JEROME WILLIAMS v. UNITED STATES OF AMERICA
No. 19-10308
United States Court of Appeals, Eleventh Circuit
January 13, 2021
[PUBLISH] D.C. Dоcket Nos. 2:16-cv-08101-CLS, 2:97-cr-00377-CLS-RRA-2
Before JORDAN, LAGOA, and BRASHER, Circuit Judges.
BRASHER, Circuit Judge:
This appeal comes to us from a post-judgment challenge to a 1998 federal sentence. The question on appeal is under what circumstances the legal landscape at the time of a defendant‘s sentencing can establish, as a matter of historical fact, that the sentencing court relied on the unconstitutionally vague residual clause of the Armed Career Criminals Act (“ACCA“) to classify a prior felony as violent and, so, to increase the defendant‘s statutory sentencing range. After the Supreme Court ruled that the ACCA‘s residual clause is unconstitutionally vague and held that its ruling was retroactive to collatеral review, Jerome Williams filed a “Johnson motion” challenging his sentence for bank robbery. He argued that case law at the time of his sentencing established that the sentencing court relied on the residual clause alone to increase his statutory
I.
In 1998, Williams was convicted of robbing a bank while carrying a firearm. The ACCA provides a statutory sentencing enhancement for certain previously convicted felons who use a firearm.
At the sentencing hearing on his bank robbery conviction, Williams did not object to the application of the ACCA, and the sentencing court never addressed why any of his previous felonies counted as violent. Applying the ACCA, the sentencing court sentenced Williams to concurrent terms of 300 months of imprisonment for bank robbery and 327 months for possession of a firearm by a convicted felon, with a consecutive term of 60 months for carrying a firearm during and in relation to a crime of violence.
After serving about 220 months of his sentence, Williams moved for leave to file his third motion under
After reviewing the then-existing legal landscape, the district court denied Williams‘s motion. The same judge who had sentenced Williams nearly two decades ago found thаt the sentencing record failed to illuminate which clause of the ACCA he had relied on to qualify the kidnapping conviction as a violent felony. As for the then-existing legal landscape, the district court concluded it was, at best, unclear which clause or clauses the sentencing court would have relied on. Instead, persuasive authority supported a
We granted Williams a certificate оf appealability on the question whether the district court erred in concluding that he had not made the requisite showing under Beeman v. United States, 871 F.3d 1215 (11th Cir. 2017), as to his kidnapping conviction under
II.
This appeal raises a mixed question of law and fact. Mixed questions of law and fact are “questions in which the historical facts are admitted or established, the rule of law is undisputed, and the issue is whether the facts satisfy the statutory standard, or to put it another way, whether the rule of law as applied to the established facts is or is not violated.” Lincoln v. Bd. of Regents of Univ. Sys. of Ga., 697 F.2d 928, 940 n.15 (11th Cir. 1983) (quoting Pullman-Standard v. Swint, 456 U.S. 273, 289 n.19 (1982)). Our review of a mixed question of law and fact depends “on whether answering it entails primarily legal or factual work.” U.S. Bank N.A. v. Village at Lakeridge, LLC, 138 S. Ct. 960, 967 (2018). We apply de novo review when the question requires a court to “expound on the law, particularly by amplifying or elaborating on a broad legal standard.” Id. We apply clear error review when the question requires a court to “marshal and weigh evidence, make credibility judgments, and otherwise address ... ‘multifarious, fleeting, special, narrow facts that utterly resist generalization.‘” Id. (quoting Pierce v. Underwood, 487 U.S. 552, 561-62 (1988)).
