparate impact liability is properly limited.” The opinion continued that housing authorities must prove their policies are “necessary” to achieve a valid interest, which mirrors the necessity requirement of this final rule.

Issue: Commenters requested that HUD provide additional guidance in the final rule concerning what may constitute substantial, legitimate, nondiscriminatory interests. A commenter cited the 1994 Interagency Policy Statement on Discrimination in Lending, which describes factors that may be relevant to the legally sufficient justification, including cost and profitability. The commenter stated that the Policy Statement contains helpful guidance for lenders and urged HUD to reference the Policy Statement in this final rule.

Others stated that the proposed rule's failure to recognize practical business considerations, including profit making, as valid interests conflicts with Inclusive Communities, which stated that disparate impact liability must be limited to ensure that “regulated entities are able to make the practical business choices and profit-related decisions that sustain a vibrant and dynamic free-enterprise system” and warned against “second-guessing” between “two reasonable approaches.” They said that defendants must be given latitude to consider market factors.

In contrast, other commenters claimed that it is best to maintain a case-by-case approach so that no justification is automatically deemed a substantial, legitimate, non-discriminatory interest despite its disparate impact on a protected class.

Commenters expressed that profit should not be a legally sufficient justification and that the proposed rule makes clear that there are no automatically valid objectives, such as maximizing profit. Commenters stated that allowing defendants to justify discriminatory policies under the guise of profit would render the discriminatory effects framework completely toothless, because it would make for-profit businesses virtually immune from challenges to their policies or practices that cause a discriminatory effect. A commenter stated that almost all discriminatory policies can be justified by profit. Commenters also stated that a profit defense would be inconsistent with disparate impact jurisprudence; run counter to HUD's mission; and encourage the continuation of profitable, but discriminatory policies. A commenter explained that courts have appropriately rejected profit and market factors as substantial, legitimate, nondiscriminatory interests in the lending arena, limiting the legitimate business justification defense to a lender's use of objective variables and practices to ascertain creditworthiness. A commenter gave as examples cases in which lenders had engaged in practices not related to creditworthiness, like subjective markup pricing, that caused disparate impacts, to show profit should not be considered a legally sufficient justification.

HUD Response: HUD does not believe listing specific valid interests is necessary or appropriate and declines to alter the text of this rule. In promulgating the 2013 Rule, HUD did not state that profit or other business considerations could never be substantial, legitimate, nondiscriminatory interests; rather, HUD declined to explicitly name increasing profits, minimizing costs, and increasing market shares as per se substantial, legitimate, nondiscriminatory interests. HUD explained that the Act covers many different types of entities and practices, and a determination of what qualifies as a substantial, legitimate, nondiscriminatory interest for a given entity is fact-specific and must be determined on a case-by-case basis. HUD agrees that factors that may be relevant to a defendant's step two burden could include cost and profitability, as HUD and other agencies stated in the 1994 Interagency Policy Statement on Discrimination in Lending. However, recognizing interests as per se legitimate would undermine the effectiveness of disparate impact liability as a tool for rooting out policies that appear to be neutral but have been adopted for discriminatory reasons. HUD notes that Inclusive Communities highlighted disparate impact's important role in uncovering such disguised animus that escapes easy classification as disparate treatment. Accordingly, this rule, like the 2013 Rule, does not list interests that would always qualify as substantial, legitimate, nondiscriminatory interests for every defendant in any context. But the rule still allows regulated entities to make the practical business choices and profit-related decisions that sustain a vibrant and dynamic free-enterprise system and does not require second guessing between two reasonable approaches. HUD thus concludes that creating per se defenses would erroneously weaken the rule and result in the dismissal of cases where the practices are not actually necessary to achieve a valid interest and, more concerning, where the seemingly valid interests put forward by defendants are acting to disguise a defendant's underlying actions that are motivated by discriminatory intent.

Issue: Commenters stated that under Inclusive Communities, defendants are only required to show that the challenged practice is related to a valid interest, not that the practice is necessary or related to a substantial interest. Other commenters disagreed and stated that replacing the “substantial legitimate non-discriminatory interest” standard with a much lower and overly broad “valid interest” standard would make it too easy for defendants to rebut allegations of discrimination, allowing insubstantial business, profit, or policy considerations to defeat meritorious disparate impact claims, and would make it virtually impossible for plaintiffs to make a step three showing.

HUD Response: Nothing in Inclusive Communities suggests that the Court endorsed lowering the burden for defendants in step two of the discriminatory effects framework. When Inclusive Communities discussed the ability of defendants to state a “valid interest”, it referred specifically to HUD's 2013 Rule and the second step of the burden shifting analysis which requires that defendant show that its policy is necessary to achieve a substantial, legitimate, and nondiscriminatory interest. HUD believes the Court in Inclusive Communities, like other courts and HUD itself, used “valid” as shorthand for the same concept that the 2013 Rule describes as “substantial, legitimate, and non-discriminatory.”

To the extent that commenters nonetheless ask HUD to substitute a “valid interest” standard out of a belief that the Court intended a lower standard rather than one synonymous with the existing step two standard, HUD believes that it would be inappropriate to do so. HUD notes that such a standard would not accord with the majority of judicial opinions concerning the Act, and so it would introduce unnecessary confusion. Furthermore, as HUD stated in its 2013 Rule, “in order to effectuate the Act's broad, remedial goal, practices with discriminatory effects cannot be justified based on interests of an insubstantial nature.”

Section 100.500(c)(3): Proving an Alternative Practice That Could Serve the Interest With a Less Discriminatory Effect

Issue: Commenters asked HUD to place the evidentiary burden at step three on defendants, rather than plaintiffs. They cited to studies showing the difficulty plaintiffs have had succeeding with discriminatory effects claims over time, as well as Second Circuit precedent and the State of California's fair housing statute, which place the burden on defendants. Commenters stated that this revision is necessary because issues of segregation and discrimination in housing and lending have not abated since the 2013 Rule and, in fact, housing is more unaffordable, many cities have seen increasing displacement of communities of color, and borrowers of color are substantially more likely than white borrowers to be denied conventional loans. These commenters also cited to the growing role of data analytics and online platforms in the housing sale and rental market, increasing risks that segments of society will be steered away from or denied housing in a way that is immune to examination of intent, and resulting in even more segregated housing patterns. These commenters cited a 2021 Harvard study finding that the gap between whites and African Americans in homeownership rate stands at 28.1 percentage points, with the gap between whites and Hispanics at 23.8 percentage points.

HUD Response: HUD declines to place the step three burden on defendants. As explained in 2013, this rule's burden-shifting scheme is consistent with the majority view of courts interpreting the Act as well as the Title VII discriminatory effects standard codified by Congress in 1991, and the discriminatory effects standard under ECOA, which borrows from Title VII's burden-shifting framework. As HUD has explained, all but one of the federal appeals courts to address the issue have placed the burden at the third step on the plaintiff. HUD additionally notes the significant overlap in coverage between ECOA, which prohibits discrimination in any aspect of a credit transaction, and the Fair Housing Act, which prohibits discrimination in housing and residential real estate-related transactions. Thus, under the rule's framework, in litigation involving claims brought under both the Fair Housing Act and ECOA, the parties and the court will not face the burden of applying inconsistent methods of proof to claims based on the same underlying facts. Having the same allocation of burdens under the Fair Housing Act and ECOA will provide for less confusion and more consistent decision making by courts. Moreover, HUD continues to believe that this framework makes the most sense because it does not require either party to prove a negative.

