Altin Shuti v. Loretta Lynch

828 F.3d 440, 2016 FED App. 0155P, 2016 U.S. App. LEXIS 12500, 2016 WL 3632539

• Court: Court of Appeals for the Sixth Circuit
• Decided: July 7, 2016
• Docket: 15-3835
• Status: Published

Opinion

OPINION

COLE, Chief Judge.

In Johnson v. United States, — U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), the Supreme Court held the Armed Career Criminal Act’s residual definition of “violent felony” void for vagueness. 18 U.S.C. § 924(e)(2)(B)(ii). In this case, we consider whether that pathmarking decision applies to the Immigration and Nationality Act’s parallel definition of “crime of violence,” a phrase that encompasses any felony that “by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.” 8 U.S.C. § 1101(a)(43)(F); 18 U.S.C. § 16(b). We conclude that the wide-ranging inquiry required by these two statutory phrases are one and the same, and therefore hold that the immigration code’s residual clause is likewise unconstitutionally vague.

I.

Petitioner Altin Bashkim Shuti, who hails from Albania, entered the United States as a lawful permanent resident in October 2008. He was 13 years old when his parents, who are now American citizens, decided to flee their home-country for fear of persecution at the hands of the Albanian Socialist Party.

Nearly six years later, in May 2014, Shuti and a few of his high-school cohorts allegedly committed a “larceny of marijuana” and “in the course of that conduct possessed a shotgun.” Shuti pleaded guilty, for his part, to the lesser offense of felony unarmed robbery, defined under Michigan law as “larceny of any money or other property” accomplished by using “force or violence against any person who is present” or “assault[ing] or put[ting] the person in fear.” Mich. Comp. Laws § 750.530. The state trial court sentenced Shuti to at least two and a half years in prison and, several months later, the Department of Homeland Security initiated removal proceedings against him.

Under the Immigration and Nationality Act (“INA”), 8 U.S.C. § 1101 et seq., the Attorney General may remove certain classes of non-citizens from this country — for instance, those who have been convicted of crimes involving moral turpitude, firearms offenses, and various drug offenses. 8 U.S.C. § 1227(a)(2). In the ordinary course, a non-citizen may apply to immigration officials for discretionary relief from removal. See, e.g., 8 U.S.C. §§ 1158 (asylum), 1229b (cancellation of removal), 1231(b)(3)(A) (withholding of removal). But aggravated felonies are different: if a non-citizen has been “convicted of an aggravated felony at any time after admission,” 8 U.S.C. § 1227(a)(2)(A)(iii), he is ineligible for most forms of discretionary relief, 8 U.S.C. §§ 1158(b)(2)(B)©, 1229b(a)(3), 1231(b)(3)(B)(iv). Removal is “virtually inevitable” in such cases. See Padilla v. Kentucky, 559 U.S. 356, 360, 130 S.Ct. 1473, 176 L.Ed.2d 284 (2010).

The term “aggravated felony” is defined expansively under the INA. Among the numerous state and federal offenses that qualify, the immigration code lists “a crime of violence (as defined in section 16 of Title 18, but not including a purely political offense) for which the term of imprisonment [is] at least one year.” 8 U.S.C- § 1101(a)(43)(F). This cross-reference, in turn, leads to the general criminal code, which defines a “crime of violence” as:

(a) an offense that has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or

(b) any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.

18 U.S.C. § 16. We deal here with the provision’s latter subsection.

In this case, the government alleged that Shuti’s Michigan conviction was an aggravated felony. Shuti acquiesced to the charge, and opted to file an application for asylum, withholding of removal, and protection under the Convention Against Torture. He also maintained that his criminal attorney “never discussed” the immigration consequences of his state court plea. But in 2015, an immigration judge denied all discretionary relief and ordered Shuti removed to Albania.

