United States v. Carpenter

781 F.3d 599, 2015 U.S. App. LEXIS 5109, 2015 WL 1412584

• Court: Court of Appeals for the First Circuit
• Decided: March 30, 2015
• Docket: 14-1286
• Status: Published

Opinion

KAYATTA, Circuit Judge.

We opine for the third time on the United States’ prosecution of Daniel Carpenter for mail and wire fraud in connection with his mishandling of client escrow funds. In 2007, we affirmed a district court order setting aside a jury verdict of guilty in favor of a new trial. United States v. Carpenter, 494 F.3d 13 (1st Cir.2007). In 2013, after a second jury also found Carpenter guilty, we reversed a district court order setting aside that verdict, and remanded for sentencing. United States v. Carpenter, 736 F.3d 619 (1st Cir.2013). Now, post-sentencing, we consider Carpenter’s direct appeal in which he argues, among other things, that the lengthy duration of this criminal proceeding violated his constitutional and statutory speedy trial rights. For the following reasons, we affirm the judgment of the district court on all grounds.

I. Background

Our 2013 opinion details the acts for which Carpenter stands convicted. In a nutshell, he told clients he would hold their money in escrow accounts for which the client would pay a fixed fee and which would cautiously generate returns of either three or six percent; then (unbeknownst to his clients) he invested the money in high-risk, high-return stock options, hoping to generate excess returns to keep for himself. His option trading fared poorly, and he lost nine million dollars in client funds. At trial, he argued unsuccessfully that he never promised that the client funds would be safe, and that he did not intend to defraud his clients when he failed to disclose his real strategy of using their money to make risky investments to see if he could hit a home run for himself.

Central to this appeal are the details of how this criminal proceeding has lasted so long.

A. The Indictment, First Trial, and Appeal (September 2004 — July 2007)

The United States indicted Carpenter in September 2004. In July 2005, Carpenter’s first trial ended with a conviction. He moved, among other things, for a new trial. In December 2005, the district court granted that motion for a new trial on the grounds that the government’s repeated use of gambling metaphors had unfairly inflamed the jury’s passions. United States v. Carpenter, 405 F.Supp.2d 85, 103 (D.Mass.2005). The government appealed, but our court affirmed the order in July 2007, remanding for a new trial. Carpenter, 494 F.3d at 13. 1

B. The Second Trial, Appeal, and Sentencing (August 2007 — March 2014)

Following remand, Carpenter’s second trial ended with another conviction in June 2008. The district court initially scheduled a sentencing hearing for September 23, 2008. The district court did not sentence Carpenter, though, until almost six years later, in March 2014. This lengthy interval provides the primary basis for Carpenter’s Sixth Amendment argument.

1.Carpenter’s Motions for Mistrial, Acquittal, and New Trial (June— November 2008)

On June 17, 2008, just before the end of the second trial, Carpenter moved for a mistrial and partial acquittal. On July 3, immediately after his second conviction, Carpenter filed a sixty-seven-page motion for acquittal or new trial. 2 Using a different legal team, he also filed two “supplemental” motions for acquittal and new trial, raising a number of additional arguments. During July and August 2008, the government and Carpenter filed eleven more motions adding to the arguments for acquittal, mistrial, and new trial, and seeking various rulings on page limits and deadlines. By August 2008, the government complained that Carpenter’s briefing totaled over one hundred pages of opening briefs, and nearly eighty pages of reply briefs.

In September, Carpenter filed a second motion for a new trial based on newly discovered evidence. In November, he filed yet a third motion for a new trial based on different newly discovered evidence. (Neither motion concerns the newly discovered evidence at issue in this appeal.)

2. The District Court Hearing on the Post-Trial Motions (December 2008)

On December 3, 2008, the district court held a hearing focusing on the June 17 mistrial motion, which was based on the argument that the government had knowingly solicited false testimony from a witness in violation of Napue v. Illinois, 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959). The court also entertained argument on the July 3 motions for acquittal or new trial. The court indicated that it would decide the other pending motions based on the written submissions.

3. Motions Related to the Merrill Lynch Civil Litigation (March— June 2009)

Before the district court ruled on the numerous, often-lengthy motions before it, Carpenter began making additional filings. Understanding these new filings requires a bit of context.

Carpenter traded his clients’ funds, in part, through an investment account with Merrill Lynch. At Carpenter’s trial, the government solicited testimony from three Merrill Lynch employees about the riskiness of Carpenter’s trading strategy. These employees, including one to whom we refer as “GL,” denied they knew Carpenter was trading third-party funds. However, GL’s testimony was impeached when a lawyer for one of Carpenter’s clients testified that Carpenter had arranged a phone call between that lawyer and GL. Phone and fax records corroborated the lawyer’s version of events. 3 Even when confronted with the phone records, however, GL continued to deny the conversation took place.

Carpenter’s primary defense to the fraud charges was his good faith: i.e., that he lacked the specific intent to defraud, because he genuinely thought he had investment discretion over the funds his company held. 4 In his view, evidence showing he was open with Merrill Lynch about the source of the funds he was trading supported his good-faith defense on the logic that he would not have been open with Merrill Lynch if he knew he was doing something wrong. The government largely demurred, calling Merrill Lynch witnesses primarily to demonstrate the riskiness of Carpenter’s approach, and devoting little attention to whether anyone at Merrill Lynch knew the source of the funds. As the prosecution told the jury, its position was that whatever Merrill Lynch employees knew was irrelevant because what mattered were the representations Carpenter made to his clients.

The district court, too, questioned the probative force of Carpenter’s argument that Merrill Lynch knew the source of the funds. It rejected a motion to acquit for insufficient evidence based on a version of this argument after both trials. Carpenter, 405 F.Supp.2d at 93-94; United States v. Carpenter, 808 F.Supp.2d 366, 378 (D.Mass.2011). Nevertheless, Carpenter continued (and continues in this appeal) to argue that any evidence showing that Merrill Lynch was aware of the source of the funds was highly relevant to his claims.

With this background in mind, we now return to the procedural chronology. In early 2009, shortly after the hearing on Carpenter’s various motions for mistrial, acquittal, and new trial, new documents began to emerge as part of a civil lawsuit against Merrill Lynch. Those documents further tended to show that Merrill Lynch employees were aware of the source of the funds Carpenter was trading.

Carpenter touted this unfolding information as bearing on his pending motions. See ECF No. 349 at 1, June 10, 2009 (noting that the new information has “a direct bearing on Carpenter’s pending post-trial motions”). On March 19, 2009, Carpenter filed an emergency motion seeking, among other things, to compel the government to acquire and review those new documents. Between March and July, Carpenter filed five reports updating the district court on the developments in the civil lawsuit against Merrill Lynch, responding to the government’s arguments, and reiterating his argument that the government should be ordered to seek out and review the new documents. This included a joint report submitted by the government and Carpenter, indicating that the government had agreed to review the new information and report “whether the government agrees that a new trial or dismissal of the charges is appropriate in light of the new evidence.” ECF No. 352 at 2, June 26, 2009.

4. Carpenter’s Motion for Expedited Hearing (October 2009)

On October 28, 2009, Carpenter moved for expedited review of his pending mo tions, albeit by filing a thirty-three-page motion supported by over three hundred pages of exhibits. ECF No. 355, October 28, 2009. Although the government had not yet reported back on the new Merrill Lynch information pursuant to the June 26 agreement, Carpenter argued that there was now more than enough information available for a ruling on his initial motions from June and July 2008. Id. at 29. Carpenter stressed that fifteen months had elapsed since he filed those initial post-trial motions. Id. at 1-2, 26, 31-32.

