The Fair Housing Act1×1. 42 U.S.C. §§ 3601–3619 (2012). (FHA) allows plaintiffs to bring suits for explicit discrimination as well as for facially neutral policies that have discriminatory effects.2×2. See Tex. Dep’t of Hous. & Cmty. Affairs v. Inclusive Cmtys. Project, Inc. (ICP I), 135 S. Ct. 2507, 2523 (2015). Discriminatory effect liability includes both the traditional disparate impact theory, which addresses adverse impacts of housing policies on a protected group, and segregative-effect claims, which allow suit when a policy perpetuates segregation.3×3. Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 937 (2d Cir. 1988). In 2015, the Supreme Court held in Texas Department of Housing & Community Affairs v. Inclusive Communities Project, Inc.4×4. 135 S. Ct. 2507. (ICP I) that disparate impact claims are cognizable under the FHA, but “barely mentioned” segregative-effect claims.5×5. Robert G. Schwemm, Segregative-Effect Claims Under the Fair Housing Act, 20 N.Y.U. J. Legis. & Pub. Pol’y 709, 714 (2017). Recognized by some courts since the 1970s, segregative-effect claims were first formally codified through a 2013 U.S. Department of Housing and Urban Development (HUD) regulation.6×6. 24 C.F.R. § 100.500(a) (2019) (stating that a facially neutral practice is illegal if it “results in a disparate impact on a group of persons or creates, increases, reinforces, or perpetuates segregated housing patterns” (emphasis added)). Segregative-effect claims require (1) segregated housing patterns in the area and (2) a practice that “creates, increases, reinforces, or perpetuates” that segregation.7×7. Id. They can succeed even if the connection between the defendant’s act and the segregative effect is more remote than would be required for disparate impact claims.8×8. Schwemm, supra note 5, at 757. Given the ICP I Court’s emphasis on plaintiffs showing “robust causality” between a defendant’s act and the discriminatory effect to establish a prima facie case,9×9. ICP I, 135 S. Ct. at 2523. the segregative-effect theory could face questions. Recently, in Inclusive Communities Project, Inc. v. Lincoln Property Company10×10. 920 F.3d 890 (5th Cir. 2019). (ICP II), a Fifth Circuit panel affirmed the district court’s dismissal of an FHA claim that relied on both disparate impact and segregative-effect theories against apartment owners for not participating in the Section 8 voucher program in white areas.11×11. Id. at 895, 909. It affirmed the dismissal of the segregative-effect claim on the ground that plaintiffs must satisfy a higher pleading standard against private defendants than against the government.12×12. See id. at 908. Although the panel was rightly concerned that HUD’s liberal construction of the segregative-effect theory could lead to widespread litigation, the panel’s approach had no basis in the statutory text. The court should have instead dismissed the segregative-effect claim because the defendants’ policy was not significantly perpetuating segregation, which would limit the theory on firmer doctrinal ground.
The Inclusive Communities Project, Inc. (ICP), is a nonprofit that “attempts to remedy harmful practices that perpetuate discrimination and segregation.”13×13. Inclusive Cmtys. Project, Inc. v. Lincoln Prop. Co., No. 17-CV-206, 2017 WL 3498335, at *2 (N.D. Tex. Aug. 16, 2017). Specifically, ICP helps its clients who receive Section 8 housing vouchers to “move to higher opportunity communities” by, for example, providing incentive payments to landlords.14×14. Id. at *4. While participation in Section 8 is voluntary for landlords,15×15. 42 U.S.C. § 1437f(o)(6)(B) (2012); Tex. Local Gov’t Code § 250.007 (2019). the owners of several residential properties in Dallas, Texas, and their property manager, Lincoln Property Company, “implemented a policy of refusing to rent to Section 8 voucher holders for 54 . . . apartment complexes” in majority-white areas.16×16. Inclusive Cmtys., 2017 WL 3498335, at *3. ICP alleged that they “do not have this policy in majority minority areas.”17×17. Id. at *1.
ICP brought a suit alleging violations of the FHA by Lincoln and the property owners.18×18. Id. ICP claimed, first, that the defendants’ policy of not accepting vouchers “causes a racially discriminatory effect.”19×19. Id. at *2 (emphasis added). Specifically, ICP claimed that the policy both “causes a disproportionate harm to Black households”20×20. Complaint at 22, Inclusive Cmtys., 2017 WL 3498335 (No. 17-cv-00206). and “perpetuates racial segregation.”21×21. Id. at 20. Second, ICP argued the defendants violated the FHA’s discriminatory treatment standard by “refusing to negotiate with or enter into leases with [it] based on the race and color of [the] Section 8 voucher clients.”22×22. Inclusive Cmtys., 2017 WL 3498335, at *2. Additionally, ICP alleged that Lincoln’s advertisements featuring the no-voucher policy violated the FHA’s prohibition on discriminatory advertising. Id. (citing 42 U.S.C. § 3604(c) (2012)). ICP also alleged that Lincoln’s refusal to rent to Section 8 voucher holders in white areas, while not maintaining that policy in less white areas, violated the disparate treatment standard. ICP II, 920 F.3d at 898.
