Why Michael Cannon is wrong about coercion in King v. Burwell

Michael Cannon of the Cato Institute has a thoroughly unconvincing rebuttal to those of us who have pointed out that his own legal gloss on the Affordable Care Act might be unconstitutionally coercive on the states. Cannon’s core error is that he repeatedly ignores the fundamental fact that health insurance markets aren’t like other markets because they are prone to adverse selection problems.

Remember, Cannon helped spearhead the legal challenge that became King v. Burwell, premised on the theory that Congress made subsidies available exclusively to states that created their own exchanges. I have argued (as have others) that such a tactic by Congress might be unconstitutionally coercive because it would threaten the states with insurance market death spirals if they refuse to comply. In a battle of statutory interpretations, the Court thus cannot sustain Cannon’s.

Cannon argues that this isn’t coercive at all under current Supreme Court precedent. His mistake, however, is brushing past the full scale of the consequences that follow if a state declines to create an exchange under King. Cannon argues that killing the individual mandate by cutting off subsidies would just impose additional costs on state residents. But this ignores the fact that on health insurance markets, these costs aren’t stagnant. Rather, these costs are dynamic and self-perpetuating until markets seize up altogether. Health care is different, and that difference makes the “choice” in King unusually coercive.

Cannon raises three separate points pushing back against the coercion argument:

Cannon #1: “The ACA’s Exchange provisions don’t penalize states. They let states make tradeoffs between taxes, jobs, and insurance coverage.”

Response: Whether Congress penalizes the states or their residents should be irrelevant under the Constitution’s federalism protections. And the tradeoff that states would have to make to decline to create an exchange under King is significantly more burdensome than Cannon admits.

“If a state fails to establish an Exchange, the ACA withholds subsidies from a state’s residents, not the state,” Cannon argues. This is overstated for two reasons. First, withholding subsidies under King not only has individual effects, but debilitating statewide effects, too. Because withholding subsidies triggers the individual mandate’s affordability exception for most consumers, state insurance markets (still obligated to comply with the ACA’s guaranteed-issue and community-rating requirements) will plunge into death spirals. That’s very much a penalty on a state as a whole, rather than just on its citizens.

Cannon goes to great lengths to contort the death spiral phenomenon into a mere imposition of costs.  Rather than call it a death spiral, he says that “withholding subsidies in uncooperative states would make the costs of the ACA’s community-rating price controls transparent to consumers, and those costs might have the effect of coercing states into implementing Exchanges.”

This is a deeply understated and incomplete formulation of what’s coercive about King. Cannon’s sleight of hand is treating the “costs” in an uncooperative state’s insurance market as if they were the costs typical of any other market. But this just isn’t the case. Health insurance markets are fundamentally different, and are singularly prone to adverse selection problems. The individual mandate is the lynchpin that secures stability in insurance markets with consumer-protecting regulations like guaranteed-issue and community rating. Removing this lynchpin doesn’t just unveil the ACA’s “true cost” (i.e., it’s cost without cross-subsidization from the young and healthy). Rather, it unleashes uncontrollable cost increases that culminate in a market collapse. That’s what would coerce the states, for the consequences of withholding subsidies impact far more than just the individuals that would have otherwise been eligible for them.

It’s also hardly clear that, for purposes of federalism constraints on congressional power, the distinction between states as sovereigns and their residents really matters. Certainly, New York v. United States suggests that it matters some. But New York itself reminds us that the underlying rationale of these federalism constraints is to protect individuals, not states. As the Court put it, “federalism secures to citizens the liberties that derive from the diffusion of sovereign power.” Drawing a bright-line and limiting coercion to direct federal salvos against state governments’ budgets would be an artificial formality that makes little sense given the whole raison d’etre of the federalism enterprise.

