City of Miami v. Wells Fargo & Co.

Case: 14-14544 Date Filed: 05/03/2019 Page: 1 of 76 [PUBLISH] IN THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT ________________________ No. 14-14544 Non-Argument Calendar ________________________ D.C. Docket No. 1:13-cv-24508-WPD CITY OF MIAMI, a Florida municipal corporation, Plaintiff - Appellant, versus WELLS FARGO & CO., WELLS FARGO BANK, N.A., Defendants - Appellees. ________________________ No. 14-14543 Non-Argument Calendar ________________________ D.C. Docket No. 1:13-cv-24506-WPD CITY OF MIAMI, a Florida Municipal Corporation, Case: 14-14544 Date Filed: 05/03/2019 Page: 2 of 76 Plaintiff - Appellant, versus BANK OF AMERICA CORPORATION, BANK OF AMERICA, N.A., COUNTRYWIDE FINANCIAL CORPORATION, COUNTRYWIDE HOME LOANS, COUNTRYWIDE BANK, FSB, Defendants - Appellees. ________________________ Appeals from the United States District Court for the Southern District of Florida ________________________ (May 3, 2019) Before MARCUS and WILSON, Circuit Judges, and SCHLESINGER, * District Judge. MARCUS, Circuit Judge: This pair of ambitious fair housing lawsuits brought by the City of Miami against major financial institutions returns to our Court after having been appealed to the Supreme Court and resolved there in Bank of American Corp. v. City of Miami, 137 S. Ct. 1296 (2017). Miami alleges that, for years, the defendant institutions, major nationwide banks, carried on discriminatory lending practices that intentionally targeted black and Latino Miami residents for predatory loans. * Honorable Harvey E. Schlesinger, United States District Judge for the Middle District of Florida, sitting by designation. 2 Case: 14-14544 Date Filed: 05/03/2019 Page: 3 of 76 The City says this resulted in disproportionate foreclosures on homeowners of those races, diminished property values in predominantly minority neighborhoods, substantially reduced tax revenue for the City, and increased expenditures by the City for municipal services. When we first heard these cases, we determined that Miami had standing under the Fair Housing Act, and that it had adequately pled proximate cause. See City of Miami v. Bank of Am. Corp., 800 F.3d 1262 (11th Cir. 2015); City of Miami v. Wells Fargo & Co., 801 F.3d 1258 (11th Cir. 2015). The Supreme Court agreed in part. It resolved the hotly contested standing issue in the City’s favor, but vacated and remanded with regard to proximate cause. See Bank of Am., 137 S. Ct. at 1305–06. The Court held that the standard that this panel had applied -- foreseeability - - was not enough on its own to demonstrate proximate cause. Id. at 1306. Instead, the Court said that proximate cause under the FHA also required “some direct relation between the injury asserted and the injurious conduct alleged.” Id. at 1306 (quotations omitted). But the Court declined to “draw the precise boundaries of proximate cause under the FHA and to determine on which side of the line the City’s financial injuries fall.” Id. It remanded the case, preferring to leave this issue open for percolation in the lower courts. See id. Today, we take up the question of how the principles of proximate cause identified by the Court’s opinion 3 Case: 14-14544 Date Filed: 05/03/2019 Page: 4 of 76 function when applied to the FHA and to the facts as alleged in the City’s complaints. At this preliminary stage in the lawsuit, we conclude that the City has adequately pled proximate cause in relation to some of its economic injuries when the pleadings are measured against the standard required by the Fair Housing Act. Proximate cause asks whether there is a direct, logical, and identifiable connection between the injury sustained and its alleged cause. If there is no discontinuity to call into question whether the alleged misconduct led to the injury, proximate cause will have been adequately pled. The question for now is whether, accepting the allegations as true, as we must, the City has said enough to make out a plausible case -- not whether it will probably prevail. Considering the broad and ambitious scope of the FHA, the statute’s expansive text, the exceedingly detailed allegata found in the complaints, and the application of the administrative feasibility factors laid out by the Supreme Court in Holmes v. Securities Investor Protection Corp., 503 U.S. 258 (1992), we are satisfied that the pleadings set out a plausible claim. The City’s pleadings meet this standard because Miami has alleged a substantial injury to its tax base that is not just reasonably foreseeable, but also is necessarily and directly connected to the Banks’ conduct in redlining and reverse- redlining throughout much of the City. This injury plausibly bears “some direct 4 Case: 14-14544 Date Filed: 05/03/2019 Page: 5 of 76 relation” to the claimed misconduct. Bank of Am., 137 S. Ct. at 1306. The injury to the City’s tax base is uniquely felt in the City treasury, and there is no risk that duplicative injuries could be pled by another plaintiff or that the apportionment of damages amongst different groups of plaintiffs would be a problem. As we see it, the City is in the best position. Indeed only the City can allege and litigate this peculiar kind of aggregative injury to its tax base. Simply put, a lawsuit commenced by an individual homeowner cannot challenge the Banks’ policies on the same citywide scale that the alleged misconduct took place on. However, the City’s pleadings fall short of sufficiently alleging “some direct relation” between the Banks’ conduct and a claimed increase in expenditures on municipal services. The complaints fail to explain how these kinds of injuries -- increases in police, fire, sanitation, and similar municipal expenses -- are anything more than merely foreseeable consequences of redlining and reverse-redlining. The Court has told us that foreseeability alone is not enough. We do not mean to suggest that the City’s claims are destined to succeed. Many questions, and many difficult questions, remain and will have to be worked out in the district court. At the motion to dismiss stage, though, we are not asking whether the complaints meet any probability requirement, only whether they plausibly allege violations of the FHA. Since we have found that they do, we allow this discrete portion of the City’s claims to proceed for now. The plaintiff 5 Case: 14-14544 Date Filed: 05/03/2019 Page: 6 of 76 has said enough to get into the courthouse and be heard. We decide nothing more today. I. Background A. The City’s Claims On December 13, 2013, the City of Miami brought three complex civil rights actions in the Southern District of Florida against several different financial institutions. One suit was filed against Bank of America Corporation, Bank of America N.A., Countrywide Financial Corporation, Countrywide Home Loans, and Countrywide Bank, FSB (collectively “Bank of America”), and another against Wells Fargo & Co. and Wells Fargo Bank, N.A. (collectively “Wells Fargo”). For simplicity, we refer to all these defendants jointly as “the Banks.” These were accompanied by another similar case against Citigroup, Inc. and related institutions. See City of Miami v. Citigroup, Inc., 801 F.3d 1268 (11th Cir. 2015). The first time this panel considered this set of cases, we heard the Citigroup case as well, but that case was not appealed to the Supreme Court. It has returned to the district court, where it has been stayed pending resolution of the other two. See Order Staying Case Pending the Supreme Court’s Disposition of Matters Now Before the Court, City of Miami v. Citigroup Inc., No. 13-cv-24510 (S.D. Fla. July 13, 2016). As a result, our opinion today concerns only Bank of America and Wells Fargo. 6 Case: 14-14544 Date Filed: 05/03/2019 Page: 7 of 76 The City alleged in considerable detail that the Banks had violated § 3604(b) and § 3605(a) of the Fair Housing Act. The first of these provisions makes it unlawful “[t]o discriminate against any person in the terms, conditions, or privileges of sale or rental of a dwelling, or in the provision of services or facilities in connection therewith, because of race, color, religion, sex, familial status, or national origin.” 42 U.S.C. § 3604(b). The second states that “[i]t shall be unlawful for any person or other entity whose business includes engaging in residential real estate–related transactions to discriminate against any person in making available such a transaction, or in the terms or conditions of such a transaction, because of race, color, religion, sex, handicap, familial status, or national origin.” Id. § 3605(a). The City alleged that the Banks had violated these provisions by intentionally engaging in discriminatory mortgage lending practices that resulted in a disproportionate and excessive number of defaults by black and Latino homebuyers and caused substantial financial harm to the City. 1 1 The City also alleged that the Banks unjustly enriched themselves by taking advantage of “benefits conferred by the City” at the same time that they engaged in unlawful lending practices which “denied the City revenues it had properly expected through property and other tax payments and . . . cost[] the City additional monies for services it would not have had to provide . . . absent [the Bank’s] unlawful activities.” Wells Fargo, 801 F.3d at 1261 (quoting Complaint for Violations of the Federal Fair Housing Act at 61, City of Miami v. Wells Fargo & Co., No. 13-24508-CIV (S.D. Fla. July 9, 2014), 2013 WL 6903725); Bank of Am., 800 F. 3d at 1267; see also Complaint for Violations of the Federal Fair Housing Act at 52–55, City of Miami v. Bank of Am. Corp., No. 13-24506-CIV (S.D. Fla. July. 9, 2014), 2013 WL 6903721 (raising the same claims and allegations). The district court held that these claims failed. City of Miami v. Bank of Am., No. 13-24506-CIV, 2014 WL 3362348 at *6–7 (S.D. Fla. July 9, 2014). We affirmed, Bank of Am., 800 F.3d at 1287; Wells Fargo, 801 F.3d at 1267, and the City did not appeal these claims to the Supreme Court. Accordingly, they are no longer in this case. 7 Case: 14-14544 Date Filed: 05/03/2019 Page: 8 of 76 The City said the Banks made a practice of systematically and intentionally refusing to extend credit to minority borrowers on the same terms as they would extend credit to non-minority borrowers and, when they did extend credit to comparably situated minority borrowers, doing so only on predatory terms, worse than the terms non-minorities received. First Amended Complaint for Violations of the Federal Fair Housing Act at 2–5, City of Miami v. Bank of Am., No. 13-cv- 24506 (S.D. Fla. Sept. 9, 2014) (“BoA FAC”); First Amended Complaint for Violations of the Federal Fair Housing Act at 2–5, City of Miami v. Wells Fargo & Co., No. 13-cv-24508 (S.D. Fla. Sept. 9, 2014) (“WF FAC”). 2 This amounted to 2 The operative complaints for our purposes are the First Amended Complaints in both cases. As we explain, infra, the City appealed the denial of motions in both cases asking the district court to reconsider the dismissal of its initial Complaints and to grant leave to file the First Amended Complaints. After this panel reversed the decisions in September 2015, litigation proceeded in the district court for some months before the Supreme Court granted the Banks’ petitions for certiorari in June 2016. On remand from this Court, the City filed what it styled “Second Amended Complaints,” (even though the First Amended Complaints had not technically been filed), and both of these were dismissed on multiple grounds, but with leave to amend. Order Granting Motion to Dismiss Second Amended Complaint, Bank of Am., No. 13-cv-24506 (Mar. 17, 2016) (No. 98); Order Granting Motion to Dismiss Second Amended Complaint, Wells Fargo, No. 13-cv-24508 (Mar. 17, 2016) (No. 77). By the time the Supreme Court granted the petition for certiorari, the City had filed Third Amended Complaints, and both Banks had moved to dismiss these as well. See Defendants’ Motion to Dismiss Third Amended Complaint with Prejudice and Request for Hearing, Bank of Am., No. 13-cv-24506 (May 16, 2016) (No. 103); Wells Fargo’s Motion to Dismiss with Prejudice the City’s Third Amended Complaint and Request for Hearing, Wells Fargo, No. 13-cv-24508 (May 24, 2016) (No 83). Both cases were then stayed pending resolution in the Supreme Court. Order Staying Case Pending the Supreme Court’s Disposition of Matters Now Before the Court, Bank of Am., No. 13-cv-24506 (July 13, 2016) (No. 128); Order Granting Unopposed Motion to Stay Further Proceedings Pending the Supreme Court’s Disposition of Matters Now Before the Court, Wells Fargo, No. 13-cv-24508 (July 13, 2016) (No. 102). Since nothing had been appealed since the Second and Third Amended Complaints were filed, the Supreme Court was explicit that it was looking to the First Amended Complaints, see Bank of Am., 137 S. Ct. at 1301, and we do the same on remand. 8 Case: 14-14544 Date Filed: 05/03/2019 Page: 9 of 76 both “redlining” (refusing to extend credit) and “reverse redlining” (extending credit on worse terms). Black and Latino borrowers were thus unable to refinance their loans effectively. BoA FAC at 4; WF FAC at 4; see also Wells Fargo, 801 F.3d at 1261; Bank of Am., 800 F. 3d at 1267. The City’s complaints detailed at considerable length the nature of these practices and characterized them as “abusive” because they resulted in loans with “unfair terms that [borrowers] could not afford.” BoA FAC at 13; WF FAC at 11; see, e.g., BoA FAC at 22 (explaining what kinds of loans the City would categorize as “predatory”); WF FAC at 34 (same). The City claims that these policies were entirely deliberate on the part of the Banks. The Amended Complaints explain that confidential witnesses -- former employees at both Bank of America and Wells Fargo -- will testify to how minorities and residents of minority neighborhoods were targeted for predatory loan terms. E.g., BoA FAC at 19; WF FAC at 29. Moreover, witnesses allegedly will testify that Wells Fargo specifically targeted Latino and African American community groups and churches (but never white churches) and that employees were assigned based on their race to make presentations to these groups. WF FAC at 30. Witnesses from Bank of America likewise will testify that they were encouraged to steer less financially savvy borrowers, often racial minorities, toward loan products that were decidedly unfavorable to the customer and different 9 Case: 14-14544 Date Filed: 05/03/2019 Page: 10 of 76 from loan products offered to white applicants, but highly favorable to the bank. BoA FAC at 19–20. According to the City, all of this violated the FHA in two stages. To start, the Banks intentionally discriminated against minority borrowers by targeting them for burdensome loans. This had a disparate impact on minority borrowers, leading, down the line, to a disproportionate number of exploitative loans in minority neighborhoods and, eventually, to a disproportionate number of foreclosures on minority-owned properties. BoA FAC at 15–19; WF FAC at 21–29. The City employed detailed regression analyses of the Banks’ self-reported data to show that a Bank of America loan in a predominantly