‘Sex’ and Religion after Bostock†
Professor of Law, University of Connecticut. and Professor of Law, Saint Louis University.
Sachin S. Pandya*
This paper reviews the U.S. Supreme Court’s opinion in Bostock v. Clayton County. There, the Court held that by barring employer discrimination against any individual “because of such individual’s . . . sex,” Title VII of the Civil Rights Act of 1964 also bars employment discrimination because an individual is gay or transgender. The paper then speculates about how much Bostock will affect how likely lower court judges will read other “sex” discrimination prohibitions in the U.S. Code in the same way, in part based on a canvass of the text of about 150 of those prohibitions. The paper also discusses the religion-based defenses that defendants may raise in response under Title VII itself, the Religious Freedom Restoration Act, and the First Amendment of the U.S. Constitution. And the paper suggests how Bostock’s effect will likely vary with the influence of Trump-appointed federal judges.
I. The Opinion
Bostock involved three lawsuits, all of which raised the question of whether Title VII’s prohibition on employer sex discrimination covers discrimination against gay or transgender individuals. In Bostock, Gerald Bostock worked for Clayton County, Georgia, as a child welfare advocate. When he joined a gay softball league, he was fired. In Zarda v. Altitude Express, Donald Zarda worked for Altitude Express as a skydiving instructor in New York City. Days after Zarda mentioned to a female customer that he was gay, he was fired. In EEOC v. R.G. & G.R. Harris Funeral Homes, Aimee Stephens, a transgender woman, worked for R.G. & G.R. Harris Funeral Homes in Michigan as a funeral director. Stephens, assigned the male sex at birth and then-presenting as a man, was fired when she told her boss that, after returning from vacation, she would be Aimee and would present as a woman. Bostock and Zarda had sued their former employers, alleging, among other claims, that they had been fired for being gay in violation of Title VII’s prohibition on employer sex discrimination. In Stephens’s case, the Equal Employment Opportunity Commission (EEOC) sued, alleging that by firing Stephens, Harris Funeral Homes had violated Title VII’s prohibition on employer sex discrimination.
The Supreme Court, in a 6–3 majority opinion by Justice Neil Gorsuch, ruled that by barring employer discrimination against any individual “because of such individual’s . . . sex,” section 703(a)(1) of Title VII also bars employment discrimination because an individual is gay or transgender. Justices Samuel Alito and Brett Kavanaugh opined in dissent.
The Court in Bostock described its task as determining “the ordinary public meaning” of section 703(a)(1) when Congress enacted Title VII in 1964. Both then and now, section 703(a)(1) declares it unlawful for an employer to “fail or refuse to hire or to discharge any individual, or otherwise to discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual’s . . . sex.” Bostock started by assuming arguendo that the term “sex” refers “only to biological distinctions between male and female.” Then, Bostock relied on two other features of section 703(a)(1)’s text.
First, the phrase “because of” denoted “the ‘simple’ and ‘traditional’ standard of but-for causation.” That test, together with the term “discriminate” (already read to require an intentional difference in treatment) implies that “an employer who intentionally treats a person worse because of sex—such as by firing the person for actions or attributes it would tolerate in an individual of another sex—discriminates against that person in violation of Title VII.”
Second, Bostock relied on the references to the “individual” in section 703(a)(1)’s text. On its own, Bostock observed, the term “discriminate” might be read to refer to “the employer’s treatment of groups rather than individuals, to see how a policy affects one sex as a whole versus the other as a whole. . . . So how can we tell which sense, individual or group, ‘discriminate’ carries in Title VII?” Bostock’s answer: “The statute . . . tells us three times . . . that our focus should be on individuals, not groups: Employers may not ‘fail or refuse to hire or . . . discharge any individual, or otherwise . . . discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual’s . . . sex.’ § 2000e–2(a)(1) (emphasis added).”
From this “ordinary public meaning” of section 703(a)(1)’s text, Bostock inferred that an employer “violates Title VII when it intentionally fires an individual employee based in part on sex. It doesn’t matter if other factors besides the plaintiff’s sex contributed to the decision. And it doesn’t matter if the employer treated women as a group the same when compared to men as a group.”
In turn, this meant that section 703(a)(1) required employers to treat “[a]n individual’s homosexuality or transgender status [as] not relevant to employment decisions. That’s because it is impossible to discriminate against a person for being homosexual or transgender without discriminating against that individual based on sex.” To show this, Bostock ran through many hypotheticals. For example, suppose an employer has two employees, one male and the other female. Both are attracted to men and are otherwise identical. If the employer fires the male employee because he is attracted to men, the employer discriminates against him for traits or conduct it accepts from the female employee. Similarly, if an employer has two employees who are female, but fires one because she was identified as male at birth, but keeps the other who was identified as female at birth, the fired “employee’s sex plays an unmistakable and impermissible role in the decision.” Where “sex” is a but-for cause, it does not matter that other factors may also have played a role in the decision. Bostock also stressed that because the ordinary public meaning of section 703(a)(1)’s text was unambiguous, it did not matter that, because of that text’s breadth, Congress in 1964 may not have foreseen that it would apply to protect gay and transgender individuals.
II. The Bostock Effect
How much will Bostock affect what the lower courts do? Justice Alito, for one, wrote that he was “virtually certain” that Bostock would have “far-reaching consequences,” citing the “over 100 federal statutes [that] prohibit discrimination because of sex.” We also believe that Bostock will make lower courts more likely to read other sex discrimination bans in the U.S. Code to protect gay and transgender individuals. But unlike Justice Alito, we have less confidence and more caveats about how much more likely.
Bostock’s core premise is that if a statute focuses on the individual and prohibits sex discrimination as section 703(a)(1) of Title VII does, it necessarily prohibits discrimination against anyone for being gay or transgender. This premise readily extends to bisexual, heterosexual, and cisgender individuals, among others, because discriminating against any such individual on that basis also necessarily makes relevant whether that individual is taken to be a man or a woman.
