Religious Pluralism, the Nondelegation Rule, and Fulton v. Philadelphia

Abstract

In Fulton v. City of Philadelphia, the United States Supreme Court ruled that the city of Philadelphia could not exclude a Catholic Social Services agency from the city’s foster care system because it would not work with same-sex couples. Access to foster care agencies, however, is a public benefit that should not involve religious qualifications interposed by a nongovernmental organization. When aided by public funds, religious organizations should not receive exemptions from nondiscrimination laws to deny potential clients services because they disagree with these clients’ religious beliefs, sexual orientations, or marital status. First, equal access to public funds is not neutral when it allows religious entities to limit the life plans of others who do not share their convictions. Second, religious entities should certainly be able to advance religion but not on the public dollar. Finally, although exemptions for these entities augment the free exercise of religion, especially when accompanied by public funds, they also amount to what has been termed a “shadow establishment”.

1. Introduction

The rights of groups or private associations versus public policy have long been a subject of contention. On the one hand, group rights are often viewed as analogous to individual rights. Groups should therefore enjoy maximum autonomy in a pluralist society as bulwarks against state power that is the primary threat to freedom. On the other hand, many contend that intermediate associations should ideally aspire to reflect the values of the state in their internal governance. Only the state is truly rational and committed to justice. As put by Jacob Levy, “From one perspective, the pure theory [of free association] and the congruence theory both treat groups as one of the more familiar categories of political philosophy: as big individuals or small states. From another, both ignore groups to focus attention on the state: the state as a unique threat to liberty, or the state as the unique locus of rationality and justice” (Levy 2015, p. 54; see pp. 47–55; see also Rosenblum 1998, pp. 36–41).

Currently, controversies involving religious groups and associations have increased in prominence. As religious diversity has increased in the United States and the law has become more secularly oriented, a number of conservative Christians have come to think of themselves as members of a persecuted minority, some sincerely and some for political reasons. Although Christians form a majority of religionists, in some quarters, debates about the rights of cultures have shifted from the rights of minorities to those of majorities. Legal developments protecting minority rights can block particular majority demands, while majorities feel vulnerable to demands that they adapt to minority rights and interests. “Yet, if minorities have a right to preserve their religious identities, why should majorities, on equal conditions, not have such a right?” (Koopmans and Orgad 2023, p. 23; see also pp. 15, 22–28).

A longtime cultural or religious majority, for example, may experience a diminution of its influence and seek to strengthen its ascendancy. As described by self-identified Christian conservative but incisive critic Paul Miller, Christian nationalists in the United States today combine the idea that nations depend upon a strong predominant culture with a defense of the dominant Anglo-Protestant or European Christian culture. On their view, this culture was an essential precondition for nation formation and is a necessary condition for its continuing sustenance (Miller 2022, p. 59; see also pp. 30–37). They “believe that Christianity should keep its historic place as the unquestioned and dominant cultural framework in the United States, and that the government should actively construct and defend that cultural framework” (p. 60; see also pp. 59, 84, 145). To put this differently, “Nationalism is an invitation to take one group’s identity and make it an official state creed … [and is] the effort by social movements to gain the prestige and resources of a government, to use that government to solidify its privileges and pursue its goals” (p. 199; see pp. 198–99, 71–77; see also Stewart 2019, pp. 5, 3–11).

In recent years, the religion clauses of the First Amendment, historically difficult to balance, have been interpreted to favor the free exercise of religion over the prohibition on a government establishment of religion. It is past time that a better balance is restored. In this paper, I propose a particular interpretation that may be useful (Loucks 2022). In Larkin v. Grendel’s Den (459 U.S. 116 [1982]), the United States Supreme Court struck down a Massachusetts Law that awarded to the governing bodies of churches and schools a veto over liquor license approvals for any business within five hundred feet of their locations. Because the government delegated public authority over licenses to religious entities, the law violated the establishment clause. This nondelegation rule could have far-reaching application against the religious use of public authority.

Here, I focus upon Fulton v City of Philadelphia (141 S. Ct. 1868 [2021]), in which the Supreme Court ruled that the city of Philadelphia could not exclude a Catholic Social Services agency from the city’s foster care system because it would not work with same-sex couples. Access to foster care agencies, however, is a public benefit that should not involve religious qualifications interposed by a nongovernmental institution. If religious organizations receive dispensations from antidiscrimination laws, they are positing a religious test, screening out beneficiaries based on their religious beliefs, sexual orientations, or marital status—and all of this with the aid of public funds. Equal access to public funds is not neutral when it allows religious entities to limit the life plans of others who do not share their convictions. Churches should certainly be able to advance religion, but not on the public dollar. In augmenting the free-exercise rights of religionists, this trend promotes what has been termed a “shadow establishment.”

2. Background

The practices of adoption and foster care agencies have recently become a focus of controversy. Prior to the first Trump administration, religious organizations that accepted government contracts prohibiting discrimination based on sexual orientation were not automatically exempt from these rules, just as they had to follow rules against discrimination based on race, sex, or national origin. In 2019, however, Trump’s Department of Labor (DOL) issued a rule allowing a religious exemption for both government contractors and closely held businesses “so long as they could point to a religious purpose that is a public part of their mission and demonstrate that they exercise their religion in order to further that purpose.” This exemption, for example, could cover religious conduct codes for employees such as those prohibiting sex outside of heterosexual marriage, or turning away job applicants because of religious objections to their marriages. “According to the DOL, the rule is to be construed in favor of the broadest possible protection for religious interests” (Rogers 2019, p. 116; see pp. 113–18).

Generally regulated by the states rather than by the federal government, in most states, adoption agencies must be licensed. In essence these agencies are deputized by the government to supply services that otherwise, at least in modern times, the government itself would need to provide. License requirements may allow states to prohibit discrimination against same-sex couples, but some states have passed exemptions permitting religious adoption agencies to reject same-sex couples wanting to adopt or foster children—but only if they do not receive public funding (Tebbe 2017, pp. 124–26). Because Catholic Charities does receive public funds, in 2006 Catholic Charities of Boston terminated child placement altogether rather than place children with same-sex couples. Catholic Charities of Illinois, however, terminated its relationship with the Catholic Church in 2011 to continue child placement on a nondiscriminatory basis. Religious adoption agencies have typically not been permitted to exclude other couples—remarried or interfaith, for instance—on theologically based grounds. But why should they be able to exclude same-sex couples, one might ask, even if they do not receive public funds? (pp. 136–37; Ball 2017, pp. 225–26; Warbelow 2019, p. 438). Religious advocates, however, argue that if faith-based entities must choose between their religious commitments and executing their missions, fewer children and families will receive services. On the other hand, religious entities may welcome public funds, such as school vouchers, for ostensibly secular purposes, as they can then divert funds formerly spent on education to advance their religious missions. Money is fungible.

Public funds that support adoption agencies may raise questions under the First Amendment’s establishment clause. The First Amendment forbids the government to make religious judgments or to provide individuals or groups with religious experiences but must nevertheless “advance norms that protect religious freedom and religious minorities in the same way they protect secular causes and groups” (Lupu and Tuttle 2014, p. 176). The question then becomes how to determine whether the government is providing religious experiences to individuals or groups along with funds that it may contribute.

