The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent


The test to textualism in King v. Burwell: A Reply to Abbe Gluck


Tuesday afternoon I had the pleasure of debating/discussing King v. Burwell with Professor Abbe Gluck at the Yale Law School. Professor Gluck is one of the more thoughtful and persistent critics of the plaintiffs' claims in King and the other cases challenging the lawfulness of the IRS rule granting tax credits for the purchase of health insurance in exchanges established by the federal government. The event, co-hosted by the Yale Federalist Society and American Constitution Society, was moderated by Linda Greenhouse, and they packed the room. (Maybe it was the food.) Alas, the event was not recorded, so you'll have to take my word that we had a vigorous yet civil and informative exchange. I think those assembled got a good sense of the competing arguments on each side in these cases.

Since I can't post a link to audio or video (and I had some time at the airport), I thought I'd instead take the opportunity to respond to Professor Gluck's contribution to the recent SCOTUSBlog symposium on King v. Burwell, "The grant in King—Obamacare subsidies as textualism's big test." It's a thoughtful piece, but I don't think I'm revealing a spoiler when I say I'm not convinced. Indeed, I think this piece helps show why the government has such a hard time mounting a textualist defense of the IRS rule.

For starters, it's interesting to note that Professor Gluck frames her piece as a defense of textualism. She writes that "King is about the proper way to engage in textual interpretation; specifically, about the interpretation of five words in a long and complex modern statute." I'd agree with that. While I don't think the argument against the IRS rule is confined to these five words ("Exchange established by the State"), Professor Gluck is correct that the meaning of this phrase should control the outcome.

Professor Gluck also maintains that "no one has to—or should—go outside the four corners of the Affordable Care (ACA) to decide it." I agree with that too (and I suppose this means Professor Gluck is no longer urging the adoption of a "CBO Canon" to resolve this case). When Michael Cannon and I embarked on our survey of the legislative history of the PPACA's exchange and tax credit provisions for our Health Matrix article, we did so because we knew many commentators (and, potentially, courts) would care about such things (and because the PPACA provides an interesting case study of unconventional lawmaking). But I have always maintained that the text should control and that the text should be interpreted in accordance with traditional norms of statutory interpretation.

From what Professor Gluck writes in her SCOTUSBlog essay, we agree on the importance of various textualist principles, such as the need to look at the context of specific textual provisions and to consider the overall structure of a statute when seeking to identify what the text means. Professor Gluck cites Justice Scalia's opinion in UARG v. EPA, a case from last term in which the Supreme Court confronted another large, complex, and at times difficult-to-parse statute: the Clean Air Act. Specifically, she cites Justice Scalia's statement that the Court must do its best to observe the "fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme." Agreed, but there's more to it than that.

What Professor Gluck omits is what else Justice Scalia has to say in the same opinion just a few pages later:

An agency has no power to "tailor" legislation to bureaucratic policy goals by rewriting unambiguous statutory terms. . . .

The power of executing laws necessarily includes both authority and responsibility to resolve some questions left open by Congress that arise during the law's administration. But it does not include a power to revise clear statutory terms that turn out not to work in practice. . . .

We reaffirm the core administrative-law principle that an agency may not rewrite clear statutory terms to suit its own sense of how the statute should operate.

Justice Scalia's UARG opinion instructs courts to both consider context and to preserve the meaning of clear statutory terms. Further, as Justice Kagan explained in another case from last term (Michigan v. Bay Mills Indian Community), courts have "no roving license, in even ordinary cases of statutory interpretation, to disregard clear language simply on the view that . . . Congress 'must have intended' something" other than what the statute's text actually says.

The scenario Justice Scalia was addressing in UARG is actually quite similar to what we see in King. The Clean Air Act has some provisions that are not entirely clear. A term like "pollutant" necessarily means pollutant, but which pollutants depends upon the context and operation of a given provision. Even when Congress refers to "any pollutant" and provides definitions for the term, there is some residual ambiguity about the scope of the term (though, again, "pollutant" must mean pollutant).

