King, Chevron, and the Age of Textualism
Abigail R. Moncrieff
Essay
95 B.U. L. Rev. Annex 9 (2015)
In the King v. Burwell oral arguments, Chief Justice John Roberts—usually one of the more active members of the Court—asked only one substantive question, addressed to the Solicitor General: “If you’re right about Chevron [deference applying to this case], that would indicate that a subsequent administration could change [your] interpretation?” As it turns out, that question was crucial to Roberts’s thinking and to the 6-3 opinion he authored, but almost all commentators either undervalued or misunderstood the question’s import (myself included). The result of Roberts’s actual thinking was an unfortunate outcome for Chevron—and potentially for the rule of law—despite the happy outcome for the Obama Administration.
King Background
The central question in King was whether Obamacare allows the Internal Revenue Service (IRS) to distribute health insurance subsidies nationwide or whether, instead, the statute limits subsidies to states that establish their own health insurance exchanges. In a 2012 rulemaking, the IRS interpreted the statute to permit subsidies in all states, regardless of the states’ exchange establishment choices. The plaintiffs argued that the IRS Rule was contrary to statute, pointing to language in an arcane provision, 26 U.S.C. § 36B, to argue that Obamacare intended to limit subsidies to establishing states. The relevant provision sets the formula for calculating subsidy amounts, and it calibrates subsidies to the cost of insurance purchased on “an Exchange established by the State,” without making reference to the fallback Exchange that the federal government set up for non-establishing states. Based on that language, the plaintiffs argued that Congress intended to use subsidies as an incentive for states to take on the thankless task of creating and managing the exchanges.
In defending the IRS Rule against the plaintiffs’ interpretation, the government’s primary rebuttal was that the overall statutory structure would fall apart if subsidies were disallowed. Obamacare rests on three interlocking pieces: market reforms, mandates, and subsidies. But the mandates are unenforceable without the subsidies, and the market reforms don’t work without the mandates. The statute was designed such that all three pieces needed to be in place for the regulatory regime to work. Removing any one of them would cause the whole thing to collapse. Furthermore, collapse of the regulatory regime through withdrawal of subsidies would not leave states in the same position they were in before Obamacare passed. Given the statute’s structure, withdrawing the subsidies would nullify the mandates while leaving the market reforms in full force and effect. But in the absence of subsidies and mandates, the market reforms would cause untold harm in individual insurance markets—the “death spirals” that everyone, including the Court, referenced throughout the arguments and in the opinions.
The Solicitor General made this case before the Court, arguing that Congress could not possibly have intended its statute to be purposefully destructive in states that refused to establish exchanges. The government also made a careful (but not particularly persuasive, to my mind) textual case that the federal exchange ought to qualify as an “Exchange established by the State” because “established by the State” was a term of art for “established in the state.” But the government never explained why Congress would use “by” as a term of art for “in”—or how the inclusion of that “term of art” added anything meaningful to the statute that would have been lacking if the provision had simply said “an Exchange established under this Act.”
Chevron Background
In an ordinary case involving an administrative agency’s interpretation of a statute, the court would assess the agency’s construction under the Chevron framework. Chevron instructs courts to defer to administrative agencies’ reasonable interpretations of ambiguous statutory terms, at least as long as the agency’s interpretation appears in a rule or order that carries “the force of law.”1 Here, the IRS had interpreted the relevant portions of Obamacare to allow subsidies nationwide, and it had, on the basis of that interpretation, promulgated a rule—indisputably with the force of law—that treated all states the same for subsidy purposes. In a statement accompanying the IRS’s final rule, the agency argued that its interpretation was “consistent with the language, purpose, and structure of section 36B and the Affordable Care Act as a whole.”2
According to straightforward understandings of Chevron, the IRS’s interpretation obviously should have triggered the Chevron framework. Chevron would then have instructed the Court to ask only two questions to determine the permissibility of the IRS Rule: (1) Is the statute silent or ambiguous on the precise question at issue, even after applying all of the traditional tools of statutory construction to discern statutory meaning? (2) If yes—if no single meaning is apparent from the statute as ordinarily construed—is the agency’s interpretation reasonable? If the Court had concluded that the statute was ambiguous and that the agency’s interpretation was reasonable, then it could have upheld the IRS Rule without holding that the agency’s interpretation was the “right” interpretation of the statute.
