United States v. Garcia-Sierra

United States Court of Appeals For the First Circuit No. 16-2503 UNITED STATES OF AMERICA, Appellee, v. ADOLFO LEÓN GARCÍA-SIERRA, Defendant, Appellant. APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF PUERTO RICO [Hon. Francisco A. Besosa, U.S. District Judge] Before Howard, Chief Judge, Kayatta, Circuit Judge, and Torresen, District Judge.* Linda A. Backiel for Appellant. Kaitlin E. Paulson, Attorney, Appellate Section, Criminal Division, U.S. Department of Justice, with whom John P. Cronan, Acting Assistant Attorney General, Rosa Emilia RodríguezVélez, United States Attorney, and Mariana E. Bauza Almonte, Assistant United States Attorney, were on brief, for Appellee. April 7, 2021 * Of the District of Maine, sitting by designation. HOWARD, Chief Judge. Adolfo León García-Sierra ("García") was convicted by a jury of conspiring to commit drug trafficking offenses and was sentenced to a term of imprisonment of 224 months and 8 days. He appeals both from his convictions and from his sentence, arguing that the erroneous admission of certain evidence tainted the verdicts and that his sentence is unreasonable. After determining that there was trial error but that it was harmless, we affirm the convictions. We do, however, remand for resentencing. I. BACKGROUND The indictment alleged a drug importation operation involving the shipment of large amounts of cocaine from South America to Puerto Rico between August 2012 and June 2014. García was charged with conspiracy to possess with intent to distribute a controlled substance in violation of 21 U.S.C. § 846 and conspiracy to import narcotics into the United States in violation of 21 U.S.C. § 963 . By way of background to García's legal challenges, we briefly sketch the central evidence presented during García's trial -- employing a "balanced-presentation" approach as we did in United States v. Rodríguez-Soler, 773 F.3d 289, 290 (1st Cir. 2014) -- as well as the sentence he received thereafter. We leave further elaboration of the facts to our discussion of each of García's claims. - 2 - A. The December 2012 Rescue at Sea The first witness at trial was Julio Ruiz, a member of the Coast Guard stationed in San Juan, Puerto Rico. Ruiz testified that in December 2012, he responded to a report made by a shipping vessel that had observed a small boat adrift about fifteen miles offshore with two persons aboard, calling for help. Ruiz located the small boat and rescued its two occupants, one of whom was García. Ruiz testified that this search-and-rescue stood out to him because the persons on board lacked documents, and the stories they told "seemed odd." They had explained that they had been on a larger ship that had sunk very quickly and that they had escaped from the sinking vessel onto the small boat on which they were found. This story seemed odd to Ruiz because the Coast Guard had received no report of a large ship in distress, because large ships do not sink quickly, and because when they do sink, they leave an oil sheen on the water, of which there was no trace. After completing the rescue, the Coast Guard transferred the two persons to the Customs and Border Protection agency. B. Testimony by Agent De Jesús This curious story was followed by the testimony of FBI Agent Juan De Jesús. De Jesús testified that around November 2012 he had been investigating an organization which, according to an unnamed source, was in the business of smuggling cocaine from South America to Puerto Rico by sea. In December of that year, De Jesús - 3 - "was advised that two persons matching the modus operandi of the organization had been rescued out at sea." De Jesús eventually interviewed each of these two persons, one of whom was García. De Jesús testified that during his interview with García, García explained to him that, prior to their rescue by the Coast Guard, he and his companion had been on a shipping vessel en route to the Dominican Republic for the purpose of transporting gasoil to Venezuela. De Jesús testified that García's shipwreck companion provided him with a similar but slightly different story: that he and García had been on a shipping vessel en route to the Dominican Republic for the purpose of transporting large amounts of cash. De Jesús proceeded to describe how García's companion eventually agreed to serve as an informant and began supplying De Jesús with ongoing updates about cocaine trafficking activity between South America and the Caribbean. De Jesús testified how the informant's assistance ultimately enabled law enforcement to seize a large shipment of cocaine upon its arrival in Puerto Rico's Guayanilla Bay in October 2013. C. The October 2013 Seizure The testimony by Agent De Jesús was followed by testimony from seven different witnesses all concerning the details of the law enforcement operation that resulted in the seizure of the cocaine shipment in Guayanilla Bay in October 2013. This evidence - 4 - is both voluminous and uncontested on appeal. It suffices for our purposes to note that the government amply demonstrated that its agents seized over 200 kilograms of bundled cocaine after members of the surveilled organization had transferred the bundles from a ship to a small stash house on shore. D. The Informant's Account Next to testify was García's shipwreck companion turned government informant. He testified that he began working as a boat captain for a cocaine trafficking organization in 2012. He described the organization's basic trafficking method: using small boats to carry bundled cocaine bricks out to a larger cargo ship waiting some twenty miles offshore, and then doing the same in reverse once the larger ship approached the target destination. The witness described several specific trafficking voyages in which he took part. He identified García as having accompanied him on one trip in late 2012. García's role was to inspect the cocaine upon arrival to ensure the cargo made it through the journey intact. According to the witness's account, he and García used a small boat to transport the cocaine from the Venezuelan coast to a large cargo ship waiting offshore. They sailed on the larger ship until they neared the Puerto Rican coast, at which time they offloaded the cocaine onto a small boat and set off in the small boat for shore. Shortly thereafter, the small boat began to take on water. The pair threw the cocaine overboard - 5 - and managed to stay afloat on the sinking vessel. They eventually caught the attention of a fishing vessel, which reported their plight to the Coast Guard, leading to the duo's rescue. The witness testified that he was subsequently questioned by Agent De Jesús and eventually agreed to serve as an informant. Finally, the witness testified that García was involved with the shipment to Guayanilla Bay in October 2013. According to the witness, although García did not accompany him on the actual voyage, García helped to organize and load the cocaine at the point of origin. When the shipment arrived in Puerto Rico, the witness informed Agent De Jesús of its location, enabling its seizure by law enforcement. E. The May 2011 Seizure The final pieces of evidence presented by the government brought the jury back in time to a smuggling incident from May 2011, before the start of the conspiracy for which García was on trial. The government had moved in limine indicating its intent to introduce this evidence for the purpose of proving García's knowledge and intent. The May 2011 activity was part of the basis of separate conspiracy charges against García which were pending resolution in other proceedings. García argued that the evidence was inadmissible. The district court did not rule on its admissibility in advance of trial but ultimately permitted the government to present the evidence to the jury. - 6 - The government employed three witnesses to do so. Two officers testified about their dramatic interception of a large shipment of cocaine which had arrived in Puerto Rico in May 2011. The third witness, a government informant (different from the one who had been rescued with García in 2012),1 testified that in May 2011 he and García were involved with a large shipment of cocaine to Puerto Rico. This witness also authenticated a recording of a phone call that he described as a conversation between himself and García about the sale of six kilograms of cocaine. F. Conviction, Sentencing, Appeal The jury convicted García, and a Presentence Investigation Report (PSR) was prepared, recommending a sentence within the range of 235 to 293 months of imprisonment. The district court imposed a sentence of approximately 224 months, reflecting a sentence at the lowest end of the recommended range with a reduction for the time that García had spent incarcerated in Colombia prior to his extradition to the United States. García appealed, challenging the introduction of "overview" testimony from several of the government's witnesses, the admission of evidence regarding the cocaine seizure in May 2011, and the reasonableness of his sentence. 