On August 20, 2019, the U.S. Department of Housing and Urban Development (HUD) published a notice of proposed rulemaking (“Proposed Rule”) seeking public comment on amendments to its regulation implementing the disparate impact theory of discrimination under the Fair Housing Act. Disparate impact does not involve intentional discrimination, instead it is a finding that a neutral policy or practice results in a disparate adverse impact on a protected class and is not supported by a business justification.
The Proposed Rule would revise the current HUD regulation to reflect the Supreme Court’s 2015 ruling in Texas Department of Housing and Community Affairs v. Inclusive Communities Project, Inc. (“Inclusive Communities”). In Inclusive Communities, the Supreme Court held that “disparate impact” discrimination is cognizable under the Fair Housing Act, but the Court went on to discuss the plaintiff’s burden to show “robust causality” between the challenged policy and a disparate impact on a protected class.
Drawing on language in Inclusive Communities, the Proposed Rule would significantly alter the burdens of plaintiffs and defendants, making it more difficult for a plaintiff to survive a motion to dismiss. Although the Proposed Rule only implements the Fair Housing Act, it raises important questions about the disparate impact doctrine under other fair lending laws, including the Equal Credit Opportunity Act. Comments are due on October 18, 2019.
The Fair Housing Act prohibits discrimination in housing based on race, color, religion, sex, disability, familial status, or national origin. In 2013, HUD promulgated a rule setting forth the requirements for a disparate impact claim under the Fair Housing Act (the “2013 Rule”). The 2013 Rule provides that disparate impact is cognizable under the Fair Housing Act and identifies the burdens of proof as follows: the plaintiff must prove that the challenged practice actually or predictably results in a disparate impact on a protected class of persons; the defendant must then prove that the practice meets a substantial, legitimate, nondiscriminatory interest. However, the plaintiff may still prevail by proving that the defendant’s goals can be achieved by another less discriminatory practice.
In Inclusive Communities, the Court acknowledged, but did not analyze, HUD’s 2013 Rule. Instead, the Court analyzed the language of the Fair Housing Act itself, and held that the Act’s prohibition on making housing “otherwise . . . unavailable” authorizes disparate impact claims. However, the Court also discussed the elements of a prima facie case, stating that a plaintiff must identify a specific policy as the cause of the disparity and establish “robust causality” between the policy and the disparity. Otherwise, the Court cautioned, defendants could be liable for discrimination they did not cause, which could lead businesses to adopt quotas and other race-based factors that raise constitutional questions.
On May 15, 2017, HUD solicited public comment on regulatory reform, seeking input on regulations that are unduly burdensome or outdated. Some commenters identified HUD’s 2013 Rule as needing revisions to more closely align the burdens of proof to the Court’s discussion in Inclusive Communities, while other commenters stated that the 2013 Rule was consistent with Inclusive Communities. In 2018, HUD published an Advance Notice of Proposed Rulemaking (ANPR) seeking public comment on whether the 2013 Rule should be amended and, if so, how. Comment was divided between those who supported the 2013 Rule and its approach to burdens of proof, including the requirement that the defendant must prove the business necessity of a challenged practice, and those who believed that the 2013 Rule’s burden-shifting approach needed revision.
HUD’s Proposed Rule
HUD now seeks comment on revisions to the 2013 Rule. HUD states that “[t]hese amendments are intended to bring HUD’s disparate impact rule into closer alignment with the analysis and guidance provided in Inclusive Communities as understood by HUD, and to codify HUD’s position that its rule is not intended to infringe on any state law for the purpose of regulating the business of insurance.”
The Proposed Rule would likely make it more difficult for a plaintiff to survive a motion to dismiss. The HUD proposal would also remove the defendant’s burden to prove a substantial business reason for the challenged policy or practice and instead would require the plaintiff to prove that the policy or practice is unnecessary to achieve a valid business objective. In addition, the Proposed Rule outlines how a defendant could defend its use of a model or other algorithm.
Prime Facie Case
The Proposed Rule would require the plaintiff to plead facts sufficient to establish five elements to make out a prima facie disparate impact case:
1. The policy or practice is “arbitrary, artificial and unnecessary to achieve a valid interest or legitimate objective,” including “a practical business, profit, policy consideration or requirement of law”;
2. A robust causal link between the policy or practice and a disparate impact on members of a protected class “which shows the specific practice is the direct cause of the discriminatory effect”;
3. The policy or practice has an adverse effect on members of a protected class “as a group”;
4. The disparity is significant; and
5. The plaintiff’s injury is directly caused by the policy or practice.
Unlike the HUD Proposed Rule, the 2013 Rule does not address the requirements for a prima facie case, and the five elements could significantly increase the plaintiff’s burden at the pleading stage. While each of the five elements can be traced to language in Inclusive Communities, the Court’s discussion of the plaintiff’s pleading requirements focused on the robust causation element.
