Sanctuary cities are likely to be an important obstacle to President-elect Donald Trump’s possible plan to round up and deport large numbers of undocumented immigrants. Several legal scholars, myself included, have argued that Supreme Court precedents forbidding federal commandeering of state and local governments will make it hard for the Trump administration to force sanctuary cities to cooperate with federal deportation efforts. Thanks to the Tenth Amendment, the federal government cannot force states to help them enforce federal law. It is also limited in its ability to use conditional spending grants to do so.

In a recent LA Times op ed, prominent lawyer David Rivkin and legal scholar Elizabeth Price Foley argue that Trump will have much broader power to force sanctuary cities to do his bidding, because the Supreme Court’s precedents banning federal commandeering of state governments supposedly do not apply to efforts to compel disclosure of information:

This “anti-commandeering” doctrine, however, doesn’t protect sanctuary cities or public universities — because it doesn’t apply when Congress merely requests information. For example, in Reno v. Condon (2000), the Court unanimously rejected an anti-commandeering challenge to the Driver’s Privacy Protection Act, which required states under certain circumstances to disclose some personal details about license holders. The court concluded that, because the DPPA requested information and “did not require state officials to assist in the enforcement of federal statutes,” it was consistent with the New York and Printz cases.

Rivkin and Foley conclude that Trump could force sanctuary cities to disclose the names and locations of undocumented immigrants, thereby facilitating deportation. Even if they are correct about this, it is important to recognize that their argument only applies to disclosure of information. It does not allow the federal government to force state and local officials to actually help apprehend and deport migrants. This is an important constraint because federal law enforcement personnel are extremely limited in number, and often need state and local cooperation to round up, detain, and deport large numbers of people.

Moreover, Rivkin and Foley’s argument only applies to information that state and local officials actually have in their possession. They could choose not to keep records on the names and locations of undocumented migrants, and not to check the immigration status of people arrested for various minor crimes and the like. Forcing local officials to gather new information goes well beyond mere disclosure and would surely qualify as commandeering. The Supreme Court said as much in its 1997 decision in Printz v. United States, where it struck down a federal law that required state officials to gather information on the backgrounds of prospective gun purchasers.

Rivkin and Foley’s analysis falls short even on its own terms. Contrary to their assumptions, Reno v. Condon does not give the federal government a blank check to compel state and local authorities to disclose information. Rather, the Court emphasized that the federal law in question was constitutional only because it “does not require [states] to enact any laws or regulations, and it does not require state officials to assist in the enforcement of federal statutes regulating private individuals” (emphasis added). The disclosures required in Reno were intended to help enforce a federal law that prevented states themselves from violating the privacy rights of citizen drivers’ license holders. By contrast, the whole point of forcing disclosure of information about undocumented migrants is precisely to “assist in the enforcement of federal statutes regulating private individuals.”

If there were indeed a general information-disclosure exception to the Tenth Amendment’s prohibition on commandeering, it would constitute a major expansion of federal power. State and local governments have extensive information about hundreds of millions of people that the federal government could use to coerce us in numerous ways. Fortunately, there is no such exception in the Constitution, or in any of the Supreme Court’s precedents. Such an exception would also enable Congress to impose massive uncompensated burdens on state governments. Ultimately, mandatory information disclosure is a form of commandeering no less than any other. It makes little sense to forbid other types of commandeering of state and local resources, but give the federal government a blank check to compel disclosure of information.

Rivkin and Foley also argue that the federal government could use conditional spending grants to pressure sanctuary cities. I agree this is a possibility. But few if any current federal grants are conditioned on cooperation with deportation efforts. Any such conditions would need to be newly enacted by Congress, would apply only to future grants, would have to be unambiguously clear, and not be so severe as to be “coercive.”. They would also have to be related to the purpose of the grant. These requirements almost certainly rule out conditioning all federal grants to sanctuary cities on cooperation with deportation efforts, though Congress could potentially condition grants to local law enforcement agencies on that basis.

In sum, Trump is likely to have a much tougher time forcing sanctuary cities to do his bidding than Rivkin and Foley suggest. This is one of a number of ways in which constitutional federalism of a sort long advocated by conservatives might help impede the Trump agenda.

UPDATE: Michael Greve comments on this post here. He argues that “[Section] 1373 does not commandeer anything at all. It merely preempts certain laws and practices.” However, for reasons discussed above, the “preemption” amounts to commandeering and can be struck down for much the same reasons as conventional commandeering.

More strangely, he contends that the federal government might be able to pull federal funds from jurisdictions that refuse to comply with Section 1373 for the following reason:

Nor is section 1373 a grant condition; it’s a free-standing prohibition, enforceable if need be by judicial injunctions. It is not easy to explain why the feds have to keep sending money, especially for law-enforcement purposes, to jurisdictions that ignore the prohibition. And suppose that a federal grant program states in bold, unmistakable letters that the recipient government shall comply with all applicable federal laws (those prohibiting the diversion of funds, say, or other forms of corruption). Is that clear enough, and sufficient to encompass 1373? I’d think so.

To my knowledge, the relevant grant programs do not in fact have such sweeping language attached to them. But if they did, it most certainly would amount to a grant condition. What is a grant condition but a mandate that funding recipients must meet certain requirements or risk losing their grants? And this sort of condition would be unenforceable because of its excessive vagueness. A requirement of compliance with “all applicable federal laws” is surely far too imprecise to qualify as the sort of “unambiguous” advance statement of conditions mandated by Supreme Court precedent. If this sort of strategy could be used to evade the requirement that conditions must be clearly stated, that rule would be essentially gutted. The feds could circumvent it anytime they wanted to simply by indicating that grant recipients must comply with all “applicable” federal law, and then enacting some separate “free-standing” rule like Section 1373.

I agree with Greve that most federal grant programs for local law enforcement should not exist in the first place. I would be happy if Congress simply repealed them. But if it wants to keep them, but condition them on compliance with Section 1373, it must do so explicitly, and clearly. And the president cannot impose such a condition on his own.