Opinion
13 C 8564
03-26-2024
PROPERTY CASUALTY INSURERS ASSOCIATION OF AMERICA, Plaintiff, v. ADRIANNE TODMAN, in her official capacity as Acting Secretary of Housing and Urban Development,[1] and the UNITED STATES DEPARTMENT OF HOUSING AND URBAN DEVELOPMENT, Defendants.
MEMORANDUM OPINION AND ORDER
REBECCA R. PALLMEYER, UNITED STATES DISTRICT JUDGE
This case addresses the U.S. Department of Housing and Urban Development (“HUD”)'s answer to a longstanding and difficult question: whether providers of homeowners insurance should face liability under the Fair Housing Act (“FHA”) for making decisions based on risk that disparately impact people of color and other protected classes. The issue has been unresolved in the courts. In 2013, HUD tackled it head-on by refusing to exempt risk-based insurance practices from its new Disparate-Impact Rule, which established a uniform legal framework for FHA discriminatory effects claims. Plaintiff Property Casualty Insurers Association of America (“PCI”) sued HUD over this decision, alleging on behalf of its members in the insurance trade that the agency's refusal to recognize an exemption was both arbitrary and capricious and contrary to law. Judge Amy St. Eve, then of this court, dismissed several of PCI's claims in 2014, but agreed that HUD had inadequately considered the issues at stake and ordered further analysis.
PCI has since merged with another insurance trade association and changed its name to the American Property Casualty Insurance Association, but continues to refer to itself as “PCI” in briefing for the sake of clarity. (See Mem. in Supp. of Pl.'s Renewed Mot. for Summ. J. [283] (hereinafter “Pl.'s Br.”) at 4 n.4.) The court will do the same.
Now, after almost a decade of political uncertainty surrounding the Rule, the matter is again before the court. HUD reinstated its 2013 Rule in substantively identical form last year; PCI argues that HUD's decision to do so remains arbitrary and capricious for the same reasons it raised earlier and more. As explained in detail below, however, HUD has supplied what was missing before: a thorough and well-reasoned explanation for its decision to allow disparateimpact claims against risk-based insurance practices under the FHA. The court therefore denies PCI's motion for summary judgment and grants HUD's cross-motion.
BACKGROUND
PCI's challenge to HUD's Rule is the latest salvo in a legal debate stretching back almost to the FHA's enactment. Much of this story was told in Judge St. Eve's 2014 decision, and the court will not repeat it in full here. See Prop. Cas. Insurers Ass'n of Am. v. Donovan, 66 F.Supp.3d 1018, 1024-34 (N.D. Ill. 2014) (hereinafter “ PCI”). But the court presents the historical context for today's decision.
I. Pre-2014 Background
The central dispute here involves risk-based pricing and underwriting of insurance, or the processes for setting rates and making coverage decisions based on an estimation of the insured party's risk of future loss. See Steven Plitt et al., Couch on Insurance §§ 1:9, 69:7, 69:14, 101:2 (3d ed.). Homeowners insurers do this by using a host of actuarial factors correlated with past losses, such as a house's construction materials, its proximity to fire hydrants, and whether it has a security system. (PCI's Statement of Facts [283-1] (hereinafter “Pl.'s SOF”) ¶ 9.) Insurers depend on risk classification to set premiums that accurately reflect the cost of providing insurance, and in turn, to remain solvent in a competitive marketplace. (Id. ¶¶ 96-97.) Not all of an insurer's practices, however, are risk-based: many other aspects of the business of insurance, such as marketing, claims processing, and payment, do not involve actuarial decision-making. See Stephen M. Dane, Race Discrimination is not Risk Discrimination: Why Disparate Impact Analysis of Homeowners Insurance Practices Is Here to Stay, Banking & Fin. Servs. Pol'y Rep., June 2014, at 4 (Admin. Record [317] (hereinafter “R.”) 004407); Huskey v. State Farm Fire & Cas. Co., No. 22 C 7014, 2023 WL 5848164, at *11 (N.D. Ill. Sept. 11, 2023).
State law regulates the insurance industry, including the process by which insurers calculate their rates. PCI, 66 F.Supp.3d at 1024-25. All states permit insurers to use risk-based pricing and underwriting methods, and many require them to do so. (Pl.'s SOF ¶ 94; R. 02915762.) In other words, states allow (or mandate) insurers to “fairly” discriminate by treating people with similar risks similarly and different risks differently. But “unfair discrimination”-charging different rates to people in the same risk group, with no basis in cost for the differential-is widely proscribed under state law. See Plitt et al., supra, §§ 69:39, 69:40, 69:44. States vary in the degree to which they require insurers to seek prior approval of proposed rates by filing them in advance with the state's insurance commissioner or other regulatory body. Id. § 2:31; see Angelo Borselli, Insurance Rates Regulation in Comparison with Open Competition, 18 Conn. Ins. L.J. 109, 126 (2011).
The federal McCarran-Ferguson Act, 15 U.S.C. §§ 1011-15, confirms the states' primacy over insurance regulation. Passed in 1945, McCarran-Ferguson prohibits federal laws of general application from being “construed to invalidate, impair, or supersede any law enacted by any State for the purpose of regulating the business of insurance,” meaning that Congress must speak clearly if it wants to intervene in this area. 15 U.S.C. § 1012(b). McCarran-Ferguson thus creates a form of “reverse preemption” that prevents enforcement of otherwise-applicable federal statutes in such a way as to interfere with states' regulatory insurance regimes. The Act is complemented by the common-law “filed-rate doctrine,” which forbids courts from revising rates that have been filed with and approved by a regulatory agency, at least in states where this is required. See Cohen v. Am. Sec. Ins. Co., 735 F.3d 601, 607 (7th Cir. 2013).
In certain ways, however, these principles may collide with another important federal policy: ensuring fair access to housing through the United States. The Fair Housing Act, enacted in 1968, makes it unlawful to “make unavailable or deny” housing, or to discriminate “in [its] terms, conditions, or privileges . . . or in the provision of services or facilities in connection therewith,” based on a person's race, gender, or other protected characteristics. 42 U.S.C. § 3604(a), (b). The Act's passage coincided with a growing understanding of homeowners insurance as an essential prerequisite to housing opportunity.Mortgage companies will typically not provide financing to borrowers who lack insurance coverage; thus, as the Seventh Circuit has recognized, “[n]o insurance, no loan; no loan, no house; lack of insurance thus makes housing unavailable.” NAACP v. Am. Family Mut. Ins. Co., 978 F.2d 287, 298 (7th Cir. 1992). HUD has long maintained that insurers' denial of this vital coverage to communities of color-a practice known as “insurance redlining”-has contributed to the problem of racial disparities in homeownership.Holding insurers liable under the FHA could change that. But whether and to what extent the FHA can properly be applied to insurers has been hotly contested for decades-both because of the federalism principles embodied in McCarran-Ferguson, and because of the fundamental nature of insurance itself.
See President's Nat'l Advisory Panel on Ins. in Riot-Affected Areas, Meeting the Insurance Crisis of Our Cities 1 (1968) (R. 032825) (“Without insurance, banks and other financial institutions will not-and cannot-make loans . . . housing cannot be constructed . . . [and] [e]fforts to rebuild our nation's inner cities cannot move forward. Communities without insurance are communities without hope.”); Dunn v. Midwestern Indem., Mid-Am. Fire & Cas. Co., 472 F.Supp. 1106, 1109-12 (S.D. Ohio 1979).
See Dunn, 472 F.Supp. at 1109 & nn. 7-8 (citing 1978 memorandum from HUD's General Counsel); Homeowners' Insurance Discrimination: Hearings before the S. Comm. on Banking, Housing and Urban Affairs, 103 Cong. 50 (1994) (statement of HUD Assistant Sec'y Roberta Achtenberg).
Since early in the FHA's history, both HUD and lower federal courts have interpreted its language to encompass disparate-impact as well as disparate-treatment theories of liability. See PCI, 66 F.Supp.3d at 1026 & n.1 (noting that eleven circuits, though not the Supreme Court or the D.C. Circuit, had so held as of 2014). In other words, FHA defendants could be liable for conduct that, while facially neutral, had the unjustified effect of denying housing to members of a protected class. This theory, however, is in tension with the very concept of risk-based insurance decision-making, which differentiates between groups of people based on objective risk factors- factors that can themselves strongly correlate with protected-class membership. See Am. Family, 978 F.2d at 290 (“Risk discrimination is not race discrimination.”). This could mean that purely risk-based insurance decisions that have a disparate impact on a protected class might violate the law. And even attempting to resolve this tension could require federal courts-rather than state insurance commissioners-to scrutinize insurers' ratemaking decisions, raising concerns under McCarran-Ferguson. In a case where plaintiffs challenged insurance policy caps as violative of the Americans with Disabilities Act, the Seventh Circuit reversed a ruling in plaintiffs' favor; the court held that “requir[ing] federal courts to determine whether limitations on coverage are actuarially sound and consistent with state law . . . obviously would interfere with the administration of the state law,” and that “[t]he states are not indifferent to who enforces their laws.” Doe v. Mut. of Omaha, 179 F.3d 557, 563-64 (7th Cir. 1999).
The question of whether insurers' risk-based practices could be subjected to disparateimpact liability under the FHA remained unsettled for many years. While HUD consistently sought to extend the FHA to the insurance industry, and passed a regulation in 1989 that swept insurers within the Act's scope, see 54 Fed.Reg. 3,232, 3,285 (Jan. 23, 1989) (codified at 24 C.F.R. § 100.70(d)(4)), the circuits split over whether insurers could be subject to both disparate-treatment and disparate-impact FHA liability, both, or neither.And while HUD took the position in Congressional hearings that its regulation contemplated liability for insurers under both theories, see Homeowners' Insurance Discrimination: Hearings Before the S. Comm. on Banking, Housing and Urban Affairs, 103 Cong. 19 (1994) (statement of HUD Assistant Sec'y Roberta Achtenberg), it never formally codified that position. See Gregory D. Squires, Why An Insurance Regulation to Prohibit Redlining?, 31 John Marshall L. Rev. 489, 499, 503-04 (1998) (describing failed attempts to promulgate a more detailed insurance regulation that might address this question); Saunders v. Farmers Ins. Exch. (“Saunders II”), 537 F.3d 961, 964 n.3 (8th Cir. 2008) (noting that “HUD ha[d] never applied a disparate impact analysis to insurers” as of 2008) (quoting Nationwide Mut. Ins. Co. v. Cisneros, 52 F.3d 1351, 1362 (6th Cir. 1995)).
In 1984, the Fourth Circuit interpreted the FHA's text to exclude insurers entirely- in part due to the concerns over state authority and the business of insurance that a contrary reading would raise. Mackey v. Nationwide Ins. Companies, 724 F.2d 419, 423 (4th Cir. 1984). After HUD took a contrary position in its 1989 regulation, two circuits (the Seventh and Sixth) partially endorsed the agency's interpretation of the statute as to disparate-treatment claims against insurers, but declined to rule on disparate-impact claims. Am. Family, 979 F.2d at 29091, 301; Nationwide Mut. Ins. Co. v. Cisneros, 52 F.3d 1351, 1359-60, 1363 (6th Cir. 1995). The Eighth Circuit also declined to decide whether the FHA permitted disparate-impact claims against insurers' risk-based practices, but held that if the FHA did allow for such a claim, it would be preempted under the McCarran-Ferguson Act in the circumstances of that case for interfering with Missouri's scheme of insurance regulation. Saunders v. Farmers Ins. Exch., 537 F.3d 961, 964-67 (8th Cir. 2008). In contrast, the Fifth Circuit held that a disparate-impact claim against an insurer was not preempted under McCarran-Ferguson, but its holding depended on a conclusion that the state insurance laws at issue did not specifically address the challenged practice. DeHoyos v. Allstate Corp., 345 F.3d 290, 298-99 & n.5 (5th Cir. 2003). District courts in several other jurisdictions reached similar results. See, e.g., Lumpkin v. Farmers Grp., Inc., No. 05-2868 MA/V, 2007 WL 6996777, at *7 (W.D. Tenn. July 6, 2007); cf. Nat'l Fair Hous. All., Inc. v. Prudential Ins. Co. of Am., 208 F.Supp.2d. 46, 61 (D.D.C. 2002) (rejecting an insurer's related argument framed in terms of states' “primary jurisdiction” over insurance regulation). The Ninth Circuit held that, in general, disparate-impact FHA claims are cognizable against insurers and do not run afoul of McCarran-Ferguson if they “would complement-rather than displace and impair-[state] law” regulating the insurance practice at issue. Ojo v. Farmers Grp., 600 F.3d 1205, 1208, 1210 (9th Cir. 2010) (en banc). However, the Ninth Circuit then certified the specific question of whether Texas's insurance code permitted the challenged practice at issue to that state's supreme court, which confirmed that it did, meaning that the FHA challenge was preempted by McCarran-Ferguson. Ojo v. Farmers Grp., 356 S.W.3d 421, 422 (Tex. 2011). In a related but distinct context, where Black plaintiffs alleged that the defendant insurer targeted and sold them life insurance policies with higher premiums and lower benefits than those sold to white purchasers in violation of 42 U.S.C. §§ 1981 and 1982, the Eleventh Circuit concluded those claims were not preempted under the McCarran-Ferguson Act. Moore v. Liberty Nat. Life Ins. Co., 267 F.3d 1209, 1222-23 (11th Cir. 2001).
