Friday, November 20, 2015

Even Mere State Refusal To Assist the Federal Government In Resettling Syrian Refugees Could Be Unlawful

by Michael Dorf

In the wake of last week's attacks on Paris, a majority of U.S. governors have announced that they will exclude Syrian refugees from their states. In light of the rigorous procedures already in place for screening refugees, the governors' proposed policy is wrongheaded, cruel, and potentially counter-productive, but I will focus here on a largely overlooked issue with respect to its legality. As others have noted, states lack the power to exclude refugees whom the federal governmernt admits, if what is meant by "exclude" are state laws or policies that forbid Syrian refugees from residing in their states. The federal decision to admit a refugee for settlement in the U.S. would preempt a state law forbidding a refugee from residing in any particular state.

But maybe (some of) the governors mean something more modest. Refugee settlement often involves cooperation of state and federal agencies, and states can refuse to cooperate with the feds. Although the federal government may purchase state cooperation from willing state sellers, Supreme Court case law does not permit the federal government to "commandeer" state or local executive officials for federal purposes.

Can the states refuse to cooperate on selective grounds? Let me set the stage with a hypothetical example. The case linked above is Printz v. United States. An interim provision of the Brady Act required local law enforcement officials to perform background checks on prospective handgun purchasers. Some local officials refused to perform those checks, and the Supreme Court invalidated the federal provision obligating them to do so as impermissible commandeering in violation of a tacit constitutional postulate of our federal system. Now the hypothetical case: Suppose that some local law enforcement official partially enforced the Brady Act by performing the background checks on non-white applicants for handguns but not on white applicants.

It is clear that this official would thereby act unconstitutionally. Even though he is not obligated to enforce the Brady Act at all--because to obligate him to do so would be impermissible commandeering--he may not selectively cooperate where the selection is based on an invidious characteristic, here race. So the fact that a state or local official may, in reliance on the anti-commandeering doctrine, refuse to cooperate at all with federal law enforcement, does not mean that state or local officials have carte blanche to refuse to cooperate selectively. The standard constitutional prohibitions on discrimination continue to apply.

Is discrimination against Syrian refugees invidious discrimination? The case law treats discrimination on the basis of national origin as tantamount to race discrimination. However, "national origin," as that term is used in the case law, refers to the country (or region) from which someone's ancestors came to the United States. Japanese-Americans who were U.S. citizens and excluded from the west coast and then sent to camps on the basis of their ancestry suffered national origin discrimination. By contrast, discrimination against people who are citizens or subjects of a particular foreign country is nationality discrimination, and at least in some contexts, such discrimination is permissible.

For example, bilateral treaties may confer special benefits or, less commonly, impose special burdens, on citizens or subjects of particular foreign countries. Likewise, during wartime, citizens or subjects of foreign powers with which the U.S. is at war may be treated as "enemy aliens" subject to various restrictions not applicable to otherwise similarly situated foreigners of other nationalities. And U.S. immigration law often treats people of different nationalities differently.

In each of the foregoing contexts, however, the decision to treat foreigners from one nation differently from foreigners from some other nation(s) is taken by the federal government in a domain in which federal power is at something like its zenith--with respect to foreign relations, war, or immigration. It is by no means clear that, in the face of a federal determination not to discriminate in some way against the nationals of a particular country, states have any power to decide to discriminate on that basis.

Yet let us suppose for the sake of argument that nationality (as opposed to national origin) discrimination by the states does not by itself trigger strict scrutiny, regardless of whether state or federal authorities engage in it. Even so, a state decision to disadvantage Syrian (but not other) refugees would be subject to strict scrutiny if that decision itself were motivated by illicit stereotypes. Gerald Neuman explained in a 2000 law review article discussing federal power to draw nationality distinctions:
If these distinctions are not defined in terms of race and are not motivated by racial prejudice--unlike the notorious Chinese exclusion laws, which had both of these characteristics--then they would not elicit heightened scrutiny under ordinary equal protection analysis.
And if the federal government may not draw nationality distinctions motivated by racial prejudice, then it follows a fortiori that the states may not do so.

Thus, we come to the nub of the question: Are the proposed state policies of non-cooperation with the federal government in resettling Syrian refugees motivated by racial prejudice (anti-Arab stereotypes) or some other presumptively unconstitutional prejudice, such as religious prejudice (anti-Muslim stereotypes)?

