As I wrote in an introductory post for Law and Liberty, the roots of King v. Burwell lay in the humble soil of federal tax law. King interpreted the terms of a provision of the Internal Revenue Code that was added by the Affordable Care Act of 2010 (ACA), popularly known as Obamacare.
What sent the case to the Supreme Court was that the law includes not one but two provisions relating to the establishment of healthcare exchanges: Section 1311 and Section 1321(c). The first concerns an exchange established by a state (that is, one of the states of the union or the District of Columbia), while the second concerns exchanges established by the U.S. Secretary of Health and Human Services. Functionally, the two provisions are mutually exclusive: The HHS Secretary may establish an exchange in a state only if the state does not elect to do so (or if the state elects to create its own but fails to meet the statutory deadline for doing so, or is deemed to be in jeopardy of failing to meet the deadline).
Those who followed the case closely may have lost their original sense of wonder that the existence of a second mode of exchange-creation would propel a case about an income tax credit formula all the way to the highest court in the land. The Supreme Court does not often hear cases about the proper construction of the Internal Revenue Code.
The premium-assistance tax credit formula, contained in Internal Revenue Code Section 36B(b)(2)(A), is available to a taxpayer whose health plan was purchased through “an Exchange established by a State under section 1311.” The formula for the amount of the credit in 36B(b)(2)(A) is very clear, abundantly clear, way clear—please take your pick. Indeed, according to the most natural reading of its operative phrase, the amount of tax credit available based on the purchase of coverage through an exchange established by HHS under Section 1321(c) is zero.
Yet the IRS promulgated regulations under which the tax credit would be the same under the statutory formula regardless of whether the coverage was obtained through a state exchange or a federally established exchange. And, in what Josh Blackman has called “a heartbreaker of a case,” the Court upheld those Treasury regulations in the face of a challenge under the Administrative Procedure Act.
Chief Justice Roberts’ opinion for the Court in King has been criticized on various grounds. What might be called the primary criticism is found in Justice Scalia’s dissent. There is no need to repeat its substance here, but it is based on a disagreement about the meaning of the statutory phrase, “an Exchange established by a State under section 1311.” I agree with the dissent’s ultimate conclusion, but my reasons are somewhat different (as I hope to explain in the third and last of this series).
Here let me take up something that flows from the Court’s opinion—I hesitate to call it secondary, because its ramifications are wide indeed. A number of us—Chris Walker at Notice & Comment, Jonathan Adler at the Volokh Conspiracy, and others—have touched how the King decision makes it impossible for a subsequent administration to revoke the current Treasury regulations and issue replacements that conform to the natural reading of the statutory language.
This is so because the majority in King went in an unexpected direction after being persuaded that 36B(b) is ambiguous. It ruled, in light of that alleged ambiguity, that “this is one of those cases” in which the mode of analysis adopted in Chevron U.S.A. v. Natural Resources Council, Inc. (1984) did not apply and that it was for the Court to interpret the ambiguous phrase. Thus, in effect, the Court took us there and back again—first ruling that a clear phrase was ambiguous but then ruling that it has a single meaning that neither the executive branch nor the lower federal courts can deviate from.
Here is the relevant passage from King:
When analyzing an agency’s interpretation of a statute, we often apply the two-step framework announced in Chevron, 467 U. S. 837. Under that framework, we ask whether the statute is ambiguous and, if so, whether the agency’s interpretation is reasonable. Id., at 842–843. This approach “is premised on the theory that a statute’s ambiguity constitutes an implicit delegation from Congress to the agency to fill in the statutory gaps.” FDA v.Brown and Williamson Tobacco Corp., 529 U. S. 120, 159 (2000). “In extraordinary cases, however, there may be reason to hesitate before concluding that Congress has intended such an implicit delegation.” Ibid.
This is one of those cases. The tax credits are among the Act’s key reforms, involving billions of dollars in spending each year and affecting the price of health insurance for millions of people. Whether those credits are available on Federal Exchanges is thus a question of deep “economic and political significance” that is central to this statutory scheme; had Congress wished to assign that question to an agency, it surely would have done so expressly. Utility Air Regulatory Group v. EPA, 573 U. S. ___, ___ (2014) (slip op., at 19) (quoting Brown and Williamson, 529 U. S., at 160). It is especially unlikely that Congress would have delegated this decision to the IRS, which has no expertise in crafting health insurance policy of this sort. See Gonzales v. Oregon, 546 U. S. 243, 266–267 (2006). This is not a case for the IRS.
It is instead our task to determine the correct reading of Section 36B . . .
Thus, the effect is not only to endorse the current Treasury regulations, but to immunize them from change. So what? might be the response. Why would a subsequent administration dream of changing them? Here is one reason: Revisiting, even reversing, the regulations upheld in King might have been a good idea because the regulations have an important spillover effect on employment. The availability of premium-assistance tax credits in a state has a huge potential impact on employers with facilities in that state.