At the same time, there are other terms and phrases in the statute, such as "100 tons-per-year" or "250 tons-per-year," that are perfectly clear. While the Court has an obligation to consider context, particularly when trying to bring precision to the definition of potentially ambiguous terms (or to identify the constraints within which an agency may operate after a finding of ambiguity), the Court must also ensure that "clear statutory terms" mean what they say. Only when a Court does both of these things can we say that it has followed textualist principles and adopted an interpretation that, as Justice Scalia further counsels, "does least violence to the text."

The parallel between UARG and King is instructive. Both involve terms ("pollutant," "Exchange") that are defined in the statute yet may yield to slightly different interpretations when placed in specific contexts. At the same time, both include phrases (the CAA's numerical emission thresholds, "established by the State") that are perfectly clear. Under UARG, and fairly traditional rules of statutory interpretation, it may be permissible to stretch or mold the understanding of the former given context and other factors, but the clear statutory text of the latter must be observed. These words cannot be written off or treated as a nullity.

The relevant provisions of the PPACA in Section 1401 (and it is provisions, plural, contrary to what many pundits claim and even if one ignores the cross-references), both make reference to Exchanges established by the State under [Section] 1311." If these provisions were referring to Exchanges, as generically defined under Section 1563, or even to "Exchanges established . . . under 1311," there would be no reason to include this additional language. (Language which, it is worth remembering, was added to multiple places at multiple times.) The law is clear about what "establish" means. This word is used throughout the text, and exchanges are identified as those "established" by states (1311) and those "established" by the federal government (1321). Further, Section 1321 instructs HHS to both "establish and operate" its exchanges. It is also clear what a State is, and that the states and the federal government are quite distinct.

Professor Gluck takes a different view. She writes:

Section 1401 can still be read literally because the section that authorizes the federal exchanges, Section 1321, provides that if a state does not establish an exchange under Section 1311, the Department of Health and Human Services (HHS) "shall . . . establish and operate such Exchange within the State." In other words, HHS must "establish" a Section 1311 exchange, which is a state exchange. Moreover, the Act defines "Exchange," with a capital E, three times in the statute as a "state" exchange. And HHS, in Section 1321, is told to establish "such [capital E] Exchange." The Court need not add or delete a single word of the ACA to reach this conclusion.

Note what her account omits, the repeated qualification that coverage months and premium amounts must both be calculated with reference to insurance purchased in an Exchange established by the State. Note also, as her account acknowledges, under Section 1321 it is HHS that is doing the establishing of the exchange. While embracing context, her interpretation ignores inconvenient-yet-perfectly-clear statutory terms.

A more complete consideration of context further illustrates the error of this interpretation. The statute defines "Exchange." Specifically, Section 1563 (the second 1563, as there are three in the PPACA) provides: "The term 'Exchange' means an American Health Benefit Exchange established under section 1311 of the Patient Protection and Affordable Care Act." Read in isolation, this definition might seem to support Professor Gluck's suggestion that all exchanges are to be treated alike for all purposes. But we should not read this, or any other provision, in isolation.

Section 1551 of the statute also provides that this definition applies "unless specifically provided for otherwise." So, when the word "Exchange" is used without qualification, we can presume the 1563 definition applies and the statute is speaking of a Section 1311 exchange. When it is modified or qualified, however, the statute instructs us to pay attention to that fact. Thus, when Section 1401 specifies that the Exchange must be one "established by the State," it does not matter that those established by the federal government under Section 1321 exchanges may be otherwise equivalent to Section 1311 exchanges for some purposes because the statute has specified that we're not simply talking about an "Exchange" or even a Section 1311 Exchange, but an "Exchange established by the State under 1311." The statute has identified an additional characteristic that these exchanges must satisfy for the purposes of these provisions.

Professor Gluck goes on to argue that other provisions of the law are "slashed to pieces" by an interpretation that insists "established by the State" means "established by the State." She cites two provisions in particular. Read in context, however, neither of these provisions proves her case, and there are other portions of the statute which undermine it.

