I am glad to see that in the wake of the cert. grants for Hobby Lobby and Conestoga Wood, there has been a frothing up of interest in the issues presented by these cases, issues that we here have been discussing for quite some time at CLR Forum. In this post, I want to address one such new claim.

Professors Nelson Tebbe and Micah Schwartzman (T&S) recently argued that an exemption from the contraception mandate under RFRA for employers like Hobby Lobby or Conestoga Wood would violate the Establishment Clause. They elaborate on their claim here and here. Many of the arguments are derived from this paper by Professor Fred Gedicks and Rebecca Van Tassell. The core of the argument is that granting an exemption from the mandate would privilege or favor religion inasmuch as it would shift the burden of purchasing contraception to third parties–i.e., the employees of the exempted corporations. The key to understanding the argument is their reliance on a Burger Court case, Estate of Thornton v. Caldor, which involved an exemption for employees from working on their Sabbath day. A Presbyterian who wished not to work on Sunday sued Caldor after the company dismissed him from a management position because he would not work Sunday. Because the law took absolutely no account of the secular interests of third parties (the employers), the law was found to violate the Establishment Clause. The “unyielding weighting in favor of Sabbath observers” resulted in a major burden on employers. T&S rely especially on this quote of Judge Learned Hand cited in Thornton: “The First Amendment … gives no one the right to insist that, in pursuit of their own interests, others must conform their conduct to his own religious necessities.” T&S (as well as Gedicks and Van Tassell) note that the principle of Thornton was restated in dicta in a more recent case, Cutter v. Wilkinson, which involved the application of RLUIPA. Justice Ginsburg, in dicta, said that in applying RLUIPA, “courts must take adequate account of the burdens a requested accommodation may impose on nonbeneficiaries.”

I think the argument is interesting, but mistaken. In truth, I have never understood Thornton very well at all and find it to be a difficult case. So I’ll start with a few basic points about exemptions and RFRA.

First, any exemption in this context will be directed toward benefiting some religious practice, and by being so directed, it will necessarily not benefit all others–i.e., “third parties.” If all choices to protect a specific form of religious exercise violate the Establishment Clause, then all exemptions for religion are Establishment Clause violations. The only thing that would be left for legislators is a law like RFRA, which accommodates religious exercise generally. Could it really be the case that the only thing the Establishment Clause permits is all or nothing? I don’t think so, and the Court has never said so. Professor Schwartzman, in other contexts, has questioned whether religion is a special category at all. If that argument were accepted and given constitutional force, then even laws like RFRA would be unconstitutional, because if the choice to protect religious exercise over non-religious ethical belief advances religion, then both specific and general accommodations are unconstitutional. The Court has not adopted that view. As Corporation of the Presiding Bishop of Church of Jesus Christ of Latter Day Saints v. Amos (1987) put it, “This Court has long recognized that the government may (and sometimes must) accommodate religious practices, and that it may do so without violating the Establishment Clause.”

Second, all exemptions burden third parties in one way or another. An exemption from laws proscribing peyote smoking imposes social costs of various kinds on third parties. An exemption from compulsory school attendance laws does so as well. An exemption for prisoners from wearing prison uniforms will burden prison officials and guards, and ultimately, everyone who is invested in a uniform system of penal justice. Indeed, one could go much further: all rights have costs that fall on third parties (you pick the context–the speech clause, Miranda rights, etc.). Thornton does not say that any time there is any shifting of burdens, the Establishment Clause is violated. Chief Justice Burger’s opinion was much, much narrower than that. It left open the possibility that a more carefully crafted Sabbath exemption law would be constitutional. That is more or less the upshot of Sherbert v. Verner (which was treated as good law by Thornton), where the Court held that a Seventh-day Adventist could not be denied unemployment compensation benefits because she refused to work on the Sabbath. In affirming that case, the Thornton Court is also affirming that it is perfectly constitutional for a state to exempt employees from Sabbath work on religious grounds, thereby imposing the costs of that exemption on third parties. All that Thornton is saying is that a law which imposes extremely severe burdens on secular interests through an “unyielding weighting of” religious interests over those other interests, and which takes no account of the secular interests at all, is constitutionally problematic. Consider an example. Under the Connecticut law at issue in Thornton, a school that is open only 5 days a week would have to provide Sabbath day exemptions to any teacher that asked for it. The burden on the school might be so severe as to impede its ability to function–compelling it even to close. The Thornton Court said that it had to “take pains not to compel people to act in the name of any religion.” (emphasis mine). It’s that kind of extreme burden on secular interests that rendered this law unconstitutional. Another obvious example might be an accommodation that interfered with a third party’s religious freedom–compelling the third party to engage in religious activities. Yet while the Court has said that “[a]t some point, accommodation may devolve into ‘an unlawful fostering of religion,'” Amos, only an extreme and absolute imposition on third party interests would justify that conclusion.