The Board of Immigration Appeals (“BIA”) affirmed. The BIA first determined that unarmed robbery was “categorically a crime of violence” as defined in 18 U.S.C. § 16(b). “[A]n individual who engages in robbery,” the BIA opined, “clearly involves a substantial risk that physical force will be used in the ordinary case.” For this proposition, the BIA relied on two analogous precedents: our decision in United States v. Mekediak, 510 Fed.Appx. 348, 353-54 (6th Cir. 2013) (applying USSG § 4B1.2(a)(2)’s definition of crime of violence to Mich. Comp. Laws § 750.530), and the Seventh Circuit’s decision in United States v. Tirrell, 120 F.3d 670, 681 (7th Cir. 1997) (applying 18 U.S.C. § 924(e)(2)(B)(ii)’s definition of violent felony to Mich. Comp. Laws § 750.530). Shuti responded that the BIA improperly comin-gled statutory definitions, but the agency skirted this minor “distinction.” Matter of Francisco-Alonzo, 26 I. & N. Dec. 594, 597-98 (BIA 2015). That would have been the last word. But while the appeal was pending, the Supreme Court handed down Johnson. Shuti argued, through supplemental briefing, that the INA’s definition of crime of violence was unconstitutionally vague in light of this intervening precedent. The BIA balked, declaring that it “do[es] not address the constitutionality of the laws [it] administer[s].” Matter of G-K-, 26 I. & N. Dec. 88, 96 (BIA 2018). Nevertheless, the agency concluded that the void-for-vagueness doctrine simply does not apply to “civil” deportation proceedings.

We now grant Shuti’s petition for review as to the “constitutional elaim[ ],” 8 U.S.C. § 1252(a)(2)(D), and vacate the order of removal. 1

II.

“No person,” the Fifth Amendment says, “shall ... be deprived of life, liberty, or property without due process of law.” The Constitution’s prohibition of vague laws springs from this well. Collins v. Kentucky, 234 U.S. 634, 638, 34 S.Ct. 924, 58 L.Ed. 1510 (1914). A statute “so vague that it fails to give ordinary people fair notice of the conduct it punishes, or so standard-less that it invites arbitrary enforcement,” Johnson, 135 S.Ct. at 2556, “violate[s] the fundamental principles of justice embraced in the conception of due process of law,” Collins, 234 U.S. at 638, 34 S.Ct. 924. The Supreme Court has, for over a hundred years, reviewed legislation of all sorts on this basis. See, e.g., Johnson, 135 S.Ct. at 2557 (sentencing enhancement); Kolender v. Lawson, 461 U.S. 352, 357-58, 103 S.Ct. 1855, 75 L.Ed.2d 903 (1983) (stop-and-identify statute); Papachristou v. City of Jacksonville, 405 U.S. 156, 162, 92 S.Ct. 839, 31 L.Ed.2d 110 (1972) (vagrancy ordinance); Int’l Harvester Co. of Am. v. Kentucky, 234 U.S. 216, 221, 34 S.Ct. 853, 58 L.Ed. 1284 (1914) (price-control statute). And, in case after case, the Court has deployed the void-for-vagueness doctrine to strike down laws that violate this “first essential of due process.” Connolly v. Gen. Const. Co., 269 U.S. 385, 391, 46 S.Ct. 126, 70 L.Ed. 322 (1926).

Johnson applied these principles to the Armed Career Criminal Act (“ACCA”), 18 U.S.C. § 924(e). Under that statute, any person who violates the federal felon-in-possession laws, 18 U.S.C. § 922(g), and has at least three prior convictions for a “violent felony” is subject to an enhanced 15-year mandatory minimum sentence. 18 U.S.C. § 924(e)(1). Congress defined violent felony, in relevant part, as any offense “punishable by imprisonment for a term exceeding one year” that “is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another.” 18 U.S.C. § 924(e)(2)(B)(ii) (emphasis added). This italicized portion of the definition is called the “residual clause,” and the Johnson Court held it void for vagueness.

To arrive at that conclusion, the Court emphasized “[t]wo features” that “conspire” to make the residual clause hopelessly vague. Johnson, 135 S.Ct. at 2557. To begin with, the clause left uncertainty about “how to estimate the risk posed by a crime” because it tied violent felony analysis “to a judicially imagined ‘ordinary case’ of a crime” instead of “real-world facts or statutory elements.” Id. And on top of that, the residual clause left uncertainty about “how much risk it takes for a crime to qualify as a violent felony” because it required application of an “imprecise ‘serious potential risk’ standard” to this “judge-imagined abstraction.” Id. at 2558. These indeterminacies combined, the Court held, to foster “more unpredictability and arbitrariness than the Due Process Clause tolerates.” Id.

