* Climenko Fellow & Lecturer on Law, Harvard Law School. Thanks to Daniel Deacon, Richard Re, and especially Eve Primus for helpful comments. Thanks also to Paul Schied and the other Columbia Law Review editors for their incredible efforts and patience in getting this Essay to print.
In January 2015, the Supreme Court directed the parties to brief and argue an additional question in Johnson v. United States: “Whether the residual clause in the Armed Career Criminal Act of 1984, 18 U.S.C. § 924(e)(2)(B)(ii), is unconstitutionally vague.”
Order in Pending Case, Johnson v. United States, No. 13-7120 (Jan. 9, 2015), available at http://sblog.s3.amazonaws.com/wp-content/uploads/2015/01/13-7120_Order_1-9-15.pdf (on file with the Columbia Law Review).
The order represents an unusual move because the defendant had not raised the vagueness issue and the Court issued the order after it had already heard argument on the question raised in the petition for certiorari. Commentators therefore view the order as a signal that the Court will likely invalidate the residual clause.
Will Baude, Supreme Court Will Consider the Constitutionality of the Armed Career Criminal Act, Wash. Post (Jan. 9, 2015), http://www.washingtonpost.com/news/
volokh-conspiracy/wp/2015/01/09/supreme-court-will-consider-the-constitutionality-of-the-armed-career-criminal-act/ (on file with the Columbia Law Review) (suggesting Justice Scalia may have gathered enough votes in support of invalidating residual clause of ACCA); Rory Little, Argument Preview: Are There (Finally) Five Votes to Declare the Residual Clause of the ACCA Unconstitutionally Vague?, SCOTUSblog (Nov. 4, 2014, 10:50 am), http://www.scotusblog.com/2014/11/argument-preview-are-there-finally-five-votes-to-declare-the-residual-clause-of-the-acca-unconstitutionally-vague/ (on file with the Columbia Law Review) (counting at least four votes for striking down residual clause); Richard M. Re, Who Made a Vague Law Vague?, Re’s Judicata (Jan. 12, 2015, 10:26 am), https://richardresjudicata.wordpress.com/2015/01/12/who-made-a-vague-law-vague/ (on file with the Columbia Law Review) (predicting Justice Scalia’s argument for striking down residual clause will prevail). This decision will have been several years in the making: The Supreme Court has had to resolve numerous circuit splits over whether various state criminal convictions qualify as prior convictions for violent felonies under the Armed Career Criminal Act (ACCA), and Justice Scalia has been calling for the Court to invalidate the residual clause for the last few years. 3 See infra notes 12–19 and accompanying text (discussing these cases).
A decision invalidating the residual clause would be significant for many reasons. The Court has not invalidated a criminal statute on vagueness grounds in over fifteen years, and ACCA is a flashpoint for many of the most pressing issues facing criminal law today—Hispanic and black offenders receive the ACCA enhancement at higher rates than white offenders do, and ACCA’s harsh mandatory minimum may lead many defendants to plead guilty to avoid more extensive prison time. 4 See infra notes 88–89 and accompanying text (discussing consequences of fifteen-year mandatory minimum).
But the practical implications of the Court’s pending decision in Johnson will depend in part on the answer to another question—if the Court invalidates ACCA’s residual clause as unconstitutionally vague, will prisoners who have already been sentenced under ACCA be able to obtain relief? This question will become relevant not only if the Court ultimately invalidates the residual clause, but it also may inform the Court’s decision to invalidate the residual clause in the first place. 5 See Peter W. Low & Joel S. Johnson, Changing the Vocabulary of the Vagueness Doctrine, 101 Va. L. Rev. (forthcoming 2015) (manuscript at 57–58) (on file with the Columbia Law Review) (noting Justices may be influenced by questions about whether ruling in Johnson would be retroactive). There are over 6,000 federal prisoners who have been sentenced under ACCA, 6 U.S. Sentencing Comm’n, Report to Congress: Mandatory Minimum Penalties in the Federal Criminal Justice System 293 (2011) [hereinafter Mandatory Minimum Penalties], available at http://www.ussc.gov/news/congressional-testimony-and-reports/mandatory-minimum-penalties/report-congress-mandatory-minimum-penalties-federal-criminal-justice-system (on file with the Columbia Law Review) (finding 2.9% of federal prisoners qualified as armed career criminals under ACCA); see also Douglas Berman, SCOTUS Orders New Briefing and Argument on ACCA’s Constitutionality in Johnson!?!?!, Sentencing Law and Policy: An Affiliate of the Law Professor Blogs Network (Jan. 10, 2015, 10:23 am), http://sentencing.typepad.com/sentencing_law_and_policy/2015/01/scotus-orders-new-briefing-and-argument-on-accas-constitutionality-in-johnson.html (on file with the Columbia Law Review) (suggesting over 5,000 prisoners sentenced under ACCA); Federal Bureau of Prisons, Statistics, http://www.bop.gov/about/statistics/population_statistics.jsp (on file with the Columbia Law Review) (last visited Jan. 25, 2015) (tallying 210,736 federal inmates). and a favorable decision in Johnson could require many of those prisoners to be resentenced. 7 It is not clear how many prisoners were subject to the enhancement under ACCA’s residual clause, specifically. ACCA defines a violent felony as any crime that “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) (2012) (emphasis added). Johnson will only determine whether the “or otherwise involves” language is unconstitutionally vague, so prisoners with three or more convictions for “burglary, arson,” “extortion,” or a crime “involv[ing] [the] use of explosives” would not benefit from Johnson. However, various doctrines and statutes limit the availability of federal collateral review—that is, review that occurs after a defendant’s conviction has become final. The proper reading of these doctrines and statutes will determine whether a decision invalidating ACCA’s residual clause will benefit most of the currently incarcerated federal prisoners who were sentenced under ACCA.
This Essay examines the impact a favorable decision in Johnson could have at the various stages of post-conviction relief for three categories of prisoners—prisoners whose convictions have not yet become final; prisoners whose convictions have become final but who have not yet filed a petition seeking post-conviction relief; and prisoners whose convictions have become final and who have already filed at least one petition seeking post-conviction relief. In doing so, it offers a reading of the relevant cases and statutes that permits any defendant sentenced under ACCA to obtain relief based on a decision invalidating the residual clause. It also highlights some underexplored statutes and doctrinal questions courts will confront as they determine which prisoners should be resentenced in light of Johnson.
Part I briefly describes the vagueness issue in Johnson and how Johnson will apply to prisoners whose convictions have not yet become final. Part II describes several judicial doctrines and statutory limitations imposed by the Anti-Terrorism and Effective Death Penalty Act (AEDPA) that may limit Johnson’s significance for prisoners whose convictions have become final. Part III describes the AEDPA provisions that may limit Johnson’s significance for prisoners who have already filed at least one petition for collateral review. Several of the doctrines and statutes are best read to permit any defendant sentenced under ACCA to obtain relief based on a favorable decision in Johnson. However, some of the case law is murkier, and some of the statutory provisions less clear. But even these doctrines and statutes permit courts to resentence incarcerated prisoners who were sentenced under ACCA, and that is how the doctrines and statutes should be read.
I. Group One: Prisoners On Direct Review
The Court has previously attempted to address the vagueness issue posed by ACCA’s residual clause, but there now appears to be a real possibility that five Justices are inclined to find the residual clause unconstitutionally vague. 8 See Little, supra note 2 (noting potential shift of Justices toward striking down residual clause). If this occurs, the implications for resentencing are clearest for those prisoners whose convictions have not yet become final: Those prisoners will be able to obtain relief based on a favorable decision in Johnson.
The void for vagueness doctrine is well established. A criminal law is unconstitutionally vague if it fails to “define the criminal offense with sufficient definiteness that ordinary people can understand what conduct is prohibited and in a manner that does not encourage arbitrary and discriminatory enforcement.” 9 Kolender v. Lawson, 461 U.S. 352, 357 (1983).
The Court is poised to decide whether ACCA’s residual clause is void for vagueness. ACCA imposes a fifteen-year mandatory minimum sentence on defendants convicted of unlawful possession of a firearm under § 922(g) who have three or more convictions for “violent felonies.” ACCA defines a violent felony as any crime that “is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another.” 10 18 U.S.C. § 924(e)(2)(B) (emphasis added). The “residual clause” refers to the text beginning “or otherwise.” 11 Chambers v. United States, 555 U.S. 122, 124 (2009). James v. United States directed courts to determine whether a conviction falls within the residual clause by using the “categorical approach.” 12 550 U.S. 192, 202 (2007). Under the categorical approach, courts “consider whether the elements of the offense are of the type that would justify its inclusion within the residual provision, without inquiring into the specific conduct of th[e] particular offender.” 13 Id.
