Religion and the State in the United States at the Turn of the Twenty-first Century

Chapter 6
Religion and the State in the United States at the Turn of the Twenty-first Century1

W. Cole Durham Jr. and Robert T. Smith

The right to freedom of religion has a long and familiar history in the United States.2 Although not specifically addressed in the original US Constitution itself, with the exception of a provision banning religious tests for public office,3 freedom of religion was identified from the beginning of American constitutional history as a foundational fundamental right by the First Amendment to the Constitution, which provides: ‘Congress shall make no law respecting an establishment of religion or prohibiting the free exercise thereof’.4

Because this language was understood for the first 150 years of its history as a constraint on federal power,5 and most church–state issues arose at the level of the individual states, federal constitutional cases interpreting the religion clauses of the First Amendment during this period were relatively rare. Contrary to what most would currently assume, the Non-Establishment clause was initially intended not as a guarantee of separation of church and state but as a constraint designed to prevent the federal government from interfering with state established churches and state religious tests for office that continued to exist in most of the states at the time the Constitution was adopted.6 The last of the state churches, that of Massachusetts, was not disestablished until 1833.7

On the Free Exercise side, case law was also thin. All the leading cases came from the federal territories before they attained statehood and while they were still ruled by Congress. The most significant cases involved federal prosecution of the nineteenth century practice of plural marriage among Mormons in the territories of Utah and Idaho.8 Not until the Supreme Court held that the Free Exercise and Non-Establishment clauses had been incorporated in the Due Process clause of the post-Civil War Fourteenth Amendment,9 and thereby been made applicable to the states in 194010 and 194711 respectively, did the modern case law interpreting these clauses begin to emerge.

While the post-incorporation evolution of case law was not instant, standard interpretations of the meaning of the two clauses had emerged by the 1960s and 1970s. Case law addressing questions of free exercise enunciated a ‘strict scrutiny’ test, according to which state action that burdened religion was impermissible unless justified by a compelling state interest that could not be advanced by less restrictive means. This ‘strict scrutiny’ or ‘compelling state interest’ test was set forth in two leading cases, Sherbert v. Verner,12 and Wisconsin v. Yoder.13 Sherbert held that a state did not have a sufficiently compelling interest to justify withholding unemployment compensation from an Adventist, who turned down otherwise available work because she refused to work on Saturday, her Sabbath. Yoder held that the state’s interest in requiring education until age 16 was not sufficiently compelling to sustain fines levied against Amish parents who refused to send children to school past the eighth grade. The effect of this strict scrutiny test was to provide strong protections for free exercise of religion, in contrast to normal judicial review, which would defer to legislative judgment unless there was no rational basis for its enactments, which is almost never the case. This did not mean that free exercise claimants invariably prevailed, but the standard placed free exercise claimants in a very strong bargaining position when working out accommodations when government programmes burdened their religious practices. The level of free exercise protection existing under the strict scrutiny test prior to Smith can be shown conceptually as in Figure 6.1.

Figure 6.1 Free exercise protection prior to Smith14

At approximately the same time, the Non-Establishment clause was given a markedly separationist interpretation. In Lemon v. Kurtzman, a case which struck down public supplements to the salaries of teachers of secular subjects in parochial schools, the US Supreme Court held that state action constituted an impermissible establishment of religion unless it met all three of the following criteria:

1. It had a secular purpose;

2. Its primary effect neither advanced nor inhibited religion; and

3. It did not result in excessive entanglement of government and religion.15

In the 1980s, Justice O’Connor introduced a somewhat different test, which focused on whether state action resulted in an endorsement of religion.16 This test placed more weight on the neutrality of the state’s posture towards religion. But in general, both tests insisted with varying degrees of strictness on a strong separation of church and state.

The standard interpretations of the two clauses began to erode in the 1980s, and in the 1990s were significantly transformed. Because accounts of the earlier case law are readily available,17 and because an understanding of the developments beginning with key cases from the 1990s is critical to understanding current religious freedom protections in the United States, this chapter focuses on developments beginning with the transformative case law of the 1990s. In particular, it focuses on the central constitutional drama in these years: namely, shifts in the applicable standards of review in free exercise and non-establishment cases. The aim is to give a concise picture of the doctrinal setting that has emerged in the opening years of the twenty-first century.

