The Divisibility of Crime

Volume 64 February 2015
The Divisibility of Crime

Jessica A. Roth†. Associate Professor of Law, Benjamin N. Cardozo School of Law.

PDF
 

Near the end of the Supreme Court’s 2012-2013 term, the Court decided Descamps v. United States, which concerned the application of the federal Armed Career Criminal Act (ACCA). The ACCA is a recidivist statute that vastly increases the penalties for persons convicted of federal firearms offenses if they have previously been convicted of certain qualifying felonies. Descamps represents the Court’s most recent word on the so-called categorical approach, which directs courts to consider the elements of a prior offense of conviction, rather than the underlying facts of the crime, in determining whether the prior conviction “counts” for purposes of applying the ACCA and other sentencing enhancements and for determining the immigration consequences of prior convictions. This Essay is the first scholarly work to track the immediate effects of Descamps and to explore its implications for the criminal law more broadly. It shows that the decision is indeed having a significant effect on criminal sentencing, resulting in a steady flow of sentencing reversals and prospectively narrowing the class of defendants eligible for sentencing enhancements based on prior convictions. But more broadly, Descamps has called attention to the statutory specificity that legislators are capable of and the adjudicative clarity that courts can promote, if there are incentives for doing so. Until now, the Court has done little to encourage either. Thus, the opinion may push courts and legislators to think more carefully and systematically about what facts must be established to constitute a particular criminal offense, how such facts are established and recorded in the context of an adjudicative proceeding, and the consequences that flow from greater or lesser specificity. Ultimately, this impact may be felt not only in the context of applying recidivist statutes and sentencing enhancements, but also in other contexts that require attention to the basis for a criminal conviction, including the doctrine governing what constituent facts of a crime require jury unanimity and claims under the Double Jeopardy Clause.

Introduction

In its 2012–2013 term, the Supreme Court’s decisions on affirmative action, voting rights, and gay marriage received the lion’s share of attention from the press and scholars. Less noticed were the Court’s decisions on issues of criminal law and procedure. Alleyne v. United States, [1][1]. Alleyne v. United States, 133 S. Ct. 2151 (2013). which held that Apprendi v. New Jersey [2][2]. Apprendi v. New Jersey, 530 U.S. 466 (2000). requires juries to find facts necessary to support the imposition of a mandatory minimum sentence, was received with little fanfare even though it directly overruled precedent (Harris v. United States). [3][3]. Harris v. United States, 536 U.S. 545 (2002). Most scholars and commentators treated Alleyne as the proverbial other shoe that everyone had been expecting to drop. Such an attitude was warranted with Alleyne, but another decision, Descamps v. United States, [4][4]. Descamps v. United States, 133 S. Ct. 2276 (2013). which involved the sentencing of recidivist offenders under the federal Armed Career Criminal Act (ACCA), has received far less attention than it is due. [5][5]. While I am not the first to note Descamps’ significance, this Essay is the first scholarly work to explore it fully. See Douglas Berman, SCOTUS Wraps Its Sentencing Docket With Another Defense Win (and Alito Dissent) In Descamps, Sentencing Law and Policy (June 20, 2013 10:11 AM), http://sentencing.typepad.com/sentencing_law_and_policy/2013/06/scotus-wraps-its-sentencing-docket-with-another-defense-win-and-alito-dissent-in-descamps.html (noting the possibility that “Descamps will prove to be the most consequential of all the Supreme Court’s criminal sentencing work this Term”).

This Essay explains why Descamps matters. In the near term, Descamps will materially affect the severity of the sentences imposed on thousands of criminal defendants per year, narrowing the class of defendants eligible for enhanced sentences pursuant to the ACCA and other sentencing enhancements like the federal Career Offender guideline [6][6]. See U.S. Sentencing Guidelines Manual § 4B1.1 (2014). For other federal sentencing guidelines that provide an enhancement based upon a qualifying prior conviction, see e.g., U.S. Sentencing Guidelines Manual § 2L1.2 (2014) (providing enhancement based on a prior conviction for the crime of unlawfully entering and remaining in the United States) and U.S. Sentencing Guidelines Manual § 2K2.1 (2014) (providing enhancement based on a prior conviction for firearms and ammunition offenses). that are dependent upon a finding of qualifying prior convictions. [7][7]. For the past several years, approximately 600 criminal defendants per year have been sentenced as Armed Career Criminals and approximately 2200 criminal defendants per year have been sentenced as Career Offenders. See U.S. Sentencing Comm’n, 2013 Sourcebook of Federal Sentencing Statistics tbl. 22 (2013) (noting that there were 582 defendants sentenced as armed career criminals and 2268 sentenced as career offenders in 2013); U.S. Sentencing Comm’n, 2012 Sourcebook of Federal Sentencing Statistics tbl. 22 (2012) (noting that there were 631 armed career criminal sentences and 2232 career offender sentences in 2012); U.S. Sentencing Comm’n, 2011 Sourcebook of Federal Sentencing Statistics tbl. 22 (2011) (noting that there were 571 armed career criminal sentences and 2257 career offender sentences in 2011); U.S. Sentencing Comm’n, 2010 Sourcebook of Federal Sentencing Statistics tbl. 22 (2010) (noting that there were 616 armed career criminal sentences and 2314 career offender sentences in 2010); U.S. Sentencing Comm’n, 2009 Sourcebook of Federal Sentencing Statistics tbl. 22 (2009) (noting that there were 697 armed career criminal sentences and 2392 career offender sentences in 2009); U.S. Sentencing Comm’n, 2008 Sourcebook of Federal Sentencing Statistics tbl. 22 (2008) (noting that there were 653 armed career criminal sentences and 2321 career offender sentences in 2008). The figures for 2014 are not yet available. Already, the decision has prompted the reversal of a number of pre-Descamps federal ACCA sentences. [8][8]. See, e.g., Olten v. United States, 134 S. Ct. 639, 639 (2013); Smith v. United States, 134 S. Ct. 258, 258 (2013); United States v. Prater, 766 F.3d 501, 523 (6th Cir. 2014); United States v. Brown, 765 F.3d 185, 196 (3d Cir. 2014); United States v. Bankhead, 746 F.3d 323, 327 (8th Cir. 2014); United States v. Howard, 742 F.3d 1334, 1349 (11th Cir. 2014); United States v. Tucker, 740 F.3d 1177, 1184 (8th Cir. 2014); United States v. Nunez-Segura, 566 F. App’x 389, 394 (5th Cir. 2014); United States v. Royal, 731 F.3d 333, 342 (4th Cir. 2013); United States v. Hemingway, 734 F.3d 323, 338 (4th Cir. 2013); United States v. Parks, 2014 U.S. Dist. LEXIS 1617, at *13–14 (W.D. Wash. Jan. 7, 2014). It also has resulted in the reversal of numerous sentences imposed pursuant to the federal Career Offender guideline, [9][9]. See, e.g., Olsson v. United States, 134 S. Ct. 530, 530 (2013); Marrero v. United States, 133 S. Ct. 2732, 2733 (2013); United States v. Brown, 765 F.3d 185, 196–97 (3d Cir. 2014); United States v. Covington, 738 F.3d 759, 767 (6th Cir. 2014); United States v. Murguia-Ochoa, 536 F. App’x 696, 699 (9th Cir. 2013). which courts have long interpreted in a manner consistent with the ACCA; [10][10]. See, e.g., United States v. Coronado, 603 F.3d 706, 708 (9th Cir. 2010); United States v. Polk, 577 F.3d 515, 518–19 (3d Cir. 2009); United States v. Seay, 553 F.3d 732, 737–39 (4th Cir. 2009); United States v. Mohr, 554 F.3d 604, 607–09 (5th Cir. 2009); United States v. Herrick, 545 F.3d 53, 56–58 (1st Cir. 2008); United States v. Gray, 535 F.3d 128, 130–31 (2d Cir. 2008); United States v. Bartee, 529 F.3d 357, 359 (6th Cir. 2008); United States v. Templeton, 543 F.3d 378, 380 (7th Cir. 2008); United States v. Williams, 537 F.3d 969, 971 (8th Cir. 2008); United States v. Tiger, 538 F.3d 1297, 1298 (10th Cir. 2008); United States v. Archer, 531 F.3d 1347, 1350–52 (11th Cir. 2008). As noted in note 7, supra, approximately 2200 federal defendants are sentenced as career offenders each year, nearly four times as many as are sentenced as armed career criminals. other provisions of the federal Sentencing Guidelines that turn on a finding of a qualifying prior felony; [11][11]. See, e.g., Ruiz-Sanchez v. United States, 134 S. Ct. 60, 60 (2013) (noting illegal reentry guidelines); United States v. Davis, 751 F.3d 769, 778 (6th Cir. 2014) (noting child pornography guidelines); United States v. Falcon, 553 F. App’x 737, 738 (9th Cir. 2014) (noting firearms and ammunition guidelines); United States v. Estrella, 758 F.3d 1239, 1254 (11th Cir. 2014) (noting illegal reentry guidelines); United States v. Martinez, 756 F.3d 1092, 1098 (8th Cir. 2014) (same); United States v. Edwards, 734 F.3d 850, 851 (9th Cir. 2013) (noting firearms and ammunition guidelines). and state sentences imposed pursuant to state recidivist laws that closely track the ACCA. [12][12]. See, e.g., People v. Wilson, 162 Cal. Rptr. 3d 43, 60 (Ct. App. 2013); State v. Dickey, 329 P.3d 1230, 1246 (Kan. Ct. App. 2014); State v. Hulsey, 2013 Kan. App. Unpub. LEXIS 861, at *29 (Kan. Ct. App. 2013). Approximately half of the states have recidivist laws—sometimes called “three strikes” laws—that are similar to the ACCA. See John Clark, James Austin & D. Alan Henry, “Three Strikes and You’re Out”: A Review of State Legislation, National Institute of Justice Research in Brief, National Institute of Justice 7–9 (Sept. 1997); Robert Heglin, Note, A Flurry of Recidivist Legislation Means: “Three Strikes and You’re Out,” 20 J. Legis. 213, 215–16 & nn.18–23 (1994) (collecting state statutes). It also has resulted in the cancellation of deportation orders premised on an immigrant’s conviction of a felony. [13][13]. See e.g., Rendon v. Holder, 764 F.3d 1077, 1090 (9th Cir. 2014); Aguilar-Turcios v. Holder, 740 F.3d 1294, 1302 (9th Cir. 2014); Lima-Lima v. Holder, 545 F. App’x 648, 650 (9th Cir. 2013); Makwana v. Att’y Gen. of the United States, 543 F. App’x 186, 190 (3d Cir. 2013); Donawa v. United States Att’y Gen., 735 F.3d 1275, 1283–84 (11th Cir. 2013).

In the long term, the significance of Descamps could be even greater. It called attention to the statutory specificity that legislators are capable of and the adjudicative clarity that courts can promote. Until now, the Court had done little to encourage either; Descamps provides incentives for both. Already, Descamps is pushing courts to think more carefully about what facts must be established to constitute a particular criminal offense, how such facts are established and recorded in the context of an adjudicative proceeding, and what consequences flow from greater or lesser specificity. [14][14]. See, e.g., Rendon, 764 F.3d at 1084–88 (considering the required elements of a California burglary conviction and the extent of jury unanimity required as to how the offense was committed); United States v. Abbott, 748 F.3d 154, 158–59 (3d Cir. 2014) (same, as to Pennsylvania drug statute); United States v. Royal, 731 F.3d 333, 340–42 (4th Cir. 2014) (same, as to Maryland assault statute); United States v. Cabrera-Gutierrez, 756 F.3d 1125, 1133–37 (9th Cir. 2013) (same, as to Oregon sexual abuse statute). As the contexts that require courts to determine whether a defendant’s prior offense of conviction “counts” multiply, prosecutors and legislators will have increasing reason to respond. And, in the course of establishing a standard for ACCA purposes, the Court deployed a new analytic term—“alternative elements”—that may assist trial courts in another important context that has long needed attention: sorting those facts that jurors must agree upon from those they need not. The Court’s new analytic term also may prompt a critical reevaluation of the doctrine governing claims under the Double Jeopardy Clause. Thus, Descamps is not only important in its own right, but also suggests the possibility of more interesting things to come for the criminal law more broadly.

I. The Descamps Decision

Descamps addressed the question of how federal courts should go about applying the Armed Career Criminal Act (“ACCA”) of 1984, codified at Title 18, United States Code, Section 924(e). Pursuant to the ACCA, the sentence of a defendant who has been convicted of certain federal firearms offenses may be significantly increased if the defendant has three prior convictions for a “violent felony” or a “serious drug offense.” [15][15]. 18 U.S.C. § 924(e)(1) (2012). At issue in Descamps was only the “violent felony” prong of the ACCA. The ACCA defines a violent felony as any felony, whether under state or federal law, that “has as an element the use, attempted use, or threatened use of physical force against the person of another,” [16][16]. 18 U.S.C. § 924(e)(2)(B)(i) (2012). that is “burglary, arson or extortion, involves use of explosives,” [17][17]. 18 U.S.C. § 924(e)(2)(B)(ii) (2012). or that “otherwise involves conduct that presents a serious potential risk of physical injury to another.” [18][18]. Id. Courts frequently refer to these three clauses as the ACCA’s “force” prong, the “enumerated felonies” prong, and the “residual clause,” respectively. See, e.g., United States v. Elliott, 757 F.3d 492, 493 (6th Cir. 2014). The Supreme Court recently ordered new briefing and argument on the question of whether the residual clause is unconstitutionally vague. See Johnson v. United States, No. 13-7120, 2015 WL 132524 (2015) (ordering briefing to be completed by Apr. 10, 2015). If a defendant has three qualifying prior convictions, he or she is considered an Armed Career Criminal for purposes of the ACCA. As a consequence, the defendant will face a mandatory minimum sentence of 15 years and a maximum of life imprisonment. This is generally a much higher sentencing range than would otherwise be applicable. For example, in Descamps’ case, he was convicted in federal court of being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g), an offense which ordinarily carries a statutory maximum penalty of 10 years in prison. [19][19]. 18 U.S.C. § 924(a)(2). Because Descamps was deemed an Armed Career Criminal, however, he was eligible for the ACCA enhancement and was sentenced to 262 months in prison—more than twice the maximum penalty he otherwise could have received. [20][20]. A defendant who is found to be an Armed Career Criminal may also face an enhanced sentencing range under the United States Sentencing Guidelines as a consequence of that designation. See U.S. Sentencing Guidelines Manual § 4B1.4 (2014) (providing a minimum offense level of 33 and a minimum criminal history category of IV for an Armed Career Criminal, and even higher levels for both if certain facts are established). Not surprisingly given these enormous consequences, defendants like Descamps frequently argue that their prior convictions do not count as prior violent felonies for ACCA purposes. [21][21]. Prior to Descamps, the Supreme Court had decided at least nine cases in the previous ten years analyzing what prior convictions count as ACCA predicate felonies. See, e.g., McNeill v. United States, 131 S. Ct. 2218, 2224 (2011) (analyzing when prior state drug convictions count as serious drug offenses for purposes of the ACCA); Sykes v. United States, 131 S. Ct. 2267, 2277 (2011) (analyzing whether a state felony vehicle flight crime qualified as a violent felony under the ACCA’s residual clause); Johnson v. United States, 559 U.S. 133, 145 (2010) (analyzing whether a prior state conviction for battery constituted a violent felony under residual clause); Chambers v. United States, 555 U.S. 122, 130 (2009) (analyzing whether a state conviction for failure to report for periodic incarceration constituted a violent felony under the residual clause); Begay v. United States, 553 U.S. 137, 145 (2008) (analyzing whether a state conviction for felony driving under the influence qualified under the residual clause); United States v. Rodriguez, 553 U.S. 377, 393 (2008) (analyzing whether a state drug conviction counted as a serious drug offense); James v. United States, 550 U.S. 192, 214 (2007) (analyzing whether a state attempted burglary statute qualified under the residual clause); Logan v. United States, 552 U.S. 23, 37 (2007) (considering the implications of a state’s failure to revoke the defendant’s civil rights on whether a conviction counted under the ACCA); Shepard v. United States, 544 U.S. 13, 27 (2005) (setting forth materials that may be consulted when a court applies modified categorical review).

Although the ACCA enumerates certain specific felonies that necessarily count as violent felonies (i.e., burglary, arson, and extortion), it does not define those crimes, and there is a great deal of variation in how the several states define them. [22][22]. The statute originally included a definition of burglary, but Congress deleted this definition in the course of amending the statute in 1986 to expand the scope of the predicate felonies that count for ACCA purposes. Although the deletion appears to have been unintentional, it has never been corrected. See Taylor v. United States, 495 U.S. 575, 588–90 (1990). Thus, in prior decisions interpreting the ACCA, the Supreme Court had construed the statute as incorporating by reference the modern “generic” version of such offenses, “roughly corresponding” [23][23]. See id. at 589. to the definitions used in the majority of states’ criminal codes, and held that only prior convictions falling within the generic version of the offense would count. To determine whether a prior conviction fell within the generic version of the offense, the Court instructed sentencing courts to engage in what the Court called a “categorical approach,” meaning courts should analyze the statutory definition of the offense of conviction, not the underlying facts of the particular case. [24][24]. Descamps, 133 S. Ct. at 2283. If that offense swept no more broadly (or was narrower) than the generic version of the offense, then a conviction for the prior offense categorically would count as an ACCA predicate felony. However, the Court also had recognized that, in certain situations involving a statute that swept more broadly than the generic version of the offense, it might be permissible for a sentencing court to engage in what the Court called “modified categorical review,” meaning it could consult certain materials to determine whether the prior conviction was for conduct falling within the generic version of the offense. At issue in Descamps was precisely when such modified categorical review was appropriate.

Descamps resolved a Circuit split on this issue and came down on the side of those Courts of Appeals that had tied the propriety of modified categorical review to the text of the statute that was the basis for the prior conviction. [25][25]. The Ninth and Sixth Circuits had applied the modified categorical approach to indivisible statutes, as explained infra, whereas the Second and First Circuits had held that the modified categorical approach applied only to divisible statutes. Compare United States v. Descamps, 466 F. App’x 563, 565 (9th Cir. 2012), and United States v. Armstead, 467 F.3d 943, 947–50 (6th Cir. 2006), with United States v. Beardsley, 691 F.3d 252, 268–74 (2d Cir. 2012), and United States v. Giggey, 551 F.3d 27, 40 (1st Cir. 2008) (en banc). The Court held that modified categorical review was permissible only when the prior conviction was pursuant to a “divisible” rather than an “indivisible” statute [26][26]. Descamps, 133 S. Ct. at 2284.—meaning that the statute explicitly set forth “alternative elements,” [27][27]. Id. at 2283–84. some of which fell within the generic offense and others of which did not. An example of a divisible statute would be a statute specifying that a person committed an offense if he unlawfully possessed a “gun, knife, or ax.” An indivisible version of the same offense would be a statute that used the broader term “weapon.”

If the statute were divisible, the sentencing court could consult a limited category of documents such as the indictment, jury instructions, verdict form, plea agreement, and plea colloquy to determine which of the statutory alternative elements provided the basis for the conviction. If the conviction were based on one of the alternatives falling within the generic version of the offense, then the prior conviction could count as a predicate felony for purposes of the ACCA. If the conviction were based on one of the statutory alternatives falling outside the scope of the generic version of the offense (or if the documents did not reveal which statutory alternative provided the basis for conviction), then the prior conviction could not count. However, the Court held that this modified categorical review was never appropriate where the prior offense of conviction was pursuant to an indivisible statute. Faced with a prior offense pursuant to an indivisible statute encompassing conduct falling both within and outside of the generic offense, courts could not consult any underlying materials in the case to determine whether the defendant’s actual conduct that gave rise to the conviction fell within the generic version of the offense. Because the lower courts in Descamps’ case (both the district court and the Ninth Circuit Court of Appeals) had applied the modified categorical approach to a statute that the Supreme Court deemed indivisible, [28][28]. The prior conviction at issue in Descamps’ case was a conviction for burglary pursuant to Cal. Penal Code Ann. § 459 (West 2010), which provided that a “person who enters” certain locations “with intent to commit grand or petit larceny or any felony is guilty of burglary.” The statute did not require that the entry have been unlawful—an element that the Supreme Court deemed a necessary component of the “generic version” of burglary. See Descamps, 133 S. Ct. at 2285, 2294 (Thomas, J., concurring) (citing 3 Wayne LaFave, Substantive Criminal Law § 21.1(a) (2d ed. 2003)). the Supreme Court reversed.

As a consequence of Descamps, the text of the criminal statute that provides the basis for a defendant’s prior conviction assumes paramount importance. Whenever a defendant’s prior conviction is based on a textually indivisible statute that is broader than the generic version of an ACCA felony, that conviction will not count for ACCA purposes, with no exceptions. Accordingly, some repeat, violent offenders will not be eligible for the enhanced penalties provided by the ACCA or other recidivist statutes, even when the underlying record makes clear that the actual offense was of a type that ordinarily would qualify as a predicate felony. Critics of the categorical approach have long complained about this anomalous result, arguing that this kind of formalism allows too many dangerous recidivists to escape the penalties Congress intended. [29][29]. See, e.g., Shepard v. United States, 544 U.S. 13, 35–36 (O’Connor, J., dissenting) (decrying the categorical approach’s “overscrupulous regard for formality,” which “forces the . . . sentencing court to feign agnosticism about clearly knowable facts” and “cannot be squared with the ACCA’s twin goals of incapacitating repeat violent offenders, and of doing so consistently notwithstanding the peculiarities of state law”) (emphasis in original). See also Descamps, 133 S. Ct. at 2302 (Alito, J., dissenting) (“The Court’s holding will . . . frustrate fundamental ACCA objectives” including that “violent, dangerous recidivists would be subject to enhanced penalties” in that “those enhanced penalties would be applied uniformly.”). Whereas some courts, like the Ninth Circuit, had liberally applied the modified categorical approach to get around this result, as set forth below in Part II, Descamps has definitively closed off that option whenever a prior conviction was based on an indivisible statute.

II. Descamps’ Refinement of the Categorical Approach

Justice Kagan’s opinion for the Court (which was joined by six Justices, including the Chief Justice) treated the case as “all but resolve[d]” [30][30]. Descamps, 133 S. Ct. at 2283. by the Court’s prior case law. [31][31]. As Daniel Richman has observed, “[e]ven for those not ordinarily interested in technical sentencing decisions, the Court’s opinion is worth reading for its miffed tone of a well-respected author confronted with an obdurate reader claiming unfamiliarity with her work.” Daniel Richman, Opinion analysis: When is a burglary not a burglary?, SCOTUSblog (June 20, 2013, 11:18 PM), http://www.scotusblog.com/2013/06/opinion-analysis-when-is-a-burglary-not-a-burglary; see also William Baude, Modified Categorical Imperative, PrawfsBlawg (June 21, 2013, 2:09 AM), http://prawfsblawg.blogs.com/prawfsblawg/2013/06/modified-categorical-imperative.html (“If you are interested in sharp-tongued judicial rhetoric, you should be following Justice Kagan” who “positively mocks the Ninth Circuit decision.”). To a certain extent, she was clearly right. In Taylor v. United States, [32][32]. Taylor v. United States, 495 U.S. 575 (1990). decided in 1990, the Court construed the ACCA to hold that the specific offenses referenced in the ACCA’s definition of “violent felony” referred to the “generic” versions of these offenses rather than any offense so titled in a state criminal code. Taylor also established the categorical approach for determining whether a state conviction qualified as a violent felony under the ACCA. The Court held that a sentencing court should look only at the elements of the prior offense of conviction, not the underlying facts of the offense, to determine if it was a violent felony. If the elements of the prior offense swept more broadly than the generic version of the offense, the prior conviction would not count. If the elements of the prior offense were essentially the same as those of the generic offense, or were narrower than the generic version, then the prior conviction would count. [33][33]. Id. at 601–02.

