United States v. Vargas

74 F.4th 673

• Court: Court of Appeals for the Fifth Circuit
• Decided: July 24, 2023
• Docket: 21-20140
• Status: Published

Opinion

Case: 21-20140 Document: 00516831915 Page: 1 Date Filed: 07/24/2023

United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit

FILED July 24, 2023 No. 21-20140 Lyle W. Cayce Clerk

United States of America,

Plaintiff—Appellee,

versus

Andres Vargas,

Defendant—Appellant.

Appeal from the United States District Court for the Southern District of Texas USDC No. 4:20-CR-80-1

Before Richman, Chief Judge, and Jones, Smith, Stewart, Elrod, Southwick, Haynes, Graves, Higginson, Willett, Ho, Duncan, Engelhardt, Oldham, Wilson, and Douglas, Circuit Judges. Stuart Kyle Duncan, Circuit Judge: †

† Ten judges join all or part of this opinion. Four join in full: Chief Judge Richman and Judges Smith, Southwick, and Engelhardt. Judges Jones and Oldham join all but Part III(C). Judges Higginson and Ho join all but Parts III(C) and III(D). Judge Willett joins all but Part III(D). Judge Wilson joins only Parts III(A) and III(D). Case: 21-20140 Document: 00516831915 Page: 2 Date Filed: 07/24/2023

No. 21-20140

Federal law imposes harsher sentences on people who commit multiple drug crimes. See U.S. Sent’g Guidelines Manual § 4B1.1(a) (U.S. Sent’g Comm’n 2018). 1 But what about conspiracies? Do people who have engaged in multiple drug conspiracies also get more prison time? That is the question before us. Our task would be easy if a statute settled the matter. No such luck: the relevant definition in the United States Sentencing Guidelines—“controlled substance offense”—does not say one way or the other whether it includes conspiracies. See § 4B1.2(b). But the official commentary says, yes, conspiracies are included. See § 4B1.2 cmt. n.1. So, the solution to our problem depends in significant measure on how much weight to give the guidelines commentary. Fortunately, the Supreme Court has told us, and the answer is: “Quite a lot.” In Stinson v. United States, 508 U.S. 36 (1993), the Court held that the guidelines commentary is “authoritative unless it violates the Constitution or a federal statute, or is inconsistent with, or a plainly erroneous reading of, that guideline.” Id. at 38. The commentary here has none of those flaws. In particular, the commentary is not “inconsistent with” the guideline merely because it mentions conspiracies and the guideline’s definition does not. So, Stinson requires us to follow the commentary. Some of our sister circuits contend the Supreme Court replaced Stinson’s highly deferential standard with a less deferential one in Kisor v. Wilkie, 139 S. Ct. 2400 (2019). 2 Others disagree and continue to apply

1 Citations are to the United States Sentencing Guidelines unless otherwise noted. 2 See United States v. Nasir, 17 F.4th 459 (3d Cir. 2021) (en banc); United States v. Campbell, 22 F.4th 438 (4th Cir. 2022); United States v. Riccardi, 989 F.3d 476 (6th Cir. 2021); United States v. Castillo, 69 F.4th 648 (9th Cir. 2023); United States v. Dupree, 57 F.4th 1269 (11th Cir. 2023) (en banc).

2 Case: 21-20140 Document: 00516831915 Page: 3 Date Filed: 07/24/2023

Stinson. 3 We agree with the second group. Stinson sets out a deference doctrine distinct from the one refined by Kisor. Until the Supreme Court overrules Stinson, then, our duty as an inferior court is to apply it faithfully. But even if we are wrong, and Kisor did alter Stinson, we would reach the same conclusion. That is because applying the traditional tools of construction—text, structure, history, and purpose—shows that the commentary reasonably reads “controlled substance offense” to include conspiracies. See Kisor, 139 S. Ct. at 2415. So, even under Kisor’s less deferential approach, we would still defer to the commentary. The sentence under review is therefore AFFIRMED. I. Background Andres Vargas tried to buy five kilograms of cocaine from an undercover agent. His plan began in January 2020, when two Mexican nationals put Vargas in touch with the agent. Vargas and a co-conspirator were to pay $125,000 in exchange for the drugs. After agreeing to meet the agent in a Wal-Mart parking lot to carry out the transaction, Vargas and his co-conspirator were arrested. Vargas later pled guilty to conspiring to possess cocaine with intent to distribute it, in violation of 21 U.S.C. §§ 846, 841(a)(1), and 841(b)(1)(B). Had this been his first offense, Vargas likely would have faced a guidelines range of 100–125 months in prison. 4 But his criminal history

3 See United States v. Lewis, 963 F.3d 16 (1st Cir. 2020); United States v. Tabb, 949 F.3d 81 (2d Cir. 2020); United States v. Moses, 23 F.4th 347 (4th Cir. 2022); United States v. Smith, 989 F.3d 575 (7th Cir. 2021), cert. denied, 142 S. Ct. 488 (2021); United States v. Maloid, 71 F.4th 795 (10th Cir. 2023). 4 Specifically, Vargas’s Presentence Report (“PSR”) calculated his base offense level as 30, based on the quantity of drugs involved. His accepting responsibility reduced

3 Case: 21-20140 Document: 00516831915 Page: 4 Date Filed: 07/24/2023

triggered an enhancement. Previously, Vargas had been convicted of (1) possessing amphetamine with intent to distribute it and (2) conspiring to possess methamphetamine with intent to manufacture and distribute it. Because these and the instant offense were classified as controlled substance offenses, Vargas was deemed a career offender under § 4B1.1, yielding a higher range of 188–235 months. 5 Vargas objected to his career offender designation, arguing that inchoate crimes, 6 such as his conspiracy convictions, do not qualify as controlled substance offenses under the definition in § 4B1.2(b). The district court overruled Vargas’s objections and sentenced him to the low end of the enhanced range: 188 months, followed by four years of supervised release. Vargas appealed. As before, he argued that conspiracies cannot qualify as controlled substance offenses because the guideline definition excludes inchoate crimes. The commentary’s inclusion of conspiracies, Vargas asserted, conflicts with the definition. A panel of this court rejected that argument. See United States v. Vargas, 35 F.4th 936 (5th Cir. 2022), vacated by 45 F.4th 1083 (5th Cir. 2022). It explained that our circuit previously “held that § 4B1.1’s career-offender enhancement lawfully includes inchoate offenses.” Id. at 938 (citing United States v. Lightbourn, 115 F.3d 291, 293 (5th

his offense level by three. Alongside this, the PSR assessed eight criminal history points, which normally yields a criminal history category of IV. That, when combined with a total offense level of 27, results in a range of 100–125 months. 5 The enhancement increased Vargas’s offense level to 31 and his criminal history category to VI. See § 4B1.1(b). 6 An “inchoate crime” is one that involves “[a] step toward the commission of another crime, the step in itself being serious enough to merit punishment.” Inchoate Offense, Black’s Law Dictionary (11th ed. 2019). The term includes conspiracies and attempts. Ibid. It does not include aiding and abetting, which “is simply a different method for demonstrating liability for the substantive offense.” United States v. Rabhan, 540 F.3d 344, 349 (5th Cir. 2008).

4 Case: 21-20140 Document: 00516831915 Page: 5 Date Filed: 07/24/2023

Cir. 1997)). This precedent bound it to affirm Vargas’s sentence, even if Kisor might have raised questions about the amount of deference due to the guidelines commentary under Stinson. Id. at 940. Vargas then petitioned for en banc rehearing, which we granted. II. Standard of Review We review a district court’s interpretation of the Sentencing Guidelines de novo. United States v. Cortez-Gonzalez, 929 F.3d 200, 203 (5th Cir. 2019). III. Discussion To qualify as a career offender under the guidelines, a defendant must have previously committed “at least two prior felony convictions of either a crime of violence or a controlled substance offense.” § 4B1.1(a)(3). 7 The guidelines define “controlled substance offense” in this way: The term “controlled substance offense” means an offense under federal or state law, punishable by imprisonment for a term exceeding one year, that prohibits the manufacture, import, export, distribution, or dispensing of a controlled substance (or a counterfeit substance) or the possession of a controlled substance (or a counterfeit substance) with intent to manufacture, import, export, distribute, or dispense. § 4B1.2(b). According to the commentary, this definition “include[s] the offenses of aiding and abetting, conspiring, and attempting to commit such

7 Additionally, the defendant must have been “at least eighteen years old” when he committed the instant offense, § 4B1.1(a)(1), and that offense must have been “either a crime of violence or a controlled substance offense,” § 4B1.1(a)(2). It is undisputed that Vargas was at least eighteen when he committed the instant offense.

5 Case: 21-20140 Document: 00516831915 Page: 6 Date Filed: 07/24/2023

offenses.” § 4B1.2 cmt. n.1. We must decide what weight, if any, to give this commentary. 8 Our discussion proceeds as follows. In Part III(A), we ask which framework—Stinson or Kisor—governs. We conclude Stinson continues to bind us. In Part III(B), we ask whether Stinson compels deference to the commentary. We conclude it does, because the commentary is not “inconsistent” with the guideline as Stinson used that term. In Part III(C), we explain that, even under Kisor’s less deferential framework, deference to the commentary is still warranted. Finally, in Part III(D), we explain why the rule of lenity does not affect our interpretation of the guidelines. III(A). Stinson, not Kisor 9 Inferior courts must follow directly applicable Supreme Court precedent that has not been overruled or modified. See, e.g., Freedom From Religion Found. v. Mack, 4 F.4th 306, 315 (5th Cir. 2021) (“We are bound to follow the Supreme Court precedent that most squarely controls our case.”). Stinson squarely applies here and has not been overruled or modified. So, follow it we must.

8 The Commission has recently proposed an amendment to the guidelines that explicitly includes inchoate offenses in the text of § 4B1.2. See Sentencing Guidelines for United States Courts, 88 Fed. Reg. 28254, 28275–76 (May 3, 2023). This amendment has been submitted to Congress for review and will take effect on November 1, 2023. Id. at 28254. In the meantime, however, we must still decide the issue before us under the current guideline. And, even though the Commission has settled this issue going forward, it “cannot, on its own, resolve the dispute about what deference courts should give to the commentary.” Dupree, 57 F.4th at 1289 n.6 (Grant, J., concurring in the judgment). 9 This Part represents the views of eleven out of sixteen judges: Chief Judge Richman, and Judges Jones, Smith, Southwick, Higginson, Willett, Ho, Duncan, Engelhardt, Oldham, and Wilson.

6 Case: 21-20140 Document: 00516831915 Page: 7 Date Filed: 07/24/2023

Stinson held that “commentary in the Guidelines Manual that interprets or explains a guideline is authoritative unless it violates the Constitution or a federal statute, or is inconsistent with, or a plainly erroneous reading of, that guideline.” 508 U.S. at 38. Commentary meeting those conditions is “binding” and “control[ling]” on courts. 10 That is so even for “unambiguous” guidelines. Id. at 44. Failing to follow the commentary thus “constitute[s] ‘an incorrect application of the sentencing guidelines.’” Id. at 43 (quoting 18 U.S.C. § 3742(f)(1)). Moreover, Stinson decided all this in a case involving commentary to the same career offender guideline at issue in the case before us today. Id. at 38 (discussing Stinson’s “career offender” sentence under “[§] 4B1.1”). As day follows night, this case is governed by Stinson. Just as in Stinson, we address commentary interpreting the career offender guideline, § 4B1.1. That commentary says that a “controlled substance offense,” as defined in § 4B1.2, includes a conspiracy to commit such an offense. See § 4B1.2 cmt. n.1. Under Stinson’s framework, that commentary “controls” unless it is “inconsistent with, or a plainly erroneous reading of” the guideline definition. 508 U.S. at 38. We can avoid applying Stinson only if the Supreme Court has overruled or modified it. See, e.g., Lefebure v. D’Aquilla, 15 F.4th 650, 660– 61 (5th Cir. 2021) (“[T]he only court that can overturn a Supreme Court precedent is the Supreme Court itself.” (citations omitted)); Nat’l Coal. for Men v. Selective Serv. Sys., 969 F.3d 546, 547 (5th Cir. 2020) (“[O]nly the Supreme Court may revise its precedent.”). No one claims Stinson has been

10 See Stinson, 508 U.S. at 42 (“[The Court of Appeals’] conclusion that the commentary now being considered is not binding on the courts was error.”); ibid. (when commentary “interpret[s]” a guideline or “explain[s] how it is to be applied,” the commentary “controls” (quoting § 1B1.7)).

7 Case: 21-20140 Document: 00516831915 Page: 8 Date Filed: 07/24/2023

overruled. Vargas argues only that Stinson was modified by a 2019 Supreme Court decision, Kisor v. Wilkie, thereby decreasing the deference due to the commentary. See 139 S. Ct. 2400. The government agrees with Vargas on this point, as do some of our sister circuits. 11 Of course, we must determine for ourselves the controlling legal framework. See, e.g., Djie v. Garland, 39 F.4th 280, 286 n.3 (5th Cir. 2022) (“As our court long ago explained, ‘it is well settled that a court is not bound to accept as controlling stipulations as to questions of law.’” (quoting Equitable Life Assur. Soc’y of U.S. v. MacGill, 551 F.2d 978, 983 (5th Cir. 1977))). Under Kisor, before a court may defer to an agency’s interpretation of its own regulation, it must “exhaust all the ‘traditional tools’ of construction” and find the regulation “genuinely ambiguous.” 139 S. Ct. at 2415. This formulation clarified the deference rule from an older decision, Bowles v. Seminole Rock & Sand Co., 325 U.S. 410 (1945) (sometimes called “Seminole Rock” 12 for short). See Kisor, 139 S. Ct. at 2415 (discussing Seminole Rock). And Kisor has been sensibly interpreted as lowering the amount of deference given to agency interpretations of regulations. 13

11 See Nasir, 17 F.4th at 471; Campbell, 22 F.4th at 444–45; Riccardi, 989 F.3d at 485; Castillo, 69 F.4th at 655–56; Dupree, 57 F.4th at 1275. Other circuits continue to apply Stinson. See Lewis, 963 F.3d at 22–24; Tabb, 949 F.3d at 87; Moses, 23 F.4th at 351–58; Smith, 989 F.3d at 584; Maloid, 71 F.4th at 805–08. For a cogent treatment of this general subject, see John S. Acton, The Future of Judicial Deference to the Commentary of the United States Sentencing Guidelines, 45 Harv. J.L. & Pub. Pol’y 349 (2022) (arguing for the continued vitality of Stinson). 12 It is also sometimes called “Auer” deference after a later case. See Auer v. Robbins, 519 U.S. 452 (1997). 13 See, e.g., Kisor, 139 S. Ct. at 2443 (Gorsuch, J., concurring in the judgment); Paul J. Larkin, Jr., Agency Deference After Kisor v. Wilkie, 18 Geo. J.L. & Pub. Pol’y 105, 118 (2020) (“The Kagan opinion lowers the reader’s expectation as to the amount of deference an agency’s rule-interpretation should receive.”).