This is owing to the fact that Section 4980H(a)—part of the ACA’s so-called employer mandate provision—imposes crushing penalties on employers who fail to offer health coverage enrollment to one or more of their full-time employees if (and only if) a full-time employee is eligible for either 1) advance payment of the tax credit or 2) a cost-sharing reduction available to individuals eligible for the tax credit. This provision has been blamed (justly or unjustly) for causing employers to dial back on full-time employment.
Section 36B, when read as the King petitioners read it, arguably allows a state to give in-state employers quasi-immunity from the employer mandate’s tax penalty by choosing not to establish a state exchange. Employees would not be able to obtain advance payments of the tax credits. That reading is consistent with a reading of Sections 1311 and 1321(c) according to which states have a choice between the advantages of premium-assistance tax credits, or the advantages of a climate more hospitable to abundant employment.
I do not suggest that lawmakers necessarily took into consideration such a differential in advantages when writing the ACA. However, a purely textual reading of Sections 1311 and 1321(c) leaves it open for the states, for their part, to judge and compare. And it is consistent with the longstanding ideological orientation of the Republican Party to give states options and to favor incentives of this type to encourage employment growth (or at least to complain that any taxation of business depresses employment).
But what of the political considerations in turning down premium-assistance tax credits by declining to establish an exchange? Wouldn’t that amount to electoral suicide? Apparently not. As we saw after Obamacare was passed, Governors in a majority of the states showed it is possible to risk turning down premium-assistance tax credits for their citizens without a voter rebellion. And, particularly if it is done early in his or her term of office, a new President with comfortable majorities in both chambers of Congress would not need to fear taking supposedly unpopular measures such as restoring the states’ ability to choose between premium-assistance tax credits or more robust employment.
Think about the air traffic controllers’ union in the first month of the Reagan administration (or, if you are British, the mining unions early in Prime Minster Thatcher’s tenure in office). It is at least possible that, under certain circumstances, a subsequent administration would have contemplated reversing the thrust of the Treasury regulations at issue in King.
There is a lot at stake here—I dare say the stakes transcend even something as big as the federalization of health care in the United States. To be specific, there is the question whether King signals the start of a new chapter in administrative law or whether it is a one-off (not to say ad hoc) decision. To paraphrase what has been said about the portion of the Court’s opinion quoted above, by having carved out an exception to Chevron, the King opinion certainly appears to have let a genie out of a bottle.
I say carved out an exception because there in fact is no existing set of “extraordinary cases” of which this case “is one.” Until King, as far as I am aware, there are no intervening cases between Chevron and today in which the Court explicitly rejected applying Chevron when reviewing a regulation adopted after notice-and-comment rulemaking. (I won’t go so far as to say that the Court doesn’t use step 1 of Chevron “creatively” to shield some matters from agency discretion. But from 1984 until fewer than two months ago, the Court has routinely “deferred” to Chevron as it were, or at least paid it lip service, in cases involving formal interpretive rulemaking.)
Despite the majority’s selective quotations from Brown and Williamson, what I have said applies to that decision as well. To be sure, Justice O’Connor’s opinion includes lengthy, highly theoretical dicta regarding various considerations that could be at play in judicial review of agency rulemaking after Chevron (and arguably before it, too). In contrast to that discussion, though, the actual holding in Brown and Williamson applies the decisive Chevron factor at step 1 of the analysis with admirable brevity.
It is worth mentioning another oddity about the King passage, quoted above, rejecting the Chevron doctrine of judicial deference to an executive agency’s interpretation of statute. That rejection is based on the conclusion that Congress’ adoption of a phrase that the Court found ambiguous was nevertheless not an implied delegation of rulemaking power. In effect, it held that a trigger for the application of Chevron was absent.
But an implicit delegation of rulemaking authority extending to the phrase “an Exchange established by a State under section 1311” arguably would be superfluous here. Arguably, it should not have been necessary for a textualist to consider whether Congress had delegated regulatory power to the Treasury Department by implication. Section 36B(g) provides that the HHS Secretary “shall prescribe such regulations as may be necessary to carry out the provisions of this section, including regulations which provide for . . . the coordination of the credit allowed under this section with the program for advance payment of the credit under section 1412 of the Patient Protection and Affordable Care Act.”
Thus, Section 36B included an explicit legislative delegation of rulemaking authority to the Treasury Secretary. And in fact this recapitulates, in real life, the second of the hypothetical statutes in Justice Scalia’s opinion for the Court in City of Arlington v. FCC (2014). His second hypothetical was intended to demonstrate that an explicit legislative grant of the power to interpret an ambiguous substantive standard is substantively identical to the legislature’s adoption of an ambiguous substantive standard when it comes to the application of Chevron.