Taylor also nodded to the possibility of what later would be termed the “modified categorical approach.” At the end of the opinion, the Court observed that, in a “narrow range of cases” [34][34]. Id. at 602. a sentencing court might consult materials beyond the statutory definition of a prior offense to determine whether it qualified as a violent felony under the ACCA. Citing the hypothetical example of a state burglary statute that included entry of an automobile as well as a building, the Court opined that the Government should be allowed to use the conviction to enhance a defendant’s sentence under the ACCA where the jury was “actually required to find all of the elements of generic burglary” [35][35]. Id. (i.e., burglary of a building, not an automobile). Thus, if the indictment or information and jury instructions indicated that the defendant was charged solely with burglary of a building, the prior conviction could count as a burglary for ACCA purposes.

In Shepard v. United States, [36][36]. Shepard v. United States, 544 U.S. 13 (2005). decided in 2005, the Court encountered precisely the type of statute alluded to in Taylor—a state burglary statute that covered entries into “boats and cars” as well as “buildings.” Drawing upon the foregoing language from the end of the Taylor decision, Shepard held that, when presented with a prior conviction pursuant to this statute, the sentencing court could consult a limited range of documents to determine whether the defendant had pleaded guilty to the “buildings” prong of the burglary statute—and thus had been convicted of generic burglary (the Court held in Taylor that generic burglary encompassed only burglary of a building)—rather than the other portions of the statute. The permissible class of documents that the sentencing judge could consider included not only charging documents and jury instructions, but also, in the case of a bench trial or a guilty plea, the judge’s formal rulings of law and findings of fact, a plea agreement, or guilty plea colloquy. [37][37]. Id. at 20. Shepard refused to expand the scope of evidentiary material further, however, specifically rejecting the use of a police report even if it had been submitted to the local court in connection with the filing of the original charges. [38][38]. Id. at 21. Thus, Shepard solidified the modified categorical approach, although (as in Taylor) that phrase appears nowhere in the Court’s opinion.

Similarly, in Johnson v. United States, [39][39]. Johnson v. United States, 559 U.S. 133 (2010). decided in 2010, the Court considered whether Florida’s battery statute categorically fell within the ACCA’s definition of a “violent felony,” defined as any crime having “as an element the use . . . of physical force against the person of another.” [40][40]. 18 U.S.C. § 924(e)(1) (2006). The Florida statute did not qualify because the Court interpreted the term “physical force” in the ACCA as requiring the application of violence and the Florida statute could be satisfied without a showing of violence. Nevertheless, reaffirming Shepard, the Court held that a sentencing court confronted with a law containing “statutory phrases that cover several different generic crimes[,] some of which require violent force, and some of which do not,” could apply the modified categorical approach “to determine which statutory phrase was the basis for the conviction.” [41][41]. Johnson, 559 U.S. at 144. In Johnson, the Court finally used the phrase “modified categorical approach,” quoting its decision in Nijhawan v. Holder, 557 U.S. 29 (2009), an immigration case in which the Court considered the application of the Taylor line of cases to the construction of certain felonies under the Immigration and Nationality Act (INA) that could provide the basis for deportation. Id. In Nijhawan, the Court ultimately rejected the use of the categorical approach for purposes of this section of the INA, holding that it was not consistent with Congressional intent. Nijhawan, 557 U.S. at 39–40. However, the Court has held that the categorical approach does apply with respect to certain other portions of the INA. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 194 (2007) (endorsing the categorical approach to interpret the term “theft offense” in the INA). Because the Florida statute was such a statute, the modified categorical approach was appropriate. [42][42]. Johnson represented the usual case in which the Government advocated for categorical review rather than modified categorical review, because the underlying state documents were inconclusive. The Government argued that the Florida statute categorically qualified as a violent felony under the ACCA, thus obviating the need for resort to the underlying documents. Johnson, 558 U.S. at 139. The Court disagreed. Id.

Thus, well before the Ninth Circuit upheld Descamps’ sentence in 2012, [43][43]. United States v. Descamps, 466 F. App’x 563 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013). it was clear—pursuant to the foregoing precedents—that a sentencing court generally should apply the categorical approach, focusing on the elements of the prior offense, to determine whether a defendant’s prior conviction qualified as a predicate violent felony under the ACCA. It was also clear that, in certain limited circumstances involving certain types of statutes, a court could apply the “modified categorical approach” and consult a limited range of documents to determine if the defendant’s prior conviction necessarily was pursuant to a particular portion of the statute that fell within the generic version of the offense. What was not completely clear—notwithstanding Justice Kagan’s assertion that the case was “all but resolved” by prior case law—was how a court should go about deciding whether a particular statute was of the type that permitted further inquiry.

Although the Court in prior decisions had suggested, mostly by way of the examples that it used, that statutory text was critical (and in Johnson, as noted supra, the Court described the sentencing judge’s objective in conducting modified categorical review as determining “which statutory phrase was the basis for conviction”), [44][44]. Johnson, 559 U.S. at 144 (emphasis added). in Descamps the Court for the first time made clear that text was decisive. For the first time, the Court used the terms “divisible statutes” and “indivisible statutes” to describe the two universes of criminal statutes that existed, as though there were no other possibilities and no overlap between the two. The Court also used the phrase “alternative elements” for the first time, to describe the distinguishing characteristic of a “divisible” statute. That is, a “divisible” statute has at least one element that is capable of being satisfied in more than one way and explicitly sets forth in its text what those alternatives are—such as a burglary statute that specifies that breaking and entering of a “building or automobile” will suffice, or a weapon possession statute that specifically applies to possession of a “gun, knife, or ax.”

The Ninth Circuit’s mistake, the Court wrote, was in trivializing the distinction between divisible and indivisible statutes on the theory that any indivisible statute could “be imaginatively reconstructed as a divisible one” [45][45]. Descamps v. United States, 133 S. Ct. 2276, 2289 (2013). and therefore that there was no meaningful difference between the two. For the Ninth Circuit, the only conceptual difference was that “[a divisible statute] creates an explicitly finite list of possible means of commission, while [an indivisible one] creates an implied list of every means of commission that otherwise fits the definition of a given crime.” [46][46]. Id. (quoting United States v. Aguila-Montes de Oca, 655 F.3d 915, 927 (9th Cir. 2011)) (alteration in original). The Ninth Circuit viewed this difference as an insufficient reason to allow modified categorical review in the case of a divisible statute but disallow it for an indivisible one. If the point was to identify the underlying basis for the defendant’s prior conviction, and thereby to determine if he qualified as an Armed Career Criminal, why not allow the sentencing court to engage in such review in both categories of cases? The Supreme Court’s answer: allowing judges to do so would raise serious constitutional concerns under the Sixth Amendment and Apprendi. [47][47]. Id. at 2288 (citing Apprendi v. New Jersey, 530 U.S. 466 (2000)).

Citing the “categorical approach’s Sixth Amendment underpinnings,” [48][48]. Id. the Court explained that a sentencing judge could enhance a defendant’s sentence based on the existence of a prior conviction without running afoul of the Sixth Amendment’s jury trial guarantee. [49][49]. Id. (citing Apprendi, 530 U.S. at 490). In Almendarez-Torres v. United States, 523 U.S. 224 (1998), the Supreme Court held that the existence of a prior conviction was a sentencing factor rather than an element and, accordingly, did not need to be included in an indictment or found by a jury. Post-Apprendi, the Court consistently has recognized Almendarez-Torres as an exception to the Apprendi line of cases, notwithstanding repeated calls from various quarters (including from within the Court, most notably from Justice Thomas) to overrule it. Descamps, 133 S. Ct. at 2294–95 (Thomas, J., concurring); James v. United States, 550 U.S. 192, 231 (2007) (Thomas, J., dissenting); Shepard v. United States, 544 U.S. 13, 27 (2005) (Thomas, J., concurring in part and concurring in the judgment). Descamps argued in the lower courts that Almendarez-Torres should be overruled and sought certiorari on that issue; the Court denied that portion of his petition. United States v. Descamps, 466 F. App’x 563, 564–65 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013). But allowing a sentencing judge to make a finding as to the nature of the prior conviction under an indivisible statute was fraught. Allowing modified categorical review in cases involving divisible statutes was consistent with Apprendi because “only divisible statutes enable a sentencing court to conclude that a jury (or judge at a plea hearing) has convicted the defendant of every element of the generic crime.” [50][50]. Descamps, 133 S. Ct. at 2290. Divisible statutes, the Court reasoned, require the prosecution to proceed, and the fact-finder to enter a decision, in a manner that together renders transparent whether or not the defendant was convicted pursuant to a statutory alternative element falling within the generic version of the offense. Indivisible statutes, by contrast, do not require such focused allegations or deliberations.

III. Implications of Descamps

Descamps makes a persuasive case for distinguishing between divisible and indivisible statutes for Sixth Amendment purposes—but only if it accurately describes how divisible statutes are charged and submitted to juries. Yet as set forth below, that is not necessarily the case. What remains to be seen is whether the decision will prompt changes in how such cases are in fact charged and litigated, and whether the Court’s use of the term “alternative elements” will have broader doctrinal implications.

A. Practical Implications

1. Single Theory Charges.

To begin with the simplest category of cases in which Descamps’ model does not accurately describe actual practices, consider again the hypothetical “gun, knife, or ax” statute. An indictment alleging a single charge of violating this divisible statute might very well track the statute and include all three of the alternative elements in the charging language, specifying only in the “to wit” clause at the end of the charging paragraph that the defendant possessed a gun, in order to give the defendant notice of the specific factual allegation underlying the charge. [51][51]. See Wayne LaFave, Jerold H. Israel, Nancy King & Orin Kerr, Criminal Procedure § 19.3(b) (5th ed. 2009 & Supp. 2013). Depending on the practice in that jurisdiction, if no evidence was presented at the trial that the defendant possessed any weapon other than a gun, the court might instruct the jury at the end of the case that, in order to convict, it must find that the defendant possessed a gun—thus following Justice Kagan’s model. But, alternatively, the judge might instruct the jury that it may convict if it finds that the defendant possessed “a gun, knife, or ax.” Either way, the verdict form likely will call for a single response—“guilty” or “not guilty” as to the single count in the indictment.

In the second scenario, is it clear that the jury necessarily found that the defendant possessed a gun? Not really—no more so than the answer “yes” to the question “would you like coffee or tea?” indicates which beverage the respondent prefers. Thus, notwithstanding the fact that the charge was pursuant to a divisible statute (and that everyone involved in the trial likely would be confident that the jury did find that the defendant possessed a gun), a subsequent sentencing judge would not be able to determine with confidence through modified categorical review that the jury necessarily convicted the defendant of a generic firearms offense. The “Descamps fix” for such a scenario is straightforward and essentially cost-free to the prosecution: streamline the indictment and seek a jury instruction mentioning only the gun.

2. Multiple Theory Charges.

The next category of case is not so simple. Consider a hypothetical case where there is evidence that the defendant possessed both a gun and a knife. Perhaps at the time of his arrest, he had a gun in his pocket and a knife in his backpack. The government alleges a single count of violating the “gun, knife, or ax” statute, and specifies in the “to wit” clause of the indictment that the defendant possessed both a gun and a knife. [52][52]. See Fed. R. Crim. P. 7(c)(1) (providing that a single count of an indictment may allege that “the means by which the defendant committed the offense are unknown or that the defendant committed it by one or more specified means”); Griffin v. United States, 502 U.S. 46, 51 (1991) (“A statute often makes punishable the doing of one thing or another . . . sometimes thus specifying a considerable number of things . . . . [T]he indictment on such a statute may allege, in a single count, that the defendant did as many of the forbidden things as the pleader chooses, employing the conjunction and where the statute has ‘or,’ . . . and it will be established at the trial by proof of any one of them.” (quoting 1 Joel Prentiss Bishop, New Criminal Procedure § 436, at 355–56 (2d ed. 1913))) (internal footnotes omitted) (internal quotation marks omitted) (emphasis omitted). Suppose some evidence is presented at trial that the defendant did not know that the knife was in the backpack (e.g., a friend asked him to carry the backpack) and some evidence suggesting that the gun in the defendant’s pocket does not meet the statutory definition of a gun (e.g., perhaps it was an antique). In other words, assuming that the statute requires knowing possession, there are possible grounds for reasonable doubt as to each potential basis for conviction. Some jurors might conclude that the defendant violated the statute by possessing a gun but not a knife, while others might reach the opposite conclusion.

If the jury were instructed that it might convict if it finds that the defendant possessed a “gun, knife, or ax,” would the defendant who is convicted of such a charge necessarily have received a jury finding that he possessed a gun? Plainly, the answer is no. Again, the charge would have been pursuant to a divisible statute, but even with modified categorical review, a subsequent sentencing judge would not be able to determine that the jury necessarily found that the defendant possessed a gun. Here, the “Descamps fix” is not so straightforward, nor is it costless to the prosecution. Should each theory of the case be alleged in a separate count of the indictment? Should the two theories remain in a single count but the jury nevertheless be instructed that it must agree unanimously (or by the requisite majority in the jurisdiction) on a single theory before it may convict? If the latter, should the verdict form require the jury to specify the prevailing theories or theories? Justice Alito called attention to some of these practical difficulties with operationalizing Descamps, [53][53]. See Descamps, 133 S. Ct. at 2298 (Alito, J., dissenting) (envisioning the case of a defendant charged with burglarizing a “floating home” pursuant to a state statute defining burglary as including unlawful entry into, inter alia, a “building” or “vessel” and asking whether, in order to convict, the jury would have to agree “whether this structure was a ‘building’ or a ‘vessel.’ . . . The Court’s answer is ‘yes.’”). Justice Alito also notes that, even if unanimity were required, the basis for the conviction would not be apparent to a subsequent sentencing court unless the jury returned “a special verdict, something that is not generally favored in criminal cases.” Id. at 2300 (citing 6 Wayne LaFave & Jerold H. Israel, Criminal Procedure § 24.10(a), at 543–44 (3d ed. 2007)). but the majority glossed over them entirely.

This second category of case is not so unusual. Under prevailing doctrine, alternative theories may be combined in a single charge, and a guilty verdict returned based on a combination of theories, if each alternative theory is properly characterized as a “means” of committing a single offense, rather than an “element” signifying a distinct crime. The Court’s two leading cases on the subject are Schad v. Arizona [54][54]. Schad v. Arizona, 501 U.S. 624 (1991). from 1991 and Richardson v. United States [55][55]. Richardson v. United States, 526 U.S. 813 (1999). from 1999. Neither opinion offers particularly clear guidance to courts as to how they should distinguish between means and elements. [56][56]. For a discussion of the indeterminacy of Schad and Richardson, see, for example, Scott W. Howe, Jury Fact-Finding in Criminal Cases: Constitutional Limits on Factual Disagreements Among Convicting Jurors, 58 Mo. L. Rev. 1, 6 (1993); Brian M. Morris, Something Upon Which We Can All Agree: Requiring a Unanimous Jury Verdict in Criminal Cases, 62 Mont. L. Rev. 1, 19–24, 31–34 (2001); Jessica A. Roth, Alternative Elements, 59 UCLA L. Rev. 170, 186–97 (2011); Stephen E. Sachs, Alternative Theories of the Crime 5–6 (May 24, 2007) (unpublished manuscript) (http://ssrn.com/abstract=1501628); Peter Westen & Eric Ow, Reaching Agreement on When Jurors Must Agree, 10 New Crim. L. Rev. 153, 153–54 (2007); see also United States v. Hurt, 527 F.3d 1347, 1356 (D.C. Cir. 2008) (observing that Schad does not provide clear guidance). What the Court has said is that legislative intent is the controlling question in determining whether a statute sets forth alternative means as opposed to elements, subject only to the barest of review for compliance with the Due Process Clause. But legislative intent in this regard (as in many others) is frequently difficult to discern. [57][57]. The Court has not provided any particularized guidance as to the appropriate methodology in determining whether a legislature intended to create means as opposed to elements. In Schad, however, the Court did cite approvingly to United States v. UCO Oil Co., 546 F.2d 833, 835–38 (9th Cir. 1976), in which the court stated that statutory language, legislative history, the nature of the proscribed conduct, and the appropriateness of multiple punishments all should bear on whether a statute should be interpreted as evidencing a legislative intent to create multiple crimes as opposed to a single crime with multiple means of commission. See Schad, 501 U.S. at 636 (citing UCO Oil Co., 546 F.2d at 835–38). The Court has never set forth a definitive test for determining when a statute has impermissibly combined as putative “means” alternatives that must, consistent with due process, be considered “elements.” [58][58]. The Court did, however, suggest certain factors that would be relevant to such a determination, including the moral equivalency of the various alternatives, whether historically they were treated as discrete crimes, and the practices of other states. For example, in Schad, the plurality opinion stated that it would be unconstitutional for a state to recast its entire criminal code as a single statute and declare each formerly separate crime a “means” of committing the offense of “crime.” Schad, 501 U.S. at 632–33. Nor has it ever held a particular statute unconstitutional on that basis. Consistent with such a permissive approach, state and federal courts have repeatedly cited Schad and Richardson in holding that a criminal statute sets forth means rather than elements. [59][59]. See, e.g., United States v. Seher, 562 F.3d 1344, 1361–62 (11th Cir. 2009) (holding that a jury need not agree on a defendant’s motive for engaging in a financial transaction, as set forth in the federal money laundering statute); United States v. Stewart, 433 F.3d 273, 319 (2d Cir. 2006) (holding that a jury need not agree on whether a defendant made a false statement or concealed a material fact, as set forth in the false statements statute); United States v. Powell, 226 F.3d 1181, 1196 (10th Cir. 2000) (holding that a jury need not agree on which verb characterizes a defendant’s conduct, as set forth in the federal kidnapping statute). See also 6 LaFave et al., supra note 53, at § 24.10(c), nn.40–41 (collecting additional state and federal cases involving multiple theories and in which courts have approved general verdicts without requiring juror agreement as to the theory). That is why, when the evidence supports them, prosecutors routinely file charges pursuant to such statutes alleging multiple theories within one count of an indictment, and jurors in such cases are not required to agree as to the means of commission—i.e., the jury can return a “patchwork” [60][60]. See Hayden J. Trubitt, Patchwork Verdicts, Different-Juror Verdicts, and American Jury Theory: Whether Verdicts are Invalidated by Juror Disagreement on Issues, 36 Okla. L. Rev. 473, 474–77 (1983) (describing how jury selection and individual jury behavior can lead to compromises that actually stymie true deliberation and discussion among jurors). or “composite” verdict. In other words, current constitutional law fosters precisely the type of adjudicative uncertainty that the Descamps Court says divisible statutes can prevent.

Under the reasoning of Descamps, a prior conviction pursuant to such a charge could qualify as a predicate felony for purposes of the ACCA in two ways: first, if each alleged theory of commission met the ACCA definition; or second, if one or more alleged theories did not fall with the ACCA definition (e.g., a gun would qualify as a “firearm” but a knife would not), if the jury was required to return a separate verdict as to each theory and found the defendant guilty pursuant to at least one of the theories meeting the ACCA definition. Thus, prosecutors may be forced to choose between, on the one hand, the short-term strategic advantage of obtaining a “composite” guilty verdict; and, on the other hand, the theoretical benefit—likely to be realized only in the future, if at all—of tagging the defendant with a conviction sure to count under the ACCA.

How prosecutors would weigh the risks and benefits of each choice would vary depending on the individual case. Generally speaking, it seems unlikely that a local prosecutor would give much thought to whether a conviction will count for ACCA purposes. But by introducing this new dynamic, Descamps may result in changes at the institutional level in how criminal charges based on multiple theories are tried. Post-Descamps, one reasonably might expect to see prosecutors’ offices, courts, and even defense counsel [61][61]. Ex ante, adjudicative clarity is a protection for defendants. Ex post, it has positive externalities for those, generally prosecutors, who want to piggyback off the prior adjudicated fact. It is perhaps not surprising then that an Internet search for materials citing Descamps reveals a number of training materials put together by defense attorneys recommending that defenders representing ACCA defendants at sentencing argue that their clients’ prior convictions were pursuant to statutory phrases that represent alternative means rather than alternative elements. See, e.g., Paresh Patel & Lisa Freeland, Post-Descamps World, http://agendas.fjc.gov/agenda/files/FedDef1410-PostDescampsWorld.pdf (last visited Jan. 26, 2015). becoming more comfortable with specific unanimity instructions and embracing more precise verdict forms. [62][62]. Even putting aside the ACCA, there are institutional reasons why prosecutors would be in favor of more specialized verdicts. See Roth, supra note 56, at 213–14 (explaining how prosecutors sometimes benefit from seeking greater clarity in the basis for a jury’s verdict). Congress could amend the ACCA to make clear that “conviction” refers to underlying conduct, not the elements of the prior offense. In his concurring opinion in Descamps, Justice Kennedy called on Congress to make such a change. Descamps v. United States, 133 S. Ct. 2276, 2294 (2013) (Kennedy, J., concurring). But it is not clear that such a change would be consistent with the Sixth Amendment. See id. at 2288 (explaining that if a court’s finding of a predicate offense went beyond merely identifying a prior conviction it would raise Sixth Amendment concerns). Assuming that the Court means what it says when it cites the categorical approach’s Sixth Amendment underpinnings in the context of criminal sentencing, a circumstance-based approach would seem to present serious constitutional problems. Courts have suggested for years that such practices would be desirable but have been reluctant to mandate them. [63][63]. See, e.g., Griffin v. United States, 502 U.S. 46, 60 (1991); id. at 61 (Blackmun, J., concurring); Schad, 501 U.S. at 645. As the number of contexts increases in which what facts necessarily were decided in a prior proceeding is of great consequence—such as for purposes of state recidivist statutes, sentencing guidelines, and deportation proceedings—prosecutors (and legislators) will have increasing reasons to respond to Descamps’ demand for greater clarity.