Third, both Thornton and a case like Texas Monthly v. Bullock seem to suggest that the burden imposed on secular interests must be state-imposed. Here the question is somewhat complicated inasmuch as the “burden” on employees is said to result from the combination of private claims and state power. Nevertheless, what these cases concerned is the alleviation of burdens on religious or secular beliefs imposed by the state.

Fourth, T&S wonder why nobody has made much of the Establishment Clause claim. But I think there is a good reason. RFRA incorporates certain limits to accommodation. That is, it would be a very rare RFRA (or RLUIPA) accommodation indeed which was constitutionally problematic under Thornton, because all RFRA (and RLUIPA) accommodations need to satisfy the substantial burden, compelling interest, least-restrictive-means threshold. The law at issue in Thornton, according to the Court required an accommodation “no matter what burden or inconvenience this imposes” on third parties. But the standard for RFRA accommodations is not, “you must grant the accommodation no matter what burden or inconvenience this imposes.” Accommodations must pass the government compelling interest threshold. If they do, they seem very much not to be violations of the Establishment Clause rule laid out in Thornton. In fact, many of the arguments about third party harms that T&S make have already been briefed by mandate advocates as part of the RFRA calculus. So they haven’t been ignored. They just haven’t been analyzed under the Thornton Establishment Clause framework, because Congress already saw to that in the statutes.

But let’s consider the Establishment Clause precedents on their own.

What we are left with in Thornton is to attempt to weigh up the gravity of the burden on third parties. Matters of degree are the key (as of course they are in RFRA). A law which takes absolutely no account of the burden to third parties, which grants to religious claimants “an absolute and unqualified right,” Thornton, and in which the burden on third parties is extremely severe, will run into Thornton problems. But, in the main, and short of such extreme disregard for third party interests, the basic Amos rule applies. Choices by legislatures to allocate burdens as they see fit, including choices that accommodate religious interests at the expense of secular interests, are not Establishment Clause violations. As the Amos Court said, “where…the government acts with the purpose of lifting a regulation that burdens the exercise of religion, we see no reason to require that the exemption comes packaged with benefits to secular entities.” Indeed, Amos is exactly a case in which the Court was dealing with a law that exempted the secular activities of a church from Title VII restrictions, imposing just the sort of third party costs that T&S think constitutionally problematic.

What about the dicta in Cutter v. Wilkinson, in which Justice Ginsburg wrote that courts need to “take adequate account of the burdens a requested accommodation may impose on nonbeneficiaries.” Does that mean, as T&S claim, that “significant burdens on non-beneficiaries” resulting from religious accommodations violate the Establishment Clause? I don’t think so, though it is a point in T&S’s favor that there is language in Cutter that puts the issue in question. The full quote from Cutter is: “Properly applying RLUIPA, courts must take adequate account of the burdens a requested accommodation may impose on nonbeneficiaries.” (emphasis mine). That sounds like it means that the strict scrutiny standard of RLUIPA and RFRA, if “properly appl[ied],” itself incorporates the Establishment Clause limits raised by cases like Thornton (see my point 4 above). What the Cutter dicta indicates is that any application of RLUIPA which completely disregarded the secular interests of third parties, crafted “an absolute and unqualified right” in favor of the religious claimants, and showed an utter indifference to any burden imposed on third parties would run intoThornton problems. But RLUIPA, “properly appl[ied], does not have those problems. And neither does RFRA.