At the federal level, Senator Cruz has proposed explicitly barring Muslim refugees from Syria, while admitting Christian refugees. That could be a legitimate distinction, if grounded in the idea that Christians, but not Muslims, face "persecution" in their home country, but of course ISIS and the Assad regime persecute (different groups of) Muslims as well as Christians. Cruz also argues that ISIS operatives might embed themselves in Muslim refugee flows but not Christian ones, which is, statistically speaking, a fair assumption, but even so, that does not warrant the use of religion (or race) as a proxy for dangerousness. After all, viewed ex ante, it might have been plausible to think that there was a greater likelihood that persons of Japanese ancestry would commit acts hostile to the United States than other persons, but our quite appropriate national repudiation of Korematsu rests partly on the basic principle that even racial, ethnic, or religious stereotypes that are, statistically speaking, based on a real correlation, are an impermissible proxy for individualized assessment.

Senator Cruz's no-Muslims rule does not have the backing of the GOP leadership, so I'll put it aside to focus on the various gubernatorial proposals to refuse to cooperate with federal resettlement of  all Syrian refugees. Even these formally religion-neutral proposals raise serious equal protection issues, insofar as they are motivated by ethnic and religious stereotypes. As with Senator Cruz, I'm willing to assume that the governors are not motivated by anti-Arab or anti-Muslim animus per se (although many of their supporters clearly are). Let's grant that the governors' concern is sincerely about security. Even so, if the security concern operates through an ethnic or religious stereotype, and has a disparate impact on that ethnic or religious group, then it is presumptively invalid (i.e., triggers strict scrutiny).

The point is easy enough to see in other contexts. Suppose that a state legislature enacts a law subjecting residents of cities with populations in excess of 100,000 people to random sobriety tests as a condition of retaining their driver's licenses. The law is facially neutral. But suppose that the reason for the law's differential treatment based on residence is that in this particular state there is only one large city, and it is the home to nearly all of the Irish-Americans in the state. Suppose further that the legislature adopted the sobriety test law because legislators were relying on a stereotype of Irish-Americans as heavy drinkers. That would make the law subject to strict scrutiny because, even though facially neutral, it was adopted for the purpose of disadvantaging Irish-Americans and it in fact does that. It doesn't matter that the legislature's ultimate purpose was the quite legitimate, indeed compelling, interest in highway safety. Nor would it matter if it turned out to be true that, statistically speaking, there was a higher rate of drunkenness among Irish-Americans than among other people in the state. The ethnic stereotype would still be a presumptively impermissible vehicle for advancing the state's interest in highway safety.

Likewise here, even if a Syrian refugee is, statistically speaking, somewhat more likely to be a terrorist disguised as a refugee than is a refugee from somewhere else, that can't justify a state's refusal to cooperate with federal authorities in settling Syrian (but not other) refugees, if the state's reasoning process runs through assumptions based on ethnicity or religion. At the very least, the governors' selective non-cooperation policies would be subject to strict scrutiny, which they would likely fail, given that the rigorous screening process for refugees is a more narrowly tailored means of advancing the admittedly compelling interest in public safety.

Conclusion: Even if limited to non-cooperation, the governors' proposals are probably unconstitutional.

Postscript: In principle, what I have just said could apply even to heightened screening procedures for Syrian refugees relative to other refugees, if the decision to impose those screening procedures runs through ethnic or religious stereotypes. It would trigger strict scrutiny for a police force to subject only members of some racial group to screening before entering public buildings, even if, statistically speaking, members of that racial group were more likely to pose a threat to public safety. So would heightened screening for refugees from Syria be unconstitutional?

Maybe. Such extra screening procedures could be upheld because the neutral justification is more plausible: There appears to be a heightened risk from Syrian (and some Iraqi) refugees because of the location of the ISIS base. Appears, perhaps, but is it true? Given that so many of the people who pose a threat are nationals of European countries, heightened screening for Syrian refugees, relative to the screeening for tourist visas for Europeans, seems like a misdirected effort. Thus, if such screening does trigger strict scrutiny, perhaps it fails the narrow tailoring requirement.