The Proposed Rule also sets forth several ways in which a defendant can defeat a plaintiff’s attempt to state a prima facie case. First, drawing on language in Inclusive Communities, the Proposed Rule provides that a defendant could defeat the plaintiff by showing its discretion is materially limited by third parties, “such as through” federal, state or local law, or court order.
More significantly, the Proposed Rule would provide three ways in which a defendant could defeat a claim based on its use of a model. In doing so, HUD explains that “[w]hile disparate impact provides an important tool to root out factors which may cause these models to produce discriminatory outputs, these models can also be an invaluable tool in extending access to credit and other services to otherwise underserved communities.” According to HUD, the proposed defenses “would demonstrate the lack of a robust causal link between the defendant’s use of the model and the alleged disparate impact. . . .”
1. The defendant could show that the model does not “rely in any material part on factors which are substitutes or close proxies for protected classes . . . and the model is predictive of risk or other similar valid objective.” The Proposed Rule’s focus on whether factors are proxies for protected classes could make it easier for a defendant to get a case dismissed if it can show that factors are facially neutral as long as the model predicts risk.
2. The defendant could show that the model is “produced, maintained or distributed by a recognized third party that determines industry standards” and that the defendant does not determine the inputs or methods, and is using the model as intended. In doing so, the Proposed Rule explains that, essentially, a defendant could show that it is not the proper party to be charged, noting that such a defendant is not likely to have access to the inputs to be able to defend a proprietary “black box” model.
3. The defendant could show that an objective third party has reviewed and validated the model and has found that it is “empirically derived and is a demonstrable and statistically sound algorithm which accurately predicts risk or other valid objectives” and does not rely on factors that are “substitutes or close proxies” for protected classes. The Proposed Rule notes that in this case, the model is not “the actual cause of the disparate impact.”
Burdens of Proof
Under the Proposed Rule, the plaintiff would also have the burden of proving each of the elements listed by a preponderance of the evidence. The Proposed Rule’s second and fifth requirements for causality could be more difficult to satisfy than the 2013 Rule’s requirement that the plaintiff prove that the policy or practice “caused or predictably will cause a discriminatory effect.”
In contrast to the 2013 Rule, the Proposed Rule would not shift the burden of proof to the defendant to prove that the policy or practice is necessary to achieve a substantial and legitimate business goal; instead, the Proposed Rule would provide that the defendant would need to produce “evidence showing that the challenged policy or practice advances a valid interest. . . .” Thus the defendant’s burden is only to produce evidence, not to provide proof, and the business interest need only be “valid,” not “substantial.” If the defendant meets its burden, the plaintiff must prove by a preponderance of the evidence that a less discriminatory policy or practice would serve the defendant’s interest “in an equally effective manner without imposing materially greater costs on, or creating other material burdens for the defendant.”
Questions for Commenters
HUD proposes several questions for commenters, including how well its Proposed Rule aligns with the decision in Inclusive Communities. HUD also requests comment on how the Proposed Rule might increase or decrease costs and economic burdens, relative to the 2013 Rule and relative to Inclusive Communities.
In addition to HUD’s specific questions, the Proposed Rule raises questions about interpretation of other fair lending laws. In this regard, Regulation B, issued by the Consumer Financial Protection Bureau (CFPB) to implement the Equal Credit Opportunity Act (ECOA), has also been interpreted by the CFPB to recognize disparate impact discrimination. ECOA is not limited to housing credit and covers nearly all types of credit, including commercial credit. In 2018, the CFPB indicated that it intended to review the disparate impact doctrine under ECOA in light of Inclusive Communities “and the Congressional disapproval of a prior Bureau bulletin concerning indirect auto lender compliance with ECOA and its implementing regulations.” While it is unclear if the CFPB is engaged in such a review, HUD’s Proposed Rule may be a precursor to the CFPB’s approach to disparate impact under the ECOA.
 84 FR 42854 (Aug. 19, 2019).
 135 S. Ct. 2507 (2015).
 42 USC 3608(a) and 3614a.
 78 FR 11460 (Feb. 15, 2013).
 24 CFR 100.500.
 135 S. Ct. 2518.
 135 S. Ct. 2523.
 82 FR 22344 (May 15, 2017)
 83 FR 28560 (June 20, 2018).
 84 FR at 42857.
 Proposed section 100.500(b)(1).
 Proposed section 100.500(b)(2).
 Proposed 100.500(c).
 Proposed 100.500(c)(1).
 84 FR at 42859.
 84 FR at 42859.
 24 CFR 100.500(c)(1).
 Proposed 100.500(d)(1)(ii).
 Proposed 100.500(d)(1)(ii).
 12 CFR part 1002, Supp. I, Comment 6(a)-2.
 CFPB Semiannual Regulatory Agenda, Fall 2018, available at https://www.reginfo.gov/public/jsp/eAgenda/StaticContent/201810/Preamble_3170.html