Such a formality would also be too easily circumvented for the Court to accept. In fact, at oral arguments in King, Justice Kennedy signaled that this very distinction matters less than Cannon thinks to the coercion inquiry. Kennedy said that the court “wouldn’t allow” Congress to impose a thirty-five mile per hour speed limit on states that don’t go along with a federal command (at 19:3). As in King, such a restriction wouldn’t impose a direct penalty on state budgets, but would instead negatively impact its residents directly. But such a negative impact would still be coercive, according to Justice Kennedy. That means that a regulation can still be coercive even if it doesn’t directly affect state budgets.

Moreover, if King contemplates the states making well-considered “tradeoffs” under the ACA, it’s hardly clear that they have been able to do so, which gets to Cannon’s second point:

Cannon #2: “Roughly half of states appear to consider those costs [of declining to create an exchange] tolerable.”

Response: These states haven’t reckoned with the full costs of declining to create an exchange under King. And just because a state chooses to accept the federal government’s punishment doesn’t make it constitutionally acceptable for Congress to pose a coercive choice in the first place.

This state acquiescence argument has appeared frequently since oral arguments. How could the exchange choice be coercive, its proponents ask, if so many states have refused to create exchanges?

The problem, as I’ve written, is that not a single state has yet embraced the full range of consequences of refusing to create an exchange under King. No state has publicly indicated that it has knowingly and voluntarily accepted an insurance death spiral as a consequence of this choice. States may be willing to take ownership of losing subsidies, and are happy to trumpet freedom from the individual mandate. Some even now claim, post hoc, that they knew they’d lose subsidies at the time they declined to create an exchange. But so far, not a single state has acknowledged the deeply destabilizing impact on their insurance markets that follows from that choice under King.

But suppose a state did claim to accept this dire consequence. Would it make a legal difference to the coercion inquiry if states (even many states) decided to take the federal government’s punishment? What if before NFIB, states had decided to end their Medicaid programs in order to avoid Obamacare’s expansion of that program? Would the threatened loss of all Medicaid funding still be coercive? In fact, Texas — one of the country’s biggest potential markets for expanded Medicaid — was making noise about abandoning Medicaid altogether to avoid the mandatory expansion even before NFIB. I doubt that the Court would permit Congress to impose an otherwise coercive choice on the states just because it wasn’t perfectly airtight coercion.

Both in theory and in practice, the number of states declining to create exchanges should have little bearing on the degree of coercion in King. If anything, the fact that so many states have made this choice in the face of the draconian implications of King’s version of the ACA points toward other doctrines that cast doubt on the petitioners’ interpretation. Maybe the thirty-six states that declined exchanges didn’t have clear notice of the consequences of such a choice, in which case the petitioners’ interpretation violates the Pennhurst doctrine. Or maybe the states simply didn’t see the elephantine implications of the petitioners’ version of the ACA because Congress hid it in a statutory mouse-hole — another legislative no-no that Justice Sotomayor raised in oral arguments (at 23:5). However you cut it, the Fantasy Affordable Care Act conjured by the King petitioners ultimately collapses in on itself.

Cannon #3: “This ‘deal’ is comparable to what the Court allowed in NFIB v. Sebelius.”

Response: The “deal” here in unlike NFIB in that it threatens grave economic harm on the states that extends beyond depriving the program’s would-be beneficiaries of insurance.

Cannon draws an analogy between the Medicaid expansion remedy in NFIB and how the ACA would operate under King: “In NFIB, the Court allowed states collectively to turn down Medicaid subsidies for as many as 16 million poor people. The Exchange provisions permit states to do the same for 16 million higher-income residents.”

Again, the coercion here is about far more than just subsidies. The yes-or-no Medicaid expansion decision didn’t threaten broader economic harm on state insurance markets. There, the consequences of state refusal were borne entirely by those eligible for expanded Medicaid.

The consequences of the choice in King aren’t nearly so cabined. Under the petitioners’ reasoning, the choice of whether or not to accept subsidies has huge economic consequences for the sustainability of state insurance markets as a whole. As I’ve explained above, foregoing these subsidies doesn’t just mean passing up on help for “16 million higher-income residents”—it also means an insurance market collapse. If a state wants a functioning health insurance market, it’s very hard to turn down subsidies. That wasn’t a consequence states had to consider three years ago under NFIB’s Medicaid expansion holding.

