The Supreme Court recently surprised many observers by upholding disparate impact claims under the Fair Housing Act (FHA) in a case called Inclusive Communities Project. 1 See Tex. Dep’t of Hous. & Cmty. Affairs v. Inclusive Cmtys. Project, Inc., 135 S. Ct. 2507, 2525 (2015). Justice Kennedy’s majority opinion ratcheted down the hostility to disparate impact analysis recently on display in his Ricci v. DeStefano opinion 2 557 U.S. 557, 584-85 (2009). and in other earlier opinions he had joined. 3 See, e.g., Smith v. City of Jackson, 544 U.S. 228, 248 (2005) (O’Connor, J., concurring); Wards Cove Packing Co. v. Antonio, 490 U.S. 642, 656-58 (1989). Nonetheless, Inclusive Communities Project’s ferocious dissents joined by four Justices 4 Inclusive Cmtys. Project, 135 S. Ct. at 2526-32 (Thomas, J., dissenting); id. at 2532-51 (Alito, J., dissenting). made clear that the battle over disparate impact is hardly abating.
The dissents focused primarily on the FHA’s language proscribing practices that inflict harm “because of race” or other protected status, 5 42 U.S.C. § 3604(a) (2013). language that recurs across the modern antidiscrimination statutes. 6 See infra note 23. They rightly observed the majority opinion’s embarrassing failure to even try grounding disparate impact liability in this text. But the dissenters went astray by insisting that such an attempt would have been futile without “torturing the English language.” 7 Inclusive Cmtys. Project, 135 S. Ct. at 2534 (Alito, J., dissenting). According to the dissents, action “because of race” can mean only one thing: action motivated by a plaintiff’s race, 8 Id.; id. at 2526-27 (Thomas, J., dissenting). the touchstone of disparate treatment liability. Therefore, they conclude, the statutory text precludes disparate impact claims, which, by definition, do not require proof of disparate treatment.
This Essay argues that, to the contrary, “because of” suggests a causal concept more expansive than even the broadest accounts of disparate treatment.Releasing “because of” from the dissenters’ straitjacket provides the textual flexibility necessary to justify disparate impact claims under the FHA and similar antidiscriminiation statutes. Indeed, a proper reading of this language does just the opposite of what the Inclusive Communities Project dissenters claim: it provides the basis for a revived antidiscrimination jurisprudence, something the Court’s liberal wing has failed to offer in recent years.
I. The “Because Of” Problem
Inclusive Communities Project rests on two pillars. The first is the FHA’s reference to practices that “otherwise make unavailable” housing units. 9 Id. at 2518-19 (majority opinion) (quoting § 3604(a)). The second is the Court’s groundbreaking 1971 decision in Griggs v. Duke Power Co., 10 Id. at 2516-18 (citing Griggs v. Duke Power Co., 401 U.S. 424 (1971)). which permitted disparate impact employment discrimination claims under Title VII of the Civil Rights Act of 1964 (Title VII). 11 42 U.S.C. § 2000e-2(a) (2013). Cross-bracing these pillars is the parallelism between the FHA’s text and Title VII’s, including Title VII’s use of the phrase “otherwise adversely affect.” 12 Inclusive Cmtys. Project, 135 S. Ct. at 2516-19 (quoting § 2000e-2(a)(2)). The dissenters agreed that Title VII and the FHA require a consistent interpretation. For them, however, Griggs was “made of sand” and should be repudiated, 13 Id. at 2526 (Thomas, J., dissenting). rather than made the foundation for extending disparate impact liability to housing. 14 Id. at 2544-45 (Alito, J., dissenting) (criticizing Griggs for its “text-free reasoning” and suggesting various narrowing interpretations). Because Congress later codified Griggs, 42 U.S.C. § 2000e-2(k) (2013), repudiating the opinion would not eliminate Title VII disparate impact liability. Cf. Ricci v. DeStefano, 557 U.S. 557, 577 (2009) (“The Civil Rights Act of 1964 did not include an express prohibition on policies or practices that produce a disparate impact.”).
The first pillar, “otherwise make unavailable,” cannot bear much weight. The full text of the operative section of the FHA is as follows:
[I]t shall be unlawful—
(a) To refuse to sell or rent after the making of a bona fide offer, or to refuse to negotiate for the sale or rental of, or otherwise make unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status, or national origin. 15 42 U.S.C. § 3604(a) (2013).