II. The 2013 Rule
That changed in 2012, when HUD announced its intent to promulgate a new standardized framework for evaluating disparate-impact claims under the FHA. This framework, based on analogous caselaw from other civil-rights statutes, involved the three-step burden-shifting approach for evaluating disparate-impact claims that was already in use by many of the courts of appeals.Commenters from the insurance industry, including PCI, quickly weighed in on HUD's rulemaking, airing the same basic concerns about applying disparate-impact liability to homeowners insurance that had been raised in court for decades. They requested that HUD exempt either risk-based insurance practices specifically, or insurance altogether, from the scope of its new Disparate-Impact Rule. (See R. 000371-83, 000454-59, 000552-56.)
Specifically, the proposed Rule required FHA disparate-impact plaintiffs to prove at the first step that a challenged housing practice has a disparate impact (or perpetuates segregation) with regard to a protected class. At the second step, the defendant would have the burden of showing that the practice was supported by legitimate and nondiscriminatory interests. At the third step, the plaintiff could show that a less-discriminatory alternative was available to serve the same interests. See Implementation of the Fair Housing Act's Discriminatory Effects Standard: Proposed Rule, 76 Fed.Reg. 70921, 70923-24 (Nov. 16, 2011).
HUD rebuffed these concerns when it issued its final Rule in early 2013. See Implementation of the Fair Housing Act's Discriminatory Effects Standard: Final Rule, 78 Fed.Reg. 11,460 (Feb. 15, 2013). Its justification for doing so, however, consumed just half of a single page in the Federal Register. HUD noted that “courts have agreed with HUD[‘s]” longstanding interpretation of the Act as applicable to insurers, citing the Ninth Circuit's decision in Ojo v. Farmers Group, 600 F.3d 1205 (9th Cir. 2010) (en banc) that McCarran-Ferguson did not per se preempt disparate-impact FHA claims, as well as the Seventh's in American Family and the Sixth's in Nationwide. Id. at 11,475. HUD did not, however, discuss the reasoning in these cases or explain in detail why they supported its position-particularly for the Ojo litigation, which ultimately ended in a finding by the Texas Supreme Court that the particular state insurance law at issue was incompatible with disparate-impact liability. Ojo v. Farmers Grp., 356 S.W.3d 421, 422 (Tex. 2011). It took the position (in a single paragraph) that McCarran-Ferguson issues do not warrant a blanket exception for insurers, and could instead be addressed on a case-by-case basis by analyzing whether each state's regulatory regime was actually incompatible with a federal disparate-impact remedy. Id. HUD rejected the industry's concerns about the fundamental nature of the insurance business in another, even shorter, paragraph, stating that insurer defendants would have the opportunity to provide “a legally sufficient justification” for their practices at step two of the burden-shifting approach. Id. HUD declined to create specific safe harbors for insurance pricing or other risk-based factors and practices in its Rule on the same basis, reasoning that insurers would have an adequate opportunity to justify these practices by litigating through the burden-shifting framework. Id. And in dismissing concerns about the costs imposed by requiring insurers to collect race and ethnicity data, HUD simply recited the steps of the burden-shifting framework and noted that the initial burden fell on the plaintiff to provide evidence of disparate impact. Id.
III. PCI's Lawsuit and the Court's 2014 Decision
PCI filed this lawsuit challenging HUD's Rule in late 2013. (See Compl. [1].) Its sevencount complaint made three basic claims: (1) HUD's decision to apply disparate-impact liability to homeowners insurance was contrary to law and in excess of the agency's jurisdiction in light of McCarran-Ferguson; (2) HUD had acted in an arbitrary and capricious manner by failing to adequately consider issues regarding McCarran-Ferguson, the filed-rate doctrine, and the nature of insurance; and (3) HUD's burden-shifting framework was both arbitrary and capricious and contrary to law. (Id. ¶¶ 71-108.) Both sides moved for summary judgment. (See PCI's Mot. for Summ. J. [20]; HUD's Mot. to Dismiss and/or for Summ. J. [30].)
In September 2014, the court dismissed PCI's lawsuit in part but granted summary judgment in its favor on several claims. PCI's facial attack on the Rule under McCarran-Ferguson, the court held, was nonjusticiable: as the Supreme Court held in Humana, Inc. v. Forsyth, McCarran-Ferguson does not impose “any sort of field preemption” and requires fact-intensive analysis to determine whether federal law directly conflicts with state regulation or whether its application would frustrate state policy or interfere with a state's administrative regime. 525 U.S. 299, 309-10 (1999). Accordingly, PCI's claim was unripe for decision in the absence of a “concrete dispute regarding a particular insurance practice” and a particular set of state laws. PCI, 66 F.Supp.3d at 1040. The court also granted summary judgment in HUD's favor on PCI's challenges to the burden-shifting framework, finding that the agency's adoption of the framework was adequately reasoned and entitled to deference. Id. at 1053.
The court found, however, that HUD had inadequately addressed the issues that commenters had identified concerning McCarran-Ferguson, the filed-rate doctrine, and the nature of insurance. Even if this case did not present a justiciable platform for addressing all of the Rule's potential conflicts with McCarran-Ferguson, the court held, HUD's one-paragraph justification for its conclusion that these conflicts should be worked out on a case-by-case basis-rather than avoided altogether via a blanket exemption or safe harbor-was flawed in several respects. HUD had failed to address “the likelihood that McCarran-Ferguson preclusion will apply” in any given case, as well as “how often McCarran-Ferguson preclusion would apply and whether it would bar entire categories of disparate impact claims against insurers.” Id. at 1048. Moreover, the agency had failed to grapple with adverse caselaw that called the viability of disparate-impact claims against insurers' risk-based practices into question. Id. at 1049. HUD's failure to separately discuss the filed-rate doctrine was also arbitrary and capricious in light of these errors. Id. at 1050.
Further, the court held, HUD had made “no effort to evaluate the substance of the insurance industry's concerns” regarding the Rule's conflict with the nature of insurance. Id. at 1051. Even if insurers could present a legitimate and nondiscriminatory basis in defense of their risk-based practices under the burden-shifting framework, HUD still needed to justify its decision to “‘require courts to engage in the Sisyphean task of working through the burden-shifting framework in each individual case when the plaintiff has no chance of success,” rather than more carefully considering arguments for particular “exemptions from the Disparate Impact Rule” or “safe harbors” for practices that plaintiffs could not effectively challenge. Id. (quoting Graoch Assocs. # 33, L.P. v. Louisville/Jefferson Cnty. Metro Hum. Rels. Comm'n, 508 F.3d 366, 376 (6th Cir. 2007)). Accordingly, the court remanded the case to HUD for further explanation. Id. at 1054.
IV. Post-Remand Developments
The most important development in the nine years that have passed since the court's remand order is the Supreme Court's decision in Texas Department of Housing and Community Affairs v. Inclusive Communities Project, Inc., 576 U.S. 519 (2015). In Inclusive Communities, the Court resolved one of the longstanding legal questions hanging over this case at its inception, by holding that disparate-impact claims are cognizable under the FHA. Left open in that ruling were a number of questions over how such claims may be properly litigated. Soon after the Court's decision, PCI submitted comments to HUD stating that Inclusive Communities created significant new limitations on disparate-impact FHA liability, and that these limits further supported exempting homeowners insurance from the Rule. (Pl.'s SOF ¶ 60; R. 004615-21.)
In October 2016, HUD issued a Supplemental Response in response to the court's remand order. Application of the Fair Housing Act's Discriminatory Effects Standard to Insurance, 81 Fed.Reg. 69,012 (Oct. 5, 2016). This Response (for which HUD did not reopen the public comment period, though PCI submitted comments anyway; see R. 004416-4621) essentially reaffirmed the agency's original position on the Rule's applicability to insurance, but did so in much greater detail. HUD noted that the FHA itself does not contain an exemption for insurance, either as enacted or as amended, and stated that creating such an exemption by regulation would be “unworkable and inconsistent with HUD's statutory mandate.” 81 Fed.Reg. at 69,013. Over the course of eight pages in the Federal Register, the Supplemental Response addressed issues with McCarran-Ferguson, the filed-rate doctrine, and the nature of insurance. Id. at 69,014-19. However, it did not substantively discuss Inclusive Communities.
PCI moved the court to amend its complaint [107] in response to the Supplemental Response and Inclusive Communities. Its proposed amendments [107-1] listed counts that were substantively identical to those in its original complaint, including several that the court had already dismissed or resolved in HUD's favor at summary judgment, as well as several new counts based on Inclusive Communities. The court granted PCI's motion in part and denied it in part. Prop. Cas. Insurers Ass'n of Am. v. Carson, No. 13 CV 8564, 2017 WL 2653069 (N.D. Ill. June 20, 2017). The court denied as futile PCI's amended counts challenging the Rule as violative of the McCarran-Ferguson Act and its counts challenging HUD's burden-shifting scheme, as these claims were foreclosed by the court's 2014 order. Id. at *8. In addition, the court denied PCI's attempt to add a direct challenge to the Rule as “contrary to law” under Inclusive Communities, finding that the Supreme Court had largely endorsed HUD's burden-shifting approach and that any issues its holding posed for applying disparate-impact liability to insurers were “best left for analysis in specific cases rather than in a facial, pre-enforcement challenge.” Id. at *8-9. The court did, however, grant PCI's motion to amend with respect to the three arbitrary-and-capricious claims on which it had previously ruled in PCI's favor, as well as an additional claim that HUD's failure to address Inclusive Communities was also arbitrary and capricious. Id. at *9 & n.4. PCI filed an amended complaint consistent with the court's order [131] and moved for summary judgment again in late 2017 [136].
At that point, after a change in presidential administrations, the case went dormant for several years. In 2017, HUD proposed to replace the 2013 Rule with a new, more limited alternative. See Reconsideration of HUD's Implementation of the Fair Housing Act's Disparate Impact Standard, 83 Fed.Reg. 28,560 (June 20, 2018); HUD's Implementation of the Fair Housing Act's Disparate Impact Standard: Proposed Rule, 84 Fed.Reg. 42,854 (Aug. 19, 2019). This litigation, now reassigned to the undersigned judge following Judge St. Eve's appointment to the Seventh Circuit [157], was stayed several times to accommodate the agency's change in leadership and new rulemaking. It appeared ripe for a resolution when, in September 2020, HUD issued a final revised Rule. HUD's Implementation of the Fair Housing Act's Disparate Impact Standard: Final Rule, 85 Fed.Reg. 60,288 (Sept. 24, 2020). But before the parties could even begin with motion practice, the new Rule was itself declared arbitrary and capricious by another federal court, who stayed its enforcement, thus leaving the 2013 Rule in place. See Mass. Fair Hous. Ctr. v. United States Dep't of Hous. & Urb. Dev., 496 F.Supp.3d 600, 611-12 (D. Mass.
Like the 2013 Rule that sparked this lawsuit, HUD's 2020 Rule did not create a categorical exemption for the insurance industry or for risk-based practices specifically. Instead, the 2020 Rule reaffirmed the position that the viability of disparate-impact claims against insurers would need to be decided on a case-by-case basis. See 85 Fed.Reg. at 60,324-26. The 2020 Rule did, however, add specific language mirroring McCarran-Ferguson's restrictions, and codified a new defense: a showing that defendants' challenged actions were “reasonably necessary to comply with” state law or other legal requirements. Id. at 60,333 (to be codified at 24 C.F.R. § 100.500(d)(1), (e)). Because the 2020 Rule never took effect, the question of what (if any) legal ramifications these changes might have had was never answered. More generally, the 2020 Rule would have sharply limited disparate-impact liability by, as one court put it, imposing “new, onerous pleading requirements on plaintiffs” and altering the burden-shifting framework in defendants' favor. Mass. Fair Hous. Ctr., 496 F.Supp.3d at 607-08.
2020). Only a few months later, another change in administrations resulted in yet another reversed course: the Biden Administration promptly directed that HUD reconsider its 2020 rulemaking. See Redressing Our Nation's and the Federal Government's History of Discriminatory Housing Policies, 86 Fed.Reg. 7,487, 7,488 (Jan. 26, 2021).
What followed was another lengthy delay, as HUD announced its intent to “recodify its previously promulgated rule” via a notice of proposed rulemaking. Reinstatement of HUD's Discriminatory Effects Standard: Proposed Rule, 86 Fed.Reg. 33,590, 33,590 (June 25, 2021). This time, unlike in 2016, HUD did reopen its public comment period, and received more than 10,000 comments on its proposal. (See HUD's Statement of Facts [295-2] (hereinafter “Def.'s SOF”) ¶ 18.) These included statements from PCI and other insurance industry representatives, as well as federal and state lawmakers and individual policyholders, that reiterated many of the same legal and policy arguments against applying the Rule to insurance. (Pl.'s SOF ¶¶ 92-114; R. 028170-29162, 029271-553.) Other commenters, however, took the opposite position and urged HUD not to reverse its original position regarding liability for the insurance industry. (E.g., R. 029163-86.)