8 Case: 21-20140 Document: 00516831915 Page: 9 Date Filed: 07/24/2023

Vargas points out that Stinson itself drew from Seminole Rock. True enough. Stinson viewed the commentary as “akin to an agency’s interpretation of its own legislative rules.” 508 U.S. at 45. So, it borrowed Seminole Rock’s rule that such an interpretation “must be given ‘controlling weight unless it is plainly erroneous or inconsistent with the [guideline].’” Ibid. (quoting Seminole Rock, 325 U.S. at 414). Because Kisor reformulated Seminole Rock, Vargas argues that Kisor necessarily reformulated Stinson, too. That is, when Kisor curtailed the deference due to an agency’s interpretation of a regulation (Seminole Rock), it also curtailed the deference due to the commentary’s interpretation of a guideline (Stinson). Some of our sister circuits have adopted this rationale. See, e.g., Dupree, 57 F.4th at 1275 (“[T]he only way to harmonize the two cases is to conclude that Kisor’s gloss on Auer and Seminole Rock applies to Stinson.”). We disagree for several reasons. First, nothing in Kisor suggests it meant to modify Stinson. Nowhere does Kisor mention the Sentencing Guidelines, the Commission, or the commentary. See 139 S. Ct. 2400. Instead, Kisor examined whether it should defer to an agency’s (specifically, the Department of Veterans Affairs’) “reasonable readings of genuinely ambiguous regulations.” Id. at 2408. Kisor did not discuss Stinson at all: it merely included Stinson in a footnote string- cite of sixteen cases described as “decisions applying Seminole Rock deference.” Id. at 2411 n.3. That footnote signals no intention to change Stinson. Quite the opposite. The footnote is merely descriptive and is not even joined by a Court majority. See id. at 2407, 2410 (only four Justices join Part II-A). What’s more, another part of Kisor—this one joined by a majority—refuses to overrule the “long line of precedents” that includes Stinson. See id. at 2422 (citing id. at 2411 nn.2–3). Far from altering Stinson, then, Kisor goes out of its way to leave it undisturbed.

9 Case: 21-20140 Document: 00516831915 Page: 10 Date Filed: 07/24/2023

Second, although Stinson borrowed from Seminole Rock, Stinson deference differs from Seminole Rock in important ways. As our Eleventh Circuit colleague has observed, the two doctrines are not “interchangeable.” Dupree, 57 F.4th at 1284 (Grant, J., concurring in the judgment). For instance, under Stinson, the commentary controls even unambiguous guidelines. See 508 U.S. at 44 (stating “commentary . . . provides concrete guidance as to how even unambiguous guidelines are to be applied in practice” (emphasis added)). Not so with Seminole Rock, which required deference only when “the meaning of the words used [was] in doubt.” 325 U.S. at 414. Another difference: under Stinson, the Commission can interpret a guideline in ways that conflict with prior judicial interpretations. See 508 U.S. at 46 (holding “prior judicial constructions of a particular guideline cannot prevent the Commission from adopting a conflicting interpretation”); accord Dupree, 57 F.4th at 1285 (Grant, J., concurring in the judgment). Not so with agencies. An agency’s interpretation cannot trump a court’s prior interpretation of an unambiguous statute. See Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 982–83 (2005). So, while Stinson drew from Seminole Rock, the two doctrines were distinct from the beginning and remain distinct today. See, e.g., Moses, 23 F.4th at 356 (“[E]ven though the two cases addressed analogous circumstances, Stinson . . . appl[ies] when courts are addressing Guidelines commentary, while Kisor applies when courts are addressing executive agency interpretations of legislative rules.”). It does not follow that refining Seminole Rock automatically refines Stinson. Third, the Sentencing Commission and administrative agencies are different animals. Yes, Stinson likened the Commission to an agency, but it cautioned that “the analogy is not precise.” 508 U.S. at 44. Quite right. The Commission is “a peculiar institution within the framework of our

10 Case: 21-20140 Document: 00516831915 Page: 11 Date Filed: 07/24/2023

Government.” Mistretta v. United States, 488 U.S. 361, 384 (1989). Unlike an executive branch agency, the Commission lodges in the judicial branch. See 28 U.S.C. § 991(a). Its seven members are appointed by the President and confirmed by the Senate, and at least three must be federal judges. Ibid. So, while “the role of other federal agencies is typically executive,” the Commission is “judicial in nature.” Moses, 23 F.4th at 355. Instead of addressing the public, as agencies do, the Commission addresses federal judges. “[I]ts Guidelines Manual, including its policy statements and commentary, is directed at providing guidance to district judges tasked with the duty of imposing an individualized sentence on a criminal defendant.” Ibid. (citing United States v. Booker, 543 U.S. 220, 245 (2005)). And unlike an agency’s gloss on its regulations, the Commission’s role in promulgating and interpreting guidelines is sanctioned by statute. 14 We therefore agree with the Fourth Circuit that “[t]hese differences justify a distinct approach in considering Guidelines commentary, on the one hand, and an agency’s interpretation of its legislative rules, on the other.” Ibid. In this vein, one other point deserves mention. A core function of the Commission is to avoid “unwarranted sentencing disparities.” See 28 U.S.C. § 991(b)(1)(B). The commentary tangibly advances that goal by “amplify[ing] and explain[ing] how the Guidelines are to be applied.” Moses, 23 F.4th at 357 (citing § 1B1.7). As more than one of our colleagues have observed, though, consistency in applying the guidelines would be frustrated

14 See generally Stinson, 508 U.S. at 40–41 (discussing Commission’s authority under the Sentencing Reform Act of 1984); see also, e.g., 28 U.S.C. § 994 (Sentencing Reform Act authorizing Commission’s guidelines and policy statements); 18 U.S.C. § 3553(b)(1) (permitting courts to consider only “sentencing guidelines, policy statements, and official commentary” in deciding whether to depart from guidelines range); 18 U.S.C. § 3742 (discussing the role of guidelines ranges in judicial review of sentences).

11 Case: 21-20140 Document: 00516831915 Page: 12 Date Filed: 07/24/2023

if Kisor governed whether the commentary controlled. 15 Only commentary to “genuinely ambiguous” guidelines would bind courts, with the rest safely ignored. If the Supreme Court meant to layer this new complexity onto an already complex system, one would expect it to say so plainly. Yet we are supposed to believe that Kisor did this in an opinion that did not even mention the Sentencing Guidelines and that refused to overrule Stinson. See Kisor, 139 S. Ct. at 2422, 2411 n.3. That is most unlikely. Ending this section where we began, we again state that it is our duty to follow squarely applicable Supreme Court precedent. Stinson is that. Distilled to its essence, Vargas’s contention is that Kisor undermined Stinson’s foundations because Stinson built on Seminole Rock. Whether that is true, though, is the Supreme Court’s business and not ours. 16 Perhaps Kisor is the coming-soon trailer for a rethinking of Stinson. Or perhaps the Sentencing Commission’s unique nature and role warrant a distinct deference doctrine untouched by Kisor. We express no view on the matter. Our job, as an inferior court, is to adhere strictly to Supreme Court precedent, whether or not we think a precedent’s best days are behind it. See Mallory v. Norfolk S. Ry. Co., 143 S. Ct. 2028, 2038 (2023) (“[A] lower court ‘should follow the case which directly controls’ . . . even if the lower court

15 See Moses, 23 F.4th at 357 (“Were we now to relegate commentary to a status where it could be considered only when the relevant Guideline is genuinely ambiguous, we would negate much of the Commission's efforts in providing commentary to fulfill its congressionally designated mission.”); Dupree, 57 F.4th at 1287 (Grant, J., concurring in the judgment) (applying Kisor would “exacerbate the degree of sentencing discrepancies”). 16 See Hohn v. United States, 524 U.S. 236, 252–53 (1998) (“[The Supreme Court’s] decisions remain binding precedent until [the Supreme Court] see[s] fit to reconsider them, regardless of whether subsequent cases have raised doubts about their continuing vitality.”); see also Dupree, 57 F.4th at 1293 (Luck, J., dissenting) (rejecting the “transitive” argument that “because X relied on Y, and Y has been clarified by Z, then X must also have been clarified by Z”).

12 Case: 21-20140 Document: 00516831915 Page: 13 Date Filed: 07/24/2023

thinks the precedent is in tension with ‘some other line of decisions.’” (quoting Rodriguez de Quijas v. Shearson/Am. Exp., Inc., 490 U.S. 477, 484 (1989))). So, we proceed to apply Stinson to the commentary at issue here. III(B). The commentary controls under the Stinson framework. 17 Under Stinson, the question is whether the commentary at issue is “inconsistent with” the applicable guideline. 508 U.S. at 38. As noted, that guideline defines a “controlled substance offense” as an “offense . . . that prohibits the manufacture, import, export, distribution, or dispensing of a controlled substance . . . or the possession of a controlled substance . . . with intent to manufacture, import, export, distribute, or dispense.” § 4B1.2(b). The commentary, in turn, states that this definition “include[s] the offenses of aiding and abetting, conspiring, and attempting to commit such offenses.” § 4B1.2 cmt. n.1. Vargas argues that, even under Stinson’s more deferential approach, the commentary clashes with § 4B1.2(b) by including conspiracies. Invoking the expressio unius or “negative-implication” canon, Vargas contends that § 4B1.2(b)’s failure to list inchoate crimes like conspiracies means they are excluded, depriving the commentary of Stinson deference. See NLRB v. SW Gen., Inc., 580 U.S. 288, 302 (2017) (“The interpretive canon, expressio unius est exclusio alterius, [means] ‘expressing one item of an associated group or series excludes another left unmentioned.’” (cleaned up)); see also Antonin Scalia & Bryan A. Garner, Reading Law: The

17 This Part represents the views of ten out of sixteen judges: Chief Judge Richman, and Judges Jones, Smith, Southwick, Higginson, Willett, Ho, Duncan, Engelhardt, and Oldham.

13 Case: 21-20140 Document: 00516831915 Page: 14 Date Filed: 07/24/2023

Interpretation of Legal Texts 107 (2012) (discussing “negative- implication canon”). This question has split the circuits. Some courts agree with Vargas. 18 Others disagree, finding no inconsistency between the commentary and § 4B1.2(b) and thus deferring to the commentary. 19 We side with the second group: the commentary is not inconsistent with the guideline definition. So, Stinson requires us to defer to the commentary. We begin by asking what Stinson means by commentary that is “inconsistent” with a guideline. Helpfully, Stinson tells us: “If . . . commentary and the guideline it interprets are inconsistent in that following one will result in violating the dictates of the other, the Sentencing Reform Act itself commands compliance with the guideline.” Stinson, 508 U.S. at 43 (emphasis added) (citing 18 U.S.C. §§ 3553(a)(4), (b)). Accordingly, Stinson criticizes courts that “refuse[] to follow commentary in situations falling short of such flat inconsistency.” Ibid. In other words, “inconsistency” demands more than merely showing that the commentary’s reading of the guideline is incorrect or implausible. Rather, there must be some irreconcilable variance (“flat inconsistency”) between the two. See, e.g., Moses, 23 F.4th at 354 (observing that Stinson

18 See Campbell, 22 F.4th at 443–46; United States v. Havis, 927 F.3d 382, 386–87 (6th Cir. 2019) (en banc) (per curiam); Castillo, 69 F.4th at 657–58; United States v. Winstead, 890 F.3d 1082, 1090–91 (D.C. Cir. 2018); see also Dupree, 57 F.4th at 1287 (Grant, J., concurring in the judgment) (arguing under Stinson that this is “a rare case of true incompatibility between commentary and its underlying guideline”). 19 See Lewis, 963 F.3d at 21–23; United States v. Richardson, 958 F.3d 151, 154–55 (2d Cir. 2020); Smith, 989 F.3d at 583–85; United States v. Merritt, 934 F.3d 809, 811–12 (8th Cir. 2019); United States v. Lange, 862 F.3d 1290, 1295 (11th Cir. 2017), overruled on other grounds by Dupree, 57 F.4th 1269; see also Dupree, 57 F.4th at 1296 (Luck, J., dissenting) (applying Stinson to conclude that conspiring to possess with intent to distribute heroin and cocaine remains a controlled substance offense).

14 Case: 21-20140 Document: 00516831915 Page: 15 Date Filed: 07/24/2023

“defined ‘inconsistent’ strictly”). Our circuit has applied Stinson in this strict fashion. For instance, we have found inconsistency where the commentary rendered a guideline functionally “inoperable.” United States v. Pringler, 765 F.3d 445, 454 (5th Cir. 2014). Yet even where there was “apparent conflict” between guideline and commentary, we understood our role under Stinson as seeking to “reconcile[]” the two to avoid the need to “declar[e] which must prevail over the other.” United States v. Clayton, 172 F.3d 347, 355 (5th Cir. 1999). 20 Such generous deference to the commentary follows from the role Congress assigned the Sentencing Commission. Stinson, 508 U.S. at 45. Under the Sentencing Reform Act, Congress gave the Commission broad authority to write, review, and revise the guidelines. See id. at 45–46 (citing 28 U.S.C. § 994(o), (w)). Importantly, the Commission can revise the guidelines in two ways—either by amending the guidelines themselves or by “amendment of the commentary.” Id. at 46. “Amended commentary,” Stinson tells us, “is binding on the federal courts even though it is not reviewed by Congress,” 21 and takes precedence even over “prior judicial constructions of a particular guideline.” Ibid. Accordingly, courts “can presume that the interpretations of the guidelines contained in the commentary represent the most accurate indications of how the Commission deems that the guidelines should be applied.” Id. at 45.

20 Other circuits have also set a high bar for finding inconsistency. See, e.g., United States v. Cramer, 777 F.3d 597, 606 (2d Cir. 2015) (following guideline over commentary where commentary would leave the guideline without “any practical effect”). 21 To be precise, revisions to the commentary are not required to be submitted to Congress for review. See 28 U.S.C. § 994(x). But, as we note below, the commentary here was reviewed by Congress. See Dupree, 57 F.4th at 1281 (Pryor, C.J., concurring) (citing Amendments to the Sentencing Guidelines for United States Courts, 54 Fed. Reg. 21348, 21379 (May 17, 1989)).

