*1 Before ED CARNES, Chief Judge, TJOFLAT, MARCUS, WILSON, WILLIAM PRYOR, MARTIN, JORDAN, ROSENBAUM, JILL PRYOR, NEWSOM, BRANCH, and HULL, [*] Circuit Judges.
NEWSOM, Circuit Judge:
The question before us is whether one of the key provisions of an important federal criminal statute, 18 U.S.C. § 924(c), is unconstitutionally vague. As relevant to our purposes, § 924(c) makes it a federal offense—punishable by a term of imprisonment ranging from five years to life—for any person to use, carry, or possess a firearm in connection with a “crime of violence.” 18 U.S.C.
§ 924(c)(1)(A). The provision challenged here—§ 924(c)(3)’s “residual clause”— defines the term “crime of violence” to mean a felony “that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.” Id . § 924(c)(3)(B).
This case is in some respects a successor to
Johnson v. United States
, 135 S.
Ct. 2551 (2015), and
Sessions v. Dimaya
,
2
matter and whether, in the abstract, “the kind of conduct that the crime involves in
the ordinary case” meets the statutory standard.
Johnson
,
On the flip side,
Johnson
and
Dimaya
also make clear—and it is common
ground here—that if § 924(c)(3)’s residual clause is instead interpreted to
incorporate what we’ll call a “conduct-based approach” to the crime-of-violence
determination, then the provision is
not
unconstitutionally vague. As its name
suggests, the conduct-based approach, in stark contrast to the categorical, focuses
not on formal legal definitions and hypothetical “ordinary case[s],” but rather on
the real-world facts of the defendant’s offense—
i.e.
, how the defendant actually
went about committing the crime in question. And as the Supreme Court
emphasized in
Johnson
—and then reiterated in
Dimaya
—there is no reason to
“doubt the constitutionality of laws that call for the application of a qualitative
standard such as ‘substantial risk’ to real-world conduct.”
Johnson
, 135 S. Ct. at
2561 (quoted in
Dimaya
,
The obvious (and decisive) question, then: Which is it here—categorical or
3
conduct-based? Because we find ourselves at this fork in the interpretive road—
the categorical approach imperiling § 924(c)(3)’s residual clause, a conduct-based
reading saving it—we invoke the canon of “constitutional doubt.” Pursuant to that
“elementary rule,” the Supreme Court has long held, “every reasonable
construction must be resorted to in order to save a statute from
unconstitutionality.”
Hooper v. California
,
Accordingly, we hold that § 924(c)(3)(B) prescribes a conduct-based approach, pursuant to which the crime-of-violence determination should be made by reference to the actual facts and circumstances underlying a defendant’s offense. To the extent that our earlier decision in United States v. McGuire , 706 F.3d 1333 (11th Cir. 2013), holds otherwise, it is overruled.
4
I A Under 18 U.S.C. § 924(c), “any person who, during and in relation to any crime of violence . . . uses or carries a firearm, or who, in furtherance of any such crime, possesses a firearm” is guilty of a federal offense and subject to a prison term ranging between five years and life. 18 U.S.C. § 924(c)(1)(A). Section 924(c) defines the term “crime of violence” as “an offense that is a felony” and—
(A) has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or (B) that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
Id. § 924(c)(3). For ease of reference—and as a way of facilitating comparisons with other similar statutes—we’ll call Subsection (3)(A) the “elements clause” and Subsection (3)(B) the “residual clause.” [1]
Importantly here, this Court held in
United States v. McGuire
that the
question whether a predicate offense qualifies as a “crime of violence” under either
subsection is one that a court “must answer ‘categorically’—that is, by reference to
the elements of the offense, and not the actual facts of [the defendant’s] conduct.”
*6
B
In 2010, Irma Ovalles was charged by information with six robbery- and carjacking-related offenses, all of which arose out of what can only be described as a three-day crime binge. As particularly relevant here, Ovalles was charged with (1) attempted carjacking in violation of 18 U.S.C. § 2119 and (2) using and carrying a firearm during a “crime of violence”—the attempted carjacking—in violation of 18 U.S.C. § 924(c)(1)(A). Ovalles entered into a written plea agreement in which she admitted that “she [was] in fact guilty” on all six counts. At her plea hearing, the government outlined the elements of each crime, and Ovalles explained that she understood what the government would have to prove should she opt to go to trial.
The government then made a comprehensive factual proffer detailing Ovalles’s involvement in the crimes. In general, the proffer summarized the evidence demonstrating that Ovalles and her co-conspirators (1) robbed a grocery store while armed with baseball bats, then (2) still wielding the bats, carjacked a Dodge Ram, then (3) carjacked a Toyota 4-Runner, pistol-whipping its owner, then (4) attempted to carjack a Chevy Venture—more on this one below—and finally (5) carjacked a Ford F-150 at gunpoint. More specifically, concerning the attempted carjacking of the Chevy Venture—during which one of Ovalles’s
6
accomplices fired an AK-47, and which therefore serves as the predicate offense for Ovalles’s § 924(c) conviction—the government’s proffer explained as follows:
They see a family getting out of a Chevy Venture in Clayton County, Georgia, and as the family is getting out of their car, these two defendants along with their co-conspirators go up to the family and demand the keys to the car and demand the car. Now, they have a baseball bat and guns with them. There’s a juvenile, a 13-year-old female, who is part of that family group of victims. They hit that juvenile in the mouth with a baseball bat. The damage to her I am sure will be addressed at sentencing. It was not—she did not go to the hospital. Let me say that.
They are demanding the keys. Somebody comes out of the apartment complex where this is happening and that person has a gun. He then confronts the assailants …. They flee, not taking the Chevy Venture, which is why it is an attempted carjacking.
The government would show that the Chevy Venture traveled in interstate commerce, that it was not made in the State of Georgia. The government would prove it was these defendants not only through their confessions as to this event, also through the victims’ testimony. They do I.D. the defendants in this particular case.
On the way out of the apartment complex … co-conspirator … Jerry Arriaga … has an AK-47 style assault rifle and he then discharges that gun several times towards the victim family, the guy who came to rescue them, who was armed, and the car, and that is the basis of Count Five, the use of a firearm during and in relation to a crime of violence.
With respect to each of the charges—including, as relevant here, the attempted-carjacking and § 924(c) counts—Ovalles stated that she had no “material disagreement with what the government sa[id] it could prove . . . .” Having heard the government’s summary of the evidence against her, Ovalles
7
pleaded guilty to each of the charged offenses, acknowledged that her pleas were voluntary, and explained that she was so pleading because she was “in fact guilty as charged in the criminal information.”
The district court thereafter sentenced Ovalles to serve 120 months in prison on the § 924(c) count—which, per the statute, the court imposed to run consecutively to the concurrent 108-month terms on the remaining charges. Ovalles did not object to her sentences, nor did she file a direct appeal.
C
Several years later, Ovalles filed a motion for relief under 28 U.S.C. § 2255
contending that her § 924(c) conviction and sentence were unconstitutional in light
of the Supreme Court’s intervening decision in
Johnson v. United States
, 135 S. Ct.
2551 (2015). In short, the Court in
Johnson
invalidated as unconstitutionally
vague the Armed Career Criminal Act’s residual clause—which, for purposes of
applying that statute’s recidivism-based sentence enhancement, defines the term
“violent felony” to include any crime that is punishable by a year in prison and that
“involves conduct that presents a serious potential risk of physical injury to
another.” 18 U.S.C. § 924(e)(2)(B)(ii). The Court voided the ACCA’s residual
clause principally based on its conclusion that the provision necessitated the
categorical approach to determining whether an underlying conviction constitutes a
“violent felony.”
See Johnson
,
8
that because § 924(c)(3)’s residual clause is “nearly identical to” the ACCA’s, Johnson ’s reasoning rendered it unconstitutional, as well. Accordingly, she argued, she was “no longer guilty of violating” § 924(c) because her predicate attempted-carjacking offense “no longer qualifie[d] as a crime of violence under § 924(c)(3)(B).”
The district court denied Ovalles’s § 2255 motion, reasoning that § 924(c)(3)’s residual clause did “not suffer from the same unpredictability” as the ACCA’s. The court subsequently granted Ovalles a certificate of appealability on the question whether § 924(c)(3)(B) is unconstitutionally vague under Johnson .
A panel of this Court affirmed the district court’s decision. For our
purposes, the panel’s opinion did two significant things. First, in accordance with
(and citing to) our earlier decision in
McGuire
, it held that the question whether
Ovalles’s attempting-carjacking offense constitutes a “crime of violence” within
the meaning of § 924(c)(3) had to be answered using the categorical approach.
See Ovalles v. United States
,
9
(1) that § 924(c)(3)’s residual clause refers not (as does the ACCA’s) to the risk of “physical injury,” but rather to the risk of “physical force,” which it thought was more precise; (2) that § 924(c)(3)(B)’s inclusion of the qualifying phrase “in the course of committing the offense”—which is absent from the ACCA—narrows the statute’s reach; and (3) that § 924(c)(3)’s residual clause isn’t plagued (and confused, as is the ACCA’s) by linkage to a disjointed hodgepodge of enumerated offenses. Id. at 1266. [2]
Not long after the panel issued its opinion, the Supreme Court decided
Sessions v. Dimaya
,
In light of the Supreme Court’s decision in Dimaya , we vacated the panel’s opinion and took this case en banc to determine (1) whether 18 U.S.C.
§ 924(c)(3)’s residual clause is unconstitutionally vague under Dimaya and (2) whether we should overrule McGuire to the extent that it requires a categorical approach to determining whether an offense constitutes a “crime of violence” within the meaning of § 924(c)(3)(B). [3]
II
At the outset, some table-setting is in order. How exactly did we get here? Why did the Supreme Court conclude in both Johnson and Dimaya that the residual clauses before it were unconstitutionally vague, and what do the decisions in those cases tell us about § 924(c)(3)’s own residual clause? Here’s how—and why, and what.
A
We begin with a deeper dive into Johnson . As already explained briefly, Johnson involved the ACCA, which prescribes a mandatory minimum 15-year *12 sentence for any person who already “has three previous convictions . . . for a violent felony . . . committed on occasions different from one another.” 18 U.S.C. § 924(e)(1). The ACCA goes on to define the term “violent felony” to mean any crime punishable by a term of imprisonment exceeding one year that—
(i) has as an element the use, attempted use, or threatened use of physical force against the person of another; or
(ii) is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another[.]
18 U.S.C. § 924(e)(2)(B). Subsection (B)(i) of the ACCA’s definitional provision
is called (as we have called § 924(c)’s parallel provision) the “elements clause,”
while Subsection (B)(ii) contains both the “enumerated-offenses clause” and
separately (and again like § 924(c)’s catch-all) the “residual clause.”
Beeman v.
United States
,
In striking down the ACCA’s residual clause as unconstitutionally vague,
the Supreme Court in
Johnson
emphasized “[t]wo features.”
12
judicially imagined ‘ordinary case’ of a crime, not to real-world facts or statutory elements,” and thus “leaves grave uncertainty about how to estimate the risk posed by a crime.” Id. at 2557–58. Indeed, the Court made clear that application of the categorical approach was the hinge on which its vagueness determination turned: “It is one thing,” the Court stressed, “to apply an imprecise ‘serious potential risk’ standard to real-world facts; it is quite another to apply it to a judge-imagined abstraction” of the sort required by the categorical approach. Id . at 2558. Continuing in the same vein, the Court reiterated that “[a]s a general matter, we do not doubt the constitutionality of laws that call for the application of a qualitative standard such as ‘substantial risk’ to real-world conduct.” Id . at 2561. But, the Court held, the categorical approach’s focus on the “idealized ordinary case” requires an “abstract inquiry” that “offers significantly less predictability than one that deals with” actual facts. Id . (internal quotation marks and citation omitted).
B
Next, Dimaya . There, the Court considered a provision of the INA that renders an alien removable if he is “convicted of an aggravated felony at any time after admission.” 8 U.S.C. § 1227(a)(2)(A)(iii). The INA goes on to define the term “aggravated felony” to include, by statutory cross-reference, “a crime of violence (as defined in section 16 of Title 18[)].” Id . § 1101(a)(43)(F). Section 16’s definition of “crime of violence,” in turn, reads a lot like the ACCA’s
13
definition of “violent felony” at issue in Johnson :
(a) an offense that has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or
(b) any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
18 U.S.C. § 16. Subsection (a) is commonly called (you guessed it) the “elements
clause” and Subsection (b) the “residual clause.”
Dimaya
,
In Dimaya , the Supreme Court voided § 16’s residual clause (again, as incorporated in the INA) as unconstitutionally vague, concluding that it shared the two features that had doomed the ACCA’s residual clause in Johnson ―namely, (1) a fuzzy “substantial risk” standard and (2) incorporation of the categorical approach to determining the violence of the underlying crime. Id. at 1213–14. Significantly, though, just as in Johnson , the Dimaya Court stressed that § 16(b)’s vagueness problem resulted principally from the categorical approach: “The [ Johnson ] Court emphasized that [the ‘serious potential risk’ standard] alone would not have violated the void-for-vagueness doctrine: Many perfectly constitutional statutes use imprecise terms like ‘serious potential risk’ (as in ACCA’s residual clause) or ‘substantial risk’ (as in § 16’s). The problem came from layering such a standard on top of the requisite ‘ordinary case’ inquiry” required by the categorical approach. Id. at 1214 (citing Johnson , 135 S. Ct. at
14
2561). As Justice Thomas summarized in his dissent—without pushback—“[t]he sole reason that the Court deem[ed] § 16(b) unconstitutionally vague is because it read[] the statute as incorporating the categorical approach,” without which “the Court ‘d[id] not doubt’ the constitutionality of § 16(b).” Id. at 1252 (Thomas, J., dissenting).
An important caveat about Dimaya ’s application of the categorical approach to invalidate § 16’s residual clause: Only a plurality of the Court concluded that the statute actually requires the categorical approach. Justice Gorsuch, who provided the decisive fifth vote, concurred separately on the assumption—but not a determination—that § 16(b) incorporates the categorical approach. Id. at 1232 (Gorsuch, J., concurring in part and concurring in the judgment) (stating that he was “proceed[ing] on the premise” that the categorical approach applied). Justice Gorsuch emphasized that he “remain[s] open to different arguments about [Supreme Court] precedent and the proper reading of language like” that found in § 16(b), and that he “would address them in another case, whether involving the INA or a different statute, where the parties have a chance to be heard and we might benefit from their learning.” Id . at 1233.
C
So what do Johnson and Dimaya portend for § 924(c), which again, as relevant here, makes it a federal offense to use, carry, or possess a firearm in
15
connection with a “crime of violence”―which again, as relevant here, means a felony offense that “by its nature, involves a substantial risk that physical force … may be used in the course of committing the offense,” 18 U.S.C. § 924(c)(3)(B)?
Allow us first to state the obvious: Section 924(c)(3)’s residual clause is
identical—in every jot and tittle—to § 16’s, which the Supreme Court struck down
in
Dimaya
. Next, the less obvious but no less true: While the panel decision in this
case offered several distinctions between § 924(c)(3)’s residual clause and the
ACCA’s—which at the time had recently been invalidated in
Johnson
—the
Supreme Court’s intervening decision in
Dimaya
(in a portion of the opinion
joined by a majority of the justices) demolished all of them. As already noted, the
panel first emphasized that § 924(c)(3)’s residual clause refers not to the risk of
“physical injury” but to the risk of “physical force,” which it said was “much more
definite.”
16
statutory indeterminacy
Johnson
described.”
Accordingly, it seems clear that if we are required to apply the categorical approach in interpreting § 924(c)(3)’s residual clause—as the panel did, per our earlier decision in McGuire , and as the Supreme Court did in voiding the residual clauses before it in Johnson and Dimaya —then the provision is done for. If, by contrast, we are not required to apply the categorical approach in interpreting § 924(c)(3)(B), then there is every reason to believe that the provision will survive, notwithstanding its incorporation of a “substantial risk” term—because, as the Supreme Court said in Johnson and then reiterated in Dimaya , there is no reason to “doubt the constitutionality of laws that call for the application of a qualitative
17
standard such as ‘substantial risk’ to real-world conduct.”
Johnson
, 135 S. Ct. at
2561 (quoted in
Dimaya
,
III
That stark divergence—in which the categorical approach dooms § 924(c)(3)’s residual clause, while a conduct-based interpretation salvages it— tees up the rule of “constitutional doubt.” Simply stated, that canon of construction provides that “[a] statute should be interpreted in a way that avoids placing its constitutionality in doubt.” Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 247 (2012).
As the Supreme Court has explained it, the constitutional-doubt canon “is a
tool for choosing between competing plausible interpretations of a statutory text,
resting on the reasonable presumption that Congress did not intend the alternative
which raises serious constitutional doubts.”
Clark v. Martinez
,
18
reasonable saving construction, a statute might be unconstitutionally vague.
See
United States ex rel. Att’y Gen. v. Delaware & Hudson Co.
,
The
question here, therefore, is whether § 924(c)(3)’s residual clause is in
fact “susceptible of multiple interpretations,”
Jennings
,
A
In assessing whether § 924(c)(3)’s residual clause truly compels the categorical approach, we begin at the beginning: Where did this “categorical approach” come from? It’s certainly not, it seems to us, the most intuitive way of thinking about a particular crime’s risk of violence. Surely the usual means of
19
considering that issue would be to account for all of the specific circumstances
surrounding the offense’s commission—
i.e.
, the actual facts. If you were to ask
John Q. Public whether a particular crime posed a substantial risk of violence,
surely he would respond, “Well, tell me how it went down—
what happened
?”