Justice Kennedy reasoned that the references to “refus[ing] to sell or rent” prohibit disparate treatment while “otherwise make unavailable” functions as a “catchall phrase looking to consequences [such as unavailability], not intent.” 16 Inclusive Cmtys. Project, 135 S. Ct. at 2519. This interpretation ignores how the catchall phrase functions as part of a list. 17 See id. at 2527 (Thomas, J., dissenting); id. at 2533-35 (Alito, J., dissenting). A list of what? Not a list of theories of discrimination, but a list of housing harms prohibited if imposed in discriminatory fashion. 18 Another FHA provision addressed to brokers and real estate agents prohibits “discriminat[ion] against any person in making available such a [housing] transaction . . . because of race, [etc.].” 42 U.S.C. § 3605(a) (2013). In that provision, “making available” acts as a generic term that includes the specific acts (sales or rentals) referenced in § 3604(a) but which receive no separate mention in § 3605(a). The FHA does not ban “refus[ing] to sell or rent” tout court, and so it is bizarre to treat “otherwise make unavailable” as generating its own legal claim. 19 Moreover, refusing to sell or rent are merely specific ways an owner might “make unavailable” a dwelling. “Refusing” and “making” are both intentional, and both bring about the result of unavailability. Instead, each listed harm is modified by “because of race, [etc.].” Neither refusing to rent nor “otherwise mak[ing] unavailable” a dwelling violates the statute unless done “because of race.” Thus, the dissents were correct that “otherwise make unavailable” is uninteresting and that the real action lies in how disparate impact liability coheres with a proscription on harms imposed “because of race.”
So what does “because of race” mean? Unfortunately, the majority opinion had nothing to say. 20 Justice Kennedy was in a tight spot, having previously accepted the dissenters’ reading of “because of” by joining a concurrence rejecting disparate impact claims of age discrimination under the Age Discrimination in Employment Act. See Smith v. City of Jackson, 544 U.S. 228, 249 (2005) (O’Connor, J., concurring). The four Justices who joined Justice Kennedy’s opinion in Inclusive Communities Project faced no such constraint, yet none wrote separately to provide a more robust analysis. Instead, it merely noted that the dissenters’ critique applies with equal force to Griggs’s interpretation of Title VII. 21 See Inclusive Cmtys. Project, 135 S. Ct. at 2519. Therefore, because stare decisis demands preserving Griggs, the dissenters must be wrong. 22 See id. at 2518. The Court did not distinguish the FHA’s “because of race” as broader than Title VII’s “because of such individual’s race,” despite having done so previously when allowing a Title IX claim by an individual who suffered retaliation for opposing sex discrimination against someone else. See Jackson v. Birmingham Bd. of Educ., 544 U.S. 167, 179 (2005). Justice Kennedy had, however, joined the Jackson dissenters in reasoning that “the natural meaning of the phrase ‘on the basis of sex’ is on the basis of the plaintiff’s sex, not the sex of some other person.” Id. at 185 (Thomas, J., dissenting). We just have no idea why.
This retreat to stare decisis is a rather uninspiring way to settle a fundamental question about the structure of antidiscrimination law. All modern antidiscrimination statutes have the same three-part structure seen in the FHA and Title VII: (1) harm (to housing, employment, or other interests) (2) inflicted “because of” (3) the plaintiff’s protected status (race, sex, age, disability, etc.). 23 See 29 U.S.C. § 623(a)(1)-(2) (2013) (age discrimination); 42 U.S.C. § 2000ff-1(a) (2013) (genetic information discrimination); 42 U.S.C. § 12112(a)-(b) (2013) (disability discrimination). As the FHA illustrates, “because of” is the common denominator, appearing in provisions that lack any reference to “discrimination” or its cognates. See, e.g., 42 U.S.C. § 3604(a) (2013); see also 42 U.S.C. § 2000e-2(a)(2) (proscribing under Title VII employment practices that inflict harm “because of such individual’s race” and without referencing “discrimination” or its cognates). The majority’s silence not only leaves Griggs vulnerable but also creates a theoretical vacuum at the center of the field. 24 Nor did any of the four Justices who joined Justice Kennedy’s opinion write separately to fill this void. Similarly useless is the majority’s passing invocation of the FHA’s statutory purpose to eradicate “discriminatory practices” that “unfairly . . . exclude.” 25 Inclusive Cmtys. Project, 135 S. Ct. at 2521-22. The reader never learns what makes something “discriminatory” or “unfair,” let alone how those concepts relate to conduct “because of race.”
