United States v. Kurt Mix

Case: 14-30837 Document: 00513099242 Page: 1 Date Filed: 06/30/2015 IN THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT No. 14-30837 United States Court of Appeals Fifth Circuit FILED UNITED STATES OF AMERICA, June 30, 2015 Lyle W. Cayce Plaintiff - Appellant Clerk v. KURT MIX, also known as Kurt E. Mix, Defendant - Appellee Appeal from the United States District Court for the Eastern District of Louisiana Before CLEMENT, PRADO, and ELROD, Circuit Judges. EDITH BROWN CLEMENT, Circuit Judge: Kurt Mix was a BP engineer involved in calculating the amount of oil spilling out of the Macondo well, the site of the Deepwater Horizon accident. He was prosecuted for deleting text messages and emails related to these calculations. He was acquitted of one count, but he was convicted of obstruction of justice for deleting a text message exchange between himself and his boss. After the verdict was rendered, Mix’s counsel discovered that the jury had been exposed to extrinsic evidence. The district court conducted a voir dire of the jurors, and it became clear that the jury foreperson, Juror 1, overheard in a courthouse elevator that other BP employees were being prosecuted. Case: 14-30837 Document: 00513099242 Page: 2 Date Filed: 06/30/2015 No. 14-30837 During a deadlock in deliberations, Juror 1 told the rest of the jury that she had overheard something that increased her confidence in voting guilty. Other members of the jury prevented her from revealing what she had overheard. The district court ordered a new trial based on the jury’s exposure to extrinsic evidence. The government appeals the grant of a new trial. We AFFIRM. FACTS AND PROCEEDINGS A. The Underlying Facts Kurt Mix was an engineer for BP. After the Deepwater Horizon accident on April 10, 2010, he was assigned to produce flow rate models to estimate how much oil was leaving the accident site each day. The public estimate was that the well was discharging 1,000 barrels of oil per day (BOPD). But Mix’s early estimates approached or exceeded 100,000 BOPD. Eventually, the Coast Guard increased the public estimate to 5,000 BOPD. Mix allegedly thought that BP wanted his estimates to match this 5,000 BOPD figure. But, even though he allegedly tried to adjust his models to reach this figure, his estimates were generally much higher. One of the consultants Mix was working with repeatedly told Mix that he thought that the actual discharge was well above 5,000 BOPD. Meanwhile, BP, working with the government, developed a plan to try to stop the oil spill. It designated this plan “Top Kill.” The idea was to pump mud into the well faster than oil was coming out, thereby sealing the well. BP held a meeting with government scientists on May 17, 2010 to discuss the plan. At the meeting, which Mix attended, scientists opined that the plan would not work if the actual flow rate was greater than 15,000 BOPD. The government alleges that Mix did not disclose that his models showed a much higher flow rate. Mix contends that his estimates were disclosed at this meeting. 2 Case: 14-30837 Document: 00513099242 Page: 3 Date Filed: 06/30/2015 No. 14-30837 Ultimately, BP attempted Top Kill between May 26 and 28. Mix was involved in this attempt, but it did not work. Both before and after BP attempted Top Kill, it issued a number of legal hold orders to Mix, advising him that he was obligated to keep any documents or information (including text messages) related to the Deepwater Horizon accident and the subsequent oil spill. BP also informed Mix about what he should do if he received a grand jury subpoena. After Mix received these notices, sometime between October 4 and 5, 2010, he deleted a text message string that he and his supervisor, John Sprague, had exchanged. Mix deleted about 331 messages. Some of these messages had been exchanged while he helped plan and carry out Top Kill. The government eventually recovered all but about 17 of the text messages. Some of the recovered messages pertained to Top Kill and flow rate estimates. B. The Trial The government alleged that Mix deleted these text messages to hide his contemporaneous thoughts about Top Kill and the flow rate so as to subvert the future grand jury proceeding that would convene to investigate the Deepwater Horizon accident. While Mix did not destroy other documents or information related to the estimated flow rate, the government alleged that he was candid with only a few people, including Sprague, who was both his supervisor and close friend. So, deleting his texts to Sprague would hide Mix’s actual thoughts about the flow rate. In contrast, Mix alleged that he deleted these messages without any nefarious intent. He argued that he was just trying to delete a candid photograph that Sprague had taken of Mix and had texted to him, but Mix unthinkingly deleted the entire text string instead. Mix pointed out that there was no evidence that he asked Sprague to delete the text message exchange from his own phone, so Mix could not have been trying to hide the messages. The district court precluded the government from directly 3 Case: 14-30837 Document: 00513099242 Page: 4 Date Filed: 06/30/2015 No. 14-30837 mentioning that Sprague had also deleted the text messages from his phone, but the government implied as much by telling the jury that some but not all