Yet, Bostock depends on concluding that, for any particular U.S. Code ban on sex discrimination, its text, like section 703(a)(1), points to (1) a focus on the individual, not the group, and (2) a relationship between the discrimination and “sex” must satisfy no more than the traditional but-for cause standard. That means that Bostock has escape hatches: Find instead that the statutory text indicates a focus on the group and not the individual or find instead that the statutory text indicates something more stringent than but-for cause.
Alternatively, a judge could find that the text’s “ordinary public meaning” is ambiguous on these two issues, and then turn to extra-textual considerations, such as substantive canons of construction, deference doctrines, or arguments about statutory purpose, to reach a different conclusion. Suppose judges can plausibly disagree (with negligible risk of professional embarrassment) about whether a sex discrimination ban’s text is ambiguous on these two issues. A judge may sincerely think the text is ambiguous in this way. A judge may deliberately call it ambiguous as a pretext to get Bostock out of the way to rule according to that judge’s ideological or other preferences. Or a judge, though trying to set aside those preferences, may unwittingly tend to take that text as ambiguous in cases where doing so lines up with those preferences. In any case, there is no accepted objective way to test whether a judge has erred in declaring the text ambiguous enough, apart from at least five Supreme Court justices saying so.
Accordingly, we expect Bostock’s effect to vary in part with how hard it is for lower court judges to write an opinion concluding that the statute’s text unambiguously focuses on the individual or the group and requires no more than but-for cause. Sometimes, the text plausibly cuts only one way. For example, consider the Title VII sex discrimination provisions not at issue in Bostock. Much as they denote a focus on the individual in section 703(a)(1), the terms “any individual” and “such individual’s” function in the same way in section 703(a)(2), as do similar uses of “any individual” and close variants under section 703’s parallel provisions for employment agencies, labor organizations, training programs, among others. What’s more, textualism lets judges rely on semantic canons of construction, including the one that says that the same terms within the same Act should be read to carry the same meaning. Thus, because Bostock reads section 703(a)(1) to focus on the individual, lower courts are likely to read the similar provisions in the rest of section 703 in the same way. After all, in the few instances in section 703 when it intended to refer to groups, Congress used the term “group.”
III. Applying Bostock outside Title VII
What about Bostock’s effect on “sex” discrimination provisions elsewhere in the U.S. Code? To roughly sketch that effect, we started with Appendix C of Justice Alito’s dissent in Bostock. There, Justice Alito purported to list the “over 100 federal statutes [that] prohibit discrimination because of sex” to support his view that “[w]hat the Court has done today––interpreting discrimination because of ‘sex’ to encompass discrimination because of sexual orientation or gender identity––is virtually certain to have far-reaching consequences.”
Unfortunately, because Appendix C does not indicate how Justice Alito or his staff identified the statutory provisions he listed, we could not reproduce it independently. For convenience, instead of taking our own census of the U.S. Code’s sex discrimination provisions, we relied on Appendix C anyway. But we focused on the statutory subsection, not the statutory section, as the unit of analysis. Thus, where Appendix C cited a provision that contained a sex discrimination prohibition in more than one of its subsections, we treated each subsection as a separate observation. Then, we further identified (1) the text denoting the requisite relationship between the discriminatory conduct and sex (e.g., “because of . . . sex,” “on the basis of . . . sex”); and (2) the text denoting who it protected from illegal discrimination (e.g., “individual,” “employee”).
The resulting dataset initially consisted of 187 observations. We dropped provisions that Justice Alito had cited in Appendix C that used the term “gender” rather than “sex.” We screened out provisions that, on their own, were statutory congressional findings, or statements of policy or principles, on the premise that, though useful for interpretation, they alone carry no independent force of law. We also dropped provisions that simply incorporated by reference another sex discrimination provision in the U.S. Code, in state law, or more generally referred to other laws that prohibited sex discrimination. We also excluded Title VII, section 703. The resulting final dataset had 151 observations.
Table 1 summarizes how the text in these provisions denote the relationship between discriminatory conduct and sex.
|on the basis of . . . sex||57|
|on the ground of . . . sex||19|
|because of . . . sex||18|
|based on . . . sex||8|
|on account of . . . sex||8|
|on the grounds of . . . sex||7|
|without regard to . . . sex||6|
|on the basis of race, sex, or ethnic groups||4|
|based upon . . . sex||3|
|because of the borrower’s sex||3|
|by reason of . . . sex||2|
|of a particular . . . sex||2|
|because of the person’s sex||1|
|entirely neutral as to the . . . sex . . . of||1|
|made or based upon difference in . . . sex||1|
|not solely be based on the . . . sex . . . of||1|
|of a specified . . . sex||1|
|on account of his or her . . . sex||1|
|on account of the . . . sex . . . of||1|
|on the basis of . . . sex . . . of||1|
|on the basis of . . . sex, opposite sex||1|
|on the basis of that person’s . . . sex||1|
|shall not consider the . . . sex . . . of||1|
|take into account . . . the . . . sex . . . of||1|
|take sex into account||1|
|without distinction as to . . . sex||1|
As Table 1 suggests, most of the sex discrimination provisions in the U.S. Code use language similar to “because of . . . sex” in section 703(a)(1) of Title VII. In turn, Bostock treats “because of” in section 703(a)(1) to mean “by reason of” or “on account of,” all equivalent ways in which Congress can denote the traditional but-for cause standard. Bostock itself also described its holding in a way that suggests that “because of such individual’s . . . sex” and “on the basis of . . . sex” are interchangeable. Thus, Bostock makes it more likely that lower courts will read the text of the other sex discrimination statutes in the same way that Bostock read section 703(a)(1) of Title VII.