*          *          *

Cannon’s rejection of the coercion argument thus doesn’t hold water. The threat levied against the states under his reading of the ACA is significantly graver than he’s willing to admit — and it’s grave precisely because of the unique problems endemic to health insurance markets.

There’s a telling moment in Cannon’s article, however. As a result of a state’s decision not to create an exchange, Cannon argues, “residents would then see lower taxes, more jobs, more hours, higher incomes, and more flexible health benefits.”

The supposed economic gains are sheer speculation by Cannon, and they are highly doubtful speculation at that given how frantically business has sought to stave off the consequences of a ruling in favor of the petitioners in King. But Cannon tips his hand with the last “benefit”: “more flexible health benefits.”

Cannon presumably envisions Red states increasing reliance on a conservative favorite: health savings accounts — personal accounts where an individual funds much of his or her own medical costs rather than counting on insurance.

This is a useful reminder that the conservative vendetta against health reform is about more than just health exchanges, government subsidies, the individual mandate, or Barack Obama. At core, it’s about rejecting the basic risk-pooling function of insurance in favor of a go-it-alone, bootstraps approach. It’s part of what Jacob Hacker calls the “personal responsibility crusade,” taking us from a society that pitches in together to protect one another and transforming us into one where you’re on your own instead. That, at heart, is what’s at stake in King v. Burwell and the continued fight for universal health reform.

Advertisements

Scott Pruitt’s King v. Burwell own-goal

Scott Pruitt, the attorney general of Oklahoma, published a response to Justice Kennedy in the Wall Street Journal following King v. Burwell oral arguments this week.  Pruitt tried to counter Kennedy’s notion that the plaintiffs’ version of the Affordable Care Act would be unconstitutionally coercive on the states — but he unwittingly proved Kennedy’s point entirely.

Pruitt dismissed Kennedy’s federalism concerns as patronizing coddling of the states.  “The states are not children that the federal government must paternalistically ‘protect’ from the consequences of their choices by rewriting statutes,” Pruitt maintained.  “In our constitutional system, states are free to make decisions and bear the political consequences, good or bad, of those choices.”

Fair enough.  But does Pruitt grasp the full consequences of a decision not to create an exchange if the ACA works as the King plaintiffs claim?  Pruitt thinks Oklahoma made the following choice: “Declining to establish a state exchange allowed Oklahoma to voice its strong political opposition to the Affordable Care Act as a whole, as well as to make a statement that it wanted neither the large-employer mandate nor the individual mandate to have effect within its borders. That was the trade-off. Oklahoma declined the premium tax credits, but freed itself of those mandates, and that was a choice the state was happy to make.”

Pruitt’s calculus leaves out the biggest cost on the scale: near-guaranteed insurance market death spirals, brought on by a market bound by the ACA’s guaranteed-issue and community-rating regulations, but freed from its individual mandate.  He leaves this out of his column for either of two reasons: he doesn’t know about it, or he knows the political fallout from trying to defend that impossible “choice.”

If he doesn’t realize the devastating implications of the plaintiffs’ theory for his state’s insurance industry, then Oklahoma’s decision wasn’t quite so clear-minded and informed, after all.  This isn’t surprising.  Though Pruitt now claims Oklahoma understood that it would lose out on subsidies, there’s no evidence that any state considered the prospect that it would lose insurance subsidies by declining to create an exchange — let alone that losing those subsidies would send its insurance market into a tailspin.

But if Oklahoma didn’t know these dire consequences, this runs right into another constitutional doctrine that cuts against the plaintiffs’ argument in King: the Pennhurst doctrine.  The Constitution requires that the states have clear notice of the terms of federal conditional spending grants.  Because states didn’t have sufficient notice of the consequences of declining to create an exchange, the Court should avoid an interpretation of the law that unconstitutionally imposes draconian consequences without providing the states with notice.

