The Circuit Split on Johnson Retroactivity

By Leah Litman
Aug 14, 2015

InJohnson v. United States, the Supreme Court held that the “residual clause” of the Armed Career Criminal Act (ACCA) is unconstitutionally vague. Defendants therefore can no longer be sentenced under the residual clause to a 15-year mandatory minimum term of imprisonment. But what about defendants who have already been sentenced under ACCA’s residual clause? I predicted in April, before Johnson was decided, that determining who can be resentenced in light of Johnson would be fraught with difficulties. The courts of appeals have begun to sort through this question, and I’ll highlight one such case in this post.

In In re Rivero, the Eleventh Circuit purported to decide an important question that affects who can be resentenced in light of Johnson—namely, whether the Supreme Court has made Johnson retroactive. Generally, new rules of constitutional law do not apply to convictions that have become final. But certain “retroactive” rules apply to convictions that have become final; prisoners can raise claims that are based on retroactive rules in post-conviction review—review that occurs after a defendant’s conviction has become final. If a prisoner has already filed one petition for post-conviction review, he may file a second or successive petition for post-conviction review only if the Supreme Court has made a rule retroactive (as opposed to a court of appeals or district court doing so).

I said that the Eleventh Circuit “purported” to decide whether the Supreme Court has made Johnson retroactive because the Eleventh Circuit’s decision is a bit quirky. Most importantly, the defendant wasn’t actually sentenced under ACCA—he was sentenced under an analogous provision of the Sentencing Guidelines (the “career-offender Guideline”). But the Eleventh Circuit “assumed” that Johnson applied to the career-offender Guideline and that the career-offender Guideline was therefore unconstitutional. Working off that assumption, the Eleventh Circuit went out of its way to disagree with the Seventh Circuit on whether the Supreme Court has made Johnson retroactive.

Rivero has thus created a potentially unnecessary circuit split, as well as some uncertainty about who can be resentenced in light of Johnson. I’ll offer some thoughts on how narrowly or broadly Rivero can be read. (Spoiler: I think it should be read pretty narrowly.) But understanding Rivero requires some background information about how ACCA and the Guidelines work; what Johnson decided; and the mechanics of post-conviction review.

ACCA, Johnson, And The Guidelines

Title 18 section 922(g) makes it unlawful for any person convicted of a felony—meaning a crime punishable by more than one year imprisonment—to possess a firearm. Section 924(a) provides that defendants convicted under §922(g) shall be “imprisoned not more than 10 years.” But ACCA—§924(e)—imposes a 15-year mandatory minimum sentence for defendants who are convicted under §922(g) and who have three or more convictions for “violent felonies.” And, prior to Johnson, ACCA defined a violent felony to include “any crime … that … otherwise involves conduct that presents a serious potential risk of physical injury to another.” (This language was referred to as the “residual clause.”)

Johnson held that the residual clause was unconstitutionally vague. Johnson reasoned: “[T]he indeterminacy of the wide-ranging inquiry required by the residual clause both denies fair notice to defendants and invites arbitrary enforcement by judges.” But Johnson made sure to say that the decision did not call into question the myriad laws that use terms like “substantial risk,” most of which “require gauging the riskiness of conduct in which an individual defendant engages on a particular occasion,” rather than the riskiness of a category of crimes.

There is, however, at least one provision that is almost entirely of a piece with ACCA’s residual clause—the career-offender Guideline of the Federal Sentencing Guidelines. The Federal Sentencing Guidelines recommend sentencing ranges based on defendants’ conduct and characteristics. Federal judges first calculate the sentencing range recommended by the Guidelines, and then choose a sentence to impose. Judges are not, however, required to impose a sentence within the recommended guidelines range.

Section 4B1.1 of the Guidelines increases a defendant’s recommended sentencing range if the defendant is a “career offender.” The Guideline defines a career offender as someone who, among other things, has “at least two prior felony convictions of … a crime of violence.” And section 4B1.2 of the Guidelines defines a “crime of violence” using precisely the same language as ACCA’s residual clause: A “crime of violence” includes any offense punishable by more than one year that “involves conduct that presents a serious potential risk of physical injury to another.” The career-offender Guideline was promulgated to implement ACCA, and courts of appeals have interpreted the career-offender Guideline to require the same “wide-ranging inquiry” the Supreme Court had used to interpret ACCA’s residual clause (and that it found unconstitutionally vague).