The dissenters would fill this void with a simple idea: “because of” means “motivated by.” 26 Id. at 2526-27 (Thomas, J., dissenting); id. at 2533-34 (Alito, J., dissenting). In other words, “because of race” limits actionable conduct to what is variously termed “disparate treatment,” 27 Id. at 2533 (Alito, J., dissenting). “intentional discrimination,” 28 Id, at 2537. or actions made with “discriminatory intent,” 29 Id. at 2546; id. at 2527 (Thomas, J., dissenting). In keeping with standard usage, the dissenters treat these all as equivalents. all of which exclude disparate impact liability. 30 While one dissent argued that disparate impact provides “evidence of disparate treatment,” id. at 2550 (Alito, J., dissenting), Griggs v. Duke Power Co. rejected the need to draw an inference of disparate treatment, 401 U.S. 424, 432 (1971). And the Inclusive Communities Project majority refused to limit the rationale for disparate impact liability to its ability to “uncover discriminatory intent.” 135 S. Ct. at 2522.
This interpretation of “because of” is not a technical feature of “legal usage” or precedent, 31 Inclusive Cmtys. Project, 135 S. Ct. at 2534 (Alito, J., dissenting). the dissenters insist. Instead, it flows from the “ordinary meaning of ‘because of.’” 32 Id. at 2533 (quoting Univ. of Tex. Sw. Med. Ctr. v. Nassar, 133 S. Ct. 2517, 2527 (2013)); id. at 2526 (Thomas, J., dissenting) (same). Nassar involved the somewhat different question of whether the employer’s motivation must be a “but-for” rather than “substantial” cause of its action. See id. at 2523 (majority opinion). The dissenters asserted that this meaning “cannot be denied.” 33 Inclusive Cmtys. Project, 135 S. Ct. at 2534 (Alito, J., dissenting). Anything else would “tortur[e] the English language.” 34 Id. As evidence, Justice Alito’s dissent cited every Washington Post article that used the phrase “because of” published the day Inclusive Communities Project was argued. 35 Id. at 2534 & n.2. Each one, he asserts, uses “because of” to “link an action and a reason for the action.” 36 Id. at 2534.
II. Motivation as One Species of Causation
The dissenters’ interpretation of “because of” is built on sand. “[A] reason for the action” 37 Id. is itself ambiguous. To identify a reason for an event is to identify one of its causes. Motivations can provide causes, but not all causes consist of motivations. If the reason a building collapsed is that an earthquake struck, nothing has been said about anyone’s motivations or intentions. This point is repeatedly borne out in Justice Alito’s own catalogue of examples, most obviously in the sentence, “[A] circuit breaker automatically opened because of electrical arcing.” 38 Id. at 2534 n.2 (alteration in original) (quoting Paul D. Blumstein, Letter to the Editor, Metro’s Safety Flaws, Wash. Post, Jan. 21, 2015, at A20).
This point about reasons as causes applies fully to human conduct. Consider the following sentence: “[T]he driver struck the pedestrian because of the pedestrian’s failure to heed the ‘Don’t Walk’ signal.” This could have the dissenters’ meaning: the pedestrian’s behavior motivated the driver to strike her. Had the pedestrian stepped into the street under different circumstances, the driver would have felt differently about her and would have avoided the collision. Absent this motivation, the driver would not have struck the pedestrian, and absent the signal disobedience, there would have been no such motivation. The driver committed disparate treatment of this pedestrian relative to an innocent one who engaged in the same physical conduct.
This disparate treatment interpretation is hardly inevitable, however. An alternative is that the pedestrian stepped into moving traffic, leaving the driver insufficient braking distance to avoid striking her regardless of how the driver felt about signal disobedience. 39 This analysis need not strip the driver of agency or responsibility. The driver may have made decisions that eliminated his or her opportunity to stop in time. Imagine a driver who obeys traffic signals but not speed limits. Such a driver would not have struck a pedestrian crossing on a walk signal and, had he not been speeding, might have avoided striking the scofflaw pedestrian. Again, this provides a “reason” for what happened. Had the pedestrian obeyed the signal, she would not have been struck. The difference is just the particular causal pathway leading from stepping off the curb to being struck; one runs through the motivations of the driver, the other does not. 40 The same distinction applies even if one construes “motivation” broadly to embrace any mental process, including subconscious ones such as “implicit bias,” that responds to the plaintiff’s protected status. See Stephen M. Rich, Against Prejudice, 80 Geo. Wash. L. Rev. 1, 45-48 (2011).