The Court recently explained this holding in Welch v. United States, — U.S. -, 136 S.Ct. 1257, 194 L.Ed.2d 387 (2016). There, the Court clarified that the residual clause’s vagueness “rests in large part on its operation under the categorical approach.” Id. at 1262. The categorical approach is an abstract mode of analysis, mandated by Congress’s focus on the historical fact of prior conviction. Taylor v. United States, 495 U.S. 575, 600, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990). To determine whether an offense is a violent felony under the residual clause, courts must consider “whether the conduct encompassed by the elements of the offense, in the ordinary case, presents a serious potential risk of injury to another.” James v. United States, 550 U.S. 192, 208, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007). This entails looking at the offense categorically — “in terms of how the law defines the offense,” not how the individual “committed it on a particular occasion.” Begay v. United States, 553 U.S. 137, 141, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008). At bottom, the Court said in Welch, the residual clause “failed not because it adopted a ‘serious potential risk’ standard but because applying that standard under the categorical approach required courts to assess the hypothetical risk posed by an abstract generic version of the offense.” Welch, 136 S.Ct. at 1262.

Johnson was no doubt a sea-change, with far-reaching precedential effects. For example, the Court has granted certiorari, vacated the judgment, and remanded for reconsideration in various sentencing cases on direct appeal. See, e.g., Richardson v. United States, — U.S. -, 136 S.Ct. 1157, 194 L.Ed.2d 164 (2016) (mem.); Holder v. United States, — U.S. -, 135 S.Ct. 2940, 192 L.Ed.2d 969 (2015) (mem.); Ball v. United States, — U.S. -, 135 S.Ct. 2933, 192 L.Ed.2d 968 (2015) (mem.). And just this term, the Court held that Johnson has retroactive effect in cases on collateral review. Welch, 136 S.Ct. at 1265; see also In re Watkins, 810 F.3d 375, 379 (6th Cir. 2015). The courts of appeals have gotten on board as well, applying Johnson to analogous residual clauses. Take our recent decision in United States v. Pawlak, 822 F.3d 902, 910 (6th Cir.2016), where we concluded that the “rationale of Johnson applies equally” to the United States Sentencing Guidelines’ residual definition of crime of violence. In addition, two other circuits have applied Johnson to immigration statutes that invoke the criminal code’s parallel definition of crime of violence. See, e.g., United States v. Hernandez-Lara, 817 F.3d 651, 652 (9th Cir. 2016) (per curiam); United States v. Vivas-Ceja, 808 F.3d 719, 720 (7th Cir. 2015); Dimaya v. Lynch, 803 F.3d 1110, 1111 (9th Cir. 2015); see also United States v. Gonzalez-Longoria, 813 F.3d 225, 235, reh’g en banc granted, 815 F.3d 189 (5th Cir. 2016). In Dimaya, for example, the Ninth Circuit concluded that Johnson’s “reasoning applies with equal force to the similar statutory language and identical mode of analysis” used in the INA’s residual definition of crime of violence. 803 F.3d at 1115.

With this legal landscape in mind, we circle back to Shuti’s constitutional challenge.

III.

Shuti maintains that the INA’s residual clause suffers from the same defects as the statute at issue in Johnson and, so too, runs afoul of the Fifth Amendment’s prohibition of vague laws. The government counsels caution. Johnson, the government tells us, was an opinion that in essence cannot be applied beyond the ACCA.

One constitutional question is presented here: is the INA’s definition of “crime of violence,” 8 U.S.C. § 1101(a)(43)(F), in combination with the criminal statute cross-referenced there, 18 U.S.C. § 16(b), unconstitutionally vague? Our review is de novo. United States v. Hart, 635 F.3d 850, 856 (6th Cir. 2011).

A.

As the Supreme Court has long recognized, the Fifth Amendment’s prohibition of vague laws is “applicable to civil as well as criminal actions.” Boutilier v. INS, 387 U.S. 118, 123, 87 S.Ct. 1563, 18 L.Ed.2d 661 (1967) (citing A.B. Small Co. v. Am. Sugar Ref. Co., 267 U.S. 233, 239, 45 S.Ct. 295, 69 L.Ed. 589 (1925)); see also Johnson, 135 S.Ct. at 2566-67 (Thomas, J., concurring in judgment). But the government suggests, as the BIA concluded, that the void-for-vagueness doctrine does not apply in deportation proceedings because they are “civil in nature.”