While PCI moved for summary judgment a third time in April 2021 [222] before HUD had finished its rulemaking, the court struck this motion and HUD's opposition and crossmotion [231] without prejudice in February 2022. (See Ord. [248].)
Finally, in March 2023, HUD promulgated a new Rule that reinstated the 2013 burdenshifting framework in full. Reinstatement of HUD's Discriminatory Effects Standard: Final Rule, 88 Fed.Reg. 19,450 (Mar. 31, 2023). HUD's response to the new round of public commentary was substantial: the response spanned almost fifty pages in the Federal Register, including seventeen devoted exclusively to comments concerning insurance, id. at 19,463-80, and another six to Inclusive Communities, id. at 19,457-62. It is this record-along with HUD's 2016 Supplemental Response, which the agency reaffirmed and incorporated by reference in its 2023 rulemaking, see id. at 19,463 n.113-that the parties now present for the court's consideration.
PCI filed a Second Amended Complaint [274] a few months after HUD issued its rule.Following yet another round of cross-motions for summary judgment [282, 295], the matter is at long last ready for the court's decision.
This amended complaint makes “technical and non-substantive amendments clarifying that [PCI's] claims in this action challenge both the 2013 Rule and the substantively identical 2023 Reinstatement Rule.” (See Joint Status Report [272] at 2.) While it reasserts all of PCI's claims, including those that the court previously dismissed from its original complaint and rejected as futile in its 2017 order on PCI's motion to amend, the parties have agreed that these claims are included solely to ensure that they are preserved for appeal. (Id. at 2-3.)
LEGAL STANDARD
As before, this request for judicial review of an agency action is governed by the standards set forth in the Administrative Procedure Act (“APA”), 5 U.S.C. § 551 et seq. Thus, PCI argues that HUD's rulemaking here violates the APA's prohibition of “arbitrary [and] capricious” agency action. Id. § 706(2)(A). In considering such a challenge, the court's review is “narrow in scope”; the court will not “substitute [its] own policy judgment for that of the agency.” Cook Cnty. v. Wolf, 962 F.3d 208, 230 (7th Cir. 2020). Rather, the court “simply ensures that the agency has acted within a zone of reasonableness and, in particular, has reasonably considered the relevant issues and reasonably explained the decision.” FCC v. Prometheus Radio Project, 592 U.S. 414, 423 (2021). As part of this explanation, the agency should “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” Encino Motorcars, LLC v. Navarro, 579 U.S. 211, 221 (2016) (quoting Motor Vehicle Mfrs. Assn. of United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).) The agency's action may be found arbitrary and capricious if it
has relied on factors which Congress had not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.Nat'l Ass'n of Home Builders v. Defs. of Wildlife, 551 U.S. 644, 658 (2007) (quoting State Farm, 463 U.S. at 43). The court's review is limited to the reasons that the agency gave itself at the time of its decision. Encino Motorcars, 579 U.S. at 224; see SEC v. Chenery Corp., 318 U.S. 80, 94-95 (1943).
DISCUSSION
I. Subject-Matter Jurisdiction
Before considering the merits of PCI's claims, the court addresses the threshold question of its jurisdiction. While Judge St. Eve held in 2014 that PCI had standing to assert its arbitrary-and-capricious claims and that they were ripe for review, see PCI, 66 F.Supp.3d at 1042-46, HUD asks the court to revisit this determination in light of changed circumstances over the intervening nine years. As explained below, the court concludes that PCI has standing and that the issues are again ripe.
A. Standing
Article III of the Constitution requires a plaintiff to have “(1) suffered an injury in fact; (2) that is fairly traceable to the challenged conduct of the defendant; and (3) that is likely to be redressed by a favorable judicial decision.” Prairie Rivers Network v. Dynegy Midwest Generation, LLC, 2 F.4th 1002, 1007 (7th Cir. 2021). Constitutional standing is a jurisdictional requirement, and while a court's prior rulings generally control later phases of litigation as the law of the case, its “ongoing obligation to assure itself of its jurisdiction means that revisiting such matters is almost always on the table.” Flynn v. FCA U.S. LLC, 39 F.4th 946, 953 (7th Cir. 2022).
An association like PCI has standing to sue on behalf of its members if “(a) its members would otherwise have standing to sue in their own right; (b) the interests it seeks to protect are germane to the organization's purpose; and (c) neither the claim asserted nor the relief requested requires the participation of individual members in the lawsuit.” Students for Fair Admissions, Inc. v. President & Fellows of Harvard Coll., 600 U.S. 181, 199 (2023) (citation omitted). In 2014, the court concluded that all three requirements of this test were met: PCI's individual members were “objects” of the Disparate-Impact Rule who faced concrete and tangible harms from the increased exposure and compliance costs that it posed, and these harms were both fairly traceable to the Rule and could be redressed by a favorable court decision on PCI's challenge. PCI, 66 F.Supp.3d at 1043-46.
HUD did not dispute the second and third elements of the associational standing test, so only the first-whether PCI's members had standing to sue in their own right-was at issue in the court's original decision. PCI, 66 F.Supp.3d at 1043. This is also the case here.
HUD argues that both of the cognizable injuries-in-fact that this court previously identified-increased exposure to disparate-impact liability and future compliance costs-have been called into question by the passage of time. It contends that, after eleven years of living with the 2013 Rule (and its 2023 reincarnation), PCI has failed to show that either of these supposedly imminent harms have in fact come to pass. The foundations for PCI's standing, HUD points out, are declarations from its members that date back to its original motion for summary judgment in 2014. (See Decl. of Peter Drogan [283-5]; Decl. of Ronald Zaleski [283-6]; Decl. of Michael Dawdy [283-7]; Decl. of Teresa C. Cracas [283-8].) But these declarations do not establish that PCI's members actually ended up suffering any increased litigation costs from a Rule-related uptick in disparate-impact lawsuits, or that they ever incurred new expenses from collecting race-based data-making these harms, in HUD's view, purely conjectural.
And even if PCI could establish such costs, HUD argues, it would be unable to prove causation or redressability. As HUD sees things, the 2013 Rule merely reaffirmed preexisting regulatory and judicial law that would already have allowed for disparate-impact claims against insurers, and the Supreme Court's 2015 Inclusive Communities decision reinforced this body of law by confirming the viability of disparate-impact claims across all jurisdictions. Thus, in HUD's view, any disparate-impact-related harms that PCI's members face are not fairly traceable to the Rule, and vacating it would do nothing to decrease their potential exposure.
It is disappointing in this context that PCI has not collected updated declarations or advanced other evidence of what real-world impacts (if any) the Rule has had on the insurance industry. Such evidence might have simplified this dispute over standing and given the court a window into the underlying merits of PCI's challenges to the Rule. But the court will not toss out PCI's claims on this basis. As in 2014, a party's standing to seek review of administrative action is usually “self-evident” where they are an “object of the action (or forgone action) at issue.” Bonacci v. TSA, 909 F.3d 1155, 1160 (D.C. Cir. 2018) (citation omitted); see also Owner-Operator Indep. Drivers Ass'n, Inc. v. Fed. Motor Carrier Safety Admin., 656 F.3d 580, 585 (7th Cir. 2011). Even if the Rule on its face applies to any provider of housing or housing-related services, HUD has made it abundantly clear from the get-go that the insurance industry is a particular target of its rulemaking. See 76 Fed.Reg. at 70,924. And while PCI's original member declarations are certainly aged, they are not necessarily stale, given the particular history of this case. HUD has reinstated its 2013 Rule in substantively identical form after nearly a decade of legal uncertainty regarding its viability and enforcement. Whether or not the litigation risks that these declarations anticipated in 2014 (see, e.g., Cracas Decl. ¶ 13) have indeed come to pass in the meantime, the arguable end of this uncertainty by way of the Rule's reincarnation presents a new flashpoint for them to manifest. Thus, while HUD's arguments as to standing have some teeth, PCI has shown a sufficiently persuasive likelihood of concrete and imminent harm to its members-again, based on the unique circumstances presented by HUD's 360-degree turn-that the court is comfortable reaching the merits.
HUD also argues that because the Rule does not explicitly require PCI's members to collect race- and ethnicity-based data, they cannot assert standing based on compliance costs that are not in fact legally mandated. It is true that a plaintiff cannot claim standing to challenge a regulation based on a misreading of what that regulation permits or requires. Elec. Priv. Info. Ctr. v. FAA, 892 F.3d 1249, 1253-54 (D.C. Cir. 2018); see Nat' Assn of Mut. Ins. Cos. v. HUD (“NAMIC”), No. CV 13-966 (RJL), __ F.Supp.3d __, 2023 WL 6142257, at *9 (D.D.C. Sept. 19, 2023) (recognizing that the Rule does not “require those engaging in housing practices to collect or use data on individuals' protected characteristics”). HUD may well be correct here: while PCI's members may end up spending more to collect race-based data in order to defend themselves from disparate-impact claims, the court cannot conclude that this is necessary as a practical matter without seeing how such litigation actually plays out. See 88 Fed.Reg. at 19,480 (noting that “[d]efendants need not present their own statistics” in a disparate-impact case, “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”). Still, defending against increased disparate-impact litigation would certainly impose some costs on PCI's members, whether or not these are incurred beforehand or after a suit has already been filed.
Nor do PCI's claims fail on causation or redressability grounds. The D.C. district court's recent opinion in NAMIC is persuasive on this point. NAMIC was another challenge to HUD's 2013 Rule brought by another insurance trade association around the same time this case was filed, but based on a different legal theory.In its September 2023 decision, the NAMIC court rejected HUD's similar arguments as to justiciability, noting that they “curiously undersell[] the Disparate-Impact Rule . . . [which] can hardly be said to have zero practical effect in light of preexisting legal duties. If it did, why promulgate it in the first place?” NAMIC, 2023 WL 6142257, at *6.
Specifically, the plaintiffs in NAMIC argued before the D.C. court-the only circuit not to have recognized disparate-impact claims as cognizable under the FHA prior to Inclusive Communities-that the Rule exceeded the scope of the FHA's text. Am. Ins. Ass'n v. HUD, 74 F.Supp.3d 30, 31-32 (D.D.C. 2014). While the district court initially ruled against HUD on this theory, id. at 47, it reversed course following the Supreme Court's 2015 decision.
This court concurs. As the history surveyed above shows, the viability of disparate-impact FHA claims against risk-based insurance practices had never been conclusively established across all circuits prior to HUD's rule. See cases cited supra note 5. HUD's goal in promulgating the Disparate-Impact Rule has always been to create a single, “consistent and predictable” standard for evaluating disparate-impact FHA claims across the country. 78 Fed.Reg. at 11,460. Even if HUD is correct that the Rule “largely codified longstanding judicial and agency consensus regarding discriminatory effects law” in most respects, 88 Fed.Reg. at 19,450, there was no such consensus here. By purporting to end this debate, the Rule makes a significant difference in the regulatory landscape-one that PCI's members are within their rights to contest.
B. Ripeness
Next, the court turns to HUD's arguments that PCI's arbitrary-and-capricious claims are not ripe for resolution. Ripeness is a concern about the proper timing of judicial intervention that is based on both constitutional and prudential considerations. Nat'l Park Hosp. Ass'n v. Dep't of Interior, 538 U.S. 803, 808-09 (2003). The test for constitutional ripeness is largely equivalent to that for Article III standing, which the court has already found to be satisfied here. See Chlorine Inst., Inc. v. Fed. R.R. Admin., 718 F.3d 922, 927 (D.C. Cir. 2013). Prudential ripeness asks whether “[1] the claim is fit for judicial decision and [2] delay will cause some hardship to the parties.” Owner-Operator Indep. Drivers Ass'n, 656 F.3d at 586 (citing Abbott Lab'ys v. Gardner, 387 U.S. 136, 149 (1967)). Challenges to final agency rules, even if brought pre-enforcement, are “presumptively reviewable” if they involve purely legal claims for the court's determination. Id. (citation omitted).
On the first prong of the Abbott prudential ripeness test, HUD appears to concede that PCI's arbitrary-and-capricious claims are purely legal, but contends that they are nevertheless unfit for judicial decision. Its argument on this point tracks the court's reasoning for finding PCI's facial challenge to the Rule under McCarran-Ferguson unripe in 2014. See PCI, 66 F.Supp.3d at 1038-41. Because PCI's challenge involves a host of conjectures as to the Rule's potential impacts and how courts might rule on future disparate-impact claims, HUD argues, the court lacks a principled basis to rule on whether its actions were in fact arbitrary and capricious and should await a more concrete dispute. In other words, HUD claims that the “data” to determine whether the Rule is likely to run afoul of McCarran-Ferguson or other laws, or disrupt insurance markets- and thus whether the decision to promulgate it was arbitrary and capricious-is unknowable before the Rule goes into effect. Further, HUD disputes that PCI's members will suffer any hardship from withholding review, since the Rule does not proscribe any direct conduct and PCI has produced no record evidence showing that its members have been forced to change any of their business practices.