The standard strict scrutiny approach to free exercise cases suffered a severe blow in the Supreme Court’s 1990 decision of Employment Division, Department of Human Resources of Oregon v. Smith.18 In that case, the Supreme Court jettisoned the standard ‘compelling state interest’ test, and held instead that any neutral law of general applicability sufficed to override free exercise claims.19 In doing so, the Court reinterpreted the Free Exercise clause, substantially narrowing the range of circumstances in which burdens on the free exercise of religion would be submitted to strict constitutional scrutiny, and correspondingly expanding the range of permissible state interference in the religious realm.20 The Court acknowledged that its decision would mean that in the usual case religious groups’ only recourse in seeking accommodations for religious beliefs would be to the political process, and that this would ‘place at a relative disadvantage those religious practices that are not widely engaged in’.21 Forgotten in this cavalier relegation of religious claims to the political process is that the key purpose of a bill of rights is to protect fundamental rights from the tyranny (and also from the insensitive inattention) of the majority.

To be sure, the revised standard was qualified in certain important respects. First, the Court made it clear that ‘the First Amendment obviously excludes all “governmental regulation of religious beliefs as such”’.22 In particular, ‘government may not compel affirmation of religious belief’;23 it may not ‘punish the expression of religious doctrines it believes to be false’;24 and it may not ‘impose special disabilities on the basis of religious views or religious status’.25 Thus, the classic bar to state interference with inner beliefs (as opposed to outer conduct) was not abandoned. Second, the Court expressly reaffirmed its line of religious autonomy cases, emphasizing that ‘government may not … lend its power to one or the other side in controversies over religious authority or dogma’.26 Third, the Court indicated that in a number of prior cases involving ‘hybrid situations’, free exercise claims, when coupled with other constitutional protections such as freedom of expression or parental rights, were sufficient to bar application of neutral and general laws.27 Fourth, the Court did not overrule use of the compelling state interest analysis in unemployment compensation cases following Sherbert where the loss of employment was not linked to criminal activity. The Court rationalized retention of strict scrutiny analysis in this context because it involved ‘individualized governmental assessment of the reasons for the relevant conduct’,28 noting that ‘where the State has in place a system of individual exemptions, it may not refuse to extend that system to cases of “religious hardship” without compelling reason’.29 Finally, implicit in the idea of ‘neutral and general laws’ is the notion that non-neutral and non-general laws that intentionally target and discriminate against religious groups or religious activities remain subject to strict scrutiny.30 In each of these situations, First Amendment protections remained what they had been prior to the Smith decision. Despite these exceptions, Smith held that in the main, ‘neutral law[s] of general applicability’ would suffice to overrule free exercise claims.

In the aftermath of this decision, there was a widely shared sense that free exercise protections had been seriously weakened, if not deliberately dismantled.31 The weakened sense of free exercise protection after Smith is illustrated in Figure 6.2.

Not only was the weakening of free exercise protections a concern, but the Court’s position seemed to ignore the fact that the major purpose of the Bill of Rights was to provide protection against tyranny of the majority and majoritarian legislation. Rather, it assumed that the law itself sets the limits on permissible variation in a pluralistic society, and that the greater accommodation and flexibility guaranteed by the strict scrutiny standard would court anarchy and lead to chaos.32 Indeed, the Court went so far as to suggest that the presumptive invalidity of burdens on free exercise presupposed by the compelling state interest test constituted a ‘luxury’ we could no longer afford.33 Critics, in contrast, complained that the protections of the Free Exercise clause had been ‘eviscerat[ed]’ in a way that ‘imperil[ed] the complete spectrum of constitutional liberties’.34 For those concerned with religious liberty, few cases were held in lower esteem.35

Today, looking back, it is becoming increasingly evident that the worst fears have not been realized. Because of strong legislative and judicial responses to Smith at both the federal and state levels, classic free exercise protections insisting on strict scrutiny of burdens on religion or at least on heightened standards of review have been reinvigorated to the point that they are perhaps even more robust where they apply than they were under pre-Smith law, though as will be seen, gaps in coverage clearly remain in the small set of states that have followed Smith.36

The subsequent developments are a testament to the resilience of free exercise of religion in the United States.37 By a nearly unanimous majority of both chambers,38 Congress passed the Religious Freedom Restoration Act of 199339 (‘RFRA’). This legislation reasserted the compelling state interest test as a matter of federal legislation. Specifically, RFRA prohibits government from imposing a substantial burden on a person’s exercise of religion even if the burden results from a rule of general applicability unless it can demonstrate that the burden ‘(1) is in furtherance of a compelling governmental interest; and (2) is the least restrictive means of furthering that … interest’.40 Congress took the position that it could impose a higher standard of religious freedom by legislation than the Court (in Smith) had concluded the Constitution required. The sense that the RFRA had restored religious freedom protections to their pre-Smith levels is illustrated in Figure 6.3.

