Re-Upping Appeasement: Religious Freedom and Judicial Politics in the 2019 Term

Nelson Tebbe Jane M.G. Foster Professor of Law, Cornell Law School and Micah Schwartzman Joseph W. Dorn Research Professor of Law, University of Virginia School of Law

Nelson Tebbe* & Micah Schwartzman**

What strategy should the more liberal justices on the U.S. Supreme Court adopt, now that they are faced with a conservative majority intent on remaking the law of religious freedom? In a series of decisions over the last few years, the Roberts Court has been slowly but radically weakening the Establishment Clause while strengthening the Free Exercise Clause. That program has presented the Supreme Court’s more liberal members with the difficult question of how to respond. Of course, one possible answer is that they should adopt no strategy at all; they should instead simply vote according to their principled interpretations of constitutional law. To the degree that they behave strategically, however, they will have to choose among competing options. How might they adjust to the reality of a Roberts Court with a transformative vision—in sum, virtually nonexistent limitations on establishment combined with powerful protections for free exercise—and with the votes to realize that vision?

In recent work, we posed this question, and we warned against one possible strategy, which we described as Establishment Clause appeasement.[1] We defined appeasement as “a sustained strategy of offering unilateral concessions for the purpose of avoiding further conflict, but with the self-defeating effect of emboldening the other party to take more assertive actions.”[2] Although we could not be completely certain, we presented evidence that liberals had engaged in appeasement in a series of religious freedom cases that the Court had decided seven to two. Specifically, we noted that some of the liberal justices had joined conservative majorities or otherwise declined to dissent.

We then warned that this strategy carried serious risks, chiefly that it could speed rather than slow the Roberts Court’s revolution in the constitutional doctrine governing religious freedom. In short, declining to dissent could embolden conservative majorities by giving them incentives to pursue more aggressive positions and by bolstering the legitimacy of their rulings. For instance, Professor Michael McConnell reasonably has pointed to the Court’s lopsided votes to bolster the conclusion that its recent decisions do not reflect partisan sentiment but instead are faithfully following the law where it leads. “One indication,” he recently wrote, “is that most of these decisions broke 7‒2 or 6‒3, instead of along the predictable 5‒4 conservative/liberal split.”[3] That pattern not only legitimizes the majority, but it also isolates the remaining dissenters, who can be portrayed as outliers. In this way, appeasement shifts the range of acceptable constitutional arguments to the right.[4]

At first glance, the 2019 Term presented mixed evidence for our descriptive claim that some justices are engaged in appeasement. On one hand, the Court extended the pattern of 7‒2 decisions in favor of religious actors in two important cases, Our Lady of Guadalupe School v. Morrisey-Berru, exempting religious schools from civil rights laws protecting certain teachers,[5] and Little Sisters of the Poor v. Pennsylvania, upholding religious and moral exemptions from contraceptive coverage requirements under the Affordable Care Act.[6] In both cases, two liberal justices joined a conservative majority, leaving just two others to dissent. On the other hand, the Court broke the pattern in another landmark decision, Espinoza v. Montana Department of Revenue, which required school choice programs to include religious schools.[7] There, the liberal justices dissented together. On this account, the 2019 Term seemed to weaken our descriptive claim that appeasement was afoot.

Moreover, our normative warning that appeasement is counterproductive also appeared to have been undermined. In cases outside the area of religious freedom, Chief Justice John Roberts joined the liberal justices to deliver important decisions on religion-inflected issues, namely LGBTQ rights[8] and reproductive freedom for women.[9] So perhaps a strategy of compromising with the Chief on religion cases, or co-opting him there, succeeded in winning his votes in other areas of law.

But on closer inspection, the story of the term may do more to support than to weaken our appeasement thesis. Consider Espinoza more closely. In 2017, some liberal justices joined a conservative majority in a religious funding case called Trinity Lutheran, perhaps hoping to head off an adverse ruling on school choice programs.[10] If so, that gambit failed in Espinoza, decisively and devastatingly. The Court issued an opinion that not only allowed government funding of religious instruction but actually required that funding in any program that benefits private schools. Considering only cases concerning the Religion Clauses, then, our descriptive claim about appeasement may well hold.

In what follows, we first reprise our appeasement argument. Then, in Part II, we present the case that Espinoza actually strengthens that analysis, and we briefly examine the two cases that more obviously support it, though we acknowledge the limitations of the evidence and the possibility of alternative accounts.[11] In Part III, we consider the term’s decisions outside the context of religious freedom, and we respond to the objection that efforts to coopt Chief Justice Roberts have proven successful, if not in cases immediately involving religion, then in the related areas of LGBTQ rights and reproductive freedom.

I. Appeasement and Its Alternatives

Appeasement is a strategy in which one party offers a series of unilateral concessions to a powerful counterparty, for the purpose of mollification but with the effect of further emboldening aggression. Although we recognize the charge of appeasement gains force from its association with the infamous Munich agreement that precipitated World War II, we disclaim any moral comparison to that episode or any other moment in history. We would prefer some other term that carries no such connotation. But no available substitute conveys the same combination of strategic purpose and counterproductive effect. We try to guard against the unwanted association by offering a careful definition of the concept and by distinguishing other decision-making strategies.[12]

While some elements of our definition are commonplace, others are less familiar. One conventional aspect is the normative assessment that appeasement leads to results that are counterproductive or self-defeating. A less common feature of our definition is that, for us, appeasement also depends on an actor’s intent or motivation. Appeasement cannot be undertaken entirely by mistake; instead, it requires a deliberate course of conduct. Of course, no one intentionally pursues a futile plan, but the part that must be purposeful is the aim to mollify an aggressor, not the failure of that effort.

Two other strategies are close cousins of appeasement. Compromise is agreement to a result that both parties consider to be nonideal but that they accept because of the fact of disagreement. Both parties recognize that something is better than nothing, and they decide to accept partial progress. Compromise can resemble appeasement insofar as it is second-best and because it often serves the purpose of peacemaking. Yet there is a crucial difference: compromise involves mutual concession, whereas appeasement is unilateral (or at least highly disproportionate). This difference may account for compromise’s positive connotation: the willingness to compromise is usually seen as a virtue.

Cooptation, the other related concept, describes an effort by one party to win over another by building longer-term relationships of trust and allegiance. The idea is that the coopted party will be persuaded to join a coalition in particular situations, and perhaps even permanently. Like appeasement, cooptation is sustained over a period of time, and it can involve asymmetric concessions at any particular moment. Cooptation can also devolve into appeasement if its practitioners are not sufficiently attentive to its prospects for success—if they stop expecting something in return for their concessions and if instead they attempt only to contain damage that might be caused by the party they are trying to coopt.