Bostock, however, also pointed to how different statutory language might have led the Court to infer differently. For example, Congress might have added the word “solely” to “indicate that actions taken ‘because of’ the confluence of multiple factors do not violate the law” or used the phrase “‘primarily because of’ to indicate that the prohibited factor had to be the main cause of the defendant’s challenged employment decision.”
This part of Bostock makes it easier for judges reading other sex discrimination statutes with these features to distinguish Bostock away. For example, when Congress immunized owners of ammonium nitrate facilities from civil liability for refusing to sell ammonium nitrate to “any person” based on a good-faith “reasonable belief” that the person wants to use it “to create an explosive device to be employed in an act of terrorism,” Congress added that the required “reasonable belief . . . may not solely be based on the . . . sex . . . of that person.” Because of the word “solely” in this provision, a lower court is now more likely to find the requisite “reasonable belief” even though the owner in part refused to sell because the person was gay or transgender.
Bostock also turned on section 703(a)(1)’s focus on the individual, not the group. Accordingly, Table 2 summarizes how the text of the other sex discrimination provisions in the U.S. Code denote the type of actors to be protected from sex discrimination.
|Office in the Corporation||2|
|Amateur Athletes, Coaches, Trainers, Managers, Administrators, and Officials||1|
|Child, Family of the Child||1|
|Individual, Position held by Individual||1|
|Individual, Officer, Employee, Agent, Director||1|
|Officer, Employee, Applicant||1|
|Person, Small Business Concern||1|
|Person, Class of Persons||1|
|Person; Owner, Officer, Director, or Employee of such Person||1|
|Sources of Qualified Applicants||1|
As Table 2 suggests, these sex discrimination provisions in the U.S. Code vary more in the way the text identifies who is protected than that text denotes the causation standard.
First, some provisions protect an “individual” from sex discrimination, either as a direct object (discriminating against “any individual”) or as the subject of the sentence (no “individual” shall be discriminated against). Bostock stressed section 703(a)(1)’s uses of the word “individual” as “tell[ing] us . . . that our focus should be on individuals, not groups.” And Bostock suggested how different statutory text might have led the Court to infer otherwise.
But not every textual difference matters. For example, we bet that lower courts will read Bostock as coming out no differently even if section 703(a)(1) had used the plural “individuals” instead of the singular “individual.” The reason: The Dictionary Act provides that, for any “Act of Congress,” unless “context” indicates otherwise, “words importing the singular include and apply to several persons, parties, or things” and “words importing the plural include the singular.” In turn, the term “context” in the Dictionary Act has been read to mean “the text of the Act of Congress surrounding the word at issue, or the texts of other related congressional Acts.” As a result, when Bostock reasoned that section 703(a)(1) focused on the individual, not the group, it would not have taken the plural form “individuals” as indicating a focus on the group over the individual. Besides, the context—here, the other words in section 703—cut the other way. In section 703, when Congress wanted to refer to the group, it used the word “group.”
Second, most of the sex discrimination provisions in the U.S. Code protect any “person” from sex discrimination, either as a direct object (discriminating against “any person”) or as the subject of the sentence (“No person” shall be discriminated against). In the U.S. Code, the default reading of “person” requires focusing on the individual and on some kinds of non-corporeal entities that law treats, in some measure, as if they act in the world as an individual could. Again, the reason is the Dictionary Act, which provides that, for any “Act of Congress, unless context indicates otherwise,” the word “‘person’ . . . include[s] corporations, companies, associations, firms, partnerships, societies, and joint stock companies, as well as individuals.” Missing from this list: the term “group,” which Congress has used when defining “person” elsewhere in the U.S. Code. Here, in the context of sex discrimination, the word “person” squarely focuses on the individual, not the group, absent more textual cues to the contrary.
Third, in a few subsections in Title 12 of the U.S. Code, sex discrimination provisions exist that use the word “groups.” For example, in disposing of assets as an appointed conservator or receiver, the Federal Deposit Insurance Corporation (FDIC) must act in a manner that “prohibits discrimination on the basis of race, sex, or ethnic groups in the solicitation and consideration of offers.”
Does this provision focus on the group and not the individual? If “race,” “sex,” and “ethnic” are adjectives that all modify “groups,” then perhaps the statute focuses on groups alone. If so read, a lawyer could wield Bostock to make it less likely that this sex discrimination provision covers gay or transgender individuals, because it covers sex groups, and thus can be read to let the FDIC treat actual and prospective offerers comparably as groups of men and women. As writers of the English language, we are skeptical of this reading. It implies that “race” here is an adjective, while ordinary English usage prefers “racial” (adjective) to “race” (noun) when describing a group (“racial group” over “race group”). Besides, we can’t find the phrase “sex groups” in the current U.S. Code or, for that matter, in any volume of the Statutes at Large. On the other hand, if only “ethnic” modifies “groups” (“ethnic groups,” a phrase Congress has used elsewhere), then the subsection’s text alone leaves it unclear whether to focus on the individual only, the group only, or both depending on case context or statutory purpose.
Fourth, what about provisions that identify the direct object of the illegal sex discrimination by a more particular category, such as “employee,” “applicant,” “member” or seeker of “membership,” “students,” or “citizen”? Again, context matters a lot. For example, in Title 36 of the U.S. Code, Congress created some national organizations for military veterans and, in so doing, often provided that the requirements for “membership” in, or to serve as “director” or “officer” of, such an organization “may not discriminate on the basis of . . . sex.” Writers of English would typically use those words to refer to how an organization treats someone who wanted to join it, or to serve as one of its directors or officers—a focus on the individual. Accordingly, a lawyer wielding Bostock can credibly argue that a court must read “the basis of . . . sex” to cover otherwise eligible gay and transgender individuals who want to join, say, the Air Force Sergeants Association. If other textual cues indicate that this sex discrimination provision focuses on the group, however, then a judge is more likely to distinguish Bostock and read the statute to let the Association disfavor any particular individual who wants to join for being gay or transgender, so long as that Association treats men and woman comparably as groups when deciding who gets in.