It is not hard to imagine an opinion invalidating ACCA’s residual clause as unconstitutionally vague. Indeed, in several dissents, Justice Scalia has essentially written the opinion already. 14 See Sykes v. United States, 131 S. Ct. 2267, 2287 (2011) (Scalia, J., dissenting); James, 550 U.S. at 215 (Scalia, J., dissenting). Since James, the Court has taken a case involving the proper interpretation of the residual clause nearly every other term. 15 Sykes, 131 S. Ct. at 2267; Chambers, 555 U.S. at 122; Begay v. United States, 553 U.S. 137 (2008). Although the Court has attempted to put different clarifying glosses on ACCA’s residual clause, none has proven particularly effective. Begay v. United States suggested that a crime falls within ACCA’s residual clause where the crime is typically committed in a “purposeful, violent, and aggressive manner,” 16 553 U.S. at 145. but United States v. Sykes maintained instead that the residual clause focuses primarily on “levels of risk.” 17 131 S. Ct. at 2275. Beginning in James, Justice Scalia indicated that he was inclined to find the residual clause unconstitutionally vague because the categorical approach “is almost entirely ad hoc.” 18 550 U.S. at 215 (Scalia, J., concurring). Since then, Justice Scalia has seized on the Court’s failure to clarify the residual clause: “What sets ACCA apart . . .—and what confirms its incurable vagueness—is our repeated inability to craft a principled test out of the statutory text.” 19 Sykes, 131 S. Ct. at 2287 (Scalia, J., dissenting); see also Re, supra note 2 (suggesting Court’s opinions construing residual clause made it vague). While the reason some Justices changed their views on ACCA may remain a mystery, every ACCA case the Court took increasingly confirmed Justice Scalia’s claim that ACCA was either incurably vague or that the Court had made it so. Given the timing of the order (three months after the case was argued), the draft opinions in Johnson may also have helped his case.
B. Prisoners on Direct Review
A decision invalidating ACCA’s residual clause as unconstitutionally vague will apply in any case where the prisoner’s conviction has not yet become final. It is black-letter doctrine that “new” rules apply to criminal cases that have not yet become final.
Griffith v. Kentucky, 479 U.S. 314, 322–23 (1986) (“But after we have decided a new rule in the case selected, the integrity of judicial review requires that we apply that rule to all similar cases pending on direct review.”).
Finality occurs when a prisoner has exhausted his direct appeal in the federal court of appeals and the Supreme Court has denied a petition for certiorari or the time to file a petition for certiorari has expired.
Greene v. Fisher, 132 S. Ct. 38, 44 (2011) (“Finality occurs when direct state appeals have been exhausted and a petition for writ of certiorari from this Court has become time barred or has been disposed of.”).
It is hard to know exactly how many ACCA offenders’ convictions have not yet become final. Sentencing reports indicate that, in a given year, almost 600 defendants will be eligible for the ACCA sentencing enhancement.
Mandatory Minimum Penalties, supra note 6, at 282–83.
It takes, on average, slightly less than a year and a half for a defendant’s conviction to become final: Once the sentencing court enters judgment, the defendant has two weeks to file an appeal;
Fed. R. App. P. 4(b)(1)(A).
the courts of appeals take, on average, a year to decide a case once an appeal has been filed;
Bureau of Justice Statistics, U.S. Courts of Appeals-Median Time Interval, tbl.B-4, available at http://www.uscourts.gov/uscourts/Statistics/JudicialBusiness/2010/appendices/
B04Sep10.pdf (on file with the Columbia Law Review). and defendants have ninety days after the court of appeals’ entry of judgment to file a petition for certiorari. 25 Sup. Ct. R. 13. These numbers suggest that a decision holding ACCA’s residual clause unconstitutionally vague could apply to around 900 prisoners whose convictions have not yet become final (the 600 sentenced this past year, plus the 300 sentenced in the latter half of the previous year).
Even if a defendant has not previously argued that ACCA’s residual clause is unconstitutionally vague, the defendant could still obtain relief by making the argument for the first time on appeal or in a petition for certiorari. Federal courts may review claims not previously argued for “plain error,” meaning errors “that affect substantial rights.” 26 Fed. R. Crim. P. 52(b). The mistaken imposition of a fifteen-year mandatory minimum clearly affects a defendant’s substantial rights by resulting in a mandatory period of incarceration that judges may not reduce. 27 See, e.g., Puckett v. United States, 556 U.S. 129, 135 (2009) (holding error affects defendant’s “substantial rights” if it “affected the outcome of the district court proceedings”); infra note 28 (listing cases finding defendant’s substantial rights affected by errors in ACCA’s application). All of the courts of appeals that have addressed the issue agree that errors regarding ACCA’s interpretation or ACCA’s application affect defendants’ substantial rights; therefore, errors regarding ACCA’s validity—i.e., ACCA’s lawful application—do as well. 28 E.g., United States v. Smith, 448 F. App’x 340, 342 (4th Cir. 2011) (finding district court’s conclusion that conviction qualified as violent felony was plain error affecting defendant’s substantial rights); United States v. Torres-Rosario, 658 F.3d 110, 116–17 (1st Cir. 2011) (same); United States v. High, 576 F.3d 429, 431 (7th Cir. 2009) (finding district court’s conclusion that conviction qualified as violent felony was plain error); United States v. Baker, 559 F.3d 443, 454–55 (6th Cir. 2009) (same where conclusion was for purposes of career offender guideline); United States v. Heikes, 525 F.3d 662, 664 (8th Cir. 2008) (finding district court’s conclusion that conviction qualified as violent felony was plain error affecting defendant’s substantial rights).
II. Group Two: Prisoners On First Collateral Review
What about prisoners whose convictions have become final? 28 U.S.C. § 2255 authorizes federal prisoners to file petitions for collateral review after their convictions have become final. Three rules—two judicial doctrines (retroactivity and procedural default) and one statutory provision (the statute of limitations)—could potentially limit whether prisoners whose convictions have become final, but who have not yet filed a petition for review under § 2255, are able to obtain relief based on a decision holding ACCA’s residual clause unconstitutionally vague. The case law on retroactivity largely points in one direction: The decision in Johnson will apply retroactively on collateral review. And because AEDPA’s statute of limitations borrows from the rules for retroactivity, the statute of limitations will not pose a bar to prisoners who file a petition for collateral review within one year of Johnson. The doctrine on procedural default, however, is much less clear. The cases do not foreclose the possibility that prisoners will be able to obtain relief based on Johnson, and some cases even affirmatively support that idea. And, this section argues, the equitable roots of procedural default suggest the doctrine should be read to permit prisoners sentenced under ACCA to obtain relief based on Johnson on collateral review.
Under Teague v. Lane, “new constitutional rules of criminal procedure will not be applicable to those cases which have become final before the new rules are announced.” 29 489 U.S. 288, 310 (1989). Although subsequent cases have applied the retroactivity analysis to decisions interpreting the scope of criminal statutes, 30 E.g., Schriro v. Summerlin, 542 U.S. 348, 351–52 (2004). a decision invalidating ACCA’s residual clause as void for vagueness would be a constitutional decision, albeit invalidating a criminal statute. Teague initially established two exceptions to the general rule against applying new rules on collateral review: New rules apply retroactively on collateral review where the rules “place certain kinds of primary, private individual conduct beyond the power of the criminal law-making authority to proscribe” or require “procedures . . . implicit in the concept of ordered liberty.” 31 489 U.S. at 307. Subsequent cases have clarified the first exception means that all new “substantive” rules apply retroactively, “includ[ing] decisions that narrow the scope of a criminal statute by interpreting its terms.” 32 Schriro, 542 U.S. at 351–52. A rule is substantive if it creates “‘a significant risk that a defendant stands convicted of an act that the law does not make criminal’ or faces a punishment that the law cannot impose upon him.” 33 Id. at 352 (quoting Bousley v. United States, 523 U.S. 614, 620 (1998)). A decision that “modifies the elements of an offense is normally substantive rather than procedural.” 34 Id. at 354.
A favorable decision in Johnson would be substantive and therefore retroactively applicable to prisoners who were sentenced under ACCA. But it will not be retroactively applicable to prisoners who were sentenced under the analogous career offender guideline.