How, then, did we get to the point where, in certain circumstances, reviewing
courts are required to ignore the real-world facts in favor of a sterile academic
inquiry into what the
Johnson
Court called “speculative,” “idealized,” “judge-
imagined abstraction[s]”?
That story follows.
1
The Supreme Court initially conceived the categorical approach in
Taylor v.
United States
,
20
“particular facts” of the defendant’s conduct, and instead adopted what the Court dubbed—and we still call—a “categorical approach.” Id . at 598–602. In explaining why the ACCA’s enumerated-offenses clause requires the categorical approach, the Taylor Court emphasized two factors—one textual, the other practical.
First, the Court concluded that when read in context, § 924(e)(2)(B)(ii) “most likely refers to the elements of the statute of conviction, not to the facts of each defendant’s conduct.” Id. at 600–01. The reason, the Court explained, is that the language of the ACCA’s operative provision, § 924(e)(1), “refers to ‘a person who ... has three previous convictions’ for—not a person who has committed— three previous violent felonies or drug offenses.” Id. at 600. Congress’s targeted focus on “convictions” rather than conduct, the Court reasoned, indicated that it “intended the sentencing court to look only to the fact that the defendant had been convicted of crimes falling within certain categories, and not to the facts underlying the prior convictions.” Id.
Second, the Taylor Court stressed that in the ACCA context, “the practical difficulties and potential unfairness of a factual approach [would be] daunting.” Id. at 601. In particular, the Court worried about the amount of evidence that might need to be introduced at a sentencing hearing in order to reconstruct the circumstances underlying a defendant’s prior (and often long-since-passed)
21
convictions. Id. Relatedly, the Court anticipated a Sixth Amendment problem that later decisions would amplify―namely, that judicial factfinding at sentencing about the real-world facts of crimes that led to prior convictions could “abridg[e a defendant’s] right to a jury trial[.]” Id. ; see also Apprendi v. New Jersey , 530 U.S. 466, 490 (2000) (“Other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt.”). In short, the Taylor Court feared that if the parties could introduce evidence bearing on the violence of the defendant’s past crimes, then sentencing proceedings might devolve into full- blown mini-trials (hence the impracticability) in which judges, rather than juries, were doing the factfinding (hence the Sixth Amendment concern). See 495 U.S. at 601–02.
For these reasons—the text’s focus on “convictions” and the impracticability (and unfairness) of effectively re-litigating the seriousness of stale crimes long after the fact—the Taylor Court concluded that for purposes of deciding whether a prior conviction constitutes a “violent felony,” the “only plausible interpretation” of § 924(e)(2)(B)(ii) is that it “generally requires the trial court to look only to the fact of conviction and the statutory definition of the prior offense,” and not to the
22
actual circumstances of the defendant’s crime. Id. at 602. [4]
2
The Supreme Court next applied the categorical approach in
Leocal v.
Ashcroft
,
3
That, for present purposes, brings us (back) to
Johnson
, in which, as already
explained, the Supreme Court applied the categorical approach in the course of
invalidating the ACCA’s residual clause.
See
24
sentence enhancement on prior crimes—the “utter impracticability” of requiring a court “to reconstruct, long after the original conviction, the conduct underlying that conviction.” Id.
4
Last came
Dimaya
, in which the Court applied the categorical approach in
striking down § 16’s residual clause—again, at least as that provision is
incorporated by the INA. A four-justice plurality concluded that § 16(b)
incorporates the categorical approach for a handful of (now increasingly familiar)
reasons. First, as in
Johnson
, the plurality noted that the government hadn’t
advocated a conduct-based approach: “To begin where
Johnson
did, the
Government once again ‘has not asked us to abandon the categorical approach in
residual-clause cases.’”
Dimaya
,
Third, the plurality explained that “[b]est read,” the text of § 16’s residual
clause incorporates the categorical approach.
Id.
Quoting the Court’s earlier
decision in
Nijhawan v. Holder
,
Simple references to a ‘conviction,’ ‘felony,’ or ‘offense,’ . . . are ‘read naturally’ to denote the ‘crime as generally committed.’ And
25
the words ‘by its nature’ in § 16(b) make that meaning all the clearer. The statute, recall, directs courts to consider whether an offense, by its nature, poses the requisite risk of force. An offense’s ‘nature’ means its ‘normal and characteristic quality.’
Dimaya
,
Finally, following Taylor and Johnson , the plurality stressed the “utter impracticability” of applying a conduct-based approach to a statute, like § 16(b), that requires consideration of prior convictions―in particular, the “daunting difficulties of accurately reconstructing, often many years later, the conduct underlying a conviction.” Id. (internal quotation marks omitted).
As already noted, Justice Gorsuch concurred separately in Dimaya , explaining that he was “proceed[ing] on the premise”—without definitively concluding—that as used in the INA, § 16(b) incorporates the categorical approach. Id. at 1232 (Gorsuch, J., concurring in part and concurring in the judgment). He gave several reasons for his circumspection: (1) “because no party [had] argued for a different way to read” the provision at issue; (2) because Supreme Court precedent (by which he presumably meant Leocal ) “seemingly require[d]” application of the categorical approach to § 16(b); and (3) “because the
26
government itself ha[d] conceded (repeatedly) that the law compels” the categorical approach in immigration-related § 16(b) cases. Id . He emphasized, though, that he would “remain open” in future cases “to different arguments about our precedent and the proper reading of language like” that found in § 16(b). Id. at 1233.
* * *
So . . . what are the takeaways? What factors have led the Supreme Court to conclude that a statute requires the categorical approach? The decisions interpreting the ACCA and § 16 reveal that the Court has historically applied the categorical approach to those statutes’ residual clauses for the following reasons:
1. because the government never asked the Court to consider a conduct-based approach ( Johnson , Dimaya );
2. because the text of those statutes’ operative provisions focused not on conduct, but rather on “convictions”—and thus, the Court reasoned, solely on formal legal elements ( Taylor , Johnson );
3. because those statutes’ definitional provisions used terms and phrases like “offense,” “felony,” and “by its nature,” which the Court concluded pointed toward a categorical (rather than conduct-based) inquiry ( Leocal , Dimaya ); 4. because those statutes lacked any reference to the underlying crime’s commission or circumstances ( Dimaya );
5. because applying the categorical approach would avoid the impracticability of requiring sentencing courts to engage in after-the-fact reconstructions of the circumstances underlying prior convictions ( Taylor , Johnson , Dimaya ); and
6. because applying the categorical approach would avoid the Sixth
27
Amendment issues that could arise from sentencing courts making findings
of fact that properly belong to juries (
Taylor
,
Johnson
,
Dimaya
).
The decisive question, it seems to us, is whether those considerations require
us to interpret § 924(c)(3)’s own residual clause to incorporate the categorical
approach—or whether, instead, the clause can “plausibly” be read to incorporate
the conduct-based approach. For reasons explained below, we conclude that
§ 924(c)(3)(B) can at the very least plausibly be read to bear a conduct-based
interpretation, and we therefore hold, pursuant to the canon of constitutional doubt,
that because the conduct-based reading spares the residual clause from the near-
certain death to which the categorical approach would condemn it, the conduct-
based approach must prevail. In so doing, we join the Second Circuit, which also
recently concluded—likewise applying the constitutional-doubt canon—that
§ 924(c)(3)(B) should be interpreted to embody the conduct-based approach.
See
United States v. Barrett
, __ F.3d ___,
B
Although it’s not particularly elegant—in fact, it’s downright clunky and more than a little repetitive—there’s really not a better way to assess whether the Supreme Court’s own stated reasons for adopting the categorical approach in the ACCA and immigration-related § 16 contexts likewise compel a categorical
28
interpretation of § 924(c)(3)(B) than simply to march through them, one by one.
1
In applying the categorical approach in both
Johnson
and
Dimaya
, the
Supreme Court “first” and most prominently noted that the government hadn’t
advocated a conduct-based interpretation.
See Johnson
,
Suffice it to say that things are very different here. In the wake of
Johnson
and
Dimaya
—and the ensuing drumbeat suggesting that application of the
categorical approach likewise imperils § 924(c)(3)’s residual clause―the
government has expressly (and at length) urged us to abandon the categorical
approach to § 924(c)(3)(B) in favor of a conduct-based interpretation.
See
Appellee’s En Banc Br. at 12–43. We have here, therefore, what the Supreme
Court lacked in both
Johnson
and
Dimaya
, and what the panel lacked in
McGuire
—namely, the benefit of the full “adversarial testing” that is so “crucial to
sound judicial decisionmaking,”
Dimaya
,
29
concurring in part and concurring in the judgment). [5]
2
A second basis that the Supreme Court has highlighted in applying the
categorical approach—derived from the text of the ACCA’s operative provision,
and specifically its reference to “convictions”―is likewise inapplicable here. As
already noted, the Court in
Johnson
(relying on and quoting its earlier decision in
Taylor
) emphasized that the ACCA’s operative clause “refers to a person who …
has three previous convictions for—not a person who has committed—three
previous violent felonies or drug offenses.”
Section 924(c)’s operative provision nowhere refers to “convictions.” See *31 18 U.S.C. § 924(c)(1)(A). Quite the opposite, in fact—it refers to conduct : It prescribes an increased term of imprisonment for “any person who, during and in relation to any crime of violence . . . uses or carries a firearm, or who, in furtherance of any such crime, possesses a firearm.” Id. That’s not dispositive, of course—§ 924(c)’s operative provision also applies to the statute’s elements clause, which all seem to agree incorporates the categorical approach. But it does demonstrate that § 924(c) lacks one of the key textual hooks that has traditionally buttressed the Supreme Court’s application of the categorical approach to statutory residual clauses.
3
Now, in fairness, there is some textual evidence that, on balance, might be thought to favor interpreting § 924(c)(3)(B) to incorporate the categorical approach. But it does not, we conclude, truly compel a categorical interpretation, especially when weighed against other textual and practical considerations.
In
McGuire
, we applied the categorical approach to § 924(c)(3)’s residual
clause because, we said, “of the statute’s terms.”
31
particular, the statute’s use of the term “offense” and the phrase “by its nature”— mandates the categorical approach. We conclude that it does not.
a
In support of its determination that § 16’s residual clause requires the
categorical approach, the
Dimaya
plurality relied on the Court’s earlier observation
in
Leocal
that § 16(b) “directs our focus to the ‘offense’ of conviction . . . rather
than to the particular facts.”
Dimaya
,
Ovalles’s position finds some support in § 924(c)(3)’s definition of “crime of violence,” which incorporates two of the three terms—“offense” and “felony”— that the Dimaya plurality highlighted: “[T]he term ‘crime of violence’ means an offense that is a felony . . . .” 18 U.S.C. § 924(c)(3). There are important counterweights, though. First, as already explained, § 924(c) nowhere uses the word “conviction,” the term that the Supreme Court has historically (going all the way back to Taylor ) emphasized as a key textual driver of the categorical approach. Second, even as to “offense” and “felony,” all the plurality said in Dimaya —echoing the Court’s earlier decision in Nijhawan —was that those terms are “naturally” read to refer to generic crimes, not that they are necessarily so read.
32
And indeed, on the very same page from which the
Dimaya
plurality took its “read
naturally” quote,
Nijhawan
explains “the linguistic fact” that “in ordinary speech
words such as ‘crime,’ ‘felony,’ ‘offense,’ and the like”
can go either way
—
“sometimes [they] refer to a generic crime . . . and sometimes [they] refer to the
specific acts in which an offender engaged on a specific occasion.” 557 U.S. at
33–34;
see also id
. at 32 (holding that statutory provision using the term “offense”
called for application of a conduct-based, rather than categorical, approach);
United States v. Hayes
,
All things considered, therefore, § 924(c)(3)’s use of the terms “offense” and “felony”―particularly when combined with the absence of the word “conviction”―would be a pretty thin reed on which to base a conclusion that the residual clause requires the categorical approach.
b
The strongest piece of evidence in favor of applying the categorical approach to § 924(c)(3)’s residual clause, it seems to us, is the provision’s use of the phrase “by its nature”: “[T]he term ‘crime of violence’ means an offense that is a felony and . . . that by its nature , involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.” 18 U.S.C. § 924(c)(3)(B) (emphasis added).
In
Leocal
, the Court relied in part on § 16(b)’s use of the same “by its
nature” phrase to conclude that the statute “requires us to look to the elements and
the nature of the offense of conviction, rather than to the particular facts relating to
petitioner’s crime.”
34
Dictionary for the proposition that “[a]n offense’s ‘nature’ means its ‘normal and characteristic quality,’” the Dimaya plurality reasoned that “§ 16(b) tells courts to figure out what an offense normally—or, as we have repeatedly said, ‘ordinarily’—entails, not what happened to occur on one occasion.” Id. at 1217– 18 (internal citation omitted).
That is undoubtedly a reasonable interpretation of § 16(b)’s language— which, again, § 924(c)(3)’s residual clause mirrors. But importantly here—where the constitutional-doubt canon is in play—it is not a necessary interpretation. There are other reasonable understandings—especially of § 924(c)(3)(B), with respect to which other interpretive considerations point in the other direction. Webster’s Third , for instance—the same dictionary that the Dimaya plurality cited in support of its categorical-approach interpretation of the “by its nature” language, alternatively defines the word “nature” to mean “the essential character or constitution of something.” Webster’s Third New International Dictionary 1507 (2002). The Oxford Dictionary of English similarly defines “nature” as “the basic or inherent features, character, or qualities of something.” Oxford Dictionary of English 1183 (3d ed. 2010). It seems to us at the very least plausible that the “something[s]” to which the term “nature” alludes could be particular acts rather than (or in addition to) the “judge-imagined abstraction[s],” Johnson , 135 S. Ct. at 2558, that underlie the categorical approach. See Barrett , __ F.3d ___, 2018 WL
35
4288566, at *13 (“[N]othing in these definitions indicates whether the offense whose inherent characteristics are to be considered is the generic crime or the particular one charged.”).
So, as Justice Thomas noted in
Dimaya
, “[o]n the one hand, the statute
might refer to the metaphysical ‘nature’ of the offense and ask whether it ordinarily
involves a substantial risk of physical force.”
Dimaya
,
To be clear, it’s no answer to say, “Sure, but Justice Thomas lost in Dimaya .” As an initial matter, he was objecting there to a plurality’s (not a majority’s) interpretation—and in particular its interpretation of a different statute, § 16(b), that lacks many of the textual, contextual, and practical features that we conclude permit a conduct-based interpretation of § 924(c)(3)’s residual clause. Moreover, and in any event, Justice Thomas’s linguistic observation about the alternative meanings of the word “nature” remains—and indeed, finds support in
36
dueling dictionary definitions. Especially in light of the inapplicability of other reasons for applying the categorical approach to § 924(c)(3)(B), we simply aren’t convinced that the phrase “by its nature” requires application of the categorical approach here.
4
In addition to focusing on what the controlling statutes
say
in assessing the
categorical-approach issue, the Supreme Court has emphasized what they
don’t say
. In
Dimaya
, for instance, the plurality reasoned that “the absence” from § 16
“of terms alluding to a crime’s circumstances, or its commission, makes a
[conduct]-based interpretation an uncomfortable fit.”
But again, § 924(c) is different. Not only (as already explained) does the statute’s operative provision refer exclusively to conduct, but its definitional provision ( i.e. , the residual clause itself) also contains conduct-based language―namely, its requirement that the risk of force arise “in the course of committing the offense.” 18 U.S.C. § 924(c)(3)(B).
5
Textual indicia aside, the Supreme Court has also emphasized practical considerations in deciding between categorical and conduct-based interpretations. As already explained, in initially devising the categorical approach in Taylor —and thereafter applying it in Johnson and Dimaya —the Supreme Court underscored the
37
“utter impracticability” of applying a conduct-based approach retrospectively to
determining the violence of prior crimes.
E.g.
,
Johnson
,
Importantly, the look-back problem doesn’t arise with respect to § 924(c), which serves an altogether different function from the statutes at issue in Johnson and Dimaya and operates differently in order to achieve that function. The ACCA
38
identifies “ previous convictions” for the purpose of applying a recidivism-based sentencing enhancement to three-time felons who later possess firearms in violation of 18 U.S.C. § 922(g). See 18 U.S.C. § 924(e)(1) (emphasis added). Similarly, § 16(b)―as incorporated by the INA―classifies certain prior convictions as “crime[s] of violence” for the purpose of rendering aliens removable. See 18 U.S.C. § 16; 8 U.S.C. § 1101(a)(43)(F); 8 U.S.C.
§ 1227(a)(2)(A)(iii).
[7]
Section 924(c), by contrast, operates entirely in the
present—it creates a new and distinct offense for any person who “during and in
relation to any crime of violence ... uses or carries a firearm, or who, in furtherance
of any such crime, possesses a firearm.” 18 U.S.C. § 924(c)(1)(A). So, unlike in
the ACCA and § 16 contexts—where courts have to look backwards in time to
consider past crimes remote from (and wholly unconnected to) the charged
*40
offense, § 924(c)(3)’s definition of a “crime of violence” is never applied to an
unrelated prior crime or conviction. Instead, in § 924(c) cases, the firearms offense
and the predicate “crime of violence” go hand-in-hand; they inherently arise out of
the same event. By dint of the statute’s plain language, the gun-related conduct
must occur “during,” “in relation to,” or “in furtherance of” the crime of violence,
and with respect to that underlying offense the actionable risk of force must occur
specifically “in the course of committing” it. In short, it’s all one big ball of
wax—the crimes are typically (as here) charged in the same indictment, and if they
are tried, they are considered by the same jury. The “utter impracticability” that
Taylor
,
Johnson
, and
Dimaya
identified—what we have called the look-back
problem—simply isn’t an issue.