1 Hereafter, we refer to this witness as "the 2011 informant" when necessary to distinguish from the one who had been rescued with García in 2012. References hereafter simply to "the informant" refer to García's shipwreck companion. - 7 - II. EVIDENTIARY CHALLENGES We tackle the evidentiary challenges first. We conclude that the trial court erred in permitting the government to solicit "overview" testimony from Agent De Jesús and in allowing it to present evidence relating to the May 2011 cocaine seizure. But we find both errors harmless and affirm García's convictions. A. Standard of Review Where an appellant objected to a district court's evidentiary ruling at trial, we review the district court's decision for abuse of discretion. United States v. Watson, 695 F.3d 159, 162 (1st Cir. 2012). If our review shows that the trial court acted outside the bounds of its discretion, then we vacate the conviction unless the error was harmless. United States v. Brown, 669 F.3d 10, 24 (1st Cir. 2012); Fed. R. Crim. P. 52(a). Where the appellant did not voice the same objection at trial, we review only for plain error. Watson, 695 F.3d at 162 . If we find that the trial court plainly erred -- in other words, erred in a way which is "clear or obvious" -- then we vacate the conviction if the appellant persuades us both that the error "affected his substantial rights" and that the error "seriously affected the fairness, integrity, or public reputation of the judicial proceedings." Id. at 163 ; see also Rosales-Mireles v. United States, 138 S. Ct. 1897, 1904-05 (2018). - 8 - B. Overview Testimony García argues that certain statements made by Agent De Jesús early in the trial constituted inadmissible "overview" testimony. See United States v. Meises, 645 F.3d 5, 13 (1st Cir. 2011); United States v. Flores-De-Jesús, 569 F.3d 8, 17 (1st Cir. 2009); United States v. Casas, 356 F.3d 104, 118-19 (1st Cir. 2004).2 Because García failed to make this objection at trial, we review for plain error. See United States v. Laureano-Pérez, 797 F.3d 45, 66-67 (1st Cir. 2015); United States v. Valdivia, 680 F.3d 33 , 47 n.10 (1st Cir. 2012). Testimony by a law enforcement agent constitutes impermissible "overview" testimony when it effectively opines that a defendant is guilty "based on the totality of information gathered" in the agent's investigation, rather than relaying the agent's first-hand experiences and observations. Meises, 645 F.3d at 15 (quoting Flores-De-Jesús, 569 F.3d at 19 ). Such opinions are impermissible coming from a lay witness,3 whose "testimony in 2 Following García's lead, we focus on the "overview" statements made by Agent De Jesús, but we acknowledge that García also argues that former officer Gerardo Torres Molino likewise provided impermissible overview testimony. Because Torres's testimony related to the investigation underlying the Rule 404(b) evidence -- which we ultimately hold was admitted in error on other grounds (as discussed below) -- we do not discuss Torres's testimony here. 3 Though not relevant here, such opinions are equally impermissible coming from an expert. See Meises, 645 F.3d at 18 n.20; Casas, 356 F.3d at 120 . - 9 - the form of an opinion is limited to one that is . . . rationally based on the witness's perception." Fed. R. Evid. 701(a); see also Meises, 645 F.3d at 15 ; Valdivia, 680 F.3d at 47 ("Because the witness is, in essence, testifying about the results of a criminal investigation before the government has presented any evidence -- often including aspects of the investigation in which he did not actually participate -- we have repeatedly admonished the use of such testimony."). Such overview testimony from a law enforcement officer remains problematic even when the specific information undergirding the officer's conclusory statements eventually comes into evidence. See Flores-De-Jesús, 569 F.3d at 17 . The problem is two-fold. First, such testimony "effectively usurp[s] the jury's role as fact-finder" by suggesting which inferences the jury should draw from the evidence appropriately before it. Meises, 645 F.3d at 16 . "[H]aving [an agent] so testify amount[s] to simply dressing up argument as evidence." Id. at 17 . Second, to the extent such testimony is a "preview of other witnesses' testimony," it functions as an endorsement by the government of the first-hand witness's account, thereby impermissibly bolstering that witness's credibility. Id.; see also Casas, 356 F.3d at 120 ("Overview testimony by government agents is especially problematic because juries may place greater weight on evidence perceived to have the imprimatur of the government."); Flores-De- - 10 - Jesús, 569 F.3d at 18 ("The overview testimony of a law enforcement official is not simply a repetition (at best) of other evidence. It is also, in effect, an endorsement of the veracity of the testimony that will follow."). Of course, as long as a law enforcement agent's testimony is limited to what he has gleaned from his first-hand observations, there is nothing wrong with the agent "describ[ing] the course of his investigation in order to set the stage for the testimony to come." Flores-De-Jesús, 569 F.3d at 19 ; see also United States v. Rosado-Pérez, 605 F.3d 48, 56 (1st Cir. 2010) ("If a proper foundation is laid, government witnesses may testify about matters within their personal knowledge."). With these principles in place, we consider the testimony which García challenges here. García primarily points to various statements made by Agent De Jesús which mention the existence and methods of "a cocaine trafficking organization" without indicating any first-hand basis for the testimony.4 De 4 Within his "overview" testimony argument, García includes objections to De Jesús's testimony regarding the informant's purported text messages with García. It remains unclear to us how this testimony arguably constitutes impermissible overview. It consists of De Jesús authenticating screen shots he had taken of text messages the informant had shown him on his phone, testifying that the informant had told him that the messages were from a conversation with García, and then reading the messages out loud. Defense counsel preserved a hearsay objection to the testimony's admission, but García provides no developed argumentation on appeal that the testimony was hearsay or the messages improperly authenticated. Even if we were to assume that the district court - 11 - Jesús relayed details of several of this organization's smuggling ventures of which he had no personal knowledge. Much of Agent De Jesús's testimony detailed information De Jesús knew only because his informant had told it to him. De Jesús further testified that he had been "advised that two persons matching the modus operandi of the organization had been rescued out at sea." He proceeded to identify these two persons as his later informant