In the first four parts of our series on the California v. Texas case currently before the Supreme Court—the latest broadside challenge to the Affordable Care Act (ACA)—we identified and evaluated a number of nitty-gritty arguments concerning the three primary areas of controversy: standing to challenge the individual insurance requirement ( sometimes called the “individual mandate”), the merits of the constitutional challenge, and the severability question that would arise were the Court to find the insurance requirement unconstitutional. In this final installment, we ruminate on some bigger-picture questions of how severability issues ought to be characterized and addressed.
Recall that a severability inquiry arises when a court enjoins as unconstitutional one provision of a multi-provision statute (for ease of exposition we use this simple case; our analysis also applies if a court declares unlawful a combination of provisions or just certain applications of one or more provisions). The question thus created is whether the enjoined provision A is “severable” from the rest of the statute (which by hypothesis is not independently unconstitutional) such that statutory remainder B-Z still enjoys status as binding law; or whether the enjoined provision A is “inseverable” from some or all of B-Z in which case the inseverable parts of B-Z are also rendered inoperative.
Jurists and scholars have heretofore characterized the severability inquiry in two fundamentally different ways—remedial and substantive—and each characterization has different implications for when and how severability analysis properly occurs. The majority view deems severability to be part of the court’s remedial inquiry: after finding provision A unconstitutional, the court considers whether to sever A from provisions B-Z as part of its process of fashioning redress for the plaintiff’s legal injury. The court might favor inseverability to avoid creating a perceived separation of powers concern arising when a statutory remainder that Congress would not have wanted to enact on its own remains operative. As a corollary, because of its presumed remedial nature, severability may be decided by the court whether or not a party raised the issue or is injured by and thus has standing to challenge any or all of the statutory remainder. The minority (but perhaps ascendant) view is that severability is a question of substantive law: once provision A is declared unconstitutional, whether A should be severed from provisions B-Z turns on an unusual form of statutory interpretation. As a corollary, because the argument that A’s downfall should also void some or all of B-Z is a substantive claim on the merits, the court may entertain such a claim only when it is presented by a party who has standing to raise it, meaning one who is injured by the provision(s) B-Z that might fall.
In our view, both the majority and minority positions are partly right and partly wrong. We agree with the minority view that severability is a substantive claim concerning the legal status of the remaining statutory provisions. Inseverability is not part of the court’s remedial calculus because it is immaterial to redressing the plaintiff’s legal injury. Rather, inseverability is a claim about the statute’s contingent meaning, testing whether Congress expressly or impliedly provided in the original statute that parts B-Z should functionally be repealed and thus become “zombie law” if A is declared unconstitutional.
But we also agree with the majority view’s conclusion that, at least sometimes, the properly-viewed-as-substantive severability question may nonetheless be addressed by the court at the same time that it fashions a remedy for the plaintiff’s legal injury. In other words, while the severability question is not itself part of the remedial calculus, it can reasonably be addressed by the court as ancillary thereto.
Problems With the Conventional Severability-as-Remedial-Inquiry Framing
At least recently, the Court (joining most scholars) has characterized the severability question as part of the remedial inquiry triggered by the underlying legal challenge. In NFIB v. Sebelius itself, five Justices (Roberts joined by Breyer and ; Ginsburg joined by Sotomayor) specifically described severability as part of the remedial inquiry, and the four joint dissenters (as explained below) acted consistently with this characterization. According to this framing, once a court decides that statutory provision A is unconstitutional, the court’s remedial power encompasses two steps.
First, the court must redress the plaintiff’s constitutional injury. Sometimes the court can simply enjoin provision A’s enforcement, and other times the court must choose between two (or more) injunctive remedies, each of which would redress the plaintiff’s mistreatment. As we explained in Parts Three and Four of this series, for example, in California the Court could remedy the (arguendo) unconstitutional mandate either by enjoining the original mandate or by enjoining the 2017 tax-zeroing provision.
Second, the court may address the fact that the injury-redressing injunction leaves on the books a statutory remainder that is not itself under constitutional attack but now has a new enforceable content which Congress never intended to exist. In the conventional story, the court should consider the separation of powers implications of letting the statutory remainder stand in this form. The joint dissent for four Justices in NFIB noted two related concerns. First, merely fashioning a remedy tailored to redress the plaintiff’s legal injury would potentially leave behind a statute unenacted and undesired by Congress, either because it no longer functions as Congress intended and/or because the loss of provision A thwarts the political deals and compromises necessary to facilitate A-Z’s passage. Of course, Congress could in theory revisit and either repeal or modify the statutory remainder, but for various reasons doing so might be politically difficult or costly such that legislative inertia functionally protects the unwanted remainder from override. (On the flip side, complete inseverability can eviscerate an entire statutory scheme based on a relatively minor constitutional defect.) And relatedly, the NFIB dissent bemoaned, the court itself could be viewed as complicit in warping the legislature’s product, functionally “assum[ing] the legislative function” by rewriting the statute and creating “its own new statutory scheme” which “can be a more extreme exercise of the judicial power than striking the whole statute.” It follows, the predominant view holds, that the court’s remedy after finding provision A unconstitutional should simultaneously take into account these potential separation of powers concerns raised by complete severability. This is best done by both redressing the plaintiff’s legal injury arising from provision A and also by considering whether to hold inseverable and thus annul the future enforceability of any and all provisions B-Z that the court believes Congress would not have wanted to stand without A.
And if the court views the separation of powers risks arising from full severance as an appropriate remedial concern, it follows that the court can address these risks now based on the plaintiff’s (arguendo) standing to challenge and seek redress from provision A, whether or not the plaintiff (or any other party) would independently have standing to complain about the operation of provisions B-Z. Requiring the court to ignore these separation of powers risks until (if ever) a party comes forward with standing to challenge B-Z would, as the NFIB dissenters noted, “be particularly destructive of sound government” and generate uncertainty as to the enforceable state of the law. Finally, as we noted in Part Four, the choice of remedy itself might also be informed by the ease of the severability analysis that different remedial options present.
Treating severability as part of the remedial inquiry holds intuitive appeal. The severability issue is triggered by finding a constitutional violation, and when there are multiple ways to redress that violation the court invokes the same method—appeal to Congress’s likely intent—both to select a plaintiff-oriented remedy and to answer the severability question. (Indeed, perhaps it is because of this methodological similarity that the Court recently but misleadingly said it applied “ordinary severability principles” to choose between two plaintiff-oriented remedies for a content-based speech regulation.)
But calling the severability question “remedial” in nature fundamentally departs from the traditional meaning of the concept in Article III adjudication. As scholar John Harrison has persuasively explained, judicial remedies are designed to address “the legal relations of the parties before the court” and redress the legal wrong done to the prevailing plaintiff, and are not designed to opine about or operate on the status of legal rules in general. Justice Thomas recently endorsed Harrison’s view, writing that “[r]emedies ‘operate with respect to specific parties,’ not ‘on legal rules in the abstract.’”
Even broadly prophylactic rules are designed to protect against constitutional injuries of the sort visited upon the plaintiff, not to address potential but unrelated legal concerns. And whether provision A’s injunction should affect the future legal effect of provisions B-Z has to do with Congress’s view of the relationship among those provisions, and nothing at all to do with the nature, scope, severity, or any other feature of the legal wrong worked by provision A. Indeed, the severability analysis is completely insensitive to whether provision A violated one person’s due process rights or three hundred million persons’ free speech rights. So there’s a large conceptual gulf between traditional plaintiff-and-legal-injury-tailored judicial remediation and the statutory analysis underpinning the severability inquiry. The two serve different functions and address different concerns.
The Substantive-Law Foundation of Severability Doctrine
Severability is better understood as a species of statutory analysis, though one quite different from conventional statutory interpretation. Severability sees the original statute, whether it presents itself this way or not, as a form of contingent legislation. In addition to containing provisions A-Z, the statute may also contain visible or latent instructions for altering its content upon the contingency that a provision is judicially declared invalid. In other words, the statute originally contains A-Z; after a court enjoins A and triggers the immanent contingency, the statute’s content shifts to either B-Z, only a subset of B . . . Z, or nothing—depending on the extent to which A is severable from the other provisions.
This framing is most visible and intuitive where Congress includes an express inseverability clause (for example, “if provision A is found unconstitutional, provisions F and M are both deemed inseverable and thus null and void”). Such a clause specifies a contingency (partial judicial invalidation) and an automatic response that instantly amends the underlying statute by repealing some or all parts of the remainder. It’s worth emphasizing that properly understood, and contrary to some judicial statements, an inseverability clause is not merely a precatory guide advising courts of Congress’s default intentions or preferences should the contingency arise. Rather, an inseverability clause is an operative statutory command that reshapes the statute’s content should the contingency arise.
Once the contingency is triggered, inseverable provisions become akin to “zombie law” as a substantive matter—they walk and talk but are no longer alive. As such, someone injured by a “zombie provision” can defend against or affirmatively sue to enjoin its application. Just as people have a substantive legal claim not to be injured by federal statutes that failed bicameralism or presentment or were subsequently repealed, so too people injured by inseverable provisions have a substantive legal claim (whether based on Article I or Fifth Amendment Due Process) not to be injured by a once-binding legal provision that is now defunct.
Understood this way, express inseverability clauses fit into a much broader family of contingent legislative acts. Rather than merely dictating which original provisions of a statute survive its partial invalidity, Congress might respond to such a contingency by springing into effect provisions that originally lay dormant. For example, Congress might have provided that should the ACA’s individual mandate be enjoined, insurance purchases would henceforth be encouraged through mass media campaigns and fast food coupons. Or Congress might respond to such a contingency by voiding or amending completely separate statutes. Recall in Part IV we noted that if the challengers are correct that the “essential” findings render the individual mandate inseverable from the rest of the ACA so the latter necessarily falls with the former, the findings arguably operate equally to annul also the entirety of ERISA and the Public Health Service Act. More generally, Justia contributor Mike Dorf insightfully discusses the general category of “fallback” laws that become operative upon judicial declarations of partial invalidity, as well as ways in which such fallback laws might themselves be unconstitutional. In all such cases, the contingency—whether a judicial declaration of invalidity or otherwise—triggers a change in the law. And judges discern the nature of that change by interpreting the substance of operative law, not by considering how to remedy the plaintiff’s legal injury triggering the change.
Of course, sometimes Congress specifies through an express severability clause that a judicial finding of partial invalidity does not change the substance of the statutory remainder. An express severability clause dictates that once provision A goes, provisions B-Z remain legally operative in their entirety. Such a clause might appear in the statute being challenged, or it might be enacted as a separate meta-rule governing all other statutes in which case it might contain narrow exceptions (for an example of a state’s generally applicable and strict severability mandate, see here; for one that allows potential exceptions, see here). Given our understanding that severability clauses are a form of contingent legislation, contrary to conventional wisdom they ought to be treated by courts as binding rather than merely precatory indicators of congressional intent.
So much for express (in)severability clauses. What if no such clause appears either in the underlying statute or in a generally applicable meta-rule? While it does so less visibly, a claim of implied inseverability follows a similar analytical path: sometimes Congress implicitly establishes a contingency plan in the original statute dictating that partial invalidity voids part or all of the statutory remainder. What might evidence such a latent plan? Well, the same things that count as evidence under current severability doctrine. Perhaps part or all of B-Z are clearly unworkable absent A (for example, where part A raises funds and part B dictates how those precise funds are spent; without A, B cannot function). Or perhaps part or all of B-Z would clearly undermine Congress’s intent absent A (for example, where part A comprehensively regulates gun ownership and part B preempts all state gun control laws; without A, B contravenes Congress’s regulatory objective). Implied inseverability contingency plans might be few and far between, such that this approach will mimic a strong presumption of severability. It’s worth noting that implied repeals of statutory provisions based on other contingencies (such as arguably inconsistent or field-preempting subsequent legislation) are similarly and strongly disfavored; Congress is generally presumed to want to preserve enacted legislation as much as possible. But again the key conceptual point is that an argument for implied severability rests on a substantive statutory claim that the judicial finding of partial invalidity triggered an implied repeal of other provisions, and not on any conceptual effort to redress the plaintiff’s injury. A person injured by a repealed-by-implication statutory provision has the same cause of action against the adverse application of zombie law as we described above.
Why Courts Can Nonetheless Consider Substantive-Law Inseverability Arguments at the Same Time That They Fashion Remedies for the Plaintiff’s Legal Injury
As we’ve explained, provisions B-Z that are inseverable from a judicially invalidated A become zombie provisions as a substantive matter—they remain “on the books” but lose their status as binding law. As with the walking dead on TV, however, zombie laws can still threaten to injure people until the laws are definitively killed. When may courts take aim?
Everyone agrees that the court finding provision A unconstitutional, after fashioning an appropriate remedy therefor, may and presumptively should right away consider a party’s claim that specific B-Z provisions that injure her must also be enjoined. The substantive inseverability claim becomes ripe and the court should treat it just as it would any other substantive claim for which the complaining party has standing.
The minority view draws the line here: as with other substantive claims, the court may not (because it lacks jurisdiction to) consider whether specific B-Z provisions should be enjoined as inseverable if they do not imminently threaten to injure a complaining party. Those potentially zombified provisions may continue to stalk the countryside unless and until a new case arises in which a party has standing to seek an injunction to kill them off.
Here’s where we part company with the minority view. Because the severability question is not itself remedial in nature, the court’s authority to address it just after invalidating provision A requires separate justification. But we view the severability and remedial questions as sufficiently intertwined for this purpose.
Recall the objection that invalidating provision A to redress the plaintiff’s legal injury, without more, raises separation of powers concerns by “judicially creating” a statutory remainder that Congress didn’t and perhaps wouldn’t support. In other contexts, we expect courts to consider whether potential remedies for constitutional wrongs might raise separation of powers concerns (constraining, for example, damages against the President or injunctions against the United States) or federalism concerns (constraining, for example, certain state court injunctions against federal officials) or even rights-based concerns (constraining, for example, injunctions that impose prior restraints). Admittedly these constitutional concerns properly influence the court’s underlying remedial order, unlike the separation of powers concerns relevant here. But all of these concerns are similarly triggered by the court’s remedial inquiry. And it similarly makes sense here for courts to ensure that their plaintiff-oriented remedies do not threaten other constitutional principles. As such, we believe, the court enjoys ancillary authority to self-police by considering (even sua sponte) whether it should enjoin other provisions along with A, even if no party to the litigation cares.
Now, whether the court should exercise that authority in a given case might turn on the circumstances. The more complicated the severability question appears to be, the more a court might feel unequipped to address it well without focused briefing by adversarial parties (or at least amici). And the separation of powers concerns raised by statutory remainders may be less urgent where the underlying statute is relatively narrow in scope and/or infrequently applied, or conversely where separate suits challenging the potentially zombified provisions will likely arise soon. In short, a court invalidating provision A (1) presumptively should entertain inseverability challenges to provisions B-Z raised by injured parties and (2) enjoys discretion to consider sua sponte other severability implications.
What Does This Mean for California v. Texas?
Assume the Supreme Court finds the individual mandate unconstitutional and enjoins its operation. The individual plaintiffs may challenge as inseverable the ACA’s insurance-reform provisions that specify minimum coverage requirements, because they claim those provisions prevent them from buying more desirable and cheaper insurance products. And the state complainants may challenge as inseverable various other provisions that constrain their behavior, such as reporting requirements. (We note also that if the Court finds the individual mandate unconstitutional and enjoins the tax-zeroing provision, an entirely different severability question arises as to the statutory remainder of the 2017 Tax Cuts and Jobs Act (of which the tax-zeroing provision was but one small part).
Beyond that, we think the Court should exercise its discretion to consider the challengers’ argument that the ACA’s essential findings constitute an express inseverability clause zombifying all other provisions of the ACA. First, these statutes are far-reaching and potentially impact a great deal of behavior. Second, it might take years to sort out inseverability on a standing-only basis because many different provisions regulate many different kinds of behavior (and some such as spending provisions may not generate cognizable Article III injuries at all). Third, the Court already should consider the plaintiffs’ express inseverability claim for those provisions that do currently injure them; that claim logically applies on an all-or-nothing basis. And finally, the express inseverability claim has been fully briefed and argued.
By contrast, we would advise the Court not to consider sua sponte any implied inseverability arguments in this case, except perhaps as applied to the specific provisions that injure the challengers. Implied severability claims are by their nature much more context-specific and nuanced, and such claims would better be left for future litigation in which the briefing pays much closer attention to the relevant functional interrelationships between the individual mandate and the myriad other ACA provisions.
In sum: if the Court invalidates and enjoins the individual mandate, the Court should consider the challengers’ substantive express inseverability claim that the entire ACA remainder should be enjoined as well. And for the reasons we sketch in Part IV (and those more comprehensively and meticulously developed by Abbe Gluck among others), the Court should emphatically refuse.