1. Prisoners Sentenced Under ACCA. — A decision invalidating ACCA’s residual clause would be substantive and therefore should apply retroactively to defendants sentenced under ACCA whose convictions have already become final. The ACCA enhancement requires additional years of imprisonment, and this results in a real substantive liability. Without the ACCA enhancement, the statutory maximum term of imprisonment for a conviction under § 922(g) is ten years, whereas with the enhancement, the statutory mandatory minimum term of imprisonment is fifteen years. 35 18 U.S.C. § 924(a)(2) (2012). A decision finding the residual clause invalid would mean that defendants subjected to the enhancement received “a punishment the law c[ould] not impose on” them—a term of years exceeding the statutory maximum for the offense they were convicted of. A decision holding the residual clause unconstitutionally vague would also “narrow the scope of a criminal statute” by reducing the applicable penalties for § 922(g) convictions. Therefore, a decision invalidating ACCA’s residual clause would be substantive and apply retroactively on collateral review.
It should not matter for retroactivity purposes that a decision invalidating ACCA’s residual clause would not change what conduct the law proscribes—§ 922(g) would still criminalize certain individuals possessing a firearm. The Court’s Sixth Amendment jurisprudence helps to show why a decision invalidating a mandatory sentencing enhancement would be substantive. Under Apprendi v. New Jersey, the Sixth Amendment requires a jury to determine whether a defendant has committed each element of an offense. 36 530 U.S. 466, 483 n.10, 490 (2000). And the Court has held that “any fact that increases the mandatory minimum” sentence is, in effect, an “element” of an offense. 37 Alleyne v. United States, 133 S. Ct. 2151, 2155, 2158 (2013). The Court reasoned that “a fact triggering a mandatory minimum . . . produces a new penalty” and thus “constitute[s] a new, aggravated crime.” 38 Id. at 2160–61. A decision altering the defendant’s eligibility for a mandatory minimum sentence therefore alters the “elements of the offense” and, for that reason, should be considered substantive for purposes of retroactivity analysis. 39 Schriro, 542 U.S. at 354. Almendarez-Torres v. United States held that a defendant’s prior conviction does not count as an element of the offense for Sixth Amendment purposes and thus the fact of a prior conviction need not be found by a jury, even where that conviction increases the defendant’s sentence. 523 U.S. 224, 243–47 (1998). But Almendarez-Torres was decided before many of the Court’s more recent cases applying Apprendi. See, e.g., Alleyne, 133 S. Ct. 2151, 2157–58 (2013); S. Union Co. v. United States, 132 S. Ct. 2344, 2349–50 (2012); United States v. Booker, 543 U.S. 220, 232 (2005); Blakeley v. Washington, 542 U.S. 296, 301–03 (2004). Facts that increase a defendant’s sentence may also be considered substantive for other purposes, even if they are not considered substantive elements of the offense for purposes of the Sixth Amendment. For example, Peugh v. United States held that advisory guidelines, which are not considered substantive for purposes of the Sixth Amendment, are substantive laws for purposes of the Ex Post Facto Clause. 133 S. Ct. 2072, 2081–85 (2013). And Schriro suggests prior convictions that increase a defendant’s sentence might count as substantive elements of an offense for purposes of retroactivity—in Schriro, one of the statutory aggravating factors that increased the defendant’s sentence to a capital sentence was his prior convictions. 542 U.S. at 351 n.3.
Additionally, the result of a favorable decision in Johnson would be that a defendant convicted under § 922(g) could not be sentenced to a term of imprisonment exceeding ten years. The decision would mean that defendants sentenced under ACCA were given a punishment (more than ten years’ imprisonment) that the law, as written, cannot impose on them.
The reasons why the void for vagueness doctrine applies to the ACCA sentencing enhancement also suggest statutory sentencing enhancements, and specifically mandatory minimum sentences, are substantive in important respects. The void for vagueness doctrine ordinarily ensures that individuals receive “fair notice of what is prohibited.” 40 E.g., United States v. Williams, 553 U.S. 285, 304 (2008). But if the Court invalidates the residual clause on void for vagueness grounds, it will be because fair-notice principles also require individuals to know what punishment the law establishes for their offense—or at least, what minimum punishment a judge is required to impose for their offense of conviction. Therefore, merely applying the void for vagueness doctrine to ACCA suggests the statutory minimum punishment for an offense is substantive in key respects. 41 See also infra notes 87–90 and accompanying text (explaining why changes in criminal law mean doctrines should focus on what punishment is authorized rather than exclusively on what conduct is prohibited).
The court of appeals cases support the conclusion that a decision invalidating the residual clause would be substantive and therefore retroactively applicable on collateral review. 42 In Johnson v. Ponton, the Fourth Circuit concluded that the Supreme Court’s decision in Miller v. Alabama, 132 S. Ct. 2455 (2012), which held that juveniles convicted of homicide cannot be sentenced to life without parole, was not substantive and therefore not retroactively applicable on collateral review. No. 13–7824, 2015 WL 924049, at *1 (4th Cir. Mar. 5, 2015). Johnson reasoned that Miller was not substantive because it “[did] not foreclose” the possibility that a juvenile could serve life without parole. Miller, 132 S. Ct. at 2469. But the mere possibility that a decision allows some hypothetical offender to receive a particular sentence does not mean a rule is not substantive; otherwise any rule interpreting the scope of a criminal statute is not substantive—Congress could just rewrite the statute to criminalize the conduct. But see Schriro, 542 U.S. at 351–52 (stating decisions narrowing scope of criminal statute apply retroactively). See also supra notes 29–34 and accompanying text (explaining why doctrines should allow for relief in cases finding sentencing enhancements invalid); infra notes 57–66 and accompanying text (same). Thus, the Fourth Circuit’s reasoning seems to be that Miller did not foreclose the possibility that the specific defendant before the court could end up with a life-without-parole sentence for the offense he has already committed. But a favorable decision in Johnson would “foreclose” a sentencer’s ability to impose more than ten years’ imprisonment for a defendant’s prior § 922(g) conviction because Congress may not impose additional penalties for previous convictions under § 922(g). Thus, a favorable decision in Johnson forecloses the possibility that prisoners sentenced under the current versions of § 922(g) and § 924(e) may receive more than ten years’ imprisonment. Indeed, every circuit that has addressed the issue has concluded that Supreme Court decisions narrowing the scope of ACCA’s residual clause apply retroactively on collateral review. 43 E.g., Jones v. United States, 689 F.3d 621, 625–26 (6th Cir. 2012) (holding substantive rules that narrow scope are retroactively applicable); Welch v. United States, 604 F.3d 408, 415 (7th Cir. 2010) (same); United States v. Shipp, 589 F.3d 1084, 1091 (10th Cir. 2009) (same); cf. Miller v. United States, 735 F.3d 141, 147 (4th Cir. 2013) (finding Fourth Circuit opinion interpreting ACCA’s scope retroactively applicable); Lindsey v. United States, 615 F.3d 998, 1000 (8th Cir. 2010) (“[T]he district court erred when it held Begay did not apply retroactively.”). United States v. Montes, 570 F. App’x 830, 831 (10th Cir. 2014), held Descamps v. United States, 133 S. Ct. 2276 (2013), not retroactive, but on the ground that it was not a “new” rule and was a mere application of existing precedent.
2. Prisoners Sentenced Under Career Offender Guideline. — A favorable decision in Johnson will not apply retroactively to prisoners who were sentenced under the career offender guideline of the Federal Sentencing Guidelines. The Federal Sentencing Guidelines suggest recommended sentencing ranges based on the defendant’s conduct and characteristics. Federal judges first calculate the sentencing range recommended by the Guidelines, and then choose a sentence to impose, but judges are not required to impose a sentence within the recommended guidelines range. One particular sentencing guideline—the career offender guideline—results in higher recommended sentencing ranges for offenders with three or more convictions for violent felonies, 44 U.S. Sentencing Guidelines Manual § 4B1.1 (2014). and defines “crimes of violence” using the same language as ACCA. 45 Compare id. § 4B1.2(a) (defining “crime of violence” as having element of physical force against another or is otherwise one of denoted crimes listed in section), with 18 U.S.C. § 924(e)(2)(B) (2012) (utilizing substantially same language to define “violent felony”). Courts interpret the phrase violent felony to mean the same thing under the Guidelines as under ACCA. 46 E.g., United States v. Ruvulcaba, 627 F.3d 218, 224 (6th Cir. 2010) (“[W]e apply the same analysis to determine whether a crime is a violent felony under the Armed Career Criminal Act and to determine whether a crime is a crime of violence under the Guidelines.”); United States v. Partee, 373 F. App’x 602, 603 (7th Cir. 2010) (“The term ‘violent felony’ under the ACCA and ‘crime of violence’ under the career offender guideline are nearly identical, and we apply the same interpretation to both provisions when determining whether a prior conviction triggers increased penalties.”).