The court rejects these arguments. As the court noted in 2014, “its determination that PCI's McCarran-Ferguson claim is not ripe for judicial review does not necessarily mean that HUD's decision to leave application of McCarran-Ferguson preclusion for a case-by-case determination was not arbitrary and capricious.” PCI, 66 F.Supp.3d at 1047. Nothing has changed since. PCI's arbitrary-and-capricious claim turns not on the actual occurrence of “contingent future events” related to the Rule's enforcement, but on the sufficiency of HUD's own predictive reasoning-how the agency itself anticipated and grappled with these possibilities-as set forth in the administrative record. Nat'l Ass'n of Home Builders v. U.S. Army Corps of Eng'rs, 417 F.3d 1272, 1282 (D.C. Cir. 2005) (citation omitted). Indeed, HUD's argument is at odds with the very concept of pre-enforcement review, which “has become the norm” for final agency rules. Owner-Operator Indep. Drivers Ass'n, 656 F.3d at 586. It is true that the “legal” questions presented here are heavily inflected by “factual” elements, given the fact-intensive nature of the McCarran-Ferguson inquiry and the Rule's potential effects on the business of insurance. But while further information on the Rule's ultimate consequences could certainly shed light on whether HUD's decision to adopt it was in fact arbitrary and capricious, that kind of hindsight is neither required by the law nor necessary to evaluate the agency's deliberative process.
HUD's claim that PCI has failed to show hardship is similarly unavailing. Even if the Rule does not specifically permit or forbid any primary conduct, its effects on disparate-impact doctrine may well be “felt immediately by those subject to it in conducting their day-to-day affairs.” Toilet Goods Ass'n, Inc. v. Gardner, 387 U.S. 158, 164 (1967). As the NAMIC court observed, this makes it “a substantive rule which as a practical matter requires [PCI's members] to adjust [their] conduct immediately.” NAMIC, 2023 WL 6142257, at *8 (citing Nat'l Park Hosp. Ass'n, 538 U.S. at 809). While HUD cites Toilet Goods for the opposite proposition-that PCI has never shown any such immediate impact on its membership-this, again, fails to account for the Rule's state of procedural limbo over the past nine years. Even if the Rule has technically been on the books all this time, its reinstatement creates a meaningful risk of increased disparate-impact litigation. This threat is all that is required to show hardship: “a person made to live in the shadow of a law that she believes to be invalid should not be compelled to wait and see if a remedial action is coming.” Owner-Operator Indep. Drivers Ass'n, 656 F.3d at 586. PCI's surviving claims are ripe for the court's decision.
II. Merits of PCI's Claims
The court thus turns to the merits: whether HUD's decision to apply the 2013 Rule, and its 2023 successor, to risk-based insurance practices was arbitrary and capricious. Two important framing points from the court's 2014 opinion are worth repeating at the outset. First, it is a bedrock principle of administrative law that agencies have discretion to select between rulemaking and case-by-case adjudication as their preferred method for making law. Shays v. Fed. Election Comm'n, 424 F.Supp.2d 100, 113 (D.D.C. 2006) (citing SEC v. Chenery Corp., 332 U.S. 194, 203 (1947)). But second, any action that an agency takes must be adequately reasoned, and the agency must weigh the pros and cons of each method before exercising this discretion. Id.
The court recognized in 2014 that HUD had the authority to create a Rule that left the extent of potential disparate-impact liability against insurers to be worked out ex post by the courts, rather than simply carving them out ex ante via an exemption or safe harbor. PCI, 66 F.Supp.3d at 1048. But the court held that, given the serious concerns that the insurance industry raised concerning McCarran-Ferguson preemption, the filed-rate doctrine, and the fundamental nature of insurance itself, HUD's perfunctory response to these concerns was arbitrary and capricious. Id. at 1049-51. In light of the serious legal and policy issues presented by its chosen approach, the court concluded, HUD needed to more carefully address the application of disparate-impact liability in the insurance industry. The question now is whether it has adequately done so.
A. Was HUD's Consideration of the McCarran-Ferguson Act Arbitrary and Capricious?
In 2014, the court concluded that in its original rulemaking, HUD had not adequately accounted for the potential clash between its Rule and McCarran-Ferguson. The court acknowledged that McCarran-Ferguson requires a case-by-case analysis to determine whether application of a federal law would conflict with a given state's regulations, see Humana, 529 U.S. at 313, but concluded that this did not excuse HUD's obligation to carefully assess the likelihood of such conflicts in any given case, in light of the obvious discordance between disparate-impact liability and risk-based insurance practices. 66 F.Supp.3d at 1048. This omission was “particularly glaring” given the weight of caselaw-in particular, the Seventh Circuit's decision in Mutual of Omaha and the Eighth's in Saunders II-suggesting that such cases against insurers might indeed routinely be preempted under McCarran-Ferguson. Id. at 1048-49.
HUD devoted three pages in its 2016 Supplemental Response and five pages in its 2023 rulemaking to addressing McCarran-Ferguson issues raised by the court and other commenters. Broadly, HUD stated that “[a]fter careful reconsideration . . . [it] continues to believe that case-by-case adjudication is preferable to creating the requested exemptions or safe harbors for insurance practices.” 81 Fed.Reg. at 69,013; see 88 Fed.Reg. at 19,474. HUD reasoned that in light of the case-by-case nature of McCarran-Ferguson analysis, such “wholesale exemptions would be overbroad” and would undermine the Act's efficacy by excluding a host of claims that might not otherwise be preempted. 88 Fed.Reg. at 19,474. It presented a number of prior court decisions that, it contended, showcased the potential viability of disparate-impact insurance claims notwithstanding McCarran-Ferguson. Id. at 19,474-75. HUD also argued that a blanket exemption would be “overbroad and quickly outdated” in light of the variation between state regulatory regimes “as well as the ways in which a single state's insurance laws can change over time ....” Id. at 19,475. It noted, further, that state insurance codes were not “hermetically sealed” and needed to be read in context with other state laws, such as fair housing and antidiscrimination protections, that might harmonize with a federal disparate-impact remedy. Id. (citing Humana, 525 U.S. at 312). Finally, HUD disputed that the Rule's burden-shifting framework categorically forbade risk-based pricing or underwriting in a way that would inevitably raise McCarran-Ferguson issues. Id. at 19,475-76.
PCI finds several faults with these explanations. First, it argues, HUD failed to meaningfully answer the court's question of “how often” McCarran-Ferguson preemption might apply. It is true that no such estimate appears in the record, either in 2016 or 2023. But in fairness to HUD, it would be extremely difficult to find-and PCI does not offer-any principled way of coming up with a precise measure of this frequency. As an initial matter, “[t]he APA imposes no general obligation on agencies to conduct or commission their own empirical or statistical studies.” Prometheus Radio Project, 592 U.S. at 427. Nor is it clear that such a study would even be reliable here given the number of variables at play. As the court previously recognized in dismissing PCI's facial attack on the Rule under McCarran-Ferguson, the likelihood of McCarran-Ferguson preemption in any given disparate-impact case would depend on a host of “imponderables,” including the nature of the insurance practice at issue, the contours of state law, and the deciding court's own analysis. PCI, 66 F.Supp.3d at 1039-40.
The only way to even begin to assess this likelihood, in other words, is by reference to the relevant jurisdiction's caselaw-whether precedent strongly suggests that disparate-impact claims against insurers are likely to fail as a matter of law in light of McCarran-Ferguson. On this score, HUD has presented a far more comprehensive survey of the legal landscape to back up its position than it did in 2013. Rather than citing a single case's reasoning (Ojo) as it did before, HUD has presented a panoply of court decisions across multiple circuits, all of which found that a disparate impact claim could proceed against insurers without necessarily running afoul of McCarran-Ferguson. HUD cites, for example, the Fifth Circuit's decision in DeHoyos v. Allstate Corp., which rejected a McCarran-Ferguson defense against a disparate-impact claim where there was no “actual state insurance law which purportedly conflicted with the application of [federal law] to the particular insurance question at issue.” 345 F.3d at 298; see 88 Fed.Reg. at 19,474.
PCI argues that this case misses the point, as do HUD's arguments on how state law can evolve and change over time. PCI is specifically requesting an exemption (or safe harbor) for risk-based practices, which-it points out-all states have either permitted or required by law for decades. (See Pl.'s SOF ¶ 94.) But HUD has also provided a sampling of cases (albeit none that are binding precedent in this circuit) supporting its position that disparate-impact claims can proceed even against risk-based practices without violating McCarran-Ferguson. In Lumpkin v. Farmers Group (“Lumpkin II”), for example, a federal district court held that “[b]ecause both the FHA and Tennessee insurance law prohibit racial discrimination, without an explicit exception for actuarially sound scoring with disparate impact, they are in harmony.” No. 05-2868 MA/V, 2007 WL 6996777, at *7 (W.D. Tenn. July 6, 2007). HUD also cites several cases that were decided after this court's 2014 ruling and that found no inherent conflict between disparate-impact liability for risk-based practices and the McCarran-Ferguson Act. See 88 Fed.Reg. at 19,463 n. 116, 19,475 nn. 220, 222. The district courts in Viens v. America Empire Surplus Lines Insurance Co., 113 F.Supp.3d 555 (D. Conn. 2015), Jones v. Travelers Casualty Insurance Co. of America, No. C-13-02390 LHK, 2015 WL 5091908 (N.D. Cal. May 7, 2015), and National Fair Housing Alliance v. Travelers Indemnity Co., 261 F.Supp.3d 20 (D.D.C. 2017) all rejected McCarran-Ferguson defenses against insurers for making adverse pricing and underwriting decisions against landlords who rented to tenants receiving Section 8 housing assistance, even though these practices were arguably also based on objective risk and cost factors.
PCI argues in a footnote that this case was wrongly decided and rests on a misreading of Tennessee law. It claims that the Lumpkin II court erroneously conflated the concepts of “unfair” discrimination based solely on race with ”fair” discrimination based on risk, and that while the former concept is banned under Tennessee's code, the latter is not only permitted but required. See Tenn. Code Ann. §§ 56-5-103(a)(1), (d), -104(1). But Lumpkin II did recognize this distinction, and concluded that the “[d]efendants read [Tennessee's law] too broadly to permit disparate impact as long as the rates are actuarially sound.” Lumpkin II, 2007 WL 6996777, at *6. In other words, Lumpkin II specifically considered, and rejected, the underlying presumption that PCI makes in this case-that if a state has enacted an insurance provision requiring insurers to consider actuarial risk factors, insurers have an unquestioned defense to a disparate impact claim. Cf. Moore, 267 F.3d at 1222 (“We are asked to assume that the abolition of one form of discrimination . . . amounts to a clear declaration by the state that all other forms of discrimination, however invidious, are acceptable. We cannot construe Alabama's scheme of insurance regulation in such a formalistic and narrow way.”).
Whether HUD has adequately addressed the cases that come out the other way-in particular, Mutual of Omaha-is a harder question. If read expansively, Mutual of Omaha establishes that the act of assessing whether an insurer's practices are “actuarially sound and consistent with state law” itself violates McCarran-Ferguson, full stop. 179 F.3d at 564. In PCI's view, this holding necessarily precludes any disparate-impact claim against an insurer's riskbased pricing or underwriting practices, as any court evaluating such a claim would inevitably need to engage in this forbidden inquiry. And while HUD cites the Eighth Circuit's decision in Saunders II in support of its contention that McCarran-Ferguson requires a case-by-case approach, that decision in fact cited Mutual of Omaha in expressing deep skepticism over whether a disparate-impact claim against risk-based practices could ever be viable. See Saunders II, 537 F.3d at 967-68 (citing Mut. of Omaha, 179 F.3d at 564).
In its rulemaking, HUD offered two responses to this adverse caselaw. First, it argues that Mutual of Omaha need not be applied as strictly as PCI urges. It cites this court's recognition that “[w]hile some states require insurers to use risk based pricing, other states merely permit riskbased pricing,” and argues on this basis that “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.” 88 Fed.Reg. at 19,476 (citing PCI, 66 F.Supp.3d at 1039-41). In addition, HUD contends that the structure of the burden-shifting framework itself demonstrates that Mutual of Omaha does not doom any possible challenge to risk-based practices. As it reasons,
[w]hile another risk-based practice could be a possible alternative at step three [of the burden-shifting paradigm], 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 .... 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.Id. (citing DeHoyos, 345 F.3d at 297 n.5).