15 Case: 21-20140 Document: 00516831915 Page: 16 Date Filed: 07/24/2023

Bearing all that in mind, we turn to Vargas’s argument that the commentary, by including conspiracies and attempts, is inconsistent with the guideline’s definition of controlled substance offense. We conclude Vargas fails to overcome the ample deference Stinson affords the commentary. First and foremost, the guideline says nothing one way or the other about conspiracies and attempts. It states only that a “controlled substance offense” means an “offense that prohibits” various drug-related activities, like manufacturing or distributing narcotics. § 4B1.2(b). In light of that, one can follow the commentary (by counting conspiracies and attempts as controlled substance offenses) without “violating the dictates of the [guideline].” Stinson, 508 U.S. at 43. The Seventh Circuit used this reasoning to find no conflict between the commentary and the § 4B1.2(b) definition. See Smith, 989 F.3d at 585 (concluding “§ 4B1.2’s Application Note 1 is authoritative and that ‘controlled substance offense’ includes inchoate offenses” (citing United States v. Adams, 934 F.3d 720, 729–30 (7th Cir. 2019))). The court saw no inconsistency because the definition “does not tell us, one way or another, whether inchoate offenses are included or excluded.” Ibid. (quoting United States v. Raupp, 677 F.3d 756, 759 (7th Cir. 2012), overruled on other grounds by United States v. Rollins, 836 F.3d 737 (7th Cir. 2016)). 22 Other circuits agree. See, e.g., United States v. Piper, 35 F.3d 611, 617 (1st Cir. 1994) (commentary not inconsistent with § 4B1.2 because it “neither excludes any offenses expressly enumerated in the guideline, nor calls for the inclusion of any

22 One case Smith relied on, Raupp, involved a “crime of violence” under § 4B1.2(a) rather than a “controlled substance offense” under § 4B1.2(b). For our purposes, though, there is no difference between the two definitions—neither says it includes or excludes inchoate offenses. Smith itself addressed a controlled substance offense under § 4B1.2(b). Ibid.

16 Case: 21-20140 Document: 00516831915 Page: 17 Date Filed: 07/24/2023

offenses that the guideline expressly excludes” (emphasis added)); United States v. Jackson, 60 F.3d 128, 131 (2d Cir. 1995) (following the commentary because guidance one way or another on inchoate crimes “does not appear in an actual guideline”). We find this reasoning sound. But some of our sister circuits do not. Relying largely on the expressio unius canon, they reason that the commentary improperly “adds” conspiracies and attempts to a definition that “clearly excludes inchoate offenses.” Winstead, 890 F.3d at 1090–92; accord Nasir, 17 F.4th at 471–72; Campbell, 22 F.4th at 444; Havis, 927 F.3d at 386; Castillo, 69 F.4th at 651; see also Dupree, 57 F.4th at 1287–88 (Grant, J., concurring in the judgment). On this view, one cannot follow the commentary without violating the guideline, making the two inconsistent. We respectfully disagree with our colleagues. Expressio unius teaches that “[t]he expression of one thing implies the exclusion of others.” Scalia & Garner, supra, at 107; see also In re Bourgeois, 902 F.3d 446, 447–48 (5th Cir. 2018) (discussing canon). “[T]he canon does not apply ‘unless it is fair to suppose that [the drafter] considered the unnamed possibility and meant to say no to it.’” United States v. Cartagena-Lopez, 979 F.3d 356, 362 (5th Cir. 2020) (quoting Marx v. Gen. Rev. Corp., 568 U.S. 371, 381 (2013)). These two inquiries are therefore helpful: “(1) Whether the statutory text communicates exclusivity, and (2) whether the included term goes hand in hand with the missing term, allowing the inference that the omission has interpretive force.” Ibid. (citing Barnhart v. Peabody Coal Co., 537 U.S. 149, 168–69 (2003)). 23 “Without these

23 By asking the second question, the dissent claims we are applying a “limited exception” to the expressio unius canon. Post, at 7. Not so. Properly applying the canon has always required asking whether the relationship between named and unnamed terms implies an intent to exclude the unnamed term. See, e.g., Marx v. Gen. Rev. Corp., 568 U.S.

17 Case: 21-20140 Document: 00516831915 Page: 18 Date Filed: 07/24/2023

clues, we cannot discern any meaning from statutory omissions.” Ibid.; see also Bourgeois, 902 F.3d at 448 (“Context may indicate that Congress did not wish for an express provision of one thing to work towards the exclusion of another.”). The expressio unius canon does not apply here. Contrary to the views of some other circuits, cf. Winstead, 890 F.3d at 1091, the definition does not “clearly exclude” conspiracies and attempts, see, e.g., Dupree, 57 F.4th at 1295 n.1 (Luck, J., dissenting) (“[S]ection 4B1.2(b)’s text does not exclude conspiracy, attempt, and aiding and abetting crimes.”). It is more accurate to say that the definition does not mention them. See Smith, 989 F.3d at 585 (the definition “does not tell us, one way or another, whether inchoate offenses are included or excluded”). Not mentioning something does not necessarily mean excluding it. The context must justify that inference. See Barnhart, 537 U.S. at 168 (the canon applies only when the “statutory listing or grouping . . . justif[ies] the inference that items not mentioned were excluded by deliberate choice”). Here it does not. Had the drafters wanted to exclude inchoate offenses from § 4B1.2(b), they could have easily said so. For instance, they could have stated that conspiracies and attempts are “excluded,” just as other guidelines “exclude” specific crimes for various reasons. See Dupree, 57 F.4th at 1295 n.1 (Luck, J., dissenting) (noting where guidelines “exclude” crimes) (citing § 3D1.1(b)(2); § 4A1.2(c)); see also § 2G2.2 cmt. n.1 (defining “distribution” not to include mere solicitation of certain material). Or they could have

371, 381 (2013) (explaining that “[t]he force of any negative implication . . . depends on context” and that “[w]e have long held that the expressio unius canon does not apply ‘unless it is fair to suppose that Congress considered the unnamed possibility and meant to say no to it’” (quoting Barnhart, 537 U.S. at 168)); Scalia & Garner, supra, at 107 (“The doctrine properly applies only when the . . . thing specified . . . can reasonably be thought to be an expression of all that shares in the grant or prohibition involved.”).

18 Case: 21-20140 Document: 00516831915 Page: 19 Date Filed: 07/24/2023

tethered the definition to the violation of specific drug laws—again, as other guidelines do. See, e.g., § 2D1.1(a)(2) (setting Base Offense Level at 38 if, inter alia, “the defendant is convicted under 21 U.S.C. § 841(b)(1)(A), (b)(1)(B), or (b)(1)(C), or 21 U.S.C. § 960(b)(1), (b)(2), or (b)(3)”). 24 Had § 4B1.2 been structured that way, one might plausibly argue that it excludes a conspiracy under 21 U.S.C. § 846 (assuming § 846 was not listed). 25 That is not how § 4B1.2 is written, however. Instead, the definition is keyed to specific drug-related actions (“manufacture, import, export, distribution, dispensing”). But not putting “conspiracy” on that list does not imply excluding it. A conspiracy is not just another drug-related activity the drafters chose to omit. Rather, a conspiracy is “[a] step toward the commission of another crime, the step in itself being serious enough to merit punishment.” Inchoate Offense, Black’s Law Dictionary (11th ed. 2019). 26 It is therefore conceptually different from the listed acts: one can conspire to commit any of them. That removes the premise for applying expressio unius—“an ‘associated group or series,’ justifying the inference that items not mentioned were excluded by deliberate choice.” Barnhart, 537 U.S. at 168 (quoting United States v. Vonn, 535 U.S.

24 See also, e.g., § 2X7.1(a)(2) (setting Base Offense Level of 16 “if the defendant was convicted under 18 U.S.C. § 555(a)); § 2A2.2(b)(7) (providing for a two-level enhancement “[i]f the defendant was convicted under 18 U.S.C. § 111(b) or § 115”). 25 To be clear, we are not saying that argument would be correct. After all, the conspiracy statute states that anyone who attempts or conspires to commit a drug crime “shall be subject to the same penalties as those prescribed for the offense, the commission of which was the object of the attempt or conspiracy.” 21 U.S.C. § 846 (emphasis added). We are saying only that the expressio unius argument for excluding conspiracies would be better if the relevant definition listed specific sections but not the conspiracy section. 26 See also, e.g., Smith v. United States, 568 U.S. 106, 110 n.3 (2013) (explaining that a “[n]arcotics conspiracy under 21 U.S.C. § 846 criminalizes ‘conspir[ing] to commit any offense’ under the Controlled Substances Act”).

19 Case: 21-20140 Document: 00516831915 Page: 20 Date Filed: 07/24/2023

55, 65 (2002)). Said another way, conspiracies do not “go[] hand in hand” with the actions listed in § 4B1.2(b), such that “the omission of [conspiracy] has interpretive force.” Cartagena, 979 F.3d at 362. 27 In addition to the expressio unius argument, some courts point to the nearby definition of “crime of violence” in § 4B1.2(a)(1). This definition, they argue, does mention “attempt,” and so one should infer that § 4B1.2(b) must exclude inchoate crimes by not mentioning them. 28 See Bittner v. United States, 598 U.S. 85, 94 (2023) (“When Congress includes particular language in one section of a statute but omits it from a neighbor, we normally understand that difference to convey a difference in meaning . . . .”). We disagree because the two sections are not parallel. “Crime of violence” is defined in terms of offenses with force-related elements—i.e., one that “has as an element the use, attempted use, or threatened use of physical

27 The dissent’s counterarguments are unavailing. Post, at 8–9. It contends that (1) “substantive and conspiracy offenses are associated items,” and (2) the Commission “chose[]” to include substantive offenses and leave conspiracies out. See ibid. (“Substantive drug crimes were chosen. Conspiracies were not.”). Respectfully, that begs the question asked by the negative-implication canon. That question is whether the list of acts in § 4B1.2(b) (“manufacture, import, [etc.]”) implies excluding inchoate forms of those acts, such as conspiracies. As we have explained, the answer is no. The dissent also claims our position amounts to inserting “pizza” into “a menu of ‘hot dogs, hamburgers, and bratwursts.’” Id. at 9 (quoting United States v. Havis, 907 F.3d 439, 450 (6th Cir. 2018) (Thapar, J., concurring), rev’d en banc, 927 F.3d 382). The dissent finds this analogy “delicious[],” ibid., but we find it undercooked. A better analogy would be whether a menu listing “milk, soy milk, and almond milk” implicitly excludes skim milk. Or whether a mother’s forbidding her child to eat “cookies” implicitly allows eating raw cookie dough. Answering those questions requires carefully considering context, just like the question before us. See Bourgeois, 902 F.3d at 448 (“Context may indicate that Congress did not wish for an express provision of one thing to work towards the exclusion of another.” (citations omitted)). 28 See Havis, 927 F.3d at 386 (arguing based on §4B1.2(a) that “the Commission knows how to include attempt crimes when it wants to”); Nasir, 17 F.4th at 471 (same); Campbell, 22 F.4th at 445 (same); Castillo, 69 F.4th at 658 (same).

20 Case: 21-20140 Document: 00516831915 Page: 21 Date Filed: 07/24/2023

force against the person of another.” § 4B1.2(a)(1). By contrast, “controlled substance offense” is not defined in terms of elements but, as noted, in terms of whether an offense “prohibits” certain drug-related actions. § 4B1.2(b). Not being parallel, the two sections shed little light on each other. In other words, we should not infer that because the authors included “attempted use of physical force” in § 4B1.2(a)(1), they must have excluded “attempted drug manufacture” from § 4B1.2(b). Maybe that would follow if § 4B1.2(b) defined a controlled substance offense as one “that has as an element the manufacture, distribution, etc., of a controlled substance.” But it does not, meaning that the premise for pitting one section against the other is lacking. See Clinchfield Coal Co. v. Fed. Mine Safety & Health Rev. Comm’n, 895 F.2d 773, 779 (D.C. Cir. 1990) (explaining that “explicit direction for something in one provision, and its absence in a parallel provision, implies an intent to negate it in the second context” (emphasis added)). Finally, according to some courts, e.g., Havis, 927 F.3d at 386, the text of § 4B1.2(b) cannot “bear the construction” that includes conspiracies and attempts, see Stinson, 508 U.S. at 46 (asking whether “the guideline which the commentary interprets will bear the construction”). Others disagree. E.g., Richardson, 958 F.3d at 155. The disagreement concerns a fine-grained inquiry into the meaning of “prohibit” in the phrase: “an offense that prohibits the manufacture, import, export, distribution, or dispensing of a controlled substance.” § 4B1.2(b) (cleaned up). Some courts contend that here “prohibit” can only mean “forbid by law.” See, e.g., Dupree, 57 F.4th at 1279. On that view, the definition would exclude conspiracies because, strictly speaking, they do not legally “forbid” the drug-related action itself but only the agreement to engage in it. See id. at 1279–80; see also Smith, 568 U.S. at 110 & n.3 (defining a narcotics conspiracy under 21 U.S.C. § 846). Others contend, however, that “prohibit” may also

21 Case: 21-20140 Document: 00516831915 Page: 22 Date Filed: 07/24/2023

carry the broader connotation of “prevent [or] hinder.” See, e.g., Richardson, 958 F.3d at 155 (alteration in original) (quoting Prohibit, Oxford English Dictionary (online ed. 2020)). On that view, the definition would include conspiracies because criminalizing agreements to engage in drug-related activities “hinders” the activities themselves. See ibid. Under Stinson deference, however, we need not say which of these two readings of “prohibit” is the correct or even the better one. All we need determine is whether the guideline can bear the commentary’s construction that includes inchoate crimes. 508 U.S. at 46. It can. For various reasons, the commentary need not have chosen the narrowest view of what constitutes an “offense that prohibits” the drug-related activities listed in § 4B1.2(b) and could have opted for a reading of the phrase broad enough to embrace inchoate crimes. See Lewis, 963 F.3d at 22. To begin with, federal law provides that those who commit inchoate drug offenses “shall be subject to the same penalties as those prescribed for the offense, the commission of which was the object of the attempt or conspiracy.” 21 U.S.C. § 846 (emphasis added). The whole purpose of the guidelines is to implement penalties for federal crimes, including violations of laws like § 846. See 28 U.S.C. § 991(b). So, it would not be at all surprising if the Commission interpreted the relevant phrase in § 4B1.2(b) to place inchoate crimes on the same footing as the underlying criminal acts. See 28 U.S.C. § 994(h) (directing the Commission “assure that the guidelines specify a sentence to a term of imprisonment at or near the maximum term authorized” for career offenders of crimes of violence or controlled substance offenses (emphasis added)). We cannot say that such an approach makes the commentary “flat[ly] inconsisten[t]” with the guideline definition. Stinson, 508 U.S. at 43.

