Does the Free Exercise Clause mandate religious exemptions to facially neutral, generally applicable laws? In Employment Div. v. Smith, 494 U.S. 872 (1990), the Supreme Court answered in the negative. This precedent has been challenged before (see Fulton v. City of Philadelphia, 593 U.S. ___ (2021), where three Justices wrote separately to question or call for the overruling of Smith), and the Court recently granted certiorari to scrutinize Smith’s scope once again (see St. Mary Catholic Parish v. Roy, No. 25-581). But as Justice Barrett cautioned in Fulton, abandoning Smith poses a pressing follow-up question: what would take its place?
This essay argues that the text, structure and history of the Free Exercise Clause and the demands of stare decisis strongly favor reaffirming Smith, and that Yoder and RFRA provide alternative frameworks that the Court should overrule and strike down respectively.
I – The evolution of Smith
In Sherbert v. Verner, 374 U.S. 398 (1963), the Court was confronted with whether or not South Carolina violated the Free Exercise Clause when it denied unemployment benefits to a Seventh-day Adventist who was fired for refusing to work Saturdays. The Court ruled that it did, but more importantly, it established a test for evaluating such Free Exercise challenges: firstly, courts must determine whether the government action posed a ‘significant burden’ on an individual’s ability to exercise their religious beliefs, and secondly, whether the state has proven that the action fulfills a ‘compelling government interest’ and that the action taken is the ‘least restrictive means’ to achieve that interest. To succeed, a claimant must establish the first prong, after which the burden shifts to the government to satisfy both elements of the second. This test was notably expanded beyond the facts of Sherbert in Wisconsin v. Yoder, 406 U.S. 205 (1972), when the Court held that the State could not compel Amish children to attend school past the eighth grade as this substantially burdened the Amish community’s ability to exercise their religious practice and did not fulfill a compelling government interest. Despite the ostensibly higher burden on the government under Sherbert, however, applications of this test often failed before the Court.[1] This demonstrates how, even before Smith, the Court was reluctant to expand the scope of Sherbert and Yoder.
From this perspective, the Court’s decision in Smith appears to have been an inevitability, and merely codified the evolution of the Court’s jurisprudence in this area. Alfred Smith and Galen Black were fired for ingesting peyote as part of Native American ceremonies, and they challenged Oregon’s denial of government benefits to them. The Court rejected this challenge in Smith, and held that the Free Exercise Clause does not require religious exemptions from neutral, generally applicable laws, and also explicitly restricted Sherbert to the context of unemployment benefits. The Court subsequently applied Smith to protect claimants in Church of the Lukumi Babalu Aye, Inc. v. Hialeah, 508 U.S. 520 (1993), when it unanimously struck down Hialeah’s ordinances prohibiting animal sacrifice, which were facially neutral but transparently targeted the Santeria religion. Importantly, this shows that even under Smith, claimants could succeed in challenging laws that were genuinely hostile toward religion.
Congress nearly unanimously (97-3 in the Senate) responded to Smith later in 1993 with the Religious Freedom Restoration Act (RFRA). The Act restored the Sherbert test across the board, requiring the federal and (at the time) state governments to demonstrate a compelling interest and least restrictive means before substantially burdening religious exercise. However, the Court struck down RFRA as applied to state and local governments in City of Boerne v. Flores, 521 U.S. 507 (1997) as exceeding Congressional power. Congress responded to Flores by enacting the Religious Land Use and Institutionalized Persons Act (RLUIPA), which applied the Sherbert test specifically to the contexts of land use regulation and prisons, domains where Congress has clearer Commerce Clause and Spending Clause authority.
In the years post Flores, the Court has applied RFRA in federal contexts (see, inter alia, Gonzales v. O Centro Espírita Beneficente União do Vegetal, 546 U.S. 418 (2006) and Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682 (2014)). Burwell v. Hobby Lobby is of particular interest in this essay, as it expanded the scope of RFRA far beyond any coherent limiting principle: the Court declared that RFRA exemptions extend to corporations and affect third-party rights (in this case, employees’ access to contraception).
