Constitutionality of Religious Restrictions on the Use of Federal Funds
CourtDepartment of Justice Office of Legal Counsel
Date FiledAugust 25, 2026
StatusPublished
📰 News Coverage: Read the LAWS.com news report on this case
Full Opinion
(Slip Opinion)
Constitutionality of Religious Restrictions on the
Use of Federal Funds
Federal statutory provisions and regulations that restrict federal funding based on either the
religious status of a program provider or anticipated religious uses of the funds are not
required under the Establishment Clause.
Because the statutory and regulatory religious restrictions applicable to programs
administered by the Department of Health and Human Services discriminate based on
anticipated religious use without any compelling governmental interest, they violate
the Free Exercise Clause. Use-based religious restrictions on generally available
federal funding programs are presumptively unconstitutional.
August 25, 2026
MEMORANDUM OPINION FOR
THE DIRECTOR OF THE OFFICE FOR CIVIL RIGHTS
DEPARTMENT OF HEALTH AND HUMAN SERVICES
Faith-based organizations across the United States deliver vital charita-
ble services—feeding the hungry, supporting families, and helping indi-
viduals overcome addiction. The federal government helps fund this work.
But the Department of Health and Human Services (“HHS”) has identified
several statutes applicable to programs and grants it administers that
restrict the use of federal funds when faith-based organizations are in-
volved, including: the Personal Responsibility and Work Opportunity Act
of 1996 (“PRWORA”), 42 U.S.C. § 604a(j); the Public Health Service
Act (“PHSA”), id. §§ 290kk-2, 300x-65(i); and the Child Care and Devel-
opment Block Grant Act of 1990 (“CCDBGA”), id. §§ 9858k(a),
9858l(a)(4).
In 2019, our Office concluded that a statutory restriction on the exten-
sion of federal loans to historically black colleges and universities “for
facilities that are predominantly used for devotional religious activity” did
not violate the Free Exercise Clause under then-prevailing Supreme Court
doctrine. Religious Restrictions on Capital Financing for Historically
Black Colleges and Universities, 43 Op. O.L.C. 191, 192 (2019) (“HBCU
Religious Restrictions”). We reached that conclusion by asking “whether
the restriction is based upon an institution’s religious status or whether it
is based upon how the federal support would be used.” Id. at 196. We
remained “mindful, however, that this area of law [was] still being devel-
oped.” Id. at 197.
1
50 Op. O.L.C. __ (Aug. 25, 2026)
HHS has asked whether it must update its framework for assessing
grants to faith-based organizations in the light of the Supreme Court’s
intervening precedent in Espinoza v. Montana Department of Revenue,
140 S. Ct. 2246 (2020), Kennedy v. Bremerton School District, 142 S. Ct.
2407 (2022), and Carson ex rel. O.C. v. Makin, 142 S. Ct. 1987 (2022),
decisions that post-dated our 2019 HBCU Religious Restrictions opinion.
See Memorandum for Josh Craddock, Deputy Assistant Attorney General,
Office of Legal Counsel, from Emily Claire Mimnaugh, Deputy General
Counsel, Department of Health and Human Services, Re: Request for
Opinion on Government Funding and Religious Use at 4–5 (July 1, 2026)
(“HHS Memorandum”). Specifically, HHS asked: (1) whether the reli-
gious restrictions applicable to HHS are required by the Establishment
Clause; and (2) if not, whether they are constitutional under the Free
Exercise Clause. See id. at 1. The answer to both questions is no. To the
extent our 2019 opinion indicated otherwise, e.g., HBCU Religious Re-
strictions, 43 Op. O.L.C. at 208–13, it no longer reflects the current state
of the law.
Our analysis has three parts. Part I describes the statutory religious re-
strictions applicable to HHS-funded programs. Part II explains why those
religious restrictions are not required by the Establishment Clause. Part III
determines that they violate the Free Exercise Clause. Although we ad-
dress the Religion Clauses separately for analytical clarity, we do not
suggest that the Clauses are in tension. Indeed, “the Clauses have ‘com-
plementary’ purposes, not warring ones where one Clause is always sure
to prevail over the others.” Kennedy, 142 S. Ct. at 2426. There is “no
conflict between the constitutional commands” of the Free Exercise and
Establishment Clauses. Id. at 2432.
I.
A.