After such a holding, however, future administrations would have been able to change course. When a court upholds an agency’s construction under Chevron, the holding allows future administrations to reverse or alter the construction at will, as long as the new construction is also reasonable. Indeed, in Chevron itself, the interpretation the Court was reviewing was a reversal of the Environmental Protection Agency’s longstanding construction of a Clean Air Act provision, and the Court deferred to it anyway. In a later case, Brand X,3 the Supreme Court made explicit its understanding that agencies may, over time, alternate among reasonable interpretations of ambiguous statutory terms.
The Chief Question
Roberts’s question at the King oral arguments was a request for confirmation that the usual pattern would apply here. If the Court upheld the IRS’s interpretation under Chevron, he asked, would a future president and a future IRS be legally allowed to take subsidies away from Americans living in non-establishing states? The Solicitor General answered that a reversal in the interpretation would likely be “unreasonable” under Chevron Step Two, but that answer ought to have been little comfort to those who prefer the Obama Administration’s approach. Agencies basically never lose at Chevron Step Two.
In the aftermath of the oral arguments, the vast majority of commentators, including most prominently Jeffrey Toobin,4 understood the Chief’s question as an indication of his interest in using Chevron as a kind of “passive virtue”5—a way for the unelected Court to avoid deciding a politically contentious question with finality. A Chevron holding would assign to the executive branch the power and duty to choose among plausible interpretations of § 36B.
At least at a superficial level, King seemed like the perfect candidate for that kind of holding. On a simplistic understanding of the case and the statute, the issue in King presented a conflict between two policy “goods”: universal insurance and federalism. The government’s argument was that Obamacare intended to give everyone in the country access to affordable insurance, which required nationwide subsidies, but the plaintiffs’ argument was that Obamacare intended to avoid a complete federal takeover of health insurance regulation, which required state-based exchanges. And according to the plaintiffs, the contingency of subsidies on exchange establishment was a necessary piece of the overall statutory commitment to preserving state power in healthcare regulation—a crucial incentive for the states to agree to establish and manage their own exchanges. The debate thus centered on a policy tension within the statute, and there was no doubt that both sides were, in a superficial sense, right. Obamacare really did try to achieve universal affordable health insurance, and it really did try to avoid a federal takeover of health insurance regulation.
When a question of statutory interpretation involves a difficult accommodation of conflicting policy goals—like the tension between economic growth and environmental protection at issue in Chevron and the superficial tension between universal insurance and federalism at issue in King—the argument for Chevron deference is particularly strong. A Chevron holding allows the Court to avoid declaring an outcome, leaving the political balancing act to a political branch: the administrative agencies. In King, a Chevron holding would have allowed the Obama Administration to prioritize universal insurance over federalism while allowing a future administration to instead prioritize federalism over universal insurance.
After the oral arguments, in all of the commentary I read (which was certainly not all, but probably was most, of the commentary that existed), everyone interpreted Roberts’s question along these lines. Roberts is generally a fan of leaving difficult policy choices to the political branches—as he made clear in his same sex marriage dissent the day after announcing King—and his question about Chevron seemed to indicate an interest in following the same strategy for Obamacare. According to the commentary, Roberts appeared inclined to vote for the government but to do so in a way that would allow future administrations to shift gears.
The commentary could not have been more wrong.
Roberts v. Roberts
In his opinion for the Court in King, Roberts argued that the relevant provision of Obamacare is ambiguous, but he refused to defer to the IRS’s interpretation under Chevron. Instead, he invoked the “major questions exception”—an infinitely flexible doctrinal escape-hatch for Chevron cases—holding that the Chevron framework is inapplicable when the question at issue is one of “deep ‘economic and political significance.’”6 Rather than letting the agency decide such a major question, Roberts held that it is “our task”—the Court’s—“to determine the correct reading of Section 36B.”7
What a strange contrast to the Chief Justice Roberts who argued, so emphatically and only a day later, that it is not the Court’s task to determine the scope of the constitutional interest in marital privacy!8 Both King and Obergefell presented delicate and contentious political issues, and in both cases, the Court had available to it valid doctrinal bases for deferring to the political branches. In King, the Court could have used Chevron to defer to the IRS, and in Obergefell, the Court could have used rational basis review or intermediate scrutiny to defer to the states, distinguishing Loving v. Virginia9 on the ground that the earlier case involved the suspect classification of race while the latter case involved, at most, an intermediate classification of gender. (Sexual orientation is still, even after Justice Kennedy’s majority opinion in Obergefell, not a suspect classification). But Roberts summarily cast aside the deferential approach in King even while arguing forcefully for deference in Obergefell.10
What explains the difference? Consider one simple explanation: It’s easier to amend a statute than it is to amend the Constitution. If the Court interprets Obamacare in a way that is politically unpopular, Congress can fix it by passing a statutory amendment through the ordinary lawmaking process. If the Court interprets the constitutional right to marital privacy in a way that is politically unpopular, the political branches can fix it only by passing a constitutional amendment, which requires a supermajority vote in Congress or the state legislatures. The need for deference is therefore stronger in a constitutional case than in a statutory case.