Issue: Commenters criticized step three of the proposed rule, stating that it enables plaintiffs to prevail even if the less discriminatory alternative practice they present is unreasonable, less practical, less productive or less effective. Commenters asked HUD to revise step three to permit plaintiffs to prevail only if there is an alternative that is equally effective, is no more costly, or can be implemented at a reasonable cost, and does not impose an undue burden on a defendant or otherwise adversely affect the defendant's non-discriminatory policies and valid interests. Otherwise, commenters said, there would be no limit on what constitutes a reasonable alternative practice allowing plaintiffs to second-guess which of two reasonable approaches should be adopted.

Commenters stated that their proposed revisions to heighten a plaintiff's burden in step three are required by or consistent with Inclusive Communities, which held that the Act is not a tool for plaintiffs to force defendants to reorder their priorities or to displace valid governmental and private priorities, and that disparate impact liability must be limited so that employers and other regulated entities are able to make practical business choices and profit-related decisions. A commenter said that step three of the proposed rule is moot in light of Inclusive Communities' recognition that re-writing governmental policies exceeds the courts' remedial powers.

Commenters further stated that an equally effective standard for prevailing at step three of the analysis is required by Wards Cove. According to the commenters, Wards Cove explicitly requires that a plaintiff demonstrate an alternative policy is an “equally effective” alternative and warns that courts should proceed with care before mandating alternative practices. Commenters said that Wards Cove further noted that cost is relevant in determining whether an alternative is equally effective.

Other commenters supported retaining step three of the proposed rule, stating that the “less discriminatory alternative” requirement is consistent with judicial precedent and Congressional intent. They stated that an “equally effective alternative” requirement is not appropriate in the housing context where the practices covered by the Act are “not readily quantifiable.”

Commenters stated that the step three burden articulated by the 2020 Rule should not be retained for various reasons. Commenters said that the 2020 Rule's requirement that plaintiffs identify an equally effective alternative created too high a burden on plaintiffs; put defendant's financial gain above ensuring access to fair housing; departed from established precedent and the core purpose of the Act without justification; lowered the burden for defendants, such that clearly meritorious claims would be dismissed and the effectiveness of disparate impact liability as an incentive to identify less discriminatory alternative practices would be severely weakened; and improperly required plaintiffs to prove that any alternative is equally effective and does not impose materially greater costs. A commenter explained that if the 2020 Rule were retained, with its increased burdens on plaintiffs at step three and reduced burdens on defendants at step two, meritorious claims would be dismissed because it would shift much of the defendant's burden of proof at step two to the plaintiff to disprove at step three, insulating from scrutiny many policies that have an unjustified discriminatory effect. A commenter noted that the 2020 Rule's language was neither consistent with nor required by Inclusive Communities. One commenter explained that under the 2020 Rule's reformulation of step three, even a policy with the most flagrantly discriminatory effects would pass legal muster so long as a less discriminatory alternative is even slightly more costly or burdensome, and even if the alternative was still significantly profitable. A commenter pointed out that the term “other material burdens” in the 2020 Rule is undefined, broad and subjective, and forces plaintiffs to obtain information that is squarely in the purview of the defendant.

HUD Response: HUD declines to modify step three of the proposed framework or adopt the 2020 standard. As HUD explained in the 2013 Rule, the framework in this rule does not allow plaintiffs to impose untenable policies upon defendants because it still requires the less discriminatory alternative to “serve the defendant's [substantial, legitimate, nondiscriminatory stated] interests.” This rule's step three continues to be consistent with the 1994 Joint Policy Statement on Discrimination in Lending, with the purpose of the Act and its goal to “eradicate discriminatory practices within a sector of the Nation's economy,” and with judicial interpretations of the Act, including Inclusive Communities. HUD also notes that a requirement that alternative policies be “equally effective” did not appear in Inclusive Communities, despite citation to the proposed source of the requirement, Wards Cove, and significant discussion of the checks on liability that have always been part of Fair Housing Act jurisprudence. HUD further notes that its position is supported by the Massachusetts Fair Housing Center court, which criticized the 2020 Rule's inclusion of this requirement as “run[ning] the risk of effectively neutering disparate impact liability” and described it as onerous and inadequately justified. HUD, based on its own experience, agrees with the district court. Moreover, as discussed elsewhere in this preamble, Wards Cove construed Title VII, and the portions cited by commenters were barely tried, even in that context, having been superseded by the Civil Rights Act of 1991. In order to avoid unnecessary confusion and uncertainty, HUD declines to abandon a well-established standard in favor of a virtually untested one.

As to other concerns that commenters suggested required revisions to step three, HUD notes that an unreasonable alternative practice that creates an undue burden on defendant would not satisfy plaintiff's three-step burden. Nor will a proposed less discriminatory alternative fail simply because there will be some amount of increased cost associated with the alternative policy. And nothing in the rule suggests that reasonable, valid policies and priorities of defendants will be second guessed or forced to be reordered. Step one of the burden shifting framework ensures that the only policies which will be examined further are ones that cause a disparate impact because of a protected characteristic. Step three ensures that any alternative policy proposed is less discriminatory and actually serves the interest the defendant has already identified in step two. Moreover, HUD disagrees that Inclusive Communities rendered step three moot by stating that re-writing governmental policies exceeds the remedial powers of courts. Inclusive Communities did not say this. Indeed, there is no statement in Inclusive Communities indicating that courts lack the authority to invalidate policies that cause unjustified discriminatory effects.

In sum, HUD believes this provision and the structure of the burden shifting framework provide sufficient protections for defendants' business interests.

Issue: A commenter stated that step three of the proposed rule conflicts with Inclusive Communities because defendants can still be held liable despite establishing that their practices are substantial, legitimate, and nondiscriminatory at step two. One commenter criticized step three as unnecessary and inviting uncertainty and continued litigation. The commenter wrote that if the challenged practice is not artificial, arbitrary, and unnecessary, plaintiffs should not be permitted to substitute their proposed practices or business judgment for defendants' practices and judgment.

HUD Response: HUD disagrees. First, step two of the burden shifting framework requires the defendant to establish that its practice is necessary to achieve a substantial, legitimate, nondiscriminatory interest of the defendant—not that the defendant establish that the practice itself is substantial, legitimate and nondiscriminatory. By step two of the analysis, the plaintiff has already established that the practice itself causes or predictably will cause a discriminatory effect. Assuming the defendant meets its step two burden, the plaintiff then must establish that a less discriminatory alternative practice exists that still serves defendant's cited interest. HUD finds nothing in Inclusive Communities indicating that it is appropriate to cut off the inquiry after the second step.

Without step three, defendants with practices that have a discriminatory effect will have little incentive to examine their policies to determine if there are less discriminatory options to achieve their goals, thus allowing practices having unjustified discriminatory effects to continue unchecked. The suggestion that there should be no third step would eliminate a full assessment as to whether the same interest could be served in a less discriminatory way. The third step allows the plaintiff to offer an alternative policy that is less discriminatory and that still accomplishes the legitimate interest identified by the defendant. If the third step were eliminated, “artificial, arbitrary, and unnecessary barriers” would remain in place despite the fact that they are unnecessary to achieve the defendant's stated purpose.

Issue: A commenter said that the third step removes the Supreme Court's requirement that the plaintiff not resort to reverse discrimination.

