The categorical approach, which is a method federal courts use to ‘categorize’ which state law criminal convictions can trigger federal sanctions, is one of the most impactful yet misunderstood legal doctrines in criminal and immigration law. For thousands of criminal offenders, the categorical approach determines whether a previous state law conviction—as defined by the legal elements of the crime—sufficiently matches the elements of the federal crime counterpart that justifies imposing harsh federal sentencing enhancements or even deportation for noncitizens. One of the normative goals courts have invoked to uphold this elements-based categorical approach is that it produces nationwide uniformity. Ironically, however, the categorical approach produces the opposite. By examining the categorical approach in different contexts, this Article shows that relying on state criminal elements has produced nonuniformity due to the variations of state law.

While scholars are increasingly weighing in, this Article contributes to the literature by applying different theories of uniformity that juxtapose the ideals of nationwide uniformity with the potential benefits of nonuniformity. This novel analysis supports several paths forward, dictated by policy preferences. First, if uniformity is to be prioritized, the elements-based categorical approach must be fundamentally redesigned to properly accomplish this goal. Second, uniformity can be responsibly abandoned by justifying the elements-based categorical approach under a different theoretical framework that acknowledges the benefits of state variation. Finally, other options might prove effective to tailor the categorical approach to the policy goals unique to different statutes, and the possible abolition of the categorical approach altogether.


The full text of this Article can be found by clicking the PDF link to the left.


The categorical approach is a method of ‘categorizing’ which state criminal convictions can trigger federal sanctions. 1 See United States v. Garcia-Santana, 774 F.3d 528, 539 (9th Cir. 2014) (stating that the categorical approach helps courts assess “[w]hat set of prior state and federal criminal convictions . . . Congress mean[t] to encompass in a provision assigning [sanctions] to such previous convictions”). For this Article’s definition of “sanctions,” see infra note 4. But lurking under-neath this simple encapsulation of the categorical approach lies a complex, muddled, and perplexing jurisprudence 2 See, e.g., Chambers v. United States, 555 U.S. 122, 126 (2009) (“This categorical approach requires courts to choose the right category. And sometimes the choice is not obvious.”). that has broad impacts in the disparate imposition of federal criminal sentencing enhancements and immigration deportations. 3 See infra notes 106–115, 137–146 and accompanying text. When Congress wrote a number of federal statutes governing the imposition of such harsh sanctions 4 This Article refers to federal criminal and civil penalties as sanctions because this term encompasses both criminal sentencing enhancements and civil immigration deportations. But see Gabriel J. Chin, Illegal Entry as Crime, Deportation as Punishment: Immigration Status and the Criminal Process, 58 UCLA L. Rev. 1417, 1441–51 (2011) (arguing that deportation constitutes quasi-punishment). —such as  criminal  sentencing  enhancements  under  the  Armed  Career  Criminal  Act (ACCA) 5 18 U.S.C. § 924(e) (2018). or deportation under the Immigration and Nationality Act (INA) 6 8 U.S.C. §§ 1101–1537 (2018). —it expressly allowed for state criminal convictions of certain crimes it enumerated in the statutes (such as murder, rape, and burglary) to serve as predicates to trigger these federal sanctions at a later time. 7 See, e.g., id. § 1101(a)(43) (applying convictions under either “Federal or State law” to aggravated felonies); Taylor v. United States, 495 U.S. 575, 588–89 (1990) (acknowledging that the ACCA could be triggered by state law convictions). The categorical approach purports to do this through an elements-based test; it compares the state criminal elements from an offender’s prior state criminal conviction to the federal elements of the crime enumerated in the statute. 8 See Mathis v. United States, 136 S. Ct. 2243, 2248–49 (2016) (noting that, when applying the categorical approach, courts “focus solely on whether the elements of the crime of conviction sufficiently match the elements of [the crime], while ignoring the particular facts of the case”). By way of example, courts utilize the categorical approach to determine if a state law’s version of murder, rape, or burglary—as defined by that state’s criminal elements—qualify under what Congress intended when it enumerated murder, rape, and burglary in the respective federal statute. 9 See, e.g., id. If the state criminal elements match or criminalize narrower conduct than the federal elements, then the state criminal conviction can serve as a predicate to impose the federal sanction. 10 See id. (explaining that, in the context of burglary, a federal conviction can be triggered by a state conviction if the state crime’s elements are “the same as, or narrower than, those of” the federal crime).