The Court responded in City of Boerne v. Flores41 that Congress lacked authority to impose the burden of providing such heightened protection of religious freedom on the states. Specifically, the Boerne Court held that in passing the RFRA, Congress had exceeded its authority under Section 5 of the Fourteenth Amendment to pass legislation aimed at enforcing the substantive rights protected by Section 1 of the Fourteenth Amendment, including free exercise of religion.42 Thus, after Boerne the level of free exercise-type protection available was reduced significantly depending on whether federal or state law applied. This is illustrated in Figure 6.4.

But nothing in Boerne prevented the states from electing to impose such heightened burdens on themselves. Accordingly, 13 states – Alabama, Arizona, Connecticut, Florida, Idaho, Illinois, Missouri, New Mexico, Oklahoma, Pennsylvania, Rhode Island, South Carolina and Texas – adopted, either by state legislation or constitutional amendment, RFRA-like norms of their own.43 Subsequently, Congress itself responded to Boerne by identifying a number of specific areas in which it did have authority under the federal Constitution to reintroduce compelling state interest standards. The result was passage of the Religious Liberty and Charitable Donation Protection Act of 199844 (RLCDPA), which prevented bankruptcy trustees from recapturing donations made to religious organizations within three months prior to declaring bankruptcy; and the Religious Land Use and Institutionalized Persons Act of 200045 (RLUIPA), which imposed heightened religious freedom protections in settings involving land use decisions, and in institutional settings (primarily prisons and mental hospitals) where religious claims might otherwise be ignored. Certain revisions were also made to the RFRA itself.46 In addition, the American Indian Religious Freedom Act Amendment of 199447 (AIRFAA) was enacted shortly after the RFRA to make the religious use of peyote by Native Americans lawful, thereby overturning the specific result in Smith. The result of these state and federal legislative enactments greatly enhance the level of free-exercise protection existing after Boerne as illustrated above in Figure 6.5.

A number of courts48 and commentators49 initially suggested that the Supreme Court might yet invalidate ‘federal’ RFRA. Because Boerne only involved a challenge to the authority of Congress to impose the heightened requirements of the RFRA on the states, it did not address the constitutionality of the RFRA in federal contexts, the question remained open whether the RFRA was valid as a limitation on government action at the federal level. This question also carried over to the legitimacy of the other federal enactments which sought to reintroduce strict scrutiny tests on the basis of other enumerated bases for Congressional authority.50 However, cases decided by the Supreme Court,51 the Federal Courts of Appeal,52 many Federal District Courts53 and other commentators54 suggest otherwise. Most significantly, in Gonzales v. O Centro Espirito Beneficente Uniao Do Vegetal,55 the Court tacitly assumed the validity of federal RFRA56 by applying the statute to create a religious exception to a federal ban on hoasca,57 a hallucinogenic tea used in a religious community’s sacramental services. The factual similarity of this case to Smith tends to confirm the notion that the Court intends to apply the RFRA faithfully in those areas where it still applies. Significantly, the Court in Gonzales adopted a particularly vigorous interpretation of the RFRA. It rejected claims that abstract federal interests, such as enforcement of criminal laws58 or treaty obligations59 could automatically trump religious freedom claims. In its view, even where such apparently strong claims are present, an individualized assessment of the federal interests and possible alternative means of achieving those interests is necessary.