Formally, Descamps does not revisit the Court’s “means” versus “element” jurisprudence from Schad and Richardson. As set forth above, under Descamps, a prosecutor may have to choose between a composite verdict and a sure ACCA predicate felony, but the decision says nothing about the constitutionality of opting for the former at the expense of the latter. And yet, by implication, Descamps suggests that Schad’s influence may be on its way out—not just as a practical matter, but doctrinally as well. [64][64]. Schad was decided by a 5-4 vote. Justice Souter wrote the main opinion for the Court for a plurality of the Justices (joined by Chief Justice Rehnquist and Associate Justices O’Connor and Kennedy). Schad, 501 U.S. at 624. Justice Scalia provided the crucial fifth vote to affirm, but made clear in a separate opinion that, in any other case, he might well find that jury unanimity was constitutionally required. Id. at 651 (Scalia, J., dissenting). Richardson, which found elements rather than means—citing constitutional concerns—was written by Justice Breyer, joined by Chief Justice Rehnquist and Justices Stevens, Scalia, Souter, and Thomas. Richardson v. United States, 526 U.S. 813, 814 (1998). Justice Kennedy dissented in Richardson, joined by Justices O’Connor and Ginsburg. Id. at 825 (Kennedy, J., dissenting). Consistent with his position in Descamps, then-Judge Alito dissented in a case decided by the Third Circuit involving the same issue as Richardson. See United States v. Edmonds, 80 F.3d 810, 829 (3d Cir. 1996) (Alito, J., dissenting). Put together with the line-up in Descamps, there now appear to be six members of the Court who might vote to require jury unanimity as to at least one statutory alternative element (Justices Kagan, Breyer, Stevens, Scalia, Thomas, and Chief Justice Roberts) and three Justices who almost certainly would take the opposite view (Justices Kennedy, Ginsburg, and Alito). Indeed, the majority opinion acknowledges the means versus element distinction but it cites only to Richardson—where the Court found elements rather than means [65][65]. See supra notes 63–64.—and not to Schad. [66][66]. Descamps, 133 S. Ct. at 2288. In addition to the majority’s choice to ignore Schad, the strongest signal of a doctrinal shift is the Court’s decision to use the phrase “alternative elements” throughout the Descamps opinion. This is a phrase the Court had never before used in a criminal opinion. [67][67]. The phrase had been used in the lower courts opinions discussing the application of the ACCA. See, e.g., United States v. Aguila-Montes de Oca, 655 F.3d 915, 963 (9th Cir. 2011); United States v. Vann, 660 F.3d 771, 789 (4th Cir. 2011). When viewed against the backdrop of Schad and Richardson (both of which were brought to the majority’s attention by Justice Alito), the choice seems significant. In Schad and Richardson, the Court painstakingly referred to statutory provisions as either “elements” or “means,” cognizant (and wary) of the consequences of characterizing a provision as the former rather than the latter. By contrast, in Descamps, the Court seems far less concerned about broadly characterizing statutory provisions as “elements.” Indeed, Descamps seems to be utilizing an implicit default rule that divisible statutes set forth “alternative elements” rather than “means.” [68][68]. See Descamps, 133 S. Ct. at 2285 n.2 (expressing surprise that the state statutes analyzed in Taylor, Shepard, and Johnson could be construed as setting forth anything but “alternative elements,” and therefore “several different . . . crimes”). The Court’s approach in determining what constitutes a divisible statute is quite formalistic. When one bears down and looks at the particular statutes involved in the Court’s ACCA cases, divisibility turns on the presence of commas and the word “or” between phrases.

B. Broader Doctrinal Implications

Going forward, the question is whether the Court will stand by this formalistic approach and its description of the different statutory phrases in divisible statutes as alternative elements in other contexts. The Court might explain in a future case that, when it referred to the statutes in Taylor, Shepard, and Johnson as including “alternative elements,” indicative of “separate crimes,” [69][69]. Id.; see id. at 2297–98 (Alito, J., dissenting) (“[L]egislatures frequently enumerate alternative means of committing a crime without intending to define separate elements or separate crimes.” (citing Schad v. Arizona, 501 U.S. 624, 636 (1991))). it did not really mean that in the sense that juries would have to agree on at least one alternative element, [70][70]. The closest the Court came to dealing with this issue in Descamps was to acknowledge Justice Alito’s point that some divisible statutes might set forth alternative means as opposed to elements. Id. at 2285 n.2. The Court expressed surprise that any of the divisible statutes in its ACCA cases to date established means as opposed to elements, and asserted that any confusion as to what constituted elements as opposed to means in future cases could be clarified by consulting the indictment and jury instructions. Id. or that a defendant could be convicted of two versions of the same offense, each containing a different alternative element. But this would require some explaining. Generally speaking, we do not think of crimes as having different elements for different purposes. We may not have a satisfactory a priori definition of what constitutes an element, [71][71]. See 6 LaFave et al., supra note 53, at § 24.10(c) (observing that the Court has never “derive[d] a uniform theory for identifying which facts are elements”); Nancy J. King & Susan R. Klein, Essential Elements, 54 Vand. L. Rev. 1467, 1497 (2001) (ultimately rejecting as “unwise” and “unattainable” the “theoretical quest for the elemental holy grail”). but we know what consequences flow from calling something an element. Elements must be alleged in the indictment. [72][72]. Hamling v. United States, 418 U.S. 87, 117 (1974). Elements must be proven to the jury (by the government) beyond a reasonable doubt. [73][73]. See Patterson v. New York, 432 U.S. 197, 210 (1977) (holding that the legislature cannot shift the burden to the defendant to disprove any element of the offense); In re Winship, 397 U.S. 358, 364 (1969) (holding that “the Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged”). Under the familiar Blockburger [74][74]. Blockburger v. United States, 284 U.S. 299 (1932). formulation, elements are facts that distinguish one offense from another, such that a defendant generally may be charged with both offenses without violating the Double Jeopardy Clause of the Fifth Amendment. [75][75]. Id. at 304. The question for future cases will be, “is an ‘alternative element’ just like a regular element for these purposes, or is it something else?”

1. Jury Unanimity.

In the context of jury unanimity, there is a good argument that alternative elements should be treated just like regular elements. After all, as the Court makes clear in Descamps, text matters. There is a difference between an indivisible statute that uses the term “weapon” and a divisible one that instead uses the list “gun, knife, or ax.” They reflect different legislative choices. The legislature that enacts the former statute enjoys the potential benefit of extending the statute’s reach to new types of weapons that the legislators might not have thought to enumerate. It also need not worry about whether the jury can agree on the type of weapon involved in any particular case. In doing so, it also takes a risk: it cannot be certain whether the court will interpret the statute as applying to a particular device. The legislature that enacts the latter type of statute exerts more control from the outset: it can be assured that the particular devices enumerated in the statutory text will be covered and others will not. [76][76]. These are the familiar tradeoffs of generality versus specificity, standards versus rules. See, e.g., John F. Manning, Second-Generation Textualism, 98 Calif. L. Rev. 1287, 1316 (2010); Frank H. Easterbrook, Statutes’ Domains, 50 U. Chi. L. Rev. 533, 546–47 (1983). (Perhaps the nunchuck lobby was influential in that debate). In the place of a single broad element, the legislature has employed a series of alternative, narrow elements, and the government must establish at least one of them before a conviction will be proper. The two different statutes are not equivalent.

This is the flipside of what the Descamps Court was getting at when it explained how the Ninth Circuit had erred in failing to accord significance to the difference between divisible and indivisible statutes on the theory that all indivisible statutes could be “hypothetically reconceive[d]” in “divisible terms” with a court supplying the missing “implied list of every means of commission that otherwise fits the definition of a given crime.” [77][77]. Descamps v. United States, 133 S. Ct. 2276, 2290 (2013). As the Court stated in Descamps, “the thing about hypothetical lists is that they are, well, hypothetical.” [78][78]. Id. at 2289, 2290. Conversely, when the legislature has enacted a list, it does not matter that the entire list hypothetically could be generalized into one broader term: at least one item on the list must be established beyond a reasonable doubt or else the requisite element has not been established. And the only way to be sure of that is to apply the familiar mechanism: juror agreement beyond a reasonable doubt. Requiring jury agreement as to at least one alternative element would be more consistent with the Court’s renewed focus, since Apprendi, on the structural role of the jury as the bulwark between the accused and the State than the contrary approach, which runs the risk of subverting the jury’s checking function. The Court seemed to be moving in this direction in Richardson, which was decided one year prior to Apprendi. Descamps may prove to be the bridge that helps complete this arc.

2. Double Jeopardy.

Descamps’ embrace of alternative elements may also lead to interesting developments in the Double Jeopardy context. If alternative elements were, in fact, indicative of separate offenses, as the Court indicated, that would suggest pursuant to Blockburger that a defendant could be punished consecutively for committing an offense pursuant to two different alternative elements. [79][79]. See United States v. Dixon, 509 U.S. 688, 696 (1993) (reaffirming the Blockburger test for constitutionality of both multiple punishments and multiple prosecutions). It would also seem to indicate that the government could prosecute a defendant by charging one statutory alternative element in one proceeding and then, especially if the government prevailed at trial, charging another alternative in a subsequent proceeding. [80][80]. Where the defendant was acquitted at a first trial, collateral estoppel may prevent the government from retrying the case, even pursuant to a different alternative element. See Dixon, 509 U.S. at 710 n.15; Ashe v. Swenson, 397 U.S. 436, 443 (1970); Nancy J. King, Portioning Punishment: Constitutional Limits on Successive and Excessive Penalties, 144 U. Pa. L. Rev. 101, 131–32 (1995). If a defendant pleaded guilty pursuant to a plea agreement, a defendant may enjoy additional protection from subsequent prosecution under the terms of the agreement, express or implied. See Santobello v. New York, 404 U.S. 257, 262–63 (1971); Daniel C. Richman, Bargaining About Future Jeopardy, 49 Vand. L. Rev. 1181, 1211–15 (1996). In this area, a theory of the jury’s function does not provide guidance; instead, one needs a theory of counts and of the purposes of the Double Jeopardy Clause. Both areas are excellent candidates for renewed focus in a post-Apprendi era. [81][81]. Many scholars have pointed out the unsatisfactory state of the Court’s current Double Jeopardy jurisprudence. See, e.g., Carissa Byrne Hessick & F. Andrew Hessick, Double Jeopardy as a Limit on Punishment, 97 Cornell L. Rev. 45, 54–55 (2011) (arguing that the Court’s Double Jeopardy jurisprudence “has largely failed to provide any real protection against multiple punishments”); Anne Bowen Poulin, Double Jeopardy and Multiple Punishment: Cutting the Gordian Knot, 77 U. Colo. L. Rev. 595, 599 (2006) (arguing that Double Jeopardy jurisprudence has become confused because of the erroneous entangling of multiple punishment and successive prosecution concerns); Jacqueline E. Ross, Damned Under Many Headings, the Problem of Multiple Punishment, 29 Am. J. Crim. L. 245, 251–53 (2002) (arguing that the Court has focused its Double Jeopardy jurisprudence on illusory problems while ignoring some real ones). The multiple-punishment problem was effectively suppressed through the United States Sentencing Guidelines and similar mandatory state sentencing regimes prior to Apprendi and Booker. Such schemes generally blunted the punitive impact of additional related counts. [82][82]. See generally Ross, supra note 81, at 253–57 (explaining how the Guidelines prevent multiple punishments for additional related counts). But in the post-Apprendi, post-Booker world of nonbinding sentencing guidelines, the problem resurfaces. Absent a constraining theory of counts, prosecutors wield enormous power over a defendant’s maximum sentence (and hence obtain additional leverage to induce a guilty plea) by deciding how many charges to include in a charging instrument. Those same charging decisions also can determine whether a defendant may be exposed to a subsequent prosecution premised on distinct charges.

Should a case arise challenging multiple punishments or multiple prosecutions based on Descamps, one resolution available to the Court (to resolve the challenge in the defendant’s favor) is to emphasize the alternative in “alternative elements.” The Court could point to the language in the opinion where it described alternative elements as creating different “versions” [83][83]. Descamps v. United States, 133 S. Ct. 2276, 2284 (2013) (describing the Court’s task in Shepard of determining which “version” of the Massachusetts burglary statute was the basis for the defendant’s conviction). of the offense and deemphasize the (more oft-repeated) language where it referred to alternative elements as signaling different crimes. But the more interesting course would be for the Court to stick to its characterization of alternative elements as elements indicative of different crimes—and then grapple anew with whether and when multiple punishments and multiple prosecutions are proper when they are sought pursuant to different offenses. In this context, as in the ACCA context, the Court has embraced a categorical approach under Blockburger, examining the elements of the offense rather than the underlying conduct involved in a particular crime (although it has periodically flirted with a conduct-based approach).

As the Court made clear in Descamps, the categorical approach has Sixth Amendment underpinnings, based on the jury trial right, in the context of the ACCA. In the context of Double Jeopardy, it is less clear what the rationale is for the categorical approach. It is arguably more easily administered than a conduct-based approach, and the Court has observed that it has a constitutional pedigree superior to that of a conduct-based approach. [84][84]. See Dixon, 509 U.S. at 704–11. In this context, however, the categorical approach leaves prosecutors’ power in stacking punishments and prosecutions largely unchecked. Thus, one open question after Descamps is whether the decision will amplify prosecutors’ power in this regard by increasing the number of offenses available to them when they charge a defendant, or conversely whether it will lead to a renewed examination of the Court’s Double Jeopardy jurisprudence. [85][85]. Another example of Descamps’ possible application would be in the context of the Sixth Amendment right to counsel. The Court has held that this right is “offense-specific,” and has adopted a categorical approach, using Blockburger’s elements test, in determining whether a defendant’s Sixth Amendment right to counsel has attached as to a particular offense at a particular point in time. See Texas v. Cobb, 532 U.S. 162, 173 (2001) (holding that “when the Sixth Amendment right to counsel attaches, it does encompass offenses that, even if not formally charged, would be considered the same offense under the Blockburger test”).

Conclusion

The Supreme Court’s decision in Descamps v. United States is the Court’s most recent opinion analyzing the application of the frequently litigated Armed Career Criminal Act. Drawing upon its prior decisions in this area, the Court reiterated that sentencing courts should apply a categorical approach in determining whether a defendant’s prior conviction constitutes a prior felony for purposes of the ACCA—meaning that courts should look at the elements of the prior offense of conviction rather than the underlying conduct. The Court also reiterated that courts may apply modified categorical review, and consult certain underlying judicial materials, to determine the basis for a conviction in certain circumstances. Descamps clarified the circumstances that permit modified categorical review. The Court held that such review is appropriate when a defendant’s prior conviction was pursuant to a divisible statute, i.e., one explicitly setting forth alternative elements, some of which fall within the generic version of the felony encompassed by the ACCA, and others of which do not. By applying modified categorical review, a sentencing court can determine which statutory alternative element provided the basis for the conviction.

In the near term, Descamps will affect the sentencing of thousands of criminal defendants, making it harder for the government to establish that a defendant has the requisite predicate felonies under the ACCA or other sentencing regimes pegged to the ACCA’s definitions of prior felonies. Many defendants (like Descamps) will have been convicted pursuant to indivisible statutes that are broader than the generic versions of the offenses incorporated into the ACCA. As to those defendants, the government effectively will be unable to establish the necessary predicate felonies for the ACCA. This is Descamps’ immediate import and, standing alone, it makes the decision quite significant. However, Descamps also has potential longer-term implications that, while less certain, are more profound. Descamps suggests that statutes like the ACCA can, and in this Court’s hands may, push us to think more carefully about what facts are required to establish a particular offense and how those facts are established and recorded in the context of an adjudicatory proceeding. In doing so, Descamps highlights not just the statutory specificity legislators are capable of, but also the adjudicative clarity that courts can promote.

The local prosecutors who handle most of the cases in the United States are unlikely to give much thought to whether a conviction will “count” for ACCA purposes in some later federal prosecution. But the clarity Descamps demands about what a defendant actually did should influence judicial retrospection in other contexts too. As the contexts multiply in which the precise basis for a prior conviction matters (and they are multiplying), prosecutors and legislators will have increased incentives to devise ways of making such a showing. Thus, we may see a change in how criminal statutes are drafted and how they are charged and submitted to juries. And, in the course of establishing a standard for ACCA purposes, the Court deployed a new analytic term—alternative elements—that may assist trial courts in another troubled context: sorting through what jurors need to agree on (elements) and what they need not (means). We have comparatively few criminal trials, but those we do have can and should do far more work at identifying what a jury of the defendant’s peers actually agreed upon. Descamps provides a framework that may help achieve that goal.


Copyright © 2015 Jessica A. Roth.

Associate Professor of Law, Benjamin N. Cardozo School of Law.

[1] Alleyne v. United States, 133 S. Ct. 2151 (2013).

[2] Apprendi v. New Jersey, 530 U.S. 466 (2000).

[3] Harris v. United States, 536 U.S. 545 (2002).

[4] Descamps v. United States, 133 S. Ct. 2276 (2013).

[5] While I am not the first to note Descamps’ significance, this Essay is the first scholarly work to explore it fully. See Douglas Berman, SCOTUS Wraps Its Sentencing Docket With Another Defense Win (and Alito Dissent) In Descamps, Sentencing Law and Policy (June 20, 2013 10:11 AM), http://sentencing.typepad.com/sentencing_law_and_policy/2013/06/scotus-wraps-its-sentencing-docket-with-another-defense-win-and-alito-dissent-in-descamps.html (noting the possibility that “Descamps will prove to be the most consequential of all the Supreme Court’s criminal sentencing work this Term”).

[6] See U.S. Sentencing Guidelines Manual § 4B1.1 (2014). For other federal sentencing guidelines that provide an enhancement based upon a qualifying prior conviction, see e.g., U.S. Sentencing Guidelines Manual § 2L1.2 (2014) (providing enhancement based on a prior conviction for the crime of unlawfully entering and remaining in the United States) and U.S. Sentencing Guidelines Manual § 2K2.1 (2014) (providing enhancement based on a prior conviction for firearms and ammunition offenses).

[7] For the past several years, approximately 600 criminal defendants per year have been sentenced as Armed Career Criminals and approximately 2200 criminal defendants per year have been sentenced as Career Offenders. See U.S. Sentencing Comm’n, 2013 Sourcebook of Federal Sentencing Statistics tbl. 22 (2013) (noting that there were 582 defendants sentenced as armed career criminals and 2268 sentenced as career offenders in 2013); U.S. Sentencing Comm’n, 2012 Sourcebook of Federal Sentencing Statistics tbl. 22 (2012) (noting that there were 631 armed career criminal sentences and 2232 career offender sentences in 2012); U.S. Sentencing Comm’n, 2011 Sourcebook of Federal Sentencing Statistics tbl. 22 (2011) (noting that there were 571 armed career criminal sentences and 2257 career offender sentences in 2011); U.S. Sentencing Comm’n, 2010 Sourcebook of Federal Sentencing Statistics tbl. 22 (2010) (noting that there were 616 armed career criminal sentences and 2314 career offender sentences in 2010); U.S. Sentencing Comm’n, 2009 Sourcebook of Federal Sentencing Statistics tbl. 22 (2009) (noting that there were 697 armed career criminal sentences and 2392 career offender sentences in 2009); U.S. Sentencing Comm’n, 2008 Sourcebook of Federal Sentencing Statistics tbl. 22 (2008) (noting that there were 653 armed career criminal sentences and 2321 career offender sentences in 2008). The figures for 2014 are not yet available.

[8] See, e.g., Olten v. United States, 134 S. Ct. 639, 639 (2013); Smith v. United States, 134 S. Ct. 258, 258 (2013); United States v. Prater, 766 F.3d 501, 523 (6th Cir. 2014); United States v. Brown, 765 F.3d 185, 196 (3d Cir. 2014); United States v. Bankhead, 746 F.3d 323, 327 (8th Cir. 2014); United States v. Howard, 742 F.3d 1334, 1349 (11th Cir. 2014); United States v. Tucker, 740 F.3d 1177, 1184 (8th Cir. 2014); United States v. Nunez-Segura, 566 F. App’x 389, 394 (5th Cir. 2014); United States v. Royal, 731 F.3d 333, 342 (4th Cir. 2013); United States v. Hemingway, 734 F.3d 323, 338 (4th Cir. 2013); United States v. Parks, 2014 U.S. Dist. LEXIS 1617, at *13–14 (W.D. Wash. Jan. 7, 2014).

[9] See, e.g., Olsson v. United States, 134 S. Ct. 530, 530 (2013); Marrero v. United States, 133 S. Ct. 2732, 2733 (2013); United States v. Brown, 765 F.3d 185, 196–97 (3d Cir. 2014); United States v. Covington, 738 F.3d 759, 767 (6th Cir. 2014); United States v. Murguia-Ochoa, 536 F. App’x 696, 699 (9th Cir. 2013).

[10] See, e.g., United States v. Coronado, 603 F.3d 706, 708 (9th Cir. 2010); United States v. Polk, 577 F.3d 515, 518–19 (3d Cir. 2009); United States v. Seay, 553 F.3d 732, 737–39 (4th Cir. 2009); United States v. Mohr, 554 F.3d 604, 607–09 (5th Cir. 2009); United States v. Herrick, 545 F.3d 53, 56–58 (1st Cir. 2008); United States v. Gray, 535 F.3d 128, 130–31 (2d Cir. 2008); United States v. Bartee, 529 F.3d 357, 359 (6th Cir. 2008); United States v. Templeton, 543 F.3d 378, 380 (7th Cir. 2008); United States v. Williams, 537 F.3d 969, 971 (8th Cir. 2008); United States v. Tiger, 538 F.3d 1297, 1298 (10th Cir. 2008); United States v. Archer, 531 F.3d 1347, 1350–52 (11th Cir. 2008). As noted in note 7, supra, approximately 2200 federal defendants are sentenced as career offenders each year, nearly four times as many as are sentenced as armed career criminals.

[11] See, e.g., Ruiz-Sanchez v. United States, 134 S. Ct. 60, 60 (2013) (noting illegal reentry guidelines); United States v. Davis, 751 F.3d 769, 778 (6th Cir. 2014) (noting child pornography guidelines); United States v. Falcon, 553 F. App’x 737, 738 (9th Cir. 2014) (noting firearms and ammunition guidelines); United States v. Estrella, 758 F.3d 1239, 1254 (11th Cir. 2014) (noting illegal reentry guidelines); United States v. Martinez, 756 F.3d 1092, 1098 (8th Cir. 2014) (same); United States v. Edwards, 734 F.3d 850, 851 (9th Cir. 2013) (noting firearms and ammunition guidelines).

[12] See, e.g., People v. Wilson, 162 Cal. Rptr. 3d 43, 60 (Ct. App. 2013); State v. Dickey, 329 P.3d 1230, 1246 (Kan. Ct. App. 2014); State v. Hulsey, 2013 Kan. App. Unpub. LEXIS 861, at *29 (Kan. Ct. App. 2013). Approximately half of the states have recidivist laws—sometimes called “three strikes” laws—that are similar to the ACCA. See John Clark, James Austin & D. Alan Henry, “Three Strikes and You’re Out”: A Review of State Legislation, National Institute of Justice Research in Brief, National Institute of Justice 7–9 (Sept. 1997); Robert Heglin, Note, A Flurry of Recidivist Legislation Means: “Three Strikes and You’re Out,” 20 J. Legis. 213, 215–16 & nn.18–23 (1994) (collecting state statutes).