Because the Guidelines are merely recommendations, they are less “substantive” than ACCA’s statutory minimum sentencing enhancement. The application of the career offender guideline, unlike the application of ACCA, does not necessarily alter the prescribed range of sentences to which a criminal defendant is exposed. 47 Alleyne v. United States, 133 S. Ct. 2151, 2162–63 (2013). And under the Court’s Sixth Amendment sentencing jurisprudence, a fact that increases a defendant’s sentence is an element of the offense only where the existence of that fact requires a judge to sentence a defendant to more time, not where it merely recommends that the judge do so. 48 Id.; see United States v. Booker, 543 U.S. 220, 223 (2005) (“If the Guidelines . . . could be read as merely advisory provisions that recommended, rather than required, the selection of particular sentences in response to differing sets of facts . . . [they] would not implicate the Sixth Amendment.”). Additionally, defendants sentenced under the career offender guideline may not have necessarily received a sentence that exceeds the term of imprisonment that is authorized by statute for their offense of conviction. 49 See infra note 73 (discussing cases where defendants sentenced under career offender guideline received sentence below statutory maximum for offense of conviction). Thus, as to these defendants, a decision in Johnson does not mean they received a sentence that the law cannot impose on them, or a sentence unauthorized by law. Most district courts have thus concluded that decisions interpreting ACCA are not retroactive, but only as to prisoners sentenced under the career offender guideline. 50 E.g., United States v. Ross, No. 06-cr-132-jcs, 2010 WL 148397, at *2 (W.D. Wis. Jan. 12, 2010) (finding Begay and Chambers not retroactive to interpretations of career offender guideline based on prior district court opinion); United States v. Holt, 677 F. Supp. 2d 1063, 1065 (W.D. Wis. 2009) (finding Begay and Chambers not retroactively applicable to interpretations of career offender guideline); United States v. Johnson, No. 04-269, 2009 WL 2611279, at *2 (D. Minn. Aug. 24, 2009) (finding Begay not retroactively applicable to interpretations of career offender guideline); United States v. Campbell, No. 6:06-812-HMH, 2009 WL 1254287, at *4 (D.S.C. May 1, 2009) (same); cf. Cadieux v. United States, No. 03-41-B-W, 2009 WL 1286421, at *7 (D. Me. May 8, 2009) (finding First Circuit opinion interpreting ACCA not intended to be retroactively applicable to defendant sentenced under ACCA). See supra notes 44–46 and accompanying text for a description of the Federal Sentencing Guidelines.
B. Procedural Default
The doctrine of procedural default bars consideration of a claim on collateral review where the defendant could have, but did not raise the claim on direct appeal. 51 E.g., United States v. Frady, 456 U.S. 152, 167–68 (1982) (“Under this standard, to obtain collateral relief based on trial errors to which no contemporaneous objection was made, a convicted defendant must show both (1) ‘cause’ excusing his double procedural default, and (2) ‘actual prejudice’ resulting from the errors of which he complains.”); Lynn v. United States, 365 F.3d 1225, 1232–33 (11th Cir. 2004) (“Because collateral review is not a substitute for a direct appeal, the general rule ha[s] developed that . . . a defendant must assert all available claims on direct appeal.”). Procedural default may affect two kinds of prisoners: (1) prisoners who have been sentenced under ACCA, and (2) prisoners sentenced under the separate, career offender guideline. 52 See supra notes 44–46 and accompanying text for a description of the Federal Sentencing Guidelines. With respect to prisoners sentenced under ACCA, the doctrine is slightly unclear and largely boils down to two issues. One is whether prisoners sentenced under ACCA should be viewed as “actually innocent” of something in light of Johnson. Two is whether federal courts would (and could) find that it is a “fundamental miscarriage of justice” for prisoners to receive a sentence above the statutory maximum for their offense of conviction. With respect to prisoners sentenced under the career offender guideline, the doctrine is clearer—these prisoners will not be able to obtain relief based on a favorable decision in Johnson.
1. Prisoners Sentenced Under ACCA. — Several courts have invoked procedural default to decline to reach claims that a defendant’s prior convictions were mistakenly classified as violent felonies in light of subsequent cases. 53 See, e.g., McKay v. United States, 657 F.3d 1190, 1192 (11th Cir. 2011) (defendant procedurally defaulted claim that he was improperly sentenced under career offender guideline); United States v. Gibson, 424 F. App’x 461, 466 (6th Cir. 2011) (same); United States v. Coley, 336 F. App’x 933, 935–36 (11th Cir. 2009) (same). However, some courts justify applying procedural default on the ground that federal collateral review is rarely available for “non-constitutional error[s]” 54 E.g., Lynn, 365 F.3d at 1232. such as the proper interpretation of “violent felony.” 55 E.g., Hunter v. United States, 559 F.3d 1188, 1189–90 (11th Cir. 2009). If the purpose of procedural default in federal collateral review is to limit collateral proceedings to claims of constitutional error, courts may choose not to use procedural default to bar claims that ACCA is unconstitutionally vague. However, most courts maintain that procedural default bars consideration of any claim whose “‘merits can be reviewed without further factual development.’” 56 E.g., Lynn, 365 F.3d at 1233 n.14 (quoting Mills v. United States, 36 F.3d 1052, 1055 (11th Cir. 1994)). Because no factual development is required to review the claim that ACCA’s residual clause is unconstitutionally vague, most courts will find that a defendant procedurally defaulted the claim if the defendant did not make the argument on appeal.
This is not the end of the matter, because courts will hear a procedurally defaulted claim if the defendant establishes “cause” and “prejudice.” 57 E.g., Massaro v. United States, 538 U.S. 500, 504 (2003). The additional term of years resulting from ACCA’s mandatory sentencing enhancement constitutes prejudice; 58 E.g., United States v. Thomas, 274 F.3d 655, 669 (2d Cir. 2001) (“It is beyond cavil that imprisonment for an additional 52 months . . . constitutes prejudice.”); cf. Lafler v. Cooper, 132 S. Ct. 1376, 1386 (2012) (“[I]neffective assistance of counsel during a sentencing hearing can result in Strickland prejudice because ‘any amount of [additional] jail time has Sixth Amendment significance.’” (quoting Glover v. United States, 531 U.S. 198, 203 (2001))). therefore, these cases will turn on whether the defendant has established “cause.” Traditionally, “cause” has included only instances where the claim’s sheer novelty meant there was no legal basis for raising the claim before, or where the claim was so obvious that counsel was constitutionally ineffective for failing to raise it. 59 E.g., Bousley v. United States, 523 U.S. 614, 622 (1998) (“[A] a claim that ‘is so novel that its legal basis is not reasonably available to counsel’ may constitute cause for a procedural default . . . .” (quoting Reed v. Ross, 468 U.S. 1, 16 (1984))); Murray v. Carrier, 477 U.S. 478, 488 (1986) (“So long as a defendant is represented by counsel whose performance is not constitutionally ineffective under the standard established in Strickland v. Washington, we discern no inequity in requiring him to bear the risk of attorney error that results in a procedural default.” (citation omitted)). This standard will not be of much use to prisoners seeking to make use of Johnson. 60 Cf., e.g., Eve Brensike Primus, Procedural Obstacles to Reviewing Ineffective Assistance of Trial Counsel Claims in State and Federal Postconviction Proceedings, 24 Crim. Just. 6, 8–10 (2009) (suggesting many claims fall somewhere in between being so novel that counsel could not have predicted or made them and being so obvious that counsel was ineffective for raising them). The claim’s legal basis is hardly novel: Justice Scalia has been suggesting ACCA is unconstitutionally vague for the last eight years, and the void for vagueness doctrine is nothing new. Courts are also unlikely to find that counsel was ineffective for failing to argue that ACCA was unconstitutionally vague because circuit precedent foreclosed the argument—every circuit has rejected the claim that ACCA’s residual clause is unconstitutionally vague. 61 E.g., United States v. Van Mead, 773 F.3d 429 (2d Cir. 2014); United States v. Johnson, 526 F. App’x 708 (8th Cir. 2013); United States v. Hawkins, 512 F. App’x 746 (10th Cir. 2013); United States v. Brown, 516 F. App’x 461 (6th Cir. 2013); United States v. Guadalupe, 493 F. App’x 146 (1st Cir. 2012); United States v. Battle, 494 F. App’x 404 (4th Cir. 2012); United States v. Jones, 689 F.3d 696 (7th Cir. 2012); United States v. Devo, 457 F. App’x 908 (11th Cir. 2012); United States v. Gibbs, 656 F.3d 180 (3d Cir. 2011); United States v. Wilson, 64 F. App’x 417 (5th Cir. 2003); United States v. Powell, 967 F.2d 595 (9th Cir. 1992). The fact that case law foreclosed the argument that ACCA is unconstitutionally vague does not excuse counsel’s failure to make it—perceived futility to raising a claim does not provide cause to excuse a procedural default. 62 Smith v. Murray, 477 U.S. 527, 536 (1986). Put another way, the claim is neither so novel that counsel could not have conceived of it (thus excusing the defendant from raising it), nor so obvious that counsel was inept and thus constitutionally ineffective for not making it (thus establishing cause to excuse the procedural default).
