Statutory Ambiguity in King v. Burwell: Time for a Categorical Chevron Rule
The fate of the Affordable Care Act1 (ACA) may turn on the precise meaning of five words tucked into 26 U.S.C. § 36B(b)(2)(A), a once-obscure provision of the law under which tax subsidies are available for “qualified health plans . . . which were enrolled in through an exchange established by the State.”2 Does this language permit the issuance of subsidies for taxpayers enrolled through exchanges created by the federal government? On one hand, it seems that Congress clearly intended to allow this situation.3 On the other hand, the text says “State”—how much clearer could it be?
During oral arguments in King v. Burwell,4 the case that will decide what those five words mean, Justice Scalia suggested that even the irrefutable existence of a congressional intent to make insurance subsidies available on federally created health care exchanges might not matter if the relevant provision of the Affordable Care Act was poorly written.5 In other words, even unambiguous legislative history may fail to save a statutory construction if the text cannot plausibly be read to support it.
Courts typically resolve questions of this sort—questions regarding the permissibility of an agency’s interpretation of a statute it administers— by employing the two-step framework provided in Chevron U.S.A., Inc. v. National Resources Defense Council, Inc.6 Chevron held that when reviewing an agency’s construction of a statute it administers, a court asks two questions. At Step One, it asks whether Congress’s “unambiguously expressed intent” is clear; if so, the matter is closed.7 If Congress’s intent is unclear, however, a court proceeds to Step Two, asking whether “the agency’s answer is based on a permissible construction of the statute.”8 In Chevron terminology, Justice Scalia’s comments during the King arguments can be reframed as suggesting that when a statute’s meaning seems textually evident, Congress’s unmistakable objective will not suffice to create ambiguity, the presence of which triggers review that is highly deferential to the agency. Cases in which a judge will perceive a conflict between clear statutory text and clear legislative history may be uncommon, but as Scalia’s comments suggest, they can and will arise. What role should unambiguous legislative history play in Chevron analysis when a statute’s text is clear? Can it create or dispel ambiguity, and/or weigh in favor or against the reasonability of an agency’s construction, and if so, when?
Though the Supreme Court has confronted these questions before, the case law has not provided a consistent rule.9 This is unfortunate; enunciating a bright-line rule for the role of legislative history where the text is clear would enhance the clarity of the Chevron test. King, which involves a provision of the ACA whose history and text are both arguably clear,10 yet point in opposite directions, highlights the need for a bright-line, categorical rule on how courts should utilize legislative history in Chevron analysis when a statute’s clear text and equally clear legislative history conflict with one another. This Essay proposes three alternative formulations for a categorical rule to govern such conflicts. One would hold statutory text categorically irrelevant when legislative intent is clear, one would hold intent categorically irrelevant when the text is clear, and the third would hold that an agency construction based either on clear text or clear intent is automatically reasonable. Any of these three bright-line rules would be acceptable, but the third would be best. This Essay concludes by recommending that the Court use King to articulate such a rule.11
One categorical approach available to the Court is to hold that when legislative history is clear, a reviewing court must read the law in light of it. Admittedly, legislative history is a highly contested method of discovering statutory meaning. Justices have disagreed over whether legislative history should be considered at Step One at all, and if so, whether it can outweigh or trump statutory text. For example, in Zuni Public School District No. 89 v. Department of Education (2007), Justice Breyer’s majority opinion examined a statute’s legislative history at Chevron Step One before turning to its text,12 and Justice Stevens’ concurrence asserted that legislative history can be determinative at Step One,13 while Justice Scalia’s dissent asserted that when a statute’s text is clear, legislative history had no place in Step One analysis whatsoever.14 However, a “legislative history trumps” approach is consistent with Chevron’s canonical reasoning that when Congress passes indeterminate statutes, it intends to delegate discretion to agencies.15 Because Chevron is premised on legislative intent, courts should be more willing to examine evidence of legislative history in Chevron analysis. If the reason courts defer to agencies in the first place is that they assume that Congress intended to delegate discretion to agencies, courts should be less willing to defer when they know that Congress did not intend to give agencies discretion. Because Chevron analysis itself turns on congressional intent, under this approach agencies would merit no deference when that intent is clear in the legislative history.16