One answer to this question appears in the five-decade-old Supreme Court decision of Lemon v. Kurtzman (403 U.S. 602 [1971]). Laws must promote a secular purpose, they cannot hinder or advance religion, and they must not lead to an excessive entanglement between government and religion. For practical purposes, Lemon was effectively discarded in Kennedy v. Bremerton School District (597 U.S. 507 [2022]), in which the Supreme Court upheld the right of a high school football coach in Washington State to pray after games at the fifty-yard line, despite the lack of a secular purpose and the fact that team members experienced subtle coercion to join their coach in prayer. To many, Lemon meant that benefits were available to nonreligious individuals and entities that were withheld from religious ones. In recent years, this issue has arisen most frequently in the context of public funds through vouchers that parents may direct to private schools, either secular or religious. “The state is constitutionally responsible for religious experience when it creates financial incentives that will steer its citizens into such experience even if they would choose to avoid it but are willing to tolerate it to attain some other valuable good” (Lupu and Tuttle 2014, pp. 100–1).

This point, however, applies to other public policy areas also. For example, InnerChange Freedom Initiative (IFI) was an explicitly Christian prison program in Iowa that provided participating prisoners with greater family contact, help with parole boards, and other material benefits. The only program meeting state requirements that applied for the proffered funds, religious or secular, it called for prisoners to abandon “false” religions and “sinful” sexual orientations. The program was optional for prisoners, but the benefits provided incentives not otherwise available. IFI was eventually struck down for violating the establishment clause (Americans United for Separation of Church and State v. Prison Fellowship Ministries (509 F. 3d 406 [8th Circuit 2007]); see also (Lupu and Tuttle 2014, pp. 106–9)). As Andrew Koppelman notes, “Yet this result came about through a process in which the state made no religious decisions, acted on the basis of generally applicable laws that were facially religion-neutral, and pursued indisputably secular purposes at every stage in the process” (Koppelman 2013, p. 110; see pp. 110–14).

From the viewpoint of Lupu and Tuttle, the government was providing religious experience through IFI. Why can the government not utilize religious experience to bring about desirable secular ends? They believe that vouchers, which provide indirect aid, are legitimate, but otherwise, “The state lacks competence to choose and control religious experience, and it threatens both its own secular identity and the jurisdictional realm of faith when it coordinates with religious authority in an attempt to maximize social outcomes” (Lupu and Tuttle 2014, p. 111; see pp. 119–12, 75). If it turned out that prisoners visited by particular prison chaplains were more compliant than those interacting with other chaplains, prison wardens presumably would not discourage other chaplains from visiting prisoners.

An additional lesson from IFI is that although the Iowa state program requirements were facially religion-neutral, they affected the religious choices of prisoners who were not Christian and/or not religious. Equal-access/equal-treatment advocates maintain that if a generally applicable law makes it more difficult for religious adherents to follow their faiths, their religious exercise has been hindered (Monsma 2002, p. 266). Although the public provision of the widest possible range of alternatives is not required, it would have been preferable for Iowa to reject IFI when no alternatives existed that offered material benefits. Benefits to and burdens upon religious exercise are determined not by laws’ intentions but by their effects. In an early equal protection case, Yick Wo v. Hopkins (118 U.S. 356 [1886]), for example, the Supreme Court invalidated a San Francisco ordinance that laundries must function only in brick or stone buildings. This regulation was theoretically neutral, but most Chinese operated wooden laundries. Regardless of intent, then, the ordinance was discriminatory in effect.

The government has historically accepted the constitutional prohibition on legislating solely on religious grounds. “What a state may not do … is declare any particular religious doctrine to be the true one, or enact laws that clearly imply such a declaration of religious truth” (Koppelman 2013, p. 84; see also pp. 90–91). This makes it difficult, however, to protect the free exercise of religion solely through neutral and generally applicable laws. Laws with secular justifications may incidentally impinge upon religious practice, sometimes positively and sometimes negatively. Alternatively, they may burden some religious practices but benefit others, as in the case of IFI. The secular purpose requirement in Lemon allowed for laws that benefit religion indirectly, but policies could not be created solely based upon religious justifications. Specifically, “There would be little left of the Fourteenth Amendment, since most forms of discrimination that the amendment forbids have at one time or another been sincerely defended on religious grounds” (Koppelman 2013, p. 96; see pp. 84–98). In the present context, the government deputizes private adoption and foster care agencies for the secular purpose of caring for children. It need not and should not, however, allow this function to be executed on the basis of sectarian religious principles.

Although Lemon’s secular purpose requirement allowed laws that surreptitiously aided religion, it was necessary to prevent the creation of policies that are based on religious justifications alone. Some public policies take shape because they are deemed to promote the common good irrespective of whether they benefit religious practice. One example appears in McGowan v. Maryland (336 U.S. 420 [1961]). Here, the Supreme Court agreed with Maryland that a uniform day of rest was good public policy from a secular perspective. Therefore, the establishment clause was not at issue as with mandatory Sunday closing laws. Gordon Babst argues that such laws “have successfully hidden the sectarian rationale for their existence behind a putatively legitimate governmental purpose” and therefore do not fulfill a publicly justified secular rationale (Babst 2002, p. 78; see pp. 72–76, 57–58). Therefore, such laws constitute a “shadow establishment” (2). The growing “shadow” expanding the scope of the free exercise of religion is also well suited to the aims of Christian nationalism.

3. Fulton v. City of Philadelphia

In 2021, the Supreme Court ruled that the city could not exclude a Catholic Social Services (CSS) agency from the city’s foster care system because it would not work with same-sex couples. This exclusion violated the free-exercise clause of the First Amendment. CSS maintained that its policies accorded with its rights to religious freedom and free speech, maintaining that it was subject to religious hostility. An ACLU lawyer noted, however, that if foster care agencies are allowed to reject otherwise-qualified families for religious reasons unrelated to their ability to care for children, thousands of children may fail to receive proper care (Liptak 2020).

Lurking behind the presenting issue in Fulton was the status of the prior case of Unemployment Division v. Smith (494 U.S. 872 [1990]). Under Smith, neutral and generally applicable laws may burden religious exercise if they are applicable to all individuals and groups whatever their religious affiliation or lack thereof. Although Philadelphia’s contract with foster care agencies prohibits discrimination, including that based on sexual orientation, the contract allows discretionary exceptions, a point that the city failed to address in this case. The agency will not certify unmarried couples or same-sex couples even if married, although it is open to single parents, both gay and straight (Liptak 2021, p. A16). Although religious conservatives hail the virtues of the two-parent family, it seems that it must be the right kind of two-parent family. Under Smith, if public policy “singles out religion either for special benefits or special liabilities” (Monsma 2002, pp. 265–66), it is not espousing neutrality between religious and nonreligious endeavors. If either benefits or burdens accrue to religious practice based on generally applicable laws, this is despite a practice being religious, not specifically because it is religious. It is the law’s effect, not its intention, that results in benefits or burdens. Under this rubric, requiring CSS to follow the city’s nondiscrimination policy is neutral because it applies equally to religious and nonreligious social service agencies.

For Smith’s critics, however, neutrality means equality of treatment for religious and nonreligious endeavors. When organizations engage in activities that promote the public good such as schools or social services, the government cannot exclude them from generally available benefits based on their religious affiliations. “Equal treatment theory would allow individuals and groups to participate fully and equally with their fellow citizens in America’s public life, without being forced either to shed or disguise their religious convictions and character” (Esbeck 1998, p. 12; see pp. 12–14). Putting this more broadly, “The pluralist approach encourages communities of conscience to preserve the institutions necessary to perpetuate their distinctive ways of life and to pass these on to future generations” (McConnell 2000, p. 105; see pp. 100–6). True neutrality requires attention not only to the intentions behind a law or public policy but also to its consequences. If a generally applicable law “make it harder for a person of devout faith to follow the tenets of his or her faith, that person’s free exercise of religion has been hindered” even if no such intent exists (Monsma 2002, p. 266). Under this interpretation, the city’s nondiscrimination policy is not neutral because it effectively burdens CSS where it would likely not burden a nonreligious agency. Equal treatment policies can be problematic, however, because religious groups receive religious exemptions that secular organizations are denied but also equal access to government funding that secular organizations receive (Gill 2019, pp. 25–31, 69–98).