and García, and identified the latter by name, by photograph, and by pointing to García in court. De Jesús testified briefly about having interviewed García sometime after his rescue at sea. Aspects of this testimony raise some of the concerns that have troubled us with previous uses of overview witnesses. See, e.g., Casas, 356 F.3d at 118-19 ; Flores-De-Jesús, 569 F.3d at 23-24, 26-27 . Agent De Jesús testified "well beyond his personal knowledge" when he discussed the existence, methods, and myriad ventures of the smuggling operation. Casas, 356 F.3d at 118 . Much erred in overruling García's objection, we would find the error harmless. The informant later testified about these same text messages and their connection to García. Moreover, De Jesús's testimony about the texts was so garbled it could hardly be said to have helped the government's case: De Jesús mistakenly read the messages out of order, he did not indicate who was the author of each message, and numerous translation issues stilted the reading. And to his botched reading of the messages, De Jesús added nothing, for he neither analyzed nor interpreted them. - 12 - of his "testimony was unquestionably hearsay. It unnecessarily anticipated testimony that [the informant] would give himself. It had the imprimatur problems we have described." Flores-De-Jesús, 569 F.3d at 23 . Most problematically, it identified García as a member of the smuggling organization with whom the informant was in communication.5 This was tantamount to "testif[ying] that . . . the defendant[] was guilty of the conspiracy charged." Id. at 24 (quoting Casas, 356 F.3d at 119 ). But De Jesús stopped short of giving his own conclusions about the role García played in the conspiracy, which we have previously considered especially problematic. See Flores-DeJesús, 569 F.3d at 24 (considering "[t]he most troubling part of Agent Toro's testimony" to be "his conclusions about the roles of the defendants in the conspiracy"); Meises, 645 F.3d at 16, 18 (holding inadmissible agent's testimony expressing "his opinions as to defendants' roles in the conspiracy" because "it was patently unfair for [Agent] Cruz to present his view of appellants' culpability"). Thus, the problem here is limited to the concern that De Jesús's testimony may have been interpreted by the jury as "vouching" for another witness's subsequent testimony. Cf. 5 During its direct examination of Agent De Jesús, the government asked whether "[the informant was] also communicating to other members of the organization," to which De Jesús responded "yes," and then stated that "he also had -- was communicating with Adolfo Léon García-Sierra." - 13 - Meises, 645 F.3d at 17-18 (considering "the problem with Cruz's testimony [to] extend beyond vouching for what the jury may perceive as a less credible witness"). Still, the problem remains that De Jesús, like the agent in Meises, "had no insight to offer the jurors based on personal knowledge of the [defendant's] inculpatory conduct. Like them, he had to rely on [an informant's] account." 645 F.3d at 16 . But we proceed on plain error review, and we cannot say that the problematic aspects of De Jesús's testimony both "affected [García's] substantial rights" and "seriously affected the fairness, integrity, or public reputation of the judicial proceedings." Watson, 695 F.3d at 163 . The factual matter provided by De Jesús's testimony otherwise came into evidence. A member of the Coast Guard had already testified about rescuing García and the other man at sea, and had noted that the story they told to explain their circumstance had struck him as implausible. And the informant eventually testified extensively about his first-hand experiences with the smuggling operation generally, and with García specifically. The remaining potential prejudice is therefore the effect of De Jesús's testimony on the jury's assessment of the informant's credibility. Without a doubt, this witness's testimony was indispensable to the government's case, and it is - 14 - impossible to divine the precise influence De Jesús's preview of the witness's account may have had on the jury's assessment of the witness's credibility. But on plain error review García bears the burden of persuading us that the errors plaguing De Jesús's testimony harmed him, see United States v. Morris, 784 F.3d 870, 874 (1st Cir. 2015), and we are not persuaded that they did. The informant stood up impressively well to cross examination. No holes whatsoever were poked in his story. And that story was consistent with other competent evidence proffered by the government: most importantly, by the Coast Guard officer's testimony and by the testimony regarding the Guayanilla seizure. Consequently, García's challenge to the admittedly problematic overview testimony provided by Agent De Jesús falters on the shoals of plain error review. It provides no basis for relief. C. Rule 404(b) Evidence We turn next to García's second evidentiary challenge, which concerns the admission of prior-bad-acts evidence under Federal Rule of Evidence 404(b). Before turning to the specifics of García's challenge, we overview the basic legal landscape. Rule 404(b) concerns the admissibility of evidence of "crimes, wrongs, or acts" other than those for which a defendant is on trial. Rule 404(b)(1) states that "[e]vidence of any other crime, wrong, or act is not admissible to prove a person’s character in order to show that on a particular occasion the person - 15 - acted in accordance with the character." Thus, Rule 404(b)(1) prohibits a particular inference one might draw from such evidence: it "forbid[s] the prosecution from asking the jury to infer from the fact that the defendant has committed a bad act in the past, that he has a bad character and therefore is more likely to have committed the bad act now charged." United States v. Moccia, 681 F.2d 61, 63 (1st Cir. 1982) (Breyer, J.). We refer to this inference as the forbidden "propensity inference." See, e.g., United States v. Varoudakis, 233 F.3d 113, 122 (1st Cir. 2000). But Rule 404(b) does not render inadmissible prior-bad-acts evidence for any other purpose, such as to prove "motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident." Fed. R. Evid. 404(b)(2). A trial court faced with a proffer of prior-bad-acts evidence "must engage in a two-step analysis" to determine whether the evidence should be admitted. United States v. Tkhilaishvili, 926 F.3d 1, 15 (1st Cir. 2019); see also United States v. MartínezMercado, 919 F.3d 91, 101 (1st Cir. 2019). First, the trial court must determine whether the evidence has a "special relevance" to an issue in the case. Tkhilaishvili, 926 F.3d at 15 (quoting Veranda Beach Club Ltd. P'ship v. W. Sur. Co., 936 F.2d 1364 , 1373 (1st Cir. 1991)); see also United States v. Sabean, 885 F.3d 27, 35 (1st Cir. 2018). "Special relevance" is a bit of a misnomer, for what step one requires is that the evidence be relevant "for - 16 - any purpose apart from showing propensity to commit a crime." United States v. Habibi, 783 F.3d 1, 2 (1st Cir. 2015) (quoting United States v. Doe, 741 F.3d 217, 229 (1st Cir. 2013)); see also Tkhilaishvili, 926 F.3d at 15 (explaining that step one of the analysis is satisfied if the evidence is relevant for a purpose other than "to show a defendant's evil inclination" (quoting Veranda Beach, 936 F.2d at 1373)). If the prior-bad-acts evidence is relevant only for the forbidden propensity inference, then the evidence is inadmissible under Rule 404(b)(1) and the inquiry ends. See, e.g., Martínez-Mercado, 919 F.3d at 101-03 . Otherwise the trial court advances to step two, the application of Federal Rule of Evidence 403. See, e.g., Habibi, 783 F.3d at 4 . Under Rule 403, the trial court may exclude the prior-bad-acts evidence if it determines in its discretion that the probative value of the evidence is substantially outweighed by any unfair prejudice. See Tkhilaishvili, 926 F.3d at 15 ; United