For over a century, 11 See infra notes 121–124 and accompanying text. courts have used the categorical approach to impact thousands of people every year, 12 See Katherine Witsman, DHS, Annual Report: Immigration Enforcement Actions: 2017, at 12 tbl.6 (2019), [] (stating that 295,364 noncitizens with criminal records were deported in 2017—down from 333,592 in 2016, 326,406 in 2015, and over 400,000 recorded annually between 2012 and 2014). These figures include all deportations on the basis of criminal convictions, without specifying deportations based alone on aggravated felony determinations. Id.; see also Jessica A. Roth, The Divisibility of Crime, 64 Duke L.J. Online 95, 97 n.7 (2015) (finding that “approximately 600 criminal defendants per year have been sentenced as Armed Career Criminals” under the ACCA); How Many Hundreds (or Thousands?) of ACCA Prisoners Could Be Impacted by a Big Ruling in Johnson?, Sent’g L. & Pol’y (June 13, 2015), [] (estimating that over 7,000 people are currently serving enhanced ACCA sentences). justifying the elements-based approach as one that would promote nationwide uniformity across all jurisdictions, 13 See Taylor v. United States, 495 U.S. 575, 590 (1990) (emphasizing uniform definitions of predicate crimes); United States v. Mayer, 560 F.3d 948, 952 (9th Cir. 2009) (Kozinski, J., dissenting) (explaining the uniformity virtues of the categorical approach); United States ex rel. Mylius v. Uhl, 210 F. 860, 862 (2d Cir. 1914) (“[T]he rule which confines the proof of the nature of the offense to the judgment is clearly in the interest of a uniform . . . administration of the law . . . .”); Laura Jean Eichten, Comment, A Felony, I Presume? 21 USC § 841(b)’s Mitigating Provision and the Categorical Approach in Immigration Proceedings, 79 U. Chi. L. Rev. 1093, 1136 (2012) (“A key goal of the categorical approach has always been the uniform administration of the law.”). Courts have also justified using the categorical approach on the basis that it serves judicial economy, but that notion has been challenged. See Mayer, 560 F.3d at 952 (Kozinski, J., dissenting) (“A great virtue of the categorical approach has been its consistency across doctrinal areas.”); Mylius, 210 F. at 862 (claiming that the categorical approach promotes “efficient administration of the law”); see also infra notes 164–172 and accompanying text. without relying on the different “technical definitions and labels” or the “vagaries of state law.” 14 Taylor, 495 U.S. at 588, 590. But the categorical approach’s reliance on state criminal elements has proven instead to be “an impediment to uniformity.” 15 United States v. Chapman, 866 F.3d 129, 137 (3d Cir. 2017) (Jordan, J., concurring); see also Cristina M. Rodríguez, Uniformity and Integrity in Immigration Law: Lessons from the Decisions of Justice (and Judge) Sotomayor, 123 Yale L.J. Forum 499, 503–04 (2014) [hereinafter Rodríguez, Uniformity and Integrity in Immigration Law] (describing the categorical approach as a “design [that] thwarts uniformity, the courts’ best efforts notwithstanding”). Therefore, because the categorical approach consistently results in disparities that are triggered by the very technicalities and differences in state law it promised to ignore, it has come time for its reconsideration.