Enacted only a few years apart, the PRWORA and PHSA provisions at
issue were both part of the “Charitable Choice” movement. See generally
42 U.S.C. §§ 604a, 290kk-1, 300x-65(i), 9858k(a), 9858l(a)(4); see also
Jonathan Friedman, Charitable Choice and the Establishment Clause,
5 Geo. J. on Fighting Poverty 103, 113 (1997). And although the CCDBGA
was not part of Charitable Choice, its funding restriction on religious
2
Constitutionality of Religious Restrictions on the Use of Federal Funds
activity and its limitation on religious decisions in employment and ad-
missions share many of the same features and motivations.
“The central objective of Charitable Choice [was] to involve faith-
based providers in anti-poverty efforts, while protecting the religious
integrity and character of those organizations.” Friedman, 5 Geo. J. on
Fighting Poverty at 113. For PRWORA, Charitable Choice meant “al-
low[ing] States to contract with religious organizations” to “administer
and provide services under” various welfare programs. 42 U.S.C.
§ 604a(a)(1)(A), (b). For the PHSA, it meant “allow[ing] religious organi-
zations to be program participants on the same basis as any other nonprof-
it private provider” in the administration of substance-abuse programs. Id.
§ 290kk-1(b). The CCDBGA also assumed that religious child-care pro-
viders would be eligible to participate on an equal basis with other private
providers. See id. § 9858l. As described in President George W. Bush’s
follow-on Faith-Based and Community Initiatives, the goal was “to enlist,
equip, enable, empower, and expand the work of faith-based and other
community organizations.” Exec. Order No. 13199, 66 Fed. Reg. 8499,
8499 (Jan. 29, 2001); see also Exec. Order No. 13279, 67 Fed. Reg.
77,141, 77,142–43 (Dec. 12, 2002); Nicole Stelle Garnett & Cameron
Grinnell, Manhattan Inst., The Persistence of the Unconstitutional Sta-
tus/Use Distinction in Federal Grants to Religious Nonprofits at 2–3
(2025), https://perma.cc/F9SD-AA66. Charitable Choice and its kin thus
reflected an affirmative policy of inclusion toward religious organizations.
Although Charitable Choice aimed to include religious institutions in
these government funding programs, it also imposed use-based religious
restrictions geared toward avoiding perceived Establishment Clause
snares. See Brian Craig Kimball, Comment, The Federal Faith-Based and
Community Initiative: A Guide for Future Legislation, 71 Miss. L.J. 241,
259–60 (2001). 1 PRWORA states that “[n]o funds provided directly to
institutions or organizations to provide services and administer programs
1 Senator John Ashcroft, one of the chief architects of Charitable Choice, argued that
the restrictions “embod[y] U.S. Supreme Court case precedent to clarify what is constitu-
tionally permissible when state and local governments cooperate with the religious and
charitable sector of society.” Friedman, 5 Geo. J. on Fighting Poverty at 114 (citation
omitted). Senator Ashcroft believed that “the focus should not be on whether sectarian
organizations receive public funds, but on whether those funds are used for religious
activities.” Id. (emphasis added).
3
50 Op. O.L.C. __ (Aug. 25, 2026)
under [PRWORA] shall be expended for sectarian worship, instruction, or
proselytization.” 42 U.S.C. § 604a(j).2 The PHSA similarly provides that
“[n]o funds provided under a designated [PHSA] program shall be expend-
ed for sectarian worship, instruction, or proselytization.” Id. § 290kk-2. A
separate provision restricts funds provided “through a grant or contract to
a religious organization to provide services under any [PHSA] substance
abuse program . . . for sectarian worship, instruction, or proselytization.”
Id. § 300x-65(i).3
The CCDBGA states that “[n]o financial assistance provided under
this subchapter, pursuant to the choice of a parent under section
9858c(c)(2)(A)(i)(I)”—that is, the option “to enroll [a] child with a child
care provider that has a grant or contract for the provision of such ser-
vices”—“or through any other grant or contract under the State plan, shall
be expended for any sectarian purpose or activity, including sectarian
worship or instruction.” Id. §§ 9858k(a), 9858c(c)(2)(A)(i)(I). Congress
deliberately excluded indirect assistance from that provision, providing
that child-care certificates “shall not be considered to be grants or con-
tracts,” and that “[n]othing in this subchapter shall preclude the use of
such certificates for sectarian child care services if freely chosen by
the parent.” Id. § 9858n(2).4 The CCDBGA also imposes an anti-
discrimination requirement on certain religious child-care providers—
those receiving 80% or more of their operating budget from government
funding—as a condition on the receipt of further funding under the pro-
gram. Such providers must not “discriminate against any individual” in
either the employment of persons who primarily work directly with chil-
dren or in admissions “because of the religion of such individual.” Id.
§ 9858l(a)(4).