African-American or Latino Miami neighborhood was 5.857 times more likely to result in foreclosure than a Bank of America loan in a non-minority Miami neighborhood. BoA FAC at 26. For a minority-neighborhood loan from Wells Fargo, foreclosure was 6.975 times more likely. WF FAC at 39. Further calculations indicated that, controlling for credit history and other factors, a black Miami borrower was 1.581 times more likely than a white borrower to receive a predatory loan from Bank of America and 4.321 times more likely to receive one from Wells Fargo. BoA FAC at 22; WF FAC at 34. A Latino borrower was 2.087 times more likely to receive such a loan from Bank of America and 1.576 times more likely to receive one from Wells Fargo. BoA FAC at 22; WF FAC at 34. Notably, even among borrowers with 10 Case: 14-14544 Date Filed: 05/03/2019 Page: 11 of 76 good credit (FICO scores over 660), black borrowers were 1.533 and 2.572 times more likely to receive a predatory loan from Bank of America and Wells Fargo, respectively. BoA FAC at 22; WF FAC at 34. For Latino borrowers with good credit, the figures were 2.137 and 1.875 times more likely. BoA FAC at 22; WF FAC at 34. The City identified similar disproportionalities on a larger scale by comparing predominantly white neighborhoods with predominantly minority ones. See BoA FAC at 23–25; WF FAC at 35–37. The complaints further alleged that the Banks’ lending practices resulted in minority borrowers suffering especially fast and frequent foreclosures, an important indicator of predatory lending practices. See BoA FAC at 29–30; WF FAC at 41. White borrowers’ average time from origination to foreclosure was 3.448 years for a Bank of America loan or 3.266 years for a Wells Fargo loan; for black and Latino borrowers the averages were 3.144 (black) and 3.090 (Latino) years from Bank of America and 2.996 (both black and Latino) from Wells Fargo. BoA FAC at 28; WF FAC at 41. Confidential witnesses from both Banks will allegedly support the claims that each had deliberately targeted black and Latino borrowers for predatory loans. E.g., BoA FAC at 19–21; WF FAC at 30–31. What’s more, according to the complaints, the Banks’ misconduct was not limited to loan origination but instead continued almost to the point of foreclosure. As the Supreme Court noted, Bank of Am., 137 S. Ct. at 1301, the predatory 11 Case: 14-14544 Date Filed: 05/03/2019 Page: 12 of 76 practices alleged by the City included denying black and Latino customers opportunities to refinance or make loan modifications on fair terms. E.g. BoA FAC at 21; WF FAC at 32. According to the complaints, both Banks would cause foreclosures when “a minority borrower who previously received a predatory loan sought to refinance the loan, only to discover that [the bank] refused to extend credit at all,” or refused to do so on the same terms they extended when “refinancing similar loans issued to white borrowers.” BoA FAC at 4; WF FAC at 4. We note that such refinancing and loan modification decisions can occur before or after a property enters default. The City alleged that it suffered both economic and noneconomic injuries. BoA FAC at 31–35; WF FAC at 44–48. In both complaints the alleged noneconomic injury was that the Banks’ conduct “adversely impacted the racial composition of the City and impaired the City’s goals to assure racial integration and desegregation and the social and professional benefits of living in an integrated society.” 3 BoA FAC at 31; WF FAC at 44. The City identified two forms of economic injuries. It claimed damages from each bank based on reduced property tax revenues, arguing that the Banks’ lending policies caused minority-owned 3 Because we find that the City adequately pled proximate cause for one of its economic injuries, we need not evaluate whether it also did so for its noneconomic injuries. The parties in their briefing have not squarely addressed these injuries, and the Supreme Court had little to say about them. We leave this matter to the district court to address in the first instance. 12 Case: 14-14544 Date Filed: 05/03/2019 Page: 13 of 76 properties to fall into unnecessary or premature foreclosure, causing them to lose substantial value and, in turn, decreasing the values of surrounding properties. BoA FAC at 32–34; WF FAC at 45–47. This reduced the City’s tax base and the tax revenue it took in. According to the City, “Hedonic regression” techniques could be used to quantify with considerable particularity what portion of its losses was attributable to the Banks’ conduct. BoA FAC at 32–34; WF FAC at 45–47. Additionally, the City claimed damages based on the cost of the increased municipal services it provided to deal with problems attributable to the foreclosed and often vacant properties -- including police, firefighters, building inspectors, debris collectors, and others. These increased services too, the City claimed, would not have been necessary if the properties had not been foreclosed upon as a result of the Banks’ discriminatory lending practices. BoA FAC at 34–35; WF FAC at 47–48. The City also asked for a declaratory judgment stating that the Banks’ conduct violated the FHA, an injunction barring the Banks from engaging in similar predatory conduct, compensatory and punitive damages, and attorneys’ fees. BoA FAC at 40–41; WF FAC at 53–54. 13 Case: 14-14544 Date Filed: 05/03/2019 Page: 14 of 76 B. District Court and 11th Circuit Decisions On July 9, 2014, the district court granted motions to dismiss by both banks. 4 City of Miami v. Bank of Am., No. 13-24506-CIV, 2014 WL 3362348 (S.D. Fla. July 9, 2014). The court determined first that the City lacked standing because, according to this Court’s opinion in Nasser v. City of Homewood, 671 F.2d 432 (11th Cir. 1982), its claims fell outside the “zone of interests” protected by the FHA. Bank of Am., 2014 WL 3362348 at *3–4. Miami had alleged “merely economic injuries” that were not “affected by a racial interest.” Id. at *4. The trial court also determined that the FHA contained a proximate cause requirement, and that this requirement had not been met because the City failed to “demonstrate that the [Banks’] alleged redlining and reverse redlining caused the foreclosures to occur.” Id. at *5. According to the district court, the City should have “allege[d] facts that isolate [the Banks’] practices as the cause of any alleged lending disparity.” Id. Additionally, the district court said that the FHA claims fell outside the statute of limitations. Id. at *6–7. The City moved for reconsideration and for leave to file amended complaints in both cases, arguing that it had standing and that it could remedy the statute of limitations issue by amending the complaint. See Bank of Am., 800 F.3d 4 The court issued a detailed opinion in the Bank of America case. It then adopted and incorporated that opinion in the Wells Fargo case. City of Miami v. Wells Fargo & Co., No. 13-24508-CIV (S.D. Fla. July 9, 2014). 14 Case: 14-14544 Date Filed: 05/03/2019 Page: 15 of 76 at 1271; Wells Fargo, 801 F.3d at 1264. The amended complaints added further details concerning noneconomic injuries suffered by the City as a result of the Banks’ discriminatory lending practices. See BoA FAC at 31; WF FAC at 44. The City also argued in its amended complaints that predatory lending had “adversely impacted the racial composition of the City,” “impaired the City’s goals to assure racial integration and desegregation,” and “frustrate[d] the City’s longstanding and active interest in securing the benefits of an integrated community.” BoA FAC at 31; WF FAC at 44. The district court denied the motion, and the City appealed to this Court. Bank of Am., 800 F.3d at 1271; Wells Fargo, 801 F.3d at 1265. We reversed and remanded in a series of opinions, the most detailed of which we issued in City of Miami v. Bank of America Corp., 800 F.3d 1262 (11th Cir. 2015). We began by addressing standing and held that the City had alleged an injury in fact sufficiently concrete and immediate to confer Article III standing. Id. at 1272. Next we addressed “statutory standing” -- a misnomer, since it really concerns “whether the plaintiff has a cause of action under the statute.” Id. at 1273 (quotation omitted). We determined that this plaintiff did because the key statutory language “swe[pt] as broadly as allowed under Article III” and placed the City within the statute’s zone of interests. Id. at 1278; see also Trafficante v. Metro 15 Case: 14-14544 Date Filed: 05/03/2019 Page: 16 of 76 Life Ins. Co., 409 U.S. 205, 209 (1972) (identifying “a congressional intention to define [FHA] standing as broadly as is permitted by Article III”). We next held that the FHA did require proximate cause, but that the City’s pleadings were adequate. Bank of Am., 800 F.3d at 1279–80, 1283. We rejected the idea that the City needed to “allege that the [Banks’] actions directly harmed the City.” Id. As we saw it, the “strict directness requirement” suggested by the Banks could not be the proper standard for the FHA because it “would run afoul of Supreme Court and Eleventh Circuit caselaw allowing entities who have suffered indirect injuries . . . to bring a claim under the FHA.” Bank of Am., 800 F.3d at 1281 (citing Gladstone, Realtors v. Village of Bellwood, 441 U.S. 91, 109–11 (1979) (allowing a village to sue a firm for discriminatory practices that caused segregation); Havens Realty Corp. v. Coleman, 455 U.S. 363, 378–79 (1982) (allowing suit by a non-profit for “impairment of its organizational mission” and “drain on its resources”); Baytree of Inverrary Realty Partners v. City of Lauderhill, 873 F.2d 1407, 1408–09 (11th Cir. 1989) (allowing a non-minority developer to challenge a zoning decision as discriminatory)). Proximate cause, we said, was “not a one-size fits-all analysis.” Id. Rather, it “can differ statute by statute.” Id. Instead, noting the Supreme Court’s “broad and inclusive” readings of the FHA, e.g., Trafficante, 409 U.S. at 209, we agreed with the City that the proper 16 Case: 14-14544 Date Filed: 05/03/2019 Page: 17 of 76 standard for proximate cause was “foreseeability.” Id. at 1281–82. Damages claims under the FHA were analogous to tort claims, we reasoned, and tort law had traditionally relied on foreseeability as the standard by which to evaluate causal relationships and to limit the expansion of liability. See id. at 1282. The City had met this standard; “it claim[ed] that the [Banks’] discriminatory lending caused property owned by minorities to enter premature foreclosure, costing the City tax revenue and municipal expenditures.” Id. at 1282. We said that “[a]lthough there are several links in that causal chain, none are unforeseeable,” and also noted that the regression analyses in the complaints made the pleadings more than just speculative. Id. We made no comment, though, on “whether the City will be able to actually prove its causal claims.” Id. at 1283 (emphasis added). That would be worked out in the district court on a factual basis far beyond the four corners of the complaint. Accordingly, we remanded the case. Id. at 1289. Bank of America and Wells Fargo each filed petitions for certiorari, which the Supreme Court granted. It consolidated the two cases. C. Supreme Court Decision The Supreme Court began by resolving the “statutory standing” question of whether the City had a cause of action under the FHA. Bank of Am., 137 S. Ct. at 1302–05. The Court agreed that its previous cases had interpreted the statute broadly for standing purposes. See id. Congress had even amended the statute 17 Case: 14-14544 Date Filed: 05/03/2019 Page: 18 of 76 after these decisions, without making any change to the relevant language, indicating its assent to the Court’s expansive reading of the original text. Id. at 1303–04. On proximate cause, though, the Supreme Court disagreed with this panel but reached no firm conclusions. It determined that foreseeability, standing alone, was not a sufficiently rigorous standard. See id. at 1306 (“[T]he Eleventh Circuit erred in holding that foreseeability is sufficient to establish proximate cause under the FHA.”). The Court explained that “foreseeability alone does not ensure the close connection that proximate cause requires.” Id. Since lending and housing policies are deeply interconnected with “economic and social life,” the Court said that “[a] violation of the FHA may . . . ‘be expected to cause ripples of harm to flow’ far beyond the defendant’s misconduct.” Id. (quoting Assoc. Gen. Contractors of Cal., Inc. v. Carpenters, 459 U.S. 519, 534 (1983)). Since “[n]othing in the statute suggests that Congress intended to provide a remedy wherever those ripples travel,” foreseeability alone could not be enough. Id. While the Court was clear that foreseeability was not enough, it was less definitive when it came to laying out what more FHA proximate cause required. Its key instructions were these: Rather [than foreseeability], proximate cause under the FHA requires “some direct relation between the injury asserted and the injurious conduct alleged.” Holmes v. Sec. Investor Protect. Corp., 503 U.S. 258, 268 (1992). A damages claim under the statute “is analogous to a 18 Case: 14-14544 Date Filed: 05/03/2019 Page: 19 of 76 number of tort actions recognized at common law,” Curtis v. Loether, 415 U.S. 189, 195 (1974), and we have repeatedly applied directness principles to statutes with “common-law foundations,” Anza v. Ideal Steel Supply Corp., 547 U.S. 451, 457 (2007). “The general tendency” in these cases is, “in regard to damages at least, is not to go beyond the first step.” Hemi Grp., LLC v. City of New York, 559 U.S. 1, 10 (2010). What falls within that “first step” depends in part on the “nature of the statutory cause of action,” Lexmark Int’l, Inc. v. Static Control Components, Inc., 134 S. Ct. 1377, 1390 (2014), and an assessment “of what is administratively possible and convenient,” Holmes, 503 U.S. at 268. Bank of Am., 137 S. Ct. at 1306 (citation formats altered). The Court remanded for further proceedings, stating that “[t]he lower courts should define, in the first instance, the contours of proximate cause under the FHA and decide how that standard applies to the City’s claims for lost property-tax revenue and increased municipal expenses.” Id. Lacking “the benefit of [the Eleventh Circuit’s] judgment” on how to apply the principles it had laid out, and with no other circuits having weighed in yet, the Court “decline[d]” to speak first. Id. In this opinion, we endeavor carefully to apply the Court’s mandate to these complaints, to determine if they plausibly state a claim under the Fair Housing Act. II. Standard of Review We review a district court’s decision to dismiss a complaint de novo “accepting the factual allegations in the complaint as true and construing them in the light most favorable to the plaintiff.” Boyd v. Warden, Holman Corr. Facility, 19 Case: 14-14544 Date Filed: 05/03/2019 Page: 20 of 76 856 F.3d 853, 863–64 (11th Cir. 2017); see also Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). Here, the district court also denied leave to amend, which we generally review for abuse of discretion; however, when, as here, leave to