26 F.4th 627

4th Cir.

2022

UNITED STATES OF AMERICA, Plaintiff - Appellee, v. ANTHONY R. BUSTER, a/k/a Anthony Raymond Buster, a/k/a Blue Brown Harlem, Defendant - Appellant.

No. 21-4101

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

February 22, 2022

PUBLISHED. Argued: December 7, 2021. Appeal from the United States District Court for the Eastern District of Virginia, at Richmond. John A. Gibney, Jr., Senior District Judge. (3:20-cr-00034-JAG)

Before GREGORY, Chief Judge, RICHARDSON and HEYTENS, Circuit Judges.

Reversed, vacated, and remanded by published opinion. Judge Heytens wrote the opinion, in which Chief Judge Gregory joined. Judge Richardson wrote a dissenting opinion.

ARGUED: Caroline Swift Platt, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Alexandria, Virginia, for Appellant. Aidan Taft Grano-Mickelsen, OFFICE OF THE UNITED STATES ATTORNEY, Alexandria, Virginia, for Appellee. ON BRIEF: Geremy C. Kamens, Federal Public Defender, Nia Ayanna Vidal, Assistant Federal Public Defender, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Alexandria, Virginia, for Appellant. Raj Parekh, Acting United States Attorney, Alexandria, Virginia, Heather Hart Mansfield, Assistant United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Richmond, Virginia, for Appellee.

TOBY HEYTENS, Circuit Judge:

The Fourth Amendment forbids “unreasonable searches and seizures.” A search is constitutionally reasonable if it is justified to protect police officers from threats posed by those who “may be armed and presently dangerous.” Terry v. Ohio, 392 U.S. 1, 30 (1968). The question here is whether that doctrine can be stretched to cover a warrantless search of a bag recently possessed by a person who was—by the time the bag was opened—handcuffed and face-down on the ground. At least on these facts, the answer is no.

I.

Because this appeal arises from the denial of a motion to suppress, we view the facts in the light most favorable to the government. United States v. Black, 707 F.3d 531, 534 (4th Cir. 2013). Although we review the district court‘s “findings of historical fact for clear error,” “[w]e review de novo the ultimate legal conclusion of whether reasonable suspicion existed to justify police action.” United States v. McCoy, 513 F.3d 405, 410 (4th Cir. 2008).

II.

A.

At 11:18 p.m. on September 22, 2019, two officers in a patrol car approached Anthony Buster as he walked along Fairfield Avenue in Richmond, Virginia. About 30 minutes earlier, the officers had responded to a report of “a domestic assault where a firearm discharged in the air” and had been looking for the assailant ever since. JA 40. The officers approached Buster for two reasons: they believed he matched witness descriptions of the assailant and that he was the person they had seen outside the victim‘s apartment earlier that evening.

After getting out of the patrol car, one officer said “Yo! Let me talk to you real quick” and motioned for Buster to come over. JA 233. Buster said “Nah,” and continued walking. JA 233. The same officer said “Yo! Hey!” and continued toward Buster. JA 234. At that point, Buster took off running but tripped and fell almost immediately.

The officers caught up with Buster while he was still on the ground and tackled him. Buster was wearing “a single-strap bag that goes across your body” whose pouch had “ended up in front of ” Buster when he fell. JA 56–57. Perceiving that Buster was clutching or reaching for the bag, the officers pulled Buster‘s arm away from the bag, pulled the bag to the rear of Buster‘s body, and handcuffed him. Buster said the bag‘s strap was choking him, so one of the officers cut the strap, grabbed the bag, and removed it from Buster‘s person. The bag felt “[h]ard to the touch,” which in the officer‘s “experience . . . indicates . . . a weapon.” JA 136. The officer opened the bag and found a gun and a box of ammunition.

The officers also peppered Buster with questions without giving him the familiar Miranda warnings. After discovering additional ammunition in Buster‘s pants pocket, an officer asked if Buster had anything else on his person. Buster responded that the only items he had were “that gun and my knife.” U.S. Ex. 1B at 11:24:40–45 p.m. The officers took Buster to the police station, where they asked more questions. Eventually, an officer realized “he had neglected to read Buster his Miranda rights” and left the room. JA 234. About ten minutes later, the officer came back, gave Buster the Miranda warnings, and elicited “‘essentially’ the same material discussed in the pre-Miranda interview.” JA 235.

B.

Buster was charged with one count of possessing a firearm after having been convicted of a felony. Soon after, he filed a motion to suppress the firearm, the ammunition, and his various statements as having been obtained in violation of the Fourth and Fifth Amendments. The district court held a hearing where the officers testified and video from their body-worn cameras was admitted into evidence.

The district court granted Buster‘s motion in part and denied it in part. By then, the government had agreed it would not seek to use many of the pre-Miranda statements, rendering moot Buster‘s request to suppress them. The court granted Buster‘s request to suppress his post-Miranda statements, concluding they were “the product of an impermissible two-step interview tactic” and thus barred by Missouri v. Seibert, 542 U.S. 600 (2004), and United States v. Mashburn, 406 F.3d 303 (4th Cir. 2005). JA 243.