Supreme Court Justice John Roberts wrote the opinion for the court in Fulton (Fulton v. City of Philadelphia, (141 S. Ct. 1898 [2021], 1874–1883), and Justices Samuel Alito, Neil Gorsuch, and Clarence Thomas joined only in the judgment, not in the opinion. Both the district and circuit courts had ruled that the city’s nondiscrimination requirement was neutral and generally applicable under Smith. Roberts noted that CSS could direct same-sex couples toward over twenty other Philadelphia agencies that would certify them. More important, Smith did not apply because the city provided for individualized exemptions at the stage when a child was actually being placed with a foster family; therefore, the policy was not neutral or generally applicable. Additionally, the availability of foster children through the city contract could not be deemed a public accommodation governed by nondiscriminatory policy, because foster-parent certification depended upon meeting certain conditions and was thus not generally available to the public.

The most interesting feature of Roberts’s opinion was that instead of using Fulton as a vehicle for overturning Smith, as many conservatives had hoped, instead, he relied on Church of the Lukumi Babalu Aye v. City of Hialeah (508 U.S. 520 [1993]). The Afro-Cuban religion of Santeria requires that animals be occasionally sacrificed, cooked, and eaten. In Hialeah, Florida, the city council forbade the ritual killing of animals not primarily designated for consumption. The council’s discussion displayed hostility toward Santeria. The Supreme Court overturned the ordinance because although the killing of animals was a permitted practice, the council prohibited it when associated with a religious ritual. Allowing this ritual does not single out Santeria; rather, its practitioners are on a level with anyone who kills and eats animals for any reason. “The free exercise clause … bars the government from prohibiting an activity because it is religious; the law cannot forbid an activity that is part of a religious practice and leave unregulated identical or closely similar activities that are not religious” (Greenawalt 2012, p. 49). John Locke makes the same point in reverse. Animal sacrifice should not be forbidden as a religious rite if animals may be killed for food. If health reasons forbid the killing of specific animals for food, however, religious sacrifice of animals may be forbidden as well (Locke [1689] 1983, p. 42). Religious beliefs are not illegitimately burdened in this context.

The same year as Lukumi, liberals and conservatives in Congress coalesced in the passage of the Religious Freedom Restoration Act (RFRA). Liberals were concerned about religious minorities who might be denied exemptions from neutral and generally applicable laws, while religious conservatives wanted to utilize the political process to lessen Smith’s burden on religious freedom (Horwitz and Tebbe 2016, p. 220; see also pp. 206–11). Under RFRA, only a compelling state interest using the least restrictive means allows a law to impose a substantial burden on the free exercise of religion (Flanders 2016, p. xx). Varied interpretations may attach to the definitions of compelling state interest, least restrictive means, and substantial burden. Putting this differently, conservatives have turned their attention from how the government might aid religion despite the establishment clause to how they might interpret the free-exercise clause to accommodate their beliefs and practices. “Rather than focusing on the public benefits of religion as a source of social cohesion and integration, more recent proponents of a more active political role for religion have suggested the need for greater religious autonomy and freedom” (Jelen 2000, p. 94). RFRA enabled Smith’s opponents to secure religious exemptions from laws that could be interpreted as neither neutral nor generally applicable.

Crucial here is the fact that in overturning the Hialeah ordinance, the court did not grant an exemption to Santeria. The court simply removed the city’s burden on the permitted practice of killing animals, rather than conferring a benefit on a religious practice. If the government had secular reasons for prohibiting the killing of animals—species endangerment or public health—both secular and religious practices would be affected. But such a civil regulation would affect religious practice only incidentally. The Hialeah ordinance singled out religious practice while exempting hunters and butchers. Therefore, it was not neutral and generally applicable and was therefore subject to strict scrutiny. Accordingly, the city would have had to demonstrate a compelling interest in the prohibition and would need to use the least restrictive means to achieve its aims. Similarly, to the court in Fulton, the city neither demonstrated a compelling interest nor used the least restrictive means to effect it.

Religious hostility or animus toward religious convictions was also the justification for the court’s decision in Masterpiece Cakeshop (Masterpiece Cakeshop v. Colorado Civil Rights Commission, 138 S. Ct. 1719 [2018]). Here, the court ruled narrowly for a Christian baker in Colorado who refused to design a cake for a same-sex couple celebrating their marriage because of the religious hostility displayed by one commissioner. The court did not address the larger question of whether the Colorado nondiscrimination ordinance governing public accommodations was applicable against business owners with religious objections. Therefore, it did not address Smith’s conclusion that neutral and generally applicable laws may burden religious exercise if the burden is incidental to the law’s purpose. Similarly, in Fulton, the court did not address the legitimacy of Philadelphia’s nondiscrimination policy in the face of foster care agencies with religious objections to same-sex couples. Instead, it focused upon the city’s contract inclusion of the availability of individualized exemptions in certain cases. In Roberts’s view, then, the city ordinance did not constitute a neutral and generally applicable law, and Smith did not apply.

In Fulton, the city’s lawyer had pointed out that the certification of foster parents was more complex than Roberts portrayed. In the standard city contract, foster care agencies identified individuals who qualified as potential foster parents. No foster family or child could be rejected because of race, religion, or sexual orientation. Only later, when specific children were placed in particular homes, might the best interests of a child require placement in one type of home over another. For example, in one case, a foster child had used a racial slur, and CSS avoided placing that child with foster parents of the race that had been disrespected. Roberts focused on that second stage, but CSS objected to the nondiscrimination requirement pertinent to the first stage (Greenhouse 2021, pp. 209–11; see also pp. 77–79). Focus on the first stage would mean that the nondiscrimination policy was neutral and generally applicable and that Smith would apply. Focus on the second stage meant that the policy was neither neutral nor generally applicable, and that therefore, it did not apply.

Although the Fulton decision was a narrow one, its implications are broad. Some of the justices wanted to use it to overrule Smith, but the case demonstrated that this would be unnecessary to ensure that religious exercise is not burdened. In both Lukumi and Masterpiece Cakeshop, a finding of hostility to religion is sufficient irrespective of whether a policy is or is not neutral or generally applicable. “If CSS could prevail, despite Smith, on the specific facts of this case, what would be the court’s justification for taking the giant step of overruling the precedent?” (Greenhouse 2021, p. 79; see also pp. 80–83, 212–15). Moreover, the imputation of hostility to religion to one party in a conflict could be used in any number of disputes in which the presence of hostility might be a matter of perception and therefore subjective. Does “targeting” religious exercise, which Justice Alito discussed in his concurrence in Fulton, refer to the terms of a law itself, or may it also refer to the motivations of those who create or apply the law? (Fulton, 1918–1922) The current Supreme Court’s tendency to favor free-exercise claims need not be grounded on the granting of individual religious exemptions.

Justice Amy Coney Barrett wrote a separate concurrence joined by Justices Brett Kavanagh and Stephen Breyer (Fulton, 1882–1883). Despite disliking Smith, she suggested that there was no need to consider overruling it in view of the plethora of questions that would arise concerning its replacement. Free-exercise jurisprudence dictated that any law burdening the free exercise of religion required strict scrutiny if government officials could issue individualized exemptions as in Philadelphia’s foster care contracts. Because in her view, the contract could not survive strict scrutiny, its terms could be equated with hostility toward religion.