But if that’s the reason for the distinction, then why have Chevron at all? That logic would justify wholesale abandonment of Chevron; it does not justify retail exceptions. And Roberts is not inclined to attack Chevron as a general matter. There was something peculiar about King that made him sideline deference in that case only.
Chevron v. Textualism
Now consider a more nuanced and complicated explanation: In both cases, Roberts wanted courts to enforce the political branches’ policy choices. In Obergefell, enforcement of the states’ political choices required rejection of the constitutional claim and application of deferential “rational basis” review. But in King, the political choice that Roberts wanted to enforce was not the IRS’s or any agency’s; it was the 2010 Congress’s. Despite the sloppily drafted statutory text, Roberts thought it clear that the 2010 Congress—the Congress that enacted Obamacare—wanted subsidies to be available nationwide. He thought it clear that the 2010 Congress, while concerned about federalism, never intended to use subsidies as an incentive for states to establish their own exchanges. A threat of withholding subsidies—which was tantamount to a threat of purposefully destroying insurance markets in non-establishing states—was too extreme, and there was no evidence beyond the sloppy text that Congress actually intended to convey such a threat.
Enforcement of Congress’s political choice—enforcement of the enacting Congress’s intended balance between universal insurance and federalism—required the Court to bind not only the Obama Administration but also all future administrations to the uniform availability of subsidies. Under the principle of legislative supremacy, clear congressional choices always trump agencies’ choices, even when Congress has delegated some interpretive power to the executive. In order to enforce legislative supremacy and to defer to Congress’s choice, Roberts needed to hold that the statute not only permits but requires the IRS to make subsidies available nationwide. Given the entirety of his opinion, that seems quite clearly to have been Roberts’s conclusion.
But why sideline Chevron to get there? Under Chevron Step One, the Court must use all “traditional tools of statutory construction” to determine whether “Congress had an intention on the precise question at issue,” and if the Court finds a clear congressional intent, then “that intention is the law and must be given effect.”11 If Roberts thought it obvious that Congress intended to make subsidies available nationwide, why did he not simply hold under Chevron Step One that Congress’s intent was clear and that the IRS’s construction was the only correct interpretation of the statute? Why did he argue, instead, that Section 36B is ambiguous but that this and future agencies deserve no deference under Chevron?
The problem is that Roberts lives in the age of textualism. Section 36B contained language—“established by the State”—that the government’s interpretation rendered superfluous. The government gave no explanation for the presence of that language other than a “term of art” argument that I found unconvincing, despite agreeing with the government’s conclusion, and Roberts apparently found it unconvincing, too. He made no reference to the “term of art” view and admitted that his construction rendered the words “by the State” superfluous. Upon that admission, however, Roberts needed to justify his decision to read three words out of the statute. His justification was that enforcement of the text would undermine the purpose—though Roberts, in his age of textualism, carefully used the term “legislative plan” rather than “legislative purpose” in making that argument.
So where’s the Chevron problem? Surely legislative purpose is a “traditional tool of statutory construction” that should be permissible at Chevron Step One. Apparently not, according to Roberts. In the age of textualism, Roberts argued, broad notions of legislative purpose—particularly if the purpose is most clearly evident from sources extrinsic to the text of the relevant provision—are admissible evidence of congressional intent only if the statutory text is ambiguous. But in the simultaneous age of Chevron, ambiguous statutory text constitutes a delegation of interpretive power to the agency, not to the courts. For a smart and pragmatic textualist in the age of Chevron, King highlights an annoying “gotcha.”