Finally, what about a provision that does not identify who it protects against sex discrimination (labeled in Table 2 as “unspecified”)? Bostock’s effect on these provisions is simply uncertain. We expect that lawyers who want to wield Bostock will search for words surrounding the provisions—be they in the same section, related sections, or in provisions of different yet related Acts of Congress—to present as a basis for inferring a focus on the individual, not the group alone. In turn, Bostock’s effect depends not only on those surrounding words, but also how likely a court is inclined to declare the statutory text ambiguous nonetheless, and if so, all the other statutory interpretation arguments to which a court may then resort, either on its own or at a lawyer’s urging. Still, if a court concludes, for whatever reason, that the provision focuses on the individual, then, thanks to Bostock, a court is more likely to read that sex discrimination provision to also cover gay and transgender individuals.
IV. Religion Defenses after Bostock
After Bostock, some employers are more likely to raise religion-based defenses to Title VII liability for discriminating against gay or transgender individuals. In dicta, Bostock pointed to three legal sources for such defenses: Title VII itself, the Religious Freedom Restoration Act, and the First Amendment. Let’s consider each in turn.
A. Title VII: Religion Organization Exemptions
Title VII affords several exemptions from liability, two of which cover employers who are religious organizations. Under section 702(a), Title VII does not apply to any “religious corporation, association, educational institution, or society with respect to the employment of individuals of a particular religion to perform work connected with” that organization’s activities. And under section 703(e), Title VII does not declare it illegal for an educational institution “to hire and employ employees of a particular religion” if it is at least substantially “owned, supported, controlled, or managed by a particular religion” or a particular religious organization, or if the educational institution’s “curriculum . . . is directed toward the propagation of a particular religion.” In other words, religious organizations, and the schools close enough to them, need not fear Title VII liability for discriminating against an individual because that individual is of a different religion.
After Bostock, if a gay or transgender individual brings an otherwise winning Title VII sex discrimination claim, how likely is an employer to escape Title VII liability with one of these exemptions? In his Bostock dissent, Justice Alito worried that these exemptions, as read by the lower courts, “provide only narrow protection.”
A lot initially depends on whether the defendant-employer qualifies as “a religious corporation, association, educational institution, or society” or as a school closely affiliated with one. To decide this, lower courts have pointed to, among other factors, whether the employer is a non-profit or for-profit entity. If an employer does qualify as a religious organization, then the exemptions are expansive, because Title VII defines “religion” to include “all aspects of religious observance and practice, as well as belief,” and because section 702(a) applies even if the employee performed only secular activities. For example, courts have held that Title VII exempts a religious organization that fires an employee for becoming pregnant after extramarital sex, provided that organization prove that it fired her because extramarital sex is inconsistent with its “particular religion.” Thus, religious employers are likely to invoke these exemptions to defeat Title VII sex discrimination liability, arguing that the employer’s “particular religion” requires conforming gender expression to the sex assigned at birth or limiting sexual intimacy to the opposite sex, and thus discriminating against a gay and transgender individual because of religion.
B. Religious Freedom Restoration Act
The Religious Freedom Restoration Act (RFRA) may provide another defense in some cases. Under RFRA, the federal government “shall not substantially burden a person’s exercise of religion” unless the government shows that applying “the burden to the person” furthers a “compelling governmental interest” and is the “least restrictive means” to further that interest. The term “person” in RFRA includes a closely-held for-profit corporation.
In Bostock, the Court noted in dicta that because RFRA “displac[es] the normal operation of other federal laws, it might supersede Title VII’s commands in appropriate cases.” After Bostock, suppose a gay or transgender individual brings an otherwise winning Title VII sex discrimination claim. In response, the defendant-employer, a corporation, raises a RFRA defense, arguing that Title VII “substantially burden[s]” its exercise of “religion” by imposing civil liability for acting consistent with a religious motivation not to employ anyone who is gay or transgender. How likely is that RFRA defense to prevail?
In federal court, such a RFRA defense is unavailable where the federal appellate court has concluded that RFRA does not apply unless the government is a party to the litigation. RFRA, by its terms, only applies where “[g]overnment” substantially burdens religious exercise. RFRA makes the “government” bear the burdens of “going forward with the evidence and of persuasion” in showing that the challenged burden on religious exercise is the least restrictive means of furthering a compelling governmental interest, which the government cannot do if it is not a party to the lawsuit. Moreover, Congress enacted RFRA to restore a Free Exercise Clause doctrine that had only applied to burdens on religious exercise imposed by the government.
What then did Bostock mean by noting in dicta that RFRA “might supersede Title VII’s commands in appropriate cases” (citing 42 U.S.C. § 2000bb-3)? Perhaps Bostock was referring to Harris Funeral Homes, where the defendant had litigated a RFRA defense that it could raise because a federal government agency (the EEOC) had initiated the Title VII lawsuit against it. If so, the citation to 42 U.S.C. § 2000bb-3 simply reminds the reader that RFRA “applies to all Federal law, and the implementation of that law, whether statutory or otherwise,” including federal government enforcement of Title VII by the EEOC.
Alternatively, perhaps Justice Gorsuch cited to 42 U.S.C. § 2000bb-3 to encourage lawyers to argue in future cases that, because RFRA applies to the “implementation” of “all Federal law,” a RFRA defense is available even in a Title VII lawsuit with only private parties. This reading, however, raises many puzzles, including whether such a RFRA defense exists if that Title VII lawsuit is in state court and why RFRA defines “government” to include a “person” only if that person is “acting under color of law.”