The nonuniformity of the categorical approach can be told as a tale of two jurisdictions. Compare the case of Arthur Taylor to that of Richard Mathis; both offenders were convicted of second-degree burglary, but Taylor was convicted under Missouri law and Mathis was convicted under Iowa law. 16 See Mathis v. United States, 136 S. Ct. 2243, 2250 (2016); Taylor, 495 U.S. at 578. Compare Mo. Rev. Stat. § 560.045 (1969) (current version at Mo. Ann. Stat. § 569.170 (West 2017)) (breaking and entering a dwelling house), with Iowa Code § 713.1 (1989) (burgling an occupied structure). The factual evidence preserved in each offender’s state court proceeding showed that both men admitted to similar conduct of having burgled buildings. 17 See United States v. Mathis, 786 F.3d 1068, 1075 (8th Cir. 2015) (finding that two of Mathis’s previous burglary convictions were for burgling garages), rev’d, 136 S. Ct. 2243; United States v. Taylor, 932 F.2d 703, 707 (8th Cir. 1991) (finding that Taylor’s previous burglary convictions were for burgling buildings, including a service station shop). Some years later, both Taylor and Mathis emerged on the radar of federal law enforcement because they both, as persons  convicted  of  felonies,  illegally  possessed  a  firearm  in  violation  of  federal law. 18 See Mathis, 136 S. Ct. at 2250; Taylor, 495 U.S. at 578. In these later federal court proceedings, federal prosecutors pursued sentencing enhancements under the ACCA based on the previous state burglary convictions. 19 See Mathis, 136 S. Ct. at 2246; Taylor, 495 U.S. at 579. But even dealing with such similar criminal conduct, the Supreme Court came to different conclusions under the categorical approach. Whereas the Court held that Missouri’s burglary statute can trigger the federal sanction, it separately held that Iowa’s burglary statute cannot. 20 See Mathis, 136 S. Ct. at 2257; Taylor, 495 U.S. at 602. Because of the minor differences in how these two sister states drafted their respective burglary statutes, the Court applied the categorical approach differently to accommodate these differences in state law. This explains how Mathis, who had the “luck” 21 See Cazarez-Gutierrez v. Ashcroft, 382 F.3d 905, 917 (9th Cir. 2004) (characterizing defendants as having been “lucky enough” to commit a crime in a more lenient state that results in less serious consequences under the categorical approach). of committing criminal conduct in Iowa, could enjoy a sentencing windfall 22 See United States v. Doctor, 842 F.3d 306, 314, 316 (4th Cir. 2016) (Wilkinson, J., concurring) (describing the categorical approach as creating sentencing windfalls); Caleb E. Mason & Scott M. Lesowitz, A Rational Post-Booker Proposal for Reform of Federal Sentencing Enhancements for Prior Convictions, 31 N. Ill. U. L. Rev. 339, 368–70 (2011) (arguing that differences in state law can result in offenders avoiding otherwise prescribed punishment under the categorical approach). But see Evan Tsen Lee, Mathis v. U.S. and the Future of the Categorical Approach, 101 Minn. L. Rev. Headnotes 263, 272–76 (2016) [hereinafter Lee, The Future of the Categorical Approach] (arguing against the windfall logic). whereas Taylor, committing similar criminal conduct across a northern border in Missouri, was not so lucky. 23 See Mathis, 136 S. Ct. at 2260–61 (Breyer, J., dissenting) (highlighting similarities between the Missouri statute in Taylor and the Iowa statute in Mathis, and criticizing the different applications of the categorical approach between cases). Two offenders who admitted to committing similar criminal conduct of burgling different buildings at the state-proceeding stage were treated differently by federal courts imposing sanctions at the federal-proceeding phase because of the differences in state law. And the disparate impact was monumental. The nonuniformity of the categorical approach can mean the difference between adding several years behind bars or receiving a lesser sentence, or even the difference between staying in the country or being deported. 24 See infra notes 106–115, 137–146 and accompanying text.