The district court granted the defendants’ motions to dismiss.23×23. Inclusive Cmtys., 2017 WL 3498335, at *12–13. Regarding the discriminatory treatment claims, the court determined that they were “mislabeled” because the allegation of a colorblind policy disproportionately hurting minorities is properly a disparate impact claim.24×24. Id. at *6. The court then dismissed the discriminatory effect claims.25×25. See id. at *9. Under the HUD regulation’s burden-shifting approach, a plaintiff must first establish a prima facie case that shows a “causal connection between the defendants’ policy and a disparate impact.”26×26. Id. at *7 (citing ICP I, 135 S. Ct. 2507, 2524 (2015)). While the court determined that the refusal to rent to voucher holders was a “policy” that could be scrutinized, it also found that ICP did not satisfy the “robust causality” test announced in ICP I because it did not detail how the policy either reduced opportunities for Black renters or “caused the racial disparity” across census tracts.27×27. See id. at *8–9. Even if ICP had met its prima facie burden, Lincoln could argue at the second stage of the burden-shifting framework that the policy was necessary to achieve “substantial, legitimate, nondiscriminatory interests.”28×28. 24 C.F.R. § 100.500(c)(2) (2019). The court believed this was met because of “[b]usiness concerns such as increased costs, administrative delays . . . , and other financial risks.”29×29. Inclusive Cmtys., 2017 WL 3498335, at *10. The framework would then allow the plaintiff to prove that the legitimate interest “could be served by another practice that has a less discriminatory effect.”30×30. 24 C.F.R. § 100.500(c)(3). Even if ICP had met its prima facie burden, it would fail at this stage, according to the court, because the alternatives offered, such as incentive payments, were not proven to address those legitimate concerns.31×31. Inclusive Cmtys., 2017 WL 3498335, at *10. Finally, the court dismissed the § 3604(c) claims because the advertisements would not suggest to “an ordinary reader that African American . . . rental applicants are discouraged.” Id. at *13.
The Fifth Circuit affirmed.32×32. ICP II, 920 F.3d at 895. Writing for the panel, Judge Engelhardt33×33. Judge Engelhardt was joined by Judge Jones. explained that the Supreme Court in ICP I had added “safeguards to incorporate into the burden-shifting framework,”34×34. ICP II, 920 F.3d at 902 (internal quotation marks omitted) (quoting Crossroads Residents Organized for Stable & Secure Residencies v. MSP Crossroads Apartments LLC, No. 16-233, 2016 WL 3661146, at *6 (D. Minn. July 5, 2016)). including a “robust causality” requirement at the prima facie stage.35×35. Id. He noted that the Court left “robust causation” undefined with respect to disparate impact claims, leading to conflicting circuit court interpretations.36×36. Id. at 903–04. One circuit’s definition requires a policy causing a disparity to also be arbitrary and unnecessary.37×37. Id. at 904 (citing Ellis v. City of Minneapolis, 860 F.3d 1106, 1112–14 (8th Cir. 2017)). The ICP II court held that refusal to participate in the voluntary Section 8 program could not be arbitrary or unnecessary unless there was actual evidence of arbitrariness, which was not present here.38×38. Id. at 907. A dissenting judge from another circuit would have required the policy itself to cause the racial disparities, as distinct from disparities resulting from a neutral policy applied against general, unequal “pre-existing conditions.”39×39. Id. at 905 (quoting Reyes v. Waples Mobile Home Park Ltd. P’ship, 903 F.3d 415, 434–35 (4th Cir. 2018) (Keenan, J., dissenting)). There would have been no claim under this definition due to preexisting segregation in the community. Under a third definition, in which any policy that affects protected classes more than others meets the standard, the defendants’ no-voucher policy would give rise to a claim.40×40. Id. at 906 (citing Reyes, 903 F.3d at 425 (majority opinion)). However, the panel concluded that such a broad definition would be inconsistent with ICP I.41×41. Id. at 906–07.