The district court denied Buster‘s motion to suppress in all other respects. The court concluded the initial stop was valid because “the officers had reasonable suspicion that Buster was the suspect in a reported domestic assault incident potentially involving a gun” and that “the pat-down of Buster‘s person and the search of his bag were reasonable” because “the officers had reason to believe they were dealing with an armed and dangerous person.” JA 240. The court also declined to suppress Buster‘s on-the-scene statement referencing the already discovered firearm, concluding it fell within the public-safety exception of New York v. Quarles, 467 U.S. 649 (1984).

After the district court‘s ruling, Buster and the government reached a plea agreement. The written agreement stated that Buster was “pleading guilty conditionally under United States v. Bundy, 392 F.3d 641 (4th Cir. 2004),” and that it “preserve[d]” Buster‘s “right to appeal the denial of his motion to suppress.” JA 249. The district court accepted the plea and sentenced Buster to 51 months of imprisonment.

III.

Before turning to the merits of Buster‘s appeal, we must address the government‘s assertion that we are powerless to do so. We disagree.

The general rule is that a valid guilty plea “waives all nonjurisdictional defects in the proceedings conducted prior to entry of the plea.” Bundy, 392 F.3d at 644. Since 1983, however, the Federal Rules of Criminal Procedure have codified a practice—known as a conditional guilty plea—allowing some defendants to “enter a plea of guilty . . . while preserving certain pretrial issues for appeal.” Id. at 645. Specifically, Rule 11(a)(2) provides that “[w]ith the consent of the court and the government, a defendant may enter a conditional plea of guilty or nolo contendere, reserving in writing the right to have an appellate court review an adverse determination of a specified pretrial motion.”

In Bundy, this Court confronted a situation where a defendant attempted to use the conditional guilty plea mechanism to appeal the denial of “a motion for production of certain documents.” 392 F.3d at 648. The Court rejected that effort, reasoning that the “discovery issue” raised by Bundy‘s motion to compel was “not case-dispositive” because a ruling in his favor would do nothing more than allow him to “see certain documents and decide whether they help his defense.” Id. And even though Bundy‘s plea agreement also purported to preserve his ability to appeal two other issues that were “the proper subjects of a conditional guilty plea“—a Fourth Amendment unreasonable search claim and a Fifth Amendment compelled self-incrimination claim—the Court concluded that the presence of the discovery issue so “taint[ed] the entire plea” as to “render[] the entire plea invalid.” Id. at 648–49. Accordingly, the Court vacated Bundy‘s convictions and remanded without reaching the merits of any of Bundy‘s claims. Id. at 649–50.

Bundy does not preclude us from considering Buster‘s claims. For one thing, unlike the discovery dispute that troubled the Court in that case, each issue preserved by Buster‘s written plea agreement satisfies Bundy‘s definition of “case-dispositive.” Echoing language from Rule 11‘s advisory committee notes, Bundy stated that an issue is “case-dispositive if . . . a ruling in the defendant‘s favor would require dismissal of the charges or suppression of essential evidence.” 392 F.3d at 648 (emphasis added); see also Fed. R. Crim. P. 11 advisory committee‘s note to 1983 amendments (contemplating conditional guilty pleas will occur “only when the decision of the court of appeals will dispose of the case either by allowing the plea to stand or by such action as compelling dismissal of the indictment or suppressing essential evidence“). The advisory committee notes repeatedly identify the denial of “suppression motions” as the quintessential situation where conditional guilty pleas are appropriate—to the point of including language underscoring that ”Rule 11(a)(2) is not limited to” such circumstances. And here, suppressing evidence is exactly—and exclusively—what Buster seeks.

The government responds that the un-Mirandized statement Buster made about the firearm at the scene is not truly essential evidence because (the government maintains) it would have prevailed at trial even without that statement. The government would thus have a reviewing court ask—before reaching the merits of any Fourth or Fifth Amendment issue preserved via Rule 11—whether a defendant who in reality chose to plead guilty after having failed to suppress a particular piece of evidence would have been convicted at a hypothetical trial where the government was unable or chose not to use the very evidence it had previously and successfully fought to keep in. Neither Bundy, the text of Rule 11, nor the advisory committee notes charge appellate courts with conducting that sort of odd counterfactual inquiry as a necessary prelude to considering an otherwise-proper appeal.¹

And there is more. Bundy specifically acknowledged its result may have been different had the various issues the defendant sought to preserve been “inextricably intertwined,” 392 F.3d at 649 (quotation marks omitted)—a description fitting this situation to a T. In Bundy, the defendant sought to plead guilty while preserving for appellate review three issues with little overlap in underlying facts or law that had been raised in three different pretrial motions requesting three different forms of relief (production of documents, suppression of evidence, and dismissal of certain counts). Id. at 644. Here, in contrast, Buster‘s written plea agreement preserved his ability to appeal the denial of a single motion requesting a single form of relief (suppression), and all the arguments that Buster sought to preserve for appeal depend on the same basic evidence and involve closely related legal doctrines. Neither of the primary aims cited in Bundy—“judicial economy” and ensuring “that the conditional guilty plea not be employed in a manner that renders appellate review difficult or impossible,” 392 F.3d at 646—would be served by artificially separating our consideration of Fourth and Fifth Amendment issues that are inextricably intertwined.