2 This restriction affects HHS’s Administration for Children and Families (“ACF”),
which administers block grants to states for the Temporary Assistance for Needy Families
(“TANF”) program, as well as the states that provide services through charitable, reli-
gious, or private organizations. 42 U.S.C. § 604a(a)(1)(A).
3 This restriction affects programs administered by the Substance Abuse and Mental
Health Services Administration that award assistance to public or private entities for the
prevention or treatment of substance abuse. 42 U.S.C. § 290kk(a). It also affects funded
states that partner with non-governmental organizations for the provision of substance-
abuse services. Id. § 300x-65(j).
4 The CCDBGA created the Child Care and Development Fund program. Its statutory
restriction affects ACF’s administration of block grants to states for child-care programs,
as well as the states themselves when they partner with child-care providers.
4
Constitutionality of Religious Restrictions on the Use of Federal Funds
Two features indicate that these restrictions were deliberately crafted to
avoid problems under then-current Supreme Court Establishment Clause
precedents—which the Court has since abrogated or undermined. See
Part II.
First, to avoid the “special Establishment Clause dangers” of direct-
payment programs, Rosenberger v. Rector & Visitors of Univ. of Va., 515
U.S. 819, 842 (1995), the restrictions apply to funds provided directly to
religious institutions rather than indirect-payment programs of “true
private choice,” Zelman v. Simmons-Harris, 536 U.S. 639, 649 (2002).
PRWORA expressly applies to “funds provided directly” to religious
institutions. 42 U.S.C. § 604a(j). The PHSA restriction similarly applies
to “funds provided under a designated program,” id. § 290kk-2, which is
defined to include “discretionary and formula grant programs adminis-
tered by the Substance Abuse and Mental Health Services Administration
[“SAMHSA”] that make awards of financial assistance to public or pri-
vate entities for the purpose of carrying out activities to prevent or treat
substance abuse,” id. § 290kk(a). In other words, the statutory restriction
applies only to programs that provide grants directly “to” private
organizations (such as religious organizations), as opposed to programs
that provide vouchers or other forms of indirect funding. Cf., e.g., id.
§ 604a(a)(1)(B). And the CCDBGA restriction applies to financial assis-
tance provided directly to “child care provider[s],” id. § 9858c(c)(2)(A)(i)(I),
and to other end recipients of federally funded grants or contracts, see id.
§§ 9858k(a), 9858l. In the same breath, it allows indirect religious uses of
child-care certificate funds, see id. § 9858n(2), tracking the private-choice
line drawn in Zelman, see 536 U.S. at 649.
Second, the restrictions prohibit only the use of funds for “sectarian
worship, instruction, or proselytization,” id. §§ 604a(j), 290kk-2; see also
id. § 9858k(a) (“sectarian purpose or activity, including sectarian worship
or instruction”), or the receipt of certain funds if providers make religious
employment or admissions decisions, id. § 9858l(a)(4). These restrictions
ostensibly enforced the Supreme Court’s “proscription against ‘funding a
specifically religious activity in an otherwise substantially secular set-
ting.’” Bowen v. Kendrick, 487 U.S. 589, 613 (1988) (alteration accepted)
(quoting Hunt v. McNair, 413 U.S. 734, 743 (1973)). Funding to “perva-
sively sectarian” institutions was presumed to further religious activity.
Id. at 610.
5
50 Op. O.L.C. __ (Aug. 25, 2026)
B.
HHS implements a similar religious restriction for many programs
through 45 C.F.R. part 87. For direct-assistance programs, the relevant
regulations “prohibit[] organizations from engaging in explicitly religious
activities . . . as part of the HHS-funded program or service.” HHS Mem-
orandum at 5 (citing 45 C.F.R. § 87.3(d)). According to the regulations,
“[t]hose activities must be offered separately in time or location and
participation must be voluntary,” while “indirect assistance is not subject
to that restriction.” Id. Part 87 also prohibits organizations receiving
direct aid from “discriminat[ing] against a program beneficiary or pro-
spective program beneficiary on the basis of religion, a religious belief, a
refusal to hold a religious belief, or a refusal to attend or participate in a
religious practice.” 45 C.F.R. § 87.3(f). HHS’s regulations additionally
include “program-specific analogues” applicable to the Substance Abuse
Prevention and Treatment Block Grant; the Projects for Assistance in
Transition from Homelessness Formula Grants; and discretionary funding
under the PHSA for SAMHSA’s substance-abuse prevention and treat-
ment services. Id. at 5–6 (citing 42 C.F.R. §§ 54.1–54.8, 54a.1–54a.8).5
Under these regulations, “[a]ny program with religious content must be
a separate and distinct program from the federally funded program, and
the distinction must be completely clear to the consumer.” Memorandum
for Healthy Marriage and Responsible Fatherhood Federal Grantees, from
Administration for Children and Families, Department of Health and
Human Services, Re: Separation of Explicitly Religious Activities at *1
(2008) (“HHS, Separation of Explicitly Religious Activities”), https://
perma.cc/ZGQ7-LN42. And if the religious organization chooses to
separately offer its course containing religious content, it will be subject
to burdensome differentiation and cost-allocation requirements. See id.