In previous work, we presented evidence that certain liberal justices were engaged in appeasement. In a series of religious freedom cases, Justices Stephen Breyer and Elena Kagan have joined conservative majorities, leaving only Justices Ruth Bader Ginsburg and Sonya Sotomayor to dissent. We described this trend in three major areas of Religion Clause doctrine: public funding, government religious symbols, and religious exemptions.[13]

Especially relevant here is the Court’s 2017 decision in Trinity Lutheran.[14] In an opinion by Chief Justice Roberts, the Court ruled that a church could not be excluded from a state program to resurface school playgrounds.[15] Missouri had a provision in its state constitution that strictly prohibited public money from flowing to religious organizations, and consequently it barred religious schools from the playground resurfacing program.[16] The Court applied strict scrutiny to that exclusion and found that the state’s desire for a particularly strict separation of church and state was not a compelling interest.[17]

Seven justices supported that result, with only Justices Ginsburg and Sotomayor dissenting.[18] Of particular interest for us, Justice Kagan joined the majority opinion in full.[19] Justice Breyer concurred in the judgment.[20] Chief Justice Roberts’s opinion therefore commanded six votes, except as to footnote 3, in which he wrote:

This case involves express discrimination based on religious identity with respect to playground resurfacing. We do not address religious uses of funding or other forms of discrimination.[21]

That crucial footnote was impossible to miss because two justices specifically refused to join it, meaning it was supported by only four votes and did not represent the opinion of the Court.[22] Justice Neil Gorsuch, joined by Justice Clarence Thomas, wrote separately to object that the status/use distinction could not be maintained.[23] That left only Justice Kagan, along with Justices Kennedy and Alito, in support of the Chief’s decision to reserve questions about funding “religious uses.”[24]

Given the lineup of votes, it is quite possible that Justice Kagan insisted on footnote 3 as a condition of joining the majority in Trinity Lutheran. She may have sought to include the footnote in order to forestall a future ruling that would require states to include religious schools in voucher programs or other school choice initiatives. Justice Kagan might have hoped that states would be permitted to exclude religious schools from such programs not on the basis of the schools’ religious status, but because the schools would use public funds to support religious uses. And, in fact, that was how commentators understood footnote 3—as an attempt to preserve voucher programs that did not include religious schools.[25]

The sense that Justice Kagan was engaged in strategic action is supported by her decision in an earlier case concerning taxpayer standing.[26] There, Justice Kagan dissented from the Court’s decision to reject taxpayer standing in an Establishment Clause challenge to a school choice program.[27] Beyond the standing issue, she seemed to articulate a vision of nonestablishment as a structural protection against government support for religion using public funds. Assuming her views had not changed, that position would have given her reason to oppose Trinity Lutheran, which approved direct tax funding of a church, albeit not for religious uses. And that, in turn, suggests that she may have acted strategically in Trinity Lutheran, conditioning her vote on the inclusion of footnote 3 in an attempt to avoid implications for school choice programs.

Justice Breyer, for his part, began his opinion with the proposition that the Establishment Clause must permit the government to include religious schools in basic social services like police and fire protection, and he reasoned that the playground resurfacing program was analogous insofar as its aid was universal and neutral in content.[28] That may well be his principled view. But he may instead have believed that a refusal to join the majority, combined with a narrow rationale, would slow the Court’s drive to require that school choice programs include religious schools. And we know that Justice Breyer has deep constitutional concerns about vouchers because he dissented passionately from Zelman v. Simmons-Harris, the decision that permitted voucher programs even when they overwhelmingly support parochial schools that engage in core religious practices.[29] Given that dissent, Justice Breyer might have concurred in Trinity Lutheran in an effort to disarm the majority by persuading its members that funding of religious organizations ought to be confined to programs involving basic services that are analogous to police and fire protection.

Appeasement strategies like these—if that is what they were—carry several interrelated risks, as we have explained.[30] First, they can have detrimental effects on outcomes. Not only are such strategies futile or ineffective, but they actually can be counterproductive, because they incentivize the other party to pursue its program more aggressively.[31]

In the Court’s religion cases, we cannot be certain whether the pattern of agreement by two liberal justices has affected outcomes. After all, the Roberts Court already has five votes for, broadly speaking, the combination of a weakened Establishment Clause and a strengthened approach to religious exemptions. But it is quite possible that the pattern that McConnell has noticed—where religious interests have prevailed in nearly all recent Supreme Court decisions—has been encouraged and supported by the other pattern he observes, namely that many of these cases have been decided seven to two.[32]

And beyond worsening outcomes, appeasement may also enhance the Court’s legitimacy. A decision that comes down seven to two carries greater authority than one that is decided by a bare majority of five. And in fact, supporters of Trinity Lutheran have not hesitated to highlight its lopsided vote.[33] Recall Professor McConnell’s claim that the Roberts Court is not partisan, which he supported by noting that its decisions are drawing more than five votes. He pointed to the pattern we are tracing as evidence of legitimacy when he wrote that “[i]n the last [twelve] cases involving religion, the religious side prevailed, sometimes by lopsided majorities.”[34] Even more powerfully, perhaps, McConnell acknowledged that “[t]he court may be political,” but he insisted that “its politics is of the middle.”[35] If some of the liberal justices are engaged in appeasement, however, then this description is inaccurate. The Court is politically polarized,[36] like the rest of the country, and its liberals are struggling to adjust to that reality.

Conversely, lopsided voting isolates any remaining dissenters, who are then vulnerable to being depicted as outliers or even radicals. For example, Justice Sotomayor wrote a powerful, ringing dissent in Trinity Lutheran that lost force when it was joined by only one of her colleagues, Justice Ginsburg.[37]

Relatedly, appeasement can affect the so-called “Overton Window” of accepted constitutional positions.[38] Liberal decisions to join conservative majorities, even for strategic reasons, can facilitate a rightward shift in the set of politically or legally feasible options, while a decision to dissent forcefully as a group can work to anchor that set or at least signal that conservatives are attempting to move the window. On a court that is as polarized as the current one, much of the struggle among the justices is over the range of what are considered plausible constitutional positions. Today’s outcomes are dictated by the five conservatives, but tomorrow’s outcomes will be affected by the scope of constitutional arguments that carry weight.

Because we build intentions into the definition of appeasement, we cannot know for sure whether Justices Breyer or Kagan followed this strategy in Trinity Lutheran. It is possible that they were simply voting their consciences and acting solely on the basis of constitutional principle and precedent, according to their own interpretations. Yet we have offered some evidence that they were engaged in strategic behavior, based on what we know about their principled positions on matters of free exercise and nonestablishment. Justice Kagan dissented in Winn, the taxpayer standing case, and Justice Breyer dissented in Zelman, the school voucher case. Along the way, they gave us indications of their views on nonestablishment in the area of government funding. Keeping those views in mind, it becomes harder to believe that they would have ruled against the state if they had commanded the necessary votes. To test the intuition, imagine that there had been a majority in favor of the state in Trinity Lutheran. Would these two have dissented?