There is, however, another circumstance where courts hear procedurally defaulted claims: fundamental miscarriages of justice. 63 E.g., United States v. Frady, 456 U.S. 152, 172 (1982). The precise scope of the fundamental miscarriage of justice exception is unclear. At the very least, the exception includes cases where a defendant establishes he is actually innocent. 64 Spencer v. United States, 773 F.3d 1132, 1139 (11th Cir. 2014); Hawkins v. United States, 706 F.3d 820, 829 (7th Cir. 2013); Sun Bear v. United States, 644 F.3d 700, 706 (8th Cir. 2011) (en banc); United States v. Pettiford, 612 F.3d 270, 275, 282 (4th Cir. 2010); Spence v. Superintendent, Great Meadow Corr. Facility, 219 F.3d 162, 171 (2d Cir. 2000). It is less clear whether a defendant who cannot establish he is actually innocent could nonetheless establish there has been a fundamental miscarriage of justice excusing his procedural default. 65 See infra note 81 and accompanying text (discussing cases on this question). Read carefully, however, the relevant decisions suggest courts could, and probably should, find defendants sentenced under ACCA have established “cause” and “prejudice” for their default, either because the defendants are actually innocent in light of Johnson, or because a sentence exceeding the statutory maximum for the offense of conviction is a fundamental miscarriage of justice. 66 See supra note 64 and accompanying text (discussing cases finding defendants sentenced under ACCA either actually innocent or victims of fundamental miscarriage of justice).
The Supreme Court’s cases on the actual innocence exception could be read to suggest the actual innocence exception is available to defendants seeking to establish they are innocent of or ineligible for the sentence they received. Schlup v. Delo held that a claim of actual innocence could excuse a defendant’s procedural default where the defendant established he “probably” did not commit the underlying offense. 67 513 U.S. 298, 323–24 (1994). Sawyer v. Whitley held that a defendant could establish actual innocence, thus excusing a procedural default, if the defendant established by “clear and convincing evidence” that he was ineligible for a capital sentence, even where he committed the underlying offense. 68 505 U.S. 333, 344–45 (1992). In Sawyer, state law provided that defendants could be sentenced to death where the state established certain “statutory aggravating circumstances.” And Sawyer held that where there is clear and convincing evidence that a statutory aggravating circumstance was inappropriately applied, the defendant is ineligible for or innocent of the death sentence and there has been a fundamental miscarriage of justice. 69 Id. Sawyer’s reasoning could apply to defendants sentenced under the residual clause: A favorable decision in Johnson would mean that there will be no evidence that ACCA’s statutory mandatory minimum applied to defendants sentenced under the residual clause. Thus, these defendants are ineligible for or innocent of more than ten years’ imprisonment, and therefore there has been a fundamental miscarriage of justice.
But Sawyer involved a death sentence, and it is unclear whether defendants may be actually innocent of noncapital sentences as well as capital ones. 70 The Court granted certiorari in Dretke v. Haley to decide whether the actual innocence exception to procedural default applied to claims that resulted in mistaken noncapital sentences, but ultimately resolved the case on other, narrower grounds. 541 U.S. 386, 388–89 (2004) (“[T]he question before us is whether this exception applies where an applicant asserts ‘actual innocence’ of a noncapital sentence. Because the District Court failed first to consider alternative grounds for relief . . . that might obviate any need to reach the actual innocence question, we vacate . . . and remand.”). Several courts of appeals have suggested a prisoner is not actually innocent where he establishes he was improperly subjected to the career offender guideline in light of subsequent precedent, and some of these cases contain broad statements that the actual innocence exception is categorically unavailable for sentencing errors that result in a noncapital sentence. 71 United States v. Nichols, 472 F. App’x 856, 857 (10th Cir. 2012) (“A defendant, however, cannot be actually innocent of a non-capital sentence.”). But, on closer inspection, these cases do not actually decide whether defendants convicted under § 922(g) (for unlawful possession of a firearm) may establish actual innocence, thus excusing a procedural default. All but one of the cases finding the actual innocence exception inapplicable did not involve defendants convicted under § 922(g). 72 See infra note 79, and text accompanying notes 92–93 (listing cases that did not involve defendants convicted under § 922(g) and describing lone exception). More importantly, in all of these cases, the defendant ultimately received a sentence below the statutory maximum for the defendant’s offense of conviction: The mistaken application of the Guidelines did not result in the defendant receiving a sentence that exceeded the statutory maximum for the offense of conviction. 73 Whiteside v. United States, 775 F.3d 180, 182 (4th Cir. 2014) (controlled substances violation with statutory maximum of forty years and sentenced to seventeen and a half years); Spencer v. United States, 773 F.3d 1132, 1142 (11th Cir. 2014) (controlled substances violation with statutory maximum of twenty years and sentenced to fewer than thirteen years); Hawkins v. United States, 706 F.3d 820, 822 (7th Cir. 2013) (assault with weapon with statutory maximum of twenty years and sentenced to nineteen years); Sun Bear v. United States 644 F.3d 700, 706 (8th Cir. 2011) (en banc) (second-degree murder with statutory maximum of life and sentenced to thirty years). By contrast, a favorable decision in Johnson will mean that prisoners who have been convicted of unlawful possession of a firearm under § 922(g) and sentenced under the residual clause will have received a sentence that exceeds the statutory maximum for the offense of conviction. Without the ACCA enhancement, the statutory maximum for a § 922(g) conviction is ten years, 74 18 U.S.C. § 924(a)(2) (2012). and with the enhancement the statutory minimum is fifteen years. 75 Id. (e)(2)(B)(ii).
Indeed, the relevant cases contain statements that generally support the idea that a defendant may be actually innocent, thus excusing a procedural default, if he was improperly sentenced to a mandatory minimum sentence. In every single one of the cases involving the career offender guideline, the courts suggested the defendant would be actually innocent if a sentencing error resulted in the defendant receiving a sentence above the statutory maximum for the defendant’s offense of conviction. 76 E.g., Whiteside v. United States, 748 F.3d 541, 557–60 (4th Cir. 2014) (Wilkinson, J., dissenting) (suggesting differences between cases involving claims of federal constitutional error and errors of federal law and cases involving “advisory Guidelines determinations”), rev’d en banc, 775 F.3d 180 (4th Cir. 2014); Spencer, 773 F.3d at 1139 (framing scope of procedural default in terms of whether prisoner was “sentenced below the statutory maximum”); Hawkins, 706 F.3d at 822–24 (framing scope of procedural default in terms of whether error results in “sentence that exceeds the statutory maximum sentence”); Sun Bear, 644 F.3d at 703–04 (suggesting there are differences between cases involving career offender guideline and cases where “application of § 924(e)(1) . . . increased the defendant’s statutory maximum sentence, which will be true with many if not most felon-in-possession convictions”); id. at 705 (“‘[A] sentence is not illegal . . . unless it exceeds the statutory maximum for the offense of conviction.’” (quoting United States v. Stobaugh, 420 F.3d 796, 804 (8th Cir. 2005))). The cases reason that where “[a] sentence . . . violates a statute”—such as where a defendant receives a sentence that exceeds the statutory maximum for his offense of conviction—this “could well be thought an error grave enough to warrant relief . . . a ‘fundamental error equivalent to actual innocence.’” 77 Brown v. Rios, 696 F.3d 638, 641 (7th Cir. 2012). This would be the case for § 922(g) offenders.