The part in favor of the T&S position is that Justice Ginsburg later says in another piece of dicta that “Our decisions indicate that an accommodation must be measured so that it does not override other significant interests.” That dictum might seem to support the T&S view. But I don’t think it does. First, that dictum as applied to the facts of Cutter deals not with all conceivable third party interests, but the state’s interest in prison security. Note the gravity of the interest that was rejected by the Court as sufficient to beat out the interest in religious accommodation: one might think that prison security and the safety of other inmates was indeed the sort of paramount third-party interest that should be well-nigh indefeasible. And yet it lost unanimously. If that sort of interest is not “significant,” then there will be many others that fail to trigger Establishment Clause concerns. Second, to the extent we are talking about third party interests, the citation for the dictum is Thornton, which, again, speaks in terms of “unyieldingly weighting” of religious reasons “above all others.” Again, if one properly applies the strict scrutiny test, there will be no Thornton problem. Just as there wasn’t in Cutter.

At this point, then, we need a clear description of the burdens that the contraception mandate is imposing and the nature of the shifting of those burdens onto third parties by RFRA.

1. The burden the mandate imposes on employers with religious objections. The burden imposed by the mandate in the case of for profit corporations is to pay directly for their employees’ contraceptive products–actions which these employers say violate their religious beliefs.

2. The burden that RFRA shifts to the employees who want contraception. Here, I will shamelessly poach from Eugene Volokh’s fine post on this topic: “Let’s be precise about the effect of an employer mandate exemption would have on employee: If the employees want certain implantation-preventing contraceptives, they would have to buy them with their own funds, rather than getting them for free through an employer-provided health insurance plan. They would thus be in essentially the same legal position as all employees are today, before the employer mandate kicks in, and as some employees will be in the future, if their employers are exempted from the mandate (for instance, because they have grandfathered plans). The employer isn’t forbidding its employees from using certain contraceptives. It’s just not paying for them.

The burden on employees would thus be a burden relative to what the ACA would provide in the absence of an exemption. The employees of a business that would get a religious exemption wouldn’t get a benefit (free implantation-preventing contraceptives) — apparently worth several hundred dollars per year — to which employees of most other businesses will be legally entitled.”

Now, back to Thornton. Is this an exemption on religious grounds that implicates the sort of extreme concerns as did Thornton? I don’t think so. The reason is plainly that employees are not being prevented from purchasing the contraception products that they desire. They are not being prevented from using their own salaries, paid by their employers, to purchase such products. So this is not the same as actually threatening a school with closure, or “compelling” it to close down, for example. It is not the same as subordinating “all secular interests in the workplace,” which is the language of Thornton. It is not the creation of “an absolute and unqualified right” that affects the core nature of employment itself. And it is not preventing anybody from buying as much contraception as they wish. The “cost” to third parties is that of not having someone else pay *directly and explicitly* for such products for them. To my mind, that is a far cry from the sort of extreme disregard of secular interests at issue in the law in Thornton. It misdescribes the holding in Thornton to claim, as do Gedicks and Van Tassell, that the Court “has tended to find an Establishment Clause violation when accommodation imposes a noticeable or perceptible increase in the marginal weight of a pre- existing burden on identifiable third parties, or creates such a burden where none previously existed.” If that were true, then accommodating Adele Sherbert would be a violation of the Establishment Clause and Amos would have come out the other way.

What makes these cases even less like Thornton is the precise nature of the contraceptive products objected to by Hobby Lobby and Conestoga Wood, something that just about everybody skips right over. The products at issue are Plan B, Ella, and IUDs. Plan B and Ella are emergency contraception, not ordinary contraception. They are not the pill. Plan B is available over the counter. Therefore, the burden of not having insurance coverage for these products is relatively low because they are emergency drugs and insurance plans for these employers will cover other forms of contraception. As to IUDs, it is true that these are more costly up front, but they may be less costly over time than the pill (I am not certain about this). The point is that the actual exemptions sought in these cases do not come close to raising the kind of “absolute and unqualified right” that got the Connecticut law in Thornton into hot water.