Even if Justices Breyer and Kagan were acting strategically in Trinity Lutheran, they might have been pursuing a strategy other than appeasement. Perhaps Justice Kagan believed she was compromising by offering the majority something it valued—her vote—in exchange for something she valued—footnote 3 and the chance to cabin the rule of the case. A narrower account is that she compromised not by voting differently from how she would have otherwise, but by accepting reasoning that she opposed in exchange for a chance to press for a less sweeping majority opinion—one that included footnote 3. And perhaps Justice Breyer similarly believed that he was giving up some of what he wanted, the chance to dissent, in order to gain something else, namely influence over how the holding of Trinity Lutheran might be understood in the future. McConnell encouraged this view when he suggested that “the court seems to reach results that very likely would carry the day in Congress on many of these issues, if Republicans and Democrats were inclined to talk to one another and compromise.”[39]

The compromise he seemed to have in mind was something like the Fairness for All bill, a piece of federal legislation introduced by Republicans that would have extended civil rights protection to LGBT citizens alongside exemptions for religious objectors.[40] But this so-called compromise garnered no real support among either Republicans or Democrats.[41] Though compromise is often admirable, it is difficult to reach in a polarized political climate.

If Justice Kagan was practicing compromise, she might have been disappointed when footnote 3 failed to draw majority support. Yet she might nevertheless have hoped that the distinction between status and use could delay or derail any attempt to extend the rule of Trinity Lutheran to school choice programs.[42]

Another possibility is that these justices were instead pursuing a strategy of cooptation, supporting the result in order to build up personal and political capital that they could spend in future cases, particularly by drawing Chief Justice Roberts away from the conservatives. After all, the Chief wrote the majority opinion in Trinity Lutheran. It is not outrageous to think that Justices Breyer and Kagan saw that decision as a chance to convince him of their reasonableness without giving up anything of comparable importance, given that he already had a majority without their votes.

In our earlier work, we took the possibility of cooptation seriously, but we questioned whether it was really happening, considering Chief Justice Roberts’s demonstrated conservatism, not least on issues of religious freedom. His constitutional politics are very different from those of justices who may have been successfully influenced by liberals in the past, such as Justices Sandra Day O’Connor, Anthony Kennedy, or David Souter. Chief Justice Roberts is the median justice, but he does not sit at the center of the Court’s political distribution, which is asymmetrically skewed to the right. In other words, he does not sit halfway between the two extremes, even though there are four justices to his right and four to his left. And the fact that someone as conservative as Chief Justice Roberts, as compared to Justice Kennedy, is now the “swing” vote signals the overall shift of the Court rightward. The prospects of cooptation are also diminished by current levels of political polarization, which are unlike anything seen during the tenures of any of those three justices (save perhaps Justice Kennedy during his final years on the Court).

Cooptation is a realistic strategy when the distribution of political positions on the Court looks like a bell curve, with most of the justices crowded near the middle and one or two on the tails. Then it is conceivable that someone on the center-left could slowly convince someone on the center-right. But when the distribution of political positions looks like a barbell—as it does on the current Roberts Court, with strong conservatives facing off against moderate liberals—then the cooptation strategy seems more ambitious, far riskier, and thus less plausible as an explanation for the decisions of the more centrist liberal justices.

The Court’s 2019 Term offers additional evidence that can help us assess these claims, albeit not definitively. Looking at what the justices have done since 2017 can help us determine whether the liberals have used strategic behavior and, if so, whether that behavior yielded any gains. That evidence seems particularly pertinent to the cooptation story not only because Chief Justice Roberts joined the liberals in several salient cases but also because he wrote the majority opinion in Espinoza. Can the term’s decisions help us determine which strategy was in play—appeasement, compromise, cooptation, or none at all—and whether any such approach was successful?

II. Evidence from the 2019 Term

A. State Funding of Religion

You might think that subsequent events should be irrelevant to an assessment of appeasement. What matters is whether the actor adopted a program of unilateral concession with the hope of mollifying a powerful counterparty. And that can be determined at the time of decision. Appeasement therefore should be identifiable ex ante, even if failure is part of its definition.

Still, subsequent events can give us insight into what happened by providing additional evidence of intentions, purposes, and motivations. One factor that we take into account when determining purpose, after all, is subsequent effect.[43] So a diagnosis may only become possible in retrospect, and that is true even if our main concern is how things looked in the first instance.[44]

Keeping that in mind, what does Espinoza tell us about whether appeasement occurred in Trinity Lutheran? The comparison is apt because Chief Justice Roberts relied heavily on his earlier opinion in Trinity Lutheran when he was finally faced with the school choice issue that the justices had been anticipating.

Espinoza involved a challenge to a program enacted by Montana’s legislature that allowed taxpayers to receive a dollar-for-dollar tax credit of up to $150 per year for contributions to a “student scholarship organization.”[45] The scholarship organization then awarded grants to students, which they could use to offset tuition at private schools. Students and their families chose the schools, which then received the grants directly from the scholarship organizations.[46] The program was designed to benefit all private schools, but it mostly aided religious schools, which outnumbered secular schools in the state.[47]

When it created the program, the legislature had directed that it be implemented in a manner consistent with Montana’s “no-aid” provision in its state constitution.[48] Like the state constitutional provision at issue in Trinity Lutheran, and like the similar provisions that exist in some forty other states, Montana’s clause required stricter separation between church and state than was required by the federal Establishment Clause. Specifically, it provided:

Aid prohibited to sectarian schools. . . . The legislature, counties, cities, towns, school districts, and public corporations shall not make any direct or indirect appropriation or payment from any public fund or monies, or any grant of lands or other property for any sectarian purpose or to aid any church, school, academy, seminary, college, university, or other literary or scientific institution, controlled in whole or in part by any church, sect, or denomination.[49]

Interpreting this provision, the Montana Department of Revenue promulgated Rule 1, which prohibited the use of these grants at religious schools.[50] After Karen Espinoza and other parents sued the department, the Montana Supreme Court issued a decision confirming the conflict between the state constitution and the school choice program.[51] But the state court also held that the department lacked authority to promulgate Rule 1 and it therefore invalidated the program altogether, as to both secular and religious schools.[52]

Chief Justice Roberts wrote an opinion reversing and remanding. He relied on Trinity Lutheran for the rule that “disqualifying otherwise eligible recipients from a public benefit ‘solely because of their religious character’ imposes ‘a penalty on the free exercise of religion that triggers the most exacting scrutiny.’”[53] He found that the Montana no-aid provision implicated that rule because it excluded religious schools from the program on the basis of their religious status. And Montana could not overcome the presumption of unconstitutionality because its interest in pursuing a stricter vision of church-state separation was not compelling. Here too, Chief Justice Roberts relied on Trinity Lutheran.[54] Finally, he rejected the argument that religious students and schools were not being discriminated against because Montana eliminated the program entirely. The Chief explained that the state court first invalidated the program based on a state constitutional provision that “expressly discriminates on the basis of religious status” and only then decided to eliminate the program as a way of effectuating its ruling.[55]

With respect to the distinction between status and use, the majority rejected the view that excluding religious schools was tantamount to defunding religious uses, such as teaching theology, praying, and holding worship services. Chief Justice Roberts responded that even if the purpose or effect of the exclusion was to defund religious use, it actually turned on the religious status of the schools.[56] Moreover, he made it clear that he was not embracing the view that “some lesser degree of strict scrutiny applies to discrimination against religious uses of government aid.”[57] He noted that Justices Gorsuch and Thomas had questioned the stability of the status/use distinction and, without disagreeing, he explained that there was no need to examine that argument because Montana had engaged in discrimination on the basis of status.