HUD Response: The commenter did not explain what it meant by reverse discrimination or how step three might cause a plaintiff to resort to such discrimination. It is possible that the commenter meant that the need to satisfy the third step causes the defendant to resort to reverse discrimination. However, there is nothing in step three, or any other part of the proposed rule that requires the plaintiff or anyone else to resort to any type of discrimination. To the contrary, step three encourages defendants to utilize practices that have the least discriminatory effect because of any protected characteristic.

Section 100.500(f) (2020 Rule Only): Limiting Damages and Other Penalties in Discriminatory Effects Cases

Issue: Commenters criticized the proposed rule for omitting provisions in the 2020 Rule that limit HUD's authority to seek damages and penalties in discriminatory effects cases. The commenters stated that the 2020 Rule was consistent with Inclusive Communities' statement that “even when courts do find liability under a disparate-impact theory, their remedial orders must be consistent with the Constitution” and that “remedial orders should concentrate on the elimination of the offending practice.” One commenter suggested this means that the rule must only allow remedial orders that completely eliminate rather than only minimize the discriminatory effect. Commenters stated further that punitive or exemplary damages should not be allowed in discriminatory effects cases, noting that courts have applied a rigorous standard in assessing whether an award of punitive damages is proper. Commenters stated that because a discriminatory effects claim does not require a showing of defendant's state of mind, this type of claim cannot meet the standard for punitive damages.

HUD Response: HUD believes that limiting or suggesting favored remedies in this rule would be contrary to the plain language of the Act and its broad remedial purpose. The Act explicitly provides for punitive and compensatory damages, civil penalties (in cases brought by the Attorney General), and injunctive relief in federal court, and actual damages, injunctive and equitable relief and civil penalties in administrative hearings. HUD does not believe that Inclusive Communities can be read to suggest that remedial orders should be the sole or favored remedy in discriminatory effects cases or that civil penalties are somehow inappropriate. Rather, the Court merely addressed what courts must keep in mind when remedial orders are issued. Nor does HUD believe Inclusive Communities can be read to limit remedial orders to only those that completely eliminate discriminatory effects. Moreover, well-established criteria in statute and in decades of judicial precedent set forth when penalties and punitive damages may be appropriate, thus preventing arbitrary awards. In any case, the availability of various remedies does not mean that they will be sought or granted in all cases; remedies are considered on a case-by-case basis.

Comments Regarding Other Defenses and Safe Harbors

Issue: Commenters stated that HUD should add a defense similar to the 2020 Rule's third-party defense allowing defendants to rebut plaintiff's prima facie case by showing that their discretion was materially limited by a third party. They said such a defense is necessary to protect defendants from being held liable for discriminatory conduct mandated by law. As an example, a commenter asserted that independent mortgage banks must follow guidelines set by federal agencies, including the government sponsored enterprises, the Department of Veterans Affairs, the Federal Housing Administration, and the Department of Agriculture, which may cause a disparate impact. The commenter stated that these mortgage banks should be granted a safe harbor in such situations. A commenter stated that Inclusive Communities requires a third-party defense because it stated that causation does not exist where the defendant's discretion is substantially limited and cited to a concurring opinion in the appellate court decision that included as an element of plaintiff's prima facie case that the defendant's policy or practice is not a result of a law that substantially limits defendant's discretion.

Commenters also stated that a third-party defense is necessary to protect insurers that conform to state laws and regulations on insurance in compliance with McCarran-Ferguson. They said that HUD cannot now say that the third-party defense is inconsistent with the Act when it said in the 2020 Rule that “in the event that unlawful discriminatory practices are mandated by statute or court order, the most effective way to eliminate the discrimination is to remove or modify the underlying statute or order that mandated the unlawful discrimination.” They also stated that HUD's proposal to remove the 2020 defense implies that HUD intends to improperly test the boundaries of its ability to preempt state regulations.

Other commenters supported HUD's proposal not to retain the third-party defense from the 2020 Rule, arguing that the defense would allow defendants to evade liability for illegal acts by showing that they complied with third party requirements that are themselves discriminatory. Commenters stated that the defense would deny plaintiffs the ability to address whether less discriminatory alternatives exist to the defendant's chosen method of meeting the third-party requirement. Commenters also stated that such a defense is contrary to the Act's preemption clause because it would prioritize compliance with local ordinances over federal civil rights obligations, even where there may be less discriminatory ways to comply with the third party's requirement, and even in exclusionary zoning cases. In addition, a commenter stated that a third-party defense would impede efforts to prevent algorithm-driven discrimination. Another commenter characterized the 2020 Rule's third-party defense as unnecessarily confusing and vague. A commenter stated that whether a party's discretion is limited by a third party should be addressed at the second step of the burden shifting framework, not at the pleading stage.

HUD Response: HUD disagrees that a third-party defense should be included in this rule. First, Inclusive Communities does not suggest that such a defense is necessary or appropriate. Inclusive Communities stated that if a plaintiff “cannot show a causal connection . . .—for instance, because federal law substantially limits the Department's discretion—that should result in dismissal of this case.” As this passage suggests, if federal law requires defendants to act in a certain manner, plaintiffs may not be able to show that defendants' actions are the cause of a discriminatory effect. Such an argument already is available to defendants under this rule in appropriate cases and does not require revisions to this rule. And as noted in the 2013 Rule, the discriminatory effects standard already permits a defendant to defend against a claim of discriminatory effects by establishing a legally sufficient justification, as specified in § 100.500. Thus, independent mortgage banks, for example, who follow federal guidelines, have multiple opportunities under the rule to defend their practices: first, at the prima facie stage, if federal guidelines, rather than the challenged practices are the cause of the discriminatory effect, and also at the second step of the burden shifting framework, by showing that the practice is truly necessary to comply with the federal guidelines.

That does not mean that, as the 2020 Rule allowed, any time defendants are subject to a third-party requirement, the plaintiff's case will necessarily fail. As other commenters explained, there may be multiple ways of complying with a third-party requirement, some of which have an unjustified discriminatory effect and some of which do not. In those cases, the defendant caused the effect by opting for one way of complying with a third-party requirement over another that does not cause such an effect. A third-party defense such as was included in the 2020 Rule would allow defendants to avoid the requirement to utilize a less discriminatory alternative to comply with a third party requirement. Indeed, if the defense could be raised at the pleading stage (as permitted by the 2020 Rule), disparate impact claims could be dismissed based on mere assertions of third party requirements, without plaintiffs having any opportunity to challenge these assertions with the benefit of discovery. Without discovery, some plaintiffs would have no means of ascertaining whether the third-party obligation actually exists, and if so, whether it is the actual, legitimate reason for defendant's policy or practice, whether that obligation actually requires the defendant to implement the policy at issue or if there is a less discriminatory way to do so. This would essentially eliminate any meaningful inquiry into steps two and three of the burden shifting framework whenever the defendant asserts that its policy was required by a third party. Such a defense is inappropriate because it presumes that discrimination may be permitted without consideration of whether the third-party requirement is itself discriminatory or whether there are non-discriminatory ways to comply with that third-party requirement. Moreover, such a defense would be inconsistent with the Act, which specifies that state and local laws requiring or permitting discriminatory housing practices are invalid.