T&S counter that the “baseline” for what is an extreme disregard for a secular interest has changed because Obamacare entitles employees to employer-paid contraception. In making this argument, they seem to adopt a claim made by Joey Fishkin that Obamacare makes employers like Hobby Lobby and Conestoga Wood into “federal agents.” Note that this claim would, if accepted, get around the third problem I noted above: namely, that the issues in Thornton and Texas Monthly only really relate to burdens imposed on secular interests by the state. If Hobby Lobby is also the state (or an agent of the state), then all private employers are agents of the state. So, at least in this context, you can extend the Thornton rule to all private corporations. Therefore, T&S claim, in this new “world which we now occupy” the burden is all the more severe on employees of religious corporations, because everybody else is now getting their contraception paid for.

But of course, as T&S acknowledge, it is a contestable question that the new dispensation that created this brand new world of private corporations that are government agents represents the current baseline. I suspect that many people would reject this description of the baseline. And even accepting a baseline of no-cost contraception for FDA approved products, is it really so clear that post-fertilization-contraception-products-only RFRA claims fall below the baseline, and so far below it to implicate burdens of constitutional gravity?

In fact, rather than talk of baselines, it seems to me that this is the heart of T&S’s claim: “At root, the constitutional conviction is that it is unfair and unconstitutional for the government to impose any substantial costs of a religious exemption on a focused and identifiable class of third parties.”

I certainly respect that this is T&S’s constitutional conviction, but it is not one that I share; it is not one that the Establishment Clause has ever been interpreted to require; and it is not reflected in the Supreme Court cases. I began this post by noting that the whole point of religious exemptions is to shift certain costs to others. That is what exemptions do–cost-shift. The rule of Thornton did not say that any “substantial cost” shifting is unconstitutional. It said that in extreme cases–such as the law it considered–accommodation may implicate Establishment Clause concerns. And as to “a focused and identifiable class of third parties,” I think that T&S are including this at least in part in order to get around the conscientious objector cases. This is an argument that one finds in the Gedicks and Van Tassell paper as well. Yet I do not understand why the Establishment Clause should be said to reflect such a distinction. Why is it, exactly, that accommodations on religious grounds are constitutionally acceptable when they damage an “unfocused” group but violations of the Establishment Clause when they impose costs on a “focused group.” And who decides what is focused and unfocused? And what is the class in this case which is “focused and identifiable”? All of the women in the workforce who wish to receive free contraception from their employers, together with all of the men in the workforce whose spouses or female dependents are on their plans and also want free contraception? That group probably numbers in the millions. Firms that do not object to providing contraceptive coverage? Again, that would be a huge group, and not particularly focused.

I also cannot agree with the argument in the Gedicks and Van Tassell paper that what the Establishment Clause proscribes is accommodations which shift costs to third parties in such a way as to affect “the private ordering of such parties–that is, in their deciding whether or how to act in some relevant way.” (24) They write that exemptions for conscientious objection would not be a “material” factor in inducing people to act one way or the other. They write: “it is difficult to imagine that permissive exemption of religious pacifists from the draft would be a factor in the decision of nonpacifists to comply with or evade the draft.” But of all the examples available, that seems the one most directly against them. In a footnote, they claim that people do all sorts of dangerous things every day, like drive a car or fly in a plane, and therefore people take risks with their lives all of the time. I don’t find that explanation for why exemptions from military service for religious people would not induce nonreligious people to seek similar exemptions plausible. Perhaps exemptions from oath-taking would have been a better example, though even there one might argue that the costs on third parties could be quite substantial, or “material.” At any rate, I still do not see the connection to the Establishment Clause limits set out in Thornton.

In sum, interesting as the argument is, I am not persuaded that a RFRA exemption would violate the Establishment Clause in these cases. I should add that the question whether RFRA compels these exemptions is, for me, a difficult one. But I don’t think the argument from the Establishment Clause works.