5. Additional Delay and Additional Motions (October 2009 — September 2011)

By June 2010, the district court had yet to rule, notwithstanding Carpenter’s renewed expression of concern about the delay. See ECF No. 360 at 3-4, January 13, 2010. In a June 17 letter, Carpenter sent the court conflicting messages: he asked it to decide the pending motions for acquittal, mistrial, or new trial, but he also advised the court that such action likely would not be necessary because Carpenter expected the government to dismiss the indictment against him. ECF No. 370 at 1, June 17, 2010 (“[I]t is our expectation that, in light of these new developments, the Government will move to dismiss the indictment with prejudice ... thereby obviating the need for the Court to resolve the motions that have been pending for two years.”).

Another half-year then passed without the government dropping the case or the district court deciding the motions that Carpenter had told the court it should, but might not need to, decide. In January, February, and July 2011, Carpenter submitted three more letters providing supplemental authority or new information, each of which also included a request for rulings on the motions. The third update also requested a status conference, noting that it had now been three years since the trial, “every day of which is alone punishment.” ECF No. 375 at 8, July 14, 2011.

C. Second Grant of New Trial and Appeal (September 1, 2011)

On September 1, 2011 — more than three years after Carpenter’s initial motion for a new trial after his second conviction in July 2008 — the district court granted Carpenter’s motion for a new trial. Carpenter, 808 F.Supp.2d at 386. The court denied Carpenter’s motion for acquittal and other related motions. Id. The government appealed (No. 11-2131), and Carpenter filed an appeal of his own (No. 11-2133), which he later moved to consolidate with the first appeal.

In March 2012, while both these appeals were still pending, Carpenter filed a motion in district court to dismiss his indictment for violating his Sixth Amendment right to a speedy trial. The district court denied the motion on jurisdictional grounds, citing the pending appeals. Carpenter then moved for this court to remand the pending appeals to allow the district court to consider his Sixth Amendment speedy trial motion. On May 3, 2013, this court denied the motion to remand, as well as the motion to consolidate the appeals, and a briefing schedule was set.

On May 23, 2013, the government moved for a forty-four-day extension to file its brief on the grounds that the assistant United States attorney who had filed the appeal in November 2011 was no longer in charge of the case, and the attorney who had taken over needed time to review the extensive record while also managing other cases with May and June deadlines. This court granted the order that day. Also on that day, Carpenter moved in this court to dismiss the indictment, arguing that the government had failed to “diligently prosecute[]” the appeal, as required by 18 U.S.C. § 3731. The basis for this claim was that the government had not moved to expedite its appeal during the nineteen months that Carpenter’s various motions were under advisement, and that it now sought an extension. This court denied the motion in July. The case proceeded through briefing and was argued on November 7, 2013. On November 25, 2013, this court reversed the district court’s grant of a third trial and remanded for sentencing. Carpenter, 736 F.3d at 632. 5

D. Sentencing (February 2014)

Prior to sentencing, Carpenter again moved to have the district court dismiss his indictment for violating the Sixth Amendment’s speedy trial clause. 6 The district court denied the motion. United States v. Carpenter, No. 04-10029-GAO, 2014 WL 691659 (D.Mass. Feb. 21, 2014). On February 26, 2014, the district court sentenced Carpenter to thirty-six months’ imprisonment, three years’ supervised release, and penalties; judgment was entered March 4. On May 23, the district court also granted the government’s motion to order forfeiture of over fourteen million dollars.

E. The Current Appeal

In this appeal, Carpenter argues that the district court erred in its February 21, 2014, order by failing to set aside his conviction and dismiss the indictment because the duration of the proceedings violated his Sixth Amendment right to a speedy trial. He also challenges: the district court’s April 8, 2008, order denying relief under the Speedy Trial Act, United States v. Carpenter, 542 F.Supp.2d 183 (D.Mass. 2008); the district court’s September 1, 2011, order denying Carpenter’s motion for acquittal on the basis of sufficiency of the evidence, and (implicitly) denying a new trial on the basis of newly discovered Merrill Lynch evidence, United States v. Carpenter, 808 F.Supp.2d 366 (D.Mass. 2011); and the district court’s March 4, 2014, sentencing order. 7

II. Standard of Review

Our court has repeatedly reviewed district court rulings on Sixth Amendment speedy trial motions for abuse of discretion. See United States v. Salimonu, 182 F.3d 63, 69 (1st Cir.1999); United States v. Santiago-Becerril, 130 F.3d 11, 21 (1st Cir.1997); United States v. Colombo, 852 F.2d 19, 21 (1st Cir.1988). This formulation of the standard varies from that used in most other circuits, which review such claims de novo, albeit while applying clear error review to the district court’s factual findings. See, e.g., United States v. Lopesierra-Gutierrez, 708 F.3d 193, 202 (D.C.Cir.2013); United States v. Velazquez, 749 F.3d 161, 174 (3d Cir.2014); United States v. Bishop, 629 F.3d 462, 466 (5th Cir.2010); United States v. Jackson, 473 F.3d 660, 664 (6th Cir.2007); United States v. Hills, 618 F.3d 619, 629 (7th Cir.2010); United States v. Summage, 575 F.3d 864, 875 (8th Cir.2009); United States v. Corona-Verbera, 509 F.3d 1105, 1114 (9th Cir.2007); United States v. Larson, 627 F.3d 1198, 1207 (10th Cir.2010); United States v. Villarreal, 613 F.3d 1344, 1349 (11th Cir.2010). Our formulation of the standard also seems in tension with both our standard for reviewing motions to dismiss under the Speedy Trial Act itself, see United States v. Valdivia, 680 F.3d 33, 38 (1st Cir.2012) (reviewing denial of a Speedy Trial Act motion “de novo as to legal rulings and for clear error as to factual findings”), and more significantly, with our recent en banc ruling that the ultimate question of whether prison officials have violated the Eighth Amendment is reviewed de novo. Kosilek v. Spencer, 774 F.3d 63, 84 (1st Cir.2014) (en banc). 8 Be that as it may, this case presents no need to resolve any fine questions regarding the standard of review because, even under de novo review, our conclusion would remain the same.

We review the denial of a Rule 29 motion for judgment of acquittal de novo, examining the evidence in the light most favorable to the verdict. United States v. Howard, 687 F.3d 13, 19 (1st Cir.2012). We review the denial of a motion for a new trial based on newly discovered evidence for manifest abuse of discretion. United States v. Wright, 625 F.2d 1017, 1019 (1st Cir.1980). Finally, we review sentencing decisions for abuse of discretion, examining the district court’s findings of fact for clear error and its interpretations of the sentencing guidelines de novo. United States v. Flores-Machi-cote, 706 F.3d 16, 20 (1st Cir.2013).

III. Analysis

A. Sixth Amendment Right to a Speedy Trial

In Barker v. Wingo, 407 U.S. 514, 530, 92 S.Ct. 2182, 33 L.Ed.2d 101 (1972), the Supreme Court adopted a balancing test for assessing claims of a violation of the Sixth Amendment right to a speedy trial. The test weighs four factors: the length of the delay, the reason for the delay, the defendant’s assertion of the right to a speedy trial, and whether the defendant has been prejudiced by the delay. Id. The Court expressly rejected a bright-line rule in favor of a “functional analysis of the right in the particular context of the case.” Id. at 522, 92 S.Ct. 2182. It did so because the remedy — dismissal of the indictment— was “unsatisfactorily severe.” Id. (“The right of a speedy trial is necessarily relative. It is consistent with delays and depends upon circumstances. It secures rights to a defendant. It does not preclude the rights of public justice.”) (quoting Beavers v. Haubert, 198 U.S. 77, 87, 25 S.Ct. 573, 49 L.Ed. 950 (1905)).