of the texts were recovered from Sprague’s phone. C. The Deliberations The jury deliberated for two partial days and one full day. During the full day (the second day of deliberations), the jury became deadlocked. The district court gave a modified Allen 1 charge near the end of that day. After two more hours of deliberations, the jury convicted Mix of obstructing justice by deleting the texts between himself and Sprague. D. The Hearing on Extrinsic Influences Mix’s counsel immediately contacted the jurors without leave of court, allegedly to obtain feedback about the defense’s failed trial strategy. The contacted jurors revealed that, during the deadlock, the jury foreperson announced that she had overheard extrinsic information in the courthouse elevator, and this information gave her comfort in voting guilty. Mix’s counsel eventually filed a motion for a new trial based upon this extrinsic influence on the jury. 2 In response to this motion for a new trial, the district court held a hearing to determine the nature of the extrinsic influence. Juror 1 testified that, about two days before the jury began deliberating, she heard an unknown man on the courthouse elevator saying that Mix “was not the only person who was 1 Allen v. United States, 164 U.S. 492, 501 (1896) (allowing a district court to give a deadlocked jury an instruction providing, among other things, “that, although the verdict must be the verdict of each individual juror, and not a mere acquiescence in the conclusion of his fellows, yet they should examine the question submitted with candor, and with a proper regard and deference to the opinions of each other”). 2 As detailed in the district court’s opinion, defense counsel acted improperly, and perhaps even unethically, in contacting the jurors. The district court ultimately excluded the affidavits that defense counsel had collected from the jurors. This issue is not before us on appeal, and the government no longer argues that defense counsel’s improper contact with the jurors justifies denying a new trial. 4 Case: 14-30837 Document: 00513099242 Page: 5 Date Filed: 06/30/2015 No. 14-30837 being prosecuted” and that “[t]here were going to be other trials” of BP employees. She also testified that she did not remember the district court’s instruction to notify it about hearing extrinsic information. She admitted that she told the other jurors that she overheard something, but she denied revealing what she had overheard. Several other jurors testified that, during the deadlock, Juror 1 told them that she overheard information in the elevator, and this information gave her comfort in voting guilty. The jurors prevented her from telling them what she overheard. While Juror 1 denied saying that the information gave her comfort in voting guilty, the district court concluded that she had in fact said so. The government does not challenge this factual finding on appeal. The government opposed Mix’s motion for a new trial. It argued that the extrinsic information overheard by Juror 1 was cumulative and that any problem was cured by the extensive jury instructions directing the jury to disregard extrinsic evidence. E. The District Court’s Order The district court ordered a new trial based on the extrinsic influence on the jury. It reasoned that Juror 1 had clearly been troubled by the information. Further, it found that other jurors were likely influenced by Juror 1’s statement that she overheard extrinsic evidence that gave her comfort in voting guilty. Finally, it noted that the jurors’ failure to report the extrinsic evidence undermined its confidence in their ability to follow other instructions (such as the instruction to disregard extrinsic evidence). STANDARD OF REVIEW “We review only for abuse of discretion a court’s handling of complaints of outside influence on the jury.” United States v. Smith, 354 F.3d 390, 394 (5th Cir. 2003). We review a district court’s grant of a new trial under Federal 5 Case: 14-30837 Document: 00513099242 Page: 6 Date Filed: 06/30/2015 No. 14-30837 Rule of Criminal Procedure 33 using the same abuse-of-discretion standard. United States v. Piazza, 647 F.3d 559, 564 (5th Cir. 2011). DISCUSSION The introduction of extraneous prejudicial information into the jury room violates a defendant’s Sixth Amendment right to an impartial jury and his Sixth Amendment right to confrontation. E.g., Parker v. Gladden, 385 U.S. 363, 364–66 (1966) (per curiam). Even one juror’s prejudice is sufficient to warrant a new trial. See id. at 366 (“[P]etitioner was entitled to be tried by 12, not 9 or even 10, impartial and unprejudiced jurors.”). The government argues that the district court abused its discretion because the extrinsic evidence was not prejudicial. We disagree. The district court did not abuse its discretion when it found that the jury was prejudiced in two separate ways. First, Juror 1 was prejudiced by hearing that other BP employees were being prosecuted. Second, the rest of the jury was prejudiced by the foreperson’s enigmatic statement that she heard outside information that gave her comfort in voting guilty. To be entitled to a new trial based on an extrinsic influence on the jury, a defendant must first show that the extrinsic influence likely caused prejudice. United States v. Sylvester, 143 F.3d 923, 934 (5th Cir. 1998). The