This brings us to the Court’s modern developments to Smith. In Fulton v. City of Philadelphia, 593 U.S. ___ (2021), the Court unanimously held that the refusal of Philadelphia to contract with a Catholic foster care agency unless they agree to certify same-sex couples as foster parents violates the Free Exercise Clause. In Masterpiece Cakeshop v. Colorado Civil Rights Commission, 584 U.S. 617 (2018) and 303 Creative LLC v. Elenis, 600 U.S. 570 (2023), the Court ruled in favor of claimants who objected to producing pro-LGBT expressive content. At first glance, these additions to the Court’s Free Exercise jurisprudence seem to undercut Smith – after all, the Court struck down facially neutral and generally applicable government action for significantly burdening the free exercise of religion. In reality, these decisions were consistent with or fell outside the scope of Smith by the Court’s own analysis. Fulton dealt with a rule which had exceptions – the existence of exceptions to the rule meant the law was not generally applicable, triggering strict scrutiny that Philadelphia could not satisfy; the ruling in Masterpiece Cakeshop was extremely narrow, and was decided on the grounds that the Commission had displayed hostility to petitioner’s beliefs, violating Smith’s own requirement of neutrality; and 303 Creative was decided on entirely separate Free Speech Clause grounds altogether.
Where does this leave Smith? As this discussion has shown, the Court has hardly whittled down Smith, but it has not meaningfully applied it either, especially in favor of the government. The best way to describe its state, therefore, is one of Sherbert-esque limbo, awaiting a decision that will do to Smith what Smith did to Sherbert – either confine it to its facts or sweep it away entirely. The question then becomes: what should this new direction be, or should the Court revive Smith?
II – RFRA
The obvious first choice is to apply RFRA’s restored Sherbert standard – the Court has already done so in cases such as Burwell v. Hobby Lobby, so formally switching to exclusively RFRA would be consistent with its existing applications. This direction, however, has two irreconcilable flaws.
At the very least, application of RFRA would create an inconsistency in the application of the Free Exercise Clause between the federal government and states.[2] As RFRA has been struck down as applied to the state governments in Flores, a test derived from RFRA could only apply to the federal government, making a double standard. This would be a significant deviation from the manner in which other Bill of Rights provisions have been incorporated against the states – Justice Brennan, writing for the Court in Malloy v. Hogan, 378 U.S. 1 (1964), made this explicit. See “[t]he Court…has rejected the notion that the Fourteenth Amendment applies to the States only a watered-down, subjective version of the individual guarantees of the Bill of Rights”, Id. at 10 (internal quotation marks omitted). Justice Alito, a key opponent of Smith on the Court, recognized this in his own landmark decision incorporating the Second Amendment against the states, McDonald v. City of Chicago, 561 U.S. 742 (2010). See Id. at 766, recognizing the “well established rule that incorporated Bill of Rights protections apply identically to the States and the Federal Government.” Therefore, an application of RFRA cannot be reconciled with the weight of the Court’s precedent on consistent Bill of Rights applications between the federal and state governments, as adopting RFRA would leave the Smith standard in place against the states.
Secondly, an examination of the Court’s precedents renders RFRA unconstitutional in its entirety. “It is emphatically the province and duty of the judicial department to say what the law is[,]” Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803). RFRA strips this exact power from the courts – it instructs the judiciary on how to interpret the Free Exercise Clause and prescribes the outcomes of cases. This violates the separation of powers doctrine, as the Court has recognized that
Congress may not dictate the outcomes of cases, or say what the law is (see United States v. Klein, 80 U.S. (13 Wall.) 128, 147 (1871), which struck down an 1870 statute for “pass[ing] the limit which separates the legislative from the judicial power” as it prescribed the rule of decision in a particular cause); this separation of powers issue is compounded by the fact that the Court ruled on this exact issue in Smith, effectively rendering RFRA a statutory attempt by Congress to overturn a constitutional decision. The constitutional structure further supports this reading through an analogy to the anti-commandeering doctrine, which forbids Congress from compelling state executive officers to administer federal law, and was the grounds on which RFRA was struck down in part in Flores: if Congress cannot compel the states in such a manner, why can it do so to the judiciary? The answer is that it cannot – this is precisely the kind of legislative override that is forbidden by the constitutional structure of three branches of government, so RFRA cannot stand.