Justice Alito wrote a separate concurrence joined by Justices Clarence Thomas and Neil Gorsuch (Fulton, 1883–1927), focusing on the court’s regrettable disinclination to overrule Smith. In the present circumstance, the city would predictably abandon its ability to grant individual exemptions and would then argue that it was protected by Smith as the lower courts had ruled. CSS would again challenge Smith in future disputes, and the Supreme Court would be back where it started. So why not overrule Smith immediately? (Fulton, 1887–1888).

Alito summarized religious exemption cases such as Sherbert v. Verner (374 U.S. 398 [1963]), Wisconsin v Yoder (406 U.S. 205 [1972]), and Thomas v. Review Board (450 U.S. 707 [1981]), concluding that “[a] regulation neutral on its face may, in its application, nonetheless offend the constitutional requirement of governmental neutrality if it unduly burdens the free exercise of religion” (Yoder, p. 220), even if its substance did not discriminate against any religion. Suggestive of the problem with IFI, a policy may exercise a disparate impact upon religious exercise even absent this intent (Fulton, 1889–1992). Alito relied on the purported ordinary meaning of the First Amendment: prohibiting the free exercise of religion has always meant “forbidding or hindering unrestrained religious practices or worship” (Fulton, 1896).

On Alito’s interpretation of the Smith court, religious practice apparently could not be restricted unless the same conduct were restricted for everyone else. According a particular right to a specific group of people—those who exercise religion—is not related to the rights of individuals outside this group. Applying this point to Lukumi, Hialeah could forbid the Santeria practice of animal sacrifice only if it also forbade the killing and eating of animals for any reason whatsoever. Applied to Masterpiece Cakeshop, Colorado could forbid discrimination based on sexual orientation only if it forbade any type of discrimination under any circumstances. Under Smith, moreover, the troublesome question arose of when facially neutral laws in fact “targeted” religious exercise. Pinning down subjective motivations is difficult and may itself be subjective (Fulton, 1918–1922). Finally, for Alito, CSS’s policy of rejecting same-sex couples was simply an expression of the conviction that marriage should only comprise a man and a woman. “In an open, pluralistic, self-governing society, the expression of an idea cannot be suppressed simply because some find it offensive, insulting, or even wounding” (1924; see 1924–1926). Religious freedom cannot survive unless this same principle also applies to religious practices. Therefore, Alito would have overruled Smith in ruling against the city of Philadelphia.

In a concurring opinion joined by Justices Thomas and Alito, Justice Gorsuch also argued that the court should have overruled Smith (Fulton, 1926–1931). The district court had ruled that agencies such as CSS qualify as public accommodations, which is why it found against CSS for discrimination. But as we have seen, Roberts’s opinion concentrated not upon the first stage of the process, in which any potential foster parents might be considered, but upon the second stage, when the agency places a specific child with a particular family. Because CSS does not consider same-sex couples, it contravenes Philadelphia’s antidiscrimination policy at the outset. At the second stage, by contrast, individualized exemptions were permissible, which allowed the court to rule that Smith did not apply. Like Alito, however, Gorsuch concluded that by ruling on narrow grounds both here and in Masterpiece Cakeshop, the court had opened itself to new rounds of litigation that could easily have been prevented if it had stepped up and overruled Smith at this opportunity.

4. Public Accommodations and Specific Cases

Although Justice Roberts stated that CSS was not a place of public accommodation open to all comers, as he focused upon the second stage, that of the placement process, Justice Gorsuch deemed that it was a public accommodation because the city contract required that at the outset all aspiring foster parents be considered without regard to religion, race, or sexual orientation. It was this policy that CSS had challenged. If CSS is a public accommodation, the city’s nondiscrimination policy was a neutral and generally applicable law, and Smith would have applied as the lower courts had ruled. CSS would have lost. Therefore, to grant CSS its desired exemption, the court would have needed to overrule Smith. To avoid this move, Roberts used Lukumi as a vehicle to rule narrowly for CSS.

According to Chris Gottlieb, however, both Roberts and Gorsuch were incorrect about whether CSS was a public accommodation in the sense of being accessible to all interested parties. Certifications as foster parents are not available to everyone who might apply, as the purpose of the certification process is to determine if given individuals are appropriate applicants. Foster care is, however, an accessible service in the sense that all children in state custody are entitled to it, and any parent whose child enters foster care is entitled to its support (Gottlieb 2022, pp. 49–59). Additionally, in my view, foster care agencies are public accommodations in the sense that the opportunity to apply as a potential foster parent is available to any interested party—or at least ought to be available although CSS disagreed. Moreover, adoption agencies are not private businesses such as restaurants and hotels, and they are often supported by public funds.

Although Fulton is the most prominent recent case concerning adoption agencies, it should be noted that some adoption agencies are making distinctions among potential clients based on these clients’ religious beliefs as well as their sexual orientations. In 2019, South Carolina’s Miracle Hill Ministries, a large foster-family provider, received a waiver from a previous rule that federally funded adoption agencies must accept all families. Miracle Hill families must “sign a doctrinal statement that marriage is between a man and a woman, and refuting the existence of transgender people altogether.” LGBT children may be fostered by families who deny their identities or may languish in institutional care without a family (Boylan 2019). One aspiring foster mother, Aimee Maddonna, was rejected because she is of the “wrong” religion—Catholic—and the agency told her that it only works with evangelical Christians. This case was litigated in both state and federal courts (Aimee Maddonna v. U.S. Dept. of Health and Human Services, 6–19-cv-3551-TMC [2020]; Boston 2020). In 2023, however, a federal court ruled against Maddonna after five years of litigation (Laser 2023).

In 2021, Michigan-based Bethany Christian Services (BCS), the largest Protestant adoption and foster agency in the U.S., announced that henceforth, it would work with LGBTQ people nationwide. Previously, some branches had agreed not to discriminate but others had continued to do so (Graham 2021; Church & State, 2021). A woman in East Nashville, Tennessee, lesbian Kelly Easter, wanted to provide a foster home for refugee children and had reached out to the Office of Refugee Resettlement (ORR) of the U.S. Department of Health and Human Services in 2020, which directed her to BCS. Learning in 2021 that BCS would now work with her, Easter was crushed to learn that BCS was not independent but was a subgrantee of the U.S. Conference of Catholic Bishops (USCCB). Therefore, BCS was prohibited from working with Easter despite its own change in policy. After a federal lawsuit was filed, however, USCCB decided that it would not object to its grantees working with a single lesbian foster parent (Church & State, 2022b). Along related lines, a same-sex couple in Texas was rejected by Catholic Charities Fort Worth because their marriage “failed to mirror the Holy Family” (Hayes 2021, pp. 7–8; Marouf v. Azar, 391 F. Supp. 3d 23 D.D.C. [2019]), although some changes were later made in the program that rendered the case moot (Marouf v. Beccera, No. 1: 18-cv-378 APM [2023]). Apparently USCCB thought that a single lesbian was preferable to individuals in a state-sanctioned same-sex relationship.

Finally, in 2022 a Jewish couple filed suit against the Tennessee Department of Children’s Services because of its funding of the Holston United Methodist Home for Children, which had turned down their adoption application on learning that these prospective parents “did not share our belief system” (Hauser 2022). Medication for a condition that might cause birth defects made biological reproduction inadvisable. The couple found an adoption-ready child with a disability in Florida and applied to the only agency in their area that handled out-of-state placements. They filled out the paperwork and prepared their home, but they were advised on the very day they were to begin their training that their religion prevented adoption through Holston. Although a state court denied their standing to sue in 2022, a state appeals court ruled in 2023 for their right to sue on grounds that religious discrimination rendered the Rutan-Rams without access to services available to Christians (Rutan-Ram v. Tennessee Department of Children’s Services, No. 22–80 III [2023]; Hayes 2022; Church & State, 2023). In 2024, the Tennessee Supreme Court denied the state’s appeal for review, which enables the case to proceed on the merits to the trial court.

