The Travel Ban and the Ontology of the Compelling Interest Test

by Michael Dorf

In my Verdict column for this week, I discusse how the Trump administration's release of Travel Ban 3.0 should affect the pending SCOTUS challenge to Travel Ban 2.0. I mostly steer clear of the mootness questions on which briefs will be filed today. I assume that even if the Court holds the current litigation moot, the validity of Ban 3.0 will be litigated in the lower courts almost immediately. My column discusses both statutory and constitutional objections to Ban 3.0, but here I will focus on the latter.

As I note in the column, the substitution of one mostly-Muslim country (Chad) for another (Sudan), the addition of some Venezuelan government officials, and the addition of a country that sends virtually none of its nationals to the US (North Korea) does not diminish the strength of the plaintiffs' prima facie case of discrimination: (1) The policy has a pronounced disparate impact on Muslims; and (2) that disparate impact is intentional, in the sense that, but for Trump's clearly displayed anti-Muslim animus, Ban 3.0 would not exist.

Therefore, the proper legal question (assuming courts get over any justiciability hurdles and arguments that ordinary constitutional principles have no bearing in the immigration context) is whether the intentionally discriminatory policy survives strict scrutiny, i.e., whether it is narrowly tailored to a compelling interest. Here I want to ask whether it makes sense to apply strict scrutiny--rather than a rule of per se invalidation--to laws and policies (such as the Travel Ban) that are motivated by animus.

In March, I asked why the Establishment Clause had emerged as the chief basis for challenging the Travel Ban. After all, there are other constitutional provisions that forbid religious discrimination, such as the Free Exercise Clause or the equal protection component of the Fifth Amendment Due Process Clause. The answer, I suggested, is that the Establishment Clause is "structural"--a limit on what the government can do rather than specifically a reservation of an individual right. That difference is potentially important, I explained, because of cases holding that non-citizens outside the US have fewer constitutional rights than persons inside the US (or maybe even no such rights). A structural provision is not a "right," so it limits the government everywhere.

But might the structural nature of the Establishment Clause have another implication? Structural provisions of the Constitution tend to have an on/off quality: they are not subject to balancing or other tests for overriding them. Congress has the power to regulate interstate commerce, but if Congress attempts to regulate beyond the scope of the commerce power or any other power, the resulting law is invalid; courts do not say that Congress cannot regulate beyond the scope of its enumerated powers unless the law passes strict scrutiny. Likewise, courts do not say that the president cannot usurp congressional authority unless his doing so survives strict scrutiny.

To be sure, even in federalism and separation-of-powers cases, there may be a kind of balancing going on. Instead of allowing a prima facie infringement to survive, the courts take account of countervailing interests in considering whether there even is a prima facie infringement. That said, however, the on/off character of structural provisions has consequences.

For example, the Court ruled in Printz v. United States that Congress categorically lacks authority to "commandeer" state and local executive officials. In dissent, Justice Stevens worried that the Court's rule could prove disastrous in a national emergency, when federal officials might need to organize and mobilize state and local relief efforts before there is time to mount a response using federal personnel. Notably, the majority did not respond by saying that the anti-commandeering principle would yield in such a case because it is defeasible under the strict scrutiny test. The dissent was correct that the rule is non-defeasible.

Note that this does not necessarily mean the dissent was right about the bottom line. In a true emergency, it could be expected that state and local officials would voluntarily cooperate with federal orders, and such voluntary cooperation would not implicate the anti-commandeering rule. But I'm not now interested in whether the majority or dissent in Printz had the better of the argument. I'm interested in the common ground--namely, that in other domains, structural constitutional provisions are not subject to being overridden.

Not so for the Establishment Clause, at least not where the particular Establishment Clause claim is that government has discriminated based on religion. Then, the compelling interest test applies. So said the SCOTUS in Larson v. Valente, where it matter-of-factly applied the test in a religious discrimination case under the Establishment Clause.

And that makes a kind of sense. Given that laws challenged as discriminating based on race, national origin, and religion trigger strict scrutiny when the underlying claim is brought pursuant to the Equal Protection Clause (or the equal protection component of the Fifth Amendment's Due Process Clause where the federal government's laws or policies are at issue), it would be odd to subject such laws discriminating on the basis of religion to a rule of per se invalidation under the Establishment Clause.

But maybe the whole compelling interest test is misguided in such cases.

In a 1997 article in the Yale Law Journal, Jed Rubenfeld argued that the compelling interest test makes sense as a means of "smoking out" illicit government motive (borrowing a phrase from the Court's affirmative action case law), but that it cannot be justified as a means of balancing costs and benefits. The real harm of a race-based classification, Rubenfeld argued, is not the classification itself but the fact that the classification instantiates white supremacy or some other caste or racial spoils system--and strict scrutiny will help us figure out whether that is a fact.

Rubenfeld's article was titled Affirmative Action and so he considered cases in which government uses an express racial classification that it justifies on remedial, diversity, or other benign grounds. He says that strict scrutiny or some other kind of heightened scrutiny can be justified in such cases, if at all, on the smoking-out theory but not the cost-benefit theory. I tend to agree.

But notice what happens if we apply Rubenfeld's theory to laws and policies--like the Travel Ban--that are facially neutral but have a disparate impact that, we are satisfied based on the evidence, was intended: Because, under Rubenfeld's theory, illicit purpose is the unconstitutional thing itself--the thing that strict scrutiny would aim to smoke out--there is no reason to apply strict scrutiny, which, by hypothesis, can only be applied in such circumstances on the rejected cost-benefit theory. Instead, I take the logic of Rubenfeld's argument to imply that if a court finds that a law or policy has a disparate impact based on an illicit criterion (such as race or religion) and that disparate impact was purposeful (i.e., the law or policy was chosen because, rather than in spite, of the disparate impact), the court should find the law or policy invalid without proceeding to a further step of applying strict scrutiny.

Yet even though I find Rubenfeld's argument largely persuasive in its own domain, I think this implication is wrong. If a court were truly convinced that Travel Ban 3.0 is narrowly tailored to serve the compelling interest in national security, I would want the court to uphold the Ban, notwithstanding the fact that it was chosen for an illicit reason and has a disparate impact based on the illicit classification.

How likely is that? Not very. We need to imagine that some law or policy is indispensable to serve some extremely important policy objective but that the government decision maker or makers did not adopt the law or policy for the purpose of serving that purpose; rather he or they adopted it for some illicit purpose; it just coincidentally confers these vital benefits.

I do not think that this is likely to be the right outcome of the application of strict scrutiny to Travel Ban 3.0, mostly because there have been exactly zero acts of terrorism perpetrated by nationals of the listed countries over the last several decades. Based on the statistics, a ban on US citizens entering the country would be more effective in protecting innocent lives.

Still, it is possible to imagine a much closer case. Suppose that after conducting the review of screening procedures, the administration had come up with something very different but a policy that still had a disparate impact on Muslims. In those circumstances, we might conclude that Trump's anti-Muslim bias was a but-for cause of the ultimate policy--because it set in motion a course of events that would not have occurred otherwise but that ultimately led to Ban 3.0--AND that the policy chosen by others who don't share Trump's anti-Muslim bias satisfies strict scrutiny. That conclusion, in the admittedly hypothetical case, leads me to think that Rubenfeld was at least partly wrong. There are circumstances in which the compelling interest test is properly conceived in cost-benefit terms rather than as a smoking-out device.