That notion is misguided. If anything, it is “well established” that the Fifth Amendment “entitles” non-citizens to due process in removal proceedings. Reno v. Flores, 507 U.S. 292, 306, 113 S.Ct. 1439, 123 L.Ed.2d 1 (1993). This includes the constitutional requirements of “fair notice” and “even-handed administration of the law.” See Papachristou, 405 U.S. at 162, 171, 92 S.Ct. 839; cf. Mellouli v. Lynch, — U.S. -, 135 S.Ct. 1980, 1987, 192 L.Ed.2d 60 (2015) (recognizing the need to “promote efficiency, fairness, and predictability in the administration of immigration law”).

Jordan v. De George clinches the matter in that regard. 341 U.S. 223, 71 S.Ct. 703, 95 L.Ed. 886 (1951). There, the Court considered a vagueness challenge to 8 U.S.C. § 1227(a)(2)(A)(ii)’s early-twentieth-century predecessor, which authorized the removal of non-citizens who have been convicted of two or more crimes involving moral turpitude. Id. at 225, 71 S.Ct. 703. “Despite the fact” that the provision at issue was “not a criminal statute,” the Court applied the “established criteria” of the void-for-vagueness doctrine “in view of the grave nature of deportation.” Id. at 231, 71 S.Ct. 703; see also Vartelas v. Holder, — U.S. -, 132 S.Ct. 1479, 1487, 182 L.Ed.2d 473 (2012) (noting the “severity” of subjecting “permanent residents” to “potential banishment”); U.S. ex rel. Guarino v. Uhl, 107 F.2d 399, 400 (2d Cir. 1939) (L. Hand, J.) (lamenting “the dreadful penalty of banishment, which is precisely what deportation means to one who had lived here since childhood”). The criminal versus civil distinction is thus “ill suited” to evaluating a vagueness challenge regarding the “specific risk of deportation.” Cf. Padilla, 559 U.S. at 365-66, 130 S.Ct. 1473 (describing how deportation proceedings are “intimately related” to, and “enmeshed” in, our criminal laws).

It should come as no surprise, then, that we have previously recognized the void-for-vagueness doctrine’s applicability “beyond criminal laws to immigration stat utes.” Mhaidli v. Holder, 381 Fed.Appx. 521, 525 (6th Cir. 2010). So too have other circuits. See, e.g., Alphonsus v. Holder, 705 F.3d 1031, 1042 (9th Cir. 2013); Arriaga v. Mukasey, 521 F.3d 219, 222 (2d Cir. 2008); Garcia-Meza v. Mukasey, 516 F.3d. 535, 536 (7th Cir. 2008). Our conclusion remains the same: because deportation strips a non-citizen of his rights, statutes that impose this penalty are subject to vagueness challenges under the Fifth Amendment. Boutilier, 387 U.S. at 123, 87 S.Ct. 1563; Jordan, 341 U.S. at 231, 71 S.Ct. 703.

B.

Like the Seventh and Ninth Circuits, we are convinced that Johnson is equally applicable to the INA’s residual definition of crime of violence. 8 U.S.C. § 1101(a)(43)(F); 18 U.S.C. § 16(b). The text of the immigration code at once compels a categorical approach to prior convictions and an imprecise analysis of possible risk. This “wide-ranging inquiry,” as with the similar statutory language in the ACCA and Sentencing Guidelines, “denies fair notice to defendants and invites arbitrary enforcement by judges.” See Johnson, 135 S.Ct. at 2557. The consistent comingling of residual-clause precedents interpreting the INA, ACCA, and Guidelines shores up our conclusion. See Pawlak, 822 F.3d at 910. Imposing the penalty of deportation under this nebulous provision, we conclude, denies due process of law.

1.

Begin with a comparison of the text. The INA provision at issue here defines a “crime of violence” as a felony that:

by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.

8 U.S.C. § 1101(a)(43)(F); 18 U.S.C. § 16(b) (emphasis added). The ACCA defines a “violent felony” as any crime punishable by imprisonment for a term exceeding one year that:

is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another.