In 2021, the Biden administration’s Department of Health and Human Services withdrew exemptions granted by the first Trump administration that allowed the use of religion by child welfare providers to discriminate against foster parents (Church & State, 2022a). Welfare providers who discriminate based on religion, especially recipients of public funds, are subject to the same objections raised by James Madison. “Who does not see that the same authority which can establish Christianity, in exclusion of all other religions, may establish with the same ease any particular sect of Christians, in exclusion of all other Sects?” (Madison 1785, p. 607). The case of Aimee Maddonna illustrates this exact point. The narrowest type of exemption for religious institutions refers to the ability to hire or retain employees based on religious faith, meaning those of the same religion as that grounding the institution. If an institution can require that its clients also subscribe to its faith, this broadens the exemption considerably. Beyond this, an institution may insist that its clients share the same interpretation of its faith that it holds itself. Maddonna was found wanting not because she was not Christian, but because she was a Catholic rather than an evangelical Christian.

The rejection of LGBT clients raises similar issues. A great many Christians believe that LGBT individuals are excellent parents. Christians who reject this belief, therefore, are subscribing to a particular interpretation of Christianity not shared by all Christians or by all other persons of faith. An agency with strict religious tests can serve only a small range of potential clients. Even if a great many agencies exist, if each subscribes to narrow criteria when screening clients, the situation becomes untenable. In Rutan-Ram, for example, the Jewish couple who applied to Holston did so because it was the only agency in their area that processed out-of-state placements. The problem of alternatives arose in the implementation of President George W. Bush’s faith-based initiative and is also exemplified in IFI as discussed above. Although social service programs receiving public funding could include religious content, secular alternatives were supposed to be available. In some situations, individuals needing aid had to turn to the closest available source for lack of information or transportation. Some religious clients might be forced to use a secular program to receive aid or might reluctantly choose a secular program rather than a religious one they disliked. Mandating alternatives does not mandate the provision of every possible alternative. “You can expect a lot of Baptist programs in Texas and a few Mormon ones, and the opposite in Utah. So what does equality mean under these circumstances?” (Goodstein 2001, p. A18).

In theory, pluralism exists as a range of options among which potential clients may choose. But if each one is exclusive, even if exclusive in different ways, clients may face undue burdens as they comb through these choices to find agencies accepting of their own identities. Faith-based agencies that exclude should not receive public funding, whether the exclusivity is based on religion, race, national origin, sexual orientation, or gender identity. In fact, in 2024 the Biden administration issued a rule that prohibits federally funded health providers from discrimination based on sexual orientation and gender identity (Weiland 2024). It also finalized rules reinstating religious freedom protections for individuals using federally funded social services such as homeless and domestic violence shelters, food banks, job training, and elder care (Church & State, 2024). In all of these contexts, provider-imposed religious restrictions limit the life plans of countless other individuals. Because they do not fit the religious criteria of these providers, these restrictions render them second-class citizens and amount to a shadow establishment of religion.

5. Establishment Issues

In Lemon v. Kurtzman (403 U.S. 602 [1971]), laws required a secular purpose, they must not have advanced or hindered religion, and they must not have led to excessive entanglement between government and religion. Public funding of adoption agencies carries a secular purpose, that of placing children without parents or guardians with foster or adoptive parents. In Fulton, because the city of Philadelphia was not itself using religious tenets to discriminate against same-sex couples, one might argue that it had fulfilled its obligations and that it was not responsible for CSS’s policies. Similarly, prison authorities in Iowa probably concluded that because it had considered all applicants for its prison program, they could not be blamed for the fact that only IFI met state requirements. Although varying interpretations exist as to the relationship between Lemon’s second and third prongs, establishment issues may exist even when the government is not directly involved.

In Larkin v. Grendel’s Den (459 U.S. 116, 123–126 [1982]), the Supreme Court struck down a Massachusetts law that gave the governing bodies of churches and schools a veto over liquor license approvals for any business within five hundred feet of their locations (117). Because the government had delegated public authority over liquor licenses to religious entities, the law violated the establishment clause. “In short, the law permitted religious institutions to determine which otherwise eligible applicants could and could not receive a state license” (Loucks 2022, p. 2281; Grendel’s Den, p. 125). First, the law was not neutral and required no accountability from churches. Second, the law advanced religion by implying public support for these religious institutions. Third, it entangled churches in a public decision-making process (pp. 120, 125–27). The court affirmed this religious nondelegation rule in Board of Education v. Grumet (512 U.S. 687 [1994]), holding that a New York law that established a school district congruent with the borders of Kiyas Joel Village for the benefit of Satmar Jewish children needing special education granted favored status for a particular religious sect (696–705; see also Loucks 2022, pp. 2281–87).

In Fulton, suggests Andrew Loucks, the court did not consider the impact of the nondelegation rule (Loucks 2022, pp. 2289–300). Although the Supreme Court has not defined the scope of this rule, “Nonetheless, the religious nondelegation rule embodies a critical bulwark against religious involvement in government,” and “clearly applies to contexts in which religious organizations directly control government functions with their religiously motivated preferences. Fulton v. City of Philadelphia was such a case” (p. 2300). CSS is a religious institution under the authority of the Philadelphia Archdiocese, even if the rule is narrowly interpreted to apply only to churches. Moreover, the delegation of foster parent certification to private agencies is neutral in that it applies to both religious and nonreligious agencies. Unlike issues of the possible provision of religious experience of indoctrination that arise in the provision of public funds to private schools, Grendel addressed “a law that allowed religious institutions to wield the government’s power to determine a third party’s access to government licenses” (p. 2303). These are public benefits that should not involve religious qualifications interposed by a nongovernmental institution. The Fulton court considered the free-exercise rights of CSS but not interference with the free-exercise rights of its clients (pp. 2301–8). As put by Zalman Rothschild (2022), “The more CSS emphasized its religious identity, religious values, and status as a corporate branch of the Catholic Church—all of which supported its free exercise claim—the more it laid the foundation for a separate constitutional challenge, not against the city’s nondiscrimination law, but against the city allowing CSS to operate as a foster agency in the first place” (2022)—at least with public funds.

From an opposing perspective, Robert Destro presents an interesting, but I believe misguided, argument for allowing religious entities receiving public funds to operate in line with their religious convictions (Destro 1998, p. 107). He cites the Constitution’s Article VI: “No religious test shall ever be required as a Qualification to any Office or public Trust under the United States.” Although he applies this clause to religious organizations’ ability to hire as they choose, his argument carries broader implications. A religious test comprises any screening device that excludes individuals from a public trust, like the administration of a government grant or contract, based on religious belief or practice. Atheists and agnostics are neither more nor less likely to abuse any trust placed in them, and we should not prefer them to religious believers (pp. 119–23). Destro argues, moreover, that second-class citizenship for religious believers is forbidden by the Fourteenth Amendment’s privileges and immunities clause and the equal protection clause. Therefore, religious organizations should not be required to abandon their religious identities to obtain treatment equal to secular organizations, lest the government ignore crucial differences between religious and secular organizations. In sum, the use of either religious or nonreligious adherence “as justification for the disparate treatment of citizens and other persons should be viewed by the courts as presumptively invalid unless the government can prove that it has an unquestionably legitimate purpose for its policy” (p. 127; see pp. 124–35).