See Barrett
, __ F.3d ___,
6
Relatedly, echoing the earlier decision in
Taylor
, the
Dimaya
plurality
acknowledged that the Court “adopted the categorical approach in part to avoid the
Sixth Amendment concerns that would arise from sentencing courts’ making
findings of fact that properly belong to juries.”
40
reviewing court could well run afoul of the rule that “[o]ther than the fact of a prior
conviction, any fact that increases the penalty for a crime beyond the prescribed
statutory maximum must be submitted to a jury, and proved beyond a reasonable
doubt,”
Apprendi
,
Yet again, that isn’t a concern here. For starters, because the jury in a §
924(c) case—unlike in, say, an ACCA case—is considering contemporaneous gun-
related and predicate offenses, its role already necessarily entails consideration of
the entire course of conduct charged as the underlying “crime of violence.” Moreover, and significantly, the government here has conceded that whether the
defendant’s predicate offense constitutes a “crime of violence” within the meaning
of § 924(c)(3)’s residual clause should be treated as a mixed question of fact and
law to be resolved by a jury.
See
Appellee’s En Banc Br. at 33–34 (citing
United
States v. Gaudin
,
41
carriage, or possession of the firearm occurred during and in relation to (or in
furtherance of) the federal offense, but also—and importantly—(4) that the federal
offense was in fact a “crime of violence.” As has the Supreme Court in similar
circumstances, we conclude that the government’s concession that, absent a plea, it
must prove and a jury must find all four elements―including that the underlying
offense qualifies as a “crime of violence”―“eliminat[es] any constitutional
concern.”
Nijhawan
,
* * *
So where does all of this leave us? With something of a mixed bag, frankly. While some of the factors to which the Supreme Court has pointed in adopting and applying the categorical approach might be thought (on balance) to favor a similar interpretation of § 924(c)(3)(B), others cut pretty decisively in the opposite direction, toward a conduct-based approach. And in constitutional-doubt land, the tie (or the toss-up, or even the shoulder-shrug) goes to the statute-saving option— which, here, is the conduct-based interpretation. So to be clear, we needn’t—and don’t—conclude that textual, contextual, and practical considerations compel a conduct-based reading of § 924(c)(3)’s residual clause. Nor, for that matter, do we even need to find that § 924(c)(3)(B) is best read to incorporate a conduct-based approach. Reasonable minds—say, for instance, the minds of the (putatively) or she . . . recklessly engages in conduct which creates a grave risk of death to a person other than himself or herself, and thereby causes the death of another person.”); Conn. Gen. Stat. § 53a-112(a)(1)(A) (“A person is guilty of arson in the second degree when, with intent to destroy or damage a building . . . he starts a fire or causes an explosion and . . . such act subjects another person to a substantial risk of bodily injury . . . .”); Mo. Stat. § 565.120(1) (“A person commits the offense of kidnapping in the second degree if he or she knowingly restrains another unlawfully and without consent so as to interfere substantially with his or her liberty and exposes him or her to a substantial risk of serious physical injury.”).
Nor are we particularly troubled by Ovalles’s argument that allowing jurors to consider a defendant’s use or possession of a firearm will inevitably transform every underlying offense into a “crime of violence” within the meaning of § 924(c)(3). The government concedes that it must separately prove—and that jurors should be charged that they must separately find—(1) that the federal offense was “a crime of violence” and (2) that the defendant used, carried, or possessed a firearm in the course of committing the underlying offense. District courts should caution jurors that they may not find that the underlying offense involved a substantial risk of physical force solely because the defendant possessed a gun.
43
reasonable judges who join this opinion—can and will disagree about that. It is
enough for us to conclude—as we think is indisputable—that § 924(c)(3)(B) is at
least “plausibl[y]” (or “fairly possibl[y]”) understood to embody the conduct-based
approach.
See Clark
,
We therefore overrule McGuire to the extent that it requires application of the categorical approach to determine whether an offense constitutes a “crime of violence” within the meaning of § 924(c)(3)(B) and hold that the crime-of-violence determination should be made, instead, using a conduct-based approach.
IV
Having jettisoned the categorical interpretation in favor of the conduct-based
approach for cases arising under § 924(c)(3)’s residual clause, we can make quick
work of the contention that the clause is unconstitutionally vague in the light of
Dimaya
. It is not. The Supreme Court has repeatedly explained—and again all
here agree—that the vagueness problem that plagued the ACCA and § 16, and that
is asserted here, is a function of the “speculative,” “idealized,” “abstract” inquiries
required by the categorical approach.
Johnson
,
V
That leaves us only to apply § 924(c)(3)(B)’s conduct-based approach to Ovalles’s case. Given the stipulated facts before us—embodied in a written plea agreement and a detailed colloquy—doing so is remarkably straightforward.
It is common ground here that in order to convict Ovalles on the § 924(c)
45
charge, the government would need to prove (or Ovalles would need to plead to) four distinct elements: (1) that Ovalles actually committed the underlying federal offense—here, the attempted carjacking; (2) that the attempted-carjacking offense constitutes a “crime of violence” within the meaning of § 924(c)(3); (3) that Ovalles knowingly used, carried, or possessed a firearm; and (4) that any use or carriage of the firearm occurred during and in relation to, or that any possession of the firearm was in furtherance of, the attempted carjacking. See 18 U.S.C. § 924(c)(1)(A). The only element in dispute here is the second: Did Ovalles’s attempted-carjacking offense, as she has admitted it actually occurred , constitute a “crime of violence”— i.e. , did her own acknowledged conduct “involve[] a substantial risk that physical force against the person or property of another may be used in the course of committing the offense”? Easy. Of course it did.
As it pertained to the attempted-carjacking count, the information charged Ovalles with “attempt[ing] to take a motor vehicle . . . from the person and presence of another, by force, violence and intimidation” with “the intent to cause death and serious bodily harm.” Ovalles then executed a written plea agreement acknowledging that she understood each charge in the information—including the attempted-carjacking charge—and that “she [was] pleading guilty because she is in fact guilty of the crimes [as] charged.” If attempting to steal a car “by force, violence and intimidation” and with “the intent to cause death and serious bodily
46
harm” (as Ovalles has admitted she did) doesn’t involve a “substantial risk” that physical force may be used, then it’s hard to imagine what does.
Here, though, there’s no need for imagination—the real-life details of Ovalles’s crime, all of which she has admitted, confirm it. We won’t restate in full the government’s factual proffer recounting Ovalles’s involvement in the three-day carjacking spree. Suffice it to say that, in general, the proffer demonstrated that Ovalles and her co-conspirators robbed a grocery store, successfully carjacked three automobiles by force, and attempted to carjack a fourth. With respect to the attempted carjacking, in particular—which, again, serves as the predicate offense for Ovalles’s § 924(c) conviction—the proffer detailed that Ovalles and her co- conspirators approached a family getting out of their minivan, demanded the keys, hit the family’s 13-year-old child in the face with a baseball bat, and then, in making their escape, fired an AK-47 assault rifle at the family and a Good Samaritan who had come to their aid. See supra at 6–7.
Especially when layered on top of Ovalles’s admission to the overtly violent charge in the information, the government’s detailed factual proffer―with which Ovalles repeatedly said she had no “material disagreement”—leads inexorably to the conclusion that the attempted carjacking at issue here constitutes a “crime of violence” within the meaning of § 924(c)(3)(B). Based on the facts to which she has expressly stipulated, there simply can be no serious dispute that Ovalles
47
recognized that her conduct posed a very real “risk” that physical force “may” be used―just, as it turns out, it was.
VI
Accordingly, we hold as follows:
1. The question whether a predicate offense constitutes a “crime of violence” within the meaning of 18 U.S.C. § 924(c)(3)(B) should be determined using a conduct-based approach that accounts for the actual, real-world facts of the crime’s commission, rather than a categorical approach.
2. To the extent that our decision in United States v. McGuire , 706 F.3d 1333 (11th Cir. 2013), required use of the categorical approach in making the crime-of-violence determination under § 924(c)(3)(B), it is overruled.
3. As interpreted to embody a conduct-based approach, § 924(c)(3)(B) is not unconstitutionally vague.
4. In light of the particular circumstances of its commission, all of which Ovalles has expressly admitted, her attempted-carjacking offense was a “crime of violence” within the meaning of § 924(c)(3)(B).
The case is REMANDED to the panel for proceedings consistent with this opinion.
48
WILLIAM PRYOR, Circuit Judge, joined by ED CARNES, Chief Judge, and TJOFLAT, NEWSOM, and BRANCH, Circuit Judges, concurring:
How did we ever reach the point where this Court, sitting en banc, must debate whether a carjacking in which an assailant struck a 13-year-old girl in the mouth with a baseball bat and a cohort fired an AK-47 at her family is a crime of violence? It’s nuts. And Congress needs to act to end this ongoing judicial charade.
I join the majority opinion in full, but I write separately to explain why our resolution of this appeal forecasts how Congress should address the vexing issue of how to punish violent recidivists under laws like the Armed Career Criminal Act [1] : by restoring the traditional role of the jury. The caselaw about how to punish recidivists has confounded the federal courts for decades and has made the resolution of this appeal tricky, but our decision also suggests a way out of the mess. Although our decision involves a contemporaneous crime and not a prior conviction, our conclusion that a jury may make findings about a defendant’s violent conduct applies with equal force to recidivist statutes. Indeed, the modern abandonment of the jury’s traditional role of making findings about prior convictions has created more problems than it has solved.
A. Recidivist Wars
Empirical research proves that Congress has good reason to punish recidivists with long sentences of imprisonment. A recent study by the United States Sentencing Commission found that “[c]areer offenders, as a group, tend to recidivate at a higher rate than non-career offenders.” [2] More specifically, “almost two-thirds . . . of career offenders released between 2004 and 2006 were rearrested” in the eight-year period after their release, while just under “one- half . . . of non-career offenders released in 2005 were rearrested” in the same period. [3] And a defendant’s criminal history as calculated under the Sentencing Guidelines is a strong predictor of future encounters with the justice system. [4] Offenders without any criminal history points are rearrested at a rate of 30.2 percent. [5] This rate jumps to 63.3 percent for offenders with five criminal history points and catapults to 81.5 percent for offenders with more than 10 criminal history points. [6] Relatedly, offenders with the lowest criminal history category are *51 rearrested at a rate of 33.8 percent, while offenders with the highest criminal history category are rearrested at a rate of 80.1 percent. [7]
Among career offenders, violent offenders, not surprisingly, pose the greatest risk to the public. For example, “drug trafficking only” career offenders recidivate at a rate of 54.4 percent, while “violent [crime] only” career offenders recidivate at a rate of 69 percent. [8] The median drug trafficker recidivates after 26 months, while the median violent offender recidivates after only 14 months. [9] The median drug trafficker commits two “[r]ecidivism [e]vents,” while the median violent offender commits three. [10] And when a drug trafficker reoffends, he is most likely to commit another drug trafficking offense, while a violent offender is most likely to commit robbery. [11] Indeed, even “mixed” career offenders who have at least one violent offense [12] recidivate at a rate of 69.4 percent and are most likely to commit assault when they recidivate. [13]
Past offenses involving a firearm are also a strong predictor of future crimes. [14] For example, “[o]ffenders whose federal offense involved firearms [are] *52 most likely to be rearrested” at a rate of 68.3 percent, and “[o]ffenders who received an enhanced sentence for a weapon . . . ha[ve] higher recidivism rates than other offenders.” [15] In short, criminals with an affinity for guns and violence are the least likely to change their ways. [16]
That recidivists merit longer sentences is hardly a new discovery. Over a century ago, the Supreme Court explained that “[t]he propriety of inflicting severer punishment upon old offenders has long been recognized in this country and in England.” [17] Accordingly, “[s]tatutes providing for such increased punishment were enacted in Virginia and New York as early as 1796, and in Massachusetts in 1804; and there have been numerous acts of similar import in many states.” [18]
In more recent years, Congress has moved to protect public safety by enacting federal crimes that provide lengthy terms of imprisonment for *53 recidivists. [19] The Armed Career Criminal Act, for example, provides a 15-year mandatory minimum sentence for a criminal convicted of unlawfully possessing a firearm or ammunition who has “three previous convictions . . . for a violent felony or serious drug offense.” [20] The Act defines a “serious drug offense” in terms of convictions under certain federal and state laws that provide a maximum sentence of 10 or more years of imprisonment. [21] And it defines a “violent felony” to include crimes that (1) have “the use, attempted use, or threatened use of physical force” as an “element” [22] —the so-called “elements clause”; (2) are “burglary, arson, or extortion, [or] involve[] use of explosives” [23] —the “enumerated offenses clause”; or (3) “otherwise . . . present[] a serious potential risk of physical injury to another” [24] —the “residual clause.”
B. The Residual Clause Strikes Back
As thoroughly chronicled in the majority opinion, years of litigation about federal recidivist statutes have weakened the penalties created by Congress. But *54 these developments have not benefited all offenders. Recidivist drug traffickers who possess a firearm are still subject to the 15-year mandatory minimum under the Armed Career Criminal Act. But after the demise of the residual clause in Johnson v. United States , [25] violent recidivists who possess a firearm—the ones akin to the violent career offenders who recidivate at a rate nearly 15 percentage points higher than recidivist drug traffickers and who tend to commit violent crimes when they do so—sometimes escape the mandatory minimum if their earlier offenses lacked a legal “element” that involves physical force, regardless of how the offender actually committed the offense. For example, an offender may have been convicted of “sexual abuse in the first degree . . . by forcible compulsion” under Alabama law for a crime that was actually violent. [26] But because the Supreme Court of Alabama has adopted a “water[ed] . . . down” interpretation of the statute that “means that [the statute] does not categorically include as an element the use, attempted use, or threatened use of physical force,” the conviction will not count as a crime of violence because “the true facts matter little, if at all, in this odd area of the law.” [27]
The same problem may exist for other offenses that can technically be committed in nonviolent ways—but rarely, if ever, are—such as kidnapping, arson, *55 and other sex crimes. Criminals who have committed violent crimes may escape the mandatory minimum sentence of the Armed Career Criminal Act because sentencing judges must “close [their] eyes” to everything but the legal definitions of prior convictions. [28] In essence, the so-called categorical approach “divorces what a habitual offender actually did from the punishment they are meant to receive.” [29] “The bizarre results” occasioned by this approach “are hard to grasp” because “the doctrine is not based in reality, but rather relies on the legal fiction that crimes are [merely] comprised of a set of elements, as opposed to the underlying criminal conduct.” [30]
I do not mean to suggest that the residual clauses of the Armed Career Criminal Act or of the federal definition of a crime of violence in recidivist statutes [31] should have survived—at least not in their original forms that required the sentencing judge to decide whether the offender’s criminal history qualified him for an increased punishment. Judges have rightfully complained that the categorical approach has taxed judicial economy. “The dockets of . . . all federal *56 courts are now clogged with [ACCA] cases,” [32] and perhaps “no other area of law has demanded more of [the courts’] resources.” [33] And as the Supreme Court explained in Johnson , the two-step process demanded by the categorical approach, in which the sentencing judge imagined the “ordinary case” of a crime and then decided “whether that abstraction present[ed] a serious potential risk of physical injury,” was vague and confusing. [34] The residual clause produced several circuit splits over whether certain crimes were sufficiently violent. [35] Indeed, the justices could not agree whether fleeing in a vehicle from law enforcement or drunk driving qualified as violent crimes, [36] eventually leading the Supreme Court to conclude that “trying to derive meaning from the residual clause . . . [was] a failed enterprise.” [37] And even the elements clause has created confusion, with members of this Court disagreeing about whether crimes like attempted murder, attempted armed robbery, robbery, aggravated assault, and felony battery are violent crimes. [38]
The problems with the categorical approach also cannot necessarily be remedied by having the sentencing judge make findings about the underlying facts of prior convictions. The Sixth Amendment requires that, “[o]ther than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury.” [39] And the specific details of a recidivist’s past crimes go beyond the simple “fact of [his] prior conviction.” [40] Indeed, some members of the Supreme Court have opined that, under the Sixth Amendment, even the bare fact of a prior conviction should be proved to a jury. [41]
C. Return of the Jury
robbery are not violent felonies);
United States v. Lee
,
[39]
Apprendi v. New Jersey
,
[40]
Id.
;
see also Mathis v. United States
,
[41]
See, e.g.
,
Dimaya
,
The combination of the need to punish violent recidivists and the problem of judicial factfinding about prior convictions yields an obvious solution for Congress: rewrite the Armed Career Criminal Act and other recidivist statutes to require that the government must prove to a jury beyond a reasonable doubt that the defendant has previously been convicted of a felony the actual commission of which involved the use, attempted use, or threatened use of physical force against another person. For example, in the case of a defendant with a prior conviction under Alabama law for first-degree sexual abuse by forcible compulsion, the prosecution could introduce evidence that the defendant in fact used violent force to subdue his victim. This evidence might include certified charging documents, stipulations, plea agreements, factual proffers, and verdict forms from the past prosecution, as well as traditional evidence such as witness testimony and physical evidence.