Although our resolution of this case is fully consistent with Bundy‘s holding and result, we acknowledge tension with broad language in the Court‘s opinion about what constitutes a case-dispositive issue. As Chief Justice Marshall reminded us long ago, however, “general expressions, in every opinion, are to be taken in connection with the case in which those expressions are used,” Cohens v. Virginia, 19 U.S. 264, 399 (1821), and this Court has directly—and recently—disclaimed the notion “that everything said in a panel opinion binds future panels,” Payne v. Taslimi, 998 F.3d 648, 654 (4th Cir. 2021). The broad language on which the government relies was not “necessary to the outcome” (id.) in Bundy—which, as noted earlier, involved an attempt to use a conditional guilty plea to obtain appellate review of a pretrial discovery motion and flagged denials of suppression motions as situations where conditional guilty pleas may be used. For that reason, we read Bundy‘s “general language” the same way courts “often read general language in judicial opinions—as referring in context to circumstances similar to the circumstances then before the Court and not referring to quite different circumstances that the Court was not then considering.” Illinois v. Lidster, 540 U.S. 419, 424 (2004).

IV.

Buster raises several objections to the district court‘s denial of his motion to suppress, but we reach only one of them. We hold that the district court erred in denying Buster‘s motion to suppress the firearm because the sole theory the government has pressed in support of that result does not apply here.

Because suppression of relevant evidence is always a “last resort, not [a] first impulse,” Hudson v. Michigan, 547 U.S. 586, 591 (2006), the Supreme Court has identified numerous grounds on which a given search may be deemed constitutionally reasonable or suppression may otherwise be denied. If officers have probable cause to believe a particular place or item contains contraband or evidence of a crime, they can get a warrant to search it. See, e.g., Riley v. California, 573 U.S. 373, 381–82 (2014). Nor is a warrant always required, including “when the exigencies of the situation make the needs of law enforcement so compelling that a warrantless search is objectively reasonable under the Fourth Amendment.” Mincey v. Arizona, 437 U.S. 385, 394 (1978) (quotation marks omitted). In addition, officers who make “a lawful arrest” may conduct a warrantless search of an “arrestee‘s person and the area within his immediate control” as an “incident to” that arrest, United States v. Davis, 997 F.3d 191, 195 (4th Cir. 2021) (quotation marks omitted), and later may conduct an “inventory search” of the arrestee‘s possessions so long as it is done “according to standardized criteria, such as a uniform police department policy” and is “performed in good faith,” United States v. Seay, 944 F.3d 220, 223 (4th Cir. 2019) (quotation marks omitted).

Even when a search was constitutionally unreasonable, the government still may have arguments for why any resulting evidence should not be suppressed. It may argue, for example, that the challenged search was conducted in good-faith reliance on a facially valid warrant or then-binding judicial precedent. See, e.g., United States v. Leon, 468 U.S. 897, 922 (1984); Davis v. United States, 564 U.S. 229, 232 (2011). The government also may argue that “evidence gathered in an otherwise unreasonable search” would have been “ultimately or inevitably discovered . . . by lawful means,” including a search incident to arrest or an inventory search. Seay, 944 F.3d at 223 (quotation marks omitted).

Here, however, the district court identified only one basis for denying Buster‘s motion to suppress the firearm found during the search of his bag, and the government has never offered any other. Specifically, the district court concluded that the “search of [Buster‘s] bag” was constitutionally reasonable under the protective search doctrine associated with Terry v. Ohio, 392 U.S. 1 (1968). JA 240. On the facts of this case, we respectfully disagree.

The Supreme Court has repeatedly emphasized that “[t]he purpose of ” the “limited search” authorized by Terry “is not to discover evidence of crime, but to allow the officer to pursue his investigation without fear of violence.” Minnesota v. Dickerson, 508 U.S. 366, 373 (1993) (quoting Adams v. Williams, 407 U.S. 143, 146 (1972)); accord Terry, 392 U.S. at 29. Indeed, Terry itself states that such searches may lawfully be conducted as part of an investigatory stop only when an officer “reasonably . . . conclude[s] in light of his experience” that “the persons with whom he is dealing may be armed and presently dangerous.” 392 U.S. at 30 (emphasis added). For that reason, “a protective search—permitted without a warrant and on the basis of reasonable suspicion less than probable cause—must be strictly limited to that which is necessary for the discovery of weapons which might be used to harm the officer or others nearby.” Dickerson, 508 U.S. at 373 (quotation marks omitted).

Although the parties dispute many things about the events of the evening, there is no disagreement about one critical fact: When the officer opened Buster‘s bag (thus beginning a “search” of the bag), Buster was handcuffed on the ground and had no access to it. Indeed, the record is clear that the officers opened the bag and examined its contents after they had tackled Buster, handcuffed him, cut the bag off his body, and “move[d] it away from his person.” U.S. Br. 9. The government offers no explanation for how the contents of the bag presented any credible threat to the officers’ safety at the time they searched it, and quickly frisking an unsecured suspect or a bag during a Terry stop is simply not the same as methodically searching the contents of a bag to which a suspect no longer has access—particularly where the suspect remained restrained and under the officers’ physical control. “Having already used significant force to secure the scene for safety purposes, the officers cannot leverage the safety rationale into a justification for a full-scale search.” United States v. Miles, 247 F.3d 1009, 1015 (9th Cir. 2001).

