Ernest Young v. United States

EASTERBROOK, Circuit Judge. In 1994 Ernest Young pleaded guilty to distributing cocaine, to possessing a firearm despite a prior felony conviction, and to carrying a gun during and in relation to a drug offense. Two years later, he filed a petition under 28 U.S.C. § 2255, asking the judge to vacate the third of these convictions, on the ground that Bailey v. United States, — U.S. -, 116 S.Ct. 501, 133 L.Ed.2d 472 (1995), revamps the understanding of 18 U.S.C. § 924(c). The district court denied the petition after concluding that the acts to which Young admitted violate the statute as it is understood in 1997. Young agreed to sell four ounces of crack cocaine to a person who turned out to be an undercover agent. Young and Eugene Smith drove to the agreed delivery point. In the car Young toted both the cocaine and a loaded gun, in order to protect himself. The agent persuaded Young to make delivery elsewhere. When the arrest occurred a few blocks away, Young had the drugs but not the weapon. Young now contends that only someone who carries a gun during and in relation to a drug offense of which he was convicted violates § 924(c). He was convicted of distributing the cocaine, but not of possessing the drug with intent to distribute it. Although he allows that he may have carried the gun in connection with the possession, he did not carry it at the time of the distribution. An alternative understanding of the argument is that only someone who has the gun on his person, or within ready reach, at the time of arrest violates § 924(c). I How Young’s contentions come before us for decision is something of a mystery. Young might have denied culpability and taken a direct appeal, but he pleaded guilty and did not appeal. It is hornbook law that § 2255 cannot be used as a belated appeal. Bailey appears to have little if any bearing on Young’s argument, and anyway the Supreme Court has not held that Bailey applies retroactively to persons whose sentences became final before its announcement on December 6, 1995. One might have expected the prosecutor to alert the district judge to these shortcomings. But the only procedural defense the prosecutor offered was that the petition was untimely under amendments to *796 § 2255 made by the Antiterrorism and Effective Death Penalty Act. This was quickly dispatched in light of our conclusion that the time limit does not apply to petitions filed before April 23, 1997. See Lindh v. Murphy, 96 F.3d 856, 866 (7th Cir.1996) (en banc), reversed on other grounds, — U.S. -, 117 S.Ct. 2059, 138 L.Ed.2d 481 (1997). Other potential defenses were ignored, in the district court and on appeal. Lest our decision be understood to decide sub silentio that the sequence may be replicated, we point out four potential obstacles. 1. Section 2255 is not a way to advance arguments that could have been presented earlier — especially not when the arguments rest entirely on a statute. See Reed v. Farley, 512 U.S. 339, 114 S.Ct. 2291, 129 L.Ed.2d 277 (1994). Although § 2255 ¶1 permits a collateral attack on the ground that “the sentence was imposed in violation of the Constitution or laws of the United States”, only a small portion of statutory claims demonstrate that the sentence or conviction is itself a violation of law. The error must be so fundamental that a “complete miscarriage of justice” has occurred. Reed, 512 U.S. at 348, 114 S.Ct. at 2297, quoting from Hill v. United States, 368 U.S. 424, 428, 82 S.Ct. 468, 471, 7 L.Ed.2d 417 (1962). Other “non-constitutional errors which could have been raised on appeal but were not, are barred on collateral review — regardless of cause and prejudice.” Bontkowski v. United States, 850 F.2d 306, 313 (7th Cir.1988). If a legal development after the conviction shows that the “conviction and punishment are for an act that the law does not make criminal”, Davis v. United States, 417 U.S. 333, 346, 94 S.Ct. 2298, 2305, 41 L.Ed.2d 109 (1974), the statutory standard is satisfied: the change of law shows why the claim could not have been made earlier, and imprisonment for a non-crime is a miscarriage of justice. Young does not contend that he is in prison for an act the law does not make criminal. He believes that he could not be convicted under § 924(e) without also being convicted of possessing cocaine with intent to distribute. That is to say, he complains that the United States did not charge him with enough crimes — for the facts adduced when his plea of guilty was accepted show that he possessed cocaine with intent to distribute that drug. Perhaps something about § 924(c) limits its operation to the use of guns in relation to other acts that lead to convictions, but this is a long way from a claim that Young’s acts were not crimes. What