States v. Hicks, 575 F.3d 130, 142 (1st Cir. 2009). With that prelude, we return to García's specific challenge. García contends that the district court failed to properly conduct each step of the Rule 404(b) analysis. As to the first step, he argues that the district court erred by failing to require the government to specify the permissible purpose for which the prior-bad-acts evidence was admissible, and he further argues that there was no such permissible purpose. García argues that - 17 - error at the second step flows from this initial error, for having failed to specify any permissible purpose for which the prior-badacts evidence was admissible, the district court was unable to weigh the probative value of that evidence against any unfair prejudice it could cause. Finally, García argues that the probative value of the Rule 404(b) evidence was indeed substantially outweighed by its unfairly prejudicial effect. As we set out to explain, we disagree with García's contention that the district court failed to conduct the two steps of the Rule 404(b) analysis, but we agree that the outcome of that analysis should have been the exclusion of the Rule 404(b) evidence under Rule 403. Determining that this error was also harmless, however, we affirm García's convictions. 1. Step One As García persistently objected to the admission of the prior-bad-acts evidence at trial, we review the district court's "ruling that [this] evidence was admitted consistent with [Rule] 404(b) . . . for abuse of discretion." United States v. Gemma, 818 F.3d 23, 35 (1st Cir. 2016); see also United States v. Moon, 802 F.3d 135, 144 (1st Cir. 2015). As we set out to explain, we detect no abuse of discretion at step one. García contends that the district court procedurally erred by failing to specify the particular purpose for which it deemed the Rule 404(b) evidence admissible. Cf. United States v. - 18 - Arias-Montoya, 967 F.2d 708, 713 (1st Cir. 1992) (stating that prior-bad-acts evidence "should not be accepted unless the government articulates with suitable precision the 'special' ground for doing so" (quoting United States v. García–Rosa, 876 F.2d 209 , 221 (1st Cir. 1989), vacated on other grounds sub nom. Rivera-Feliciano v. United States, 498 U.S. 954 (1990))). We disagree. García never requested that the district court clarify the permissible purpose for which the Rule 404(b) evidence was admissible, and the record before us indicates that it was clear that the government offered this evidence to prove García's knowledge and intent.6 While the judge never explicitly stated The government's notice of intent stated that the type of — 6 — prior-bad-acts evidence which it would seek to admit "tend[s] to prove that a required state of mind existed at the time required by the instant accusations, establishing the existence of significant probative value." The notice further pointed out that "[t]he instant offenses require both that possession be knowing and with intent to distribute." The defense can hardly complain that this notice failed to communicate the purpose for which the government purported to offer the prior-bad-acts evidence, as it stated in its responsive motion that "[t]he United States is evidently offering this evidence to prove 'knowledge and intent.'" Moreover, when the court asked the government at trial to clarify "the issue here that makes this 404(b) testimony relevant," the government replied, "I think there is an issue of known possession and intent -- or an issue of lack of mistake. And, hence, we presume brother counsel is going to say, 'Oh, they were rescued. By [sic] my guy didn't know anything about these drugs or anything that was happening.'" And in a later colloquy between the judge and defense counsel over the evidence's admissibility, the defense acknowledged that it understood that the evidence was being offered to prove "[k]nowledge or intent." - 19 - the specific purpose for which he considered the prior-bad-acts evidence admissible, this alone is not error. See United States v. Donovan, 984 F.2d 507, 511 (1st Cir.) (stating that "explicit findings under Rule 404(b) . . . are not an invariable prerequisite to the admission of Rule 404(b) evidence"), reheard on other grounds, United States v. Aversa, 984 F.2d 493 (1st Cir. 1993). Because the government had made clear the purposes for which it was offering the prior-bad-acts evidence, the trial court acted within its discretion by not expressly stating the permissible relevance for which it deemed the evidence admissible. As to whether the prior-bad-acts evidence was in fact relevant to García's knowledge or intent, we agree with the government that it was arguably relevant to García's knowing participation in the cocaine smuggling operation.7 Through the 7 We set intent aside. The government has failed to provide us with any case-specific explanation for how the prior-bad-acts evidence offered here was relevant to García's intent. Cf. United States v. Henry, 848 F.3d 1, 9 (1st Cir. 2017) (emphasizing that district courts must "carefully consider the proponent's assertion of why a prior [bad act] has special relevance and examine whether, in the particular case-specific circumstances, the proponent is simply attempting to disguise propensity evidence by artificially affixing it with the label of a permitted Rule 404(b)(2) purpose"); see also Martínez-Mercado, 919 F.3d at 102 (cautioning that "the relevance of a prior conviction admitted to prove 'intent' . . . may rest on little more than propensity" (quoting Henry, 848 F.3d at 15 (Kayatta, J., concurring))); United States v. Lynn, 856 F.2d 430, 436 (1st Cir. 1988) (noting that, absent the forbidden propensity inference, "the probative worth of [the defendant's] conviction toward proving his intent to commit the instant offense is difficult to conceptualize"). - 20 - prior-bad-acts evidence, the government sought to show that García had previously smuggled cocaine from South America to Puerto Rico by sea. If credited, this evidence would tend to decrease the likelihood that García was ignorant of the illicit purpose of the sea voyage on which he had embarked in December 2012, an argument the government may have fairly anticipated given that García had told Agent De Jesús that the purpose of the ill-fated trip had been to transport gasoil. Cf. United States v. Robles-Alvarez, 874 F.3d 46, 51 (1st Cir. 2017) (considering evidence of prior drug smuggling trip relevant where defense might have argued "that the appellant's mere presence with [a co-conspirator] on the voyages was not sufficient to support a conviction"). García complains that he never actually defended against the charges based on a lack of knowledge. But knowledge was an element of the crimes charged that the government had to prove, and nowhere did García "express a clear and unequivocal intention to remove" the issue of knowledge from the trial. United States v. Garcia, 983 F.2d 1160, 1174 (1st Cir. 1993) (emphasis omitted). Absent an "offer to stipulate" or its practical equivalent, evidence of García's knowledge was relevant to the case and not barred by Rule 404(b). Id.; see also Henry, 848 F.3d at 9 ("A defendant's failure to argue lack of knowledge . . . does not 'remove th[at] issue[] from the case.'" (quoting United States v. Pelletier, 666 F.3d 1, 6 (1st Cir. 2011))). - 21 - In short, the district court permissibly considered the prior-bad-acts evidence "specially" relevant for proving García's knowledge. See, e.g., United States v. Lopez-Cotto, 884 F.3d 1, 13-14 (1st Cir. 2018). There was no abuse of discretion at step one. 2. Step Two Though it was acceptable for the district court to consider the prior-bad-acts evidence proffered by the government probative as to knowledge and thus not barred by Rule 404(b)(1), the court was nonetheless obliged to consider whether the admission of this evidence would unfairly prejudice García and to exclude the evidence if its probative