Destro’s argument seems logical from one perspective, but nevertheless, it is problematic. When religious organizations are accorded exemptions from generally applicable antidiscrimination laws in accordance with the desires of equal-treatment advocates, the government is itself discriminating against individuals and organizations disadvantaged by these policies. It is these religious organizations themselves that are applying a religious test, rejecting employees and possible beneficiaries because they reject their religious convictions, sexual orientations, or marital status. And these nonprofits are using public funds to do so, using their authority to discriminate based on reasons unrelated to the aims of the government grants or contracts that they have received. They are in effect conferring second-class citizenship on people of whom they disapprove. The government’s complicity is not sanitized because it is not directly engaged in this action. The nondelegation rule in Grendel’s Den affected churches, which do not receive public funds, and schools, which may have been public or private. But adoption agencies receiving public funds have even less justification for making religious distinctions while carrying out a government function. This point resonates with Locke’s caution in A Letter Concerning Toleration that although a religious entity is entitled to reject members who act contrarily to its principles, “No private Person has any Right, in any manner, to prejudice another Person in his Civil Enjoyments, because he is of another Church or Religion” (Locke [1689] 1983, p. 31; see pp. 30–31). Limiting the life plans of other people who do not subscribe to specific religious principles certainly falls under this rubric.

Nonprofits should be required to adhere to nondiscrimination policies in providing public services even if they do not receive public funding. “It is one thing …. to say that religious beliefs should determine religious care of the faithful, and quite another to say that they should determine the provision of non-profit goods and services for the general public, whether or not the state is subsidizing those services” (Lever 2017, p. 236, italics original; see pp. 236–37). If a nonprofit turns away those who do not accept its religious teachings, the government should provide alternatives, “if necessary by removing any subsidies it gives to agencies that discriminate on doctrinal grounds, and giving these to religious or secular bodies with more inclusive practices” (p. 237). Otherwise, although the government would not directly engage in religious discrimination, it is enabling religious organizations to do so. Destro believes that when the government enforces nondiscrimination laws, it is positing a religious test that indicates who shall be awarded grants and contracts. Destro, however, would allow the government to delegate the application of religious tests to subsidized nonprofits that can then apply their own. Although the government would not intend this outcome, equal-treatment advocates themselves remind us that attention not only to a policy’s intentions but also to its effects is necessary to neutrality toward religion.

In this context, Fulton focused on the fact that when children were placed with specific families, the city’s nondiscrimination law allowed adoption agencies to consider the best interests of a child in determining this placement. Because this sort of discrimination was allowed at this stage, discrimination among potential foster parents at the first stage in the process should be acceptable as well. These are different types of discrimination, however. At the second stage, it is the child’s welfare that is at stake. At the first stage, it is the religious convictions of the provider, the adoption agency, that are in play. Fulton allows the provider, CSS, to apply its own test unrelated to the mission of matching children with the most appropriate adoptive or foster parents.

The nondelegation rule potentially applies to other settings as well. Religious communities are often helpful to immigrants in adjusting to life in the United States by providing both communities ties and resources such as legal help and job leads. One avenue lies in the Temporary Religious Worker Visa (the R-1), under which immigrant members of religious nonprofits for at least two years may secure employment through their churches, synagogues, temples, or other religious nonprofits. Although the work is supposed to be religiously oriented, religious organizations have exploited the availability of this visa, mainly to provide labor. The problem is that the R-1 visa “application process is enabled through the religious organization, giving them the power to hold, or withhold, employment.” No minimum wage applies to these workers. Moreover, the R-2 visa is available for the spouse and/or children of these workers. R-1 workers may fear reporting possible abuses for fear of being deported along with their families (Americans United 2024). In these settings, religious organizations are wielding public authority with minimal oversight, if any, as to whether they are adhering to the terms of these visas.

The Supreme Court in 1987 upheld a broad right to church autonomy in all activities when it allowed the Church of Jesus Christ of Latter-day Saints to fire a janitor at an LDS gym for failing to qualify for temple activities. The court ruled that the 1972 amendments to Title VII of the 1964 Civil Rights Act exempted all citizens, not only religious ones, from restrictions on hiring. “A law is not unconstitutional because it allows churches to advance religion, which is their very purpose. For a law to have forbidden ‘effects’ …, it must be fair to say that the government itself has advanced religion through its own activities and influence” (Corporation of the Presiding Bishop v. Amos, 483 U.S. 327 [1987], p. 337, emphasis original). Although some justices worried about enhanced religious influence in secular contexts, they were also concerned that a detailed analysis of specific activities to decide on their religiosity would promote excessive entanglement with religion. Moreover, it might “chill” religious activity by inducing religious organizations to refrain from some activities that they considered religious lest they attract litigation. Therefore, religious nonprofits should be granted a “sphere of deference” encompassing activities that are likely religious (pp. 343–45).

Churches should certainly be able to advance religion but not on the public dollar. When a religious organization advances its beliefs with the help of public funds, it is indirectly allowing the state to participate in the advancement of religion. A seldom-noticed downside exists to the equal treatment regime. Neutrality as equal access is exemplified by Widmar v. Vincent (454 U.S. 263 [1981]), in which the Supreme Court ruled that a state university’s religious student organizations could not be denied access to publicly funded campus buildings for their meetings when secular student organizations were admitted. Widmar’s expansion of generally available benefits to religious institutions and individuals did not rely on the free-exercise clause, but rather on equal access. Some cases have relied upon the First Amendment’s free-speech clause (Rosenberger v. Rector of the University of Virginia, 115 S. Ct. 2510 [1995]. The court ruled that the university could not refuse to subsidize a student religious group’s publication when it had funded the publications of other student groups, as the refusal constituted viewpoint discrimination. Justice Souter observed, however, that the court had ignored the establishment clause by treating the case as a free-speech issue. This ignored the Constitution’s special treatment of religion. The private exercise of religion deserves protection but not by awarding public funds (Davis 1998, pp. 140–41; see pp. 137–42).

Derek Davis implies that ironically, sometimes, we must disadvantage religious practice in order to protect religion’s special status. As we have seen, equal-access advocates rely on the free-exercise clause because they believe that the free exercise of religion is denied to religious interests when they are excluded from opportunities available to nonreligious interests. But under this rubric, “It is difficult to see what jurisdiction, if any, would remain to the establishment clause other than forbidding an official establishment of a government-sponsored religion, or the express preference of one religion over another” (Davis 1998, p. 146). The establishment clause may be used to hamper the free exercise of religion, but a capacious interpretation of the free exercise clause may also vitiate the establishment clause. Davis concludes, “If equal treatment prevails, one’s religious beliefs would be protected equally and on the same level as one’s right to attend a particular university, live in a certain neighborhood, or express a particular political opinion. In short, one’s religious faith would be on a par with every other worldview and life belief,” and no rationale would exist for exemptions from taxes or other government regulations (p. 154; see pp. 145–57).

We may infer that for Davis, equal access for religious interests means equal treatment under the establishment clause, but because the courts ignore troubling establishment issues. When this is linked with special protections in free-exercise cases, as Rogers Smith notes, the combination does not promote equality, but instead “compound[s] the advantages already enjoyed by the religious viewpoints that are most widely shared and the most politically potent in the United States” (Smith 1998, p. 183). Smith’s point reminds one of the Christian nationalist desires to maintain European Christian culture as the unquestioned cultural framework of the United States. Classifying Rosenberger as a free-speech case means that “there is no stopping point to support religion other than that such support is also available to non-religious causes” (Joppke 2015, p. 119). Comparing Smith and Rosenberger, Christian Joppke continues, “Under legal secularism, religious minorities had been strongly protected under the Free Exercise Clause. Now the situation was the reverse. Minorities had become vulnerable under a diminished Free Exercise Clause, while majorities had their free go under an obliterated Establishment Clause” (p. 120; see pp. 118–20; Gill 2019, pp. 93–96). Ignoring troubling establishment issues in the attempt to protect free exercise is the very essence of the shadow establishment.