18 U.S.C. § 924(e)(2)(B)(ii) (emphasis added).

While not a perfect match, these provisions undeniably bear a textual resemblance. See, e.g., Johnson v. United States, 559 U.S. 133, 140, 130 S.Ct. 1265, 176 L.Ed.2d 1 (2010) (noting that 18 U.S.C. § 16’s definition of crime of violence is “very similar” to the ACCA’s definition of violent felony); Nijhawan v. Holder, 557 U.S. 29, 37, 129 S.Ct. 2294, 174 L.Ed.2d 22 (2009) (conceding that “the ‘aggravated felony1 statute ... resembles [the] ACCA in certain respects”); Chambers v. United States, 555 U.S. 122, 133 n.2, 129 S.Ct. 687, 172 L.Ed.2d 484 (2009) (Alito, J., concurring in judgment) (recognizing that “18 U.S.C. § 16(b) ... closely resembles ACCA’s residual clause”). In both statutes, Congress has focused on the fact of prior “conviction,” compare 8 U.S.C. § 1227(a)(2)(A)(iii), with 18 U.S.C. § 924(e)(1), and in both residual provisions Congress has asked whether the crime possibly “involves” too much “risk” of harm, compare 18 U.S.C. § 16(b), with 18 U.S.C. § 924(e)(2)(B)(ii).

An identical mode of analysis flows from this plain reading of the text. Both residual clauses require a categorical approach to prior convictions. To be sure, the categorical approach has “historically” been used to determine “whether a state conviction renders an alien removable under the immigration statute.” See Mellouli, 135 S.Ct. at 1986-87; see also Moncrieffe v. Holder, — U.S.-, 133 S.Ct. 1678, 1684-85, 185 L.Ed.2d 727 (2013); Nijhawan, 557 U.S. at 33-38, 129 S.Ct. 2294; Gonzales v. Duenas-Alvarez, 549 U.S. 183, 185-87, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007). And the two residual provisions deploy the ordinary case method in precisely the same fashion. Compare James, 550 U.S. at 209, 127 S.Ct. 1586 (holding that attempted burglary “satisfies the requirements of [18 U.S.C.] § 924(e)(2)(B)(ii)’s residual provision” because it is an offense “that, by its nature, presents a serious potential risk of injury to another”) (emphasis added), with Leocal v. Ashcroft, 543 U.S. 1, 10, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004) (holding that burglary “would be covered under [18 U.S.C.] § 16(b) ... because [the offense], by its nature, involves a substantial risk that the burglar will use force against a victim”) (emphasis added).

An imprecise analysis of the possible risk of harm posed by this abstraction ensues. Compare Begay, 553 U.S. at 143-45, 128 S.Ct. 1581 (holding that driving under the influence of alcohol is not a “violent felony” under 18 U.S.C. § 924(e)(2)(B)(ii) because the offense is not “roughly similar ... in degree of risk posed” to crimes that “typically” involve “purposeful, violent, and aggressive” conduct) (emphasis added), with Leocal, 543 U.S. at 9-11, 125 S.Ct. 377 (holding that driving under the influence of alcohol is not a “crime of violence” because 18 U.S.C. § 16(b)’s risk standard “naturally suggests” more than the “merely accidental or negligent conduct”) (emphasis added). Neither term — “substantial” in the INA or “serious” in the ACCA — “sets forth [objective] criterion” to determine how much risk it takes to qualify as a crime of violence or violent felony. See James, 550 U.S. at 219, 127 S.Ct. 1586 (2007) (Scalia, J., dissenting) (using Leocal as an example of how “courts might vary dramatically in their answer”).

In short, both provisions combine indeterminacy about “how to measure the risk posed by a crime” and “how much risk it takes for the crime to qualify” as a crime of violence or a violent felony. See Johnson, 135 S.Ct. at 2557-58. We cannot avoid the conclusion that the INA’s residual clause falls squarely within Johnson’s core holding.

2.

Confirmation comes readily. Consider the insidious comingling of precedents in this context: as Judge Kozinski of the Ninth Circuit has explained it, “[t]he interoperability of the [categorical approach] means that precedents can be mixed and matched, regardless of which statute was at issue in which case.” See United States v. Mayer, 560 F.3d 948, 952 (9th Cir. 2009) (Kozinski, J., dissenting from denial of rehearing en banc). In other words, INA cases can be applied to the ACCA, ACCA cases can be applied to the Guidelines, and Guidelines cases can be applied to the INA. See id.