value was substantially outweighed by its prejudicial effect. See Fed. R. Evid. 403; Tkhilaishvili, 926 F.3d at 15 . Because García objected to the admission of the prior-bad-acts evidence at trial under Rule 403, we review the district court's admission of the prior-bad-acts evidence notwithstanding that objection for abuse of discretion. See Gemma, 818 F.3d at 35 . At the outset, we reject García's contention that the district court failed to conduct Rule 403 balancing at all. Although the court did not do so explicitly, the record indicates that the court performed the requisite balancing implicitly. See United States v. Breton, 740 F.3d 1, 15 (1st Cir. 2014) ("[T]he absence of an express Rule 403 finding . . . does not mean the - 22 - district judge failed to perform this analysis."). Prior to the introduction of the evidence on the third day of trial, there was an extensive discussion on its admissibility. It is clear from this discussion that the court understood García's objection to the evidence to be rooted in Rule 403. Moreover, the court's questions to counsel indicate that it considered both the prejudice and the probative value of this evidence. It is also evident from the court's subsequent questions that it continued to consider the value and prejudice of the evidence as it was presented. Nonetheless, we disagree with the outcome of the court's implicit balancing. The prior-bad-acts evidence had marginal permissible relevance; it was lengthy, confusing, and unaccompanied by sufficient guidance from the court; and (as is often the case with prior-bad-acts evidence) the potential for prejudice was fairly obvious. To explain, we first review the prior-bad-acts evidence which the government presented. We then describe its limited permissible probative value, how that value was overshadowed by its potential for prejudice and confusion, and how this prejudice and confusion was not mitigated by the instructions provided to the jury. The government used three witnesses to present the prior-bad-acts evidence. First, the two officers described their surveillance of a drug trafficking organization's members and activities. They testified in detail about an incident in which - 23 - a cargo van that they had been surveilling fled inspection at a ferry terminal. The high-speed chase which ensued ended only when the fleeing van lost control and crashed in a field. The officers testified that they recovered from the van 525 kilograms of cocaine wrapped in black plastic bags. Photographs of the van, the ferry, and the cocaine were all introduced into evidence. The officers' testimony did not connect the May 2011 cocaine seizure to García, whom the officers never mentioned. The government instead attempted to make this connection through the 2011 informant. This witness testified that he knew García personally. He further testified that around May 2011 there was a "negotiation" regarding 525 kilograms of cocaine with which García was involved. The witness stated that García helped to coordinate the transportation of this cocaine to Puerto Rico. A recording was then played aloud of a phone conversation which the witness described as between himself and García. The conversation concerned the sale of six kilograms of cocaine. The Rule 404(b) presentation concluded with defense counsel cross-examining the witness about his own criminal activity and plea agreements. Federal Rule of Evidence 403 allows a court to "exclude relevant evidence if its probative value is substantially outweighed by a danger of," among other things, "unfair prejudice, confusing the issues, [or] misleading the jury." "Unfair prejudice 'speaks to the capacity of some concededly relevant evidence to - 24 - lure the factfinder into declaring guilt on a ground different from proof specific to the offense charged.'" United States v. DiRosa, 761 F.3d 144, 153 (1st Cir. 2014) (quoting Old Chief v. United States, 519 U.S. 172, 180 (1997)). As discussed above, evidence that García had been previously involved in maritime cocaine smuggling was relevant to García's knowledge that the voyage he had embarked on in December 2012 had as its true purpose the illicit smuggling of cocaine, not the innocent transportation of gasoil. But the government introduced ample evidence establishing this knowledge apart from the prior-bad-acts presentation -- namely, the informant's extensive testimony about both his conversations with García concerning the smuggling operation and his observations of García participating in the smuggling operation, as well as photographs of the informant's text messages with García pertaining to the failed December 2012 voyage. This independent evidence of García's knowledge undermined the marginal value of the prior-bad-acts evidence to the government's case. See Varoudakis, 233 F.3d at 123 (finding that prior-bad-acts evidence ought to have been excluded under Rule 403 where "the government did not need the [prior-bad-acts] evidence to prove [the defendant's] knowledge"); cf. Lynn, 856 F.2d at 436 (questioning the probative value of prior-bad-acts evidence offered to show intent where "the government would have succeeded in proving intent should the jury - 25 - believe the testimony of [other witnesses], rendering negligible their [sic] need to show intent by the prior bad acts"). And while the probative value of the prior-bad-acts evidence was thus relatively low, the prejudice worked by its admission was comparatively high due to the danger it presented of confusing the jury and luring it into forbidden propensity reasoning. The Rule 404(b) evidence related to a separate cocainesmuggling operation not straightforwardly connected to García. Consequently, its introduction resulted in a "mini-trial" through which the government sought to establish that García really was implicated in the May 2011 cocaine seizure. See United States v. Gilbert, 229 F.3d 15, 24 (1st Cir. 2000) (rejecting government's challenge to the exclusion of prior-bad-acts evidence in part because district court's "concerns about the extent to which [whether the prior bad act had even occurred] would have to be litigated during the course of trial" were warranted). The priorbad-acts evidence presented in this case spanned three witnesses and two days of trial, it sparked the introduction of photographs and the playback of a recorded phone call, and it required considerable detective work by the jury to draw from the disjointed pieces of this colorful presentation the conclusion that García had been involved with the May 2011 cocaine shipment. Moreover, to avoid improper use of this evidence, the jury would have had to close its mind to the all-too tempting - 26 - inference that because García had been involved with a cocaine smuggling conspiracy in 2011 he was more likely to have been involved with a cocaine smuggling conspiracy in 2012-2014. See Varoudakis, 233 F.3d at 123 ("[T]he more the prior bad act resembles the crime, the more likely it is that the jury will" fall into forbidden propensity reasoning); Lynn, 856 F.2d at 436 ("The ordinary inference here would seem very close to the inference the Rule was designed to avoid."). Instead, the jury would have had to infer from all the evidence presented of the May 2011 cocaine seizure only that García knew that he was not out transporting gasoil when he was rescued in December 2012. In the circumstances of this case, the permitted use of the Rule 404(b) evidence was much less natural and intuitive than the forbidden propensity use, adding to the danger of unfair prejudice. Sometimes careful limiting instructions can cure the prejudice that would otherwise render inappropriate the introduction of prior-bad-acts evidence. See, e.g., United States v. Manning, 79 F.3d 212, 217 (1st Cir. 1996) (stating that "[t]he district court minimized any prejudicial impact of the prior drug dealing evidence by instructing the jury, contemporaneously and again in its final