amend was denied “on the grounds of futility,” we review the denial de novo “because it is a conclusion of law that an amended complaint would necessarily fail.” Id. (“An amendment is considered futile when the claim, as amended, would still be subject to dismissal.”). Essentially, then, we must consider de novo whether the City’s amended complaints, which the district court declined to entertain, have stated a claim. This is in line with the review conducted by the Supreme Court, Bank of Am., 137 S. Ct. at 1302, and with our review when we first heard these cases, Bank of Am., 800 F.3d at 1271. For this case to proceed past a motion to dismiss, we need not find that the Banks’ actions in fact proximately caused the plaintiff’s injuries; we must find that the City plausibly alleged that they did so. We evaluate whether each complaint “contain[s] sufficient factual matter . . . to ‘state a claim to relief that is plausible on its face.’” Iqbal, 556 U.S. at 678 (quoting Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007)). The standard is “plausibility”; it is decidedly not a “probability requirement.” Id. Still, plausibility requires that allegations push past the line of showing “a sheer possibility” and “[w]here a complaint pleads facts that are ‘merely consistent with’ a defendant’s liability, it ‘stops short of the line 20 Case: 14-14544 Date Filed: 05/03/2019 Page: 21 of 76 between possibility and plausibility of entitlement to relief.’” Id. (quoting Twombly, 550 U.S. at 557). Facts are taken as true so long as they are not a “formulaic recitation of the elements” or “conclusory.” Id. at 681 (quoting Twombly, 550 U.S. at 555). III. Analysis Prosser and Keeton observed, of proximate cause, that “[t]here is perhaps nothing in the entire field of law which has called forth more disagreement, or upon which the opinions are in such a welter of confusion.” W. Page Keeton et al., Prosser and Keeton on the Law of Torts § 41, at 263 (5th ed. 1984). The Supreme Court has not directed the lower courts to propose a sweeping standard for proximate cause that would apply across all cases or to define the word “direct” for all time. Indeed, we are not looking for (and could not find) “a black-letter rule that will dictate the result in every case.” Holmes, 503 U.S. at 272 n.20. Rather, we must evaluate where, as a matter of judicial economy and policy, we ought to draw the line for this particular cause of action and these particular claims. As the Court has said, proximate cause “label[s] generically the judicial tools used to limit a person’s responsibility for the consequences of that person’s own acts. . . . [P]roximate cause reflects ideas of what justice demands, or of what is administratively possible and convenient.” Id. at 268 (quotation omitted). 21 Case: 14-14544 Date Filed: 05/03/2019 Page: 22 of 76 We begin our analysis, then, with the Supreme Court’s instructions. Proximate cause under the FHA certainly requires foreseeability, which is present here but is not enough. See Bank of Am. 137 S. Ct. at 1305–06. Proximate cause also requires “some direct relation between the injury asserted and the injurious conduct alleged.” Id. at 1306 (quoting Holmes, 503 U.S. at 268). Our task today is to determine what plaintiffs must do in order to plausibly allege “some direct relation” and to evaluate whether the City’s complaints have met that standard. At the outset, we consider the central phrase “some direct relation.” There is give in the joints between “some direct relation” and “some direct causation.” These are not identical concepts, and so when, for example, Bank of America suggests that Miami’s injuries “were not caused ‘directly’ by a loan,” it may not be presenting the question in a way that precisely and accurately reflects the Court’s instruction. We might agree that the injuries were not “caused directly by a loan” and yet still find “some direct relation” between the injury and the statutory violations. Causation is “the act or process of causing,” and to “cause” something is “to serve as cause or occasion of” or “to bring [it] into existence.” Webster’s Third New International Dictionary 356 (2002). Relation, on the other hand is “the mode in which one thing or entity stands to another, itself, or others,” or “a logical bond.” Id. at 1916. We are considering “direct relation,” as a critical aspect of “proximate cause,” so some palpable causation is required. We ought not forget, 22 Case: 14-14544 Date Filed: 05/03/2019 Page: 23 of 76 though, that foreseeability, while insufficient on its own, remains a requirement and ensures some causal connection. “Some direct relation,” then, works to guarantee that there is a “logical bond” between violation and injury. Put another way, while foreseeability ensures “cause,” “some direct relation” ensures that the cause is sufficiently “proximate.” Further, the law requires “some direct relation” not any quantifiable amount of it. The standard is softened by the modifier “some,” meaning, “of an unspecified but appreciable or not inconsiderable quantity, amount, extent or degree.” Id. at 2171. The requirement is therefore somewhat easier to meet than if the Court had said we needed to find “a direct relation.” We are also aware that our analysis must be tied to the Fair Housing Act in a specific way. The Supreme Court twice emphasized that the policy judgments that shape our proximate cause analysis will necessarily depend on the FHA. For starters, “what falls within [the] ‘first step,’” to which proximate cause is “general[ly]” constrained, will “depend[] in part on the ‘nature of the statutory cause of action.’” Bank of Am., 137 S. Ct. at 1306. Second, our task is to “define, in the first instance, the contours of proximate cause under the FHA and” to apply that standard to the City’s claims. Id. (emphasis added). After thoroughly reviewing the City’s complaints, we are satisfied that we can find “some direct relation,” a meaningful and logical continuity, between the 23 Case: 14-14544 Date Filed: 05/03/2019 Page: 24 of 76 City’s tax revenue injury and the Banks’ conduct by following the four guiding principles the Court outlined in Bank of America. We begin by considering (a) “what falls within [the] ‘first step’” of the causal chain, as we are aware of “‘[t]he general tendency’ in these cases . . . ‘not to go beyond [that] first step.’” Id. (quoting Hemi Grp., 559 U.S. at 10). What falls within the first step will, we’re told, depend on (b) “the nature of the statutory cause of action,” and (c) “an assessment of what is administratively possible and convenient.” Id. Finally, since the common law is the basis for the direct relation requirement, we also look to (d) the FHA’s common-law antecedents to the extent that we can. Id. We consider each of these principles in turn. A. An intervening step does not vitiate proximate cause The Court has also told us that “[t]he general tendency in these cases, in regard to damages at least, is not to go beyond the first step.” Bank of Am. 137 S. Ct. at 1306 (quotations omitted). Both Banks, in their briefing on remand, make much of this “first step” analysis. It is clear, though, that proximate cause does not always cut off at the first step after a violative act. A general tendency is not the same as a hard and fast rule that dictates the outcome in every case, and Supreme Court precedent shows that an intervening step will not vitiate proximate cause in all instances. What is more important, precedent reveals, is the certainty with which we can say the injury is fairly attributable to the statutory violation. An 24 Case: 14-14544 Date Filed: 05/03/2019 Page: 25 of 76 extended causal chain often makes attribution more difficult, but may not be the final word. 1. Supreme Court precedent makes crystal clear that an intervening step does not necessarily mean proximate cause has not been plausibly alleged. Lexmark International, Inc. v. Static Control Components, Inc., 134 S. Ct. 1377 (2014), is a good starting point precisely because the chain of causation there was complex and involved more than one simple step. The basic facts were these: Lexmark manufactured and sold both laser printers and toner cartridges for those printers. Lexmark, 134 S. Ct. at 1383. Other companies would buy used cartridges, then refurbish and resell them. Id. The plaintiff company, Static Control, made and sold the components that the refurbishers needed to operate. Id. at 1384. Lexmark would have preferred that its customers return used cartridges directly to Lexmark, so that it could refurbish and resell them itself. See id. at 1383. Static Control sued under the Lanham Act, alleging that Lexmark falsely and misleadingly led Lexmark’s customers to believe that they were legally obligated to return spent cartridges to Lexmark and also led the refurbishers to believe that their businesses were illegal. Id. at 1384. This, said Static Control, violated the Lanham Act’s prohibition on “false or misleading representation of fact . . . in commercial advertising or promotion [that] misrepresents the nature, characteristics, [or] 25 Case: 14-14544 Date Filed: 05/03/2019 Page: 26 of 76 qualities . . . [of] another person’s goods, services, or commercial activities.” Id. (quoting 15 U.S.C. § 1125(a)). The Supreme Court granted certiorari in Lexmark to resolve issues of “statutory standing” under the Lanham Act, including proximate cause. See id. at 1385, 1390. The central question was “whether the harm alleged has a sufficiently close connection to the conduct the statute prohibits.” Id. An “intervening step of consumer deception” between Lexmark’s alleged conduct and the harm sustained by Static Control was not fatal to the claim. Id. at 1391. The Court said that a Lanham Act plaintiff could demonstrate proximate cause by “show[ing] . . . injury flowing directly from the deception wrought by the defendant’s advertising,” including, critically, “deception of consumers caus[ing] them to withhold trade from the plaintiff.” Id. (emphasis added). Common-law principles supported the idea that “a plaintiff [could] be directly injured” even when a misrepresentation was made to a third party. Id. Liability was not without limits, though. A company injured only by a third party’s “inability to meet [its] financial obligations” could not sue. Id. at 1392 (quotation omitted). The Court acknowledged, as it would again in Bank of America, the “general tendency” not to go “beyond the first step.” Id. at 1394; see Bank of Am., 137 S. Ct. at 1306. In the circumstances presented by Lexmark, though, the nature of the conduct and the harm sustained persuaded the Court against following this 26 Case: 14-14544 Date Filed: 05/03/2019 Page: 27 of 76 general tendency. See Lexmark, 134 S. Ct. at 1394. The Court explained its reluctance to invariably limit liability to the first step this way: [T]he reason for that general tendency is that there ordinarily is a “discontinuity” between the injury to the direct victim and the injury to the indirect victim, so that the latter is not surely attributable to the former (and thus also to the defendant’s conduct), but might instead have resulted from “any number of other reasons.” Id. (quoting Anza, 547 U.S. at 458–59). This “discontinuity” is evaluated in terms of logical relations and policy judgments. See Anza, 547 U.S. at 459–60. In Lexmark, if the refurbishers sold fewer cartridges, “it would follow more or less automatically” that Static Control suffered a proportional injury, “without the need for any ‘speculative . . . proceedings’ or ‘intricate, uncertain inquiries.’” Lexmark, 134 S. Ct. at 1394 (quoting Anza, 547 U.S. at 459–60). There was no “discontinuity” problem in Lexmark because of the close connection between Static Control and the remanufacturers it supplied. “Any false advertising that reduced the remanufacturers’ business necessarily injured Static Control as well.” Id. There was “something very close to a 1:1 relationship” between harm to remanufacturers and harm to Static Control because the false advertising hurt Static Control in direct proportion to how much it hurt the remanufacturers. Id. In other words, harm to Static Control was “so integral an aspect of the [violation] alleged” that the Court went beyond the first step in the causal chain. Id. 27 Case: 14-14544 Date Filed: 05/03/2019 Page: 28 of 76 Still, Lexmark cabined liability by requiring “economic or reputational injury flowing directly from the [defendant’s] deception.” Id. at 1391. This requirement would not be met “when the deception produces injuries to a fellow commercial actor that in turn affect the plaintiff.” Id. By way of example, the Court suggested that a competitor whose business failed because of false advertising could sue the false advertiser, but that competitor’s landlord could not sue the false advertiser for the value of rent payments he could no longer collect. See id. The landlord’s injury suffered from a “discontinuity” that Static Control’s injury did not. See id. at 1394. All of this goes to show that intervening steps in a causal chain cannot automatically and invariably end the analysis. If they could, there would be no meaningful way to distinguish Static Control from the hypothetical landlord. Each is only injured if the defendant’s direct competitor (the refurbishers on the actual facts of the case) is injured first. The chain of causation is an important part of the analysis, but the real deciding factor, at least for this Lanham Act case, was whether there was a discontinuity between violation and injury, or, put another way, whether the injurious conduct “necessarily injured [the plaintiff] as well” as the first-step victim, in such a way that the first-step victim was “not [a] more immediate victim[]” than the plaintiff. Id. Thus we may look to whether the injury 28 Case: 14-14544 Date Filed: 05/03/2019 Page: 29 of 76 is “surely attributable” to the statutory violation; we do not simply count the steps in the causal chain. Id. Bridge v. Phoenix Bond & Indemnity Co., 553 U.S. 639 (2008), a RICO mail fraud case, makes the same point. In a suit between bidders at County- operated tax lien auctions, the defendants were accused of filing fraudulent documents in order to increase their success at the auctions at the expense of the plaintiffs. See Bridge, 553 U.S. at 642–44. To the extent that a chain of causation was discussed at all, it was explained in terms of reliance. See id. at 657–58. The defendants argued that the County, not the plaintiffs, had relied on their misrepresentations. Id. at 648. The Court held that the plaintiffs’ injuries were proximately caused by the defendants’ misrepresentations even though the plaintiffs had not themselves relied on the fraudulent filings -- “first-party reliance” was not “necessary to ensure that there was a sufficiently direct relationship between the defendant’s wrongful conduct and the plaintiff’s injury” to satisfy proximate cause. Id. at 657–58 (emphasis added). The “direct relationship” in Bridge turned on a connection between harm and injury that was non-formalistic and disconnected from step-counting. The decision instead rested on policy considerations, see id. at 654–55, and on common-law principles, see id. at 656–57. The Court never mentioned a “first step” default rule, but it’s clear that the plaintiffs’ harm would not have fallen in 29 Case: 14-14544 Date Filed: 05/03/2019 Page: 30 of 76 the first step, since the County had to rely on the defendants’ misrepresentations and then had to conduct auctions on these fraudulent terms before the plaintiffs could be injured. Instead, Bridge observed that the Second Restatement of Torts “does not say that only those who rely on the misrepresentation” -- that is, those at the front of the chain of causation -- “can suffer a legally cognizable injury.” Bridge, 553 U.S. at 656 (citing 3 Restatement (Second) of Torts §548A (1976)). While neither Lexmark nor Bridge arose under the FHA, their approach to the first-step rule is instructive. Both decisions acknowledge that “under common- law principles, a plaintiff can be directly injured by a misrepresentation even where” someone else relied on it. Lexmark, 134 S. Ct. at 1391 (citing Bridge, 553 U.S. at 639). This is a foundational principle of proximate cause; it is not an idiosyncratic wrinkle associated with a particular statute or type of claim. E.g., Blue Shield of Virginia v. McCready, 457 U.S. 465, 482–84 (1982) (discussing “antitrust injury” as a requirement of proximate cause under certain statutes). 