The battle lines over the Affordable Care Act were drawn sharply today over an exotic question of statutory interpretation that has vast implications for the survival of the ACA as we know it. The two key decisions are Halbig v. Burwell, where Judge Thomas Griffith and Senior Judge Ray Randolph of the D.C. Circuit held that individuals who purchased their health insurance through the federal exchanges were not entitled to the subsidies for those who purchase their coverage through exchanges established by the states. Senior Judge Harry Edwards saw in this decision a “not-so-veiled attempt to gut” the ACA.
In the parallel Fourth Circuit decision in King v. Burwell, a unanimous court thought that the case was indeed close, but then deferred to the decision of HHS on the scope of the provision, relying on the time-honored proposition of the Supreme Court in Chevron USA v. National Resources Defense Council (1984), which holds that, when a statutory text lacks a plain meaning, the courts should defer to the statutory administrator to resolve the ambiguity between the two rival interpretations.
The issue has momentous significance because in some 36 states — through which over half the present enrollees have obtained their coverage — the exchanges are owned and operated by the federal government, not the states. Any judicial decision that knocks out these subsidies will lead to a two-tier system, which in turn will lead to a collapse of the overall program (not to mention a huge level of unpardonable dislocation to those individuals who thought they had coverage but now discover that after the fact they do not). Today’s split decisions create an intolerable level of uncertainty that will only end when the United States Supreme Court decides the case, which it should do on an expedited basis.
These long and learned opinions should not obscure the fact that at the root of the case is a simple question: Do the words an “exchange established by a State” cover an exchange that is established by the federal government “on behalf of a state”? To the unpracticed eye, the two propositions are not synonyms, but opposites. When I do something on behalf of myself, it is quite a different thing from someone else doing it on my behalf. The first case involves self-control. The second involves a change of actors. It is not, moreover, that the federal government establishes the exchange on behalf of a state that has authorized the action, under which case normal principles of agency law would apply. Quite the opposite: the federal government decides to act because the state has refused to put the program into place. It is hard to see, as a textual matter, why the two situations should be regarded as identical when the political forces at work in them are so different. Under the so-called “plain meaning approach”, there is no need to look further. The text does not authorize the subsidies for these transactions, so it is up to Congress to fix the mess that it created in 2010.
Or so the argument of the majority in Halbig goes. Administrative law, however, is a strange subject in which deference is given to the administrator in the case of ambiguity, which can arise, it is commonly claimed, when the statutory language is placed into its larger context. In this case, that context includes a phrase that allows the federal government to set up “such exchange,” from which the inference might be drawn that any exchange that the federal government sets up should be treated for all purposes as if it were a state exchange—a proposition that leaves it unclear why the specific language that relates to the subsidies does not incorporate that understanding. One of the sad features of the original Chevron decision was that it imported ambiguity into a statute whose operative provision was clear by using precisely this tactic: find a different section that can be read in tension with the operative position and allow the administrator to pick between inconsistent readings.
The first criticism, therefore, of the government’s position is that it is too driven by the Chevron precedent. An issue of this magnitude should not be decided one way in a Democratic administration only for it to be fair game for reversal in a Republican administration. This is a question of law that should be decided by courts, which resolve the ambiguity the best that they can. The administrators are themselves inevitable partisans in these cases, so it makes no sense to defer on questions of legal interpretation where they do not have access to any materials that are not fully available to judges.
Yet the Chevron rule is so ingrained that no circuit court judge would be prepared to depart from that rule if it turns out that the statute does not have a plain meaning. Ultimately, the position of Griffiths is that the meaning was plain (enough) to carry the day. On balance, he was right — notwithstanding the strong counterassault, which comes in three parts.
The first argument is that the context and structure of the act suggests another meaning. This position is derived from Justice O’Connor’s excellent opinion in FDA v. Brown & Williamson, where the question was whether tobacco should be treated as a drug subject to FDA regulation under a statutory provision that stated any substance counted as a drug if was “intended to affect the structure or any function of the body.” That phrase is really broad; indeed, nonsensically so. Justice O’Connor reached the conclusion that was consistent with FDA precedents on the point that it is absurd to sweep into this literal definition any substance for which “there was no claim of therapeutic or medical benefit,” which no one made for cigarettes. Indeed, it would be absurd to think that the FDA should conduct clinical trials to see how tobacco cures cancer.
Yet we are light years from that situation here because it is not incoherent to run a more limited program with the intent to drive states to form these exchanges. Indeed, that was just what was done with respect to the Medicaid mandate, where the effort was to cut out all benefits from pre-existing Medicaid programs if the states did not sign up for the new program—a position that was ultimately rejected by the Supreme Court in NFIB v. Sebelius (which also upheld the individual mandate — which now, of course, has been waived without clear executive authority).
A second point of contention concerns the interaction of legislative history with text. On that subject, there was stunningly little material to go on: only a single statement by Max Baucus in the Senate hearing that the legislation “conditions” the willingness of the state to set up the exchange. What is striking about the defenders of the government is that they do not cite any language that cuts in the opposite direction, but only claim that there is nothing in the legislative history that demonstrates the point. In this connection, however, the single statement by Baucus looms large, both because of his central role in the design of the statute and because there is nothing written on the other side of the issue.
Next it is said by Judge Edwards that the ACA had as its central purpose the extension of coverage to virtually all Americans, which could not be done if the subsidies were denied to people who enrolled on the federal exchange. But the difficulty with that argument is that legislation has multiple purposes, and, although he derides, he does not refute the alternative view: the statutory design was intended to give the states a strong incentive to create their exchanges so that the federal government did not have to expend its resources to do so. There is, again, little in the debates to resolve this question, but, by the same token, there is no explanation as to why this provision was inserted if it was not intended to have that effect. It would have been simple enough to draft a provision saying that everyone gets the subsidy no matter whether they enroll on the state or federal exchanges.
It is with decidedly mixed emotions that I conclude that the supposed ambiguity is not strong enough to displace the textual simplicity of the Griffith argument. It is really intolerable to first drive people out from private coverage and then pull out the rug from their federal coverage. What a miscarriage of justice. Sadly, however, those issues are not decisive on this question of statutory interpretation. On balance, I have to conclude that Judge Griffith’s opinion looks correct. The text seems to be clear and nothing else seems strong enough to displace it. It is an open question whether the Supreme Court will agree, as its precedents are sufficiently muddy that we live in a world of “anything goes.” My guess is that Griffith’s position will prevail 5-4 in the Supreme Court on a straight conservative-liberal split. This is, to be sure, an odd and unhappy way to make public policy.