For example, if a faith-based organization runs a federally supported
soup kitchen, can organizers say a prayer before they serve the meal? 6 Not
5 Other HHS Charitable Choice regulations, which apply to programs authorized under
the Community Services Block Grant Act, raise materially similar constitutional issues.
See 45 C.F.R. pt. 1050.
6 See White House Off. of Faith-Based & Cmty. Initiatives, Guidance to Faith-Based
and Community Organizations on Partnering with the Federal Government at 2 (2006)
(“GWB Guidance”), https://perma.cc/GUK2-MTU6.
6
Constitutionality of Religious Restrictions on the Use of Federal Funds
unless the “explicitly religious activit[y]” of prayer is “separate[] in time
or location” from the funded services. 45 C.F.R. § 87.3(d). Similarly, a
religious organization that receives federal funds to help unemployed
people improve their job skills could not convey their religious teachings
about the value and dignity of work. 7 And a religious organization that
teaches a federally funded course on healthy marriage and responsible
fatherhood could not endorse any religious messages in its program. 8 We
understand that such federally mandated bifurcation has proven challenging
to small faith-based organizations, and may deter religious institutions
from participating altogether in federally funded programs. 9
II.
Three strands of Supreme Court precedent illuminate the history of use-
based restrictions on government funding, each of which was rooted in—
or absorbed by—the now-defunct precedent of Lemon v. Kurtzman, 403
U.S. 602 (1971). We now describe the Lemon test and the three lines of
precedent that grew out of it. We then explain why Lemon was ultimately
“abandoned” in favor of a framework based on the original meaning of
the Establishment Clause. Kennedy, 142 S. Ct. at 2427. Applying that
framework, we conclude that the PRWORA, PHSA, and CCDBGA re-
strictions are not required by the Establishment Clause.
A.
Our Nation has a mixed history of placing legal restrictions on the pro-
vision of government funds to religious institutions. Some parts of this
history involve “shameful” restrictions based on religious status. Mitchell
v. Helms, 530 U.S. 793, 828 (2000) (plurality opinion). Other parts of that
history—relating to restrictions based on religious use—were more well-
intentioned but ultimately misguided. The PRWORA, PHSA, and
7 See GWB Guidance at 11.
8 Memorandum for Healthy Marriage and Responsible Fatherhood Federal Grantees,
from Administration for Children and Families, Department of Health and Human Ser-
vices, Re: Curricula Use and Compliance with 45 CPR Part 87 at 1–2 (Jan. 23, 2008),
https://perma.cc/UKL8-RTZ6.
9 See Dep’t of Just., Eradicating Anti-Christian Bias within the Federal Government:
The 2026 Report by the Task Force to Eradicate Anti-Christian Bias at 111 (2026),
https://perma.cc/Y38H-P86A.
7
50 Op. O.L.C. __ (Aug. 25, 2026)
CCDBGA restrictions grew out of that history, enacted in response to the
Supreme Court’s development of its Religion Clauses jurisprudence over
the course of the twentieth century.
1.
Some of the earliest religious restrictions were born of “prejudice,”
Espinoza, 140 S. Ct. at 2271 (Alito, J., concurring), and sought to bar
government funding to religious institutions based on their religious
status. The most notorious of these was the so-called Blaine Amendment.
Proposed in 1875 by Speaker of the House James G. Blaine, this amend-
ment “would have amended the Constitution to bar any aid to sectarian
institutions.” Mitchell, 530 U.S. at 828 (plurality opinion). Because “the
amendment arose at a time of pervasive hostility to the Catholic Church
and to Catholics in general, . . . it was an open secret that ‘sectarian’ was
code for ‘Catholic.’” Id.; see also Locke v. Davey, 540 U.S. 712, 723 n.7
(2004). Although the Blaine Amendment passed the House, it fell two
votes short of the necessary two-thirds majority in the Senate and was
never referred to the states for ratification. See Espinoza, 140 S. Ct. at
2268 (Alito, J., concurring).