We begin our analysis by determining the length of delay at issue. In undertaking that determination, we consider and reject the government’s argument that the Sixth Amendment places no limit on the length of post-conviction proceedings. We next closely examine the reasons for the delay, the extent to which Carpenter sought greater speed, and the nature of any prejudice caused to Carpenter by the delay.

1. Length of the Challenged Delay 9

The first factor, length of delay, is both a triggering mechanism for the rest of the analysis, and a factor in that analysis. United States v. Souza, 749 F.3d 74, 81 (1st Cir.2014). While Carpenter’s brief often cites the ten years that passed from indictment to sentencing, he does not claim that the pace of the proceedings was undue at all times. For example, he has no bone to pick with the speed with which the case moved forward from the indictment in September 2004 to the end of his first trial in July 2005. Nor does he suggest any delay in the district court taking five months to hear and decide the motions Carpenter filed in the wake of that first trial. Certainly the proceedings moved apace up to December 2005, when the district court entered orders denying Carpenter’s motion for acquittal and granting him a new trial.

Carpenter’s complaint about the pace of proceedings finds its first toehold with the government’s decision in January 2006 to appeal the district court’s granting of a new trial. That appeal sidetracked the case for just over twenty months until the mandate denying the appeal issued in September 2007. Second, Carpenter complains about the pace of proceedings between the end of the second trial in June 2008 and the entry, in September 2011, of the district court’s order setting aside the second jury’s verdict and ordering a third trial. Finally, Carpenter complains about the twenty-six months consumed by the government’s successful appeal of the order setting aside the second jury verdict.

The government concedes the foregoing calculations of delay. It argues, however, that the latter two time periods should play no role in our Sixth Amendment analysis because they postdated the June 2008 guilty verdict that our court ultimately sustained. Describing this passage of time as, at worst, a delay in sentencing, the government urges this court to follow the Second Circuit in United States v. Ray, 578 F.3d 184, 198-99 (2d Cir.2009). In that case, the court opined that “the harms arising from delayed sentencing ... are quite different from those animating the Speedy Trial Clause.” Id. at 198. The court concluded that the Constitution protects defendants from sentencing delay through the Fifth Amendment, not the Sixth. Id. at 199.

We decline to adopt that conclusion. Although neither the Supreme Court nor this circuit has held that the Sixth Amendment applies to post-conviction delay, both have assumed so arguendo. See Pollard v. United States, 352 U.S. 354, 361, 77 S.Ct. 481, 1 L.Ed.2d 393 (1957); United States v. Nelson-Rodriguez, 319 F.3d 12, 60 (1st Cir.2003) (noting that most circuits that had considered the issue had either held or assumed the same). It is no doubt true that the concerns arising from pre-trial delay — when a person presumed to be innocent stands under the shadow of accusation — are not identical to those arising from post-conviction delay. The difference, though, is not quite as great as it may seem, given that a guilty verdict is not yet final until appeals are exhausted. Moreover, our Sixth Amendment analysis itself recognizes the difference in the pre- and post-verdict time frames, in that the required balancing of interests includes an assessment of the extent to which delay causes prejudice. Thus, we see no reason to depart from the majority view that assumes that the Sixth Amendment also protects against post-trial delay.

Having thus rejected the government’s attempt to excise from our Sixth Amendment analysis the five years that passed between the second jury verdict and the decision of this court sustaining that verdict, we turn our attention to asking, initially, whether any or all of the delays that Carpenter challenges are sufficiently ordinary so as to terminate our Sixth Amendment analysis. “Until there is some delay which is presumptively prejudicial, there is no necessity for inquiry into the other factors.... ” Barker, 407 U.S. at 530, 92 S.Ct. 2182. In so doing, we accept Carpenter’s position that “[ejach of these periods should be assessed separately.” At the same time, we also consider any actual delay cumulatively.

We need not tarry in making this initial inquiry. Delay of around one year is considered presumptively prejudicial, and the presumption that delay prejudices the defendant “intensifies over time.” Doggett v. United States, 505 U.S. 647, 652 and n. 1, 112 S.Ct. 2686, 120 L.Ed.2d 520 (1992). Given that the periods of time here each well exceed one year, and cumulatively exceed six years, we think it practical to proceed to examining the reasons for that delay.

2. Reason for Delay

The second Barker prong, the reason for delay, is “often considered the focal inquiry.” United States v. Trueber, 238 F.3d 79, 88 (1st Cir.2001). As discussed above, Carpenter challenges three periods of delay: (1) the twenty months occupied by the government’s failed appeal after the first grant of a new trial in December 2005; (2) thirty-four months of the roughly three years it took the district court to rule on Carpenter’s motion for acquittal or new trial after his second conviction in June 2008, and (3) the twenty-six months consumed by the government’s second, successful appeal of the district court’s grant of a new trial in September 2011. We now consider each in turn.

a. The Government’s First Appeal

Carpenter first argues that the time consumed by the first appeal constitutes unwarranted delay because the government’s position on appeal was weak. In general, delay caused by interlocutory review does not cut against the government. United States v. Loud Hawk, 474 U.S. 302, 312-15,106 S.Ct. 648, 88 L.Ed.2d 640 (1986) (“Given the important public interests in appellate review ... it hardly need be said that an interlocutory appeal by the Government ordinarily is a valid reason that justifies delay.”) (internal citation omitted). The Loud Hawk court noted, however, that a “tangential or frivolous” appeal would weigh heavily against the government, so courts should consider the strength of the government’s position on the appealed issue, the importance of the issue to the case, and (in some cases) the seriousness of the crime to determine whether an appeal should cut against the government. Id. at 315, 106 S.Ct. 648.

These factors all cut strongly against concluding that the time consumed by the first appeal constituted unjustified delay. The government’s closing comments that led to the grant of a new trial did not even elicit a contemporaneous objection from defense counsel or rebuff from the court itself. While the district court in its discretion concluded that the comments were sufficiently prejudicial as to require a new trial, that same court, in its February 2014 order rejecting Carpenter’s speedy trial motion, described the government’s appeal of that grant of a new trial to be “legitimate and justifiable.” Carpenter, 2014 WL 691659 at *2. Most notably, the panel hearing the appeal was split, with one judge finding persuasive the government’s argument that the district court erred in applying too strict a standard in assessing the impact of closing comments that failed to draw a contemporaneous objection. United States v. Carpenter, 494 F.3d 13, 29 (1st Cir.2007) (Campbell, J., dissenting). However one defines the category of reasonably strong appeals, it likely includes an appeal that garnered an actual vote by a circuit court judge in favor of the appeal.

The appealed order had set aside a jury verdict convicting Carpenter of very serious charges. A successful appeal would have ended (i.e., also shortened) the case. Focusing his argument on the fact that the government chose to appeal at all, Carpenter otherwise raises no objection to the duration of the appeal. We therefore reject his contention that the time consumed by the appeal constituted unjustified delay of any type relevant to our analysis.

b. The Government’s Second Appeal

As for the government’s appeal of the second grant of a new trial in September 2011, Carpenter cannot argue that the government’s position, which actually prevailed, was weak. He argues instead that the appeal took too long — twenty-six months in total — because the government failed to diligently prosecute the appeal. Nineteen months elapsed between the government’s filing of a second notice of appeal in September 2011 and when a briefing schedule was set in May 2013. Carpenter also points to the government’s successful motion, after the briefing schedule was set, for a forty-four-day extension to allow new counsel to familiarize herself with the record.