instructions, about the proper use of prior bad act evidence"). But for limiting instructions to be "suitably prophylactic" in the Rule 404(b) context, they must guide the - 27 - jury's attention away from the forbidden propensity inference by clearly directing it toward the specific permissible relevance that the prior-bad-acts evidence has to the case. Sabean, 885 F.3d at 38 , 35 (quoting United States v. Mehanna, 735 F.3d 32, 64 (1st Cir. 2013)) (considering sufficiently curative court's instructions to the jury "that the government was offering the [Rule 404(b)] testimony "as evidence of what the Government says is the defendant's motive to commit the [crimes charged]" and that the evidence was to be considered only for this specified, limited purpose); United States v. Newsom, 452 F.3d 593, 606 (6th Cir. 2006) (urging district courts to include in jury instructions "the specific factor named in the rule that is relied upon to justify admission of the [prior bad] acts evidence" (quoting United States v. Johnson, 27 F.3d 1186, 1194 (6th Cir.1994))); see also Pattern Criminal Jury Instructions for the District Courts of the First Circuit, § 2.06, cmt. 3 (updated July 28, 2014) ("Courts should encourage counsel to specify and limit the purpose or purposes for which prior act evidence is admitted . . . . Instructions for purposes other than that for which the specific evidence was admitted should not be given."). Here, the trial court twice cautioned the jury about the limited proper use of the prior-bad-acts evidence. Before the 2011 informant testified, the court told the jury that the priorbad-acts evidence was not to be used to infer García's propensity - 28 - for criminal behavior, but was only to be used "to show that he may have had a motive, an opportunity, an intent, or to prove preparation, plan, knowledge, identity, absence of mistake or lack of accident in what he did." In giving the jury its final charge, the court provided a similarly all-encompassing instruction: You have heard evidence that the Defendant previously committed acts similar to those charged in this case. You may not use this evidence to infer that, because of his character, the Defendant carried out the acts charged in this case. You may, however, consider this evidence only for the limited purpose of deciding whether the Defendant had the state of mind or intent necessary to commit the crimes charged in the indictment, or whether the Defendant had a motive or the opportunity to commit the acts charged in the indictment, or whether the Defendant acted according to a plan or in preparation for commission of a crime, or whether the Defendant committed the acts he is on trial for by accident or mistake. This latter instruction was specifically requested by defense counsel, and the omnibus nature of both instructions tracks the explanations of the permissible relevance of the prior-bad-acts evidence provided several times to the court by the government at sidebar. Given that both parties promoted these instructions, it is certainly understandable that the district court provided them. Nonetheless, the overinclusive nature of the instructions prevented them from focusing the jury's attention on the one permissible use of the prior-bad-acts evidence in this case: to - 29 - prove that Garcia likely would not have confused a ship transporting gasoil with a ship engaged in smuggling a large shipment of cocaine. Without case-specific guidance on how to otherwise use the prior-bad-acts evidence, it was all too likely that the jury would engage in the forbidden but intuitive propensity reasoning. See Varoudakis, 233 F.3d at 125 (explaining that the propensity inference "is not rejected because character is irrelevant; on the contrary, it is said to weigh too much with the jury and to so overpersuade them as to prejudge one with a bad general record and deny him a fair opportunity to defend against a particular charge" (quoting Michelson v. United States, 335 U.S. 469 , 475–76 (1948))). Neither instruction effectively trained the jury's attention on the narrow, permissible use of the prior-badacts evidence in this case (to establish knowledge), and so we are doubtful that the instructions prevented misuse of the prior-badacts evidence or dispelled confusion about that evidence's proper relevance to the crimes charged. Consequently, we cannot consider these instructions "suitably prophylactic," Sabean, 885 F.3d at 38 (quoting Mehanna, 735 F.3d at 64 ); they did not effectively mitigate the prejudice posed by the prior-bad-acts evidence admitted in this case.8 To be clear, the misstep we identify is not the provision — 8 — of the instructions, in which both parties had a hand, but the admission of the overly prejudicial, minimally valuable prior-badacts evidence, whose prejudice the instructions did not cure. - 30 - In sum, "[t]he propensity danger of the [prior-bad-acts] evidence was unmistakable" here, and substantially outweighed its limited probative value. Varoudakis, 233 F.3d at 124 . The district court exceeded the bounds of its discretion when it implicitly determined otherwise. 3. Harm We turn to whether the erroneous admission of the Rule 404(b) evidence in this case prejudiced García. "An error will be treated as harmless only if it is 'highly probable' that the error did not contribute to the verdict." United States v. Kilmartin, 944 F.3d 315, 338 (1st Cir. 2019) (quoting United States v. Fulmer, 108 F.3d 1486, 1498 (1st Cir. 1997)). To analyze whether an error was harmless we must divine from "the record as a whole . . . the probable impact of the improper evidence upon the jury." Fulmer, 108 F.3d at 1498 (quoting United States v. Melvin, 27 F.3d 703, 708 (1st Cir. 1994)). In doing so, we consider factors such as "the centrality of the tainted material, its uniqueness, its prejudicial impact, the uses to which it was put during the trial, [and] the relative strengths of the parties' cases." Kilmartin, 944 F.3d at 338 (quoting United States v. Piper, 298 F.3d 47, 57 (1st Cir. 2002)). In a criminal case, the crucial factor is typically "the strength or weakness of the government's evidence of guilt" less the improperly admitted evidence. Kilmartin, 944 F.3d at 338 . - 31 - The admission of the prior-bad-acts evidence in this case, though error, was harmless. The government's case against García was strong. García was found on a small sinking boat off the coast of Puerto Rico without any plausible explanation for how he had gotten there, and his boat-mate provided a detailed, firsthand narrative that explained the occurrence as one of a series of cocaine smuggling ventures in which García had participated. His mate's testimony was corroborated in other respects by records of his text messages (which the mate described as a discussion between himself and García about García's having met with one of the smuggling organization's leaders to discuss the failed 2012 operation), and by the successful cocaine seizure at Guayanilla, and it was in all respects uncontradicted. We therefore consider it "highly probable" that the admission of the evidence pertaining to the May 2011 cocaine seizure was not a determinative factor in the jury's guilty verdict. See id. D. Cumulative Error Finally, García argues that his convictions should be vacated based on the collective impact of the evidentiary missteps in his trial. We accept that the cumulative prejudicial effect of independently innocuous trial errors may warrant a new trial. See, e.g., United States v. Peña-Santo, 809 F.3d 686, 702 (1st Cir. 2015) ("[I]ndividual errors, insufficient in themselves to necessitate a new trial, may in the aggregate have a more - 32 - debilitating effect." (quoting United States v. Sepúlveda, 15 F.3d 1161 , 1195–96 (1st Cir. 1993))). But here, for the same reasons that we find each of the evidentiary errors which we have identified harmless, we find them collectively so as well. The convictions are therefore