Here is Professor Gluck's first example:

Section 36B(f)(3) requires "[e]ach Exchange (or any person carrying out 1 or more responsibilities of an Exchange under section 1311(f)(3) or 1321(c)" to report the premiums doled out. Section 1321 is the federal exchange provision, and so this section is rendered meaningless if the federal exchanges have no subsidies.

This is a very selective description of the reporting provisions added in the HCERA reconciliation bill. Yes, this provision requires the reporting of tax credits, but it requires the reporting of lots of things, including the level and duration of coverage, premiums "without regard to any credit," information on those covered by the policy, and more. This information matters to the federal government whether or not tax credits are issued. As the D.C. Circuit explained in its Halbig opinion, "Even if credits are unavailable on federal Exchanges, reporting by those Exchanges still serves the purpose of enforcing the individual mandate-a point the IRS, in fact, acknowledged in promulgating a recent regulation, 26 C.F.R. § 1.6055-1(d)(1)."

A premise of Professor Gluck's claim is that it would be nonsensical to apply these reporting requirements where there are no tax credits. Yet these provisions apply to all insurance purchases on exchanges, and not all those buying insurance on exchanges are eligible for tax credits. Moreover, the provision applies to "any person carrying out 1 or more responsibilities of an Exchange," so even those involved with exchanges that have nothing to do with tax credits must report relevant information. So, even under her interpretation, these provisions apply to entities and transactions where tax credits are absent. Finally, this provision—unlike Section 1401—makes explicit reference to both Section 1311 and Section 1321 exchanges, which would be completely unnecessary if, as Professor Gluck claims, the two types of exchanges are equivalent for any and all purposes.

Her next example highlights what many see as the most problematic statutory provision for the plaintiffs' case.

Section 1312(f) provides that only "qualified individuals" can purchase on an Exchange but defines a qualified individual as one who "resides in the State that established the Exchange." Failure to understand a federally operated exchange as the legal equivalent of a state exchange would mean that federal exchanges have no customers.

Here it's worth remembering that definition of exchanges I noted above. This provision (Section 1312(f)) is one of several identifying the requirements of Section 1311 exchanges. The reference to "an Exchange" under the Section 1563 is presumed to be a Section 1311 exchange which, under Section 1311, is an exchange established by the state. Read in context, this provision—like other parts of Section 1311, 1312, and 1313 that detail requirements of Section 1311 exchanges—speaks to the establishing state and is identifying a requirement of a state-established exchange.

Section 1321, however, while requiring the federal government to "establish and operate" the "required exchange" does not create an exchange "established by the State." In this way, Section 1321 takes us out of the default definition of an exchange. This exchange is not that defined under Section 1563 because it is otherwise specified as being one "established" under a different provision by a different entity. Therefore, we need not read Section 1312(f)'s "qualified individual" provision as directly binding on the Section 1321 exchange. Further confirmation of this reading comes from the fact that Section 1321 goes on to provide HHS with the authority to replicate the various requirements for Section 1311 exchanges. That is, Section 1321 anticipates that the requirements for Section 1311 exchanges cannot all, automatically apply to Section 1321 exchanges, so it gives HHS the power to fill those gaps.

Is this account of Section 1312(f) persuasive? Consider two more things. First, this interpretation, unlike that offered in defense of the IRS rule, preserves the plain meaning of "established by the State," and thus does less violence to the overall text of the PPACA than one which would give this phrase no meaning at all. Second, this provision—Section 1312(f)—at worst, is ambiguous. But if this provision is ambiguous, that tells us nothing about the meaning of Section 1401, and it certainly does not justify ignoring the text of Section 1401. As courts have noted time and again with other large, complex statutes, ambiguity in one provision does not render the entire statute ambiguous. Under Chevron, what matters is whether there is ambiguity on the precise question at issue.

Pretending that "established by the State" does not mean "established by the State" creates other problems when one looks at the statute as a whole. This phrase is used in several provisions throughout the PPACA. In each case, these provisions serve to facilitate coordination between state exchanges and other programs or to provide incentives for state action. Thus the use of "Exchange established by the State" in these provisions is not anomalous at all.

In some of these provisions, however, adopting Professor Gluck's interpretation—that "Exchange established by the State" means both those established by the States and those established by the federal government—creates some anomalous results.