This argument is logically appealing, but the court cannot predict how much purchase it will ultimately have without knowing more about how disputes against insurers' risk-based practices will actually unfold in practice. PCI argues that HUD's hypothetical is untethered from reality, and that in the vast majority of disputes over insurers' risk-based practices, courts will need to gauge the validity of the insurer's claims of actuarial objectivity in order to determine whether those risk-based practices constitute a legally sufficient business justification at steps two and three. This prediction seems plausible and finds at least some support in statements HUD makes elsewhere in its rulemaking that call the objectivity of insurers' purportedly “riskbased” practices into question. As justification for its decision not to create a “risk-based” safe harbor, the agency suggests that “[e]ven practices such as ratemaking that are largely actuarially-based can incorporate an element of non-actuarially-based subjective judgment or discretion under state law,” and that “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.'” 88 Fed.Reg. at 19,468 (citing Nat'l Fair Hous. All. v. Prudential Ins. Co. of Am., 208 F.Supp.2d 46, 60 (D.D.C. 2002)). The court addresses the merits of this argument below, but notes here that expecting federal courts, rather than state insurance commissioners, to draw this line between subjectivity and objectivity runs perilously close to what Mutual of Omaha forbids.
See discussion infra Section II.C.
This leaves HUD's second major ground for disregarding Mutual of Omaha: that it simply disagrees with the case's holding and is not bound to follow it in crafting a nationwide rule. The agency notes that Mutual of Omaha and like cases are not binding outside their circuits, and 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.” Id. at 19,476 (citing Nat'l Cable & Telecomms. Ass'n v. Brand X Internet Servs., 545 U.S. 967, 982, 983-84 (2005)). Because circuit caselaw is already inconsistent on this issue and could evolve further, HUD argues, it should not be obligated to create a carveout for all risk-based insurance practices based on a single Seventh Circuit holding.
On its own, this explanation is unsatisfying. In its 2014 ruling, this court did not direct HUD to address why Mutual of Omaha is not binding on other courts, but whether its reasoning was sufficiently persuasive that most if not all other courts would follow its lead-making a preemptive exemption appropriate. See PCI, 66 F.Supp.3d at 1048-49. That said, HUD has shown that not all other jurisdictions have in fact strictly adhered to Mutual of Omaha's stringent rule against federal-court review of purportedly actuarial insurance practices.In particular, HUD cites the Fifth Circuit's holding in DeHoyos that, “[i]n engaging in the unremarkable task of determining whether specific conduct falls within the ambit of federal civil rights law, a court would no more become a ‘super actuary' than the court becomes a ‘super entrepreneur' each time the court must determine whether a discriminatory practice constitutes a business necessity.” 345 F.3d at 297 n.5.; see also, e.g., Prudential, 208 F.Supp.2d at 60-61; Lumpkin, 2007 WL 6996777, at *7; Viens, 113 F.Supp.3d at 572-73; Jones, 2015 WL 5091908, at *5; Travelers Indem., 261 F.Supp.3d at 35 n.4; cf. Moore v. Liberty Nat. Life Ins. Co., 267 F.3d 1209, 1222 (11th Cir. 2001) (reaching a similar conclusion as to life insurance). This court and other district courts in this circuit are bound by Mutual of Omaha (as HUD concedes, see 88 Fed.Reg. at 19,476), and are thus barred from evaluating at least some aspects of insurers' risk-based practices. But case law cited by HUD at least shows that Mutual of Omaha's underlying reasoning does not compel equivalent results elsewhere. And even within this circuit, it is worth asking whether the McCarran-Ferguson concerns raised in Mutual of Omaha are so inevitable that federal courts cannot even attempt to reconcile them.
Some of this can be traced back to deeper schisms in the underlying caselaw. For example, HUD points out that the Second Circuit has expressed skepticism over whether McCarran-Ferguson applies at all to “subsequently enacted civil rights legislation,” a position that the Seventh Circuit has repudiated. 88 Fed.Reg. at 19,474 n.208 (citing Viens, 113 F.Supp.3d at 572). Compare Spirt v. Teachers Ins. & Annuity Ass'n, 691 F.2d 1054, 1065 (2d Cir. 1982), with Am. Family, 978 F.2d at 294-95.
Consider, for instance, Huskey v. State Farm Fire & Cas. Co., No. 22 C 7014, 2023 WL 5848164 (N.D. Ill. Sept. 11, 2023), a recent disparate-impact challenge in this district to an insurer's claims-processing policies. In denying the defendant's motion to dismiss based on, inter alia, McCarran-Ferguson, the district court held that “disparate-impact liability under the FHA, as applied, seems to complement [rather than frustrate] Illinois insurance law,” and that while “calculating the relative value of insurance policies is off-limits under the McCarran Ferguson Act . . . [t]here is no indication that State Farm's liability would hinge on its practices' actuarial soundness or consistency with Illinois law.” Id. at *11. The Huskey court was prepared to revisit the issue if it later became clear through discovery that liability “depends on the actuarial soundness of its claims-handling policies in a way that interferes with Illinois's insurance policy or administrative regime.” Id. (emphasis added). Huskey postdates HUD's 2023 rulemaking and cannot supply a post hoc justification for the agency's actions, see Chenery, 318 U.S. at 95, but it offers a useful illustration of the scope of Mutual of Omaha's holding in practice. At minimum, it shows that Mutual of Omaha has no bearing on insurers' practices that are clearly not riskbased. And it suggests that, even where an insurer's risk-based practices are called into question, a court should at least consider whether examination of these practices would actually frustrate state law before simply assuming the question is foreclosed by Mutual of Omaha. See Huskey, 2023 WL 5848164, at *11 (“While the Seventh Circuit's reading of the McCarran-Ferguson Act is prophylactic to a degree, State Farm's arguments veer into the realm of field preemption. The Act is not so expansive.”) (citations omitted).
This is arguably underscored by the difference of opinion among the states over the Rule's implications for their policy schemes. HUD points to state antidiscrimination laws, several of which allow for equivalent disparate-impact claims against insurers, as evidence that the Rule does not per se conflict with states' regulatory regimes. 88 Fed.Reg. at 19,475. It cites Viens, Jones, and Toledo Fair Housing Center v. Nationwide Mutual Insurance Co., 94 Ohio Misc.2d 151, 157, 704 N.E.2d 667, 670 (Ohio Ct. Com. Pl. 1997), which all found that a disparate-impact remedy complemented, rather than frustrated, their respective states' insurance laws, as well as public commentary expressing that “the proposed rule does not undermine the state regulation of insurance and presents no conflict with McCarran-Ferguson.” Id. at 19,474-75 & nn. 220-22. To provide a categorical exemption, HUD argues, “would deprive all states of this federal support in addressing discriminatory insurance practice-even those states that welcome or depend on such support.” Id. at 19,475; see also Am. Family, 978 F.2d at 295 (“Duplication is not conflict.”). HUD's position may read too much into the state laws themselves: in cases involving potentially conflicting statutes, it cannot always be presumed that the legislature anticipated their collision in advance. In other words, state legislatures may have been moved in one session to adopt insurance regulation, and in another to adopt antidiscrimination legislation, without any thought whatsoever as to how the two might interact. Then, absent further legislative action to resolve this conflict, it falls to courts to hammer out a workable compromise. Cf. Epic Sys. Corp. v. Lewis, 584 U.S. 497, 510 (2018) (“When confronted with two [statutes] allegedly touching on the same topic, this Court is not at ‘liberty to pick and choose among [them]' and must instead strive ‘ “to give effect to both.” '”) (citing Morton v. Mancari, 417 U.S. 535, 551 (1974)).
As Mutual of Omaha declared, “[t]he states are not indifferent to who enforces their laws,” 179 F.3d at 564, but there is reason in this case to consider the states' own comments on this issue. On one side, attorneys general from fourteen states and the District of Columbia have weighed in as amici to voice their support for HUD's decision not to exempt the insurance industry from its Rule. (See States' Brief as Amici Curiae in Sup. of Defs.' Mot. for Summ. J. [309] (hereinafter “States' Amicus Supporting HUD”).) This includes Illinois itself, which takes the position-notwithstanding Mutual of Omaha-that Illinois law “imposes disparate impact liability on insurance companies” and that the Rule “complement[s]” rather than “frustrat[es]” this administrative regime. (Id. at 11-14.) On the other, insurance commissioners from four stateshave submitted an amicus brief arguing that their regulatory regimes would be disrupted by such a remedy. (See Brief of State Ins. Comm'rs in Supp. of Pl.'s Renewed Mot. for Summ. J. [292] (hereinafter “States' Amicus Opposing HUD”).)
The full list is: Illinois, California, Colorado, Delaware, Hawaii, Maryland, Massachusetts, Minnesota, Nevada, New Jersey, New York, North Carolina, Oregon, Washington, and the District of Columbia.
These states are: Idaho, Montana, Louisiana, and Oklahoma.
This diversity of opinion is significant, in the court's view. PCI tries to argue that the mere presence of complementary state fair-housing laws does not eliminate Mutual of Omaha's concern over “who decides”-that the task of how to reconcile these laws with insurers' risk-based practices must be left to state insurance regulators, not federal courts. But it is HUD's case-by- case approach, not PCI's requested blanket exemption, that presumes less about what state regulators want and need. It is clear from the administrative record and amicus submissions that several states want no part of a federal disparate-impact remedy that might disturb their administrative regimes. (See States' Amicus Opposing HUD at 7-10.) In these states, McCarran-Ferguson may very well end up preempting the Rule's application to certain risk-based insurance practices-a point that HUD acknowledges in its rulemaking. 88 Fed.Reg. at 19,474. A state that wishes to join this group in the future, after reassessing the Rule's effects in practice, “need only say so” by amending its regulations. Am. Family, 978 F.2d at 297.
PCI also casts doubt on HUD's supporting amici attorneys general in its briefing, noting that they are not joined in their brief by any state insurance commissioners. But it offers no reason for the court to question the authority of these attorneys general to represent their states' policy interests.
The opposing amici states argue in their brief that HUD's actions were also arbitrary and capricious because the agency erroneously determined that the Rule did not have “federalism implications,” which would have required HUD to “consult[] with State and local officials” as part of its rulemaking. (States' Amicus Opposing HUD at 13 (citing Exec. Order No. 13,132, § 6(c), 64 Fed.Reg. 43,255, 43,257-58 (Aug. 4, 1999))). In its 2023 rulemaking, HUD concluded that any potential federalism implications were “based solely on the assertion that this rule would interfere with McCarran-Ferguson,” 88 Fed.Reg. at 19,497, which is borne out by the administrative record. (See R. 029137 (comment by PCI's successor APCIA raising Executive Order 13,132 in the context of potential McCarran-Ferguson issues)). Because HUD addressed McCarran-Ferguson in depth, as discussed above, the fact that it did not consult state regulators is not an independent basis for discarding the Rule.
Other states, however, would welcome this added layer of federal protection in advancing their fair-housing objectives. (See States' Amicus Supporting HUD at 13-14.) These states also allow or permit risk-based decision-making under their insurance laws, but are willing to coordinate these laws with federal-court enforcement of the FHA and accept the policy results. The McCarran-Ferguson Act, as interpreted in Humana, gives them that prerogative. As both HUD and its supporting amici persuasively argue, it would be a perverse result indeed to instead use McCarran-Ferguson-a law that demands respect for states' individual policy choices-to cut off these divergent preferences at the pass. Cf. FTC v. Travelers Health Ass'n, 362 U.S. 293, 298-300 (1960) (observing that the McCarran-Ferguson Act was not intended to allow a state to “regulate activities carried on beyond its own borders”).
PCI offers several other objections to HUD's discussion on McCarran-Ferguson that merit only brief discussion. First, PCI argues that HUD arbitrarily failed to address how costly it would be for insurers to litigate the McCarran-Ferguson issue on a case-by-case basis. To a large extent, this argument tracks PCI's broader objection to the cost of case-by-case litigation that it raises elsewhere. The court will address this issue in greater depth below, but for now, it suffices to say that the agency's decision was not clearly irrational in light of the legal arguments already addressed. If, as HUD concluded, disparate-impact claims have a meaningful chance of surviving preemption, its decision to allow these claims their day in court was a reasonable policy trade-off.
See discussion infra Section II.C.
Second, PCI argues that the Supreme Court's intervening decision in West Virginia v. EPA, 597 U.S. __, 142 S.Ct. 2587 (2022) renders HUD's reaffirmation of a case-by-case approach for resolving McCarran-Ferguson issues arbitrary and capricious. (See Mot. for Leave to File Notice of Suppl. Authority [251].) West Virginia, however, is inapplicable here. As an initial matter, that case involved a dispute over the level of deference owed to an agency's interpretation of its enabling statute, not-as here-a question of whether the agency acted in an arbitrary and capricious manner during a rulemaking. Id. at 2599-2600. HUD's underlying statutory authority to promulgate the Rule (and to apply it to the insurance industry) is not an issue before the court at this time.