Free exercise resilience is also evident in state courts. Particularly when one considers that prior to Smith, state jurisprudence on free exercise issues had tended to move in lockstep with federal precedents,60 state supreme courts have shown surprising indifference to Smith. To date, only three state high courts – Maryland,61 new Jersey62 and Wyoming63 – have followed the Smith line of reasoning in interpreting their state constitutions. In contrast, a total of 11 states – Alaska,64 hawaii,65 indiana,66 Maine,67 Massachusetts,68 Michigan,69 Minnesota,70 new york,71 Ohio,72 Washington73 and Wisconsin74 – have interpreted their state constitutions’ free exercise and religious liberty provisions to require strict scrutiny or some other analysis more restrictive than Smith. Of the 10 other state high courts that have addressed free exercise claims, four of those courts – California,75 Colorado,76 Utah77 and Vermont78 – have expressly declined the opportunity to follow Smith and have deferred this decision until a future case. The remaining six high courts that have heard free exercise challenges – Iowa,79 Mississippi,80 Montana,81 Nebraska,82 Nevada83 and Virginia84 – did not explicitly indicate whether they were interpreting the state constitution or the federal Free Exercise clause when they applied Smith principles. The high courts from the remaining 13 states that have not passed state RFRA protections – Arkansas,85 Delaware,86 Georgia,87 Kansas,88 Kentucky,89 Louisiana,90 New Hampshire,91 North Carolina,92 North Dakota,93 Oregon,94 South Dakota,95 Tennessee96 and West Virginia97 – have not yet had an occasion to decide whether to reject or follow Smith. However, many of these jurisdictions have high court precedents antedating Smith that apply strict scrutiny standards in protecting free exercise. Significantly, many of the state courts that have refused to follow Smith as a matter of state constitutional law were in precisely this situation when the opportunity arose to consider the issue.98 Whether they have done so because they regard the earlier precedent as binding or as the more persuasive option, the result is that many of the as yet undecided states are likely to retain the strict scrutiny approach, though the fusion of federal and state standards in many of these cases leaves some uncertainty as to the applicable state standard in a post-Smith/Boerne setting.99 Moreover, in some of these jurisdictions, intermediate courts have adopted a strict scrutiny interpretation of state free exercise provisions.100 A state-by-state tally of each state’s approach to free exercise is illustrated in Figure 6.6.

As indicated by Figure 6.6, the bottom line is that at the level of the highest courts in the various states, only three states have expressly followed Smith in interpreting their state constitutions; six have followed Smith, but in contexts that left it uncertain whether the decision was applying federal or state law; four have expressly postponed decision of this issue, and 13 have not squarely faced the issue since Smith was decided. In contrast, a total of 24 states have expressly rejected Smith and have heightened but not strict scrutiny – 13 by statute or constitutional amendment, and 11 by judicial decree. Pre-Smith precedents in many of the undecided states still point towards a strict scrutiny approach. This means that only 4.9 per cent (14.9 million) of the US population live in states that expressly follow Smith,101 whereas 57.8 per cent (175.8 million) live in the 24 states that have expressly rejected Smith.102 The remaining 37.1 per cent (112.8 million) of the people live in the remaining 23 states103 where strict or heightened scrutiny is at least potentially applicable either because their high courts have not explicitly held that the Smith doctrine constitutes the appropriate interpretation of the relevant state constitutional norms or because the high courts of those states have not yet had an occasion to decide whether or not they will follow Smith. At a minimum, in these undecided states, a religious claimant would still have the ability to make a strict scrutiny argument under state constitutional law until the high court expressly rules otherwise.104 Given that many undecided states had cases applying a strict scrutiny standard before Smith was decided, and some have had intermediate rulings to the same effect, it seems reasonable to expect that a substantial number of the undecided states are likely to retain that approach.

Moreover, when one considers that federal RFRA and other federal religious freedom protection enactments cover several of the most important domains in which religious freedom claims are likely to arise, it seems clear that strict scrutiny protection remains available for a very high percentage of potential claims. Because nearly all of the US population live in a state where a strict scrutiny argument is controlling, or can at least be asserted in good faith, the post-Smith interpretation of many states’ constitutions indicates that a significant additional measure of free exercise protection has been re-established. This is illustrated in Figure 6.7.

Not only does the analysis above indicate that strict scrutiny is potentially applicable to the majority of free exercise claims in the United States, but the assertion of strict scrutiny today is arguably more robust than it was prior to Smith. Before that time, free exercise claims were based entirely on federal and state constitutional grounds. Even with the use of the strict scrutiny standard claimants still had the burden of proving that otherwise valid legislation posed a constitutional violation. With the passage of federal and state RFRAs, as well as RLUIPA and other federal statutory free exercise based protections, claimants today can make claims under the more certain footing of a modern statutory remedy.105 As explained by the Court in Gonzales, the RFRA contemplates a ‘focused’ inquiry that requires the government to establish that its compelling interest is not merely generally applicable, a standard that would be fairly easy to meet, but that its compelling interest must apply to ‘the particular claimant whose sincere exercise of religion is being substantially burdened’.106 Because of this more difficult statutory standard, the ‘RFRA makes clear that it is the obligation of the courts to consider whether exceptions are required under the test set forth by Congress’.107 The possibility of enhanced free exercise protections arising from federal and state statutory rights is illustrated in Figure 6.8.