Justice Breyer dissented and made two points, only the first of which drew Justice Kagan’s vote. First, he argued that Montana should be understood to have excluded only religious uses and not to have discriminated on the basis of religious status. “There is no dispute,” he noted, “that religious schools seek generally to inspire religious faith and values in their students.”[58] Montana’s decision not to fund them was the only practicable way for it to prevent tax dollars from flowing to core religious uses like instruction, prayer, and worship. States should have latitude—“play in the joints” between the Religion Clauses—to avoid the “religiously inspired political conflict and division” that so often accompanies government support for religion.[59]

Second, and writing now only for himself, Justice Breyer reiterated his view that interpreting and applying the Religion Clauses is a complex endeavor that is ineluctably fact-sensitive and irreducible to legal rules.[60] In cases concerning the tension between free exercise and nonestablishment, in particular, “there is no test-related substitute for the exercise of legal judgment.”[61] Instead, judges must consider the values animating the clauses and seek to vindicate them in particular contexts.

Justice Kagan also signed the dissent from Justice Ginsburg, who argued that the Montana court’s decision to eliminate the program altogether could not have entailed any discrimination on the basis of religion.[62] Justice Ginsburg rejected as “imaginary” the majority’s view that the state court had first invalidated the program based on a religious classification and only then eliminated the program.[63] Instead, the Montana court simply struck a program that violated the state constitution.

What are we to make of these votes by Justices Breyer and Kagan? Justice Breyer may have thought that Trinity Lutheran was different because it involved a public benefit akin to police and fire protection, whereas Espinoza was about school aid, a traditional concern of the Establishment Clause. But from what he actually said, it seems that both he and Justice Kagan had pinned their hopes on Trinity Lutheran footnote 3 and its distinction between religious status and use. If that is the case, and if they thought that Justice Kagan’s agreement to join the majority was part of a symmetrical compromise that had yielded something of value in an attempt to forestall the foreseeable school choice decision, then they were seriously mistaken.

Espinoza came only three years after Trinity Lutheran. And when it came, the Court flattened the status/use distinction, preserving it only as a technicality. In so doing, the majority effectively invalidated the no-aid provision in Montana and strongly signaled that similar provisions in dozens of other states are unconstitutional.

Stepping back, then, it appears that Justice Kagan may have spent capital in Trinity Lutheran in exchange for a footnote that did little to avoid or even meaningfully slow an adverse decision on school choice programs. To the extent that Justice Breyer hoped that footnote 3 would avoid a ruling that school choice programs must include religious schools, as his later decision suggested, he too gained nothing by giving ground in the earlier case. Together, they helped to legitimate Trinity Lutheran in asymmetric fashion, without impeding the Roberts Court’s march toward not only allowing but requiring state funding of religious schools.

Again, we cannot be certain of any of this, given the importance of intentions and motivations to assessments of instrumental decisionmaking. But these justices’ opinions and votes in Espinoza do provide some evidence that they had been relying on the unstable distinction between religious status and use to head off such a result. Maybe they made their decisions in both cases on the basis of pure legal principle. And, technically, the Court has not yet held that a state cannot prohibit its school choice program from funding religious uses. So it could be argued that Justice Kagan’s gambit in Trinity Lutheran has at least delayed a holding along those lines. As long as governments define religious uses without reference to the religious identity of the school, they can still argue for excluding such uses from school choice programs.[64]

But the evidence also supports the alternative view that they were attempting to manage the Roberts Court majority—and that they failed. As Justice Breyer wrote, with the agreement of Justice Kagan, Montana’s exclusion was tantamount to an exclusion of religious uses. They seem to have recognized that the game was up. When Espinoza arrived, they could only object on narrow, unsatisfying grounds. The time for ringing dissents and declarations of principle had passed.

B. Religious Exemptions

If Espinoza challenged our thesis, the Court’s other two religion opinions appeared to support it. Both Our Lady of Guadalupe and Little Sisters were decided by the lineup of seven justices that had become familiar in religious freedom cases. Did these decisions bolster the claim of appeasement?

Little Sisters concerned regulations promulgated by the Trump administration that exempted employers who objected on religious or moral grounds to contraceptive coverage requirements adopted pursuant to the Affordable Care Act (ACA). The Department of Health and Human Services had required all employers that provided health insurance for their workers to include cost-free coverage for all approved forms of female contraception. Religious employers complained that providing such coverage violated their beliefs. Initially, for-profit corporations raised objections under the Religious Freedom Restoration Act (RFRA), and the Court granted them exemptions in Burwell v. Hobby Lobby Stores, Inc.[65] Nonprofit employers then complained that the regulatory accommodation given to them, which required that they certify their objection to providing coverage, still imposed a substantial burden by triggering contraception coverage and thereby making them complicit.[66] Before that dispute could be resolved, President Trump was elected. His administration promulgated new rules that created sweeping and categorical exemptions for employers, whether for-profit or non-profit, who objected on religious or moral grounds to covering contraception.[67] Not only that, but women affected by the exemptions were not to be compensated in any way.[68] According to agency estimates, between 70,500 and 126,400 women were expected to lose contraception coverage as a result of the religious and moral exemptions.[69] State governments challenged the new rules and won a nationwide injunction from the lower courts.[70]

Justice Thomas, writing for a five-justice majority, held that the federal agencies responsible for creating the religious and moral exemptions at issue had exercised authority given to them by Congress under the ACA and that they had satisfied the procedural requirements of the Administrative Procedure Act.[71] He also held that the agencies acted properly in considering the demands of RFRA, although the Court did not reach the merits of religious freedom claims raised under that statute.[72] Ultimately, the Court reversed the court of appeals and remanded for further consideration. Notably, the Court also dissolved the nationwide injunction that had prevented the religious and moral exemptions from going into effect.[73]

Justice Kagan, joined by Justice Breyer, concurred in the judgment. She disagreed with the majority that the ACA authorized the rule in express terms, but she also disagreed with the dissent that the statute obviously did not authorize the action. Faced with ambiguity, she concluded that the Court ought to have deferred to the agencies’ interpretation of their own authorizing statutes.[74]

Yet Justice Kagan also expressed doubts about whether the regulations should stand. Although the circuit court had struck the regulations because they were procedurally invalid, it had not considered whether they were substantively invalid insofar as they were “arbitrary” or “capricious” under the APA.[75] Justice Kagan therefore encouraged the lower courts to consider that question, and she suggested that the rules may well be irrational, both because the agencies exempted employers who did not object on religious grounds to the existing accommodation for nonprofits, and because the rules exempted publicly traded corporations that might not have rights under existing interpretations of RFRA.[76]