22 Case: 21-20140 Document: 00516831915 Page: 23 Date Filed: 07/24/2023

Furthermore, the commentary’s broader reading of “prohibit” syncs with how ordinary English speakers would use the term. See United States v. Billups, 850 F.3d 762, 765–66 (5th Cir. 2017) (noting that the guidelines are “subject to the ordinary rules of statutory construction” and so “in the absence of a statutory definition, we give terms their ordinary meaning” (citations omitted)). Suppose a university’s code of conduct provides: “The university prohibits cheating on exams.” A week before finals, a professor discovers that a group of students has concocted an elaborate plan to cheat on their tests. The plot is thwarted before the students can act. Would finding the students guilty of violating the code of conduct be “flatly inconsistent” with the code’s “prohibition” on cheating? Of course not. Moreover, the history of the current definition strongly supports the commentary’s inclusion of inchoate offenses. See Scalia & Garner, supra, at 432, 440 (contrasting disfavored “legislative history” with “statutory history,” meaning “[t]he enacted lineage of a statute, including prior laws, amendments, codifications, and repeals”). 29 The two prior versions, in 1987 and 1988, cross-reference specific drug crimes, along with a catch-all for “similar offenses.” 30 The commentary explained that both definitions included “aiding and abetting, conspiring, or attempting to

29 See also In re Crocker, 941 F.3d 206, 213 (5th Cir. 2019) (“Enacted revisions in the wording of statutes are part of ‘statutory history,’ not ‘the sort of unenacted legislative history that often is neither truly legislative (having failed to survive bicameralism and presentment) nor truly historical (consisting of advocacy aimed at winning in future litigation what couldn’t be won in past statutes).’” (quoting BNSF Ry. Co. v. Loos, 139 S. Ct. 893, 906 (2019) (Gorsuch, J., dissenting))). 30 See § 4B1.2(2) (1987) (defining “controlled substance offense” as “an offense identified in 21 U.S.C. §§ 841, 952(a), 955, 955a, 959; §§ 405B and 416 of the Controlled Substances Act as amended in 1986, and similar offenses”); § 4B1.2(2) (1988) (defining “controlled substance offense” as “an offense identified in 21 U.S.C. §§ 841, 845b, 856, 952(a), 955, 955a, 959; and similar offenses”) (emphases added).

23 Case: 21-20140 Document: 00516831915 Page: 24 Date Filed: 07/24/2023

commit such offenses, and other offenses that are substantially equivalent to the offenses listed.” § 4B1.2(2) cmt. n.2 (1987); see also § 4B1.2(2) cmt. n.2 (1988) (similar). Today’s definition was born the following year, 1989, replacing the cross-references with a broader reference to “federal or state law prohibiting the manufacture, [etc.] . . . of a controlled substance.” § 4B1.2(2) (1989) (cleaned up). And, of course, the commentary continued to explain that inchoate offenses were included. Id. cmt. n.1. Keep in mind, moreover, that this commentary—as is often the case—passed through notice and comment and was submitted to Congress. See Dupree, 57 F.4th at 1281 (Pryor, C.J., concurring) (explaining that, “in practice, the Commission ordinarily uses the same procedure to revise the commentary as it does to revise the Guidelines” and that “[t]he application note [to § 4B1.2(b)] is an apt example” (citing § 4B1.2 cmt. n.2 (1987); Amendments to the Sentencing Guidelines for United States Courts, 54 Fed. Reg. at 21379)). In other words, Congress had the opportunity to consider the inchoate offense commentary when it reviewed the current form of § 4B1.2(b). In light of this history, we would be hard pressed to say that the current iteration of § 4B1.2(b) “cannot bear” the Committee’s longstanding construction that it—like the two prior versions—includes inchoate offenses. *** We sum up. Stinson tells us to treat the commentary to a guideline as “binding” unless “following one will result in violating the dictates of the other.” Stinson, 508 U.S. at 43. As explained in this Part, we do not find that kind of “flat inconsistency” between the guideline definition of controlled substance offense and the commentary’s view that the definition includes conspiracies. Ibid. We accordingly defer to the commentary under Stinson.

24 Case: 21-20140 Document: 00516831915 Page: 25 Date Filed: 07/24/2023

III(C). Even under Kisor, we would defer to the commentary. 31 Alternatively, we will assume that Kisor did modify the Stinson framework, as some of our sister circuits hold. See Nasir, 17 F.4th at 471; Campbell, 22 F.4th at 444–45; Riccardi, 989 F.3d at 485; Castillo, 69 F.4th at 655–56; Dupree, 57 F.4th at 1275. Even under that less deferential framework, however, we would still defer to the commentary’s view that the definition of “controlled substance offense” includes inchoate crimes. Kisor clarified when a court must defer to an agency’s interpretation of its own regulation. “First and foremost,” courts should not defer “unless the regulation is genuinely ambiguous.” Kisor, 139 S. Ct. at 2415 (first citing Christensen v. Harris County, 529 U.S. 576, 588 (2000); and then citing Seminole Rock, 325 U.S. at 414). That threshold inquiry demands scrutiny of the regulation, not merely “wav[ing] the ambiguity flag just because [a court] found the regulation impenetrable on a first read.” Ibid. Specifically, “a court must ‘carefully consider’ the text, structure, history, and purpose of a regulation,” and, in view of that, conclude that “the interpretive question still has no single right answer.” Ibid. (cleaned up) (quoting Pauley v. BethEnergy Mines, Inc., 501 U.S. 680, 707 (1991) (Scalia, J., dissenting)). And even then, “the agency’s reading must still be ‘reasonable,’” meaning it must fall “within the zone of ambiguity the court has identified after employing all its interpretive tools.” Id. at 2415–16 (quoting Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 515 (1994)). Next, a court must assure itself that “the character and context of the agency interpretation entitles it to controlling weight.” Id. at 2416. For

31 This Part represents the views of six out of sixteen judges: Chief Judge Richman, and Judges Smith, Southwick, Willett, Duncan, and Engelhardt.

25 Case: 21-20140 Document: 00516831915 Page: 26 Date Filed: 07/24/2023

instance, the interpretation must be “the agency’s ‘authoritative’ or ‘official position,’ rather than any more ad hoc statement not reflecting the agency’s views.” Ibid. (quoting United States v. Mead Corp., 533 U.S. 218, 257–59 & n.6 (2001) (Scalia, J., dissenting)). The interpretation must also “in some way implicate [the agency’s] substantive expertise.” Id. at 2417. Finally, the interpretation “must reflect [the agency’s] ‘fair and considered judgment.’” Ibid. (quoting Christopher v. SmithKline Beecham Corp., 567 U.S. 142, 155 (2012)). This means an agency will not earn deference for “a merely ‘convenient litigating position’” or for a “new interpretation” that unfairly disrupts expectations, such as “when an agency substitutes one view of a rule for another.” Id. at 2418 (first quoting Christopher, 567 U.S. at 155; then citing Long Island Care at Home, Ltd. v. Coke, 551 U.S. 158, 170 (2007); and then citing Thomas Jefferson, 512 U.S. at 515). If an agency’s interpretation survives this gauntlet, Kisor tells courts to afford the agency “significant leeway to say what its own rules mean,” thus “enabl[ing] the agency to fill out the regulatory scheme Congress has placed under its supervision.” Ibid. In other words, courts are to defer to the agency’s interpretation. With those principles in mind, we analyze the guideline and commentary at issue under the Kisor framework. (1). Text, structure, history, and purpose Considering the text, structure, history, and purpose of § 4B1.2(b)’s definition of “controlled substance offense” leads us to conclude that the definition is genuinely ambiguous. See id. at 2415. Specifically, the definition does not “directly or clearly address” whether inchoate offenses are included or excluded. Id. at 2410 (plurality op.).

26 Case: 21-20140 Document: 00516831915 Page: 27 Date Filed: 07/24/2023

a. Text Section 4B1.2(b)’s text poses this question: does an “offense that prohibits” various drug-related activities include a conspiracy to commit those same activities? Or to put the question in terms of Vargas’s criminal history: do conspiracies to possess cocaine and meth with intent to distribute them count as “offense[s] that prohibit[] the possession of a controlled substance with intent to distribute” under § 4B1.2(b)? We conclude that the text does not settle this question. As we have already observed, the definition says nothing about conspiracies. It is silent on the subject. See Smith, 989 F.3d at 585 (the definition “does not tell us, one way or another, whether inchoate offenses are included or excluded” (quoting Raupp, 677 F.3d at 759)) (and collecting cases). 32 Yes, § 4B1.2(b) lists various actions (“manufacture, import, export, distribution, or dispensing of a controlled substance”), and, yes, “conspiracy” is not one of them. But that does not mean conspiracies are excluded, as some courts hold. See, e.g., Winstead, 890 F.3d at 1092 (concluding the definition “clearly excludes inchoate offenses”). That is because, as already noted, a conspiracy differs conceptually from the actions listed in § 4B1.2(b). It is not just another drug-related activity the drafters left out. A conspiracy is, instead, a punishable step toward committing any of the listed activities. See Inchoate Offense, Black’s Law Dictionary (11th ed. 2019); see also Smith, 568 U.S. at 110 & n.3. So, the fact that § 4B1.2(b) does not list “conspiracy” alongside

32 See also Piper, 35 F.3d at 617 (noting that because § 4B1.2(b) does not “expressly exclude[]” conspiracies, “the Sentencing Commission’s inclusion of conspiracy convictions is most accurately viewed as interstitial”).

27 Case: 21-20140 Document: 00516831915 Page: 28 Date Filed: 07/24/2023

“manufacture” does not support the inference that the definition excludes conspiracies. The expressio unius canon does not work that way. See Chevron U.S.A. Inc. v. Echazabal, 536 U.S. 73, 81 (2002) (“The canon depends on identifying a series of two or more terms or things that should be understood to go hand in hand,” thus “supporting a sensible inference that the term left out must have been meant to be excluded.” (emphasis added) (citing Earl T. Crawford, The Construction of Statutes 337 (1940))); Vonn, 535 U.S. at 65 (the canon presumes “a commonly associated group or series” with one member “left unmentioned”); Scalia & Garner, supra, at 107 (expressio unius “must be applied with great caution, since its application depends so much on context”). Moreover, as discussed below, when the guidelines differentiate inchoate from substantive offenses, they do so expressly. See infra III(C)(1)(b). Our sister circuits have also clashed over the meaning of “prohibit” in § 4B1.2(b). Compare Dupree, 57 F.4th at 1279, with Richardson, 958 F.3d at 155. As noted, some read “prohibit” to mean “forbid by law” (i.e., “criminalize”) the activities listed in § 4B1.2(b). Presumably, this reading would exclude conspiracies because they criminalize agreeing to do the activities, not doing the activities themselves. See Iannelli v. United States, 420 U.S. 770, 777 (1975) (“Traditionally the law has considered conspiracy and the completed substantive offense to be separate crimes.”). Others read “prohibit” to include “hindering” the listed activities, a reading that easily brings in conspiracies and attempts. See Richardson, 958 F.3d at 155. With utmost respect for our colleagues, parsing the verb “prohibit” is not a helpful way to answer this question. We consider it unlikely that, while writing § 4B1.2(b), the drafters said to themselves: “How should we convey that a ‘controlled substance offense’ excludes conspiracies and

28 Case: 21-20140 Document: 00516831915 Page: 29 Date Filed: 07/24/2023

attempts? Let’s try this: use the verb prohibit, and make the direct object only particular drug-related actions but not ‘agreements’ or ‘attempts’ to do those actions. That will make it crystal clear.” 33 Even if we assume the drafters took that awkward approach, however, the resulting phrase (“offense that prohibits [drug-related activities]”) fails to do the trick. Recall our hypothetical about a university whose rules “prohibit cheating on exams.” See supra III(B). Whether that rule is violated by a failed cheating plot is not going to be settled by picking among dictionary definitions of “prohibit.” Cf. Smith v. United States, 508 U.S. 223, 241–42 (1993) (Scalia, J., dissenting) (explaining that just because a word can be used a certain way does not mean that it was used that way). Rather, one would have to consider the broader structure, history, and purpose of the university’s rules to see whether a “conspiracy to cheat” sensibly falls within the rule. See Kisor, 139 S. Ct. at 2415. To those considerations we now turn. b. Structure We have already discussed some of the structural considerations in our Stinson analysis, supra III(B), so we only mention them briefly here. When the guideline authors want to exclude crimes expressly, they know how. See, e.g., § 3D1.1(b)(2); § 4A1.2(c). Yet the authors did not take that approach in § 4B1.2(b), suggesting they did not want to exclude inchoate

33 We agree with the dissent that “[t]he operative question is not what the authors of § 4B1.2(b) ‘said to themselves,’ but what they included in the text of the Guidelines.” Post, at 12. But we still have to discern what the text means. One tool for doing that is to make reasonable inferences about how normal English speakers use words. See Biden v. Nebraska, 143 S. Ct. 2355, 2379–80 (2023) (Barrett, J., concurring) (explaining that the meaning of a congressional command, like a parent’s instruction to a babysitter, depends on reasonable inferences about the speaker’s intent); see also Bostock v. Clayton County, 140 S. Ct. 1731, 1767 (2020) (Alito, J., dissenting) (“[W]hen textualism is properly understood, it calls for an examination of the social context in which a statute was enacted . . . .”).

29 Case: 21-20140 Document: 00516831915 Page: 30 Date Filed: 07/24/2023

crimes. See Dupree, 57 F.4th at 1295 n.1 (Luck, J., dissenting). Or again, when the authors want to tie a guideline to the violation of certain crimes, they know how. See, e.g., § 2D1.1(a)(2). Yet the authors did not take that approach in § 4B1.2(b), suggesting they did not want to exclude the sections on conspiracies or attempts. Finally, the statutory backdrop makes inchoate drug offenses “subject to the same penalties” as the underlying offense. 21 U.S.C. § 846. Given that, why suppose that the authors would exclude drug conspiracies from the career offender calculus? If the authors wanted to do that, we would expect express language instead of implication. See, e.g., Jackson, 60 F.3d at 133 (finding it “relevant” to commentary’s validity “that Congress has manifested its intent that drug conspiracies and underlying offenses should not be treated differently . . . [by] impos[ing] the same penalty for a narcotics conspiracy conviction as for the substantive offense” (citing 21 U.S.C. § 846)). 34 This view also finds support in how the guidelines address penalties for individual inchoate crimes. Unless a guideline “expressly” provides otherwise, see § 2X1.1(c), the rule is that attempts or conspiracies have the same base offense level as the substantive offense. See § 2X1.1(a); see also § 2X1.1 cmt. 2 (“Under § 2X1.1(a), the base offense level [for solicitation, attempt, or conspiracy] will be the same as that for the substantive offense.”). To be sure, attempts and conspiracies may get a 3-level decrease in certain circumstances, but not if defendants completed their roles or were thwarted

34 As discussed, supra III(B), we infer precious little from the inclusion of “attempt” in § 4B1.2(a)(1), the neighboring “crime of violence” definition. That provision and § 4B1.2(b) are apples and oranges and so (to mix metaphors) do not shed much light on each other.

30 Case: 21-20140 Document: 00516831915 Page: 31 Date Filed: 07/24/2023

from doing so. See § 2X1.1(b)(1), (b)(2). 35 And the commentary states that this reduction will not be warranted “[i]n most prosecutions for conspiracies or attempts.” 36 This treatment of inchoate crimes supports our reading of § 4B1.2(b) in two ways. First, the guidelines generally penalize conspiracies and attempts the same as the substantive offense. See 2X1.1(a) & cmt. 2; see also 21 U.S.C. § 846 (making inchoate drug crimes “subject to the same penalties” as underlying offense). That supports reading § 4B1.2(b) as counting inchoate crimes toward the career offender designation. At a minimum, it offers no reason to think § 4B1.2(b) unambiguously excludes inchoate crimes. Second, when the guidelines mean to distinguish penalties for inchoate and substantive crimes (as they do in reducing the offense level for some conspiracies and attempts in § 2X1.1(b)), they do so explicitly. Unlike § 2X1.1(b), however, § 4B1.2(b) is silent on the matter. In sum, the guideline’s structure does not suggest that the § 4B1.2(b) definition excludes conspiracies and attempts.