In arguing to the contrary, the government relies on the searching officer‘s testimony that, when she removed the bag from Buster‘s body, she noticed it “was hard to the touch,” which, “[i]n [her] experience . . . indicates . . . a weapon.” JA 135–36. But even assuming the officer had reasonable suspicion that the bag contained a weapon (a point we need not decide), that fact alone could not generate reasonable suspicion that Buster was “presently dangerous” after he was already restrained and no longer had access to the bag. Terry, 392 U.S. at 30.²

The government‘s reliance on the officers’ suspicion that Buster may have discharged a firearm earlier in the evening fails for the same reason. Even if the officers had reasonable suspicion that was true (another point we need not decide), the likelihood that the officers would find a firearm or other weapon in the bag has no bearing on the justification for a protective frisk after Buster had been separated from the bag and no longer had access to it. Cf. Davis, 997 F.3d at 200 (search incident to arrest doctrine did not justify search of a backpack “not within reaching distance of ” the arrestee when the officer searched it).

We emphasize the limits of our holding. We do not address situations where a firearm was found on a suspect‘s person or a bag was opened before a suspect was subdued or while they were still within reach of the bag. Cf. United States v. Walker, 615 F.3d 728, 730, 733–34 (6th Cir. 2010) (approving limited Terry search of bag where armed robbery suspect was not restrained and officers “by no means had the scene under control or their safety secure“). This case does not present (and the parties have not raised) any question about whether or when officers may search a bag before returning it once a Terry stop concludes. Cf. id. at 734 (describing situation where the “only alternative” to a protective frisk was “to give a suspect access to a potential weapon” located “in an un-searched bag“). Because the government has never so argued, we do not consider whether at some point the officers might have acquired probable cause to arrest Buster for assault or some other offense and, if they did, whether the district court‘s decision not to suppress the firearm could have been justified on some other ground. We hold only—but importantly—that a doctrine authorizing a limited warrantless search to protect officer safety cannot be stretched to cover situations where there is no realistic danger to officer safety. Accordingly, we reverse the district court‘s denial of Buster‘s motion to suppress the firearm.

That leaves only the issue of remedy. Although the government argues that any error in denying Buster‘s motion to suppress his on-the-scene statement about the firearm would have been harmless, it makes no such claim about the district court‘s failure to suppress the firearm itself. The government‘s silence on this point makes sense given the circumstances of this case, where the underlying charge is predicated on Buster possessing a firearm we now hold must be suppressed. For that reason, we need not address whether—and, if so, how—harmless-error analysis applies in the context of conditional guilty pleas. Compare United States v. Lustig, 830 F.3d 1075, 1087 (9th Cir. 2016) (court reviewing a conditional guilty plea may deem an error harmless if “the government has proved beyond a reasonable doubt that the erroneously denied suppression motion did not contribute to the defendant‘s decision to plead guilty“), with id. at 1092 (Watford, J., concurring) (arguing that “harmless error analysis . . . has no place in this context” and that if a reviewing court “does anything other than affirm in full the district court‘s denial of [the defendant‘s] suppression motion, he is entitled to withdraw his guilty plea without more“). Instead, we hold only that this firearm should have been suppressed and that Buster must be given an opportunity to withdraw his guilty plea. See Fed. R. Crim. P. 11(a)(2) (“A defendant who prevails on appeal may . . . withdraw the plea.“).³

* * *

The government‘s motion to vacate the guilty plea and remand for further proceedings without reaching the merits is denied. The district court‘s order denying Buster‘s motion to suppress the firearm is reversed, the judgment of conviction is vacated, and the case is remanded for further proceedings consistent with this opinion.

SO ORDERED

RICHARDSON, Circuit Judge, dissenting:

In United States v. Bundy, this court held that conditional guilty pleas are valid under Rule 11(a)(2) only if all the issues they preserve are “case-dispositive“—so central to the government‘s case that a ruling for the defendant would effectively require dismissal of the charges against him. 392 F.3d 641, 648 (4th Cir. 2004). Bundy‘s rule is textually baseless, pragmatically unjustified, and entirely binding on this panel. And it commands that Buster‘s attempt to preserve a non-dispositive Miranda issue for appeal renders his entire conditional plea invalid. But rather than apply Bundy faithfully to the facts before us, the majority warps Bundy‘s core holding and adds to it a new “inextricably-intertwined” exception that promises to vex litigants for decades to come. While I would fully support abandoning Bundy through the proper channel—an en banc rehearing—I cannot condone the majority‘s sly revision of it here. I respectfully dissent.

I.

For most of our nation‘s history, a defendant‘s guilty plea would waive his right to appeal the resolution of pretrial motions. See Tollett v. Henderson, 411 U.S. 258, 267 (1973). This, when coupled with the defendant‘s inability to bring an interlocutory appeal on such issues, see Di Bella v. United States, 369 U.S. 121, 131–32 (1962), meant that “a defendant who has lost one or more pretrial motions will often go through an entire trial simply to preserve the pretrial issues for later appellate review.” Fed. R. Crim. P. 11 advisory committee note to 1983 Amendment. To combat this waste of judicial resources, some (but not all) circuits permitted defendants to plead guilty conditionally. Id.; compare United States v. Burke, 517 F.2d 377, 379 (2d Cir. 1975) (permitting conditional pleas so long as the prosecutor does not object to their entry), with United States v. Matthews, 472 F.2d 1173, 1174 (4th Cir. 1973) (holding that a guilty plea “swallows up” pretrial errors).