The US ‘Non-Establishment clause’, which provides that ‘Congress shall make no law respecting an establishment of religion’108 has also undergone significant revision in recent years. As noted earlier,109 the test that had been worked out by the early 1970s (the so-called ‘Lemon test’) held that state action constitutes an impermissible establishment of religion unless:

1. It has a secular purpose;

2. Its primary effect neither advances nor inhibits religion; and

3. It does not result in excessive entanglement of government and religion.110

This test raised a high if sometimes curvilinear and even ‘serpentine’111 wall between church and state, and tended as applied to enforce a strong separation of church and state. The test was supplemented in the 1980s by Justice O’Connor’s ‘endorsement’ test, which held that an establishment clause violation occurred where state action had either the ‘actual purpose to endorse or disapprove religion’ or ‘the practice under review in fact conveys a message of endorsement or disapproval’.112 In effect, her test reconceived the first two prongs of the Lemon test as assessing whether a reasonable observer would perceive a government programme as having a subjective intention or an objective effect of endorsing religion. As a practical matter, the Lemon test probably made most sense in the context of assessing financial aid; the endorsement test was more helpful in cases assessing use of religious symbols in public settings. The field was further confused by at least one case that relied primarily on historic practices. Thus, in Marsh v. Chambers,113 the Supreme Court sustained the practice of having a chaplain say a prayer at the opening of the sessions of a state legislature because the Congress that adopted the First Amendment had a similar practice.

In fact, not only was the wall of separation serpentine; it was never impermeable. Already at the end of the nineteenth century, the Court had held that it was permissible to provide direct aid to a legal entity controlled by a religious community because it had a secular purpose and a separate legal existence from the religious community.114 In the late 1940s, in the first case holding that the Non-Establishment clause was applicable to the states via the Fourteenth Amendment, the Court sustained a programme reimbursing parents of children attending parochial schools for costs of bus transportation they incurred.115 Subsequently, the Court upheld a statute requiring school authorities to lend textbooks to secondary school pupils attending parochial schools.116 State support of secular aspects of religiously-affiliated university education has generally withstood Non-Establishment clause scrutiny.117

Various forms of indirect aid have also been sustained. Thus, in Walz v. Tax Commission,118 the Supreme Court sustained property tax exemptions given to religious organizations. In support of its decision, the Court noted the neutral character of the exemption statute, which benefited both religious and non-religious organizations,119 as well as long historical practice120 and the fact that tax exemption helped avoid entanglement of religion and state.121 Similarly, in Mueller v. Allen,122 the Court sustained a state personal income tax deduction for transportation, textbook and tuition expenses that benefited children attending parochial schools. There is a debate about whether tax exemptions constitute a de facto albeit indirect subsidy, or merely reflect a domain outside the normal tax base that the state lacks jurisdiction to tax. Tax expenditure analysts assume that they are a subsidy, but there are weighty arguments (beyond the purview of this chapter) to the contrary.123

On the other hand, the case that crystallized the Lemon test struck down a statute that provided publicly funded supplements to the salaries of teachers of secular subjects in parochial schools.124 In the Court’s view, this type of subsidy caused excessive entanglement between church and state because the legislature could not assume that parochial school teachers would avoid infusing their secular subjects with religious content, and the necessary monitoring to assure secularism in teaching would result in excessive entanglement.125 In general, the Lemon test made it difficult to sustain any aid to programmes provided within parochial schools because of their pervasively sectarian environment, and the sense that any aid would advance the overall religious mission of the institutions.

The difficulty with these tests is that they often obscured rather than clarified the actual principles that were actually deciding the cases. In an article summarizing the state of religious liberty law as it had evolved by the mid-1980s, Professor Douglas Laycock identified a set of six inconsistent theories that had been supported over the years by varying Supreme Court justices.126 Thus, some Supreme Court justices favoured a ‘no aid to religion’ theory, and held, as in Lemon