Beyond this, HUD's rulemaking does not contravene the broader principles expressed in West Virginia. The West Virginia Court disfavored agency actions that make “decisions of vast economic and political significance” based on “novel reading[s]” of statutes without clear Congressional authorization. Id. at 2605. And the Court has otherwise warned against agency “intru[sions] into an area that is the particular domain of state law.” Ala. Ass'n of Realtors v. Dep't of Health & Hum. Servs., 594 U.S. __, 141 S.Ct. 2485, 2489 (2021). HUD's actions here do not violate those prohibitions. Courts and lawmakers have been aware of the close relationship between homeowners insurance and fair housing objectives for decades. See 88 Fed.Reg. at 19,467 n.155 (outlining the “long and well documented” history of “discrimination in the homeowners insurance industry,” including multiple Congressional hearings); Am. Family, 978 F.2d at 298; Nationwide, 52 F.3d at 1351. HUD has consistently asserted its authority to regulate this important determinant of housing opportunity for over forty years, and Congress has never second-guessed this assertion. HUD's case-by-case approach does not trample on states' traditional prerogative to regulate the insurance industry and does not assume that McCarran-Ferguson will have no impact here. Compare 88 Fed.Reg. at 19,474 (acknowledging that “[s]ome discriminatory effects claims against insurers will be preempted under McCarran-Ferguson and some will not,” depending in part on the state law at issue), with Ala. Ass'n of Realtors, 141 S.Ct. at 2489 (striking down the CDC's attempt to regulate landlord-tenant relationships across all states through a single nationwide eviction moratorium).
See Dunn, 472 F.Supp. at 1109 & n.7 (noting that HUD has interpreted the FHA to prohibit insurance redlining since at least 1978); Am. Family, 978 F.2d at 300 (noting that Congress granted HUD rulemaking authority in its 1988 amendments to the FHA “with knowledge” of this interpretation); Implementation of the Fair Housing Amendments Act of 1988: Final Rule, 54 Fed.Reg. 3232, 3285 (Jan. 23, 1989) (codified at 24 C.F.R. § 100.70(d)(4)) (exercising this authority to codify HUD's interpretation of the FHA as prohibiting unequal provision of property or hazard insurance).
In sum, the court finds that HUD's explanation as to McCarran-Ferguson is within the “zone of reasonableness” necessary to pass muster. Prometheus Radio Project, 592 U.S. at 423. The McCarran-Ferguson problem is undoubtedly tricky, but HUD on its second try has shown how enough future minds could differ on its resolution to justify a case-by-case approach. Just as importantly, HUD engaged with a broad array of stakeholders in reaching this decision and considered their competing arguments. Some have raised legitimate concerns about the Rule's effect on state insurance regulations, but others have argued that the Rule would work no mischief in their administrative schemes. HUD's justification for allowing this question to be worked out one state at a time was well-supported by the principles of federalism for which McCarran-Ferguson stands.
B. Was HUD's Consideration of the Filed-Rate Doctrine Arbitrary and Capricious?
PCI's next challenge to HUD's Rule rests on the “filed-rate doctrine.” This common-law rule “forbids courts from invalidating or modifying rates that have been filed with regulatory agencies.” Schilke v. Wachovia Mortg., FSB, 820 F.Supp.2d 825, 835 (N.D. Ill. 2011), aff'd sub nom. Cohen v. Am. Sec. Ins. Co., 735 F.3d 601 (7th Cir. 2013). The doctrine has two animating concerns: a “historical antipathy to rate setting by courts, deemed a task they are inherently unsuited to perform competently,” and “a policy of forbidding price discrimination” between litigants and non-litigants. Arsberry v. Illinois, 244 F.3d 558, 562 (7th Cir. 2001); see also AT&T v. Cent. Off. Tel., Inc, 524 U.S. 214, 221-23 (1998). In its 2014 opinion, the court found the former concern “very similar to the purposes of the McCarran-Ferguson Act,” and noted that “where the filed-rate doctrine applies, the McCarran-Ferguson Act almost certainly applies too.” PCI, 66 F.Supp.3d at 1049. Thus, HUD's failure to adequately address the insurance industry's concerns regarding McCarran-Ferguson meant that it had also inadequately addressed concerns over the filed-rate doctrine. Id.
On remand, HUD addressed these concerns twice: first over the course of two pages in its 2016 Supplemental Response, and then in another two pages in its 2023 Reinstatement, which largely reiterated the same arguments made in 2016. 81 Fed Reg. at 69,017-18; 88 Fed.Reg. at 19,477-78. The agency pointed out that the doctrine has never been successfully used to defeat an FHA claim, and that several courts have rejected past attempts by insurers to do so. Id. at 19,478 (citing Saunders v. Farmers Ins. Exch. (“Saunders I”), 440 F.3d 940, 946 (8th Cir. 2006); DeHoyos, 345 F.3d at 297 n.5; Lumpkin v. Farmers Grp., Inc. (“Lumpkin I”), No. 05-2868 MA/V, 2007 WL 6996584, at *7-8 (W.D. Tenn. Apr. 26, 2007)). HUD described the “fit” between the doctrine and disparate-impact FHA liability as “attenuated, at best,” since disparate-impact claims “do not challenge the reasonableness of the insurance rates but rather their discriminatory effects.” 81 Fed.Reg. at 69,018 (citing Lumpkin I, 2007 WL 6996584, at *8). It also noted that, unlike McCarran-Ferguson preemption (which arises from federal statutory law), interpreting the filed-rate doctrine as allowing for state ratemaking regulations to trump federal antidiscrimination law would “stand the Supremacy Clause on its head.” Id. (quoting Perryman v. Litton Loan Servicing, LP, No. 14-CV-02261-JST, 2014 WL 4954674 (N.D. Cal. Oct. 1, 2014)). Finally, HUD called the underlying coherence of the doctrine itself into question, citing statements that “the law on the filed rate doctrine is extremely creaky,” Id. (citing Town of Norwood v. New England Power Co., 202 F.3d 408, 420 (1st Cir. 2000)), and that the doctrine is “weak and forcefully criticized,” Id. (citing Cost Mgmt. Servs. v. Wash. Nat'l Gas Co., 99 F.3d 937, 946 (9th Cir. 1996)). In a similar vein to its McCarran-Ferguson arguments, HUD reasoned that the doctrine's application is highly fact-dependent and may vary significantly based on the “particular state's ratemaking structures,” making a case-by-case approach more appropriate for accommodating these variations. Id.
HUD's first line of defense-that the Rule does not implicate the filed-rate doctrine at all- is unpersuasive. The agency splits hairs in its rulemaking and briefing by claiming that disparateimpact claims challenge the effects of discriminatory rates and the practices used to reach them, not their basic “reasonableness.” As PCI correctly points out, the caselaw on the doctrine has been extended to encompass both challenges to the reasonableness of rates and to the manner in which they were adopted. See, e.g., Square D Co. v. Niagara Frontier Tariff Bureau, Inc., 476 U.S. 409, 410 (1986) (applying doctrine to bar claims “alleg[ing] that rates filed with the Interstate Commerce Commission . . . were fixed pursuant to an agreement forbidden by the Sherman Act”). The doctrine's underlying concerns-disuniformity in rates charged to consumers, and courts' lack of institutional competency to set rates-could potentially arise in a disparate-impact FHA suit, since a court could be asked to evaluate (and potentially invalidate) the insurer's rate and the factors used to calculate it over the course of the burden-shifting process.
HUD's second argument regarding the Supremacy Clause, however, has more force. The filed-rate doctrine's origins can be traced back to the Supreme Court's decision in Keogh v. Chicago & Northwest Railway, 260 U.S. 156 (1922). The Keogh Court's holding rested on the principle that the antitrust laws require injury to “business or property” to establish standing, and the plaintiff could not prove that he was entitled to a lower rate. Id. at 163. While Keogh involved a challenge under federal law to a federal tariff-a rate filed with the Interstate Commerce Commission-its rule has since been applied to also bar federal-law challenges to rates set by state agencies. See, e.g., S. Branch LLC v. Commonwealth Edison Co., 46 F.4th 646, 650 (7th Cir. 2022); Wegoland Ltd. v. NYNEX Corp., 27 F.3d 17, 20 (2d Cir. 1994); Taffet v. S. Co., 967 F.2d 1483, 1494 (11th Cir. 1992). Importantly, though, these cases involved claims under either antitrust laws or the federal Racketeer Influenced and Corrupt Organizations Act (“RICO”)-both of which contain the same “business or property” requirement for standing. See 15 U.S.C. § 15(a); 18 U.S.C. § 1964(c).
In Saunders I, the Eighth Circuit held that Keogh's rule was inapplicable to a challenge to state-regulated insurance rates under the FHA. It held that, unlike the challenges to rates filed with federal agencies in Keogh and Square D, “the Supremacy Clause tips any legislative competition in favor of the federal anti-discrimination statutes” over “rates filed with a state regulatory agency.” 440 F.3d at 944. This issue, the Eighth Circuit reasoned, did not come into play in RICO or antitrust cases because of the “no-injury principle” inherent in those laws: a RICO or antitrust plaintiff would have no more standing to challenge a filed state rate than they would a federal rate. Id.; see Taffett II, 967 F.2d at 1494 (grounding the lack of distinction between federal and state rates in this rationale). But “standing to sue under . . . the [FHA] is far broader” and extends to any injury caused by a discriminatory housing practice. Saunders I, 440 F.3d at 944; see 42 U.S.C. §§ 3602(i), 3613(a) . Thus, the Saunders I court concluded that “the judicially created filed rate doctrine [should not] restrict Congress's broad grant of standing to seek judicial redress for race discrimination.” 440 F.3d at 946.
While PCI urges that Saunders I is “inconsistent with the weight of caselaw” holding that the filed-rate doctrine applies to state agency rates, see 88 Fed.Reg. at 19,477, it only cites RICO and antitrust cases in support of this point; it provides no authority-and the court has found none itself-to question the Eighth Circuit's more specific conclusion as to the FHA. The Saunders I court ultimately concluded that the determinative question in clashes between federal civil-rights laws and state-regulated insurance prices was not the filed-rate doctrine, but McCarran-Ferguson preemption, id at 945-a matter that, as explained here, HUD has adequately addressed.
Even failing this-assuming arguendo that Saunders I was wrongly decided, and that FHA claims against insurers may in fact be barred by the filed-rate doctrine-HUD has made a solid argument for handling such claims on a case-by-case basis. By its very name, the “filed-rate doctrine” comes into play only if a rate is actually filed with the relevant regulatory agency. And as both HUD and its supporting state amici note, states are all over the map on whether and when they require insurers to file. Some use a “prior approval” system that requires preclearance before an insurer can even begin to use their proposed rate, others follow a “use and file” approach where the insurer must file their rates within a set amount of time after deploying them, and still others do not require state filing whatsoever. See 81 Fed.Reg. at 69,018 n.92 (citing 2 Nat'l Ass'n of Ins. Comm'rs, Compendium of State Laws on Insurance Topics § II-PA-10-21 (2011)); States' Amicus Supporting HUD at 9 (citing Borselli, supra, at 126). Illinois, for example, is a “use and file” state, and while it does require insurers to submit their rates to its Department of Insurance, the Seventh Circuit has questioned whether the filed-rate even applies to propertyinsurance claims in this state, as “it is not at all clear that the Department has the authority to approve or disapprove” these rates once filed. 88 Fed.Reg. at 69,018 n.92 (citing Cohen, 735 F.3d at 607); see 50 Ill. Admin. Code §§ 754.10(a)(2), 754.40(a). In such states, the filed-rate doctrine may be a nonissue for FHA disparate-impact litigants.
This holding is in tension with the general majority rule that "[i]t is the filing of the tariffs, and not any affirmative approval or scrutiny by the agency, that triggers the filed rate doctrine.” S. Branch LLC, 46 F.4th at 653 (7th Cir. 2022) (citing Town of Norwood, 202 F.3d at 419); see also Goldwasser v. Ameritech Corp., 222 F.3d 390, 402 (7th Cir. 2000); In re Title Ins. Antitrust Cases, 702 F.Supp.2d 840, 853 (N.D. Ohio 2010); but see Brown v. Ticor Title Ins. Co, 982 F.2d 386, 394 (9th Cir. 1992) (finding filed-rate doctrine inapplicable in “[t]he absence of meaningful state review”). Regardless, there is clearly enough texture in states' administrative ratemaking schemes-and in the caselaw interpreting them-to make a uniform exemption inappropriate.
The court thus accepts HUD's explanation as to the filed-rate doctrine. The agency has presented substantial evidence that both the caselaw on the doctrine and the structures of states' administrative schemes vary significantly from jurisdiction to jurisdiction, and that at least some jurisdictions have found it completely inapplicable to FHA claims. This diversity makes it impossible to conclude as a general matter that the doctrine will be outcome-determinative in any given case. Accordingly, HUD's conclusion-that the filed-rate doctrine does not warrant a blanket exemption for risk-based pricing and underwriting of insurance-was not arbitrary and capricious.
C. Was HUD's Consideration of the Nature of Insurance Arbitrary and Capricious?
PCI's next challenge to HUD's Rule sounds not in law, but in policy. The argument is a familiar one: that risk-based decision-making and disparate-impact liability are fundamentally incompatible, and that attempting to mix them will sow chaos in the insurance markets. The Seventh Circuit recognized this conflict in NAACP v. American Family Mutual Insurance: “Insurance works best when the risks in the pool have similar characteristics.... Risk discrimination is not race discrimination. Yet efforts to differentiate more fully among risks may produce classifications that could be generated by discrimination.” Am. Family, 978 F.2d at 290. In its 2014 opinion, the court agreed with PCI that HUD's original rulemaking had failed to meaningfully engage with this thorny issue. See PCI, 66 F.Supp. at 1051.