Part of the difficulty lies in conceptualizing what, exactly, prisoners sentenced under ACCA are actually innocent of. Courts could characterize prisoners sentenced under ACCA’s residual clause as actually innocent of a particular kind of offense conduct—the pattern of “violent” crimes referenced in federal law. But it is not clear that the pattern of “violent” crimes referenced in federal law is part of the “offense” for which the defendant has been convicted. 78 See Low & Johnson, supra note 5, at 56–57 (suggesting it is not part of offense conduct because it is prior crime and therefore not traditionally considered element of offense). The defendant was convicted of conduct—possession of a firearm—that remains unlawful under § 922(g), whether or not the defendant previously committed three or more violent crimes. Courts could instead depict these prisoners as actually innocent of their sentence because they received a punishment that the law as written cannot impose on them: a term of imprisonment exceeding ten years. But the doctrine is unclear on whether a defendant may establish he is actually innocent of a noncapital sentence. 79 See, e.g., Sawyer v. Whitley, 505 U.S. 333, 339 (1992) (“[T]he miscarriage of justice exception is concerned with actual as compared to legal innocence.”); United States v. Addonizio, 442 U.S. 178, 186–87 (1979) (distinguishing case on ground that “[t]he subsequent development . . . was a change in the substantive law that established that the conduct for which petitioner had been convicted and sentenced was lawful”). But this reads too much into the relevant statements, which could fairly be considered dicta. See, e.g., Sawyer, 505 U.S. at 344–45 (permitting defendants to establish actual innocence based on absence of statutory aggravating circumstances, despite statement regarding “actual as opposed to legal” innocence, because “[b]oth the elements of the crime and statutory aggravating circumstances in Louisiana are used to narrow the class of defendants eligible for the death penalty”); Addonizio, 442 U.S. at 186–87 (refusing to find actual innocence based on change in “the way in which the court’s judgment and sentence would be performed” in limited context of parole violation). See also infra notes 87–90 and accompanying text (explaining why actual innocence should no longer be conceived solely in terms of crime of conviction). Courts could instead phrase their determination in terms of “miscarriage[s] of justice,” and find that a miscarriage of justice occurs where “[a]n unlawful or illegal sentence is . . . imposed . . . in excess of statutory authority.” 80 Sun Bear, 644 F.3d at 705; see also Brown, 696 F.3d at 641 (“A sentence that violates a statute, as distinct from a sentence permitted by a statute though more severe than authorized by the guidelines, could well be thought an error grave enough to warrant relief in a habeas corpus proceeding—a ‘fundamental error equivalent to actual innocence,’ . . . .” (internal citation omitted)). But it is unclear whether the fundamental miscarriage of justice exception extends to anything other than cases of actual innocence. 81 Compare, e.g., Sawyer, 505 U.S. at 339 (“[T]he miscarriage of justice exception is concerned with actual as compared to legal innocence.”), and Herrera v. Collins, 506 U.S. 390, 404 (1993) (“This . . . fundamental miscarriage of justice exception, is grounded in the ‘equitable discretion’ of habeas courts to see that federal constitutional errors do not result in the incarceration of innocent persons.”), with Hill v. United States, 368 U.S. 424, 428 (1962) (referring to “complete miscarriage of justice[s]” cognizable on collateral review). When the Court equates fundamental miscarriages of justice with actual innocence it often does so to emphasize the narrowness of the exception. See, e.g., Schlup v. Delo, 513 U.S. 298, 321 (1995) (“To ensure that the fundamental miscarriage of justice exception would remain ‘rare’ and would only be applied in the ‘extraordinary case’ . . . this Court explicitly tied the miscarriage of justice exception to the petitioner’s innocence.”). But it would not unduly expand the fundamental miscarriage of justice exception to include prisoners who have been sentenced to a term exceeding the statutory maximum for their offense of conviction—judges rarely invalidate sentencing enhancements or statutory mandatory minimums on constitutional grounds. See Philip P. Frickey, Interpretation on the Borderline: Constitution, Canons, Direct Democracy, 1 J. Legis. & Pub. Pol’y 105, 129 (1997) (noting void for vagueness doctrine rarely invoked).
Despite these difficulties, there are sound reasons why courts should not narrow the actual innocence or fundamental miscarriage of justice exception such that it applies only to cases where the defendant’s conduct has been declared lawful or beyond the reach of criminal law. Sawyer held that a defendant may be actually innocent of a sentence, 82 See Sawyer, 505 U.S. at 344–45 (permitting defendants to establish actual innocence based on absence of statutory aggravating circumstances because “[b]oth the elements of the crime and statutory aggravating circumstances in Louisiana are used to narrow the class of defendants eligible for the death penalty”). and while Sawyer involved a death sentence, subsequent cases have undermined the extent to which various sentencing doctrines are limited to the death penalty context. 83 See, e.g., Miller v. Alabama, 132 S. Ct. 2455, 2460 (2012) (holding juveniles convicted of homicide cannot be sentenced to life without parole); Graham v. Florida, 130 S. Ct. 2011, 2030 (2010) (holding juveniles convicted of nonhomicide offenses cannot be sentenced to life without parole); Carol S. Steiker & Jordan M. Steiker, Miller v. Alabama: Is Death (Still) Different?, 11 Ohio St. J. Crim. L. 37, 39 (2013) (“The net-effect of ‘juveniles are different’ and ‘LWOP is different’ might be the functional end of ‘death is different.’”). The Court’s Sixth Amendment jurisprudence also points in this direction. Under Apprendi v. New Jersey, a jury must determine whether a defendant has committed each element of an offense. 84 530 U.S. 466, 483 n.10, 490 (2000). Early on, the Court held that any fact that makes the defendant eligible for a death sentence by statute is an element of the offense. 85 Ring v. Arizona, 536 U.S. 584, 589 (2002) (“Capital defendants, no less than noncapital defendants, we conclude, are entitled to a jury determination of any fact on which the legislature conditions an increase in their maximum punishment.”). Subsequent cases have extended that rule to apply to noncapital sentences provided for by statute. Now, any fact that governs whether a statutory minimum sentence applies—including a noncapital additional term of years—is an element of the offense that must be found by a jury. 86 Alleyne v. United States, 133 S. Ct. 2151, 2155 (2013).
Additionally, the increasing scope of criminal law’s reach makes an actual innocence exception focused exclusively on conduct that has been declared lawful somewhat antiquated. 87 See, e.g., William J. Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 512 (2001) (“Criminal law is both broad and deep: a great deal of conduct is criminalized, and of that conduct, a large proportion is criminalized many times over. I believe these propositions would be accepted by anyone who read an American criminal code, state or federal.”). Many criminal justice issues today, including overcriminalization, are framed in terms of the problems with lengthy sentences and mandatory minimum sentences, rather than with new definitions of crime. 88 See, e.g., Rachel E. Barkow, Federalism and the Politics of Sentencing, 195 Colum. L. Rev. 1276, 1278–79 (2005) (noting “get tough” on crime rhetoric increasingly leads to “harsher sentences”); Rachel E. Barkow, Our Federal System of Sentencing, 58 Stan. L. Rev. 119, 120 (2005) (“Congress has responded to high-profile local crimes not only with new federal laws, but also with longer sentences for existing laws.”). If one key issue in criminal law is that prisoners are unjustly sentenced to excessive terms of imprisonment, these prisoners could sensibly be thought of as innocent of their sentences. 89 See, e.g., Darryl K. Brown, Can Criminal Law Be Controlled?, 108 Mich. L. Rev. 971, 990 (2010) (noting “much of America’s contemporary policy of extreme punishment arises from sentencing” and that “the biggest cause of overpunishment is, simply . . . excessive sentences”); Erik Luna, The Overcriminalization Phenomenon, 54 Am. U. L. Rev. 703, 718–19 (2005) (noting failure to recognize overcriminalization as “broad phenomenon encompassing a multiplicity of concerns” including lengthy sentences); Malcolm C. Young, Special Interests, Principles, and Sentencing Reform in America, 96 J. Crim. L. & Criminology 1509, 1510 (2006) (“In pursuit of a claim of near-innocence for the alleged victims of ‘overcriminalization,’ the authors of these chapters sidestep the central issue in sentencing, which is simply what amount of punishment a society should impose on wrongdoers.”). It may have made some sense three decades ago to think of actual innocence or fundamental miscarriages of justice solely in terms of whether a person’s conduct was criminal. But the same is not true today, when many think of similar criminal justice issues in terms of the sentences imposed, rather than what crimes are defined by law. 90 See supra note 88 and accompanying text (noting criticism of lengthy sentences); cf. Carol S. Steiker, Second Thoughts About First Principles, 107 Harv. L. Rev. 820, 825 (1994) (urging scholars to think about scope of various criminal procedure doctrines in terms of “historical changes” and “changes in law enforcement institutions”).