Assuming a Title VII defendant can raise a RFRA defense, it must show that Title VII or its implementation “substantially burdens” the defendant’s conduct; that conduct is an “exercise of religion”; and the religious motivation for that conduct is sincerely held. How a RFRA defense to Title VII would fare on the merits is unclear. Past RFRA litigation tells us little. From July 2014 up through 2018, the federal district courts decided 115 RFRA claims on the merits, but only seven of those were employment cases (about six percent, with plaintiffs winning in four and losing in three). Still, RFRA’s definition of “religion” is broad, providing that religious “exercise” need not be “compelled by, or central to, a system of religious belief.” Accordingly, we expect the lower courts to accept most assertions that the conduct at issue is an exercise of “religion.” In contrast, taking the case law as a guide, we expect more disputes among litigants over RFRA’s “substantial burden” and whether imposing it is the “least restrictive means” to advance a “compelling” government interest.
To illustrate, consider the fate of the RFRA defense in Harris Funeral Homes. The defendant funeral home was a for-profit corporation that Thomas Rost owned and operated. In the Sixth Circuit, the funeral home argued that Title VII, as applied to prohibit it from firing Ms. Stephens, was a “substantial burden” on Rost’s religious exercise of running the funeral home to serve grieving people. The Sixth Circuit considered and rejected two alleged substantial burdens.
First, Rost did not suffer a “substantial burden” on the ground that letting Stephens wear a skirt-suit to work would distract grieving families and thereby obstruct Rost’s ability to serve them. This assumed that customers would perceive Stephens as a man in woman’s attire and be disturbed by a transgender funeral director. It was, however, at least a “material question of fact as to whether [Rost’s] clients would actually be distracted.” More importantly, as a matter of law, “a religious claimant cannot rely on customers’ presumed biases” to establish a RFRA substantial burden.
Second, Rost did not suffer a “substantial burden” on the ground that Rost had to either provide female attire to Stephens or let her wear female attire to work—which he believed to be religiously forbidden—or go out of business. Although Rost “currently provides his male employees with suits and his female employees with stipends to pay for clothing,” no law or religious motivation required Rost to provide that benefit, and the record did not show that benefit was “necessary to attract workers.” Moreover, the court accepted as sincere Rost’s belief that he would “’violate God’s commands’” by letting Stephens “represent herself as a woman,” “because it would make him ‘directly involved in supporting the idea that sex is a changeable social construct rather than an immutable God-given gift.’” Nonetheless, the court found no RFRA “substantial burden” as a result, because as a matter of law, “bare compliance with Title VII—without actually assisting or facilitating Stephens’s transition efforts—does not amount to an endorsement of Stephens’s views.”
Finally, the Sixth Circuit held that, in any case, the EEOC showed that any such “substantial burden” furthers a “compelling governmental interest” and is the “least restrictive means” to further that interest. If the EEOC could not enforce Title VII against the funeral home for firing Stephens, it could not advance its compelling interest of combating workplace discrimination. And Title VII liability was the least restrictive means to enforce that compelling interest here. For example, neither a gender-neutral dress code, nor an “equally-burdensome” sex-specific dress code, sufficed as lesser restrictive alternatives, because Rost’s sex-stereotyping applied not just to what Stephens wanted to wear, but Stephens’s appearance and behavior more generally.
C. The First Amendment and the Ministerial Exception
The Religion Clauses of the First Amendment bar applying employment discrimination statutes “to claims concerning the employment relationship between a religious institution and its ministers.” After Bostock, suppose a gay or transgender individual brings an otherwise winning Title VII sex discrimination claim, and the defendant-employer raises this “ministerial exception” defense to defeat Title VII liability. How likely is that defense to prevail?
In short, because the Court has adopted a case-by-case approach to the issue of who counts as a “minister,” a lot depends on how easily lawyers and judges can analogize to the case characteristics of prior rulings on the ministerial exception defense. Relevant factors include whether the entity and the potential minister considered the person a minister, whether that person had a distinct role within that entity related to its religious mission, how much religious training the role required, and whether the person’s job duties included conveying the entity’s religious message or carrying out its religious mission. The title “minister” and its equivalents, and the associated formal religious training, are not dispositive. For example, in Our Lady of Guadalupe School v. Morrissey-Berru, the U.S. Supreme Court wrote that the ministerial exception applied to “employment disputes involving teachers at religious schools who are entrusted with the responsibility of instructing their students in the faith,” thus apparently increasing that defense’s scope.
V. The Trump Judges
Bostock’s effect depends not only how lower court judges read Bostock and the text of sex discrimination statutes, but also on those judges’ ideological and personal preferences about gender, sexuality, and religion. As best as can be measured, judge ideology matters to case outcomes, though it is often hard to disentangle how much it matters relative to other motivations or influences.
In recent years, federal judges have been more openly appointed based on their apparent ideological preferences, on the premise that those preferences will substantially affect how those judges will rule. As of September 2020, President Trump had appointed over 200 judges to serve on the main Article III federal courts (the U.S. Supreme Court, the thirteen U.S. Courts of Appeal, and the federal district courts), or about a quarter of the active federal judges on those courts. Most Trump appointees to the federal appellate courts had ties to the Federalist Society and were chosen as part of a process that weighted heavily their conservative bona fides.
If those lower court judges’ ideological preferences include disapproval of individuals who depart from heterosexual or cisgender norms, how much will that affect what those judges do with Bostock?