government then bears the burden of proving the lack of prejudice. Id. The government can do so by showing there is “no reasonable possibility that the jury’s verdict was influenced by the extrinsic evidence.” United States v. Davis, 393 F.3d 540, 549 (5th Cir. 2004). We deal with each component of this burden- shifting framework in turn. 6 Case: 14-30837 Document: 00513099242 Page: 7 Date Filed: 06/30/2015 No. 14-30837 A. Mix Met His Initial Evidentiary Burden Mix met his initial evidentiary burden of showing that prejudice was likely, both as to the information overheard by Juror 1 and as to the information that she relayed to the rest of the jury. i. Information Overheard by Juror 1 The government argues that the information overheard by Juror 1 was innocuous. It argues that the only plausible effect of the overheard information was that Juror 1 would feel better about convicting Mix, a relatively low-level BP employee, given that potentially higher-level BP employees were also being prosecuted. Because Mix had no right to jury nullification, this information was not prejudicial. But we agree with Mix that the overheard information lent credence to the government’s theory of the case, which was that Mix deleted the texts to hide the fact that he knew that he was misrepresenting the flow rate. Mix, in contrast, argued that he thoughtlessly deleted the texts. Knowing that other individuals were being prosecuted could bolster the government’s theory in at least two ways. First, Mix’s scheme to cover up his texts would only work if his supervisor, Sprague, had also deleted the text messages. Thus, if Sprague was also being prosecuted for the same misconduct, that would tend to bolster the government’s theory that Mix deleted the texts as part of a scheme rather than by mistake. Indeed, the district court specifically prohibited the government from presenting evidence that Sprague also deleted the texts, finding that the probative value of this evidence was outweighed by its potential for prejudice. Nonetheless, the jury could have inferred that Sprague deleted the texts based on the fact that the government stated that it recovered some but not all of the deleted texts from Sprague’s phone. The overheard information would tend to buttress the implication that Sprague had also deleted the texts as part of a scheme with Mix. 7 Case: 14-30837 Document: 00513099242 Page: 8 Date Filed: 06/30/2015 No. 14-30837 Second, Mix would have had a stronger motive to delete the text messages if he had acted unlawfully in misrepresenting the flow rate calculation. The jurors knew that other people had also allegedly misrepresented the calculation. Evidence that others were being prosecuted for this misrepresentation would increase Mix’s motive to cover up his own misrepresentation. The government argues that Juror 1 could not have reasonably inferred that the other prosecuted BP employees were Sprague or others involved in misrepresenting the estimated flow rate. But that was the obvious inference, given that Juror 1 overheard the information in the courthouse elevator during Mix’s trial, which was a trial that centered around allegedly false flow rate calculations and deleted text messages. The government also argues that evidence about Mix’s motive is irrelevant because it was not an element of the crime. But, while the government did not have to prove why Mix wanted to subvert the grand jury, it had to prove that he deleted the text messages with the intent of subverting future grand jury proceedings. The government attempted to prove this intent partly by showing that Mix had a motive to subvert the grand jury proceedings. Finally, independent of the government’s theory of the case, the fact that Mix was part of BP, an organization that perpetrated allegedly criminal acts, could lead the jury to find him guilty by association. Of course, such a finding would be improper, but the jury instructions did not dwell on this point. Thus, the district court did not abuse its discretion in finding that the overheard information was likely prejudicial. The information that Juror 1 received was just the kind of information that could affect a juror, particularly given the government’s theory of the case. 8 Case: 14-30837 Document: 00513099242 Page: 9 Date Filed: 06/30/2015 No. 14-30837 ii. Information Relayed by Juror 1 The government argues that the information that Juror 1 relayed to the rest of the jury was not extrinsic. The government asks us to imagine a scenario where a juror pretends that she overheard information that gave her comfort in voting guilty, but, in fact, she has not overheard any information. The government argues that this situation would simply involve the juror’s strategy to persuade other jurors and therefore would be an intrinsic rather than extrinsic injection of information into the deliberation room. We reject this argument. A juror’s reference to non-record evidence that directly pertains to the particular case being considered is by definition an injection of extrinsic information, even if it is fabricated. The government also argues that the intrusion on the rest of the jury was de minimis, given that Juror 1 did not actually reveal the extrinsic information. But, given the nature of the communication (the foreperson vouching for Mix’s guilt based on unrevealed extrinsic evidence), the district court did not abuse its discretion by holding that prejudice was likely. Moreover, objective evidence indicated that the jury was indeed affected by Juror 1’s statement. The statement was made by the foreperson during a deadlock, and, within two hours, the jury returned a guilty verdict. 