Restrictions on funds provided to “sectarian” institutions nevertheless
made their way into many state laws, “often as a condition of entering the
Union.” Id. at 2268–69. For example, Montana’s enabling act required
“[t]hat provision shall be made for the establishment and maintenance of
systems of public schools . . . free from sectarian control.” Act of Feb. 22,
1889, ch. 180, § 4, 25 Stat. 676, 677; see also Locke, 540 U.S. at 723–24
n.7 (recording that Washington State’s enabling act contained nearly
identical language). These state enabling acts “complet[ed] the unfinished
work of the failed Blaine Amendment,” and as of 2020, 38 states still had
“little Blaine Amendments.” Espinoza, 140 S. Ct. at 2269, 2271 (Alito, J.,
concurring) (citations omitted); see also Zelman, 536 U.S. at 721 (Breyer,
J., dissenting) (observing how anti-Catholic sentiment “played a signifi-
cant role” in the refusal to “help pay for ‘sectarian’ (i.e., Catholic) school-
ing”).
Such status-based religious restrictions are understood to have “a
shameful pedigree that we do not hesitate to disavow.” Mitchell, 530 U.S.
at 828 (plurality opinion). And the Supreme Court has made abundantly
clear that “disqualifying otherwise eligible recipients from a public
8
Constitutionality of Religious Restrictions on the Use of Federal Funds
benefit ‘solely because of their religious character’ imposes ‘a penalty on
the free exercise of religion that triggers the most exacting scrutiny.’”
Espinoza, 140 S. Ct. at 2255 (quoting Trinity Lutheran Church of
Columbia, Inc. v. Comer, 582 U.S. 449, 462 (2017)). They are rightly con-
sidered “presumptively unconstitutional.” HBCU Religious Restrictions,
43 Op. O.L.C. at 196. “No matter how narrowly drawn, a religious-funding
restriction stemming from ‘hostility toward religion’ is unconstitutional.”
Id. at 212 (quoting Locke, 540 U.S. at 721).
2.
Other religious restrictions—such as those that are the subject of this
opinion—are more subtle. They seek to place limits on the religious use
of government funds. Unlike the status-based restrictions discussed above,
these use-based restrictions are not necessarily animated by religious
prejudice. As alluded to above, “[w]hat evidence we have . . . suggests
that Congress’s motive for these restrictions was likely grounded in a
legitimate desire to conform the statute to the Supreme Court’s then-
prevailing Establishment Clause precedent, not in religious animus.” Id. at
213.
In Lemon, the Supreme Court “ambitiously attempted to find a grand
unified theory,” Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2087
(2019) (plurality opinion), that would provide a “one-size-fits-all test” for
applying the Establishment Clause, Shurtleff v. City of Boston, 142 S. Ct.
1583, 1603–04 (2022) (Gorsuch, J., concurring in the judgment). The
Court “gleaned” three tests for identifying an Establishment Clause viola-
tion from “the cumulative criteria developed by the Court over many
years”:
(1) Does the challenged government action “have a secular legislative
purpose”?
(2) Is the challenged government action’s “principal or primary effect
. . . one that neither advances nor inhibits religion”?
(3) Does the challenged government action “foster an excessive gov-
ernment entanglement with religion”?
Lemon, 403 U.S. at 612–13 (cleaned up). Later, the Court “further . . .
refined” these tests to ask simply “whether the challenged governmental
9
50 Op. O.L.C. __ (Aug. 25, 2026)
practice either has the purpose or effect of ‘endorsing’ religion.” County
of Allegheny v. ACLU, Greater Pittsburgh Chapter, 492 U.S. 573, 592
(1989); see also id. at 628 (O’Connor, J., concurring in part and concur-
ring in the judgment) (explaining that “the endorsement test asks the right
question”).
Lemon’s three questions braid three distinct strands of Supreme Court
precedent. These in turn influenced the adoption of use-based restrictions
on government funding. Many such restrictions fell within the Supreme
Court’s “play in the joints” jurisprudence. Others were crafted to reflect
the Supreme Court’s “pervasively sectarian” limitation. Still others sought
to walk the line drawn by the Court between the “direct” and “indirect”
funding of religious institutions. We address each strand of precedent in
turn.
a.
Even before Lemon, the Supreme Court described a certain “play in the
joints” between the Establishment Clause and the Free Exercise Clause.