For example, Section 1311(f) authorizes states to establish "one or more subsidiary exchanges." Pursuant to this, 1311(f)(3)(A) provides that "A State may elect to authorize an Exchange established by the State under this section to enter into an agreement with an eligible entity to carry out 1 or more responsibilities of the Exchange." If "established by the State" means an exchange that is, in fact, established by the State, this makes perfect sense. But if it also means an Exchange established by the federal government it creates the anomalous situation that a state that elected not to create its own exchange is deciding whether the federal government is allowed to enter into contracts to create subsidiary exchanges.

Then there's Section 2001, the so-called "maintenance of effort" provision. It provides that states are not allowed to tighten existing eligibility standards under the State's Medicaid program until the date on which the Secretary determines that an Exchange established by the State under section 1311″ is operational. A state that fails to comply risks losing its Medicaid funds. There are two ways to read this provision. Under one, this provision says to states: You can't throw people off of your Medicaid program by tightening existing eligibility rules until you've created an exchange. That makes sense, as it prevents states from creating a coverage gap, even if leveraging Medicaid in this way is constitutionally suspect under NFIB.

Now consider the alternative read. If the "Exchange established by the State" could be either a state or federal exchange, the provision is problematic. This is because for the state that does not create an exchange and wishes to alter its eligibility requirements, continued receipt of Medicaid money is conditioned on the federal government getting around to launching an exchange. The state's flexibility to alter its own program is not contingent on what the state does, but on what the federal government does. Thus the state's receipt of Medicaid funds is held hostage to federal action.

To understand why this is so problematic, it is worth revisiting the Medicaid portions of the NFIB v. Sebelius oral argument. If there was one thing that concerned the justices more than a potentially coercive use of conditional spending controlled by a clear statutory requirement, it was the prospect of a potentially coercive use of conditional spending controlled by administrative discretion. In the former case, the state has some certainty, and the notion that conditional grants of funds are like contracts is observed. In the latter case, however, all a state can do is cross its fingers and hope the federal government does that which is required to unlock their funds and programmatic flexibility. So however problematic one believes Section 2001 when one believes "established by the State" means established by the State, the alternative reading of this phrase is even worse.

Next let's turn to Section 2201. Here, under Section 2201(a), the PPACA states that "a condition of . . . receipt of any Federal financial assistance under section 1903(a) for calendar quarters beginning after January 1, 2014" is meeting several requirements. One of those is that the state must "establish procedures" for (among other things)

ensuring that the State agency responsible for administering the State plan under this title (in this section referred to as the 'State Medicaid agency'), the State agency responsible for administering the State child health plan under title XXI (in this section referred to as the 'State CHIP agency') and an Exchange established by the State under section 1311 of the Patient Protection and Affordable Care Act utilize a secure electronic interface sufficient to allow for a determination of an individual's eligibility for such medical assistance, child health assistance, or premium assistance, and enrollment in the State plan under this title, title XXI, or a qualified health plan, as appropriate;

So, the state must "ensur[e] that . . . an Exchange established by the State under section 1311 . . . utilize[s] a secure electronic interface." As applied to exchanges that are, in fact, established by the State, this makes sense. As applied to federal exchanges, however, this provision makes no sense. How can a state be expected to "establish procedures" to ensure that a federal exchanges—those established and operated by the federal government—"utilize a secure electronic interface sufficient to allow for a determination of an individual's eligibility"? States can certainly establish such procedures for the exchanges they create and control but states cannot tell federal entities what to do. And yet this is an obligation imposed on states as a condition of continuing to receive Medicaid funds. Perhaps this provision too is unenforceable post-NFIB, but there is no question it only makes sense if an "Exchange established by the State" is, in fact, an Exchange established by the State.

A few other tidbits, Professor Gluck repeats the error made by others that there are three definitions of "Exchange" in the PPACA. That is untrue. There is only one real definition, that which is provided in Section 1563. Indeed, one of the others referenced is nothing approaching a definition at all. It is, as it is expressly labeled, a requirement of 1311 exchanges. It says as much and operates as such. Indeed, were it otherwise (as my co-author explained) for-profit exchanges could be transmuted into not-for-profits exchanges, and that would be quite absurd (and defeat the clear purpose of that provision to boot).