This illustrates the incompatibility between an elements-based categorical approach and nationwide uniformity. Any state-to-federal sanctioning regime that relies so heavily on state criminal elements will struggle to achieve nationwide uniformity. The federal sanctions applied at a later date will always be wholly dependent on the differences between various state laws. Any promise or commitment that the courts have made to establish nationwide uniformity has proven to be illusory in practice. 25 See infra notes 99–110, 137–146 and accompanying text. As a result, courts have fallen into the very trap they sought to avoid: The different technicalities and vagaries of state law govern the imposition of federal sanctions. 26 See Wayne A. Logan, Contingent Constitutionalism: State and Local Criminal Laws and the Applicability of Federal Constitutional Rights, 51 Wm. & Mary L. Rev. 143, 145–47 (2009) [hereinafter Logan, Contingent Constitutionalism] (discussing the national nonuniformity of federal constitutional rights in criminal procedure due to differences in state law); see also Letter from James Madison to Thomas Jefferson (Oct. 24, 1787), in The Founders’ Constitution 644, 646 (Phillip B. Kurland & Ralph Lerner eds., 1987) (stating that “mutability” of state laws represented a “serious evil”).

This problem arises, in part, because the courts’ broken promises are premised on the semantic imprecision of the ambiguous ideal of “uniformity.” The irony of the term “uniformity” is that it is not uniform. It enjoys a diversity of meanings, some of which can lead down different policy paths. This Article focuses on two such meanings most commonly used by the courts to justify the categorical approach. First is uniformity in terms of application. This type of uniformity ensures that the same rule, or set of rules, is applied the same way to all cases within applicable boundaries. 27 See, e.g., United States v. Torres, 923 F.3d 420, 424–25 (5th Cir. 2019) (citing the uniform application across different areas of law); Amos v. Lynch, 790 F.3d 512, 518 (4th Cir. 2015) (citing Taylor v. United States, 495 U.S. 575, 590–91 (1990)) (stating that the categorical “approach allows federal laws to be applied uniformly”); Rodríguez, Uniformity and Integrity in Immigration Law, supra note 15, at 506 (“[T]he best courts can do in their oversight . . . is to promote a kind of consistency in legal interpretation to guide law enforcement and administration. By consistency, I mean a predictable approach to resolving immigration law questions that pushes administrative actors to adhere to consistent legal standards . . . .”); see also 6 Charles Gordon, Stanley Mailman, Stephen Yale-Loehr & Ronald Y. Wada, Immigration Law and Procedure § 71.05[2][b] (Matthew Bender & Co., Inc. rev. ed. 2013) (highlighting the importance of employing a “uniform rule” to determine which state convictions qualify as aggravated felonies). Second is uniformity in terms of outcomes. This uniformity ensures that similar sanctions are meted out for similar cases. 28 See, e.g., Gerbier v. Holmes, 280 F.3d 297, 311–12 (3d Cir. 2002) (stating “that permanent resident aliens who are in like circumstances, but for irrelevant and fortuitous factors, [should] be treated in a like manner” (quoting Francis v. Immigr. & Naturalization Serv., 532 F.2d 268, 273 (2d Cir. 1976))); Rebecca Sharpless, Clear and Simple Deportation Rules for Crimes: Why We Need Them and Why It’s Hard to Get Them, 92 Denv. U. L. Rev. 933, 943 (2015) [hereinafter Sharpless, Deportation Rules] (noting that “[t]he categorical analysis rests on the bedrock principle[] that fairness requires that immigrants convicted of the same offenses be treated uniformly”). So whereas uniformity of application is more concerned with similar means, uniformity of outcomes is more concerned with similar ends. While these uniformities are distinct in theory, they often share overlapping principles and goals in practice when utilized to build legitimacy in legal regimes. 29 See infra notes 188–194 and accompanying text.

By recognizing these distinct uniformities, this Article argues that the elements-based categorical approach fails both standards because of its dependence on state law. Not only do differences in state criminal elements trigger different applications of the categorical approach, 30 See infra section II.A. but they also trigger different outcomes among otherwise similar offenses. 31 See infra section II.B. The categorical approach has failed to meet the very metric that courts developed it to achieve.