But in 1983, Federal Rule of Criminal Procedure 11(a)(2) was amended to expressly allow conditional guilty pleas: “With the consent of the court and the government, a defendant may enter a conditional plea of guilty or nolo contendere, reserving in writing the right to have an appellate court review an adverse determination of a specified pretrial motion.” The Rule‘s text does not limit the issues that the government and the defendant—with district court‘s consent—may preserve via a conditional plea.

Nonetheless, we imposed such a requirement (and a strict one at that) in Bundy. There we held that “a valid guilty plea preserves for appellate review only case-dispositive pretrial issues.” Bundy, 392 F.3d at 647. Therefore, “[a] district court should reject any conditional guilty plea that purports to preserve for appellate review pretrial issues that, in the district court‘s own judgment, are not fully case-dispositive.” Id. Bundy then strictly defined “case-dispositive“: “The disposition of a pretrial issue is case-dispositive if (1) a ruling in the defendant‘s favor would require dismissal of the charges or suppression of essential evidence, or (2) a ruling in the Government‘s favor would require affirming the conviction. In short, there should be no trial after the specified issues are resolved by the court of appeals.” Id. at 648.

One searches in vain for any textual basis for Bundy‘s strict limitation on conditional pleas. Instead, Bundy appears to have derived its rule from decisions in other circuits that either predate the 1983 enactment of Rule 11(a)(2) or are themselves devoid of any textual analysis. See id. at 646–47. In doing so, it advanced two policy justifications for its atextual creation: judicial economy and adequacy of review. But even on their own terms neither support the rule Bundy imposed.

First, the Bundy court worried that permitting a conditional plea on minor issues might stall the case, citing concerns expressed in Rule 11‘s advisory committee note. Id. at 645–46. But the advisory committee did not say that a judicial case-dispositive rule was necessary to address that concern; instead, it concluded that the government-approval requirement was adequate to avoid a deluge of dilatory conditional pleas. Fed. R. Crim. P. 11 advisory committee note to 1983 Amendment. That reasoning makes sense. It may be true that the government does not have a direct incentive to protect judicial economy. But it does have strong parallel incentives to avoid excessive appeals. In most cases, the government‘s interest in preserving its own resources and avoiding unjustified delay will motivate it to not sanction appeals over trivial issues.

There may be exceptions. And mistakes may be made. But vacating and remanding when the government, the defendant, and the court agree to permit a conditional plea fails to promote judicial economy. To the contrary, rejecting this conditional plea appears to demand unnecessary proceedings below and, likely, a second appeal. When a non-dispositive issue is intentionally preserved (like in Bundy), it is likely because the issue is salient to the defendant, perhaps irrationally so. In that instance, Bundy‘s rule requires the defendant to go to trial solely to preserve the issue, leading to a considerable waste of resources to put on an unnecessary trial.

Second, Bundy worried that, unless our appellate court restricts the permissible subjects of a conditional appeal, we may be forced to pass on an issue not yet adequately developed. 392 F.3d at 646–47. But it is unclear why our self-doubt about resolving some issues justifies setting aside pleas authorized by Rule 11, agreed to by both parties, and blessed by the district court.¹ After all, we have a “virtually unflagging” obligation to decide cases properly before us, even if they might be hard sometimes. Lexmark Int‘l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 126 (2014). If a party or the district court doesn‘t think we will be able to reach an adequate decision based on the record as it exists, they don‘t have to agree to the plea.² But once they have, we have a duty to do our best to resolve the preserved issues.

And even if adequacy of review were a real concern, it is unclear why case-dispositiveness is a reasonable proxy for adequate-reviewability. For example, the Rule 11 advisory committee identified a speedy trial violation as an issue that is “best considered only after the relevant facts have been developed at trial.” Fed. R. Crim. P. 11 advisory committee note to 1983 Amendment. Yet a speedy trial violation requires the automatic dismissal of the indictment, meaning that it is case-dispositive. See Barker v. Wingo, 407 U.S. 514, 522 (1972). On the other hand, motions to suppress evidence may be resolved based on undisputed facts, yet not be case-dispositive.

