Assume that in California v. Texas, the pending challenge to the Affordable Care Act (ACA or Act), the Supreme Court holds that the 2017 Tax Cuts and Jobs Act’s (TCJA) zeroing out of the potential tax penalty for failure to maintain the Act’s specified health insurance coverage rendered unconstitutional the part of the Act providing that individuals “shall” maintain insurance coverage. (We shall hereafter call this provision “the insurance requirement” although some refer to it—and we ourselves have referred to it in other Parts ‒ as a component of the Act’s “mandate.”) As we noted at the end of Part Three of this series, the Court will next need to fashion an appropriate remedy for the constitutional violation. Most litigants and analysts assume that the Court would declare the insurance requirement unconstitutional and enjoin its effect—an odd kind of relief to fathom since after 2017 the government doesn’t impose any consequences on the failure to procure insurance anyway. But we suggested another recourse would be to enjoin the 2017 amendment (essentially restoring the original tax-penalty schedule) to return the ACA to the form the Court upheld in 2012 in National Federation of Independent Business (NFIB) v. Sebelius.
Choice of Remedies if the Insurance Requirement is Unconstitutional
Under current remedial doctrine, where (as here) there are multiple ways the Court might remedy an unconstitutional edict, the Court takes its cue from congressional intent ‒ assuming it has some basis to infer what Congress’s counterfactual intent would be had it known its actually intended handiwork was unconstitutional. Just last term, the Court noted that when a statutory amendment interacts with the existing underlying statute to create a constitutional violation, the Court traditionally “has treated the original, pre-amendment statute as the ‘valid expression of the legislative intent’” and cures the constitutional problem by enjoining the amendment only. Applied here, this approach supports curing any constitutional problem with the insurance requirement by enjoining the 2017 zeroing-out provision rather than by enjoining the 2010 requirement itself.
Yet in our view, the Court’s current tendency to focus only on Congress’s counterfactual intent to determine the proper remedy is too narrow-minded. At least in some cases, other equitable principles deserve some consideration as well. First, as one of us has written, sometimes a constitutional norm lurks in the background and argues in favor of one remedy over another. For example, in last term’s Barr v. American Association of Political Consultants, Inc. (the case quoted above), the Court should have at least considered that remedying a content-based speech discrimination by enjoining the initial statute (which itself suppressed communication in a broad but content-neutral way) rather than enjoining the amendment (which partly lifted the suppression in a content-based way) would have promoted the norm that more speech is better than less. However, such a norm is not obviously implicated in the ACA context (both the insurance requirement and the tax might be said to inhibit liberty, albeit in different ways).
A second equitable principle does seem relevant here. We also believe that the Court should at least consider whether choosing one remedy over the other better rewards a successful plaintiff for suing, a practice that over time encourages rather than discourages meritorious litigation. By definition, both plausible injunctions in the ACA context will redress the plaintiff’s legal injury, but perhaps one injunction will better redress the plaintiff’s injury-in-fact. This equitable principle has not seemed to move the Court in the past (indeed, not in Barr where the plaintiff sued to lift a discriminatory restriction on its speech but merely secured a broader restriction on others’ speech). But the concern is particularly sharp in the ACA case, where one remedy would arguably make the successful individual plaintiffs (hereafter Hurley) not just unimproved in a tangible sense but actually materially worse off than they were before filing suit: if the Court enjoins the 2017 tax-zeroing, Hurley will have prevailed in his constitutional challenge at the direct expense of his pocketbook.That is surely a disincentive for individuals to assert constitutional claims.
And here’s a new and third remedial principle to consider: where possible, avoid constitutional thickets. As noted in earlier Parts of this series, when an ambiguous statute fairly admits of two different interpretations, and one of those interpretations risks constitutional trouble, the Court should adopt the safer alternative even if it represents a less natural reading of the statute. As we explained in Part One of this series, this Ashwander principle drove Chief Justice Robert’s saving interpretation of the tax penalty in NFIB; and as we explained in Part Three, the same principle should lead the Court to reject the underlying constitutional challenge here. But if the Court disagrees and invalidates the insurance requirement on the merits, we think the same principle applies as well to the Court’s remedial choice.
How might this choice be informed by Ashwander avoidance values? If the Court enjoins either the 2010 requirement or the 2017 tax-zeroing provision, the Court will face a broader question of severability: whether the 2010 insurance requirement is severable from the rest of the ACA, or whether the 2017 tax-zeroing provision is severable from the rest of the TJCA (or indeed the broader bill of which it is part). In our view, the severability question itself, at bottom, raises constitutional concerns (even if congressional wishes are part of the analysis). Most obviously, the question raises separation of powers concerns: the Court will end up either leaving in place a partial statute that Congress didn’t really enact, or invalidating swaths of a statute that Congress did and fully had the power to enact. But more fundamentally, as we will explain in our forthcoming Part Five, in many cases the severability question also raises a due process concern: in our view, an important reason an inseverable provision should be enjoined (perhaps most appropriately in a subsequent case) is because otherwise an individual will be subject to a statutory provision that Congress did not intend (under these second-best circumstances) to have the force of law. Put differently, litigants claiming inseverability are essentially arguing that liberty or property is being denied on account of a non-law ‒ or perhaps “Zombie-law.” Seen in this light, a difficult severability question is yet another form of constitutional thicket that ought to be avoided if “fairly possible.” So if Remedy A would trigger a very complicated and close severability analysis and Remedy B would trigger an easy and obvious severability analysis, this distinction itself provides reason to prefer Remedy B.
In the ACA context, at least upon superficial analysis, Remedy B (enjoining the 2017 tax repeal) wouldn’t seem to raise a difficult question at all; the tax-zeroing provision appears conceptually and functionally disjoined from other provisions of the TCJA, and enjoining it alone does not seem foreclosed by an applicable inseverability clause. As we explain below, Remedy A (enjoining the insurance requirement) should also lead the Court to sever that provision and uphold the rest of the ACA. But the challengers at least put up a fight here, and the answer, while we think it clear, is not snap-your-fingers simple, largely because the challengers argue that there is, in effect, an applicable inseverability clause. So even if we’re correct in our ultimate conclusion ‒ but especially if we’re not ‒ the extended Ashwander principle provides strong reason to simplify things and enjoin the 2017 tax repeal to avoid this entire inquiry.
The General Doctrine of Severability
Severability becomes an issue in California v, Texas if the Court finds: (1) that plaintiffs have standing; (2) that the word “shall” in the insurance requirement became obligatory after 2017; (3) that an obligatory meaning of “shall” is beyond congressional power; and (4) that the proper remedy is declaring the insurance requirement invalid and enjoining its operation. The Court would then have to address the challengers’ $64,000 question—namely, does the rest of the ACA fall because it is inextricably linked to the insurance requirement?
In the proceedings below, the district court ruled for the challengers down the line and concluded that the insurance requirement could not be severed and thus the entire ACA was rendered invalid. On appeal, the Fifth Circuit agreed that the requirement became mandatory after 2017 and was constitutionally invalid in that form, but on the issue of severability it remanded the case to the district court for it to “employ a finer-toothed comb . . . and conduct a more searching inquiry into which provisions of the ACA Congress intended to be inseverable from the mandate” and thus should also be enjoined along with the requirement. The Fifth Circuit also instructed the district court to take a closer look at how broad the scope of any remedy should be. The Supreme Court granted review before the district court performed these analyses.
Hurley, Texas (shorthand for the state challengers), and the Solicitor General (or “SG” for the intervening United States) argue that the insurance requirement cannot be severed and thus the entire ACA must fall. Notably, the SG’s position on severability has not been consistent. Although it took less extreme positions in the lower courts, the SG now agrees with the challengers that the entire ACA must come down. Its argument proceeds in two steps: first, the insurance requirement is inseverable from the ACA’s guaranteed-issue and community-rating provisions (though the SG acknowledges these would be constitutionally valid standing alone); and second, every other part of the Act is tied to one or more of these three sets of provisions so that nothing in the Act is severable and salvageable.
In general terms, the judicial inquiry into severability reflects a court’s effort to cabin the effect of its invalidation of a particular statutory provision upon the remainder of the statute. “When confronting a constitutional flaw in a statute,” the Court has explained in one case, “we try to limit the solution to the problem” by severing “problematic portions while leaving the remainder intact.” The Court has said there is a “well established” severability framework even as it has described the framework in different ways. Many cases invoke a two-part test that asks first whether the remaining statutory provisions can still operate in the way Congress designed them and second whether Congress would have enacted the remaining provisions standing alone. In a recent formulation of this test, a unanimous Court (in an opinion by Justice Clarence Thomas) explained: “[w]e ordinarily give effect to the valid portion of a partially unconstitutional statute” if (1) that portion “remains fully operative as a law” and (2) it is not evident from the statutory text and context that Congress would have preferred no statute at all to a partially valid statute. These two framework elements are related because both go to the ultimate question of “determin[ing] what Congress would have intended in light of the Court’s constitutional holding.” Severability thus involves a counterfactual inquiry: figuring out what Congress would have done had it known some portion of a statute would later be invalidated. Courts may consider a range of evidence in determining what Congress’s preference would have been. Express severability or inseverability clauses are generally very good evidence but they are not always dispositive and the Court has made clear that “[i]n the absence of a severability clause . . . Congress’s silence is just that—silence—and does not raise a presumption against severability.” Indeed and importantly, nearly all recent cases either explicitly or implicitly deploy a presumption (which we argue below ought to be a very strong presumption) that Congress would have wanted to preserve independently unobjectionable portions of a statute.
A fair-minded application of previously established severability principles to California v. Texas would lead to preserving the remainder of the ACA.
The ACA’s defenders argue that the obvious and compelling indication of congressional preference is what Congress actually did in 2017. We agree that 2017, rather than 2010, is the proper focus of the inquiry into congressional intent. It was, after all, the 2017 Congress that (under the challengers’ theory) introduced the constitutional infirmity; we think it proper to ask what the Congress at the time the constitutional flaw arose would have wanted the rest of the Act to look like had it foreseen that courts would determine that zeroing-out of the tax penalty made the insurance requirement constitutionally invalid.(We suppose someone could argue that today’s Congress, in 2020, is the one whose counterfactual intent we should consult on the ACA, and that Congress’s views about health coverage in the midst of a pandemic might be quite different from its views even a year earlier, but we don’t explore that possibility here.)And when we look at what Congress did in 2017, we see that it left in place every other provision of the ACA. For the ACA’s defenders, that would seem to end the matter.
But this is too quick. The challengers’ severability theory, which must be engaged on its own terms, is essentially that Congress, after eliminating the tax penalty, was counting on the mandatory nature of the insurance requirement (and the resulting sense of legal obligation among the citizenry) to keep the number of people procuring insurance high enough to make the other parts of the Act function as intended. Whether that theory is plausible—as a matter of the post-amendment statutory text or as a matter of reasonable factual economic expectation by Congress—is another matter, one on which we opine below. But the ACA’s defenders must confront the theory head-on. They cannot avoid the challengers’ argument, which says that invalidating the requirement undermines Congress’s 2017 actual intent, simply by observing that Congress could have repealed, but did not repeal, the entire ACA when it acted three years ago.
The challengers’ argument, in more detail, is that statutory “findings” the 2010 Congress adopted function as a statutory inseverability clause with respect to the insurance requirement and the ACA’s guaranteed-issue and community-rating provisions. The challengers emphasize Congress’s use of the word “essential” in several of the findings concerning the relationship between broadened individual coverage and these other ACA provisions. In particular, the challengers point to Congress’s 2010 finding that by increasing the number of healthy individuals in the insured pool, the insurance requirement is “essential to creating effective health insurance markets in which improved health insurance products that are guaranteed issue and do not exclude coverage of pre-existing conditions can be sold” (emphasis added). Because these “major” ACA provisions are thus operationally dependent on the insurance requirement, the former must fall with the latter. And once these “major” ACA provisions fall, the challengers say, none of the rest of the ACA can be salvaged, not because the rest of the Act is covered by anything resembling an inseverability clause, but because, according to the SG, “it is evident that Congress would not have enacted the [rest of the Act] without the individual mandate and the guaranteed-issue and community-rating provisions.” Importantly, the challengers contend that because the 2017 Congress did not alter or repeal the 2010 findings, it shared the view that should the insurance requirement be invalidated, the “major” provisions, and thus the entire ACA, must go.
There are several difficulties with this argument.
First, the ACA’s findings clearly do not constitute the sort of “express” inseverability clause to which courts presumptively give effect. Congress knows how to write an inseverability clause in plain language. It did not do so here: the statutory findings make no mention at all of severability. The challengers do not explain why in 2010 or in 2017, when Congress (on their account) wanted courts not to sever unconstitutional provisions of the ACA, Congress failed to include an express inseverability clause but instead expected legislative findings to function as such. Nor do the challengers point to any other case in which the Court has deemed statutory findings of the kind Congress included in the ACA to serve as an express inseverability clause. That Congress in 2017 would have understood the 2010 findings to constitute an inseverability provision seems especially unlikely given NFIB: no member of the Court in NFIB argued that the ACA’s findings were in effect an inseverability clause (or thought the ACA otherwise contained an inseverability provision).
Accordingly, on the issue of severability, at most the 2010 findings could serve as one piece of evidence of congressional intent. Even then, the question remains: intent as to what? The challengers say: intent that the entire ACA must fall. But that too is quite unlikely given what the findings actually say. A more plausible understanding is that Congress in 2010 articulated legal conclusions to support the insurance requirement as an exercise of Congress’s Commerce Clause power as the Court construed that power in seminal cases such as United States v. Lopez, United States v. Morrison, and Gonzales v. Raich. The “findings” begin with Paragraph 1, which closely tracks the language of these Commerce Clause cases. It states: “The individual responsibility requirement provided for in this section . . . is commercial and economic in nature, and substantially affects interstate commerce, as a result of the effects described in paragraph (2).” Paragraph (2), titled “[e]ffects on the national economy and interstate commerce,” then includes ten sub-paragraphs of findings, several of which (beginning with the first that “[t]he requirement regulates activity that is commercial and economic in nature”) also track the Court’s language in Commerce Clause jurisprudence.
Texas argues that the congressional findings, made in 2010 and preserved in 2017, are the “law of the land.” But this is a misplaced label if we understand Congress’s findings not as any “law” but simply as an explanation of the constitutional basis for its power to enact a statute. NFIB did not agree with Congress’s “finding” that the insurance requirement was a permissible regulation of something that was “economic” and that “substantially affects interstate commerce” (indeed, five justices rejected the very argument). Whether in 2017 Congress retained its 2010 findings because it continued to view the insurance requirement as a proper exercise of its Commerce Clause power (an argument that, as we explained in Part One, remains available even though three still-sitting Justices rejected that conclusion), because the Senate passed the 2017 amendment under reconciliation procedures (allowing a simple majority insulated from filibustering to approve budget-related legislation) that did not permit repeal of the findings, or gave the question no further thought—the preservation of a legal conclusion from Congress is quite different from a statement of its preference as to severability, let alone from a “law.”
Next, the challengers emphasize the word “essential” in the findings. But if the findings are understood as evidence of congressional preference, the presumption of severability surely informs how best to read them. Just as “shall” should not be understood as a command if there is another “fairly possible” meaning that avoids a constitutional difficulty, reasonable constructions of “essential” that comport with the presumption of severability should be preferred.
Indeed, the same underlying premise—courts should sidestep invalidation of Congress’s legislative enactments if they can find a reasonable interpretation of Congress’s words that sustains federal enactments—lies at the core of so-called Ashwander avoidance principles and must also drive severability doctrine. In other words, if we choose plausible but second-best interpretations of statutes under Ashwander to avoid invalidation-by-unconstitutionality, shouldn’t we do the same to avoid invalidation-by-inseverability? For this basic reason, absent an explicit inseverability clause or a patent unworkability of remaining parts of a statute, the Court should indulge a very strong presumption of severability provided Congress’s text can fairly be read to preserve parts of laws that are by themselves constitutionally unobjectionable. Some Justices, such as Justice Brett Kavanaugh and those who joined him in Barr, seem to intuit the propriety of this strong presumption of severability, but none of them seems yet to appreciate the linkage between Ashwander principles and severability that make the strong presumption of severability the best way to cohere doctrine generally.
The challengers deem “essential” to mean indispensable, such that the remainder of the statute simply cannot exist without the insurance requirement. But that is not the only fairly possible or even the most natural interpretation of “essential.” Here, McCulloch v. Maryland is instructive. Recall that Chief Justice John Marshall, interpreting the Necessary and Proper Clause, wrote that “necessary” does not always mean “an absolute physical necessity, so strong, that one thing to which another may be termed necessary, cannot exist without that other” but instead the term “frequently imports no more than that one thing is convenient, or useful, or essential to another.” Borrowing from Marshall, the word “essential” (like necessary) may “admit[] of all degrees of comparison, and is often connected with other words which increase or diminish the impression the mind receives of the urgency it imports. A thing may be . . . [essential], very . . . [essential], absolutely or indispensably . . . [essential].” Congress nowhere found the individual insurance requirement “indispensably essential,” and thus one might well understand it as being merely useful or conducive to the other healthcare reforms Congress implemented.
And even if one adopts the challengers’ reading that essential means indispensable, we would need still ask: essential to what? The challengers repeatedly say essential to the operation of insurance markets. But that overlooks the actual text of the findings. As noted, the 2010 statute says that the insurance requirement “is essential to creating effective health insurance markets in which improved health insurance products that are guaranteed issue and do not exclude coverage of pre-existing conditions can be sold.” 42 U.S.C. 18091(2)(I) (emphasis added). Congress in 2017 might have agreed that the insurance requirement was essential (in the strong sense) to creating markets in 2010 but fully understood that seven years later the requisite markets were by that time up and running so that the requirement was no longer essential to their continued success.
Further, a statement of Congress’s understanding of market operations (in 2010 that perhaps can be said to be repeated in 2017) does not support the further conclusion that those understandings were a but-for cause of Congress’s enactment of the rest of the ACA. Congress might have sought well-functioning markets but nonetheless been willing to settle for moderately well-performing ones ‒ even as it issued findings of a relationship between the insurance requirement and the markets. That Congress thought the insurance requirement indispensable to markets doesn’t show that it wouldn’t have wanted the other ACA provisions to remain in place if the insurance requirement fell.
We also note that the challengers’ argument that after 2017 the statutory findings must be understood the very same way they were understood in 2010 (the SG, for example, says that the 2017 amendment “does not alter what Congress said in the ACA about how these three provisions are inextricably intertwined”) is in tension with their argument that the meaning of “shall,” though retained with the 2017 amendment, did necessarily change in its meaning ‒ from (as construed in NFIB) a choice between purchasing insurance and paying a tax to (after 2017) an outright command to purchase insurance. The challengers do not explain why the Court should give the insurance requirement a new reading after 2017 but must give the findings relating to it the very same meaning as in 2010.
Finally, the challengers do not acknowledge the full reach of their approach. While the challengers repeatedly invoke the findings that deem the insurance requirement essential to the markets the ACA sought to create, they do not cite the related finding that the insurance requirement is also “essential” to the operation of other federal laws that regulate health insurance, notably the Employee Retirement Income Security Act (ERISA). We have a hard time imagining that the challengers take the position that Congress in 2010 and 2017 decided that if the insurance requirement were to fall, it should drag ERISA down with it. Yet if ERISA survives, why not the other ACA provisions as to which the insurance requirement is equally described as “essential?”
More on this provocative notion concerning the effect of severability analysis on other statutes—and related questions about how severability analysis in one case with one set of parties ought best to be conceptualized when it has carryover effects to other cases involving other parties—in our next and final installment.