The unusual passage of nineteen months between the filing of the appeal and the setting of a briefing schedule was not, however, due to any fault of the government. Rather, it was Carpenter’s own cross-appeal and related motions that slowed down the government’s appeal. As discussed above, after the government filed its notice of appeal on September 27, 2011, Carpenter on September 29 filed his own notice of appeal, which challenged the denial of his motions for acquittal and mistrial. On November 7, 2011, he moved to consolidate that appeal with the government’s appeal. In March 2012, he also moved in the district court to dismiss the indictment on Sixth Amendment grounds. When the district court in May 2012 denied that motion, citing the pending appeals, Carpenter moved for this court to remand the case to the district court to rule on the Sixth Amendment speedy trial motion. We denied this motion in May 2013, and also dismissed Carpenter’s cross-appeal for lack of jurisdiction. 10 At that point, with Carpenter’s own predicate motions resolved in due course, a briefing schedule for the government’s appeal was set.

Carpenter is correct that the government’s request for an extension then slowed things down by forty-four days, but the fact that this extension was warranted is evidenced by this court’s granting the government’s motion for an extension and denying Carpenter’s motion to dismiss for lack of diligent prosecution. With that final motion resolved, the case proceeded apace to argument on November 7, 2013, and a decision reversing the district court’s grant of a new trial on November 25, 2013. Carpenter, 736 F.3d at 632.

Carpenter cites no authority to suggest that the government should have sought expedited briefing, much less that it should have done so while Carpenter’s own motion to remand was pending. Yet another motion in a motion-laden case could hardly have helped the goal of swift resolution. We therefore agree with the government that the district court did not abuse its discretion in determining that the second appeal occasioned no unwarranted delay in the conclusion of the case.

c. The Time Between the Second Verdict and the Second New Trial Order

As the district court itself recognized, see Carpenter, 2014 WL 691659 at *2, Carpenter’s argument acquires some traction when we turn to the roughly thirty-eight months that passed between the end of the second trial in June 2008 and the district court’s granting Carpenter’s motion for a second new trial in September 2011. 11 A portion of that time passed in what can fairly be described as normal course. The parties agreed to a schedule for post-trial briefs, a hearing was set for October and then postponed to December 3, 2008, based on requests by both parties and without objection. Had nothing else been at issue, nor anything else filed, one would normally have expected a decision on the admittedly extensive and heavily briefed motions for acquittal or new trial by the June 2009 anniversary of the trial. Instead, no ruling issued until September 2011.

The district court admitted that it bore at least some responsibility for this “regrettable” delay, but also cited Carpenter’s many motions as the main culprit. Id. at *2-3. It is well-established that it cuts heavily against a defendant’s speedy trial claims when his own motions contribute to the delay. Loud Hawk, 474 U.S. at 316-17, 106 S.Ct. 648 (“Having sought the aid of the judicial process and realizing the deliberateness that a court employs in reaching a decision, the defendants are not now able to criticize the very process which they so frequently called upon.”) (internal quotation marks omitted); United States v. Worthy, 772 F.3d 42, 49 (1st Cir.2014); United States v. Muñoz-Franco, 487 F.3d 25, 60-61 (1st Cir.2007); Nelson-Rodriguez, 319 F.3d at 61; United States v. Muñoz-Amado, 182 F.3d 57, 62 (1st Cir.1999); United States v. Gibson, 353 F.3d 21, 22-23 (D.C.Cir.2003).

In this case, in addition to the numerous motions that were pending at the end of the second trial, Carpenter made twenty filings between the June 2008 jury verdict and the December 2008 hearing on his motions. As described in the facts section of this opinion, these filings included two separate sets of motions for acquittal or new trial filed by two separate legal teams and which advanced numerous, distinct theories of error; two additional motions for a new trial based on newly discovered evidence; and lengthy responses to the government’s responses. Many were very substantial. 12 Presumably he wanted the district court to review them all, and of course the court had to review the government’s responses.

Carpenter defends his avalanche of filings as justified. But that is beside the point. The filings cut against Carpenter in our Sixth Amendment analysis not because they were weak or otherwise not justified. Rather, they cut in this manner because they support the district court’s statement that the need to consider the filings reasonably consumed time.

Carpenter then points out that the district court did not rely on the post-hearing filings regarding Merrill Lynch when it finally granted Carpenter a new trial in September 2011. That is, if the district court granted Carpenter’s 2008 motion without regard to any of the arguments raised in subsequent motions, why did it not do so closer to 2008? 13 The problem with this argument is that it suggests the district court would have known at the outset the grounds on which its decision would rest. This expects too much of the district court. Carpenter’s Merrill Lynch motions began in March 2009, three months after the motion hearing. As he continued updating the court, he explicitly stated that this new information had “a direct bearing” on his pending motions. ECF No. 349 at 1, June 10, 2009. His regular updates suggested a fast-changing situation that could likely lead to more motions. At least initially, the district court cannot be faulted for holding off on making a ruling while the information was still evolving and Carpenter was insisting it was relevant.

This argument loses some force over time, however, particularly after the civil trial against Merrill Lynch concluded in July 2009. Although Carpenter continued to provide updates based on newly decided cases and the damages phase of the civil trial, after October 2009 they became less frequent, and Carpenter began pressing the district court to rule on his motions. See ECF No. 355 at 1, October 28, 2009. Granted, the unusually large number of motions, as well as their length, makes it unsurprising that the district court would still need an unusually long time to give them full consideration. By the time of Carpenter’s October 2009 motion, however, sixteen months had passed since the end of trial, and Carpenter had formally moved the district court to make a decision on the motions that had been argued in December 2008. Once the anniversary of that hearing had passed, the reason for any continuing delay can no longer be attributed primarily to Carpenter.

We therefore conclude that there was an unwarranted delay of some twenty-one months in the progress of this case, from roughly December 2009 to September 1, 2011. 14 Nevertheless, the prosecution played no role in this unwarranted delay. Nor was the district court using delay in any hostile manner. When delay is the result of negligence and not bad faith, it weighs less heavily in the balancing called for by Barker. Barker, 407 U.S. at 531, 92 S.Ct. 2182 (“A deliberate attempt to delay the trial in order to hamper the defense should be weighted heavily against the government. A more neutral reason such as negligence or overcrowded courts should be weighted less heavily but nevertheless should be considered....”); see Santiago-Becerril, 130 F.3d at 22.

3. Assertion of the Right

The third prong is the assertion of the speedy trial right, in particular the “frequency and force” with which the defendant objected to delay. Barker, 407 U.S. at 529, 92 S.Ct. 2182. Courts look with some skepticism at assertions of speedy trial rights made by defendants who contribute to the delay, and are particularly skeptical of those who raise the issue for the first time in a motion to dismiss. See, e.g., id. at 534-35, 92 S.Ct. 2182 (“Barker did not want a speedy trial.... While he hoped to take advantage of the delay in which he had acquiesced, and thereby obtain a dismissal of the charges, he definitely did not want to be tried.”); Santiago-Becerril, 130 F.3d at 22. And like the other factors, assertion of the right is not in itself decisive. See Muñoz-Franco, 487 F.3d at 60-61 (finding no violation despite the fact that the appellants three times raised speedy trial objections).

Carpenter’s first motion to dismiss on Sixth Amendment speedy trial grounds came in March 2012, after the government had appealed his second grant of a new trial. However, he began pressing for action on his pending motions for acquittal or new trial much earlier, when he moved for expedited hearing in October 2009, and he continued to stress the delay in subsequent filings. Thus, in our view, the district court’s conclusion that Carpenter’s assertion of the right had been “spotty at best,” Carpenter, 2014 WL 691659 at *4, is not quite accurate, at least in the time period from late 2009 to 2011, when his undecided motion was pending.

Of course, on June 17, 2010, Carpenter informed the court that the whole case might go away, and that the court might not need to decide the post-trial motions. It is reasonable to think that a busy trial judge, so informed, might attend to other matters rather than re-engaging with a voluminous set of motions. Nevertheless, this does not significantly detract from the fact that during the period of unwarranted delay, Carpenter otherwise steadily pressed the district court for action.