35 “If a conspiracy, decrease by 3 levels, unless the defendant or a co-conspirator completed all the acts the conspirators believed necessary on their part for the successful completion of the substantive offense or the circumstances demonstrate that the conspirators were about to complete all such acts but for apprehension or interruption by some similar event beyond their control.” § 2X1.1(b)(2). The treatment for attempts is substantially the same. § 2X1.1(b)(1). 36 See § 2X1.1, cmt. background (“In most prosecutions for conspiracies or attempts, the substantive offense was substantially completed or was interrupted or prevented on the verge of completion by the intercession of law enforcement authorities or the victim. In such cases, no reduction of the offense level is warranted.”).

31 Case: 21-20140 Document: 00516831915 Page: 32 Date Filed: 07/24/2023

c. History Next, we consult the history of the contested provision. See Kisor, 139 S. Ct. at 2415 (courts must consider, inter alia, a regulation’s “history” to discern its meaning). Like the structural inquiry, § 4B1.2(b)’s history resists the conclusion that it excludes inchoate crimes. To the contrary, the definition’s history argues for including them. Many courts that have interpreted § 4B1.2(b), both pre- and post- Kisor, have not considered the provision’s history and have stopped with the supposedly “plain” text. 37 But Kisor reminds us that language itself may gather meaning from its history. See id. at 2415–16; see also Scalia & Garner, supra, at 352 (contrasting statutory history, defined as “[t]he enacted lineage of a statute, including prior laws, amendments, codifications, and repeals,” with disfavored legislative history). In the 1987 version of the guidelines, a “controlled substance offense” was initially defined as “an offense identified in 21 U.S.C. §§ 841, 952(a), 955, 955a, 959; §§ 405B and 416 of the Controlled Substance Act as amended in 1986, and similar offenses.” § 4B1.2(2) (1987). In commentary, the Commission clarified that this definition “include[d] aiding and abetting, conspiring, or attempting to commit such offenses, and other offenses that are substantially equivalent to the offenses listed.” Id. § 4B1.2(2) cmt. n.2. And this clarification had adequate textual basis in the definition itself; conspiring to commit any of the delineated offenses surely constituted a

37 See Dupree, 57 F.4th 1269 (no engagement with history); Campbell, 22 F.4th 438 (same); Nasir, 17 F.4th 459 (same); Havis, 927 F.3d 382 (same); Castillo, 69 F.4th 648 (same); Winstead, 890 F.3d 1082 (same). Our dissenting colleagues likewise decline to consider the definition’s history.

32 Case: 21-20140 Document: 00516831915 Page: 33 Date Filed: 07/24/2023

“similar offense” as any of the underlying crimes. 38 Alongside this, the commentary also explained that equivalent state offenses were included. Ibid. In 1988, the definition of “controlled substance offense” remained largely the same. But this time it did not refer to the Controlled Substance Act, instead defining the term as “an offense identified in 21 U.S.C. §§ 841, 845b, 856, 952(a), 955, 955a, 959; and similar offenses.” § 4B1.2(2) (1988). As before, this definition delineated a list of statutory violations, followed by the broadening term “similar offenses.” The 1988 commentary continued to clarify that “similar offenses” referred to inchoate crimes, as well as state analogues of the listed federal statutes. See id. § 4B1.2(2) cmt. n.2. Finally, in 1989, the guidelines replaced the cross-references to federal statutes with the current reference to “federal or state law prohibiting the manufacture, import, export, or distribution of a controlled substance,” while removing the “similar offenses” catch-all. § 4B1.2(2) (1989). 39 Significantly, the commentary continued to include inchoate offenses. Id. § 4B1.2(2) cmt. n.1. This throws light on the 1989 amendment. The prior “similar offenses” language embraced both equivalent state offenses and inchoate offenses. That language was now replaced with “prohibiting” language that referred only to a series of drug-related activities. Yet the inclusion of both (1) equivalent state law offenses and (2) inchoate offenses was understood to inhere in the new definition—albeit with the state law clarification now in the

38 Vargas’s attorney conceded this point at oral argument. U.S. Court of Appeals for the Fifth Circuit, 21-20140 USA v. Vargas, January 24, 2023, YouTube, at 7:10, https://youtu.be/RVjwkCV5M9c?t=430. 39 Although the Commission subsequently made minor alterations to the definition, this amendment created what is “substantially, and substantively, its current form.” United States v. Ruth, 966 F.3d 642, 652 (7th Cir. 2020).

33 Case: 21-20140 Document: 00516831915 Page: 34 Date Filed: 07/24/2023

guideline itself. In other words, a strong inference exists that the shift from “similar offenses” to “prohibiting” meant to hold constant the inclusion of both (1) equivalent state law offenses and (2) inchoate offenses. Thus, if inchoate offenses are “similar offenses” to the actually completed crimes, they are also “offenses that prohibit” the completed crimes. 40 d. Purpose Finally, Kisor tells us to consider the “purpose of a regulation.” 139 S. Ct. at 2415. The point of the career-offender enhancement is obvious: to give longer sentences to people who are more culpable because they have committed multiple drug crimes. It is equally obvious that this rationale extends to inchoate drug crimes. People who conspire or attempt to distribute drugs have also committed drug crimes, just like people who do the actual distributing. Considering purpose, then, counsels in favor of reading § 4B1.2(b) to include inchoate offenses. See Piper, 35 F.3d at 617 (including inchoate offenses accords with the Commission’s “oft-demonstrated preoccupation with punishing drug traffickers sternly”). The statutory background also supports including inchoate drug offenses along with completed offenses. As already noted, in the subchapter

40 One final practical note on history. Most amendments to the commentary today go through notice and comment and submission to Congress for review. See Dupree, 57 F.4th at 1281 (Pryor, C.J., concurring). The commentary to § 4B1.2 was no different. See Amendments to the Sentencing Guidelines for United States Courts, 54 Fed. Reg. at 21379. This is despite the fact that revisions to the commentary are not required to undergo the same process as guideline revisions. See 28 U.S.C. § 994(x). Nevertheless, Congress had the opportunity to consider the inchoate offense commentary at the same time that it reviewed what is substantively the current form of § 4B1.2 itself. This supplements the already extensive historical record that favors a reading of the guideline that is consistent with the inclusion of inchoate offenses.

34 Case: 21-20140 Document: 00516831915 Page: 35 Date Filed: 07/24/2023

on drug crimes, federal law provides that anyone “who attempts or conspires to commit any offense defined in this subchapter shall be subject to the same penalties as those prescribed for the offense, the commission of which was the object of the attempt or conspiracy.” 21 U.S.C. § 846 (emphasis added); see also 18 U.S.C. § 2(a) (“Whoever commits an offense against the United States or aids, abets, counsels, commands, induces or procures its commission, is punishable as a principal.”). By including inchoate offenses in § 4B1.2(b), the Commission honors these statutory commands. A glance at the purposes of sentencing is also instructive. Sentences are meant “to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense.” 18 U.S.C. § 3553(a)(2)(A). Sentences deter, protect the public, and rehabilitate. Id. § 3553(a)(2)(B)–(D). The guidelines are required to reflect these purposes. See 28 U.S.C. § 991(b)(1)(A). They are also meant to “provide certainty and fairness in meeting the purposes of sentencing, avoiding unwarranted sentencing disparities among defendants with similar records who have been found guilty of similar criminal conduct.” Id. § 991(b)(1)(B); see also Moses, 23 F.4th at 357 (noting that the Commission “promulgated commentary specifically to satisfy that purpose” (emphasis omitted)). Reading the § 4B1.2(b) definition to include inchoate offenders serves all these purposes. See Piper, 35 F.3d at 617 (including inchoate offenses comports with the “spirit” and “aim” of the guideline). One who repeatedly conspires to distribute drugs is more culpable than a one-time conspirator— just as someone who distributes drugs repeatedly is more culpable than a one- time distributor. It makes perfect sense to punish both repeat offenders more harshly because both are more culpable. Doing so “reflect[s] the seriousness of the offense,” “promote[s] the respect for the law,” and “provide[s] just punishment.” See 18 U.S.C. § 3553(a)(2)(A)–(D).

35 Case: 21-20140 Document: 00516831915 Page: 36 Date Filed: 07/24/2023

Or compare two offenders: one with two drug distribution convictions and a conspiracy drug conviction, and another with three drug distribution convictions. Recognizing that these two offenders are similarly situated for career offender purposes “provide[s] certainty and fairness” and “avoid[s] unwarranted sentencing disparities among defendants with similar records.” See id. § 991(b)(1)(B). In short, reading § 4B1.2(b) to include inchoate offenses best harmonizes with the purposes the text seeks to advance. (2). The commentary’s character and context Kisor also instructs us to consider whether “the character and context of the agency interpretation entitles it to controlling weight.” Kisor, 139 S. Ct. at 2416 (citing Christopher, 567 U.S. at 155). Vargas does not contest this part of the analysis, but we briefly run through it for the sake of completeness. First, the commentary to § 4B1.2(b) was “actually made by the [Commission].” Ibid. More than just an “ad hoc statement not reflecting the [Commission’s] views,” the commentary represents the Commission’s “authoritative” and “official position.” Ibid. (quoting Mead, 533 U.S. 257– 59). The commentary was, after all, officially promulgated. See Amendments to the Sentencing Guidelines for United States Courts, 54 Fed. Reg. at 21379. Next, the commentary doubtless “implicate[s] [the Commission’s] substantive expertise.” Kisor, 139 S. Ct. at 2417. Finally, the commentary undeniably reflects the Commission’s “fair and considered judgment.” Ibid. (quoting Christopher, 567 U.S. at 155). It has remained substantively identical for decades and thus can hardly be said to be a “convenient litigating position” or “post hoc rationalization advanced to defend past agency action from attack.” Ibid. (cleaned up). Thus, the commentary’s character and context show that it is owed deference under Kisor.

36 Case: 21-20140 Document: 00516831915 Page: 37 Date Filed: 07/24/2023

III(D). Rule of Lenity 41 We conclude with a brief note on the rule of lenity. Although Vargas does not raise the issue, some judges have argued that the rule of lenity should resolve ambiguity in a guideline in the criminal defendant’s favor. See, e.g., Nasir, 17 F.4th at 472–74 (Bibas, J., concurring). We doubt that the rule of lenity applies to the guidelines, however. And even if it does, it applies only in the face of “grievous ambiguity,” a standard not met here. See United States v. Castleman, 572 U.S. 157, 172–73 (2014) (citation omitted). The rule of lenity is animated by purposes that do not apply to merely advisory guidelines. First, the rule reflects concerns about fair notice and due process of law: “[F]air warning should be given to the world in language that the common world will understand, of what the law intends to do if a certain line is passed.” McBoyle v. United States, 283 U.S. 25, 27 (1931). This confines ambiguous criminal statutes within their clear scope. But with respect to the guidelines, fair notice “is not at issue because the Guidelines ‘do not bind or regulate the primary conduct of the public.’” United States v. Wright, 607 F.3d 708, 719 (11th Cir. 2010) (Pryor, J., concurring) (quoting Mistretta, 488 U.S. at 396). Second, the rule of lenity reinforces the separation of powers by preventing courts from expanding vague statutes. “[B]ecause of the seriousness of criminal penalties, and because criminal punishment usually represents the moral condemnation of the community, legislatures and not courts should define criminal activity.” United States v. Bass, 404 U.S. 336, 348 (1971). Understandably, when the guidelines were mandatory and thus

41 This Part represents the views of eight out of sixteen judges: Chief Judge Richman, and Judges Jones, Smith, Southwick, Duncan, Engelhardt, Oldham, and Wilson.

37 Case: 21-20140 Document: 00516831915 Page: 38 Date Filed: 07/24/2023

“ha[d] the force and effect of laws,” Booker v. United States, 543 U.S. 220, 234 (2005), lenity influenced interpretation of the guidelines’ ambiguous provisions. See Wright, 607 F.3d at 718 (Pryor, J., concurring). In a post- Booker world, however, the guidelines are advisory. United States v. Smith, 977 F.3d 431, 435 (5th Cir. 2020) (citation omitted). In that world, separation-of-powers concerns are vitiated because now-advisory guidelines do not usurp the congressional prerogative to ordain punishments for criminal offenses. Ibid. One of our Third Circuit colleagues, while acknowledging that the guidelines are advisory, argues that courts should “still attend to the rule [of lenity] and its animating purposes” because the guidelines “exert a law-like gravitational pull on sentences.” Nasir, 17 F.4th at 474 (Bibas, J., concurring). Our dissenting colleagues share that view. Post, at 16–17. With respect, we disagree because determining whether lenity applies to the guidelines should be based on their legal, not practical, effects. Our view finds support in the Supreme Court’s decision in Beckles v. United States, 580 U.S. 256 (2017). There, the Court held that the guidelines are not susceptible to void-for-vagueness challenges precisely because their now-advisory role does not implicate concerns about vagueness. See id. at 266–67. It was the guidelines’ advisory status, not their “gravitational pull,” that influenced whether the vagueness doctrine applied. We apply the same logic to the rule of lenity. And doing so leads us to believe that it “no longer applies to the purely advisory Guidelines.” Smith, 977 F.3d at 435. But even if lenity does apply, the guideline at issue here is not “grievously ambiguous.” See Wooden v. United States, 142 S. Ct. 1063, 1075 (2022) (Kavanaugh, J., concurring). Lenity only comes into play “when a criminal statute contains a ‘grievous ambiguity or uncertainty,’ and ‘only if, after seizing everything from which aid can be derived,’ the Court ‘can make

38 Case: 21-20140 Document: 00516831915 Page: 39 Date Filed: 07/24/2023

no more than a guess as to what Congress intended.’” Ocasio v. United States, 578 U.S. 282, 295 n.8 (2016) (quoting Muscarello v. United States, 524 U.S. 125, 138–39 (1998)). This differs from the threshold level of ambiguity needed to trigger Kisor deference, which applies when a regulation is only “genuinely ambiguous.” Kisor, 139 S. Ct. at 2415 (emphasis added). The § 4B1.2(b) definition of controlled substance offense may be ambiguous, but it is not “grievously” so. As our application of Kisor shows, the commentary advances a reading of that text that is reasonable, that finds ample support in the broader structure of the guidelines, and that is consistent with the way the Commission has always read the definition in its various iterations. So, even if lenity applies, it still would not overcome the deference due to the commentary. III(E). Summary 42 Stinson squarely applies to the guidelines commentary at issue here and was not overruled or modified by Kisor. As an inferior court, then, we must apply Stinson. Under its framework, the commentary is binding because it does not “violat[e] the dictates of” § 4B1.2(b). Stinson, 508 U.S. at 43. Accordingly, we reaffirm our longstanding precedent that inchoate offenses like conspiracy are included in the definition of “controlled substance offense.” See Lightbourn, 115 F.3d 291. Alternatively, even under Kisor we would defer to the commentary. The definition’s text, structure, history, and purpose show that the commentary takes a reasonable view of a genuinely ambiguous guideline.