6. Free-Exercise Issues

As noted above, in Fulton, Justice Alito objected that under the Smith rule, neutral and generally applicable laws can impede the free exercise of religion even if this is not intended and is only an incidental effect. Specifically, he took issue with the idea, as he explained it, that religious practice could not be restricted unless the same conduct were also restricted for everyone else. A right accorded to a particular group of people—in this case those who exercise religion—should not affect or be affected by the rights of individuals outside this group (Fulton, 1896–1897). The rights of any group of people, however, are always comparative as these relate to the rights of others.

In the early part of this century, LGBT rights were much less favorably regarded than they are currently. A caveat often put forward was that legislation protecting the civil rights of those with a nontraditional sexual orientation amounted to granting special rights to LGBT individuals. These skeptics did not realize that the majority already enjoyed these rights, and that legislation was necessary only to put them on an equal footing with others. The same can be said of marriage equality. Committed nontraditional couples wishing to marry now have opportunities equal to those of traditional couples. Before marriage equality, the views of some opponents amounted to the idea that “expanding a right to a new group deprives the rest of us of our right to deny that right to others” (Chait 2009, p. 2). When Proposition 8 in California in 2008 rescinded the right to marry that same-sex couples had been granted months earlier, one argument “was that providing access to marriage to gay couples would reduce the rights available to others,” such as those of parents who wanted to shield their children from knowledge of same-sex relationships. “From this vantage point, an essential selling point of Proposition 8 was that it would protect people from the excesses of extending rights to gay couples” (Feldblum 2009, emphasis original; see pp. 221–28; Gill 2012, pp. 178–83). Certainly, the white supremacists who battled civil rights advocates tooth and nail after Reconstruction and during Jim Crow thought that rights accorded to persons of color diminished their own rights to deny them. Not only did Jim Crow curtail the life plans of persons of color, it also constituted an establishment of white supremacy. Although in southern states, it was the law and therefore not in shadow, in much of the country it was informal and therefore could be considered a shadow establishment.

These examples may seem extreme, but they illustrate the point. Smith ruled that Oregon could penalize members of the Native American Church for using peyote in religious ceremonies by denying them unemployment benefits, as the penalty was neutral and generally applicable to both religious and nonreligious users of peyote. Before RFRA, discussed above, the burden of proof fell upon religious adherents to show that a compelling state interest existed for permitting their religious practices. After RFRA, the burden was on the government to demonstrate a compelling state interest in curtailing a practice, and if this existed, the government must utilize the least restrictive means of doing so. RFRA changed the relationship between religious and nonreligious interests. Under equal access, religious groups could receive the same benefits as secular groups, but after RFRA, they could receive exemptions from generally applicable laws denied to secular groups. Well before RFRA, however, according to Michael McConnell, “The government must be religion-blind except when it accommodates religion” (McConnell 1992, p. 177, emphasis original; see pp. 168–69, 175–77, 184–86; Gill 2019, pp. 25–31).

Because of the special status of religious belief and practice in American culture, Samuel Marcosson suggests that legislative exemptions from generally applicable laws for granting public funds to religious groups are permissible provided that certain conditions are met. The government should “impose eligibility requirements that demand that participants be full members of the social contract that recognizes the full, equal citizenship of all persons protected by antidiscrimination laws.” Both religious and nonreligious groups should be eligible for government contracts and programs. But “Asking for the exemption represents a fundamental (and hence ironic) decision to foreswear the obligations imposed by those self-same values, and those who cannot do so cannot then turn around and demand to enjoy the benefits of our society’s commitment to neutrality and equality, as we put into practice the values of the Fourteenth Amendment’s” equal protection guarantee (Marcosson 2009, p. 151). A modicum of deference to religious liberty interests is acceptable, but constitutional values “need not and should not yield completely merely because the ‘religion card’ is played” (pp. 152–53). The observance of antidiscrimination laws should be a condition of eligibility for all organizations. From Marcosson’s perspective, a denial of public funds would be automatic for organizations that want exemptions from these very laws like CSS in Philadelphia.

Others argue, however, that despite the demand for special treatment, without it, “It’s obvious, on reflection, that … religious organizations would lose the right to define their organizational mission enjoyed by secular organizations that receive public funds.” He suggests that all private associations should be exempt from antidiscrimination laws, “whenever necessary to preserve their distinctive character” (Rosen 2001, pp. 16, 17). Both Rosen and Marcosson would treat religious and secular organizations the same. But where Marcosson would require religious organizations to adhere to equality norms, Rosen would ease the path for secular organizations by also excusing them from these requirements. This resolution would provide a broad interpretation of freedom of conscience across the board. It would also allow all voluntary associations to operate under whatever policies they desired, even if these conflicted with constitutional values. Whether we raise the status of those at the bottom or lower those at the top, equality may be achieved either way. And either route complicates Alito’s claim that rights accorded to one group have no impact on the rights of those outside. In the present context, according rights to some groups to use their own religious convictions to curtail the life plans of others who do not share their beliefs constitutes a shadow establishment.

Alito’s second major concern was how to determine whether a seemingly neutral law was targeting religious exercise. Subjective motivations behind the creation of a law can be difficult to discern, and assessments of these motivations may be subjective as well. Consideration of motivation, however, arises in a great deal of litigation. If an individual injures or kills someone, for example, the type of charge against the perpetrator and the disposition of the case will certainly take motivation into account. Courts are or should be prepared to enter into these considerations. In both Lukumi and Masterpiece Cakeshop, the Supreme Court readily found animus toward religious exercise and ruled in favor of the plaintiffs. Issues concerning targeting must often be sorted out on a case-by-case basis, and this complexity need not render targeting considerations irrelevant.

Alito’s final concern in Fulton was that CSS’s policy of rejecting same-sex couples as foster parents was simply free expression of the conviction that marriage should only comprise a man and a woman. However, the policy involved not only free expression, but also a decision that impacted the actual lives of would-be foster parents. An open and pluralistic society must support free expression, but additional issues arise when expression becomes action, whether the root of these actions is religious or nonreligious. Religious beliefs and the actions that flow from them deserve a high degree of protection, but actions should not automatically receive a pass simply because they stem from religious belief.

A different argument is available to foster care and adoption agencies that want to focus upon certain types of clients to the exclusion of others. The interests of both children entering foster care and their communities of origin may be better served when agencies can support and maintain these children’s original cultural and community ties. Orthodox Jewish foster children, for example, are used to a strict religious upbringing, and it may be appropriate for an agency to recruit Orthodox Jews as foster parents. Chris Gottlieb suggests that because LGBT youth are overrepresented in the foster care system, they may be advantaged by agencies that specifically recruit LGBT foster parents, although without rejecting straight applicants. “To the extent it is recognized as beneficial to establish foster care agencies that are purposely not diverse because they seek to specialize in providing services to a particular community—whether that be based on sexual orientation, race, ethnicity, religion, or culture—it should be acknowledged that such specialization is at odds with banning discrimination in the certification of foster parents if we include in the definition of discrimination … referring potential foster parents to different foster care agencies” (Gottlieb 2022, p. 74; see pp. 66–68, 72–77). A prime example is the Indian Welfare Act of 1978, which gives preference to Native American foster parents for Native American children, although this has become controversial precisely because of the conflict between nondiscrimination and pluralism. Is pluralism better achieved by diversity within foster care agencies or by diversity among them? Either way, however, it seems preferable in my view to favor foster parents who belong to a religious, cultural, or sexual minority when there are fewer of these potential foster parents, especially when children of one of these groups are overrepresented in the foster care system.