That phrase originated in Walz v. Tax Commission of the City of New
York, when the Supreme Court upheld a tax exemption provided to “reli-
gious organizations for religious properties used solely for religious
worship.” 397 U.S. 664, 666 (1970). The Court coined that now-famous
phrase by reference to government actions that were perceived not to
“sponsor[]” or “interfere” with religion or to “have the effect of doing
so.” Id. at 669. It identified the “general principle” that the First Amend-
ment “will not tolerate either governmentally established religion or
governmental interference with religion.” Id. “Short of those expressly
proscribed governmental acts,” the Court explained, there was space for
“a benevolent neutrality which will permit religious exercise to exist
without sponsorship and without interference.” Id. The tax exemption in
Walz fell within this “play in the joints” because it did not violate the
“course of constitutional neutrality.” Id. Or, put differently, “the trans-
cendent value of free religious exercise in our constitutional scheme
leaves room for ‘play in the joints’ to the extent of cautiously delineated
secular governmental assistance to religious [institutions].” Norwood v.
Harrison, 413 U.S. 455, 469 (1973).
The Court eventually glossed the “play in the joints” principle to mean
that “there are some state actions permitted by the Establishment Clause
10
Constitutionality of Religious Restrictions on the Use of Federal Funds
[that are] not required by the Free Exercise Clause.” Locke, 540 U.S. at
718–19; see also Cutter v. Wilkinson, 544 U.S. 709, 719–20 (2005). But
“[a]t some point, accommodation may devolve into ‘an unlawful fostering
of religion.’” Corp. of the Presiding Bishop of the Church of Jesus Christ
of Latter-Day Saints v. Amos, 483 U.S. 327, 334–35 (1987) (quoting
Hobbie v. Unemp. Appeals Comm’n of Fla., 480 U.S. 136, 145 (1987)).
Indeed, the entire “play in the joints” enterprise was animated by a desire
to “avoid excessive entanglement” between government and religion—a
touchstone of the Lemon test. Walz, 397 U.S. at 669–70, 674; see also
Amos, 483 U.S. at 335 (noting that Walz “became the third element of the
Lemon test”).
But what started as a shield to protect government programs from
Establishment Clause challenges eventually turned into a sword. Govern-
ments started to argue that the “play in the joints” meant that, just like
they were free to include religious organizations in funding programs,
they were also free to exclude religious organizations whose use of
government funds involved religious uses. The State of Washington
successfully pushed this argument to defend its constitutional provision
“prohibiting even indirectly funding religious instruction” through a gen-
eral scholarship program “that w[ould] prepare students for the ministry.”
Locke, 540 U.S. at 719. The States of Missouri, Montana, and Maine
raised similar “play in the joints” arguments, though each without
success. See Trinity Lutheran, 582 U.S. at 458; Espinoza, 140 S. Ct. at
2254; see also Carson, 142 S. Ct. at 2009–10 (Breyer, J., dissenting).
Many remaining use-based restrictions can be understood in the histori-
cal context of this “play in the joints” framework. According to that
principle, a government had some leeway to either include or exclude
religious activities in programs it funded. See Locke, 540 U.S. at 719.
Use-based restrictions—such as those incorporated into PRWORA, the
PHSA, and the CCDBGA—may simply reflect a policy of exclusion,
premised on the government’s so-called “Establishment Clause interest.”
Good News Club v. Milford Cent. Sch., 533 U.S. 98, 113 (2001). 10 When
10 Even when the Supreme Court applied the “play in the joints” framework, it repeat-
edly emphasized that states do not have a compelling interest “in achieving greater
separation of church and State than is already ensured under the Establishment Clause of
the Federal Constitution.” Widmar v. Vincent, 454 U.S. 263, 276 (1981). And in Carson,
the Court reemphasized that “an interest in separating church and state more fiercely than
11
50 Op. O.L.C. __ (Aug. 25, 2026)
viewed in this light, some have argued that “States enjoy a degree of
freedom to navigate the [Religion] Clauses’ competing prohibitions,”
which “includes choosing not to fund certain religious activity where
States have strong, establishment-related reasons for not doing so.” Carson,
142 S. Ct. at 2004 (Breyer, J., dissenting).
b.
A second line of precedent prohibited government funds from
“flow[ing] to an institution in which religion is so pervasive that a sub-
stantial portion of its functions are subsumed in the religious mission or
when it funds a specifically religious activity in an otherwise substantially
secular setting.” Hunt, 413 U.S. at 743. This “pervasively sectarian”
limitation grew out of a fear that “direct government aid might have” the
“primary effect of advancing religion.” Bowen, 487 U.S. at 610 (citation
omitted); cf. Tilton v. Richardson, 403 U.S. 672, 685–87 (1971) (plurality
opinion); Comm. for Pub. Educ. & Religious Liberty v. Nyquist, 413 U.S.
756, 773–74 (1973). “Aid normally may be thought to have a primary
effect of advancing religion,” the Court explained, “when it funds a
specifically religious activity in an otherwise substantially secular set-
ting.” Hunt, 413 U.S. at 743.
