NYU law professor Rick Hills has written a thoughtful response to my post on federalism arguments in King v. Burwell:

Alas, I find that Ilya Somin over at Volokh’s is, in the context of King v. Burwell, back-pedaling furiously away from NFIB’s anti-coercion principle. As explained by Abby Moncrieff’s excellent amicus brief, that principle is relevant to King in the form of an “avoidance canon.” The Affordable Care Act, on Abby’s reading, should not be construed to put state governments to the choice of either setting up healthcare exchanges are having their citizens lose access to tax subsidies. Such a choice is precisely similar to forcing Texas to choose between either carrying out the feds’ greenhouse gas policy (on one hand) or having the feds impose extra-onerous burdens on Texas’ private industry (on the other).
As I explain after the jump, Abby is absolutely correct, at least if NFIB Part IV(B) is given a reasonably broad reading. Yet I find that Ilya Somin is dismissing these worries about coercion by offering a bizarrely narrow reading of NFIB’s anti-coercion principle.
Is Ilya just another fair-weather federalist who forgot his decentralizing principles as soon as his immediate political or ideological interests cut the other way? I think that the problem is better understood as a strategic rather than moral problem: Libertarians and conservatives do not want to sacrifice their litigation priorities for the sake of general principles of federalism until they get a credible commitment from liberals like Abby that the liberals will follow suit when liberal priorities are threatened by robust decentralization. After the jump, I’ll explain how King presents a golden opportunity for such a grand bargain.

I am not doing any “backpedaling” from NFIB, because its anti-coercion principles, like those of previous Supreme Court precedents in this field, apply only to federal grants to state governments, not to grants to private individuals. Chief Justice Roberts’ controlling opinion in NFIB explicitly states that “the legitimacy of attaching conditions to federal grants to the States depends on the voluntariness of the States’ choice to accept or decline the offered package.” It does not apply that principle to subsidies offered to individuals and does not repudiate the distinction between the two made in New York v. United States, which I discussed in my previous post. Rick notes a 1937 case that suggests that the coercion theory may apply in regulatory cases, without actually deciding the issue. But that possibility was repudiated in the much more recent New York decision. As far as Supreme Court precedent is concerned, there is no coercion problem in cases where the federal government incentivizes the states to do something by regulating private parties or denying them subsidies. Even if there were such a problem, we can’t solve it by adopting an interpretation of the ACA under which the federal government actually has more power and the states have less choice than they would under the plaintiffs’ view.

If we set precedent aside, and ask what courts should do to move towards the correct interpretation of the Constitution, I would gradually move away from the coercion approach and enforce a more limited (and more originalist) theory of what counts as spending for the “general welfare” – a position I espoused during the NFIB litigation, and long before. Thus, contra Rick, I am in no way “soft-pedaling federalism ideals that, in other contexts, [I] endorse.” I have never endorsed the coercion approach as the right way to limit the federal spending power, and have only been willing to support it in limited circumstances as a lesser evil compared to virtually unlimited federal spending authority. And since I view it as a dubious second- or third-best strategy for limiting unconstitutional applications federal spending power, I see no reason to use it to justify actual expansions of federal power in cases where that application is not required by existing precedent.

Rick also proposes a kind of grand bargain between left and right, under which both would agree to support a broad interpretation of the anti-coercion principle. That might be less bad than some other possible options. I am willing to accept imperfect bargains in cases where the available alternatives are even worse. But, as Rick himself notes, it’s highly unlikely that any such bargain will be accepted by both sides and consistently applied. Still, when and if such a bargain actually becomes feasible, I would be happy to support it if there are no better realistically feasible options – even if that would mean accepting the federal government’s interpretation of the ACA in King (which, as I said in my earlier post, could well be correct for reasons unrelated to federalism and coercion). But it seems improbable that any such opportunity will emerge between now and the time when the Court decides King.

UPDATE: Rick has posted a well-reasoned and thoughtful response to this post here, focusing mostly the correct interpretation of New York v. United States:

Ilya construes too broadly New York [v. United States’] imprimatur for conditional preemption of state laws. Of course, New York permits Congress to offer state officials the option of escaping federal preemption of state law by regulating according to federal standards. But New York makes perfectly clear that this choice cannot be compelled. In New York’s words, “[w]here Congress encourages state regulation rather than compelling it, state governments remain responsive to the local electorate’s preferences in the state’s place” (emphasis added). In short, contrary to Ilya’s claim, New York is not a nonsensically formalistic opinion that carefully limits one sort of commandeering but gives carte blanche for another functionally identical variety.

Rick is correct that New York differentiates between federal laws that merely “encourage” state regulation and those that “compel” it. But this distinction between encouragement and compulsion depends in large part on whether the law in question imposes conditions on the receipt of federal funds by state governments or instead subsidizes or regulates private parties. As Justice O’Connor writes in the majority opinion, in the latter situation “[t]he affected States are not compelled by Congress to regulate, because any burden caused by a State’s refusal to regulate will fall on [private citizens], rather than on the State as a sovereign.”

This distinction may be flawed. But, contra Rick, it is not “nonsensically formalistic.” As I noted in my previous post on this issue, if federal laws become coercive any time they indirectly impose major burdens on state governments by regulating private parties, a wide range of federal legislation could potentially be challenged. And the magnitude and effect of such indirect burdens are likely more difficult for courts to measure and assess than the withholding of federal funds from state governments.

Moreover, if we are going to eliminate seemingly formalistic distinctions in this area of law, we must also get rid of the distinction between explicitly stated regulatory conditions and federal laws that “only” implicitly impose painful choices on state governments, thereby forcing them to change important policies. For example, federal income tax laws greatly diminish the pool of income available for state governments to tax, and thereby massively reduces their potential revenue, forcing them to restrict their spending much more than they might have to otherwise. The “coercive” impact is far greater than the effects of denial of the ACA insurance subsidies at issue in King v. Burwell. If we abandon “formalistic” distinctions and focus only on the actual real-world effects of federal laws on state governments, federal income tax laws are far more coercive than the plaintiffs’ interpretation of the ACA. The same would be true of many other important federal laws.