42 This Part represents the views of five out of sixteen judges: Chief Judge Richman, and Judges Smith, Southwick, Duncan, and Engelhardt.

39 Case: 21-20140 Document: 00516831915 Page: 40 Date Filed: 07/24/2023

Because it is undisputed that the other Kisor preconditions are met, defer we must. Finally, the rule of lenity does not affect our analysis of the guidelines. IV. Conclusion 43 Vargas was properly sentenced as a career offender under § 4B1.1(a) because he was guilty of three controlled substance offenses as defined by § 4B1.2(b) and its accompanying commentary. We therefore AFFIRM Vargas’s sentence.

43 This Part represents the views of ten out of sixteen judges: Chief Judge Richman, and Judges Jones, Smith, Southwick, Higginson, Willett, Ho, Duncan, Engelhardt, and Oldham.

40 Case: 21-20140 Document: 00516831915 Page: 41 Date Filed: 07/24/2023

Andrew S. Oldham, Circuit Judge, joined by Jones, Circuit Judge, concurring in part: I agree with the majority that we are bound by Stinson v. United States, 508 U.S. 36 (1993). But even if an inferior court could reconsider Stinson in light of subsequent Supreme Court decisions, the relevant case is United States v. Booker, 543 U.S. 220 (2005)—not Kisor v. Wilkie, 139 S. Ct. 2400 (2019). I therefore join all but Part III.C of the majority opinion. I. Start with Stinson. The key premise of the case is that the Guidelines bind district courts. The Court began its analysis by stating: “As we have ob- served, ‘the Guidelines bind judges and courts in the exercise of their uncon- tested responsibility to pass sentence in criminal cases.’” Stinson, 508 U.S. at 42 (quoting Mistretta v. United States, 488 U.S. 361, 391 (1989)). From there, the Court noted that it had already extended this premise to the Sen- tencing Commission’s policy statements: The principle that the Guidelines Manual is binding on federal courts applies as well to policy statements. In Williams v. United States, 503 U.S. 193, 201 (1992), we said that “[w]here . . . a policy statement prohibits a district court from taking a specified action, the statement is an authoritative guide to the meaning of the applicable Guideline.” Id. at 42. From these two premises (i.e., the binding nature of the Guidelines and the binding nature of the Guidelines’ policy statements), the Stinson Court concluded that the Guidelines’ commentary was also binding: Commentary which functions to “interpret [a] guideline or ex- plain how it is to be applied,” USSG § 1B1.7, controls, and if failure to follow, or a misreading of, such commentary results in a sentence “select[ed] . . . from the wrong guideline range,”

41 Case: 21-20140 Document: 00516831915 Page: 42 Date Filed: 07/24/2023

Williams, 503 U.S. at 203, that sentence would constitute “an incorrect application of the sentencing guidelines” under 18 U.S.C. § 3742(f)(1). . . . Our holding in Williams dealing with policy statements applies with equal force to the commentary before us here. Id. at 42–43. Thus, the cornerstone of the Stinson regime is the binding nature of the Guidelines. Of course, the Booker Court held that the Guidelines were not binding on federal courts. See 543 U.S. at 245. So if we were free to predict what the Supreme Court would do today, one might reasonably guess that Stinson would fall. Of course, we are not so free because “it is th[e Supreme] Court’s prerogative alone to overrule one of its precedents.” State Oil Co. v. Khan, 522 U.S. 3, 20 (1997). II. Even if we were at liberty to update Stinson with a subsequent Supreme Court decision, I do not understand why we would choose Kisor to do the updating. Given that Booker renders the Guidelines themselves advisory, why would we apply “Kisor deference” or any other kind of “deference” to the Guidelines’ commentary? Consider by analogy the Federal Rules of Civil Procedure. The Federal Rules and the Guidelines share similar promulgation procedures. See 28 U.S.C. §§ 2072–2074 (Federal Rules); 28 U.S.C. § 994(p), (x) (Sentencing Guidelines). Both go through public notice-and-comment and then are submitted to Congress. If Congress fails to act, both go into effect at a set date. The Federal Rules and the Guidelines are often promulgated with notes from the committees that helped draft them. For the former, an advisory rules committee appointed by the Supreme Court (“Advisory

42 Case: 21-20140 Document: 00516831915 Page: 43 Date Filed: 07/24/2023

Committee”) can promulgate “committee notes.” For the latter, the Sentencing Commission can promulgate “commentary.” Such notes and commentary can be sent to Congress, but no statute requires it. 1 And the commentary at issue in this case was, in fact, submitted to Congress. See Amendments to the Sentencing Guidelines for United States Courts, 54 Fed. Reg. 21348, 21379 (May 17, 1989). 2 The Advisory Committee’s notes to the Federal Rules are not entitled to Seminole Rock, Auer, or any other sort of deference. Instead, judges treat the Advisory Committee’s notes like legislative history. See 4 Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 1029 (4th ed.) (noting committee notes “provide something akin to a ‘legislative history’ of the rules”); Stinson, 508 U.S. at 43 (referenc- ing the Advisory Committee’s notes as akin to legislative committee re- ports); Tome v. United States, 513 U.S. 150, 160 (1995) (employing the notes as a “useful guide in ascertaining the meaning of the Rules” and a “respected source of scholarly commentary”). Judges can refer to them, but they need not defer to them. In the words of Justice Scalia: The Advisory Committee’s insights into the proper interpre- tation of a Rule’s text are useful to the same extent as any

1 As to the committee notes to the Federal Rules, see, e.g., Letter from John G. Roberts, Jr., Chief Justice, to Nancy Pelosi, Speaker of the House of Representatives, Supreme Court of the United States Orders – Term Year 2021 (Apr. 11, 2022) (transmitting notes). As to the Guidelines’ commentary, see United States v. Dupree, 57 F.4th 1269, 1280–81 (11th Cir. 2023) (en banc) (W. Pryor, C.J., concurring) (noting Guidelines’ commentary is often submitted to Congress even if not required by statute). 2 Stinson rejected an analogy between the Guidelines commentary and the Advisory Committee Notes because the commentary at issue in Stinson “was issued well after the guideline . . . had been promulgated” so it could not capture the drafter’s intent. 508 U.S. at 44. But that’s not what happened in this case; the commentary was submitted to Congress alongside the Guideline. And as Stinson acknowledged, “much commentary [is] issued at the same time as the guideline it interprets.” Ibid.

43 Case: 21-20140 Document: 00516831915 Page: 44 Date Filed: 07/24/2023

scholarly commentary. But the Committee’s intentions have no effect on the Rule’s meaning. Even assuming that we and the Congress that allowed the Rule to take effect read and agreed with those intentions, it is the text of the Rule that controls. Krupski v. Costa Crociere S. p. A., 560 U.S. 538, 557 (2010) (Scalia, J., con- curring). In a post-Booker world, one could reasonably argue that the commen- tary to the Guidelines should not receive any deference that the Advisory Committee’s notes to the Federal Rules do not. Deference to the former but not the latter would be particularly incongruous because, unlike the Guide- lines, the Federal Rules are binding on federal courts. Well hold on, you might say, even after Booker the Guidelines are still binding in some sense. True, the Supreme Court has said that district courts must start their sentencing decisions by calculating the appropriate Guide- lines range. See, e.g., Peugh v. United States, 569 U.S. 530, 541 (2013) (“[D]istrict courts must begin their analysis with the Guidelines and remain cognizant of them throughout the sentencing process.” (quotation omitted)). But that does not make the Guidelines binding in the same sense as a federal statute, the Federal Rules, or even a lawfully promulgated federal regulation. We have an abundance of post-Booker cases holding that “even if the correct guidelines range was not considered,” the error is harmless upon a convincing showing that “the district court would have imposed the same sentence had it not made the error, and . . . that it would have done so for the same reasons it gave at the prior sentencing.” United States v. Guzman-Rendon, 864 F.3d 409, 411 (5th Cir. 2017) (emphasis added) (quotation omitted); see also, e.g., United States v. Reyna-Aragon, 992 F.3d 381, 388 (5th Cir. 2021); United States v. Redmond, 965 F.3d 416, 420–22 (5th Cir. 2020); United States v. Cas- tro-Alfonso, 841 F.3d 292, 298–99 (5th Cir. 2016); United States v. Richardson, 676 F.3d 491, 511–12 (5th Cir. 2012); United States v. Ibarra-Luna, 628 F.3d

44 Case: 21-20140 Document: 00516831915 Page: 45 Date Filed: 07/24/2023

712, 714 (5th Cir. 2010). But of course, no district court could avoid reversal by announcing that it found a federal statute, Federal Rule, or lawful regula- tion merely advisory and that it would reach the same result in any event.

* * *

Post-Booker, the world is topsy-turvy. The Sentencing Guidelines are not binding, but the commentary is. The Federal Rules are binding, but the Advisory Committee’s notes are not. Regardless, until the Supreme Court overrules Stinson, we are bound to follow it.

45 Case: 21-20140 Document: 00516831915 Page: 46 Date Filed: 07/24/2023

Jennifer Walker Elrod, Circuit Judge, joined by Stewart, Haynes, Graves, Wilson, and Douglas, Circuit Judges, dissenting in part and dissenting from the judgment: ∗ Whether Kisor modified Stinson is an unusually thorny question of vertical stare decisis, as evidenced by the growing division among the federal circuits.1 But regardless of whether we proceed under Stinson or Kisor, the commentary at issue here deserves no deference. Whatever way one looks at it—through the lens of consistency between the relevant Guideline and its commentary (Stinson) or ambiguity in the Guideline itself (Kisor)—the Guidelines definition of “controlled substance offense” does not include conspiracy convictions as the commentary contends. I therefore take no position on Part III(A) of the majority opinion. But I dissent from Parts III(B) and (C). The commentary fails under both Stinson and Kisor. Separate and apart from my views on Stinson, Kisor, and Application Note 1 to U.S.S.G. § 4B1.2(b), I also dissent from the plurality’s holding in Part III(D) because it improperly cabins a “time-honored” interpretive canon. Cargill v. Garland, 57 F.4th 447, 470 (5th Cir. 2023) (en banc). I agree with my colleagues on other circuits who would apply the rule of lenity when interpreting the Sentencing Guidelines. See United States v. Campbell, 22 F.4th 438, 446 (4th Cir. 2022) (observing that “the rule of lenity . . . has some force” when interpreting the Guidelines (internal quotation marks and

∗ Judge Graves joins in full. Judges Stewart, Haynes, and Douglas join as to Sections I–III. Judge Wilson joins as to Sections II and III. 1 See United States v. Castillo, 69 F.4th 648 (9th Cir. 2023) (applying the Kisor framework); United States v. Dupree, 57 F.4th 1269 (11th Cir. 2023) (en banc) (same); United States v. Riccardi, 989 F.3d 476 (6th Cir. 2021) (same); United States v. Nasir, 17 F.4th 459 (3d Cir. 2021) (en banc) (same). But see United States v. Maloid, __ F.4th __, 2023 WL 4141073 (10th Cir. June 23, 2023) (applying Stinson); United States v. Moses, 23 F.4th 347 (4th Cir. 2022) (same). Both parties here agree that Kisor applies to Guidelines commentary. Blue Br. 16; Red Br. 29.

46 Case: 21-20140 Document: 00516831915 Page: 47 Date Filed: 07/24/2023

citation omitted)); United States v. Nasir, 17 F.4th 459, 472–74 (3d Cir. 2021) (Bibas, J., concurring) (same); United States v. Winstead, 890 F.3d 1082, 1092 n.14 (D.C. Cir. 2018) (same). I The majority opinion follows the Tenth and Fourth Circuits in holding that Stinson continues to govern judicial deference to the Sentencing Commission’s commentary to its own Guidelines. In doing so, it rejects the conclusion of the Eleventh, Ninth, Sixth, and Third Circuits that Kisor modified the Stinson framework. The majority opinion chiefly relies on differences between the administrative agencies at issue in Kisor and the Sentencing Commission at issue in Stinson. I certainly agree that the Commission is no administrative agency. As “a sort of hybrid body that does not fit squarely within any of the three branches of government,” it has aptly been described as an “odd duck” in our tripartite system. United States v. Havis, 907 F.3d 439, 442, 443 (6th Cir. 2018), rev’d en banc on other grounds. See also Mistretta v. United States, 488 U.S. 361, 384 (1989) (describing the Commission as a “peculiar institution”). That being said, Stinson and Kisor are clearly related. Stinson held that the standard of deference then-applicable to an administrative agency’s interpretations of its own legislative rules also applies to the Sentencing Commission’s interpretations of its Guidelines. Stinson v. United States, 508 U.S. 36, 45 (1993). Stinson then adopted a formulation of that standard that Kisor has now deemed a “reflexive” “caricature of the doctrine.” Kisor v. Wilkie, 139 S. Ct. 2400, 2415 (2019). The argument that Kisor’s correctives are more sweeping than the majority opinion supposes may therefore have some merit.