Tasking the jury with determining recidivism is consistent with the common law. “Habitual offender laws like the ACCA enjoy a long tradition in this country that dates back to colonial times.” [42] And“[a]t common law, the fact of prior convictions had to be charged in the same indictment charging the underlying crime . . . and submitted to the jury for determination along with that crime.” [43] In *59 1967, the Supreme Court acknowledged that “[t]he common-law procedure for applying recidivist statutes . . . , which requires allegations and proof of past convictions in the current trial, is, of course, the simplest and best known procedure.” [44] And earlier decisions of state courts left little doubt that “a verdict of the jury finding the prior conviction . . . [was] essential to the power of the court to impose the increased punishment” absent a statute that displaced the common law. [45]
“[T]he right to have a jury decide prior-offense status . . . was the law in virtually every federal and state jurisdiction, from the Founding past World War II.” [46] For example, a former Texas recidivism statute provided enhanced penalties when the prosecutor proved to the jury that the offender “ha[d] been before convicted of the same [felony] offense . . . or one of the same nature.” [47] And a jury finding was more than a formality. Under a former Indiana statute, “the previous convictions, sentences, and imprisonments [had to] be described specifically, and the jury [had to] find that the defendant was convicted, sentenced, and imprisoned *60 in the instances described, and not otherwise.” [48] Indeed, in Kelley v. State , [49] the Supreme Court of Indiana held that the prosecution could not rely on “a certified transcript of a judgment . . . [that did] not describe the crime for which [the defendant] was convicted, but merely recite[d]” that the defendant was sentenced to between one and seven years of imprisonment. [50]
The common-law method for proving prior convictions ordinarily permitted the prosecutor to rely on a broad array of evidence if the defendant refused “to stipulate as to the prior conviction[] and thereby relieve the State of the necessity of . . . adducing proof before the jury of such prior convictions.” [51] For example, in Crocker v. State , [52] after a defendant accused of “robbery with a prior conviction for robbery” contested the fact of his first conviction, the prosecutor introduced “records from the penitentiary,” “testimony of one of the officers involved in the prior case,” and a “comparison of finger prints of the [defendant] with the finger prints taken in connection with the prior offense.” [53] And in Dozier v. State , [54] the *61 jury was tasked with weighing “certified copies of the judgment and sentence,” “records of the Texas Department of Correction, including fingerprints,” and “expert testimony identifying [the fingerprints] as identical with those of the [defendant].” [55] Factfinding by juries was considered essential to recidivism statutes.
And still today, there are states that require factfinding by juries to sentence an offender under their own recidivist statutes. Under Indiana’s current habitual offender statute, [56] “habitual offender proceedings are treated as substantive criminal trials. The State must prove the allegations beyond a reasonable doubt” to a jury. [57] The test for determining whether a prior conviction from another jurisdiction qualifies as a predicate offense under Indiana’s statute does not turn on the elements of the offense, but on whether the acts that resulted in the out-of-state conviction would have constituted a predicate offense “if they had been committed in Indiana.” [58]
California also uses a conduct-based approach in cases in which it is not possible to determine whether an offense committed in another jurisdiction would qualify as a predicate felony through an inspection of the elements of the offense *62 alone. [59] Although California law permits judges to make the factual findings required to determine that the conduct that resulted in an out-of-state conviction would have been a predicate offense if committed in the state, a California court “may not rely on its own independent review of record evidence to determine what conduct ‘realistically’ led to the defendant's conviction,” but must instead rely on “those facts that were already necessarily found by a prior jury in rendering a guilty verdict or admitted by the defendant in entering a guilty plea.” [60]
Within our own Circuit, Alabama [61] and Georgia [62] likewise require the state to prove that the conduct underlying an offender’s prior out-of-state conviction would have been a predicate offense under the state’s recidivist statute if committed in the state, although both Alabama [63] and Georgia [64] permit judges to make the requisite factual findings. So in advocating a return to the jury’s traditional role in determining recidivism at the federal level, I am not advancing a *63 proposal untested by contemporary experience. I am instead proposing a solution that has proved workable in practice in several states.
Some may object that jurors’ knowledge of past convictions will lead them to engage in propensity reasoning, [65] but the Supreme Court long ago held that the “use of prior convictions in [a] . . . criminal trial . . . [is not] so egregiously unfair upon the issue of guilt or innocence as to offend” the guarantee of due process of law. [66] In any event, the details of past convictions are unlikely to reach the jury. [67] To begin, about 97 percent of federal prosecutions end in guilty pleas. [68] And those few defendants who go to trial can stipulate to their past convictions and limit any proceedings to the charged offenses. [69]
The supposed unfairness to defendants of admitting proof of prior convictions is the same concern that motivated the failed experiment of allowing judges to assess recidivism. When the Supreme Court decided Spencer v. Texas [70] in 1967 and upheld the common-law jury method against a challenge under the Fourteenth Amendment, [71] the majority opined that it “might well agree” that other methods of assessing recidivism, such as “leaving that question to the court,” would be “faire[r]” than placing a defendant’s criminal history before the jury. [72] Justice Stewart also wrote in his concurring opinion that, “[i]f the Constitution gave [him] a roving commission to impose upon the criminal courts of [the states his] own notions of enlightened policy, [he] would not join the [majority] opinion” because other “recidivist procedures . . . are far superior to those utilized [under the common law].” [73] And then-contemporary academic commentators critiqued the common-law method and identified alternatives, including “determination of [recidivist] status by the judge.” [74] But here we are. The cure of judicial factfinding has proved worse than the disease it was supposed to treat.
Another objection might be that compiling evidence of prior convictions will be burdensome. But with the advent of electronic records and other advantages of modern technology, the task of reconstructing the details of prior convictions will today be far easier than it was under the original common-law procedure and, over time, will become easier still. [75] Because most state and federal prosecutions end in guilty pleas, the relevant details of prior convictions ordinarily will be preserved in factual proffers and other plea records. [76] And in the light of the rapid rate at which many violent offenders recidivate, other relevant evidence likely will be fresh. [77]
In any event, Congress has some options for ensuring that all violent recidivists remain subject to the mandatory minimum penalty. As explained above, the elements clause may fail to capture offenders who commit broadly defined crimes in violent ways. But the facts of their prior crimes, when ascertainable, should not be immune from consideration when they commit new crimes. supra note 45, at 340–41 & n.21 (critiquing the common-law method, but also acknowledging that judicial factfinding “deprives the defendant of a jury trial on the issue”).
[75]
Cf. Crocker
,
[76] See Bureau of Justice Assistance, U.S. Dep’t of Justice, Plea and Charge Bargaining 1 (2011) (“[A]bout 90 to 95 percent of both federal and state court cases are resolved through [plea bargaining].”).
[77] See U.S. Sentencing Comm’n, Report to the Congress , supra note 2, at 42 (finding a “[m]edian [t]ime to [r]ecidivism” of 14 months).
That this country comprises 50 states with different criminal codes all but ensures that a formalist approach to defining a crime of violence—administered only by judges—will be either over- or underinclusive and almost certainly will be vulnerable to persistent litigation like that which felled the residual clause. For example, a recent bill introduced in the United States Senate proposes to apply the mandatory minimum to defendants with “[three] or more previous serious felony convictions,” namely “any conviction . . . for an offense . . . punishable by imprisonment for a statutory maximum term of not less than 10 years.” [78] But this definition is both too narrow and too broad. It would exclude the offense of “felony battery” under Florida law, [79] a third-degree felony with a maximum term of imprisonment of five years, [80] even though this crime requires the infliction of “great bodily harm, permanent disability, or permanent disfigurement.” [81] It also would exclude convictions for “domestic battery by strangulation” under Florida law [82] and for “assault[ing] another and inflict[ing] substantial bodily harm” under Minnesota law. [83] But it would include the Florida offense of “[u]nlawful *67 possession or use of a fifth wheel,” [84] a second-degree felony with a maximum term of imprisonment of 15 years. [85] It would include a conviction under Utah law for “the intentional sale of five or more unlawful telecommunication devices within a six-month period,” [86] also a second-degree felony with a maximum term of imprisonment of 15 years. [87] And it would include the Massachusetts offense of knowing possession of “ten or more pieces of false money . . . with intent to utter or pass the same as true,” a crime punishable by a life sentence. [88] In the light of the substantial recidivism differences between violent and nonviolent offenders, Congress should ensure that their punishments are based on the nature of their past convictions and not the potential sentences for those convictions.
Make no mistake—Congress must eventually do something. The need to punish violent recidivists is just as strong today as it was when the Supreme Court decided Spencer over 50 years ago. [89] The only question is what Congress should do. Restoring the common-law role of the jury is the right place to start.
*68 MARTIN, Circuit Judge, dissenting:
As United States Circuit Judges, we have been given great power and privilege. And our positions call upon us to decide the fate of many people who have neither. In a nation that incarcerates a larger percentage of its population than almost all others, federal judges devote much time to examining (and reexamining) the sentences imposed on people serving time in our federal and state prisons. The interpretation the majority of this en banc Court gives to the sentencing statute at issue here, which gives no relief for Irma Ovalles, presents the opportunity to review the development of this Circuit’s sentencing jurisprudence in recent years. My review reveals a body of law that has relentlessly limited the ability of the incarcerated to have their sentences reviewed. Decisions of this Court have left only a narrow path to relief for those serving sentences longer than the law now allows. Yet this narrow path is not mandated by decisions of the Supreme Court or by Acts of Congress. Indeed, this Court has withheld relief from prisoners even when precedent counsels otherwise.
As did my colleagues in the majority, I begin with the landmark Supreme
Court decision in Johnson v. United States,
Yet at every turn, this Court erred in ways that stopped prisoners from getting their sentences reviewed and prevented people who had meritorious claims from getting relief. Ms. Ovalles’s case is a “successor” to Johnson, Maj. Op. at 2, and it is the latest in this line of decisions. Judge Jill Pryor’s dissent shows how the majority strays from the plain text of the statute and from Supreme Court precedent. I fully join her opinion. I write separately to provide the context of the Eleventh Circuit’s response to Johnson. This is not the first time this Court has mistakenly applied Johnson, nor the first time our mistake will leave an unlawful *70 sentence intact. Indeed, today’s en banc decision, like others before it, promises to have lasting effects for many prisoners.
I. JOHNSON AFTERMATH
When the Supreme Court invalidates a statute that was the basis for
sentences being served by thousands of federal prisoners, the first question
becomes, who benefits? Do we simply quit relying on this unconstitutional statute
for those who will be sentenced in the future, or do we go back and give relief to
those who were sentenced under the flawed statute in the past? There are rules
governing who gets relief. See generally Teague v. Lane,
1060 (1989). Generally, people sentenced under a flawed statute more than a year
before the Supreme Court announces a new rule like the one in Johnson get relief
based on the new rule (retroactive relief) if the Supreme Court has told us they can.
See 28 U.S.C. § 2255(f)(3), (h); In re Henry,
One statutory tool for a person seeking relief from an unlawful federal prison sentence is found in 28 U.S.C. § 2255. Provisions of this statute were intended to (and do) limit opportunities for prisoners to get courts to review problems with their sentences. Prisoners may contest their sentences once as a matter of right under § 2255, but they must do it within one year of their sentences becoming final. Id. § 2255(f). This one-year statute of limitations goes by quickly *71 indeed for most federal prisoners, who are routinely sentenced to serve decades- long prison terms. Section 2255 does allow prisoners to bring a second or successive challenge to their sentences after that first year, but it is much harder. For starters, prisoners seeking relief when they are years into serving their sentence must first come to a court of appeals to get permission. Id. §§ 2244(b)(3)(A), 2255(h). There are other barriers to getting a second chance at relief in federal court, but for purposes of this discussion, perhaps the most important is that the statute gives inmates no ability to contest a decision from a panel of this Court telling them they cannot file a second or successive petition. Id. § 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.”). For this “seeking permission” process, prisoners are stuck with the answer we give them, whether our answer is right or wrong.
As one would expect, after the Supreme Court decided Johnson, thousands
of prisoners tried to get permission from our Court (and others) to challenge their
sentences. This is understandable because many of them had been sentenced under
a statute they now knew was, in part, unconstitutional. To the best of my
knowledge, our Court applied Johnson to give relief to inmates exercising their
direct appeal rights, and to those who were within the one-year time limit for
*72
bringing their first collateral attack. E.g., Mays v. United States,
However, prisoners who were already more than a year into serving sentences impacted by Johnson have had a hard time getting their cases considered in the Eleventh Circuit. Shortly after Johnson was decided, this Circuit declined to
allow the Supreme Court’s new rule in Johnson to serve as the basis for a second
or successive motion under § 2255. In re Franks,
2016), abrogated by Welch v. United States,
As I discussed in my dissent in Franks, the Supreme Court’s decisions in
Schriro v. Summerlin,
Turns out we were wrong. Just shy of a year after Johnson issued, the
Supreme Court decided Welch v. United States,
Fortunately, the Supreme Court was quick to correct this Court’s erroneous
decision on retroactivity—Welch issued just over eight months after Rivero and
three months after Franks. But in the meantime our Court had turned away dozens
of prisoners seeking authorization to file second or successive § 2255 motions
*75
based on Johnson. See In re Robinson,
Thus, in a compressed timeframe, the Court began reviewing thousands of applications from inmates seeking to file a second or successive § 2255 petition based on Johnson’s retroactivity. The statute requires these applications to make a “prima facie” showing that they are entitled to relief. 28 U.S.C. §§ 2244(b)(3)(C), 2255(h). I will now review how this Circuit has implemented this statutory requirement for Johnson applications.
II. INSTITUTING MERITS REVIEW OF APPLICATIONS SEEKING
AUTHORIZATION TO FILE SECOND OR SUCCESSIVE § 2255
MOTIONS RAISING JOHNSON CLAIMS
Again, when prisoners apply to circuit courts for authorization to file a
second or successive habeas motion, the governing statute limits our review of the
application to determining only whether the prisoner has made a “prima facie
showing” that his proposed motion “contain[s] . . . a new rule of constitutional law,
made retroactive to cases on collateral review by the Supreme Court, that was
previously unavailable.” 28 U.S.C. §§ 2244(b)(3)(C), 2255(h). I had occasion to
*76
discuss more fully in my concurrence in In re Williams,
Up until the rush of Johnson filings, this Court only reviewed filings from
prisoners asking for permission to bring a second or successive petition to see
whether the prisoner had made a prima facie case. We did this because, again, this
is the task the statute assigns us. See, e.g., In re Moss,
Cir. 2013) (explaining that the panel’s conclusion that the prisoner had made a prima facie showing was “a limited determination” and the District Court would need to do a de novo review); see also Jordan v. Sec’y, Dep’t of Corr., 485 F.3d
1351, 1357–58 (11th Cir. 2007) (holding that District Courts must review de novo
whether a petitioner has actually satisfied the requirements to file a second or
successive application because the statute “restricts us to deciding whether the
petitioner has made out a prima facie case of compliance with the § 2244(b)
requirements”); In re Joshua,
Before the rush of Johnson filings, we also recognized the practical challenges of conducting a merits review by looking only at the filings asking for permission to bring the action:
When we make that prima facie decision we do so based only on the petitioner’s submission. We do not hear from the government. We usually do not have access to the whole record. And we often do not have the time necessary to decide anything beyond the prima facie question because we must comply with the statutory deadline. See § 2244(b)(3)(D) (requiring a decision within 30 days after the motion is filed).
Jordan,
Yet after Johnson, this Court began doing exactly what we had previously
explained the constraints on our prima facie review would not permit: deciding the
merits of a prisoner’s Johnson claim. See, e.g., In re Thomas,
(11th Cir. 2016). And the Court did this based on the limited filings, which the *78 prisoner understood was just a permission-seeking document intended to allow him to present the merits of his claims in District Court.
The questions that come up about sentences after Johnson are complicated. Every ACCA sentence is based on the individual criminal history of the people serving them. Under ACCA, a prisoner who had in the past committed three or more crimes that qualified as either a “serious drug crime” or a “violent felony” got a sentence that had to be at least fifteen years long. 18 U.S.C. § 924(e). Most are longer. Yet ACCA allows no more than a ten-year sentence for a prisoner who had fewer than three such earlier criminal convictions. Id. § 924(a)(2). Thus, the devil is in the details. Johnson required the federal courts to do the tedious work of examining the criminal histories of people who have been convicted of violating a seemingly endless list of state crimes, defined by many different states in many different ways. In many cases, this Court undertook this complicated review based merely on the prisoner’s application to file a second or successive § 2255 motion.
In so doing, this Court effectively reimposed sentences on these inmates
here at the court of appeals level, without ever allowing more thorough District
Court review. We turned away prisoners seeking District Court review by ruling
that their criminal history that had qualified for a fifteen-year plus sentence still
qualified them for the longer sentence under a part of ACCA that survived
Johnson. See, e.g., In re Hires,
As it turned out, the language the Supreme Court invalidated in Johnson also
appeared in other federal sentencing statutes, so our Court began to get filings from
prisoners sentenced under those statutes as well. For example, prisoners like Ms.
Ovalles also sought to challenge sentences imposed under 18 U.S.C. § 924(c), the
statute at issue in this case. Section 924(c) contains language very similar to that
declared unconstitutional in Johnson and “identical” to that declared
unconstitutional in Sessions v. Dimaya,
Also as a result of Johnson’s holding that the language of ACCA was
unconstitutional, it should have surprised no one that our Court would begin to see
challenges to sentences imposed under identical language in the U.S. Sentencing
Guidelines. USSG § 4B1.2(a). Yet here again, where prisoners sought to challenge
their sentences under this guideline, our Court denied them the chance to present
their claims in District Court. In a direct appeal from a sentencing, with full
adversarial testing, this Court said Johnson did not apply because the sentences
were imposed under a Guideline system that was advisory. United States v.