The City of Arlington hypothetical, not unlike Internal Revenue Code Section 36B, included a substantive requirement in one subsection and in a subsequent subsection granted authority to issue regulations implementing the substantive standard. The Chief Justice, writing in dissent in City of Arlington, allowed that, yes, when it came to that second hypothetical, “the answer is easy. The majority’s hypothetical Congress has spoken clearly and specifically in Section 2 of the Act about its delegation of authority to interpret Section 1.”
The Chief Justice’s opinion for the Court in King arguably retreats from that “clear-and-specific” position. And quite apart from the echo of City of Arlington, given the particular language of Section 36B(g), it is difficult to imagine how the Treasury would issue regulations providing for the coordination of tax credits and advance payments without at least indirectly ruling on whether the credits were available within the territory of a given exchange. If Treasury believed the phrase “an Exchange established by a State under Section 1311” was ambiguous, regulations that coordinated premium tax credits would have been possible if and only if Treasury interpreted the phrase either the way the six justices did or according to its tenor.
Be all of that as it may, however the genie got habeas relief, he or she is a free agent now. And it is by no means clear that this genie is well-intentioned or even predictable. Judgment on those questions must be deferred until we see whose wishes the genie grants (and how the wishes are interpreted).
To drop the Thousand and One Nights metaphor, we might ask, how will the six justices of the King majority coalesce in the future in applying King’s exception to Chevron? For example, did Justices Ginsburg, Kagan, and/or Sotomayor refrain from concurring in the judgment on grounds other than Roberts’ because they felt a six-justice majority would have a beneficial effect on further efforts to construe other portions of the ACA (which the Court itself regards as less than perfectly written)?
Let me turn to a related issue. I do not think the basis for refusing to apply Chevron really can be the ostensible one—that the Internal Revenue Service has no expertise regarding health insurance policy. Josh Blackman in National Review trusts that the ostensible reason is the real reason, but I have my doubts. After all, the Internal Revenue Code is replete with provisions that act as incentives or disincentives on a variety of unusual and/or specialized and technical activities outside the ken of tax experts.
Examples: Section 45K allows a tax credit for the production of alternative energy, including from “geopressurized brine.” Section 129 excludes from income certain qualifying dependent care provided by an employer. Section 168(c) allows a five-year depreciation schedule for “qualified technological equipment,” including “high technology medical equipment.” The IRS regulates with respect to those matters even though it is not expected to know much, if anything, about fracking, reading Berenstain Bears books to toddlers, or what constitutes “high technology” medical equipment (thermometers with digital read-outs?).
Of course, the question whether Chevron should have been applied in King is inseparable from the question whether the phrase “established by a State under section 1311” is ambiguous. Consider, once again, the round trip on which Chief Justice Roberts’ opinion takes us. It purports to demonstrate, first, that the phrase does not necessarily refer exclusively to exchanges set up by states of the Union and/or the District of Columbia. It then says the phrase cannot possibly be interpreted by the executive branch or by the judiciary other than to refer to any exchange created under the auspices of the ACA, whether state-created or federally created, whether established under section 1311 or under section 1321(c).
Thus the “ambiguous” phrase ends up with a decidedly determinate and fixed meaning. To accomplish this feat, the opinion relies on the proposition that Congress considered establishment under Section 1321(c) to be the equivalent of establishment under 1311, at least for purposes of the premium-assistance tax credit allowed under Internal Revenue Code Section 36B.
Here is the crucial language on that issue:
By using the phrase “such Exchange,” Section [1321(c)] instructs the Secretary to establish and operate the same Exchange that the State was directed to establish under Section . . . In other words, State Exchanges and Federal Exchanges are equivalent—they must meet the same requirements, perform the same functions, and serve the same purposes. Although State and Federal Exchanges are established by different sovereigns, Sections  and  do not suggest that they differ in any meaningful way. A Federal Exchange therefore counts as “an Exchange” under Section 36B.
It seems to me this equivalence proposition, rather than any lack of expertise on the part of the IRS, is the real basis of the King holding. The lack-of-expertise point is likely a makeweight; to really get where the opinion goes, it is necessary to accept the Chief Justice’s premise that the Affordable Care Act must be interpreted according to its overarching purpose, as revealed by its architecture.
We are left to wonder which other provisions of Obamacare, and in fact which other statutes, might have characteristics that militate against the application of Chevron.
One last observation. Although I cannot agree with the logic of the equivalence holding, I nonetheless sympathize with the dilemma facing the King majority. King is first and foremost, as I said at the outset, a tax case. It’s specifically about an income tax credit and income tax credits, by their very nature, must be determinate amounts.