4. Prejudice

The prejudice prong seeks to protect three interests: avoidance of oppressive pretrial incarceration, minimizing anxiety and concern, and limiting the possibility that the defense will be impaired. Doggett, 505 U.S. at 654, 112 S.Ct. 2686 (quoting Barker, 407 U.S. at 532, 92 S.Ct. 2182).

A defendant must struggle to satisfy the prejudice prong after conviction, when two of the three factors relevant to the prejudice analysis — excessive pre-trial incarceration and impairment of an effective defense — are of little or no relevance. See Pérez v. Sullivan, 793 F.2d 249, 256 (10th Cir.1986). Thus, Carpenter cannot argue that this delay extended any pretrial detention because there was none: Carpenter was released on personal recognizance on February 24, 2004, the day of his arraignment. Carpenter likewise can offer no evidence that any delay prejudiced his defense. Indeed, a quicker pace may have deprived him of the basis for his forays concerning the Merrill Lynch documents.

Carpenter’s argument instead focuses on the anxiety he suffered throughout the proceedings. His brief, supported by record materials, describes a “living hell” of lost business opportunities, finan cial stress, sleeplessness, panic attacks, and the like. He points, however, to no opportunities that would not have been lost as well in the wake of a speedier conviction. And while anxiety about the outcome of post-conviction motions and appeals is no doubt real, anxiety is a normal part of the pendency of criminal charges. It therefore becomes a sign of prejudice only when “undue pressures” exist. Muñoz-Franco, 487 F.3d at 61 (quoting Santiago-Becerril, 130 F.3d at 22-23).

While Carpenter argues convincingly that he has suffered great stress throughout the proceedings, he does not demonstrate why his anxiety was greater than that suffered by many other defendants, other than that it continued longer. See United States v. Colombo, 852 F.2d 19, 26 (1st Cir.1988) (noting that “[t]he passage of time alone ... is not conclusive evidence of prejudice”). While it may be possible that post-conviction delay could result in prejudice by shifting the time period in which a defendant serves his sentence, Carpenter makes no such argument here. In sum, while the length of delay causes us to presume some prejudice, we find nothing in this record to establish that Carpenter suffered a type of prejudice that would take on added weight in our assessment of the constitutionality of that delay.

5. Weighing the Factors

While the travel of the case as a whole was remarkable, its length arose almost entirely because the district court exercised (and exceeded in one instance) its discretion in granting Carpenter relief from verdicts against him. In the end, we have an unjustified delay of roughly twenty-one months, which occurred after a guilty verdict was returned and without any meaningful fault of the government. While the delay was unfortunate, it did not impair the defense, create any undue pressure, or result in any period of incarceration.

Carpenter points to no precedent for setting aside a guilty verdict in such circumstances, nor are we aware of any. Rather, precedent points otherwise. In Katz v. King, 627 F.2d 568 (1st Cir.1980), we rejected a claim that a four-month delay between the completion of trial and the entry of the verdict violated the defendant’s right to a speedy trial. In so doing, we noted that other courts have “found delays in sentencing of up to twenty-nine months not to be excessive.” Id. at 576 (citing United States v. Campisi, 583 F.2d 692, 694 n. 5 (3d Cir.1978)). We also rejected a challenge to a fourteen-month delay between conviction and sentencing, relying “most importantly” on the diminished possibility of prejudice in the post-conviction time frame. Nelson-Rodriguez, 319 F.3d at 61 (“[T]he courts have great reluctance to find a speedy trial deprivation where there is no substantial and demonstrable prejudice.”); cf. Worthy, 772 F.3d at 49-50 (affirming the district court’s rejection of a Sixth Amendment claim based on a twenty-three-month pre-trial delay in light of the “complexity of the case, [the defendant’s] own responsibility for and acquiescence in the delay, and the absence of any cognizable prejudice”).

While each case must be evaluated in the context of its own circumstances, see Barker, 407 U.S. at 533, 92 S.Ct. 2182, the complete lack of any precedent for throwing out a guilty verdict when the defendant has not demonstrated prejudice (beyond the considerable anxiety felt by many criminal defendants post-conviction) strongly supports our own conclusion that, under Barker, Carpenter’s Sixth Amendment rights have not been violated. We do not reject the notion that post-conviction delay might give rise to a remedy, perhaps on mandamus review (which Car penter never sought). Cf. Dolan v. United States, 560 U.S. 605, 616-17, 130 S.Ct. 2533, 177 L.Ed.2d 108 (2010) (indicating that mandamus could be used to compel a district court to hasten a ruling under certain rare circumstances). However, the relief Carpenter seeks — dismissal of the indictment for which he has twice been found guilty — is unwarranted here.

B. Speedy Trial Act

Carpenter next makes a more technical, statutory version of a speedy' trial claim. He argues that the district court erred in its April 8, 2008, order by not dismissing the indictment for purportedly violating the Speedy Trial Act, 18 U.S.C. §§ 3161-3174, during the period of time between the district court’s first grant of a new trial on December 15, 2005, and setting the date for that trial on November 26, 2007. United States v. Carpenter, 542 F.Supp.2d 183, 183-85 (D.Mass.2008). This circuit reviews a denial of a statutory speedy trial claim de novo as to legal rulings, and for clear error as to factual findings. United States v. Valdivia, 680 F.3d 33, 38 (1st Cir.2012).

The Speedy Trial Act imposes a seventy-day deadline on bringing a defendant to trial, which normally runs from the later of the filing of the information or indictment, or the first appearance of the defendant. See 18 U.S.C. § 3161(c)(1). However, a specific provision covers retrials: section 3161(e) provides that if the defendant is to be tried again “following a declaration by the trial judge of a mistrial or following an order of such judge for a new trial,” or “following an appeal or a collateral attack,” that new trial must commence within seventy days “from the date the action occasioning the retrial becomes final.” Id. § 3161(e). In addition, section 3161(h) provides a list of eight kinds of delay that “shall be excluded ... in computing the time within which the trial of any such offense must commence.” This list includes “delay resulting from any interlocutory appeal.” Id. § 3161(h)(1)(C).

In Carpenter’s case, the district court ordered a new trial on December 15, 2005. The government appealed thirteen days later on January 9, 2006. After this court affirmed the new trial order, fifty-eight more days elapsed before the status conference in which the district court set a new trial date and granted a so-called “ends-of-justice” continuance until that date, as is allowed by 18 U.S.C. § 3161(h)(7)(A). From this timeline, Carpenter makes two arguments for why the government violated the Speedy Trial Act.

1. When the “Action Occasioning Retrial” Becomes Final

The parties’ first disagreement concerns when, under section 3161(e), the “action occasioning retrial becomes final” and the seventy-day clock begins to run in a case where the district court set aside a verdict and ordered a new trial, the prosecution appealed that order, and the appellate court confirmed the order. Carpenter argues that the clock began to run upon the district court’s new trial order; the government counts from the date on which the court of appeals’ mandate issued. Under Carpenter’s view, the thirteen pre-appeal days added to the fifty-eight post-appeal days exceed by one day the seventy-day limit.

Neither the Supreme Court nor our circuit has interpreted the phrase “the date the action occasioning the retrial becomes final” as applied to a district court’s retrial order that is affirmed on appeal. Here, though, we require no precedent to answer the question posed. Rather, we look to the statutory language, which we read as plainly providing that the seventy days starts when the appellate mandate affirming the district court order issues, 15 thereby rendering that order final. We read the language in this manner primarily because there is no reason to have used the term “becomes final” if the drafters actually meant the date the challenged order was entered.