It is difficult to believe that Justices Breyer and Kagan accepted the legality of the new rules as a matter of ideal interpretation. After all, each joined Justice Ginsburg’s powerful dissent in Hobby Lobby, where she concluded that the business corporation did not have a right under RFRA to an exemption from the contraception mandate.[77] And that was in a case where the Court presumed—correctly, as it turned out—that women would receive alternate coverage when the Obama administration extended the nonprofit accommodation to companies like Hobby Lobby. But under the Trump administration’s rules, as Justice Kagan expressly acknowledged in her opinion, the situation was markedly worse—employees would lose contraception coverage altogether, causing harm to tens of thousands of women.[78] Moreover, in Hobby Lobby, Justice Kagan, joined by Justice Breyer, wrote separately to bracket the question of whether business corporations were covered by RFRA, meaning that their reasoning in that case depended solely on the substantive application of the statute and not on the simple proposition that for-profit entities were not covered by RFRA.[79] Given all of the above, it is difficult to understand their opinion in Little Sisters except as guided by strategic considerations.[80]

In Our Lady of Guadalupe, the Court significantly expanded the scope of the ministerial exception, which it had constitutionalized several years earlier in Hosanna-Tabor Evangelical Lutheran Church & School v. E.E.O.C.[81] The question in Our Lady of Guadalupe was whether two lay teachers working at Catholic elementary schools counted as “ministers” for the purpose of triggering the exception.[82] Both teachers had won their civil rights cases in the Ninth Circuit,[83] but the Supreme Court reversed. In an opinion by Justice Alito, the Court held that the teachers were “ministers” for purposes of the Constitution, even though they were not ordained (as women cannot be in the Catholic church) and even though they primarily taught secular subjects.[84]

Kristen Biel instructed fifth graders in all their academic subjects, including a religion curriculum that she taught from a workbook chosen by the school administration for thirty minutes per day, four days a week. She was present for prayers twice a day, but did not lead them; likewise, she took her students to a monthly mass for the entire school but she had no role in the worship service.[85] Agnes Morrissey-Berru, the other plaintiff, taught fifth or sixth grade, including all academic subjects. Like Biel, she taught religion from a prescribed textbook and escorted her students to worship services. She did lead her students in prayer, planned a monthly service, and produced an Easter performance by her students each year.[86] Yet neither teacher was trained as a clergy member, neither had a liturgical title, and neither held herself out as clergy.[87] Both were fired for reasons that they said were discriminatory—one on the basis of age, the other on the basis of disability.[88]

Justice Alito ruled that the schools had a constitutional right to terminate the teachers, even if they actually discriminated and even though the Catholic faith does not claim that discrimination on the basis of age or disability is theologically required. He reasoned that the ministerial exception protected the church’s ability to make employment decisions concerning “those holding certain important positions with churches and other religious institutions.”[89] Though Justice Alito was not completely clear on why the Constitution required this extraordinary latitude, his opinion emphasized the importance of “independence in matters of faith and doctrine and in closely linked matters of internal government.”[90]

In determining who counted as a minister for these purposes, Justice Alito said that “[w]hat matters, at bottom, is what an employee does.”[91] And because “[r]eligious education is vital to many faiths,”[92] performing that function is a critical factor in determining whether an employee is covered under the ministerial exception and therefore is deprived of civil rights protections. In the cases of Biel and Morrissey-Berru, there was “abundant record evidence that they both performed vital religious duties,” including instructing their students in the faith, praying with them, attending services with them, and preparing them for “other religious activities.”[93]

Justices Breyer and Kagan signed the majority opinion, once again leaving Justices Sotomayor and Ginsburg as the only two dissenters.[94] But neither Justice Breyer nor Justice Kagan wrote to explain their votes. Perhaps that was because they agreed fully with Justice Alito’s reasoning; after all, Justice Kagan had joined Justice Alito’s concurrence in the Court’s previous ministerial decision case.[95] So perhaps they were with him as a matter of legal principle.

But their decision to sign the majority opinion in Our Lady of Guadalupe has left commentators puzzled because the decision greatly expands a constitutional exception from important civil rights laws, and it does so even though religious actors are not supposed to be able to claim exemptions from general laws under the Court’s governing free exercise rule.[96] That Justices Breyer and Kagan joined the majority in Hosanna-Tabor has long been a mystery, and that they joined this decision raises even more questions.[97] For instance, Professor Noah Feldman observes that:

It’s remarkable that the two pragmatist liberal justices, Justices Stephen Breyer and Elena Kagan, joined that opinion. For Breyer and Kagan to take this step suggests that they may have been trying to show that they’re willing to cross traditional liberal lines to avoid a 5‒4 decision—hence protecting the court from the perception of deep ideological division.[98]

Feldman points out that Justice Alito’s “church autonomy doctrine could be extremely far-reaching,” and he credits Justice Sotomayor’s argument in dissent that religious employers will be incentivized to define their missions broadly in order to avoid employment laws. That, in turn, will “allow religious institutions to make [hiring] decisions on the basis of a prohibited animus that had nothing to do with their religious beliefs.”[99] Knowing all this, why would Justices Breyer and Kagan join Justice Alito’s opinion and give it added legitimacy?

One possibility, which Feldman implicitly recognizes, is that they were acting tactically. They knew that Justice Alito had the votes to rule in favor of the schools even without them. So perhaps they joined his decision so they could help to shape it. After all, Justice Alito’s decision was not as radical as it might have been. He recognized that his ruling “does not mean that religious institutions enjoy a general immunity from secular laws.”[100] Language like that might have been the product of an agreement. Moreover, the majority did not go as far as Justices Thomas and Gorsuch, who wrote separately to argue that courts must defer to religious organizations’ claims of who counts as a minister. As Professors Ira Lupu and Robert Tuttle rightly point out, the majority’s refusal to accept an even more expansive view was significant.[101] Perhaps this result was a consequence of negotiations with Justices Breyer and Kagan.

But if Justices Breyer and Kagan influenced the majority opinion in this way, it may not have been worth the tradeoff. After all, Justice Alito ended the sentence quoted above by saying that the exception “does protect [religious institutions’] autonomy with respect to internal management decisions that are essential to the institution’s central mission.”[102] There is little in Our Lady of Guadalupe to restrain the Roberts Court from going further, for example, by exempting Catholic universities from labor laws that protect the ability of faculty or graduate students to unionize, to use Feldman’s hypothetical.[103] So the strategy, if there was one, raised risks similar to those that materialized in Espinoza. The liberal justices may have given up their votes to stave off worse outcomes, only to suffer further and more significant losses in future cases.

III.  Appeasement or Compromise?

Perhaps the benefits of strategic action by center-left justices were gained not in religious freedom cases, but in other areas of law. If so, their decisions would not have been futile or counterproductive after all. Professor Feldman suggests that Justices Breyer and Kagan engaged in productive compromise, if only in a loose way, and that they won an important victory in the term’s decision concerning LGBTQ rights.[104] In Bostock, after all, Justice Gorsuch and Chief Justice Roberts joined the liberals to extend Title VII, the main federal employment discrimination law, to cover LGBTQ employees.[105] That was a landmark ruling with potentially far-reaching implications for civil rights and antidiscrimination law.