[13] See e.g., Rendon v. Holder, 764 F.3d 1077, 1090 (9th Cir. 2014); Aguilar-Turcios v. Holder, 740 F.3d 1294, 1302 (9th Cir. 2014); Lima-Lima v. Holder, 545 F. App’x 648, 650 (9th Cir. 2013); Makwana v. Att’y Gen. of the United States, 543 F. App’x 186, 190 (3d Cir. 2013); Donawa v. United States Att’y Gen., 735 F.3d 1275, 1283–84 (11th Cir. 2013).

[14] See, e.g., Rendon, 764 F.3d at 1084–88 (considering the required elements of a California burglary conviction and the extent of jury unanimity required as to how the offense was committed); United States v. Abbott, 748 F.3d 154, 158–59 (3d Cir. 2014) (same, as to Pennsylvania drug statute); United States v. Royal, 731 F.3d 333, 340–42 (4th Cir. 2014) (same, as to Maryland assault statute); United States v. Cabrera-Gutierrez, 756 F.3d 1125, 1133–37 (9th Cir. 2013) (same, as to Oregon sexual abuse statute).

[15] 18 U.S.C. § 924(e)(1) (2012).

[16] 18 U.S.C. § 924(e)(2)(B)(i) (2012).

[17] 18 U.S.C. § 924(e)(2)(B)(ii) (2012).

[18] Id. Courts frequently refer to these three clauses as the ACCA’s “force” prong, the “enumerated felonies” prong, and the “residual clause,” respectively. See, e.g., United States v. Elliott, 757 F.3d 492, 493 (6th Cir. 2014). The Supreme Court recently ordered new briefing and argument on the question of whether the residual clause is unconstitutionally vague. See Johnson v. United States, No. 13-7120, 2015 WL 132524 (2015) (ordering briefing to be completed by Apr. 10, 2015).

[19] 18 U.S.C. § 924(a)(2).

[20] A defendant who is found to be an Armed Career Criminal may also face an enhanced sentencing range under the United States Sentencing Guidelines as a consequence of that designation. See U.S. Sentencing Guidelines Manual § 4B1.4 (2014) (providing a minimum offense level of 33 and a minimum criminal history category of IV for an Armed Career Criminal, and even higher levels for both if certain facts are established).

[21] Prior to Descamps, the Supreme Court had decided at least nine cases in the previous ten years analyzing what prior convictions count as ACCA predicate felonies. See, e.g., McNeill v. United States, 131 S. Ct. 2218, 2224 (2011) (analyzing when prior state drug convictions count as serious drug offenses for purposes of the ACCA); Sykes v. United States, 131 S. Ct. 2267, 2277 (2011) (analyzing whether a state felony vehicle flight crime qualified as a violent felony under the ACCA’s residual clause); Johnson v. United States, 559 U.S. 133, 145 (2010) (analyzing whether a prior state conviction for battery constituted a violent felony under residual clause); Chambers v. United States, 555 U.S. 122, 130 (2009) (analyzing whether a state conviction for failure to report for periodic incarceration constituted a violent felony under the residual clause); Begay v. United States, 553 U.S. 137, 145 (2008) (analyzing whether a state conviction for felony driving under the influence qualified under the residual clause); United States v. Rodriguez, 553 U.S. 377, 393 (2008) (analyzing whether a state drug conviction counted as a serious drug offense); James v. United States, 550 U.S. 192, 214 (2007) (analyzing whether a state attempted burglary statute qualified under the residual clause); Logan v. United States, 552 U.S. 23, 37 (2007) (considering the implications of a state’s failure to revoke the defendant’s civil rights on whether a conviction counted under the ACCA); Shepard v. United States, 544 U.S. 13, 27 (2005) (setting forth materials that may be consulted when a court applies modified categorical review).

[22] The statute originally included a definition of burglary, but Congress deleted this definition in the course of amending the statute in 1986 to expand the scope of the predicate felonies that count for ACCA purposes. Although the deletion appears to have been unintentional, it has never been corrected. See Taylor v. United States, 495 U.S. 575, 588–90 (1990).

[23] See id. at 589.

[24] Descamps, 133 S. Ct. at 2283.

[25] The Ninth and Sixth Circuits had applied the modified categorical approach to indivisible statutes, as explained infra, whereas the Second and First Circuits had held that the modified categorical approach applied only to divisible statutes. Compare United States v. Descamps, 466 F. App’x 563, 565 (9th Cir. 2012), and United States v. Armstead, 467 F.3d 943, 947–50 (6th Cir. 2006), with United States v. Beardsley, 691 F.3d 252, 268–74 (2d Cir. 2012), and United States v. Giggey, 551 F.3d 27, 40 (1st Cir. 2008) (en banc).

[26] Descamps, 133 S. Ct. at 2284.

[27] Id. at 2283–84.

[28] The prior conviction at issue in Descamps’ case was a conviction for burglary pursuant to Cal. Penal Code Ann. § 459 (West 2010), which provided that a “person who enters” certain locations “with intent to commit grand or petit larceny or any felony is guilty of burglary.” The statute did not require that the entry have been unlawful—an element that the Supreme Court deemed a necessary component of the “generic version” of burglary. See Descamps, 133 S. Ct. at 2285, 2294 (Thomas, J., concurring) (citing 3 Wayne LaFave, Substantive Criminal Law § 21.1(a) (2d ed. 2003)).

[29] See, e.g., Shepard v. United States, 544 U.S. 13, 35–36 (O’Connor, J., dissenting) (decrying the categorical approach’s “overscrupulous regard for formality,” which “forces the . . . sentencing court to feign agnosticism about clearly knowable facts” and “cannot be squared with the ACCA’s twin goals of incapacitating repeat violent offenders, and of doing so consistently notwithstanding the peculiarities of state law”) (emphasis in original). See also Descamps, 133 S. Ct. at 2302 (Alito, J., dissenting) (“The Court’s holding will . . . frustrate fundamental ACCA objectives” including that “violent, dangerous recidivists would be subject to enhanced penalties” in that “those enhanced penalties would be applied uniformly.”).

[30] Descamps, 133 S. Ct. at 2283.

[31] As Daniel Richman has observed, “[e]ven for those not ordinarily interested in technical sentencing decisions, the Court’s opinion is worth reading for its miffed tone of a well-respected author confronted with an obdurate reader claiming unfamiliarity with her work.” Daniel Richman, Opinion analysis: When is a burglary not a burglary?, SCOTUSblog (June 20, 2013, 11:18 PM), http://www.scotusblog.com/2013/06/opinion-analysis-when-is-a-burglary-not-a-burglary; see also William Baude, Modified Categorical Imperative, PrawfsBlawg (June 21, 2013, 2:09 AM), http://prawfsblawg.blogs.com/prawfsblawg/2013/06/modified-categorical-imperative.html (“If you are interested in sharp-tongued judicial rhetoric, you should be following Justice Kagan” who “positively mocks the Ninth Circuit decision.”).

[32] Taylor v. United States, 495 U.S. 575 (1990).

[33] Id. at 601–02.

[34] Id. at 602.

[35] Id.

[36] Shepard v. United States, 544 U.S. 13 (2005).

[37] Id. at 20.

[38] Id. at 21.

[39] Johnson v. United States, 559 U.S. 133 (2010).

[40] 18 U.S.C. § 924(e)(1) (2006).

[41] Johnson, 559 U.S. at 144. In Johnson, the Court finally used the phrase “modified categorical approach,” quoting its decision in Nijhawan v. Holder, 557 U.S. 29 (2009), an immigration case in which the Court considered the application of the Taylor line of cases to the construction of certain felonies under the Immigration and Nationality Act (INA) that could provide the basis for deportation. Id. In Nijhawan, the Court ultimately rejected the use of the categorical approach for purposes of this section of the INA, holding that it was not consistent with Congressional intent. Nijhawan, 557 U.S. at 39–40. However, the Court has held that the categorical approach does apply with respect to certain other portions of the INA. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 194 (2007) (endorsing the categorical approach to interpret the term “theft offense” in the INA).

[42] Johnson represented the usual case in which the Government advocated for categorical review rather than modified categorical review, because the underlying state documents were inconclusive. The Government argued that the Florida statute categorically qualified as a violent felony under the ACCA, thus obviating the need for resort to the underlying documents. Johnson, 558 U.S. at 139. The Court disagreed. Id.

[43] United States v. Descamps, 466 F. App’x 563 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013).

[44] Johnson, 559 U.S. at 144 (emphasis added).

[45] Descamps v. United States, 133 S. Ct. 2276, 2289 (2013).

[46] Id. (quoting United States v. Aguila-Montes de Oca, 655 F.3d 915, 927 (9th Cir. 2011)) (alteration in original).

[47] Id. at 2288 (citing Apprendi v. New Jersey, 530 U.S. 466 (2000)).

[48] Id.

[49] Id. (citing Apprendi, 530 U.S. at 490). In Almendarez-Torres v. United States, 523 U.S. 224 (1998), the Supreme Court held that the existence of a prior conviction was a sentencing factor rather than an element and, accordingly, did not need to be included in an indictment or found by a jury. Post-Apprendi, the Court consistently has recognized Almendarez-Torres as an exception to the Apprendi line of cases, notwithstanding repeated calls from various quarters (including from within the Court, most notably from Justice Thomas) to overrule it. Descamps, 133 S. Ct. at 2294–95 (Thomas, J., concurring); James v. United States, 550 U.S. 192, 231 (2007) (Thomas, J., dissenting); Shepard v. United States, 544 U.S. 13, 27 (2005) (Thomas, J., concurring in part and concurring in the judgment). Descamps argued in the lower courts that Almendarez-Torres should be overruled and sought certiorari on that issue; the Court denied that portion of his petition. United States v. Descamps, 466 F. App’x 563, 564–65 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013).

[50] Descamps, 133 S. Ct. at 2290.

[51] See Wayne LaFave, Jerold H. Israel, Nancy King & Orin Kerr, Criminal Procedure § 19.3(b) (5th ed. 2009 & Supp. 2013).

[52] See Fed. R. Crim. P. 7(c)(1) (providing that a single count of an indictment may allege that “the means by which the defendant committed the offense are unknown or that the defendant committed it by one or more specified means”); Griffin v. United States, 502 U.S. 46, 51 (1991) (“A statute often makes punishable the doing of one thing or another . . . sometimes thus specifying a considerable number of things . . . . [T]he indictment on such a statute may allege, in a single count, that the defendant did as many of the forbidden things as the pleader chooses, employing the conjunction and where the statute has ‘or,’ . . . and it will be established at the trial by proof of any one of them.” (quoting 1 Joel Prentiss Bishop, New Criminal Procedure § 436, at 355–56 (2d ed. 1913))) (internal footnotes omitted) (internal quotation marks omitted) (emphasis omitted).

[53] See Descamps, 133 S. Ct. at 2298 (Alito, J., dissenting) (envisioning the case of a defendant charged with burglarizing a “floating home” pursuant to a state statute defining burglary as including unlawful entry into, inter alia, a “building” or “vessel” and asking whether, in order to convict, the jury would have to agree “whether this structure was a ‘building’ or a ‘vessel.’ . . . The Court’s answer is ‘yes.’”). Justice Alito also notes that, even if unanimity were required, the basis for the conviction would not be apparent to a subsequent sentencing court unless the jury returned “a special verdict, something that is not generally favored in criminal cases.” Id. at 2300 (citing 6 Wayne LaFave & Jerold H. Israel, Criminal Procedure § 24.10(a), at 543–44 (3d ed. 2007)).

[54] Schad v. Arizona, 501 U.S. 624 (1991).

[55] Richardson v. United States, 526 U.S. 813 (1999).

[56] For a discussion of the indeterminacy of Schad and Richardson, see, for example, Scott W. Howe, Jury Fact-Finding in Criminal Cases: Constitutional Limits on Factual Disagreements Among Convicting Jurors, 58 Mo. L. Rev. 1, 6 (1993); Brian M. Morris, Something Upon Which We Can All Agree: Requiring a Unanimous Jury Verdict in Criminal Cases, 62 Mont. L. Rev. 1, 19–24, 31–34 (2001); Jessica A. Roth, Alternative Elements, 59 UCLA L. Rev. 170, 186–97 (2011); Stephen E. Sachs, Alternative Theories of the Crime 5–6 (May 24, 2007) (unpublished manuscript) (http://ssrn.com/abstract=1501628); Peter Westen & Eric Ow, Reaching Agreement on When Jurors Must Agree, 10 New Crim. L. Rev. 153, 153–54 (2007); see also United States v. Hurt, 527 F.3d 1347, 1356 (D.C. Cir. 2008) (observing that Schad does not provide clear guidance).

[57] The Court has not provided any particularized guidance as to the appropriate methodology in determining whether a legislature intended to create means as opposed to elements. In Schad, however, the Court did cite approvingly to United States v. UCO Oil Co., 546 F.2d 833, 835–38 (9th Cir. 1976), in which the court stated that statutory language, legislative history, the nature of the proscribed conduct, and the appropriateness of multiple punishments all should bear on whether a statute should be interpreted as evidencing a legislative intent to create multiple crimes as opposed to a single crime with multiple means of commission. See Schad, 501 U.S. at 636 (citing UCO Oil Co., 546 F.2d at 835–38).

[58] The Court did, however, suggest certain factors that would be relevant to such a determination, including the moral equivalency of the various alternatives, whether historically they were treated as discrete crimes, and the practices of other states. For example, in Schad, the plurality opinion stated that it would be unconstitutional for a state to recast its entire criminal code as a single statute and declare each formerly separate crime a “means” of committing the offense of “crime.” Schad, 501 U.S. at 632–33.

[59] See, e.g., United States v. Seher, 562 F.3d 1344, 1361–62 (11th Cir. 2009) (holding that a jury need not agree on a defendant’s motive for engaging in a financial transaction, as set forth in the federal money laundering statute); United States v. Stewart, 433 F.3d 273, 319 (2d Cir. 2006) (holding that a jury need not agree on whether a defendant made a false statement or concealed a material fact, as set forth in the false statements statute); United States v. Powell, 226 F.3d 1181, 1196 (10th Cir. 2000) (holding that a jury need not agree on which verb characterizes a defendant’s conduct, as set forth in the federal kidnapping statute). See also 6 LaFave et al., supra note 53, at § 24.10(c), nn.40–41 (collecting additional state and federal cases involving multiple theories and in which courts have approved general verdicts without requiring juror agreement as to the theory).

[60] See Hayden J. Trubitt, Patchwork Verdicts, Different-Juror Verdicts, and American Jury Theory: Whether Verdicts are Invalidated by Juror Disagreement on Issues, 36 Okla. L. Rev. 473, 474–77 (1983) (describing how jury selection and individual jury behavior can lead to compromises that actually stymie true deliberation and discussion among jurors).

[61] Ex ante, adjudicative clarity is a protection for defendants. Ex post, it has positive externalities for those, generally prosecutors, who want to piggyback off the prior adjudicated fact. It is perhaps not surprising then that an Internet search for materials citing Descamps reveals a number of training materials put together by defense attorneys recommending that defenders representing ACCA defendants at sentencing argue that their clients’ prior convictions were pursuant to statutory phrases that represent alternative means rather than alternative elements. See, e.g., Paresh Patel & Lisa Freeland, Post-Descamps World, http://agendas.fjc.gov/agenda/files/FedDef1410-PostDescampsWorld.pdf (last visited Jan. 26, 2015).

[62] Even putting aside the ACCA, there are institutional reasons why prosecutors would be in favor of more specialized verdicts. See Roth, supra note 56, at 213–14 (explaining how prosecutors sometimes benefit from seeking greater clarity in the basis for a jury’s verdict). Congress could amend the ACCA to make clear that “conviction” refers to underlying conduct, not the elements of the prior offense. In his concurring opinion in Descamps, Justice Kennedy called on Congress to make such a change. Descamps v. United States, 133 S. Ct. 2276, 2294 (2013) (Kennedy, J., concurring). But it is not clear that such a change would be consistent with the Sixth Amendment. See id. at 2288 (explaining that if a court’s finding of a predicate offense went beyond merely identifying a prior conviction it would raise Sixth Amendment concerns). Assuming that the Court means what it says when it cites the categorical approach’s Sixth Amendment underpinnings in the context of criminal sentencing, a circumstance-based approach would seem to present serious constitutional problems.

[63] See, e.g., Griffin v. United States, 502 U.S. 46, 60 (1991); id. at 61 (Blackmun, J., concurring); Schad, 501 U.S. at 645.

[64] Schad was decided by a 5-4 vote. Justice Souter wrote the main opinion for the Court for a plurality of the Justices (joined by Chief Justice Rehnquist and Associate Justices O’Connor and Kennedy). Schad, 501 U.S. at 624. Justice Scalia provided the crucial fifth vote to affirm, but made clear in a separate opinion that, in any other case, he might well find that jury unanimity was constitutionally required. Id. at 651 (Scalia, J., dissenting). Richardson, which found elements rather than means—citing constitutional concerns—was written by Justice Breyer, joined by Chief Justice Rehnquist and Justices Stevens, Scalia, Souter, and Thomas. Richardson v. United States, 526 U.S. 813, 814 (1998). Justice Kennedy dissented in Richardson, joined by Justices O’Connor and Ginsburg. Id. at 825 (Kennedy, J., dissenting). Consistent with his position in Descamps, then-Judge Alito dissented in a case decided by the Third Circuit involving the same issue as Richardson. See United States v. Edmonds, 80 F.3d 810, 829 (3d Cir. 1996) (Alito, J., dissenting). Put together with the line-up in Descamps, there now appear to be six members of the Court who might vote to require jury unanimity as to at least one statutory alternative element (Justices Kagan, Breyer, Stevens, Scalia, Thomas, and Chief Justice Roberts) and three Justices who almost certainly would take the opposite view (Justices Kennedy, Ginsburg, and Alito).

[65] See supra notes 63–64.

[66] Descamps, 133 S. Ct. at 2288.

[67] The phrase had been used in the lower courts opinions discussing the application of the ACCA. See, e.g., United States v. Aguila-Montes de Oca, 655 F.3d 915, 963 (9th Cir. 2011); United States v. Vann, 660 F.3d 771, 789 (4th Cir. 2011).

[68] See Descamps, 133 S. Ct. at 2285 n.2 (expressing surprise that the state statutes analyzed in Taylor, Shepard, and Johnson could be construed as setting forth anything but “alternative elements,” and therefore “several different . . . crimes”).

[69] Id.; see id. at 2297–98 (Alito, J., dissenting) (“[L]egislatures frequently enumerate alternative means of committing a crime without intending to define separate elements or separate crimes.” (citing Schad v. Arizona, 501 U.S. 624, 636 (1991))).

[70] The closest the Court came to dealing with this issue in Descamps was to acknowledge Justice Alito’s point that some divisible statutes might set forth alternative means as opposed to elements. Id. at 2285 n.2. The Court expressed surprise that any of the divisible statutes in its ACCA cases to date established means as opposed to elements, and asserted that any confusion as to what constituted elements as opposed to means in future cases could be clarified by consulting the indictment and jury instructions. Id.

[71] See 6 LaFave et al., supra note 53, at § 24.10(c) (observing that the Court has never “derive[d] a uniform theory for identifying which facts are elements”); Nancy J. King & Susan R. Klein, Essential Elements, 54 Vand. L. Rev. 1467, 1497 (2001) (ultimately rejecting as “unwise” and “unattainable” the “theoretical quest for the elemental holy grail”).

[72] Hamling v. United States, 418 U.S. 87, 117 (1974).

[73] See Patterson v. New York, 432 U.S. 197, 210 (1977) (holding that the legislature cannot shift the burden to the defendant to disprove any element of the offense); In re Winship, 397 U.S. 358, 364 (1969) (holding that “the Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged”).

[74] Blockburger v. United States, 284 U.S. 299 (1932).

[75] Id. at 304.

[76] These are the familiar tradeoffs of generality versus specificity, standards versus rules. See, e.g., John F. Manning, Second-Generation Textualism, 98 Calif. L. Rev. 1287, 1316 (2010); Frank H. Easterbrook, Statutes’ Domains, 50 U. Chi. L. Rev. 533, 546–47 (1983).

[77] Descamps v. United States, 133 S. Ct. 2276, 2290 (2013).

[78] Id. at 2289, 2290.

[79] See United States v. Dixon, 509 U.S. 688, 696 (1993) (reaffirming the Blockburger test for constitutionality of both multiple punishments and multiple prosecutions).

[80] Where the defendant was acquitted at a first trial, collateral estoppel may prevent the government from retrying the case, even pursuant to a different alternative element. See Dixon, 509 U.S. at 710 n.15; Ashe v. Swenson, 397 U.S. 436, 443 (1970); Nancy J. King, Portioning Punishment: Constitutional Limits on Successive and Excessive Penalties, 144 U. Pa. L. Rev. 101, 131–32 (1995). If a defendant pleaded guilty pursuant to a plea agreement, a defendant may enjoy additional protection from subsequent prosecution under the terms of the agreement, express or implied. See Santobello v. New York, 404 U.S. 257, 262–63 (1971); Daniel C. Richman, Bargaining About Future Jeopardy, 49 Vand. L. Rev. 1181, 1211–15 (1996).

[81] Many scholars have pointed out the unsatisfactory state of the Court’s current Double Jeopardy jurisprudence. See, e.g., Carissa Byrne Hessick & F. Andrew Hessick, Double Jeopardy as a Limit on Punishment, 97 Cornell L. Rev. 45, 54–55 (2011) (arguing that the Court’s Double Jeopardy jurisprudence “has largely failed to provide any real protection against multiple punishments”); Anne Bowen Poulin, Double Jeopardy and Multiple Punishment: Cutting the Gordian Knot, 77 U. Colo. L. Rev. 595, 599 (2006) (arguing that Double Jeopardy jurisprudence has become confused because of the erroneous entangling of multiple punishment and successive prosecution concerns); Jacqueline E. Ross, Damned Under Many Headings, the Problem of Multiple Punishment, 29 Am. J. Crim. L. 245, 251–53 (2002) (arguing that the Court has focused its Double Jeopardy jurisprudence on illusory problems while ignoring some real ones).

[82] See generally Ross, supra note 81, at 253–57 (explaining how the Guidelines prevent multiple punishments for additional related counts).

[83] Descamps v. United States, 133 S. Ct. 2276, 2284 (2013) (describing the Court’s task in Shepard of determining which “version” of the Massachusetts burglary statute was the basis for the defendant’s conviction).

[84] See Dixon, 509 U.S. at 704–11.

[85] Another example of Descamps’ possible application would be in the context of the Sixth Amendment right to counsel. The Court has held that this right is “offense-specific,” and has adopted a categorical approach, using Blockburger’s elements test, in determining whether a defendant’s Sixth Amendment right to counsel has attached as to a particular offense at a particular point in time. See Texas v. Cobb, 532 U.S. 162, 173 (2001) (holding that “when the Sixth Amendment right to counsel attaches, it does encompass offenses that, even if not formally charged, would be considered the same offense under the Blockburger test”).

