What does “because of race” 1 E.g., 42 U.S.C. § 3604(a) (2014). mean in an antidiscrimination statute like the Fair Housing Act of 1968 (FHA)? 2 Pub. L. No. 90-284, 82 Stat. 81, tit. VII (codified as amended at 42 U.S.C. §§ 3601-19). The FHA prohibits discrimination based on race, color, religion, national origin, sex, familial status, and disability. See, e.g., 42 U.S.C. § 3604(a), (f). I will use the example of race here for ease of exposition, but the same points apply to the other protected statuses under the FHA. The question arose last Term in Texas Department of Housing & Community Affairs v. Inclusive Communities Project, Inc., the case in which the Supreme Court recognized a disparate impact cause of action in the FHA. 3 135 S. Ct. 2507, 2525 (2015). In dissent, Justice Alito asserted that allowing the statute to reach beyond intentional discrimination required attributing an untenable meaning to the statutory phrase “because of”—or, in Justice Alito’s words, “torturing the English language.” 4 Id. at 2534 (Alito, J., dissenting). Justice Thomas joined Justice Alito’s dissent but also wrote separately to disavow the textual interpretation underpinning the decision in Griggs v. Duke Power Co., 401 U.S. 424 (1971). See Inclusive Cmtys. Project, 135 S. Ct. at 2526-28 (Thomas, J., dissenting) (asserting that “because of” in Title VII means “motivated by”). The majority did not directly answer that point, relying instead on the way these same words were interpreted by two Supreme Court cases on Title VII of the Civil Rights Act of 1964 (Title VII) and the Age Discrimination in Employment Act of 1967 (ADEA). 5 See Inclusive Cmtys. Project, 135 S. Ct. at 2519 (“Both Title VII and the ADEA contain identical ‘because of’ language, and the Court nonetheless held those statutes impose disparate-impact liability.” (citations omitted)). See generally Smith v. City of Jackson, 544 U.S. 228 (2005) (ADEA); Griggs, 401 U.S. 424 (Title VII). In The Many Meanings of “Because of”: A Comment on Inclusive Communities Project, Noah Zatz takes on the textual challenge set out by the dissent. 6 Noah D. Zatz, The Many Meanings of “Because of”: A Comment on Inclusive Communities Project, 68 Stan. L. Rev. Online 68 (2015), https://www.stanfordlawreview.org/online/because-of. He convincingly argues that the phrase “because of” is far broader than “motivated by” and can readily encompass causal relationships that would support a disparate impact claim.
I agree with Zatz that the words “because of” present no impediment to a disparate impact claim under the Fair Housing Act. But I arrive at that conclusion after grappling with a point that Zatz’s analysis neglects: that, in ordinary speech, the set of causal relationships evoked by a phrase like “because of race” depends on the type of event to which those words are attached. For instance, there are fewer plausible causal relationships between attribute and act implicated in the sentence “A slammed the door in B’s face because of B’s race” than in the sentence “B had trouble finding an apartment because of his race.” Recognizing that different kinds of acts inflect the meaning of “because of” in different ways sheds new light on the arguments of both the majority and the dissent in Inclusive Communities—showing the dissent to be less wrong and the majority more right in their respective textual approaches than Zatz gives them credit for.
I. Acts and Attributes
The Fair Housing Act and other antidiscrimination laws formulate prohibitions by using the words “because of” to connect a protected attribute or status with specified acts or omissions. For example, the Fair Housing Act makes it unlawful “[t]o 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.” 7 42 U.S.C. § 3604(a). Although the enumerated acts cause harm because of status, liability is not triggered by the causal link between status and harm alone; rather, it is triggered by the causal link between status and particular harm-causing acts. 8 For this reason, I would quibble slightly with Zatz’s articulation of antidiscrimination law’s “three-part structure” comprising “(1) harm . . . (2) ‘because of’ . . . [and] (3) . . . protected status.” Zatz, supra note 6, at 70, 73. However, as I explain below, his formulation does capture well what antidiscrimination law aspires to do: break the causal link between status and harm. See infra note 17 and accompanying text.
This distinction matters. To say that a person experienced a harm “because of” a particular attribute opens up a very broad set of causal possibilities. Consider the sentence “B could not find an apartment on the north side of town because of his race.” The harm is the not-finding of the apartment, and it could be caused by B’s race in any number of ways: apartment managers and landlords refused to transact with B; realtors refused to show B any apartments on that side of town; B lacked transportation to the area because of racially motivated infrastructure decisions; B lacked access to racially correlated word of mouth networks that would allow him to learn about available apartments; B lacked funds to bid on an apartment in the area because of past racial discrimination in schooling or employment; B feared intimidation by racists if he ventured into that part of town; and so on.
When “because of race” is instead linked to another party’s act or omission, the range of causal possibilities depends on the type of act or omission involved. Some acts, such as “refused to rent,” represent overtly willful decisions. For such a willful decision to be “because of race” means, in ordinary language, that the decision itself is influenced or motivated by race in some way. Thus, the sentence “A refused to rent to B because of B’s race” is most naturally read to mean A’s refusal was somehow motivated by B’s race, whether consciously or not. 9 Proof problems in establishing unconscious motivations represent one rationale for disparate impact liability. See Inclusive Cmtys. Project, 135 S. Ct. at 2522.