Thus, if one accepts that the phrase “an Exchange established by a State under section 1311” does not necessarily mean only an exchange established by a state that has chosen to proceed under Section 1311, then there is a very narrow but legitimate basis for agreeing with the majority decision that the phrase cannot be subject to executive branch interpretation—that, in other words, the tax credit cannot be toggled on and off with each change of executive branch policy decisions. Income tax credits cannot be like Schrodinger’s cat. Once the phrase “an Exchange established by a State under section 1311” does not necessarily mean only that, it is necessary, for the sound administration of the revenue laws, that it mean any exchange, state or federal.
This is not the Court’s rationale for rejecting Chevron, of course. However, it fits the case as a phenomenon. Any member of the King majority, except perhaps the Chief Justice himself given his authorship of the King opinion, can distinguish King in the future by pointing to the narrower, tax-certainty-based rationale. The genie’s freedom might be temporary.
 King v. Burwell, Slip Opinion at 4, 20. Advance payment of premium-assistance tax credits are one of two mechanisms for financing individual coverage for low- to middle-income taxpayers. In addition to premium-assistance tax credits, the ACA authorizes payments to insurance issuers, referred to as “cost-sharing reduction payments.” These apply to silver-level plan coverage obtained by qualified taxpayers with income between 100 percent and 250 percent of the federal poverty line, provided that the qualified taxpayer is “eligible” for a premium tax credit. See ACA Section 1402 and subsection (f)(2). See also http://obamacarefacts.com/insurance-exchange/cost-sharing-reduction-subsidies-csr/. Thus, if an individual is not eligible for a premium tax credit, he or she is also ineligible for a cost-sharing reductions.
However, a textualist could conclude that eligibility for a premium tax credit equal to zero is not precisely the same as ineligibility for a premium tax credit. Cf. the requirement in Code Section 36B(f)(3)(C), Exchanges to report the aggregate amount of “any advance payment” of premium tax credits “or reductions under [ACA] section 1412.” (emphases added), and Code Section 4980H(a)(2), which provides for an employer mandate tax triggered by employee’s receipt of premium tax credit “or” cost-sharing reduction (emphasis added). Thus, a reasonable argument could be made that if an individual is entitled to a premium tax credit, even if it is equal to zero, cost-sharing reductions would apply to his or her silver-level plan.
 See Health Insurance Premium Tax Credit, 77 Fed. Reg. 30,377 (May 23, 2012); see, also, Prop. Treas. Reg. §§ 1.36B-1 et seq., Fed. Reg. Vol. 76, No. 159, p. 50931 (August 8, 2011).
 To keep the length of this post within bounds, I have assumed away other impediments to revoking and replacing the regulations. But see Motor Vehicle Manufacturers Ass’n v. State Farm Mutual Ins. Co., 463 U. S. 29 (1983).
 See ibid, 529 U.S. at 133.“[W]e find that Congress has directly spoken to the issue here, and precluded the FDA’s jurisdiction to regulate tobacco products.” Moreover, the analysis in Brown and Williamson ultimately is textual not consequential: Tobacco is not a drug within the meaning of the FDA’s organic statute, Brown and Williamson holds, because, under the literal words of the statute, everything within the category “drug” must be capable of being rendered “safe and effective”—an attribute everyone agreed tobacco lacked. Thus, for all its meandering, Brown and Williamson is based on the statutory text and nothing else.
 Justice Ginsburg recently said of the majority opinion in Obergefell v. Hodges (2015) that she would have emphasized the equal protection argument had she written the opinion, but she joined Justice Kennedy’s opinion for the Court rather than writing separately for the sake of having a single majority opinion for the Court. See also, for example, Michigan v. Bay Mills Indian Community, No. 12-515 (May 27, 2015), where Justice Kagan wrote for the Court,
Congress can abrogate [tribal] immunity [of Indian nations] as and to the extent it wishes. If Congress had authorized this suit, Bay Mills would have no valid grounds to object. But Congress has not done so: The abrogation of immunity in Indian Gaming R Act] applies to gaming on, but not off, Indian lands. We will not rewrite Congress’s handiwork.
Slip Opinion at 21 (emphasis added). Justices Kennedy, Breyer, and Sotomayor, along with the Chief Justice, joined in Justice Kagan’s opinion for the Court. See also, for example, Baker and Botts L.L.P. v. Asarco LLC, No. 14-103 (June 15, 2015). Justice Sotomayor, concurring in part and concurring in the judgment, noted the impropriety of judicial interpretation of a portion of the Bankruptcy Code based on policy reasons. It is interesting to note that Justice Kennedy joined in the Court’s opinion in Baker and Botts, which states in part that “we would lack the authority to rewrite the statute even if we believed” that policy reasons called for a non-textual interpretation, and added that “Our job is to follow the text, even if doing so will supposedly undermine a basic objective of the statute . . .” Slip Opinion at 12-13 (internal quotation marks omitted). I leave it to others to decide whether these positions can be reconciled with the conclusions reached in King v. Burwell.