RFRA raises another, Establishment Clause-related concern. Take the example of Burwell v. Hobby Lobby: had it been an atheist claimant, RFRA would have granted them no exemption to the mandate at hand. Or, take Flores: “[i]f the historic landmark on the hill in Boerne happened to be a museum or an art gallery owned by an atheist, it would not be eligible for an exemption from the city ordinances that forbid an enlargement of the structure,” City of Boerne v. Flores, 521 U.S. 507, 536-537 (1997) (Stevens, J., concurring). Therefore, by mandating religious but not irreligious exemptions, Congress passed a “law respecting an establishment of religion,” U.S. Const. amend. I, cl. 1 – the exact sort of law the Establishment Clause prohibits. Justice Stevens made this very argument, see Id., and it is an independent reason why RFRA ought to be declared unconstitutional.
III – Sherbert, Yoder, and Smith
While the test established in RFRA cannot be applied by the judiciary, could this sort of a regime find refuge in one of the Court’s precedents on this matter? For this, we must turn to the merits of the issue at hand in Sherbert, Yoder, and Smith.
A – Merits
The Free Exercise Clause (hereinafter Clause) provides that “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof,” U.S. Const. amend. I” (emphasis added). To what extent does this restriction apply? While the Clause “obviously excludes all governmental regulation of religious beliefs as such[,]” Smith, supra, at 872, the extent of its protection for conduct incidental to religious belief is more limited. This is especially the case with laws not targeting religion. Take, for example, jury duty – this well-recognized civic obligation could plausibly trample on an individual’s political or religious objections to being forced to endorse a legal verdict which violates their moral conscience. Does this government action violate those individuals’ rights under Free Speech and Free Exercise Clauses, and by extension, mandate such exemptions to jury duty? It is far-fetched to think so. The Court’s precedents prior to Sherbert, Yoder, and Smith actually affirm this reading, contrary to Justice Alito’s opinion in Fulton. See Reynolds v. United States, 98 U. S. 145, 166-167 (1879), “[c]an a man excuse his practices…because of his religious belief? To permit this would be to make the professed doctrines of religious belief superior to the law of the land, and in effect to permit every citizen to become a law unto himself”; and Minersville School Dist. Bd. of Ed. v. Gobitis, 310 U. S. 586, 594-595 (1940), “[t]he mere possession of religious convictions which contradict the relevant concerns of a political society does not relieve the citizen from the discharge of political responsibilities.”[3]
Justice Alito also raises the objection in Fulton that such a reading of the Clause would contradict the absolute nature of other Bill of Rights provisions, namely the Sixth and Seventh Amendments. [4] However, these are far less analogous to the Clause than other First Amendment clauses and the Second Amendment, given that these concern civil liberties, unlike the Sixth and Seventh, which pertain to rights in the context of criminal and civil cases. A look at the relevant precedents further reveals the consistency of a less strict reading. With respect to limitations on the Bill of Rights, look no further than District of Columbia v. Heller, 554 U.S. 570 (2008), in which the Opinion of the Court (which Justice Alito joined) recognized that “[l]ike most rights, the right secured by the Second Amendment is not unlimited.” As for the test itself, see United States v. O’Brien, 391 U.S. 367 (1968), in which the Court held that laws that are content-neutral, further a substantial government interest, and are the least restrictive means of furthering that interest may be upheld even if they bar conduct incidental to speech (in this case, burning draft cards) – this test is extremely similar to the one established in Smith, making Smith the doctrinally consistent reading.
Finally, the history of ‘free exercise’ legislation demonstrates that the limited nature of the Clause was understood at the time of the founding. These provisions were often restricted to government action taken directly against religion (see, e.g., Maryland Act Concerning Religion of 1649, Rhode Island Charter of 1663, and New Hampshire Constitution) or discriminatory action (e.g. New York Constitution), not facially neutral, generally applicable laws. A contrary reading to the one in Smith, therefore, would run afoul of the founders’ vision of the Clause.
In sum, as a matter of text, history and precedent, Smith provides the better reading of the Free Exercise Clause.