Matchett,
such thin pleadings, and with no adversarial testing was the wrong approach. I was also coming to understand that some of our merits rulings were wrong, and there was little to no recourse for inmates who got bum rulings.
[2] Though I advocated that Johnson applied to the advisory guidelines, see In re Clayton, 829 F.3d
1254, 1256 (Martin J., concurring in the result), our Court’s holding to the contrary carried the
day. See Beckles v. United States,
During this time, many panels, including some I served on, chose to publish
our rulings on these applications. See, e.g., In re Smith,
In a short time span, our Court got thousands of authorization applications raising Johnson claims. But once any panel published a decision holding an inmate’s past conviction still counts as a predicate conviction under the ACCA, it became easier to dispose of new filings. A panel receiving a new application to file a second or successive petition could quickly cite to a published decision rejecting someone else’s application based on its declaration that the two had identical predicate convictions that had been ruled a “violent felony” even after *82 Johnson. Unfortunately, several of these published merits decisions were just plain wrong. That means not only did the mistaken decision deny relief for the inmate who brought the case, but it will continue to require denial of relief for others who would rightly be entitled to it.
III. EXAMPLES OF INCORRECT MERITS DECISIONS ON SECOND OR SUCCESSIVE APPLICATIONS In the rush to conduct these merits reviews of applications to file a second or successive habeas petition within the 30-day statutory time frame, based only on a form filed by a usually uncounseled prisoner, it should come as no surprise that our Court made some mistaken rulings. I think it worthwhile to discuss three examples of such mistakes in more detail here, to aid in understanding the impact of this Court’s chosen method for responding to prisoners seeking relief after Johnson. Unless we overrule them en banc or the Supreme Court corrects our errors, the merits decisions we made under the constraints I have discussed will continue to bar relief for prisoners with meritorious claims.
A. Florida Robbery and Aggravated Assault
In In re Hires, a panel denied Mr. Hires authorization to file a second or
successive habeas motion based on an earlier conviction imposed on him in
Florida.
To be clear, Turner was also wrongly decided. In United States v. Golden,
As for Florida armed robbery, in Dowd, this Court summarily stated in one
sentence that Florida armed robbery “is undeniably” a violent felony and cited to
ACCA’s elements clause.
B. Johnson’s Applicability to the Mandatory Guidelines
See id. at 1351–52. (discussing change in Florida law in 1997). In another case, Judge Jordan
showed that intermediate Florida courts were affirming robbery convictions despite conduct that
did not involve an act causing the victim to fear death or great bodily harm, calling into question
Lockley’s soundness. See United States v. Lee,
As I’ve mentioned, the United States Sentencing Guidelines contained
language identical to what the Supreme Court ruled unconstitutionally vague in the
ACCA statute. This Court upheld guideline sentences imposed in reliance on this
language under the advisory guideline regime. Matchett,
Guidelines—whether mandatory or advisory—cannot be unconstitutionally vague
because they do not establish the illegality of any conduct and are designed to
assist and limit the discretion of the sentencing judge.” Id. But this is not what
Matchett said. Matchett recognized that the vagueness doctrine, and thus Johnson,
*87
applied to “criminal statutes that define elements of a crime or fix punishments.”
Matchett,
This Court alternatively held in Griffin that Welch did not make Johnson
retroactive on all collateral challenges. The Griffin panel said Johnson was
retroactive only where it applied to result in “a substantive change of law [by]
*88
alter[ing] the statutory range of permissible sentences.” Griffin,
Welch explained that a new rule “has a procedural function” if it “alters only
the procedures used to obtain the conviction.”
Even with these flaws in Griffin, it is binding precedent in our Circuit, and we are required to deny leave to file a second or successive petition to any prisoner seeking relief from a sentence imposed under the mandatory guideline system.
C. Florida Manslaughter
My third example of a merits decision that missed the mark came in In re
Burgest,
But Burgest’s reasoning ignored Eleventh Circuit precedent that requires us to apply the categorical approach when deciding whether a prior conviction is a qualifying offense for sentencing enhancement purposes. See Lockley, 632 F.3d at 1241–42 (noting “robbery” is enumerated as a “crime of violence” in the Guideline commentary to § 4B1.2 and stating “[w]here, as here, the Guidelines specifically designate a certain offense as a ‘crime of violence,’ we compare the elements of *90 the crime of conviction to the generic form of the offense as defined by the States, learned treatises, and the Model Penal Code”). When it failed to use the categorical approach, the panel merely relied on the label Florida uses. This practice was explicitly rejected by this Court in Palomino Garcia. See 606 F.3d at 1327–29 (rejecting the government’s argument that a 16-level enhancement under Guideline § 2L1.2 should be affirmed merely because the State called the offense “aggravated assault” and the Guideline definition enumerated “aggravated assault” as a “crime of violence”). A blind reliance on how any given state law labels a crime leads to prisoners who engaged in similar conduct being sentenced differently. That is because conduct criminalized in one state may not be in another. For example, if Florida labeled the act of crossing the street outside of the crosswalk “manslaughter,” that would be enough for the Burgest panel to conclude jaywalking was a “crime of violence.” Our precedent should have stood to defeat this result.
Where other circuits used the categorical approach, they found that Florida’s
manslaughter statute encompasses more acts than those captured in the generic
definition. See United States v. Mendoza-Padilla,
Hires, Griffin, and Burgest are examples of how this Court’s approach to
second or successive applications has harmed prisoners serving long sentences
imposed under ACCA or the Sentencing Guidelines. The Court’s approach in
these cases disregarded our own precedent and rules in search of a quick end to the
flood of applications filed by prisoners raising Johnson claims. How did this
happen? For one thing, I believe the Court tried to get to the merits of these cases
*92
too fast. The statute for reviewing these applications sets a 30-day limit for ruling
on them. 28 U.S.C. § 2244(b)(3)(D). Once our Court decided to get to the merits
at the mere “seeking permission” stage, we should have joined seven of our sister
circuits and held that the 30-day timeframe was not mandatory. See In re Johnson,
Our nation has experienced an explosion in its prison population over the last twenty years or so. [5] The Supreme Court’s decision in Johnson required federal judges to take a close look at the sentences of thousands of people incarcerated in federal prison. For many inmates, we did not. In our haste, this Court made mistakes (myself included) that always seemed to work against the prisoner. Yet this Court has not just failed to correct its mistakes. It has also acted to set these incorrect decisions in stone.
IV. BARRING MULTIPLE APPLICATIONS TO FILE § 2255 MOTIONS
During the two or so months after the Supreme Court decided Welch, some prisoners were savvy enough to apply to seek relief more than once. This Court then undertook to put an end to this practice, and in doing so, adopted an interpretation of AEDPA that is far removed from the text of the statute. And beyond our Court’s departure from what the statute requires, it also overlooked Supreme Court precedent counseling a path quite different from what this Court chose.
Until recently, federal prisoners at least had the comfort of knowing that
when they were denied leave to file second or successive habeas motions, that
denial was without prejudice and “there [was] no rule against filing multiple
applications for leave to file a successive § 2255 petition.” See In re Anderson,
In order to reach this result, the Baptiste panel first determined § 2244(b)(1)
applies to § 2255 motions, even though § 2244(b)(1) only expressly refers to
petitions for habeas corpus relief filed by state prisoners under 28 U.S.C. § 2254.
(11th Cir. 2016) (Rosenbaum and Jill Pryor, JJ., concurring in result). I add only
that my best understanding is that Baptiste’s interpretation is contrary to Supreme
Court precedent. In Magwood v. Patterson,
The Supreme Court also explained in Magwood that § 2254’s “requirement
of custody pursuant to a state-court judgment distinguishes § 2254 from other
statutory provisions authorizing relief from constitutional violations—such as
§ 2255, which allows challenges to the judgments of federal courts.” 561 U.S. at
333,
Nonetheless, the judges of this Court are bound to apply Baptiste. The harm
it does is not hard to see. For example, I will return briefly to Mr. Hires’s case. In
re Hires,
In this way, Baptiste requires a prisoner to bear all consequences of any mistake this Court makes in his case. It would be one thing if this were what AEDPA required of us. But it is not. Baptiste itself was wrongly decided.
To the extent that Baptiste was issued as a work-load management tool for
the courts, it is not necessary. As discussed, AEDPA already allows a very narrow
path for courts to revisit decisions on authorization applications. See 28 U.S.C. §
2244(b)(3)(E) (barring appeals, petitions for rehearing, and petitions for a writ of
certiorari from the grant or denial of authorization to file a second or successive
petition). In fact, the only approach this Court has identified is a decision to rehear
the case nostra sponte by an en banc court. See In re Lambrix,
V. A FINAL ROADBLOCK TO RELIEF
I end with the Court’s recent refusal to reconsider United States v. Beeman,
2018). In Beeman, a panel of this Court held that to prove a Johnson claim, a
prisoner had to prove that “it was use of the residual clause that led to the
sentencing court’s enhancement of his sentence.”
Of course, before Johnson District Courts did not say at sentencing which
ACCA clause they relied on because nothing in the law required them to. See
Beeman,
Beeman said its historical-fact test was necessary to preserve the “burden of
proof.”
For those rare prisoners who somehow made it past this Court’s review of
their authorization applications and through the District Court’s front doors, they
will face one last, likely fatal, roadblock. District Courts will now decide whether
prisoners should get the benefit of Johnson without being able to consider
developments in that law intended to help them evaluate who qualifies as a violent
*100
repeat offender. In the end, of the thousands of inmates who filed authorization
applications raising potentially meritorious Johnson claims, very few will ever get
a full review of the merits of their claims and even fewer will get relief. See
Williams,
VI. TODAY’S DECISION IN CONTEXT
As this review shows, our Court has made many missteps after Johnson. It all began, in my view, when we started conducting merits review of prisoners’ cases when all they wanted was permission to file a second or successive § 2255 motion in District Court. We compounded that error when we started publishing those decisions, establishing them as precedent. Then we locked in our mistakes by adopting an interpretation of § 2255, not in keeping with the text of the statute, to prevent prisoners from filing more than one application to file a second or successive petition, even where we wrongly denied their first. And then we denied those prisoners who managed to get full District Court review of their sentences the ability to use current Supreme Court precedent to show they had been sentenced under the ACCA residual clause. At every turn, our rulings put obstacles in the paths of prisoners trying to have their sentences reviewed. These roadblocks mean some prisoners in our Circuit will serve unconstitutional sentences.
Today’s majority en banc opinion decides that the Supreme Court’s decisions in Johnson and Dimaya did not invalidate § 924(c)’s residual clause. Again, I fully join Judge Jill Pryor’s dissent explaining why this is wrong. I note with regret that today’s decision is but a piece of the precedent I reviewed here. This Court has mishandled the application of Johnson to many people who should have benefitted from it, and we do so again today.
My final observation about the majority’s en banc ruling against Ms. Ovalles is to note that the majority opinion makes much of the fact that the government has asked us to abandon the categorical approach in interpreting § 924(c)(3)(B). See Maj. Op. at 29–30. Judge Jill Pryor’s dissent explains why this consideration should not factor into our analysis of the statute at issue. Jill Pryor Dissent at 144– 45. I would add that, when deciding whether Johnson was retroactive, we paid no heed to the government’s concession that it was. See supra at 1–2. If we are going to defer to the government’s view, we should do so whether it advocates for or against relief for the prisoner.
I respectfully dissent from the majority opinion, particularly as it evidences this Court’s failure to properly review and correct unlawful sentences in the wake of Johnson.
*102 JILL PRYOR, Circuit Judge, with whom WILSON, MARTIN, and JORDAN, Circuit Judges, join, dissenting:
This case is about 18 U.S.C. § 924(c), a frequently charged federal criminal
statute, and how one recent Supreme Court case in particular,
Sessions v. Dimaya
,
Before I get to why I think the majority is wrong, let me point out where
we agree. We courts are loath to strike down acts of Congress—and appropriately
so. When a law passed by Congress may be read to have more than one plausible
*103
or “fairly possible” meaning, and one interpretation would cause the law to be
upheld as constitutional while another would require it to be struck down, we are
“obligated to construe the statute” so as to uphold it.
I.N.S. v. St. Cyr.
, 533 U.S.
289, 300 (2001). This is the canon of constitutional avoidance, or constitutional
doubt, as the majority calls it. The Supreme Court recently clarified a significant
limitation on the use of this rule of statutory construction, however: “In the
absence of more than one plausible construction, the canon simply has no
application.”
Jennings v. Rodriguez
,
This case—with all its textual analysis, discussion of canons of statutory construction, and parsing of precedent—may come across like a purely academic exercise. In reality, it is anything but. People who are serving sentences of five *104 years to life under § 924(c) will get no relief from this Court even though the Supreme Court held that an identically-worded statute was so vague that its enforcement violated the right to due process under law. For the reasons I explain in more detail below, I respectfully dissent. [1]
* * *
Section 924(c) makes it a federal crime to use or carry a firearm “during and
in relation to any crime of violence or drug trafficking crime” or to possess a
firearm in furtherance of such a crime. 18 U.S.C. § 924(c)(1)(A). A “crime of
violence,” under the provision at issue in this case, is one that “by its nature,
involves a substantial risk that physical force against the person or property of
another may be used in the course of committing the offense.”
Id
. § 924(c)(3)(B).
The punishment for violating § 924(c) is from five years to life in prison,
depending on the nature of the firearm and how it is used, carried, or possessed.
When reviewing this same crime of violence definition in another federal statute,
18 U.S.C. § 16(b), in the face of a challenge that it was unconstitutionally vague,
the Supreme Court in
Dimaya
applied the “categorical approach” to interpreting
the statute.
See Dimaya
,
No one disputes that if we must use the categorical approach to interpret § 924(c)(3)(B), the statute is unconstitutional. See Maj. Op. at 2. The majority argues that we should employ the canon of constitutional avoidance to save § 924(c)(3)(B) because the statute can plausibly be interpreted to incorporate a “conduct-based approach”—one that looks to the “real-world facts of the defendant’s offense” instead of how the law defines the crime or what conduct is entailed in the ordinary case. Maj. Op. at 3. The problem with the majority’s approach is that its alternative way of reading § 924(c)(3)(B) to permit a conduct- based approach is simply not plausible when we remain faithful to the text of the statute.
So how does the majority get there? Sifting through Supreme Court cases applying the categorical approach to ACCA and § 16, the majority identifies six factors that led the Supreme Court to apply the categorical approach in those *106 instances. It then applies those factors to § 924(c)(3)(B) and—finding many of them missing or inapplicable—determines that this statute plausibly could be read differently. But only three of the majority’s factors concern the text. And the majority’s analysis of its factors, both textual and extra-textual, is in my view deeply flawed. If we follow what the Supreme Court said about the text of the other two statutes and apply it to § 924(c) as a whole and in context, we find that § 924(c)(3)(B) presents an even stronger case for applying the categorical approach than the other statutes—so strong that no other reading is plausible.
I.
To understand why the approach the Supreme Court used in Dimaya , Johnson , and other precedent interpreting similarly-worded legislation controls the outcome of this case, some background on that precedent and the other legislation is unavoidable. In this section I first discuss three statutes, all of which Congress enacted or revised as part of an overhaul of the criminal code in the mid-1980s: the statute at issue here, 18 U.S.C. § 924(c), particularly its definition of “crime of violence”; a definition of “crime of violence” found in 18 U.S.C. § 16 that is applied in numerous criminal and noncriminal contexts; and ACCA, 18 U.S.C. § 924(e). Next I review a line of Supreme Court and circuit decisions that have explained the necessity of construing these statutes using the categorical approach. Finally, I discuss the decisions striking as unconstitutionally vague similar *107 language in ACCA and identical language in § 16, as well as decisions framing the limits of the canon of constitutional avoidance. These decisions bear directly on whether § 924(c)(3)(B) is unconstitutionally vague.
A. The Statutes
In the mid-1980s, Congress, as part of a movement to get tougher on crime,
passed the Comprehensive Crime Control Act of 1984 (the “CCCA”), which
overhauled the federal criminal code for the first time in over half a century. Pub.
L. No. 98-473, 98 Stat. 1976 (1984);
see also
S. Rep. No. 98-225, at 2 (1983)
(discussing the CCCA’s purpose of “restor[ing] a proper balance between the
forces of law and the forces of lawlessness”). The CCCA “broadly reformed the
federal criminal code in such areas as sentencing, bail, and drug enforcement, and
. . . added a variety of new violent and nonviolent offenses.”
Leocal
,
The statute at issue here, 18 U.S.C. § 924(c), revised in the CCCA, [2] criminalizes and imposes mandatory enhanced sentences for using or carrying a firearm “during and in relation to any crime of violence or drug trafficking crime” or possessing a firearm in furtherance of such a crime. 18 U.S.C. § 924(c)(1)(A). The “crime of violence or drug trafficking crime” must be committed “during”— that is, at the same time as—the firearm use or possession. “[D]rug trafficking crime” is defined in § 924(c) as “any felony punishable under the Controlled Substances Act” and a couple of other federal drug laws. Id. § 924(c)(2). “[C]rime of violence”—the focus of our inquiry—has a two part definition:
For purposes of this subsection the term “crime of violence” means an offense that is a felony and—
(A) has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or (B) that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
Id. § 924(c)(3). Subsection (c)(3)(A) of this definition is known as the “elements clause,” and subsection (c)(3)(B) is known as the “residual clause.” This case is about the residual clause. The residual clause acts as a catchall for violent crimes *109 that do not meet the elements clause definition, so long as those crimes satisfy the residual clause’s criteria.