A keen observer might suggest this prohibition against funding “perva-
sively sectarian” entities depends on the same status-based distinctions
that the Supreme Court called “shameful.” Mitchell, 530 U.S. at 828
(plurality opinion). But the Court long stressed that “the proposition that
the Establishment Clause prohibits any program which in some manner
aids an institution with a religious affiliation has consistently been rejected,”
Hunt, 413 U.S. at 742, and that “th[e] Court has never held that religious
institutions are disabled by the First Amendment from participating in
publicly sponsored social welfare programs,” Bowen, 487 U.S. at 609; see
also Mitchell, 530 U.S. at 826–29 (plurality opinion); HBCU Religious
Restrictions, 43 Op. O.L.C. at 205–06.
Instead, the “pervasively sectarian” limitation primarily aimed at pro-
hibiting religious uses. Some religious institutions could be trusted to use
funds for secular purposes, the Court reasoned, yet it found “a risk that
the Federal Constitution cannot qualify as compelling in the face of the infringement of
free exercise.” 142 S. Ct. at 1998 (cleaned up).
12
Constitutionality of Religious Restrictions on the Use of Federal Funds
direct government funding, even if it is designated for specific secular
purposes, may nonetheless advance [a] pervasively sectarian institution’s
‘religious mission.’” Bowen, 487 U.S. at 610 (emphasis added). Because
money is fungible, in other words, granting government funding directly
to a “pervasively sectarian” institution—as opposed to merely a religiously
affiliated organization—risked having the “primary effect of advancing
religion.” Id. at 609–10. The “proscription against ‘funding a specifically
religious activity in an otherwise substantially secular setting’” likewise
aimed to avoid advancing religion. Id. at 613 (alteration accepted) (quot-
ing Hunt, 413 U.S. at 743).
c.
A third line of cases purported to distinguish between direct and indi-
rect funding to religious organizations. The Supreme Court emphasized
the “special Establishment Clause dangers where the government makes
direct money payments to sectarian institutions.” Rosenberger, 515 U.S.
at 842 (emphasis added); see also Bowen, 487 U.S. at 610; Hunt, 413 U.S.
at 743. This was informed by a baseline that “[t]he Establishment Clause
. . . prevents a State from enacting laws that have the ‘purpose’ or ‘effect’
of advancing or inhibiting religion.” Zelman, 536 U.S. at 648–49 (citation
omitted).
To evaluate whether the government act had a forbidden purpose or
effect, the Court drew a distinction between “government programs that
provide aid directly to religious [institutions]” and “programs of true
private choice, in which government aid reaches religious [institutions]
only as a result of the genuine and independent choices of private individ-
uals.” Id. at 649. In these “indirect” cases, the Supreme Court explained
that if “a government aid program is neutral with respect to religion, and
provides assistance directly to a broad class of citizens who, in turn, direct
government aid to religious schools wholly as a result of their own genu-
ine and independent private choice,” then “the program is not readily
subject to challenge under the Establishment Clause.” Id. at 652. That was
because “[t]he incidental advancement of a religious mission, or the
perceived endorsement of a religious message, is reasonably attributable
to the individual recipient, not to the government.” Id. The direct-indirect
distinction, then, was thought to reflect both the first and second elements
of Lemon, as captured by the endorsement test. See id. at 669–70
13
50 Op. O.L.C. __ (Aug. 25, 2026)
(O’Connor, J., concurring) (arguing that the direct-indirect cases “focus[]
on a narrow question related to the Lemon test” and merely represent “a
refinement” of that test).
For this reason, the Supreme Court observed that “[a]ny Establishment
Clause objection to the scholarship program” in Espinoza was “particularly
unavailing because the government support makes its way to religious
schools only as a result of Montanans independently choosing to spend
their scholarships at such schools.” 140 S. Ct. at 2254. And in Carson, the
Court similarly observed that “a neutral benefit program in which public
funds flow to religious organizations through the independent choices of
private benefit recipients does not offend the Establishment Clause.” 142
S. Ct. at 1997. These cases made clear that if funding reached a religious
institution through the choices of independent private actors, then the
Establishment Clause did not care how the funding was ultimately used.
3.
For many years, Lemon was the focus of fierce criticism for its “short-
comings,” which “present[ed] particularly daunting problems.” Am.