B – Stare decisis
Our analysis of these cases cannot stop at the merits, however, as it contemplates – nay, necessitates – the overruling of the Court’s precedent. This, of course, is not to be taken lightly, as “with great power [to overrule precedent] there must also come—great responsibility,” Kimble v. Marvel Entertainment, LLC, 576 U.S. 446, 465 (2015) (quoting Amazing Fantasy No. 15: “Spider-Man”). However, reaffirming Smith requires neither overruling Sherbert nor disturbing its core holding – Smith itself confined Sherbert to the unemployment compensation context, leaving it intact within that domain. Neither does striking down RFRA in its entirety invoke stare decisis concerns – for an analogous circumstance, see Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), which overruled Chevron v. NRDC, 467 U.S. 837 (1984), but just as Loper Bright‘s overruling of Chevron did not require revisiting every decision that had applied Chevron deference, striking down RFRA need not disturb the outcomes of cases decided under it, as none of those decisions addressed RFRA’s constitutionality. Therefore, the stare decisis implications are mitigated.
Furthermore, by Justice Alito’s own concession in Fulton, the Court has recognized that stare decisis is “at its weakest when we interpret the Constitution because our interpretation can be altered only by constitutional amendment or by overruling our prior decisions[,]” Agostini v. Felton, 521 U. S. 203, 235 (1997), and that it applies with “perhaps least force of all to decisions that wrongly denied First Amendment rights.” Janus v. AFSCME, 585 U.S. 878, 917 (2018). Nevertheless, let us consider two stare decisis factors – in addition to Smith’s greater consistency with precedent discussed in III-A – which cut in favor of Smith and against Yoder, the latter of which must be overruled due to the breadth of its scope and evident inconsistency with Smith.
Workability. Yoder requires courts to make two judgments for which it is ill-suited. First, courts must determine the centrality of a practice to a religion in order to make a judgement on whether or not a government action substantially burdens that individual’s practice. The Court has recognized that “[i]t is not within the judicial ken to question the centrality of particular beliefs or practices to a faith, or the validity of particular litigants’ interpretations of those creeds.” Hernandez v. Commissioner, 490 U. S. 680, at 699. Second, it requires courts to make a policy judgement about whether or not granting an exemption will hinder the government’s ability to fulfill the interest at hand. This is a field in which courts lack expertise – see Smith, supra, at 889-890, “it is horrible to contemplate that federal judges will regularly balance against the importance of general laws the significance of religious practice.” Furthermore, there is significant dispute with respect to the meaning of the ‘hybrid-rights’ distinction applied to Yoder by Smith.[5] Therefore, the standard established in Yoder is an impractical one for courts to apply.
Reliance interests. The interests of the government with respect to Smith far outweigh those at stake in Yoder. Religious groups such as the Amish can seek exceptions from legislatures directly – this process has already demonstrated its effectiveness given the passage of RFRA and RLUIPA. The government, on the other hand, is exclusively reliant on Smithto prevent the onslaught of litigation that overruling Smith would invite – claimants could pursue exceptions from, inter alia, compulsory military service, the payment of taxes, health and safety regulation, and compulsory vaccination laws.[6]
* * *
The Court’s jurisprudence in this area is muddled, but the course most consistent with text, history, and stare decisis is not a novel one. It is simply to return to the standard laid out by Smith and clarify this standard by overruling Yoder and striking down RFRA. Any other path would not answer the question of what should replace Smith – it would only prove that nothing should.
[1] See United States v. Lee, 455 U.S. 252 (1982), requiring an Amish employer to pay Social Security taxes; Goldman v. Weinberger, 475 U.S. 503 (1986), rejecting a Jewish Air Force officer’s Free Exercise challenge to wear a yarmulke with his uniform; Bowen v. Roy, 476 U.S. 693 (1986), allowing the government to use Social Security numbers internally despite a religious objection; O’Lone v. Estate of Shabazz, 482 U.S. 342 (1987), where the Court upheld prison work schedules overriding Muslim prayer; and Lyng v. Northwest Indian Cemetery Protective Association, 485 U.S. 439 (1988), allowing the government to build a road through sacred land.
[2] Recall that the Free Exercise Clause was incorporated against the states in Cantwell v. Connecticut, 310 U.S. 296 (1940).
[3] While Gobitis was overruled, the underlying principle survives (e.g. through Reynolds).
[4] See Fulton, supra (slip op. at 25-26) (Alito, J., concurring in the judgement).
[5] At any rate, this attempted distinction is unconvincing, as Yoder primarily focused on the Free Exercise Clause rather than the parental rights issue.
[6] For further illustrations, see Smith, supra, at 889.