We would add that Chief Justice Roberts also joined the liberal justices in June Medical Services v. Russo, the decision striking down a Louisiana law that would have eliminated nearly all the abortion providers in that state.[106] Both Bostock and June Medical were “religion inflected,” since they pitted core civil rights protections against political and legal views that were generally, if not invariably, grounded in traditional religious beliefs. Perhaps, then, Justices Breyer and Kagan understood the “gravitational effects that the different cases decided in a given term have on one another,” as Professor Feldman puts it.[107] If so, they were engaged in forging a judicial compromise that yielded significant benefits for those with liberal commitments.

This account might well be correct—again, we cannot be certain without knowing more about the justices’ purposes or intentions. The most we can say is that there is some evidence to support it. But there also are alternative explanations for that evidence. The Chief and Justice Gorsuch, who joined the liberals in Bostock, must have known that American public opinion favored extending civil rights protections to LGBTQ people.[108] Sooner or later, federal lawmakers would enact legislation to make the necessary changes to antidiscrimination law. By anticipating and preempting that development, the Roberts Court bolstered its legitimacy, strengthening the case that it is an apolitical body that simply follows the law. That was precisely the message that Justice Gorsuch’s “textual” opinion was designed to convey.

It is unlikely that Justice Gorsuch was engaged in any kind of deal-making, however atmospheric. And it is plausible to interpret Chief Justice Roberts as acting to promote the Court’s institutional legitimacy by bringing its jurisprudence into line with the clear trajectory of American public opinion.[109] His was also the sixth vote, once Justice Gorsuch had made up his mind. On this alternate account, then, the Court ruled the way it did in Bostock for its own independent reasons, not because of how Justices Breyer or Kagan had voted in religious freedom cases. And that supports the view that the liberal justices were not seeing the benefits of compromise or cooptation.

Combined with its decisions in the religion cases, again, the Court seems to be adopting something like the Fairness for All bill that has stalled in Congress.[110] The gist of that proposed legislation is civil rights protections for LGBTQ citizens combined with exemptions for religious believers. Crucially, the bill does not protect its carefully calibrated provisions from RFRA claims. This is not a formula that any Democratic lawmakers have endorsed, and for good reason, because it grants religious exemptions from civil rights laws protecting LGBTQ citizens that are stronger than exemptions that currently exist in civil rights laws protecting any other vulnerable group.[111] Nor would Democrats embrace a bill that allows a balance between civil rights and religious freedom to be upended by RFRA. It is possible that Justices Breyer and Kagan have agreed to a compromise along these lines, but is seems more likely that the conservative majority of the Roberts Court is independently crafting a doctrine that reflects its own institutional interests and jurisprudential commitments.[112]

What about June Medical, where Chief Justice Roberts joined the four liberals, instead of cutting back on abortion rights?[113] The Chief concurred in the judgment, even though he had dissented in Whole Woman’s Health, a 2016 decision invalidating a nearly identical regulation in Texas.[114] He explained that although he continued to think that Whole Woman’s Health was wrongly decided, that decision was binding precedent and had to be followed.[115]

But, of course, the Supreme Court could have overturned its own precedent, and it appeared that Chief Justice Roberts had the votes to do so, after Justice Kavanaugh had replaced Justice Kennedy. Moreover, as a practical matter, precedents are not rigidly followed by the justices, to put it mildly.[116] It is true that the Chief in particular is more inclined to weaken than overrule cases he believes to be wrong.[117] But Whole Woman’s Health was a relatively recent decision that pertained to unusual facts. Even Chief Justice Roberts could have voted to overrule it without much trouble. Could something else explain his decision?

It is possible that the Chief was responding to overtures that Justices Breyer and Kagan had been making carefully in other cases, particularly but not only during the 2019 Term. Perhaps their votes on the ministerial exception and the contraception mandate succeeded in building goodwill that Chief Justice Roberts repaid in June Medical. That would be a story of successful cooptation, or compromise,[118] or “gravitational effects,”[119] and it has some plausibility.

Yet it is also plausible to think that Chief Justice Roberts had institutional reasons for voting as he did. The Fifth Circuit had flouted Whole Woman’s Health in its ruling in June Medical.[120] Not for the first time, the Fifth Circuit had ruled according to what it thought the law ought to be in the area of reproductive freedom, rather than what it was.[121] Or so the Chief, as the official head of the judicial branch of the federal government, might well have thought. On this account, Chief Justice Roberts was policing lower courts—sending them a clear message that precedent binds lower courts and must be followed by them, regardless of its hold on the Supreme Court itself.

Overall, then, there is reason to doubt that the liberal justices’ concessions in religion cases yielded gains in other areas of law. Bostock and June Medical can be explained in other ways. If liberal justices engaged in appeasement, or in strategic action of another type, it is far from obvious that they have anything to show for it, and there are countervailing reasons for concern that their concessions have contributed to real setbacks in the law governing religious freedom.


Political positions are now distributed on the Court in much the same way they are in ordinary politics—they are polarized. That distribution is regrettable, because it exacerbates conflict and inhibits cooperative self-government. It is also asymmetric, as many have noted, with conservatives taking positions that are more extreme than those of their liberal counterparts. Asymmetric polarization characterizes the Court, as it does constitutional discourse more generally, outside of courts.[122]

Given this situation, what are liberal and progressive constitutional actors to do? One option is simply to vote their conscience—interpret the Constitution in the manner that is most justified and that best fits existing traditions. But they may reasonably believe that doing so consistently will only exacerbate polarization, while yielding no gains, especially now that the Roberts Court has five reliable conservatives. So liberal justices may be inclined to engage in some degree of strategic behavior. They could compromise in a loose manner without engaging in actual horse-trading, they could attempt to coopt likely allies by slowly building relationships of trust, or they could appease in the hope of forestalling worse outcomes.

Although we cannot prove that the liberal justices have engaged in appeasement, we have presented evidence from the Court’s most recent religious freedom cases that is consistent with such a strategy. And we have warned against the dangers of selecting that option. No one likes polarization, perhaps least of all in constitutional politics, which concern fundamental democratic principles, including those governing the relationship between government and religion. But unilateral concessions may do more to exacerbate than to ameliorate our divisions, while risking the legitimation of decisions that might otherwise have warranted principled dissents.

* Jane M.G. Foster Professor of Law, Cornell Law School.

** Hardy Cross Dillard Professor of Law, University of Virginia School of Law. For helpful comments on previous versions, we thank Michael Dorf, Linda Greenhouse, Marty Lederman, Leah Litman, Ira Lupu, Melissa Murray, James Nelson, Richard Schragger, and Robert Tuttle.

[1] Micah Schwartzman & Nelson Tebbe, Establishment Clause Appeasement, 2019 Sup. Ct. Rev. 271.

[2] Id. at 272.

[3] Michael McConnell, On Religion, the Supreme Court Protects the Right to Be Different, N.Y. Times (July 9, 2020).