is more, in Broadway v. United States, 104 F.3d 901, 902-04 (7th Cir.1997), we held that the Davis standard is not satisfied by a claim that Bailey altered the understanding of the “carry” element in § 924(c) or the “in relation to” component of that crime. Broadway like Young was convicted of carrying a gun during and in relation to a drug offense; the same judge who denied Young’s petition granted Broadway’s; we reversed, holding among other things that, because Bailey concerned only the “use” element of § 924(c), § 2255 does not permit a collateral attack in cases that turn on the other elements. Four days after we issued Broadway, the district judge denied Young’s petition on the merits, without mentioning Broadway; 53 days after Broadway was released, the district judge issued a certificate of appealability in Young’s case, again without mentioning our decision. And the United States Attorney likewise has ignored Broadway. Limitations on the scope of § 2255 relief do not affect the subject-matter jurisdiction of the federal courts, so the United States has forfeited the benefits of Reed, Davis, and Broadway. Perhaps there is an answer to our doubts, based on eases such as Lee v. United States, 113 F.3d 73 (7th Cir. 1997), but Lee is in tension with Broadway and the cases on which Broadway relied. Given the prosecutor’s forfeiture, now is not the time to reconcile the approaches of different panels. 2. Both Young and the district judge treat Bailey as reopening every conviction under § 924(c), no matter how long ago that conviction may have been obtained and no matter what issues the petitioner now wants to litigate. Yet even if we get past the limited scope of § 2255, other problems remain — not least the question whether Bailey is retroactive. Changes in the law apply to cases that were already “final” (as Young’s was) only if they place primary conduct be *797 yond the power of the state to regulate or establish procedures that are implicit in the concept of ordered liberty. Teague v. Lane, 489 U.S. 288, 305-10, 109 S.Ct. 1060, 1072-75, 103 L.Ed.2d 334 (1989). Whatever one thinks about how Bailey’s reinterpretation of the “use” element fits the Teague model there can be no doubt that nothing remotely close to satisfying Teague has occurred for the other aspects of § 924(c). Bailey has nothing to say about a claim of the kind Young makes, which deals with the relation between the § 924(e) charge and the other counts of the indictment. Cases such as Eaglin v. Welborn, 57 F.3d 496, 498-99 (7th Cir.1995) (en bane), show that Teague is not a jurisdictional obstacle, however, and as in Eaglin the importance of the issue on the merits leads us not to use the power to invoke a Teague bar unbidden. 3. Young pleaded guilty to the § 924(c) charge. A litigant who can show cause and prejudice may be relieved from a forfeiture, such as a failure to present claims at trial or on appeal. But a litigant who waives a legal claim cannot obtain relief later. United States v. Olano, 507 U.S. 725, 732-34, 113 S.Ct. 1770, 1776-78, 123 L.Ed.2d 508 (1993). A conviction based on a waiver cannot be in “error” and therefore cannot be annulled. A guilty plea is a waiver, and as long as the plea was voluntary (and entered with effective assistance of counsel) it is conclusive on all factual and legal issues other than a contention that the very initiation of the proceedings violated the Constitution — a claim that Young does not advance. See United States v. Broce, 488 U.S. 563, 574-76, 109 S.Ct. 757, 765-66, 102 L.Ed.2d 927 (1989). Two decisions of this court, Stanback v. United States, 113 F.3d 651 (7th Cir.1997), and Lee v. United States, supra, conclude that claims based on Bailey may be advanced notwithstanding pleas of guilty. These decisions do not assist Young, whose claim is not based on Bailey. But the logic of Stanback and Lee nonetheless could be pertinent: each case characterized a guilty plea as an admission of the facts alleged in the indictment, but not a waiver of a defense that the acts do not violate the law. Lee, 113 F.3d at 75; Stanback, 113 F.3d at 654 (“In electing to plead guilty ... the defendant waives a challenge to the facts underlying the charge, but he does not waive the right to contest whether those facts are sufficient to constitute a crime.”). Unfortunately, neither Lee nor Stanback discusses Broce — or, for that matter, any of the many other decisions in which the Supreme Court treats a guilty plea as a waiver of legal defenses. The tenth circuit in Broce took the same position as our panels in Lee and Stanback: that a guilty plea admits the facts but does not waive legal defenses. 