A district court may resolve a Johnson motion by making a finding of historical fact that the sentencing court did or did not rely on a specific clause of the ACCA. See Beeman, 871 F.3d at 1224 n.5 (“What we must determine is a historical fact: was [the defendant] in 2009 sentenced solely per the residual clause?“); United States v. Pickett, 916 F.3d 960, 967 (11th Cir. 2019) (vacating and remanding because “[t]he district court obviously is in a better position than we аre to evaluate what likely happened ..., especially since we are remanding this case to the very judge who initially sentenced [the defendant]“). We would review such a finding for clear error. But because Williams‘s Johnson motion relies exclusively on the state of the law in 1998 and the district court resolved it by reference to legal principles alone, the issue before this Court is a mixed question of law and fact that entails primarily legal work. Accordingly, we review this mixed question of law and fact de novo. Our decision to apply a de novo standard of review under the circumstances of this case is consistent with authority from other circuits. Sеe Dimott v. United States, 881 F.3d 232, 236 (1st Cir. 2018) (“We review de novo the district courts’ denials of [the defendants‘] habeas petitions” under Johnson.); Golinveaux v. United States, 915 F.3d 564, 568 (8th Cir. 2019) (“Determining the legal environment requires a ‘legal conclusion’ about the controlling law at the time of sentencing,” which the court reviews de novo.); United States v. Copeland, 921 F.3d 1233, 1242 (10th Cir. 2019) (“our review of a district court‘s denial of a [Section] 2255 Johnson claim is de novo[,] unless the court conducted an evidentiary hearing from which it made findings“).
III.
Before we address the specifics of Williams‘s argument, a little background is necessary. The ACCA defines the kind of crimes that count as a “violent felony” in three ways. First, a felony may qualify as violent under the elements clause of the ACCA because it “has as an element the use, attempted use, or threatened use of physical force against the person of another.”
The Supreme Court held in Johnson that the residual clause is unconstitutionally vague such that “imposing an increased sentence under the residual clause of the Armed Career Criminal Act violates the Constitution‘s guarantee of due process.” Johnson, 576 U.S. at 606. But the Court emphasized that its “decision does not call into question application of the Act to the four enumerated offenses, or the remainder of the Act‘s definition of a violent felony.” Id. The Supreme Court made its decision in Johnson retroactive in Welch, which means the Court‘s decision in Johnson applies even to convictions and sentences that have long been final. Welch, 136 S. Ct. at 1268.
Thus, because of Johnson and Welch, Williams has filed a Section 2255 motion that seeks resentencing nearly two decades after his original sentence was imposed. Everyone agrees that Williams is due to be resentenced if the sentencing court relied on the residual clause of the ACCA to classify his federal kidnapping conviction as “violent.” To that end, Williams must establish that the residual clause “actually adversely affected the sentence he received” because “more likely than not[] it was use of the residual clausе that led to the sentencing court‘s enhancement of his sentence.” Beeman, 871 F.3d at 1221-22. And the sentencing court must have relied only on the residual clause in qualifying the felony as violent. Id. at 1221. Williams‘s Johnson claim fails “[i]f it is just as likely that the sentencing court relied on the elements ... clause, solely or as an alternative basis for the enhancement.” Id. at 1222. Williams can prove this historical fact through direct evidence, such as “comments or findings by the sentencing judge,” or circumstantial evidence, such as “statements in the PSR” or “the law ... at the time of sentencing.” Id. at 1224 nn.4-5.