To illustrate, consider Stuart Kyle Duncan, appointed in 2018 by President Trump to the federal court of appeals for the Fifth Circuit. In United States v. Varner, a pre-Bostock case, Judge Duncan, writing a majority opinion (for himself and Judge Jerry E. Smith, a Reagan appointee), ruled that a district court could not consider a transgender woman prisoner’s request to change the name on that prisoner’s judgment of confinement from “Norman Varner” to “Kathrine Nicole Jett.” In the appeal, the prisoner-appellant, proceeding pro se, had filed a two-sentence motion (titled “Motion to Use Female Pronouns When Addressing Appellant”): “I am a woman and not referring to me as such leads me to feel that I am being discriminated against based on my gender identity. I am a woman—can I not be referred to as one?” Denying her request, Judge Duncan wrote that the law did not require anyone to refer to “gender-dysphoric litigants with pronouns matching their subjective gender identity”; if a court were to so require, it “may unintentionally convey its tacit approval of the litigant’s underlying legal position”; and it would be “quixotic” for federal judges to order use of “a litigant’s preferred pronouns,” given the complexity of “such neologisms” in other possible cases.
Suppose we infer from Varner’s content and tone that Judge Duncan tends to prefer cisgender over transgender individuals, all else equal, for whatever reason. If so, we might expect that, as a result, Judge Duncan, either deliberately or unwittingly, is more likely to distinguish Bostock away in cases where transgender individuals bring claims of sex discrimination under federal law or more likely to accept religion-based defenses to those cases, all else equal. And the more other Trump appointees share this tendency, the more likely that they too will rule, vote, and write opinions accordingly.
Yet, this effect on Bostock will also likely vary with how Trump appointees comprise particular federal appellate courts. For example, President Trump has appointed six of the twelve active judges on Eleventh Circuit, six of the seventeen active judges on the Fifth Circuit, three of the fifteen active judges on the Fourth Circuit, but none of the active judges on the First Circuit. In any particular appeal, the odds of a Trump-appointee majority on a three-judge panel vary accordingly, and with that, what that panel will do with Bostock.
In this essay, we speculated about Bostock’s effect by pointing to the text of other sex-discrimination bans in the U.S. Code, the contours of possible religion-based defenses, and Trump appointees to the federal judiciary. Despite Justice Ginsburg’s recent death, five of the six justices in the Bostock majority remain on the Court. As a result, Bostock will likely persist as precedent, even if Justice Ginsburg’s successor prefers Justice Alito’s Bostock dissent (or disfavors Bostock on other grounds) and does not feel bound to Bostock by stare decisis. Nor do we suspect that such a successor, if appointed, would affect how hard lawyers work to distinguish Bostock away based on textual differences or work to make it easier for religion-based defenses to prevail. With Justice Ginsburg still alive, those lawyers would likely have made such arguments anyway. But, if her successor would readily accept such arguments, those lawyers are more likely to succeed, if only because they would have to convince Justice Gorsuch or Justice Roberts, not both, when the issue ultimately comes before the Court.
* Professor of Law, University of Connecticut.
** Professor of Law, Saint Louis University.
 Bostock v. Clayton Cty. Bd. of Comm’rs, 140 S. Ct. 1731 (2020).
 Bostock v. Clayton Cty. Bd. of Comm’rs, 723 F. App’x 964 (11th Cir. 2018); Zarda v. Altitude Express, Inc., 883 F.3d 100 (2d Cir. 2018); E.E.O.C. v. R.G. &. G.R. Harris Funeral Homes, Inc., 884 F.3d 560 (6th Cir. 2018).
 Bostock, 140 S. Ct. 1731 (2020)
 Id. at 1754 (Alito, J., with whom Thomas, J., joins, dissenting); id. at 1822 (Kavanaugh, J., dissenting).
 Id. at 1738.
 42 U.S.C. § 2000e-2(a)(1); Civil Rights Act of 1964, Pub. L. No. 88–352, § 703(a)(1), 78 Stat. 241, 255.
 Bostock, 140 S. Ct. at 1739 (so assuming “because nothing in our approach to these cases turns on the outcome of the parties’ debate [on this issue], and because the employees concede the point for argument’s sake”).
 Id. at 1740.
 Id. at 1741.
 Id. at 1741–42.
 Id. at 1742. On this point, Bostock argued that its reasoning was consistent with Title VII precedent. Id. at 1743–44 (discussing Phillips v. Martin Marietta Corp., 400 U.S. 542 (1971); Los Angeles Dep’t of Water & Power v. Manhart, 435 U.S. 702 (1978); and Oncale v. Sundowner Offshore Servs., Inc. 523 U.S. 75 (1998)).
 Bostock, 140 S. Ct. at 1749–52.
 Id. at 1778 (Alito, J., dissenting).
 Compare Bostock, 140 S. Ct. at 1749 (Title VII’s legislative history, though relevant for reading “ambiguous statutory language,” has “no bearing” here, because “no ambiguity exists about how Title VII’s terms apply to the facts before us”) with id. at 1763 (Alito, J., dissenting) (“The Court’s excuse for ignoring everything other than the bare statutory text is that the text is unambiguous and therefore no one can reasonably interpret the text in any way other than the Court does. . . . [T]o say that the Court’s interpretation is the only possible reading is indefensible.”).
 Brett M. Kavanaugh, Fixing Statutory Interpretation, 129 Harv. L. Rev. 2118, 2134–2144 (2016); Ward Farnsworth, Dustin F. Guzior & Anup Malani, Ambiguity about Ambiguity: An Empirical Inquiry into Legal Interpretation, 2 J. of Legal Analysis 257 (2010).
 42 U.S.C. § 2000e-2(a)(2).
 See id. § 2000e-2(b).
 Id. § 2000e-2(c)(1) (“any individual”); id. § 2000e-2(c)(2)(“any individual”, “such individual’s”); id. § 2000e-2(c)(3) (“an individual”).
 Id. § 2000e-2(d) (“any individual”).
 E.g., id. § 2000e-2(f) (exemption for actions “with respect to an individual who is a member of the Communist Party of the United States”).
 E.g., Gustafson v. Alloyd Co., 513 U.S. 561, 570 (1995).
 42 U.S.C. § 2000e-2(j) (not requiring employer to grant “preferential treatment to any individual or to any group”); id. § 2000e-2(n)(2)(B) (“members of a group”).