affirmed. III. SENTENCING CHALLENGE García argues that his sentence is unreasonable on two grounds. First, he argues that the district court erred by imposing a supervisory role enhancement without identifying evidence that García played a supervisory role in the conspiracy. Second, García argues that there is an unwarranted disparity between his sentence and those of his codefendants. Though we find the second claim to lack merit, we agree with García that the record does not support the imposition of the supervisory role enhancement. A. Standard of review Since García also raised before the district court the challenges to his sentence based upon both the supervisory role enhancement and the purportedly unwarranted sentence-length disparity, his claims are preserved on appeal. García fashions these claims as both procedural and substantive challenges to his sentence. When considering a preserved claim that a sentence is the result of procedural error, we review the district court's - 33 - "interpretations and applications of the [sentencing] guidelines" de novo, its judgment-calls for abuse of discretion, and its factual findings for clear error. United States v. FloresQuinoñes, 985 F.3d 128, 133 (1st Cir. 2021) (quoting United States v. Reyes-Torres, 979 F.3d 1, 7 (1st Cir. 2020)); see also United States v. Reyes-Santiago, 804 F.3d 453, 468 (1st Cir. 2015). When faced with a preserved claim that a sentence is substantively unreasonable, we review for abuse of discretion. Flores-Quiñones, 985 F.3d at 133 . Under this deferential standard, we will affirm a sentence as reasonable so long as the sentencing court's rationale is "plausible" and the sentence is "defensible." United States v. Gierbolini-Rivera, 900 F.3d 7, 12 (1st Cir. 2018) (quoting United States v. Martin, 520 F.3d 87, 96 (1st Cir. 2008)); see also Reyes-Santiago, 804 F.3d at 468 . B. Supervisory role enhancement As recommended in the PSR, the district court imposed a two-level enhancement to García's base offense level pursuant to §3B1.1(c) of the United States Sentencing Commission Guidelines Manual (Nov. 2016) ("the Guidelines" or "U.S.S.G."), which allows for such an increase "if the defendant was an organizer, leader, manager, or supervisor" in the offense.9 We set out the relevant 9 Section 3B1.1(c) applies to "criminal activity involv[ing] at least two, but fewer than five, complicit individuals." United States v. Ilarraza, 963 F.3d 1, 13 (1st Cir. 2020)(quoting United States v. Al-Rikabi, 606 F.3d 11, 14 (1st Cir. 2010)). Since the - 34 - legal principles first and then address García's particular arguments. The determination of which type of role a defendant played in an offense is a factual one, reversible only if clearly erroneous. See Ilarraza, 963 F.3d at 13 ; see also United States v. Cortez-Vergara, 873 F.3d 390, 393 (1st Cir. 2017). At sentencing, "[t]he government bears the burden of proving that an upward role-in-the-offense adjustment is appropriate in a given case." Al-Rikabi, 606 F.3d at14. To properly impose the upward adjustment, the sentencing court must be satisfied that a preponderance of the evidence supports the government's claim that the defendant acted as an organizer, leader, manager, or supervisor in the commission of the offense. Id.; see also United States v. Medina, 167 F.3d 77, 79 (1st Cir. 1999). A supervisory or managerial role is evidenced by some "manifestation of authority" on the part of the defendant. Savarese, 686 F.3d at 20.10 The authority possessed by the drug-smuggling conspiracy at issue in this case involved nineteen codefendants, it is unclear why the PSR recommended the aggravating role enhancement pursuant to §3B1.1(c) rather than pursuant to §§3B1.1(a) or (b), which are expressly applicable to more extensive criminal activity. But the government has not challenged the sentence, and the use of §3B1.1(c) benefits rather than harms the defendant. Therefore, we analyze whether García qualified for an enhancement under §3B1.1(c) without regard for the extensivity of the underlying criminal activity. See United States v. Savarese, 686 F.3d 1 , 21 n.16 (1st Cir. 2012). 10The Guidelines generally distinguish "leadership [or] organizational role[s] from one[s] of mere management or - 35 - defendant may be fairly minimal; "a defendant need not be at the top of a criminal scheme to be a manager or supervisor." United States v. Goldberg, 105 F.3d 770, 777 (1st Cir. 1997). "[W]e have even held that, in some circumstances, the government need only show by a preponderance of the evidence 'that the defendant exercised authority or control over another participant on one occasion.'" Savarese, 686 F.3d at 20 (quoting United States v. García–Morales, 382 F.3d 12, 20 (1st Cir. 2004)); see also United States v. Cruz, 120 F.3d 1, 3-4 (1st Cir. 1997) (en banc). Yet not all roles arguably termed "supervisory" warrant an enhancement under §3B1.1(c). The enhancement is proper only where the defendant exercised some degree of authority or control over another criminal actor; that the defendant may have managed or supervised a particular criminal activity is insufficient. United States v. Prange, 771 F.3d 17, 34 (1st Cir. 2014); see also FloresDe-Jesús, 569 F.3d at 35 . supervision," U.S.S.G. §3B1.1, cmt. n.4, with the latter roles denoting less responsibility, id. at cmt. background. Under §3B1.1(c), however, the distinction lacks a difference, since a defendant found to have played any of these four roles warrants the same two-level enhancement to his base level offense. Accordingly, even though the PSR adopted by the sentencing court found that "the defendant was considered to be a leader in the organization," it was not error to impose the enhancement as long as the defendant could have been found to be merely a manager or supervisor. We therefore focus on whether this lower bar was met. - 36 - With this background in mind, we turn to García's arguments. First, García argues that it was error for the sentencing court to fail to make specific findings to support the upward adjustment based on García's alleged supervisory role. Second, García argues that it was error for the court to find that García had a supervisory role in the smuggling organization, contending that the trial evidence established at most that he was a mere "stevedore, loading and unloading cargo at both ends of a venture." There is some force to the first claim, but our agreement with the second determines the appropriate relief here. "Without reasonably specific findings or some satisfactory surrogate in the record, we are unable to engage in meaningful review to determine whether the decision [to impose a role-in-the-offense enhancement] was clearly erroneous." Medina, 167 F.3d at 80 . Of course, "sentencing judges need not explain their reasoning in exquisite detail, especially when the reasons are 'evident from the record.'" United States v. Zehrung, 714 F.3d 628, 631 (1st Cir. 2013) (quoting United States v. Stella, 591 F.3d 23, 28 (1st Cir. 2009)). "But . . . in the end we must be able to figure out what they 'found and the basis for the findings to the extent necessary to permit effective appellate review.'" Id. at 632 (quoting United States v. Van, 87 F.3d 1, 3 (1st Cir. 1996)). As the Supreme Court has stressed, "[b]y articulating reasons, even if brief, the sentencing judge . . . - 37 - assures reviewing courts (and the public) that the sentencing process is a reasoned process." Rita v. United States, 551 U.S. 338, 357 (2007). Particularly where the underlying facts of a case involve multiple transactions and a web of participants, and where the PSR "does not even minimally focus on the specific considerations necessary" to support a finding that a defendant occupied an aggravating role in the offense, "it is necessary that the district judge make sufficient findings to articulate the rationale" for the aggravating role enhancement. United States v. Catano, 65 F.3d 219, 230 (1st Cir. 