The majority also affirmed the dismissal of the segregative-effect claims. The panel distinguished this case from Huntington Branch, NAACP v. Town of Huntington,42×42. 844 F.2d 926 (2d Cir. 1988). which involved “indefensible government policies that operated to perpetuate segregation by unreasonably restricting private construction.”43×43. ICP II, 920 F.3d at 908. The ICP II majority feared that imposing “affirmative housing obligations”44×44. Id. on private actors would “be contrary to the cautionary standards”45×45. Id. at 909. of ICP I. Notably, the panel conceded that Huntington Branch considered a claim past the prima facie stage, but reasoned that segregative-effect claims against a private defendant required “a heavier pleading burden.”46×46. Id. at 909 n.11. Additionally, the majority disagreed that the disparate treatment claims were “mislabeled,” but instead dismissed them for being “vague and conclusory.” Id. at 910–11. Similarly, the disparate treatment claim against Lincoln was dismissed because the allegation was “unclear.” Id. at 911–12. Finally, the majority affirmed the dismissal of the advertising claim. Id. at 912.
Judge Davis concurred in part and dissented in part. He dissented from the dismissal of both types of discriminatory effect claims.47×47. Id. at 913 (Davis, J., concurring in part and dissenting in part). He concurred in the dismissal of the disparate treatment and advertising claims. Id. at 912–13. In recounting the Supreme Court’s disparate impact rulings, Judge Davis argued that “robust causation” required only identifying a specific policy causing the effect, and that ICP “clearly challenges a policy . . . of not renting to voucher holders.”48×48. Id. at 919. He argued that this policy “causes the exclusion of the predominantly Black voucher population”49×49. Id. from the buildings, so the disparate impact claim should not have been dismissed.50×50. See id. at 920. He also noted that a segregative-effect claim only requires a plaintiff to show “existing patterns of segregation and that the challenged policy will perpetuate” it.51×51. Id. at 922. Judge Davis then accused the majority of requiring plaintiffs to prove “that the policy caused the initial segregation.”52×52. Id. Additionally, he was not persuaded that the defendants should be protected by the voluntary nature of the Section 8 program because discrimination based on voucher status is “not listed as an exemption” in the FHA.53×53. Id. at 924. The Fifth Circuit denied a petition for rehearing en banc.54×54. Inclusive Cmtys. Project, Inc. v. Lincoln Prop. Co., 930 F.3d 660, 660 (5th Cir. 2019).
As the circuit split on the definition of “robust causality” shows, the Court’s holding in ICP I has proven difficult for lower courts to apply consistently.55×55. Claire Williams, Note, Inclusive Communities and Robust Causality: The Constant Struggle to Balance Access to the Courts with Protection for Defendants, 102 Minn. L. Rev. 969, 989 (2017) (noting ICP I “failed to provide sufficient guidance to lower courts” and “created confusion”). While Justice Kennedy’s majority opinion seemingly affirmed the viability of segregative-effect claims,56×56. Schwemm, supra note 5, at 714 & n.26. its emphasis on “robust causality” calls into question whether the segregative-effect claim’s more lax causation requirements remain acceptable. In this doctrinal uncertainty, courts face a dilemma: the HUD regulation, if applied literally, threatens potentially widespread litigation for property owners in segregated areas; but requiring that a defendant had caused the preexisting segregation might effectively preclude all liability. The Fifth Circuit’s response to this puzzle in ICP II — heightened pleading standards for private parties — has no basis in the statutory text. Instead, the Fifth Circuit should have addressed these concerns by requiring at the prima facie stage that the challenged decision significantly perpetuates segregation. This solution would more faithfully interpret the FHA and ICP I decision and would build on a test used in some earlier cases.
Under a literal reading of the segregative-effect doctrine, any action of a property owner in a segregated area that results in merely a reflection of existing residential patterns would meet the prima facie standard for a segregative-effect claim.57×57. The HUD regulation makes liable a practice that “creates, increases, reinforces, or perpetuates” segregation. 24 C.F.R. § 100.500(a) (2019) (emphasis added). See also Schwemm, supra note 5, at 743 n.183 (describing HUD’s general acceptance of this broad reading). Indeed, the HUD regulation codifying the doctrine establishes no minimum for how significant the effect on segregation must be for a viable claim.58×58. Schwemm, supra note 5, at 743 n.183. In this case, if the segregative-effect claim succeeded, it would “effectively mandate a landlord’s participation in the voucher program,”59×59. ICP II, 920 F.3d at 909. despite Congress’s choice to make it voluntary.60×60. See, e.g., Salute v. Stratford Greens Garden Apartments, 136 F.3d 293, 300 (2d Cir. 1998). Yet the theory could go further: any housing provider in a segregated area who declined any minority applicant could face a segregative-effect claim that could get past the prima facie stage.61×61. Schwemm, supra note 5, at 762–64. “The only apparent difficulty at the prima-facie-case stage would be showing that the defendant’s action caused a substantial effect on local segregation,” assuming such a requirement existed. Id. at 763. While such a low prima facie standard could rely on the burden-shifting framework to avoid actual liability for status quo action,62×62. See id. at 742 n.179. as Justice Alito worried, “the costs of litigation . . . may push cost-conscious defendants to settle,” and risk-averse parties “may let race drive their decisionmaking in hopes of avoiding litigation altogether.”63×63. ICP I, 135 S. Ct. 2507, 2549–50 (2015) (Alito, J., dissenting) (first quoting Bell Atl. Corp. v. Twombly, 550 U.S. 544, 559 (2007); and then citing Ricci v. DeStefano, 557 U.S. 557, 563 (2009)).