Carpenter’s argument to the contrary turns on section 3161(h)(1)(C), which provides that the time for “any” interlocutory appeal is “excluded” from Speedy Trial Act calculations. Carpenter argues that because this latter provision covers “any” interlocutory appeal, Congress intended to account for an appeal after a new trial order by “excluding” the time of the appeal from the seventy days that began to run on the date of the district court order, and not by postponing the beginning of the seventy-day period until the court of appeals affirms the order.

Carpenter’s argument faces an uphill battle, given that the text of section 3161(h)(1)(C) does not address the question of when the seventy-day clock begins to run. The strongest argument in support of Carpenter’s position (albeit one that Carpenter didn’t make) is that section 3161(e) contains a provision that “[t]he periods of delay enumerated in section 3161(h) are excluded in computing the time limitations specified in this section.” If the seventy-day clock doesn’t even start running until the conclusion of the appeal, why incorporate the exclusion for interlocutory appeals under section 3161(h)(1)(C)?

The answer is that our reading of section 3161(e) does not render entirely unnecessary the need to have a tolling period for interlocutory appeals in cases where new trial orders are affirmed on appeal. It is entirely possible for an appeals court to affirm a new trial order, triggering the seventy-day clock under section 3161(e), and for a party to file a different interlocutory appeal before the seventy days expires. In that case, the seventy days would start when the retrial order became final — either when it was entered for cases that are not appealed, or when the appellate court’s mandate issued for cases that are — and would be tolled for the duration of any subsequent appeal. 16

This plain reading of section 3161(e) is supported by Congress’ use of identical language in section 3161(d)(2). That section addresses the scenario in which a district court dismisses an indictment, but an appeals court causes it to be reinstated. In such a case, the only possible candidate for the “action occasioning retrial” is the action of the appeals court. Yet section 3161(d)(2) also includes a statement that “the periods of delay enumerated in section 3161(h) are excluded.” Clearly, that incorporation of section 3161(h) in section 3161(d)(2) cannot mean that the seventy-day clock begins to run before the appeal is concluded. The wholesale incorporation of section 3161(h) into section 3161(d), then, simply covers the possibility that some other interlocutory appeal might cause some need to toll the running of the seventy days. 17 We read it to do the same in section 3161(e).

Our interpretation is also consistent with the Guidelines to the Administration of the Speedy Trial Act, As Amended, issued by a United States Judicial Conference committee. 106 F.R.D. 271, 282 (1984) (“[I]f an appeal or petition for certiorari is filed, the action occasioning the retrial should not be considered final until the appeal or petition has been disposed of.”) The government also points to legislative history tending to suggest that the “becomes final” language was added to clarify that the seventy days were triggered by the conclusion of any appeals. See Anthony Partridge, Legislative History of Title I of the Speedy Trial Act of 19?% 80-82 (Fed. Judicial Center 1980).

The circuit court opinions Carpenter cites as support for his reading are not to the contrary, despite language seeming to suggest otherwise. See United States v. Pitner; 307 F.3d 1178, 1182-83 (9th Cir.2002) (“interlocutory appeals interrupt the seventy day period; they do not start it running”) (internal quotation marks omitted); United States v. Rivera, 844 F.2d 916, 919 (2d Cir.1988). 18 (“[T]he 70-day period ... started ... when the mistrial was declared, and the speedy trial clock resumed on the date the exclusion allowed for an interlocutory appeal ... ended.”) In both cases, the actual issue concerned an interlocutory appeal of an order other than the mistrial orders that made a’new trial necessary. In Rivera, there was actually no dispute — or material issue — concerning when the seventy-day clock started to run. Rivera, 844 F.2d at 919-22. In Pitner, there was such a dispute, but we would have reached the same result under our reading of section 3161(e). There, the district court declared a mistrial because the jury was deadlocked, and the prosecution did not (and likely could not) appeal. Pit-ner, 307 F.3d at 1180. The “action occasioning the retrial” was thus the mistrial order, the finality of which was never delayed or suspended. Months later, the defendant appealed from the denial of a motion to dismiss on double jeopardy grounds. Just as we would, the Ninth Circuit held that the time taken by that interlocutory appeal was simply to be excluded in calculating the seventy days that began running when the mistrial was declared. Id. at 1182. To the extent the court explained that holding in terms that went beyond the facts presented, we disagree for the reasons stated above.

For these reasons, we have little trouble concluding that when a party appeals a district court order granting a new trial, the action occasioning the retrial becomes final when the mandate of the appellate court issues. Here, that happened when this circuit upheld the grant of a néw trial fifty-eight days before the district court set a new trial date and granted an ends-of-justice continuance for the time before trial.

2. The Sufficiency of the Ends-of-Justice Findings

Carpenter next argues that the district court’s November 28, 2007, ends-of-justice continuance, which excluded from Speedy Trial Act calculations all time between the status conference setting the trial date and the commencement of trial, was invalid because the district court did not at the time make specific findings in the record as to why it was granting the continuance, as required by 18 U.S.C. § 3161(h)(7)(A). This provision allows the district court to grant a continuance on the basis of findings that “the ends of justice served by taking such action outweigh the best interest of the public and the defendant in a speedy trial.” Id. It further provides that delay resulting in the continuance is not excludable “unless the court sets forth, in the record of the case, either orally or in writing, its reasons for finding that the ends of justice” are served in light of statutory factors defined in section 3161(h)(7)(B). 19 Id. § 3161(h)(7)(A). One of the factors is whether failure to grant a continuance would unreasonably deny the defendant the ability to obtain counsel or continuity of counsel, or would deny counsel for either party time to prepare. Id. § 3161(h)(7)(B)(iv).

The Supreme Court has emphasized the importance of on-the-record findings, albeit in the context of a case where the defendant, at the district court’s urging, waived for all time his speedy trial rights, and the government argued that the prosecution could be salvaged by granting a retroactive ends-of-justice continuance. Zedner v. United States, 547 U.S. 489, 507-08, 126 S.Ct. 1976, 164 L.Ed.2d 749 (2006). Zedner held that findings must be made “if only in the [trial] judge’s mind” by the time the continuance is granted, and must be entered in the record by the time the district court denies the motion to dismiss on STA grounds. Id. at 506-07, 126 S.Ct. 1976. Because the district court “entered” its reasons for granting a continuance into the record through the order denying the motion to dismiss, Carpenter, 542 F.Supp.2d at 183-84, as well as in the hearing on that motion, Carpenter focuses his argument on whether the district court “made” the findings at the time it granted the motion.

This argument fails. The transcript of the January 24, 2008, hearing in which the district court granted the continuance demonstrates the court carefully considered why and for how long the government and Carpenter’s counsel would be unavailable in light of countervailing considerations such as the availability of witnesses. It rejected Carpenter’s request to put off scheduling a trial until a status conference in March, and it ultimately set a trial date earlier than the one Carpenter’s counsel requested. The record is clear that the district court balanced counsels’ schedules with the public’s interest in a speedy trial, and thus its decision represents a determination that granting a continuance served the ends of justice.

C. Sufficiency of the Evidence and Jury Instructions

Carpenter next argues that the district court erred when it denied Carpenter’s motion for acquittal in which he argued that the government did not prove that he had an affirmative duty to disclose anything to the investors, which is a necessary element of a theory of fraud by omissions. This argument fails because the government did not prosecute a theory of fraud by omission: its theory was that the marketing materials and agreements contained misleading affirmative statements. 20 As the district court noted, “[ajrguing that the defendant omitted material information necessary to make the affirmative statements not misleading did not transform the case from one of affirmative misrepresentations to a case of misrepresentation solely by reason of omission.” Carpenter, 2014 WL 691659 at *5. In its 2013 decision, this court shared the district court’s view of the prosecution’s theory. See Carpenter, 736 F.3d at 623-24 (noting that the government’s theory was that the marketing materials “effectively promised” the exchangors’ funds would be kept safe). Under the theory of misleading affirmative statements, there was no need to prove the elements of a pure failure-to-disclose case.