47 Case: 21-20140 Document: 00516831915 Page: 48 Date Filed: 07/24/2023

Those are the relevant inputs. What we should do with them is another matter. As faithful middle managers, “[w]e are bound to follow the Supreme Court precedent that most squarely controls our case.” Freedom from Religion Found. v. Mack, 4 F.4th 306, 315 (5th Cir. 2021). But it is not always clear which case squarely controls. Old cases are sometimes overruled or modified without fanfare. See, e.g., id. (applying Town of Greece v. Galloway, 572 U.S. 565 (2014), rather than Lemon v. Kurtzman, 403 U.S. 602 (1971), even though Galloway did not expressly overrule Lemon); Trump v. Hawaii, 138 S. Ct. 2392, 2423 (2018) (taking the “opportunity to make express what is already obvious” by explicitly stating that Korematsu v. United States, 323 U.S. 214 (1944) is no longer good law). At the same time, we are not infrequently admonished to “leave to [the Supreme Court] the prerogative of overruling its own decisions.” Rodriguez de Quijas v. Shearson/Am. Exp., Inc., 490 U.S. 477, 484 (1989). See also Mallory v. Norfolk S. R. Co., 143 S. Ct. 2028, 2038 (2023) (criticizing the Pennsylvania Supreme Court for treating an on-point Supreme Court decision as having been implicitly overruled); Hohn v. United States, 534 U.S. 236, 252–53 (1998) (“Our decisions remain binding precedent until we see fit to reconsider them, regardless of whether subsequent cases have raised doubts about their continuing vitality.”). We would benefit from further guidance in this area. But we need not determine whether Stinson or Kisor applies today because Vargas prevails under either framework. II Assuming arguendo that Stinson controls, “[i]t does not follow that commentary is binding in all instances.” Stinson, 508 U.S. at 43. Stinson deference is not absolute. It incorporates a fail-safe for commentary that is “plainly erroneous or inconsistent” with the Sentencing Guidelines

48 Case: 21-20140 Document: 00516831915 Page: 49 Date Filed: 07/24/2023

themselves. Id. at 45. That backstop exists for precisely this situation, in which the commentary at issue purports to change the meaning of a term. See United States v. Riccardi, 989 F.3d 476, 493 (6th Cir. 2021) (Nalbandian, J., concurring in part) (“Stinson requires that commentary interpret the guidelines, not contradict or add to them.”). A The career offender designation is a three-strikes rule. It means a significantly extended Guidelines sentence for adult offenders who return to court for sentencing on a new felony “crime of violence” or “controlled substance offense” after having already received two prior felony “crime of violence” or “controlled substance offense” convictions. U.S.S.G. § 4B1.1(a). When Andres Vargas appeared at his sentencing hearing in April of 2021, he already had a conviction for a substantive drug offense and a conviction for conspiracy to commit another drug offense to his name. Because Vargas was an adult, and because he was presently being sentenced for another conspiracy drug offense, the Government sought to designate him a career offender on the theory that he was being sentenced for his third “controlled substance offense.” The problem with that theory is that it is belied by the plain text of the Guidelines, which defines the term “controlled substance offense” to include only substantive drug offenses and not conspiracies. Guidelines § 4B1.2(b) states that a “controlled substance offense” is: [A]n offense under federal or state law, punishable by imprisonment for a term exceeding one year, that prohibits the manufacture, import, export, distribution, or dispensing of a controlled substance (or a counterfeit substance) or the possession of a controlled substance (or a counterfeit

49 Case: 21-20140 Document: 00516831915 Page: 50 Date Filed: 07/24/2023

substance) with intent to manufacture, import, export, distribute, or dispense. Because the Guideline only identifies substantive drug crimes as qualifying offenses, Vargas only has one strike. Not three. Two of his convictions—his prior conspiracy conviction as well as the conspiracy conviction giving rise to this appeal—do not count. That should be the end of the analysis. It is well established that a “definition which declares what a term ‘means’ . . . excludes any meaning that is not stated.” Burgess v. United States, 553 U.S. 124, 130 (2008) (citation omitted). Therefore, the commentary’s attempt to add conspiracy offenses, see § 4B1.2(b) cmt. n.1, must be disregarded as “plainly erroneous or inconsistent” with the Guidelines. Stinson, 508 U.S. at 45. Many of our sister circuits have arrived at the same conclusion. United States v. Castillo, 69 F.4th 648, 652 (9th Cir. 2023) (“The text of § 4B1.2(b) does not identify conspiracy to commit any of the offenses as such an offense.”); United States v. Dupree, 57 F.4th 1269, 1277 (11th Cir. 2023) (en banc) (“The definition does not mention conspiracy or attempt or any other inchoate crimes.”); Nasir, 17 F.4th at 471 (“The guideline does not even mention inchoate offenses.”); Campbell, 22 F.4th at 444 (“The text of U.S.S.G. § 4B1.2(b) contains a lengthy definition . . . that does not mention attempt offenses.”); United States v. Havis, 927 F.3d 382, 386 (6th Cir. 2019) (en banc) (“The guideline expressly names the crimes that qualify as controlled substance offenses . . . none are attempt crimes.”); Winstead, 890 F.3d at 1091 (“Section 4B1.2(b) presents a very detailed ‘definition’ of controlled substance offense that clearly excludes inchoate offenses.”). B The majority opinion itself acknowledges that § 4B1.2(b) “says nothing one way or the other about conspiracies and attempts.” Ante, at 16.

50 Case: 21-20140 Document: 00516831915 Page: 51 Date Filed: 07/24/2023

But while I would read this omission as an exclusion, the majority opinion treats it as an invitation. The logic seems to be that if the Guidelines are not explicit in renouncing a qualifying offense, the Sentencing Commission is free to include it by means of its commentary. It does not take a great stretch of the imagination to see the pitfalls of a rule that writes the Sentencing Commission that kind of blank check. Cf. Castillo, 69 F.4th at 663 (noting “grave constitutional concerns” raised by deference to Application Note 1); Campbell, 22 F.4th at 446 (warning that deference to Application Note 1 permits “circumvention of the checks Congress put on the Sentencing Commission”) (citation omitted); Havis, 927 F.3d at 386–87 (same). It is also difficult to square with the mechanism set up by Congress. Biden v. Nebraska, 143 S. Ct. 2355, 2379 (2023) (Barrett, J., concurring) (explaining that “[c]ontext . . . includes common sense”). There is little point in subjecting the Guidelines to notice and comment and congressional review if the Commissioners may unilaterally add to those Guidelines through its commentary. Castillo, 69 F.4th at 663. That is why Stinson described Guidelines commentary as having the limited “functional purpose of . . . assist[ing] in the interpretation and application” of the Guidelines. Stinson, 508 U.S. at 45. Only the Guidelines are “promulgate[d] by virtue of an express congressional delegation of authority for rulemaking” and accompanied by the requisite constitutional safeguards. Id. at 44. In any event, no rule of deference or construction can hide the fact that Application Note 1 is adding something altogether new to the Guidelines definition of “controlled substance offense.” Cf. Havis, 927 at 386 (concluding that the commentary impermissibly “add[ed] an offense not listed in the guideline”); Dupree, 57 F.4th at 1278 (agreeing). The majority opinion makes several attempts to avoid the plain meaning of the Guideline’s text, but none are persuasive.

51 Case: 21-20140 Document: 00516831915 Page: 52 Date Filed: 07/24/2023

1 The majority opinion spills a great deal of ink in an effort to excuse the text of § 4B1.2(b) from the negative-implication canon: the common-sense semantic rule reflecting the shared intuition of English speakers that “[t]he expression of one thing implies the exclusion of others.” Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 107. It points to a limited exception to this rule under which “the canon does not tell us that a case was provided for by negative implication unless an item unmentioned would normally be associated with the items listed.” Barnhart v. Peabody Cole Co., 537 U.S. 149, 169 n.12 (2003). See also Chevron U.S.A. Inc. v. Echazabal, 536 U.S. 73, 81 (2002) (explaining that negative implication “canon depends on identifying a series of two or more terms or things that should be understood to go hand in hand, which is abridged in circumstances supporting a sensible inference that the term left out must have been meant to be excluded”). This exception is not without its critics. Justice Scalia accused it of being not only “unheard-of,” but “absurd,” “since it means that the more unimaginable an unlisted item is, the more likely it is not to be excluded.” Barnhart, 537 U.S. at 180, 181 (Scalia, J., dissenting). 2 More to the point, this

2 The majority opinion also cites the uncontroversial principle that the negative implication canon applies only when the things specified “can reasonably be thought to be an expression of all that shares in the grant or prohibition involved.” Ante, at 17 n. 23 (citing Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 107). Thus, “[t]he sign outside the restaurant ‘No dogs allowed’ cannot be thought to mean that no other creatures are excluded.” Reading Law, 107. “On the other hand, the sign outside a veterinary clinic saying ‘Open for treatment of dogs, cats, horses, and all other farm and domestic animals’ does suggest (by its detail) that” unmentioned animals will not be treated. Id. Here the detail and specificity of § 4B1.2(b) places it in the latter category and communicates that the definition is exhaustive. Moreover, § 4B1.2(b)’s status as the definition of a term is itself good reason to conclude that the definition communicates the complete scope of that term.

52 Case: 21-20140 Document: 00516831915 Page: 53 Date Filed: 07/24/2023

case bears no resemblance to those in which we have applied the limited exception to the negative implication canon. Consider the example cited by the majority opinion, United States v. Cartagena-Lopez, 979 F.3d 356 (5th Cir. 2020). There we held that a statutory provision excluding a defendant’s time in jail from the calculation of his supervised release term did not abrogate the common law fugitive tolling doctrine, under which a defendant is also precluded from tolling a period during which he is on the run. Id. at 362–63. In doing so, we relied on the fact that the relevant statutory provision was “not a standalone tolling provision” and did not purport to be about “tolling per se,” but rather tolling in a particular circumstance. Id. at 362. That circumstance—being a fugitive—was not so closely related to going back to jail for another crime that we could read the enumeration of the latter reason for tolling to exclude tolling for the former reason. Id. We also relied on the bedrock principle that “[i]n interpreting statutes, we presume that Congress is aware of the common law and does not undertake to change it lightly.” Id. at 362–63. The fugitive tolling doctrine is a centuries-old rule rooted in the “widely recognized” principle that a convict should not benefit from his wrongdoing. Id. at 362. Because implied changes to such longstanding rules are “disfavored,” we declined to read the Act to “render the common law of parole obsolete.” Id. at 363. The Guidelines definition of “controlled substance offense” could not be more different. It is a standalone provision with no common law pedigree and which purports to provide the complete definition of the term. What is more, substantive and conspiracy offenses are associated items. They go hand in hand because they are both offenses in the pool of crimes from which the Sentencing Commission could have pulled when deciding which offenses would qualify as a “controlled substance offense” for

53 Case: 21-20140 Document: 00516831915 Page: 54 Date Filed: 07/24/2023

purposes of the career offender designation. Substantive drug crimes were chosen. Conspiracies were not. Judge Thapar puts the point deliciously: Interpreting a menu of “hot dogs, hamburgers, and bratwursts” to include pizza is nonsense. Nevertheless, that is effectively what the government argues here when it says that we must apply deference to a comment adding to rather than interpreting the Guidelines. Havis, 907 F.3d at 450 (Thapar, J., concurring). The majority opinion’s contrary conclusion relies on a basic distinction between substantive and conspiracy offenses. While a substantive offense is its own completed crime, a conspiracy offense is “[a] step toward the commission of another crime.” Ante, at 19 (quoting Inchoate Offense, Black’s Law Dictionary (11th ed. 2019)). Fair enough. Conspiracy “has ingredients, as well as implications, distinct from the completion of the unlawful project.” Pinkerton v. United States, 328 U.S. 640, 644 (1946). But no two crimes are exactly the same. The question for our purposes is whether a conspiracy to commit a drug offense is so different from a substantive drug offense that we would not expect a reasonable English speaker to place conspiracies in the category of drug offenses that might possibly be included in the definition of the term “controlled substance offense.” I do not think that is a close call. To be clear, conspiracy offenses are distinct offenses—not simply a way of committing a substantive offense. That is why a defendant can be charged with two separate offenses—one count of conspiracy and one count for the substantive offense—when he completes a drug crime. Pinkerton, 328 U.S. at 643. It is difficult to understand how a definition could enumerate offenses relating to “import, export, [and] distribution,” as well as “dispensing” and “possession,” but not account for one of the most (perhaps the most?) frequently charged federal crimes: conspiracy.

54 Case: 21-20140 Document: 00516831915 Page: 55 Date Filed: 07/24/2023

2 The plain reading of § 4B1.2(b) to include only substantive drug offenses is bolstered by its neighboring provisions. Housed within the same Guideline is the definition of the term “crime of violence,” the other predicate category of offense which can lead to a career offender designation. Unlike “controlled substance offense,” the definition of “crime of violence” explicitly includes attempt crimes. See U.S.S.G. § 4B1.2(a) (“The term ‘crime of violence’ means any offense . . . that — (1) has as an element the use, attempted use, or threatened use of physical force against the person of another.” (emphasis added)). When faced with adjacent definitions, one of which expressly includes a category and one of which does not, the ordinary reader draws the inference that the omission is meaningful. That is why “[a]textual judicial supplementation is particularly inappropriate when, as here,” the drafter has demonstrated “that it knows how to adopt the omitted language or provision.” Rotkiske v. Klemm, 140 S. Ct. 355, 361 (2019). The majority opinion rejects the significance of this context on the ground that the two definitions are not parallel. While a “crime of violence” is defined by reference to its elements, the term “controlled substance offense” is defined in terms of what it prohibits. Ante, at 20–21. Right off the bat, this may be a distinction without a difference. A criminal offense that prohibits certain conduct can only do so by making the commission of that conduct an element of the offense. In any event, the majority opinion is missing the forest for the trees. The semantic point is that, when drafting these two neighboring definitions, the Sentencing Commission used explicit language to indicate the inclusion of inchoate offenses in the definition of the term “crime of violence.” Why

55 Case: 21-20140 Document: 00516831915 Page: 56 Date Filed: 07/24/2023

would it change its tactics to rely on silent implication when constructing the very next definition? 3 The majority opinion also indulges the Government’s “creative dictionary use” of the term “prohibit.” Dupree, 57 F.4th at 1288 (Grant, J., concurring in the judgment). The Government maintains that “prohibit” could mean “prevent or hinder.” Thus, it says, § 4B1.2(b) can be read to include conspiracies because the criminalization of conspiracies to commit drug offenses would “hinder” the commission of the substantive drug offenses named in the Guideline. The Government’s reading violates one of the most basic rules of statutory construction: “Words are to be understood in their ordinary, everyday meanings.” Reading Law at 69. See also Nebraska, 143 S. Ct. at 2378 (Barrett, J., concurring) (“To strip a word from its context is to strip that word of its meaning.”). The question is not whether “prohibit” could possibly mean “hinder” in some unlikely hypothetical. Instead, the question is whether the word carries that meaning in context. It does not. Indeed, like Judge Grant, “I personally cannot think of any context where ‘prohibit’ naturally means ‘hinder.’” Dupree, 57 F.4th at 1288. That includes the university code hypothetical suggested by the majority opinion. Ante, at 23. As a preliminary matter, university codes typically do distinguish between inchoate and substantive violations. 3 Aside

3 See Edward N. Stoner II & John Wesley Lowery, Navigating Past the “Spirit of Insubordination:” A Twenty-First Century Model Student Conduct Code With a Model Hearing Script, 31 J.C. & U.L. 1, 27 (2004) (“Any student found to have committed or to have attempted to commit the following misconduct is subject to the disciplinary sanctions outlined in Article IV: (1) Acts of dishonesty, including . . . (a) cheating, plagiarism, or other forms of academic dishonesty.”).