781 F.2d 792, 796 (10th Cir.1986) (en banc). That view did not prevail: In holding that the admissions in a guilty plea “go only to the acts constituting the [crime],” ... the Court of Appeals misapprehended the nature and effect of the plea. A guilty plea “is more than a confession which admits that the accused did various acts.” Boykin v. Alabama, 395 U.S. 238, 242, 89 S.Ct. 1709, 1711, 23 L.Ed.2d 274 (1969). It is an “admission that he committed the crime charged against him.” North Carolina v. Alford, 400 U.S. 25, 32, 91 S.Ct. 160, 164, 27 L.Ed.2d 162 (1970). By entering a plea of guilty, the accused is not simply stating that he did the discrete acts described in the indictment; he is admitting guilt of a substantive crime. Broce, 488 U.S. at 570, 109 S.Ct. at 762. Unless a guilty plea waives legal arguments and consents to punishment, an Alford plea would be impossible — for the defendant in Alford denied that he committed the crime charged, yet the Court held that his plea could be the basis of conviction and confinement. The Court observed in Broce that even a later decision holding unconstitutional a penalty the defendant pleaded guilty to avoid does not reopen the case. 488 U.S. at 571-72, 109 S.Ct. at 763-64, summarizing Brady v. United States, 397 U.S. 742, 90 S.Ct. 1463, 25 L.Ed.2d 747 (1970). In Broce itself the Court concluded that legal developments after the date of the plea do not permit the defendant to obtain relief. When pleading guilty, a defendant surrenders his right to make legal arguments. Usually he gets a lower sentence in *798 exchange. Young, who received a total of 180 months’ imprisonment, obtained a substantial reduction from the “mandatory” life sentence he was facing under 21 U.S.C. § 841(b)(1)(A). (The quotation marks recognize the fact that mandatory sentences aren’t, if the prosecutor agrees to drop the charges that lead to these sentences.) If the law allowed the defendant to get off scot free in the event the argument later is shown to be a winner, then the defendant could not get the reduction in the first place. Every plea would become a conditional plea, with the (unstated) condition that the defendant obtains the benefit of favorable legal developments, while the prosecutor is stuck with the original bargain no matter what happens later. That approach destroys the bargain, and the prospect of such an outcome will increase the original sentence. Conditional pleas are authorized by Fed. R.Crim. P. 11(a)(2), but only if the defendant first raises the legal issue in a pretrial motion and the prosecutor consents to the procedure. The approach of Stanback and Lee is incompatible with these limitations (the defendants did not make motions; the prosecutors did not consent); and the need for the conditional-plea apparatus is one more indication that a plea of guilty does more than admit the facts alleged. If that were all it did, Rule 11(a)(2) would be pointless. Stanback and Lee treated an effort to get out of a guilty plea as no different fundamentally from an argument on appeal that jury instructions are plain error. Both panels expressed the view that, if the law in the circuit at the time is adverse to the defendant’s legal position, the defendant need not take any step to preserve his argument for later review, and it is correspondingly proper to revisit the subject de novo after the law becomes favorable. That is a problematic statement even for cases in which the defendant is convicted at trial over his vigorous opposition. See Johnson v. United States, — U.S.-, 117 S.Ct. 1544, 137 L.Ed.2d 718 (1997); United States v. Frady, 456 U.S. 152, 102 S.Ct. 1584, 71 L.Ed.2d 816 (1982). Johnson, which was decided after Stanback and Lee, rejects an argument similar to the one those cases adopted. Someone who wants to take advantage of a legal point must put it forward, Johnson explained, even if the law of the circuit is against it, unless the new approach shows that the original decision was plain error — and an error even though “plain” does not support relief if the district court’s decision might have been justified on the fuller record that could have been built had the issue been brought to the court’s attention. What is more, plain error review is designed, as Olano observes, for forfeiture cases. Waiver means that there is no error at all. A guilty plea waives every issue except a claim that the court lacks power to adjudicate. Broce, 488 U.S. at 574r-76, 109 S.Ct. at 765-66. Young does not argue that this exception applies to him' — but then why should he, given the prosecutor’s inattention to the problem? Once again the prosecutor’s forfeiture, by inducing the defendant to omit a line of argument that might have stilled our doubts, takes the issue off the table. We need not decide whether to overrule Stan-back and Lee or to place the subject before the court era banc; it is enough for now to flag the apparent discrepancy between their approach and that of the Supreme Court. 