[4] Schwartzman & Tebbe, supra note 1, at 302‒03.

[5] Our Lady of Guadalupe Sch. v. Morrisey-Berru, 140 S. Ct. 2049 (2020).

[6] Espinoza v. Mont. Dep't of Revenue, 140 S. Ct. 2246 (2020).

[7] Id.

[8] Bostock v. Clayton Cty., 140 S. Ct. 1731 (2020).

[9] June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103 (2020).

[10] Trinity Lutheran Church of Colum. v. Comer, 137 S. Ct. 2012 (2017); see Schwartzman & Tebbe, supra note 1, at 288‒95.

[11] One alternative we consider below is that these liberal justices agreed to join conservative outcomes or opinions in order to influence the reasoning of the opinions. A narrower version is that they did not vote differently from how they otherwise would have, but they accepted reasoning that they otherwise would have resisted in order to shape the majority’s opinions. Both of these accounts attribute strategic reasoning to the justices, but they identify the strategy as one of compromise rather than appeasement.

[12] This part draws on Schwartzman & Tebbe, supra note 1, at 273‒76, 304‒11.

[13] Id. at 276‒301.

[14] Trinity Lutheran, 137 S. Ct. 2012.

[15] Id. at 2025.

[16] Id. at 2017.

[17] Id. at 2021, 2024.

[18] Id. at 2016.

[19] Id.

[20] Id. at 2026‒27.

[21] Id. at 2024 n.3.

[22] Id. at 2017.

[23] See id. at 2025‒26 (Gorsuch, J., concurring).

[24] Id. at 2016.

[25] See, e.g., Frank Ravitch, Symposium: Trinity Lutheran and Zelman—Saved by Footnote 3 or a Dream Come True for Voucher Advocates?, SCOTUSBlog (June 26, 2017).

[26] See Schwartzman & Tebbe, supra note 1, at 292 (discussing Ariz. Christian Sch. Tuition Org. v. Winn, 563 U.S. 125 (2011)).

[27] See Winn, 563 U.S. at 148 (Kagan, J., dissenting).

[28] Trinity Lutheran, 137 S. Ct. at 2026 (Breyer, J., concurring in the judgment).

[29] Zelman v. Simmons-Harris, 536 U.S. 639, 717 (2002) (Breyer, J., dissenting).

[30] See Schwartzman & Tebbe, supra note 1, at 301‒04.

[31] Id. at 302.

[32] McConnell, supra note 3.

[33] See, e.g., Richard W. Garnett & Jackson C. Blais, Religious Freedom and Recycled Tires: The Meaning and Implications of Trinity Lutheran, 2016‒2017 Cato Sup. Ct. Rev. 105, 121 (2017) (“[I]t is striking and significant that a seven-justice majority, in a roiling political environment and a case that is at least adjacent to the culture-war arena, ruled that the Constitution requires the disbursal of funds to a church for its school.”).

[34] McConnell, supra note 3.

[35] Id.

[36] See Schwartzman & Tebbe, supra note 1, at 271 n.1 (collecting sources).

[37] See Trinity Lutheran Church of Colum. v. Comer, 137 S. Ct. 2012, 2041 (2017) (Sotomayor, J., dissenting).

[38] In Jack Balkin’s memorable terms, appeasement can influence which interpretations are considered “off the wall” rather than “on the wall.” Jack Balkin, From Off the Wall to On the Wall: How the Mandate Went Mainstream, Atlantic (June 4, 2012).

[39] McConnell, supra note 3.

[40] Fairness for All Act, H.R. 5331, 116th Cong. (2019).

[41] See Micah Schwartzman, Judicial Compromise and Political Uncertainty (or, What If You’re Not Sugar Ray Robinson?), Balkinization (July 22, 2020); Kelsey Dallas, Five Years Ago, Utah Passed Landmark Legislation on LGBTQ and Religious Rights. Why Didn’t Other States Follow Its Lead?, Deseret News (Mar. 11, 2020).

[42] A related possibility is that liberal justices have restrained their votes or reasoning in order to preserve collegiality amongst the justices or to promote the Court’s institutional legitimacy. Insofar as these moves are designed to build capital that the justices can spend later, or in other cases, they may fall into our categories of compromise or cooptation. But insofar as they attend to these institutional concerns without careful attention to success in achieving better reasoning or better outcomes, their decision-making strategies may shade from compromise or cooptation into appeasement, which is often defined in terms of seeking conflict avoidance and aiming to mollify other parties. And to the extent justices vote in ways that they otherwise would not, or join opinions with reasoning that they would otherwise reject, they may also violate role-based duties of sincerity and candor, which are integral to the normative legitimacy of adjudication. See Micah Schwartzman, Judicial Sincerity, 94 Va. L. Rev. 987, 1022‒24 (2008) (criticizing claims that promoting collegiality and legitimacy can justify insincere judicial decisions).


[43] See Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 266 (1977) (setting out factors, including the effects of official actions, for determining their intent or purpose).

[44] See, e.g., Paul Kennedy, A Time to Appease, 108 Nat’l Int. 7, 13 (2010) (“Certainty about such matters only comes, I suspect, with hindsight; and there we are wise, because we know what happened.”).

[45] Espinoza v. Montana Dep't of Revenue, 140 S. Ct. 2246, 2251 (2020).

[46] Id.

[47] Espinoza v. Mont. Dep’t of Revenue, 435 P.3d 603, 607 (Mont. 2018) (“[M]ost [institutions that grantees attended] were religiously-affiliated private schools.”).

[48] Espinoza, 140 S. Ct. at 2252.

[49] Mont. Const. art. X, § 6(1).

[50] Espinoza, 140 S. Ct. at 2252.

[51] See Espinoza, 435 P.3d at 612.

[52] Id. at 614.

[53] Espinoza, 140 S. Ct. at 2255 (quoting Trinity Lutheran Church of Colum. v. Comer, 137 S. Ct. 2012, 2021 (2017)).

[54] Id. at 2260.

[55] Id. at 2262.

[56] Id. at 2257.

[57] Id.

[58] Id. at 2285 (Breyer, J., dissenting).

[59] Id. at 2287.

[60] Id. at 2289‒92.

[61] Id. at 2291 (quoting Van Orden v. Perry, 545 U.S. 677, 700 (2005) (Breyer, J., dissenting)).

[62] Id. at 2278 (Ginsburg, J., dissenting).

[63] Id. at 2280.

[64] See, e.g., Carson v. Makin, 401 F. Supp. 3d 207, 211 (D. Maine 2019) (upholding Maine’s school choice program even though it excludes “sectarian schools” and relying on Trinity Lutheran footnote 3).

[65] Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682, 692‒93 (2014).

[66] Zubik v. Burwell, 136 S. Ct. 1557 (2016) (per curiam).

[67] Little Sisters of the Poor Saints Peter & Paul Home v. Pennsylvania, 140 S. Ct. 2367, 2377‒78 (2020).

[68] Id. at 2403 (Ginsburg, J., dissenting).

[69] Id. at 2401.

[70] Id. at 2378‒79 (majority opinion).