56 Case: 21-20140 Document: 00516831915 Page: 57 Date Filed: 07/24/2023

from that, section 4B1.2(b) is not a university code. It is a definition made by reference to the federal criminal code. Whatever might be true of other genres, federal criminal law does not use “prohibit” when it means “hinder.” Cf. Nebraska, 143 S. Ct. at 2378 (2023) (Barrett, J., concurring) (explaining that “[b]ackground legal conventions . . . are part of [a] statute’s context”). Such literalism defies what “every lawyer and citizen knows[:] criminal law is not suggestive—it either bans conduct or it allows it.” Dupree, 57 F.4th at 1288. See also Havis, 927 F.3d at 386 n.4 (“[T]he guideline’s boilerplate use of the term ‘prohibits’ simply states the obvious: criminal statutes proscribe conduct.”); Campbell, 22 F.4th at 448 (agreeing and adding that “[i]nterpreting ‘prohibits’ to include anything that makes the outlawed conduct more likely to occur would sweep into criminal statutes a vast swath of conduct based on a secondary dictionary definition”); Dupree, 57 F.4th at 1278–79 (agreeing). 4 At another point, the majority opinion appears to derive some import from what the Sentencing Commission likely “said to themselves” while drafting § 4B1.2(b). Ante, at 28. The operative question is not what the authors of § 4B1.2(b) “said to themselves,” but what they included in the text of the Guidelines. “Men intend what they will; but it is only the laws that they enact which binds us.” Antonin Scalia, Common-Law Courts in a Civil Law System: The Role of United States Federal Courts in Interpreting th e Constitution and Laws, in A Matter of Interpretation 3, 17 (new ed. 2018). 4

4 The majority’s thought experiment is self-defeating in any event. It is just as unlikely that the drafters said to themselves: “How should we convey that ‘controlled substance offense’ includes conspiracies and attempts. Let’s try this: use the verb prohibit, because a secondary definition of that term is hinder and offenses that criminalize conspiracies clearly hinder the substantive offense.”

57 Case: 21-20140 Document: 00516831915 Page: 58 Date Filed: 07/24/2023

* * * “When a statute includes an explicit definition, we must follow that definition.” Tanzin v. Tanvir, 141 S. Ct. 486, 490 (2020) (quoting Digital Realty Trust, Inc. v. Somers, 138 S. Ct. 767, 777 (2018)). I would apply that rule here to hold—contra Application Note 1—that the Guidelines definition of “controlled substance offense” does not include conspiracies. III The commentary also fails under the Kisor framework, and on largely the same grounds. After all, Kisor sets a higher bar than Stinson. If something fails Stinson, it necessarily fails Kisor. Under Kisor, we first ask whether U.S.S.G. § 4B1.2(b) is ambiguous. For all of the reasons stated above, it is not. The career-offender Guideline “just means what it means—and the court must give it effect.” Kisor, 139 S. Ct. at 2415. Kisor instructs that “a court should not afford Auer deference unless the regulation is genuinely ambiguous.” Id. Moreover, “before concluding that a rule is genuinely ambiguous, a court must exhaust all the ‘traditional tools’ of construction.” Id. The “court must carefully consider the text, structure, history, and purpose of a regulation, in all the ways it would if it had no agency to fall back on.” Id. (quotation marks and alterations omit- ted). 5

5 That is not to say that ambiguity alone is sufficient to trigger deference under Kisor. Even in the presence of an ambiguous rule or regulation, an agency interpretation is only entitled to deference if it is “reasonable.” Kisor, 139 S. Ct. at 2415. “In other words, it must come within the zone of ambiguity the court has identified after employing all its interpretive tools.” Id. at 2415–16. Assuming an agency can check both of these boxes, “[s]till, we are not done—for not every reasonable agency reading of a genuinely ambiguous rule should receive Auer deference.” Id. at 2416. At the final step, “a court must make an independent inquiry into whether the character and context of the agency

58 Case: 21-20140 Document: 00516831915 Page: 59 Date Filed: 07/24/2023

Section 4B1.2(b) expressly identifies those crimes that qualify as con- trolled substance offenses, and inchoate crimes are not among them. Camp- bell, 22 F.4th at 442 (“The text of U.S.S.G § 4B1.2(b) does not state or in any way indicate that aiding and abetting, conspiracy, and attempt are ‘con- trolled substance offense[s].”); Nasir, 17 F.4th at 471 (“The Guideline does not even mention inchoate offenses.”); Havis, 927 F.3d at 386 (“[T]he plain language of § 4B1.2(b) says nothing about attempt crimes.”); Winstead, 890 F.3d at 1089 (“As is apparent, neither the crime of attempting to distribute drugs nor attempted possession with intent to distribute drugs is included in the guideline list.”). Under Kisor as well as Stinson, courts are not at liberty to rely on Guidelines commentary that adds to the plain text of the Guide- lines. IV The plurality opinion misapprehends the rule of lenity by unneces- sarily disclaiming any role for lenity in interpreting the Sentencing Guide- lines. Assuming the Kisor framework, and that the career-offender enhance- ment were somehow ambiguous, I would apply the rule of lenity to resolve that ambiguity in Vargas’s favor. A Kisor requires courts to “exhaust all the ‘traditional tools’ of con- struction” before concluding that a legal text is ambiguous. Id. No tool of construction is more “time-honored” than the rule of lenity. Cargill v. Gar- land, 57 F.4th 447, 471 (5th Cir. 2023) (en banc) (quoting Liparota v. United

interpretation entitles it to controlling weight.” Id. That is, the “regulatory interpretation must be one actually made by the agency;” “must in some way implicate its substantive expertise;” and “must reflect ‘fair and considered judgment.’” Id. at 2416, 2417.

59 Case: 21-20140 Document: 00516831915 Page: 60 Date Filed: 07/24/2023

States, 471 U.S. 419 (1985)). Indeed, the rule is “not much less old than con- struction itself.” United States v. Wiltberger, 18 U.S. 76, 95 (1820). In its most basic formulation, the rule of lenity resolves uncertainty in favor of the criminal defendant. See, e.g., Liparota, 471 U.S. at 427; see also Reading Law at 296. That typically means declining to defer to the executive- branch’s interpretation of criminal prohibitions and penalties. See generally United States v. Hamilton, 46 F.4th 389, 398 n.2 (5th Cir. 2022); Thomas Z. Horton, Lenity Before Kisor: Due Process, Agency Deference, and the Interpreta- tion of Ambiguous Penal Regulations, 54 Colum. J.L. & Soc. Probs. 629, 632– 33, 640–44, 664–66 (2021) (discussing lenity’s historical provenance and ex- plaining the canon’s applicability). I do not see any reason to exempt the Sentencing Guidelines from the longstanding rule of lenity. Nor does Kisor require that result. The applica- tion of the rule of lenity to Kisor is consistent with the theoretical underpin- nings of both doctrines. See Carter v. Welles-Bowen Realty, Inc., 736 F.3d 722, 731 (6th Cir. 2013) (Sutton, J., concurring) (making a similar point about the relationship between Chevron and the rule of lenity). As indicated by the plurality opinion, the rule of lenity is commonly understood to be grounded in principles of fair notice and the constitutional separation of powers. Ante at 37. The rule reflects the judgment that “fair warning should be given to the world in language that the common world will understand, of what the law intends to do if a certain line is passed.” McBoyle v. United States, 283 U.S. 25, 27 (1931). And it respects the constitutional ballast that “the power of punishment is vested in the legislative, not in the judicial department. It is the legislature, not the Court, which is to define a crime, and ordain its punishment.” Wiltberger, 18 U.S. at 95. The constitutional separation of powers is clearly implicated when the Sentencing Commission is permitted to resolve ambiguity in its own

60 Case: 21-20140 Document: 00516831915 Page: 61 Date Filed: 07/24/2023

Guidelines. The Guidelines themselves are subject to Congressional over- sight and the strictures of the Administrative Procedure Act. Mistretta, 488 U.S at 393–94. Not so the commentary. See Castillo, 69 F.4th at 663. The Commission may unilaterally increase the Guidelines range for disfavored criminal conduct by issuing vague Guidelines and then adopting expansive readings of those Guidelines in its commentary, or by doing the same to open- ended Guidelines already in place. See Havis, 907 F.3d at 450 (6th Cir. 2018) (Thapar, J., concurring) (noting that deference “incentivizes agencies to reg- ulate ‘broadly and vaguely’ and later interpret those regulations self-serv- ingly, all at the expense of the regulated”) (quoting Perez v. Mortg. Bankers Ass’n, 575 U.S. 92, 111 (2015) (Scalia, J., concurring in the judgment)). Def- erence without the rule of lenity thus permits “the same agency to make the rules and interpret the rules.” Id. at 452. 6 The plurality opinion is unimpressed by this concern because the Guidelines are now advisory. Ante at 37. That is cold comfort to the criminal defendant. As a descriptive matter, “the now-advisory nature of the Guide- lines does not render the limits on the Commission’s rulemaking power any less important.” Havis, 907 F.3d at 443. Last year, 67.8 percent of all of- fenders received sentences that were within the Guidelines range or justified by a reason for departure provided by the Guidelines Manual. U.S. Sentenc- ing Comm’n, 2022 Annual Report 9. Even when a district court chooses to deviate from the Guidelines, the Guidelines exert a significant anchoring effect on its sentencing decision. See

6 Chief Judge William Pryor, a former commissioner on the United States Sentencing Commission, has suggested that “the Commission could shore up the authority of its commentary without substantially modifying its practice by moving what normally goes in the commentary to the main text of the Guidelines in future revisions.” Dupree, 57 F.4th at 1281 (Pryor, C.J., concurring).

61 Case: 21-20140 Document: 00516831915 Page: 62 Date Filed: 07/24/2023

Molina-Martinez v. United States, 578 U.S. 189, 198–199 (2016) (“The Guide- lines are the framework for sentencing and anchor the district court’s discre- tion.” (internal quotation marks and alterations omitted)). In formulating a sentence, a district court must always begin by correctly calculating the Guidelines range. And the court must consider that range in making its final determination—justifying any deviation from the Guidelines with “suffi- ciently compelling” reasons. Gall v. United States, 552 U.S. 38, 50 (2007). “So just as a runner’s starting position influences the time in which he fin- ishes the race, a defendant’s sentence depends in part on what the Guidelines range is, even if that range is nonbinding.” Havis, 907 F.3d at 444. Turning to the purposes behind Kisor deference, the Supreme Court’s holding in Kisor rests chiefly on its belief that Congress wishes for agencies to have interpretive authority over laws in their particular sphere of opera- tion. Kisor, 139 S. Ct. at 2413 (reasoning that Congress is “attuned to the comparative advantages of agencies over courts in making” policy choices in specialized areas). That is because “[a]gencies (unlike courts) have ‘unique expertise,’ often of a scientific or technical nature, relevant to applying a reg- ulation ‘to complex or changing circumstances.’” Id. But this comparative expertise is not implicated in the project of interpreting the Sentencing Guidelines. Interpretation of criminal laws is one of the quintessential func- tions of a judge. “It is emphatically the province and duty of the judicial de- partment to say what the law is.” Marbury v. Madison, 5 U.S. 137, 177 (1803). Kisor offers two additional justifications for deference: that agencies are best positioned to understand their own regulations, Kisor, 139 S. Ct. at 2412, and “the well-known benefits of uniformity in interpreting genuinely ambiguous rules,” id. at 2413. These interests apply when a sentencing court undertakes to interpret the Guidelines. But to my mind they cannot over- come the interests on the other side of the ledger. The rule of lenity impli- cates structural concerns lying at the very foundation of our criminal justice

62 Case: 21-20140 Document: 00516831915 Page: 63 Date Filed: 07/24/2023

system, while Kisor is designed to answer pragmatic problems in administer- ing the humdrum rules and regulations of daily life. It would be folly on the order of Esau’s to exchange constitutional structural safeguards and our rich tradition of lenity simply to optimize the modern administrative state. I agree with my judicial colleagues in other circuits who have concluded that the rule of lenity has force in interpreting the Sentencing Guidelines. See Winstead, 890 F.3d at 1092 n.14 (“We are inclined to believe that the rule of lenity . . . has some force” in interpreting the Sentencing Guidelines.); Campbell, 22 F.4th at 446 (agreeing); Nasir, 17 F.4th at 472–74 (Bibas, J., concurring); Havis, 907 F.3d at 451 (Thapar, J., concurring). 7 B One final note. We have previously recognized the existence of two competing “standards for whether a statute is sufficiently ambiguous to trig- ger the rule of lenity.” Cargill, 57 F.4th at 469. On one view, “the rule of lenity does not apply when a law merely contains some ambiguity or is diffi- cult to decipher,” but instead applies only in the face of “grievous ambigu- ity.” Wooden v. United States, 142 S. Ct. 1063, 1075 (Kavanaugh, J., concur- ring). That is, when ambiguity persists after every other tool of construction has been tried. Id. On the second view, by contrast, lenity comes into play from the start to resolve all reasonable doubts in favor of the criminal defend- ant. Id. at 1084 (Gorsuch, J., concurring in the judgment). See also Nebraska,

7 The case for lenity’s application here is only buttressed by the fact that many states apply the rule in connection with their state sentencing guidelines. See Commonwealth v. Rossetti, 186 N.E.3d 729, 742 (Mass. 2022); State v. Weatherwax, 392 P.3d 1054, 1060 (Wash. 2017); State v. Spencer, 248 P.3d 256, 276 (Kan. 2011); State v. Maurstad, 733 N.W.2d 141, 148 (Minn. 2007); Commonwealth v. Shiffler, 879 A.2d 185, 195 (Pa. 2005); State v. Rife, 789 So.2d 288, 294 (Fla. 2001); Scott v. State, 720 A.2d 291, 295 (Md. 1998); State v. Anaya, 933 P.2d 223, 233 (N.M. 1996); People v. District Court, Second Judicial Dist., 713 P.2d 918, 922 (Colo. 1986).

63 Case: 21-20140 Document: 00516831915 Page: 64 Date Filed: 07/24/2023

143 S. Ct. at 2376 (Barrett, J., concurring) (describing lenity’s role as “break- ing a tie between equally plausible interpretations of a statute”). We have also recognized that “the Supreme Court does not appear to have decided which of these standards governs.” Cargill, 57 F.4th at 469. Yet in Part III(D) a plurality of judges on this court take a side, stating that lenity only comes into play in the face of grievous ambiguity—the proverbial Gordian knot of interpretive problems. As with the Stinson vs. Kisor ques- tion, I would not resolve this issue at this time. In this case “it does not mat- ter which standard applies because the rule of lenity applies even under the more stringent ‘grievously ambiguous’ condition.” Cargill, 57 F.4th at 469.

* * * To be designated a career offender is of no small moment for the criminal defendant. The Guidelines assign all career offenders to the highest possible Criminal History Category and drastically augment their offense level. Once the sentencing table operates its multiplying effect, many career offenders find themselves with a Guidelines sentence at or near the maximum penalty permitted by statute. Case in point, Vargas’s career offender designation increased his Guidelines sentence from a likely range of 100–125 months to a range of 188–235 months. Yet Vargas would not be considered a career offender in the Third, Fourth, Sixth, Ninth, Eleventh, and D.C. Circuits—meaning that his sentence would likely be at least five years shorter had he been convicted in one of those jurisdictions. We should not countenance that kind of disparity in the federal system. Such disparities will continue for many criminal defendants until the Supreme Court provides us with much needed guidance. I respectfully dissent.