One particularly puzzling argument advanced by Bundy is that “[t]he adequacy of the factual record is especially important for harmless-error analysis” and therefore, “[p]ermitting conditional guilty pleas to preserve non-case-dispositive pretrial issues for appeal would undermine harmless-error analysis.” 392 F.3d at 647. But it is unclear why we are in a better position to determine whether an issue is case-dispositive than we are to determine whether an issue is harmless. The two are basically the same question, just with different thresholds. Bundy‘s standard is essentially a harmful-error test, where we must determine whether the exclusion of the evidence would be fatal to the government‘s case, as compared to a harmless-error test where we determine whether it was inconsequential to that case. In that sense, I agree with the majority that the government is asking us to engage in a bizarre “hypothetical trial where the government was unable or chose not to use the very evidence it had previously and successfully fought to keep in.” Majority Op. 7. But the majority is wrong to place the fault with the government; Bundy itself requires that “odd counterfactual inquiry.” Majority Op. 7.³

In short, Bundy created a judicial doctrine unsupported by either the text of Rule 11 or sound policy. And that doctrine is binding on this panel. One panel of this court cannot overrule another. McMellon v. United States, 387 F.3d 329, 332 (4th Cir. 2004) (en banc). Panels may well disagree. But we have decided that an opinion produced by a panel states rules of law that can—indeed must—be followed. So we must adhere to prior panel decisions no matter how strongly we disagree: “We must follow a prior panel decision even if it had abysmal reasoning, put forward unworkable commands, engendered no reliance interests, lacked consistency with other decisions, and has been undermined by later developments.” Payne v. Taslimi, 998 F.3d 648, 654 n.2 (4th Cir. 2021).

And adhering to Bundy here produces a clear result: Buster‘s plea must be vacated. Buster sought to preserve two issues for appeal: (1) a motion to suppress the gun and ammunition found during the stop based on the Fourth Amendment; and (2) a motion to suppress his un-Mirandized statement that he had “that gun and my knife” based on the Fifth Amendment. The first issue involves the “suppression of essential evidence” and is case-dispositive; the government cannot possibly prove that Buster possessed a firearm that had travelled in interstate commerce that night if the gun and ammunition are suppressed.⁴ The second issue is not case-dispositive. Even if it the statement about the

gun is suppressed, the gun itself provides the evidence necessary to prove that Buster possessed a firearm that had travelled in interstate commerce, and other evidence from that night provides strong proof that he did so knowingly. The statement might be icing on the prosecution‘s cake, but it is not essential to a conviction and so is not case-dispositive.

Bundy is also clear on the effect of preserving both dispositive and non-dispositive issues in a conditional plea: “the presence of one non-case-dispositive issue in this conditional plea renders the entire plea invalid.” 392 F.3d at 649. So, because Buster‘s plea tried to preserve a non-dispositive Miranda issue, his entire plea is invalid under Bundy and must be vacated.

II.

The majority reaches the contrary outcome, holding that Buster‘s conditional plea was valid. In doing so, the majority purports to follow Bundy. But it does so by twisting the language of Bundy to reach the answer it wants, rather than faithfully applying it.

The majority‘s analysis starts out well enough, identifying Bundy‘s core holding that only “case-dispositive” issues may be preserved, and that an issue is “case-dispositive if . . . a ruling in the defendant‘s favor would require dismissal of the charges or suppression of essential evidence.” 392 F.3d at 648. The obvious next question for someone truly interested in discerning the meaning of Bundy‘s holding would be: What is “essential evidence“? That is a question the majority never bothers to ask, probably because it doesn‘t like Bundy‘s answer. That answer can be found, in fact, in the very next sentence of the Bundy opinion: “In short, there should be no trial after the specified issues are resolved by the court of appeals.” Consistent with that explanation, the concluding sentence of that paragraph summarizes: “The question for the district court [on whether to approve a conditional plea] is whether the court of appeals’ resolution of the issue specified in the plea would end the case one way or the other.” Id. These statements leave no doubt: when the court said that denial of suppression motions could be preserved only if they involved “essential evidence,” it meant that they must involve essential evidence. Not any evidence. Not material evidence. Not important evidence. Essential evidence—evidence so “absolutely necessary” or “indispensable” that its suppression would force the government to dismiss the charge, thus disposing of the case. Essential, Random House Webster‘s College Dictionary (2000).

Instead, the majority implies (though it doesn‘t expressly hold) that any suppression motion necessarily involves essential evidence and so is case-dispositive. Majority Op. 6 (holding that Bundy is satisfied because “suppressing evidence is exactly—and exclusively—what Buster seeks“). That interpretation amazingly manages to redefine “essential” as “nonessential” and “dispositive” as “non-dispositive.” And it requires the majority to dismiss as dicta all of Bundy‘s language clarifying what “case-dispositive” and “essential evidence” mean.

But Bundy‘s “broad language” defining its case-dispositive rule is far from dicta. It is true, of course, that not every word in a panel decision is binding; “peripheral” remarks that do not represent the court‘s “full and careful consideration” are dicta and are not controlling. Payne, 998 F.3d at 654. But we aren‘t dealing here with peripheral remarks detached from Bundy‘s holding. We are dealing with language defining what Bundy‘s holding actually means. That definition is no doubt broader than the majority would like. But that doesn‘t transform those statements into dicta.

The only statements strictly “necessary” to resolve a case are “affirmed” and “reversed.” Yet all agree that a court‘s holding encompasses more than this judgment; it includes the court‘s ratio decidendi—the chain of reasoning necessary to tie the facts of the case to the decision reached. Dicta is what remains: those statements that “could have been deleted without seriously impairing the analytical foundation of the holding.” Payne, 998 F.3d at 654 (citations omitted). Though that distinction is often difficult to make, see generally Michael Abramowicz & Maxwell Stearns, Defining Dicta, 57 Stan. L. Rev. 953 (2005), it is essential that we exercise care in doing so. If we simply dismiss any statement we disagree with as ‘mere dicta,’ we are no longer acting as judges, but as policymakers—deciding which words from precedent we like and selectively applying them. See Michael C. Dorf, Dicta and Article III, 142 U. Pa. L. Rev. 1997, 2029–30 (1994).