The complexities of navigating the nonuniformity of the categorical approach have demanded more of the Supreme Court’s docket in recent years in its attempt to clarify these complications; the Court has decided over twenty such cases in the past thirteen years, 32 See Shular v. United States, 140 S. Ct. 779, 787 (2020); United States v. Davis, 139 S. Ct. 2319, 2336 (2019); Quarles v. United States, 139 S. Ct. 1872, 1880 (2019); Stokeling v. United States, 139 S. Ct. 544, 555 (2019); United States v. Stitt, 139 S. Ct. 399, 407–08 (2018); Beckles v. United States, 137 S. Ct. 886, 897 (2017); Mathis v. United States, 136 S. Ct. 2243, 2257 (2016); Welch v. United States, 136 S. Ct. 1257, 1268 (2016); Johnson v. United States, 135 S. Ct. 2551, 2563 (2015); United States v. Castleman, 572 U.S. 157, 171 (2014); Descamps v. United States, 570 U.S. 254, 277–78 (2013); Moncrieffe v. Holder, 569 U.S. 184, 190 (2013); Sykes v. United States, 564 U.S. 1, 15–16 (2011); McNeill v. United States, 563 U.S. 816, 823 (2011); Johnson v. United States, 559 U.S. 133, 145 (2010); Nijhawan v. Holder, 557 U.S. 29, 41–43 (2009); Chambers v. United States, 555 U.S. 122, 130 (2009); United States v. Rodriquez, 553 U.S. 377, 393 (2008); Begay v. United States, 553 U.S. 137, 148 (2008); Logan v. United States, 552 U.S. 23, 37 (2007); James v. United States, 550 U.S. 192, 208–09 (2007). with two more cases pending during this 2020–2021 Term alone. 33 See United States v. Borden, 769 F. App’x 266 (6th Cir. 2019), cert. granted, 140 S. Ct. 1262 (Mar. 2, 2020) (No. 19-5410) (mem.) (addressing the application of the ACCA under the categorical approach to mens rea elements); Perieda v. Barr, 916 F.3d 1128 (8th Cir. 2019), cert. granted, 140 S. Ct. 680 (Dec. 18, 2019) (No. 19-438) (mem.) (addressing whether a criminal attempt conviction qualifies as a deportable crime under categorical approach). Increasingly, lower court judges have voiced complaints that the categorical approach has become one of the most judicially taxing issues burdening the federal bench. 34 See infra notes 167–173 and accompanying text. It comes as no surprise, then, that there has also been a sharp increase in scholarly interest both praising the merits of the categorical approach 35 See, e.g., Jennifer Lee Koh, The Whole Better than the Sum: A Case for the Categorical Approach to Determining the Immigration Consequences of Crime, 26 Geo. Immigr. L.J. 257, 295–99 (2012) [hereinafter Koh, The Whole Better than the Sum] (arguing the benefits of the categorical approach include a clear doctrine and judicial efficiency); Rebecca Sharpless, Toward a True Categorical Elements Test: Taylor and the Categorical Analysis of Crimes in Immigration Law, 62 U. Mia. L. Rev. 979, 1031–34 (2008) [hereinafter Sharpless, Toward a True Categorical Elements Test] (same); see also Amit Jain & Phillip Dane Warren, An Ode to the Categorical Approach, 67 UCLA L. Rev. Discourse 132, 150 (2019) (arguing the merits of uniformity produced by the categorical approach); Fatma Marouf, A Particularly Serious Exception to the Categorical Approach, 97 B.U. L. Rev. 1427, 1470 (2017) (same). and/or proposing solutions to fix what has become nearly unworkable. 36 See, e.g., Rachel E. Barkow, Categorical Mistakes: The Flawed Framework of the Armed Career Criminal Act and Mandatory Minimum Sentencing, 133 Harv. L. Rev. 200, 207–09 (2019) [hereinafter Barkow, Categorical Mistakes] (arguing that Congress’s statutory design of the ACCA is problematic for incorporating state law, and that Congress should have instead given discretion to the Sentencing Commission); Sheldon A. Evans, Punishing Criminals for Their Conduct: A Return to Reason for the Armed Career Criminal Act, 70 Okla. L. Rev. 623, 649–52 (2018) [hereinafter Evans, Punishing Criminals] (examining the inconsistencies of basing ACCA sentencing enhancements on state law); Iris Bennett, Note, The Unconstitutionality of Nonuniform Immigration Consequences of “Aggravated Felony” Convictions, 74 N.Y.U. L. Rev. 1696, 1720–24 (1999) (explaining that differences between states’ criminal laws lead to discrepancies in the imposition of federal immigration sanctions).