The phrase “because of” refers directly to causal concepts, not mental ones. Indeed, explicit references to motivations or intentions are absent from the principal provisions of the FHA and other statutes discussed in Inclusive Communities Project. Motivation, though, is not opposed to causation in the way that courts and commentators too often oppose disparate treatment to disparate impact in terms of intent versus effects. Rather, a plaintiff’s protected status may affect a defendant’s motivations, which may affect the defendant’s behavior that harms the plaintiff. Motivations are one way to satisfy the formula (1) harm, (2) because of, (3) protected status, just not the only way.
For example, consider an FHA claim by a tenant who suffers severe racial harassment by other tenants. Now imagine that the landlord, despite having notice of the situation, refuses to modify its practices that facilitate or fail to prevent the harassment. On the one hand, the tenant’s housing conditions suffer because of her race. On the other, the landlord’s conduct may not be motivated by her race; it simply might want to avoid getting involved in a dispute between tenants. Under analogous circumstances of workplace harassment, courts uniformly rely on causation, not motivation, to treat Title VII’s “because of” requirement as satisfied; 41 See, e.g., Dunn v. Wash. Cty. Hosp., 429 F.3d 689, 691 (7th Cir. 2005); Noah D. Zatz, Managing the Macaw: Third-Party Harassers, Accommodation, and the Disaggregation of Discriminatory Intent, 109 Colum. L. Rev. 1357, 1377-80, 1400-03 (2009). the weight of FHA authority does the same. 42 See Neudecker v. Boisclair Corp., 351 F.3d 361, 364-65 (8th Cir. 2003); Martinez v. Cal. Inv’rs XII, No. CV 05-7608-JTL, 2007 WL 8435675, at *6 (C.D. Cal. Dec. 12, 2007); see also Davis ex rel. LaShonda D. v. Monroe Cty. Bd. of Educ., 526 U.S. 629, 633, 644-45 (1999) (allowing Title IX suits against schools for failing to prevent student-on-student harassment). The dissenters’ interpretation of “because of” could not allow this.
III. Legal Precedent For Causation Without Motivation
Additional examples demonstrate that my causal interpretation of causal language is as familiar to antidiscrimination law as to the Washington Post. Consider the Americans with Disabilities Act of 1990 (ADA): “No covered entity shall discriminate against a qualified individual on the basis of disability in regard to job application procedures, the hiring, advancement, or discharge of employees, employee compensation, job training, and other terms, conditions, and privileges of employment.” 43 42 U.S.C. § 12112(a) (2013). The ADA’s employment discrimination provision follows the standard formula: (1) harm (hiring, discharge, etc.), (2) causation (“on the basis of”), (3) protected status (disability). So what does this mean?
The ADA’s text further specifies that “the term ‘discriminate . . . on the basis of disability’ includes . . . not making reasonable accommodations to the known physical or mental limitations of an otherwise qualified individual with a disability.” 44 Id. § 12112(b)(5)(A). Thus, Congress explicitly treated the ADA’s provision for nonaccommodation liability as a species of the more general concept of discrimination “because of” or “on the basis of.” 45 Antidiscrimination law generally uses “because of,” “on the basis of,” and “based on” as equivalents. See, e.g., 42 U.S.C. § 2000e(k) (2013); id. § 2000e-2(e); Smith v. City of Jackson, 544 U.S. 228, 238-39 (2005) (plurality opinion); Smith, 544 U.S. at 246 (Scalia, J., concurring in part and concurring in the judgment). Indeed, the 1988 amendments to the FHA had done much the same thing when prohibiting disability discrimination. 46 42 U.S.C. § 3604(f)(1) (making it unlawful to “discriminate . . . because of a handicap”); id. § 3604(f)(3)(B) (specifying that “discrimination includes . . . a refusal to make reasonable accommodations”). All this is completely consistent with the causal interpretation given above but defies the restriction of “because of” to “motivated by.” 47 See Zatz, supra note 41, at 1400-03, 1406-14 (arguing that disparate treatment, third-party harassment, and nonaccommodation liability all rely on the same causal conception of discriminatory injury).