That principle is on perfect display in cases where the BIA has applied ACCA precedents like James (now overruled) to the immigration code. Matter of Francisco-Alonzo is illustrative. 26 I. & N. Dec. at 597-98. In that decision, the BIA was tasked with deciding whether felony battery was a “crime of violence” and, therefore, an “aggravated felony” under the INA. Id. at 596. In conducting this analysis, the BIA concluded that it must employ the “James ‘ordinary case’ analysis,” and further opined that the immigration code defines crime of violence in “terms similar to the [ACCA’s] residual clause.” Id. at 598, 600. And on that basis, the agency concluded it was proper to “appl[y] James” — and other circuit level precedents interpreting the ACCA and Sentencing Guidelines — for the proposition that felony battery “meets the ‘risk of injury’ requirement” in the immigration code. Id. at 599-601. We have leaned on ACCA and Guidelines precedents in like manner. See, e.g., United States v. Stout, 706 F.3d 704, 708-09 (6th Cir. 2013); Van Don Nguyen v. Holder, 571 F.3d 524, 529-30 (6th Cir. 2009); Patel v. Ashcroft, 401 F.3d 400, 408 (6th Cir. 2005).

For a case study in vagueness, look no further than the BIA’s decision below. In this very case, the BIA invoked the INA’s residual definition of crime of violence to query whether unarmed robbery “clearly involves a substantial risk that physical force will be used in the ordinary case.” James, 550 U.S. at 208, 127 S.Ct. 1586. Then, relying on our previous application of the Guidelines’ residual clause to Michigan’s offense of unarmed robbery, the agency concluded that its categorical abstraction “fit[ ] comfortably within the [residual definition] of ‘crime of violence.’ ” Mekediak, 510 Fed.Appx. at 354. We have, of course, held that exact provision void for vagueness. As we stated in Pawlak, “[g]iv-en our reliance on the ACCA for guidance in interpreting [USSG] § 4B1.2, it stretches credulity to say that we could apply the residual clause of the Guidelines in a way that is constitutional, when courts cannot do so in the context of the ACCA.” Pawlak, 822 F.3d at 911 (quoting United States v. Madrid, 805 F.3d 1204, 1211 (10th Cir. 2015)). So too with the INA.

C.

The government takes issue with our conclusion. It seeks refuge in a few textual differences between the INA and the ACCA that, in its view, foreclose application of Johnson. Failing that, the government attempts to narrowly characterize Johnson’s holding and precedential effect. These points are all well taken, though we think they are, ultimately, distinctions without a difference.

To start, the government suggests that the ACCA’s enumerated-crimes clause was a decisive factor in Johnson. The INA’s lack of a prefatory list should, in its view, put an end to our inquiry. But the existence of a prefatory “list of examples,” though surely confusing, was not determinative of the Court’s vagueness analysis. See Johnson, 135 S.Ct. at 2558, 2561. Rather, the Court’s “wide-ranging inquiry” holding was the “[m]ore important[ ]” aspect. See id. at 2557, 2561. At any rate, the INA’s lack of an enumerated-crimes clause actually makes its residual clause a “broad[er]” provision, as it “eover[s] every offense that involved a substantial risk of the use of ‘physical force against the person or property of another.’ ” See Begay, 553 U.S. at 144, 128 S.Ct. 1581.

In a similar vein, the government argues that the INA’s residual clause provides a sufficiently definite standard because its text focuses on the risk that “force” may be used in the ordinary case of “committing the offense.” This distinction, the government claims, renders the risk analysis somehow less uncertain. See Leocal, 543 U.S. at 10 n. 7, 11, 125 S.Ct. 377. We are hard pressed to accept these textual distinctions. Even though the INA refers to the risk that “force may be used,” rather than the risk that potential “injury might occur,” Johnson is equally applicable. The reason is simple: a marginally narrower abstraction is an abstraction all the same.