 E.g., Adams ex. rel. Kasper v. School Board, 968 F.3d 1286, 1305 (11th Cir. 2020) (applying Bostock to Title IX of the Civil Rights Act, 20 U.S.C. § 1681(a)).
 Bostock v. Clayton County, 140 S. Ct. 1731, 1778 (2020) (Alito, J., dissenting); see id. at 1791–1796 (Appendix C).
 Bostock, 140 S. Ct. at 1753 (“today’s holding—that employers are prohibited from firing employees on the basis of homosexuality or transgender status”) (emphasis added).
 Id. at 1739 (citations omitted).
 6 U.S.C. § 488f (emphasis added).
 For a provision with text cutting the other way, see 28 U.S.C § 994(d) (US Sentencing Commission “shall assure that the guidelines and policy statements are entirely neutral as to the . . . sex . . . of offenders”) (emphasis added).
 Bostock, 140 S. Ct. at 1740.
 Id. at 1740‒41.
 1 U.S.C. § 1.
 Rowland v. Cal. Men’s Colony, Unit II Men’s Advisory Council, 506 U.S. 194, 199 (1993).
 See supra note 25.
 1 U.S.C. § 1.
 E.g., 7 U.S.C. § 2702(b); 12 U.S.C. § 3752(7); 15 U.S.C. § 8702(15); 22 U.S.C. § 2797c(a)(8)(A); 22 U.S.C. § 8531(4)(a); 42 U.S.C. § 2014(s); 52 U.S.C. § 30101(11).
 While whomever can disfavor someone because of the sex assigned to that human being, in ordinary English-language writing, no one typically assigns companies and corporations a sex. Apple, Inc. and ExxonMobil are neither male nor female. Still, Congress sometimes writes sex discrimination provisions also to protect a non-corporeal entity, as well as human beings related to a non-corporeal entity, in some way. E.g., 50 U.S.C. § 4842(a)(1)(B) (requiring regulations prohibiting any “United States person” who, intending to support a foreign country’s boycott against any country “friendly” to the U.S., discriminates “against any United States person on the basis of . . . sex . . . of that person or of any owner, officer, director, or employee of such person”); 15 U.S.C. § 633(b)(1) (Small Business Administration “shall not discriminate on the basis of sex . . . against any person or small business concern applying for or receiving assistance from the Small Business Administration”).
 12 U.S.C. § 1821(d)(13)(E)(iv).
 We searched the Westlaw database of the current U.S. Code Annotated (“TE(‘sex groups’)”) and the Hein Online database of all the volumes of the U.S. Statutes at Large (“sex groups”). Both searches yielded zero results.
 E.g., 42 U.S.C. § 247b-4(c)(2) (“racial and ethnic groups”).
 E.g., 36 U.S.C. §§ 20204(b), 20205(c) (Air Force Sergeants Association).
 Cf. 36 U.S.C. § 220522(a)(9) (an amateur sports organization is eligible to be recognized as the national governing body only if governing board “members are selected without regard to . . . sex, except that, in sports where there are separate male and female programs, it provides for reasonable representation of both males and females on the board of directors or other governing board”).
 Bostock v. Clayton County, 140 S. Ct. 1731, 1754 (2020).
 42 U.S.C. § 2000e-1(a).
 Id. § 2000e-2(e)(2).
 Section 703(e) of Title VII also exempts any employer, religious organization or not, that discriminates “on the basis of . . . religion . . . in those certain instances where religion . . . is a bona fide occupational qualification reasonably necessary to the normal operation of that particular business or enterprise.” Id. § 2000e-2(e)(1).
 Bostock, 140 S. Ct. at 1781 (footnote omitted).
 E.g., Spencer v. World Vision, Inc., 633 F.3d 723, 724 (9th Cir. 2011); LeBoon v. Lancaster Jewish Cmty. Ctr. Ass’n, 503 F.3d 217, 226 (3d Cir. 2007); see also EEOC Compliance Manual § 12-II(C)(1)(2020).
 42 U.S.C. § 2000e(j).
 Compare Civil Rights Act of 1964, Pub. L. No. 88–352, § 702, 78 Stat. 241, 255 (“of its religious activities”) with Equal Employment Opportunity Act of 1972, Pub. L. No. 92–261, § 3, 86 Stat. 103, 103–04 (amending section 702: “of its activities”).
 See generally Darian B. Taylor, Validity, Construction, and Application of Civil Rights Act of 1964 (42 U.S.C.A. §§ 2000e-1(a), 2000e-2(e)(2)) Exempting Activities of Religious Organizations from Operation of Title VII Equal Employment Opportunity Provisions, 6 A.L.R. Fed. 3d Art. 6, §§ 19–20 (2015 & Supp. 2020) (compiling cases).
 42 U.S.C. § 2000bb-1(a), (b).
 Burwell v. Hobby Lobby Stores, 573 U.S. 682, 719 (2014).
 Bostock v. Clayton County, 140 S. Ct. 1731, 1754 (citing 42 U.S.C. § 2000bb-3). The Bostock and Zarda defendants had not raised RFRA, and the defendant in Harris Funeral Homes had not sought review of the Sixth Circuit’s ruling on its RFRA defense. See id.
 Listecki v. Official Comm. of Unsecured Creditors, 780 F.3d 731, 736–37 (7th Cir. 2015); General Conference Corp. of Seventh-Day Adventists v. McGill, 617 F.3d 402, 410 (6th Cir. 2010). But see Hankins v. Lyght, 441 F.3d 96, 103 (2d Cir. 2006) (RFRA defense available in ADEA lawsuit brought by private plaintiff, because ADEA was also “enforceable” by EEOC); id. at 114–15 (Sotomayor, J., dissenting); and Rweyemamu v. Cote, 520 F.3d 198, 204 n.2 (2d Cir. 2008) (dicta disfavoring RFRA analysis in Hankins).