Williams argues that more likely than not the sentencing court classified his federal kidnapping conviction as a violent felony under the residual clause. He makes two main arguments. First, Williams argues that it is unlikely the sentencing court relied on any other clause. Everyone agrees that the second clause for enumerated offenses has no application here, and Williams argues that case law at the time of his sentencing makes it unlikely the sentencing court relied on the elements clause. Williams argues that, before he was sentenced in 1998, we had held that a defendant could be convicted of federal “kidnapping by inveiglement” without using force as long as the defendant “formed the intent to use forcible action, in the event his deception failed, to complete the kidnapping.” United States v. Boone, 959 F.2d 1550, 1555, 1557 (11th Cir. 1992); see also United States v. Lewis, 115 F.3d 1531, 1535 (11th Cir. 1997) (listing elements of federal kidnapping). And we had instructed sentencing courts applying the ACCA at the time to look “only to the statutory
The government‘s response is three-fold. First, it argues that Williams is reading too much into Boone and Lewis, which concerned jury instructions and a guilty plea, not sentencing. Second, the government cites persuasive authority to argue that the sentencing court just as likely used the elements clause to categorize the federal kidnapping conviction as “violent.” The Second Circuit, for example, had relied on the elements clause to declare a federal kidnapping conviction to be “violent,” reasoning it “cannot be questioned” that “the crime of kidnapping involves the threatened use of physical force against a person and is thus a crime of violence under this statute.” United States v. Patino, 962 F.2d 263, 267 (2d Cir. 1992). The government also points out that we had held that federal kidnapping is a crime of violence under the elements clause of an identically worded Sentencing Guidelines provision. United States v. Salemi, 26 F.3d 1084, 1087 (11th Cir. 1994). Third, and in any event, the government argues that Williams has not met his burden of proof merely by showing the elements clause would not likely have justified the sentence; he must show that the residual clause alone led tо the sentencing court‘s enhancement as a matter of historical fact. Case law that merely suggests the elements clause would not have properly applied in a hypothetical case, the government argues, is not enough to establish that the residual clause did apply in this case.
We agree with the government. Because no case-specific evidence exists to suggest which clause the sentencing court relied on, the only available evidence is the state of the law at the time of Williams‘s sentencing. But this category of circumstantial evidence tells us very little.
In 1998, our most on-point precedent was Salemi. There, we reversed a district court when it held that federal kidnapping was not a crime of violence under the elements clause of a sentencing guideline. Salemi, 26 F.3d at 1087. At the time, the Sentencing Guidelines were “binding on judges” and carried “the force and effect of laws.” United States v. Booker, 543 U.S. 220, 234 (2005). And the guideline in Salemi, like the ACCA, provided that a crime would count as violent if it had as “an element the use, attempted use, or threatened use of physical force.” 26 F.3d at 1087 (quoting
Based on these precedents and the equally ambiguous precedents from our sister circuits, we cannot say that the sentencing judge did not rely in part on the elements clause when he sentenced Williams. A sentencing judge looking at the legal landscape in 1998 could have thought that he would be reversed like the sentencing judge in Salemi if he held that federal kidnapping did not satisfy the ACCA‘s identical elements clause. Or he could have treated the issue as one of first impression, held that there was no force element to a federal kidnapping, and relied exclusively on the residual clause. Had a hypothetical sentencing court confronted this question in the 1990s, we cannot say how it would have answered it. Thus, it is doubly difficult for us to conclude that the particular sentencing court in Williams‘s case—as a matter of historical fact—answered it by not applying the elements clause.
We previously faced a similarly uncertain legal landscape in Pickett, where the defendant argued that his felony offense qualified under the residual clause and it was “uncertain at best” whether it also qualified under the elements clause. 916 F.3d at 964. We held that such a motion fails if it was “just as likely that the sentencing court relied on the elements or enumerated offenses clause, solely or as an alternative basis for the enhancement.” Id. at 963 (quoting Beeman, 871 F.3d at 1222). Even if the residual clause were the “most obvious clause under which the convictions qualified,” it “does not necessarily mean even by implication that the elements clause could not also have been relied on.” Id. at 965.