The Divisibility of Crime

Volume 64 February 2015
The Divisibility of Crime

Jessica A. Roth†. Associate Professor of Law, Benjamin N. Cardozo School of Law.

PDF
 

Near the end of the Supreme Court’s 2012-2013 term, the Court decided Descamps v. United States, which concerned the application of the federal Armed Career Criminal Act (ACCA). The ACCA is a recidivist statute that vastly increases the penalties for persons convicted of federal firearms offenses if they have previously been convicted of certain qualifying felonies. Descamps represents the Court’s most recent word on the so-called categorical approach, which directs courts to consider the elements of a prior offense of conviction, rather than the underlying facts of the crime, in determining whether the prior conviction “counts” for purposes of applying the ACCA and other sentencing enhancements and for determining the immigration consequences of prior convictions. This Essay is the first scholarly work to track the immediate effects of Descamps and to explore its implications for the criminal law more broadly. It shows that the decision is indeed having a significant effect on criminal sentencing, resulting in a steady flow of sentencing reversals and prospectively narrowing the class of defendants eligible for sentencing enhancements based on prior convictions. But more broadly, Descamps has called attention to the statutory specificity that legislators are capable of and the adjudicative clarity that courts can promote, if there are incentives for doing so. Until now, the Court has done little to encourage either. Thus, the opinion may push courts and legislators to think more carefully and systematically about what facts must be established to constitute a particular criminal offense, how such facts are established and recorded in the context of an adjudicative proceeding, and the consequences that flow from greater or lesser specificity. Ultimately, this impact may be felt not only in the context of applying recidivist statutes and sentencing enhancements, but also in other contexts that require attention to the basis for a criminal conviction, including the doctrine governing what constituent facts of a crime require jury unanimity and claims under the Double Jeopardy Clause.

Introduction

In its 2012–2013 term, the Supreme Court’s decisions on affirmative action, voting rights, and gay marriage received the lion’s share of attention from the press and scholars. Less noticed were the Court’s decisions on issues of criminal law and procedure. Alleyne v. United States, [1][1]. Alleyne v. United States, 133 S. Ct. 2151 (2013). which held that Apprendi v. New Jersey [2][2]. Apprendi v. New Jersey, 530 U.S. 466 (2000). requires juries to find facts necessary to support the imposition of a mandatory minimum sentence, was received with little fanfare even though it directly overruled precedent (Harris v. United States). [3][3]. Harris v. United States, 536 U.S. 545 (2002). Most scholars and commentators treated Alleyne as the proverbial other shoe that everyone had been expecting to drop. Such an attitude was warranted with Alleyne, but another decision, Descamps v. United States, [4][4]. Descamps v. United States, 133 S. Ct. 2276 (2013). which involved the sentencing of recidivist offenders under the federal Armed Career Criminal Act (ACCA), has received far less attention than it is due. [5][5]. While I am not the first to note Descamps’ significance, this Essay is the first scholarly work to explore it fully. See Douglas Berman, SCOTUS Wraps Its Sentencing Docket With Another Defense Win (and Alito Dissent) In Descamps, Sentencing Law and Policy (June 20, 2013 10:11 AM), http://sentencing.typepad.com/sentencing_law_and_policy/2013/06/scotus-wraps-its-sentencing-docket-with-another-defense-win-and-alito-dissent-in-descamps.html (noting the possibility that “Descamps will prove to be the most consequential of all the Supreme Court’s criminal sentencing work this Term”).

This Essay explains why Descamps matters. In the near term, Descamps will materially affect the severity of the sentences imposed on thousands of criminal defendants per year, narrowing the class of defendants eligible for enhanced sentences pursuant to the ACCA and other sentencing enhancements like the federal Career Offender guideline [6][6]. See U.S. Sentencing Guidelines Manual § 4B1.1 (2014). For other federal sentencing guidelines that provide an enhancement based upon a qualifying prior conviction, see e.g., U.S. Sentencing Guidelines Manual § 2L1.2 (2014) (providing enhancement based on a prior conviction for the crime of unlawfully entering and remaining in the United States) and U.S. Sentencing Guidelines Manual § 2K2.1 (2014) (providing enhancement based on a prior conviction for firearms and ammunition offenses). that are dependent upon a finding of qualifying prior convictions. [7][7]. For the past several years, approximately 600 criminal defendants per year have been sentenced as Armed Career Criminals and approximately 2200 criminal defendants per year have been sentenced as Career Offenders. See U.S. Sentencing Comm’n, 2013 Sourcebook of Federal Sentencing Statistics tbl. 22 (2013) (noting that there were 582 defendants sentenced as armed career criminals and 2268 sentenced as career offenders in 2013); U.S. Sentencing Comm’n, 2012 Sourcebook of Federal Sentencing Statistics tbl. 22 (2012) (noting that there were 631 armed career criminal sentences and 2232 career offender sentences in 2012); U.S. Sentencing Comm’n, 2011 Sourcebook of Federal Sentencing Statistics tbl. 22 (2011) (noting that there were 571 armed career criminal sentences and 2257 career offender sentences in 2011); U.S. Sentencing Comm’n, 2010 Sourcebook of Federal Sentencing Statistics tbl. 22 (2010) (noting that there were 616 armed career criminal sentences and 2314 career offender sentences in 2010); U.S. Sentencing Comm’n, 2009 Sourcebook of Federal Sentencing Statistics tbl. 22 (2009) (noting that there were 697 armed career criminal sentences and 2392 career offender sentences in 2009); U.S. Sentencing Comm’n, 2008 Sourcebook of Federal Sentencing Statistics tbl. 22 (2008) (noting that there were 653 armed career criminal sentences and 2321 career offender sentences in 2008). The figures for 2014 are not yet available. Already, the decision has prompted the reversal of a number of pre-Descamps federal ACCA sentences. [8][8]. See, e.g., Olten v. United States, 134 S. Ct. 639, 639 (2013); Smith v. United States, 134 S. Ct. 258, 258 (2013); United States v. Prater, 766 F.3d 501, 523 (6th Cir. 2014); United States v. Brown, 765 F.3d 185, 196 (3d Cir. 2014); United States v. Bankhead, 746 F.3d 323, 327 (8th Cir. 2014); United States v. Howard, 742 F.3d 1334, 1349 (11th Cir. 2014); United States v. Tucker, 740 F.3d 1177, 1184 (8th Cir. 2014); United States v. Nunez-Segura, 566 F. App’x 389, 394 (5th Cir. 2014); United States v. Royal, 731 F.3d 333, 342 (4th Cir. 2013); United States v. Hemingway, 734 F.3d 323, 338 (4th Cir. 2013); United States v. Parks, 2014 U.S. Dist. LEXIS 1617, at *13–14 (W.D. Wash. Jan. 7, 2014). It also has resulted in the reversal of numerous sentences imposed pursuant to the federal Career Offender guideline, [9][9]. See, e.g., Olsson v. United States, 134 S. Ct. 530, 530 (2013); Marrero v. United States, 133 S. Ct. 2732, 2733 (2013); United States v. Brown, 765 F.3d 185, 196–97 (3d Cir. 2014); United States v. Covington, 738 F.3d 759, 767 (6th Cir. 2014); United States v. Murguia-Ochoa, 536 F. App’x 696, 699 (9th Cir. 2013). which courts have long interpreted in a manner consistent with the ACCA; [10][10]. See, e.g., United States v. Coronado, 603 F.3d 706, 708 (9th Cir. 2010); United States v. Polk, 577 F.3d 515, 518–19 (3d Cir. 2009); United States v. Seay, 553 F.3d 732, 737–39 (4th Cir. 2009); United States v. Mohr, 554 F.3d 604, 607–09 (5th Cir. 2009); United States v. Herrick, 545 F.3d 53, 56–58 (1st Cir. 2008); United States v. Gray, 535 F.3d 128, 130–31 (2d Cir. 2008); United States v. Bartee, 529 F.3d 357, 359 (6th Cir. 2008); United States v. Templeton, 543 F.3d 378, 380 (7th Cir. 2008); United States v. Williams, 537 F.3d 969, 971 (8th Cir. 2008); United States v. Tiger, 538 F.3d 1297, 1298 (10th Cir. 2008); United States v. Archer, 531 F.3d 1347, 1350–52 (11th Cir. 2008). As noted in note 7, supra, approximately 2200 federal defendants are sentenced as career offenders each year, nearly four times as many as are sentenced as armed career criminals. other provisions of the federal Sentencing Guidelines that turn on a finding of a qualifying prior felony; [11][11]. See, e.g., Ruiz-Sanchez v. United States, 134 S. Ct. 60, 60 (2013) (noting illegal reentry guidelines); United States v. Davis, 751 F.3d 769, 778 (6th Cir. 2014) (noting child pornography guidelines); United States v. Falcon, 553 F. App’x 737, 738 (9th Cir. 2014) (noting firearms and ammunition guidelines); United States v. Estrella, 758 F.3d 1239, 1254 (11th Cir. 2014) (noting illegal reentry guidelines); United States v. Martinez, 756 F.3d 1092, 1098 (8th Cir. 2014) (same); United States v. Edwards, 734 F.3d 850, 851 (9th Cir. 2013) (noting firearms and ammunition guidelines). and state sentences imposed pursuant to state recidivist laws that closely track the ACCA. [12][12]. See, e.g., People v. Wilson, 162 Cal. Rptr. 3d 43, 60 (Ct. App. 2013); State v. Dickey, 329 P.3d 1230, 1246 (Kan. Ct. App. 2014); State v. Hulsey, 2013 Kan. App. Unpub. LEXIS 861, at *29 (Kan. Ct. App. 2013). Approximately half of the states have recidivist laws—sometimes called “three strikes” laws—that are similar to the ACCA. See John Clark, James Austin & D. Alan Henry, “Three Strikes and You’re Out”: A Review of State Legislation, National Institute of Justice Research in Brief, National Institute of Justice 7–9 (Sept. 1997); Robert Heglin, Note, A Flurry of Recidivist Legislation Means: “Three Strikes and You’re Out,” 20 J. Legis. 213, 215–16 & nn.18–23 (1994) (collecting state statutes). It also has resulted in the cancellation of deportation orders premised on an immigrant’s conviction of a felony. [13][13]. See e.g., Rendon v. Holder, 764 F.3d 1077, 1090 (9th Cir. 2014); Aguilar-Turcios v. Holder, 740 F.3d 1294, 1302 (9th Cir. 2014); Lima-Lima v. Holder, 545 F. App’x 648, 650 (9th Cir. 2013); Makwana v. Att’y Gen. of the United States, 543 F. App’x 186, 190 (3d Cir. 2013); Donawa v. United States Att’y Gen., 735 F.3d 1275, 1283–84 (11th Cir. 2013).

In the long term, the significance of Descamps could be even greater. It called attention to the statutory specificity that legislators are capable of and the adjudicative clarity that courts can promote. Until now, the Court had done little to encourage either; Descamps provides incentives for both. Already, Descamps is pushing courts to think more carefully about what facts must be established to constitute a particular criminal offense, how such facts are established and recorded in the context of an adjudicative proceeding, and what consequences flow from greater or lesser specificity. [14][14]. See, e.g., Rendon, 764 F.3d at 1084–88 (considering the required elements of a California burglary conviction and the extent of jury unanimity required as to how the offense was committed); United States v. Abbott, 748 F.3d 154, 158–59 (3d Cir. 2014) (same, as to Pennsylvania drug statute); United States v. Royal, 731 F.3d 333, 340–42 (4th Cir. 2014) (same, as to Maryland assault statute); United States v. Cabrera-Gutierrez, 756 F.3d 1125, 1133–37 (9th Cir. 2013) (same, as to Oregon sexual abuse statute). As the contexts that require courts to determine whether a defendant’s prior offense of conviction “counts” multiply, prosecutors and legislators will have increasing reason to respond. And, in the course of establishing a standard for ACCA purposes, the Court deployed a new analytic term—“alternative elements”—that may assist trial courts in another important context that has long needed attention: sorting those facts that jurors must agree upon from those they need not. The Court’s new analytic term also may prompt a critical reevaluation of the doctrine governing claims under the Double Jeopardy Clause. Thus, Descamps is not only important in its own right, but also suggests the possibility of more interesting things to come for the criminal law more broadly.

I. The Descamps Decision

Descamps addressed the question of how federal courts should go about applying the Armed Career Criminal Act (“ACCA”) of 1984, codified at Title 18, United States Code, Section 924(e). Pursuant to the ACCA, the sentence of a defendant who has been convicted of certain federal firearms offenses may be significantly increased if the defendant has three prior convictions for a “violent felony” or a “serious drug offense.” [15][15]. 18 U.S.C. § 924(e)(1) (2012). At issue in Descamps was only the “violent felony” prong of the ACCA. The ACCA defines a violent felony as any felony, whether under state or federal law, that “has as an element the use, attempted use, or threatened use of physical force against the person of another,” [16][16]. 18 U.S.C. § 924(e)(2)(B)(i) (2012). that is “burglary, arson or extortion, involves use of explosives,” [17][17]. 18 U.S.C. § 924(e)(2)(B)(ii) (2012). or that “otherwise involves conduct that presents a serious potential risk of physical injury to another.” [18][18]. Id. Courts frequently refer to these three clauses as the ACCA’s “force” prong, the “enumerated felonies” prong, and the “residual clause,” respectively. See, e.g., United States v. Elliott, 757 F.3d 492, 493 (6th Cir. 2014). The Supreme Court recently ordered new briefing and argument on the question of whether the residual clause is unconstitutionally vague. See Johnson v. United States, No. 13-7120, 2015 WL 132524 (2015) (ordering briefing to be completed by Apr. 10, 2015). If a defendant has three qualifying prior convictions, he or she is considered an Armed Career Criminal for purposes of the ACCA. As a consequence, the defendant will face a mandatory minimum sentence of 15 years and a maximum of life imprisonment. This is generally a much higher sentencing range than would otherwise be applicable. For example, in Descamps’ case, he was convicted in federal court of being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g), an offense which ordinarily carries a statutory maximum penalty of 10 years in prison. [19][19]. 18 U.S.C. § 924(a)(2). Because Descamps was deemed an Armed Career Criminal, however, he was eligible for the ACCA enhancement and was sentenced to 262 months in prison—more than twice the maximum penalty he otherwise could have received. [20][20]. A defendant who is found to be an Armed Career Criminal may also face an enhanced sentencing range under the United States Sentencing Guidelines as a consequence of that designation. See U.S. Sentencing Guidelines Manual § 4B1.4 (2014) (providing a minimum offense level of 33 and a minimum criminal history category of IV for an Armed Career Criminal, and even higher levels for both if certain facts are established). Not surprisingly given these enormous consequences, defendants like Descamps frequently argue that their prior convictions do not count as prior violent felonies for ACCA purposes. [21][21]. Prior to Descamps, the Supreme Court had decided at least nine cases in the previous ten years analyzing what prior convictions count as ACCA predicate felonies. See, e.g., McNeill v. United States, 131 S. Ct. 2218, 2224 (2011) (analyzing when prior state drug convictions count as serious drug offenses for purposes of the ACCA); Sykes v. United States, 131 S. Ct. 2267, 2277 (2011) (analyzing whether a state felony vehicle flight crime qualified as a violent felony under the ACCA’s residual clause); Johnson v. United States, 559 U.S. 133, 145 (2010) (analyzing whether a prior state conviction for battery constituted a violent felony under residual clause); Chambers v. United States, 555 U.S. 122, 130 (2009) (analyzing whether a state conviction for failure to report for periodic incarceration constituted a violent felony under the residual clause); Begay v. United States, 553 U.S. 137, 145 (2008) (analyzing whether a state conviction for felony driving under the influence qualified under the residual clause); United States v. Rodriguez, 553 U.S. 377, 393 (2008) (analyzing whether a state drug conviction counted as a serious drug offense); James v. United States, 550 U.S. 192, 214 (2007) (analyzing whether a state attempted burglary statute qualified under the residual clause); Logan v. United States, 552 U.S. 23, 37 (2007) (considering the implications of a state’s failure to revoke the defendant’s civil rights on whether a conviction counted under the ACCA); Shepard v. United States, 544 U.S. 13, 27 (2005) (setting forth materials that may be consulted when a court applies modified categorical review).

Although the ACCA enumerates certain specific felonies that necessarily count as violent felonies (i.e., burglary, arson, and extortion), it does not define those crimes, and there is a great deal of variation in how the several states define them. [22][22]. The statute originally included a definition of burglary, but Congress deleted this definition in the course of amending the statute in 1986 to expand the scope of the predicate felonies that count for ACCA purposes. Although the deletion appears to have been unintentional, it has never been corrected. See Taylor v. United States, 495 U.S. 575, 588–90 (1990). Thus, in prior decisions interpreting the ACCA, the Supreme Court had construed the statute as incorporating by reference the modern “generic” version of such offenses, “roughly corresponding” [23][23]. See id. at 589. to the definitions used in the majority of states’ criminal codes, and held that only prior convictions falling within the generic version of the offense would count. To determine whether a prior conviction fell within the generic version of the offense, the Court instructed sentencing courts to engage in what the Court called a “categorical approach,” meaning courts should analyze the statutory definition of the offense of conviction, not the underlying facts of the particular case. [24][24]. Descamps, 133 S. Ct. at 2283. If that offense swept no more broadly (or was narrower) than the generic version of the offense, then a conviction for the prior offense categorically would count as an ACCA predicate felony. However, the Court also had recognized that, in certain situations involving a statute that swept more broadly than the generic version of the offense, it might be permissible for a sentencing court to engage in what the Court called “modified categorical review,” meaning it could consult certain materials to determine whether the prior conviction was for conduct falling within the generic version of the offense. At issue in Descamps was precisely when such modified categorical review was appropriate.

Descamps resolved a Circuit split on this issue and came down on the side of those Courts of Appeals that had tied the propriety of modified categorical review to the text of the statute that was the basis for the prior conviction. [25][25]. The Ninth and Sixth Circuits had applied the modified categorical approach to indivisible statutes, as explained infra, whereas the Second and First Circuits had held that the modified categorical approach applied only to divisible statutes. Compare United States v. Descamps, 466 F. App’x 563, 565 (9th Cir. 2012), and United States v. Armstead, 467 F.3d 943, 947–50 (6th Cir. 2006), with United States v. Beardsley, 691 F.3d 252, 268–74 (2d Cir. 2012), and United States v. Giggey, 551 F.3d 27, 40 (1st Cir. 2008) (en banc). The Court held that modified categorical review was permissible only when the prior conviction was pursuant to a “divisible” rather than an “indivisible” statute [26][26]. Descamps, 133 S. Ct. at 2284.—meaning that the statute explicitly set forth “alternative elements,” [27][27]. Id. at 2283–84. some of which fell within the generic offense and others of which did not. An example of a divisible statute would be a statute specifying that a person committed an offense if he unlawfully possessed a “gun, knife, or ax.” An indivisible version of the same offense would be a statute that used the broader term “weapon.”

If the statute were divisible, the sentencing court could consult a limited category of documents such as the indictment, jury instructions, verdict form, plea agreement, and plea colloquy to determine which of the statutory alternative elements provided the basis for the conviction. If the conviction were based on one of the alternatives falling within the generic version of the offense, then the prior conviction could count as a predicate felony for purposes of the ACCA. If the conviction were based on one of the statutory alternatives falling outside the scope of the generic version of the offense (or if the documents did not reveal which statutory alternative provided the basis for conviction), then the prior conviction could not count. However, the Court held that this modified categorical review was never appropriate where the prior offense of conviction was pursuant to an indivisible statute. Faced with a prior offense pursuant to an indivisible statute encompassing conduct falling both within and outside of the generic offense, courts could not consult any underlying materials in the case to determine whether the defendant’s actual conduct that gave rise to the conviction fell within the generic version of the offense. Because the lower courts in Descamps’ case (both the district court and the Ninth Circuit Court of Appeals) had applied the modified categorical approach to a statute that the Supreme Court deemed indivisible, [28][28]. The prior conviction at issue in Descamps’ case was a conviction for burglary pursuant to Cal. Penal Code Ann. § 459 (West 2010), which provided that a “person who enters” certain locations “with intent to commit grand or petit larceny or any felony is guilty of burglary.” The statute did not require that the entry have been unlawful—an element that the Supreme Court deemed a necessary component of the “generic version” of burglary. See Descamps, 133 S. Ct. at 2285, 2294 (Thomas, J., concurring) (citing 3 Wayne LaFave, Substantive Criminal Law § 21.1(a) (2d ed. 2003)). the Supreme Court reversed.

As a consequence of Descamps, the text of the criminal statute that provides the basis for a defendant’s prior conviction assumes paramount importance. Whenever a defendant’s prior conviction is based on a textually indivisible statute that is broader than the generic version of an ACCA felony, that conviction will not count for ACCA purposes, with no exceptions. Accordingly, some repeat, violent offenders will not be eligible for the enhanced penalties provided by the ACCA or other recidivist statutes, even when the underlying record makes clear that the actual offense was of a type that ordinarily would qualify as a predicate felony. Critics of the categorical approach have long complained about this anomalous result, arguing that this kind of formalism allows too many dangerous recidivists to escape the penalties Congress intended. [29][29]. See, e.g., Shepard v. United States, 544 U.S. 13, 35–36 (O’Connor, J., dissenting) (decrying the categorical approach’s “overscrupulous regard for formality,” which “forces the . . . sentencing court to feign agnosticism about clearly knowable facts” and “cannot be squared with the ACCA’s twin goals of incapacitating repeat violent offenders, and of doing so consistently notwithstanding the peculiarities of state law”) (emphasis in original). See also Descamps, 133 S. Ct. at 2302 (Alito, J., dissenting) (“The Court’s holding will . . . frustrate fundamental ACCA objectives” including that “violent, dangerous recidivists would be subject to enhanced penalties” in that “those enhanced penalties would be applied uniformly.”). Whereas some courts, like the Ninth Circuit, had liberally applied the modified categorical approach to get around this result, as set forth below in Part II, Descamps has definitively closed off that option whenever a prior conviction was based on an indivisible statute.

II. Descamps’ Refinement of the Categorical Approach

Justice Kagan’s opinion for the Court (which was joined by six Justices, including the Chief Justice) treated the case as “all but resolve[d]” [30][30]. Descamps, 133 S. Ct. at 2283. by the Court’s prior case law. [31][31]. As Daniel Richman has observed, “[e]ven for those not ordinarily interested in technical sentencing decisions, the Court’s opinion is worth reading for its miffed tone of a well-respected author confronted with an obdurate reader claiming unfamiliarity with her work.” Daniel Richman, Opinion analysis: When is a burglary not a burglary?, SCOTUSblog (June 20, 2013, 11:18 PM), http://www.scotusblog.com/2013/06/opinion-analysis-when-is-a-burglary-not-a-burglary; see also William Baude, Modified Categorical Imperative, PrawfsBlawg (June 21, 2013, 2:09 AM), http://prawfsblawg.blogs.com/prawfsblawg/2013/06/modified-categorical-imperative.html (“If you are interested in sharp-tongued judicial rhetoric, you should be following Justice Kagan” who “positively mocks the Ninth Circuit decision.”). To a certain extent, she was clearly right. In Taylor v. United States, [32][32]. Taylor v. United States, 495 U.S. 575 (1990). decided in 1990, the Court construed the ACCA to hold that the specific offenses referenced in the ACCA’s definition of “violent felony” referred to the “generic” versions of these offenses rather than any offense so titled in a state criminal code. Taylor also established the categorical approach for determining whether a state conviction qualified as a violent felony under the ACCA. The Court held that a sentencing court should look only at the elements of the prior offense of conviction, not the underlying facts of the offense, to determine if it was a violent felony. If the elements of the prior offense swept more broadly than the generic version of the offense, the prior conviction would not count. If the elements of the prior offense were essentially the same as those of the generic offense, or were narrower than the generic version, then the prior conviction would count. [33][33]. Id. at 601–02.