But perhaps Pruitt and Oklahoma do understand these consequences.  If so, omitting prospective market death spirals from his response to Justice Kennedy is telling, for it all but concedes that accepting such grave devastation to Oklahoma’s insurance industry would be impossible to defend.  Imagine a political figure having to explain to her state’s citizens that she induced complete dysfunction in the insurance industry and jeopardized the stability of their health insurance in order to defy President Obama.

Perhaps Pruitt can make a plausible case that missing out on millions of dollars in federal subsidies is worth it to escape the individual and employer mandates.  But there’s no tenable way to justify accepting the resultant economic carnage that would occur if a state refused to create an exchange under the petitioners’ argument.  Especially with healthcare and business stakeholders pleading with Red states to comply with Obamacare rather than tempt market chaos.  (Remember when the ACA was supposed to be an economic killer?)

So Pruitt’s response to Justice Kennedy only serves to illustrate Kennedy’s very point.  If one reckons with the full consequences of the plaintiffs’ version of the law, the consequences of snubbing the federal government are far too draconian to seriously entertain.  By omitting the death spiral from his cost-benefit calculus for Oklahoma, Pruitt’s silence admits what the Supreme Court is catching on to: the plaintiffs’ version of the law is too coercive to embrace.

King v. Burwell & constitutional avoidance

On Wednesday, the Supreme Court heard oral arguments in the much-anticipated ObamaCare case, King v. Burwell. The big story coming out of arguments was federalism: namely, the repeated concerns that Justices Kennedy and Sotomayor raised about whether the plaintiffs’ interpretation of the ACA raises problems under the constitutional rule against coercion of the states.

I’ve written about this theory against the plaintiffs’ argument repeatedly, so it was great to see it gain traction before the Court. But I wanted to take some time to walk through the details of how applying constitutional avoidance to disfavor the plaintiffs’ potentially coercive interpretation of the law would work in practice.

First, what role could constitutional avoidance play in the Court’s analysis? To start, King is fundamentally about whether the IRS has legal authority under the ACA to issue tax credit subsidies to individuals who purchase insurance on federal exchanges. This is a Chevron inquiry, as the Court must determine whether the IRS’s interpretation is reasonable under the meaning of the law.

To determine what’s a reasonable interpretation of the ACA, the Court will need to marshal its typical methods of statutory interpretation. One of these methods is the canon of constitutional avoidance. Under this canon, the Court will avoid an interpretation of the law that raises constitutional problems and adopt a competing interpretation.

In the King oral arguments, Justice Kennedy in particular signaled that the plaintiffs’ interpretation raises “a serious constitutional problem” under the prohibition on federal coercion of the states. The question then becomes whether there is an alternative interpretation for the Court to adopt.

The alternative interpretation does not need to be particularly compelling – it need only be “fairly possible.” Nor does it need to be the “most natural interpretation,” as Chief Justice Roberts reiterated in the 2012 individual mandate case. “[E]very reasonable construction must be resorted to, in order to save a statute from unconstitutionality,” he told us.

The upshot is that constitutional avoidance lowers the bar for the government. And this bar had already been lowered under Chevron. With avoidance, the government need not even prove that its interpretation is a particularly reasonable reading of the law — it must only be “fairly possible.”

In fact, based on a pair of recent applications of avoidance by the Court, the bar for the government under avoidance might be even lower. Take the 2009 case NAMUDNO v. Holder, which raised constitutional issues about Section 5 preclearance under the Voting Rights Act. A Texas water utility district (called “Northwest Austin”) argued that preclearance was unconstitutional, or in the alternative, that Northwest Austin should be allowed to bail out of preclearance.

Bailing out didn’t appear to be an option, however, because the VRA limited bailout to “political subdivisions” of covered states — and “political subdivision” was defined as counties, parishes, and other subdivisions that conducted voter registration.  As a water utility district, Northwest Austin was none of these things.