3 This evidence further supports a holding that prejudice was likely. The district 3 The government disputes that Juror 1 made her statement after the district court gave the modified Allen charge. But the district court found that the jury convicted Mix two hours after Juror 1 made her extrinsic statement, which means that it found that the statement was made after it issued the Allen charge. Further, one of the jurors testified that Juror 1 told the jury about the extrinsic information either at the end of the day that they received the Allen charge or on the morning after that. Three other jurors also testified that they believed that she made the statement after they had notified the district court of the deadlock. So the district court was justified in inferring that Juror 1 made her statement after or shortly before the Allen charge, meaning that the jurors resolved their deadlock and came to their decision within two hours of hearing her statement. 9 Case: 14-30837 Document: 00513099242 Page: 10 Date Filed: 06/30/2015 No. 14-30837 court therefore did not abuse its discretion in holding that the presumption of prejudice was triggered. B. The Government Did Not Prove Lack of Prejudice Because Mix met his initial evidentiary burden, the government was required to prove that the extrinsic evidence did not prejudice the jury—i.e., that there was “no reasonable possibility” that the jury was influenced by the extrinsic evidence. Davis, 393 F.3d at 549. The district court did not abuse its discretion in holding that the government failed to discharge this burden. i. Cumulative Evidence The government argues that the information Juror 1 overheard was cumulative of evidence that was introduced during other parts of the proceedings. Given that the other jurors never heard the content of the overheard statement, this argument only affects prejudice as to Juror 1, not prejudice as to the rest of the jury. The government points to three places in the record showing that the jury heard about other cases related to the Deepwater Horizon accident. First, it points to an instruction saying that there had been “substantial publicity about the Deepwater Horizon incident and other legal cases arising from that incident.” Second, it points to a question the district court posed to an FBI agent, asking whether she had testified “[f]ive times for all of the issues before that grand jury relating to BP including other defendants and other criminal cases, or are you talking about Mr. Mix?” 4 Third, it points to the same FBI agent’s testimony that she had testified before the grand jury “not only for this case, but other aspects of Deepwater Horizon.” But none of these particular instances indicate that the jury heard that the government was actively 4 Mix points out that the FBI agent did not have the chance to answer this question because Mix’s counsel withdrew his preceding question. 10 Case: 14-30837 Document: 00513099242 Page: 11 Date Filed: 06/30/2015 No. 14-30837 prosecuting other defendants and bringing them to trial. At most, the jury heard that there were other cases about the incident (whether civil or criminal) and that the grand jury had considered whether to indict others (whether or not it had actually returned indictments, and whether or not the government had later dropped the cases). Thus, the information overheard by Juror 1 went beyond what she heard at trial in that it specified that other BP employees were actively being prosecuted, not just that the grand jury had considered whether to indict them. Further, only the third instance involves the jury actually hearing evidence of other cases; the first two instances were not evidence but rather were instructions or questions by the court. The government also argues that it would have been “unsurprising and innocuous” to learn that other defendants were being prosecuted for the Deepwater Horizon accident, given that it “killed 11 men and caused the largest oil spill in history.” But we do not believe this was necessarily a commonsense inference. Apparent accidents often do not result in criminal liability, even if they are accompanied by major losses of life and property. Thus, the district court did not abuse its discretion in concluding that the government did not carry its burden of proving harmlessness based on the cumulative nature of the information Juror 1 overheard. ii. Jury Instructions The government also argues that the jury instructions cured any prejudice. The district court did not apply the presumption that the jurors followed their instructions because they violated an instruction by failing to bring the extrinsic evidence to the court’s attention. Admittedly, the jurors other than Juror 1 may have reasonably believed that they did not have to bring the information to the district court’s attention because they did not hear the substance of the extrinsic evidence. In addition, the other jurors prevented Juror 1 from revealing the content of the extrinsic information she overheard, 11 Case: 