In King, the answer to the statutory question is perfectly obvious if you’re willing to take off the textualist’s blinders to legislative purpose. But if you’re a modern textualist judge, the only permissible way to take off your blinders is to argue first that the statutory text is ambiguous. But if you argue that the statutory text is ambiguous, then you have to defer to the agency. But if you defer to the agency, then future administrations will be allowed to implement an interpretation of the statute that you know, despite your blinders, is actually inconsistent with congressional intent—as Roberts went out of his way to confirm at oral arguments. But if you allow future administrations to violate an extrinsically obvious but textually ambiguous legislative intent, you will have violated your obligation to enforce legislative supremacy. But, but, but… How to escape?
For Roberts, the answer was the major questions exception to Chevron deference. As noted above, the major questions exception is infinitely flexible. The Court has never articulated a standard for distinguishing “major” from “minor” questions other than its oblique reference to “economic and political significance.” And pretty much anything Congress legislates could satisfy a standard of “economic and political significance.” If a question were not economically and politically significant, Congress would not have spent its time passing a statute about it. The standard is—or could easily become—tautological.
The major questions exception reminds me of the opening scene in Disney’s Aladdin, in which a street merchant is trying to sell a device of unknown usefulness: “Combination hookah and coffee maker!” he announces. “Also makes julienned fries!” Then he bangs the device on the table, proclaiming, “It will not break! It will not [sproing.] It broke.”12
With enough applications, the major questions exception could subsume Chevron. Roberts needed an escape hatch from his Chevron-textualism bind, and he used the most readily apparent and flexible one available—the one that also makes julienned fries! In so doing, however, he banged the major questions exception on the table, adding to the list of “major questions” invocations in a way that could ultimately undermine his own and Chevron’s commitment to judicial deference.
An Escape from the Escape
I have a modest suggestion. When the correct answer to a statutory question is obvious to everyone who’s willing to be honest about congressional intent—but the obvious answer cannot be proved by applying the “acceptable” tools of statutory construction (however the “acceptable” toolbox is delimited in a given era)—judges should simply assert the obvious answer regardless. And they should do so openly and notoriously at Chevron Step One, without pulling Houdini-like escapes from the legalistic straight-jackets of their interpretive philosophies.
Textualism has made tremendous headway in the last several decades, and not without good cause. When the answer to a statutory question is genuinely not obvious, judges promote predictability and transparency—also known as “the rule of law”—by limiting their inquiry, as much as possible, to the text of the statute. Chevron has similarly served as a powerful and sensible restraint on courts. Judges serve rule of law and separation of powers values by deferring to reasonable administrative interpretations of ambiguous statutory terms.
But judges do not serve any legitimate values by ignoring obviously correct resolutions of statutory cases simply because the text of the statute is inartfully drafted, and judges undermine important rule of law values by refusing deference to agencies for no reason other than their desire to reach beyond the “acceptable” toolbox of statutory construction and their sense that they cannot do so under Chevron Step One.
In King, there was an ever-growing mountain of evidence—all extrinsic to the statutory text and structure—that the 2010 Congress intended to make subsidies available nationwide,13 that the plaintiffs in the case were relatively uninterested in the case’s outcome,14 and that the lawyers and think tanks funding the lawsuit were primarily interested in destroying rather than enforcing the statutory scheme.15 There was no extrinsic evidence on the other side. The only evidence for the plaintiffs’ view was the text. The plaintiffs’ lawyers invented a plausible fairy tale to explain their interpretation—an extended daydream of some nonexistent but sensible Congress that intended to use subsidies as an incentive for states to establish exchanges—but the story was transparently fictional. Everyone knew that the real 2010 Congress had no such intent.
The first expositors of the plaintiffs’ statutory argument, Jon Adler and Michael Cannon, pointed to one statement of Senator Ben Nelson’s that might have supported their fairy tale version of congressional intent,16 but in all of my conversations with them in academic debates and on Twitter, Adler and Cannon fiercely denied that they were pointing to Nelson’s statement as evidence of actual intent. To their mind, textualism demanded an intent-free, robotic interpretation of the statutory text, and actual congressional intent mattered not at all. But that’s crazy. “Words no longer have meaning”17 if not understood as representations of their speaker’s intent.