Section 924(c) is one of several CCCA statutes to address crimes of
violence. Another is 18 U.S.C. § 16, which “provide[s] . . . a general definition of
the term ‘crime of violence’ to be used throughout the [CCCA],” including “for
defining the elements of particular [criminal] offenses.”
Leocal
,
The term “crime of violence” means— (a) an offense that has as an element the use, attempted use, or threatened
use of physical force against the person or property of another, or (b) any other offense that is a felony and that, by its nature, involves a
substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
18 U.S.C. § 16. Section 16’s residual clause, § 16(b), which is “essentially identical to § 924(c)(3)’s residual clause at issue here,” Maj. Op. at 10, was struck down as unconstitutional in Dimaya .
Since passage of the CCCA, § 16’s general “crime of violence” definition
“has . . . been incorporated into a variety of statutory provisions, both criminal and
noncriminal.”
Leocal
,
In fact, for the vast majority of instances where § 16’s definition is incorporated into the criminal code, the “crime of violence” element is committed at the same time as the offense’s other elements. For example, 18 U.S.C. § 25 criminalizes and provides penalties for an adult “who intentionally uses a minor to commit a crime of violence,” as that term is defined in § 16. A person who violates § 25 commits the “crime of violence” element at the same time as the *111 offense’s other elements. [3] And so many incorporations of § 16 function very similarly to the incorporation of the “crime of violence” definition in § 924(c)(3)— the elements and residual clauses—into § 924(c)(1)—which contains the offense’s other elements: using, carrying, or possessing a firearm during (at the same time as) and in relation to (to facilitate) the predicate crime of violence or drug trafficking crime. [4]
Yet another (and for our purposes, the last) CCCA statute to address crimes of violence is ACCA. ACCA, enacted as part of the CCCA, imposes a mandatory 15-year term of imprisonment on a person convicted of being a felon in possession of a firearm who previously was convicted of three serious drug offenses, violent felonies, or both. ACCA defines “violent felony” like this:
[A]ny crime punishable by imprisonment for a term exceeding one year . . . that—
(i) has as an element the use, attempted use, or threatened use of physical force against the person of another; or (ii) is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another[.]
Id. § 924(e)(2)(B). Like § 924(c) and § 16, ACCA has an elements clause, subsection (i), and it had—until the Supreme Court struck it down as unconstitutionally vague—a residual clause, which in subsection (ii) began with “or otherwise . . . .” The balance of subsection (ii) is known as the “enumerated crimes clause.”
Section 924(c), § 16, and ACCA, all either enacted or revised as part of one comprehensive piece of legislation, are notably similar in their text. All three statutes, until the majority’s decision today, have been analyzed using the same “categorical approach.” I discuss this approach in the next subsection.
B. The Categorical Approach that Has Governed All Three Statutes
When Congress passed the CCCA, in which it revised § 924(c) to include a
“crime of violence” definition and enacted § 16 and ACCA, it did so with the
expressed intention to capture certain crimes—not conduct—as crimes of violence.
See generally Taylor
,
When in Taylor the Supreme Court first described the categorical approach, it did so in the context of ACCA’s enumerated crimes clause. The Court based its holding—that a categorical, rather than a factual or conduct-based, approach must be used—primarily on ACCA’s text as a whole, but also on ACCA’s legislative history and because construing it with a conduct-based approach would lead to practical problems in criminal proceedings. The Court first explained that ACCA’s text “generally supports” a categorical approach because it refers to “convictions,” not commissions of an offense, and because its elements clause refers to a statute’s “element[s],” not to any particular facts or conduct. Id. “Read in this context,” the Court said, the enumerated crimes clause should be read categorically. Id. at 600-01. Second, the Court noted, “the legislative history of [ACCA] shows that Congress generally took a categorical approach to predicate offenses.” Id. at 601. Third, the Court observed that “the practical difficulties and potential unfairness of a factual approach are daunting.” Id. Specifically, the Court worried that a conduct-based approach would lead to mini-trials where the government would seek to prove to a jury, and the defense would seek to rebut, the circumstances of prior—maybe even long prior—offenses. Id. at 601-02. The Court also doubted that a sentencing court, rather than a jury, would be able to make these findings because doing so might draw a Sixth Amendment challenge. Id. at 602.
After
Taylor
, the Supreme Court held that the categorical approach must be
applied to ACCA’s elements and residual clauses as well.
See Curtis Johnson v.
United States
,
In contrast to its multi-factored reasoning in
Taylor
, when the Supreme
Court first construed § 16, it applied the categorical approach based
solely
on the
text of the statute, without reference to legislative history or practical concerns.
And it said this construction was not merely optimal, but required. Specifically, in
Leocal
, a unanimous Supreme Court explained that “[i]n determining whether the
petitioner’s conviction falls within the ambit of § 16, the statute directs our focus to
*116
the ‘offense’ of conviction.”
Leocal
,
When it came to construing § 924(c)(3), in
United States v. McGuire
this
Court used the same textual analysis to explain why the categorical approach must
be used.
We employ this categorical approach because of the statute’s terms: It asks whether [the defendant] committed “an offense” that “has as *117 an element the use, attempted use, or threatened use of physical force against the person or property of another,” or that “ by its nature , involves a substantial risk that physical force against the person or property of another may be used.”
Id. (emphasis in original) (quoting 18 U.S.C. § 924(c)(3)). McGuire held that § 924(c), just like § 16, by its terms required a categorical approach. Since McGuire , our Court has continued to apply the categorical approach to § 924(c)(3)(B) until today, and we have been in good company doing so. All but two circuits apply the categorical approach to § 924(c). [6]
C. Johnson , Dimaya , and the Role of the Canon of Constitutional
Avoidance
The majority agrees that if the categorical approach must be used, then based on Dimaya § 924(c)’s residual clause is unconstitutionally vague. The question is whether the Supreme Court’s decisions—spanning from 1990 to 2018—that the categorical approach must be applied to language like § 924(c)’s can be side- stepped due to the canon of constitutional avoidance. In this section I discuss the Supreme Court’s reasoning in these decisions. I also discuss the evolution of the *118 canon of constitutional avoidance in this setting and how a minority of the Justices have advocated unsuccessfully for its use to save ACCA’s and § 16(b)’s residual clauses.
After wrestling with ACCA’s residual clause on several occasions, the
Supreme Court in
Johnson
struck it as void for vagueness.
Justice Alito dissented in Johnson . He argued that the Court should have employed the canon of constitutional avoidance to save ACCA’s residual clause by adopting a conduct-based approach like the majority adopts here. Id. at 2578 (Alito, J., dissenting). Justice Alito noted that ACCA’s residual clause referred to “ conduct that presents a serious potential risk of physical injury to another” and therefore plausibly could refer “to things done during the commission of an offense that are not part of the elements needed for conviction.” Id. (emphasis added). The majority rejected his approach, noting that the government had not urged the Court to abandon the categorical approach and explaining that “ Taylor had good reasons to adopt” it, including the statute’s text and practical problems resulting from a conduct-based approach. See id. at 2562 (majority opinion).
When
Johnson
was decided, though, it was an open question whether extra-
textual factors—such as the government’s concessions and practical
considerations—could play a role in the decision whether to invoke the canon of
constitutional avoidance to save a statute. In
Zadvydas v. Davis
, the Supreme
Court construed an INA statute providing that a noncitizen under an order of
removal whom the government has not removed within the 90-day statutory
removal period “‘may be detained beyond the removal period.’”
This year, however, the Supreme Court put any lack of clarity in
Zadvydas
to rest. In
Rodriguez
, the Court clarified that “[t]he canon of constitutional
avoidance comes into play
only
when, after the application of ordinary textual
analysis, the statute is found to be susceptible of more than one construction. In
the absence of more than one plausible construction, the canon simply has no
application.”
Rodriguez
,
Later in the term, the Supreme Court struck § 16’s residual clause as
unconstitutionally vague.
See Dimaya
,
Again, as in
Johnson
, a dissent argued that the Court should have employed
the canon of constitutional avoidance to save the residual clause.
Id.
at 1254-59
(Thomas, J., dissenting). Justice Thomas, joined by Justices Kennedy and Alito,
announced that “[t]he text of § 16(b) does not require a categorical approach,”
id.
at 1254, despite
Leocal
’s holding that the text of § 16(b) “requires” a categorical
approach.
Leocal
,
II.
Section 924(c)’s residual clause is void for vagueness for precisely the same
reasons that its twin, § 16(b), is void for vagueness. First, the categorical
approach—which the Supreme Court said must be applied to text identical to
§ 924(c)(3)(B)—requires courts to imagine the ordinary case of an offense. And
“[n]othing [in the statute] helps courts to perform that task.”
Dimaya
, 138 S. Ct. at
1215 (majority opinion). Second, the “substantial risk” standard in § 924(c)(3)(B)
is indeterminate at best, creating “uncertainty about the level of risk that makes a
crime ‘violent.’”
Id.
Combined, these two features of § 924(c)(3)(B) create
“‘more unpredictability and arbitrariness than the Due Process Clause tolerates.’”
Id.
at 1216 (quoting
Johnson
,
The majority, in an attempt to salvage § 924(c)’s residual clause, tosses out the categorical approach in the name of constitutional avoidance. The majority *124 acknowledges that the categorical approach must be applied to the definitions of drug trafficking crime and of crime of violence in § 924(c)(3)’s elements clause, but holds that a conduct-based approach at least plausibly applies to the definition of crime of violence in § 924(c)(3)’s residual clause. In adopting a conduct-based approach to § 924(c)(3)(B)—and thus overruling in part our decision in McGuire —the majority contorts the plain text of the statute and reads similar structure and language differently within the same statute. The majority does so in disregard of the Supreme Court’s unequivocal statement in Leocal that language identical to § 924(c)(3)(B) requires a categorical approach—that is, it cannot plausibly be read another way. And the majority does so in disregard of the Supreme Court’s admonition that in deciding whether to employ the canon of constitutional avoidance we must look to a statute’s text alone. The majority effectively rewrites the statute to avoid having to strike it down.
To reach this result, the majority distills from
Taylor
and its progeny a
multi-factor test for deciding whether we should apply the categorical approach to
§ 924(c)(3)(B). Only three of the majority’s six factors are text-based, and they
appear seemingly on an even playing field with three other extra-textual
“considerations.” Maj. Op. at 27-28. Even if these extra-textual considerations
mattered before
Rodriguez
, they do not matter now. What we know now is that if
a purely textual analysis leads to only “one plausible construction, the canon [of
*125
constitutional avoidance] simply has no application.”
Rodriguez
,
A. The Text of Section 924(c) Permits Only One Reading: that the
Categorical Approach Is Required.
The question we must ask in this case is whether the text of § 924(c)(3)(B) plausibly can be read to support a conduct-based approach rather than a categorical one. If not, the inquiry ends, and § 924(c)(3)(B) is unconstitutional. The answer is that it cannot. Section 924(c)’s text, read as a whole, in its statutory context, and with the Supreme Court’s guidance, cannot plausibly be read to permit a conduct- based approach to the residual clause.
First of all, the Supreme Court has already told us that a categorical
approach is the only approach we can take to this text. In
Leocal
, the Supreme
Court unanimously explained that the text of § 16(b)—“identical” to the language
in § 924(c)(3)(B), Maj. Op. at 10, 16——“
requires
us to look to . . . the nature of
the offense of conviction, rather than to the particular facts relating to [the] crime”
because the text includes “offense” in conjunction with “by its nature.”
Leocal
,
The reasoning behind Leocal ’s holding applies with at least equal force to § 924(c). To recap, § 924(c)’s “crime of violence” definition looks like this:
For purposes of this subsection the term “crime of violence” means an offense that is a felony and—
(A) has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or (B) that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.
18 U.S.C. § 924(c)(3). “[O]ffense that is a felony,” language that required a
categorical approach to § 16(b), is part of the definition of “crime of violence” in
both the residual clause
and
the elements clause, and no one disputes that the latter
requires a categorical approach. “[W]here . . . Congress uses similar statutory
language and similar statutory structure in two adjoining provisions, it normally
intends similar interpretations.”
Nijhawan v. Holder
,
There is additional textual support for a categorical reading of
§ 924(c)(3)(B) beyond the “crime of violence” definition. The categorical
approach also applies, without a doubt, to the other possible predicate offense, a
*128
“drug trafficking crime.”
See
18 U.S.C. § 924(c)(2). That the categorical
approach applies to the drug trafficking crime definition as well as the elements
clause’s crime of violence definition is evidence that the same approach must also
apply to the residual clause’s crime of violence definition.
Nijhawan
, 557 U.S. at
39;
see U.S. Bank of Or. v. Indep. Ins. Agents of Am., Inc.
,
The text of § 924 as a whole further supports a categorical reading of
§ 924(c). Section 924 also contains ACCA, § 924(e). The statutes are unique in
establishing mandatory minimum sentences for federal firearms offenses.
See
U.S.
Sentencing Comm’n,
Mandatory Minimum Penalties for Firearms Offenses in the
Federal Criminal Justice System
2 (2018). The Supreme Court has, without
exception, construed ACCA (including, when it was in effect, its residual clause)
to require a categorical approach. Taking an even broader look at the statutory
scheme, we know that § 924(c), ACCA, and § 16 all were created or amended as
part of the same legislative action, the CCCA. The use of very similar residual
clauses in these statutes is more evidence still that Congress intended the same
categorical meaning for § 924(c)(3)(B).
See Rodriguez
,
None of the majority’s three text-based factors convinces me that § 924(c)(3)(B) plausibly can be read to support a conduct-based approach. Nor am I persuaded that we can divine the meaning of one clause of a statute by breaking it down word by word and isolating those words from their surrounding context. See Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts § 2, at 56 (2012) (“The words of a governing text are of paramount concern, and what they convey, in their context, is what the text means.”). Indeed, by separating the text of § 924(c) into three “considerations” that are hardly separate ideas, the majority performs a sleight of hand that, the Supreme Court has told us on numerous occasions, has no place in textual analysis. See, e.g. , Rodriguez , 138 S. Ct. at 845 . Even picked apart, though, the words do not support majority’s interpretation.
The majority’s three textual considerations or factors are: “the text of [ACCA’s] operative provisions focused not on conduct, but rather on ‘convictions’—and thus . . . solely on formal legal elements” (factor 2 of the majority’s six-factor test); ACCA’s and § 16’s “definitional provisions used terms and phrases like ‘offense,’ ‘felony,’ and ‘by its nature,’ which . . . pointed toward a categorical (rather than conduct-based) inquiry” (factor 3); and “those statutes *130 lacked any reference to the underlying crime’s commission or circumstances” (factor 4). Maj. Op. at 27. I address them in turn.
1. The “Operative Provision” of Section 924(c) Does Not Permit a Conduct-Based Approach to the Residual Clause.
In its analysis of § 924(c)’s text, the majority first notes that what it deems ACCA’s “operative provision”—“[a]ny person who violates [18 U.S.C. § 922(g)] and has three previous convictions . . . for a violent felony or a serious drug offense, or both”—refers to “convictions” rather than underlying conduct, yet there is no reference to “convictions” in § 924(c)’s “operative provision,” § 924(c)(1). Rather, the majority says, § 924(c)(1) “refers to conduct ,” Maj. Op. at 30-31 (emphasis in original), pointing to the language that “any person who, during and in relation to any crime of violence . . . uses or carries a firearm, or who, in furtherance of any such crime, possesses a firearm,” id. (quoting § 924(c)(1)(A)). According to the majority, this language offers strong support for applying a conduct-based approach to the statute’s residual clause.
I find the majority’s argument unpersuasive for at least two reasons. First,
setting aside for now comparisons to other statutes, the language in § 924(c)(1) that
the majority insists “refer[s] exclusively to conduct,” Maj. Op. at 37, cannot bear
the weight the majority places on it. Section 924(c)(1) makes it a federal crime to
use, carry, or possess a firearm “during and in relation to” one of
three
categories
of crimes: drug trafficking crimes, elements-clause crimes of violence, or residual-
*131
clause crimes of violence. If the “during and in relation to” language supported a
conduct-based approach to the residual clause, it necessarily would do the same for
the drug trafficking and elements-clause definitions. After all, the same words—
“during and in relation to”—must be ascribed the same meaning for each category
of crime listed in § 924(c), whether a drug trafficking crime, an elements- clause
crime of violence, or a residual-clause crime of violence. “To give these same
words”—[“during and in relation to”]—“a different meaning for each category”—
[the elements- and residual-clause definitions of crime of violence and the drug
trafficking crime definition]—“would be to invent a statute rather than interpret
one.”
Clark v. Martinez
,
Second, and relatedly, by comparing § 924(c)(3)(B) to ACCA the majority is
in some sense setting up a straw man, because ACCA is not the best analogue to
§ 924(c)(3); § 16 is. As the
Dimaya
plurality noted, § 16’s text, even “more
plainly” than ACCA’s, requires a categorical approach.
Dimaya
, 138 S. Ct. at
1218 (plurality opinion);
see id.
at 1235-36 (Roberts, C.J., dissenting) (reaffirming
Leocal
). So to compare and contrast ACCA’s language with § 924(c)’s is a bit
*132
disingenuous when a more similar statute exists. The majority does not look to
§ 16(b) when examining this textual factor, perhaps because the majority confines
all discussion of § 16(b) to its incorporation into the INA, where it, like ACCA’s
residual clause, defined a prior conviction. But § 16(b) in its numerous other
applications functioned just as § 924(c)(3)(B) does, yet § 16(b) “require[d]” a
categorical approach.