In re Rivero: The Guidelines

That brings us to Rivero. Rivero was convicted of conspiring to possess with intent to distribute cocaine and using a firearm in furtherance of a drug trafficking offense. (He was also originally convicted of being a felon in possession of a firearm—the predicate offense for the ACCA mandatory minimum—but the Eleventh circuit had previously vacated his §922(g) conviction.) Based on Rivero’s prior convictions, the district court sentenced him using the career-offender Guideline (in particular, using the “serious potential risk” provision of the career-offender Guideline).

It therefore might be the case that Rivero simply doesn’t have a Johnson claim because the career-offender Guideline remains valid after Johnson. (I should note that the Sentencing Commission is currently deciding whether to retain or revise the career-offender Guideline in light of Johnson.) I’m not sure whether a vagueness challenge to the career-offender Guideline should succeed. There are good arguments for such a challenge. A defendant’s calculated Guidelines range is both important and sticky: District courts must first accurately calculate the Guidelines range—their failure to do so is reversible error—and, perhaps for this reason, last year 46% of sentences fell within the recommended Guidelines range. And a recent Supreme Court case about the Ex Post Fact Clause recognized these features of the Guidelines.

But the vagueness/due process analysis is different for the Guidelines than for ACCA. Although Johnson decided that the language in the career-offender Guideline was unconstitutionally vague, courts aren’t required to sentence a defendant within the Guidelines range. The Guidelines are advisory—they’re recommendations, and that make the constitutional calculus somewhat different. Even if the career-offender Guideline language is so vague that courts can’t meaningfully distinguish between prior convictions that count as violent felonies and prior convictions that do not, their inability to do so is of less consequence for defendants sentenced under the Guidelines than for defendants sentenced under ACCA. Under ACCA, courts were required to sentence defendants to more time if the defendant had three or more prior convictions for violent felonies. Under the Guidelines, it was merely recommended that they do so. (It is because the Guidelines are advisory that they do not count as substantive elements of a criminal offense for purposes of the Sixth Amendment.)

So a vagueness/due process analysis of the career-offender Guideline would, in most cases, have to take account of the advisory nature of the Guidelines. But it turns out, the Guidelines weren’t actually advisory for Rivero. The Guidelines haven’t always been advisory: Before the Supreme Court’s decision in United States v. Booker, the Guidelines were “mandatory”—courts were required to impose sentences within the Guidelines range, and Rivero was sentenced when the Guidelines were mandatory. This makes Rivero look a lot more like the prisoners who were sentenced under ACCA than defendants who were sentenced under the advisory Guidelines. But there are still differences between mandatory Guidelines and ACCA. For example, even when the Guidelines were mandatory, the Guidelines authorized courts to reduce a defendant’s recommended sentencing range if the court determined the defendant’s criminal history “substantially over-represent[ed] the seriousness of the defendant’s criminal history or the likelihood that the defendant will commit other crimes.” That is, courts could depart from the career-offender Guidelines “mandatory” sentencing range, but they can’t do the same with respect to ACCA’s mandatory minimum. (As a further aside, the Eleventh Circuit, on appeal, rejected Rivero’s argument that the district court erred by not departing downward from the Guidelines range under this provision. The court also affirmed Rivero’s conviction after Booker was decided, but didn’t remand for resentencing in light of Booker.)

Whether Johnson applies to the Guidelines—and particularly the mandatory Guidelines—is, I think, an interesting and difficult question. But here’s the rub: The Eleventh Circuit (or at least the majority opinion) didn’t really discuss any of this complexity. It merely noted the defendant was sentenced under the “mandatory” Sentencing Guidelines, and it “assume[d] that … Johnson applies to the residual clause of section 4B1.2(a)(2) of the Sentencing Guidelines.”