Textualism’s rejection of intent is—and should always be—a rejection of flimsy extrinsic evidence, cherry-picked empirical arguments, and conclusory judicial assertions of actual intent. It is not—and should not ever become—an instruction to judges to adopt any plausible story of congressional intent that fits the text, even if all of the extrinsic evidence in the world indicates that the asserted story of congressional intent is empirically wrong. King became a close case—and Chevron became a weaker doctrine—because the textualist rejection of flimsy extrinsic evidence of intent expanded dangerously into the territory of abandoning intent itself.
Roberts’s opinion is a resounding victory for Obamacare; it is a modest victory for smart structural interpretation in the age of textualism; it is a timid victory for pragmatic and economic realism. But it is not an unambiguous triumph of rationality over legalism.
Roberts closed his opinion with this observation: “Congress passed the Affordable Care Act to improve health insurance markets, not to destroy them.”18 As everyone including Adler and Cannon and Scalia, Thomas, and Alito recognized, the plaintiffs’ interpretation would have threatened to destroy health insurance markets in 34 states while the IRS’s interpretation sought to improve them. That should have been good enough for Chevron Step One.
Obvious congressional intent should be the law whether obvious from the text or not. Otherwise, the law becomes “pure applesauce.”19 (Whatever that means).
[1] United States v. Mead Corp, 533 U.S. 218 (2001).
[2] Health Insurance Premium Tax Credit, 77 Fed. Reg. 30377, 30378 (proposed May 23, 2012) (to be codified at 26 C.F.R. pts. 1 and 602).
[3] Nat’l Cable & Telecommunications Ass’n v. Brand X Internet Servs., 545 U.S. 967 (2005).
[4] Jeffrey Toobin, Did John Roberts Tip His Hand?, The New Yorker (Mar. 4, 2014), http://www.newyorker.com/news/daily-comment/did-john-roberts-tip-his-hand.
[5] Alexander M. Bickel, The Least Dangerous Branch: The Supreme Court at the Bar of Politics (2d ed. 1986).
[6] King v. Burwell, No. 14-114, slip op. at 8 (U.S. June 25, 2015) (quoting FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 159 (2000)).
[7] Id.
[8] Obergefell v. Hodges, No. 14-556, slip op. at 1-29 (U.S. June 26, 2015) (Roberts, C.J., dissenting).
[9] 388 U.S. 1 (1967).
[10] Obergefell, slip op. at 2 (Roberts, C.J., dissenting) (“[A] State’s decision to maintain the meaning of marriage that has persisted in every culture throughout human history can hardly be called irrational.”).
[11] Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843 n.9 (1984).
[12] Aladdin (Walt Disney Pictures), available at https://www.youtube.com/watch?v=Cd7aik82JyA.
[13] See, e.g., Robert Barnes, Supreme Court Case on Key Obamacare Provision Takes up this Senator’s Account, Wash. Post (Jan. 28, 2015), http://www.washingtonpost.com/politics/courts_law/supreme-court-case-on-key-obamacare-provision-takes-up-this-senators-account/2015/01/28/339ca646-a6fc-11e4-a2b2-776095f393b2_story.html; Robert Pear, 4 Words Imperil Health Law; All a Mistake, Its Writers Say, N.Y. Times, May 26, 2015, at A1.
[14] Louise Radnofsky, Jess Bravin, & Brent Kendall, Health-Law Challenger’s Standing in Supreme Court Case Is Questioned, Wall St. J. (Feb. 6, 2015, 6:52 PM), http://www.wsj.com/articles/health-law-challengers-standing-in-supreme-court-case-is-questioned-1423264458.
[15] Jeffrey Toobin, Hard Cases, The New Yorker (Mar. 9, 2015), http://www.newyorker.com/magazine/2015/03/09/hard-cases-jeffrey-toobin.
[16] Jonathan H. Adler & Michael F. Cannon, Taxation Without Representation: The Illegal IRS Rule to Expand Tax Credits Under the PPACA, 23 Health Matrix 119, 149 n.109 (2013).
[17] King v. Burwell, No. 14-114, slip op. at 2 (U.S. June 25, 2015) (Scalia, J., dissenting).
[18] King, slip op. at 21 (Roberts, C.J., dissenting).
[19] King, slip op. at 10 (Scalia, J., dissenting).
Click here for the PDF.