Of New Rules and New Applications of Old Rules
by Michael C. Dorf
My latest Verdict column discusses last week's SCOTUS opinion in Cruz v. Arizona. In a 5-4 opinion for the Court by Justice Sotomayor (joined by the other Democratic appointees plus CJ Roberts and Justice Kavanaugh), the Court held that it had the authority to review and then reversed the Arizona Supreme Court's rejection of John Montenegro Cruz's challenge to his death sentence. The trial judge had failed to inform the sentencing jury that rejection of a death sentence would result in life imprisonment without parole (LWOP). That clearly violated Simmons v. South Carolina. However, the Arizona Supreme Court held that Cruz could not bring a second state post-conviction claim because his case didn't fit a state law exception for when there "has been a significant change in the law that, if applicable to the defendant’s case, would probably overturn the defendant’s judgment or sentence.”
That was quite dubious because Cruz pointed to Lynch v. Arizona--a SCOTUS ruling after Cruz's sentencing that held that the courts of Arizona had no sound basis for their failure to apply Simmons. Not good enough, said the Arizona Supreme Court: Lynch was a change in the application of pre-existing law, not a change in the law.
Justice Sotomayor's opinion says that the distinction between changes in the law and changes in the applications of pre-existing law was inconsistent with the Arizona courts' prior cases. In light of this inconsistency, the state court construction of the procedural rule governing the availability of a second or successive postconviction petition was not an adequate basis for defeating Supreme Court review of the underlying federal question. As I explain in the column, that's fair, but I would have preferred a holding that the distinction between changes in the law and changes in applications of the law would be an unreasonable basis for denying federal rights, even if the Arizona courts applied it consistently.
The column lays out the foregoing argument in greater detail and with more background. Here I want to pivot to discuss the distinction the Arizona Supreme Court drew in wider context. As I'll explain, one can point to a sensible conceptual difference between changes in the law and changes in the law's application, but it's hard to see why this distinction should make a practical difference in most settings.
Let's start conceptually. If the legislature enacts or amends a statute, that's a change in the law; if a court holds that a statute that formerly was understood to mean X in circumstance A actually will now mean Y in circumstance A, that's a change in the application of the law. So far so good.
But note that even in the foregoing abstract example, we might also say that the court's ruling--changing the application from X to Y--could also be characterized as a change in the law itself, not just in the application of the law. And in most circumstances it would be sensible to regard the change that way. After all, generally when we have reason to care about changes in the law it's because we want to provide some protection for reliance interests. If a legislature repealed a law that said X and replaced it with a law that said Y, that would be a change in the law. So likewise, if a court changes its interpretation of some law from X to Y, that should also count as a change in the law. If we think people were entitled to rely on the law being X, that's true whether it was X in virtue of a statute or in virtue of a court decision about what some statute means.
To be sure, there are some circumstances in which it might make a difference whether some operative legal rule appears in a statute rather than only in a judicial decision. But notably, that's not generally true for purposes of protecting reliance interests. There, the question is simply whether people and organizations were entitled to rely on the old rule. Of course, one could stipulate that reliance on statutes but not interpretations or applications of statutes is reasonable (or vice-versa), but that's simply a way of leveraging the inevitable circularity that always attends efforts to protect reliance interests. There's nothing inherent in the difference between legislated changes and judge-made changes that makes the former changes in law but the latter mere changes in the application of the law.
That's all pretty abstract. Let me give a concrete example. Federal habeas corpus courts give substantial deference to state court findings of fact. In Teague v. Lane, the Supreme Court held (originally in a plurality opinion, later endorsed in majority opinions) that so-called "new rules" of constitutional law would not apply retroactively on habeas; if a state court applied the pre-existing precedents, prisoners wouldn't be freed based on the state court's failure to anticipate precedents that break new ground. In Wright v. West, the Court held that notwithstanding deferential review with regard to facts and what amounted to deferential review as to law via Teague, state court applications of law to fact would continue to be reviewed de novo per the earlier holding of Brown v. Allen. The resulting regime was bizarre: deference to state court factual determinations; deference to state court legal determinations; but in between, no deference to applications of law to fact.
The Justices in the majority in Wright v. West pushed back by saying that Teague did not actually adopt a principle of deference on legal questions, only a principle of non-retroactivity. Once one determines which law applies--old or new--there is no deference in its application. Put in terms of the current topic, what Teague barred was federal habeas relief where the petitioner sought a new rule but not where the petitioner merely sought the application of an old rule.
But with due respect, that makes little sense in terms of the policies Teague aimed to vindicate--namely, not to penalize state courts for making good faith efforts to apply the law as it existed at the time of the defendant's trial, sentencing, and appeal. A subsequent federal court grant of habeas can be equally surprising and thus upsetting of state court reliance on prior precedent if it applies an old rule in a surprising way or if it applies a new rule. Accordingly, when Congress more or less codified Teague in a statute, it obligated deference across the board, thus overruling Wright via statute.
I should be clear that I disagree with just how deferential towards state courts federal habeas courts are supposed to be. The threshold for what counts as an "old" rule under Teague (a rule of law is old only if it is "dictated by precedent") is, in my view, too high; likewise, the deference required by statute and court decisions interpreting the statutory provisions allow for too many uncorrected constitutional violations, in my view. But whatever quantum of deference one thinks appropriate, that should not vary based on the ostensible distinction between new rules and new applications of old rules.
As I hope my critique of both Wright v. West and the Arizona Supreme Court in Cruz illustrates, my objection to the distinction between new rules and new applications of old rules (or between changes in the law and changes in the application of the law) is non-partisan. In Wright, SCOTUS deployed the distinction to benefit a habeas petitioner, a decision that codes as liberal. In Cruz, the AZ Supreme Court deployed the distinction to disadvantage a capital defendant, a decision that codes as conservative. Both uses of the distinction are wrong.
Finally, this exercise should cast some doubt on a point I mostly assumed earlier in this essay--that the distinction between new rules and new applications of old rules makes even conceptual sense. It does if what we mean is that we can distinguish when different institutional actors are involved. If only the legislature can create rules, whereas only the courts can apply them, then we have a ready distinction--but we don't need the new-rule/new-application distinction then; we can simply rely on the institutional distinction.
Where there is no institutional difference, any distinction between a new rule and a new application of an old rule will be stipulative. Perhaps one might say that a certain threshold must be crossed before one says we have moved from the territory of new applications to that of new rules. But that threshold will be arbitrary. If one draws the distinction at all, it should be with regard to the values meant to be served by distinguishing old from new in any particular context.