Taylor also nodded to the possibility of what later would be termed the “modified categorical approach.” At the end of the opinion, the Court observed that, in a “narrow range of cases” [34][34]. Id. at 602. a sentencing court might consult materials beyond the statutory definition of a prior offense to determine whether it qualified as a violent felony under the ACCA. Citing the hypothetical example of a state burglary statute that included entry of an automobile as well as a building, the Court opined that the Government should be allowed to use the conviction to enhance a defendant’s sentence under the ACCA where the jury was “actually required to find all of the elements of generic burglary” [35][35]. Id. (i.e., burglary of a building, not an automobile). Thus, if the indictment or information and jury instructions indicated that the defendant was charged solely with burglary of a building, the prior conviction could count as a burglary for ACCA purposes.

In Shepard v. United States, [36][36]. Shepard v. United States, 544 U.S. 13 (2005). decided in 2005, the Court encountered precisely the type of statute alluded to in Taylor—a state burglary statute that covered entries into “boats and cars” as well as “buildings.” Drawing upon the foregoing language from the end of the Taylor decision, Shepard held that, when presented with a prior conviction pursuant to this statute, the sentencing court could consult a limited range of documents to determine whether the defendant had pleaded guilty to the “buildings” prong of the burglary statute—and thus had been convicted of generic burglary (the Court held in Taylor that generic burglary encompassed only burglary of a building)—rather than the other portions of the statute. The permissible class of documents that the sentencing judge could consider included not only charging documents and jury instructions, but also, in the case of a bench trial or a guilty plea, the judge’s formal rulings of law and findings of fact, a plea agreement, or guilty plea colloquy. [37][37]. Id. at 20. Shepard refused to expand the scope of evidentiary material further, however, specifically rejecting the use of a police report even if it had been submitted to the local court in connection with the filing of the original charges. [38][38]. Id. at 21. Thus, Shepard solidified the modified categorical approach, although (as in Taylor) that phrase appears nowhere in the Court’s opinion.

Similarly, in Johnson v. United States, [39][39]. Johnson v. United States, 559 U.S. 133 (2010). decided in 2010, the Court considered whether Florida’s battery statute categorically fell within the ACCA’s definition of a “violent felony,” defined as any crime having “as an element the use . . . of physical force against the person of another.” [40][40]. 18 U.S.C. § 924(e)(1) (2006). The Florida statute did not qualify because the Court interpreted the term “physical force” in the ACCA as requiring the application of violence and the Florida statute could be satisfied without a showing of violence. Nevertheless, reaffirming Shepard, the Court held that a sentencing court confronted with a law containing “statutory phrases that cover several different generic crimes[,] some of which require violent force, and some of which do not,” could apply the modified categorical approach “to determine which statutory phrase was the basis for the conviction.” [41][41]. Johnson, 559 U.S. at 144. In Johnson, the Court finally used the phrase “modified categorical approach,” quoting its decision in Nijhawan v. Holder, 557 U.S. 29 (2009), an immigration case in which the Court considered the application of the Taylor line of cases to the construction of certain felonies under the Immigration and Nationality Act (INA) that could provide the basis for deportation. Id. In Nijhawan, the Court ultimately rejected the use of the categorical approach for purposes of this section of the INA, holding that it was not consistent with Congressional intent. Nijhawan, 557 U.S. at 39–40. However, the Court has held that the categorical approach does apply with respect to certain other portions of the INA. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 194 (2007) (endorsing the categorical approach to interpret the term “theft offense” in the INA). Because the Florida statute was such a statute, the modified categorical approach was appropriate. [42][42]. Johnson represented the usual case in which the Government advocated for categorical review rather than modified categorical review, because the underlying state documents were inconclusive. The Government argued that the Florida statute categorically qualified as a violent felony under the ACCA, thus obviating the need for resort to the underlying documents. Johnson, 558 U.S. at 139. The Court disagreed. Id.

Thus, well before the Ninth Circuit upheld Descamps’ sentence in 2012, [43][43]. United States v. Descamps, 466 F. App’x 563 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013). it was clear—pursuant to the foregoing precedents—that a sentencing court generally should apply the categorical approach, focusing on the elements of the prior offense, to determine whether a defendant’s prior conviction qualified as a predicate violent felony under the ACCA. It was also clear that, in certain limited circumstances involving certain types of statutes, a court could apply the “modified categorical approach” and consult a limited range of documents to determine if the defendant’s prior conviction necessarily was pursuant to a particular portion of the statute that fell within the generic version of the offense. What was not completely clear—notwithstanding Justice Kagan’s assertion that the case was “all but resolved” by prior case law—was how a court should go about deciding whether a particular statute was of the type that permitted further inquiry.

Although the Court in prior decisions had suggested, mostly by way of the examples that it used, that statutory text was critical (and in Johnson, as noted supra, the Court described the sentencing judge’s objective in conducting modified categorical review as determining “which statutory phrase was the basis for conviction”), [44][44]. Johnson, 559 U.S. at 144 (emphasis added). in Descamps the Court for the first time made clear that text was decisive. For the first time, the Court used the terms “divisible statutes” and “indivisible statutes” to describe the two universes of criminal statutes that existed, as though there were no other possibilities and no overlap between the two. The Court also used the phrase “alternative elements” for the first time, to describe the distinguishing characteristic of a “divisible” statute. That is, a “divisible” statute has at least one element that is capable of being satisfied in more than one way and explicitly sets forth in its text what those alternatives are—such as a burglary statute that specifies that breaking and entering of a “building or automobile” will suffice, or a weapon possession statute that specifically applies to possession of a “gun, knife, or ax.”

The Ninth Circuit’s mistake, the Court wrote, was in trivializing the distinction between divisible and indivisible statutes on the theory that any indivisible statute could “be imaginatively reconstructed as a divisible one” [45][45]. Descamps v. United States, 133 S. Ct. 2276, 2289 (2013). and therefore that there was no meaningful difference between the two. For the Ninth Circuit, the only conceptual difference was that “[a divisible statute] creates an explicitly finite list of possible means of commission, while [an indivisible one] creates an implied list of every means of commission that otherwise fits the definition of a given crime.” [46][46]. Id. (quoting United States v. Aguila-Montes de Oca, 655 F.3d 915, 927 (9th Cir. 2011)) (alteration in original). The Ninth Circuit viewed this difference as an insufficient reason to allow modified categorical review in the case of a divisible statute but disallow it for an indivisible one. If the point was to identify the underlying basis for the defendant’s prior conviction, and thereby to determine if he qualified as an Armed Career Criminal, why not allow the sentencing court to engage in such review in both categories of cases? The Supreme Court’s answer: allowing judges to do so would raise serious constitutional concerns under the Sixth Amendment and Apprendi. [47][47]. Id. at 2288 (citing Apprendi v. New Jersey, 530 U.S. 466 (2000)).

Citing the “categorical approach’s Sixth Amendment underpinnings,” [48][48]. Id. the Court explained that a sentencing judge could enhance a defendant’s sentence based on the existence of a prior conviction without running afoul of the Sixth Amendment’s jury trial guarantee. [49][49]. Id. (citing Apprendi, 530 U.S. at 490). In Almendarez-Torres v. United States, 523 U.S. 224 (1998), the Supreme Court held that the existence of a prior conviction was a sentencing factor rather than an element and, accordingly, did not need to be included in an indictment or found by a jury. Post-Apprendi, the Court consistently has recognized Almendarez-Torres as an exception to the Apprendi line of cases, notwithstanding repeated calls from various quarters (including from within the Court, most notably from Justice Thomas) to overrule it. Descamps, 133 S. Ct. at 2294–95 (Thomas, J., concurring); James v. United States, 550 U.S. 192, 231 (2007) (Thomas, J., dissenting); Shepard v. United States, 544 U.S. 13, 27 (2005) (Thomas, J., concurring in part and concurring in the judgment). Descamps argued in the lower courts that Almendarez-Torres should be overruled and sought certiorari on that issue; the Court denied that portion of his petition. United States v. Descamps, 466 F. App’x 563, 564–65 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013). But allowing a sentencing judge to make a finding as to the nature of the prior conviction under an indivisible statute was fraught. Allowing modified categorical review in cases involving divisible statutes was consistent with Apprendi because “only divisible statutes enable a sentencing court to conclude that a jury (or judge at a plea hearing) has convicted the defendant of every element of the generic crime.” [50][50]. Descamps, 133 S. Ct. at 2290. Divisible statutes, the Court reasoned, require the prosecution to proceed, and the fact-finder to enter a decision, in a manner that together renders transparent whether or not the defendant was convicted pursuant to a statutory alternative element falling within the generic version of the offense. Indivisible statutes, by contrast, do not require such focused allegations or deliberations.

III. Implications of Descamps

Descamps makes a persuasive case for distinguishing between divisible and indivisible statutes for Sixth Amendment purposes—but only if it accurately describes how divisible statutes are charged and submitted to juries. Yet as set forth below, that is not necessarily the case. What remains to be seen is whether the decision will prompt changes in how such cases are in fact charged and litigated, and whether the Court’s use of the term “alternative elements” will have broader doctrinal implications.

A. Practical Implications

1. Single Theory Charges.

To begin with the simplest category of cases in which Descamps’ model does not accurately describe actual practices, consider again the hypothetical “gun, knife, or ax” statute. An indictment alleging a single charge of violating this divisible statute might very well track the statute and include all three of the alternative elements in the charging language, specifying only in the “to wit” clause at the end of the charging paragraph that the defendant possessed a gun, in order to give the defendant notice of the specific factual allegation underlying the charge. [51][51]. See Wayne LaFave, Jerold H. Israel, Nancy King & Orin Kerr, Criminal Procedure § 19.3(b) (5th ed. 2009 & Supp. 2013). Depending on the practice in that jurisdiction, if no evidence was presented at the trial that the defendant possessed any weapon other than a gun, the court might instruct the jury at the end of the case that, in order to convict, it must find that the defendant possessed a gun—thus following Justice Kagan’s model. But, alternatively, the judge might instruct the jury that it may convict if it finds that the defendant possessed “a gun, knife, or ax.” Either way, the verdict form likely will call for a single response—“guilty” or “not guilty” as to the single count in the indictment.

In the second scenario, is it clear that the jury necessarily found that the defendant possessed a gun? Not really—no more so than the answer “yes” to the question “would you like coffee or tea?” indicates which beverage the respondent prefers. Thus, notwithstanding the fact that the charge was pursuant to a divisible statute (and that everyone involved in the trial likely would be confident that the jury did find that the defendant possessed a gun), a subsequent sentencing judge would not be able to determine with confidence through modified categorical review that the jury necessarily convicted the defendant of a generic firearms offense. The “Descamps fix” for such a scenario is straightforward and essentially cost-free to the prosecution: streamline the indictment and seek a jury instruction mentioning only the gun.

2. Multiple Theory Charges.

The next category of case is not so simple. Consider a hypothetical case where there is evidence that the defendant possessed both a gun and a knife. Perhaps at the time of his arrest, he had a gun in his pocket and a knife in his backpack. The government alleges a single count of violating the “gun, knife, or ax” statute, and specifies in the “to wit” clause of the indictment that the defendant possessed both a gun and a knife. [52][52]. See Fed. R. Crim. P. 7(c)(1) (providing that a single count of an indictment may allege that “the means by which the defendant committed the offense are unknown or that the defendant committed it by one or more specified means”); Griffin v. United States, 502 U.S. 46, 51 (1991) (“A statute often makes punishable the doing of one thing or another . . . sometimes thus specifying a considerable number of things . . . . [T]he indictment on such a statute may allege, in a single count, that the defendant did as many of the forbidden things as the pleader chooses, employing the conjunction and where the statute has ‘or,’ . . . and it will be established at the trial by proof of any one of them.” (quoting 1 Joel Prentiss Bishop, New Criminal Procedure § 436, at 355–56 (2d ed. 1913))) (internal footnotes omitted) (internal quotation marks omitted) (emphasis omitted). Suppose some evidence is presented at trial that the defendant did not know that the knife was in the backpack (e.g., a friend asked him to carry the backpack) and some evidence suggesting that the gun in the defendant’s pocket does not meet the statutory definition of a gun (e.g., perhaps it was an antique). In other words, assuming that the statute requires knowing possession, there are possible grounds for reasonable doubt as to each potential basis for conviction. Some jurors might conclude that the defendant violated the statute by possessing a gun but not a knife, while others might reach the opposite conclusion.

If the jury were instructed that it might convict if it finds that the defendant possessed a “gun, knife, or ax,” would the defendant who is convicted of such a charge necessarily have received a jury finding that he possessed a gun? Plainly, the answer is no. Again, the charge would have been pursuant to a divisible statute, but even with modified categorical review, a subsequent sentencing judge would not be able to determine that the jury necessarily found that the defendant possessed a gun. Here, the “Descamps fix” is not so straightforward, nor is it costless to the prosecution. Should each theory of the case be alleged in a separate count of the indictment? Should the two theories remain in a single count but the jury nevertheless be instructed that it must agree unanimously (or by the requisite majority in the jurisdiction) on a single theory before it may convict? If the latter, should the verdict form require the jury to specify the prevailing theories or theories? Justice Alito called attention to some of these practical difficulties with operationalizing Descamps, [53][53]. See Descamps, 133 S. Ct. at 2298 (Alito, J., dissenting) (envisioning the case of a defendant charged with burglarizing a “floating home” pursuant to a state statute defining burglary as including unlawful entry into, inter alia, a “building” or “vessel” and asking whether, in order to convict, the jury would have to agree “whether this structure was a ‘building’ or a ‘vessel.’ . . . The Court’s answer is ‘yes.’”). Justice Alito also notes that, even if unanimity were required, the basis for the conviction would not be apparent to a subsequent sentencing court unless the jury returned “a special verdict, something that is not generally favored in criminal cases.” Id. at 2300 (citing 6 Wayne LaFave & Jerold H. Israel, Criminal Procedure § 24.10(a), at 543–44 (3d ed. 2007)). but the majority glossed over them entirely.

This second category of case is not so unusual. Under prevailing doctrine, alternative theories may be combined in a single charge, and a guilty verdict returned based on a combination of theories, if each alternative theory is properly characterized as a “means” of committing a single offense, rather than an “element” signifying a distinct crime. The Court’s two leading cases on the subject are Schad v. Arizona [54][54]. Schad v. Arizona, 501 U.S. 624 (1991). from 1991 and Richardson v. United States [55][55]. Richardson v. United States, 526 U.S. 813 (1999). from 1999. Neither opinion offers particularly clear guidance to courts as to how they should distinguish between means and elements. [56][56]. For a discussion of the indeterminacy of Schad and Richardson, see, for example, Scott W. Howe, Jury Fact-Finding in Criminal Cases: Constitutional Limits on Factual Disagreements Among Convicting Jurors, 58 Mo. L. Rev. 1, 6 (1993); Brian M. Morris, Something Upon Which We Can All Agree: Requiring a Unanimous Jury Verdict in Criminal Cases, 62 Mont. L. Rev. 1, 19–24, 31–34 (2001); Jessica A. Roth, Alternative Elements, 59 UCLA L. Rev. 170, 186–97 (2011); Stephen E. Sachs, Alternative Theories of the Crime 5–6 (May 24, 2007) (unpublished manuscript) (http://ssrn.com/abstract=1501628); Peter Westen & Eric Ow, Reaching Agreement on When Jurors Must Agree, 10 New Crim. L. Rev. 153, 153–54 (2007); see also United States v. Hurt, 527 F.3d 1347, 1356 (D.C. Cir. 2008) (observing that Schad does not provide clear guidance). What the Court has said is that legislative intent is the controlling question in determining whether a statute sets forth alternative means as opposed to elements, subject only to the barest of review for compliance with the Due Process Clause. But legislative intent in this regard (as in many others) is frequently difficult to discern. [57][57]. The Court has not provided any particularized guidance as to the appropriate methodology in determining whether a legislature intended to create means as opposed to elements. In Schad, however, the Court did cite approvingly to United States v. UCO Oil Co., 546 F.2d 833, 835–38 (9th Cir. 1976), in which the court stated that statutory language, legislative history, the nature of the proscribed conduct, and the appropriateness of multiple punishments all should bear on whether a statute should be interpreted as evidencing a legislative intent to create multiple crimes as opposed to a single crime with multiple means of commission. See Schad, 501 U.S. at 636 (citing UCO Oil Co., 546 F.2d at 835–38). The Court has never set forth a definitive test for determining when a statute has impermissibly combined as putative “means” alternatives that must, consistent with due process, be considered “elements.” [58][58]. The Court did, however, suggest certain factors that would be relevant to such a determination, including the moral equivalency of the various alternatives, whether historically they were treated as discrete crimes, and the practices of other states. For example, in Schad, the plurality opinion stated that it would be unconstitutional for a state to recast its entire criminal code as a single statute and declare each formerly separate crime a “means” of committing the offense of “crime.” Schad, 501 U.S. at 632–33. Nor has it ever held a particular statute unconstitutional on that basis. Consistent with such a permissive approach, state and federal courts have repeatedly cited Schad and Richardson in holding that a criminal statute sets forth means rather than elements. [59][59]. See, e.g., United States v. Seher, 562 F.3d 1344, 1361–62 (11th Cir. 2009) (holding that a jury need not agree on a defendant’s motive for engaging in a financial transaction, as set forth in the federal money laundering statute); United States v. Stewart, 433 F.3d 273, 319 (2d Cir. 2006) (holding that a jury need not agree on whether a defendant made a false statement or concealed a material fact, as set forth in the false statements statute); United States v. Powell, 226 F.3d 1181, 1196 (10th Cir. 2000) (holding that a jury need not agree on which verb characterizes a defendant’s conduct, as set forth in the federal kidnapping statute). See also 6 LaFave et al., supra note 53, at § 24.10(c), nn.40–41 (collecting additional state and federal cases involving multiple theories and in which courts have approved general verdicts without requiring juror agreement as to the theory). That is why, when the evidence supports them, prosecutors routinely file charges pursuant to such statutes alleging multiple theories within one count of an indictment, and jurors in such cases are not required to agree as to the means of commission—i.e., the jury can return a “patchwork” [60][60]. See Hayden J. Trubitt, Patchwork Verdicts, Different-Juror Verdicts, and American Jury Theory: Whether Verdicts are Invalidated by Juror Disagreement on Issues, 36 Okla. L. Rev. 473, 474–77 (1983) (describing how jury selection and individual jury behavior can lead to compromises that actually stymie true deliberation and discussion among jurors). or “composite” verdict. In other words, current constitutional law fosters precisely the type of adjudicative uncertainty that the Descamps Court says divisible statutes can prevent.

Under the reasoning of Descamps, a prior conviction pursuant to such a charge could qualify as a predicate felony for purposes of the ACCA in two ways: first, if each alleged theory of commission met the ACCA definition; or second, if one or more alleged theories did not fall with the ACCA definition (e.g., a gun would qualify as a “firearm” but a knife would not), if the jury was required to return a separate verdict as to each theory and found the defendant guilty pursuant to at least one of the theories meeting the ACCA definition. Thus, prosecutors may be forced to choose between, on the one hand, the short-term strategic advantage of obtaining a “composite” guilty verdict; and, on the other hand, the theoretical benefit—likely to be realized only in the future, if at all—of tagging the defendant with a conviction sure to count under the ACCA.

How prosecutors would weigh the risks and benefits of each choice would vary depending on the individual case. Generally speaking, it seems unlikely that a local prosecutor would give much thought to whether a conviction will count for ACCA purposes. But by introducing this new dynamic, Descamps may result in changes at the institutional level in how criminal charges based on multiple theories are tried. Post-Descamps, one reasonably might expect to see prosecutors’ offices, courts, and even defense counsel [61][61]. Ex ante, adjudicative clarity is a protection for defendants. Ex post, it has positive externalities for those, generally prosecutors, who want to piggyback off the prior adjudicated fact. It is perhaps not surprising then that an Internet search for materials citing Descamps reveals a number of training materials put together by defense attorneys recommending that defenders representing ACCA defendants at sentencing argue that their clients’ prior convictions were pursuant to statutory phrases that represent alternative means rather than alternative elements. See, e.g., Paresh Patel & Lisa Freeland, Post-Descamps World, http://agendas.fjc.gov/agenda/files/FedDef1410-PostDescampsWorld.pdf (last visited Jan. 26, 2015). becoming more comfortable with specific unanimity instructions and embracing more precise verdict forms. [62][62]. Even putting aside the ACCA, there are institutional reasons why prosecutors would be in favor of more specialized verdicts. See Roth, supra note 56, at 213–14 (explaining how prosecutors sometimes benefit from seeking greater clarity in the basis for a jury’s verdict). Congress could amend the ACCA to make clear that “conviction” refers to underlying conduct, not the elements of the prior offense. In his concurring opinion in Descamps, Justice Kennedy called on Congress to make such a change. Descamps v. United States, 133 S. Ct. 2276, 2294 (2013) (Kennedy, J., concurring). But it is not clear that such a change would be consistent with the Sixth Amendment. See id. at 2288 (explaining that if a court’s finding of a predicate offense went beyond merely identifying a prior conviction it would raise Sixth Amendment concerns). Assuming that the Court means what it says when it cites the categorical approach’s Sixth Amendment underpinnings in the context of criminal sentencing, a circumstance-based approach would seem to present serious constitutional problems. Courts have suggested for years that such practices would be desirable but have been reluctant to mandate them. [63][63]. See, e.g., Griffin v. United States, 502 U.S. 46, 60 (1991); id. at 61 (Blackmun, J., concurring); Schad, 501 U.S. at 645. As the number of contexts increases in which what facts necessarily were decided in a prior proceeding is of great consequence—such as for purposes of state recidivist statutes, sentencing guidelines, and deportation proceedings—prosecutors (and legislators) will have increasing reasons to respond to Descamps’ demand for greater clarity.