D. Motion for New Trial Because of Newly Discovered Evidence

Carpenter next challenges the district court’s denial of his motion for a new trial based on the Merrill Lynch documents that appeared after his second trial. As described above, those documents indicated that Merrill Lynch knew that Carpenter was investing other people’s money, and thus supported Carpenter’s claim that he did not hide that fact from Merrill Lynch. All of this, he claims, would have turned the tide on his good faith defense. In its 2011 order, the district court did not buy this argument. 21 Carpenter, 808 F.Supp.2d at 379-86. We now review that decision. 22

“Ordinarily we will affirm the trial court’s denial of a new trial” based on claims of newly discovered evidence “unless the court has manifestly abused its discretion.”. United States v. Wright, 625 F.2d 1017, 1019 (1st Cir.1980). This remains the case even when the district court did not clearly articulate its reasons for denying the motion, which, because of the complexities introduced by the voluminous filings in this case, is the situation presented here. Id.; United States v. Connolly, 504 F.3d 206, 212 (1st Cir.2007). Although the standard is somewhat heightened when the government knowingly presents false testimony, the district court rejected the claim that the government did so in any way that affected the integrity of the trial, and we agreed. 23 See Carpenter, 736 F.3d at 630-31. Thus, we reject Carpenter’s argument that a heightened standard is called for in this case.

A district court may grant a motion for a new trial based on newly discovered evidence if (1) the evidence was unknown or unavailable to the defendant at the time of trial; (2) failure to learn of it was not because of lack of due diligence; (3) the evidence is material, and not merely cumulative or impeaching; and (4) it will probably result in acquittal upon retrial. Wright, 625 F.2d at 1019. Here, we need not discuss the first two prongs, because Carpenter’s claims founder on the latter two. See United States v. Hernández-Rodríguez, 443 F.3d 138, 143 (1st Cir.2006) (noting that “we have no discretion to grant a motion for a new trial if any one of the four factors is lacking”).

In the first trial, Carpenter’s attorney presented strong impeachment evidence that GL, one of Carpenter’s brokers at Merrill Lynch, was lying when he said he had never spoken with Patterson, the lawyer of one of the investors. Yet the jury returned a guilty verdict. Second and more importantly, the entire “good faith” argument was, at best, something of a bank-shot: whether or not Carpenter told Merrill Lynch that Carpenter was managing and investing the funds of his clients said very little about whether Carpenter believed the representations that Carpenter made to his clients; presumably most investment managers disclose to their brokers that the funds they invest belong to others. That hardly proves the good faith of statements made by the manager to the investors (other than a statement that the broker would know that the funds belonged to investors).

Of course, Carpenter might well have recognized that, to the extent Merrill Lynch knew third-party investors were involved, the likelihood of further inquiry by Merrill Lynch increased. So in that sense we do not suggest that the evidence would have been irrelevant. Rather, we hold only, on abuse of discretion review, that it was not so probative as to have mandated a new trial.

E. Carpenter’s Sentence

We need not tarry long on Carpenter’s argument that the district court abused its discretion with a sentence of thirty-six months’ imprisonment, well below the recommended sentence of fifty-one to sixty-three months; In justifying its downward departure, the district court stated that it wanted to avoid sentencing disparities, and presented data that First Circuit fraud sentences tend to be in the two- to three-year range. Nevertheless, Carpenter challenges both the procedural and substantive reasonableness of the sen tence. This court reviews sentencing decisions for abuse of discretion, reviewing findings of fact for clear error and its interpretations of the sentencing guidelines de novo. United States v. Flores-Machicote, 706 F.3d 16, 20 (1st Cir.2013).

Carpenter argues that his sentence was procedurally flawed because the district court considered two impermissible factors: the effect of the long proceedings on the investors, and the fact that a grand jury recently found probable cause to indict Carpenter in Connecticut. As for the first, the district court’s only reference to the effect of the long proceedings on the exchangors came in a single sentence explaining why the length of the proceedings were not a factor in Carpenter’s favor, not as a reason for imposing a sentence higher than would otherwise have been imposed. As for the second, the district court must consider a defendant’s “history and characteristics,” 18 U.S.C. § 3553(a)(1), and this circuit has stated that in doing so, it may consider brushes with the law, such as arrests, that have not resulted in convictions. See Flores-Machicote, 706 F.3d at 21. However, Carpenter correctly notes that the government does not cite to a case that directly considers the role of indictments on unrelated charges in sentencing.

For the proposition that the fact of indictment should not be considered in sentencing, Carpenter cites United States v. Williams, 22 F.3d 580, 581-82 (5th Cir.1994). In Williams, a defendant pled guilty to providing about ten grams of cocaine to an undercover agent as part of a deal that dismissed a conspiracy charge. Id. at 581. The district court sentenced the defendant based on the entire amount of drugs sold by the conspiracy, as quantified in the indictment. Id. The Fifth Circuit held that the fact of the indictment was an impermissible factor, but that this was harmless error because other evidence supported a larger role in the conspiracy. Id. at 582.

We need not delve into the issue of when an indictment may be considered, however. Even if the indictment was an impermissible factor, there is no indication that it played a “significant” role in the sentence. See United States v. Mangual-Garcia, 505 F.3d 1, 16 (1st Cir.2007) (quoting United States v. Haack, 403 F.3d 997, 1004 (8th Cir.2005)).

Carpenter next argues his sentence was substantively unreasonable because the district court failed to sufficiently consider the fact that the investors’ losses were at least partially caused by the stock market downturn in 2000. Wben a defendant challenges a within-guidelines sentence, he bears the “heavy burden” of marshaling “fairly powerful mitigating reasons and persuading] us that the district judge was unreasonable.” United States v. Madera-Ortiz, 637 F.3d 26, 30 (1st Cir.2011) (internal quotation marks omitted). Here, Carpenter challenges a sentence that is not within the recommended guidelines, it is actually below those guidelines, and he fails to meet this burden.

Under the United States Sentencing Guidelines, a “key determinant” of the sentence for fraud claims is the amount of loss, which is a proxy for the seriousness of the offense. United States v. Rostoff, 53 F.3d 398, 405 (1st Cir.1995); see U.S.S.G. § 2F1.1. Here, the sentence was calculated based on the amount the investors actually lost, which was over nine million dollars.

Carpenter is correct that when a loss is the product of multiple factors, district courts may depart from the guidelines. See United States v. Gregorio, 956 F.2d 341, 345-46 (1st Cir.1992). However, Carpenter is the reason that the investors’ money was exposed (and greatly so) to market risk in the first place. If he had promised to buy fire insurance for his clients but failed to do so, he could hardly cite the fire as a cause for mitigation when their homes burned down. The district court was not unreasonable in failing to find that the stock market downturn did not justify a downward departure. 24

F. Forfeiture Order

Because both parties now agree that the forfeiture order is not properly before this court, we do not reach this issue.

IV. Conclusion

For the foregoing reasons, we affirm the conviction and sentence.

1 . Carpenter then petitioned for certiorari from the denial of his cross-appeal of the district court’s denial of his motion for acquittal. The Supreme Court denied that petition in February 2008. Carpenter v. United States, 552 U.S. 1230, 128 S.Ct. 1443, 170 L.Ed.2d 275 (2008).

2 . The grounds in this motion were (a) that the government presented insufficient evidence to convict under 18 U.S.C. §§ 1341 and 1343, (b) that the government had again unfairly prejudiced the defendant, and (c) erroneous and prejudicial rulings on the evidence. However, the district court did not actually decide based on any of these arguments, but instead focused on various ways that the government Overstated the evidence and focused excessively on Carpenter's greed. United States v. Carpenter, 808 F.Supp.2d 366, 380 (D.Mass.2011).