Issue: Commenters supported HUD's proposed removal of the “outcome prediction” defense that was inserted into the 2020 Rule, which was designed to shield certain policies based on algorithms from disparate impact liability with a “results-based approach.” Commenters stated that the final rule's framework was the appropriate method of analyzing discriminatory effects claims involving algorithmic and machine learning technologies. A commenter noted that disparate impact litigation is a key mechanism for redressing discrimination in light of the increase of algorithms, which bring risks for perpetuating or amplifying patterns of discrimination through biased development, biased inputs, or bias arising from automatic adaptations from artificial intelligence. Another commenter stated that the potential for disparate impact liability protects borrowers and encourages lenders using these technologies to innovate in ways that expand access to credit. A commenter stated that the rule should clarify in the preamble that the Act applies to entities that rely on algorithms.

Commenters expressed numerous concerns about the defense. Commenters stated that this defense would have the practical effect of foreclosing many disparate impact claims based on algorithms and models and would shield such defendants from liability. Commenters stated that in creating this defense, the 2020 Rule impermissibly created exemptions for predictive models in the lending and insurance industry that have no basis in the Act or any other source of authority, because the Act does not grant HUD authority to create safe harbors or exceptions from discriminatory effects liability, and no court has ever held that entire categories of policies or practices that might otherwise be subject to challenge are exempt from such liability. Commenters also noted that HUD had previously explained why categorical exemptions from disparate impact liability are undesirable.

A commenter noted that the defense would shield a wide range of discriminatory policies and practices, because many discriminatory models would qualify for an exemption because of the 2020 Rule's novel “similarly situated individuals” analysis, for which there is no basis as a matter of law or as a matter of fact. As one commenter explained, it would be easy for defendants to show that a challenged policy or practice is intended to predict an outcome because that is what any predictive model claims to do, and it would be easy to show that the prediction represents a valid interest. Another commenter stated that the defense is based on an outdated academic theory of discrimination that relies on statistical disparities to absolve defendants of liability, without acknowledging the possibility that the defendant's policies contributed to the disparities.

Commenters noted that the defense would make it very difficult for plaintiffs to demonstrate that an alternative, less discriminatory policy would result in the same outcome, without imposing materially greater costs or other material burdens because there is no standard or agreed upon definition of algorithmic predictive performance or accuracy. Commenters also stated that the defense would make it difficult for plaintiffs to prevail because of the proprietary nature of algorithms; without knowing what the algorithm is and how it works, it is nearly impossible to demonstrate what the 2020 Rule requires: that the practice has a disproportionately adverse effect on members of the protected class, that there is a robust causal link between the algorithm and this adverse effect, and that this effect is significant.

Commenters also described the defense as ambiguous and difficult for parties and courts to apply. Another commenter described this and other defenses as confusing and harmful, noting that the defense would obfuscate discrimination in lender models and algorithmic systems.

Moreover, commenters noted that the defense was promulgated without public notice and comment and was not a “logical outgrowth” of the 2019 Proposed Rule, thereby violating the APA.

HUD Response: HUD agrees with the commenters that the defense is not appropriate to include in this rule. Upon HUD's consideration of these comments and in light of HUD's experience interpreting and handling cases under the Act, HUD has determined that the outcome prediction defense is unclear and not found in any case law. The rule properly describes the framework to be used in cases involving algorithms and machine learning in housing and housing-related transactions. The defense, if retained, could in practice improperly exempt many housing-related practices that are increasingly reliant upon algorithms and automated processes that rely on outcome predictions, such as lending practices, from liability under a disparate impact standard. The defense would be inconsistent with HUD's repeated finding, including in the 2020 Rule, that “a general waiver of disparate impact law for the insurance industry would be inappropriate.” And although unclear, with its novel “compared to similarly situated individuals not part of the protected class” language, it appears that this defense would suggest using comparators that are, in HUD's experience, inappropriate, and would fail to consider the reasons why disparities are observed. At the very least, the defense introduces unnecessary confusion into disparate effects doctrine.

Issue: In addition to insurers, various other commenters requested safe harbors or exemptions. Commenters stated HUD should develop safe harbors for those who “followed rules set out by HUD in developing their operating policies.” Another commenter seemed to similarly request specific protections for public housing agencies (PHAs) that have policies that are consistent with HUD rules for operation for federally assisted housing, are in compliance with otherwise legitimate laws, are approved for use in federally insured housing, or are for the purpose of eligibility criteria for enhancing housing opportunities for protected classes or other under-housed persons.

Commenters suggested that HUD should have, as a safe harbor, a process where HUD provides concrete guidance to housing providers so that the housing providers do not have to wait until litigation to discover that their policy may violate the Act. Another commenter requested a safe harbor for entities that implement written policies that identify non-discriminatory goals, explain how the policy is reasonably calculated to achieve that goal, and conclude that the policy does not impose a greater burden on members of protected classes than it does on the wider population. Some commenters requested a safe harbor for credit unions that limit membership based on statutory requirements, explaining that while a disparate impact claim already would fail under the proposed rule because credit unions are legally unable to lend outside their membership, litigating these cases, even by just filing a motion to dismiss, is costly, particularly for small credit unions. They stated further that Inclusive Communities stated that where a causal connection between a policy and a disparate impact cannot be shown because federal law substantially limits discretion of the defendant, dismissal is appropriate.

Other commenters said no safe harbors should be provided for policies and practices that have discriminatory effects by limiting housing opportunities for protected groups. Commenters stated that the 2013 Rule, 2016 Supplement, and this rule appropriately apply a case-by-case disparate impact analysis to all housing related industries and agreed that this approach is consistent with the Act.

HUD Response: HUD agrees with commenters that advocate for a case-by-case approach rather than safe harbors. As explained above, HUD believes that it does not have the authority to create exemptions that do not appear in the statute. Moreover, even if a court were to find that HUD had such authority, HUD believes that a case-by-case approach appropriately implements HUD's obligations to enforce the Act to redress discrimination that exists in an entire sector of the economy and to affirmatively further fair housing. Moreover, safe harbors are unnecessary as regulated entities can defend themselves utilizing the second step of the burden shifting framework. Regulated entities may also use the burden shifting framework to assess their own existing policies and practices as well as new policies and practices that are under consideration in order to ascertain whether they may cause an unjustified discriminatory effect. HUD also emphasizes that entities' purported compliance with program specific rules does not guarantee compliance with the Fair Housing Act, and that the Fair Housing Act's mandate is to refrain from discrimination—including refraining from using policies or practices with unjustified discriminatory effects. The rule provides clarity as to how HUD and a court would analyze such a claim, allowing regulated entities to better comply with their obligations under the Act and prevent unjustified discriminatory effects in the first place. HUD notes further that Inclusive Communities provides no support for any exemptions; the passage cited by commenters merely explains that courts dismiss lawsuits pursuant to the pleading standards in the Federal Rules of Civil Procedure.

Comments Regarding Additional Explanations, Examples and Guidance

Issue: Commenters suggested that HUD should provide additional examples of practices that can have unjustified discriminatory effects in the rule, its Preamble, or in future guidance. They suggested that HUD: note in the rule that policies or practices that result in the benign neglect of people with disabilities can have discriminatory effects; provide examples of specific zoning ordinances or other policies that restrict manufactured housing and may have a discriminatory effect; and discuss criminal records screening practices as examples of policies that may have an unjustified discriminatory effect on protected classes. One commenter suggested that HUD should outline less discriminatory alternatives to eviction in cases where a housing provider has a policy of evicting the household for certain types of lease violations.