Take Leocal’s discussion of burglary as an example. There, the court held that burglary is a “classic example” of a crime of violence. Id. at 10, 125 S.Ct. 377. On one view, it is “[t]he fact that an offender enters a building ... [that] creates the possibility of a violent confrontation between the [burglar] and an occupant.” See Taylor, 495 U.S. at 588, 110 S.Ct. 2143; see also Leocal, 543 U.S. at 10, 125 S.Ct. 377. But as the Johnson Court subsequently pointed out, assessing the level of risk posed by the ordinary case of burglary is an entirely speculative enterprise. One can just as easily imagine a run-of-the-mill burglar who breaks into a seemingly empty home, hears the occupants stirring, and runs away without confrontation. Cf . Johnson, 135 S.Ct. at 2558; James, 550 U.S. at 211, 127 S.Ct. 1586 (Scalia, J., dissenting). As with the ACCA, the INA’s residual definition of crime of violence fails to provide a “reliable way to choose between these competing accounts,” regardless of its focus on the risk that force may be used. See Johnson, 135 S.Ct. at 2558. And, as noted earlier, the theoretical distinction between these statutes has been erased in practice. See Mayer, 560 F.3d at 952 (Kozinski, J., dissenting from denial of rehearing en banc); cf. Johnson, 559 U.S. at 140, 130 S.Ct. 1265 (equating 18 U.S.C. § 924(e)(2)(B)’s definition of “violent” with the “substantial force” standard used in the INA). The interoperability of the categorical approach in these cases may have been its virtue, but the taint of its indeterminacy is also its downfall.

The government does not endeavor to distinguish away Johnson’s, core holding, nor can it. The application of an imprecise risk-based standard to a hypothetical ordinary case of the crime “does not comport with the Constitution’s guarantee of due process.” See Johnson, 135 S.Ct. at 2558, 2560. Recognizing as much, the government claims that Johnson was a narrow decision, one that specifically avoided calling other federal laws into question. “As a general matter,” the Court said, “we do not doubt the constitutionality of laws that call for the application of a qualitative standard such as ‘substantial risk’ to real-world conduct.” Id. at 2561. But our holding is plainly consistent with this disclaimer. The INA’s residual clause, as described above, does not call for courts to “gaug[e] the riskiness of conduct in which an individual defendant engages on a particular occasion” or for the application of such a standard “to real-world conduct.” See id. (emphasis added). The immigration code, rather, mandates a categorical mode of analysis that deals with “an imaginary condition other than the facts.” See id. (quoting Int’l Harvester Co., 234 U.S. at 223, 34 S.Ct. 853); see also Leocal, 543 U.S. at 7, 11, 125 S.Ct. 377.

The government persists, however, arguing that our recent decision in United States v. Taylor, 814 F.3d 340 (6th Cir. 2016), forecloses Shuti’s constitutional challenge to the INA’s residual clause. To the contrary, we find Taylor wholly consistent with our conclusion. There, we held that 18 U.S.C. § 924(c)’s definition of crime of violence was not unconstitutionally vague. Id. at 375-76. That conclusion, we think, makes perfect sense because the statute at issue in Taylor is a criminal offense and “creation of risk is an element of the crime.” See Johnson, 135 S.Ct. at 2557. As the Johnson Court determined, no doubt should be cast upon laws that apply a qualitative risk standard to “real-world facts or statutory elements.” See id. at 2557, 2561 (emphasis added). Unlike the ACCA and INA, which require a categorical approach to stale predicate convictions, 18 U.S.C. § 924(c) is a criminal offense that requires an ultimate determination of guilt beyond a reasonable doubt — by a jury, in the same proceeding. This makes all the difference. And as district courts have engaged with 18 U.S.C. § 924(c) on the front lines, they have often “appl[ied] the substantial risk element ... to the actual conduct in the present case.” See United States v. Checora, No. 2:14cr457DAK, 155 F.Supp.3d 1192, 1203, 2015 WL 9305672, at *9 (D. Utah Dec. 21, 2015); see also United States v. Prickett, No. 3:14-CR-30018, 2015 WL 5884904, at *2 (W.D. Ark. Oct. 8, 2015).

We understand Taylor, then, as applying Johnson’s real-world conduct exception to uphold the constitutionality of 18 U.S.C. § 924(c)(3)(B). See Taylor, 814 F.3d at 376 (noting that “[t]he jury found that Taylor murdered Luck in the course of committing two crimes of violence”) (emphasis added). Besides, the government’s reading of Taylor has been undercut by the Supreme Court’s intervening decision in Welch. As the Court made clear this term, the ACCA’s vagueness “rests in large part on its operation under the categorical approach.” Welch, 136 S.Ct. at 1262. That residual clause did not fail for the reasons latched onto by the government. See Taylor, 814 F.3d at 376-78. Rather, it failed “because applying [the serious potential risk] standard under the categorical approach required courts to assess the hypothetical risk posed by an abstract generic version of the offense.” See Welch, 136 S.Ct. at 1262. Taylor did not have the benefit of the Court’s guidance in this regard. Any dictum in that decision, purporting to address the constitutionality of the INA’s residual clause, is simply that.