Legion, 139 S. Ct. at 2080–81 (plurality opinion); see also County of
Allegheny, 492 U.S. at 655–56 (Kennedy, J., concurring in the judgment
in part and dissenting in part) (observing in 1989 that “[p]ersuasive criti-
cism of Lemon has emerged” and collecting cases). A growing consensus
found that Lemon “bypassed any inquiry into the Clause’s original mean-
ing,” “ignored longstanding precedents,” and “produced only chaos.”
Shurtleff, 142 S. Ct. at 1604 (Gorsuch, J., concurring in the judgment).
And for decades, Supreme Court justices had cast doubt on whether “the
Lemon test [fit] in the larger scheme of Establishment Clause jurispru-
dence” at all. Van Orden v. Perry, 545 U.S. 677, 686 (2005) (plurality
opinion).
But “[l]ike some ghoul in a late-night horror movie that repeatedly sits
up in its grave and shuffles abroad, after being repeatedly killed and
buried,” Lemon continued to “stalk[] [the Court’s] Establishment Clause
jurisprudence.” Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist.,
508 U.S. 384, 398 (1993) (Scalia, J., concurring in the judgment); see also
id. at 398–99 (counting opinions from justices repudiating Lemon). Thus,
while we were forced to conclude in 2019 that, under existing precedent,
“antiestablishment interests might justify a use-based religious-funding
14
Constitutionality of Religious Restrictions on the Use of Federal Funds
restriction,” HBCU Religious Restrictions, 43 Op. O.L.C. at 208, we
nevertheless observed in 2021 that “the Court ha[d] retreated” from
“Lemon-based analysis” in the contexts of religious displays and legisla-
tive prayers, validating a “historical approach” that is “more consistent
with the Nation’s constitutional tradition and the original meaning of the
Establishment Clause,” Religious Seasonal Decorations in Federal Gov-
ernment Buildings, 45 Op. O.L.C. __, at *11 (Jan. 15, 2021) (“Seasonal
Decorations”).
In 2022, the Court drove a stake through Lemon’s heart. The Supreme
Court in Kennedy definitively rejected Lemon for representing an “ahis-
torical approach to the Establishment Clause” that the “Court long ago
abandoned,” along with “its endorsement test offshoot.” Kennedy, 142
S. Ct. at 2427. Kennedy’s conclusive abrogation of Lemon produced a
“sea change” in Establishment Clause jurisprudence. Nathan v. Alamo
Heights Indep. Sch. Dist., 173 F.4th 576, 591 (5th Cir. 2026) (en banc).
Even if the Supreme Court has not overruled each of Lemon’s progeny
“by name,” we cannot blindly adhere to faulty precedent when its entire
ratio decidendi has been eviscerated. Id. Any remaining “doubt” we had
about the status of these and other Lemon-era precedents, see HBCU
Religious Restrictions, 43 Op. O.L.C. at 205, was resolved by Kennedy,
142 S. Ct. at 2427.
B.
The Supreme Court has now “instructed that the Establishment Clause
must be interpreted by ‘reference to historical practices and understand-
ings’” from the time the First Amendment was ratified. Id. at 2428 (quoting
Town of Greece v. Galloway, 572 U.S. 565, 576 (2014)). This approach
requires an “analysis focused on original meaning and history,” id., be-
cause looking to history can “reveal[] . . . the contemporaneous under-
standing of ” the Establishment Clause, Lynch v. Donnelly, 465 U.S. 668,
673–74 (1984) (examining “an unbroken history of official acknowledg-
ment by all three branches of government of the role of religion in American
life from at least 1789”). It is to this historically based analysis that we
now turn.
To identify the historical practices and understandings relevant to the
Establishment Clause, the Supreme Court has pointed us toward the
“hallmarks of religious establishments the framers sought to prohibit
15
50 Op. O.L.C. __ (Aug. 25, 2026)
when they adopted the First Amendment.” Kennedy, 142 S. Ct. at 2429 &
n.5; see also Shurtleff, 142 S. Ct. at 1609–10 (Gorsuch, J., concurring in
the judgment) (identifying the hallmarks of a religious establishment). 11
This Office has long looked to such historical practices to inform its
views of the Religion Clauses, and criticized the Lemon-era failure to
“engage in any serious analysis of the historical understanding of estab-
lishment.” Memorandum for the Files, from Attorney-Adviser, Office of
Legal Counsel, Re: Historical Understanding of Religion Clauses at 36
(Mar. 17, 2004) (“File Memorandum”). As this Office has explained, “the
concept of a national ‘establishment’ at the time of the Founding did not
refer to any government action touching religion, but rather to a distinc-
tive form of support, enforcement, and control of religious faith.” Seasonal
Decorations at *8. “The state religious establishments that the Founders
sought to protect from federal interference” often “involve