In practice, however, judicial review of federal immigration decisions, if any, would be much more deferential than review of a state decision not to cooperate in federal efforts to settle refugees admitted by the federal government (after screening). Accordingly, there is little likelihood that such heightened federal screening would be invalidated.


Joe said...

Perhaps, we can compare Truax v. Raich with Scalia's dissent in Arizona v. U.S.

Thanks for the post.

Greg said...

The problem with the Irish sobriety example is that often it is quite difficult to determine the reason that the legislature passes a law. Suppose legislator A proposes that we subject Irish Catholics to random sobriety tests. Legislator B rightly recognizes that would be unconstitutional, and proposes instead that they should instead focus on large cities, based purely on city size statistics. Legislator A recognizes that Legislator B's proposal would disproportionately affect Irish Catholics, and signs on for that reason. Now, both Legislators A and B go to Legislator C and request their vote, A with a focus on how it affects the Irish, and B with how it matches accident per population statistics.

Given that convoluted history, how would we determine if the law was intentionally targeting Irish Catholics? Certainly Legislator A was, but Legislator B wasn't. Is Legislator C's vote "fruit of the poisonous tree" because of being lobbied by A, even if her decision is based on B's population arguments? Is Legislator B's vote also "fruit of the poisonous tree" because the idea of random sobriety tests was originally proposed in an unconstitutional manner? Does it even require vote-counting to determine legislative "intent" or is there some other method used?

I would argue this resistance to Syrian refugees is even worse than the example because the legislators themselves probably harbor a mix of justifications, some based on anti-muslim bias and others based on legitimate concerns about ISIS activity within Syria's borders. In that kind of mixed-motive environment, it seems exceedingly hard to establish that a law was enacted primarily due to anti-muslim bias.

This isn't to say what I think restricting Syrian refugees is a good idea, just the practical complexities of trying to invalidate restrictions in the courts, and more generally of invalidating laws based on legislative "intent."

Shag from Brookline said...

I was born in Boston (1930) and attended public school theres grades 2-12, college and law school there. Boston was largely Irish and Catholic back then. Neighborhoods were often described by Cathoc parishes, even by non-Cathoics such as me. Over several decades, Irish Catholic politicians prevailed in Boston elections. But the Commonwealth was controlled by Protestant Yankees and at the state level legislation was enacted whereby appointments to certain Agencies in Boston were made by the Governor, usually a Protestant or a non-Irish Catholic. It was well known that this was a means to limit Boston's elected Irish Catholic officials. The legislation made no reference to Irish or Catholics. To my knowledge the Supreme Court never addressed this along the lines discussed by Mike. This was before the Warren Court got into full swing. The public in MA needed no special scrutiny to know what was going on. But Irish politicians had a saying going way back before they made significant inroads that continues to this day: "Don't get mad; get even." Some of these limits on Boston have been lifted as Boston became more diversified as did the rest of the state. The Brahmins had fear of the then "other." There's always another "other." So I have no problem with Mike's Irish example, as the Irish in that example are the "other." Xenophobes look for the "other" du jour.

Michael C. Dorf said...

Greg is right that subjective motive tests are notoriously difficult to apply, and for that reason, they are often criticized, but this is what the constitutional case law pretty clearly requires in this context. See the Village of Arlington Heights case:

Joe said...

Legislative actions are not quite as opaque as the first comment posits though it can be at various instances. Weighing everything into the balance, including purpose and effect, as Shag's example suggests, it is possible to determine in various cases if wrongdoing is present. Intent or purpose alone is not all that matters, but at least as a factor in the analysis, it can be useful and determined reasonably enough.

Shag's example or some voter id law perhaps provides a real life example that can trump some theoretical bit of confusion. Mixed motives do not suddenly make things innocent here & the one cited is of limited help -- what of other concerns? Why singling out one group while others also have a terrorist threat? And, yes, other things factor in there, such as discriminatory effects. See also, Windsor v. U.S., where purpose and effect mattered, or various Establishment Clause cases.

Subjective tests (though I'm not sure if that is even required -- we can argue for some objective test of what a reasonable person would think the legislature was doing) are tricky but they are applied and in this context, it isn't QUITE as hard as all that in numerous cases.

Sam Rickless said...

Great post, Mike. Makes me think of the reasoning in Yick Wo.