A significance requirement is not generally part of segregative-effect doctrine, but it would likely avoid these problems. While disparate impact claims must show a “substantial” disparity, most segregative-effect cases do not allude to any similar requirement.64×64. Schwemm, supra note 5, at 742. Only a few cases have cited a significance requirement. See Davis v. N.Y.C. Hous. Auth., 166 F.3d 432, 438 (2d Cir. 1999) (“The proper standard . . . is whether the [policy] will significantly perpetuate segregation . . . .”); Metro. Hous. Dev. Corp. v. Village of Arlington Heights (Arlington Heights II), 558 F.2d 1283, 1291 (7th Cir. 1977); Ave. 6E Invs., LLC v. City of Yuma, No. 09-cv-00297, 2013 WL 2455928, at *7 (D. Ariz. June 5, 2013), rev’d on other grounds, 818 F.3d 493 (9th Cir. 2016) (rejecting a segregative-effect claim because the effect was not “significant enough”). However, as the Second Circuit noted when it did apply a significance requirement in a segregative-effect case, Huntington Branch hints at such a test,65×65. Davis, 166 F.3d at 438 (citing Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 938 (2d Cir. 1988)). as does the Supreme Court’s language in a school desegregation case.66×66. Id. (citing Gilmore v. City of Montgomery, 417 U.S. 556, 568 (1974)). Embracing this shift would resolve the policy problem of unlimited litigation for property owners undertaking status quo behaviors that have only a de minimis effect on broader patterns. For example, in ICP II, the plaintiffs hypothesized that “3 to 5” units at each complex would be affected,67×67. Brief for Appellant at 38, ICP II, 920 F.3d 890 (5th Cir. 2019) (No. 17-10943). which might not satisfy a significance requirement when “hundreds of other multifamily complexes” in the area also do not accept vouchers.68×68. Id. at 15.
A significance requirement would also align segregative-effect doctrine with the ICP I Court’s emphasis on establishing “robust causality.” The emphasis on causation is a challenge for segregative-effect claims, which have “required much less in the way of proof of a causal connection between the defendant’s action and continuing segregation.”69×69. Schwemm, supra note 5, at 757. Successful segregative-effect claims will need more than the ICP II dissent’s test of “identify[ing] an offending policy,”70×70. ICP II, 920 F.3d at 921 (Davis, J., concurring in part and dissenting in part). which appears inconsistent with ICP I’s statement that defendants are protected “from being held liable for racial disparities they did not create.”71×71. ICP I, 135 S. Ct. 2507, 2523 (2015) (citing Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 653 (1989)). To be sure, some successful segregative-effect plaintiffs could show robust causation. For example, Lincoln distinguished Mhany Management, Inc. v. County of Nassau,72×72. 819 F.3d 581 (2d Cir. 2016). in which a developer successfully challenged a zoning policy preventing low-income housing, because the policies in question “prevented construction that would have . . . added as many as 101 minority households” and had limited affordable housing for years.73×73. Brief of Appellees at 19, ICP II, 920 F.3d 890 (5th Cir. 2019) (No. 17-10943).
In fact, a significance requirement is an appropriate translation of “robust causality” from the traditional disparate impact context into the context of segregation. As the distinction between this case and Mhany suggests, whether the segregative effect is “significant” is connected to the idea of “causation.” If a policy is preventing racial minorities from moving to a white area in significant numbers, it can be fairly described as a “cause” of continued segregation in a way that policies of a smaller scale cannot be. This intuition is made more clear by comparing a significance requirement with one definition of “robust causality” discussed by the ICP II panel — that a policy must cause the “pre-existing conditions” that lead to the effect.74×74. ICP II, 920 F.3d at 905 (discussing definition articulated by the dissent in Reyes v. Waples Mobile Home Park Ltd. Partnership, 903 F.3d 415 (4th Cir. 2018)). Since “pre-existing” segregation often occurred generations ago due to concerted effort,75×75. See generally Richard Rothstein, The Color of Law: A Forgotten History of How Our Government Segregated America (2017). such a definition would be inapt for segregative-effect claims. Clarifying the existence of a significance requirement would provide the safeguards that ICP I requires without resorting to a cramped definition of robust causation.