In a classic nonaccommodation claim, a worker loses a job because she cannot use a tool, and she cannot use the tool because of her disability. The employer does not care why the worker cannot use the tool; any worker who cannot use the tool is excluded. Thus, the worker loses a job “because of” her disability in the causal sense, even though the employer is not motivated by her disability in the sense contemplated by disparate treatment liability. 48 See Parker v. Columbia Pictures Indus., 204 F.3d 326, 338 (2d Cir. 2000). Again, this remains true even on the most expansive conceptions of disparate treatment that include implicit bias.
Not only has Congress explicitly used “because of” in the way the dissenters declared unthinkable, but lower courts rely on this causal analysis to do doctrinal work. They have rejected liability where the failure to accommodate established no causal connection between disability and injury. That happens either when the failure to accommodate did not cause the injury 49 Parker v. Sony Pictures Entm’t, Inc., 260 F.3d 100, 108 (2d Cir. 2001). or when the plaintiff’s disability was not the reason an accommodation was needed. 50 See Cheryl L. Anderson, What Is “Because of the Disability” Under the Americans with Disabilities Act?: Reasonable Accommodation, Causation, and the Windfall Doctrine, 27 Berkeley J. Emp. & Lab. L. 323, 349-56 (2006). Thus, “[a]lthough ‘discriminate’ is defined in very broad terms, that expansive definition does not change the requirement that to be actionable the discrimination must be ‘because of the disability.’” 51 Felix v. N.Y.C. Transit Auth., 324 F.3d 102, 105 (2d Cir. 2003); accord Columbia Pictures Indus., 204 F.3d at 337.
Title VII’s “because of” language also has been construed to go beyond “motivated by” while remaining faithful to a causal interpretation. Most notably, the Supreme Court did just that in its first encounter with reasonable accommodation of religion. Trans World Airlines, Inc. v. Hardison considered a claim by a plaintiff who had been fired for refusing to work on Saturdays. 52 432 U.S. 63, 69 (1977). This refusal stemmed from the plaintiff’s religious beliefs, but the employer would have responded the same way to any worker’s refusal to work on Saturdays, irrespective of his or her religion or reason for refusing. 53 Id. at 82 (“There has been no suggestion of discriminatory intent in this case.”). Thus, the worker’s religion did not motivate the employer’s conduct, yet the worker’s religion did cause that conduct. 54 Congress elided this distinction when, after the Hardison dispute arose, it amended Title VII to define “religion” to include “religious . . . practice[s]” like not working on Saturdays. 42 U.S.C. § 2000e(j) (2013). The resulting legal fiction characterizes the employer’s nonaccommodation as disparate treatment, see EEOC v. Abercrombie & Fitch Stores, Inc., 135 S. Ct. 2028, 2033 (2015), notwithstanding the employer’s indifference to the worker’s religion, see Zatz, supra note 41, at 1428-29. His religion was the reason he refused to work, which was the reason the employer fired him. The Court held that this scenario could run afoul of Title VII’s ban on “discriminat[ion] . . . because of such individual’s . . . religion,” 55 Hardison, 432 U.S. at 74-76, 76 n.11 (discussing § 2000e-2(a)(1)). The Court held nonaccommodation liability to be a “defensible construction” of the pre-amendment statute. Id. at 76 n.11. Moreover, it did so under section 2000e-2(a)(1), id. at 74, not section 2000e-2(a)(2), which contains the “otherwise adversely affect” language. See also Zatz, supra note 41, at 1428-29 (making similar points about Hardison). Nonetheless, other considerations defeated the plaintiff’s claim. Hardison, 432 U.S. at 84. though it did not explain why.
This short Essay has not tried to establish affirmatively how the FHA and similar statutes provide for disparate impact claims. More modestly, I have tried to demolish the notion that their “because of” language necessarily stands in the way. The Inclusive Communities Project dissents are simply wrong that “because of” can mean nothing other than “motivated by.” They are wrong as a matter of plain English, and they are wrong as a matter of relevant legal usage. Instead, the statutes’ core causal concepts open the door to a more robust antidiscrimination law, one in which discriminatory intent is relevant but not essential. Charting that path is a broader project underway in other work. 56 See, e.g., Noah D. Zatz, Disparate Impact and the Unity of Equality Law (July 15, 2015) (unpublished manuscript) (on file with author).
The Author thanks Seth Williams for research assistance and Lee Fennell for comments on a draft.