Not all acts embody willful decisions of this sort, however; some merely describe what the actor did in the world. Zatz’s example of the driver who struck the pedestrian because the latter ignored a “Don’t Walk” signal shows this clearly. 10 Zatz, supra note 6, at 72. The pedestrian’s failure to heed the sign caused the driver to strike her, but it is unlikely that the driver was motivated to strike her because of this sign-disregarding attribute. 11 Id. A verb like “struck” does not imply the kind of overtly willful decision that is necessarily motivated by the attribute lying on the other side of “because of.” To see the distinction, modify Zatz’s example to the following: “The driver refused to brake to avoid the pedestrian because she ignored the ‘Don’t Walk’ signal.” In this case, we would readily conclude that the nonbraking was motivated by the pedestrian’s failure to heed the sign.
The reason is simple: refusal, in ordinary usage, is not just preceded by, but actually embodies, a conscious decision. And there are few ways that a decision (as opposed to a mere action) can be caused by race other than being motivated by it. 12 Dispersed or aggregate decisionmaking might present exceptions, for instance, where a refusal is carried out by a party other than the one who is motivated by race. I thank Noah Zatz for discussions on this point. By contrast, excluding someone from housing (as opposed to slamming a door), allowing an outcome (as opposed to insisting on it), and making something unavailable (as opposed to refusing to provide it) merely describe results in the world that the actor brought about—results which can be caused by race without being motivated by race. Conscious decisions produce these acts, to be sure, just as conscious decisions (to get in the car, to drive on a particular route at a particular speed) precede a driver’s striking of the pedestrian. But, as in the pedestrian case, there is room for the attribute to cause the act without also causing any decision that preceded the act.
The dissent’s argument about the plain meaning of “because of” is fatally flawed for the reasons Zatz identifies, but it does have punch in a subset of cases—those in which the words link a willful, decision-embedding act with a protected attribute. This is not to concede that such a textual analysis determines the final meaning of statutory language in general, or provisions of the Fair Housing Act in particular. 13 For example, the interaction of words with “because of” must be informed by statutory usages, which may diverge from popular usages. See, e.g., 42 U.S.C. § 3604(f)(3) (2014) (defining “discrimination” to include not accommodating disabilities). Instead, I am making a narrower point: If one looks to standard English usage for guidance, there are real differences in how different verbs inflect the meaning of “because of” when linked to protected attributes. Not all actions evoke the same set of causal possibilities.
II. Otherwise Make Unavailable or Deny
Justice Kennedy, writing for the majority in Inclusive Communities, relies on 42 U.S.C. § 3604(a)’s clause “otherwise make unavailable or deny” in concluding that the FHA includes a disparate impact cause of action. On his reading, that language “refers to the consequences of an action rather than the actor’s intent.” 14 Tex. Dep’t of Hous. & Cmty. Affairs v. Inclusive Cmtys. Project, Inc., 135 S. Ct. 2507, 2518 (2015). Zatz, however, thinks the clause “cannot bear much weight” because, like the other varieties of housing refusal, it is connected to “because of.” 15 See Zatz, supra note 6, at 69-70. On this account, all of the items in 42 U.S.C. § 3604(a)’s list of prohibited actions must textually stand or fall together as far as disparate impact is concerned. My view is different. A phrase like “otherwise make unavailable or deny,” when combined with the words “because of,” might plausibly correspond to a broader set of causal possibilities than refusals to rent, sell, or negotiate.
It is true denying a dwelling, or otherwise making it unavailable, involves voluntary acts (or, in some cases, omissions), as does refusing to rent, sell, or negotiate. But “make unavailable” operates at a broad level of generality to identify an act based on the results that its actor brought about in the world, not based on the state of mind that the actor had in performing it. As such, the act can occur because of an attribute without being motivated by that attribute. Some examples will clarify.
Example 1: Confusing Directions. A has a meeting in an unfamiliar city. She receives confusing directions to the meeting from her boss and misses the meeting as a result. When A’s boss asks A why she was not available to field client questions at the meeting, A replies: “It was because of the confusing directions you gave me. You made me unavailable for that meeting because of those confusing directions!”
Here, it is clear that the hapless A was not motivated to miss the meeting because of the confusing directions, nor did the confusing directions motivate any decision by her boss. Yet her statement is perfectly comprehensible: she was made unavailable for the meeting by the confusing nature of the directions. Had the directions not been confusing, she would have been available.
Example 2: Rain Check. A baseball stadium has a retractable roof that can be put up in case of inclement weather. The roof’s high cost was justified by the claim that once the roof was installed, the playing field would never again be unavailable because of rain. Accordingly, a contractual provision specifies fines for the stadium operator if her acts or omissions in operating the roof make the field unavailable because of rain. A fine was imposed on the operator when the retractable roof was not put up in time to avoid a rain delay. 16 See generally Patrick Dorsey, Miami Marlins Experience Opening Day Rain Delay Despite Having Retractable Roof, ESPN (Apr. 6, 2015), http://es.pn/1MZYtVw (describing such a shortfall in roof operations).