3 . This exchange was the basis for Carpenter’s June 2008 motion for mistrial on the grounds that the government had knowingly offered perjured testimony in violation of Napue, 360 U.S. at 269, 79 S.Ct. 1173. The district court denied the motion and in 2013 we affirmed on the grounds that although the employee seems to have testified falsely, the government made all necessary disclosures and the defense was able to vigorously cross-examine the employee. Carpenter, 736 F.3d at 630-31.

4 . After his first trial, for example, Carpenter moved for acquittal on the grounds that the government had failed to disprove his good faith. Carpenter, 405 F.Supp.2d at 93-94.

5 . On September 3, 2013, Carpenter also filed a certiorari petition seeking review of both the district court's denial of Carpenter's motion for acquittal, and this court’s decision dismissing Carpenter's appeal for lack of jurisdiction. That petition was denied. Carpenter v. United States, - U.S. -, 134 S.Ct. 901, 187 L.Ed.2d 776 (2014).

6 . In January, he also moved for dismissal for violation of the Speedy Trial Act, renewing the same argument he made at the close of the second trial in 2008, which the district court at that time denied. United States v. Carpenter, 542 F.Supp.2d 183 (D.Mass.2008). The district court denied the motion in open court on January 28, 2014.

7 .. Carpenter also initially appealed the May 23 forfeiture order, but both parties now agree that this order will be the subject of a separate appeal.

8 . Of course, any tension is mitigated in part by the fact that even under review for abuse of discretion, an error in identifying the correct legal standard is by its nature an abuse of discretion. See Highmark Inc. v. Allcare Health Mgmt. Sys., Inc., - U.S. -, 134 S.Ct. 1744, 1748 n. 2, 188 L.Ed.2d 829 (2014) (“The abuse-of-discretion standard does not preclude an appellate court's correction of a district court’s legal or factual error: ‘A district court would necessarily abuse its discretion if it based its ruling on an erroneous view of the law or on a clearly erroneous assessment of the evidence.’ ”) (quoting Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 405, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990)).

9 . We use the unqualified word "delay” in this opinion to refer to the entire period of time between two events, recognizing that only unjustified delay that fails the Barker balancing test violates the Sixth Amendment. See Barker, 407 U.S. at 533, 92 S.Ct. 2182 (analyzing whether "the length of delay between arrest and trial” violates the Sixth Amendment).

10 . The court determined that the orders from which Carpenter appealed were not appeal-able collateral orders. As mentioned above, Carpenter's petition for certiorari challenging this determination was denied. Carpenter, 134 S.Ct. at 901.

11 . Carpenter made the new tried motion in July 2008, thirty-eight months before September 2011. However, he asks the court to focus on only the thirty-four months between the hearing on the new trial motion on December 3, 2008, and the granting of that motion on September 1, 2011.

12 . Some of the more lengthy filings include the initial sixty-seven-page motion for acquittal or new trial; a response to the government’s opposition to that motion, totaling forty-eight pages with exhibits; and a reply memorandum in support of his mistrial motion totaling seventy-six pages with exhibits.

13 . Carpenter also argues in his reply brief that the district court could not have considered the new evidence in preparation for the September 2011 order because the district court stated in the December 2008 hearing that the government’s behavior with regard to this evidence did not violate its obligations under Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963). However, Carpenter at no time suggested that he was presenting the Merrill Lynch evidence solely as a Brady claim, and in fact explicitly stated multiple times in his 2009 submissions that this information was relevant to his pending motions for retrial and acquittal.

14 . Putting a finer point on this estimate would require comparative empirical evidence not readily available and would, in any event, be irrelevant to our conclusions.

15 . An appeal ends for Speedy Trial Act purposes when the mandate issues. United States v. Rush, 738 F.2d 497, 509 (1st Cir.1984).

16 . For example, the prosecution might appeal from the granting of a motion to exclude before the new trial commences, or a defendant might appeal the denial of a double jeopardy motion before the second trial, as the defendant did in United States v. Pitner, 307 F.3d 1178, 1182-83 (9th Cir.2002).

17 .Granted, an appeal covered by section 3161(d) is a direct'appeal, and thus section 3161(h)(1)(C) does not pose precisely the same question of apparent redundancy as an interlocutory appeal of a new trial order under section 3161(e). Nevertheless, it still demonstrates how section 3161(h)(1)(C) plays an important role in the statutory framework unrelated to the question of when the clock begins to run, which suggests Congress did not intend it to obliquely define the starting point of the seventy-day clock in a retrial case under section 3161(e).

18 . Carpenter also cites to United States v. Ginyard, 572 F.Supp.2d 30, 36 (D.D.C.2008), for the sentence “an interlocutory appeal interrupts, but does not restart the running of the clock.” The issue in Ginyard — was whether an extension provision in section 3161(e) applies retroactively' — is even further removed from the question here.

19 . The factors are (i) whether failure to grant the continuance would make the continuation of the proceeding impossible or result in a miscarriage of justice, (ii) whether the case is particularly unusual or complex, (iii) whether a grand jury proceeding is unusual or corn-plex, or (iv) in a case that is not particularly unusual or complex, whether failure to grant the continuance would unreasonably deny the ability to obtain counsel, or the continuity or preparation of that counsel. 18 U.S.C. § 3161(h)(7)(B).

20 . Carpenter first raised the issue of whether the government’s theory was misrepresentation or omission as one of three issues in a July 2008 supplemental motion for acquittal. This motion was summarily denied in the September 1, 2011, order granting Carpenter a new trial. Carpenter, 808 F.Supp.2d at 386. The order did not specifically discuss whether the government had, in fact, offered a theory of fraud by omission at trial. However, in a separate motion for a new trial, he recast the same basic argument as a challenge to the indictment, arguing that the government charged him with affirmative misrepresentation, but at trial argued fraud by omission. It is in this context that the district court determined that the theory was one of affirmative misrepresentation.

21 . In the September 1, 2011, order, the district court explicitly rejected the argument that the Merrill Lynch evidence constituted grounds for a judgment of acquittal, and also did not list it as one of the grounds on which it granted a new trial. Carpenter, 808 F.Supp.2d at 378-79.

22 . The district court did not expressly consider these claims as a motion for a new trial based on newly discovered evidence, but did state in a February 10, 2014, status conference that its September 1, 2011, order was an implicit denial on those grounds. The reason for this approach is as follows: Carpenter began introducing emerging evidence from the Merrill Lynch trial in March 2009. He presented this information as relevant to his pending motions, not as the basis for a new motion. In fact, when the government argued that Carpenter was, in essence, making an argument for a new trial based on newly discovered evidence, Carpenter expressly denied that he was.

When the district court granted Carpenter's motion for a new trial on September 1, 2011, it summarily denied all other pending motions. Carpenter, 808 F.Supp.2d at 386. After we reversed the second grant of a new trial, Carpenter argued in a February 10, 2014, status conference that one of his submissions regarding Merrill Lynch — the October 28, 2009, "Memorandum in Support” of his pending motions (ECF No. 355) — was an undecided motion for a new trial. The district court instead held Carpenter to his prior position that the new information was part of his previous motions and not a distinct motion, and treated the argument as having been implicitly decided against Carpenter in the September 1, 2011 motion. It did so over the government’s contention that Carpenter had waived any argument for a new trial based on newly discovered evidence.

23 . We decline Carpenter's invitation to revisit this determination, which he argues is warranted because the new evidence even more strongly suggests that GL’s testimony was not tó be believed. The fact remains that the government presented evidence from which a jury could conclude that the questionable portion of his testimony was not to be believed.

24 . The district court was also not unreasonable for failing to depart downward on the grounds that Carpenter has already suffered for his crime.