In Pickett, the defendant relied on two opinions from this Court, neither of which was binding. The first was an unpublished table decision stating that the offense at issue “did not categorically satisfy the elements clаuse.” Id. (citing United States v. Rozier, 37 F. App‘x 499 (11th Cir. 2002)). The second was an unpublished opinion holding that the offense at issue was not always a violent felony—based only on another circuit‘s holding that the offense did not always satisfy the elements clause. Id. (citing United States v. Wright, 181 F. App‘x 914 (11th Cir. 2006)). We explained that, legally, “[n]othing in this collection of cases amounts to binding precedent” and that, practically, it was doubtful the district court even saw the table decision. Id. The government in turn relied on dicta from this Court that an offense was a crime of violence under an analogous elements clause in the Sentencing Guidelines. Id. at 965-66. After reviewing these decisions, we found they offered a “very weak
The legal landscape here is as uncertain as it was in Pickett. As in Pickett, we conclude that, “[f]aced with this uncertain precedential landscape,” the sentencing court “likely would have quickly determined” that Williams‘s federal kidnapping conviction “qualified under the residual clause.” Id. at 966. “[B]ut we do not know what else it might have thought.” Id. These authorities do not establish “either what the court thought about the elements clause—or even whether the court thought about it at all.” Id. Of course, “[w]ith the residual clause plainly available, the district court would not have needed to consider the elements clause.” Id. But that is beside the point. As a matter of historical fact, “we genuinely do not know what actually happened.” Id.
The otherwise smooth application of Pickett to this case does have one wrinkle. Unlike in Pickett, Williams argues that if the sentencing court had relied on the elements clause, it would have been an error under binding precedents from this Court, namely our view of the federal kidnapping statute and our categorical approach to the elements clause. Our dissenting colleague agrees with Williams. He says that Pickett is “easily distinguishable” because here “Supreme Court and Eleventh Circuit precedent required the use of the categorical approach under the elements clause of the ACCA (Taylor and Oliver), and Eleventh Circuit precedent held that federal kidnapping could be accomplished without the use, attempted use, or threatened use of physical force (Boone and Lewis).” Dissenting Op. at 24.
We are unpersuaded that our 1990s case law meaningfully distinguishes this case from Pickett. This is so for two reasons.
First, Williams and our dissenting colleague too quickly discount our decision in Salemi, in which we held that federal kidnapping was violent under the elements clause of an identically worded sentencing guideline. By clouding our case law on this point, Salemi greatly diminishes the relevance of Boone and Lewis. To be certain, Salemi is distinguishable because it concerned a sentencing guideline, not a statute. It is possible—maybe even highly likely—that we would hold today that federal kidnapping does not satisfy the elements clause of the ACCA. See United States v. Gillis, 938 F.3d 1181, 1210 (11th Cir. 2019) (holding, as a matter of first impression, that federal kidnapping is not a crime of violence under
Second, because Williams аnd our dissenting colleague‘s approach requires carefully parsing circa-1998 precedents, it necessarily sheds too little light on the historical question at issue here. Absent authority that would have compelled a particular result in 1998, Williams cannot meet his burden of proof through case law alone. The question for us is, as a matter of historical fact, on which clause did the sentencing court hang Williams‘s sentence? Because there is no clear precedent
Williams argues that this result is unfair and arbitrary, but we disagree. Our case law in this area “reflects longstanding and fundamental interests in finality.” Beeman, 871 F.3d at 1223. Williams has raised a very particular kind of claim—a retroactive claim about an unconstitutionally vague sentencing enhancement. At his sentencing two dеcades ago, Williams could have raised a constitutional claim or litigated the statutory issue of whether the elements clause could justify the enhancement. But, only now, after more than eighteen years, has Williams challenged whether his kidnapping conviction counts as a violent felony. For Williams‘s collateral attack to succeed, he needs to establish that the sentencing court committed a specific kind of error, a retroactive constitutional one. It is not enough that the sentencing court might have committed a statutory error—the application of the elements clause in a case that did not warrant it—because that error would not be retroactive on collateral review.