Applied to King, this approach would require the Court to permit the provision of subsidies on federally created exchanges. In King, the legislative history strongly suggests that Congress intended to make subsidies available on state- and federally-created exchanges alike. The Congressional Budget Office, which played a central role in crafting the ACA by scoring its costs, never even considered the possibility that insurance subsidies would not be available in every state.17 Rather, their analyses assumed that subsidies would be available nationwide.18 Congressional staffers who helped write the ACA insist that Congress did not intend to require states to create their own exchanges in order for their residents to receive subsidies.19 Lest this be dismissed as post hoc rationalizing, statements of senior staffers made as Congress deliberated over the law support this claim.20 Notably, an earlier draft of the bill did threaten states with a denial of tax credits—but for failing to comply with employer responsibility rules, not for refusing to establish an exchange.21 In any case, this punitive provision was conspicuously absent from the final bill altogether.22
Contrary readings of the ACA’s legislative history that purport to show that Congress intended to condition the availability of health insurance subsidies on a state’s establishment of an exchange are erroneous. Thomas Christina, the lawyer credited with first noticing that the ACA may limit subsidies to state-created exchanges, originally believed that this language might merely be an unintended drafting error. Only later did conservative activists craft an account of congressional intent to coerce states into establishing their own exchanges.23 In fact, a Republican-backed bill in 2011—by which time multiple states had manifested their intention to refuse to create their own exchanges—assumed that subsidies would be available in every state under the ACA.24 Remarks by Representative Paul Ryan made in 2010 also suggest that he believed that the law would make insurance subsidies available in all fifty states.25
Those who argue that the ACA does not permit the issuance of subsidies to those on federally created exchanges make a number of arguments that Congress in fact intended to withhold subsidies on federally created exchanges. These arguments, however, ignore the contexts in which the evidence they marshal arose. One is that the Senate Finance Committee, which played a major role in drafting the provisions of the ACA that provide for the exchanges, lacked jurisdiction over “non-group health insurance markets,” and needed to condition the availability of tax credits, which fall within its jurisdiction, on states’ creation of exchanges to legislate on the subject.26 In support of this claim, they quote Senator Max Baucus, then-Chairman of the Committee, as saying that the bill “conditions” tax credits on states establishing exchanges.27 But as one commentator has noted, Baucus’ quote had nothing to do with federally created exchanges, but with whether the Committee had jurisdiction to consider an unrelated piece of malpractice reform legislation.28 Moreover, the Committee did not need to limit the availability of tax credits to users of state-created exchanges to bring insurance exchanges within its jurisdiction; extending the availability of credits to users of any exchange, state or federal, would have brought the subject within its reach.29 In fact, this argument does not just fail, but backfires: if the ACA limited the availability of tax credits to users of state-created exchanges, the Senate Finance Committee would have lacked jurisdiction over the federally created exchanges.
Some also point to a statement made by MIT economist Jonathan Gruber, often described as the “architect” of the ACA,30 claiming that subsidies would not be available to users of federally created exchanges.31 Gruber certainly played an important role in designing the economic underpinnings of the Massachusetts health care reform law upon which the ACA is largely based, but his role in crafting the ACA itself was fairly minimal. He was a paid consultant who merely “provided the White House and sometimes its congressional allies with data . . . that they could use to devise policies or to defend their positions in public.”32 Second, his claim contradicts everything else that he himself has said about the law at other times, which indicate that the law should not be understood to make subsidies available only in states that establish their own insurance exchanges.33 It would be dishonest to credit Gruber’s comments only when they weigh against the availability of subsidies on the federal exchanges but not when they weigh in favor of it.