HUD Response: HUD declines to insert examples of practices that may specifically have unlawful discriminatory effects into this rule. This rule is designed to provide a framework to help entities and courts assess whether a policy or practice may have an unjustified discriminatory effect, not to establish a list of practices that may be unlawful under a discriminatory effects theory. HUD has already issued Guidance on some topics, including certain criminal records screening practices and certain zoning practices —that may have unjustified discriminatory effects on protected classes.

While HUD believes the rule provides a sufficiently clear framework under which specific practices can be evaluated, HUD will consider issuing more guidance as it deems appropriate.

Issue: A commenter urged HUD to consider providing separate guidance related to the use of algorithms, artificial intelligence, and machine learning.

HUD Response: HUD appreciates this request and believes that the rule provides the appropriate framework for evaluating discriminatory effects liability for all claims under the Act, including as applied to algorithms, artificial intelligence, and machine learning. However, given the rapid evolution in this field, HUD will consider in the future whether to adopt more detailed guidance expanding on those particular types of claims.

Issue: A commenter asked HUD to specify that a one-off action, like the decision of a private developer to construct a new building in one location rather than another, is insufficient to establish disparate impact liability. Other commenters opposed this change, noting that a single zoning decision or single application of a zoning standard often results in or is, in fact, a community's policy or practice and can have wide discriminatory impacts. They noted that a “single event” limitation would essentially sanction many discriminatory zoning actions, even where Inclusive Communities specifically called suits targeting “zoning laws and other housing restrictions . . . that function to unfairly exclude minorities from certain neighborhoods without any sufficient justification . . . [as] resid[ing] at the heartland of disparate impact liability.”

HUD Response: HUD believes that a so-called “one-off” action may, in certain cases, be sufficient to establish a practice that has an unjustified discriminatory effect. As noted throughout this preamble, HUD continues to find that discriminatory effects claims should be assessed on a case-by-case basis. A “one-off” exception would tend to protect siting decisions that may have been influenced by a community's desire to keep out people of a certain race, or against people with disabilities. And HUD agrees with commenters that this limitation may pose obstacles to meritorious zoning cases. HUD notes that an individual siting decision by a private housing developer, or a single zoning decision by a locality, will not result in an unjustified discriminatory effect liability so long as the developer or locality has a legally sufficient justification for that decision. HUD believes this fully protects localities and the individual siting decisions of private housing developers. Furthermore, while the Court in Inclusive Communities noted in dicta that it would be difficult to prove that a developer's one-time decision to build in one location rather than another was a policy that caused a disparate impact, it did not go so far to say that such a scenario could never succeed under a disparate impact theory.

Issue: A commenter suggested that in the preamble to the final discriminatory effects rule and in separate guidance, HUD should outline potentially less discriminatory alternatives to eviction in cases where a housing provider has a policy of evicting the household for certain types of lease violations.

HUD Response: As discussed above, HUD believes that each discriminatory effects claim should be assessed on a case-by-case basis, including what less discriminatory alternatives might exist. HUD declines to provide additional guidance in this regulation but will consider in the future whether such guidance may be appropriate.

Issue: Commenters asked HUD to make revisions to various guidance documents including the 2016 Office of General Counsel Guidance entitled “Application of Fair Housing Act Standards to the Use of Criminal Records by Providers of Housing and Real Estate-Related Transactions” (2016 Guidance) and a 2011 internal HUD memorandum for HUD's Fair Housing and Equal Opportunity headquarters and field staff entitled “Assessing Claims of Housing Discrimination against Victims of Domestic Violence under the Fair Housing Act (FHAct) and the Violence Against Women Act (VAWA)”. Other commenters stated that the 2016 Guidance provided clarity.

HUD Response: While guidance is beyond the scope of this rulemaking, HUD will consider at a later date whether any revisions to guidance documents may be necessary or helpful.

Comments Regarding Effects of the Proposed Rule

Issue: Commenters stated that regulated entities will face increased litigation risks under the proposed rule. They said that the “specter of disparate-impact litigation” could discourage businesses from undertaking the activities that ensure a well-functioning housing market, undermining the Act's purpose and the free-market system. A commenter stated that the 2013 Rule created an uncertain legal environment where any adverse impact that a practice may have on a protected group invited the threat of a lawsuit over its discriminatory effect. As an example, this commenter stated that people of color are more likely to be tenants than homeowners, so the proposed rule invites tenant advocates to assert that any rule or policy that is adverse to actual or prospective renters may have a discriminatory effect while citing little or no statistical or evidentiary basis. Commenters stated that reinstating the 2013 Rule will increase litigation costs, with one commenter saying that this is due to unclear, overly burdensome, and duplicative standards, saying that HUD itself recognized this in the 2019 Proposed Rule. A commenter also noted that the proposed rule would impose additional burdens on entities administering Community Development Block Grant disaster relief funding.

HUD Response: HUD disagrees that this rule will increase litigation risks in a manner that interferes with the free-market system and undermines the purpose of Act. This rule does not restrict valid free-market activity, but rather only regulates policies that have unjustified discriminatory effects and encourages businesses to develop and implement policies that achieve their substantial legitimate purposes in the least discriminatory manner. HUD believes the rule will further the goal of a vibrant, integrated, and open housing market. Furthermore, the rule does not create any new liability or burdens for businesses or declare any activity per se unlawful. It merely prescribes a method for evaluating liability under the Act. HUD believes that litigation and burdens concerning the Act is more properly attributable to the Act, rather than the rule. HUD notes that commenters' concerns are undermined by the fact that the rule has been in place since 2013 and in HUD's experience, no such increased litigation has occurred.

Issue: A commenter critiqued the rule for increasing the threat of challenges to certain types of landlord practices that the commenter characterized as justified by practical business decision making. For instance, the commenter stated, the rule allows landlords to be sued for occupancy restrictions that are stricter than state or local codes even though these restrictions advance housing providers' legitimate interests in limiting wear and tear, minimizing operational costs, and addressing issues with safety, overcrowding, and noise in multifamily properties. According to the commenter, advocates have increasingly used the rule to challenge occupancy restrictions with HUD and state agencies. The commenter also said that the rule allows lawsuits against landlords who purchase buildings that are largely populated by persons with certain protected characteristics and institute new rules that are facially neutral but have a disparate impact on protected classes, such as requiring existing tenants to provide valid government issued identification or requiring tenants to pay rent through direct deposit from a bank account. The commenter also mentioned the 2016 Guidance on criminal records, alleging that it has invited costly challenges to all criminal records screening, and requires property owners to set up a “mini parole board” to review applicants with criminal records, even if there are legitimate reasons not to rent to persons with certain types of criminal records, including lesser offenses like disorderly conduct, illegal drug use, and nuisance conduct. Finally, the commenter stated that advocates are using the 2013 Rule against landlords who choose not to participate in the Section 8 program. The commenter described one class action case in which it represented the new owner of a building who decided to not accept Section 8 vouchers and wanted to raise rents to pay for costly improvements to the building, which allegedly had a discriminatory effect on residents with disabilities, African Americans and Hispanics who used Section 8 vouchers or could not afford higher rent. The commenter said that the claim survived a summary judgment motion because of the possibility that the plaintiffs might be able to show a less discriminatory alternative, such as raising rents less, doing less to improve the property, or looking for some public funding or subsidies to allow the owners to get a return on investment without discontinuing participation in Section 8 and raising rents.