2. Turning to our case, the Banks argue that the City fails to allege proximate cause simply because its injuries fall outside of the “first step.” So wooden a rule would be inconsistent with precedent as well as with the oft-noted statement that, when it comes to proximate cause, it is “virtually impossible to announce a black- letter rule that will dictate the result in every case.” Assoc. Gen., 549 U.S. at 536 30 Case: 14-14544 Date Filed: 05/03/2019 Page: 31 of 76 (citing Blue Shield, 457 U.S. at 477 n.13; Palsgraf v. Long Island R. Co., 162 N.E. 99, 103 (N.Y. 1928) (“[W]hat is a proximate cause, depend[s] in each case upon many considerations . . . .”) (Andrews, J., dissenting)); see also Lexmark, 134 S. Ct. at 1390 (“The proximate-cause inquiry is not easy to define . . .”); Bridge, 553 U.S. at 659 (“[P]roximate cause is generally not amenable to bright-line rules.”); Holmes, 503 U.S. at 272 n.20. Proceeding beyond a first step here is consistent with the instruction that we stop at the first step only as a “general tendency,” and that we also consider the nature of the cause of action, principles of continuity, and administrative feasibility. Bank of Am., 137 S. Ct. at 1306. Wells Fargo nevertheless argues that the City’s claims cannot survive proximate cause analysis because they are wholly “contingent,” and points to Holmes v. Securities Investor Protection Corp., 503 U.S. 258 (1992), as an example of a case where proximate cause failed “because the [plaintiffs’] injuries were ‘purely contingent on the harm suffered by [third parties].’ [Holmes, 503 U.S.] at 271.” But Holmes established no such rule. Rather, it explicitly identified three distinct considerations by which proximate cause should be evaluated. Holmes, 503 U.S. at 269. We discuss these in considerable detail, infra, but for now it is enough to observe that contingency on third parties is not among them. 31 Case: 14-14544 Date Filed: 05/03/2019 Page: 32 of 76 The Court would have had no recourse to policy considerations if it could have disposed of Holmes merely based on the involvement of a third party. 5 Indeed, if Wells Fargo were right about the application of the first-step rule here, this case would be over at the pleading stage simply because the chain of causation involves more than a single step. And there would have been no reason for the Supreme Court to remand this case back to our Court for further proximate cause analysis. The Supreme Court, too, was obviously aware that individual homeowners were involved in this case. Again, if that alone were enough to bar proximate cause, the Court would hardly have sought the input of the lower federal courts. The essential point for us then is that the “general tendency” to stop at the first step is just that, a general tendency, not an inexorable rule. If, in fact, there were a hard and fast “only the first step” rule limiting liability, a plaintiff homeowner who was forced into foreclosure on account of a predatory bank loan that violated the Fair Housing Act would never be able to plausibly allege that the foreclosure was proximately caused by the bank’s predation. By the Banks’ lights, there are two critical steps in the chain of 5 We also reject the application of Anza v. Ideal Steel Supply Corp, 547 U.S. 451, 457 (2006), that Wells Fargo has offered in its briefing on remand. It cites Anza for its proposed rule that proximate cause cannot exist where the causes of a plaintiff’s injury are distinct from the violative conduct. But this amounts to simply assuming the answer to our question. Whether the causes of the City’s injury are too distinct and too remote from the violative conduct is at the heart of the proximate cause inquiry. Furthermore, the Bank’s suggested rule overlooks that Anza applied the factors drawn from Holmes. Anza, 547 U.S. at 459–60. 32 Case: 14-14544 Date Filed: 05/03/2019 Page: 33 of 76 causation between the act of redlining and foreclosure: the middle and distinct step being a homeowner’s default. So inexorable a rule would even bar the homeowner from seeking redress for the foreclosure under the FHA, since the foreclosure only occurs after the homeowner takes the independent step of failing to make payments on the predatory loan. If, arguendo, we were to count steps, the Banks’ arguments would still be unconvincing. For starters, the Banks overstate the length of the causal chain by reading the complaints unfavorably to the City, ignoring, among other things, allegations that Bank of America “fail[ed] to offer refinancing or loan modifications to minority customers on fair terms” and that Wells Fargo “limit[ed] the ability of minority borrowers to refinance out of the same predatory loans that they previously received from the Bank.” BoA FAC at 21; WF FAC at 32. These alleged torts and acts of redlining occurred after -- perhaps long after -- the bad loans were originated, and may have occurred far down the causal chain, shortly before foreclosure. To take one example of their step-counting, Wells Fargo says that injury to the City from lost property-tax revenue falls six links down the causal chain and depends on the actions of five independent parties. This count ignores the complaints’ allegations that distinct FHA violations occurred when homeowners already having financial difficulties were attempting to refinance or otherwise modify their loans, and instead counts only from the point of loan 33 Case: 14-14544 Date Filed: 05/03/2019 Page: 34 of 76 origination. Because loan modification and refinancing can occur after a borrower defaults, FHA violations of this type could be far more closely connected to the City’s lost revenue. At the very least there might only be one step between the denial of refinancing and a bank’s foreclosing on a property already in default. Additionally, the Bank counts a reduction in a home’s market value and a concomitant tax assessment of that property as comprising two distinct steps in the causal chain and speculates that foreclosures are not carried out by the Banks themselves but by “a fourth party (presumably the mortgage servicers).” The complaints make no mention of an independent mortgage servicer, and in fact suggest that the Banks were involved in refinancing and loan modifications, presumably serving themselves as the loans’ mortgage servicers. By ignoring these unfavorable facts in the pleadings, and by generally refusing to draw inferences in the City’s favor, this accounting fails to read the complaints in the light most favorable to the City, as we must at the motion to dismiss stage. Additionally, Wells Fargo’s count is undermined by the very principles animating the general tendency in tort to stop at the first step. We generally stop there because, as the time frame is extended and the chain of causation becomes more attenuated, additional variables (maybe even causes) may intervene, making it more difficult to fairly attribute fault to the tortfeasor. As the Supreme Court put it in this case, “the housing market is interconnected with economic and social 34 Case: 14-14544 Date Filed: 05/03/2019 Page: 35 of 76 life,” and “[a] violation of the FHA may . . . be expected to cause ripples of harm to flow far beyond the defendant’s misconduct.” Bank of Am., 137 S. Ct. at 1306 (quotations omitted). Thus, a further problem with the Banks’ count is that the variety of factors that we worry might independently explain a homeowner’s foreclosure (like the loss of a job or spiking health care costs) all occur at the front end of the step- counting, between the act or acts of redlining and the foreclosure, not later. Once we have reached increased foreclosures on a neighborhood or citywide basis, it seems to us that the path to the City’s substantially decreased tax base is clear, direct and immediate; we can discern no obvious intervening roadblocks. Virtually all the independent variables posited by the Banks occur before foreclosure. We can identify the same kind of continuity here as in Lexmark: if the Banks’ predatory lending practices injured homeowners and led to foreclosures on a massive scale, these injuries inflicted on multiple homeowners in the same city must almost surely have injured the City as well. There is no discontinuity in this portion of the causal chain. Thus, since the risks of multiple independent variables after foreclosure are not likely, and are not many, even if we were to count steps in the way the Banks have suggested, the principles animating the general first-step rule do not support the Banks’ calculation of the post-foreclosure causal chain. 35 Case: 14-14544 Date Filed: 05/03/2019 Page: 36 of 76 The Banks’ step-counting is self-evidently conducted so as to identify as many steps as possible. We might just as easily place the same injury at the second or third step: First, a bank extends predatory loans in violation of the FHA. Second, homeowners default. Third, the bank forecloses and the property values plummet, necessarily reducing the City’s tax base and injuring its fisc. The chain will be shorter still if struggling homeowners sought to refinance and then faced swift foreclosures when fair terms were not extended. This count, which draws inferences in favor of the City, is decidedly more appropriate for the motion to dismiss stage. At the very least, the ease with which we can count far fewer steps reinforces our view that step-counting is of limited value and cannot alone settle the challenging questions of proximate cause here. B. The text and history of the FHA suggest a far-reaching statute The Supreme Court also instructs us to consider “the nature of the statutory cause of action.” Id. As we see it, the FHA has a broad remedial purpose, is written in decidedly far-reaching terms, and is perfectly capable of accommodating the type of aggregative causal connection that the City has identified between the Banks’ alleged misconduct and financial harm to Miami. Proximate cause is only one means by which Congress sets the scope of liability. Thus, for example “Congress might express limits by defining the parties who may sue, by using affirmative defenses, by providing limited remedies, by narrowly proscribing 36 Case: 14-14544 Date Filed: 05/03/2019 Page: 37 of 76 prohibited conduct, or by defining statutory terms.” Sandra F. Sperino, Statutory Proximate Cause, 88 Notre Dame L. Rev. 1199, 1236 (2013). The result, especially for complex and repeatedly amended statutes like the FHA, is an “interconnected web of congressional judgments about how and when liability should be limited.” Id. To develop a proximate cause standard without taking the full statutory scheme into account would be “to intrude upon” this web and to ignore the will of Congress. See id. Most obviously, the “[t]he language of the [FHA] is broad and inclusive.” Trafficante v. Metro. Life Ins. Co., 409 U.S. 205, 209 (1972). It provides for suit by “an aggrieved person,” 42 U.S.C. § 3613(a)(1)(A), (c)(1), expansively defined as “any person who claims to have been injured by a discriminatory housing practice,” § 3602(i) (emphasis added). As we’ve said before, “‘any’ means all.” Jones v. Waffle House, Inc., 866 F.3d 1257, 1267 (11th Cir. 2017) (emphases added). This language facilitates suits by a broad range of potential plaintiffs, and the Court has told us that it evinces not merely “a congressional intent to confer standing broadly,” Bank of Am., 137 S. Ct. at 1303, but “a congressional intention to define standing as broadly as it is permitted by Article III,” Trafficante, 409 U.S. at 209. The Court has also noted that later Congresses amended the FHA with full knowledge of the Court’s broad readings and without changing the language the Court had construed so broadly. Bank of Am., 137 S. Ct. at 1303. 37 Case: 14-14544 Date Filed: 05/03/2019 Page: 38 of 76 The Fair Housing Act also prohibits a wide range of conduct. Thus, it is a violation “to refuse to sell or rent . . . or to refuse to negotiate . . . or otherwise make unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status, or national origin” as well as “[t]o discriminate against any person in the terms, conditions, or privileges of sale or rental of a dwelling, or in the provision of services or facilities in connection therewith” for any of the same reasons. 