The incongruity between the statute’s seemingly-plain textual meaning and Congress’ intent to make subsidies available nationwide is that Congress was simply sloppy. Congress intended the ACA to serve as a first draft, whose language would be cleaned up prior to the passage of a final version.34 However, the election of Senator Scott Brown, a Republican, to replace Ted Kennedy deprived Democrats of a filibuster-proof majority in the Senate, forcing them to use the ACA, which was effectively a first draft, as the final product.35 This account of the ACA’s drafting, if true, explains how Congress could have enacted a law whose plain text seemingly runs contrary to its obvious intent—precisely the scenario that the categorical Chevron rules proposed here would resolve.
Alternatively, the Court can rule that legislative history is categorically irrelevant to Chevron analysis. Many, including Justice Scalia, have argued that legislative history should not be considered at Chevron Step One.36 There are several reasons to disregard legislative history in performing Chevron analysis. One is that Congress can be said to have a general macro-intent to delegate discretion to agencies when it passes ambiguous statutes, since it knows that “indeterminacy gives rise to discretion.”37 Then there are the traditional arguments against using legislative history—the impossibility of discerning “intent” in a collective body,38 the unreliability of traditional indicia of legislative intent,39 and the illegitimacy of giving legal status to the unenacted intent of particular Members of Congress,40 which was never subject to bicameralism and presentment.41 Statutory text may also reflect judgments not only about substantive goals but also the manner in42 and extent to which43 those goals should be pursued, which resorting to legislative history may undermine. In the context of King, for instance, one could argue that the ACA reflects a congressional judgment not only about whether qualified taxpayers should receive insurance subsidies, but about the precise means by which those subsidies are made available—through exchanges established by the states themselves. Finally, notwithstanding its language about congressional intent and agency expertise, perhaps Chevron is best understood as an effort to simplify a doctrine that had grown convoluted over time;44 explicitly incorporating an analysis of legislative history into Chevron analysis would undermine its hard-won elegance. The government would lose King under this “text trumps” rule; “State,” after all, means “State.”
Both categorical approaches outlined above offer the primary advantage of bright-line rules—clarity. The presence of either unambiguous legislative intent or unambiguous text is completely determinative, telling a judge exactly how the statute must be construed. At the same time, however, they also reflect the shortcomings of a bright-line approach. First, both the ‘legislative history trumps’ and ‘text trumps’ rules reduce agency discretion. Settling a question of statutory interpretation at Chevron Step One freezes a statute’s meaning until it is amended, preventing the agency from reaching a different conclusion at a later time due to new scientific or technical knowledge, changing circumstances on the ground, or the emergence of a new political majority. Moreover, both rules feel jurisprudentially incomplete, as both ignore something important about what law is. It seems wrong to say that a statute’s text or the intent that animates it can be categorically irrelevant to determining its meaning. As a theoretical matter, both text and intent inform a law’s meaning because both are basic, constitutive elements of what law is. An approach that would require judges to turn a blind eye to either threatens to endorse a stunted notion of law itself.45
There is a categorical third approach that avoids these difficulties—when text and legislative history are clear yet in conflict, a statute should be considered per se ambiguous, and a permissible construction relying on either text or legislative history should be considered per se reasonable. In effect, clashes between clear text and legislative history would result in the agency winning every time, provided only that the agency’s construction makes sense either as a matter of textualism or of legislative intent. This approach strikes a fair balance between text- and intent-oriented interpretive modalities,46 respecting (if not adopting) the reasoning undergirding both approaches. Cases where a statute’s text and legislative history sharply conflict are likely to present especially tricky questions of statutory interpretation. The “agency wins” rule acknowledges that an interpretation contradicting either a statute’s clear text or history cannot be unambiguously correct. It is also the jurisprudentially richest approach because it recognizes the valid claims to “lawfulness” of both textual and historical appeal.47 In other words, as a methodological matter, an interpretive claim premised on assertions about a legal document’s text or the intent of those who created it is a proper legal argument, of the sort that judges and other legal interpreters may validly consider. Moreover, giving controlling weight to either statutory text or legislative history would constrain an agency’s interpretive discretion, whereas the categorical “agency wins” rule is consistent with Chevron’s purpose of empowering agencies with broad policymaking latitude.48