[71] Id. at 2386.

[72] Id. at 2382‒83.

[73] Id. at 2373.

[74] Id. at 2397 (Kagan, J., concurring in the judgment) (“Chevron deference was built for cases like these.”) (citing Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842–43 (1984)).

[75] Id. at 2398 (quoting 5 U.S.C. § 706(2)(A)).

[76] Id. at 2398‒99. The Court’s decision in Hobby Lobby only gives closely-held business corporations protection under RFRA, bracketing whether publicly traded businesses might also be shielded by the statute. Burwell v. Hobby Lobby Stores, Inc, 573 U.S. 682, 717 (2014).

[77] Hobby Lobby, 573 U.S. at 739‒40 (Ginsburg, J., dissenting).

[78] Little Sisters of the Poor, 140 S. Ct. at 2399 (Kagan, J., concurring in the judgment) (“[T]he rule’s overbreadth causes serious harm, by the Departments’ own lights.”).

[79] Hobby Lobby, 573 U.S. at 772 (Kagan, J., dissenting).

[80] Cf. Noah Feldman, Why Supreme Court Liberals Joined Conservatives on Religion, Bloomberg (July 8, 2020).

[81] Hosanna-Tabor Evangelical Lutheran Church & School v. E.E.O.C., 565 U.S. 171, 190 (2012).

[82] Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049, 2055 (2020).

[83] Id. at 2058‒59.

[84] Id. at 2066.

[85] Id. at 2058‒59.

[86] Id. at 2056‒57.

[87] Id. 2058‒59.

[88] Id.

[89] Id. at 2060.

[90] Id. at 2061.

[91] Id. at 2064.

[92] Id.

[93] Id. at 2066.

[94] Id. at 2051.

[95] See Hosanna-Tabor Evangelical Lutheran Church & School v. E.E.O.C., 565 U.S. 171, 198 (2020) (Alito, J., concurring).

[96] See Employment Div., Dep't of Human Res. v. Smith, 494 U.S. 872 (1990).

[97] Ira C. Lupu & Robert W. Tuttle, The Mystery of Unanimity in Hosanna-Tabor Evangelical Lutheran Church & School v. EEOC, 20 Lewis & Clark L. Rev. 1265 (2017). Lupu and Tuttle resolve the mystery by arguing that the justices share a commitment to avoiding ecclesiastical questions, and that Hosanna-Tabor’s ministerial exception is grounded in that commitment. They therefore would argue that the liberal members of the Court support the ministerial exception for reasons that are principled, not strategic.

[98] Feldman, supra note 80.

[99] Id.

[100] Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049, 2060 (2020).

[101] See Ira C. Lupu & Robert W. Tuttle, Commentary, The 2020 Ministerial Exception Cases: A Clarification, Not a Revolution, Take Care Blog (July 8, 2020) (“That the Court majority did not follow the Thomas-Gorsuch path is of profound consequence to the future of Religion Clause principles. Their broad view of the autonomy of religious institutions would give such institutions grounds for ignoring a wide variety of legal norms.”).

[102] Our Lady of Guadalupe, 140 S. Ct. at 2060 (emphasis added).

[103] Feldman, supra note 80; see also Richard Schragger & Micah Schwartzman, Against Religious Institutionalism, 99 Va. L. Rev. 917, 946 (arguing that “there is no centrally defined core institutional mission of the church on which to build a limited account of institutional autonomy”).

[104] Feldman, supra note 80 (“[I]t could be argued that, by joining the conservatives [in Our Lady of Guadalupe], Breyer and Kagan helped bring about the result in the LBTQ [sic] case.”).

[105] Bostock v. Clayton Cty., 140 S. Ct. 1731 (2020).

[106] June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103 (2020).

[107] Feldman, supra note 80.

[108] Justin McCarthy, Slim Majority in U.S. Favors New LGBT Civil Rights Laws, Gallup (June 13, 2019).

[109] See Kent Greenfield & Adam Winkler, Did John Roberts Doom Supreme Court Reform with His Decisions?, Hill (July 8, 2020). It is also possible that Chief Justice Roberts voted with the majority, once he knew how Justice Gorsuch was voting, so that he could have the power to assign the opinion to Justice Gorsuch. (The senior justice in the majority has the power to assign the opinion.) On that theory as well, Chief Justice Roberts voted with the majority for his own reasons, independent of whatever Justices Breyer or Kagan had done in the past.

[110] See Andrew Koppelman, Supreme Court Rulings Make the World Safer for Both LGBT People and Religious Freedom, USA Today (July 21, 2020); Mark Movsesian, The Roberts Court Attempts a Compromise, First Things (July 15, 2020); David French, The Supreme Court Tries to Settle the Religious Liberty Culture War, Time (July 14, 2020).

[111] See Nelson Tebbe, Religious Freedom in an Egalitarian Age 152‒57 (2017); Schwartzman, supra note 42.

[112] To be fair, the Fairness for All bill has also been opposed from the right, but that does not mean that the Roberts Court majority is not pursuing a similar settlement. See Ryan T. Anderson, “Fairness for All” Is Well Intentioned But Inadequate and Misguided, Heritage Found. (Dec. 7, 2019).

[113] June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103 (2020).

[114] Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292 (2016).

[115] June Medical, 140 S. Ct. at 2133.

[116] See Frederick Schauer, Stare Decisis—Rhetoric and Reality in the Supreme Court, 2018 Sup. Ct. Rev. 121, 131.

[117] See Richard M. Re, Narrowing Precedent in the Supreme Court, 114 Colum. L. Rev. 1861 (2014). This pattern seems to have held in June Medical. Chief Justice Roberts affirmed the holding of Whole Woman’s Health, but rejected its reasoning and, in the process, may have laid the groundwork for further narrowing of reproductive rights. See Melissa Murray, The Symbiosis of Abortion and Precedent, 134 Harv L. Rev. (forthcoming 2020); Melissa Murray, The Supreme Court’s Abortion Decision Seems Pulled From The ‘Casey’ Playbook, Wash. Post (June 29, 2020) (“In this way, Roberts’s decision in June Medical Services does to Whole Woman’s Health what Casey did to Roe. It preserves the outer shell of the earlier decision while gutting its substance.”); Leah Litman, June Medical as the New Casey, Take Care Blog (June 26, 2020).

[118] McConnell, supra note 3.

[119] Feldman, supra note 80.

[120] June Med. Servs. L.L.C. v. Gee, 905 F.3d 787 (5th Cir. 2018) cert. granted, 140 S. Ct. 35, 204 L. Ed. 2d 1193 (2019), and cert. granted, 140 S. Ct. 35, 204 L. Ed. 2d 1193 (2019), and rev'd sub nom. June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103 (2020).

[121] Whole Woman’s Health itself was another example.

[122] See Joseph Fishkin & David E. Pozen, Asymmetric Constitutional Hardball, 118 Colum. L. Rev. 915 (2018); Rick L. Hasen, Polarization and the Judiciary, 22 Ann. Rev. Pol. Sci. 261 (2019).