What The Eleventh Circuit Decided In Rivero

In re Rivero instead chose to address other issues that affect who may be resentenced in light of Johnson—specifically, whether Johnson is “retroactive” and whether the Supreme Court has “made” it retroactive. Generally, “new” constitutional rules do not apply to cases that have already become final. A conviction becomes final when a prisoner has exhausted his direct appeal in the federal court of appeals and the Supreme Court has denied a petition for certiorari or the time to file a petition for certiorari has expired. New rules apply “retroactively” to cases that have already become final if the new rule is “substantive” or announces a “watershed” rule of criminal procedure.

After a prisoner’s conviction has become final, a prisoner can raise claims and challenge his conviction in a post-conviction proceeding. But the Anti-terrorism and Effective Death Penalty Act (AEDPA) limits prisoners’ ability to file petitions for post-conviction relief. In particular, title 28 section 2255(h)(2) permits prisoners to file second or successive petitions under very limited circumstances, including where the petition contains “a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court, that was previously unavailable.”

Relying on this provision, Rivero sought to bring a second or successive post-conviction petition based on Johnson. And operating under the assumption that “Johnson applies to … the Sentencing Guidelines,” the Eleventh Circuit went on to decide whether the Supreme Court had made Johnson retroactive.

The Eleventh Circuit first held that the “new rule announced in Johnson is substantive rather than procedural”—that is, Johnson announced a new rule of constitutional law that can be raised in post-conviction proceedings. I’m on board so far. Second, the Eleventh Circuit held that the Supreme Court had not “made” the decision in Johnson retroactive through its prior cases. Of note, the Eleventh Circuit reasoned that the Supreme Court’s prior retroactivity decisions did not make Johnson retroactive because “the rule announced in Johnson does not meet the criteria the Supreme Court uses to determine whether the retroactivity exception for new substantive rules applies.”

I think it’s a hard question whether the Supreme Court has “made” Johnson retroactive. The answer turns in part on how you read the Court’s decision in Tyler v. Cain. Tyler said that the Court can “make” a rule retroactive only through “holdings,” rather than dicta, but Justice O’Connor’s concurrence, which was supported by the four dissenting Justices, said that the Court could “make” a rule retroactive “through multiple holdings that logically dictate the retroactivity of the new rule.” And, using this logic, the Seventh Circuit (in a case actually involving ACCA) held that other Supreme Court cases logically dictated that Johnson was retroactive. Judge Jill Pryor would have done the same in Rivero.

I’m less convinced by the Eleventh Circuit’s suggestion that a new rule can be “substantive”—as the Eleventh Circuit held Johnson was—but at the same time not satisfy “the retroactivity exception for new substantive rules.” These are two sides of the same coin: A rule is retroactively applicable in post-conviction proceedings if it’s substantive—that is the retroactivity exception for new substantive rules. I don’t think there’s any conceptual space between the two. The confusion may stem from the fact that the Supreme Court has used different language to describe “substantive” rules—sometimes it has said a rule is “substantive” if it “narrows the scope of a criminal statute”; other times it has said a rule is “substantive” because it “places particular conduct or persons covered by a statute beyond the State’s power to punish.” My reading of these cases is that a rule is substantive either if it narrows the scope of a criminal statute, or if it says that a State can’t punish certain conduct or persons. If that’s right, Johnson is clearly retroactive because the decision narrowed the scope of a criminal statute—the ACCA enhancement no longer applies to defendants with three or more convictions for “crimes that otherwise involve conduct that presents a serious risk of potential injury to another.”

How To Read Rivero

So what to make of the Eleventh Circuit’s decision in Rivero? I think subsequent courts will have some latitude in deciding what Rivero actually held. (And prisoners raising Johnson claims will have some latitude in describing what Rivero held.) For example, did Rivero hold that Johnson is retroactively applicable in post-conviction proceedings? On the one hand, Rivero said that Johnson announced a new substantive rule, quoting a prior Eleventh Circuit case about retroactivity. On the other hand, the Eleventh Circuit said that Johnson did not fall within the retroactivity exception for new substantive rules. But, it said this in the course of explaining why the Supreme Court had not “made” Johnson retroactive for purposes of §2255(h)(2)—that is, for reasons that only apply to second or successive petitions for post-conviction review.