So let us carefully consider the decision in Bundy. That defendant entered a conditional plea preserving for appeal certain pretrial decisions, including the denial of a discovery motion. Bundy, 392 F.3d at 644. The result of Bundy was the vacatur of that plea. Id. at 650. So what reasoning linked those facts to that result? First, the court held that only “fully case-dispositive” issues could be preserved for appeal. Id. at 647. Second, the court defined case-dispositive: a ruling favorable to the defendant must “require dismissal of the charges or suppression of essential evidence,” such that there would be “no trial after the specified issues are resolved.” Id. at 648. Third, it applied this definition of “case-dispositive” to the discovery issue, observing that a “favorable ruling on this issue will not guarantee dismissal of Count Two.” Id.5 Finding that the discovery issue was therefore not “case-dispositive” under its definition of that term, the court vacated the plea. Id. at 650.

The majority seeks to redefine that second step out of existence, narrowing Bundy to a case about “a pretrial discovery motion.” Majority Op. 8. The Bundy court could have done that. It could have limited its decision to discovery issues. But it didn‘t. To the contrary, it held that “a valid conditional guilty plea preserves for appellate review only case-dispositive pretrial issues.” Id. at 647. We cannot now substitute a narrower holding Bundy could have advanced for the one it actually did. Brown v. Board of Education is about more than a school in Kansas. 347 U.S. 483, 495 (1954) (holding that “in the field of public education the doctrine of ‘separate but equal’ has no place“). Marbury v. Madison is about more than William Marbury‘s commission. 5 U.S. (1 Cranch) 137, 180 (1803) (holding that “a law repugnant to the Constitution is void“). And Bundy is about more than a discovery dispute. 392 F.3d at 648 (holding that the resolution of an issue in a conditional plea must “end the case one way or the other“). Bundy stated a rule of law applicable to all conditional pleas: they must preserve only issues that would dispose of the case. The Miranda issue preserved here would not dispose of the case. And so Bundy prohibits including it in the plea.

The majority also puts forth an alternative rationale for its decision: a postulated exception to the Bundy rule that permits us to review non-case-dispositive issues that are “inextricably intertwined” with case-dispositive ones. Ironically, while casting aside Bundy‘s core holding as dicta, the majority relies heavily on Bundy‘s discussion of this possible exception, which was dicta. Bundy never held that the exception exists, but simply found that if it exists it did not apply under the facts presented there. 392 F.3d at 649. Yet the majority today affirms the exception‘s existence and concludes that it is satisfied.

Unlike its attempt to redefine Bundy‘s holding, the majority has the authority to mint this new exception. Bundy left the door open, and we may walk through it if we wish. But I think it unwise. Because, as a practical matter, the majority‘s approach only makes Bundy—already a trap for unwary defendants and prosecutors—even more complex and hazardous. Future parties to an incipient plea agreement now need to embark on a multi-step test to decide what they can and cannot preserve for appeal. First, they need to decide whether each issue is “case-dispositive.”⁶ Then, the parties must proceed to the new Bundy-Buster Step 2: Are the issues that are not “case-dispositive” nonetheless “inextricably intertwined” with those that are? The majority does not say what it means for two issues to be “inextricably intertwined.” At best, litigants may look to the four

probative factors identified by the majority in this case: (1) the two issues appealed were raised in the same motion; (2) they would warrant the same general type of remedy; (3) they rest on the “same basic evidence“; and (or?) (4) they involve “closely related legal doctrines.”

But this “standard” raises even more questions. What does it mean for two issues to turn on “the same basic evidence“? Here, the only “basic evidence” relevant to the majority‘s Fourth Amendment ruling is that Buster was handcuffed and on the ground at the time of the search. The Miranda issue, by contrast, depends on whether the question eliciting Buster‘s incriminating statement was justified by the officer‘s need to protect himself while searching a now-standing Buster. So the evidence pertinent to each issue is not really the same. I suppose the two events occurred during the same police interaction—is that what it takes for two issues to depend on the “same basic evidence“?

And what counts as “closely related legal doctrines“? Here, Buster seeks to suppress two different types of evidence based on different provisions of the U.S. Constitution—the Fourth and Fifth Amendments. Are those two doctrines “closely related” just because they both involve constitutional rules of criminal procedure? What if Buster‘s appeal also involved a deprivation of post-indictment counsel in violation of the Sixth Amendment? Is that doctrine also “closely related“? What about warrant defects? Speedy trial violations? Double jeopardy? We don‘t know. And neither, more importantly, do defendants, prosecutors, or district courts.

If the Bundy-Buster doctrine were based on the text of Rule 11 or the common law, litigants and district courts might have some way of predicting the answers to these thorny questions. But because both Bundy‘s rule and the majority‘s judicial exception were created from whole cloth, litigants and courts can only guess the answers and hope that we agree with them on appeal.⁷ While the majority, like Bundy before it, justifies its decision as in the interest of “judicial economy,” in truth it will only create more confusion and waste for future defendants, prosecutors, and judges.