The courts of appeals have also generally coalesced around the idea that a defendant would be actually innocent if a sentencing error resulted in the defendant receiving a sentence above the statutory maximum for the defendant’s offense of conviction. 91 See supra note 76 (discussing these cases). United States v. Pettiford is the one exception. 92 612 F.3d 270 (4th Cir. 2010); see supra note 73 (listing cases where statutory maximum exceeded erroneous sentence). The analysis in the preceding paragraphs arguably suggests Pettiford was wrongly decided. There, the defendant convicted under § 922(g) tried to argue he was improperly subjected to the ACCA enhancement in light of subsequent precedent, and maintained that his actual innocence provided cause to excuse his procedural default. Pettiford rejected the argument on the ground that actual innocence applies only where the defendant argues he did not commit the previous crimes, not to claims that the defendant’s previous crimes do not constitute violent felonies under ACCA. 93 612 F.3d at 284. But even Pettiford could be read not to foreclose relief for prisoners sentenced under ACCA. Pettiford suggested the actual innocence exception was unavailable because defendants should not use federal collateral review to relitigate their prior convictions, and particularly prior state convictions. 94 Id. at 282 (“To begin with, a federal sentencing proceeding is not ordinarily an appropriate forum in which to challenge the validity of a prior state conviction.”). In Pettiford, the defendant sought to argue what his state conviction was “for” in light of various state-court documents. 95 See Brief of Appellee at 19–22, United States v. Pettiford, 612 F.3d 270 (2010) (No. 09-4119) 2009 WL 4396281 (relying on charging documents and statement of probable cause to challenge grounds of conviction). However, the claim that ACCA is unconstitutional does not require courts to look into what happened in the earlier-in-time conviction or sentencing proceedings; it concerns only the general validity and lawfulness of ACCA’s residual clause. Pettiford also stressed a defendant is not actually innocent where courts arrive at a mistaken “legal classification” of the defendant’s predicate crimes. 96 612 F.3d at 284. But the claim that ACCA is unconstitutionally vague is not an argument about a mistaken legal classification; it is an argument about an invalid and unenforceable one.
2. Prisoners Sentenced Under the Career Offender Guideline. — The equitable roots of procedural default may lead some to think the fundamental miscarriage of justice exception also applies to defendants sentenced under the career offender guideline, thus excusing these defendants’ procedural default. Procedural default is an equitable doctrine, and when the doctrine applies to federal prisoners, it is rooted only in concerns of finality rather than comity or federalism (as is the case when the doctrine applies to state prisoners). 97 See Martinez v. Ryan, 132 S. Ct. 1309, 1318 (2012) (“The rules for when a prisoner may establish cause to excuse a procedural default are elaborated in the exercise of the Court’s discretion. These rules reflect . . . equitable judgment[s].”); United States v. Frady, 456 U.S. 152, 167–68 (1982) (establishing cause-and-prejudice standard for federal prisoners); cf. Kendall Turner, Note, A New Approach to the Teague Doctrine, 66 Stan. L. Rev. 1159, 1173–74 (2014) (arguing federalism and comity considerations do not justify Teague’s application to federal prisoners on federal collateral review). But there are considerable costs to permitting every defendant sentenced under the career offender guideline to obtain resentencing in light of Johnson, and the doctrine likely forecloses these prisoners’ ability to obtain relief.
The equities may be different in cases where a defendant was sentenced under a guideline whose language is unconstitutionally vague than in cases where a defendant was sentenced under a misinterpretation of the Guidelines. Consider the different finality interests in these two kinds of cases. Many decisions maintain collateral review is unavailable for Guidelines errors because every Guidelines application affects the defendant’s sentence, so every case (re)interpreting the Guidelines could require resentencing hundreds or thousands of defendants. 98 Spencer v. United States, 773 F.3d 1132, 1144 (11th Cir. 2014) (“[F]inality is a concern . . . for a first motion to vacate a sentence”); Whiteside v. United States, 775 F.3d 180, 186–87 (4th Cir. 2014) (noting 80,000 prisoners sentenced each year); Hawkins v. United States, 706 F.3d 820, 824 (7th Cir. 2013) (same). These fears may have been especially acute in cases involving the career offender guideline, which was constantly in flux in light of the Supreme Court cases reinterpreting ACCA. However, courts rarely find Federal Sentencing Guideline language to be unconstitutionally vague. For a court to reach that conclusion, the Guidelines language must appear word for word in a criminal statute, and a court must find that statute void for vagueness. And applying the fundamental miscarriage of justice exception to cases where defendants were sentenced under a guideline whose language has been declared unconstitutionally vague would not necessarily mean the exception applies to any case involving an incorrect interpretation or application of the Guidelines.
However, it is unlikely that many courts will conclude the fundamental miscarriage of justice exception allows defendants sentenced under the career offender guideline to be resentenced in light of Johnson. Several decisions suggest the actual innocence exception is categorically unavailable for Guidelines errors where the defendant received a sentence below the statutory maximum applicable to his offense.
Sun Bear v. United States, 644 F.3d 700, 704–05 (8th Cir. 2011) (suggesting Guidelines errors are not cognizable on collateral review because they do not result in sentences “in excess of statutory authority”); United States v. Peterman, 249 F.3d 458, 462 (6th Cir. 2001) (“Courts have generally declined to collaterally review sentences that fall within the statutory maximum.”).
Additionally, although the interests in finality are less powerful in cases involving unconstitutionally vague guidelines, they are still substantial: Many more defendants are subjected each year to the career offender guideline than to ACCA—approximately 2,000 offenders each year receive the career offender guideline enhancement, whereas only 600 offenders each year are sentenced under ACCA.
Guidelines estimates suggest over 2,000 offenders each year are subjected to the career offender guideline. This number does not differentiate between defendants convicted of unlawful firearms possession in violation of § 922(g) rather than other offenses. U.S. Sentencing Comm’n, Quick Facts: Career Offender, available at http://
www.ussc.gov/sites/default/files/pdf/research-and-publications/quick-facts/Quick_Facts_Career_Offender.pdf (on file with the Columbia Law Review) (last visited Mar. 2, 2015). Courts would therefore have to resentence thousands of more defendants if courts conclude that defendants sentenced under the career offender guideline qualify as actually innocent for purposes of procedural default.
C. Statute of Limitations
Section 2255(f) establishes a one-year statute of limitations for filing a petition for collateral review which runs from the later of two dates: the date on which the conviction became final, or “the date on which the right asserted was initially recognized by the Supreme Court, if that right has been newly recognized by the Supreme Court and made retroactively applicable to cases on collateral review.” The retroactivity analysis suggests prisoners’ petitions for collateral review would be timely if the prisoner files the petition within a year of a decision holding ACCA’s residual clause unconstitutional. 101 Cf. Whiteside, 775 F.3d at 183–85 (holding court of appeals decision regarding ACCA did not constitute “facts supporting the claim or claims presented” restarting limitations period if decision was not retroactively applicable under § 2255(f)(4)). The circuits agree the statute of limitations is restarted under § 2255(f) when the Supreme Court recognizes the right, even if the Court does not explicitly make the right retroactive. 102 See Narvaez v. United States, 674 F.3d 621, 625–26 (7th Cir. 2011) (finding petition timely because it was filed within one year of decision court of appeals found retroactively applicable); Fischer v. United States, 285 F.3d 596, 600 (7th Cir. 2002) (finding statute of limitations begins even though it is “[d]istrict and appellate court . . . opinions ‘holding’ that a decision applies retroactively to cases on collateral review.”); cf. United States v. Jones, 689 F.3d 696, 626–28 (7th Cir. 2012) (holding statute of limitations restarted once Begay was decided); Lindsey v. United States, 615 F.3d 998, 999–1000 (8th Cir. 2010) (suggesting same); Welch v. United States, 604 F.3d 408, 413–15 (7th Cir. 2010) (same). Additionally, AEDPA’s statute of limitations is tolled for claims of actual innocence. 103 McQuiggin v. Perkins, 133 S. Ct. 1924, 1928 (2013). Although McQuiggin specifically concerned the statute of limitations in 28 U.S.C. § 2244, there is little reason to think it would be tolled for innocence under that provision, which is applicable to state prisoners, but not under § 2255, which is applicable to federal prisoners. Thus, if Johnson means that prisoners sentenced under ACCA’s residual clause are actually innocent, their actual innocence could excuse their untimely petitions for collateral review.
III. Group Three: Prisoners On Second or Successive Petitions For Collateral Review
The greatest barrier to relief may be the one applicable to a third category of prisoners—prisoners whose convictions have become final and who have previously filed at least one petition for review under 28 U.S.C. § 2255. But even this provision could be interpreted to allow prisoners sentenced under ACCA to obtain relief through second or successive petitions for collateral review. Indeed, several courts of appeals have interpreted the provision to allow prisoners to raise claims that they were mistakenly sentenced as a career offender under ACCA.