Leocal
,
A look at some examples of § 16’s incorporation into the criminal code proves both points. By examining § 16’s incorporation into the criminal code, we see how conduct-based elements and categorical elements exist side-by-side without the former allowing a conduct-based reading of the latter. And we see just how similar § 16(b)’s incorporations are to § 924(c)(3)(B)’s incorporation into § 924(c)’s “operative provision,” and thus why § 16 is a better textual analogue for § 924(c)(3)(B) than ACCA is.
Let’s take, for example, 18 U.S.C. § 25, which, as I mentioned in Part I.A.,
criminalizes an adult’s intentional use of a minor to commit a crime of violence.
To convict a defendant under § 25, a jury must find that the defendant
intentionally
*133
used a minor to commit the crime of violence, an element that necessarily will
require examination of the defendant’s conduct. But the “crime of violence”
element of the statute necessarily is categorical,
see Leocal
,
2. The Phrases “Offense that Is a Felony” and “By its Nature” Require a Categorical Approach to Section 924(c)’s Residual Clause.
Next, the majority examines the words “offense,” “felony,” and “by its nature,” concluding that although in some contexts they may signal a categorical reading, in this context they do not necessarily. I am unpersuaded.
The majority says that “offense” plausibly can be read to refer to the
defendant’s conduct, pointing to other contexts in which “offense” denotes
conduct. In the case the majority relies primarily upon,
Nijhawan
, the Supreme
Court recognized “the linguistic fact” that in ordinary speech words such as
“offense” or “felony” can refer to a person’s conduct rather than a statute of
conviction.
an offense that— (i) involves fraud or deceit in which the loss to the victim or victims exceeds §10,000; or
(ii) is described in section 7201 of Title 26 (relating to tax evasion)
in
which
the revenue loss to the Government exceeds $10,000[.]
8 U.S.C. § 1101(a)(43)(M) (emphasis added). The Supreme Court explained that
the second of these definitions has to refer to the actual circumstances of the
petitioner’s offense because “[t]here is no offense ‘described in section 7201 of
title 26’ that has a specific loss amount as an element.”
Nijhawan
,
As I explained above, applying this same logic to § 924(c)(3) compels a
categorical reading of the statute. Like subsection (a)(43)(M)’s “offense . . . in
which” language, the prefatory phrase “offense that is a felony” in § 924(c) is
incorporated into the definitions for
both
the elements clause
and
the residual
clause. If we start with the uncontroversial proposition that “offense that is a
felony” is categorical in the elements clause and then apply the Supreme Court’s
reasoning in
Nijhawan
, we must conclude that Congress intended “offense that is a
felony” to be categorical for both crime of violence definitions in § 924(c)(3).
See
Nijhawan
,
Aside from its reading of “offense,” the majority appears to acknowledge that the phrase “by its nature” is best read to require a categorical approach. But, *137 says the majority, Justice Thomas, in dissent in Dimaya , makes some good points about the meaning of the word “nature” that make a conduct-based reading possible. “Nature,” the majority says, might mean “the essential character or constitution of something,” and that “something” might be “particular acts” rather than the elements of a crime of conviction. Maj. Op. at 35. The problem with the majority’s argument is that the text of § 924(c) tells us precisely what that “something” is: when the statute says “by its nature,” its clearly refers to “offense that is a felony.” [12] And for the reasons I just explained, “offense that is a felony” must be read categorically. Isolating the word “nature” and claiming that it is amenable, out of context, to more than one meaning does little to inform the meaning of § 924(c)(3)(B) when § 924(c) is read as a whole.
The majority also credits Justice Thomas’s suggestion that “the words ‘by its nature,’ ‘substantial risk,’ and ‘may’ would mean only that an offender who engages in risky conduct cannot benefit from the fortuitous fact that physical force was not actually used during his offense.” Maj. Op. at 36 (quoting Dimaya , 138 S. Ct. at 1254). Taken separately, I agree that these words and phrases could mean that. But together? No way.
Again, the residual clause defines “crime of violence” as “an offense that is a felony and . . . that by its nature, involves a substantial risk that physical force . . . may be used” during commission of the offense. 18 U.S.C. § 924(c)(3)(B). The “substantial risk” requirement on its own means that an offender who engages in risky conduct cannot benefit from the fact that force was not actually used, so the addition of “may” would serve no purpose other than to abstract the defendant’s actual conduct one step further. Of course, we cannot read words out of a statute in the name of constitutional avoidance. See Gustafson v. Alloyd Co. , 513 U.S. 561, 574 (1995) (emphasizing that courts must “avoid a reading [of a statute] which renders some words altogether redundant”).
3. Section 924(c)(3)(B)’s “In the Course of Committing the Offense” Language Does Not Support a Conduct-Based Approach.
Third, in its reach to find another meaning in the statute, the majority
considers as a factor the presence or absence of “any reference to the underlying
crime’s commission or circumstances,” a factor that it says it distills from
Dimaya
.
Maj. Op. at 27. The majority argues that § 924(c)’s residual clause “is different”
from § 16(b) because § 924(c)(3)(B) says “in the course of committing the
offense,” which it deems to be “conduct-based language.”
Id.
at 37. This is flat
wrong—both statutes contain “in the course of committing the offense.”
Therefore, the statutes are not different at all. And if this temporal language
*139
somehow suggested a conduct-based approach, it likewise would do so for the
identically worded § 16(b), and we know from
Leocal
that that is not the case.
We also know from the Supreme Court’s decisions applying the categorical
approach to ACCA that the fact that the risk must arise in the course of committing
the offense does not admit of a conduct-based approach. In
James
, the Supreme
Court explained that the “risk” that triggered ACCA’s residual clause must arise
“while the crime is in progress”—that is, in the course of committing the offense.
James
,
* * *
The majority’s strained reading of § 924(c)’s residual clause does not, in my
view, even approach plausible. Under
Rodriguez
, we must end the analysis there.
The statute cannot be construed to avoid a plain violation of due process. And yet,
“[u]nable to find sure footing in the statutory text, the [g]overnment and [the
majority] pivot away from the plain language and raise a number of practical
concerns. These practical considerations are meritless and do not justify departing
from the statute’s clear text.”
Pereira
,
B. The Majority’s Extra-textual Factors Do Not Support Application of a
Conduct-Based Approach.
Even if we could set aside the text of the statute (which of course we may not do), the majority’s three “practical considerations” lacking a textual hook support no different reading of § 924(c)’s residual clause. To review, these three practical considerations are because: in Johnson and Dimaya the government “never asked the Court to consider a conduct-based approach” (factor 1); “applying the categorical approach would avoid the impracticability of requiring sentencing courts to engage in after-the-fact reconstructions of the circumstances underlying prior convictions” (factor 5); and “applying the categorical approach would avoid the Sixth Amendment issues that could arise from sentencing courts making findings of fact that properly belong to juries” (factor 6). Maj. Op. at 27-28.
Before I get into the majority’s factors, I pause to note that although the
majority claims to have distilled its factors from
Taylor
and its progeny, absent
from the majority’s consideration is one of
Taylor
’s three reasons for applying the
categorical approach to ACCA: that “the legislative history of the enhancement
statute shows that Congress generally took a categorical approach to predicate
offenses.”
The Senate stated that the purpose of its revision was to “ensure that all
persons who commit federal crimes of violence, including those crimes
set forth in
*142
statutes
which already provide for enhanced sentences for their commission with a
dangerous weapon, receive a mandatory sentence.”
Id.
(emphasis added). The
Senate sought to do so in response to the Supreme Court’s body of cases that had
“negated [§ 924(c)]’s use in cases
involving statutes
, such as the bank robbery
statute and assault on a federal officer statute.”
Id.
at 312 (emphasis added).
Congress explained that “[t]hese are precisely
the type of
extremely dangerous
offenses for which a mandatory punishment for the use of a firearm is most
appropriate,” so it overrode the court decisions to ensure that certain
statutes
fell
within the purview of § 924(c).
Id.
(emphasis added);
see also United States v.
Robinson
,
Relatedly, our court and nearly every other federal court in the nation has
consistently applied the categorical approach to § 924(c)(3), and Congress has not
once sought to intervene—despite the fact that, as evidenced by the above
legislative history, Congress previously has substantially revised the statute in
response to the federal courts’ construction of it. “The claim to adhere to case law
is generally powerful once a decision has settled statutory meaning” because
*143
“unlike in the context of constitutional interpretation, the legislative power is
implicated, and Congress remains free to alter what [a] [c]ourt has done.”
Patterson v. McLean Credit Union
,
The canon of constitutional avoidance “rest[s] on the reasonable
presumption that Congress did not intend the alternative which raises serious
constitutional doubts. The canon is thus a means of giving effect to congressional
intent, not of subverting it.”
Martinez
,
I now turn to the majority’s extra-textual factors.
1. Whether the Government Has Asked Us to Abandon the Categorical Approach Has No Relevance to the Constitutional Avoidance Analysis. The first factor the majority considers is that here, unlike in Johnson and Dimaya , the government has argued for a conduct-based approach: “[T]he government has expressly (and at length) urged us to abandon the categorical approach to § 924(c)(3)(B) in favor of a conduct-based interpretation.” Maj. Op. at *144 29. First, this description strikes me as incomplete at best. For years, and even after Johnson , the government consistently has urged that we apply a categorical approach to § 924(c). It was only after our court invited the government to argue for abandonment of the categorical approach in supplemental panel briefing and en banc briefing in this case that the government changed course. And even then, in its en banc brief the government’s primary argument was not that we should overrule McGuire , but rather that we should affirm Ms. Ovalles’s § 924(c) conviction on the ground that her predicate offense satisfies the elements clause. See En Banc Br. of the Appellee USA at ii. The government argued only in the alternative that we should overrule McGuire insofar as it held that a categorical approach must be applied to § 924(c)’s residual clause.
Second, and more importantly, as the majority itself observes, this factor is not really “interpretive,” id. at 29; and so it is “an odd place to start in interpreting a statute.” Id. Under Rodriguez it is clear that the government’s suggestion (or lack of it) that we should apply a conduct-based approach to a statute is irrelevant to whether the text of the statute plausibly can be read to permit such an approach.
2. The Majority’s Practical Problems and Sixth Amendment Factors in Reality Constitute One Irrelevant Factor that Fails to Support the Majority’s Position.
The majority’s fifth factor—the supposed lack of practical problems
associated with a conduct-based approach—is yet another variety of extra-textual
*145
consideration that the Supreme Court in
Rodriguez
told us we cannot consider.
See
Rodriguez
,
It certainly was possible to apply a conduct-based approach to ACCA’s residual clause and § 16(b) and still comply with the Sixth Amendment. See W. Pryor Concurrence at 8-9. Take ACCA’s residual clause, for example. Say a defendant convicted of burglary three times in 2013 was charged in 2014 with being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g). He elected to have his case tried to a jury. The government could secure an ACCA enhancement based on these burglary convictions so long as it proved to the jury beyond a reasonable doubt that the defendant’s commission of the burglaries involved a serious potential risk of injury. (Maybe in each case the defendant was still inside a home when the homeowners returned, and the homeowner in each case could testify to that before the jury.) Securing the defendant’s ACCA- enhanced sentence should be relatively straightforward (if somewhat time- consuming) because the prior convictions were close in time to the § 922(g) charge. Now assume those burglary convictions were 10 years old. Maybe the government could scrape together enough evidence to prove the risk element of *146 those long-ago crimes beyond a reasonable doubt, maybe not. But this is not a Sixth Amendment problem. It is a practical problem for the prosecution that arises because of the Sixth Amendment jury trial right. Applying a conduct-based approach to ACCA would not have violated the Sixth Amendment; it merely would have led to fewer ACCA-enhanced sentences because the government’s burden to prove old facts to a jury could turn out to be difficult to satisfy in practice. [13]
The same is true for § 16(b) as incorporated by the INA—a conduct-based approach would merely lead to fewer removals. If a factfinder was tasked with determining whether a noncitizen had committed a crime with conduct that involved a substantial risk that physical force may be used, the factfinder’s ability to make this finding could well depend on how long ago the crime was committed.
The majority’s fifth and sixth factors, then, amount to one concern: that given the Sixth Amendment’s dictates there may be serious practical hurdles to proving prior convictions.
Turning to that concern, the majority distinguishes ACCA and § 16, on the
one hand, from § 924(c), on the other, because by requiring the examination of
*147
prior convictions the two former statutes pose these practical hurdles, while the
latter does not. Yes, it may be difficult to reconstruct the facts underlying the
predicate convictions because of the passage of time and possible loss of witnesses
and records; conversely, because § 924(c) is always applied to a contemporaneous
offense, there is no such practical difficulty. This is true for ACCA and § 16 as
incorporated into the INA. But what of § 16 in its numerous other applications?
Many other, primarily criminal, statutes incorporate the “crime of violence”
definition in § 16 as an element that must satisfied at the same time as the statute’s
other elements. In those applications, no reconstruction of old facts is necessary
because, as with § 924(c)—and I use the majority’s words here—“it’s all one big
ball of wax.” Maj. Op. at 40. As I emphasized above, “§ 16 is a criminal statute,
and it has both criminal and noncriminal applications,” so we cannot solely look to
its applications in the INA when examining whether it is similar to or different
from § 924(c).
Leocal
,
And, contrary to the majority’s suggestion, practical problems do abound with a conduct-based approach to § 924(c)(3)(B), even if those problems are not completely coextensive with problems that might arise by using a conduct-based *148 approach to ACCA’s residual clause and § 16(b) as incorporated into the INA. As amici curiae the National Association of Criminal Defense Lawyers and Families Against Mandatory Minimums explain: “Under a fact-based approach, whether the charged predicate offense satisfies § 924(c)(3) would turn not on whether the defendant committed the predicate offense, but how he committed it.” Br. of Amici at 11. In some § 924(c)(3)(B) cases (like my ACCA residual clause example above), proof of what made the defendant’s conduct risky (that is, how he committed the predicate offense) will not be essential to satisfy the elements of the predicate offense. In those cases, will prosecutors know whether the offense ultimately will satisfy the residual clause, so as to facilitate informed charging decisions? Will defendants and defense counsel know whether the charged predicate will satisfy the residual clause, so as to enable informed plea bargaining decisions? I doubt it.
What’s more, given the abstract nature of the residual clause inquiry I am not nearly as confident as the majority is that instructions can be fashioned to guide juries appropriately on—or that juries are well equipped to decide—what kind of conduct satisfies § 924(c)’s residual clause. Under a conduct-based approach, which the majority concedes must be based on facts found by a jury looking at the defendant’s conduct, the jury’s inquiry is divorced, in at least four ways, from the conduct it is tasked with examining. First, the jury must ignore any conduct that *149 forms the basis for other charged offenses and focus only on the conduct involved in committing the § 924(c) predicate offense. This could prove difficult if, to use as an example Ms. Ovalles’s case—in which she and her codefendants robbed a store, carjacked three vehicles, and attempted to carjack a fourth—the predicate offense (the attempted carjacking) occurred alongside other criminal conduct. Then, the text of § 924(c)(3)(B) provides two additional abstractions: “substantial risk” that physical force “may” be used. The majority says no problem—juries routinely decide what kind of conduct satisfies a substantial risk standard or what constitutes physical force. But “substantial risk” is only one abstraction. How does a jury decide whether conduct presents a substantial risk that physical force “may” be used? That is, how does a jury decide what conduct satisfies an abstraction (“may”) of already abstract conduct (“substantial risk”)? The majority has no answer for this. [14] Fourth, and further abstracting a jury’s inquiry is that a defendant’s use, carrying, or possession of a gun is insufficient, without more, to satisfy the risk standard in § 924(c)’s residual clause—or else the crime of violence element would be meaningless. A jury therefore must decide whether a defendant’s predicate offense conduct only, separate and independent from the *150 defendant’s possession of a gun, presented a substantial risk that physical force may be used, that doubly abstract standard. We presume that juries follow instructions, true, but we cannot presume that a jury will be able to make sense of an inquiry this abstracted from the facts of the case. Nor can we presume, given these realities, that a district court will be able to provide sufficiently concrete instructions to guide the jury. [15]
Finally, the majority’s ultimate disposition of Ms. Ovalles’s appeal proves my point about practical difficulties by highlighting another problem with a conduct-based approach: appeals. The majority, sitting as an appellate court, affirms Ms. Ovalles’s § 924(c)(3)(B) conviction even though Ms. Ovalles did not admit, and no fact-finder found, that her conduct created a substantial risk that physical force may have been used in the course of committing the predicate offense of attempted carjacking. The majority decides as a matter of law that the risk element has been satisfied by examining the elements of the predicate offense to which she pled guilty. See Maj. Op. at 45-47 (explaining that Ms. Ovalles’s plea to the elements of attempted carjacking satisfy the elements clause and that “the *151 real-life details of [her] crime” only “confirm it”). If § 924(c)(3)(B) is to have a new conduct-based “substantial risk” element, it often will not be the case that we can look to the defendant’s admissions to the elements of the predicate offense to determine that this conduct element is satisfied. This case is a prime example.
To “confirm” its conclusion that the elements of the attempted carjacking offense to which Ms. Ovalles pled guilty satisfied the conduct-based residual clause element, the majority considers the totality of Ms. Ovalles’s and her co- conspirators’ conduct. But under the majority’s own holding, whether the defendant engaged in conduct that satisfies the “substantial risk” standard is an element of the § 924(c) offense that must be decided by a jury or admitted by the defendant. Again, Ms. Ovalles never admitted that her attempted carjacking involved a substantial risk that the use of force might be used. Perhaps, without saying so, the majority is performing a harmless error analysis—an analysis the government never asked us to perform and on which the government bears the burden.