HUD Response: HUD disagrees that the commenter's concerns merit revising the rule. The examples provided by the commenter are well within the types of cases that may or may not state a valid claim under the Act and may be decided through the rule's framework. While the practices cited by the commenter that have a discriminatory effect on protected classes may advance legitimate interests, these practices are still illegal if, for example, they are not necessary to achieve substantial, legitimate, nondiscriminatory interests of the defendant or there is a different practice that advances that same interest but has a less discriminatory effect. This is the essential framework of discriminatory effects liability that has developed in case law, whether the rule exists or not. For example, as noted in the 2013 Rule, even decades prior to the rule, unreasonable restrictions on occupancy that impose a discriminatory effect on families with children will result in liability. Furthermore, the 2016 Guidance on criminal records sets out this well-established discriminatory effects framework. It does not require a “mini parole board” but rather posits that an individualized review of a person with a criminal record is likely to have a less discriminatory effect than a policy that imposes an automatic categorical ban. To the extent the commenter disagrees with that assessment, its quarrel is with the Guidance's particularized application of this rule and not with the more general principles this rule sets out.

In sum, HUD believes that this rule strikes a reasonable balance, in accordance with the Act and with caselaw, between allowing policies that permit landlords to advance their interests, even if those policies disproportionately adversely impact protected classes, while requiring landlords to demonstrate that their policies are necessary to advance those interests.

Issue: A commenter stated that the proposed rule violates the constitutional principle of separation of powers because § 100.500(c) is an attempt by HUD to dictate rules of judicial procedure and evidence to the judicial branch. The commenter said the rule would unnecessarily produce complication throughout the federal courts because they have no obligation to use a standard dictated by the executive branch, and different courts will decide to follow the rule (or not) in different ways. The commenter requested that HUD rescind all provisions that address judicial standards of review, rules of procedure, and evidence. The commenter continued that HUD's reliance on Congress's delegation of certain authority under section 3608(a) is improper because that statutory provision does not mention the judiciary, standards of review, rules of procedure or evidence, or any directives for HUD to assume a role that is clearly the province of the judiciary.

HUD Response: The final rule sets out a framework for analyzing and proving cases under a theory of discriminatory effects and does not amend or establish rules of judicial or civil procedure or evidence. It remedies concerns expressed by many commenters that the 2020 Rule infringed upon the judicial branch by, for example, setting pleading standards, establishing a confusing burden that appeared to contradict the Federal Rules of Civil Procedure, and defining “plausibility” in such a way as to preclude substantively meritorious claims. HUD agrees that it does not have the authority to amend pleading standards, to modify the defenses under Federal Rule of Civil Procedure 12, or the rules of evidence. It does, however, have authority to “make rules . . . to carry out” the Act, including the prohibition of discrimination in housing. That is precisely what HUD is doing here.

Comments Regarding the Administrative Procedure Act

Issue: Commenters disagreed about whether the proposed rule violates the Administrative Procedure Act (APA). A commenter stated that HUD's reasoning for the proposed rule, that “the practical effect of the 2020 Rule's amendments is to severely limit HUD's and plaintiffs' use of the discriminatory effects framework in ways that substantially diminish that frameworks' effectiveness,” does not satisfy APA requirements because HUD neither provided the essential facts on which its conclusion was based nor explained how those facts justified that conclusion, instead making a conclusory statement. Another commenter stated that reverting to the 2013 Rule, which was promulgated before Inclusive Communities, without adequately explaining the reversal would violate the APA.

Other commenters stated that the 2020 Rule violated the APA and that the proposed rule would rightfully reinstate the standard set forth in the 2013 Rule. Commenters stated that changes made by the 2020 Rule were arbitrary and capricious and contrary to law. Commenters also stated that the 2020 Rule failed to explain why it was deviating from legal standards and failed to address that courts have easily applied existing disparate impact case law and Inclusive Communities. Commenters also stated that the 2020 Rule violated the APA by failing to address numerous comments about the negative effects the rule would have, namely on plaintiffs' ability to successfully challenge housing discrimination in accordance with the Act.

HUD Response: HUD believes that the proposed rule and this final rule fully comply with the requirements of the APA. HUD disagrees with the assertion that HUD did not explain its proposed rule in light of Inclusive Communities. The proposed rule directly and thoroughly explained HUD's reason for believing that the 2013 Rule was consistent with and in fact supported by Inclusive Communities. The proposed rule also explained why HUD believed that the 2020 Rule was deficient. There, HUD provided a number of different reasons why it was proposing to change course from the position it had taken in 2020 and was proposing recodification of the 2013 Rule, not just that “the practical effect of the 2020 Rule's amendments was to severely . . . diminish the [discriminatory effects] framework's effectiveness.” While the commenter characterized this statement as conclusory, HUD explained that this belief was “based on HUD's experience investigating and litigating discriminatory effects cases.” HUD further explained that its experience informed “it that many of the points made by commenters opposing the 2020 Rule and the Massachusetts District Court are correct, including that the changes the 2020 Rule makes, such as amending pleading standards, changing the burden shifting framework, and adding defenses, all favoring respondents, will at the very least introduce unnecessary confusion and will at worst make discriminatory effects liability a practical nullity.” Further, the APA requires in relevant part that a proposed rule make reference to the legal authority under which the rule is proposed and include either the terms or substance of the proposed rule or a description of the issues involved, all of which HUD did in the proposed rule. Now, in this final rule, HUD has again explained that the Act vests HUD with the requisite authority, and has further explained why, having reconsidered the 2020 Rule, HUD is finalizing its proposal to recodify the 2013 Rule.

Furthermore, this final rule explains HUD's position after consideration of the comments HUD received on the proposed rule. As explained in the proposed rule, the facts that spurred HUD's decision to recodify the 2013 Rule include the consistent concerns expressed through thousands of public comments regarding the effect of the 2020 Rule's changes on disparate impact jurisprudence and protected classes, the concerns raised by the court in Massachusetts Fair Housing Center, HUD's own experience in interpreting and applying the Act, which indicated that these criticisms are correct, and HUD's determination after examining case law that several provisions of the 2020 Rule were inconsistent with the purpose of the Act and judicial and agency precedent. HUD expounds upon its reasoning in this final rule in responding to specific comments.

HUD is not ignoring facts or circumstances that underlay HUD's 2020 Rule; rather, HUD is acknowledging its change in position, drawing on its experience in different ways than it did in the 2020 Rule, drawing on case law that did not exist when the 2020 Rule was promulgated, relying on other case law that the 2020 Rule downplayed and/or ignored, and drawing on new public comments about the final version of the 2020 Rule that did not exist when HUD decided to issue the 2020 Rule.

Comments Regarding HUD's Findings and Certifications

Issue: Commenters stated that HUD inappropriately and incorrectly assumed that reinstating the 2013 Rule “would not have federalism implications,” and asserted that HUD should have consulted with state regulators as required by Executive Order 13132 and acknowledged that the rule would interfere with state law in violation of McCarran-Ferguson in all or nearly all cases.

HUD Response: HUD stands by its certification that this rule—like the 2013 Rule it recodifies and the 2020 Rule it rescinds—does not have federalism implications. These commenters' assertion that this rule is inconsistent with Executive Order 13132 is based solely on the assertion that this rule would interfere with states' ability to regulate insurance in violation of McCarran-Ferguson. As discussed extensively above, HUD disagrees with this assertion and finds that this rule does not interfere with state insurance laws and is consistent with McCarran-Ferguson. Therefore, this rule has no federalism implications. The existing relationship between the Act and McCarran-Ferguson, and therefore the Act and state insurance law, remains the same before and after this rule. Section 6 of Executive Order 13132 only requires consultation with the states when there are federalism implications, when a regulation has “substantial direct effect on the States”; therefore, HUD has no obligation to consult with state regulators.