1995). Here, as in Catano, "[n]either the PSR nor the sentencing transcript discusses [the defendant's] involvement or identifies why" the defendant was considered a manager or supervisor in the cocaine smuggling conspiracy. Id. At the sentencing hearing, García personally and through counsel argued that García did not have a supervisory role in the smuggling organization. The district court did not directly respond but stated that García's base offense level was increased two levels pursuant to §3B1.1(c) "[b]ecause he was an organizer, leader, manager, or supervisor in the criminal activity," and that "the pre-sentence investigation report satisfactorily reflects the components of Mr. García's offense by considering its nature and circumstances." The PSR, in turn, states in support of the §3B1.1(c) enhancement only that "[a]ccording to trial notes and discovery - 38 - reports, the defendant was considered to be a leader in the organization." This is curious, because elsewhere the PSR has a section entitled "Roles of the Members of the Conspiracy," and though it lists the conspiracy's "leaders," that list does not include García. Instead, García is listed as a "transporter." The only other facts in the PSR pertinent to García's role are found in the section "Acceptance of Responsibility." There, the PSR states that García "noted that he did not give orders to anybody," "acknowledge[d] that he supported [a co-conspirator]," and "noted that he did not supervise anybody." That neither the PSR nor the district court offered any explanation for rejecting García's claim that he was neither a leader nor a supervisor in the organization complicates our review, for a defendant being sentenced "is entitled to reasoned findings, on a preponderance standard, not to an appellate court's assumptions drawn free-form from an inscrutable record." Catano, 65 F.3d at 230 (quoting United States v. McDowell, 918 F.2d 1004 , 1012 n.8 (1st Cir. 1990)). The government argues that the dearth of fact-finding in support of the §3B1.1(c) enhancement does not warrant remand in this case because the basis for the enhancement was "evident from the record," Zehrung, 714 F.3d at 631 (quoting Stella, 591 F.3d at 28 ), obviating the need for the district court to explain its reasoning. We disagree. - 39 - There is no evidence in the record to support the government's claim that García oversaw any of the organization's workers. Rather, the portions of the record cited by the government show only that García packaged and waterproofed the drugs and chatted with others about the operation. The evidence which most strongly supports the government's characterization of García is the informant's testimony that one of the conspiracy's leaders had told him that García "was going to take care of organizing the entire [October 2013] trip." But when asked on direct to clarify what García was organizing, the informant replied "the cocaine, what I was going to be carrying." And the government points to no evidence showing that García ever directed a single person to perform a single task for the conspiracy. See United States v. Fuller, 897 F.2d 1217, 1221 (1st Cir. 1990); United States v. Altagracia Castillo, 145 F. App'x 683, 685 (1st Cir. 2005); Cf. García–Morales, 382 F.3d at 20 . We cannot agree with the government that it was evident from this record that the supervisory role enhancement was warranted. While it appears that García was in charge of certain tasks, like preparing the cocaine for the voyage and inspecting it upon arrival, the government's burden was to show by a preponderance of the evidence that García exercised authority or control over other participants in the smuggling venture. See Prange, 771 F.3d at 34 ; cf. Flores-De-Jesús, 569 F.3d at 35 - 40 - (stating that evidence that defendant "ke[pt] the drug point wellstocked and collect[ed] the proceeds . . . is insufficient to establish the requisite control over another criminal actor that our case law requires"). This the government has not done. Nor could it, there being no evidence in the record that García managed or supervised at least one other person. Thus, the role-in-the-offense finding is clearly erroneous, and resentencing is warranted. See, e.g., Al-Rikabi, 606 F.3d at 16 (vacating sentence and remanding for resentencing where district court's finding that defendant "was an organizer, leader, manager, or supervisor" under U.S.S.G. §3B1.1(c) was clearly erroneous); United States v. Ramos-Paulino, 488 F.3d 459, 464 (1st Cir. 2007) (same). C. Disparities Finally, García argues that his sentence is unreasonable because it is significantly longer than the sentence imposed upon several of his co-conspirators. See 18 U.S.C. § 3553 (a)(6) ("The court, in determining the particular sentence to be imposed, shall consider . . . the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct."). Indeed, García's approximately 224-month sentence is six-and-a-half years longer than any of his codefendants' sentences and more than double the sentence imposed on the conspiracy's head honcho, Francisco Merán Montero. This - 41 - disparity is problematic, García avers, because many of these codefendants were bigger fish in the criminal syndicate than García and responsible for smuggling greater quantities of cocaine. We review the consistency of García's sentence with § 3553(a)(6) for abuse of discretion, see United States v. Bedini, 861 F.3d 10, 21 (1st Cir. 2017); United States v. Acevedo, 824 F.3d 179, 186 (1st Cir. 2016), and we find no such abuse here. "[A] sentence may be 'substantively unreasonable because of the disparity with the sentence given to a codefendant.'" Reyes-Santiago, 804 F.3d at 467 (quoting United States v. ReverolRivera, 778 F.3d 363, 366 (1st Cir. 2015)). But "'[a] well-founded claim of disparity' must compare 'apples . . . to apples.'" Bedini, 861 F.3d at 21 (quoting United States v. Mateo–Espejo, 426 F.3d 508, 514 (1st Cir. 2005)). Consequently, a disparity claim will not succeed if there are "material differences between [the complaining defendant's] circumstances and those of their more leniently punished confederates." United States v. GalindoSerrano, 925 F.3d 40, 52 (1st Cir. 2019) (quoting Reyes-Santiago, 804 F.3d at 467 ). Here, there is no undue disparity. The fact that García's co-conspirators received shorter sentences than García is fully explained by their materially different circumstances. Whereas García went to trial, his codefendants pled guilty. See, e.g., Galindo-Serrano, 925 F.3d at 52 ; Bedini, 861 F.3d at 21-22 ; - 42 - see also Reyes-Santiago, 804 F.3d at 467 (noting that it is "permissible [to] distin[guish] between co-defendants who go to trial and those who plead guilty, between those who cooperate and those who do not" (internal citations omitted)). The plea deals accepted by García's codefendants include drug-quantity stipulations which account for much of the disparity García complains of. And "to the extent [García] is challenging factbargaining in general . . . th[is] argument[] also fail[s]. We have upheld the practice generally." United States v. Hall, 557 F.3d 15, 21 (1st Cir. 2009); see also United States v. Rodriguez, 162 F.3d 135, 152-53 (1st Cir. 1998). Finally, because García never requested a downward variance based upon the disparity, cf. Robles-Alvarez, 874 F.3d at 52-53 , but merely used the disparity argument in support of his recommended Guidelines calculation, and because the reason for the disparity was apparent from the plea agreements, the district court acted within its discretion by declining to specifically address García's disparity argument at sentencing, see United States v. Rivera-Morales, 961 F.3d 1, 20 (1st Cir. 2020). As either a procedural or a substantive challenge, therefore, García's disparity-based objection to his sentence fails. IV. CONCLUSION For the foregoing reasons, we affirm the convictions but vacate the sentence and remand for resentencing. - 43 -