4. Young commenced this collateral attack after April 23,1996, and it is therefore governed by the Antiterrorism and Effective Death Penalty Act. (The crimes and conviction preceded the new statute, but, under the reasoning of portions of our Lindh opinion that the Supreme Court did not disturb, the dates of these events do not affect application of the AEDPA. The new rules apply to all petitions filed after the law came into force.) One difference between the AEDPA and pri-or law is that an appeal by the prisoner in a case under § 2255 is permissible only if a judge issues a certificate of appealability. 28 U.S.C. § 2253(c)(1)(B). The new statute continues: (2) A certificate of appealability may issue under paragraph (1) only if the applicant has made a substantial showing of the denial of a constitutional right. (3) The certificate of appealability under paragraph (1) shall indicate which specific issue or issues satisfy the showing required by paragraph (2). *799 The district court issued a certificate of ap-pealability that does not comply with either of these subparagraphs. The district judge wrote that the “only appealable issue is that of the meaning of the use or carry language of 18 U.S.C. § 924(c).” This “issue” is neither “specific” nor based on the Constitution. Even if a “specific issue” may be teased out of the two-page narration in which the district judge explains that many questions under § 924(c) are unsettled in this circuit, it is hard to see how Young could satisfy § 2253(c)(2), for his claim is based entirely on the statute, while the AEDPA conditions appeal on “a substantial showing of the denial of a constitutional right” (emphasis added). Young apparently believes that any misapplication of a law violates the Constitution. Yet it should not be necessary to repeat the refrain that there is a gulf between a statutory and a constitutional error. Gilmore v. Taylor, 508 U.S. 333, 342, 113 S.Ct. 2112, 2117-18, 124 L.Ed.2d 306 (1993); Pulley v. Harris, 465 U.S. 37, 41, 104 S.Ct. 871, 874, 79 L.Ed.2d 29 (1984); Smith v. Phillips, 455 U.S. 209, 221, 102 S.Ct. 940, 948, 71 L.Ed.2d 78 (1982); cf. Archie v. Racine, 847 F.2d 1211, 1215-18 (7th Cir.1988) (en banc) (collecting eases). Nunez v. United States, 96 F.3d 990 (7th Cir.1996), applies this principle to § 924(c) eases in particular, holding that an incorrect understanding of “use” or “carry” does not violate any rule of constitutional law. The question in Nunez was whether a claim based on Bailey could be raised in a second or successive petition for collateral review, which like a certificate of appealability is limited to constitutional claims. We held that it may not be. The same reasoning means that a Bailey claim — or any other statutory claim — presented in an initial petition under the approach of Davis, and rejected by the district judge, may not be appealed. Davis itself shows why. Justice Rehnquist’s dissent in Davis — where, recall, the claim was that a decision of the Supreme Court after Davis’s conviction established that his acts were not crimes — argued that the petition should have been rejected because it was not based on the Constitution. The majority might have replied that any legally unwarranted conviction violates the due process clause. But the Court did not say this. Instead it emphasized that § 2255 permits relief when the sentence violates “the Constitution or laws of the United States” (emphasis added). See 417 U.S. at 342-45, 94 S.Ct. at 2302-04. The difference between § 2255 ¶ 1 and § 2253(c)(2) is clear: one authorizes relief when the sentence violates the Constitution or laws of the United States, and the second authorizes appeal when there has been a substantial showing that the sentence violates the Constitution. If the district court denies a petition based on a statutory issue, § 2253(c)(2) precludes an appeal. Once again, however, the United States has not noticed the problem. This problem is the closest of the four to being a jurisdictional obstacle. The absence of a certificate of appealability precludes an appeal; should an erroneously issued certificate be treated the same as the lack of a certificate? We think not. The certificate is a screening device, helping to conserve judicial (and pros-ecutorial) resources. The obligation to identify a specific issue concentrates the parties’ attention (and screens out weak issues); the limitation to constitutional claims also reduces the number of appeals while simultaneously removing a category of claim that under Davis and Reed has poor prospects. Once a certificate has issued, however, the case proceeds to briefing and decision; the resources have been invested. It