Formally, Descamps does not revisit the Court’s “means” versus “element” jurisprudence from Schad and Richardson. As set forth above, under Descamps, a prosecutor may have to choose between a composite verdict and a sure ACCA predicate felony, but the decision says nothing about the constitutionality of opting for the former at the expense of the latter. And yet, by implication, Descamps suggests that Schad’s influence may be on its way out—not just as a practical matter, but doctrinally as well. [64][64]. Schad was decided by a 5-4 vote. Justice Souter wrote the main opinion for the Court for a plurality of the Justices (joined by Chief Justice Rehnquist and Associate Justices O’Connor and Kennedy). Schad, 501 U.S. at 624. Justice Scalia provided the crucial fifth vote to affirm, but made clear in a separate opinion that, in any other case, he might well find that jury unanimity was constitutionally required. Id. at 651 (Scalia, J., dissenting). Richardson, which found elements rather than means—citing constitutional concerns—was written by Justice Breyer, joined by Chief Justice Rehnquist and Justices Stevens, Scalia, Souter, and Thomas. Richardson v. United States, 526 U.S. 813, 814 (1998). Justice Kennedy dissented in Richardson, joined by Justices O’Connor and Ginsburg. Id. at 825 (Kennedy, J., dissenting). Consistent with his position in Descamps, then-Judge Alito dissented in a case decided by the Third Circuit involving the same issue as Richardson. See United States v. Edmonds, 80 F.3d 810, 829 (3d Cir. 1996) (Alito, J., dissenting). Put together with the line-up in Descamps, there now appear to be six members of the Court who might vote to require jury unanimity as to at least one statutory alternative element (Justices Kagan, Breyer, Stevens, Scalia, Thomas, and Chief Justice Roberts) and three Justices who almost certainly would take the opposite view (Justices Kennedy, Ginsburg, and Alito). Indeed, the majority opinion acknowledges the means versus element distinction but it cites only to Richardson—where the Court found elements rather than means [65][65]. See supra notes 63–64.—and not to Schad. [66][66]. Descamps, 133 S. Ct. at 2288. In addition to the majority’s choice to ignore Schad, the strongest signal of a doctrinal shift is the Court’s decision to use the phrase “alternative elements” throughout the Descamps opinion. This is a phrase the Court had never before used in a criminal opinion. [67][67]. The phrase had been used in the lower courts opinions discussing the application of the ACCA. See, e.g., United States v. Aguila-Montes de Oca, 655 F.3d 915, 963 (9th Cir. 2011); United States v. Vann, 660 F.3d 771, 789 (4th Cir. 2011). When viewed against the backdrop of Schad and Richardson (both of which were brought to the majority’s attention by Justice Alito), the choice seems significant. In Schad and Richardson, the Court painstakingly referred to statutory provisions as either “elements” or “means,” cognizant (and wary) of the consequences of characterizing a provision as the former rather than the latter. By contrast, in Descamps, the Court seems far less concerned about broadly characterizing statutory provisions as “elements.” Indeed, Descamps seems to be utilizing an implicit default rule that divisible statutes set forth “alternative elements” rather than “means.” [68][68]. See Descamps, 133 S. Ct. at 2285 n.2 (expressing surprise that the state statutes analyzed in Taylor, Shepard, and Johnson could be construed as setting forth anything but “alternative elements,” and therefore “several different . . . crimes”). The Court’s approach in determining what constitutes a divisible statute is quite formalistic. When one bears down and looks at the particular statutes involved in the Court’s ACCA cases, divisibility turns on the presence of commas and the word “or” between phrases.

B. Broader Doctrinal Implications

Going forward, the question is whether the Court will stand by this formalistic approach and its description of the different statutory phrases in divisible statutes as alternative elements in other contexts. The Court might explain in a future case that, when it referred to the statutes in Taylor, Shepard, and Johnson as including “alternative elements,” indicative of “separate crimes,” [69][69]. Id.; see id. at 2297–98 (Alito, J., dissenting) (“[L]egislatures frequently enumerate alternative means of committing a crime without intending to define separate elements or separate crimes.” (citing Schad v. Arizona, 501 U.S. 624, 636 (1991))). it did not really mean that in the sense that juries would have to agree on at least one alternative element, [70][70]. The closest the Court came to dealing with this issue in Descamps was to acknowledge Justice Alito’s point that some divisible statutes might set forth alternative means as opposed to elements. Id. at 2285 n.2. The Court expressed surprise that any of the divisible statutes in its ACCA cases to date established means as opposed to elements, and asserted that any confusion as to what constituted elements as opposed to means in future cases could be clarified by consulting the indictment and jury instructions. Id. or that a defendant could be convicted of two versions of the same offense, each containing a different alternative element. But this would require some explaining. Generally speaking, we do not think of crimes as having different elements for different purposes. We may not have a satisfactory a priori definition of what constitutes an element, [71][71]. See 6 LaFave et al., supra note 53, at § 24.10(c) (observing that the Court has never “derive[d] a uniform theory for identifying which facts are elements”); Nancy J. King & Susan R. Klein, Essential Elements, 54 Vand. L. Rev. 1467, 1497 (2001) (ultimately rejecting as “unwise” and “unattainable” the “theoretical quest for the elemental holy grail”). but we know what consequences flow from calling something an element. Elements must be alleged in the indictment. [72][72]. Hamling v. United States, 418 U.S. 87, 117 (1974). Elements must be proven to the jury (by the government) beyond a reasonable doubt. [73][73]. See Patterson v. New York, 432 U.S. 197, 210 (1977) (holding that the legislature cannot shift the burden to the defendant to disprove any element of the offense); In re Winship, 397 U.S. 358, 364 (1969) (holding that “the Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged”). Under the familiar Blockburger [74][74]. Blockburger v. United States, 284 U.S. 299 (1932). formulation, elements are facts that distinguish one offense from another, such that a defendant generally may be charged with both offenses without violating the Double Jeopardy Clause of the Fifth Amendment. [75][75]. Id. at 304. The question for future cases will be, “is an ‘alternative element’ just like a regular element for these purposes, or is it something else?”

1. Jury Unanimity.

In the context of jury unanimity, there is a good argument that alternative elements should be treated just like regular elements. After all, as the Court makes clear in Descamps, text matters. There is a difference between an indivisible statute that uses the term “weapon” and a divisible one that instead uses the list “gun, knife, or ax.” They reflect different legislative choices. The legislature that enacts the former statute enjoys the potential benefit of extending the statute’s reach to new types of weapons that the legislators might not have thought to enumerate. It also need not worry about whether the jury can agree on the type of weapon involved in any particular case. In doing so, it also takes a risk: it cannot be certain whether the court will interpret the statute as applying to a particular device. The legislature that enacts the latter type of statute exerts more control from the outset: it can be assured that the particular devices enumerated in the statutory text will be covered and others will not. [76][76]. These are the familiar tradeoffs of generality versus specificity, standards versus rules. See, e.g., John F. Manning, Second-Generation Textualism, 98 Calif. L. Rev. 1287, 1316 (2010); Frank H. Easterbrook, Statutes’ Domains, 50 U. Chi. L. Rev. 533, 546–47 (1983). (Perhaps the nunchuck lobby was influential in that debate). In the place of a single broad element, the legislature has employed a series of alternative, narrow elements, and the government must establish at least one of them before a conviction will be proper. The two different statutes are not equivalent.

This is the flipside of what the Descamps Court was getting at when it explained how the Ninth Circuit had erred in failing to accord significance to the difference between divisible and indivisible statutes on the theory that all indivisible statutes could be “hypothetically reconceive[d]” in “divisible terms” with a court supplying the missing “implied list of every means of commission that otherwise fits the definition of a given crime.” [77][77]. Descamps v. United States, 133 S. Ct. 2276, 2290 (2013). As the Court stated in Descamps, “the thing about hypothetical lists is that they are, well, hypothetical.” [78][78]. Id. at 2289, 2290. Conversely, when the legislature has enacted a list, it does not matter that the entire list hypothetically could be generalized into one broader term: at least one item on the list must be established beyond a reasonable doubt or else the requisite element has not been established. And the only way to be sure of that is to apply the familiar mechanism: juror agreement beyond a reasonable doubt. Requiring jury agreement as to at least one alternative element would be more consistent with the Court’s renewed focus, since Apprendi, on the structural role of the jury as the bulwark between the accused and the State than the contrary approach, which runs the risk of subverting the jury’s checking function. The Court seemed to be moving in this direction in Richardson, which was decided one year prior to Apprendi. Descamps may prove to be the bridge that helps complete this arc.

2. Double Jeopardy.

Descamps’ embrace of alternative elements may also lead to interesting developments in the Double Jeopardy context. If alternative elements were, in fact, indicative of separate offenses, as the Court indicated, that would suggest pursuant to Blockburger that a defendant could be punished consecutively for committing an offense pursuant to two different alternative elements. [79][79]. See United States v. Dixon, 509 U.S. 688, 696 (1993) (reaffirming the Blockburger test for constitutionality of both multiple punishments and multiple prosecutions). It would also seem to indicate that the government could prosecute a defendant by charging one statutory alternative element in one proceeding and then, especially if the government prevailed at trial, charging another alternative in a subsequent proceeding. [80][80]. Where the defendant was acquitted at a first trial, collateral estoppel may prevent the government from retrying the case, even pursuant to a different alternative element. See Dixon, 509 U.S. at 710 n.15; Ashe v. Swenson, 397 U.S. 436, 443 (1970); Nancy J. King, Portioning Punishment: Constitutional Limits on Successive and Excessive Penalties, 144 U. Pa. L. Rev. 101, 131–32 (1995). If a defendant pleaded guilty pursuant to a plea agreement, a defendant may enjoy additional protection from subsequent prosecution under the terms of the agreement, express or implied. See Santobello v. New York, 404 U.S. 257, 262–63 (1971); Daniel C. Richman, Bargaining About Future Jeopardy, 49 Vand. L. Rev. 1181, 1211–15 (1996). In this area, a theory of the jury’s function does not provide guidance; instead, one needs a theory of counts and of the purposes of the Double Jeopardy Clause. Both areas are excellent candidates for renewed focus in a post-Apprendi era. [81][81]. Many scholars have pointed out the unsatisfactory state of the Court’s current Double Jeopardy jurisprudence. See, e.g., Carissa Byrne Hessick & F. Andrew Hessick, Double Jeopardy as a Limit on Punishment, 97 Cornell L. Rev. 45, 54–55 (2011) (arguing that the Court’s Double Jeopardy jurisprudence “has largely failed to provide any real protection against multiple punishments”); Anne Bowen Poulin, Double Jeopardy and Multiple Punishment: Cutting the Gordian Knot, 77 U. Colo. L. Rev. 595, 599 (2006) (arguing that Double Jeopardy jurisprudence has become confused because of the erroneous entangling of multiple punishment and successive prosecution concerns); Jacqueline E. Ross, Damned Under Many Headings, the Problem of Multiple Punishment, 29 Am. J. Crim. L. 245, 251–53 (2002) (arguing that the Court has focused its Double Jeopardy jurisprudence on illusory problems while ignoring some real ones). The multiple-punishment problem was effectively suppressed through the United States Sentencing Guidelines and similar mandatory state sentencing regimes prior to Apprendi and Booker. Such schemes generally blunted the punitive impact of additional related counts. [82][82]. See generally Ross, supra note 81, at 253–57 (explaining how the Guidelines prevent multiple punishments for additional related counts). But in the post-Apprendi, post-Booker world of nonbinding sentencing guidelines, the problem resurfaces. Absent a constraining theory of counts, prosecutors wield enormous power over a defendant’s maximum sentence (and hence obtain additional leverage to induce a guilty plea) by deciding how many charges to include in a charging instrument. Those same charging decisions also can determine whether a defendant may be exposed to a subsequent prosecution premised on distinct charges.

Should a case arise challenging multiple punishments or multiple prosecutions based on Descamps, one resolution available to the Court (to resolve the challenge in the defendant’s favor) is to emphasize the alternative in “alternative elements.” The Court could point to the language in the opinion where it described alternative elements as creating different “versions” [83][83]. Descamps v. United States, 133 S. Ct. 2276, 2284 (2013) (describing the Court’s task in Shepard of determining which “version” of the Massachusetts burglary statute was the basis for the defendant’s conviction). of the offense and deemphasize the (more oft-repeated) language where it referred to alternative elements as signaling different crimes. But the more interesting course would be for the Court to stick to its characterization of alternative elements as elements indicative of different crimes—and then grapple anew with whether and when multiple punishments and multiple prosecutions are proper when they are sought pursuant to different offenses. In this context, as in the ACCA context, the Court has embraced a categorical approach under Blockburger, examining the elements of the offense rather than the underlying conduct involved in a particular crime (although it has periodically flirted with a conduct-based approach).

As the Court made clear in Descamps, the categorical approach has Sixth Amendment underpinnings, based on the jury trial right, in the context of the ACCA. In the context of Double Jeopardy, it is less clear what the rationale is for the categorical approach. It is arguably more easily administered than a conduct-based approach, and the Court has observed that it has a constitutional pedigree superior to that of a conduct-based approach. [84][84]. See Dixon, 509 U.S. at 704–11. In this context, however, the categorical approach leaves prosecutors’ power in stacking punishments and prosecutions largely unchecked. Thus, one open question after Descamps is whether the decision will amplify prosecutors’ power in this regard by increasing the number of offenses available to them when they charge a defendant, or conversely whether it will lead to a renewed examination of the Court’s Double Jeopardy jurisprudence. [85][85]. Another example of Descamps’ possible application would be in the context of the Sixth Amendment right to counsel. The Court has held that this right is “offense-specific,” and has adopted a categorical approach, using Blockburger’s elements test, in determining whether a defendant’s Sixth Amendment right to counsel has attached as to a particular offense at a particular point in time. See Texas v. Cobb, 532 U.S. 162, 173 (2001) (holding that “when the Sixth Amendment right to counsel attaches, it does encompass offenses that, even if not formally charged, would be considered the same offense under the Blockburger test”).

Conclusion

The Supreme Court’s decision in Descamps v. United States is the Court’s most recent opinion analyzing the application of the frequently litigated Armed Career Criminal Act. Drawing upon its prior decisions in this area, the Court reiterated that sentencing courts should apply a categorical approach in determining whether a defendant’s prior conviction constitutes a prior felony for purposes of the ACCA—meaning that courts should look at the elements of the prior offense of conviction rather than the underlying conduct. The Court also reiterated that courts may apply modified categorical review, and consult certain underlying judicial materials, to determine the basis for a conviction in certain circumstances. Descamps clarified the circumstances that permit modified categorical review. The Court held that such review is appropriate when a defendant’s prior conviction was pursuant to a divisible statute, i.e., one explicitly setting forth alternative elements, some of which fall within the generic version of the felony encompassed by the ACCA, and others of which do not. By applying modified categorical review, a sentencing court can determine which statutory alternative element provided the basis for the conviction.

In the near term, Descamps will affect the sentencing of thousands of criminal defendants, making it harder for the government to establish that a defendant has the requisite predicate felonies under the ACCA or other sentencing regimes pegged to the ACCA’s definitions of prior felonies. Many defendants (like Descamps) will have been convicted pursuant to indivisible statutes that are broader than the generic versions of the offenses incorporated into the ACCA. As to those defendants, the government effectively will be unable to establish the necessary predicate felonies for the ACCA. This is Descamps’ immediate import and, standing alone, it makes the decision quite significant. However, Descamps also has potential longer-term implications that, while less certain, are more profound. Descamps suggests that statutes like the ACCA can, and in this Court’s hands may, push us to think more carefully about what facts are required to establish a particular offense and how those facts are established and recorded in the context of an adjudicatory proceeding. In doing so, Descamps highlights not just the statutory specificity legislators are capable of, but also the adjudicative clarity that courts can promote.

The local prosecutors who handle most of the cases in the United States are unlikely to give much thought to whether a conviction will “count” for ACCA purposes in some later federal prosecution. But the clarity Descamps demands about what a defendant actually did should influence judicial retrospection in other contexts too. As the contexts multiply in which the precise basis for a prior conviction matters (and they are multiplying), prosecutors and legislators will have increased incentives to devise ways of making such a showing. Thus, we may see a change in how criminal statutes are drafted and how they are charged and submitted to juries. And, in the course of establishing a standard for ACCA purposes, the Court deployed a new analytic term—alternative elements—that may assist trial courts in another troubled context: sorting through what jurors need to agree on (elements) and what they need not (means). We have comparatively few criminal trials, but those we do have can and should do far more work at identifying what a jury of the defendant’s peers actually agreed upon. Descamps provides a framework that may help achieve that goal.


Copyright © 2015 Jessica A. Roth.

Associate Professor of Law, Benjamin N. Cardozo School of Law.

[1] Alleyne v. United States, 133 S. Ct. 2151 (2013).

[2] Apprendi v. New Jersey, 530 U.S. 466 (2000).

[3] Harris v. United States, 536 U.S. 545 (2002).

[4] Descamps v. United States, 133 S. Ct. 2276 (2013).

[5] While I am not the first to note Descamps’ significance, this Essay is the first scholarly work to explore it fully. See Douglas Berman, SCOTUS Wraps Its Sentencing Docket With Another Defense Win (and Alito Dissent) In Descamps, Sentencing Law and Policy (June 20, 2013 10:11 AM), http://sentencing.typepad.com/sentencing_law_and_policy/2013/06/scotus-wraps-its-sentencing-docket-with-another-defense-win-and-alito-dissent-in-descamps.html (noting the possibility that “Descamps will prove to be the most consequential of all the Supreme Court’s criminal sentencing work this Term”).

[6] See U.S. Sentencing Guidelines Manual § 4B1.1 (2014). For other federal sentencing guidelines that provide an enhancement based upon a qualifying prior conviction, see e.g., U.S. Sentencing Guidelines Manual § 2L1.2 (2014) (providing enhancement based on a prior conviction for the crime of unlawfully entering and remaining in the United States) and U.S. Sentencing Guidelines Manual § 2K2.1 (2014) (providing enhancement based on a prior conviction for firearms and ammunition offenses).

[7] For the past several years, approximately 600 criminal defendants per year have been sentenced as Armed Career Criminals and approximately 2200 criminal defendants per year have been sentenced as Career Offenders. See U.S. Sentencing Comm’n, 2013 Sourcebook of Federal Sentencing Statistics tbl. 22 (2013) (noting that there were 582 defendants sentenced as armed career criminals and 2268 sentenced as career offenders in 2013); U.S. Sentencing Comm’n, 2012 Sourcebook of Federal Sentencing Statistics tbl. 22 (2012) (noting that there were 631 armed career criminal sentences and 2232 career offender sentences in 2012); U.S. Sentencing Comm’n, 2011 Sourcebook of Federal Sentencing Statistics tbl. 22 (2011) (noting that there were 571 armed career criminal sentences and 2257 career offender sentences in 2011); U.S. Sentencing Comm’n, 2010 Sourcebook of Federal Sentencing Statistics tbl. 22 (2010) (noting that there were 616 armed career criminal sentences and 2314 career offender sentences in 2010); U.S. Sentencing Comm’n, 2009 Sourcebook of Federal Sentencing Statistics tbl. 22 (2009) (noting that there were 697 armed career criminal sentences and 2392 career offender sentences in 2009); U.S. Sentencing Comm’n, 2008 Sourcebook of Federal Sentencing Statistics tbl. 22 (2008) (noting that there were 653 armed career criminal sentences and 2321 career offender sentences in 2008). The figures for 2014 are not yet available.

[8] See, e.g., Olten v. United States, 134 S. Ct. 639, 639 (2013); Smith v. United States, 134 S. Ct. 258, 258 (2013); United States v. Prater, 766 F.3d 501, 523 (6th Cir. 2014); United States v. Brown, 765 F.3d 185, 196 (3d Cir. 2014); United States v. Bankhead, 746 F.3d 323, 327 (8th Cir. 2014); United States v. Howard, 742 F.3d 1334, 1349 (11th Cir. 2014); United States v. Tucker, 740 F.3d 1177, 1184 (8th Cir. 2014); United States v. Nunez-Segura, 566 F. App’x 389, 394 (5th Cir. 2014); United States v. Royal, 731 F.3d 333, 342 (4th Cir. 2013); United States v. Hemingway, 734 F.3d 323, 338 (4th Cir. 2013); United States v. Parks, 2014 U.S. Dist. LEXIS 1617, at *13–14 (W.D. Wash. Jan. 7, 2014).

[9] See, e.g., Olsson v. United States, 134 S. Ct. 530, 530 (2013); Marrero v. United States, 133 S. Ct. 2732, 2733 (2013); United States v. Brown, 765 F.3d 185, 196–97 (3d Cir. 2014); United States v. Covington, 738 F.3d 759, 767 (6th Cir. 2014); United States v. Murguia-Ochoa, 536 F. App’x 696, 699 (9th Cir. 2013).

[10] See, e.g., United States v. Coronado, 603 F.3d 706, 708 (9th Cir. 2010); United States v. Polk, 577 F.3d 515, 518–19 (3d Cir. 2009); United States v. Seay, 553 F.3d 732, 737–39 (4th Cir. 2009); United States v. Mohr, 554 F.3d 604, 607–09 (5th Cir. 2009); United States v. Herrick, 545 F.3d 53, 56–58 (1st Cir. 2008); United States v. Gray, 535 F.3d 128, 130–31 (2d Cir. 2008); United States v. Bartee, 529 F.3d 357, 359 (6th Cir. 2008); United States v. Templeton, 543 F.3d 378, 380 (7th Cir. 2008); United States v. Williams, 537 F.3d 969, 971 (8th Cir. 2008); United States v. Tiger, 538 F.3d 1297, 1298 (10th Cir. 2008); United States v. Archer, 531 F.3d 1347, 1350–52 (11th Cir. 2008). As noted in note 7, supra, approximately 2200 federal defendants are sentenced as career offenders each year, nearly four times as many as are sentenced as armed career criminals.

[11] See, e.g., Ruiz-Sanchez v. United States, 134 S. Ct. 60, 60 (2013) (noting illegal reentry guidelines); United States v. Davis, 751 F.3d 769, 778 (6th Cir. 2014) (noting child pornography guidelines); United States v. Falcon, 553 F. App’x 737, 738 (9th Cir. 2014) (noting firearms and ammunition guidelines); United States v. Estrella, 758 F.3d 1239, 1254 (11th Cir. 2014) (noting illegal reentry guidelines); United States v. Martinez, 756 F.3d 1092, 1098 (8th Cir. 2014) (same); United States v. Edwards, 734 F.3d 850, 851 (9th Cir. 2013) (noting firearms and ammunition guidelines).

[12] See, e.g., People v. Wilson, 162 Cal. Rptr. 3d 43, 60 (Ct. App. 2013); State v. Dickey, 329 P.3d 1230, 1246 (Kan. Ct. App. 2014); State v. Hulsey, 2013 Kan. App. Unpub. LEXIS 861, at *29 (Kan. Ct. App. 2013). Approximately half of the states have recidivist laws—sometimes called “three strikes” laws—that are similar to the ACCA. See John Clark, James Austin & D. Alan Henry, “Three Strikes and You’re Out”: A Review of State Legislation, National Institute of Justice Research in Brief, National Institute of Justice 7–9 (Sept. 1997); Robert Heglin, Note, A Flurry of Recidivist Legislation Means: “Three Strikes and You’re Out,” 20 J. Legis. 213, 215–16 & nn.18–23 (1994) (collecting state statutes).