Commenters raised these concerns yet again during HUD's 2023 rulemaking. They noted that the insurance industry “is predicated on setting rates and making underwriting decisions based on relevant, mathematical, and objective risk factors that accurately predict loss”; that this has been a “bedrock principle of state insurance regulation for more than 150 years”; and 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.” 88 Fed.Reg. at 19,466. And they warned that forcing insurers to abandon risk-based decision-making would cause adverse selection by “overcharging low-risk customers and likely driving them from the markets,” resulting in higher premiums for everyone. Id.
HUD responded to these arguments in its 2023 Reinstatement. It argued that the industry's concerns were overblown and that the Rule did not seek to interfere with legitimate and necessary risk-based practices. HUD cited the “long and well documented” history of discrimination in the homeowners insurance industry, beginning with overt racial discrimination and evolving into more “covert” forms over time. 88 Fed.Reg. at 19,467-68 & nn.155. As an example, HUD pointed out that insurers once denied communities of color access to insurance based on purportedly neutral factors like property age, value, and location, even though “data demonstrated that such restrictions were not justified by risk of loss.” Id. at 19,468 & nn.156-57. This history, HUD argued, demonstrated how many seemingly “objective” actuarial factors are in fact infected by “non-actuarially-based subjective judgment” and may be subject to “discretion under state law,” such as state statutes that allow insurers to rely on “judgment factors” in ratemaking. Id. at 19,468. Even factors that are genuinely objective “can vary by context” in their relevance and “may change over time as societal behaviors evolve, new technologies develop, and analytical capabilities improve.” Id. at 19,469. And many insurance practices-“such as marketing, claims processing, and payment”-involve no actuarial decision-making whatsoever. Id. at 19,468; see Huskey, 2023 WL 5848164, at *11. HUD reiterated the point that the “business necessity” defense under the burden-shifting framework protects objective risk-based practices for which there are no less-discriminatory alternatives. 88 Fed.Reg. at 19,467.
HUD also disputed concerns that the Rule would cause insurance markets to fail. HUD again asserted that the Rule simply reaffirms jurisprudence that already allowed for disparateimpact suits, and that its first ten years of existence have not seen the tidal wave of litigation that insurers predicted. Id. at 19,468. It also claimed that several states-namely, Connecticut, California, and Ohio-separately provide for disparate-impact liability against insurers under their own laws, and that the insurance industry has not collapsed in these jurisdictions. Id. at 19,467. And it noted that “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.” Id.
To summarize HUD's response, then, the agency asserts that the Rule does not per se prohibit discrimination in rate setting based upon risk of loss. It instead simply allows litigants to challenge whether insurance practices are truly founded in risk, and if so, whether there are effective less-discriminatory alternatives that do not harm the insurer's business. And according to the agency, the Rule's foundations in existing law undermine contentions that it will lay waste to the insurance industry.
In its earlier ruling, this court was unmoved by HUD's argument that the burden-shifting framework by itself would afford the opportunity for insurers to justify their “risk-based” practices as legitimate business necessities. See PCI, 66 F.Supp.3d at 1050-51. Commenters raised the issue again in 2023, making the claim that any “less discriminatory alternative” identified by a plaintiff at step three “would necessarily correspond to a different risk than the factor at issue, identified through actuarially sound methodology.” 88 Fed.Reg. at 19,467. PCI picks up this torch in its briefing, arguing that HUD's claim that insurers can choose among “alternative risk- based practices” is fanciful and that the Rule will require “many if not most risk-based practices . . . to be eliminated from the underwriting process.” (Pl.'s SOF ¶¶ 96-97 (quoting R. 028694).)
But this time around, HUD has added a new layer of analysis largely absent from its earlier rulemaking-the notion that “risk” is not a monolith. Rather, the business of insurance involves a myriad of different practices, not all of them governed by strict objectivity. At the outside, some practices (like marketing and claims processing) rely little or not at all on actuarial analysis. See 88 Fed.Reg. at 19,468; Huskey, 2023 WL 5848164, at *11. Venturing further inward, other practices (such as ratemaking and underwriting) aspire towards objectivity but may in fact incorporate a significant level of subjective discretion-both in their initial selection of risk factors, and in opportunities for business considerations to “override” actuarial calculations at a later stage in the process. See 88 Fed.Reg. at 19,468, 19,472-73; Dane, supra, at 4-5 (R. 004407-08) (noting that “[u]nderwriting guidelines are typically not the result of careful, statistical studies . . . [but] [r]ather . . . are often based on hunches and subjective stereotypes about classes of consumers and types and geographic location of property,” and that state laws often allow rate filers to “modif[y], or ignore[]” actuarial conclusions based on “business judgment”). And at the “core” of the industry, some risk-based factors may genuinely be the only way for insurers to price and provide their products in a cost-effective manner, even if these factors necessarily correlate with race. HUD denies that the Rule poses any danger to these core factors. See 88 Fed.Reg. at 19,467. Its Rule is instead meant to give litigants the ability to parse out these distinctions and ensure that the factors insurers rely on are in fact legitimate ones.
This added nuance also makes HUD's theory perfectly compatible with American Family. That case declined to “resolve the question” of whether to allow FHA disparate-impact claims against insurers in light of the complex policy issues presented by the risk-based nature of insurance. Am. Family, 978 F.2d at 290; see 88 Fed.Reg. at 19,471 n.186 (making this same point). HUD has presented a theory for doing so here that, though not indisputable, is within the bounds of reasoned decision-making.
In light of this, PCI has not shown that it was unreasonable for HUD to refuse to carve out an exemption for “actuarial or risk-based calculations.” (Pl.'s SOF ¶ 100). As the agency points out, there appears to be no clear or principled way of defining such an exemption ex ante, and there will likely always some slippage at the margins where subjectivity (and, potentially, bias) creeps in. 88 Fed.Reg. at 19,469. HUD's historical summary drives this point home. It shows that the homeowners insurance industry has previously raised risk as a shield for purportedly “objective” factors (like property location and age) without clearly substantiating that these factors had any meaningful relationship to the cost of providing insurance. Id. at 19,468 & n.156-57. There are countless other ways in which similar issues with insurers' methodologies might arise in the future. Drawing this line will always require highly fact-intensive analysis. HUD's conclusion that this can best be done on a case-by-case basis was not unreasonable.
One example that HUD provides is “price optimization,” or the use of quantitative tools to set an optimal price point for goods or services based on the seller's business considerations. Insurers are increasingly using price optimization to deviate their rates from actuarial cost models based on the price elasticity of demand (the measure of how sensitive customers will be to a price change). See 88 Fed.Reg. at 19,468 (citing Casualty Actuarial & Statistical (C) Task Force, Price Optimization White Paper, Nat'l Ass'n of Ins. Comm'rs 9 (Nov. 19, 2015), https://content.naic.org/sites/default/files/inline-files/committeesccatfrelated price optimizationwhitepaper.pdf). HUD does not drive this point fully home by explaining how price optimization specifically results in discriminatory effects on protected classes under the FHA. But the example still supports the more general point that ”[t]here is not, and never has been, an absolute homage to actuarial outcomes in the business or regulation of insurance.” Dane, supra, at 6 (R. 004409).
As the court has already acknowledged, inviting federal courts to engage in this analysis raises a wholly separate set of issues. PCI's arguments on this score reveal complications for HUD: using the burden-shifting framework to assess which practices are sufficiently objective and necessary to constitute a legitimate “business justification” relieves concerns over the fundamental nature of insurance, but sharpens the institutional competence and federalism concerns embedded in McCarran-Ferguson and the filed-rate doctrine. Courts have for years wrestled with the genuine tension between these principles. Having reviewed the administrative record, however, the court is satisfied that the agency has presented a well-reasoned case for letting the courts-in dialogue with state regulators-continue to play a role in working out this balance.
See discussion supra Sections II.A-.B.
PCI is correct that, as this court observed, litigation of these issues can be costly, and should not proceed at all in cases where the plaintiffs have no reasonable chance to prevail. PCI, 66 F.Supp.3d at 1051 (citing Graoch, 508 F.3d at 375-76). But the agency did consider such costs to insurers in its most recent rulemaking, and presented a valid case for why these claims would not be fruitless. See 88 Fed.Reg. at 19,478. As the agency's cited sources illustrate, insurers have “been subject to discriminatory effects liability since well before the 2013 Rule,” albeit not consistently and not across all jurisdictions. See 88 Fed.Reg. at 19,463 n.116, 19,464 n.132 (citing cases). At least some of these cases found it worthwhile to interrogate whether certain practices-like setting underwriting requirements based on factors like maximum dwelling age, minimum value, or replacement cost-evaluated risk “fairly and objectively” or were “not purely risk-based.” Prudential, 208 F.Supp.2d at 60; see also Toledo Fair Hous. Ctr., 94 Ohio Misc.2d at 163 (denying summary judgment based on “substantial questions of fact as to whether the [defendant's] underwriting guidelines in question are justified by business necessity”). HUD further notes that, while it “is unaware of any trial on the merits of a discriminatory effects claim against an insurer,” many of these past legal challenges have survived motions to dismiss or motions for summary judgment and ultimately resulted in settlements or consent orders through which insurers agreed to adjust their practices in order to reduce potential harms to protected classes.88 Fed.Reg. at 19,471 n.187. In light of this varied body of caselaw, HUD did not act unreasonably in refusing to create a “categorical bar[]” based on a “generalized application of the burden-shifting framework.” Graoch, 508 F.3d at 376.
As a corollary point to its cost argument, PCI also complains that the Rule will require insurers to collect data on race and ethnicity in order to better defend against future disparate-impact suits. (See, e.g., Pl.'s SOF ¶ 121.) The court has already expressed its skepticism of this argument in addressing the justiciability of PCI's claims. See discussion supra Section I.A. Nothing in the Rule requires collection of such data. See 24 C.F.R. § 100.500; 88 Fed.Reg. at 19,480; NAMIC, 2023 WL 6142257, at *9. In a disparate-impact challenge, it is the plaintiff who has the burden of presenting such evidence at step one, and a defendant may rebut this evidence in “numerous other ways, including by showing that the data put forward by plaintiff is incorrectly or wrongly analyzed,” or by proving legitimate business necessity. 88 Fed.Reg. at 19,480. And if the Rule does spur PCI's members to collect at least some additional data on the racial impacts of their policies, such data could be relevant in a disparate-treatment challenge as well and could help insurers reduce costs from disparate-impact suits over time. See 88 Fed.Reg. at 19,469 (noting that “[t]his kind of self-examination . . . is intended not to lead to liability, but rather to protect entities from liability”).
See, e.g., Consent Decree, United States v. Nationwide Mut. Ins. Co., No. C2-97-291 (Mar. 10, 1997), https://www.justice.gov/crt/housing-and-civil-enforcement-cases-documents-367 (requiring insurer to revise policies making coverage decisions based on maximum dwelling age or minimum value); DeHoyos v. Allstate Corp., 240 F.R.D. 269, 276 (W.D. Tex. 2007) (approving settlement requiring insurer to change credit scoring formula to reduce disparate impact on minority policyholders and applicants); Press Release, Nat'l Fair Hous. All., National Fair Housing Alliance Settles Disparate Impact Lawsuit with Travelers Indemnity Company (Feb. 23, 2018), https://nationalfairhousing.org/travelers (describing settlement prohibiting defendant insurer from considering rentals to Section 8 voucher holders in determining insurance coverage for rental properties).
For the same reasons, HUD adequately supported its conclusion that applying the Rule to risk-based insurance practices will not cause widespread market failures. Not all of HUD's points on this front hit the mark: as already noted, the absence of increased litigation under the Rule itself cannot be disentangled from its political instability over the past decade. Accordingly, the agency's repeated refrain that the Rule simply reaffirms preexisting caselaw fails to realistically acknowledge the possible consequences of its reinstatement. And while HUD cites several cases suggesting that disparate-impact claims against insurers are viable under certain states' laws, this is too small a dataset to draw any meaningful conclusions about the basic nature of the industry. Further, while the agency makes an offhand reference to other risk-based contexts in which disparate-impact liability has proven workable, such as mortgage lending, it does not substantiate this point with any sources or show how these industries are comparable to the insurance industry's deep dependence on risk-based methodologies.
HUD specifically cites the decisions in Viens, Jones, and Toledo Fair Housing Center as evidence that “some states specifically provide for discriminatory effects liability against insurers under state law, further undermining the claim that providing for such liability as a matter of federal law threatens the fundamental nature of the industry.” 88 Fed.Reg. at 19,467 & n.153 (first citing Viens, 113 F.Supp.3d at 573 n.20; then citing Jones, 2015 WL 5091908, at *5; and then citing Toledo Fair Hous. Ctr., 94 Ohio Misc.2d at 157-59, 704 N.E.2d at 670-72). These cases do support the more abstract proposition that disparate-impact liability can complement, rather than frustrate, states' administrative schemes, but they do not provide sufficient foundation for the empirical point that HUD tries to make here.
