Severability as an Erie Question

I appreciate the opportunity to post this month. To begin, I'll share some thoughts about the doctrine of severability, which seems to be receiving greater attention lately in light of its potential role in the Supreme Court's upcoming decision on the Affordable Care Act.

I see the doctrine as a particularly attractive point of discussion for a couple of reasons. On one hand, it is important--severability figures centrally in a broad array of state and federal constitutional litigation, and is powerful because the viability of large statutory schemes can hinge entirely on whether an unconstitutional provision is severable. If the Court invalidates the individual mandate in the health care litigation, for example, severability will decide whether the mandate's invalidity brings down the entire Affordable Care Act. High stakes, indeed.

On the other hand, severability is a deeply vexing doctrine. Most agree that the Court's historical approach to it is unsatisfying. In some cases the Court decided severance questions without explanation. In others the Court articulated one doctrine but seemed to apply another. Moreover, many have disagreed with the substantive content of the announced doctrine; a handful of articles suggest varying reforms, including abolishing severability altogether. 

I think there is another aspect to the jurisprudence that is equally intriguing but less recognized, and it is one that frames severability as a question of vertical choice of law. In other words, when a federal court invalidates part of a state statute, whose law determines whether the unconstitutional portion is severable? Should it be the federal law of the reviewing court, or the law of the state whose statute is under review? 

Unlike other questions about severability, the choice-of-law question has a clear answer in the cases. It is not, however, the answer that one might expect. Historically, the Court held the severability of state statutes in federal court to be a matter of state law, in accordance with Erie's principle of judicial federalism. But in some relatively recent decisions, principally Ayotte v. Planned Parenthood of Northern New England, the Court seems to have reversed course by establishing that severability is a matter of general federal common law. Thus, it appears that federal courts must today apply the Court's federal doctrine regardless of whether a federal or state statute is at issue.

This aspect of the doctrine seems problematic for a couple of reasons. First, it seems problematic as a matter of stare decisis, for it rather inexplicably rejects a practice that had been settled ever since Erie. Second, it seems problematic as a matter of judicial federalism, for it ensures that general federal common law displaces a large body of diverse state law without any discernable constitutional authorization or supporting federal interest. 

I've been writing a piece--available here--that identifies and explains these problems in more detail, and also proposes a solution. Briefly stated, I think that the federalization of severability for state statutes is appropriate only in limited circumstances. Federal courts should generally have to apply state law to decide the severability of state statutes, and should be able to override the applicable state-law doctrine only when applying it would require severance in a manner that exceeds the limits of Article III judicial power. I think this approach would better honor both Erie and Article III, and also harmonize with the vast majority of the Court's modern precedent.