is too late to narrow the issues or screen out weak claims. Perhaps a motion to dismiss an appeal on the ground that a certificate was improperly issued would serve some function. But once the briefs have been written and the case heard, there is little point in scrutinizing the certificate of appealability. An obligation to determine whether a certificate should have been issued, even if the parties do not present this issue for decision- — a step entañed by the conclusion that a proper certificate is a jurisdictional requirement— would increase the complexity of appeals in coñateral attacks and the judicial effort required to resolve them, the opposite of the legislative plan. So we proceed to the merits as the parties have presented them. *800 II The portion of § 924(c)(1) that matters to this case reads: Whoever, during and in relation to any crime of violence or drug trafficking crime ... for which he may be prosecuted in a court of the United States, uses or carries a firearm, shall, in addition to the punishment provided for such crime of violence or drug trafficking crime, be sentenced to imprisonment for five years____ The pistol Young carried is a “firearm”. Both distribution of cocaine and possession of cocaine with intent to distribute are “drug trafficking crimes” under the definition in § 924(c)(2). Young slides back and forth between an argument that he did not “carry” the gun and an argument that he did not carry the gun “in relation to” the distribution of cocaine. Each branch of the argument starts from the observation that at the time of the distribution (and arrest) he and the gun were at least a city block apart. And each branch suffers from the same defect: § 924(c) is not limited to guns being carried at the moment of arrest (or distribution). Young “canned” the gun on his person from his residence to the car, and the gun was “carried” (= transported) by the car while he was in it. See United States v. Baker, 78 F.3d 1241, 1247 (7th Cir.1996) (defining “carry” as “transport”). During the movement from house to car, and the movement by car to meeting place, the gun served the purpose of protection, so that the carrying was “in relation to” the “drug trafficking crime” of possessing cocaine with intent to distribute that drug. By the time Young emerged from the ear— with drugs, but without gun — the crime had been completed. What he did later could not absolve him of criminal responsibility. Whether he distributed the cocaine on the spot, a few blocks away, in Bogotá a year later, or decided he didn’t like the situation and took the drug home again, made no difference. Section 924(c)(1) refutes Young’s contention that the gun must be carried “in relation to” a drug crime of which the defendant is convicted. “Whoever, during and in relation to any crime of violence or drug trafficking crime ... for which he may be prosecuted in a court of the United States, uses or carries a firearm” (emphasis added) receives a five-year term. Possession of cocaine with intent to distribute is a crime for which Young could have been prosecuted. Although the five years is “in addition to” the punishment for the drug offense, this does not imply that prosecution for the offense to which the gun was related is essential. Suppose Young had been prosecuted under § 924(c) alone. It is easy to add five years to zero; 0 + 5 = 5. Some parts of the criminal code, particularly the recidivist offenses, make one crime ride piggyback on another conviction. For example, the armed career criminal statute, § 924(e)(1), covers any person who “has three previous convictions ... for a violent felony or a serious drug offense”; the contrast between this language and § 924(e)(1) is stark. Many statutes build on predicate offenses whether or not the defendant is or has been convicted of the other crimes. Consider the Racketeer Influenced and Corrupt Organizations Act: the crime defined in 18 U.S.C. § 1962 is the operation of an enterprise through a “pattern of racketeering activity”; § 1961(1) defines “racketeering activity” as the commission of any of a long list of crimes. A person may be convicted under § 1962 whether or not he is or has been convicted of the predicate racketeering activities. Sedi-ma, S.P.R.L. v. Imrex Co., 473 U.S. 479, 488-93, 105 S.Ct. 3275, 3280-83, 87 L.Ed.2d 346 (1985). Similarly the continuing criminal enterprise statute, 21 U.S.C. § 848, heaps punishment on anyone who “violates any provision of this subchapter” if the “violation is part of a continuing series of violations” over which the defendant was a supervisor. 