The Court nonetheless granted Northwest Austin bailout eligibility in order to avoid ruling on the constitutionality of Section 5.  It went so far as to brazenly read the statutory definition of “political subdivision” into statutory surplussage, holding that “all political subdivisions—not only those described in [the definition]—are eligible to file a bailout suit.”

This was an extremely aggressive application of constitutional avoidance.  The Court disposed of the requirement that a competing interpretation of the statute be fairly possible, simply observing that the Court’s “usual practice is to avoid the unnecessary resolution of constitutional questions” — seemingly regardless of whether there is a plausible interpretation to latch onto.

In comparison to NAMUDNO, King presents an opportunity for a much more straightforward application of avoidance doctrine.  Each side marshals a tangle of textual arguments, and the government’s argument that the phrase “such Exchange” means subsidy-eligible federal exchanges is, at the very least, fairly possible.  Adopting that interpretation doesn’t require doing nearly the kind of violence to the statutory text that the Court committed in NAMUDNO in order to stave off ruling on the VRA for four more years.

In fact, King is probably also an easier avoidance case than NFIB was.  In order to rule upon the mandate as a tax, Roberts had to go against both the text of Section 5000A, which called the individual mandate a “penalty,” and the repeated, adamant public statements of the Obama administration and members of Congress, insisting that the mandate was not a tax.

Some commentators suggest that the government’s interpretation of the law would not be the appropriate remedy for a constitutional avoidance holding. Josh Blackman argues that the principled way of applying avoidance would be to invalidate all subsidies on every exchange. Randy Barnett suggests that the Court “strike down the federal insurance regulations that allegedly create the ‘death spiral’ and threaten to ‘destroy’ state insurance markets unless states set up exchanges” rather than “judicially authoriz[ing] billions of dollars in subsidies that Congress refused to authorize.”

There are a number of problems with these analyses. For one thing, Barnett’s “judicial authorization” argument presupposes that Congress did not authorize subsidies on federal exchanges — the very issue King is trying to answer. What’s more, Congress has not attempted to override the IRS’s decision to give subsidies on federal exchanges. If Congress thought the IRS got the law wrong and was spending money that it didn’t authorize, it could have amended the ACA to clarify that subsidies are only available on state exchanges and curb the IRS’s discretion.

The analyses also upend the very purpose of avoidance. The canon of avoidance is grounded on a presumption that Congress intends to legislate consistently with constitutional principles. To toss out all federal subsidies would seriously disregard this presumption and undermine the very purpose of constitutional avoidance in the first place.

These solutions also do far more violence to the text of the statute than is necessary. They would excise the potential constitutional defect with a bazooka rather than a scalpel. In fact, they disregard the presumption in favor of severability, where the Court presumes that Congress meant for as much of a law to stand as possible where a portion is found to be constitutionally problematic. Significantly, this presumption stands even when a statute does not contain an express severability clause — which the ACA famously does not.

Neither of these solutions comport with how the Court handles remedies for laws that violate the anti-coercion principle. Looking at the controlling remedy in NFIB is instructive here. The unconstitutional part of the Medicaid expansion was the threat that states would lose all funding for “old” Medicaid if they didn’t enact “new” Medicaid. The Court didn’t strike down all funding for old or new Medicaid. Rather, it simply removed the coercive threat. It de-leveraged the Medicaid expansion by decoupling old Medicaid funding from state decisions about new Medicaid.

The analogous remedy in King would be to unwind the death spiral threat if states fail to enact exchanges. In the ACA, federal subsidies act as a key that unlocks insurance market protections (the employer and individual mandates). So to remove the coercive threat, the Court must decouple the subsidies from the states’ decision about whether to establish an exchange.

This does not require striking down all tax subsidies or invalidating the ACA’s insurance market regulations. It simply requires making subsidies available entirely independently of a state’s decision to create or not create a health exchange.

That’s how constitutional avoidance would likely play out in King. It’s a heavier thumb on the scale against the plaintiffs’ argument, and it ultimately preserves the status quo of insurance subsidies universally available and consumer-protecting regulations universally enforced across the country.