Will we uphold other convictions on appeal under this same flawed logic?
Many defendants like Ms. Ovalles were convicted when
McGuire
required use of
the categorical approach. Under
McGuire
, the district court was tasked with
deciding as a matter of law whether the defendant’s charged predicate offense
satisfied the residual clause.
III.
The majority claims to be avoiding one constitutional problem—the unconstitutional vagueness of § 924(c)(3)(B)—with a practical solution— abandoning the categorical approach as applied to that statute. The law does not permit the majority’s approach, and the text of § 924(c) does not permit the majority’s resolution. Even under the majority’s flawed framework, neither the government’s position, the dictates of the Sixth Amendment, nor practical considerations support application of a conduct-based approach over a categorical one. And there is every indication that the majority’s practical solution will breed a host of new practical problems.
Our displeasure with the categorical approach and the results of its
application to residual clauses does not permit us to jettison it. The categorical
approach is dictated by the text of the statute and Congress’s intent to impose
increased penalties based on the violation of certain predicate statutes. If Congress
wants to change course, or enact a residual clause that comports with the dictates
of due process, it can readily do so.
See McCarthan v. Director of Goodwill
Indust.-Suncoast, Inc.
,
In my view, the majority’s erroneous decision perpetuates unconstitutional sentences for inmates sentenced under § 924(c)’s residual clause in this circuit. If I am correct that Leocal and Dimaya require a categorical approach to interpreting § 924(c)(3)(B), these inmates—who should be entitled to relief—are serving between five years and life in prison beyond what our Constitution allows. And we will have perpetuated that injustice by failing to apply faithfully the Supreme Court’s directives.
I would not throw out the plain text of § 924(c)(3)(B) to save it from unconstitutionality. I would leave it to Congress save it. Respectfully, I dissent.
Notes
[*] Senior Circuit Judge Hull elected to participate in this decision, pursuant to 28 U.S.C. § 46(c).
[1] We have at times referred to these as the “use-of-force” and “risk-of-force” clauses,
respectively.
See Ovalles v. United States
,
[2] The panel separately held that Ovalles’s attempted-carjacking offense qualifies as a “crime of
violence” under the elements clause,
see
18 U.S.C. § 924(c)(3)(A).
Ovalles
,
[3] In a 28 U.S.C. § 2255 proceeding—as always—we review questions of law, like those
presented here,
de novo
.
Devine v. United States
,
[4] Although
Taylor
involved the ACCA’s enumerated-offenses clause, the Supreme Court later
extended the categorical approach (albeit without explanation) to the ACCA’s residual clause.
See James v. United States
,
[5] Although the parties in McGuire dickered over whether a pure categorical approach or a “modified” categorical approach should govern § 924(c)(3)(B), they agreed that some form of categorical approach applied. No one urged the Court to apply a conduct-based approach. See McGuire Appellant’s Br. at 17–19; McGuire Appellee’s Br. at 17–18. 30
[6] The dissent thinks it inconceivable that that the word “offense” could require the categorical
approach for cases arising under § 924(c)(3)’s elements clause and yet, at the same time, permit a
conduct-based approach for cases arising under the residual clause.
See
Dissenting Op. of J.
Pryor at 127, 134–35. It cites
Nijhawan
for the proposition that “where . . . Congress uses
similar statutory language and similar statutory structure in two adjoining provisions, it normally
intends similar interpretations”—and from that premise reasons that because the term “offense”
supports application of the categorical approach to the elements clause, it “must” do so with
respect to the residual clause, as well.
Id
. at 26 (quoting
Nijhawan
,
[7] A point of clarification: The dissent charges—as if it were a bad thing—that our opinion
“focuses solely on § 16’s incorporation into the INA.” Dissenting Op. of J. Pryor at 110;
see
also id.
at 123 n.8. But of course it does—by design and with good reason. We have so
“qualif[ied our] references to § 16(b),”
id
. at 110, because
Dimaya
’s holding is likewise limited.
See
[8] We are unpersuaded by Ovalles’s assertion (which the dissent echoes, see Dissenting Op. of J. Pryor at 148–49) that a conduct-based approach would require hopelessly complex and ineffective jury instructions about the crime-of-violence element. The district court can simply instruct jurors that they must find beyond a reasonable doubt that the underlying offense―the defendant’s commission of which they must already have found in order to satisfy the first element―“involve[d] a substantial risk that physical force against the person or property of another may [have] be[en] used in the course of committing the offense.” 18 U.S.C. § 924(c)(3)(B). There is nothing remarkable about asking jurors to make that sort of risk determination—and, if necessary, requiring judges to instruct jurors on the meaning of terms like “substantial” and “physical force.” That’s exactly how similar questions have been resolved for centuries and are resolved every day in courts throughout the country. See Gaudin , 515 U.S. at 511–15; see also, e.g. , Ala. Code § 13A-6-2(a)(2) (“A person commits the crime of murder if he 42
[9] Needless to say, we reject the dissent’s charge that we have “reach[ed]” out to “effectively rewrite[]” § 924(c)(3)’s residual clause in order “to avoid having to strike it down.” Dissenting Op. of J. Pryor at 124, 137. To the contrary, we have simply acknowledged that the interpretive question is a close one and, accordingly—and pursuant to time-honored canons of construction—saved it from the trash heap. It is the dissent, by contrast, that doggedly insists on the one and only reading of § 924(c)(3)(B) that guarantees its invalidation.
[10] One brief word in conclusion: The dissent repeatedly criticizes our reliance on what it calls
“extra-textual factors” in determining whether § 924(c)(3)(B) can plausibly be read to bear a
conduct-based interpretation. In particular, the dissent objects to the first, fifth, and sixth factors
examined above.
See
Dissenting Op. of J. Pryor at 103, 119–20, 124, 139–40, 143–44. The
reason, the dissent says, is that under the Supreme Court’s recent decision in
Jennings v.
Rodriguez
,
[1] 18 U.S.C. § 924(e).
[2] U.S. Sentencing Comm’n, Report to the Congress: Career Offender Sentencing Enhancements 39 (2016).
[3] Id.
[4] See U.S. Sentencing Comm’n, Recidivism Among Federal Offenders: A Comprehensive Overview 5, 18–19, 27 (2016).
[5] Id. at 18.
[6] Id.
[7] Id. at 19.
[8] U.S. Sentencing Comm’n, Report to the Congress , supra note 2, at 42.
[9] Id.
[10] Id.
[11] Id.
[12] See id. at 38.
[13] Id. at 42.
[14] See U.S. Sentencing Comm’n, Recidivism Among Federal Offenders , supra note 4, at 20.
[15] Id.
[16] Id. ; see also U.S. Sentencing Comm’n, The Past Predicts the Future: Criminal History and Recidivism of Federal Offenders 12 (2017) (“Offenders convicted of a firearms offense had the highest rearrest rate (68.4%]), followed by offenders convicted of a violent offense (64.1%) . . . .”).
[17]
Graham v. West Virginia
,
[18]
Id.
(collecting cases);
see also Spencer v. Texas
,
[19] See, e.g. , H.R. Rep. No. 98-1073, at 2 (1984) (relying on “extensive studies on recidivism” to justify an earlier version of the Armed Career Criminal Act). See generally U.S. Sentencing Comm’n, Mandatory Minimum Penalties for Firearms Offenses in the Federal Criminal Justice System (2018).
[20] 18 U.S.C. § 924(e)(1).
[21] Id. § 924(e)(2)(A).
[22] Id. § 924(e)(2)(B)(i).
[23] Id. § 924(e)(2)(B)(ii).
[24]
Id.
,
invalidated by Johnson v. United States
,
[25]
[26] Ala. Code § 13A-6-66(a)(1).
[27]
United States v. Davis
,
[28] Id. at 595.
[29] Sheldon A. Evans, Punishing Criminals for Their Conduct: A Return to Reason for the Armed Career Criminal Act , 70 Okla. L. Rev. 623, 626 (2018).
[30] Id. at 645
[31]
See Sessions v. Dimaya
,
[32]
United States v. Vann
,
[33]
United States v. Aguila-Montes de Oca
,
[34]
[35] See id. at 2560 (collecting cases).
[36]
See Sykes v. United States
,
[37]
Johnson
,
[38] See, e.g. , Hylor v. United States , ___ F.3d ____, No. 17-10856, slip op. at 10–13 (11th Cir. July 18, 2018) (Jill Pryor, J., concurring) (opining that attempted murder and attempted armed
[42] Evans, Punishing Criminals for Their Conduct , supra note 29, at 628.
[43]
Almendarez-Torres
,
[44]
Spencer
,
[45]
State v. Findling
,
[46] King, Sentencing and Prior Convictions , supra note 18, at 553.
[47] 1925 Tex. Crim. Stat. art. 62;
see also Spencer
,
[48]
Kelley v. State
,
[49]
[50]
Id.
at 455;
cf. Carter v. Commonwealth
,
[51]
Crocker v. State
,
[52]
[53] Id. at 393.
[54]
[55] Id. at 82.
[56] Ind. Code § 35-50-2–8.
[57]
Moore v. State
,
[58]
Weiss v. State
,
[59]
See People v. Gallardo
,
[60] Id.
[61]
See Skinner v. State
,
[62]
See Walker v. Hale
,
[63] See Ala. R. Crim. P. 26.6(a) (“Except in death penalty cases and in cases involving offenses committed prior to January 1, 1980, the judge shall impose the sentence in all cases.”).
[64]
See Brown v. State
,
[65] See, e.g. , Dubroff, supra note 45, at 337 (citing the “prejudice engendered by proving guilt for the present offense simultaneously with the issue of recidivism” under the common-law method).
[66]
Spencer
,
[67] See generally Nancy J. King, Juries and Prior Convictions: Managing the Demise of the Prior Conviction Exception to Apprendi, 67 SMU L. Rev. 577 (2014) (outlining procedures that preserve the right to a jury trial in recidivism proceedings while limiting prejudice to defendants).
[68] See U.S. Sentencing Comm’n, 2017 Sourcebook of Federal Sentencing Statistics fig. C, https://www.ussc.gov/research/sourcebook-2017.
[69]
See Pitcock v. State
,
[70]
[71] Id. at 559.
[72] Id. at 566–67.
[73] Id. at 569 (Stewart, J., concurring); see also id. (observing that many states had recently modified or abandoned the common-law method).
[74] David S. Sidikman, Note, The Pleading and Proof of Prior Convictions in Habitual Criminal Prosecutions , 33 N.Y.U. L. Rev. 210, 215–16 (1958) (capitalization omitted) (explaining that judicial assessment “should be preferred over the common-law” method and that “a jury trial should not be considered essential where the issue is conviction of prior offenses”); cf. Radice,
[78] Restoring the Armed Career Criminal Act, S. 3335, 115th Cong. § 2 (2018).
[79] Fla. Stat. § 784.041(1).
[80] Id. § 775.082(3)(e).
[81] Id. § 784.041(1)(b).
[82] Id. § 784.041(2)(a).
[83] Minn. Stat. § 609.223 subd. 1.
[84] Fla. Stat. § 812.0147.
[85] See id. § 775.082(3)(d).
[86] Utah Code § 76-6-409.8(2).
[87] Id. § 76-3-203(2).
[88] Mass. Gen. Laws ch. 267, § 17.
[89]
See Spencer
,
[1] In Saint Fleur, I agreed with the panel opinion that Mr. Saint Fleur’s § 924(c) conviction
“look[ed] to be valid despite Johnson.” Saint Fleur,
[3] Judge Wilson and I discussed this issue in greater depth in our concurring opinions in Williams,
[4] I did understand at the time, however, that this Court’s decision in United States v. Lockley,
[5] See E. Anne Carson, Bureau of Justice Statistics, Prisoners in 2016 1 (Aug. 7, 2018), https://www.bjs.gov/content/pub/pdf/p16.pdf.
[1] After
Johnson
was decided but before the Supreme Court issued its decision in
Dimaya
, a panel
of this court rejected Ms. Ovalles’s constitutional vagueness challenge to the residual clause of
§ 924(c).
See Ovalles v. United States
,
[2] “Prior to 1984, § 924(c) provided for the imprisonment of one who uses a firearm to commit
any felony. In 1984, as part of the [CCCA], Congress revised that rule and provided that the
firearm penalty should apply only to those who engaged in a crime of violence.”
United States v.
Cruz
,
[3] Other incorporations of § 16 in this manner abound. See, e.g. , 18 U.S.C. § 119(a)(1) (criminalizing knowingly making public restricted personal information about certain persons performing official duties “with the intent to threaten, intimidate, or incite the commission of a crime of violence against that covered person” or a member of the person’s family); 18 U.S.C. § 842(p)(2)(A) (forbidding the teaching or demonstrating of how to make an explosive device with the intent that the teaching “be used for, or in furtherance of, an activity that constitutes a Federal crime of violence”); 18 U.S.C. § 1952(a)(2) (criminalizing interstate travel with intent to “commit any crime of violence to further any unlawful activity”); 18 U.S.C. § 1956(a)(1), (c)(7)(B)(ii) (incorporating § 16’s crime of violence definition into the money laundering statute); 18 U.S.C. § 1959(a)(4) (enhancing penalties for violent crime in aid of racketeering where a defendant threatens to commit a crime of violence); 18 U.S.C. § 2261(a)(1) (criminalizing interstate travel to commit a crime of violence against a domestic partner); 18 U.S.C. § 3663A(c)(1)(A)(i) (incorporating § 16’s crime of violence definition into the Mandatory Victims Restitution Act, which requires an award of restitution if the defendant is convicted of a crime of violence). Section 16 also is incorporated into the definitions of other noncriminal statutes. See, e.g. , 34 U.S.C. § 30503(a)(1)(A) (providing, in the federal hate crimes statute, federal assistance for the prosecution of any “crime of violence”); 34 U.S.C. § 12361(d)(2)(A) (incorporating § 16’s crime of violence definition into the definition of “crime of violence motivated by gender” in the Violence Against Women Act).
[4] I use “elements” here to describe the set of facts and circumstances the government must prove
to obtain a conviction. I recognize that the Supreme Court has characterized § 924(c) as having
only two elements.
See, e.g.
,
Smith v. United States
,
[5] The majority therefore is incorrect when it says that the Supreme Court “conceived” the categorical approach in Taylor , see Maj. Op. at 20, because in Taylor the Supreme Court explained that the statute “always” has embodied the categorical approach. In other words, the categorical approach is not a pure judicial creation; rather, it is a judicial explanation of congressional intent.
[6]
See United States v. Taylor
,
[7]
Dimaya
resulted in a somewhat fractured decision. Justice Kagan, writing for a majority of the
Court, struck § 16(b) as unconstitutionally vague.
[8] Although
Dimaya
interpreted § 16(b) as incorporated into the INA specifically, the plurality
expressly acknowledged that, “as we have said before, § 16(b) is a criminal statute with
applications outside the immigration context,” so there was “no ground for discovering a novel
interpretation of § 16(b),” as applied to the INA.
[9] When the majority says that “[o]nly a plurality of the [Supreme] Court concluded that [§ 16(b)]
actually
requires
the categorical approach,” Maj. Op. at 15, it fails to account for
Leocal
. No
doubt realizing this shortcoming, the majority rationalizes that “the
Leocal
Court didn’t provide a
detailed explanation” when it held that § 16(b)’s language requires a categorical approach. Maj.
Op. at 23. But the Court did provide an explanation—a textual one.
See supra
at 14-15. We are
not free to ignore the Supreme Court’s decisions, whatever their level of detail. Nor, for that
matter, are we free to ignore
Leocal
even if we might question whether its unanimous
proclamation that § 16 requires a categorical approach remains good law in light of
Dimaya
’s
invalidation of § 16(b). “The [Supreme] Court has told us, over and over again, to follow any of
its decisions that directly applies in a case, even if the reasoning of that decision appears to have
been rejected in later decisions and leave to that Court the prerogative of overruling its own
decisions.”
Evans v. Sec’y, Fla. Dep’t of Corr.
,
[10] At the same time, the majority ignores the highest and best guidance we have—the Supreme Court’s unanimous statement in Leocal that the precise words we are tasked with interpreting in this case require a categorical approach.
[11] The majority also cites
United States v. Hayes
,
[12] In holding that ACCA’s residual clause must be read with a categorical approach and defined by the “ordinary case” of an offense, the Supreme Court described an offense that would fall within the residual clause as one “that, by its nature , presents a serious risk of injury to another”—categorical language—even though ACCA does not contain that phrase. James , 550 U.S. at 209.
[13] The majority describes its concerns in terms of avoiding sentencing-judge-found facts, but of course the same concern exists in the § 924(c) context: even under a conduct-based approach, the Sixth Amendment would not permit sentencing judges to find the facts necessary to secure a conviction. Thus, the real concern is whether it is practical to have a jury find the facts necessary to determine whether the defendant has committed a crime of violence.
[14] It is no answer to say that the word “may” in the statute does not create an inquiry distinct
from the jury’s inquiry about “substantial risk” because “[i]t is our duty to give effect, if
possible, to every clause and word of a statute.”
Duncan v. Walker
,
[15] That is not all. Deeming the § 924(c)(3) inquiry a fact question for the jury rather than a
question of law for the court easily could lead to inconsistent verdicts across cases. Identical
conduct that one jury considers to be conduct presenting a substantial risk that physical force
may be used might not, to another jury, present such a risk. The categorical approach prevents
such disparities.
See Moncrieffe
,