 42 U.S.C. § 2000bb-1(a).
 Id. §§ 2000bb-1(b), 2000bb-2(3).
 Hankins, 441 F.3d at 114–15 (Sotomayor, J., dissenting).
 See 42 U.S.C. § 2000bb(b) (RFRA’s purposes: “(1) to restore the compelling interest test as set forth in Sherbert v. Verner, 374 U.S. 398 (1963) and Wisconsin v. Yoder, 406 U.S. 205 (1972), and to guarantee its application in all cases where free exercise of religion is substantially burdened; and (2) to provide a claim or defense to persons whose religious exercise is substantially burdened by government”) (emphasis added).
 E.E.O.C. v. R.G. &. G.R. Harris Funeral Homes, Inc., 884 F.3d 560, 584 (6th Cir. 2018) (“[I]f Stephens had initiated a private lawsuit against the Funeral Home to vindicate her rights under Title VII, the Funeral Home would be unable to invoke RFRA as a defense because the government would not have been party to the suit.”).
 42 USC § 2000bb-3(a); see Little Sisters of the Poor Saints Peter & Paul Home v. Pennsylvania, 140 S. Ct. 2367, 2383 (2020) (42 U.S.C. § 2000bb-3(a) covers “regulations implementing the [Affordable Care Act] contraceptive mandate”).
 Hankins, 441 F.3d at 115 (Sotomayor, J., dissenting) (“Read in conjunction with the rest of the statute, [42 U.S.C. § 2000bb-3] simply requires courts to apply RFRA ‘to all Federal law’ in any lawsuit to which the government is a party.”).
 Shruti Chaganti, Why the Religious Freedom Restoration Act Provides a Defense in Suits by Private Plaintiffs, 99 Va. L. Rev. 343, 357 (2013) (reasoning that government action includes the imposition of legal rules “to be enforced by private plaintiffs,” citing, for example, New York Times v. Sullivan, 376 U.S. 254 (1964); that RFRA applies to the “implementation” of federal law; and therefore that “private plaintiffs suing over defendants’ exercises of religion are enforcing, or ‘implement[ing],’ a government-imposed burden on religion”) (footnote omitted).
 See City of Boerne v. Flores, 521 U.S. 507 (1997) (declaring RFRA unconstitutional as applied to States); Religious Land Use and Institutionalized Persons Act of 2000, Pub. L. No. 106–274, § 7(a)(1), (b) 114 Stat. 803, 806 (amending RFRA by striking “a State, or a subdivision of a State” in definition of “government” and striking “and State” in 42 U.S.C. § 2000bb-3(a)).
 See 42 U.S.C. § 2000bb-2(1),(2) (“any branch, department, agency, instrumentality, and official (or other person acting under color of law) of the United States” or the District of Columbia, Puerto Rico, and the federal territories and possessions).
 Meredith Abrams, Empirical Analysis of Religious Freedom Restoration Act Cases in the Federal District Courts Since Hobby Lobby, 4 Colum. Hum. Rts. L. Rev. Online 55, 71–72 tbl. 1–2 (2019).
 See 42 U.S.C. § 2000bb-2(4).
 E.E.O.C. v. R.G. &. G.R. Harris Funeral Homes, Inc., 884 F.3d 560, 585–597 (6th Cir. 2018).
 Id. at 585. No party disputed that Rost’s religious motivation was sincere. Id.
 Id. at 586–87. Cf. 29 CFR § 1604.2(a)(1)(iii), (2) (EEOC guideline that, unless necessary for “authenticity or genuineness . . . e.g., an actor or actress,” Title VII’s “bona fide occupational qualification” exception for “sex” discrimination does not apply to “refusal to hire an individual because of the preferences of coworkers, the employer, clients or customers”).
 Harris Funeral Homes, 884 F.3d at 587–88.
 Id. at 590 (citations omitted).
 Id. at 591–93.
 Id. at 593–94.
 Hosanna-Tabor Evangelical Lutheran Church and School v. E.E.O.C., 565 U. S. 171, 188 (2012).
 Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049, 2055 (2020).
 See Allison P. Harris & Maya Sen, Bias and Judging, 22 Ann. Rev. of Pol. Sci. 241 (2019); Lee Epstein & Jack Knight, Reconsidering Judicial Preferences, 16 Ann. Rev. of Pol. Sci. 11 (2013).
 Biographical Directory of Article III Federal Judges, 1789-Present, Fed. Jud. Ctr., (last visited Sept. 26, 2020).
 Rebecca R. Ruiz et al., A Conservative Agenda Unleashed on the Federal Courts, N.Y. Times (March 14, 2020).
 E.g., Letter of Am. Bar Ass’n to U.S. Senate Judiciary Comm. re: Nomination of Lawrence J.C. VanDyke to the U.S. Court of Appeals for the Ninth Circuit (Oct. 29, 2019) (“Some interviewees raised concerns about whether Mr. VanDyke would be fair to persons who are gay, lesbian, or otherwise part of the LGBTQ community. Mr. VanDyke would not say affirmatively that he would be fair to any litigant before him, notably members of the LGBTQ community.”) The Senate confirmed VanDyke’s appointment to the Ninth Circuit on December 11, 2019.
 United States v. Varner, 948 F.3d 250 (5th Cir. 2020).
 Id. at 259 (Dennis, J., dissenting).
 Id. at 254–58 (footnote omitted). In dissent, Judge James L. Dennis, a Clinton appointee, stated that he would have granted the request, noting, as the majority opinion had, that “though no law compels granting or denying such a request, many courts and judges adhere to such requests out of respect for the litigant’s dignity.” Id. at 260 (citations omitted).
 Fed. Jud. Ctr., supra note 81.