42 U.S.C. § 3604(a)–(b). It also expressly forbids racial preferences in housing notices and advertisements, misrepresentation of housing availability based on race, and “representations regarding the entry or prospective entry into the neighborhood” of members of protected classes. Id. § 3604(c)–(e). Most relevant for the City’s claims, it is also “unlawful for any person or other entity whose business includes engaging in residential real estate–related transactions to discriminate against any person in making available such a transaction, or in the terms or conditions of such a transaction,” including in loans “for purchasing, constructing, improving, repairing, or maintaining a dwelling.” Id. § 3605(a). Again, this is far-reaching, and takes aim at discrimination that might be found throughout the real estate market and throughout the process of buying, maintaining, or selling a home. The injury to the City is as valid and as cognizable under the FHA as the injury to the homeowners. See Bank of Am., 137 S. Ct. at 1303 (finding the City 38 Case: 14-14544 Date Filed: 05/03/2019 Page: 39 of 76 has standing). And cities have been allowed to sue for similar injuries in the past, as in Gladstone Realtors v. Village of Bellwood, 441 U.S. 91 (1979), where a city alleged it was injured when segregative racial steering practices by two real estate firms redounded to “the economic and social detriment of the citizens of [the] village.” Id. at 95. The village said “racial steering effectively manipulate[d] the housing market” because marketing homes in certain areas only to buyers of certain races served to dramatically reduce the number of potential buyers. Id. at 109–10. The Court identified “profound” consequences in neighborhoods impacted by this practice, including financial consequences. Id.; see id. at 111 & n.24. Not only did the village have standing to challenge racial steering, but the Court even said that “[a] significant reduction in property values directly injures a municipality by diminishing its tax base.” Id. at 110–11 (emphasis added). The phrase “directly injures” was not used in a causal sense, but nonetheless suggests that the Court has not traditionally considered this type of injury to be overly attenuated or nebulous. Moving beyond the text, in evaluating proximate cause for a statutory cause of action, the Court has repeatedly directed us to consider a statute’s remedial aims -- Congress’s priorities in passing it. See, e.g., Blue Shield, 457 U.S. at 478 (considering “the relationship of the injury alleged with those forms of injury about which Congress was likely to have been concerned”); see also Lexmark, 134 S. Ct. 39 Case: 14-14544 Date Filed: 05/03/2019 Page: 40 of 76 at 1393 (identifying “diversion of sales to a direct competitor” as possibly “the paradigmatic direct injury from false advertising”); Assoc. Gen. Contractors of Cal., Inc. v. California State Council of Carpenters, 459 U.S. 519, 538 (1983) (following Blue Shield in evaluating whether an injury fell “squarely within the area of congressional concern,” (quoting Blue Shield, 457 U.S. at 484)); cf. Holmes, 503 U.S. at 274 (“[W]e fear that RICO’s remedial purposes would more probably be hobbled than helped . . . .” if the plaintiff’s proximate cause standard were adopted). The legislative history of the FHA is less detailed than most. Congress passed it through “a truncated legislative process in which no committee reports were issued.” Rodney A. Smolla, Federal Civil Rights Acts § 3:10, at 546–47 (3d ed. 2017). A legislative assistant who worked on the legislation recalled after the Act’s passage that this was “a time when riots threatened to close down every major city in the country,” and that the Act was presented as a response; it was hoped that “the law as a teacher might overcome the ignorance and fear of whites which previously had blocked attempts to lower a black-white barrier.” Jean Eberhart Dubofsky, Fair Housing: A Legislative History and a Perspective, 8 Washburn L. J. 149, 154 (1969). The legislative history that we do have and the context in which the FHA arose both comport with a proximate cause standard that can accommodate claims like the City’s. Thus, for example, the Supreme Court 40 Case: 14-14544 Date Filed: 05/03/2019 Page: 41 of 76 has recently recounted how, in response to the turmoil of the mid-60s, President Johnson convened the National Advisory Commission on Civil Disorders (the “Kerner Commission”), which “identified residential segregation and unequal housing and economic conditions in the inner cities as significant, underlying causes of the social unrest.” Tex. Dept. of Hous. & Cmty. Affairs v. Inclusive Cmtys. Project, Inc., 135 S. Ct. 2507, 2516 (2015). “The Commission concluded that ‘[o]ur Nation is moving toward two societies, one black, one white—separate and unequal.’” Id. (quoting Report of the National Advisory Commission on Civil Disorders 1 (1968)). In passing the Fair Housing Act, Congress explicitly took aim at these broad social ills -- maladies that were being felt on a citywide scale. Congress did not just target individual discriminatory landlords, but sought to reshape in meaningful ways the landscape of American cities. Senator Walter Mondale, the chief sponsor of the bill that would become the FHA, explained that the Act was intended to end housing segregation as a means of bringing about racial equality more broadly. The legislation’s goals were directed at the neighborhood level and indeed were aimed beyond the housing market: [O]vert racial discrimination remains in one major sector of American life—that of housing. . . . [F]air housing is one more step toward achieving equality in opportunity and education . . . . The soundest, long-range way to attack segregated schools is to attack the segregated neighborhood. 41 Case: 14-14544 Date Filed: 05/03/2019 Page: 42 of 76 114 Cong. Rec. 3421 (Feb. 20, 1968); see also id. at 3422 (“[I]n truly integrated neighborhoods people have been able to live in peace and harmony -- and both Negroes and whites are the richer for the experience.”). A co-sponsor asked, “As segregation continues to grow . . . will not the cities which house the majority of the nation’s industrial and commercial life find themselves less and less able to cope with their problems, financially and in every other way?” Id. at 2988 (Feb. 14, 1968) (statement of Sen. Brooke). The legislation, as its sponsors saw it, was designed to extend through chains of causation. Among other things, its sponsors quite specifically referenced harm to a city’s tax base, arising out of discrimination in the housing market. Senator Mondale put it this way: Declining tax base, poor sanitation, loss of jobs, inadequate educational opportunity, and urban squalor will persist as long as discrimination forces millions to live in the rotting cores of central cities. Id. at 2274 (Feb. 6, 1968) (emphasis added). It does not go too far to suggest, at least at a high order of abstraction, that in setting a standard for proximate cause the FHA looks far beyond the single most immediate consequence of a violation. Limiting the proximate cause calculus in that way would not be consonant with the powerful remedial purposes animating the bill. The Act took aim at “the segregated neighborhood” in general, not just at the prejudiced building owner. 42 Case: 14-14544 Date Filed: 05/03/2019 Page: 43 of 76 This is all to say that, notwithstanding issues of proof that might arise later in litigation, the FHA was written in broad terms and was aimed at broad problems. We can discern no reason to think as a general matter that the City’s claims are out of step with the “nature of the statutory cause of action” and the remedial scheme that Congress created. C. Tracing causation here is not administratively infeasible Perhaps the most important step in the proximate cause analysis in this case is consideration of “what is administratively possible and convenient.” Bank of Am., 137 S. Ct. at 1306 (quoting Holmes, 503 U.S. at 268). This is because administrative feasibility is most closely connected to the policy judgments upon which proximate cause standards necessarily depend. See Holmes, 503 U.S. at 268 (“[P]roximate cause reflects ideas of what justice demands, or of what is administratively possible and convenient.” (internal quotations omitted)). In no small part, we conclude that the City has adequately -- that is to say, plausibly -- pled proximate cause because we find it entirely practicable and not unduly inconvenient for the courts to handle damages like the City’s tax revenue injury. Cases like this will never be among the simplest that our courts see, but the federal courts regularly handle complex and high-stakes cases, so complexity alone is no reason to dismiss a case on the pleadings. As pled, the City’s injury is ascertainable with a sufficient degree of precision for some of its economic 43 Case: 14-14544 Date Filed: 05/03/2019 Page: 44 of 76 injuries, specifically the harm to its tax base. Since the City’s injuries are unique to its treasury, no other plaintiff will plead the same injuries, or attempt to recover the same funds. The City is in the best position to allege and litigate this peculiar kind of injury, to deter future violations and, theoretically, to actually remedy its distinctive injury. In Holmes v. Securities Investor Protection Corp., 503 U.S. 258 (1992), the Court conducted a proximate cause analysis focused on administrative feasibility. Holmes involved the Securities Investor Protection Corporation (SIPC), a federally created non-profit corporation empowered to sue broker-dealers who “ha[ve] failed or [are] in danger of failing to meet [their] obligations to [their] customers.” Id. at 261 (quoting 15 U.S.C. § 78eee(a)(3)). In Holmes, SIPC alleged that broker- dealers manipulated securities in “a ‘pattern of racketeering activity’” that violated RICO, 18 U.S.C. §§ 1962, 1961(1), (5). Holmes, 503 U.S. at 263. The case asked whether SIPC, “neither a purchaser nor a seller of securities” could sue under RICO even though its claims were, in essence, the same claims that could have been brought by investors under Rule 10b–5, under which SIPC lacked standing. Id. at 265 n.7. RICO allows for a civil action by “[a]ny person injured in his business or property by reason of a violation of section 1962.” Id. at 265 (quoting 18 U.S.C. § 1964(c)). SIPC claimed that it should be allowed to sue under this provision 44 Case: 14-14544 Date Filed: 05/03/2019 Page: 45 of 76 “because it [was] subrogated to the rights of those customers of the broker-dealers who did not purchase manipulated securities.” Id. at 270. The chain of causation was thus the following: The defendants manipulated securities and broker-dealers bought manipulated securities. The broker-dealers sold some of their customers these manipulated securities, and their value plummeted when the fraud was exposed. As a result, the broker-dealers could not meet their obligations to their customers, including to some who had not even bought manipulated securities. It was these nonpurchasing customers’ rights that SIPC argued it was subrogated to. See id. at 270–71. Granting, “arguendo” that SIPC was subrogated as it claimed, Holmes, 503 U.S. at 271, the Court found that “the conspirators’ conduct did not proximately cause the nonpurchasing customers’ injury,” id. at 270. The Court pointed to three distinct reasons why “directness of relationship” was “one of [the] central elements” of “Clayton Act causation.” Holmes, 503 U.S. at 269. First, less direct injuries are harder to attribute to a violation. Id. Second, claims at different levels of remove complicated the apportionment of harm among plaintiffs at different levels. Id. Third, directly injured plaintiffs usually can be counted on to sue and deter wrongdoing. Id. at 269–70. All three reasons for directness were said to apply “with equal force” to antitrust and RICO suits. Id. at 270. The Court then considered the reasons again in the context of the particular claims and found that each indicated that SIPC’s claims fell outside the proper 45 Case: 14-14544 Date Filed: 05/03/2019 Page: 46 of 76 limits of proximate causation. Id. at 272–74. It would appear, from the Court’s analysis, that these are not only reasons why “directness of relationship” is required, but are also important factors against which to measure the directness of the relationship. When applied to the City’s claims against the Banks, these factors strongly suggest that the City has plausibly alleged “some direct relation” between the Banks’ alleged conduct and the injury to the City’s tax revenue. We cannot say the same for the increased-expenditures injury, though. Here, we review how each of the Holmes factors applies to the City’s economic injuries, distinguishing, when relevant, between the tax-revenue injury and the increased-expenditures injury. Finally, we explain how the Holmes factors provide a clear means by which to cabin liability and prevent suits by more remotely injured parties. 1. Holmes’s first factor says that directness is required because “the less direct an injury is, the more difficult it becomes to ascertain the amount of a plaintiff’s damages attributable to the violation, as distinct from other, independent, factors.” Holmes, 503 U.S. at 269. In Holmes, this factor decidedly cut against finding a sufficiently direct relationship because it would have required the district court “to determine the extent to which [the nonpurchasing customers’] inability to collect from the broker-dealers was the result of the alleged conspiracy to manipulate, as 46 Case: 14-14544 Date Filed: 05/03/2019 Page: 47 of 76 opposed to” any other reasons why the broker-dealers might not have money on hand (such as “poor business practices,” and the “failure to anticipate developments in the financial markets”). Id. at 273. On this first factor, the City’s tax-revenue injury fares well, and its municipal-expenditures injury fares poorly. The City has plausibly alleged that it can present an analysis of reduced tax revenue that is precise enough to avoid difficulties with isolating the role of the Banks’ alleged violations. It has not done so for its municipal expenditures. In its complaints, the City of Miami’s tax-revenue injury seeks damages “based upon reduced property tax revenues resulting from (a) the decreased value of the vacant properties themselves, and (b) the decreased value of properties surrounding the vacant properties.” BoA FAC at 31; WF FAC at 44. According to the City, “[r]outinely maintained property tax and other data allow for the precise calculation of the property tax revenues lost by the City as a direct result of particular [bank] foreclosures.” BoA FAC at 32; WF FAC at 45. This “Hedonic regression methodology,” the City asserts, “can be used to quantify precisely the property tax injury to the City caused by [the Banks’] discriminatory lending practices and resulting foreclosures in minority neighborhoods.” BoA FAC at 34; WF FAC at 47. The City does not go so far as to conduct this analysis and attach the results to its pleadings, but its First Amended Complaints nonetheless suffice to describe 47 Case: 14-14544 Date Filed: 05/03/2019 Page: 48 of 76 the analysis in far more than speculative or conclusory fashion. The City has explained in some detail how this analysis works: Homes in foreclosure tend to experience a substantial decline in value . . . [F]oreclosure properties and the problems associated with them likewise cause especially significant declines in surrounding property values because the neighborhoods become less desirable. . . . Routinely maintained property tax and other data allow for the precise calculation of the property tax revenues lost by the City as a direct result of particular [bank] foreclosures. Using a well-established statistical regression technique that focuses on effects on neighboring properties, the City can isolate the lost property value attributable to [the Banks’] foreclosures and vacancies from losses attributable to other causes, such as neighborhood conditions. This technique, known as Hedonic regression, when applied to housing markets, isolates the factors that contribute to the value of a property by studying thousands of housing transactions. Those factors include the size of a home, the number of bedrooms and bathrooms, whether the neighborhood is safe, whether neighboring properties are well-maintained, and more. Hedonic analysis determines the contribution of each of these house and neighborhood characteristics to the value of a home. BoA FAC at 32–33; WF FAC at 45–46. 