How Should SCOTUS Read one of the Few Provisions of Federal Law that Favors Habeas Petitioners?

Last week, the Supreme Court heard oral argument in Guerrero v. Johnson. The case involves the proper interpretation of a provision of the Antiterrorism and Effective Death Penalty Act (AEDPA), enacted in 1996, that limits the circumstances under which a person can successfully bring a federal court habeas corpus petition on the ground that their state court conviction or sentence was unconstitutional. This area of the law is maddeningly technical, but I shall try to make it accessible to non-specialists without dumbing it down too much.

The relevant statutory provision (28 U.S.C. § 2244) generally forbids a person in state custody from bringing more than one federal habeas petition. Subsection (b)(2)(A) makes an exception if the petitioner brings a claim that wasn't the subject of a previous habeas petition if "the claim relies on a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court, that was previously unavailable." The exception is designed for circumstances in which the Supreme Court issues a decision that was not easy to anticipate, and therefore it would be unfair to deny a habeas petitioner the benefit of the precedent simply because their lawyer did not anticipate it.

One such decision was Atkins v. Virginia, the 2002 ruling that held that persons with intellectual disability are ineligible to be executed. Atkins was undoubtedly a new rule of constitutional law when it was decided. That's because the Supreme Court has generally said that any rule that was not "dictated by precedent" is a new rule. Not only wasn't Atkins dictated by precedent; it contradicted and essentially overturned the 1989 precedent of Penry v. Linaugh.

As a general matter, the new rule cases make it difficult for persons in state custody to successfully bring federal habeas petitions. That's because the Court's decision in Teague v. Lane and another provision of AEDPA that builds on it generally forbid the grant of relief to habeas petitioners who rely on new rules. By precluding habeas for petitions based on new rules and then defining a new rule very broadly, the Court-made law and AEDPA set a high bar for habeas relief.

However, Teague itself recognized an exception to the new-rule ban for cases in which the new rule the habeas petitioner seeks makes the underlying conduct constitutionally protected. For example, if someone had been convicted for "homosexual sodomy" before the Supreme Court's 2003 decision in Lawrence v. Texas, they could rely on Lawrence as the basis for habeas relief because it protected their primary conduct. The Court has also held that the Teague exception applies where the new rule makes a class of persons ineligible for a particular punishment.

It's the Teague exception that 28 U.S.C. § 2244(b)(2)(A) invokes. If a habeas petitioner is relying on a new rule, but it's the kind of new rule that can be invoked because it either protects the underlying conduct or categorically rules out the particular penalty, then a habeas petitioner can bring the claim in a second or successive petition because its newness would have made it previously unavailable. And because Atkins makes a class of people--those with intellectual disability--ineligible for the death penalty, it qualifies.

But here's the rub: Dexter Johnson, the habeas petitioner and respondent in the Supreme Court case argued last week, was convicted and sentenced to death for crimes he committed in 2006, four years after Atkins was decided by the Supreme Court. How then can he argue that Atkins is "a new rule . . . that was previously unavailable?"

The key is that Johnson contends that Atkins itself effectively changed when the American Psychological Association changed its criteria for intellectual disability with the release of the 5th edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V). Whereas previously an IQ of 70 was the cutoff, the new standards were holistic in a way that indicated that Johnson could qualify even though he didn't qualify under the old standard. Because that occurred after Johnson's initial petition, his lawyers now argue that Atkins as filtered through the DSM-V amounted to a new rule and that Johnson's claim therefore was previously unavailable.

During last week's oral argument, the Democratic appointees all expressed sympathy for Johnson's position, each on slightly different grounds. The version that seems most straightforward is Justice Kagan's suggestion that one should think of Atkins as really Atkins I and Atkins II. Because Johnson is relying on Atkins II, she indicated, he fits the exception.

With the exception of Chief Justice Roberts, who uncharacteristically did not ask any questions, the conservative Justices pushed back in various ways. There was some technical back-and-forth about when exactly a new rule (whether formal or de facto) can be announced and still count as unavailable. There was also a suggestion by several Justices that whatever the answer to that question might be, Atkins can't be a new rule because it was on the books when Johnson committed his crime. In this view, the DSM-V change is just like any other change in facts: that might affect how the rule gets applied in particular cases but doesn't make Atkins itself new again.

For my part, I agree with Justice Kagan's framing, although I recognize that there are matters of degree having to do with the extent to which the Atkins rule actually incorporates expert psychiatric opinion. The more one thinks of Atkins as an empty shell, the more one should be inclined to the Kagan view. By contrast, the more one thinks of Atkins itself as the rule and matters like changes in successive versions of the DSM as akin to ever-changing facts of the sort relevant to applying any rule of law, the less one should be inclined to the Kagan view.

I found one suggestion from the conservatives particularly unpersuasive. Justice Thomas asked whether Marbury v. Madison, which was decided in 1803, could count as a new rule under the theory advanced by Johnson's lawyer. This reductio is easy to fend off, however. Marbury is not the kind of case that says anything at all about whether a category of conduct is constitutionally protected or makes a class of persons ineligible for a kind of punishment. Nor does it incorporate an external standard in the way that Atkins does. There are very few cases that have both of those features. In fact, Atkins appears to be the only one. So the suggestion that under Johnson's theory everything old can be new again is way off base.

In the end, the case comes down to how to read § 2244(b)(2)(A). A highly formalistic reading of "new rule" would dispose of Johnson's case on the ground that Atkins isn't new. But what if one attempts to interpret the § 2244(b)(2)(A) exception purposefully? The evident purpose of the exception is to avoid the unfairness of barring a habeas petitioner with a presumptively valid claim simply because they didn't bring it at a time when they could not reasonably have been expected to appreciate that it could succeed.  AEDPA overall closes off most avenues to habeas. It would add insult to injury for the Supreme Court to construe one of its few provisions that aim at fairness to habeas petitioners as yet another closed door.

-- Michael C. Dorf