AEDPA greatly limits the availability of second or successive petitions for collateral review: Prisoners may only file a second or successive petition where a panel certifies that newly discovered evidence proves the defendant innocent of the offense, or certifies there is “a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court, that was previously unavailable.” 104 28 U.S.C. § 2255(h) (2012). Unlike the statute of limitations applicable to initial petitions for collateral review, the gateway to second petitions for collateral review opens only once the rule was made retroactive by the Supreme Court. 105 E.g., Tyler v. Cain, 533 U.S. 656, 662 (2001) (interpreting similarly worded provision in § 2244, which is applicable to state-court prisoners, in this way); In re Jackson, 776 F.3d 292, 293 (5th Cir. 2015) (per curiam) (stating statute of limitations restarts only if litigant identifies “‘a Supreme Court decision that either expressly declares the collateral availability of the rule . . . or applies the rule in a collateral proceeding’”). This would mean that prisoners subjected to ACCA’s sentencing enhancement could only file second or successive petitions once the Supreme Court explicitly recognizes that its decision invalidating ACCA’s residual clause applies retroactively. 106 This assumes the Court does not state or make the rule retroactively applicable to collateral review in Johnson, which it could only do in dicta. It is unclear whether the Court could “make” a rule retroactive for purposes of the statute of limitations through dicta. Tyler arguably suggests it could not: “The only way the Supreme Court can, by itself, ‘lay out and construct’ a rule’s retroactive effect, or ‘cause’ that effect ‘to exist, occur, or appear,’ is through a holding.” 553 U.S. at 663.
There is, however, another avenue these prisoners may use to obtain relief. 28 U.S.C. § 2241 authorizes federal courts to grant writs of habeas corpus, and § 2241 does not contain any of the limitations applicable to petitions for review under § 2255. However, there are limits on when prisoners may use § 2241 to evade § 2255’s limitations: Section 2255(e) (the “savings clause”) only permits prisoners to utilize § 2241 where “the remedy [provided by § 2255] is inadequate or ineffective to test the legality of [the prisoner’s] detention.”
Section 2255(e)’s savings clause permits prisoners sentenced under ACCA who have already filed one petition for collateral review to obtain relief under § 2241. Section 2255’s limitations on second or successive petitions mean that defendants sentenced under ACCA will linger in prison—possibly for a term of imprisonment that exceeds the statutory maximum for their offense—because they must wait to file a petition for collateral review until the Supreme Court explicitly recognizes that Johnson applies retroactively. 107 The Court could conceivably do so either implicitly or explicitly (in dicta) in Johnson. See supra note 106 (discussing Court’s capabilities in this area). That makes the § 2255 remedy nonexistent—and thus inadequate or ineffective—to test the legality of these prisoners’ detentions. 108 Some may think this creates an impermissible work-around to Tyler’s rule that the Supreme Court must make a rule retroactive before a second or successive petition is permitted. But § 2241 is an explicit workaround to limitations contained in § 2255, as § 2255(e) specifies. However narrow the savings clause may be, it should, at a minimum, allow defendants to obtain relief where they were subjected to an unconstitutionally vague mandatory sentencing enhancement that resulted in a sentence that exceeds the statutory maximum for their offense of conviction.
The few cases interpreting the savings clause confirm this intuition. Moore v. Warden, FCC Coleman-Medium held that § 2255’s limitations on second or successive petitions made the § 2255 remedy inadequate for a defendant who was mistakenly subjected to the ACCA enhancement in light of subsequent precedent. 109 568 F. App’x 838, 840–41 (11th Cir. 2014). Indeed, in several cases “the government concede[d] that a claim that a sentencing error [that] resulted in a sentence longer than the statutory maximum may be brought in an initial § 2255 motion or, if that remedy is foreclosed by § 2255(h), in a § 2241 petition by virtue of the savings clause.” 110 Gilbert v. United States, 640 F.3d 1293, 1306 (11th Cir. 2011) (en banc); see also Brown v. Rios, 696 F.3d 638, 640 (7th Cir. 2012) (“The government has conceded that Brown can use the habeas corpus statute to challenge the legality of his sentence.”); cf. Bryant v. Warden, 738 F.3d 1253, 1274, 1277–79 (11th Cir. 2013) (allowing claim that defendant was wrongly sentenced under ACCA to proceed under § 2241). Even cases finding the savings clause inapplicable—thus preventing recourse to § 2241—note that the savings clause permits litigants to sidestep § 2255(h)’s limitations on second or successive petitions “[w]hen a change of law, retroactively applicable, shows that a prisoner . . . has received an illegally high sentence.” 111 Webster v. Caraway, 761 F.3d 764, 767 (7th Cir. 2014), vacated by order granting reh’g en banc, 769 F.3d 1194 (7th Cir. 2014). This would be the case for defendants convicted under § 922(g) and sentenced under ACCA: Without the ACCA enhancement, the statutory maximum term of imprisonment for a § 922(g) conviction is ten years, but ACCA requires a minimum term of fifteen years’ imprisonment.
The same may not be true for defendants who were subjected to the career offender guideline’s enhancement. Gilbert v. United States held that a prisoner could not challenge his sentence under § 2241 where a misinterpretation of the career offender guideline resulted in a sentence that did not exceed the statutory maximum for the defendant’s offense of conviction. 112 640 F.3d at 1293. Although a decision finding a guideline’s language unconstitutionally vague may be different from a decision reinterpreting a guideline, it is unclear if that difference allows prisoners to make use of the savings clause. 113 Brown, 696 F.3d at 641 (suggesting availability of relief under § 2241 depends on whether defendant argues there has been a guideline or statutory error).
Some of the issues raised by a decision invalidating ACCA’s residual clause require some further analysis. In addition to the three categories of defendants discussed in this Essay, there is a fourth category of prisoners who may seek relief in the event of a favorable decision in Johnson: Defendants who pled guilty to other offenses in the shadow of the ACCA enhancement, believing the enhancement was both valid and applied to their case. 114 See, e.g., Benjamin Levin, Guns And Drugs 31–52 (Feb. 25, 2015) (unpublished manuscript) (on file with the Columbia Law Review) (suggesting ACCA creates incentives to plea); see also Ethan Davis, Comment, The Sentence Imposed Versus the Statutory Maximum: Repairing the Armed Career Criminal Act, 118 Yale L.J. 369, 377 (2008) (same). It is not clear whether these prisoners will be able to obtain relief based on a favorable decision in Johnson. It is not particularly well established when a plea deal is invalid on the ground that the defendant entered into the agreement based on a mistaken belief about the law. 115 E.g., Bousley v. United States, 523 U.S. 614, 618–19 (1998) (suggesting plea would be invalid if defendant showed “that neither he, nor his counsel, nor the court correctly understood the elements of the crime with which he was charged”); United States v. Pierre, 120 F.3d 1153 (11th Cir. 1997) (finding plea invalid because defendant mistakenly believed he preserved speedy trial objection). Given that over ninety-six percent of federal defendants plead guilty, whether these defendants may obtain relief may be the real test of Johnson’s significance. 116 Missouri v. Frye, 132 S. Ct. 1399, 1407 (2012).
The order directing the parties in Johnson to address whether ACCA is unconstitutionally vague has positioned the Court to dramatically affect federal sentencing. Federal sentencing reports estimate the average sentence for offenders who qualified as armed career criminals but were not subject to the ACCA’s enhancement was 122 months’ imprisonment, whereas offenders who received the enhancement were sentenced to an average 210 months’ imprisonment—a difference of over seven years. 117 Mandatory Minimum Penalties, supra note 6, at 292. And as the introduction mentioned, ACCA is a flashpoint for many of the modern issues in criminal law. 118 Id.; see also Levin, supra note 114, at 31–52 (collecting research describing ACCA’s effects).
In addition to the prisoners who may or would have been sentenced under ACCA, a favorable decision in Johnson should also significantly affect those who have already been sentenced under ACCA. A decision finding ACCA’s residual clause unconstitutional, will raise a multitude of questions about various statutes and doctrines that govern resentencing. But these statutes and doctrines should not bar prisoners who have been sentenced under ACCA from obtaining relief based on a favorable decision in Johnson. Prisoners on direct review, first collateral review, and even successive collateral review all possess colorable arguments for resentencing should the Court find the residual clause of ACCA void for vagueness.
Preferred Citation: Leah M. Litman, Residual Impact: Resentencing Implications of Johnson’s Potential Ruling on ACCA’s Constitutionality, 115 Colum. L. Rev. Sidebar 55 (2015), https://live-columbia-law-review.pantheonsite.io/Residual_ Litman