The setting for Fulton, Gottlieb observes, is markedly different from the milieu in which Smith was decided. In Smith, conservative Justice Anton Scalia wrote the decision against the freedom of Native Americans to smoke peyote in religious ceremonies, lest each conscience be “a law unto itself” (Smith, p. 890; see also Reynolds v. United States, 98 U.S. 145 [1879]). The dissenters argued that fundamental rights, including minority religious practices by Amish, Jews, and Jehovah’s Witnesses, should be “beyond the vicissitudes of political controversy” and the power of majorities (p. 903). Liberals in 1990 wanted to protect the free exercise of religion by religious minorities, which contributed to their support for RFRA. Today, however, it is conservatives who want to protect religious liberties for evangelical Christians and Roman Catholics, who are not oppressed minorities but who feel like oppressed groups amidst a more diverse culture. “Advocates for the rights of the historically oppressed LGBTQ community are pressing a majoritarian-empowering view because LGBTQ rights have successfully garnered the support of a majority of voters, while the religious Right, which historically has been more privileged and has at times been aligned with the oppression of people of color, is now at the vanguard of defending minority rights” (Gottlieb 2022, p. 70; see pp. 69–71). Christian nationalism is a prime example of this shift.

Exemption cases illustrate the weakness in Alito’s argument that rights accorded to one group should not affect or be affected by the rights of people outside this group. In Sherbert v. Verner (374 U.S. 398 [1963]), a key exemption case, the Supreme Court overruled the denial of unemployment compensation to a Seventh Day Adventist for refusing to work on her Sabbath of Saturday. From an equal-access standpoint, when the government grants exemptions for Sabbath observance, this does not violate neutrality but instantiates neutrality by considering religious differences. Workers who must choose between their religious beliefs and making a living are disadvantaged in comparison to those without religious duties (Sandel 1990, pp. 88–90; see also Nussbaum 2008, pp. 135–40). In Thornton v. Calder (105 S. Ct. 2914 [1985]), on the other hand, the court invalidated a Connecticut law that awarded a weekly day off to Sabbath observers. The court believed that a Sabbath law advantaged the observant over the nonobservant, as the former could choose a weekend day off where the latter could not.

For the Sherbert court, Sabbath observance is a duty, and exempting the observant from work on their Sabbath without endangering unemployment benefits does not burden others. To the observant, work on the Sabbath is not “available” in the sense that workers can opt into it.

Sabbath-observant workers have less choice than do the nonobservant, who can accept work at any time. Therefore, they are burdened by secular expectations as to their flexibility. From Thornton’s perspective, Sabbath observance is a choice like any other. Awarding comparatively more freedom to observant workers does burden those who cannot use religious observance to justify choice. Observant workers are not burdened but simply not advantaged based on their religious beliefs over those who have no choice in a day off.

Applying these cases to Fulton, the law could extend exemptions to all adoption agencies, religiously or secularly oriented, to honor reservations they might have as to the types of foster or adoptive parents with whom they wanted to work. That is, nondiscrimination policies would not apply. Alternatively, the law could be altered to allow only religious exemptions, as CSS would prefer. Only religious workers might experience their desire to exclude LGBT clients as a religious duty, as per Sherbert. Finally, the law could prohibit discrimination by any entity, as the city of Philadelphia attempted to do. If no adoption agency were exempt, none would be advantaged on the basis of its religious convictions, as per Thornton, over those who might have secularly based rationales for preferring some types of families over others. These comparisons demonstrate that the rights of the religiously observant do affect and are affected by the relative position of those who are not. Under either interpretation of these cases, one group loses vis-à-vis others who are outside the group. McGowan v. Maryland, for example, by allowing the establishment of a uniform day of rest, disadvantaged those whose Sabbaths fell on days other than Sunday. This led to Sherbert, which in turn disadvantaged those whose preferences for days off were not based on a religious rationale.

7. Conclusion

Some public policies take shape because they are deemed to promote the common good irrespective of whether they benefit religious practice, such as the uniform Sabbath law unsuccessfully challenged in McGowan v. Maryland. If one disagrees—as I do—that shadow establishments are good public policy, the added benefit to religious practice will be controversial in many quarters. One of Alito’s objections to Smith was that a facially neutral law may be devised that applies to both secular and religious conduct but that is subtly targeted against religious conduct. McGowan is the reverse. That is, a facially neutral and generally applicable ruling seemingly benefits both secular and religious activities, but its benefits flow principally to religious ones—which are those of the majority religion in the United States. In Smith, religious practice may be burdened, whereas in McGowan, it benefits.

Other examples exist of public policies interpretated by some as neutral that can benefit both religious and nonreligious practitioners but that in my view favor the religious in ways that contravene free and equal citizenship. Mentioned above was the first Trump administration’s 2017 rule allowing religious exemptions for government contractors and closely held businesses “as long as they could demonstrate a religious purpose that is a public part of their mission and demonstrate that they exercise their religion in order to further that purpose,” and aimed at “the broadest possible protection of religious interests” (Rogers 2019, p. 116). In 2017, that administration issued rules that greatly expanded employers’ rights, both religious and nonreligious, to exclude contraceptive coverage from employee health insurance. These exclusions could stem either from sincerely held religious beliefs or from general moral convictions (Pear and Ruiz 2017, p. A14). The first Trump administration in 2018 initiated a Conscience and Religious Freedom Division within the Department of Health and Human Services “with the stated goal of ensuring that health care workers and institutions are never forced to deliver medical services they object to” (Hafner 2018, p. A16). The religious freedom of religious nonprofits and their workers impacts the life plans of untold numbers of individuals who may not share the convictions of these agencies. The fact that the government allows adoption agencies—or religious entities as in Grendel’s Den—to impose their own religious criteria does not exculpate them from responsibility. This practice transforms an ostensible free-exercise benefit into a burden on the life plans of many others. It relegates them to second-class citizenship thereby constituting a shadow establishment.

Finally, not all establishment efforts are in shadow, including those of many Christian nationalists. In June of 2025, for example, members of the Southern Baptist Convention (SBC) voted overwhelmingly to call for the overturning of marriage equality, legalized in 2015 by Obergefell v. Hodges (135 S. Ct. 2584 [2015]). The SBC, the U.S.’s largest Protestant denomination, “is often seen as a bellwether for conservative evangelicalism writ large” (Graham 2025). Southern Baptists are going on record and playing a long game, following the example of anti-abortionists. This effort is part of a larger measure, “On Restoring Moral Clarity through God’s Design for Gender, Marriage, and the Family,” addressing funding for Planned Parenthood, parental rights, and “fairness in female athletic competition.” The resolution maintains that it is lawmakers’ responsibility “to pass laws that reflect the truth of creation … and to oppose any law that denies or undermines what God has made plain through nature and Scripture” (Dias 2025). In 2023, the SBC expelled several churches with female pastors and is currently working toward expanded restrictions on women in church leadership. The SBC is a private organization, but like other conservative religious entities, it “reflects a movement within conservative Christianity to see that laws align with their set of biblical values, and a political commitment to pursue those goals.” This example of Christian nationalism would restrict the life plans of those who do not support these goals, it would meld public policy with sectarian religious convictions, and it would propel the shadow establishment out of the closet.