Finally, the “agency wins” approach serves values associated with both judicial minimalism, the idea that courts should decide less rather than more,49 and representation-reinforcement, the idea that courts should generally defer to legislation produced by a fair and open democratic process.50 It does so by deciding interpretive questions in a way that maximizes dissatisfied parties’ ability to seek correction through the political process. If conflicting text and legislative history mean that a statute is per se ambiguous, then these cases—which, like King, might be politically charged—will be resolved at Chevron’ssecond step. The court’s inquiry at this step involves determining only whether an agency’s interpretation was reasonable, not unambiguously correct.51 Finding ambiguity at Step One thus leaves agencies free to change their mind during subsequent presidential administrations, ultimately entrusting the issue to the realm of electoral politics.52 True, dissatisfied parties can also seek legislative change, but legislative modifications to existing statutes are “infrequent” and involve “high transactions costs,”53 while presidential elections occur frequently and on a fixed schedule. There is also reason to believe that voters are more likely to vote based on their views of “general governmental policies” in presidential elections than in congressional elections.54 The “agency wins” rule therefore preserves politically delicate issues for the political process to resolve, whereas both the “legislative history trumps” and the “text trumps” alternatives would freeze a statute’s “unambiguous” meaning in place. Under the “agency wins” approach, a future Republican president could completely jettison the Obama Administration’s reading of section 36B(b)(2)(A) and adopt her own. In Chevron cases such as King,the outcomes of which will inevitably engender significant political repercussions, the Court may find an approach that would let it rule for the agency while leaving future presidents with the final word on a statute’s meaning especially attractive.55 Indeed, during oral arguments in King, Chief Justice Roberts suggested that he might take precisely this approach to the ACA.56
Going forward, the “agency wins” rule would do the most to resolve legislative history’s uncomfortable existence within current Chevron doctrine. This approach differs from existing Chevron approaches because it avoids the problem of determining the precise amount of weight to assign legislative history. Using even clear legislative history to find statutory meaning can be problematic because interpreters still must decide how far to stretch a statute’s text to accommodate it. This question, inherently difficult in ordinary cases, is especially vexing when a statute’s text is also clear, as section 36B(b)(2)(A) arguably is. In contrast, the “agency wins” rule accords respect to congressional intent in an eminently manageable way by employing legislative history to find not statutory clarity, which would require reconciling inconsistent text and history, but statutory ambiguity, which requires no such thing. Although it may not seem very respectful to assume that Congress writes laws poorly, it is surely more respectful for a court to find statutory ambiguity at Step One than to find that the statute unambiguously means something other than what Congress clearly intended it to mean. It is also consistent with the assumption that Congress means to write ambiguously in order to delegate politically hot decisions to the agencies that are best situated to make them. This assumption is perhaps a legal fiction, but it is at least a respectful one. Even judges who, unlike Scalia, are comfortable using legislative history would benefit from dispensing with the need to nebulously “balance” it against statutory text when clear text and clear intent conflict.
While this essay endorses the “agency wins” approach, any of the three would make for a helpful doctrinal development. It is more important that the Court adopt some categorical rule to give meaningful guidance to lower courts, agencies, and litigants when clear statutory text and legislative history contradict one another. King, which showcases a stark clash between textual and historical clarity, presents a particularly suitable vehicle to announce a rule for such cases. By directing courts to find per se ambiguity at Chevron Step One whenever a statute’s clear text and legislative history conflict, such a rule would help courts manage their workloads by avoiding the need to balance the two, and would improve the efficacy and flexibility of agency decision-making going forward.
Bradley Silverman is a member of the Yale Law School J.D. Class of 2016. He is gratefully indebted to Robert Black for inspiring and influencing him over their many conversations on these topics, as well as for Robert’s wonderful friendship.
Preferred Citation: Bradley Silverman, Statutory Ambiguity in King v. Burwell: Time for a Categorical Chevron Rule, 125 Yale L.J. F. 44 (2015), http://www.yalelawjournal.org/forum/statutory-ambiguity-in-king-v-burwell-time-for-a-categorical-chevron-rule.