This Article contributes a novel analysis to the literature on the categorical approach by juxtaposing the theoretical framework of nationwide uniformity with the political ideals of a federalist system of government. Building from the costs and benefits of each, the Article argues that there are unique paths forward that would restructure the categorical approach to solve its problem of nonuniformity. The first path prioritizes nationwide uniformity as a worthwhile and achievable pursuit. If either uniformity of application or outcomes is to be salvaged, the categorical approach must be redesigned to properly accommodate these uniformities. The most viable option to maximize both uniformities would be a conduct-based categorical approach. Such an approach would rely on the underlying criminal conduct of the offender’s state court conviction, and then determine if such conduct—regardless of whether it was committed in Missouri or Iowa—would fit within the federal elements of the enumerated crime. 37 See Evans, Punishing Criminals, supra note 36, at 666–67 (arguing that a conduct-based approach would solve nonuniformity between jurisdictions). By excising reliance on state law criminal elements, a conduct-based categorical approach will impose sanctions according to criminal conduct, which would mitigate nonuniformity that comes with state law variety. 38 Id.

But there is another path forward, one that recognizes that while a conduct-based categorical approach holds a greater likelihood of coming closer to uniformity than the current elements-based approach, it will not eliminate all disparity. And if perfect uniformity cannot be achieved, 39 See Amanda Frost, Overvaluing Uniformity, 94 Va. L. Rev. 1567, 1606 (2008) (arguing that the detriments of nationwide nonuniformity of federal law are overstated). the current elements-based categorical approach can be retained, albeit under a different justification. This Article argues that the practical and theoretical benefits of federalism can serve to justify the disparate application and outcomes of the elements-based categorical approach. Embracing the natural differences that arise within a federal system that accounts for and encourages differences in state law according to the preferences of that polity would justify the disparate treatment of criminal offenders who choose to violate the laws of that state, and must then accept the different costs that come with such state citizenship.

This Article proceeds in Part I by focusing on the ACCA and the INA, two key areas in which applying the categorical approach has the greatest effect on defendants and the doctrine’s development. The examples illustrate how the differences in states’ criminal statutes trigger different applications of the categorical approach and, in turn, trigger different outcomes for similar criminal conduct. Part II serves as the primary contribution of the Article by examining how the categorical approach fails to uphold courts’ promises of uniformity. By comparing courts’ logic with that of the different underpinnings of uniformity in application and uniformity in outcomes, it becomes nearly impossible to achieve either uniformity under the current elements-based categorical approach. Part III presents potential remedies to the categorical approach by exploring several paths with different views on uniformity. First, if nationwide uniformity is an achievable goal, this Article argues that the best option is to transition to a conduct-based categorical approach that removes the disparities created by differences in state law. Uniformity might also be achieved with a separate-sovereign system, allowing federal sanctions to be imposed for only previous federal—not state—convictions. Second, if nationwide uniformity is to be abandoned, the elements-based categorical approach can be salvaged, but under the different theoretical and practical justifications that may come by embracing the benefits of having variety in state law under a federal system of government. Embracing nonuniformity also allows a novel analysis of tailoring the categorical approach, pursuing different approaches for different statutes to serve the unique policy goals of each. And finally, the challenges of salvaging uniformity, and possibly redesigning the categorical approach, require consideration of complete abolition of the doctrine altogether.