A significance requirement is on firmer ground than the panel’s instinct to “impose a heavier pleading burden on ICP’s efforts to require private defendants to take . . . affirmative action,”76×76. ICP II, 920 F.3d at 909 n.11. which has no statutory basis. To be sure, a segregative-effect claim has never succeeded against a private defendant.77×77. Schwemm, supra note 5, at 749. Landmark segregative-effect cases distinguished reversing government policies from compelling “affirmative steps” by a private actor,78×78. Arlington Heights II, 558 F.2d 1283, 1287 (7th Cir. 1977); see also Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 940 (2d Cir. 1988). which would be a “massive judicial intrusion on private autonomy.”79×79. Arlington Heights II, 558 F.2d at 1293. However, the FHA neither expressly nor implicitly exempts private actors from liability for policies with segregative effects.80×80. See Maia Hutt, Note, This House Is Not Your Home: Litigating Landlord Rejections of Housing Choice Vouchers Under the Fair Housing Act, 51 Colum. J.L. & Soc. Probs. 391, 421 (2018) (noting the FHA is “very specific in delineating exceptions”). Indeed, the statute’s definition of “discriminatory housing practice” makes no distinction between public and private actors.81×81. See 42 U.S.C. § 3602(f) (2012).
Importantly, the significance requirement also does not depend on the force of the Obama-era HUD regulation. This matters because, first, the ICP I decision itself may have revised the 2013 regulation. Despite the United States claiming that the regulation was entitled to Chevron deference,82×82. Brief for the United States as Amicus Curiae Supporting Respondent at 15, ICP I, 135 S. Ct. 2507 (2015) (No. 13-1371). the Court’s opinion in ICP I was silent on the question.83×83. Given the Court’s emphasis on not construing disparate-impact liability so broadly as to create “serious constitutional concerns,” ICP I, 135 S. Ct. at 2523, the decision not to extend deference might reflect the application of the constitutional avoidance canon before deferring, see Edward J. DeBartolo Corp. v. Fla. Gulf Coast Bldg. & Constr. Trades Council, 485 U.S. 568, 575 (1988). As some scholars have noted, it seems unlikely that the Court’s interpretation, which relied largely on the purpose and legislative history of the FHA, reflects a “clear” reading of the FHA.84×84. One scholar describes the opinion as avoiding the “better reading of the statute” by relying on “shaky” textual analysis likely because of the “unusually high stakes” of the case. Ryan D. Doerfler, High-Stakes Interpretation, 116 Mich. L. Rev. 523, 567 (2018); see also Robert G. Schwemm, Fair Housing Litigation After Inclusive Communities: What’s New and What’s Not, 115 Colum. L. Rev. Sidebar 106, 110 (2015) (“Justice Kennedy seemed to rely most heavily on the need for an expansive reading of the FHA to help accomplish its goal . . . .”); Alison Tanner, Note, Live and Learn: Using the Fair Housing Act to Advance Educational Opportunity for Parenting Students, 105 Geo. L.J. 1453, 1471–72 (2017) (noting the dissent’s textual argument was “not directly answered by the majority,” id. at 1472). However, the Court arguably could have found the statute ambiguous, deferred to HUD, and still upheld disparate-impact liability with fewer interpretive strains. The fact that the Court did not do so, as the ICP II panel implied in a footnote,85×85. ICP II, 920 F.3d at 903 n.6. suggests that the Court intended to set aside the regulation in favor of an approach with “adequate safeguards.”86×86. ICP I, 135 S. Ct. at 2523. Second, the regulation seems unlikely to be in place for much longer. Specifically, HUD has proposed a rule deleting the codification of the segregative-effect claim.87×87. HUD’s Implementation of the Fair Housing Act’s Disparate Impact Standard, 84 Fed. Reg. 42,854, 42,862 (proposed Aug. 19, 2019) (to be codified at 24 C.F.R. pt. 100).
With the Justice who wrote the enigmatic ICP I decision for a five-vote majority no longer on the Court, “[w]hether HUD’s recognition of this . . . theory will survive after ICP [I] is an open question.”88×88. Schwemm, supra note 84, at 122. A significance requirement presents a way forward for the segregative-effect theory consistent with ICP I, while avoiding the practical problem of excessive litigation risk.