Issue: Commenters stated that the proposed rule would have an impact on regulated entities and that the proposed rule does not pass a basic cost benefit analysis. Another commenter stated that the proposed rule would eliminate economic burdens that the 2020 Rule imposed by removing ambiguity and uncertainty that would have led to expensive litigation and dispute resolution and would have imposed these expenses on plaintiffs, state governments, the public, and attorneys general.

HUD Response: HUD agrees that 2020 Rule would have been burdensome if it had not been enjoined and that the proposed rule does not impose a significant economic impact, as further explained in HUD's certification. HUD stands by its certification that this rule will not have a significant economic impact. Because the rule does not change decades-old substantive law articulated by HUD and the courts, but rather formalizes a clear, consistent, nationwide standard for litigating discriminatory effects cases under the Fair Housing Act, it adds no additional costs to housing providers and others engaged in housing transactions. Rather, HUD believes that the rule will simplify compliance with the Fair Housing Act's discriminatory effects standard and decrease litigation associated with such claims by clearly allocating the burdens of proof and how such burdens are to be met.

Other Comments

Issue: A commenter stated that the proposed rule is unnecessary because credit unions have already carefully structured their policies to comply with the Act.

HUD Response: HUD appreciates the efforts of regulated entities to comply with the Act and its implementing regulations. However, HUD disagrees that any alleged current compliance provides a basis for retracting the rule or providing exemptions, particularly where policies may change in the future. HUD believes—and many commenters have stated—that this rule is a necessary tool for ensuring both new and continued compliance.

Issue: A commenter asserted that plaintiffs sometimes use the disparate impact framework to bring costly and lengthy litigation which is resolved without a court finding. A commenter suggested HUD include an extensive examination of disparate impact cases relating to residential lending activity from the standpoint of any actual discriminatory findings and court judgments and provide an accounting of cases brought in class action form, examining and reporting any monetary awards actually being delivered to the purported class.

HUD Response: This comment is outside the scope of the proposed rule because it is a criticism of plaintiffs who bring cases based on a disparate impact theory of liability and/or the disparate impact theory itself, rather than the final rule. HUD does not believe that conducting such an examination or finding that most disparate impact cases settle before a judicial determination, would inform any changes HUD should or should not make to the proposed rule. For these reasons, HUD declines to conduct such an examination in this final rule.

IV. Severability

Consistent with the requirements of the Administrative Procedure Act, HUD has carefully responded to all public comments received in response to its notice of proposed rulemaking. HUD has determined that the discriminatory effects standard and burden-shifting framework in this rule appropriately implement, and are fully consistent with, the Fair Housing Act and governing law, including Inclusive Communities. Furthermore, HUD's decision to not create exemptions for any industry covered by the Fair Housing Act is also fully consistent with the plain language of the Act and governing law, including the McCarran-Ferguson Act. As explained in 2013, 2016, and 2020, as well as in greater detail above, HUD is declining to provide any exemptions, including for the insurance industry, in whole or in part, including because HUD lacks the authority to create such exemptions under the Act. Further, declining to provide exemptions for certain industries furthers congressional intent by effectuating the Act's broad remedial purpose.

Through this rule, HUD is taking two separate actions. First, HUD rescinds the 2020 Rule, removing 24 CFR 100.500 and the second and third sentences of 24 CFR 100.5(b), thus nullifying the 2020 Rule and eliminating any and all legal effect that the 2020 Rule could have. Second, HUD adds a new 24 CFR 100.500 and a new second sentence to 24 CFR 100.5(b). The new language in both sections is identical to the language in those sections of the Code of Federal Regulations that took effect on March 18, 2013, which HUD refers to throughout this preamble as “the 2013 Rule.” HUD intends the language promulgated today to be the only operative rule.

HUD intends these separate actions to be legally severable. In particular, in the event that any portion of § 100.500 or § 100.5(b) of this final rule is held to be invalid or unenforceable, HUD intends that the rescission of the 2020 Rule be unaffected. HUD believes that it would be more consistent with the plain language and legislative history of the Act for the Code of Federal Regulations to contain no language regarding discriminatory effects liability and for litigants to rely on existing jurisprudence than for any provision of the 2020 Rule to remain in effect. HUD has made this determination for all the reasons described elsewhere in this preamble, including that the 2020 Rule is inconsistent with such jurisprudence. In addition to rescinding the 2020 Rule for the reasons described more fully in this preamble, HUD's rescission will serve to resolve three pending lawsuits, all of which challenge the 2020 Rule as arbitrary and capricious and inconsistent with Inclusive Communities and other case law. Moreover, having no rule in place at all regarding discriminatory effects would be workable, as precedent proves; for decades prior to the 2013 Rule, there was no HUD rule on discriminatory effects liability, and litigants relied on caselaw.

HUD also intends that the rule be treated as severable in its applications to certain industries. Litigation brought by the insurance industry regarding the 2013 Rule is ongoing. One of those cases, decided in the context of the 2013 Rule, has already upheld the rule's burden-shifting framework for analyzing discriminatory effects claims as a reasonable interpretation of the Act, but also held that HUD had not adequately explained why case-by-case adjudication was preferable to using its rulemaking authority to provide exemptions or safe harbors related to homeowners insurance. To resolve that suit, HUD issued the 2016 Supplemental Explanation. The plaintiff filed an amended complaint and that litigation is pending. HUD believes, as described in greater detail above, that discriminatory effects liability can be properly applied to the insurance industry and that doing so is fully consistent with the Act's plain language and broad remedial purpose. However, should a court decide that the insurance (or any other) industry or certain types of insurance (or other) claims should be exempt from the Rule, HUD intends that this final rule remain in effect and apply to all other actors and claims covered by the Act. Moreover, in the event of such a court decision, this final rule would still function sensibly with respect to others covered by the Act, as nothing in this final rule's applicability to the insurance (or any other) industry affects its applicability to others covered by the Act.

V. Findings and Certifications

Regulatory Review—Executive Orders 13563 and 12866

Executive Order 13563 (“Improving Regulation and Regulatory Review”) directs agencies to propose or adopt a regulation only upon a reasoned determination that its benefits justify its costs, emphasizes the importance of quantifying both costs and benefits, of harmonizing rules, of promoting flexibility, and of periodically reviewing existing rules to determine if they can be made more effective or less burdensome in achieving their objectives. Under Executive Order 12866 (“Regulatory Planning and Review”), a determination must be made whether a regulatory action is significant and therefore, subject to review by the Office of Management and Budget (“OMB”) in accordance with the requirements of the order. This rule was determined to be a “significant regulatory action” as defined in section 3(f) of Executive Order 12866 (although not an economically significant regulatory action, as provided under section 3(f)(1) of the Executive Order).

In its proposed rule, HUD invited comments on whether any further analysis was needed to assess the impact of the rule, given the fact that the rule would simply be retaining the status quo and would therefore have no new impact on regulated entities. Specifically, HUD explained: “[b]ecause the 2020 Rule never took effect, and therefore did not affect the obligations of any regulated entities, this proposed rule is only recodifying the 2013 Rule and will have no impact on regulated entities except to affirm that the 2013 Rule remains in effect. Furthermore, the 2013 Rule itself had little direct effect on regulated entities because it only “formalize[d] the longstanding interpretation of the Fair Housing Act to include discriminatory effects liability” and “[was] not a significant departure from HUD's interpretation to date or that of the majority of federal courts.” HUD stated further that it did not believe that additional analysis was needed on this point but invited comment.