6 This gives us a clear idea of the final analysis -- based on empirical data drawn from thousands of housing transactions, the City will calculate the impact of the Banks’ foreclosures on property values in redlined (and reverse-redlined) areas of Miami, controlling for other variables and isolating the impact of the redlining. The City points to studies in which this 6 Wells Fargo says that the City pleads “that it could conduct a study that might be able to show that foreclosures affect property values.” (emphases in original). The City’s pleading does not hedge its allegations in this way, and discounting the feasibility of its analysis on the front end is inconsistent with our obligation to draw all reasonable inferences in the City’s favor at this preliminary stage. 48 Case: 14-14544 Date Filed: 05/03/2019 Page: 49 of 76 methodology has produced the kind of results the City will need to present. BoA FAC at 33–34; WF FAC 46–47. The complaints conclude, then, that “[a]pplication of such Hedonic regression methodology . . . can be used to quantify precisely the property tax injury to the City caused by [the Banks’] discriminatory lending practices and resulting foreclosures.” BoA FAC at 34; WF FAC at 47. The pleadings strongly suggest, then, that the tax-revenue injury to the City attributable to the Banks’ alleged discriminatory practices is readily calculable. Under Twombly and Iqbal a mere statement that “[the Banks] caused the loss of property tax revenue” would be wholly conclusory and would be disregarded. But by indicating and explaining at considerable length the kind of analysis that would be conducted to quantify the loss of revenue attributable to discriminatory lending, the City has plausibly alleged a calculable harm and has made more than a formulaic recitation of how the causation requirement will be met. There could be a battle of experts down the line over whether the regression analysis really shows what the City says it does, but, as we see it, that would be for a later stage of the litigation. The complaints’ allegation that regression analysis can pinpoint causation is sufficient at this stage and helps account for the concerns raised by this first factor. The plausibility of hedonic regression analysis has a direct bearing on how “difficult it [is] to ascertain the amount of [the City’s] damage attributable to the 49 Case: 14-14544 Date Filed: 05/03/2019 Page: 50 of 76 violation, as distinct from other, independent, factors,” Holmes, 503 U.S. at 269, and thus helps determine “what is administratively possible and convenient,” in terms of damages calculation, Bank of Am., 137 S. Ct. at 1306. Here, the City has plausibly pled that it is practicable and convenient to quantify the property tax injury’s causal connection to the Banks’ policies. In addition to the studies identified by the City, we note in passing there has been a substantial amount of writing (including legal literature) that has discussed or employed hedonic regression analysis as a practicable and effective way of calculating impacts on real estate values.7 One author at the University of Wisconsin School of Business has written that “[o]ver the past three decades, hedonic estimation has clearly matured from a new technology to become the standard way economists deal with housing heterogeneity.” 8 Likewise, the Organisation for Economic Co-operation and Development’s (OECD) Handbook on Residential Property Prices Indices (RPPI) calls hedonic regression “probably the best method that could be used in order to construct quality RPPIs for various types of property.” 9 While we may 7 See, e.g., Yun-Chien Chang, Economic Value or Fair Market Value: What Form of Takings Compensation is Efficient?, 20 SUP. CT. ECON. REV. 35, 52–54 (2012); Yun-Chien Chang & Lee Anne Fennell, Partition and Revelation, 81 U. CHI. L. REV. 27, 35 n.35(2014); Randall Akee, Checkerboards and Coase: The Effect of Property Institutions on Efficiency in Housing Markets, 52 J.L. & ECON. 395, 402–03 (2009). 8 Stephen Malpezzi, Hedonic Pricing Models: A Selective and Applied Review, in HOUSING ECONOMICS AND PUBLIC POLICY 67, 87 (Tony O’Sullivan & Kenneth Gibb eds., 2003). 9 OECD, Hedonic Regression Methods, in HANDBOOK ON RESIDENTIAL PROPERTY PRICE INDICES 50, 57 (2003). 50 Case: 14-14544 Date Filed: 05/03/2019 Page: 51 of 76 not engage in a Daubert analysis at this early stage, that hedonic regression analysis has been favorably referenced and employed elsewhere and over several decades cuts against the suggestion at the motion to dismiss stage that it is a wholly implausible invention of the City’s or that apportioning blame through its application would be so challenging that the complaints cannot state a claim. At a more basic level, the aggregative nature of the City’s claims also helps eliminate any discontinuity between the statutory violation and the injury. See Lexmark, 134 S. Ct. at 1394. Here, the claimed violation is a “pattern or practice of reverse redlining” leading to a “disproportionately high rate of foreclosure.” BoA FAC at 30; WF FAC at 43. Even if each individual act of redlining does not bear a one-to-one proportional relationship to Miami’s loss of tax revenue, see Lexmark, 134 S. Ct. at 1394, since intervening circumstances affect which individual properties go into foreclosure, in the aggregate it has been plausibly alleged that the impact of redlining and reverse-redlining can be identified with precision and deterred. Similarly, in Blue Shield, the Court identified “the physical and economic nexus between the alleged violation and the harm to the plaintiff” as a factor to consider when applying proximate cause to the Clayton Act. Blue Shield, 457 U.S. at 478. The scale of the misconduct matches the scale of the injury. 51 Case: 14-14544 Date Filed: 05/03/2019 Page: 52 of 76 Since the Supreme Court spoke to this issue in Bank of America, two district courts analyzing similar FHA claims by municipalities against lending institutions have relied on the availability of regression analysis in finding that proximate cause had been plausibly alleged. Most similarly and most recently, a district court in the Northern District of California entertained a similar case brought by the City of Oakland against Wells Fargo for predatory lending practices. City of Oakland v. Wells Fargo Bank, N.A., No. 15-cv-04321, 2018 WL 3008538 (N.D. Cal. June 15, 2018). 10 In denying a motion to dismiss that court recognized, as we have, that “[w]hile the fact of aggregative injury itself does not obviate proximate cause analysis,” the proffer of “a specific statistical analysis in regard to [the City’s] property tax injury” adequately supports the idea that there was “an alleged provable and quantifiable causal link between the defendant’s conduct and plaintiff’s injury” notwithstanding the suggested length of the causal chain. Id. at *8. But when it came to a municipal-expenditures injury -- similar to Miami’s -- the district court ruled that since “Oakland ha[d] not proffered any statistical 10 After this decision, the district court in the Oakland case certified two questions for interlocutory appeal to the Ninth Circuit: “(1) Do Oakland’s claims for damages based on the injuries asserted in the [First Amended Complaint] satisfy on a motion to dismiss proximate cause required by the FHA?” and “(2) Is the proximate-cause requirement articulated in City of Miami limited to claims for damages under the FHA and not to claims for injunctive or declaratory relief?” Order Granting Defendant’s Motion to Amend Order to Include Certification for Interlocutory Appeal, City of Oakland, No. 15-cv-04321, 2018 WL 7575537 at *2. The parties’ briefs before the Ninth Circuit are due in June and July of 2019. City of Oakland v. Wells Fargo & Co., No. 19-15169 (9th Cir. Mar. 26, 2019), ECF No. 10. 52 Case: 14-14544 Date Filed: 05/03/2019 Page: 53 of 76 analyses comparable to those in the property-tax analysis,” the first Holmes factor counted against them and problems of multiple causation abounded. Id. at *10. Those claims were dismissed without prejudice. Id. A district court in Philadelphia found that city had “adequately [pled] proximate cause for . . . non- economic injuries” stemming from FHA violations in part by relying on “a regression analysis of [bank] loan data” that had demonstrated Latino and African- American borrowers were 2.641 and 4.147 times more likely to go into foreclosure than similarly situated white borrowers. City of Philadelphia v. Wells Fargo & Co., No. 17-2203, 2018 WL 424451 (E.D. Pa. Jan. 16, 2018). In describing its second claimed economic injury -- increased municipal expenditures for police, fire, sanitation, and the like resulting from widespread foreclosures -- the City does not do as well when measured against the first Holmes factor. When it comes to this injury, the City’s complaints fail to explain how we can ascertain with any level of detail or precision which expenditures will be attributable to the Banks. It is self-evident that this must be accomplished somehow because the entire increase in municipal expenditures over any time period cannot possibly be fairly attributed to the Banks’ conduct. Intervening causes and independent variables will inevitably run up this measure of damages because the City’s expenditures occur at some obvious level of remove from the foreclosures that it says cause them. They are further down the chain, to put it in 53 Case: 14-14544 Date Filed: 05/03/2019 Page: 54 of 76 step-counting terms. There is nothing in the pleadings that suggests the plaintiff will be able to sort out the extent to which these damages are attributable to the Banks’ misconduct. The City’s pleadings on increased expenditures are in stark contrast to their pleadings about tax revenue. The increased expenditures pleadings state, in almost conclusory fashion, that “the City is required to provide increased municipal services” at foreclosed properties and that “these services would not have been necessary if the properties had not been foreclosed upon.” BoA FAC at 34; WF FAC at 47. The complaints then proceed to list types of expenditures: police, fire, building code enforcement, and the like. If any direct connections exist between the foreclosure and any of these expenditures, the City has not explained them, and they are not so obvious as to be self-evident. We also see no explanation of how the City will identify the amount of increase attributable to the foreclosures or to the Banks’ conduct. In pleading the tax-revenue injury, however, the City explains in considerable detail how hedonic regression analysis will help pinpoint the attributable loss. This is not to say that hedonic regression analysis, specifically, should have been referenced again in relation to the expenditure injury in order to satisfy the first Holmes factor. Rather we suggest only that a pleading in a case like this must make clear how a plaintiff will demonstrate that an injury is 54 Case: 14-14544 Date Filed: 05/03/2019 Page: 55 of 76 attributable to a defendant. Hedonic regression serves this role with regard to the tax-revenue injury, but may not be the only way this could ever be accomplished.11 The first Holmes factor, addressing the attributability of the injury to the defendant’s conduct, is where the two economic injuries alleged by the City are most different in terms of finding a “direct relation.” The tax-revenue injury has been pled with detailed explanations of statistical proof, which serve to plausibly allege that the courts will be able to work out just what the Banks can be fairly held accountable for. The municipal-expenditure injury fails to do this. Statistical proofs are not the only method of proof, but, when we readily discern a discontinuity between misconduct and injury, for the plaintiff to plausibly plead proximate cause, it must allege something more with which to bridge the gap and demonstrate that the defendant’s liability will bear “some direct relation” to its 11 A few other district courts have illustrated the point. In three lawsuits against banks alleging the same kinds of practices that Miami has alleged, Cook County, Illinois claimed damages based on “out-of-pocket costs it . . . incurred in processing the discriminatory foreclosures, such as additional funding for the Cook County Sheriff to serve foreclosure notices and for the Circuit Court of Cook County to process the deluge of foreclosures.” County of Cook v. Bank of Am. Corp., No. 14 C 2280, 2018 WL 1561725 at *7 (N.D. Ill. Mar. 30, 2018); see also County of Cook, Ill. v. HSBC N. Am. Holdings, 314 F. Supp. 3d 950, 956 (N.D. Ill. 2018); County of Cook, Ill. v. Wells Fargo & Co., 314 F. Supp. 3d 975, 982 (N.D. Ill. 2018). Those out-of-pocket costs were found to have plausibly been proximately caused by the banks because these expenditures are automatic results of foreclosures. See HSBC, 314 F. Supp. 3d at 962; Wells Fargo, 314 F. Supp. 3d at 984; Bank of Am., 2018 WL 1561725 at *7. These were the only claimed damages that survived a motion dismiss in the Northern District of Illinois, as all three courts there found that Cook County had failed to adequately plead proximate cause for the kinds of tax-revenue and municipal-expenditure injuries that Miami pleads here. See HSBC, 913 F. Supp. 3d at 962–64; Wells Fargo, 314 F. Supp. 3d at 988; Bank of Am., 2018 WL 1561725 at *5. Miami pled no out-of-pocket costs closely tied to foreclosures along these lines. 