Williams also argues that, in requiring him to come forward with “clear precedent showing that the court could only have used one clause or another,” Pickett, 916 F.3d at 964, we are applying something higher than a more-likely-than-not standard. Again, we disagree. The root problem here is that Williams is relying on circumstantial evidence which, when unclear, has little to no bearing on the ultimate issue. As a hypothetical, imagine a jury trying to determine who was at fault in a car accident. As in this case, the sole evidence is circumstantial—a local business‘s security footage, focusеd not on the intersection where the accident occurred, but on the street a few yards away. If the footage shows the traffic suddenly stop shortly before the time of the accident, the jury might be persuaded that the cars had stopped at a red light that the offending car ran, causing the accident. If the footage shows an empty street or stalled rush hour traffic, the circumstantial evidence does not tell the factfinder much about the disputed issue at all. The law at the time of Williams‘s sentencing is more like these latter examples than the former. Because the authorities that Williams cites are not clear, they fail to shed light on what the sentencing court did as a matter of historical fact.
We have held that “[i]t is no more arbitrary to have a movant lose in a [Section] 2255 proceeding because of a silent record than to have the [g]overnment lose because of one.” Beeman, 871 F.3d at 1224. When a case turns on an issue of historical fact, the available evidence, whether traffic footage or the legal landscape of yesteryear, must satisfy the burden of proof on the disputed issue. If the evidence is silent or in equipoise, then the party with the burden fails. So it is here.
IV.
Because the district court did not err in denying Williams‘s motion, the district court‘s judgment is AFFIRMED.
JORDAN, Circuit Judge, dissenting:
I agree with my collеagues in the majority—and with the Eighth and Tenth Circuits—that de novo review is appropriate
I
Mr. Williams was convicted in 1998 of, among other crimes, possession of a firearm by a convicted felon. At his sentencing, the district court found that Mr. Williams had previously committed three violent felonies as defined in the ACCA, and therefore enhanced his sentence. See
One of those prior felonies was federal kidnapping, in violation of
After Mr. Williams was sentenced, the Supreme Court struck down the ACCA‘s residual clause in Johnson v. United States, 576 U.S. 591 (2015), and later made that ruling retroactive in Welch v. United States, 136 S. Ct. 1257 (2016). So, if Mr. Williams can establish that the district court classified his federal kidnapping conviction as a violent felony under only the ACCA‘s unconstitutional residual clause, he can have his sentence vacated and be resentenced. See
II
For Mr. Williams to succeed, he must show, by a preponderance of evidence, that the district court relied only on the ACCA‘s residual clause with respect to his federal kidnapping conviction. See Beeman v. United States, 871 F.3d 1215, 1221-22 (11th Cir. 2017). The preponderance of evidence standard does not require Mr. Williams to make that showing to a high degree of certainty. Instead, “the preponderance-of-the-evidence standard results in a roughly equаl allocation of the risk of error between litigants.” Grogan v. Garner, 498 U.S. 279, 286 (1991). Accordingly, Mr. Williams meets his evidentiary burden so long as the evidence “tip[s] the scales just one little bit in [his] favor.” Blossom v. CSX Transp., Inc., 13 F.3d 1477, 1479 (11th Cir. 1994).
To meet that burden, Mr. Williams can rely on direct evidence. Or he can rely on circumstantial evidence, such as the legal landscape at the time of sentencing. See Beeman, 871 F.3d at 1224 nn.4-5. Circumstantial evidence is “[e]vidence based on inference.” Black‘s Law Dictionary 698 (11th ed. 2019). See also United States v. Henderson, 693 F.2d 1028, 1031 (11th Cir. 1982). The tests for the sufficiency of direct and circumstantial evidence are identical. See United States v. Doe, 661 F.3d 550, 560 (11th Cir. 2011).