[13] See e.g., Rendon v. Holder, 764 F.3d 1077, 1090 (9th Cir. 2014); Aguilar-Turcios v. Holder, 740 F.3d 1294, 1302 (9th Cir. 2014); Lima-Lima v. Holder, 545 F. App’x 648, 650 (9th Cir. 2013); Makwana v. Att’y Gen. of the United States, 543 F. App’x 186, 190 (3d Cir. 2013); Donawa v. United States Att’y Gen., 735 F.3d 1275, 1283–84 (11th Cir. 2013).

[14] See, e.g., Rendon, 764 F.3d at 1084–88 (considering the required elements of a California burglary conviction and the extent of jury unanimity required as to how the offense was committed); United States v. Abbott, 748 F.3d 154, 158–59 (3d Cir. 2014) (same, as to Pennsylvania drug statute); United States v. Royal, 731 F.3d 333, 340–42 (4th Cir. 2014) (same, as to Maryland assault statute); United States v. Cabrera-Gutierrez, 756 F.3d 1125, 1133–37 (9th Cir. 2013) (same, as to Oregon sexual abuse statute).

[15] 18 U.S.C. § 924(e)(1) (2012).

[16] 18 U.S.C. § 924(e)(2)(B)(i) (2012).

[17] 18 U.S.C. § 924(e)(2)(B)(ii) (2012).

[18] Id. Courts frequently refer to these three clauses as the ACCA’s “force” prong, the “enumerated felonies” prong, and the “residual clause,” respectively. See, e.g., United States v. Elliott, 757 F.3d 492, 493 (6th Cir. 2014). The Supreme Court recently ordered new briefing and argument on the question of whether the residual clause is unconstitutionally vague. See Johnson v. United States, No. 13-7120, 2015 WL 132524 (2015) (ordering briefing to be completed by Apr. 10, 2015).

[19] 18 U.S.C. § 924(a)(2).

[20] A defendant who is found to be an Armed Career Criminal may also face an enhanced sentencing range under the United States Sentencing Guidelines as a consequence of that designation. See U.S. Sentencing Guidelines Manual § 4B1.4 (2014) (providing a minimum offense level of 33 and a minimum criminal history category of IV for an Armed Career Criminal, and even higher levels for both if certain facts are established).

[21] Prior to Descamps, the Supreme Court had decided at least nine cases in the previous ten years analyzing what prior convictions count as ACCA predicate felonies. See, e.g., McNeill v. United States, 131 S. Ct. 2218, 2224 (2011) (analyzing when prior state drug convictions count as serious drug offenses for purposes of the ACCA); Sykes v. United States, 131 S. Ct. 2267, 2277 (2011) (analyzing whether a state felony vehicle flight crime qualified as a violent felony under the ACCA’s residual clause); Johnson v. United States, 559 U.S. 133, 145 (2010) (analyzing whether a prior state conviction for battery constituted a violent felony under residual clause); Chambers v. United States, 555 U.S. 122, 130 (2009) (analyzing whether a state conviction for failure to report for periodic incarceration constituted a violent felony under the residual clause); Begay v. United States, 553 U.S. 137, 145 (2008) (analyzing whether a state conviction for felony driving under the influence qualified under the residual clause); United States v. Rodriguez, 553 U.S. 377, 393 (2008) (analyzing whether a state drug conviction counted as a serious drug offense); James v. United States, 550 U.S. 192, 214 (2007) (analyzing whether a state attempted burglary statute qualified under the residual clause); Logan v. United States, 552 U.S. 23, 37 (2007) (considering the implications of a state’s failure to revoke the defendant’s civil rights on whether a conviction counted under the ACCA); Shepard v. United States, 544 U.S. 13, 27 (2005) (setting forth materials that may be consulted when a court applies modified categorical review).

[22] The statute originally included a definition of burglary, but Congress deleted this definition in the course of amending the statute in 1986 to expand the scope of the predicate felonies that count for ACCA purposes. Although the deletion appears to have been unintentional, it has never been corrected. See Taylor v. United States, 495 U.S. 575, 588–90 (1990).

[23] See id. at 589.

[24] Descamps, 133 S. Ct. at 2283.

[25] The Ninth and Sixth Circuits had applied the modified categorical approach to indivisible statutes, as explained infra, whereas the Second and First Circuits had held that the modified categorical approach applied only to divisible statutes. Compare United States v. Descamps, 466 F. App’x 563, 565 (9th Cir. 2012), and United States v. Armstead, 467 F.3d 943, 947–50 (6th Cir. 2006), with United States v. Beardsley, 691 F.3d 252, 268–74 (2d Cir. 2012), and United States v. Giggey, 551 F.3d 27, 40 (1st Cir. 2008) (en banc).

[26] Descamps, 133 S. Ct. at 2284.

[27] Id. at 2283–84.

[28] The prior conviction at issue in Descamps’ case was a conviction for burglary pursuant to Cal. Penal Code Ann. § 459 (West 2010), which provided that a “person who enters” certain locations “with intent to commit grand or petit larceny or any felony is guilty of burglary.” The statute did not require that the entry have been unlawful—an element that the Supreme Court deemed a necessary component of the “generic version” of burglary. See Descamps, 133 S. Ct. at 2285, 2294 (Thomas, J., concurring) (citing 3 Wayne LaFave, Substantive Criminal Law § 21.1(a) (2d ed. 2003)).

[29] See, e.g., Shepard v. United States, 544 U.S. 13, 35–36 (O’Connor, J., dissenting) (decrying the categorical approach’s “overscrupulous regard for formality,” which “forces the . . . sentencing court to feign agnosticism about clearly knowable facts” and “cannot be squared with the ACCA’s twin goals of incapacitating repeat violent offenders, and of doing so consistently notwithstanding the peculiarities of state law”) (emphasis in original). See also Descamps, 133 S. Ct. at 2302 (Alito, J., dissenting) (“The Court’s holding will . . . frustrate fundamental ACCA objectives” including that “violent, dangerous recidivists would be subject to enhanced penalties” in that “those enhanced penalties would be applied uniformly.”).

[30] Descamps, 133 S. Ct. at 2283.

[31] As Daniel Richman has observed, “[e]ven for those not ordinarily interested in technical sentencing decisions, the Court’s opinion is worth reading for its miffed tone of a well-respected author confronted with an obdurate reader claiming unfamiliarity with her work.” Daniel Richman, Opinion analysis: When is a burglary not a burglary?, SCOTUSblog (June 20, 2013, 11:18 PM), http://www.scotusblog.com/2013/06/opinion-analysis-when-is-a-burglary-not-a-burglary; see also William Baude, Modified Categorical Imperative, PrawfsBlawg (June 21, 2013, 2:09 AM), http://prawfsblawg.blogs.com/prawfsblawg/2013/06/modified-categorical-imperative.html (“If you are interested in sharp-tongued judicial rhetoric, you should be following Justice Kagan” who “positively mocks the Ninth Circuit decision.”).

[32] Taylor v. United States, 495 U.S. 575 (1990).

[33] Id. at 601–02.

[34] Id. at 602.

[35] Id.

[36] Shepard v. United States, 544 U.S. 13 (2005).

[37] Id. at 20.

[38] Id. at 21.

[39] Johnson v. United States, 559 U.S. 133 (2010).

[40] 18 U.S.C. § 924(e)(1) (2006).

[41] Johnson, 559 U.S. at 144. In Johnson, the Court finally used the phrase “modified categorical approach,” quoting its decision in Nijhawan v. Holder, 557 U.S. 29 (2009), an immigration case in which the Court considered the application of the Taylor line of cases to the construction of certain felonies under the Immigration and Nationality Act (INA) that could provide the basis for deportation. Id. In Nijhawan, the Court ultimately rejected the use of the categorical approach for purposes of this section of the INA, holding that it was not consistent with Congressional intent. Nijhawan, 557 U.S. at 39–40. However, the Court has held that the categorical approach does apply with respect to certain other portions of the INA. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 194 (2007) (endorsing the categorical approach to interpret the term “theft offense” in the INA).

[42] Johnson represented the usual case in which the Government advocated for categorical review rather than modified categorical review, because the underlying state documents were inconclusive. The Government argued that the Florida statute categorically qualified as a violent felony under the ACCA, thus obviating the need for resort to the underlying documents. Johnson, 558 U.S. at 139. The Court disagreed. Id.

[43] United States v. Descamps, 466 F. App’x 563 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013).

[44] Johnson, 559 U.S. at 144 (emphasis added).

[45] Descamps v. United States, 133 S. Ct. 2276, 2289 (2013).

[46] Id. (quoting United States v. Aguila-Montes de Oca, 655 F.3d 915, 927 (9th Cir. 2011)) (alteration in original).

[47] Id. at 2288 (citing Apprendi v. New Jersey, 530 U.S. 466 (2000)).

[48] Id.

[49] Id. (citing Apprendi, 530 U.S. at 490). In Almendarez-Torres v. United States, 523 U.S. 224 (1998), the Supreme Court held that the existence of a prior conviction was a sentencing factor rather than an element and, accordingly, did not need to be included in an indictment or found by a jury. Post-Apprendi, the Court consistently has recognized Almendarez-Torres as an exception to the Apprendi line of cases, notwithstanding repeated calls from various quarters (including from within the Court, most notably from Justice Thomas) to overrule it. Descamps, 133 S. Ct. at 2294–95 (Thomas, J., concurring); James v. United States, 550 U.S. 192, 231 (2007) (Thomas, J., dissenting); Shepard v. United States, 544 U.S. 13, 27 (2005) (Thomas, J., concurring in part and concurring in the judgment). Descamps argued in the lower courts that Almendarez-Torres should be overruled and sought certiorari on that issue; the Court denied that portion of his petition. United States v. Descamps, 466 F. App’x 563, 564–65 (9th Cir. 2012), cert. denied in part and granted in part, 133 S. Ct. 90, rev’d, 133 S. Ct. 2276 (2013).

[50] Descamps, 133 S. Ct. at 2290.

[51] See Wayne LaFave, Jerold H. Israel, Nancy King & Orin Kerr, Criminal Procedure § 19.3(b) (5th ed. 2009 & Supp. 2013).

[52] See Fed. R. Crim. P. 7(c)(1) (providing that a single count of an indictment may allege that “the means by which the defendant committed the offense are unknown or that the defendant committed it by one or more specified means”); Griffin v. United States, 502 U.S. 46, 51 (1991) (“A statute often makes punishable the doing of one thing or another . . . sometimes thus specifying a considerable number of things . . . . [T]he indictment on such a statute may allege, in a single count, that the defendant did as many of the forbidden things as the pleader chooses, employing the conjunction and where the statute has ‘or,’ . . . and it will be established at the trial by proof of any one of them.” (quoting 1 Joel Prentiss Bishop, New Criminal Procedure § 436, at 355–56 (2d ed. 1913))) (internal footnotes omitted) (internal quotation marks omitted) (emphasis omitted).

[53] See Descamps, 133 S. Ct. at 2298 (Alito, J., dissenting) (envisioning the case of a defendant charged with burglarizing a “floating home” pursuant to a state statute defining burglary as including unlawful entry into, inter alia, a “building” or “vessel” and asking whether, in order to convict, the jury would have to agree “whether this structure was a ‘building’ or a ‘vessel.’ . . . The Court’s answer is ‘yes.’”). Justice Alito also notes that, even if unanimity were required, the basis for the conviction would not be apparent to a subsequent sentencing court unless the jury returned “a special verdict, something that is not generally favored in criminal cases.” Id. at 2300 (citing 6 Wayne LaFave & Jerold H. Israel, Criminal Procedure § 24.10(a), at 543–44 (3d ed. 2007)).

[54] Schad v. Arizona, 501 U.S. 624 (1991).

[55] Richardson v. United States, 526 U.S. 813 (1999).

[56] For a discussion of the indeterminacy of Schad and Richardson, see, for example, Scott W. Howe, Jury Fact-Finding in Criminal Cases: Constitutional Limits on Factual Disagreements Among Convicting Jurors, 58 Mo. L. Rev. 1, 6 (1993); Brian M. Morris, Something Upon Which We Can All Agree: Requiring a Unanimous Jury Verdict in Criminal Cases, 62 Mont. L. Rev. 1, 19–24, 31–34 (2001); Jessica A. Roth, Alternative Elements, 59 UCLA L. Rev. 170, 186–97 (2011); Stephen E. Sachs, Alternative Theories of the Crime 5–6 (May 24, 2007) (unpublished manuscript) (http://ssrn.com/abstract=1501628); Peter Westen & Eric Ow, Reaching Agreement on When Jurors Must Agree, 10 New Crim. L. Rev. 153, 153–54 (2007); see also United States v. Hurt, 527 F.3d 1347, 1356 (D.C. Cir. 2008) (observing that Schad does not provide clear guidance).

[57] The Court has not provided any particularized guidance as to the appropriate methodology in determining whether a legislature intended to create means as opposed to elements. In Schad, however, the Court did cite approvingly to United States v. UCO Oil Co., 546 F.2d 833, 835–38 (9th Cir. 1976), in which the court stated that statutory language, legislative history, the nature of the proscribed conduct, and the appropriateness of multiple punishments all should bear on whether a statute should be interpreted as evidencing a legislative intent to create multiple crimes as opposed to a single crime with multiple means of commission. See Schad, 501 U.S. at 636 (citing UCO Oil Co., 546 F.2d at 835–38).

[58] The Court did, however, suggest certain factors that would be relevant to such a determination, including the moral equivalency of the various alternatives, whether historically they were treated as discrete crimes, and the practices of other states. For example, in Schad, the plurality opinion stated that it would be unconstitutional for a state to recast its entire criminal code as a single statute and declare each formerly separate crime a “means” of committing the offense of “crime.” Schad, 501 U.S. at 632–33.

[59] See, e.g., United States v. Seher, 562 F.3d 1344, 1361–62 (11th Cir. 2009) (holding that a jury need not agree on a defendant’s motive for engaging in a financial transaction, as set forth in the federal money laundering statute); United States v. Stewart, 433 F.3d 273, 319 (2d Cir. 2006) (holding that a jury need not agree on whether a defendant made a false statement or concealed a material fact, as set forth in the false statements statute); United States v. Powell, 226 F.3d 1181, 1196 (10th Cir. 2000) (holding that a jury need not agree on which verb characterizes a defendant’s conduct, as set forth in the federal kidnapping statute). See also 6 LaFave et al., supra note 53, at § 24.10(c), nn.40–41 (collecting additional state and federal cases involving multiple theories and in which courts have approved general verdicts without requiring juror agreement as to the theory).

[60] See Hayden J. Trubitt, Patchwork Verdicts, Different-Juror Verdicts, and American Jury Theory: Whether Verdicts are Invalidated by Juror Disagreement on Issues, 36 Okla. L. Rev. 473, 474–77 (1983) (describing how jury selection and individual jury behavior can lead to compromises that actually stymie true deliberation and discussion among jurors).

[61] Ex ante, adjudicative clarity is a protection for defendants. Ex post, it has positive externalities for those, generally prosecutors, who want to piggyback off the prior adjudicated fact. It is perhaps not surprising then that an Internet search for materials citing Descamps reveals a number of training materials put together by defense attorneys recommending that defenders representing ACCA defendants at sentencing argue that their clients’ prior convictions were pursuant to statutory phrases that represent alternative means rather than alternative elements. See, e.g., Paresh Patel & Lisa Freeland, Post-Descamps World, http://agendas.fjc.gov/agenda/files/FedDef1410-PostDescampsWorld.pdf (last visited Jan. 26, 2015).

[62] Even putting aside the ACCA, there are institutional reasons why prosecutors would be in favor of more specialized verdicts. See Roth, supra note 56, at 213–14 (explaining how prosecutors sometimes benefit from seeking greater clarity in the basis for a jury’s verdict). Congress could amend the ACCA to make clear that “conviction” refers to underlying conduct, not the elements of the prior offense. In his concurring opinion in Descamps, Justice Kennedy called on Congress to make such a change. Descamps v. United States, 133 S. Ct. 2276, 2294 (2013) (Kennedy, J., concurring). But it is not clear that such a change would be consistent with the Sixth Amendment. See id. at 2288 (explaining that if a court’s finding of a predicate offense went beyond merely identifying a prior conviction it would raise Sixth Amendment concerns). Assuming that the Court means what it says when it cites the categorical approach’s Sixth Amendment underpinnings in the context of criminal sentencing, a circumstance-based approach would seem to present serious constitutional problems.

[63] See, e.g., Griffin v. United States, 502 U.S. 46, 60 (1991); id. at 61 (Blackmun, J., concurring); Schad, 501 U.S. at 645.

[64] Schad was decided by a 5-4 vote. Justice Souter wrote the main opinion for the Court for a plurality of the Justices (joined by Chief Justice Rehnquist and Associate Justices O’Connor and Kennedy). Schad, 501 U.S. at 624. Justice Scalia provided the crucial fifth vote to affirm, but made clear in a separate opinion that, in any other case, he might well find that jury unanimity was constitutionally required. Id. at 651 (Scalia, J., dissenting). Richardson, which found elements rather than means—citing constitutional concerns—was written by Justice Breyer, joined by Chief Justice Rehnquist and Justices Stevens, Scalia, Souter, and Thomas. Richardson v. United States, 526 U.S. 813, 814 (1998). Justice Kennedy dissented in Richardson, joined by Justices O’Connor and Ginsburg. Id. at 825 (Kennedy, J., dissenting). Consistent with his position in Descamps, then-Judge Alito dissented in a case decided by the Third Circuit involving the same issue as Richardson. See United States v. Edmonds, 80 F.3d 810, 829 (3d Cir. 1996) (Alito, J., dissenting). Put together with the line-up in Descamps, there now appear to be six members of the Court who might vote to require jury unanimity as to at least one statutory alternative element (Justices Kagan, Breyer, Stevens, Scalia, Thomas, and Chief Justice Roberts) and three Justices who almost certainly would take the opposite view (Justices Kennedy, Ginsburg, and Alito).

[65] See supra notes 63–64.

[66] Descamps, 133 S. Ct. at 2288.

[67] The phrase had been used in the lower courts opinions discussing the application of the ACCA. See, e.g., United States v. Aguila-Montes de Oca, 655 F.3d 915, 963 (9th Cir. 2011); United States v. Vann, 660 F.3d 771, 789 (4th Cir. 2011).

[68] See Descamps, 133 S. Ct. at 2285 n.2 (expressing surprise that the state statutes analyzed in Taylor, Shepard, and Johnson could be construed as setting forth anything but “alternative elements,” and therefore “several different . . . crimes”).

[69] Id.; see id. at 2297–98 (Alito, J., dissenting) (“[L]egislatures frequently enumerate alternative means of committing a crime without intending to define separate elements or separate crimes.” (citing Schad v. Arizona, 501 U.S. 624, 636 (1991))).

[70] The closest the Court came to dealing with this issue in Descamps was to acknowledge Justice Alito’s point that some divisible statutes might set forth alternative means as opposed to elements. Id. at 2285 n.2. The Court expressed surprise that any of the divisible statutes in its ACCA cases to date established means as opposed to elements, and asserted that any confusion as to what constituted elements as opposed to means in future cases could be clarified by consulting the indictment and jury instructions. Id.

[71] See 6 LaFave et al., supra note 53, at § 24.10(c) (observing that the Court has never “derive[d] a uniform theory for identifying which facts are elements”); Nancy J. King & Susan R. Klein, Essential Elements, 54 Vand. L. Rev. 1467, 1497 (2001) (ultimately rejecting as “unwise” and “unattainable” the “theoretical quest for the elemental holy grail”).

[72] Hamling v. United States, 418 U.S. 87, 117 (1974).

[73] See Patterson v. New York, 432 U.S. 197, 210 (1977) (holding that the legislature cannot shift the burden to the defendant to disprove any element of the offense); In re Winship, 397 U.S. 358, 364 (1969) (holding that “the Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged”).

[74] Blockburger v. United States, 284 U.S. 299 (1932).

[75] Id. at 304.

[76] These are the familiar tradeoffs of generality versus specificity, standards versus rules. See, e.g., John F. Manning, Second-Generation Textualism, 98 Calif. L. Rev. 1287, 1316 (2010); Frank H. Easterbrook, Statutes’ Domains, 50 U. Chi. L. Rev. 533, 546–47 (1983).

[77] Descamps v. United States, 133 S. Ct. 2276, 2290 (2013).

[78] Id. at 2289, 2290.

[79] See United States v. Dixon, 509 U.S. 688, 696 (1993) (reaffirming the Blockburger test for constitutionality of both multiple punishments and multiple prosecutions).

[80] Where the defendant was acquitted at a first trial, collateral estoppel may prevent the government from retrying the case, even pursuant to a different alternative element. See Dixon, 509 U.S. at 710 n.15; Ashe v. Swenson, 397 U.S. 436, 443 (1970); Nancy J. King, Portioning Punishment: Constitutional Limits on Successive and Excessive Penalties, 144 U. Pa. L. Rev. 101, 131–32 (1995). If a defendant pleaded guilty pursuant to a plea agreement, a defendant may enjoy additional protection from subsequent prosecution under the terms of the agreement, express or implied. See Santobello v. New York, 404 U.S. 257, 262–63 (1971); Daniel C. Richman, Bargaining About Future Jeopardy, 49 Vand. L. Rev. 1181, 1211–15 (1996).

[81] Many scholars have pointed out the unsatisfactory state of the Court’s current Double Jeopardy jurisprudence. See, e.g., Carissa Byrne Hessick & F. Andrew Hessick, Double Jeopardy as a Limit on Punishment, 97 Cornell L. Rev. 45, 54–55 (2011) (arguing that the Court’s Double Jeopardy jurisprudence “has largely failed to provide any real protection against multiple punishments”); Anne Bowen Poulin, Double Jeopardy and Multiple Punishment: Cutting the Gordian Knot, 77 U. Colo. L. Rev. 595, 599 (2006) (arguing that Double Jeopardy jurisprudence has become confused because of the erroneous entangling of multiple punishment and successive prosecution concerns); Jacqueline E. Ross, Damned Under Many Headings, the Problem of Multiple Punishment, 29 Am. J. Crim. L. 245, 251–53 (2002) (arguing that the Court has focused its Double Jeopardy jurisprudence on illusory problems while ignoring some real ones).

[82] See generally Ross, supra note 81, at 253–57 (explaining how the Guidelines prevent multiple punishments for additional related counts).

[83] Descamps v. United States, 133 S. Ct. 2276, 2284 (2013) (describing the Court’s task in Shepard of determining which “version” of the Massachusetts burglary statute was the basis for the defendant’s conviction).

[84] See Dixon, 509 U.S. at 704–11.

[85] Another example of Descamps’ possible application would be in the context of the Sixth Amendment right to counsel. The Court has held that this right is “offense-specific,” and has adopted a categorical approach, using Blockburger’s elements test, in determining whether a defendant’s Sixth Amendment right to counsel has attached as to a particular offense at a particular point in time. See Texas v. Cobb, 532 U.S. 162, 173 (2001) (holding that “when the Sixth Amendment right to counsel attaches, it does encompass offenses that, even if not formally charged, would be considered the same offense under the Blockburger test”).