In a last ditch effort, the government suggests that the INA’s definition of crime of violence has not generated widespread confusion or proven unworkable in practice. This is patently not the case. See Padilla, 559 U.S. at 378, 380, 130 S.Ct. 1473 (Alito, J., concurring in judgment) (noting that the aggravated felony inquiry under the INA is “complicated by ... significant variations” among “Immigration and Customs Enforcement, the [BIA], and [courts of appeals] and district courts considering immigration-law and criminal-law issues”). Even so, the government’s argument ignores the realities of judicial review. We find it entirely unsurprising that the INA has generated less conflicting case law than the ACCA, as there are more criminal appeals than petitions for review of immigration orders. The Supreme Court’s docket is almost entirely discretionary, see Singleton v. Commissioner, 439 U.S. 940, 942, 99 S.Ct. 335, 58 L.Ed.2d 335 (1978) (Stevens, J., opinion respecting denial of certiorari), and the courts of appeals have narrow jurisdiction over petitions for review of immigration orders, compare 28 U.S.C. § 1291, with 8 U.S.C. § 1252. At any rate, the government mistakes a correlation for causation; conflicting judicial interpretations only provide ex post “evidence of vagueness.” Johnson, 135 S.Ct. at 2558; see also Sykes v. United States, 564 U.S. 1, 131 S.Ct. 2267, 2286, 180 L.Ed.2d 60 (2011) (Scalia, J., dissenting).

Determining whether a particular offense is an aggravated felony is already “quite complex.” See Padilla, 559 U.S. at 377-78, 130 S.Ct. 1473 (Alito, J., concurring in judgment). The INA’s residual definition of “crime of violence” makes that inquiry hopelessly indeterminate. From a non-citizen’s perspective, this provision substitutes guesswork and caprice for fair notice and predictability. If the residual clause cannot be applied in a “principled and objective” manner by judges, see Johnson, 135 S.Ct. at 2558, we fail to see how non-citizens and their counsel will be able to anticipate the immigration consequences of criminal convictions, see Mellouli, 135 S.Ct. at 1987; Padilla, 559 U.S. at 366, 130 S.Ct. 1473.

Shuti is set to begin “a life sentence of exile from what has [been his] home” since age 13, deprived of his “established means of livelihood,” and separated from “his family of American citizens.” See Jordan, 341 U.S. at 243, 71 S.Ct. 703 (Jackson, J., dissenting). Before imposing this penalty, the Due Process Clause requires more definite standards. We therefore find the INA’s residual definition of “crime of violence,” 8 U.S.C. § 1101(a)(43)(F); 18 U.S.C. § 16(b), void for vagueness.

IV.

The petition for review is granted, the order of removal is vacated, and the case is remanded to the BIA for further proceedings consistent with this opinion.

1 . We consider Shuti's constitutional challenge despite his concession of removability. See Hanna v. Holder, 740 F.3d 379, 387 (6th Cir. 2014). And for good reason: "What suffices for waiver depends on the nature of the right at issue.” New York v. Hill, 528 U.S. 110, 114, 120 S.Ct. 659, 145 L.Ed.2d 560 (2000); see also Singleton v. Wulff, 428 U.S. 106, 121, 96 S.Ct. 2868, 49 L.Ed.2d 826 (1976) (holding that prudential restrictions on appellate review are "left primarily to the discretion of the courts of appeals, to be exercised on the facts of individual cases”). Shuti’s concession of removability under the INA says nothing of whether those statutory provisions are valid under the Constitution. Further, enforcing Shuti’s concession in this instance would be inconsistent with an ordinary understanding of issue waiver. Several factors suggest that review is appropriate here: (1) Shuti's claim is a pure question of law; (2) Johnson was an intervening change in law; (3) Shuti actually raised his Johnson claim before the BIA; and (4) this Court has explicit jurisdiction to review constitutional claims, while the BIA lacks such authority. See Hayward v. Cleveland Clinic Found., 759 F.3d 601, 615 (6th Cir. 2014).