Although she says we should focus on the text, Professor Gluck cites language from the NFIB v. Sebelius joint dissent endorsing the government's understanding of how exchanges would operate. Yet this was not an issue in the case beyond the question of severability, and on that question the dissenters' view did not prevail. It is not a holding of the court. Further, the joint dissent is hardly an authoritative account of the PPACA's text. Indeed, the dissent makes the indisputably false claim that that "The ACA requires each State to establish a health-insurance 'exchange.'" The law does not require any such thing. Indeed, under well-established principles, it couldn't.

Professor Gluck's purpose in citing the NFIB dissent, beyond making it awkward for its authors to take a different view of the exchange provisions today, is to suggest that it would be absurd to think Congress would have enacted a statute that could be, in large part, undone due to state intransigence. Yet Congress has done that before, and even did it in the ACA. Medicaid relies upon state cooperation, both as initially enacted and as modified in the ACA. The Supreme Court modified the terms of the deal in NFIB, but it did not alter the fundamental premise that, with Medicaid, Congress offers states a deal, and if states refuse, then Congress's purpose is frustrated.

Professor Gluck has elsewhere argued that the exchange provisions are different from Medicaid because Congress created a fallback: a federal exchange. Yet that's what Congress does all the time in "cooperative federalism" statutes, while simultaneously providing states with incentives to act. So, for instance, under the Clean Air Act, if a state refuses to implement the program, the federal government will step in. In addition, the recalcitrant state foregoes some benefits (funding for environmental programs and highways), and also suffers some sanctions (more onerous offset requirements).

As the CAA is structured, if only a few states refuse to cooperate, it is no big deal, as the EPA can take over for them and provide the desired fallback. Yet if thirty-six states were to refuse to implement the CAA's provisions, the core of the statute would come crashing down, as the federal EPA does not have the resources to take on the permitting and enforcement obligations currently shouldered by states. Nowhere close. So Congress has indisputably done what defenders of the IRS rule say it would never do: condition benefits on state cooperation in a program that will not function as Congress had hoped should a sufficient number of states refuse to hum the federal government's tune.

The structure Congress adopted in the exchange provisions does not undermine the plaintiff's arguments either. All the eligibility requirements for tax credits—that insurance be purchased on an exchange, that it be an exchange "established by the State," that the individual fall within the requisite income range, etc.—are contained in the same place: Section 1401. And these limitations are precisely where Congress has put similar limitations on the receipt of tax credits for other forms of health insurance before. So there is nothing odd, awkward or unprecedented with what Congress did here, as it is completely in line with what it's done before.

Professor Gluck reminds the reader that,

Justice Scalia's own statutory interpretation treatise argues (at pages 63 and 168) that "there can be no justification for needlessly rendering provisions in conflict if they can be interpreted harmoniously," and that statutory provisions should not be interpreted to render them ineffective or superfluous.

Precisely. However poorly the PPACA may be written, it is Professor Gluck's interpretation, and that interpretation urged by the government, that renders the repeated phrase "established by the State" superfluous and creates unnecessary conflict within the statute. And it is this interpretation that seeks to deprive clear statutory terms of no meaning. Thus it is this interpretation that does the most violence to the statutory text that Congress actually enacted into law.

At this point it should be clear that the case against the IRS rule is not based upon one phrase read in isolation or a single statutory provision ripped from context. The argument is based upon a consideration of the entire PPACA—the one that Congress wrote and enacted into law, not the one that others may have preferred. This argument considers how the law's various provisions interrelate and the context in which these provisions appear. In sum, it is an argument based upon well-established approaches to statutory interpretation, and it is an argument that most would readily accept were we not talking about the PPACA.

So, by all means, confine the inquiry in King to the text of the PPACA, for it is the text—and not any unexpressed intentions—that are the law. If that is what the Supreme Court does, it should conclude that the IRS tax credit rule is unlawful. By July of next year we will see if that is what the Court does.