Here, it is quite unlikely that the rain motivated the stadium operator to neglect to put up the roof. Nonetheless, the operator’s failure to put up the roof made the playing field unavailable because of rain. Rain caused the field’s unavailability when coupled with the operator’s roof-positioning choices even though the rain did not cause those roof-positioning choices.
Example 3: Shortchanging Women. A moderator rarely calls on women during the question and answer period following a talk. This is because the women in the group are shorter on average than the men, and the moderator, whose vision is usually partly occluded by the podium, does not notice the women when they raise their hands (though he could easily make adjustments to overcome this issue). A university policy designed to improve gender equity at academic events prohibits refusing to call on, or otherwise making participation unavailable to, any person because of sex. The moderator is reprimanded under this policy.
Here, the moderator is not refusing to call on women because of sex. But he is nonetheless making participation unavailable to women because of sex by allowing a factor that is correlated with sex (height) to determine who gets called on and who does not. But for sex, all participants would be called on at the same rate. The moderator’s way of running the seminar thus makes participation unavailable because of sex, even though the sex of the participants does not motivate the moderator to run the seminar in that way.
III. Doing This by Doing That
The examples above provide intuitive support for the idea that “otherwise make unavailable or deny” tends to encompass a broader set of causal possibilities when linked to “because of” than does an act of outright refusal. But there is another way to explain the point. A formulation like “otherwise make unavailable” naturally lends itself to an explanatory clause that describes how one is making something unavailable, whereas other actions may not readily invite that elaboration. Compare the following two statements that might be made about a facially neutral policy, adopted without discriminatory intent, that nonetheless has a disparate impact:
Statement 1: A refused to rent to B because of race by relying on selection criteria that systematically favor white applicants.
Statement 2: A made the apartment unavailable to B because of race by relying on selection criteria that systematically favor white applicants.
Statement 2 more naturally bears the weight of the last clause, “by relying on selection criteria that systematically favor white applicants.” Indeed, one can easily imagine the shortened sentence “A made the apartment unavailable to B because of race” eliciting the question, “how?” It is far more difficult to imagine someone following up the statement “A refused to rent to B because of race” with a question like “how?” because the statement already explains how this result came about: A was motivated by B’s race to refuse the rental. In other words, a neutral policy that has a disparate racial impact can be readily understood as a way of making a dwelling unavailable because of race, but it cannot be so readily understood as a way of refusing to transact because of race.
One can of course argue that A is in a sense refusing to rent to B because of race by choosing selection criteria that favor white persons. But this seems like an aggressive imputation of motive to someone who is (by hypothesis) acting for neutral reasons. We might similarly think it an aggressive imputation of motive to say of a colleague who misses a faculty meeting because of a conflicting engagement “that is just her way of refusing to attend”—at least in the absence of any evidence that the conflict was undertaken to avoid the faculty meeting. Yet we might say that by undertaking her conflicting commitment she is, “in effect, refusing to attend.” The catchall phrase “otherwise make unavailable or deny” makes an “effective refusal” as unlawful as an actual refusal when it comes to withdrawing housing opportunities.
That the law would reach effective refusals is hardly surprising. Antidiscrimination law’s core aspiration is to erase the causal arrow that links status to harm, 17 In this aspirational sense, Zatz’s articulation of antidiscrimination law’s three-part formulation of harm, “because of,” and status, rings true. See Zatz, supra note 6, at 70, 73. to produce a society in which those with a particular status are not, for that reason, arbitrarily deprived of opportunities. Yet the law can pursue that goal only by regulating the acts and omissions of human beings—and only that subset of human beings who are presently alive and in a position to have their conduct regulated. Telling this limited group to refrain from making decisions that are motivated by race is not nearly enough to keep individuals from continuing to suffer harm because of race, given all that has already happened and all that continues to happen at the hands of those beyond the law’s reach. Requiring those regulable parties to refrain from acts and omissions that are the functional equivalent of decisions motivated by race—ones that make opportunities unavailable because of race—helps to chip away at the causal connection between status and opportunity.
In this brief reply, I have not attempted to provide a full account or defense of the disparate impact cause of action in the FHA, nor to unpack the inner workings of antidiscrimination law. Instead, I to have drawn well-deserved attention to Zatz’s important essay and to have advanced in some way the causal inquiry that he has helpfully opened up. The nub of the debate over disparate impact is whether an act that causes harm because of protected status must also have been motivated by that protected status in order to be actionable. A textual analysis can take us only so far in answering that question. But as Zatz’s cogent essay shows us, and as this reply has tried to emphasize in a somewhat different way, the FHA’s textual “because” readily extends to effects.
I thank Christopher Fennell, Lior Strahilevitz, Noah Zatz, and the students in my Fair Housing seminar at the University of Chicago Law School for helpful comments and conversations. Research support from the Stuart C. and JoAnn Nathan Faculty Fund and the Harold J. Green Faculty Fund is also gratefully acknowledged.