* * *

All of this is not to say that we should be stuck with Bundy‘s stringent judge-created rule forever. Though one panel of our court cannot overrule another, the entire court, sitting en banc, can. Jones v. Angelone, 94 F.3d 900, 905 (4th Cir. 1996). But in the meantime, Bundy remains binding law we should faithfully apply. So I would apply Bundy to these facts, bar Buster‘s appeal, and give a full-throated endorsement for overturning Bundy during reconsideration en banc.⁸ Instead, the majority goes to great lengths to refashion the Bundy rule to suit its own liking. But just as in Bundy, there is nothing in Rule 11 to justify the majority‘s redefinition of “case-dispositive” or its new inextricably intertwined exception. So its decision today simply adds made-up alterations and exceptions to a made-

up rule. I decline to join in this tailoring of the emperor‘s new clothes, and respectfully dissent.

Notes

Although such an approach finds support in certain non-precedential opinions, see, e.g., , we expressly disclaim it here. We also make clear that whether a particular conditional guilty plea comports with has no impact on our appellate jurisdiction because is simply a gloss on Rule 11 and “any jurisdictional requirements found in [federal rules] exist by virtue of the Constitution and statutes, not the rules themselves.” ; accord (stressing the need to avoid “profligate” use of the term “jurisdiction“—“a word of many, too many, meanings” (quotation marks omitted)); (Becker, J., concurring) (explaining that “[t]he issue of jurisdiction is separate from the issue . . . of whether the rules provide a procedural mechanism by which” a court may decide a particular issue). A district court exercising a discretionary power may, at times, abuse that discretion, including by failing to recognize “judicially-recognized factors limiting its discretion.” . But that does not give us carte blanche to impose any limitations we feel like on the district court. “Courts are the mere instruments of the law, and can will nothing. When they are said to exercise a discretion, it is a mere legal discretion, a discretion to be exercised in discerning the course prescribed by law; and, when that is discerned, it is the duty of the Court to follow it.” (Marshall, C.J.). Here, Rule 11(a)(2) opens our doors to conditional pleas predicated on pretrial rulings; we cannot close those doors to all but a select few defendants.

Although the “plain feel” doctrine may be relevant when an officer “discovers contraband through the sense of touch during an otherwise lawful search,” , the officer never testified that she had reason to believe the bag contained contraband (as opposed to a weapon) and the government acknowledges that it has never made any such argument here. Oral Arg. 36:00–36:16. Indeed, is it our general rule that a party who has knowingly and voluntarily requested the district court to take a particular action cannot later complain to us that the action it requested was an abuse of discretion. See ; . created a seemingly unprecedented exception to this “invited-error doctrine,” in which we must reverse an action that both parties requested the district court take as an abuse of that court‘s discretion.

Should the government choose to forge ahead with this case, the district court may need to address whether today‘s holding affects the admissibility of Buster‘s on-the-scene statement about “that gun and my knife.” Because of its resolution of the threshold questions, the district court did not consider whether that statement would be admissible if the gun itself were suppressed, and the parties have not briefed that issue on appeal. “[M]indful that we are a court of review, not of first view,” , we leave any such questions to the court of first instance. This argument for limiting conditional pleas was also considered and rejected by the advisory committee. It acknowledged that applying harmless error analysis absent a full trial record would be difficult. Fed. R. Crim. P. 11 advisory committee note to 1983 Amendment. But it concluded that this cost did not justify barring conditional pleas.

One might even argue that the motion to suppress the gun and ammunition is not case-dispositive. The government might try to prove its case based on Buster‘s confession that he possessed “that gun“—a statement he did not seek to suppress as part of his Fourth Amendment motion. But the government could not introduce the gun itself or any evidence obtained from it, so there would be no way to prove that the referenced “gun” met the legal definition of a firearm or that it had moved in interstate commerce, essential elements under § 922(g).

The court also considered two other pretrial motions preserved by Bundy—who pleaded guilty to possession of an unregistered firearm. . The first was a motion to dismiss because the gun-registration law was unconstitutional, which the court observed would mandate dismissal as a matter of law. The second was a motion to suppress all evidence seized from a search of Bundy‘s house, including the firearm at issue. Notably, the court did not decide that this motion was automatically case-dispositive merely because it was a suppression motion. Rather, it held that the motion was case-dispositive only after determining that “a favorable ruling on these issues would require suppression of essential evidence and would likely lead to dismissal of the firearms charge in Count Two.” (emphasis added).

Even applying that first step will be daunting, since the majority opinion makes clear that “case-dispositive” no longer means “dispositive of the case,” yet what it does mean remains elusive.

Until today, those parties could at least rely on some of our unpublished decisions for guidance. See, e.g., . But the majority expressly disavows those decisions.

Because Buster‘s attempt to preserve the non-dispositive Miranda issue voids his entire plea, I would not reach the merits of the case-dispositive Fourth Amendment issue. Yet I note that our circuit has held that a search of a suspect‘s bag much like the one here was a valid frisk. See . I also emphasize the majority‘s comment that the government might have succeeded if it sought to justify the search on another theory; the government should now be on notice that it needs to put forth all the alternative theories that justify a search.

Read the detailed case summary