55 Case: 14-14544 Date Filed: 05/03/2019 Page: 56 of 76 actions. On the increased-expenditure injury, the City has not plausibly pled its case. 2. Holmes’s second factor posits that “recognizing claims of the indirectly injured would force courts to adopt complicated rules apportioning damages among plaintiffs removed at different levels of injury from the violative acts, to obviate the risk of multiple recoveries.” Holmes, 503 U.S. at 269. In Holmes, this was a problem because a trial court would “have to find some way to apportion the possible respective recoveries by the broker-dealers and the customers, who would otherwise each be entitled to recover the full treble damages.” Id. at 273. Antitrust cases also emphasize “the risk of duplicative recovery engendered by allowing every person along a chain of distribution to claim damages” from a single violation, and have identified this as a powerful reason to allow suit only by more directly injured parties. Blue Shield, 457 U.S. at 474–75 (citing Ill. Brick Co. v. Illinois, 431 U.S. 720, 745 (1977); Hawaii v. Standard Oil Co. of Cal., 405 U.S. 251 (1972)). Again, no such problem is presented in this case: plainly, the injuries to the City’s treasury are not shared by any other possible plaintiff. These economic injuries, whether or not they can ultimately be recovered for, could only be alleged by the City. The City’s theory, therefore, presents no concern about double 56 Case: 14-14544 Date Filed: 05/03/2019 Page: 57 of 76 recovery or “piggy-backing,” precisely because its injuries are unique. Substantially decreased tax revenue and increased municipal costs are injuries that affect the interests of the City alone, so there could not plausibly be any difficulty in apportioning them between multiple plaintiffs. The harm and damages pled by individual homeowners suing under the FHA would be entirely different. These plaintiffs could sue for actual and punitive damages, injunctive and equitable relief, or attorney’s fees. See 42 U.S.C. § 3613(c). In this Circuit they could also seek damages based on “anger, embarrassment, and emotional distress.” Banai v. Sec’y, U.S. Dep’t of Hous. & Urban Dev. ex rel. Times, 102 F.3d 1203, 1207 (11th Cir. 1997). Many courts have permitted FHA plaintiffs to recover pecuniary damages, including moving costs and forfeited security deposits. E.g., Belcher v. Grand Reserve MGM, LLC, 269 F. Supp. 3d 1219, 1239 (M.D. Ala. 2017). But no court could allow a homeowner to recover for harm to a city’s treasury. The City, on the other hand, has claimed it was financially harmed in two ways -- injury to its tax base and tax revenue because of reduced property value, and increased expenditures on City services in certain neighborhoods. The damages that these harms could support do not overlap in any way with or duplicate the damages that individual homeowners may have sustained, because the harms are entirely separate. The City’s injuries would be “passed on” forms of the homeowners’ injuries if, for example, it were claiming it lost tax revenue 57 Case: 14-14544 Date Filed: 05/03/2019 Page: 58 of 76 because homeowners who were overpaying on their mortgages were subsequently unable to pay their taxes. But this is not the case. The City alleges instead that the Banks’ misconduct had a direct, necessary, and immediate impact on its fiscal health at the neighborhood and citywide level. The homeowners’ losses are obviously related to the losses that Miami allegedly incurred, but they are independent. Injury to the City’s treasury is a distinct injury, it is not purely derivative of misfortunes visited on homeowners. Even if we were to assume that each victim of discrimination were to successfully bring an individual claim -- a wholly implausible hypothesis -- the City would still have an independent set of claims based on the independent harm that it suffered. The claims would be linked only by shared facts surrounding the violation. 3. Holmes’s final factor builds on the first two: “[T]he need to grapple with these problems,” the Court said, referring to the first two factors, “is simply unjustified by the general interest in deterring injurious conduct, since directly injured victims can generally be counted on to vindicate the law as private attorneys general, without any of the problems attendant upon suits by plaintiffs injured more remotely.” Id. at 269–70. Thus, in Holmes, the directly injured broker-dealers “could be counted on to bring suit for the law’s vindication,” and so it was less necessary for SIPC to sue on behalf of the broker-dealers’ customers. 58 Case: 14-14544 Date Filed: 05/03/2019 Page: 59 of 76 Id. at 273. In Bridge, the same factor cut in the other direction; the plaintiffs were injured because they had less success at county-run auctions when the defendants cheated, and their harm was “the direct result of [defendant’s] fraud” in part because, even though the plaintiffs didn’t themselves rely on any fraudulent assertions, “no more immediate victim is better situated to sue.” Bridge, 553 U.S. at 658. The third Holmes factor has been applied to antitrust statutes as well. The availability of more directly injured plaintiffs influenced the Court’s determination that “directness” between the plaintiff’s injury and the defendant’s action was lacking in Associated General Contractors of California, Inc. v. Carpenters, 459 U.S. 519, 540–41 (1983). There, a union alleged that the defendant employer association had “coerced certain third parties . . . to enter into business relationships with nonunion firms” in a way that violated the Clayton Act. Id. at 520–21. This hurt unionized firms and, thus, hurt the plaintiff union. Id. The Court found that the union had not alleged a sufficiently direct injury, noting that: The existence of an identifiable class of persons whose self-interest would normally motivate them to vindicate the public interest in antitrust enforcement diminishes the justification for allowing a more remote party such as the Union to perform the office of a private attorney general. Denying the Union a remedy on the basis of its allegations in this case is not likely to leave a significant antitrust violation undetected or unremedied. Id. at 542. 59 Case: 14-14544 Date Filed: 05/03/2019 Page: 60 of 76 In these kinds of cases it is preferable for more directly injured plaintiffs to take the lead in acting as private attorneys general because those plaintiffs will be able to deter future violations more effectively and efficiently. When more directly injured parties can sue for the full swath of harm caused by statutory violations, the threat of suit by these parties has the maximum possible deterrent effect on potential violators, and it’s simply not worthwhile to also allow minimally deterring suits by plaintiffs farther down the causal chain. Cf. Southern Pac. Co. v. Darnell-Taenzer Lumber Co., 245 U.S. 531, 534 (1918) (“The [defendant] ought not to be allowed to retain his illegal profit, and the only one who can take it from him is the one that alone was in relation with him, and from whom the [defendant] took the sum.”) (Holmes, J.). Thus, in antitrust cases the courts have long expressed a “concern for the reduction in the effectiveness of those suits if brought by indirect purchasers with a smaller stake in the outcome than that of direct purchasers suing for the full amount of the overcharge.” Ill. Brick Co., 431 U.S. at 745; see also Lexmark, 134 S. Ct. at 1391 (establishing, for the Lanham Act, a rule barring suit by “commercial parties who suffer merely as a result of [a direct] competitor’s ‘inability to meet [its] financial obligations’” (quoting Anza, 547 U.S. at 458)). For this reason, the federal courts have rejected the monopolist’s “passing on” defense -- the argument that mid-stream purchasers charged monopoly prices are 60 Case: 14-14544 Date Filed: 05/03/2019 Page: 61 of 76 not harmed because they can “pass on” any overcharges to consumers. See Richard A. Posner, Economic Analysis of Law 395–96 (9th ed. 2014). Consumers cannot sue over monopoly pricing, but this does not mean there is a meaningful reduction in enforcement. See id. at 396 (“[Consumers’] right [to sue] is less valuable because, being at once more remote . . . and more numerous, consumers are less efficient antitrust enforcers.”). The same logic surrounding deterrence does not apply here. While the City may be a step further down the chain of causation, the homeowners are much more numerous and less well suited to prosecute a global claim. The City’s suit could achieve better deterrence because it aims to recover for a larger injury sustained on a citywide basis and to remedy a different, broader violation than a homeowner’s suit would. The City alleges widespread redlining and reverse-redlining policies conducted by the Banks. The City’s suit challenges the entire policies. A discrimination claim by an individual homeowner would challenge only a bank’s discriminatory action against that homeowner. Conceivably the banks could face suits by many homeowners who were discriminated against, or who lost property value because their neighbors’ houses were foreclosed on, but these suits wouldn’t serve to condemn the pattern or policy. Individual homeowners would also face most of the same causation problems -- presumably a variety of factors contributed to any individual foreclosure -- but the regression analysis that can isolate the 61 Case: 14-14544 Date Filed: 05/03/2019 Page: 62 of 76 impact of redlining on the neighborhood scale could not solve this problem on the individual level because of the diversity of individual circumstances. Moreover, even if homeowners were better-situated to vindicate the same wrongdoing -- and that is plainly not the case here -- the additional assumption that others will pursue the lawsuits and achieve deterrence is nowhere near as likely true under the FHA as under the other statutes. Antitrust and, to a lesser extent, civil RICO suits are typically brought by sophisticated business entities, often with deep pockets, and those plaintiffs also have the added incentive of realizing treble damages. In Holmes and Associated General Contractors, the Court left the job of deterring statutory violations to broker-dealers and businesses using union labor. These parties were subject to complex regulatory environments and undoubtedly had counsel in place prepared to sue to vindicate their rights. Here, instead, the alternative plaintiffs are homeowners whose homes were foreclosed upon. Compared with the sophisticated parties involved in the RICO and antitrust cases, it seems far less likely that these injured parties could be counted on to file suit. While it’s possible that some might, it seems exceedingly unlikely that more than a handful at most would do so, and we have seen precious little reason to believe that many have. The vast majority of individual cases of discriminatory lending cases are left unpursued and unaddressed. In Associated General, the Court observed that “[d]enying the [plaintiff] a remedy on the basis of its allegations in this case 62 Case: 14-14544 Date Filed: 05/03/2019 Page: 63 of 76 [was] not likely to leave a significant [statutory] violation undetected or unremedied.” 459 U.S. at 542. That’s less plausible here. As a result, the peculiarities surrounding the claims raised in this case make it distinct from an antitrust case where a mid-stream purchaser can bring essentially the same action against a monopolist as the ultimate consumer would. The question of efficacy is harder to resolve here because, although the homeowners are arguably closer in the chain, they are too numerous and diffuse to be counted on for deterrence. They are better situated only in a narrow, literal sense. The City, on the other hand, has plausibly alleged an injury, calculable in the aggregate, that bears a direct relation to the Banks’ policies, applied in the aggregate, which allegedly violated the statute. 4. The Holmes factors also help cabin proximate cause against other less- directly injured plaintiffs. Writing separately in Bank of America, Justice Thomas suggested that Miami’s theory of causation would require that neighboring homeowners who were not foreclosed on but whose property values fell could also sue the banks. 137 S. Ct. at 1312 (Thomas, J., concurring in part and dissenting in part). Still other concerns have been raised about the corner grocer who may have lost business on account of foreclosures, the utility company, or maybe even real estate brokers working on commission to resell properties in the area. The Court’s 63 Case: 14-14544 Date Filed: 05/03/2019 Page: 64 of 76 own guidance from Holmes, attuned as it is to the administrative practicalities, can go a long way toward allaying these concerns. Application of the Holmes factors reasonably establishes that the City’s injury is logically bound up with the Banks’ alleged conduct in meaningful ways missing from injuries to these other putative claimants. To begin with the first factor, corner grocers or neighboring property owners, unlike the City, would have substantially more difficulty plausibly alleging that they could calculate damages attributable to the Banks’ actions with a reasonable degree of certainty. By contrast, hedonic regression techniques are most effective when applied on a neighborhood or citywide level, precisely because individual variations among homeowners average out when foreclosures are considered in the aggregate. Indeed the amended complaints specifically cite to a study showing that each foreclosure in Chicago was “responsible for an average decline of approximately 1.1% in the value of each single-family home within an eighth of a mile.” BoA FAC at 33; WF FAC at 46. The average decline can be used to calculate the impact of foreclosures on property value across a neighborhood on account of the large sample size. A neighboring homeowner, on the other hand, would be affected only by homes closest to his own, and these might not accurately reflect the citywide average, in terms of causation or value. And a corner grocer might be affected by the surrounding few blocks of homes, but 64 Case: 14-14544 Date Filed: 05/03/2019 Page: 65 of 76 this may be wholly unrepresentative of citywide trends, making causal attribution to redlining all the more difficult. Put another way, in step-counting terms, the grocer or the utility company is demonstrably further down the causal chain because many independent variables can enter the equation after foreclosure occurs. The City will necessarily be harmed as soon as the foreclosures occur. For the grocer, though, a nearby home’s foreclosure does not necessarily mean anything. The causal chain for the grocer is longer, more attenuated, and more of a problem for proximate cause analysis because there is a powerful discontinuity between the foreclosures and the grocer’s loss of profits. It would therefore be exceedingly difficult to trace causation from the Banks’ predatory practices through the foreclosures to the grocers’ diminished profits. The corner grocer, or the utility company, is not affected when a home is foreclosed or when property values go down. These parties are not hurt until individual homeowners decide to spend less money at that grocer, decide not to pay the electric bill, or move away. Thus, there are many more third parties, and many more intervening steps when it comes to the corner grocer or the utility company. When a homeowner is given a bad, discriminatory loan, the injury to one of these more distant plaintiffs is in no sense “automatic[],” Lexmark, 134 S. Ct. at 1394; it therefore becomes harder to attribute back to the Banks. 