« Fitbits and the Fourth Amendment | Main | Bad presidents and fictional schools »

Monday, April 18, 2016

The Subtle But Serious Flaw in the Supreme Court's Welch Ruling

This morning, to just about no one's surprise, the Supreme Court held in Welch v. United States that its June 2015 decision in Johnson v. United States--invalidating on vagueness grounds the "residual clause" of the Armed Career Criminal Act--is "substantive," and may therefore be retroactively enforced by federal prisoners even through collateral post-conviction review. In English, even for those prisoners whose direct appeals were already over, Welch allows them to pursue relief based upon Johnson, which could mean reduced sentences (and, potentially, immediate release) for hundreds--if not thousands--of federal prisoners. And to its credit, the Court really hustled--handing down the decision just 19 days after argument, presumably with an eye toward the one-year statute of limitations for Johnson-based claims (which runs on June 26). So far, so good.

But as I've blogged about at some length previously (and explained in a short piece in the latest issue of the Federal Sentencing Reporter), the problem Johnson raised in the lower courts had two distinct layers to it: The obvious one (is Johnson retroactive), and the far more subtle one (how can lower courts tell when new Supreme Court rulings like Johnson are retroactive). Whether because of the compressed schedule, the short-handed Court, or a more general disinclination to reach that issue, Welch says absolutely nothing about this second layer--and, in the process, does nothing to avoid the same mess that Johnson produced in the lower courts from recurring in the future.

In a nutshell, here's the problem: For prisoners who have already had one bite at the post-conviction apple, the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) imposes harsh restrictions on second-or-successive claims for post-conviction relief. Not only must applicants first obtain the permission of the relevant court of appeals, but permission will only be granted, as relevant here, if the applicant can show that he can likely benefit from "a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court, that was previously unavailable." In other words, for second-or-successive petitioners, the new rule doesn't just have to be retroactive; it has to have been "made retroactive" by the Supreme Court.

In Tyler v. Cain, the Supreme Court clarified that only it, and not the lower courts, can "ma[k]e" a new rule retroactive. But the Justices were unclear on how they so provide. For instance, Justice O'Connor's concurring opinion suggested that, if a Supreme Court ruling at Tprovides that all "substantive" new rules are retroactive, then a Supreme Court decision at Tarticulating a new substantive rule is "made retroactive" by the earlier ruling at T0. But Justice Thomas's majority opinion was more rigid, suggesting that the Court has to expressly make the new rule retroactive (presumably after it is handed down), and the lower courts have divided over which is the correct framework. Thus, after Johnson, a majority of circuits held that the Court's pre-Johnson holdings with regard to the retroactivity of substantive rulings necessarily "made" Johnson retroactive; two circuits held that Johnson wasn't substantive in the first place (and therefore wasn't retroactive); and the Tenth Circuit, most significantly, held that, even if Johnson was substantive, the Supreme Court had not "made" it retroactive. In other words, even the courts that agreed Johnson was "substantive" disagreed over whether that meant that the Supreme Court had already "made" it retroactive.

As I've argued before, it would be easy enough for the Supreme Court to resolve this split--and the open question raised by the distinction between the Thomas and O'Connor opinions in Tyler. And the post-Johnson mess was a perfect opportunity to do so, since Johnson is the exact kind of ruling likely to fall right in the middle of the split (that is, a ruling that is almost certainly substantive, but where there's any room for doubt). But rather than clarifying whether the Thomas or O'Connor approach is the correct one for lower courts to follow going forward, all the Court held today was that Johnson is substantive, and can therefore be retroactively enforced even by second-or-successive applicants, full stop.

To be sure, some lower courts might read into Welch the conclusion that all substantive holdings are retroactive, but the Supreme Court had already held exactly that (in Schriro v. Summerlin), and that didn't stop the Tenth Circuit from still requiring a separate ruling that expressly "made" Johnson retroactive. The Supreme Court provided that ruling this morning in Welch, but did nothing to prevent a court like the Tenth Circuit from demanding the same thing the next time the Court hands down a new substantive rule--or to thereby prevent the exact same headaches that Johnson caused from recurring. As a result, Welch will go into the books as a small but important win for federal prisoners, but a gigantic missed opportunity for clarity in future retroactivity cases.

[Full disclosure: I signed onto an amicus brief in support of Welch, albeit on a different issue--whether the Teague retroactivity framework even applies to federal prisoners.]

Posted by Steve Vladeck on April 18, 2016 at 11:49 AM in Steve Vladeck | Permalink


Steve -- wouldn't any comment in Welch about the "made retroactive" issue have been pure dicta, such that it still wouldn't have stopped the Tenth Circuit and other like-minded courts in future cases? Also, trying to resolve that issue might well have fractured the 8-Justice majority, and thus slowed down the issuance of the opinion.

Posted by: Hash | Apr 18, 2016 12:34:24 PM

Hi Hash! Obviously, I agree with you about _why_ the Court punted, as the post suggests. But I'm not sure it would have mattered if clarification of Tyler came as dicta. After all, the AEDPA question the lower courts have to ask is what it means for the Supreme Court to "ma[k]e" a new rule retroactive. Even in dicta, the views of a majority of the Justices on the _answer_ to that question would, I have to think, be rather persuasive to courts like the Tenth Circuit. And unlike in contexts like Williams v. Taylor, the fact that such a conclusion came in dicta would not itself bar relief, since it's not a "clearly established law" issue.

Posted by: Steve Vladeck | Apr 18, 2016 1:28:56 PM

Well, Thomas's opinion in Tyler was for a *majority* of the Court, and even that didn't seem to stop most CAs. After all, he was pretty clear that the test for whether prior holdings have already "made" a new rule retroactive is whether the prior holdings "necessarily dictate" or "logically dictate" retroactivity of the new rule. So here, since the court has held that all substantive rules are retroactive, the question is whether the prior cases "dictate" the conclusion that Johnson is substantive. And it's pretty hard to see how they do, since the prior formulations of "substantive" rules didn't quite reach the constitutional invalidation of a sentencing statute on vagueness grounds. While I think the majority's position in Welch is by far the better conclusion for *extending* the category of substantive rules, I don't think one can say that the dissent's position was "necessarily" or "logically" foreclosed by precedent, even if it admittedly adopted a fairly artificial distinction of Bousley.

So, looping back to the dicta point, the only reason that dicta here would have more of an effect than Tyler itself is because it is dicta by the *current* court, which they presumably would adhere to. But, of course, by the time that the Court next creates a new rule whose "substantive" nature is at least arguably not dictated* by precedent, it probably won't be the same court any more, so any dicta in Welch would be no more persuasive to future courts than the majority opinion in Tyler.

Posted by: Hash | Apr 18, 2016 7:35:45 PM

Welch wasn't a second or successive petition, so I struggle to see how dicta about retroactivity and the standard for second and successive relief could have been worked into the opinion. The whole point of picking Welch as the vehicle instead of one of the original petitions, I think, was that Schiro et al. didn't really logically dictate retroactivity here, so the Court had to make Johnson retroactive in a first petition.

Posted by: Asher Steinberg | Apr 18, 2016 8:07:22 PM

Hash -- I'm not sure your characterization of Tyler is accurate; I think a number of lower courts (and commentators) find five votes between O'Connor's concurrence and the dissent, and so it seems to me there's more than just the identity of the Justices behind why dicta in a 7-1 or 6-2 (or even 5-3) ruling might have more persuasive force on the lower courts than whichever view of Tyler is the correct one.

Asher -- Perhaps this goes back to my earlier objection to the Court taking Welch over the other vehicles (http://prawfsblawg.blogs.com/prawfsblawg/2016/01/vehicle-problems-vs-unusual-vehicles-the-supreme-courts-bizarre-cert-grant-in-welch.html). But I still think the Justices were well aware of the second circuit split here, and could have at least dropped a footnote _after_ holding that Johnson is retroactive to the effect of "In addition, because it is, as we hold today, a new rule of substantive constitutional law, Johnson had been "made retroactive" by our prior decisions holding that such rules are per se retroactive." That might've been dicta, but it sure would've been useful dicta!

Posted by: Steve Vladeck | Apr 19, 2016 9:53:16 AM

Fair enough, Steve, though it seems to me the contrary characterization of the majority opinion in Tyler is wishful thinking, given the following unambiguous passage:

Justice Breyer observes that this Court can make a rule retroactive over the course of two cases. See post, at 672— 673 (dissenting opinion). We do not disagree that, with the right combination of holdings, the Court could do this. But even so, the Court has not made Cage retroactive. Multiple cases can render a new rule retroactive *only if the holdings in those cases necessarily dictate retroactivity of the new rule.*

If Justice O'Connor didn't agree with the "necessarily dictate" standard, she shouldn't have joined this paragraph of the *majority* opinion.

Thus, while substantive rules are "per se retroactive", there was no case *dictating* that Johnson was *necessarily* "substantive."

Posted by: Hash | Apr 19, 2016 4:26:38 PM

Hash -- Thanks for this exchange; let me just push back on one more thread and see if it helps to illuminate either our differences or our agreement:

I agree with you 100% that, "while substantive rules are 'per se retroactive', there was no case *dictating* that Johnson was *necessarily* 'substantive.'"

But if I'm a lower-court judge and I conclude that Johnson _is_ substantive, doesn't it follow at _that_ point that it has been "made retroactive" by those Supreme Court cases holding that substantive rules are per se retroactive? In other words, Thomas says the Supreme Court has to "necessarily dictate" the _retroactivity_ of the new rule, not the _substantiveness_ of the new rule.

My beef is not with the Fifth and Eleventh Circuits (although, as Welch shows, I think they got the substantive/procedural question wrong); it's with the Tenth Circuit for holding that, even if Johnson _is_ substantive, it hasn't been "made retroactive." In my view, that analysis is not at all compelled by Thomas's Tyler opinion; is arguably inconsistent with O'Connor's concurrence; and is not remotely affected by Welch.

Do you agree with that, or is your view that, insofar as the new rule has to be "made retroactive" by the Supreme Court, that also means that it has to be "made substantive"?

Posted by: Steve Vladeck | Apr 19, 2016 6:01:54 PM

We do disagree. My view is that a rule has been "made retroactive" *by the Supreme court* only if *the supreme court itself* has rendered *all* of the requisite holdings *dictating* the conclusion that the ruling is retroactive.

That includes the holding dictating the conclusion that the rule is substantive. And for good reason: the sct could *disagree* with the lower court about whether the new rule is "substantive," and given that possibility it doesn't make sense to say that the sct has already "made retroactive" the new rule.

Posted by: Hash | Apr 19, 2016 9:07:03 PM

To be less abstract: you think ca10 should have granted permission to file a successive petition. But, if sct had reached the opposite result in Welch, such a petition would obviously be improper. Which proves that, until Welch, *the sct* had not "made retroactive" Johnson.

Posted by: Hash | Apr 19, 2016 9:11:22 PM

Two final thoughts:

1. On your view, it's not even clear why the sct has to have itself recognized the right at all. If a lower court can provide the critical holding that the rule is substantive, why can't it also provide the critical holding about the rule -- and then simply declare that the new rule it has recognized is substantive and thus purportedly has been "made retroactive" by sct.

2. Conversely, why can't a lower court supply the critical holding that a new procedural rule is watershed (since sct has held all watershed rules are retroactive)?

Bottom line: if it's still reasonably debatable whether a new rule is substantive, then the sct has not itself yet "made retroactive" that rule.

Posted by: Hash | Apr 19, 2016 9:31:16 PM

Post a comment