More importantly, for whom did Rivero say that Johnson isn’t retroactive? (Or, more precisely, for whom did Rivero say that the Supreme Court had not “made” Johnson retroactive?) I think there’s a pretty good argument that Rivero only held that the Supreme Court had not made Johnson retroactive as to defendants sentenced under the career-offender Guideline. The Eleventh Circuit reasoned that the dissent’s “assumption”—which it also made—that “Johnson also applies to the residual clause of the career offender” Guideline “makes clear that precedents of the Supreme Court do not necessarily dictate” that Johnson is retroactive. The Supreme Court could not have made Johnson retroactive to Rivero’s case, the Eleventh Circuit reasoned, because “[t]he Supreme Court has never held that the Sentencing Guidelines are subject to a vagueness challenge. And four of our sister circuits have held that the Sentencing Guidelines—whether mandatory or advisory—cannot be unconstitutionally vague.” In other words, the Supreme Court had not “made” Johnson retroactive to cases involving the career-offender Guidelines because it had never addressed a vagueness challenge to the Guidelines.

This is a pretty limited reading of Rivero, and it’s probably informed by my view that Johnson is retroactive (because it satisfies the retroactivity exception for substantive rules). I also tend to think there are avenues for prisoners to raise Johnson claims in second or successive petitions—whether that’s because the Court has “made” Johnson retroactive, or because the “savings clause” allows prisoners to raise Johnson claims under §2241 if they can’t bring a second or successive petition.

At bottom, I think the Eleventh Circuit probably shouldn’t have decided questions about Johnson’s retroactivity without first addressing whether it was even necessary to do so—that is, without first deciding whether Johnson even applied to defendants sentenced under the career-offender Guideline. That is all the more true given that the Eleventh Circuit seemed to be of the view that it is an easy call that Johnson does not apply to the career-offender Guideline—the court noted that the Supreme Court has never entertained a vagueness challenge to the Guidelines, and that four courts of appeals held that Guidelines, even mandatory ones, could never be unconstitutionally vague. (I can’t help but note that the Eleventh Circuit’s reasoning for why the Supreme Court has not made Johnson retroactive—which assumes that Guidelines aren’t amenable to vagueness challenges—is in tension with its threshold assumption that Johnson applies to the Guidelines. And whether the Guidelines—especially mandatory versions—are amenable to vagueness challenges is, in my opinion, a more difficult question after Booker and Johnson.) But because the Eleventh Circuit didn’t address whether Johnson applies to the career-offender Guideline, we’re left with a potentially unnecessary circuit split and a fair amount of uncertainty about who, exactly, the Eleventh Circuit has said cannot be resentenced in light of Johnson.

I’ll close with one other unhappy note. AEDPA greatly circumscribes the Supreme Court’s ability to review the Eleventh Circuit’s determinations. Under title 28 section 2244(b)(3)(E), “the grant or denial of an authorization by a court of appeals to file a second or successive application shall not be appealable and shall not be the subject of a petition for rehearing or for a writ of certiorari.” In Felker v. Turpin, the Supreme Court upheld this provision against a constitutional challenge, noting that §2244 did not repeal the Court’s jurisdiction to entertain “original” petitions for habeas corpus—petitions that are filed directly in the Supreme Court (under §2241). An “original” petition in Rivero would make the same argument that was made in the court of appeals, but the Court wouldn’t technically be reviewing the judgment of the court of appeals.

But “original” habeas petitions aren’t a full substitute for petitions for certiorari. “Original” habeas petitions are rarely granted. Felker didn’t decide whether the statutory limitations on courts of appeals’ (and district court’s) authority to grant habeas petitions also apply to the Supreme Court’s authority to grant original habeas petitions under §2241. (The prisoner in Felker had been convicted in state court; therefore the relevant statutory limitations were those contained in §2254 rather than §2255. And §2255(e) at least seems to contemplate that a court would sometimes be able to grant a writ under §2241 even though it could not do so under §2255.) But Felker did note that, under the Supreme Court’s own rules, original petitions are rarely granted—they are granted only in “exceptional circumstances.” I’m not sure what those exceptional circumstances would be, but getting the Court to hear an original petition in Rivero may be an insurmountable barrier.