Instead, it is the agency's review of the history and caselaw on disparate-impact claims that satisfies the court. PCI's parade of horribles about the Rule's possible effects on the insurance business merely resurrects claims that insurers have been making for decades to ward off liability. See 88 Fed.Reg. at 19,468 n.158; see also, e.g., Nevels v. W. World Ins. Co., Inc., 359 F.Supp.2d 1110, 1123 (W.D. Wash. 2004) (rejecting claims that “applying the FHA to surplus lines insurers would ‘wreak havoc on the states' regulatory scheme, restrict the availability of insurance coverage, and put the courts in the improper role of making complex decisions on insurance rates and coverage'”). The Rule could have at least some of these ill effects, but any prediction on this front is ultimately just that, and HUD acted within the scope of its expertise by concluding, based on relevant experience, that the Rule did not pose an existential threat to the industry.
Here, in its second bite at the apple, HUD has supplied a “logical rationale” for declining to provide insurers with a categorical exemption under the Rule. The court declines to “substitute [its] own policy judgment for that of the agency.” Wolf, 962 F.3d at 230. It is certainly possible that filtering insurers' risk-based practices through a disparate-impact analysis will impose costs on the industry that would not otherwise exist. But achieving the FHA's remedial goal of “provid[ing], within constitutional limitations, for fair housing throughout the United States” can be costly for any sector subject to the Act, from landlords to mortgage lenders to local governments. 42 U.S.C. § 3601. HUD did not abuse its discretion in reaching the conclusion that it did here.
D. Was HUD's Consideration of Inclusive Communities Arbitrary and Capricious?
PCI's final attack on HUD's Rule goes beyond the grounds laid out in the court's 2014 decision. Citing the Supreme Court's intervening decision in Inclusive Communities, PCI argues that HUD's choice to maintain and ultimately reinstate the 2013 Rule was additionally arbitrary and capricious in light of the new limits on disparate-impact liability that the Supreme Court allegedly established in that decision.
As a brief recap, in its 2017 order addressing PCI's motion to amend its complaint, this court refused to allow PCI to add a new claim that HUD's Rule was “contrary to law” under the FHA. PCI, 2017 WL 2653069, at *8-9. The court did, however, allow PCI to add a claim concerning Inclusive Communities and to address “how, if at all, Inclusive Communities affects the various issues the Court ordered PCI to consider on remand.” Id. at *9 n.4. The court's analysis here, then, focuses on that question.
In Inclusive Communities, the Supreme Court held that while disparate-impact claims were cognizable under the FHA, they should be limited to challenges against “artificial, arbitrary, and unnecessary barriers.” 576 U.S. at 543 (quoting Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971)). The Court further cautioned that “[c]ourts should avoid interpreting disparate-impact liability to be so expansive as to inject racial considerations into every housing decision.” Id. And the Court held that plaintiffs who base their claim on “a statistical disparity” must satisfy a “robust causality” requirement by “point[ing] to a defendant's policy or policies causing that disparity.” Id. at 542.
The practical effect of these phrases, and their relationship to HUD's Rule, is still a matter of dispute. On one hand, a number of federal courts-including this court in 2017-have concluded that Inclusive Communities did not disturb the Rule's basic three-step burden-shifting approach. PCI, 2017 WL 2653069, at *8-9 (finding that the Court “did not identify any aspect of HUD's burden-shifting approach that required correction”); see also, e.g., de Reyes v. Waples Mobile Home Park Ltd. P'ship, 903 F.3d 415, 424, 432 n.10 (4th Cir. 2018); MHANY Mgmt. Inc. v. Cnty. of Nassau, 819 F.3d 581, 618 (2d Cir. 2016); Ave. 6E Invs., LLC v. City of Yuma, 818 F.3d 493, 512-13 (9th Cir. 2016). On the other hand, the Fifth Circuit held that Inclusive Communities “undoubtedly announce[d] a more demanding test than that set forth in the HUD regulation,” but acknowledged a lack of consensus on the precise contours of this test. Inclusive Cmtys. Project v. Lincoln Prop. Co., 920 F.3d 890, 902-04 (5th Cir. 2019).
In its 2023 Reinstatement, HUD (unsurprisingly) took the position that “the 2013 Rule is consistent with Inclusive Communities” and denied any need to amend its language based on the Supreme Court's decision. 88 Fed.Reg. at 19,459. The agency addressed both comments related to Inclusive Communities in general (over the course of six pages in the Federal Register, see id. at 19,457-62), and comments related to the decision's implications for the insurance industry in particular (over an additional two pages, see id. at 19,469-70). HUD noted that the Supreme Court did not specifically address insurance in its decision, and that its expansive description of the FHA's purpose-to “eradicate discriminatory practices within a sector of our Nation's economy”-counseled against categorically exempting any particular industry from the Rule. Id. at 19,464 (citing Inclusive Cmtys., 576 U.S. at 539); see also id. at 19,469 n.166. The Reinstatement rejected the notion that requiring insurers to consider protected traits in the rating and underwriting process would run afoul of Inclusive Communities' warning against “injecting race into housing decisions.” Id. at 19,469. And it argued that Inclusive Communities' supposed “limits” could be read as restating the existing requirements in the burden-shifting framework, rather than separate and heightened new pleading standards-meaning that insurers could still raise “business necessity” as a defense to disparate-impact suits. Id. at 19,461-62, 19,470.
The court will not vacate or remand HUD's 2023 Rule based on its discussion of Inclusive Communities. To do so would be inconsistent with both this court's earlier (2017) ruling and with the “large body of case law holding that insurers . . . can be held liable under the FHA,” which “Inclusive Communities does not call . . . into question.” Travelers Indem., 261 F.Supp.3d at 29; see also, e.g., Huskey, 2023 WL 5848164, at *7-9; NAMIC, 2023 WL 6142257, at *8-12. HUD did not, in its Final Rule, include a discrete section specifically responding to the court's 2017 directive to consider Inclusive Communities' relevance to the other issues presented on remand- namely, the McCarran-Ferguson Act, the filed-rate doctrine, and the nature of insurance. There is, nevertheless, enough discussion of equivalent topics over the course of the agency's rulemaking to find the requisite “rational connection” satisfied. State Farm, 463 U.S. at 43.
PCI raises three specific challenges to HUD's Rule based on phrases in the Inclusive Communities decision, but none move the court. First, PCI claims that the kinds of risk-based insurance practices for which it is seeking an exemption are not the kinds of “artificial, arbitrary, and unnecessary” barriers that Inclusive Communities says the FHA was designed to prevent. 576 U.S. at 543. But this language is consistent with HUD's description of the Rule's intended use: to sift out objective and necessary risk-based practices from those that are unsupported by actuarial science or replaceable by a less discriminatory alternative. See 88 Fed.Reg. at 19,470; Travelers Indem., 261 F.Supp.3d at 30 (“[I]nsurance policies can create exactly the type of ‘artificial, arbitrary, and unnecessary barriers' to housing that disparate-impact liability is suited to address.”).
In any event, Inclusive Communities quoted the “artificial, arbitrary, and unnecessary” language from its earlier decision in Griggs, 401 U.S. at 431, and it is not clear that the Court meant this as an additional pleading requirement rather than simply a recital of established disparate-impact principles. See 88 Fed.Reg. at 19,462 (reasoning that this language merely referenced a “short-hand formulation for the type of policy that traditionally has been held to create an unjustified discriminatory effect”); Sw. Fair Hous. Council, Inc. v. Maricopa Domestic Water Improvement Dist., 17 F.4th 950 n.9 (9th Cir. 2021) (declining to address whether such a requirement exists); but see Ellis v. City of Minneapolis, 860 F.3d 1106, 1114 (8th Cir. 2017) (requiring plaintiff to plead facts sufficient to establish existence of “artificial, arbitrary, and unnecessary” policy). Insurers can certainly raise the “artificial, arbitrary, and unnecessary” language as a defense in jurisdictions that have taken a more expansive view of this language, and may well prevail on this basis in litigation-but the state of the caselaw here does not warrant a blanket nationwide exemption on this basis.
Second, PCI argues that the Rule clashes with Inclusive Communities' “robust causality requirement” by allowing plaintiffs to hold insurers liable “based solely on a showing of a statistical disparity.” 576 U.S. at 540. The court does not read the Rule as permitting such a result. HUD acknowledged in its 2023 rulemaking that “the [R]ule do[es] not use the precise words ‘robust causality'” but also observed that “[w]hat Inclusive Communities requires is [not particular language but] that a court's examination of causality be robust.” 88 Fed.Reg. at 19,461. Thus, under the burden-shifting framework, a plaintiff must “link a specific practice to a current or predictable disparity” as part of the prima facie case. Id.; see 24 C.F.R. § 100.500(c)(1); NAMIC, 2023 WL 6142257, at *11; see also Travelers Indem., 261 F.Supp.3d at 34 (holding that plaintiffs had pleaded sufficient facts to establish that insurer's policy of refusing to cover properties with Section 8 tenants was causally linked to racial and gender-based disparities).
PCI also briefly argues that HUD “staked out internally inconsistent positions” on Inclusive Communities' “policy or policies” requirement, by stating at one point in its rulemaking that a “specific practice” is required, 88 Fed.Reg. at 19,461, but later taking the position that plaintiffs could challenge “the decision-making process as a whole,” Id. at 19,485. The court sees no obvious inconsistency in these two statements. A regulated entity's overarching decision-making process can, if sufficiently pervasive, be a challengeable “policy” under Inclusive Communities. See Cnty. of Cook v. Bank of Am. Corp., No. 14 C 2280, 2018 WL 1561725, at *9 (N.D. Ill. Mar. 30, 2018), aff'd on other grounds, 78 F.4th 970 (7th Cir. 2023) (noting that there is no requirement that “the ‘specific' practice challenged in a disparate-impact claim must be limited to a single component”).
And, PCI claims, it will be impossible to prove that insurers' practices are the cause of discriminatory effects, because insurers' discretion is limited by state law. But the diversity of both state regulatory schemes and individual insurers' risk-based practices, discussed earlier, undermines this argument. See 88 Fed.Reg. at 19,472-73, 19,475; States' Amicus Supporting HUD at 9. Even if all states permit or require risk-based practices under their laws, as PCI repeats several times, “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.” 88 Fed.Reg. at 19,472. For example, many states allow insurers to override actuarial calculations based on subjective “judgment factors.” Id. at 19,473. Whatever added causality standard Inclusive Communities sets forth beyond the existing requirements of the burden-shifting framework (if any), it goes too far to presume that all claims against insurers would necessarily fail this standard.
Finally, PCI's argument that HUD arbitrarily failed to consider Inclusive Communities' guidance against “inject[ing] racial characteristics into every housing decision” goes nowhere. 576 U.S. at 543. As both this court and the NAMIC court have already noted, the DisparateImpact Rule does not “require those engaging in housing practices to collect or use data on individuals' protected characteristics.” NAMIC, 2023 WL 6142257, at *9; see discussion supra Section 1.A. Even if the Rule does, in practice, spur insurers to collect this kind of data to safeguard themselves against litigation risk, “[t]his 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.” 88 Fed.Reg. at 19,469; see NAMIC, 2023 WL 6142257, at *9 (deeming the plaintiff's similar argument “a complaint about any sort of disparate-impact liability that might apply to insurance practices”). Voluntary self-analysis is nothing like the “racial quotas” that troubled the Inclusive Communities Court-indeed, the Court specifically noted that “awareness of race” could be used to “encourage revitalization of communities that have long suffered the harsh consequences of segregated housing patterns.” 576 U.S. at 545. Given the long history of discriminatory insurance redlining that has contributed to these patterns, as HUD summarized at length in its rulemaking, see 88 Fed.Reg. at 19,467 n.155, it was reasonable for the agency to conclude that this language in Inclusive Communities does not warrant an insurance exemption.
CONCLUSION
Reconciling risk-based insurance practices with disparate-impact liability under the FHA implicates questions of fairness, efficiency, and federalism that have troubled both courts and commentators for decades. HUD's summary decision to apply its Disparate-Impact Rule to the insurance industry in 2013 gave short shrift to this complex and nuanced debate.
But the explanation that HUD puts before this court now-after nine years, multiple administrative proceedings, and dozens of pages in the Federal Register-is considerably more complete. This time around, HUD considered the concerns that the court posed, engaged with relevant stakeholders, marshaled others' expertise as well as its own, and documented its reasoning in detail. That HUD ultimately reached the same conclusion does not make this outcome less legitimate, any more than the court's 2014 remand mandated any particular outcome. If the court had believed that no amount of reappraisal could have cured the defects in HUD's reasoning, it would simply have vacated the rule. Instead, it gave the agency an opportunity to take a “hard look” at the relevant issues and present a well-reasoned case for applying disparate-impact liability to the insurance industry. That is all the arbitrary-and-capricious standard demands, and HUD has amply done so here.
Accordingly, the court denies PCI's motion for summary judgment [282] and grants HUD's cross-motion for summary judgment [295].