Some commenters stated that the rule does not pass a cost benefit analysis, but they did not explain why this was so. Nor did they address HUD's explanation in the proposed rule as to why a deeper assessment of the impact of the rule was unnecessary. HUD continues to believe that this rule will provide significant benefits, while having no new impact on regulated entities, for the reasons explained earlier and summarized below.

As explained in 2013, a “uniform rule would simplify compliance with the Fair Housing Act's discriminatory effects standard, and decrease litigation associated with such claims. By providing certainty in this area to housing providers, lenders, municipalities, realtors, individuals engaged in housing transactions, and courts, this rule would reduce the burden associated with litigating discriminatory effect cases under the Fair Housing Act by clearly establishing which party has the burden of proof, and how such burdens are to be met. With a uniform standard, entities are more likely to conduct self-testing and check that their practices comply with the Fair Housing Act, thus reducing their liability and the risk of litigation. A uniform standard is also a benefit for entities operating in multiple jurisdictions. Also, legal and regulatory clarity generally serves to reduce litigation because it is clearer what each party's rights and responsibilities are, whereas lack of consistency and clarity generally serves to increase litigation. For example, once disputes around the court-defined standards are eliminated by this rule, non-meritorious cases that cannot meet the burden under § 100.500(c)(1) are likely not to be brought in the first place, and a respondent or defendant that cannot meet the burden under § 100.500(c)(2) may be more inclined to settle at the pre-litigation stage.” And as HUD explains both in this rule and the proposed rule, Inclusive Communities did not disrupt this long-standing case law or the 2013 Rule; rather, it affirmed it, citing to HUD's 2013 Rule multiple times with approval. The Court articulated long-standing limitations on the scope of disparate impact liability, which HUD had already accounted for in the 2013 Rule.

When deciding whether to enact this rule, HUD also considered whether any part of the 2020 Rule should be retained, which is evidenced by our discussion of various parts of the 2020 Rule elsewhere. It decided that no substantive portion of the 2020 Rule should be incorporated into this rule. Only three additional illustrations of discriminatory practices under the Act at § 100.70(d) are incorporated from the 2020 Rule, which were not specifically objected to by commenters and present no substantive change from the 2013 Rule. The 2020 Rule would impose significant costs to the agency, the public, and regulated entities while affording little, if any, benefit. As described in further detail elsewhere in this preamble, the 2020 Rule introduced new and confusing standards, including standards not found anywhere in case law, that were largely untested. Accordingly, the 2020 Rule would require regulated entities to spend more resources attempting to ascertain what the 2020 Rule means and how to defend against any potential claims, as well as increased spending that could last for years as courts try to interpret what the 2020 Rule means. Relatedly, entities that are covered by the Fair Housing Act have a serious reliance interest in the 2013 Rule, which has been in place for ten years. Conversely, these entities should have little to no reliance interest in the 2020 Rule, which never went into effect.

Furthermore, HUD's experience investigating and litigating discrimination cases under various regulatory frameworks informs that the 2020 Rule would make it significantly more difficult, almost impossible, to bring a discriminatory effects claim, and significantly more difficult to provide sound guidance to housing providers attempting to comply with the Act, at great cost to the agency in terms of its mission and its resources. Extra staff time would need to be spent to determine how to apply the 2020 rule to current cases being investigated and new cases that will be filed, and to determine how to address various guidance documents for the public and grantees which have been issued based on the 2013 Rule. Additionally, allowing unlawful discrimination to go unchecked and unremedied because of burdensome and confusing pleading and proof standards would come at great cost to the public who, as the Act mandates, are entitled to equal access to housing throughout the country.

Regulatory Flexibility Act

The Regulatory Flexibility Act (“RFA”) (5 U.S.C. 601 et seq.) generally requires an agency to conduct a regulatory flexibility analysis of any rule subject to notice and comment rulemaking requirements, unless the agency certifies that the rule will not have a significant economic impact on a substantial number of small entities. This rule amends the Code of Federal Regulations to accurately reflect HUD's discriminatory effects regulation as it currently exists. As a result, all entities, big and small, have a responsibility to comply with the law.

As discussed above, this Rule will continue to apply the 2013 Rule, which has been in effect uninterrupted for ten years. HUD concludes, as it did when it published the 2013 Rule, that the majority of entities, large or small, currently comply and will remain in compliance with the Fair Housing Act. All entities, large and small, have been subject to the Fair Housing Act for over fifty years and subject to the 2013 Rule for ten years. For the minority of entities that have failed to institutionalize methods to avoid engaging in illegal housing discrimination and plan to come into compliance as a result of this rulemaking, the costs will be the costs of compliance with a preexisting statute and regulation. This rule does not change substantive obligations; it merely recodifies the regulation that more accurately reflects the law. Any burden on small entities is simply incidental to the pre-existing requirements to comply with this body of law. Furthermore, HUD anticipates that this rule will eliminate confusion for all entities, including small Fair Housing Advocacy organizations, by ensuring HUD's regulations accurately reflect current standards. Accordingly, the undersigned certifies that this rule will not have a significant economic impact on a substantial number of small entities. HUD invited comment on this certification in the proposed rule. HUD did not receive any comments providing analysis of the number of small entities which commenters believe may be affected by this regulation. Some commenters stated that application of discriminatory effects law to the business of insurance would harm small businesses. HUD has responded to these comments in this rule.

Environmental Impact

This rule sets forth nondiscrimination standards. Accordingly, under 24 CFR 50.19(c)(3), this rule is categorically excluded from environmental review under the National Environmental Policy Act of 1969 (42 U.S.C. 4321).

Executive Order 13132, Federalism

Executive Order 13132 (entitled “Federalism”) prohibits an agency from publishing any rule that has federalism implications if the rule either: (i) imposes substantial direct compliance costs on state and local governments and is not required by statute, or (ii) preempts state law, unless the agency meets the consultation and funding requirements of section 6 of the Executive Order. This rule does not have federalism implications and does not impose substantial direct compliance costs on state and local governments or preempt state law within the meaning of the Executive Order.

Unfunded Mandates Reform Act

Title II of the Unfunded Mandates Reform Act of 1995 (2 U.S.C. 1531–1538) (“UMRA”) establishes requirements for federal agencies to assess the effects of their regulatory actions on state, local, and tribal governments, and on the private sector. This rule does not impose any federal mandates on any state, local, or tribal governments, or on the private sector, within the meaning of the UMRA.

List of Subjects in 24 CFR Part 100

• Aged
• Civil rights
• Fair housing
• Incorporation by reference
• Individuals with disabilities
• Mortgages, and Reporting and recordkeeping requirements

For the reasons discussed in the preamble, HUD amends 24 CFR part 100 as follows:

PART 100—DISCRIMINATORY CONDUCT UNDER THE FAIR HOUSING ACT

1. The authority citation for 24 CFR part 100 continues to read as follows:

Authority: 42 U.S.C. 3535(d), 3600–3620.

Subpart A—General

2. Amend § 100.5 by revising paragraph (b) and removing paragraph (d) to read as follows:

Subpart B—Discriminatory Housing Practices

3. In § 100.70, paragraph (d)(5) is republished to read as follows:

Subpart G—Discriminatory Effect

4. Revise § 100.500 to read as follows: