At the height of the COVID‑19 pandemic, the United States Supreme Court resolved a decades-long debate about the Free Exercise Clause of the First Amendment.1See, e.g., Christopher C. Lund, Second-Best Free Exercise, 91 Fordham L. Rev. 843, 847 (2022) (“The Free Exercise Clause is in the middle of a remarkable transformation.”); Thomas C. Berg, Religious Freedom Amid the Tumult, 17 U. St. Thomas L.J. 735, 735–36 (2022) (calling the issue “[t]he most extraordinary religious-freedom question of the last two years”). Not in a lengthy opinion evaluating the complex doctrine, but in a four-page per curiam decision.2See generally Tandon v. Newsom, 141 S. Ct. 1294 (2021) (per curiam). In just a few lines, the Court shifted the free-exercise framework, holding that religious observers are entitled to exemptions from laws that exempt “comparable” secular conduct.3Id. at 1296. More specifically, the Court held that such laws are subject to strict scrutiny. But if a law treats secular conduct better than “comparable” religious conduct, it almost inevitably fails strict scrutiny. More on that later. See infra note 338 and accompanying text.
What makes religious and secular conduct comparable has since become the subject of considerable discussion.4See, e.g., Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1921 (2021) (Alito, J., concurring) (noting the difficulty of “[i]dentifying appropriate comparators”); Justin Collings & Stephanie Hall Barclay, Taking Justification Seriously: Proportionality, Strict Scrutiny, and the Substance of Religious Liberty, 63 B.C. L. Rev. 453, 457 (2022) (noting that “scholars and jurists alike legitimately debate what religious and secular conduct is analogous”); Lund, supra note 1, at 849 n.41 (“Of course, the crucial thing will be deciding which secular activities count as equivalents.”); Cass R. Sunstein, Our Anti-Korematsu, 2021 Am. J.L. Equal. 221, 235 (2021) (“[W]hat kinds of factual demonstration are necessary, exactly, to show relevant similarity?”); Zalman Rothschild, Free Exercise’s Lingering Ambiguity, 11 Cal. L. Rev. Online 282, 284–85, 284 n.14 (2020) (pointing out practical difficulties in comparing religious and secular conduct); Mark Storslee, The COVID‑19 Church-Closure Cases and the Free Exercise of Religion, 37 J.L. & Religion 72, 78 (2022) (commenting on “just how hard it can be to identify which activities are truly ‘comparable,’ at least when the underlying controversy relates to highly technical questions like disease transmission”). New cases posing hard questions seem to arise daily. Is a movie theater comparable to a church?5See, e.g., S. Bay United Pentecostal Church v. Newsom, 140 S. Ct. 1613, 1613 (2020) (Roberts, C.J., concurring); S. Bay United Pentecostal Church v. Newsom, 141 S. Ct. 716 (2021). Is declining a vaccine for medical reasons comparable to declining for religious reasons?6See, e.g., Doe v. San Diego Unified Sch. Dist., 19 F.4th 1173, 1176 (9th Cir. 2021), reh’g en banc denied, 22 F.4th 1099 (9th Cir. 2022); We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 285 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). Is a school affinity group that limits membership based on gender identity comparable to a religious student organization that limits membership based on sexual orientation?7See, e.g., Fellowship of Christian Athletes v. San Jose Unified Sch. Dist. Bd. of Educ., 46 F.4th 1075, 1084 (9th Cir. 2022).
Despite framing the key question in terms of religious and secular comparators, the Court offered little guidance on how to compare religious and secular acts and entities.8See sources cited supra note 4. It thus gave district judges virtually free reign, and the results in recent years have been disturbingly partisan.9See, e.g., Zalman Rothschild, Free Exercise Partisanship, 107 Cornell L. Rev. 1067, 1082 (2022) (finding that from 2016 through 2020, “Democratic-appointed judges sided with the government 93% of the time and with religious plaintiffs 7% of the time, while Republican-appointed judges sided with the government 44% of the time and with religious plaintiffs 56% of the time (a 49% differential)”); see id. at 1071 (arguing that “the devolution towards partisanship has been enabled by ambiguity in the relevant free exercise doctrine”). Even so, commentators have paid little attention to the questions of who must prove or disprove comparability, and how.10Two commentators recently put the issue mildly, noting that “[i]t isn’t entirely clear, for instance, who bears the burden of empirical uncertainty—the rights claimant or the legislature.” Collings & Barclay, supra note 4, at 474. This Article focuses on those questions.
The Free Exercise Clause says that “Congress shall make no law . . . prohibiting the free exercise” of religion.11U.S. Const. amend. I. Courts have interpreted this to mean that some religious observers are entitled to exemptions from laws that restrict their religious conduct.12See, e.g., Sherbert v. Verner, 374 U.S. 398, 409–10 (1963); Wisconsin v. Yoder, 406 U.S. 205, 236 (1972); Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021). The exemptions framework arguably deviates from the text of the First Amendment, which begins that “Congress shall make no law.” U.S. Const. amend. I (emphasis added); see Nicholas Quinn Rosenkranz, The Subjects of the Constitution, 62 Stan. L. Rev. 1209, 1266 (2010) (“If Congress violates the Free Exercise Clause by making a law prohibiting the free exercise of religion, then it must be that the violation happens when Congress makes such a law.” ); Philip A. Hamburger, A Constitutional Right of Religious Exemption: An Historical Perspective, 60 Geo. Wash. L. Rev. 915, 938 (1992) (“Rather than suppose that civil laws will in some respects prohibit the free exercise of religion and that exemptions will be necessary, the First Amendment assumes Congress can avoid enacting laws that prohibit free exercise.”); Gerard V. Bradley, Beguiled: Free Exercise Exemptions and the Siren Song of Liberalism, 20 Hofstra L. Rev. 245, 306 (1991) (“‘Congress shall make no law . . . .’ This means that a class of legislation is forbidden. A class is definable by foreclosing legislative adoption of truth claims of one or another church. But the conduct exemption does not forbid a class of legislation.”). But see Stephanie H. Barclay, The Historical Origins of Judicial Religious Exemptions, 96 Notre Dame L. Rev. 55, 120 (2020) (arguing that “[t]his textual interpretation . . . overlooks the historical context in which equitable exemptions arose and evolved”). As the Supreme Court said more than a century ago, however, the First Amendment doesn’t “make the professed doctrines of religious belief superior to the law of the land.”13Reynolds v. United States, 98 U.S. 145, 167 (1879). More recently, the Court held that the First Amendment doesn’t exempt religious observers from “neutral, generally applicable law[s]” that incidentally infringe on their “religiously motivated action.”14Emp. Div. v. Smith, 494 U.S. 872, 881 (1990). If a law isn’t neutral and generally applicable vis-à-vis religion, it warrants strict scrutiny.15See Church of the Lukumi Babalu Aye v. City of Hialeah, 508 U.S. 520, 531 (1993). That is, religious observers are entitled to exemptions from laws that burden their religious exercise and aren’t neutral and generally applicable, unless the government can prove that such laws are narrowly tailored to achieving compelling state interests.16Id.; see also Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793, 861–62 (2006).
Unsurprisingly, the crucial issue in many exemptions cases is whether the law or state action in question is neutral and generally applicable.17See, e.g., Fulton, 141 S. Ct. at 1877; Does v. Mills, 16 F.4th 20, 30 (1st Cir. 2021); Agudath Israel of Am. v. Cuomo, 983 F.3d 620, 631 (2d Cir. 2020); McTernan v. City of York, 564 F.3d 636, 648 (3d Cir. 2009); Canaan Christian Church v. Montgomery County, 29 F.4th 182, 199 (4th Cir. 2022); Cornerstone Christian Schs. v. Univ. Interscholastic League, 563 F.3d 127, 137 (5th Cir. 2009); Roberts v. Neace, 958 F.3d 409, 415 (6th Cir. 2020); St. Augustine Sch. v. Evers, 906 F.3d 591, 595 (7th Cir. 2018); New Doe Child #1 v. United States, 901 F.3d 1015, 1025 (8th Cir. 2018); Parents for Priv. v. Barr, 949 F.3d 1210, 1234 (9th Cir. 2020); 303 Creative LLC v. Elenis, 6 F.4th 1160, 1183 (10th Cir. 2021), cert. granted in part, 142 S. Ct. 1106 (2022); Eternal Word Television Network, Inc. v. Sec’y of U.S. Dep’t of Health & Hum. Servs., 818 F.3d 1122, 1165 (11th Cir. 2016), vacated, 2016 U.S. App. LEXIS 24382 (11th Cir. May 31, 2016); Archdiocese of Washington v. Washington Metro. Area Transit Auth., 897 F.3d 314, 331 (D.C. Cir. 2018). In its short per curiam opinion in 2021, the Supreme Court held that a law isn’t generally applicable with respect to religion, and “therefore trigger[s] strict scrutiny under the Free Exercise Clause, whenever [it] treat[s] any comparable secular activity more favorably than religious exercise.”18Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). Comparability is “judged against the asserted government interest that justifies the regulation at issue.”19Id. Religious and secular acts are “comparable” if they “undermine the government’s asserted interests in a similar way.”20Fulton, 141 S. Ct. at 1877. A state can’t prohibit the slaughter of animals for sanitation reasons but then exempt secular forms of slaughter that undermine the sanitation interest as much or more than religious killings that remain prohibited.21See Church of the Lukumi, 508 U.S. at 535–36. Nor may a police department that prohibits officers from growing beards exempt officers with medical conditions but decline to exempt Muslim officers, when the two sets of exemptions would detract from the purpose of the rule in a similar way.22See Fraternal Order of Police Newark Lodge No. 12 v. City of Newark, 170 F.3d 359, 366 (3d Cir. 1999).
But when a law exempts certain secular conduct, how can courts assess whether a requested religious exemption would “undermine the government’s asserted interests in a similar way”?23Fulton, 141 S. Ct. at 1877. What interests can the government assert? Who bears the burden to show or disprove comparability? What evidence is relevant to that inquiry?24Professor Cass Sunstein introduced some of these questions in a short paper in 2021. See Sunstein, supra note 4, at 235. (“[W]hat kinds of factual demonstration are necessary, exactly, to show relevant similarity?”). Since the Supreme Court formally adopted the new general-applicability standard in 2021, these questions have become all the more critical in exemptions cases.25Some have criticized the comparability test as “unworkable” given the challenges of comparing religious and secular activities. See, e.g., Zalman Rothschild, Individualized Exemptions, Vaccine Mandates, and the New Free Exercise Clause, 131 Yale L.J.F. 1106, 1121 n.59 (2022) (arguing that “a test predicated on identifying similarly situated secular and religious comparators is unworkable”). Without denying these challenges, this Article comments on how courts can, should, and have grappled with them.
Even so, courts rarely address them explicitly and have reached inconsistent conclusions.26See Collings & Barclay, supra note 4, at 457 (noting that “scholars and jurists alike legitimately debate what religious and secular conduct is analogous,” but “many courts do not”). The Supreme Court has said that religious claimants bear the burden to show that laws are “not ‘neutral’ or ‘generally applicable’” (i.e., that “comparable” secular activity is exempt).27See Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2421–22 (2022) (emphasis added) (“Under this Court’s precedents, a plaintiff may carry the burden of proving a free exercise violation in various ways, including by showing that a government entity has burdened his sincere religious practice pursuant to a policy that is not ‘neutral’ or ‘generally applicable.’”); see also 303 Creative LLC v. Elenis, 6 F.4th 1160, 1186 n.8 (10th Cir. 2021), cert. granted in part, 142 S. Ct. 1106 (2022). But in addressing California’s COVID-19 restrictions, the Court commented on the State’s failure “to explain why it could not safely permit at-home worshipers to gather in larger numbers” despite permitting larger gatherings in commercial settings (i.e., why the religious gatherings at issue were incomparable to the exempt secular gatherings).28Tandon v. Newsom, 141 S. Ct. 1294, 1297 (2021).
Courts also have struggled to ascertain what the relevant governmental interests are. While the government’s “asserted” interests count under recent Supreme Court doctrine,29Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021); Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 545 (1993). questions remain about which assertions are relevant (e.g., pre-enactment explanations vs. post-hoc rationalizations), and how broad and numerous the asserted interests can be.30See, e.g., Does v. Mills, 142 S. Ct. 17, 20 (2021) (Gorsuch, J., dissenting) (arguing “that only the government’s actually asserted interests as applied to the parties before it count—not post-hoc reimaginings of those interests expanded to some society-wide level of generality”).
Finally, courts have applied inconsistent evidentiary standards in assessing comparability. In declining to exempt religious observers from a COVID-19 vaccine mandate, one court emphasized that the mandate’s medical exemptions were “likely to be more limited in number than religious exemptions,” and thus that the requested religious exemptions would undermine the state’s COVID-19 measures more than the medical exemptions.31We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 286 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). A judge in another vaccine-mandate case rejected the same reasoning, concluding that “an influx of religious accommodation requests,” compared to a small number of medical exemptions, “is not a valid reason to deny First Amendment rights.”32U.S. Navy SEALs 1-26 v. Biden, 578 F. Supp. 3d 822, 838 (N.D. Tex. 2022).
These doctrinal inconsistencies are becoming more pronounced as lower courts attempt to apply the new standard.33See, e.g., Rothschild, supra note 9, at 1071 (arguing that “the devolution towards partisanship has been enabled by ambiguity in the relevant free exercise doctrine”). This Article identifies these often-implicit discrepancies and proposes ways courts might approach these issues in future cases. In other words, this Article attempts to answer the “how” questions—namely, how courts can and should align the new comparability standard with the purposes of the general applicability test. Part I traces the origins of general applicability in the free exercise context. Part II examines the competing interpretive theories, the approach the Supreme Court recently adopted, the practical implications of the new standard, and the reasons general applicability matters. Part III focuses on how courts have been applying the recent doctrine and how they should determine whether secular and religious exemptions would undermine state interests in similar ways. This Article concludes by suggesting that the best way to assess religious and secular comparators is through a burden-shifting framework familiar in the antidiscrimination context, as well as a permissive evidentiary inquiry.
There are many important questions this Article doesn’t address. It has little to say about whether general applicability should matter for free exercise purposes, or whether general applicability should turn on the comparison of religious and secular activities. Although some of the following discussion inevitably implicates such questions, they aren’t this Article’s focus. Courts and commentators have debated these issues for decades.34For a discussion of whether general applicability is a proper test under the Free Exercise Clause, compare Michael W. McConnell, Free Exercise Revisionism and the Smith Decision, 57 U. Chi. L. Rev. 1109, 1117–18 (1990) [hereinafter McConnell, Revisionism] (arguing that the Free Exercise Clause calls for religious exemptions from some generally applicable laws) and Michael W. McConnell, The Origins and Historical Understanding of Free Exercise of Religion, 103 Harv. L. Rev. 1409, 1512 (1990) [hereinafter McConnell, Origins] (contending that early state constitutions and statutes supported the view that the Free Exercise Clause allows for religious exemptions from generally applicable laws) with William P. Marshall, Correspondence on Free Exercise Revisionism – In Defense of Smith and Free Exercise Revisionism, 58 U. Chi. L. Rev. 308, 320 (1991) (arguing that the Free Exercise Clause doesn’t require exemptions from generally applicable laws) and Hamburger, supra note 12, at 948 (arguing that “[a] more general examination of religious freedom in late eighteenth-century America reveals that a constitutional right of religious exemption was not even an issue in serious contention among the vast majority of Americans”). For a discussion of what general applicability means, see infra Section II.A. The Supreme Court answered the first question in 1990,35See Emp. Div. v. Smith, 494 U.S. 872, 881–82 (1990). and declined to revisit that decision in 2021.36See Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1878 (2021). It answered the second question at the height of the COVID-19 pandemic.37See Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). For better or worse, neutral, generally applicable laws don’t trigger strict scrutiny under the Free Exercise Clause.38Id. And laws are generally applicable unless they treat religion worse than comparable secular conduct.39Id. These are the waters of modern free exercise. The question now is how to navigate.40Nor does this Article spend substantial time addressing whether specific secular and religious activities are comparable. As this Article contends, such questions are case-specific and context-dependent. The more important issues are how courts can, do, and should assess religious and secular comparators in any given case.
I. General Applicability Origins
The concept of general applicability emerged in the free-exercise context in the last fifty years.41See Christopher C. Lund, A Matter of Constitutional Luck: The General Applicability Requirement in Free Exercise Jurisprudence, 26 Harv. J.L. & Pub. Pol’y 627, 627–30 (2003). The concept exists in other constitutional contexts, too. See generally Jeffrey M. Shaman, Rules of General Applicability, 10 First Amend. L. Rev. 419 (2012) (discussing the concept of general applicability in the contexts of equal protection, free speech, and free exercise). Before then, the Free Exercise Clause typically was read to require exemptions from certain laws that infringed on the exercise of religion, regardless of whether those laws were neutral and generally applicable.42McConnell, Origins, supra note 34, at 1511. In a series of cases at the end of the twentieth century, the doctrine changed.43Lund, supra note 41, at 633–34.
A. United States v. Lee
In the 1970s, an Amish farmer refused to pay social security taxes or participate in the social security system.44United States v. Lee, 455 U.S. 252, 254 (1982). The IRS fined him more than $27,000 for unpaid taxes. He paid $91, the amount owed for the first quarter of 1973, then sued the United States in federal court for a refund, claiming that the imposition of the social security tax violated his free-exercise rights.45Id. at 254–55. The case reached the Supreme Court, which accepted his belief that his faith forbade making social security payments or receiving benefits.46Id. at 257. But the Court held that the government had “justif[ied] [this] limitation on religious liberty by showing that it [was] essential to accomplish[ing] an overriding governmental interest.”47Id. at 257–58. .
Justice John Paul Stevens concurred in what would become crucial language in the development of free exercise jurisprudence.48Lund, supra note 41, at 630 n.12. He began by observing that the Court’s “constitutional standard suggest[ed] that the Government always bears a heavy burden of justifying the application of neutral general laws to individual conscientious objectors.”49Lee, 455 U.S. at 262 (Stevens, J., concurring). He disagreed with that structure, arguing that a religious objector should “shoulder the burden of demonstrating that there is a unique reason for allowing him a special exemption from a valid law of general applicability.”50Id. He agreed with the Court’s reasoning “that the difficulties associated with processing other claims to tax exemption on religious grounds” justified rejecting the petitioner’s claim, but only because “this reasoning support[ed] the adoption of a different constitutional standard than the Court purport[ed] to apply.”51Id. at 263. That is, “[t]he Court’s analysis support[ed] a holding that there [was] virtually no room for a ‘constitutionally required exemption’ on religious grounds from a valid tax law that [was] entirely neutral in its general application.”52Id.
B. Employment Division v. Smith
Almost a decade later, the Court adopted Justice Stevens’s view in what would become one of its most significant free exercise decisions. In the 1980s, the State of Oregon prohibited possession of “controlled substance[s]” except when prescribed by medical practitioners.53Emp. Div. v. Smith, 494 U.S. 872, 874 (1990). Alfred Smith and Galen Black were fired when they used the controlled substance, peyote, for sacramental purposes at a ceremony of the Native American Church.54Id. The State denied them unemployment benefits because they had been fired for what counted as work-related misconduct.55Id. They brought a free-exercise challenge in state court, and their case reached the Supreme Court of Oregon.56Id. at 875. The court concluded they were entitled to unemployment benefits because the State’s reason for the “misconduct” rule—to preserve the financial integrity of the compensation fund—didn’t justify infringing on their religious practice.57Id.
The United States Supreme Court reversed.58Id. at 890. It held that the compelling interest test—strict scrutiny—doesn’t apply to “neutral, generally applicable law[s]” that burden “religiously motivated action.”59Id. at 881. “We have never held,” it explained, “that an individual’s religious beliefs excuse him from compliance with an otherwise valid law prohibiting conduct that the State is free to regulate.”60Id. at 878–79. It opined that “precisely because we value and protect religious divergence, we cannot afford the luxury of deeming presumptively invalid, as applied to the religious objector, every regulation of conduct that does not protect an interest of the highest order.”61Id. at 888.
C. Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah
Several years after Smith, the Court had its first opportunity to clarify the scope of its holding. The case, Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah,62508 U.S. 520 (1993). arose when a Santeria church, which practiced animal sacrifice, leased land in a Florida city.63Id. at 525–26. In response to “concern” from members of the community “that certain religions may propose to engage in practices which are inconsistent with public morals, peace, or safety,” the city council convened and enacted several ordinances.64Id. at 526. The ordinances prohibited “sacrific[ing] any animal within the corporate limits of the City.”65Id. at 528. They applied to any group that killed, slaughtered or sacrificed animals for “any type of ritual.”66Id. at 527. They exempted slaughtering by “licensed establishments” of animals “specifically raised for food purposes,” including kosher slaughter.67Id. at 528, 536.
The church brought a free-exercise challenge, and the case reached the Supreme Court.68Id. at 523. The Court held that the ordinances violated the church’s free-exercise rights.69Id. at 547. First, it explained that, “[a]t a minimum, the protections of the Free Exercise Clause pertain if the law at issue discriminates against some or all religious beliefs or regulates or prohibits conduct because it is undertaken for religious reasons.”70Id. at 532. Citing Smith, the Court explained that “if the object of a law is to infringe upon or restrict practices because of their religious motivation, the law is not neutral, and it is invalid unless it is justified by a compelling interest and is narrowly tailored to advance that interest.”71Id. at 533 (citations omitted). Because the ordinances “exclude[d] almost all killings of animals except for religious sacrifice” and permitted “killings that [were] no more necessary or humane” than Santeria slaughters, they were both overinclusive and underinclusive with respect to their purported objective.72Id. at 536. The Court thus determined that the ordinances weren’t neutral.73Id. at 538.
Second, the Court held that the ordinances weren’t generally applicable. Although it declined to “define with precision the standard used to evaluate whether a prohibition is of general application,” it held that the ordinances failed the test for the same reasons they weren’t neutral.74Id. at 543. It noted, for example, that the city “ha[d] not explained why commercial operations that slaughter ‘small numbers’ of hogs and cattle”—which were exempt from the ordinances—did “not implicate its professed desire to prevent cruelty to animals and preserve the public health.”75Id. at 545.
Because the ordinances weren’t neutral and generally applicable, the Court applied strict scrutiny.76See id. at 531. It noted that “[a] law that targets religious conduct for distinctive treatment or advances legitimate governmental interests only against conduct with a religious motivation will survive strict scrutiny only in rare cases.”77Id. at 546. Even if the city had compelling interests for the ordinances, they weren’t narrowly tailored because “[t]he proffered objectives [were] not pursued with respect to analogous nonreligious conduct” and “could [have been] achieved by narrower ordinances that burdened religion to a far lesser degree.”78Id.
Justice Antonin Scalia, the author of Smith, concurred. He urged that a
defect of lack of neutrality applies primarily to those laws that by their terms impose disabilities on the basis of religion (e.g., a law excluding members of a certain sect from public benefits), whereas the defect of lack of general applicability applies primarily to those laws which, though neutral in their terms, through their design, construction, or enforcement target the practices of a particular religion for discriminatory treatment.79Id. at 557 (Scalia, J., concurring) (citations omitted).
II. General Applicability Theory
In determining whether laws are generally applicable under Smith and Lukumi, courts must make two decisions. First, they must decide what makes a law generally applicable. Second, they need to figure out how to ascertain general applicability under the substantive standard. All the while, they must keep in mind the reasons why general applicability matters. The Supreme Court recently answered the first question, as discussed below in Section A. The second question, outlined in Section B, remains unresolved and is this Article’s primary focus. Section C discusses the normative theory underlying the general applicability test—in other words, the goals general applicability aims to achieve.
A. Substantive Theory (the “What”)
What does generally applicable mean? Since Smith and Lukumi, many courts and commentators have disagreed about when laws are generally applicable. One view is that a law is generally applicable unless it is intended to discriminate against religion.80See, e.g., James M. Oleske, Jr., Lukumi at Twenty: A Legacy of Uncertainty for Religious Liberty and Animal Welfare Laws, 19 Animal L. 295, 335 (2013) (“The selective-exemption rule is properly viewed as a [narrow] tool to guard against the type of intentional discrimination prohibited by the Free Exercise Clause . . . .”); Lund, supra note 41, at 639 (“[M]any commentators have argued, often forcefully, that the general applicability inquiry should be interpreted as simply being a prohibition on intentional discrimination.”); Mark Tushnet, The Redundant Free Exercise Clause?, 33 Loy. U. Chi. L.J. 71, 71 & n.3, 72 (2001) (arguing that, under Smith and Lukumi, the Free Exercise Clause “protects only against statutes that target religious practice”). After all, the Court in Lukumi spoke in discrimination terms when it called general applicability “[t]he principle that government, in pursuit of legitimate interests, cannot in a selective manner impose burdens only on conduct motivated by religious belief . . . .”81Church of the Lukumi, 508 U.S. at 543 (emphasis added). The Supreme Court seemed to endorse the intentional discrimination view shortly after Lukumi, in City of Boerne v. Flores.82521 U.S. 507, 534 (1997). See Oleske, supra note 80, at 331 (“Having struck down a religious-exemption law enacted by Congress because it was not properly aimed at intentional discrimination, it would be odd indeed for the Court to approve a judicial religious-exemption rule that is not aimed at intentional discrimination.”); Thomas C. Berg, Can Religious Liberty Be Protected As Equality?, 85 Tex. L. Rev. 1185, 1192–93, 1197 (2007) (reviewing Christopher L. Eisgruber & Lawrence G. Sager, Religious Freedom and the Constitution (2007)) (“The Court in Boerne, in finding RFRA disproportionate to Free Exercise Clause violations, compared the statute to a relatively narrow constitutional prohibition against laws reflecting ‘bigotry,’ ‘animus,’ or ‘hostility’ toward the burdened faith.”). The Court opined that the Religious Freedom Restoration Act (“RFRA”), enacted in response to Smith, attempted to “substantive[ly] alter” Smith’s holding by creating religious exemptions to laws “without regard to whether they had the object of stifling or punishing free exercise.”83Flores, 521 U.S. at 534. The Court in City of Boerne held that RFRA was not a proper use of Congress’s power under Section Five of the Fourteenth Amendment because it called for strict scrutiny of laws not “motivated by religious bigotry.”84Id. at 535. The Court thus suggested that a law doesn’t warrant strict scrutiny under Smith—and therefore is generally applicable—unless it is “motivated by religious bigotry.”85Id. Some courts have adopted versions of this interpretation.86See, e.g., St. John’s United Church of Christ v. City of Chicago, 502 F.3d 616, 637 (7th Cir. 2007) (deeming a government action neutral and generally applicable because it was taken despite—not because of—its effect on religion); KDM ex rel. WJM v. Reedsport Sch. Dist., 196 F.3d 1046, 1051 (9th Cir. 1999) (no free exercise violation because the state action “could hardly be said to reflect a purpose to suppress religion or religious conduct” (brackets and internal quotation marks omitted)); Strout v. Albanese, 178 F.3d 57, 65 (1st Cir. 1999) (no free exercise violation where “substantial animus” didn’t “motivate the law in question”).
Still, the “discriminatory intent” view has received criticism.87See, e.g., Douglas Laycock & Steven T. Collis, Generally Applicable Law and the Free Exercise of Religion, 95 Neb. L. Rev. 1, 10 (2016) (“The Free Exercise Clause protects religious observers against unequal treatment, regardless of targeting, motive, or an improper object.” (internal quotation marks and citations omitted)). As Lukumi reiterated, Smith prescribed strict scrutiny for laws that aren’t neutral and generally applicable.88Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 531–32 (1993). While discriminatory intent usually undermines neutrality,89Discriminatory intent might not necessarily undermine neutrality, since an inept legislature theoretically could fail in the attempt to enact a law with disparate effects. See id. at 558–59 (Scalia, J., concurring) (“Had the Hialeah City Council set out resolutely to suppress the practices of Santeria, but ineptly adopted ordinances that failed to do so, I do not see how those laws could be said to ‘prohibit the free exercise’ of religion.”). “general applicability” refers to the effect of the law rather than the intent of the legislature. A law may be generally applicable (or not) regardless of its underlying purposes. This is evident in Lukumi’s distinct discussions of neutrality and general applicability. Whereas the former concerned the “object” of a law,90Id. at 533 (majority opinion). the latter involved inequitable “results.”91Id. at 542; see also Douglas Laycock, The Supreme Court and Religious Liberty, 40 Cath. Law. 25, 28 (2000) (“Whatever else it may be, Lukumi is not a motive case. The lead opinion explicitly relies on the city’s motive to exclude a particular religious group—and that part of the opinion has only two votes. So whatever the holding is, it is not a holding about motive.”).
Other courts and commentators have endorsed what could be called a “substantial underinclusion” interpretation. Under this view, a law isn’t generally applicable if it is so riddled with secular exceptions that it effectively applies only to religious conduct.92See, e.g., Ward v. Polite, 667 F.3d 727, 740 (6th Cir. 2012) (“At some point, an exception-ridden policy takes on the appearance and reality of a system of individualized exemptions, the antithesis of a neutral and generally applicable policy and just the kind of state action that must run the gauntlet of strict scrutiny.”); Parents for Priv. v. Barr, 949 F.3d 1210, 1236 (9th Cir. 2020) (“The correct inquiry here is whether, in seeking to create a safe, non-discriminatory school environment for transgender students, the Student Safety Plan selectively imposes certain conditions or restrictions only on religious conduct.”); Olsen v. Mukasey, 541 F.3d 827, 832 (8th Cir. 2008) (“General applicability does not mean absolute universality. Exceptions do not negate [general applicability].”); Frederick Mark Gedicks, The Normalized Free Exercise Clause: Three Abnormalities, 75 Ind. L.J. 77, 114 (2000) (“[U]nder the Smith doctrine, a religiously neutral law does not fail the test of general applicability merely by being modestly or even substantially underinclusive; rather, the law must be so dramatically underinclusive that religious conduct is virtually the only conduct to which the law applies.”). This approach relies on Lukumi’s admonition that “inequality results when a legislature decides that the governmental interests it seeks to advance are worthy of being pursued only against conduct with a religious motivation.”93Church of the Lukumi, 508 U.S. at 542–43 (emphasis added). Under this interpretation, a law can violate general applicability even without intentional discrimination, but only if it allows “substantial” secular conduct while prohibiting comparable religious exercise.94See id. at 543 (noting that a law isn’t generally applicable when the “underinclusion is substantial, not inconsequential”); see also Stormans, Inc. v. Wiesman, 794 F.3d 1064, 1079 (9th Cir. 2015). Appealing as this category may be on an intuitive level, however, it is susceptible to practical criticism insofar as it turns on degrees of underinclusiveness and the meaning of nebulous terms such as “substantial.”95See, e.g., Richard F. Duncan, Free Exercise Is Dead, Long Live Free Exercise: Smith, Lukumi and the General Applicability Requirement, 3 U. Pa. J. Const. L. 850, 876 (2001) (arguing that a law is generally applicable despite containing exemptions for secular conduct if “the degree of underinclusion does not appear to be substantial”).
At the other end of the spectrum is the theory the Court recently adopted, often called the “most-favored-nation” interpretation.96See Douglas Laycock, The Remnants of Free Exercise, 1990 S. Ct. Rev. 1, 49 (interpreting the general applicability test as requiring “that religion get something analogous to most-favored nation status”). Professor Nelson Tebbe recently labeled a version of the most-favored nation interpretation—or, perhaps more accurately, a theory of which the most-favored-nation interpretation is a version—the “equal value” theory. See generally Nelson Tebbe, The Principle and Politics of Equal Value, 121 Colum. L. Rev. 2397, 2459–60 (2021). Under this theory, a law isn’t generally applicable if it makes implicit value comparisons between religious and secular motives.97See Laycock, supra note 96, at 51. A legislature can’t say, for example, that a secular reason for noncompliance with a law is more important than a religious reason.98See, e.g., Duncan, supra note 95, at 875 (“The decision of the legislature to value secular conduct that is not expressly protected by the constitution more than analogous religiously motivated conduct is precisely the kind of unequal treatment that should be the minimum standard for constitutional protection of the free exercise of religion.”). This means that a law isn’t generally applicable with respect to religion if it exempts any form of secular conduct that detracts from its purpose as much as the prohibited religious conduct would.99See id. at 862. In other words, a law “lacks general applicability if it prohibits religious conduct while permitting secular conduct that undermines the government’s asserted interests in a similar way.”100Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021).
This is true even of laws that “treat some comparable secular businesses or other activities as poorly as or even less favorably than the religious exercise at issue.”101Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). “The question is not whether one or a few secular analogs are regulated,” but “whether a single secular analog is not regulated.”102Laycock & Collis, supra note 87, at 22. Commentators sometimes use the analogy of racial favoritism in employment.103See id. at 26. According to this argument, “the exercise of religion is entitled to be treated like the best-treated secular analog,” just as “[m]inority employees are entitled to be treated as well as the best-treated race, not merely as well as some other badly treated race.”104Id. Because the Constitution provides special solicitude for religious exercise,105See, e.g., Michael W. McConnell, Neutrality Under the Religion Clauses, 81 Nw. U. L. Rev., 146, 152 (1986) (arguing that, in light of the Free Exercise Clause, “[r]eligious conscience . . . stands on a different constitutional footing than other moral or political disagreements with governmental policy”). the argument goes, the government must place religious conduct in any preferred category it creates for analogous secular conduct.106This view also has received significant criticism. See, e.g., James M. Oleske, Jr., Free Exercise (Dis)honesty, 2019 Wis. L. Rev. 689, 729 (noting that “to constitutionally compel religious exemptions from even modestly underinclusive laws that bear no indicia of discriminatory intent has been critiqued by a number of commentators and rejected by several lower courts” (footnotes omitted)). First, as noted, Lukumi included language suggesting at least that, to defeat general applicability, substantial secular conduct must be permitted. See Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 543 (1993) (noting that a law isn’t generally applicable when the “underinclusion is substantial, not inconsequential”). Second, as discussed, the Court in Flores suggested that laws are generally applicable vis-à-vis religion unless they evince “religious bigotry.” City of Boerne v. Flores, 521 U.S. 507, 534 (1997). Finally, as some commentators point out, it would be strange to say a law “prohibits the free exercise of religion” because it contains a secular exemption but wouldn’t “prohibit the free exercise of religion” if it simply prohibited the secular conduct, too. See Lund, supra note 41, at 629 (arguing that the most-favored-nation approach makes general applicability “a matter of constitutional luck” that “depends on random, arbitrary factors”).
Imagine that a police department prohibits all officers from growing beards because it wants to foster a uniform appearance on the force. If a Muslim officer seeks to grow a beard for religious reasons, the police department can refuse so long as the no-beard rule is rationally related to a legitimate state interest. There’s no question that the rule covers all officers and evinces no discriminatory intent, so it is generally applicable under any theory.
But if the police department were to exempt officers with medical conditions from the no-beard rule, the rule would no longer be generally applicable under the most-favored-nation approach because a medical exemption would compromise the uniform appearance of the force as much as a religious exemption. By allowing the medical exemption, the department would have acknowledged that some individual concerns outweigh its interest in uniformity, and religion, as the “most favored” reason for an exception, therefore must be exempt, also. This was then-Judge Samuel Alito’s reasoning in Fraternal Order of Police Newark Lodge No. 12 v. City of Newark.107170 F.3d 359 (3d Cir. 1999). There, the Third Circuit concluded that the police department’s “decision to allow officers to wear beards for medical reasons undoubtedly undermine[d] the [d]epartment’s interest in fostering a uniform appearance through its ‘no-beard’ policy.”108Id. at 366. The Third Circuit technically didn’t apply the most-favored-nation interpretation because it decided the case on the ground that the medical exemption evinced discriminatory intent, which would have been unnecessary under the most-favored-nation approach. See id. at 365.
The Supreme Court adopted the most-favored-nation interpretation in a per curiam decision in Tandon v. Newsom.109141 S. Ct. 1294, 1296 (2021) (per curiam); see also Lund, supra note 1, at 857 (“In Tandon, the Court formally adopts the ‘most favored nation’ approach to the concept of general applicability—an approach we saw back in the Third Circuit’s decision in Fraternal Order.”); Tebbe, supra note 96, at 2420 (“Any ambiguity over whether the Court had adopted equal value was resolved in Tandon.”); Leah M. Litman, Disparate Discrimination, 121 Mich. L. Rev. 1, 21 (2022) (“Tandon appeared to embrace what some scholars called the ‘most-favored nation’ theory of discrimination for free exercise claims; under that theory, a law or policy ‘discriminates’ against religion if it treats a comparable nonreligious entity better than religious entities.”); Rothschild, supra note 25, at 1112–13 (observing that the Supreme Court recently adopted the most-favored-nation theory); Aaron Tang, Who’s Afraid of Carson v. Makin?, 132 Yale L.J.F. 504, 525–26 (2022) (“[T]he Court has recently shifted its free-exercise doctrine to afford greater protection to religious claimants by embracing a narrower conception of what counts as a generally applicable law, deeming even a single instance of more favorable treatment for a secular activity sufficient to trigger strict scrutiny.”). The Court reiterated the new standard in Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021), and Kennedy v. Bremerton School District, 142 S. Ct. 2407, 2422 (2022). In 2021, in the midst of the COVID-19 pandemic, California promulgated regulations restricting social interactions.110Tandon v. Newsom, 992 F.3d 916, 917 (9th Cir. 2021). The regulations limited private, in-home gatherings to three households.111Id. at 918. Two California residents seeking to hold in-home religious services with more than three households challenged the regulations on free-exercise grounds.112Tandon v. Newsom, 517 F. Supp. 3d 922, 946 (N.D. Cal. 2021). They argued that the regulations weren’t neutral and generally applicable because they didn’t apply the same three-household limit to commercial entities such as salons, retail stores, indoor restaurants, and train stations.113Id. at 963.
The district court denied the claimants’ request to preliminarily enjoin the regulations, explaining that “[t]he State’s private gatherings restrictions appl[ied] to all gatherings, whether religious or secular.”114Id. at 975. Only commercial gatherings were treated differently.115Id. at 963. The Court of Appeals for the Ninth Circuit declined to stay the regulations pending appeal, pointing to the district court’s further “conclu[sion] that the State reasonably distinguishe[d] in-home private gatherings from the commercial activity [the claimants] assert[ed] [was] comparable.”116Tandon, 992 F.3d at 925.
The Supreme Court reversed.117Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). It held that the regulations likely weren’t neutral and generally applicable because they applied a less restrictive standard to what the Court deemed “comparable” secular gatherings.118Id. at 1297. The Court explained that “whether two activities are comparable for purposes of the Free Exercise Clause must be judged against the asserted government interest that justifies the regulation at issue,” meaning “the risks various activities pose, not the reasons why people gather.”119Id. at 1296. According to the Court, “the Ninth Circuit did not conclude that [the comparator commercial] activities pose[d] a lesser risk of transmission than [the] applicants’ proposed religious exercise at home.”120Id. at 1297.
As discussed below,121See discussion infra Section III.B.1.a. Tandon rested on several questionable premises and created more uncertainty than it resolved. Even so, it cemented the most-favored-nation interpretation as the prevailing theory of general applicability under the Free Exercise Clause.122Lund, supra note 1, at 857. Laws aren’t generally applicable if they exempt secular conduct that is “comparable” to the prohibited religious conduct.123Fulton v. City of Philadelphia, 141 S. Ct. 1869, 1872 (2021) (reiterating the Supreme Court’s holding in Smith that a law also isn’t generally applicable “if it invites the government to consider the particular reasons for a person’s conduct by providing a mechanism for individualized exemptions”); see also Rothschild, supra note 25, at 1120–21.
B. Procedural Theory (the “How”)
How can parties establish or refute comparability? Ascertaining the substantive law isn’t enough. Just as important is determining the steps the parties must take to achieve their goals under the law. Anyone who watches legal dramas on television knows that criminal defendants are “innocent until proven guilty.”124See Coffin v. United States, 156 U.S. 432, 458–59 (1895); see also Stephen E. Arthur & Robert S. Hunter, Federal Trial Handbook: Criminal § 31:1 (2022–2023 ed.) (“The defendant in a criminal case is presumed to be innocent until proven guilty beyond a reasonable doubt.”); Law and Order: SVU: Blinded (NBC television broadcast Nov. 13, 2007) (“Everyone is innocent until proven guilty.”). Lawyers and nonlawyers alike understand that the prosecution bears the “burden of proof.”125See Taylor v. Kentucky, 436 U.S. 478, 485 (1978) (“[T]he rule about burden of proof requires the prosecution by evidence to convince the jury of the accused’s guilt.” (internal citations omitted)); Law and Order, SVU: True Believers (NBC television broadcast Nov. 2, 2011) (“[T]he burden of proof is still on the state.”). The burdens and presumptions, in turn, can mean everything. They tell us what happens when the evidence is lacking.126See Tex. Dep’t of Cmty. Affs. v. Burdine, 450 U.S. 248, 255 n.8 (1981) (“The word ‘presumption’ properly used refers only to a device for allocating the production burden.” (internal quotations and citations omitted)); Fed. R. Evid. 301 (“[T]he party against whom a presumption is directed has the burden of producing evidence to rebut the presumption.”). If the record is sparse or inconclusive, criminal defendants can’t be convicted.127Coffin, 156 U.S. at 459.
The burdens and presumptions in constitutional cases are just as impactful. Usually, courts presume that laws are constitutional.128See United States v. Carolene Prods. Co., 304 U.S. 144, 148, 152 n.4 (1938). A person challenging a law as unconstitutional thus bears the burden to demonstrate why.129See, e.g., Pac. States Box & Basket Co. v. White, 296 U.S. 176, 185 (1935). Almost a century ago, however, the Supreme Court announced a “narrower scope for operation of the presumption of constitutionality when legislation appears on its face to be within a specific prohibition of the Constitution.”130Carolene, 304 U.S. at 152 n.4.
Consider discrimination claims under the Fourteenth Amendment. When a law applies to some groups but not others, a person in a restricted group can challenge that law under the Equal Protection Clause.131U.S. Const. amend. XIV; see also Engquist v. Or. Dep’t of Agric., 553 U.S. 591, 601 (2008) (“Our equal protection jurisprudence has typically been concerned with governmental classifications that affect some groups of citizens differently than others.” (internal quotation marks and citations omitted)). If the law expressly distinguishes between people along “suspect class” lines, such as race, it is presumptively unconstitutional and triggers strict scrutiny.132See, e.g., Grutter v. Bollinger, 539 U.S. 306, 326 (2003) (“[A]ll racial classifications imposed by government must be analyzed by a reviewing court under strict scrutiny. This means that such classifications are constitutional only if they are narrowly tailored to further compelling governmental interests.” (internal quotation marks and citations omitted)). To establish a compelling justification for such classifications, the government essentially must prove that people in the restricted and unrestricted classes aren’t similarly situated.133See, e.g., Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 228 (1995); see also id. at 246 (Stevens, J., dissenting). That is, the government must demonstrate differences between the groups that provide compelling justifications for treating them differently.134See Grutter, 539 U.S. at 327 (noting that “strict scrutiny must take relevant differences into account,” and that “[w]hen race-based action is necessary to further a compelling governmental interest, such action does not violate the constitutional guarantee of equal protection so long as the narrow-tailoring requirement is also satisfied.” (internal quotation marks and citations omitted)).
Laws that classify based on innocuous or “benign” characteristics, by contrast, remain presumptively constitutional and trigger rational basis review.135See Mass. Bd. of Ret. v. Murgia, 427 U.S. 307, 312–14 (1976). For example, a law requiring police officers to retire by a certain age doesn’t violate the Equal Protection Clause if it rationally relates to a legitimate state interest.136Id. at 314–15. In such cases, it is the challenger who must prove that the restricted and unrestricted groups are similarly situated in all relevant respects.137See, e.g., FCC v. Beach Commc’ns, Inc., 508 U.S. 307, 314–15, (1993) (“On rational-basis review, a classification in a statute . . . [bears] a strong presumption of validity, and those attacking the rationality of the legislative classification have the burden to negative every conceivable basis which might support it.” (internal quotation marks and citation omitted)).
The presumption of constitutionality attaches in these cases because benign classifications necessarily appear in almost every law.138See David Bogen, Generally Applicable Laws and the First Amendment, 26 Sw. U. L. Rev. 201, 241 (1997) (“All laws classify.”); Collings & Barclay, supra note 4, at 466 (“Strict scrutiny . . . stands in sharp contrast to rational basis review, under which a law is presumed constitutional . . . .”). Equal-protection claimants alleging selective mistreatment under facially neutral policies likewise must show that others who are similarly situated have received better treatment.139See, e.g., Fincher v. Town of Brookline, 26 F.4th 479, 486 (1st Cir. 2022) (“To prove discrimination, a plaintiff can identify and relate specific instances where persons situated similarly in all relevant aspects were treated differently.” (internal quotation marks and citations omitted)); Hu v. City of New York, 927 F.3d 81, 91 (2d Cir. 2019) (“To prevail on such a claim, a plaintiff must prove that . . . the person, compared with others similarly situated, was selectively treated . . . .” (internal quotation marks and citations omitted)).
Some have argued that this shouldn’t be a threshold showing, but that an equal protection claimant should need to show only that two groups or individuals have been treated differently, at which point the nature of the claimed disparate treatment—and thus the level of scrutiny—will dictate which party must prove or disprove that the groups or individuals are similarly situated. See, e.g., Giovanna Shay, Similarly Situated, 18 Geo. Mason L. Rev. 581, 624 (2011) (“In sum, ‘similarly situated’ analysis is not a preliminary showing required to proceed to equal protection review.”); Scott E. Rosenow, Note, Heightened Equal-Protection Scrutiny Applies to the Disparate-Impact Doctrine, 20 Tex. J. on C.L. & C.R. 163, 172 (2015) (“Supreme Court practice reveals that the only required threshold for stating an equal-protection claim is that an official act treats one person differently than another.”). In the rational basis cases, where claimants bear the burden, laws are much more likely to survive than in strict scrutiny cases where the government bears the burden.140See Raphael Holoszyc-Pimentel, Note, Reconsidering Rational-Basis Review: When Does Rational Basis Bite?, 90 N.Y.U. L. Rev. 2070, 2071–72 (2015); City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 455 (2002) (Souter, J., dissenting).
Allocating the burdens and presumptions is just as important in free-exercise cases. In his concurrence in Lee, for example, Justice Stevens criticized the majority for requiring the government to carry “a heavy burden of justifying the application of neutral general laws to individual conscientious objectors.”141United States v. Lee, 455 U.S. 252, 262 (1982) (Stevens, J., concurring). He argued that a religious objector should “shoulder the burden of demonstrating that there is a unique reason for allowing him a special exemption from a valid law of general applicability.”142Id. Although Justice Stevens cared largely about what evidence would suffice to carry the relevant burdens, he saw the allocation of the burdens and presumptions as significant.143Id. at 263 nn.2–3. The government might not be able to demonstrate a sufficient justification for taxing the Amish claimant, but the claimant couldn’t prove the government lacked a sufficient justification.144Id. at 261–62.
In Smith, the Supreme Court latched on to Justice Stevens’s argument, explaining that “we cannot afford the luxury of deeming presumptively invalid, as applied to the religious objector, every regulation of conduct that does not protect an interest of the highest order.”145Emp. Div. v. Smith, 494 U.S. 872, 888 (1990). In Tandon, however, the Court criticized the Ninth Circuit for failing to “requir[e] the State to explain why it could not safely permit at-home worshipers to gather in larger numbers while using precautions used in secular activities.”146Tandon v. Newsom, 141 S. Ct. 1294, 1297 (2021) (per curiam) (emphasis added). In that sense, the Court seemed to say that California bore the burden to prove that the exempt secular conduct was incomparable to the requested religious exemption.
Akin to the concept of presumptions and burdens is the notion of deference. In free exercise and other constitutional cases, the outcome often turns on whether the government has a sufficient justification for its actions.147See, e.g., Collings & Barclay, supra note 4, at 455 (“[E]very prominent framework available for adjudicating religious liberty claims adopts justification as its nominal core requirement.”); John Witte, Jr., Joel A. Nichols & Richard W. Garnett, Religion and the American Constitutional Experiment 138 (5th ed. 2022) (“As every law student learns, the standard of review a court employs, or the level of deference it accords when reviewing official action, is very often outcome-determinative.”). This requires courts to determine whether the factual assertions underlying those justifications deserve deference.148See Collings & Barclay, supra note 4, at 498. The Supreme Court has held, for example, that epidemiological evidence supporting state-enacted vaccine mandates is entitled to some level of deference, because “the police power of a State must be held to embrace, at least, such reasonable regulations established directly by legislative enactment as will protect the public health and the public safety.”149Jacobson v. Massachusetts, 197 U.S. 11, 25 (1905) (upholding a smallpox vaccine mandate against a Due Process challenge). And when the government chooses to prosecute a particular person for a crime, that decision warrants a degree of “judicial deference” because prosecutorial discretion is a component of the executive power.150United States v. Armstrong, 517 U.S. 456, 465 (1996) (discussing the decision to prosecute a defendant, and noting that “to dispel the presumption that a prosecutor has not violated equal protection, a criminal defendant must present clear evidence to the contrary” (internal quotation marks and citations omitted)).
Early in the COVID-19 pandemic, the Supreme Court faced a string of free-exercise cases requiring it to determine how much deference is due in particular circumstances. In the first case, decided in May 2020, the Court declined to enjoin a California executive order limiting attendance at public gatherings and prohibiting “places of worship” from exceeding twenty-five percent capacity or one hundred attendees.151S. Bay United Pentecostal Church v. Newsom (South Bay I), 140 S. Ct. 1613, 1613 (2020). Concurring, Chief Justice John Roberts observed that “the Order exempt[ed] or treat[ed] more leniently only dissimilar activities, such as operating grocery stores, banks, and laundromats, in which people neither congregate in large groups nor remain in close proximity for extended periods.”152Id. (Roberts, C.J., concurring). His reasoning is worth quoting at length:
Our Constitution principally entrusts the safety and the health of the people to the politically accountable officials of the States to guard and protect. When those officials undertake to act in areas fraught with medical and scientific uncertainties, their latitude must be especially broad. Where those broad limits are not exceeded, they should not be subject to second-guessing by an unelected federal judiciary, which lacks the background, competence, and expertise to assess public health and is not accountable to the people.153Id. at 1613–14 (internal quotation marks and citations omitted).
In 2021, the Court changed course and enjoined a different California order insofar as it prohibited indoor worship services.154S. Bay United Pentecostal Church v. Newsom (South Bay II), 141 S. Ct. 716, 716 (2021). Chief Justice Roberts reiterated his view that “federal courts owe significant deference to politically accountable officials with the ‘background, competence, and expertise to assess public health,’”155Id. at 716 (Roberts, C.J., concurring) (quoting South Bay I, 140 S. Ct. at 1614). but concluded that California’s latest order “appear[ed] to reflect not expertise or discretion, but instead insufficient appreciation or consideration of the interests at stake.”156Id. at 717. Justice Elena Kagan dissented, arguing that the Court’s injunction “displace[d] the judgments of experts about how to respond to a raging pandemic,” and thereby required the State to “treat worship services like secular activities that pose[d] a much lesser danger.”157Id. at 720 (Kagan, J., dissenting).
Professor Cass Sunstein has endorsed the Court’s reversal of course in these COVID-19 cases, arguing that too deferential a standard might allow states to trounce constitutional rights in the face of crises.158Sunstein, supra note 4, at 235 (“To know the risks associated with various buildings and institutions, we need to answer several questions. Does it matter if people are together for ten minutes, or thirty, or sixty? How much does that matter? How much does proximity matter? If people speak together or sing together, what are the incremental risks? What happens, exactly, in drug stores and grocery stores, and how does it compare to what happens in churches and synagogues?”). Declining to accord such deference, he argues, “can reasonably be seen as a kind of anti-Korematsu—as a strong signal of judicial solicitude for constitutional rights and of judicial willingness to protect against discrimination, even under emergency circumstances in which life is on the line.”159Id. at 222 (footnote omitted). This admittedly “harsh” comparison has caught on, leading others to warn against applying the “doggedly deferential reasoning” of Korematsu v. United States160323 U.S. 214 (1944). In Korematsu, the Supreme Court upheld Fred Korematsu’s conviction for violating an order interning United States citizens of Japanese ancestry, based largely on deference to military authorities about national security issues. Id. at 218–19. to free-exercise cases.161Collings & Barclay, supra note 4, at 501.
The upshot is that the allocation of presumptions and burdens is often as important as ascertaining the substantive law. This is certainly true in free-exercise cases, where substance and methodology work in tandem. Unfortunately, courts are able to—and often do—manipulate the procedural levers to reach partisan outcomes.162See, e.g., Lund, supra note 1, at 859–60 (discussing “the risk of manipulation”); Rothschild, supra note 9, at 1071 (“Left unconstrained by indeterminate doctrine, judges adjudicating these cases have been able to decide them according to their political preferences.”); Tebbe, supra note 96, at 2482 (“Unfortunately, but unsurprisingly, [the prevailing general applicability doctrine is] subject to nonideal execution. Egalitarians who are attracted to the ideal therefore should pause before promoting it in practice, where it has been applied according to a particular politics.”). Perhaps because these maneuvers don’t involve the more controversial—and academically attractive—substantive questions (the “what” questions), courts are able to allocate presumptions and burdens implicitly, if the parties even bother to dispute them in the first place. Given the Supreme Court’s recent endorsement of a substantive theory—the most-favored-nation approach—answering the “how” questions is more important now than ever.
C. Normative Theory (the “Why”)
Why does general applicability matter? That is, why do courts care about religious and secular comparators? Deciding what general applicability means (the “what”) and the relevant procedural framework (the “how”) requires courts to make some version of these normative judgments. General applicability isn’t written in the First Amendment. It is a judicial standard adopted to preserve what the Smith Court saw as the most likely meaning of the Free Exercise Clause.163See Emp. Div. v. Smith, 494 U.S. 872, 878 (1990) (“It is a permissible reading of the text . . . to say that if prohibiting the exercise of religion . . . is not the object of [a law] but merely the incidental effect of a generally applicable and otherwise valid provision, the First Amendment has not been offended.”); see also McConnell, Revisionism, supra note 34, at 1115 (observing that, according to Smith, “the Free Exercise Clause does not conclusively resolve whether the provision requires exemptions from generally applicable laws”). When courts decide which substantive interpretation is “correct,” and what practical standards they “ought” to adopt, they need to consider the reasons the Court announced general applicability as a threshold test in the first place. When this Article discusses how courts have attempted to ascertain comparability, and when it proposes alternatives, it relies on the goals underlying the general applicability standard. This isn’t to say these goals are necessarily worthwhile, but merely that they undergird any application of the general applicability test.
The Court in Smith adopted general applicability as a threshold requirement largely to limit the universe of laws susceptible to free exercise challenges.164See, e.g., Lund, supra note 1, at 847 (calling Smith the beginning of “a jurisprudence whose master principle was that courts should not give religious exemptions”); Ira C. Lupu, The Trouble with Accommodation, 60 Geo. Wash. L. Rev. 743, 759 (1992) (arguing that the Court’s “central objectives” in Smith were “discouraging free exercise litigation and freeing courts from the federal constitutional obligation to weigh state interests against the impact upon religion worked by state policies”); Duncan, supra note 95, at 853 (criticizing Smith for endorsing this position). It decided the case within the familiar framework of constitutional presumptions, explaining that “we cannot afford the luxury of deeming presumptively invalid, as applied to the religious objector, every regulation of conduct that does not protect an interest of the highest order.”165Smith, 494 U.S. at 888. Given the diverse array of religious beliefs in the United States, the potential consequences of subjecting laws to strict scrutiny at the behest of individual religious observers, at least as a constitutional matter, concerned the Court.166Id. In formulating the neutrality and general applicability requirements, the Court thus invoked its warning from more than a century earlier that the Free Exercise Clause doesn’t “make the professed doctrines of religious belief superior to the law of the land.”167Id. at 879 (quoting Reynolds v. United States, 98 U.S. 145, 166–67 (1879)). It observed that applying strict scrutiny to all laws that burden religious exercise “would be courting anarchy.”168Id. at 888. The Court then listed “civic obligations of almost every conceivable kind” that it anticipated the neutrality and general-applicability barriers would insulate against attack.169Such as “compulsory military service,” “the payment of taxes,” “health and safety regulation such as manslaughter and child neglect laws,” “compulsory vaccination laws,” “traffic laws,” “social welfare legislation such as minimum wage laws,” “child labor laws,” “animal cruelty laws,” “environmental protection laws,” and “laws providing for equality of opportunity for the races.” Id. at 888–89. These weren’t assurances that laws in each category would always satisfy neutrality and general applicability,170Lukumi demonstrated as much when the Court determined that the animal cruelty law identified in Smith actually wasn’t neutral and generally applicable. See Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 564–66 (1993) (Souter, J., concurring). but their inclusion emphasizes the Court’s desire to limit the universe of laws susceptible to legitimate free-exercise challenges. In the words of one commentator, Smith signaled the beginning of “a jurisprudence whose master principle was that courts should not give religious exemptions.”171Lund, supra note 1, at 847.
Ultimately, this Article accepts that Smith remains the controlling law and that the most-favored-nation interpretation is the prevailing theory of general applicability. It contends that general applicability was meant to avoid free-exercise challenges to “civic obligations of almost every conceivable kind.”172Smith, 494 U.S. at 888. Keeping these considerations in mind, the next Part discusses the practical questions that remain in assessing general applicability through an analysis of religious and secular comparators.
III. General Applicability Comparators
Under the most-favored-nation interpretation, cases like the police department’s no-beard rule might be easy to resolve in favor of religious challengers for substantially the same reasons that the Third Circuit gave in Fraternal Order. But by declaring the medical exemption “undoubtedly” comparable to the requested religious exemption, the court elided the “how” questions that would become important in harder cases. It didn’t say, for example, how courts should make the relevant comparisons, who should bear the burdens, the level of proof required, or the level of deference that should be accorded the government’s stated reasons for granting secular exemptions. Put simply, it didn’t say how to determine when a secular exemption is “comparable” to the religious conduct at issue (i.e., how to determine whether a law “permit[s] secular conduct that undermines the government’s asserted interests in a similar way”).173Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021). The Supreme Court provided little helpful guidance in Tandon, and lower courts have continued to struggle with the standards.174Commentators also have noted the difficulty of creating a coherent standard. See, e.g., Rothschild, supra note 25, at 1121 n.59 (arguing that “a test predicated on identifying similarly situated secular and religious comparators, is unworkable”); Rothschild, supra note 4, at 285 (“Virtually every entity and activity will be both similar and dissimilar to other entities and activities depending on the level of generality at which one analyzes them, resulting in the possibility that almost any secular exception can give rise to a constitutional right to a religious exception.”).
Developing a coherent understanding of the modern doctrine is more important now than ever. The exemptions cases are socially divisive, and the amorphous nature of the law has led to outcomes that seem more political than principled.175See, e.g., Rothschild, supra note 9, at 1068 (noting that “in deciding free exercise challenges by religious plaintiffs to COVID‑19 lockdown orders, 0% of Democratic-appointed judges sided with religious plaintiffs, 66% of Republican-appointed judges sided with religious plaintiffs, and 82% of Trump-appointed judges sided with religious plaintiffs”); Tebbe, supra note 96, at 2482 (“Unfortunately, but unsurprisingly, [the prevailing general applicability doctrine is] subject to nonideal execution. Egalitarians who are attracted to the ideal therefore should pause before promoting it in practice, where it has been applied according to a particular politics.”).
This Article doesn’t wade into the lengthy debate on whether general applicability is the proper standard or whether the Supreme Court has properly interpreted the test.176See We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 273 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). Under Tandon, laws aren’t generally applicable if they exempt secular conduct while prohibiting religious conduct that would do “comparable” damage to the government’s interests.177Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). Rather, this Article proposes a practical implementation of the most-favored-nation theory and its “comparability” test that comports with the goals of the general-applicability requirement. This Part considers how courts can (A) ascertain the “government’s asserted interests,” and (B) determine whether religious exemptions would undermine those interests “in a similar way” to exempt secular conduct. It then (C) proposes how a court might best evaluate religious and secular comparators in a case resembling Smith.
A. Ascertaining the Relevant Interests
In determining whether religious and secular exemptions would undermine the government’s interests in similar ways, the first issue is ascertaining the relevant interests.
In Lukumi, the Supreme Court considered whether religious and secular animal killings were comparable vis-à-vis the City’s “professed desire to prevent cruelty to animals and preserve the public health.”178Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 545 (1993) (emphasis added). The Court in Tandon held that “whether two activities are comparable for purposes of the Free Exercise Clause” is “judged against the asserted government interest that justifies the regulation at issue.”179Tandon, 141 S. Ct at 1296 (emphasis added). And in Fulton v. City of Philadelphia,180141 S. Ct. 1868 (2021). it explained that religious and secular acts are comparable if they “undermine the government’s asserted interests in a similar way.”181Id. at 1877 (emphasis added). These statements suggest that the government’s characterization of its interests warrants some deference.182See Kane v. De Blasio, 19 F.4th 152, 166 (2d Cir. 2021) (assessing the general applicability of a mask mandate “[v]iewed through the lens of the City’s asserted interest in stemming the spread of COVID‑19”); see also Tebbe, supra note 96, at 2420–21 (“[Tandon] suggested that the Court would accept, at [the general applicability] stage of the analysis, the government’s ‘asserted’ reasons without asking whether they represented its actual reasons—in other words, that it was applying something like deferential review—but that may change as the Court thinks through this new doctrine.” (footnote omitted)).
On the other hand, there must be limits to this deference, since the “government’s asserted interests” are, in one sense, to accomplish exactly what its laws accomplish. Just as it is circular to say that “a law is always generally applicable to the objects to which it applies,”183Lund, supra note 41, at 640; see also Laycock & Collis, supra note 87, at 16 (“Every law applies to everything it applies to.”). it is circular to say that the interest justifying a law is the prevention of the things the law prevents.184Laycock, supra note 96, at 31 (arguing that it would be “entirely circular” to say “that the challenged law defined the relevant category and that the challenged law was generally applicable to that category”); Lund, supra note 1, at 854 (arguing that “what counts as the ‘rule,’ and what counts as the ‘exception,’ tacitly depend on the level of generality in how things are framed”). A police department with a rule against beards for nonmedical reasons can’t assert an interest in “preventing officers from growing beards for non-medical reasons.” The government must say why it prevents or requires certain conduct. The police department must explain why it forbids beards at all, and why it nevertheless allows beards in some circumstances.185A similar question is “what counts as the ‘rule,’ and what counts as the ‘exception.’” Lund, supra note 1, at 854. One commentator uses the following example to explain:
A city might write [a] policy in these terms: “There will be no in-person gatherings in buildings while the buildings are fit with improved ventilation systems. However, hospitals are exempt from this requirement.” That version of the policy would clearly implicate the equal value [or most-favored-nation] theory—there are regulated entities and exempted activities. But a city could just as easily rewrite that policy to accomplish the same results by saying, “There will be no in-person gatherings in buildings that do not offer life-saving care while the buildings are fit with improved ventilation systems.” That version of the policy regulates the same entities and exempts the same activities as the preceding policy with the exemption, which means it could raise the same questions about whether the government has devalued religious exercise. Litman, supra note 109, at 30.
As Professor Litman acknowledges, however, both of these hypothetical laws exempt hospitals. Although only the first law expressly exempts hospitals, the second law does so functionally. It merely uses narrowing language (applying only to “buildings that do not offer life-saving care”) instead of explicit language of exemption (“hospitals are exempt”). While the example shows why courts must carefully parse certain laws to determine whether they include exceptions, it doesn’t (at least, on its own) undermine the most-favored-nation framework. .
To further complicate things, many laws target interests that aren’t explicit in their text.186See, e.g., Wendy K. Olin, Note, Constitutional Survival Camp: What Are the Chances that the General Applicability Test Will Make It?, 68 S. Cal. L. Rev. 1029, 1051 (1995) (“For the Court to determine the object of a law under the general applicability test, it must necessarily consider the legislative motive underlying the law. To discern the true object of a law, it is necessary to know why the legislature felt the need to adopt it, and that object may run much deeper than that stated in the text of the law itself.”). And the government often has multiple reasons for its actions, some of which may be general and others specific. These complexities confounded the Ninth Circuit in 2021 when a school district in California mandated that students and staff receive a COVID-19 vaccine unless they qualified for a medical exemption.187Doe v. San Diego Unified Sch. Dist., 19 F.4th 1173, 1175–76 (9th Cir. 2021), reh’g en banc denied, 22 F.4th 1099 (9th Cir. 2022). The court denied a religious student’s request for an injunction pending appeal, explaining that the medical exemption served the “primary” interest for imposing the mandate—protecting student “health and safety.”188Id. at 1178, 1182. In the court’s view, religious claimants could take the vaccine without risking their health and safety, whereas the vaccine itself would compromise the health and safety of medical claimants.189Id. at 1177–79, 1181–82. A medical exemption therefore would undermine the government’s asserted interest less than a religious exemption.190Id. at 1178. The dissent argued that religious and medical exemptions would “pose identical risks to the government’s asserted interest,” which the dissent framed as “‘ensur[ing] the highest-quality instruction in the safest environment possible for all students and employees’ by preventing the transmission and spread of COVID-19.”191Id. at 1184 (Ikuta, J., dissenting). The dissent contended that allowing a religious claimant to remain unvaccinated would undermine the goal of stemming COVID-19 just as much as allowing a medical claimant to remain unvaccinated.192Id.
The case thus turned in large part on what the district’s “asserted interest” actually was.193Id. That is, an interest in protecting student “health and safety,” generally, or creating a COVID-19-safe environment, specifically.194Id. at 1184–85. In a footnote, the majority rejected the dissent’s “narrower formulation,” reasoning that “the interest the District emphasize[d] most frequently in the record with respect to the student vaccination mandate [was] protecting the ‘health and safety’ of students.”195Id. at 1178 n.5 (majority opinion). Whether California should have been permitted to assert such a broad interest is another issue,196See infra Section III.A.2. but the important point for now is that the court opted for the broader interest because it was the government’s most frequently asserted version.197Doe, 19 F.4th at 1178 n.5.
Whether one ultimately agrees with the majority or dissent, frequency was an odd focus. As noted, laws often have multiple goals. Claimants always can contest whether laws actually serve each asserted goal, or whether religious exemptions actually would undermine each goal. But claimants can’t simply ignore interests that their conduct would uniquely jeopardize on the ground that their conduct wouldn’t uniquely jeopardize other interests. As one commentator has argued, a “law [that] is found to be underinclusive for [one] purpose” may “still satisfy the test of general applicability so long as the law addresses at least one legitimate governmental purpose and is not underinclusive with respect to that purpose.”198Duncan, supra note 95, at 878. This makes sense, since the government should be able to decline a religious exemption that would uniquely undermine any legitimate state interest, even if the exemption wouldn’t undermine certain other interests.
Courts consider multiple governmental interests all the time in the parallel context of RFRA.199See, e.g., Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal, 546 U.S. 418, 426 (2006); Holt v. Hobbs, 574 U.S. 352, 363, 365 (2015). In Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal,200546 U.S. 418 (2006). the Court assessed the government’s evidence that allowing religious uses of a hallucinogenic drug would (1) cause health risks to religious users and (2) lead to diversion into the recreational market.201Id. at 426. And in Holt v. Hobbs,202574 U.S. 352 (2015). the Court considered whether banning beards in prisons would prevent inmates from (1) smuggling contraband and (2) disguising their identities.203Id. at 363, 365. It would make no sense to deem religious drug use comparable to, say, medical use, just because both uses would compromise people’s health, if religious use would lead to substantially more recreational diversion.204This is certainly plausible. Even assuming, for example, that religious uses of a hallucinogen would lead to the same number of overdoses as medical uses, medical administrations of a drug in physician’s office might be much harder to divert for recreational resale than unmonitored religious uses. See Gonzales, 546 U.S. at 437. Nor would it make sense to deem a dermatological exception to the no-beard rule comparable to a religious exception simply because they would equally undermine the contraband interest, if only the religious exception would undermine the disguise interest.205This too is plausible. The prison in Holt was concerned that inmates with beards could shave to quickly change their appearance and access restricted areas. See Holt, 574 U.S. at 365. Inmates with medical conditions preventing them from shaving likely would be unable to do so as easily.
Another key question is timing. When has the government asserted its interests? Do the asserted interests incorporate what lawmakers said before the law was enacted? Are the interests limited to what’s written in the law’s text? Can the government formulate its interests in response to requests for religious exemptions or even litigation? What if, after its enactment, a law provides more benefits than those originally intended?
The first question—whether to consider preenactment statements—concerns legislative history.206Whether courts can consider legislative history in Free Exercise cases has been the subject of debate in the Free Exercise context. See, e.g., Stormans, Inc. v. Selecky, 586 F.3d 1109, 1131 (9th Cir. 2009) (“While the analysis of legislative history is proper in the equal protection context, the law is unsettled regarding the scope of its consideration in the free exercise arena.”). In his Lukumi concurrence, Justice Scalia noted the practical difficulties of attempting to “determine the singular ‘motive’ of a collective legislative body.”207Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 558 (1993) (Scalia, J., concurring) (citations omitted); see also Bogen, supra note 138, at 236 (“Any discussion of the purpose of a law runs into the quandary that statutes are the products of multimember bodies whose members may have both different motives (why they want the statute enacted) and different goals (how they want the statute to apply) for the same vote.”). Justice Anthony Kennedy, by contrast, would have considered “the historical background of the decision under challenge.”208Church of the Lukumi, 508 U.S. at 540 (citing Village of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 267–68 (1977)). Justice Kennedy’s view prevailed decades later, at least with respect to neutrality, when he wrote for the Court in Masterpiece Cakeshop, Ltd. v. Colorado Civil Rights Commission209138 S. Ct. 1719 (2018). that “[f]actors relevant to the assessment of governmental neutrality include . . . ‘the legislative or administrative history, including contemporaneous statements.’”210Id. at 1731 (citation omitted).
At least some lower courts also have considered contemporaneous legislative statements in assessing general applicability. In another vaccine mandate case, the Second Circuit accepted New York’s asserted interests in seeking both to “prevent the spread of COVID-19 in healthcare facilities,” and to “reduce the risk of staffing shortages that [could] compromise the safety of patients and residents even beyond a COVID-19 infection.”211We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 285 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). The court emphasized that “[t]he State [had] identified these objectives in the Regulatory Impact Statement accompanying [its] emergency rulemaking,” and the claimants had “not point[ed] to any evidence suggesting that the interests asserted [were] pretextual or should otherwise be disregarded in the comparability analysis.”212Id. The court thus viewed New York’s official statement at the time of enactment as reflecting its “asserted” interests.213Id.
At the other end of the spectrum are instances where the government articulates an interest for the first time in responding to a free exercise challenge. Justice Neil Gorsuch has argued “that only the government’s actually asserted interests as applied to the parties before it count—not post-hoc reimaginings of those interests expanded to some society-wide level of generality.”214Does v. Mills, 142 S. Ct. 17, 20 (2021) (Gorsuch, J., dissenting). This principle is familiar in the context of heightened scrutiny, where “[t]he justification must be genuine, not hypothesized or invented post hoc in response to litigation.”215United States v. Virginia, 518 U.S. 515, 533 (1996); see also Shaw v. Hunt, 517 U.S. 899, 908 n.4 (1996) (“To be a compelling interest, the State must show that the alleged objective was the legislature’s ‘actual purpose . . . .’”); Miss. Univ. for Women v. Hogan, 458 U.S. 718, 730 (1982) (“[A]lthough the State recited a ‘benign, compensatory purpose,’ it failed to establish that the alleged objective is the actual purpose underlying the discriminatory classification.” (footnote omitted)).
But Justice Gorsuch’s insistence that “post-hoc reimaginings” don’t count in assessing general applicability doesn’t follow from the heightened scrutiny cases. The general-applicability inquiry is a threshold test to determine whether strict scrutiny applies, not a form of heightened scrutiny itself. If anything, the comparability test is more akin to rational basis review, where courts may consider not only the legislature’s asserted purposes, but also hypothetical justifications. This is because the Supreme Court “has never insisted that a legislative body articulate its reasons for enacting a statute.”216U.S. R.R. Ret. Bd. v. Fritz, 449 U.S. 166, 179 (1980).
The general-applicability test isn’t simply rational basis review, of course, so the analogy isn’t perfect. But the “plausible reasons” approach still seems more sensible than the “actually asserted reasons” standard Justice Gorsuch proposes. Smith requires courts to determine whether a law is or isn’t generally applicable,217Emp. Div. v. Smith, 494 U.S. 872, 893 (1990) (O’Connor, J., concurring). not whether it would be generally applicable if limited to the reasons the legislature actually gave at the time of enactment. As the Court noted in Lukumi, neutrality concerns “the object of a law,”218Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 533 (1993) (citing Smith, 494 U.S. at 878–79). whereas general-applicability concerns “[unequal] results.”219Id. at 542–43. While the government’s stated interests at the time of enactment may be relevant to whether a law is generally applicable, the actual legislative motive, to the extent it can be discerned, isn’t dispositive of the interests the law serves.
Say, for example, that to reduce overdoses from an addictive drug, a state prohibits the use of that drug except in designated, state-sanctioned facilities where users can be monitored and discharged when they become sober.220See, e.g., Overdose Prevention Centers Averted 59 Overdoses in First Three Weeks of Operation, NYC Health (Dec. 21, 2021), https://perma.cc/D37X-HUFF (announcing that “two Overdose Prevention Centers in New York City [were] the first publicly recognized sites to open in the United States”). Imagine a religious observer later seeks an exemption, arguing that limited sacramental use of the drug in churches wouldn’t lead to overdoses either. Say also that, since the statute’s enactment, illegal, recreational uses of the drug have led to numerous car accidents, whereas uses in state facilities haven’t resulted in accidents because users are monitored as they leave. Finally, assume that such monitoring wouldn’t be feasible in the sacramental context. Is the religious exemption comparable to the secular exemption just because both would similarly preserve the law’s original articulated purpose of stemming overdoses? In other words, are potential sacramental uses and state-facility uses analogous for general-applicability purposes, just because neither would lead to many overdoses? Or may a court consider the post-hoc accident-prevention rationale, too?
It seems that Justice Gorsuch would say only the overdose interest is relevant, because that’s the only reason the legislature gave for enacting the statute. But this is hard to justify, since there is a legitimate reason to exempt the secular conduct that doesn’t apply to the religious conduct. The law is generally applicable without a religious exemption, regardless of whether it would be generally applicable if it served only the original legislative purpose. And were post-enactment interests off limits, religious claimants could uniquely undermine benefits of laws simply because those benefits weren’t obvious or expressed at the time of enactment.
Ultimately, in assessing whether secular and religious exemptions would undermine the government’s asserted interests in similar ways, courts should consider interests asserted at any time. Those interests still must be legitimate, but it shouldn’t matter when the government first articulates them. Further, the government should be able to assert multiple interests. After all, under the modern approach, general applicability is an effects-oriented inquiry, not an effort to ascertain discriminatory intent.221See, e.g., Laycock & Collis, supra note 87, at 10 (arguing that “‘[t]he Free Exercise Clause “protect[s] religious observers against unequal treatment,”‘ regardless of targeting, motive, or an improper object” (second alteration in original) (footnote omitted) (quoting Church of the Lukumi, 508 U.S. at 542)).
Justice Gorsuch also has argued that the government shouldn’t be permitted to “expand” its asserted interests “to some society-wide level of generality,” such as protecting “‘health and safety.’”222Does v. Mills, 142 S. Ct. 17, 20 (2021) (Gorsuch, J., dissenting) (quoting Yellowbear v. Lampert, 741 F.3d 48, 57 (10th Cir. 2014)). As he wrote when he was a judge on the Tenth Circuit, “[t]he more abstract the level of inquiry, often the better the governmental interest will look,” resulting in “individual interest[s] appear[ing] the less significant.”223Yellowbear, 741 F.3d at 57 (internal quotation marks and citations omitted); see also Michael W. McConnell & Richard A. Posner, An Economic Approach to Issues of Religious Freedom, 56 U. Chi. L. Rev. 1, 53 (1989) (“Not infrequently, courts weigh the particular burden on free exercise in the individual case against a powerful but abstract governmental interest that makes the free exercise claim appear insignificant.”); J. Morris Clark, Guidelines for the Free Exercise Clause, 83 Harv. L. Rev. 327, 330–31 (1969) (“The purpose of almost any law can be traced back to one or another of the fundamental concerns of government: public health and safety, public peace and order, defense, revenue. To measure an individual interest directly against one of these rarified values inevitably makes the individual interest appear the less significant.”). Commentators have raised the inverse concern, too—that construing interests at a high level of generality can unfairly aid free-exercise claimants.224See, e.g., Lund, supra note 1, at 855 (“One can always get the religious claim to win if one raises the level of generality sufficiently.”); Rothschild, supra note 9, at 1102 (discussing how the level of generality determines whether the government interest or the individual interest is given more weight in the analysis).
But these concerns pertain to the compelling-interest test that arises under strict scrutiny, not to the modern formulation of the threshold general-applicability inquiry.225See Yellowbear, 741 F.3d at 57 (opining that “the most important question we confront in the compelling interest inquiry concerns the level of generality at which our analysis should proceed” (emphasis added)). The compelling-interest test in the free-exercise context historically has involved interest-balancing and value judgments.226See, e.g., Emp. Div. v. Smith, 494 U.S. 872, 882–83 (1990) (referring to strict scrutiny in the Free Exercise context as a “balancing test”); see also Clark, supra note 223, at 329 (recognizing the Supreme Court’s adoption of “a duty to weigh the damage to an individual’s freedom of conscience against the harm to the state’s legislative scheme” and contending that “[f]ew commentators . . . have doubted the need of the law to strike some balance between the two”); Oleske, supra note 106, at 712 (observing that the “consensus seems to be that [the Court] was applying a ‘balancing test’”). In assessing general applicability under the Tandon approach, however, the importance of the government’s interest is largely beside the point; the religious conduct underlying the requested exemption must be considered just as important as whatever secular conduct the existing exemption preserves.227See Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). The “individual interest” in religious exercise never runs the risk of appearing “less significant,”228Yellowbear, 741 F.3d at 57. because that conclusion is legally impermissible.229See Laycock & Collis, supra note 87, at 23 (arguing that the Free Exercise Clause prohibits value judgments); see also Collings & Barclay, supra note 4, at 467 (contending that Smith “openly and energetically eschewed any meaningful form of interest-balancing”). The only question is whether the secular and religious conduct would undermine the government’s interests, whatever those interests are, in similar ways. The government’s interests must have been asserted and must be “legitimate,”230See, e.g., Cent. Rabbinical Cong. of U.S. & Can. v. N.Y.C. Dep’t of Health & Mental Hygiene, 763 F.3d 183, 197 (2d Cir. 2014) (“A law is . . . not generally applicable if it is substantially underinclusive such that it regulates religious conduct while failing to regulate secular conduct that is at least as harmful to the legitimate government interests purportedly justifying it.” (emphasis added)). but that’s it. If a religious exemption undermines any asserted and legitimate governmental interest more than a secular exemption, the breadth of that interest hardly matters at the general-applicability stage.
Return to the police department’s no-beard rule in Fraternal Order.231Fraternal Ord. of Police Newark Lodge No. 12 v. City of Newark, 170 F.3d 359, 360 (3d Cir. 1999). Although the department framed its interest as preserving a uniform appearance, the scope of the asserted interest was essentially irrelevant. No matter what, exempting religious claimants wouldn’t have undermined that interest more than exempting medical claimants.232Leaving aside the discussion later in this Article about comparing the number of potential exemptions. See infra Section III.B.3.a. Say the department had framed the interest narrowly, such as “preserving a uniform appearance on the force.” Allowing officers to have beards for medical reasons obviously would have undermined that interest in a similar way to a religious exemption. Alternatively, say the department had framed its interest broadly, such as “protecting the community.” Allowing officers to grow beards for religious reasons might have undermined that interest by compromising the professional appearance of the police and thereby diminishing civilians’ respect for officers. But a medical exemption would have had the exact same effect. It still would have resulted in officers with beards. No matter the reason for the no-beard rule—broad or narrow—medical and religious exemptions would have undermined that rule in similar ways; namely, some officers would have beards.
Likewise, religious conduct that would uniquely undermine any of a law’s legitimate purposes, whether broad or narrow, isn’t analogous to secular conduct that would leave any of those purposes intact. If religious conduct would compromise a broad, abstract interest in “health and safety” more than the exempt secular conduct, the two activities aren’t comparable. Consider a murder statute that has an exception for self-defense.233See, e.g., United States v. Toledo, 739 F.3d 562, 566, 568 (10th Cir. 2014) (holding that a federal defendant charged with murder under 18 U.S.C. §§ 1111, 1153 was entitled to a jury instruction on self-defense). The statute undoubtedly serves an interest in “preventing killings.”234See William Sweet & Paul Groarke, Bentham, Jeremy: Classical School, in Encyclopedia of Criminological Theory 90–91 (Francis T. Cullen & Pamela Wilcox eds., 2010) (explaining that the fundamental purpose of law is to deter crime); Paul H. Robinson & Markus D. Dubber, The American Model Penal Code: A Brief Overview, 10 New Crim. L. Rev. 319, 324–25 (2007) (explaining how, from its inception, the Model Penal Code aimed to deter criminal conduct). But an exception for religiously-motivated killings235This isn’t entirely hypothetical. See, e.g., Rachel A. Ruane, Comment, Murder in the Name of Honor: Violence Against Women in Jordan and Pakistan, 14 Emory Int’l L. Rev. 1523, 1566–67 (2000) (“Some legislators and politicians in Jordan and Pakistan claim that the traditional practice of honor killing is a genuine manifestation of a community’s culture and religion and, as such, it may not be subjected to scrutiny from an international human rights perspective.”); Sarah Alsabti, Honor Killing and the Indigenous Peoples: Cultural Right or Human Right Violation?, 45 Denv. J. Int’l L. & Pol’y 457, 458–59 (2017) (noting that some religions practice human sacrifice). arguably would undermine that interest to the same degree as the self-defense exception, because both would allow certain killings. Were the purpose of the murder statute limited to this formulation, the most-favored-nation interpretation would favor allowing people to kill others for religious reasons.236See Eugene Volokh, A Common-Law Model for Religious Exemptions, 46 UCLA L. Rev. 1465, 1540–41 (1999) (arguing against the most-favored-nation interpretation because, for example, “[e]ven the bans on intentional homicide have exceptions,” and it is “perfectly proper” for “the legislature [to] value the exempted secular [killings] more highly” than religious killings (quoting Laycock, supra note 96, at 51)). Framing the relevant interest as “preventing unjustified killings”—as some have suggested237See Jonathan J. Kim & Eugene Temchenko, Constitutional Intolerance to Religious Gerrymandering, 18 Conn. Pub. Int. L.J. 1, 47 (2018) (arguing that religious claimants would not be entitled to exemptions from murder statutes under the most-favored-nation theory because “the exemptions to intentional homicide law only include justified homicide”).—wouldn’t avoid this issue. Rather, this framing would solicit the same impermissible value-judgment of whether religious killings are as “justified” as self-defense killings.238This formulation also would be circular. Essentially, it would define the purpose of the law as preventing the things the law prevents and permitting the things the law permits.
But our intuitions against this outcome can be reconciled with the most-favored-nation approach.239No matter how they would formulate the general applicability test, commentators tend to assume that murder statutes remain generally applicable despite including self-defense exceptions. See, e.g., Volokh, supra note 236, at 1540 (rejecting the most-favored-nation interpretation because “the presence of these exceptions doesn’t justify a general license for religious objectors to . . . kill”). A murder statute may serve the narrow interest of preventing killings, but it may also serve broader interests such as preserving a societal sense of peace.240See Goals and Outcomes, Center for Homicide Research, https://perma.cc/FNQ6-UEXV. Allowing religious killings would undermine this interest more than allowing self-defense killings, which can occur only when a reasonable person would perceive an imminent and grave danger.241See, e.g., Middleton v. McNeil, 541 U.S. 433, 434 (2004) (per curiam) (explaining that the crime of murder in California is the “unlawful killing of a human being . . . with malice aforethought,” and malice is negated if a person kills because of fear of imminent peril (alteration in original)); Conn. Gen. Stat. Ann. § 53a-19(a) (“[D]eadly physical force may not be used unless the actor reasonably believes that such other person is (1) using or about to use deadly physical force, or (2) inflicting or about to inflict great bodily harm.”); Wis. Stat. Ann. § 939.48(1) (“The actor may not intentionally use force which is intended or likely to cause death or great bodily harm unless the actor reasonably believes that such force is necessary to prevent imminent death or great bodily harm to himself or herself.”). Even when self-defense killings are permitted, people can’t legally kill one another based on purely subjective motives that a “reasonable person” wouldn’t share.242See, e.g., Middleton, 541 U.S. at 436–38. A religious exception to a murder statute would disrupt this sense of peace, because it would deprive citizens of the confidence that they won’t be killed for other people’s internal reasons. This doesn’t mean religious motives are less “valuable” in some abstract sense than self-defense motives; it merely recognizes that religious and secular killings aren’t comparable vis-à-vis a murder statute simply because they equally compromise its narrower purpose.
Finally, secular and religious practices that equally undermine a law’s broad purposes sometimes disproportionately undermine its narrower purposes. The Ninth Circuit recently elided this nuance when it enjoined a school’s enforcement of an antidiscrimination policy in Fellowship of Christian Athletes v. San Jose Unified School District Board of Education.24346 F.4th 1075 (9th Cir. 2022), vacated, No. 22-15827, 2023 WL 248320 (9th Cir. Jan. 18, 2023). The school revoked a Christian student organization’s official status when the organization required its leaders to attest that homosexuality is an “impure lifestyle” and that “marriage is exclusively the union of one man and one woman.”244Id. at 1082–84. The school’s policy prohibited official clubs from discriminating based on sexual orientation, gender identity, race, and certain other characteristics, which the Christian organization’s rule clearly violated.245Id. at 1084. The court held that the rule wasn’t generally applicable, however, because the school allowed affinity groups such as the “Big Sisters/Little Sisters” club to limit membership based on protected characteristics such as gender identity.246Id. at 1084, 1093. According to the court, the rule thus exempted “comparable” secular organizations from the antidiscrimination rule.247Id. at 1081.
In reaching its conclusion, the court appeared to treat the interest underlying the policy as singularly broad (e.g., preventing disparate treatment based on protected characteristics).248Id. at 1093. But the court didn’t consider that the policy might serve another, narrower interest (e.g., preventing stigmatization of certain protected classes of people).249See, e.g., Masterpiece Cakeshop, Ltd. v. Colo. C.R. Comm’n, 138 S. Ct. 1719, 1729 (2018) (“[S]igns saying ‘no goods or services will be sold if they will be used for gay marriages’ . . . would impose a serious stigma on gay persons.”); Obergefell v. Hodges, 576 U.S. 644, 678 (2015) (discussing “[d]ignitary wounds”); Nelson Tebbe, The Principle and Politics of Liberty of Conscience, 135 Harv. L. Rev. 267, 284 (2021) (“Importantly, one purpose driving civil rights law is to combat status degradation of persons vulnerable to structural injustice.”). Affinity groups supporting women, such as Big Sisters/Little Sisters, arguably don’t stigmatize those who can’t join. Such groups don’t exclude based on perceptions that nonmembers are inferior or “impure.”250See Fellowship, 46 F.4th at 1082–84. But a religious group that calls homosexuality an “impure lifestyle” arguably stigmatizes a protected class. It potentially undermines an antistigmatization purpose of the policy that the Big Sisters/Little Sisters club doesn’t. All this is to say, the school in the Ninth Circuit case might have enforced its policy exclusively against the religious group because the religious group uniquely undermined the school’s legitimate interest in preventing stigmatization and dignitary harms. The Ninth Circuit didn’t broach the issue.251See id. at 1088–99.
The Ninth Circuit’s decision also highlights Tandon’s impact on another category where the scope of an asserted interest is crucial: cases where religious views compete with antidiscrimination laws. Take Title VII of the Civil Rights Act, which prohibits discriminatory employment decisions but exempts businesses with fewer than fifteen employees.252See 42 U.S.C. § 2000e(b). A federal district court in Maryland recently deemed Title VII generally applicable because the claimant’s business and businesses with fewer than fifteen employees were not “reasonably comparable institutions.”253Doe v. Cath. Relief Servs., No. CCB-20-1815, 2022 WL 3083439, at *7 (D. Md. Aug. 3, 2022), reh’g granted in part, 2023 WL 155243 (D. Md. Jan. 11, 2023). A federal district court in Texas, by contrast, held that Title VII isn’t generally applicable because the small-business exception undercuts the governmental “interest in eradicating all forms of discrimination” just as much as a religious exception.254Bear Creek Bible Church v. EEOC, 571 F. Supp. 3d 571, 613 (N.D. Tex. 2021).
What explains the different results in the Maryland and Texas cases? One possible answer is interest-framing. The EEOC in the Texas case apparently articulated only one, broad purpose for Title VII—”eradicating all forms of discrimination.”255Id. While this may be one of Title VII’s goals,256Although there are many “forms of discrimination” (e.g., of unprotected groups) that the statute doesn’t even purport to prohibit. however, this framing seems unnecessarily general. Title VII also helps provide equal access to employment for members of protected groups, protects people in those groups from experiencing discrimination and the attendant dignitary harms, and reduces economic inequality.257Andrew Koppelman, Gay Rights, Religious Accommodations, and the Purposes of Antidiscrimination Law, 88 S. Cal. L. Rev. 619, 627–29 (2015); see Tebbe, supra note 249, at 284 (“Importantly, one purpose driving civil rights law is to combat status degradation of persons vulnerable to structural injustice.”).
While the small-business exception certainly undermines the equal-access purpose to a degree,258See, e.g., Colin A. Devine, A Critique of the Secular Exceptions Approach to Religious Exemptions, 62 UCLA L. Rev. 1348, 1378 (2015) (“[Such] laws are underinclusive with respect to the state’s interest in eliminating discrimination . . . .”); Duncan, supra note 95, at 880 (“[In the context of housing discrimination,] the allowance of secular exemptions ‘is substantial evidence that religious exemptions would not threaten the statutory scheme.’” (quoting Laycock, supra note 96, at 50)). it may not generate as much discrimination and inequality as a religious exception that applies to employers of any size. Allowing the occasional small business to discriminate undermines Title VII’s goals to a degree, but allowing companies with thousands of employees to discriminate on religious grounds could be substantially more disruptive. This rationale persuaded the court in the Maryland case.259Doe v. Cath. Relief Servs., No. CCB-20-1815, 2022 WL 3083439, at *7 (D. Md. Aug. 3, 2022) (“CRS offers no . . . relatively close comparison, instead asking this court to find that — for the purposes of laws against employment discrimination — businesses with 15 or fewer employees, the United States government, and bona-fide tax-exempt private membership clubs are secular activities comparable to the religious activity of a social services nonprofit with over 7000 employees.”), reh’g granted in part, 2023 WL 155243 (D. Md. Jan. 11, 2023). Further, a religious exception likely would cause more otherwise-protected employees and prospective employees to experience discrimination. This is because job applicants can determine whether businesses have fewer than fifteen employees, whereas they likely can’t easily learn the religious views of each prospective employer. They are thus more likely to seek employment from, and face discrimination by, employers invoking the religious exception.
Ultimately, Justice Gorsuch’s focus on the breadth of the interests might be useful in a balancing test where value judgments are required. But that focus misses the point in the modern general-applicability context, where such judgments are impermissible.260See Duncan, supra note 95, at 875 (“The decision of the legislature to value secular conduct that is not expressly protected by the constitution more than analogous religiously motivated conduct is precisely the kind of unequal treatment that should be the minimum standard for constitutional protection of the free exercise of religion.”); see also Tebbe, supra note 96, at 2420 (“Any ambiguity over whether the Court had adopted equal value was resolved in Tandon.”).
Despite missing the mark with respect to identity and scope, Justice Gorsuch was right about one thing. Under the most-favored-nation theory, the question is whether the religious and secular activities would similarly undermine the purpose of the law itself, not whether the exempt secular activity itself serves a valuable countervailing purpose.261See Tebbe, supra note 96, at 2399. If a COVID‑19 restriction on building occupancy exempts grocery stores but not churches, the question is whether grocery stores spread COVID‑19 more than churches, not whether grocery stores further some other interest that churches don’t (e.g., providing sustenance).262The Court’s COVID‑19 cases before Tandon posed an issue like this, but the justices who wrote seemed to treat the laws in those cases as non-neutral because they expressly applied a standard to “houses of worship.” See Roman Cath. Diocese of Brooklyn v. Cuomo, 141 S. Ct. 63, 66 (2020) (per curiam); S. Bay United Pentecostal Church v. Newsom (South Bay I), 140 S. Ct. 1613, 1613 (2020); S. Bay United Pentecostal Church v. Newsom (South Bay II), 141 S. Ct. 716, 717–18 (2021) (Gorsuch, J., concurring). Taken to the extreme, this means that a secular exception that would save lives is just as likely to warrant strict scrutiny as a secular exception that would merely prevent minor inconveniences.263This sometimes leads to counterintuitive results. Exempting grocery stores from an occupancy limit might trigger strict scrutiny if grocery stores would generate as many COVID‑19 cases as churches, whereas exempting salons might not trigger strict scrutiny if salons would generate fewer cases. Again, this Article’s purpose isn’t to question the most-favored-nation theory, but to guide its application. If a vaccine mandate exempts people who would die from the vaccine, the most-favored-nation interpretation says that any sincerely-asserted religious harm must be considered equally dire.264Cf. Laycock & Collis, supra note 87, at 23 (“The transitive law applies; if medicine is more important than uniformity, and uniformity is more important than religion, then medicine is more important than religion. Whether explicit or implicit, that is the value judgment that the Free Exercise Clause prohibits.”). Religion is always as important as the exempt secular conduct.265This issue was ostensibly complicated in the “essential business” cases, where states allowed businesses they deemed “essential” to operate more fully than “houses of worship.” See, e.g., South Bay I, 140 S. Ct. at 1614 (Kavanaugh, J., dissenting); South Bay II, 141 S. Ct. at 718 (Gorsuch, J., dissenting). .
When the Ninth Circuit declined to exempt religious claimants from the school vaccine mandate because the comparator medical exception furthered a crucial health and safety interest,266Doe v. San Diego Unified Sch. Dist., 19 F.4th 1173, 1178–79 (9th Cir. 2021), reh’g en banc denied, 22 F.4th 1099 (9th Cir. 2022). the court arguably erred by “focus[ing] on the reasons for the [medical] exemption rather than the asserted interest that justifie[d] the mandate.”267Id. at 1185 (Ikuta, J., dissenting). The medical exemption undermined the primary interest for the mandate because it increased the number of unvaccinated people, and thus increased the viral risk, even though it furthered a different health and safety interest—namely, protecting people from the more severe medical consequences of taking a contraindicated vaccine.268Id. In that sense, the medical exception allowed the very harm the vaccine mandate was meant to curtail, but that compromise was worth it to further a countervailing interest in preventing adverse reactions. The case appeared more challenging because the interest underlying the medical exception also involved health and safety.269Id. at 1178 n.5 (majority opinion). But the health goals of the medical exception had to be treated as equal to the religious interest underlying the claimants’ requested religious exemption. The question should have been whether the medical and religious exemptions similarly undermined the health and safety goals underlying the vaccine mandate itself.
This isn’t to say that the Ninth Circuit got the ultimate answer wrong. If, for example, the vaccine would have been less effective for those with medical contraindications, then a religious exemption would have undermined the effectiveness of the mandate more than the medical exception. And, as discussed below, religious exemptions appeared likely to uniquely compromise the purpose of the mandate insofar as they would have been more numerous and clustered.270See infra notes 417–420 and accompanying text.
Ultimately, states should be able to assert whatever interests they want, since they are the entities that enact the rules in the first place. Those interests must be “legitimate,”271See, e.g., Cent. Rabbinical Cong. of U.S. & Can. v. N.Y.C. Dep’t of Health & Mental Hygiene, 763 F.3d 183, 197 (2d Cir. 2014) (“A law is . . . not generally applicable if it is substantially underinclusive such that it regulates religious conduct while failing to regulate secular conduct that is at least as harmful to the legitimate government interests purportedly justifying it.” (emphasis added)). but that’s it.272See, e.g., Collings & Barclay, supra note 4, at 472 (“Whatever the formula, laws rarely fail at [the legitimacy] step. In modern liberal democracies, governments do not often pursue openly an obviously improper end.”). States should also be able to identify multiple interests, since laws often serve multiple goals. And states should not need to have asserted those interests at any particular time before the litigation, since laws don’t always list everything they are designed to accomplish, and additional benefits may emerge after enactment. On the other hand, the asserted interests must be those that underlie the laws themselves, not the interests that underlie the secular exemptions.
One point is worth reiterating. Recognizing that the government can assert as many interests as it wants at any level of generality isn’t acknowledging a loophole for savvy litigators. The government can’t assert interests that its laws don’t actually serve.273See, e.g., id. If it did, exempting religious conduct wouldn’t undermine those interests at all, and would thus be comparable to secular exceptions that also wouldn’t undermine those interests. And if a law actually does serve multiple purposes, and a religious exemption actually would uniquely undermine one of those purposes, there is good reason to deny the religious exemption. This may mean most laws that include secular exceptions are generally applicable. But that’s exactly what Smith suggests.274See supra Section II.C.; infra Section III.B.1.b.
B. Proving or Disproving Comparability
Once courts ascertain the relevant governmental interests, how can they determine whether the exempt secular conduct and proposed religious conduct would undermine those interests in similar ways? Should courts defer to the government’s assertions of what activities are comparable? Who bears the burden to show or disprove comparability? Is a requested religious exemption comparable to a secular exception if the specific claimant’s religious exercise would undermine the purpose of the law no more than a single secular exemption? Or should courts assess whether exempting all potential religious claimants would undermine a law’s purposes more than granting the secular exemption to all eligible applicants? Courts have given inconsistent answers to these questions, both explicitly and implicitly. This Section considers the issues of (1) Burdens and (2) Evidence.
This Section discusses (a) how courts have allocated the burden of proving or disproving comparability, and (b) a possible methodology that would create consistency and bring the doctrine into line with the purposes of the general applicability standard.
a. Burdens in Practice
In reversing the lower courts, the majority in Tandon reached two questionable conclusions, even under the most-favored-nation interpretation of general applicability that the Court endorsed.275See generally Tandon v. Newsom, 141 S. Ct. 1294 (2021) (per curiam). First, in criticizing the Ninth Circuit for failing to “requir[e] the State to explain why it could not safely permit at-home worshipers to gather in larger numbers while using precautions used in secular activities,”276Id. at 1297. the Court ignored that the State provided precisely this explanation. As the Ninth Circuit noted:
[T]he district court found that the State reasonably concluded that when people gather in social settings, their interactions are likely to be longer than they would be in a commercial setting; that participants in a social gathering are more likely to be involved in prolonged conversations; that private houses are typically smaller and less ventilated than commercial establishments; and that social distancing and mask-wearing are less likely in private settings and enforcement is more difficult.277Tandon v. Newsom, 992 F.3d 916, 925 (9th Cir. 2021).
The district court, in turn, had relied on evidence the State submitted, and the religious claimants didn’t “dispute any of [its] findings” on appeal.278Id. Justice Kagan pointed this out in her dissent, arguing that “the Court ha[d] no warrant to ignore the record.”279Tandon, 141 S. Ct. at 1298 (Kagan, J., dissenting). In a previous free exercise case where the district court had reached the opposite conclusion, Justice Alito—a member of the Tandon per curiam—had argued that “the Court of Appeals [likely] failed to accord the District Court’s findings appropriate deference.” See Stormans, Inc. v. Wiesman, 136 S. Ct. 2433, 2438 (2016) (Alito, J., dissenting from denial of certiorari).
But perhaps more importantly, the Supreme Court muddled the threshold question of who bore the burden to show whether the commercial businesses were comparable to the private religious gatherings at issue.280Tandon, 141 S. Ct. at 1297. The Ninth Circuit had held that the religious claimants bore the “burden of showing that the regulation trigger[ed] strict scrutiny by regulating religious activities more strictly than comparable secular activities.”281Tandon, 992 F.3d at 927. It considered, for example, “whether [the claimants] ha[d] shown that rallies and protests [were] comparable secular activities.”282Id. at 928. And it noted that the claimants had “not explain[ed] why salons should be considered analogous secular conduct” and had “point[ed] to nothing in the record to support that comparison.”283Id. at 925 n.8. The Ninth Circuit thus placed the relevant burden on the religious claimants, so any evidentiary deficiencies supported—not undermined—general applicability.
The Supreme Court, by contrast, seemed to place the burden on the government, criticizing the Ninth Circuit for failing to “requir[e] the State to explain why it could not safely permit at-home worshipers to gather in larger numbers.”284Tandon, 141 S. Ct. at 1297 (emphasis added). The Court then summarily concluded that “California treats some comparable secular activities more favorably than at-home religious exercise.”285Id. The Court thus implicitly rejected the Ninth Circuit’s allocation of the burdens.
Lower courts, both before and after Tandon, also have inconsistently allocated the burdens of showing or disproving comparability.286Recent scholarship shows that courts in other countries also differ as to whether they “require governments to shoulder the evidentiary burden of establishing the relevant risk of allowing the religious exercise” or instead “deferentially assume such risk on the government’s ipse dixit.” Collings & Barclay, supra note 4, at 489. Some, like the Ninth Circuit in Tandon,287Tandon, 992 F.3d at 927. have placed the burden on religious claimants, at least implicitly. For example, the Second Circuit seemed to do so in a 2021 case involving New York’s emergency COVID-19 vaccination requirement for healthcare workers.288We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 273 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). In that case, various healthcare workers subject to the mandate sued, arguing “that because the State ha[d] afforded a medical exemption to its requirement, the Free Exercise Clause require[d] the State also to afford a religious exemption.”289Id. The court began by explaining that the claimants bore the burden to disprove neutrality and general applicability.290Id. at 281. Although it noted that the State had ”presented evidence that raise[d] the possibility that the exemptions [were] not comparable in terms of the ‘risk’ that they pose[d],” it denied the plaintiffs’ request for a preliminary injunction because the “sparse” record did “not support a conclusion that [the] [p]laintiffs ha[d] borne their burden of demonstrating that the medical exemption . . . and the religious exemption sought [were] likely comparable.”291Id. at 286, 288. This suggests that the claimants bore the burden of showing comparability, because, otherwise, the “sparse” record easily could have compelled the opposite conclusion.292Because the case arose at the preliminary injunction stage, the claimants needed to show they were substantially likely to succeed on the merits. But the court still had to decide which party bore the burdens on the underlying Free Exercise questions. See Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal, 546 U.S. 418, 429 (2006) (“[T]he burdens at the preliminary injunction stage track the burdens at trial.”); Ashcroft v. ACLU, 542 U.S. 656, 666 (2004) (holding that where “the Government bears the burden of proof on the ultimate question,” the claimants “must be deemed likely to prevail” at the preliminary injunction stage unless the evidence supports the government’s position). Because the court concluded that the religious claimants bore the burden to prove comparability, a sparse record meant that the claimants weren’t likely to succeed on the merits. See We The Patriots, 17 F.4th at 281 (“Plaintiffs must show that they are likely to succeed on their claim that [the vaccine mandate] is not a neutral or generally applicable rule. If they succeed at that step, the burden shifts to the State to show that it is likely to succeed in defending the challenged Rule under strict scrutiny.”). If the government had borne the burden of proving incomparability, by contrast, the sparse record would have favored the claimants.
Likewise, the Ninth Circuit reaffirmed its position, post-Tandon, that religious claimants bear the burden of showing secular comparators. In late 2022, the court affirmed the dismissal of a religious therapist’s challenge to a Washington law against practicing “conversion therapy, which seeks to change an individual’s sexual orientation or gender identity.”293Tingley v. Ferguson, 47 F.4th 1055, 1063–64 (9th Cir. 2022). The therapist argued that “gender-affirming therapy,” which the law allowed, similarly undermined the state’s interests by causing “the very types of psychological harms Washington [said] it want[ed] to eliminate by prohibiting conversion therapy.”294Id. at 1088–89 (internal quotation marks omitted). The court rejected this purported comparator because the law addressed “scientifically documented increased risk[s] of suicide and depression,” and the therapist provided only “anecdotal reports of ‘regret’” from gender-affirming therapy.295Id. at 1089.
The Tenth Circuit likewise placed the burden on the religious claimant in 303 Creative LLC v. Elenis2966 F.4th 1160 (10th Cir. 2021).—a case the Supreme Court recently reviewed on free-speech grounds297The Supreme Court granted certiorari on “[w]hether applying a public-accommodation law to compel an artist to speak or stay silent violates the Free Speech Clause of the First Amendment.” 303 Creative LLC v. Elenis, 142 S. Ct. 1106 (2022) (mem.).—where a website designer refused298As the court noted, the designer “ha[d] not yet offered wedding website services,” but merely intended to. 303 Creative, 6 F.4th at 1172. The court reached the merits on the ground that the designer had shown “a credible fear of prosecution.” Id. at 1173. to create websites for same-sex marriages in violation of a Colorado antidiscrimination law.299Id. at 1170. The court rejected the free exercise challenge, holding that the designer bore the burden to show that the law wasn’t neutral and generally applicable.300Id. at 1186 & n.8. As to general applicability, the court noted that the designer had “provide[d] no examples where Colorado permitted ‘secular-speakers’ to discriminate against LGBT consumers,” and thus had “fail[ed] to show that Colorado disfavor[ed] similarly-situated ‘religious-speakers.’”301Id. at 1186.
Finally, in a pre-Tandon case, the Northern District of New York placed the burden of proving comparability on religious plaintiffs, the Association of Jewish Camp Operators and various parents, who had asked the court to enjoin a COVID-19 restriction banning children’s overnight camps.302Ass’n of Jewish Camp Operators v. Cuomo, 470 F. Supp. 3d 197, 220 (N.D.N.Y. 2020). The court explained that although the restriction exempted college dormitories, and some overnight camps used similar facilities, the plaintiffs had “provided no evidence regarding the sleeping arrangements for the overnight camps themselves,” and “fail[ed] to mention whether individuals would be required to wear masks during camp activities, including sleeping.”303Id. The court thus concluded “that none of the specific explicit exemptions” that the claimants identified were “sufficiently comparable to permitting an overnight camp for the purposes of a general-applicability analysis.”304Id. at 222. Other courts appear to have reached similar results. For example, in Doe v. Catholic Relief Services, the court deemed Title VII generally applicable because although the claimant “pointed to secular institutions not covered by Title VII,” it failed to “point to reasonably comparable institutions.” No. CCB-20-1815, 2022 WL 3083439, at *7 (D. Md. Aug. 3, 2022). In Leone v. Essex County Prosecutor’s Office, a state prosecutor’s office denied an employee’s “request to work from home indefinitely for religious reasons,” citing its office scheduling policy. No. 21-12786 (SDW) (ESK), 2021 WL 4317240, at *1 (D.N.J. Sept. 23, 2021). A judge in the District of New Jersey concluded that the policy was generally applicable and thus didn’t violate the Free Exercise Clause. Id. at *6. In a footnote, the judge rejected the employee’s assertion “‘on information and belief’ that prosecutors ha[d] been allowed to work from home occasionally,” in part because the employee “ha[d] not provided any specific examples.” Id. at *6 n.13. See also Does 1-11 v. Bd. of Regents of Univ. of Colo., No. 21-CV-2637, 2022 WL 252320, at *6 (D. Colo. Jan. 27, 2022) (requiring religious claimants to show comparability); Swanson ex rel. Swanson v. Guthrie Indep. Sch. Dist. No. I-L, 135 F.3d 694, 698 n.3 (10th Cir. 1998) (noting that Free Exercise claimants seeking an exemption from a public school’s in-person attendance requirement did “not appear to have made a sufficient evidentiary showing” that the policy’s secular exemptions undermined the school’s interest in retaining funding).
Other pre-Tandon cases appear to reflect the opposite allocation of the burdens. For example, in Central Rabbinical Congress of the United States & Canada v. N.Y.C. Department of Health & Mental Hygeiene,305763 F.3d 183 (2d Cir. 2014). the Second Circuit applied strict scrutiny to a city ordinance that prohibited performing oral suction during circumcisions, a practice that allegedly had been spreading infections.306Id. at 198. Orthodox Jewish organizations in New York asked the court to preliminarily enjoin the ordinance for violating their free-exercise rights.307Id. at 192. After determining that the ordinance wasn’t neutral, the Second Circuit considered general applicability as an alternate basis for its holding.308Id. at 196. It explained that, on “the sparse record at th[e] preliminary stage,” it could not “conclude that [the ordinance was] generally applicable.”309Id. at 197. Unlike in the vaccine mandate case, therefore, the “sparse” record supported—not undermined—a finding of comparability.
Likewise, in Blackhawk v. Pennsylvania,310381 F.3d 202 (3d Cir. 2004). the Third Circuit appeared to place the burden of disproving comparability on the government.311Id. at 211. Pennsylvania’s Game and Wildlife Code required people seeking to possess certain animals to obtain permits.312Id. at 205. A Native American religious claimant sought an exemption to possess black bears without paying the permitting fee, citing his religious beliefs.313Id. At the summary judgment stage, the Third Circuit agreed with the claimant that the Game and Wildlife Code wasn’t generally applicable, in part because it contained categorical exceptions for zoos and circuses.314Id. at 211. Pennsylvania had asserted two “main interests” behind the law: earning money through permit fees, and discouraging the keeping of wild animals in captivity where captivity would not benefit the wildlife populations.315Id. Oddly, Pennsylvania doesn’t appear to have asserted an interest in safety. Id. An exemption for private ownership of wild animals by untrained religious claimants seems more likely to undermine a safety interest than exemptions for zoos and circuses, which presumably have staff trained to manage dangerous animals. The court first concluded that “[t]he state’s interest in raising money [was] undermined by any exemption, and the Commonwealth ha[d] not argued, much less shown, that religiously based exemptions, if granted, would exceed the exemptions for qualifying zoos and circuses.”316Blackhawk, 381 F.3d at 211. The court also observed that Pennsylvania “ha[d] not explained how circuses, whether nationally recognized or not, provide tangible benefits for animals living in the wild in Pennsylvania.”317Id. The court thus placed the burden on the state to disprove that the secular exemptions and requested religious exemption would similarly undermine the purposes of the law.
b. Burdens in Theory
These cases highlight that the burden question is far from resolved, even after Tandon. Although many courts agree that religious claimants bear the burden to disprove neutrality and general applicability,318See, e.g., Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2421–22 (2022) (“Under this Court’s precedents, a plaintiff may carry the burden of proving a free exercise violation in various ways, including by showing that a government entity has burdened his sincere religious practice pursuant to a policy that is not ‘neutral’ or ‘generally applicable.’”); 303 Creative LLC v. Elenis, 6 F.4th 1160, 1186 n.8 (10th Cir. 2021) (“[I]t is Appellants’ burden to show, at the very least, a triable issue of material fact that [the statute] is not neutral or generally-applicable.”), cert. granted in part, 142 S. Ct. 1106 (2022); We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 281 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). others seem to require the government to prove that religious and secular exemptions aren’t comparable.319See e.g., Tandon v. Newsom, 141 S. Ct. 1294, 1296–97 (2021) (per curiam); Cent. Rabbinical Cong. of U.S. & Can. v. N.Y.C. Dep’t of Health & Mental Hygiene, 763 F.3d 183, 197 (2d Cir. 2014); Blackhawk, 381 F.3d at 211. The lingering uncertainty warrants a normative and doctrinal question: Who should bear the burden to show or disprove comparability?
To begin with a relatively uncontroversial point, the government should bear at least a limited explanatory burden. Once a claimant shows that a law limits a particular religious practice and has at least one secular exception, the government should need to explain the reasons for the law and why it can’t simply extend the exception to the claimant’s religious exercise. Lawmakers presumably know why their laws contain secular exceptions but not religious exceptions. And if they can’t explain, they shouldn’t object to simply granting claimants’ requests. In any event, religious observers shouldn’t need to guess at legislative goals. In Blackhawk, therefore, the Third Circuit was right to conclude that religious and secular exemptions similarly undermined Pennsylvania’s monetary interests to the extent “the Commonwealth ha[d] not argued . . . that religiously based exemptions, if granted, would exceed the exemptions for qualifying zoos and circuses.”320Blackhawk, 381 F.3d at 211 (emphasis added). It also was right to emphasize that the state “ha[d] not explained how circuses, whether nationally recognized or not, provide tangible benefits for animals living in the wild in Pennsylvania.”321Id. (emphasis added). If the State truly didn’t offer any reason for treating religious conduct differently, it failed to carry its explanatory burden.322Ostensibly, this is all the Supreme Court did in Tandon. As noted, the Court criticized the Ninth Circuit for failing to “requir[e] the State to explain why it could not safely permit at-home worshipers to gather in larger numbers.” Tandon, 141 S. Ct. at 1297 (emphasis added). But the Court appears to have meant “explain” in the evidentiary sense rather than the literal sense, since the State literally “explained” in its Ninth Circuit brief:
Private gatherings in which individuals “have social connections to one another and are coming together for the purpose of being together” generally bring people “together for a longer time” and are more likely to involve prolonged conversations, both of which increase the risk of transmission . . . . In addition, private homes tend to have poorer ventilation, and distancing and mask-wearing are less likely in private gatherings . . . . Finally, restrictions are more difficult to enforce in private gatherings . . . because, as the motions panel observed, “ensuring public-facing businesses comply with  regulations is a fundamentally different task from regulating conduct in private homes, which government authorities cannot simply enter at will.”
State Appellees’ Answering Brief at 26–27, Tandon v. Newsom, 992 F.3d 916 (9th Cir. 2021) (No. 21-15228), 2021 WL 1499787, at *26–27. In any case, the Court deemed the commercial and religious gatherings comparable without any substantive analysis and despite the State’s evidence to the contrary. Whether the Supreme Court’s reversal in Tandon is viewed as an explicit or implicit reallocation of the relevant burden, therefore, that reallocation is difficult to deny.
But once the government articulates its reasons, religious claimants should bear the evidentiary burden to disprove those explanations.323Two commentators recently proposed an alternate framework where courts would “requir[e] governments to shoulder [the] evidentiary burden.” Collings & Barclay, supra note 4, at 496, 519–20. As noted, a law can fail the general-applicability test under the Supreme Court’s recent interpretation if it contains a single secular exception, and even if there is no evidence of discriminatory intent.324See Laurence H. Tribe, American Constitutional Law § 5–16, at 956 (3d ed. 2000) (“[A] law that is not neutral or that is not generally applicable can violate the Free Exercise Clause without regard to the motives of those who enacted the measure.”); Laycock, supra note 96, at 50–51 (“It is not that the legislature was consciously trying to harm religion when it failed to create a religious exemption. Rather, it is that such a discriminatory pattern of exemptions shows that the legislature’s goals do not require universal application, and that the legislature values the exempted secular activities more highly than the constitutionally protected religious activities.”). Nearly all laws include exceptions,325See Devine, supra note 258, at 1350 (“Nearly all laws have exceptions.”); Michael W. McConnell, The Problem of Singling Out Religion, 50 DePaul L. Rev. 1, 3 (2000) (“[F]ew statutes are genuinely applicable across the board, without exceptions . . . .”); Volokh, supra note 236, at 1540 (“[V]irtually all laws . . . contain many secular exceptions.”). so nearly all laws are susceptible to general applicability challenges.326See Rothschild, supra note 25, at 1113 (“[N]early every rule impinges on at least some individuals’ religious sensibilities and has at least one secular exemption . . . .”). And as the Court recognized in Smith, we live in a “cosmopolitan nation made up of people of almost every conceivable religious preference,”327Emp. Div. v. Smith, 494 U.S. 872, 888 (1990) (quoting Braunfeld v. Brown, 366 U.S. 599, 606 (1961)). so nearly all laws also have the potential to burden someone’s exercise of religion.328See Rosenkranz, supra note 12, at 1267; Christopher L. Eisgruber & Lawrence G. Sager, The Vulnerability of Conscience: The Constitutional Basis for Protecting Religious Conduct, 61 U. Chi. L. Rev. 1245, 1298 (1994) (“[T]he deep concerns of religious believers can differ sharply from each other and from widely shared secular concerns . . . .”). Courts, in turn, are reluctant to scrutinize the sincerity of claimants’ professed religious beliefs.329Frederick Mark Gedicks, “Substantial” Burdens: How Courts May (and Why They Must) Judge Burdens on Religion Under RFRA, 85 Geo. Wash. L. Rev. 94, 98 (2017) (“Challenging a claimant’s sincerity requires the government to argue and the courts to hold that claimants are lying about their beliefs—an ‘inquisitor-like’ tactic for which lawyers and judges have little appetite.”). But see Nathan S. Chapman, Adjudicating Religious Sincerity, 92 Wash. L. Rev. 1185, 1231–40 (2017) (arguing that there are methods courts can use to evaluate sincerity). And once sincerity is established—or, more likely, assumed330See, e.g., United States v. Lee, 455 U.S. 252, 257 (1982) (“We therefore accept [the claimant’s] contention that both payment and receipt of social security benefits [were] forbidden by the Amish faith.”); 303 Creative LLC v. Elenis, 6 F.4th 1160, 1181 (10th Cir. 2021) (“To be clear, we, like the Dissent, do not question Appellants’ ‘sincere religious beliefs’ or ‘good faith.’”), cert. granted in part, 142 S. Ct. 1106 (2022); Bible Believers v. Wayne County, 805 F.3d 228, 256 (6th Cir. 2015) (“Plaintiff Israel testified that he was required ‘to try and convert non-believers, and call sinners to repent’ due to his sincerely held religious beliefs. We do not question the sincerity of that claim.”).—those “beliefs need not be acceptable, logical, consistent, or comprehensible to others in order to merit First Amendment protection.”331Thomas v. Rev. Bd. of Ind. Emp. Sec. Div., 450 U.S. 707, 714 (1981). The result is that almost all laws are susceptible to free exercise challenges under the modern framework.332See Douglas Laycock, The Broader Implications of Masterpiece Cakeshop, 2019 BYU L. Rev. 167, 173 (“If a law with even a few secular exceptions isn’t neutral and generally applicable, then not many laws are.”); Rothschild, supra note 25, at 1113 (“[N]early every rule impinges on at least some individuals’ religious sensibilities and has at least one secular exemption . . . .”).
Placing the evidentiary burden on claimants to prove comparability thus would avoid the concern the Court expressed in Smith when it adopted the general-applicability standard. As the Court cautioned, “we cannot afford the luxury of deeming [generally applicable laws] presumptively invalid,” forcing the government to track down evidence at the behest of observers of “every conceivable religious preference.”333Emp. Div. v. Smith, 494 U.S. 872, 888 (1990) (quoting Braunfeld v. Brown, 366 U.S. 599, 606 (1961)). To be sure, many have criticized Smith, and at least some Supreme Court justices would likely vote to overrule it. See Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1912 (2021) (Alito, J., concurring) (arguing that Smith should be overruled); Douglas Laycock & Thomas C. Berg, Protecting Free Exercise Under Smith and After Smith, 2021 Cato Sup. Ct. Rev. 33, 33–34 (“[In Fulton, t]hree concurring justices, in an opinion by Justice Samuel Alito, argued at length for overruling Smith [and] two others, in an opinion by Justice Amy Coney Barrett, suggested that Smith was mistaken but that they were hesitant to overrule it without knowing what would replace it.”); see generally McConnell, Revisionism, supra note 34 (criticizing Smith). Requiring the government to disprove comparability would, in a similar sense, presume most laws aren’t generally applicable.334See Texas Dep’t of Cmty. Affs. v. Burdine, 450 U.S. 248, 255 n.8 (1981) (“The word ‘presumption’ properly used refers only to a device for allocating the production burden.” (quoting Fleming James, Jr. & Geoffrey C. Hazard, Jr., Civil Procedure § 7.9, at 255 (2d ed. 1977)); Fed. R. Evid. 301 (“[T]he party against whom a presumption is directed has the burden of producing evidence to rebut the presumption.”). Such a presumption would allow religious claims to slip through a window adjacent to the door the Court closed in Smith.335Lund, supra note 1, at 847 (“[Smith and its progeny were] a jurisprudence whose master principle was that courts should not give religious exemptions.”).
Placing the burden on the government also would create a strange paradox. As the Court explained in Lukumi, a law fails the narrow-tailoring element of strict scrutiny when the government’s “proffered objectives are not pursued with respect to analogous nonreligious conduct.”336Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 546 (1993). Under Tandon, however, “government regulations are not neutral and generally applicable, and therefore trigger strict scrutiny under the Free Exercise Clause, whenever they treat any comparable secular activity more favorably than religious exercise.”337Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam) (first emphasis added). If the government must disprove comparability at the general-applicability stage, it essentially must satisfy strict scrutiny to avoid strict scrutiny.338Courts and commentators are well aware of this paradox. See, e.g., Rothschild, supra note 25, at 1113–14 (“The very logic that implicates strict scrutiny—that a secular interest or entity is exempt, but a religious one is not—automatically locks in the conclusion that the lack of an exemption for religion is either not compelling, not narrowly tailored, or both.”); Michael C. Dorf, Under-Reacting to SCOTUS Theocracy, Dorf on L. (Dec. 2, 2020), https://perma.cc/9KWK-LMM8 (“[U]nder Smith, one only gets to the narrow tailoring inquiry of heightened scrutiny after determining that a law discriminates against religion. If a court can use the narrow tailoring inquiry itself to ascertain whether a law discriminates against religion, then the court has effectively overruled Smith.”); Tebbe, supra note 96, at 2450 (“[B]y the time anyone asks whether a government policy is narrowly tailored to a compelling interest, they will have already determined that it was underinclusive with respect to any such interest.”); Alan Brownstein, Protecting Religious Liberty: The False Messiahs of Free Speech Doctrine and Formal Neutrality, 18 J.L. & Pol. 119, 201 (2002) (“Of course, after we reject the state’s asserted purposes under our first standard of review and conclude that the law is underinclusive, there may not be all that much work left for the second step of strict scrutiny review.”); Legacy Church, Inc. v. Kunkel, 472 F. Supp. 3d 926, 1036 (D.N.M. 2020) (“Asserting that government action that burdens religions is subject to rational basis unless . . . the action is not the least restrictive means towards achieving the government’s stated end is a tautology, one that erases the distinction between the standards of review.”), aff’d sub nom. Legacy Church, Inc. v. Collins, 853 F. App’x 316 (10th Cir. 2021). Requiring religious claimants to establish comparability would avoid this paradox. And although the government is unlikely ever to prevail at the strict scrutiny stage once a claimant shows comparability, this is exactly what the Supreme Court envisioned in Lukumi.339Church of the Lukumi, 508 U.S. at 546 (“A law that targets religious conduct for distinctive treatment or advances legitimate governmental interests only against conduct with a religious motivation will survive strict scrutiny only in rare cases.”). Thus, the Third Circuit in Blackhawk was wrong to emphasize that Pennsylvania had not “shown” that the religious and secular activities in question weren’t analogous.340Blackhawk v. Pennsylvania, 381 F.3d 202, 211 (3d Cir. 2004). It should have been enough for the state to explain the difference, at which point the claimant, to proceed, should have been required to prove otherwise.
The Supreme Court’s jurisprudence under the Equal Protection Clause also suggests that claimants should bear the burden to show comparability.341The Court in Lukumi noted that “[n]eutrality in its application requires an equal protection mode of analysis.” Church of the Lukumi, 508 U.S. at 540 (quoting Walz v. Tax Comm’n of N.Y.C., 397 U.S. 664, 696 (1970) (Harlan, J., concurring)). For the reasons below, a different type of equal protection analysis may prove useful in gauging general applicability. In important respects, assessing comparability for free-exercise purposes resembles the “similarly situated” analysis in selective-enforcement and rational-basis equal-protection cases.342See, e.g., Stormans, Inc. v. Wiesman, 794 F.3d 1064, 1079, 1084–85 (9th Cir. 2015) (conducting similar analysis for free exercise, selective-enforcement, and equal protection claims). First, the key questions are similar. A law is generally applicable for free-exercise purposes unless the exempt secular conduct is “comparable” to the proposed religious conduct.343See Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam); Does v. Mills, 16 F.4th 20, 29 (1st Cir. 2021) (“To be generally applicable, a law may not selectively burden religiously motivated conduct while exempting comparable secularly motivated conduct.”). Likewise, a law treats classes or individuals equally for equal-protection purposes unless it effectively exempts other classes or individuals who are comparable to (i.e., “similarly situated” with) those it restricts.344See United States v. Armstrong, 517 U.S. 456, 465 (1996). Second, benign-classification and selective-enforcement cases, like general-applicability cases, arise from facially neutral laws.345See Alpha Delta Chi-Delta Chapter v. Reed, 648 F.3d 790, 804 (9th Cir. 2011).
In the equal-protection context, claimants bear the burden to prove comparability in such cases because there is reason to presume that the disparate effects of benign classifications and neutral laws stem from rational distinctions.346See, e.g., D.B. ex rel. Kurtis B. v. Kopp, 725 F.3d 681, 686 (7th Cir. 2013) (noting “the presumption of rationality that applies to government classifications” (quoting Flying J Inc. v. City of New Haven, 549 F.3d 538, 546 (7th Cir. 2008))). Because nearly all laws classify,347Bogen, supra note 138, at 241 (“All laws classify.”). it would be infeasible to presume laws invalid by virtue of classification alone. And many laws can’t be enforced against all violators, much less at the same rate,348See, e.g., Wayte v. United States, 470 U.S. 598, 607 (1985) (“In our criminal justice system, the Government retains ‘broad discretion’ as to whom to prosecute.” (quoting United States v. Goodwin, 457 U.S. 368, 380 n.11 (1982))). so it would be infeasible to presume enforcements invalid by virtue of disparate treatment alone.349See, e.g., Armstrong, 517 U.S. at 464–65 (holding that “to dispel the presumption that a prosecutor has not violated equal protection, a criminal defendant must present clear evidence to the contrary”—i.e., evidence that the decision to prosecute was “based on an unjustifiable standard such as race, religion, or other arbitrary classification” (internal quotation marks omitted)). Thus, in the equal-protection context, the government bears the burden to disprove comparability only when it classifies for presumptively invalid reasons.350See United States v. Carolene Prods. Co., 304 U.S. 144, 148, 152 n.4 (1938). The government must prove, for example, that it has good reason to treat people differently based on race, since the Court has held that race presumptively isn’t a good reason to treat people differently.351See, e.g., id.
In free-exercise cases, neutral laws that contain exemptions are closer to laws with benign classifications in the equal-protection context.352Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 540 (1993) (discussing parallels between free exercise and equal protection cases). Because “[a]ll laws are selective to some extent,”353Id. at 542. courts also have good reason to presume that laws with secular exemptions remain generally applicable. After all, the Supreme Court in Smith suggested that most “civic obligations” are generally applicable despite including secular exceptions.354See Emp. Div. v. Smith, 494 U.S. 872, 888–89 (1990). In fact, the Court in Smith listed numerous laws with secular exceptions as examples of general applicability.355See id. (citing, e.g., Gillette v. United States, 401 U.S. 437 (1971) (involving a military draft statute that contained exemptions for, among other things, people with family members who had been killed while serving); Funkhouser v. State, 763 P.2d 695 (Okla. Crim. App. 1988) (involving a manslaughter statute with an exception for “justifiable homicide”); Prince v. Massachusetts, 321 U.S. 158 (1944) (involving a child-labor statute that exempted “the sale or delivery of newspapers”)). But see Laycock, supra note 332, at 174 (“Smith’s discussion of Sherbert [v. Verner] implies that not many laws are neutral and generally applicable.”).
The equal-protection cases also offer some insight into a workable analytical framework. Courts often analyze equal-protection claims under a burden-shifting test.356See, e.g., 42 U.S.C. § 2000e–2(k); Naumovski v. Norris, 934 F.3d 200, 214 (2d Cir. 2019) (“[C]ourts already employ the [Title VII burden-shifting] framework to analyze § 1983 [equal protection] claims.”). This isn’t to say evaluating comparability in the general applicability context would be as straightforward as assessing comparability between employees in the employment discrimination context, but merely that employment discrimination law provides some guidance as to a useful framework. See Rothschild, supra note 4, at 286 (arguing that comparisons in the employment discrimination context are easier than in the Free Exercise context because, in determining general applicability, “courts often must compare widely different entities and activities” and “do not have the advantage of a specific workplace context that brings the two comparators into the kind of close proximity that makes comparisons—though still challenging—significantly more tenable”). In employment-discrimination cases, for example, a plaintiff must establish, as part of a prima facie case, that the employer took an adverse employment action.357See McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 (1973). The burden then shifts to the employer to articulate—but not prove—that the action was taken for legitimate, nondiscriminatory reasons.358See Tex. Dep’t of Cmty. Affs. v. Burdine, 450 U.S. 248, 259–60 (1981) (“In summary, the Court of Appeals erred by requiring the defendant to prove by a preponderance of the evidence the existence of nondiscriminatory reasons for terminating the respondent and that the person retained in her stead had superior objective qualifications for the position. When the plaintiff has proved a prima facie case of discrimination, the defendant bears only the burden of explaining clearly the nondiscriminatory reasons for its actions.” (footnote omitted)). If the employer provides such reasons, the burden shifts back to the employee to prove that those reasons are pretextual.359See McDonnell Douglas, 411 U.S. at 798; see also Naumovski, 934 F.3d at 215 (“In other words, a § 1983 plaintiff must establish that the employer’s stated non-discriminatory reason is either false or inadequate to support the adverse employment action.”). The employee may do so by identifying a similarly situated coworker whom the employer treated more favorably.360See McDonnell Douglas, 411 U.S. at 804.
A version of this framework would be useful in free-exercise cases. The prima facie case would entail the claimant needing to show (1) a sincere religious belief, (2) that the law in question burdens the exercise of that belief, and (3) that the law exempts certain secular conduct. The burden would then shift to the government to articulate the purposes of the law and why the religious conduct would undermine any of those purposes more than the exempt secular conduct. The religious claimant would then need to prove that a religious exemption wouldn’t undermine those purposes any more than the exempt secular conduct. The claimant, as in the employment context, would need to establish that the secular and religious conduct would undermine the government’s asserted interests in a similar way (i.e., that the religious and secular groups are similarly situated).361See Jonathan J. Marshall, Note, Selective Civil Rights Enforcement and Religious Liberty, 72 Stan. L. Rev. 1421, 1457 (2020) (advocating for requiring religious claimants in selective enforcement cases to make “a prima facie showing that there are other, similarly situated, nonreligious actors who are violators under the government’s theory of the case but who have not faced enforcement”).
The evidentiary burden on religious claimants according to this framework would be heavy. In many cases, claimants likely would be unable to produce evidence undermining the government’s reasons for treating religion differently. The point is, that’s consistent with the general-applicability doctrine. The general-applicability test resulted from the Court’s concern that too many “civic obligations” were susceptible to attack by individual religious believers.362See Emp. Div. v. Smith, 494 U.S. 872, 888–89 (1990). This concern stemmed from the notion that the First Amendment doesn’t “make the professed doctrines of religious belief superior to the law of the land.”363Reynolds v. United States, 98 U.S. 145, 167 (1879). A burden-shifting framework familiar in the antidiscrimination context would help avoid this pitfall under the Court’s recent interpretation of general applicability.
To be sure, the analogy between free exercise and equal protection has limits. The Equal Protection Clause protects against discrimination, whereas the Free Exercise Clause protects the exercise of religion, generally.364Compare U.S. Const., amend. I, with U.S. Const., amend. XIV. Justice Sandra Day O’Connor pointed this out in her concurrence in Smith, where she observed that “the Free Exercise Clause protects values distinct from those protected by the Equal Protection Clause.”365Smith, 494 U.S. at 901 (O’Connor, J., concurring). After all, a law that treats everyone the same still can prohibit some people’s exercise of religion.366See, e.g., McConnell, supra note 105, at 167 (“Neutrality often, indeed usually, serves the cause of religious liberty; but when it does not, a proper appreciation for the needs of religious practitioners requires something more.”). For example, a complete ban on alcohol would be neutral and generally applicable, but it would, in a sense, still “prohibit the free exercise”367U.S. Const. amend. I. of religious observers seeking to use sacramental wine.368See McConnell, supra note 105, at 152 & n.28. Some have argued that “it would be a mistake to conflate” the free-exercise analysis “with traditional equal protection doctrine.”369Eisgruber & Sager, supra note 328, at 1297 (“[Because] equal protection jurisprudence assumes that governmental behavior operating to disadvantage a vulnerable group is either purposefully designed to accomplish that result or is innocent of any constitutionally cognizable harm,” it differs from an ideal Free Exercise doctrine, which, according to the authors, would “condemn behavior that lies in a middle ground between purposeful discrimination and unintended disparate harm.”); see also Bogen, supra note 138, at 243 (rejecting an equal protection approach to Free Exercise cases because, unlike the Free Exercise Clause, “[t]he language of the Equal Protection Clause is language of classification”). But see Bernadette Meyler, The Equal Protection of Free Exercise: Two Approaches and Their History, 47 B.C. L. Rev. 275, 276 (2006) (“[I]t is precisely such a claim for the free exercise of religion that falls within the core of what equal protection of the laws has meant historically.”).
Further, “[p]roof of racially discriminatory intent”—or another impermissible motive—”is required to show a violation of the Equal Protection Clause,”370See Village of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 265 (1977). whereas discriminatory motive is unnecessary to disproving general applicability under the modern approach.371See, e.g., Laycock & Collis, supra note 87, at 10 (“The Free Exercise Clause protects religious observers against unequal treatment, regardless of targeting, motive, or an improper object.” (internal quotation marks, alterations, and footnote omitted)). Determining whether people are similarly situated for equal-protection purposes is merely the first step in the process of discerning discriminatory intent.372See, e.g., Domina v. Van Pelt, 235 F.3d 1091, 1099 (8th Cir. 2000) (“The threshold inquiry in an equal protection case is whether the plaintiff is similarly situated to others who allegedly received preferential treatment.”). In the free-exercise context, by contrast, the comparability analysis can be the only step in determining whether laws are generally applicable (i.e., whether they apply equally to religious and secular conduct).373See Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam).
For better or worse, however, general applicability is the standard in free-exercise cases, and, as of Tandon, comparability between secular and religious conduct is the crucial test.374See id. Whether two entities are analogous thus is an element in free-exercise cases, just as it is an element in equal-protection cases. That comparability is a means in the latter (i.e., a step toward showing intentional discrimination) and an end in the former (i.e., evidence that religion is being treated differently) doesn’t matter for purposes of allocating the burdens. In that sense, the burden-shifting framework in equal-protection cases applies just as well in the free-exercise context. While it would be wrong to equate the goals of free exercise and equal protection, they are useful comparators to the extent they both incorporate a “similarly-situated” analysis. This accords with the view of the most-favored-nation theory that “[a] stringently interpreted general-applicability rule can be understood as implementing a nondiscrimination requirement in the face of complexity.”375Laycock & Collis, supra note 87, at 26. In assessing alleged comparators at the general-applicability stage, courts have good reason to look to equal-protection cases.
A better argument against presuming incomparability stems from solicitude for minority faiths. Underrepresented religions are less politically powerful than some secular interests, so when the government enacts laws, it may be more likely to accommodate well-represented secular groups than idiosyncratic religious ones.376See, e.g., id. at 25 (“If secular interest groups burdened by the regulation get themselves exempted, they have no reason to oppose the regulation, and religious minorities are left standing alone.”); McConnell, supra note 105, at 152 (“Rarely will a neutral rule be passed or enforced that conflicts with the religious beliefs of the majority. Minority faiths are not so fortunate.”); Ry. Express Agency v. New York, 336 U.S. 106, 112 (1949) (Jackson, J., concurring) (“[T]here is no more effective practical guaranty against arbitrary and unreasonable government than to require that the principles of law which officials would impose upon a minority must be imposed generally.”). In such cases, the failure to include religious exceptions may be the result of inadvertence rather than considered judgment.377See, e.g., Abner S. Greene, Three Theories of Religious Equality . . . and of Exemptions, 87 Tex. L. Rev. 963, 986 (2009) (book review) (“[V]ulnerable groups (usually minorities) will sometimes have a claim of right against deliberate or inadvertent majoritarian neglect . . . .”). But see Zalman Rothschild, Religious Minority Status Upended: The Tale of a Hasidic Town, L.A. Rev. of Books (Feb. 17, 2022), https://perma.cc/G6DR-NKYB (highlighting evidence that at least some minority religious groups actually hold significant political sway). Secular conduct may end up exempt, while religious conduct may not, because politically-powerful groups want exceptions for the secular conduct they value, not because secular exceptions would undermine their goals less than religious exceptions. This concern is especially salient when the religious conduct at issue is politically disfavored, such as animal sacrifice in Lukumi.378Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 545–46 (1993) (“The ordinances have every appearance of a prohibition that society is prepared to impose upon Santeria worshippers but not upon itself. This precise evil is what the requirement of general applicability is designed to prevent.” (internal quotation marks, alterations, and citations omitted)).
Still, requiring the government to carry an explanatory burden would ameliorate these concerns. If there is no reason to treat religion differently, the government won’t be able to articulate legitimate reasons for omitting religious exceptions. This is apparently what happened in Blackhawk, where Pennsylvania didn’t even argue that religious exemptions would undermine its interests more than the secular exceptions to the permit requirement.379Blackhawk v. Pennsylvania, 381 F.3d 202, 211 (3d Cir. 2004). . There may be cases in which the government attempts to fabricate illegitimate reasons for declining religious exemptions,380See, e.g., Austin v. U.S. Navy Seals, 142 S. Ct. 1301, 1305–06 (2022) (Alito, J., dissenting) (suggesting that the Navy denied religious exemptions to COVID‑19 vaccines in bad faith). but if the omissions truly are inadvertent, this shouldn’t happen often. In any event, the fabricated reasons should be relatively easy for claimants to dispel at the evidentiary stage. Finally, to the extent the failure to include religious exceptions stems from discriminatory intent, disfavored religious groups would be more likely to find evidence undermining neutrality.381See, e.g., Church of the Lukumi, 508 U.S. at 540 (concluding that city ordinances evinced hostility toward the Santeria practice of animal sacrifice); Masterpiece Cakeshop, Ltd. v. Colo. C.R. Comm’n., 138 S. Ct. 1719, 1731 (2018) (“[T]he Court must draw the inference that Phillips’ religious objection [to serving a same-sex wedding] was not considered with the neutrality that the Free Exercise Clause requires.”). For these reasons, placing the evidentiary burden on religious claimants still makes most sense, given that nearly all laws include exceptions382See McConnell, supra note 325, at 3 (“[F]ew statutes are genuinely applicable across the board, without exceptions and without consideration of individual cases.”); Volokh, supra note 236, at 1540 (“[V]irtually all laws . . . contain many secular exceptions.”). and will inevitably infringe on some religious practices.383Emp. Div. v. Smith, 494 U.S. 872, 888 (1990).
In sum, although an evidentiary burden might be difficult for religious claimants to carry in many cases, that result is consistent with Smith. While Smith remains the law, the government can’t be required to provide evidentiary support for its legislative decisions at the behest of individual religious claimants who can show nothing more than that the laws in question, like nearly all laws, have exceptions.384See id. at 890 (“[T]hat unavoidable consequence of democratic government must be preferred to a system in which each conscience is a law unto itself or in which judges weigh the social importance of all laws against the centrality of all religious beliefs.”). Once the government articulates the reasons for the secular exceptions and why those exceptions don’t apply to religion, religious claimants should bear the burden to prove that religious exemptions in fact would undermine those reasons no more than the exempt secular conduct. In that respect, modern general-applicability doctrine may be administrable through a burden-shifting framework similar to that used in equal-protection and employment-discrimination cases.385Several commentators contend that the government should bear the evidentiary burden to disprove comparability. Some propose an alternative to Smith’s general applicability test that would “requir[e] governments to shoulder [the] evidentiary burden.” Collings & Barclay, supra note 4, at 496, 519–20. Another advocates “a presumption in favor of religious plaintiffs in the form of an evidentiary burden on the government” whenever a law contains secular exemptions and burdens someone’s religious exercise. See Note, Constitutional Constraints on Free Exercise Analogies, 134 Harv. L. Rev. 1782, 1796 (2021). The first of these proposals suggests a test to replace the neutrality and general applicability inquiry, not a way to ascertain general applicability. Regardless of whether this is right as a normative matter, or even as an interpretation of the Free Exercise Clause, those issues aren’t the focus of this Article. To the extent the second proposal reflects a view of how courts should ascertain comparability under the current general applicability framework, this Article advocates a different view for the reasons above. .
Allocating the burdens is only half the battle. Next, courts must weigh the evidence. Although this exercise depends on the facts of given cases, there are some threshold questions that arise routinely. One, numerosity, is a recurring issue that continues to divide courts.386See infra notes 388–423 and accompanying text. The second, uncertainty, is seldom discussed, but is becoming increasingly important.387See infra notes 424–435 and accompanying text.
a. Evidence of Numerosity
In deciding whether religious exemptions would undermine governmental interests more than secular exceptions, courts are divided as to whether they may consider the number of potential exemptions in each category, or, by contrast, must consider the effect of only the specifically-requested religious exemption.388Compare U.S. Navy Seals v. Biden, 578 F. Supp. 3d 822, 838 (N.D. Tex. 2022) with We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 286 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). In other words, is the question the effect of one religious exemption compared to one secular exemption? Or should courts compare the likely universe of religious exemptions to the likely universe of secular exemptions?
Justice Gorsuch would limit the inquiry to the government’s “interests as applied to the parties before” the court.389Does v. Mills, 142 S. Ct. 17, 20 (2021) (Gorsuch, J., dissenting). In his view, “the relevant question [at the general applicability stage] involves a one-to-one comparison between the individual seeking a religious exemption and one benefiting from a secular exemption.”390Dr. A v. Hochul, 142 S. Ct. 552, 556 (2021) (Gorsuch, J., dissenting). Some courts have endorsed this position. In U.S. Navy SEALs v. Biden,391578 F. Supp. 3d 822 (N.D. Tex. 2022). for instance, a judge in the Northern District of Texas rejected the government’s justification that “there [were] only seven permanent medical exemptions” to its vaccine mandate, whereas “there [were] more than three thousand pending requests for a religious exemption.”392Id. at 838 (internal quotation marks omitted). The court reasoned that “an influx of religious accommodation requests is not a valid reason to deny First Amendment rights.”393Id. In another vaccine case, Air Force Officer v. Austin,394588 F. Supp. 3d 1338 (M.D. Ga. 2022). a district judge in Georgia reached a similar conclusion that “[t]he general applicability test doesn’t turn on a numbers game,” and that “[a]ll it takes is one.”395Id. at 1355 (internal quotations marks and citation omitted).
The Second Circuit reached the opposite result in We The Patriots.396See We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 287 (2d Cir.) (“The record before us contains only limited data regarding the prevalence of medical ineligibility and religious objections, but what data we do have indicates that claims for religious exemptions are far more numerous.”), opinion clarified, 17 F.4th 368 (2d Cir. 2021). It noted that the evidence “indicate[d] that medical exemptions [were] likely to be more limited in number than religious exemptions,” that “it [might] be feasible for healthcare entities to manage the COVID-19 risks posed by a small set of objectively defined and largely time-limited medical exemptions,” and that “it could pose a significant barrier to effective disease prevention to permit a much greater number of permanent religious exemptions.”397Id. at 286. The court thus expressed “doubt that, as an epidemiological matter, the number of people seeking exemptions is somehow excluded from the factors that the State must take into account.”398Id. at 287. Likewise, in Blackhawk, the Third Circuit considered whether additional religious exemptions resembling the claimant’s request “would exceed the exemptions for qualifying zoos and circuses.”399Blackhawk v. Pennsylvania, 381 F.3d 202, 211 (3d Cir. 2004).
No matter which side is right as a constitutional matter, there’s no denying that numbers are relevant to the government’s interests in some cases.400See, e.g., Gabriel O. Aitsebaomo, The Nonprofit Hospital: A Call for New National Guidance Requiring Minimum Annual Charity Care to Qualify for Federal Tax Exemption, 26 Campbell L. Rev. 75, 85 (2004) (explaining that exemptions to federal taxes burden the government by lowering its revenues). If the government enacts laws to prevent the spread of a virus or earn money, the number of people exempt from those laws directly affects the relevant goals.401Id. at 84–85. While a single religious exemption might undermine those goals to the same extent as a single secular exemption, the legislature obviously isn’t so limited in deciding which exceptions to write into the rules and which to leave out. And granting a religious exemption to one claimant on free-exercise grounds may create a precedent requiring similar exemptions for similar claimants.402See, e.g., Olsen v. DEA, 878 F.2d 1458, 1464 (D.C. Cir. 1989) (declining a religious exemption for marijuana use, in part because “[t]he DEA would have no warrant to contain the exemption to a single church or religion”).
Even so, the Supreme Court has hinted at endorsement of a view similar to the position in the Texas vaccine cases. For example, it said in Fulton that religious and secular acts are comparable if they “undermine the government’s asserted interests in a similar way.”403Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021) (emphasis added). It didn’t say secular acts are comparable if they undermine the government’s asserted interest to a similar degree, in which case the number of likely exemptions would seem more relevant. And in Tandon, the Court considered the risks of allowing the specific “applicants’ proposed religious exercise at home.”404Tandon v. Newsom, 141 S. Ct. 1294, 1297 (2021) (per curiam). But see We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 287 (2d Cir.) (“The Supreme Court’s discussion in [Tandon], which compared the risks posed by groups of various sizes in various settings, suggests the appropriateness of considering aggregate data about transmission risks.”), opinion clarified, 17 F.4th 368 (2d Cir. 2021).
As discussed, Justice Gorsuch has provided the clearest endorsement of this position, albeit in dissent, arguing that “the estimated number of those who might seek different exemptions is relevant,” if at all, only during “the application of strict scrutiny.”405Dr. A v. Hochul, 142 S. Ct. 552, 556 (2021) (Gorsuch, J., dissenting). In his view, the “general applicability test doesn’t turn on that kind of numbers game.”406Id. Some district judges have followed the reasoning in Justice Gorsuch’s dissent, coming dangerously close to treating it as precedential. See, e.g., Air Force Officer v. Austin, 588 F. Supp. 3d 1338, 1355 (M.D. Ga. 2022) (relying on Justice Gorsuch’s dissent to conclude that “[t]he general applicability test doesn’t turn on a numbers game,” and that “[a]ll it takes is one” (quoting Dr. A, 142 S. Ct. at 556 (Gorsuch, J., dissenting)) (internal quotation marks and alterations omitted)). He would, as noted, apply “a one-to-one comparison between the individual seeking a religious exemption and one benefiting from a secular exemption.”407Dr. A, 142 S. Ct. at 556.
But Justice Gorsuch’s refusal to consider the numbers is more familiar in the context of strict scrutiny than it is at the general applicability-stage. Under RFRA and its counterpart, the Religious Land Use and Institutionalized Persons Act (“RLUIPA”), the Court has “require[d] the Government to demonstrate that the compelling interest test is satisfied through application of the challenged law ‘to the person’—the particular claimant whose sincere exercise of religion is being substantially burdened.”408Holt v. Hobbs, 574 U.S. 352, 363 (2015) (emphasis added) (quoting Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682, 526–27 (2014)). At that stage, the Court has “scrutiniz[ed] the asserted harm of granting specific exemptions to particular religious claimants” and “look[ed] to the marginal interest in enforcing the challenged government action in that particular context.”409Id. (internal quotation marks omitted); see also Muhammad v. Wheeler, 171 F. Supp. 3d 847, 856 (E.D. Ark. 2016) (“The proper focused inquiry under the RLUIPA is whether denying halal meat to Mr. Muhammad, not all ADC inmates, furthers a compelling government interest.”); Gholston v. Powell, No. 17-CV-00479, 2019 WL 5067201, at *9 (M.D. Ga. July 23, 2019) (similar), report and recommendation adopted, 2019 WL 4305507 (M.D. Ga. Sept. 11, 2019). The Court also has rejected the purportedly “compelling interest” that allowing a single religious exemption would invite more religious-exemption requests.410Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal, 546 U.S. 418, 436 (2006) (“The Government’s argument echoes the classic rejoinder of bureaucrats throughout history: If I make an exception for you, I’ll have to make one for everybody, so no exceptions.”). In Fulton, the Court incorporated this requirement into the compelling-interest test under the Free Exercise Clause.411There is some reason to question the basis for this decision. First, the claimant-specific language appears in RFRA and RLUIPA, but not in the Free Exercise Clause. See Religious Freedom Restoration Act, 42 U.S.C. § 2000bb-1(b) (“Government may substantially burden a person’s exercise of religion only if it demonstrates that application of the burden to the person—(1) is in furtherance of a compelling governmental interest; and (2) is the least restrictive means of furthering that compelling governmental interest.” (emphasis added)); Religious Land Use and Institutionalized Persons Act, 42 U.S.C. § 2000cc-1(a) (similar). Second, in United States v. Lee, the Supreme Court declined to grant a Free Exercise exemption to the social security tax largely because of the concern that “it would be difficult to accommodate the comprehensive social security system with myriad exceptions flowing from a wide variety of religious beliefs.” 455 U.S. 252, 259–60 (1982). In deeming the application of Philadelphia’s antidiscrimination law unconstitutional, it held that the question was “not whether the City ha[d] a compelling interest in enforcing its non-discrimination policies generally, but whether it ha[d] such an interest in denying an exception to [the religious claimant].”412Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1881 (2021).
Of course, the compelling-interest test takes place only after the court determines that the law in question isn’t neutral and generally applicable.413Id. at 1878 (noting that the first inquiry is whether the law is generally applicable). Only then does the government need to prove that it has a compelling interest in refusing the specific exemption the religious claimant seeks.414Id. at 1881. By contrast, the general-applicability test under Smith and Lukumi seems to call for a comparison of the categories of conduct the law covers and the categories it exempts.415See Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 542 (1993). In Lukumi, for instance, the Court emphasized that the “categories of selection are of paramount concern.”416Id.; see also Blackhawk v. Pennsylvania, 381 F.3d 202, 211 (3d Cir. 2004) (comparing categorical exemptions for zoos and circuses to a hypothetical category of religious exemptions). In other words, a law is generally applicable if a categorical religious exemption would undermine its purposes more than the categorical secular exemption it already includes.
The vaccine mandate cases are a perfect example. The states in those cases asserted that granting large numbers of religious exemptions would undermine the purposes of the mandates substantially more than granting a small number of medical exemptions.417See e.g., We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 286 (2d Cir.), opinion clarified, 17 F.4th 368 (2d Cir. 2021). That is, exempting numerous religious observers, in clustered areas, would cause the virus to spread more rapidly and widely than granting discrete medical exemptions.418Id. These concerns may be precisely why some states decline to include religious exemptions in immunization mandates. As the Second Circuit noted in We The Patriots, it would be odd if “the number of people seeking exemptions [were] somehow excluded from the factors that the State must take into account in assessing the relative risks . . . .”419Id. at 287. The vaccine mandate in that case was generally applicable because exempting the small category of people with medical needs was likely to undermine its goals less than exempting the large category of potential religious applicants.420One commentator has used a version of the numerosity principle to argue that secular exemptions to trespass laws don’t defeat general applicability. See Lund, supra note 41, at 643 (“[G]ranting a religious exception would do significantly more harm to the rule than the existing secular exceptions,” because “highly context-dependant [sic] and individual-specific exceptions to the trespass statute” aren’t likely to “produce a continuing flow of people onto a particular property,” whereas a religious exemption “not limited to a particular time, place, or individual” would “drive a truck through this narrow hole”).
Still, by focusing on the religious exercise of the applicants, Tandon indicated that the categorical interpretation is at least limited to the type of religious conduct for which the specific claimants seek an exemption.421See Tandon v. Newsom, 141 S. Ct. 1294, 1296–98 (2021) (per curiam). This stems from the simple idea that different religious practices affect governmental interests to different extents. A religious claimant seeking an exemption to use a hallucinogenic drug for a limited, private ceremony differs from a religious claimant seeking to use a hallucinogenic drug while driving a school bus. If the private-use claimant seeks an exemption, that person should need to demonstrate only that the number of other people seeking to use the drug for such ceremonies wouldn’t undermine the purposes of the ban more than an existing secular exception. The claimant wouldn’t need to introduce evidence of the number of people seeking exemptions for religious uses that would more severely undermine state interests, such as people seeking to use hallucinogens in public spaces. This means that a law with a secular exception might remain generally applicable despite prohibiting some forms of religious exercise but not others. This approach best accords with Tandon’s focus on the “applicants’ proposed religious exercise.”422Id. But see We The Patriots, 17 F.4th at 287 (“The Supreme Court’s discussion in [Tandon], which compared the risks posed by groups of various sizes in various settings, suggests the appropriateness of considering aggregate data about transmission risks.”).
Ultimately, Smith, Lukumi, and Tandon suggest that, in cases where numbers are relevant, the key figure is how many people would seek to engage in the same sort of religious conduct as that proposed by the specific claimants before the court.423See Tandon, 141 S. Ct. at 1297; Emp. Div. v. Smith, 494 U.S. 872, 888 (1990); Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 545–46 (1993). These categories are defined by the extent to which the specific claimants’ religious activities would undermine the government’s interests.
b. Evidence of Uncertainty
One more category of evidence, similar in a sense to numerosity, is “uncertainty.” This category reflects how religious beliefs often are unpredictable, subjective, and diverse.424See, e.g., McConnell, Revisionism, supra note 34, at 1139 (“The ideal of free exercise is counter-assimilationist; it strives to allow individuals of different religious faiths to maintain their differences in the face of powerful pressures to conform.”); Smith, 494 U.S. at 888 (noting the difficulty in accommodating the diversity of religious beliefs in the United States); United States v. Lee, 455 U.S. 252, 259–60 (1982) (“[I]t would be difficult to accommodate the comprehensive social security system with myriad exceptions flowing from a wide variety of religious beliefs.”). Religious exemptions therefore may be harder to grant, manage, and police. Such practical barriers to accommodating religion don’t necessarily arise with respect to otherwise similar secular conduct. This isn’t to say religious conduct is less worthy of exemption than secular conduct—again, a conclusion the most-favored-nation theory forbids425See Laycock & Collis, supra note 87, at 26.—but merely that, as a practical matter, some religious conduct is more difficult to exempt than secular conduct that otherwise would be comparable.
Consider the vaccine mandate cases. If a state exempts anyone who can provide a doctor’s note, it has an easy enforcement mechanism.426See We The Patriots, 17 F.4th at 286 (“[I]t may be feasible for healthcare entities to manage the COVID‑19 risks posed by a small set of objectively defined and largely time-limited medical exemptions.”). People can’t get the medical exemption without a physician attesting to the requisite circumstances.427See, e.g., Doe v. San Diego Unified Sch. Dist., 19 F.4th 1173, 1178 (9th Cir. 2021) (“The medical exemption is limited to students with contraindications or precautions recognized by the Centers for Disease Control and Prevention or the vaccine manufacturer, and the request must be certified by a physician.”), reh’g en banc denied, 22 F.4th 1099 (9th Cir. 2022). The exemption therefore is easy to limit, no matter the lengths to which people would go to avoid compliance. Although the state can require people to affirm the sincerity of their beliefs,428On rare occasions, courts have rejected nominal religious objections to immunization on the grounds that the objections actually stemmed from nonreligious motives. See, e.g., Check ex rel. MC v. N.Y.C. Dep’t of Educ., No. 13-CV-791, 2013 WL 2181045, at *2 (E.D.N.Y. May 20, 2013) (rejecting a religious objection to a state vaccine requirement because the evidence showed that the objection actually stemmed from health concerns); Farina v. Bd. of Educ. of N.Y., 116 F. Supp. 2d 503, 513 (S.D.N.Y. 2000) (similar). such beliefs, unlike medical evidence, are subjective.429In part for this reason, “scholars have long questioned the wisdom and constitutionality of adjudicating religious sincerity.” Chapman, supra note 329, at 1188. There are no readily administrable objective tests to separate people seeking the exemption for genuine religious reasons from those with ulterior motives.430Dorit Rubinstein Reiss, Thou Shalt Not Take the Name of the Lord Thy God in Vain: Use and Abuse of Religious Exemptions from School Immunization Requirements, 65 Hastings L.J. 1551, 1570 (2014) (arguing that, in vaccine objection cases, states have “very limited tools to police [religious] exemptions and prevent abuse”). This isn’t to say that people claiming sincere religious beliefs are more likely to fabricate claims than people claiming medical need. Rather it assumes that false medical and religious claims would be just as likely, all else being equal.431Nor does this view question the inherently circular and thus indisputable point that religious believers have sincere religious beliefs. As noted, it simply means that people without sincere religious beliefs or genuine medical conditions can more easily fake beliefs than medical evidence. It simply means medical need is susceptible to objective proof, whereas religious beliefs necessarily aren’t.432In Sherbert v. Verner, the Supreme Court rejected the argument that the state had a compelling interest in preventing “the filing of fraudulent claims by unscrupulous claimants feigning religious objections,” explaining that “there [was] no proof whatever to warrant such fears of malingering or deceit.” 374 U.S. 398, 407 (1963). But this was at the strict scrutiny stage where the government bore the burden of proof. Id. at 406. Insincere claimants would have an easier path to deception via fabricated religious claims than fabricated medical claims.
The administrative uncertainty that attends religious exemptions is especially pronounced in cases where the requested exemptions are desirable for nonreligious reasons. For example, many people likely didn’t (and don’t) want to comply with COVID-19 vaccine and mask mandates, completely irrespective of religion.433According to the Pew Research Center in March 2022, “[t]wo-thirds of U.S. adults say most people who claim religious objections to a COVID‑19 vaccine ‘are just using religion as an excuse to avoid the vaccine,’ while about a third (31%) say they think the objectors ‘sincerely believe getting a COVID‑19 vaccine is against their religion.’” Justin Nortey, Americans Skeptical About Religious Objections to COVID‑19 Vaccines, But Oppose Employer Mandates, Pew Rsch. Ctr. (Mar. 31, 2022), https://perma.cc/KZ6L-XXHE. In fact, some people objected to COVID‑19 vaccination for religious reasons that appeared equally applicable to other vaccines they already had received. See We The Patriots USA, Inc. v. Hochul, 17 F.4th 266, 287 (2d Cir.) (noting that vaccines that healthcare-worker claimants already had received had “connections to the same fetal cell lines that form[ed] the basis for [those claimants’] religious objections” to the COVID‑19 vaccine), opinion clarified, 17 F.4th 368 (2d Cir. 2021). In such cases, the likelihood of false claims naturally rises.434See John D. Grabenstein, What the World’s Religions Teach, Applied to Vaccines and Immune Globulins, 31 Vaccine 2011, 2019 (2013) (studying religious objections to vaccinations and concluding that many actually “reﬂected concerns about vaccine safety, not matters of theology”); Reiss, supra note 430, at 1573 (“[S]ources examining the reasons for not vaccinating suggest that the more common reasons given are not religious, but generally ill-founded safety concerns.”). If the number of exemptions would be relevant to the interest behind the laws in question, the importance of objectivity rises as well. A state might have good reason to limit exemptions to cases where applicants can provide proof of need beyond their own assertions. Again, this isn’t to say the religious reason for a given exemption is less important than a secular reason; it merely means a secular exception requiring objective evidence might undermine a law’s purposes less than a subjective religious exemption susceptible to false claims.
The upshot is that potential uncertainty and administrative costs very well may be relevant at the general applicability stage, and courts shouldn’t ignore those concerns. If the government declines a religious exemption in part because, unlike a secular exception, it would create unique public uncertainty or be difficult to administer or enforce, courts should consider such assertions and require religious claimants to produce contrary evidence.
Once these threshold questions are resolved, courts must undertake the similarly daunting task of making the actual comparisons. Many more questions will arise during that inquiry, such as when two entities are comparable enough, despite marginal, relevant differences.435See, e.g., Elim Romanian Pentecostal Church v. Pritzker, 962 F.3d 341, 346 (7th Cir. 2020) (“It would be foolish to pretend that worship services are exactly like any of the possible comparisons . . . .”), cert. denied, 141 S. Ct. 1753 (2021); Rothschild, supra note 4, at 285 & n.15 (noting the difficulty of comparing the degrees to which religious and secular activities undermine certain interests). It is easy to imagine a challenging case where a religious exemption to a vaccine mandate would result in a negligibly higher transmission risk than a medical exemption. Assuming the court is willing to consider aggregates—as this Part has urged—a medical exception that would result in ten unvaccinated people might undermine the mandate’s purpose differently from a religious exemption sought by 1,000, but might undermine that purpose “in a similar way” to a religious exemption sought by fifteen. Such tricky questions would arise on the margins, but they are well within the province of courts and factfinders. Difficult line-drawing is sometimes necessary, but the occasional difficult case isn’t a good reason to reject a test outright. Purely doctrinal considerations such as whether to consider numerosity and uncertainty are less familiar to courts, but no less important. For the reasons discussed, both should be fair game.
C. Comparing Religious and Secular Activities
So, how should courts compare religious and secular activities? To bring the issue full circle, consider a case like Smith.436See generally Emp. Div. v. Smith, 494 U.S. 872 (1990). Imagine that a state—like Oregon in 1990—has a controlled substances law that prohibits peyote use except when prescribed by a doctor. Religious claimants seeking to use peyote for sacramental purposes sue under the Free Exercise Clause requesting a religious exemption. First, the claimants must show that they would use peyote for sincere religious reasons—in other words, that the controlled substances law burdens their religious exercise.437See Thomas v. Rev. Bd. of Ind. Emp. Sec. Div., 450 U.S. 707, 717 (1981). Then they must show that the law isn’t neutral and generally applicable vis-à-vis religion.438See 303 Creative LLC v. Elenis, 6 F.4th 1160, 1186 n.8 (10th Cir. 2021), cert. granted in part, 142 S. Ct. 1106 (2022). They can contest general applicability, under Tandon, by pointing out that the controlled substances law exempts people with a medical prescription.439See Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). This essentially puts the question to the state: Is there a difference between the impact of the medical and proposed religious exemptions, as judged against the purposes of the controlled substances law? If not, the law isn’t generally applicable.
The state then must carry an explanatory burden. It must articulate legitimate reasons for prohibiting sacramental use that don’t apply equally to medical use. It could point to concerns, for example, that medical use is less likely than religious use to lead to peyote abuse and illegal trafficking; that medical use is safer because it is typically controlled and monitored; or that religious use would be relatively difficult to regulate, perhaps in light of the number of prospective religious users.
What the state can’t do, under Tandon and the most-favored-nation interpretation, is rely on the reasons for the medical exception itself. That is, the state can’t say that the medical exception is more important than the requested religious exemption because some people would face serious health risks without using controlled substances. Such reasoning would implicitly compare the medical and religious consequences of prohibiting peyote use, which, as discussed, is precisely the sort of value-judgment the most-favored-nation interpretation rejects.440See supra notes 96–120 and accompanying text.
Assuming the state articulates at least one legitimate reason to deny the religious exemption that wouldn’t also apply to the medical exception, the burden shifts to the religious claimants to disprove the state’s explanation.441See supra Section III.B.1. The claimants must introduce evidence that sacramental peyote use wouldn’t cause the identified problems more than medical use. If the state’s rationale is that medical uses are less likely to lead to peyote abuse than religious uses, the claimants could point to evidence of abuse stemming from medical uses.442See, e.g., Devine, supra note 258, at 1375–76 (“Prescription drugs cause the increased possession, trafficking, and abuse of illegal drugs.” (footnotes omitted)). If the state’s articulated rationale reflects concerns about potential diversion of peyote into the recreational market, the claimants could introduce evidence that prescription peyote also results in such diversion,443Cf. Gonzales v. Raich, 545 U.S. 1, 40 (2005) (Scalia, J., concurring) (“[M]arijuana that is grown at home and possessed for personal use is never more than an instant from the interstate market—and this is so whether or not the possession is for medicinal use or lawful use under the laws of a particular State.”). or that religious uses haven’t resulted in diversion when permitted in the past.444Cf. Gonzales v. O Centro Espirita Beneficente Uniao do Vegetal, 546 U.S. 418, 426 (2006) (“[The religious claimants] emphasized the thinness of any market for [the requested substance], the relatively small amounts of the substance imported by the church, and the absence of any diversion problem in the past.”). If the state explains that people would fabricate religious claims to obtain peyote, the claimants could introduce evidence that recreational use of peyote would be rare,445See, e.g., Bob Prue, Prevalence of Reported Peyote Use 1985–2010 Effects of the American Indian Religious Freedom Act of 1994, 23 Am. J. on Addictions 156, 157 (2014) (“Peyote use within the general population is uncommon.”). and thus that the universe of potential insincere claimants would be negligible.
None of this is to say the ultimate evidentiary evaluation will be easy. And even if the claimants fail to muster any evidence in the peyote context, there’s no doubt that such evaluations will be difficult in other free exercise cases. But evaluating evidence is familiar terrain for courts and juries. The crucial inquiry is where to begin. When a religious observer requests an exemption, the court and parties need to know who is responsible for going out and tracking down the evidence. And they need to know what evidence is relevant.
Religious exemptions always have been controversial.446See, e.g., Kent Greenawalt, Exemptions: Necessary, Justified, or Misguided? 1–5 (2016); Ira C. Lupu, Hobby Lobby and the Dubious Enterprise of Religious Exemptions, 38 Harv. J.L. & Gender 35, 36–39 (2015). The COVID-19 distancing, vaccine, and mask mandate cases brought this controversy back into both legal and popular discourse.447See, e.g., Tebbe, supra note 96, at 2398; Berg, supra note 1, at 735–36 (“The most extraordinary religious-freedom question of the last two years has involved the bans and limitations on in-person worship gatherings pursuant to public-health orders designed to slow transmission of COVID‑19.”); Nortey, supra note 433. The problem is that the Supreme Court’s most recent statement of the general-applicability standard left many questions unanswered.448See generally Lund, supra note 1; Berg, supra note 1. Under the newly-endorsed, most-favored-nation approach, the crucial question is whether religious and secular activities are “comparable.”449Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021) (per curiam). Comparability turns on how the competing activities affect governmental interests.450Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1877 (2021). But the Supreme Court hasn’t said how to determine whether religious and secular conduct would have similar effects, who bears the burden to show or disprove comparability, or what evidence is relevant to that inquiry. Commentators have since lamented these practical dilemmas.451See, e.g., Rothschild, supra note 4, at 285 & n.15; Sunstein, supra note 4, at 235–36. Courts, in turn, have used the uncertainty surrounding these “how” questions to subtly manipulate the procedural levers in free exercise cases.452See, e.g., Rothschild, supra note 4, at 285–86. Despite calls to overrule Employment Division v. Smith in 2021, it didn’t happen.453See generally Fulton, 141 S. Ct. 1868. Instead, the Supreme Court articulated a new substantive standard.454See Tandon, 141 S. Ct. at 1296–97; Fulton, 141 S. Ct. at 1877. Now, the “how” questions need answers.
This Article attempts to engage with some of these issues. If a religious claimant can establish a burden on religious exercise and identify a secular exception, the courts first must identify the government’s “asserted” interests. It doesn’t matter how broad these interests are, how numerous, or when the government first articulates them, so long as the interests are legitimate. The government then should bear the burden to articulate why accommodating religious conduct would undermine its asserted interests more than the already-exempt secular conduct. If the government does so, religious claimants should bear the evidentiary burden to show that religious accommodations would in fact detract from those interests no more than the exempt secular conduct. In relevant cases, claimants would need to demonstrate that the entire category of similar religious exemptions wouldn’t undermine the government’s interests more than the entire category of secular exemptions, and that the category of religious exemptions wouldn’t create unique numerosity or uncertainty problems.
For as long as general applicability remains the threshold test, courts must stay faithful to its purpose. That means they can’t foist an evidentiary burden on the government whenever individual religious claimants seek exemptions from neutral laws that happen to accommodate limited categories of secular conduct. We live in a “cosmopolitan nation made up of people of almost every conceivable religious preference,”455Emp. Div. v. Smith, 494 U.S. 872, 888 (1990) (quoting Braunfeld v. Brown, 336 U.S. 599, 606 (1961)). and “[n]early all laws have exceptions.”456Devine, supra note 258, at 1350; see McConnell, supra note 325, at 3; Volokh, supra note 236, at 1540. Smith said “we cannot afford the luxury of deeming presumptively invalid, as applied to the religious objector, every regulation of conduct that does not protect an interest of the highest order.”457Smith, 494 U.S. at 888. Likewise, courts shouldn’t presume that religious and secular activities are “comparable,” when doing so would invite religious objections to “civic obligations of almost every conceivable kind.”458Id.