14-30837 Document: 00513099242 Page: 12 Date Filed: 06/30/2015 No. 14-30837 which supports the inference that the jurors did heed some of the court’s instructions. Nonetheless, the district court did not abuse its discretion in holding that the prejudice was not cured by generic instructions to disregard extrinsic information. At the very least, Juror 1 failed to follow the instruction to report exposure to extrinsic information to the district court. Further, by discussing their deliberations with Mix’s counsel after they rendered the verdict, at least two of the jurors did not follow the district court’s instruction not to discuss their deliberations during interviews without the district court’s permission. Given the lapses of at least three jurors in following the district court’s instructions, the court did not abuse its discretion in concluding that the generic instructions about extrinsic evidence did not cure the prejudice. iii. Weight of the Evidence The government now argues that the evidence against Mix was so overwhelming that the extrinsic information was irrelevant to his conviction. This argument fails. 1. Forfeiture First, the government did not urge this point below, so it is forfeited. The government contends that it raised this argument by citing United States v. Ruggiero, 56 F.3d 647 (5th Cir. 1995), a case holding that a district court should evaluate the weight of the evidence when determining whether to grant a new trial based on an extrinsic influence on a jury. But the government did not actually quote, discuss, or in any way raise this weight-of-the-evidence factor when it cited Ruggiero. Moreover, at a motions hearing, the government explicitly informed the district court that another case “corrected the standard that was used in Ruggiero.” That is, the government essentially disclaimed its 12 Case: 14-30837 Document: 00513099242 Page: 13 Date Filed: 06/30/2015 No. 14-30837 reliance on Ruggiero. 5 While the government later cited Ruggiero in a district court brief, this citation was only used to support an argument that the district court was prohibited by the Federal Rules of Evidence from considering whether an extrinsic influence had a subjective effect on a juror’s mind. The government cannot now rely on its general citation of Ruggiero for an evidentiary proposition to preserve its new argument regarding the weight of the evidence. The government also argues that it preserved this point by raising arguments related to the weight of the evidence in response to a motion for acquittal and a motion for a new trial based on the manifest injustice of a verdict that is against the great weight of the evidence. But, for a motion for acquittal, the “relevant question is whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” United States v. Valle, 538 F.3d 341, 344 (5th Cir. 2008) (quoting Jackson v. Virginia, 443 U.S. 307, 319 (1979)). Granting a motion for a new trial based on the weight of the evidence generally requires the district court to find that the evidence “preponderates sufficiently heavily against the verdict such that a miscarriage of justice may have occurred.” United States v. Tarango, 396 F.3d 666, 672 (5th Cir. 2005) (quoting United States v. Lincoln, 630 F.2d 1313, 1319 (8th Cir. 1980)). In contrast, the evidence does not have to preponderate heavily against guilt to grant a new trial based on an extrinsic influence on the jury. Instead, the relevant question is whether the government can prove that the weight of 5 At oral argument, the government argued that it offered this disclaimer because Ruggiero presumed prejudice rather than imposing an initial evidentiary burden upon the defendant, and a subsequent case abrogated this portion of Ruggiero. But the government did not limit its disclaimer before the district court to the evidentiary burden issue. 13 Case: 14-30837 Document: 00513099242 Page: 14 Date Filed: 06/30/2015 No. 14-30837 the evidence was strong enough that the extrinsic influence was harmless. See Davis, 393 F.3d at 549 (stating that government bears burden of proving harmlessness, and weight of evidence is one factor in that analysis). Thus, the fact that the government made weight-of-the-evidence arguments in response to other motions that involve completely different evidentiary standards did not imply that the government was making a similar argument in response to Mix’s motion for a new trial based on extrinsic influence. Finally, the government argues that it can only waive claims, not arguments, by failing to raise them below. But the government clearly can forfeit arguments by failing to raise them. See United States v. Jones, 132 S. Ct. 945, 954 (2012) (holding that argument not raised by government below was forfeited). Admittedly, the government might be entitled to plain error review of its forfeited argument. See United States v. Castillo, 386 F.3d 632, 637 (5th Cir. 2004) (applying plain error review in criminal appeal by government). But the government does not argue that this court should apply plain error review, so it has waived this point. See United States v. Griffith, 522 F.3d 607, 610 (5th Cir. 2008) (“[T]he failure to raise an issue on appeal constitutes waiver of that argument.”). 