21 U.S.C. § 848(e)(1), (2). The defendant need not be convicted of those other crimes. United States v. Markowski, 772 F.2d 358, 361 (7th Cir.1985). Section 924(c) is similar in concept, requiring proof of a predicate offense without requiring conviction for that offense. Recasting the point as an argument that he did not “carry” the gun does not assist Young. He did not have the gun on his person when he was nabbed, but § 924(c) is *801 not limited to persons who carry their guns at the moment of arrest. United States v. Molina, 102 F.3d 928 (7th Cir.1996), and United States v. Shelby, 121 F.3d 1118 (7th Cir.1997), establish this point. In each of these eases, the police found drugs and a gun together in a “trap” — a secret compartment in the car. In each ease the defendant was empty handed when arrested. Nonetheless, we held in each ease that the trier of fact could infer that the defendant carried the gun in relation to the drugs at an earlier time, perhaps when the two were placed together into the container, or perhaps when the car transported owner, drugs, and gun simultaneously while the gun remained accessible to one who knew about the trap. In Molina the trap was on the driver’s side, and hence more accessible than in Shelby, where the trap was on the other side of the ear, but we held that this difference did not matter. Both drugs and gun remained accessible enough that a jury could conclude that defendant carried both, while the car was moving. Young satisfies the most restrictive possible definition of “carrying” — having a weapon on one’s person (“packing a gun”). He carried the gun to the car. He also carried the gun in the looser sense of transportation. He moved the gun from apartment to car, and the car transported Young, gun, and drugs to the delivery location. The self-protection motive establishes the relation between drugs and gun. If on arriving at the scene Young had thrown the gun down a sewer, or locked it in the trunk, before delivering the drugs, this would not have mattered; the crime had been committed. Suppose, however, Young were apprehended empty-handed, and we knew only that Smith’s car, which Young arrived in, contained a gun. Those stark facts would make the inferential chain more problematic. Did Young carry the gun to the car? Did he know it was in the car? Could he have obtained possession? What relation, if any, was there between gun and drugs (or between defendant and gun)? Cases of that kind — where the record discloses only the location of drugs and gun at the time of arrest — have given rise to divergent views within this circuit. Molina and Shelby strongly suggest that even a gun safely locked away could support an inference that the defendant carried the gun in relation to the drugs; other opinions, such as United States v. Cooke, 110 F.3d 1288 (7th Cir.1997), and Stanback, 113 F.3d at 658, concentrate their attention on the question whether the defendant was carrying the gun (at least constructively, in the sense of having ready access) when he was taken into custody. Perhaps the difference between these opinions is one more of emphasis or advocacy than of substance. Prosecutors often argue § 924(c) cases by stressing the location of the gun at the time of arrest, for that is the only time its location can be established with precision. When eases are presented with this emphasis, it is not surprising that the opinion devotes its analysis to resolving the parties’ dispute, rather than to asking whether the evidence would permit a rational trier of fact to conclude that the defendant carried the gun in relation to a drug crime at some earlier moment. But in principle the latter inquiry is the right one, for reasons we have already covered. Once the defendant has committed all of the elements of the § 924(c) crime, it doesn’t matter how securely locked up, or how far away, the gun becomes. None of this circuit’s opinions holds that it is impermissible to use evidence gathered at the time of arrest to draw inferences about whether the gun was “carried” earlier. But the sense of unease some of them convey is both understandable and appropriate. Bailey tells us that to “use” a gun one must actively employ it; “possession” is not “use.” Likewise, “possess” differs from “carry”. United States v. Jackson, 103 F.3d 561 (7th Cir.1996). If facts that do no more than establish constructive possession demonstrate that the defendant has “carried” the gun, then Bailey has been undone; “possession” as a basis of liability moves from the “use” element to the “carry” element, and Bailey accomplishes nothing substantive. Facts that do no more than show constructive possession thus are weak support for a conclusion that the defendant “carried” a gun. Stir in a few additional facts, however, and the inference of carrying can become compel *802 ling. Suppose a gun and a distribution quantity of drugs are found together, locked in the trunk of the defendant’s car. How