65 Case: 14-14544 Date Filed: 05/03/2019 Page: 66 of 76 The application of the second Holmes factor also strongly suggests why any injury sustained by the corner grocer or by the utility company would not be proximately caused by the Banks’ conduct. These are classic “passed on” injuries. The Court has consistently said that “where the alleged violations [are] linked to the asserted harms only through the [directly injured parties’] inability to meet their financial obligations” proximate cause will not be found. Anza, 547 U.S. at 458; see also Lexmark, 1391 S. Ct. at 1391 (referencing “the electric company,” specifically, as an entity in this sort of position); Holmes, 503 U.S. at 271. Individual homeowners have financial obligations to the City in the form of property taxes, but the City’s claim is not that it is harmed because property taxes went unpaid. Rather, the City says that property values decreased and that therefore it was not entitled to as much property-tax revenue. The City’s loss is not due to anyone lacking “the wherewithal to pay,” Holmes, 503 U.S. at 271, but to the fact that redlining means the City is not able to collect nearly as much as it would have, had the alleged FHA violations not occurred in the first place. The third Holmes factor -- the reliability with which a plaintiff could remedy harm or achieve deterrence -- also strongly suggests that we need not fear a deluge of suits by grocers, neighbors, or power companies. These potential plaintiffs, even if injured, would be in a far weaker position and far less likely than the City to achieve deterrence or to remedy the entirety of the harm caused by the Banks’ 66 Case: 14-14544 Date Filed: 05/03/2019 Page: 67 of 76 alleged violations. Their claims would be disjointed, aimed at remedying FHA violations only on particular blocks, in contrast to the City’s unique ability to attack the Banks’ whole pattern or practice on a municipal level. Any claims by the corner grocer or the neighboring homeowner would be further out in the chain of causation than the City is, and these claims would lack the broad scope that allows the City to achieve maximal deterrence. Quite simply, they would face far tougher versions of all the challenges the City faces, but with none of the offsetting advantages. As we see it, the factors articulated by the Supreme Court in Holmes cabin this decision and keep the floodgates closed. But these factors are not the only reason our ruling does not extend liability beyond what is reasonable. Foreseeability remains a real part of the proximate cause calculation and also will function to cut off liability in many instances. Dan B. Dobbs, Paul T. Hayden & Ellen M. Bublick, The Law of Torts § 199, at 686 (2d ed. 2011) (“The defendant must have been reasonably able to foresee the kind of harm that was actually suffered by the plaintiff . . . .”). We are exceedingly dubious that courts would expect lenders to reasonably foresee unending ripples of harm that overwhelm the judicial branch. 67 Case: 14-14544 Date Filed: 05/03/2019 Page: 68 of 76 D. Common-law antecedents require direct relation The Court has also told us that the FHA’s common-law antecedents are a primary reason why “proximate cause under the FHA requires ‘some direct relation between the injury asserted and the injurious conduct alleged.’” Bank of Am., 137 S. Ct. at 1306. As a result, we know that “[a] damages claim under the [FHA] ‘is analogous to a number of tort actions recognized at common law,’” and that “directness principles” thus apply. Id. We also observed, in our previous opinions, that FHA damages claims “have long been analogized to tort claims.” See Bank of Am., 800 F.3d at 1282. These analogues, however, only take us so far. They do not help flesh out the meaning of “direct relation,” even though they are the basis for imposing the requirement. The Court identified some common-law claims that are analogues for FHA claims in Curtis v. Loether, 415 U.S. 189 (1973). See Bank of Am., 137 S. Ct. at 1306 (citing Curtis for the proposition that an FHA damages claim is analogous to certain common-law torts). There, the Court was evaluating whether the Seventh Amendment, which guarantees jury trials for “suits at common law,” guaranteed a jury in an FHA suit. U.S. CONST. amend VII; see Curtis, 415 U.S. at 190. The Court said it did, because FHA claims for housing discrimination were comparable to “the common-law duty of innkeepers not to refuse temporary lodging to a 68 Case: 14-14544 Date Filed: 05/03/2019 Page: 69 of 76 traveler without justification,” “an action for defamation or intentional infliction of mental distress,” or “law[s] of insult and indignity.” Id. at 195 n.10. The identification of these common-law antecedents does not get us too far. We lack any clear indication that Congress had these common-law claims in mind when drafting the FHA, and so we are reluctant draw too much from them beyond the “some direct relation” requirement. For one thing, we would not know which common-law claim to begin with, since we do not see the obvious correspondence to the common law the Court has identified elsewhere. Thus, for example, this case is not like Bridge where common-law mail fraud was an obvious precursor to RICO mail fraud. See Bridge, 553 U.S. at 652. Other statutory causes of action have more tangible foundations in particular common-law claims. The Court does not define what exactly counts as “common- law foundations” but RICO and the antitrust statutes are the paradigmatic examples. See, e.g., Bridge, 533 U.S. at 651 (“[W]hen Congress established in RICO a civil cause of action for a person ‘injured . . . by reason of’ a ‘conspir[acy],’ it meant to adopt . . . well-established common-law civil conspiracy principles” (quoting Beck v. Prupis, 529 U.S. 494, 504 (2000))); Holmes, 503 U.S. at 267 (“[C]ourts had read § 7 [of the Sherman Act] to incorporate common-law principles of proximate causation.”). In Anza v. Ideal Steel Supply Corp., 547 U.S. 451, 457 (2006), a RICO case, the Court also said that “directness principles” 69 Case: 14-14544 Date Filed: 05/03/2019 Page: 70 of 76 apply to statutes with “common-law foundations.” Bank of Am., 137 S. Ct. at 1306 (quoting Anza, 547 U.S. at 457). Anza repeated reasoning from Holmes comparing RICO’s civil suit provision to the Clayton Act’s. Anza, 547 U.S. at 457 (citing Holmes, 503 U.S. at 267–68). The connection between the Clayton Act and RICO, laid out in Holmes, was foundational for proximate cause analysis under RICO. See, e.g., Hemi Group, 559 U.S. at 8–10 (citing Holmes, 503 U.S. 258); Bridge, 553 U.S. at 653–55 (citing Holmes, 503 U.S. 258); Anza 547 U.S. at 457 (citing Holmes, 503 U.S. at 267–68). In fact, RICO’s civil suit provision was modeled on the antitrust statutes of the early 20th century, so standards of proximate cause that applied to early Clayton and Sherman Act cases can be readily applied to RICO as well. See Holmes, 503 U.S. at 268; see also Assoc. Gen. Contractors, 459 U.S. at 531–35. Even without a deep dive into legislative history, the connection between the statutes is obvious because their civil suit provisions are almost identical. The Clayton Act reads: [A]ny person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue therefor . . . and shall recover threefold the damages by him sustained, and the cost of suit, including a reasonable attorney’s fee. 15 U.S.C. § 15. Likewise, RICO reads: Any person injured in his business or property by reason of a violation of section 1962 of this chapter may sue therefor in any appropriate United States district court and shall recover threefold the damages he 70 Case: 14-14544 Date Filed: 05/03/2019 Page: 71 of 76 sustains and the cost of the suit, including a reasonable attorney’s fee . . . . 18 U.S.C. § 1964(c). Both statutes thus say that “any person” “injured in his business or property” by a violation of the statute has a cause of action and can recover treble damages plus attorney’s fees. The common-law foundations of the Fair Housing Act are less obvious insofar as they relate to proximate cause. For starters, the FHA does not employ the language that the Clayton Act shares with RICO. It authorizes suit by an “aggrieved person,” 42 U.S.C. § 3613(a)(1)(A), expansively defined as “any person who (1) claims to have been injured by a discriminatory housing practice; or (2) believes that such person will be injured by a discriminatory housing practice that is about to occur.” Id. § 3602(i). Indeed, under the language employed by Congress in the FHA, the aggrieved person need not be injured specifically “in his business or property.” Moreover, the remedies are distinct from the relief available for a violation of the Clayton Act or RICO -- treble damages are not available, and the court “may allow” attorney’s fees, but under the FHA it is not required to, as it would be under RICO or the Clayton Act. See 42 U.S.C. § 3613(c). These textual differences make it harder to assume that the legislators drafting the FHA were drawing on the same version of proximate cause that was used in these earlier statutes. We are still able to borrow notions about proximate 71 Case: 14-14544 Date Filed: 05/03/2019 Page: 72 of 76 cause from the common law for the FHA, but it strikes us as inappropriate simply to assume that the FHA necessarily incorporated the nature and form of proximate cause as it was employed in these other statutes for which we can trace a much more direct provenance. And, since understandings of proximate cause at common law evolved over time, we would need to know which version of common–law proximate cause a statute adopted. See Sperino, supra, at 1225 (“The one thing that is certain about proximate cause is that its underlying goals are contested and evolving.”). Different statutes may not have incorporated the same common-law standard; the Palsgraf case and the First Restatement of Torts both postdated passage of the Sherman Act but had become highly foundational texts for understanding proximate cause by the time RICO was drafted. Id. The evolving shape of proximate cause at common law further complicates any analysis that would depend entirely upon incorporating a common-law standard from a statute that predated the FHA. Congress might have brought common-law standards to bear on the new problems it addressed in the 1960s, but the fact that we may identify some common-law analogue does not necessarily mean that this was the case. The Fair Housing Act does not employ language drawn from the older federal statutes with connections to the common law, and, as best as we can tell, this cause of action does not have an unmistakable common-law antecedent. As a result, while 72 Case: 14-14544 Date Filed: 05/03/2019 Page: 73 of 76 considering the common law may be valuable and informative, and while the statute’s common-law antecedents do require a plaintiff plausibly to allege “some direct relation,” we are unable to discern any further lessons from the common law that bear on our analysis. IV. Conclusion In sum, we hold that there is “some direct relation” between the City’s tax- revenue injuries and the Bank’s alleged violations of the FHA. The Supreme Court has never held that the presence of an intervening causal step or the involvement of a third party necessarily bars a finding of proximate cause as a matter of law, and we decline to establish so hard and fast a rule today. The City’s detailed allegations on its tax-revenue injury are sufficient for a variety of reasons: the FHA is a broad and ambitious statute which employs a standard of proximate cause that facilitates its operation; Congress meant for the FHA to be a big solution to a big problem -- housing segregation generated by intentional racial discrimination; for half a century courts have read the FHA broadly and Congress has repeatedly assented to these readings; the impact of the Banks’ redlining can readily be identified at a citywide or neighborhood level, so even if there were not “something very close to a 1:1 relationship,” Lexmark, 134 S. Ct. at 1394, the City has plausibly explained how it will calculate damages in a reasonably precise way; the injury is profoundly different from any injuries 73 Case: 14-14544 Date Filed: 05/03/2019 Page: 74 of 76 suffered by a homeowner; and there’s no reason to think that homeowners would make for more efficient plaintiffs. In the absence of any palpable concerns about unadministrable litigation, we ought not to unduly constrain the remedial effect of the FHA. Put differently, the City has adequately pled proximate cause when it comes to its tax-base injury because the Banks’ redlining and reverse-redlining practices bear some direct relation to the City’s fiscal injuries. There is a logical and direct bond between discriminatory lending as a pattern and practice applied to neighborhoods throughout the City and the reduction in property values. Third parties are involved, but the harm to the City is not contingent on their actions when considered in the aggregate. There is no discontinuity between the violation and the harm. Bad loans in the aggregate will mean foreclosures in the aggregate, which will mean loss of property value and a reduction in the tax base. An individual home might go under for a variety of causes, but when discriminatory lending practices pervade a neighborhood or a city, the city’s fisc will necessarily be affected because we know some number of homes will go under, and some number of properties will lose value. When we consider the claimed violations and the injuries sustained in the aggregate -- that is, at the scale that the complaints present them -- it becomes clearer that injuries to the city’s treasury are necessary, direct, and immediate results of these kinds of FHA violations. 74 Case: 14-14544 Date Filed: 05/03/2019 Page: 75 of 76 By many of these same measures, however, the City’s increased municipal expenditures injury fails, and so the district court was correct to find that proximate cause for this injury had not been adequately pled in the complaints it was reviewing. The City’s increased expenditures have not been plausibly presented as directly and automatically resulting from the Banks’ alleged conduct. Even though we think foreclosures follow directly from the Banks’ conduct, there is too much opportunity in the causal chain between foreclosure and increased expenditures for intervening actors and causes to play a role, and there has been no explanation by the City of how we might conceivably isolate the injury attributable to the Banks. Thus we have identified this alleged injury as being too remote to satisfy proximate cause. We repeat that we have not answered every outstanding question in this case. Much remains to be determined as the litigation proceeds, but this case is before this Court only on a motion to dismiss. It is not our role at this stage to decide whether hedonic regression analysis can actually identify injuries attributable to violations with sufficient accuracy, or indeed whether the banks actually made decisions based on race as opposed to socioeconomic factors that may correlate with race or with other considerations. Today we have done only two things. Broadly, we have worked out in some detail what proximate cause requires in an FHA suit. More specifically, we have evaluated whether these 75 Case: 14-14544 Date Filed: 05/03/2019 Page: 76 of 76 complaints against these Banks plausibly allege that Miami’s injuries to its tax base were directly related to the violations they describe. We simply find that the operative complaints explain in a plausible fashion how the claimed tax-revenue injuries bear “some direct relation” to the misconduct that the City is challenging. This harm to Miami, as pled, is not just foreseeable but, when measured in the aggregate, is directly related to the pattern of unlawful behavior the City has alleged. The City has plausibly alleged a violation of the FHA and has stated a claim in its First Amended Complaints. Accordingly we conclude that the district court improvidently dismissed the FHA claims in their entirety and ought to have granted the City leave to amend its complaints, since amendation would not have been futile.12 The cases are remanded to the district court for further proceedings consistent with this opinion. REVERSED and REMANDED 12 We leave it to the district court to determine which complaints should be operative for its purposes, or whether to grant the City leave to file new ones. See supra n.2. 76