2. The Merits Even if we reached the weight-of-the-evidence argument, however, we would still affirm. True, the district court found that the evidence was substantial enough to deny a directed judgment of acquittal or a new trial based on insufficient evidence. But, probably because of the government’s failure to brief the issue, it did not explicitly decide whether the evidence was so one-sided that the otherwise prejudicial extrinsic evidence was harmless. Reviewing the record evidence, this case is not that one-sided. Mix vigorously contested that he deleted the text messages with the intent of stymying the grand jury investigation. He presented evidence that, before he deleted the 14 Case: 14-30837 Document: 00513099242 Page: 15 Date Filed: 06/30/2015 No. 14-30837 text string, the last text message he received from Sprague was “a picture of himself [at a meeting] sitting there smiling in his pink shirt.” Mix argued that he unthinkingly deleted the whole string when he only wanted to delete that picture. Moreover, he presented some contested evidence that his team had disclosed the worst-case flow rate estimate of 87,000 BOPD at the Top Kill meeting, which would tend to refute the government’s theory that he deleted the text messages because he had misrepresented the true estimates at that meeting. Mix also presented evidence that all of the supposedly important information contained in the deleted text messages was preserved on his laptop, so deleting the texts would not have hidden his flow rate work. Similarly, he presented evidence that on September 27, 2010, when BP’s document collection vendor met with Mix to collect his electronic data, Mix brought his iPhone with him to the meeting. Mix did not delete the Sprague text message string until October 4 or 5, so the text messages were still on Mix’s iPhone at the time that he met with the document collection vendor. Finally, Mix presented relatively strong evidence that, contrary to the government’s theory, he had been hopeful that Top Kill would work. 6 If the jury had believed Mix’s evidence, it likely would not have found that he “corruptly” deleted the text messages, as was necessary for his conviction. The extrinsic statement overheard by Juror 1 would tend to undermine Mix’s theory that he innocently deleted the text message. Further, Juror 1’s vouching for Mix’s guilt based on extrinsic evidence could have resulted in the jury voting to convict, despite otherwise harboring some doubts about Mix’s intent in deleting the text messages. Thus, even if the government had 6 The text messages that Mix exchanged with Sprague during Top Kill leave the impression that Mix hoped and sometimes thought that the operation was going to work, meaning that he did not know that the flow rate was over 15,000 BOPD. 15 Case: 14-30837 Document: 00513099242 Page: 16 Date Filed: 06/30/2015 No. 14-30837 preserved this weight-of-the-evidence argument, it has failed to prove harmlessness. iv. Procedural Error Finally, the government contended at oral argument that the district court abused its discretion by failing to consider the three Ruggiero factors, and this failure warrants automatic reversal. These three Ruggiero factors are “the content of the extrinsic material, the manner in which it came to the jury’s attention, and the weight of the evidence against the defendant.” Ruggiero, 56 F.3d at 653 (quoting United States v. Luffred, 911 F.2d 1011, 1014 (5th Cir. 1990)). A district court must consider these three factors when deciding whether an extrinsic influence was harmless. Davis, 393 F.3d at 549. But, here, the district court did not commit reversible error by failing to explicitly discuss these factors. 1. Forfeiture First, the government failed to preserve this argument because, below, it did not argue for the consideration of the three Ruggiero factors; indeed, it did not even mention them. As we described above, a single citation to Ruggiero was not enough to properly raise the three-factor test, particularly after the government expressly and unqualifiedly told the district court that Ruggiero was no longer good law. 2. The Merits Even if the government did not forfeit its argument, the government has not proven harmlessness based on the three Ruggiero factors. First, as explained above, the content of the extrinsic evidence was not inherently harmless. Second, the manner in which the information came to the jurors’ attention does not suggest harmlessness. Juror 1 overheard the extrinsic evidence before all of the evidence had been introduced. Cf. Ruggiero, 56 F.3d at 653 (upholding denial of new trial and noting that juror “heard the extrinsic 16 Case: 14-30837 Document: 00513099242 Page: 17 Date Filed: 06/30/2015 No. 14-30837 evidence only after all of the evidence had been introduced”). Also, while there is no indication that the man in the elevator purposely exerted an extrinsic influence on Juror 1, Juror 1 purposely exerted an extrinsic influence on the rest of the jury by telling them that she had overheard something. Cf. id. (upholding denial of new trial and noting that juror did not tell the rest of the jury about extrinsic evidence until after it determined the verdict). Third, as explained above, the weight of the evidence against Mix was not so overwhelming as to compel a finding of harmlessness. CONCLUSION For the foregoing reasons, we AFFIRM the district court’s order granting Mix’s motion for a new trial. 17