did the gun get there? It is possible that some stranger put it there and then locked the trunk. The defendant might be able to show that he gave the keys to a friend and did not own or touch the gun that turned up in the trunk. But in the usual case it is likely that the defendant himself put both drugs and gun in the trunk — and, if so, the defendant “carried” the gun “during” a drug crime such as possession with intent to distribute. Tele-portation is not an option. (The “relation” between drugs and gun found in the same cache is too clear for comment. See Molina, 102 F.3d at 932.) Our cases permit the trier of fact to infer that the defendant culpably carried this gun, once attention has been drawn to the strength of this inferential process rather than the mere location of the gun when the arrest occurred. Young relies heavily on United States v. Malcuit, 104 F.3d 880 (6th Cir.1997), which differs from the law of this circuit by requiring the prosecution to show that the defendant was carrying the gun at, or close in time to, the moment of his arrest. Malcuit delivered five pounds of marijuana from his car. A search after his arrest turned up another 13 pounds in the trunk, $16,000 under the front seat, and a gun in a zippered gym bag on the back seat. The sixth circuit got off on the wrong foot by reviewing the record de novo (104 F.3d at 884) rather than asking whether a rational trier of fact could have found that the defendant committed each element of the offense. The appellate judges chose to accept testimony that Malcuit’s father gave him the gun for reasons unrelated to drugs, and that the weapon played no role in the drug business, although surely a rational jury could have disbelieved this assertion by two interested witnesses (Malcuit and his father). The court continued: During the transaction for which he was convicted of distributing approximately five pounds of marijuana, Malcuit was outside of his vehicle, except for a brief interlude in which Officer Fulmer described Malcuit re-entering “the passenger compartment of his [own] vehicle” to retrieve the paper bag containing the marijuana. The evidence presented to the trial court relating to Malcuit’s re-entering his vehicle is not particularized enough for this court to determine whether the firearm would have been “within effortless reach” at that point. Malcuit was convicted of possessing marijuana with intent to distribute it as well as for the actual distribution to Crispin. Apparently he transported the weapon in the car along with the marijuana he intended to distribute at some future time. The proof disclosed no connection between the possession of marijuana for ultimate distribution and the transportation of the weapon. Rather, there is at most evidence of “mere transportation” of the weapon and “mere possession” of the marijuana. We do not believe the fact that the weapon was transported under these circumstances satisfies the “carry” element of § 924(c)(1), even though it was transported “during” the time Malcuit possessed the marijuana for later distribution. 104 F.3d at 885-86 (bracketed addition in original; citation omitted). It is easy to see why Young embraces this analysis. But we think it deeply flawed — and not only because the appellate judges usurped the role of the trier of fact. These passages confuse the “carry” element with the “relation” element, and they assume a negative answer to what is the real issue: whether events at the time of the distribution may support an inference that the defendant carried the weapon sometime earlier. Malcuit is inconsistent with the law of this circuit and with the appropriate role of appellate courts. We decline to follow it. Several other courts likewise decline to read into “carry” an accessibility-at-the-time-of-arrest requirement. See United States v. Cleveland, 106 F.3d 1056, 1067-68 (1st Cir.1997); United States v. Mitchell, 104 F.3d 649, 653-54 (4th Cir.1997); United States v. Muscarello, 106 F.3d 636, 638 (5th Cir.1997). (Malcuit’s status in its home court is shaky, too; the sixth circuit vacated the panel opinion and granted rehearing en banc on June 12, 1997.) Any other circuit tempted to limit § 924(c) to events at the time of arrest — see United States v. Cruz-Rojas, 101 F.3d 283, 285-86 (2d Cir.1996); *803 United States v. Hernandez, 80 F.3d 1253, 1257-58 (9th Cir.1996) — cannot count us as an ally. Young might have tried to persuade a trier of fact that the gun he carried with him when he set out to deliver cocaine lacked a relation to the impending sale. Instead he pleaded guilty. He conceded that what could have been branded “just inferences” was true. There was a relation between the gun he carried and his drug dealing. His conviction is entirely proper. Affirmed.