III
Because the record in Mr. Williams’ criminal case does not indicate which of the ACCA‘s clauses the district сourt relied on to classify his federal kidnapping
At the time of Mr. Williams’ 1998 sentencing, we (and the Supreme Court) had instructed district courts to employ a categorical approach to determine whether an offense constituted a violent felony under the ACCA‘s elements clause, i.e., by “looking only to the statutory definitions of the prior offense.” United States v. Oliver, 20 F.3d 415, 418 (11th Cir. 1994) (quoting Taylor v. United States, 495 U.S. 575 (1990)). See also United States v. Gonzalez-Lopez, 911 F.2d 542, 546 (11th Cir. 1990). As a result, for federal kidnapping to constitute a violent felony under the ACCA‘s elements clause, its statutory elements must have required the use, attempted use, or threatened use of physical force. See
We had also ruled by 1998 that federal kidnapping did not require the use, attempted use, or threatened use of physical force. For example, in United States v. Boone, 959 F.2d 1550, 1555 (11th Cir. 1992), we held that federal kidnapping could be accomplished through inveiglement (i.e., seduction) or decoy given the text of
The majority argues that in Boone we held that “a kidnapping by inveiglement requires the alleged kidnapper to have formed the intent to usе forcible action.” Boone 959 F.2d at 1558. But the suggestion that Boone is unclear as it relates to an elements clause analysis fails for two reasons. First, the “force” discussed in Boone can be either psychological or physical, see id., while the ACCA‘s elements clause requires that physical force be an essential element of the predicate offense. Second, what we held in Boone was that kidnapping by inveiglement or deception requires the kidnapper to have subjectively intended to employ force (psychological or physical) if the inveiglement or deception were to fail. See id. at 1555-58. Logically, the victim need not be aware of the kidnapper‘s subjective intent to employ force if the victim were to discover that he is being invеigled or deceived. See id. at 1556. In short, Boone establishes that federal kidnapping does not require the use, attempted use, or threatened use of physical force.
Additionally, in United States v. Lewis, 115 F.3d 1531, 1535 (11th Cir. 1997), we identified the elements of federal kidnapping under
In the Eleventh Circuit, therefore, a district court in 1998 could not have properly used the ACCA‘s elements clause to classify federal kidnapping as a violent felony. After all, district courts are “presumed to know the law and apply it in making their decisions.” Walden v. Arizona, 497 U.S. 639, 653 (1990). Accord United States v. $242,484.00, 389 F.3d 1149, 1155 (11th Cir. 2004)
To recap, as things stood in 1998, Supreme Court and Eleventh Circuit precedent required the use of the categorical approach under the elements clause of the ACCA (Taylor and Oliver), and Eleventh Circuit precedent held that federal kidnapping could be accomplished without the use, attempted use, or threatened use of physical force (Boone and Lewis). Those cases render United States v. Pickett, 916 F.3d 960, 965-66 (11th Cir. 2019), where the relevant legal landscape was void of binding precedent, easily distinguishable.
IV
The majority reasons that United States v. Salemi, 26 F.3d 1084, 1087 (11th Cir. 1994), which held that federal kidnapping was a violent crime under
First, Salemi relied on Application Note 2 to
Second, Salemi—a guideline decision—could not overrule Boone, an earlier opinion explaining how federal kidnapping could be accomplished without the use, attempted use, or threatened use of force. See United States v. Steele, 147 F.3d 1316, 1318 (11th Cir. 1998) (en banc) (explaining the prior panel precedent rule). And to the extent that Salemi may have caused some confusion, Lewis later confirmed that federal kidnapping did not have as an element the use, attempted use, or threatened use of force. Boone and Lewis were binding precedent in 1998 and remained more relevant to an elements clause analysis than Salemi.
In my view, it is more likely than not that the district court relied only on the residual clause in Mr. Williams’ case to conclude that federal kidnapping was a violent felony under the ACCA. Given the legal landscape in 1998, that is an easier inference to make than the alternatives—that the district court (i) did not apply the categorical approach (contrary to Supreme Court and Eleventh Circuit precedent); (ii) found that the use, attempted use, or threatened use of physical force was an essential statutory element of federal kidnapping (contrary to Eleventh Circuit precedent); or (iii) otherwise erred in applying, or failing to apply, existing law (contrary to the presumption that district courts know and follow the law).
V
Under de novo review, I conclude that the legal landscape in 1998 tips the scales
