United States Court of Appeals For the First Circuit No. 20-1400

UNITED STATES OF AMERICA,

Appellant,

v.

JOSEPH BAPTISTE,

Defendant, Appellee.

No. 20-1401

UNITED STATES OF AMERICA,

Appellant,

v.

ROGER RICHARD BONCY,

Defendant, Appellee.

APPEALS FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MASSACHUSETTS [Hon. Allison D. Burroughs, U.S. District Judge]

Before

Howard, Chief Judge, Thompson, Circuit Judge, and Arias-Marxuach, District Judge.*

* Of the District of Puerto Rico, sitting by designation. Alexia R. De Vincentis, Assistant United States Attorney, with whom Brian C. Rabbitt, Acting Assistant Attorney General, Robert A. Zink, Acting Deputy Assistant Attorney General, Jeremy R. Sanders, Appellate Counsel, Fraud Section, Criminal Division, United States Department of Justice, and Andrew E. Lelling, United States Attorney, were on brief, for appellant. Daniel N. Marx, with whom William W. Fick, Amy Barsky, and Fick & Mark LLP, were on brief, for appellee Baptiste. Jay A. Yagoda, with whom Jared E. Dwyer and Greenberg Traurig, P.A., were on brief, for appellee Boncy.

August 9, 2021 THOMPSON, Circuit Judge.

Overview

Meet Roger Boncy and Joseph Baptiste. Boncy once served

as chairman and CEO of a U.S.-based investment company called Haiti

Invest, LLC. And Baptiste once sat on that company's board of

directors. We use the past tense, because everything changed when

the feds accused them of conspiring to bribe Haitian officials

into approving an $84 million port project in that country — one

involving cement factories, a shipping-vessel repair station, an

international transshipment station, and a power plant (among

other things). Prosecutors tried them jointly. And each had their

own lawyer. We will save lots of details about the trial and its

aftermath for later. But for now it is enough to note the

following.

The government claimed (based in large part on

undercover recordings played at trial) that Baptiste and Boncy

solicited money from undercover agents (posing as investors in

Haitian infrastructure ventures), which they promised to funnel to

Haitian bureaucrats through a Baptiste-controlled nonprofit that

supposedly helped Haiti's poor — 5% of project costs would be

allocated to bribe Haitian authorities. And as a further way to

grease the project's skids, the duo — again according to the

government's theory — promised to pay off Haitian officials with

- 3 - campaign contributions, offers of future jobs, and money to fund

their favorite social programs. At the trial's end, the jury

convicted them of conspiring to violate the Foreign Corrupt

Practices Act and the Travel Act (count 1), and convicted Baptiste

(but not Boncy) of violating the Travel Act (count 2) and

conspiring to violate the Money Laundering Act (count 3).1

After firing his original attorney and hiring a new

lawyer, Baptiste moved under Criminal Rule 33 for a new trial on

the counts of conviction based on (according to the motion)

ineffective assistance of counsel under the Sixth Amendment.2

Simplified somewhat, and as relevant here: 1 the Foreign Corrupt Practices Act criminalizes bribing foreign officials, see 15 U.S.C. § 78dd-2(a); the Travel Act criminalizes traveling in "foreign commerce" with an intent to commit an "unlawful activity," see

(a)(3); and the Money Laundering Act criminalizes transferring funds from the United States to another country with the intent to bribe a foreign official, see

(a)(2)(A). 2 Rule 33 reads in full: (a) Defendant's Motion. Upon the defendant's motion, the court may vacate any judgment and grant a new trial if the interest of justice so requires. If the case was tried without a jury, the court may take additional testimony and enter a new judgment. (b) Time to File. (1) Newly Discovered Evidence. Any motion for a new trial grounded on newly discovered evidence must be filed within 3 years after the verdict or finding of guilty. If an appeal is pending, the court may not grant a motion - 4 - Likewise invoking Criminal Rule 33, Boncy asked for a separate new

trial on the count of conviction because (the motion argued)

Baptiste's lawyer's "ineffective[ness]" influenced how the jury

"view[ed] . . . both defendants" and so impaired his (Boncy's)

Fifth Amendment "due process right" to a "fair" proceeding. The

government opposed both motions.

Following an evidentiary hearing, the district judge

found that Baptiste had shown deficient performance of counsel and

that the cumulative effect of counsel's deficiencies caused him

(Baptiste) prejudice. Not only that, but the judge also found

Baptiste's attorney's shortcomings prejudiced Boncy by (among

other things) requiring "Boncy's counsel . . . to play an outsized

role at trial rather than pursue his preferred defense strategy."

And noting that a joint trial of alleged coconspirators is

presumptively appropriate and that "severance [was] not

warranted," the judge ordered a joint retrial in the interest of

"justice" because neither defendant got "a fair" first trial — the

for a new trial until the appellate court remands the case. (2) Other Grounds. Any motion for a new trial grounded on any reason other than newly discovered evidence must be filed within 14 days after the verdict or finding of guilty. See Fed. R. Crim. P. 33 (emphasis added). - 5 - significance of the "justice" buzzword (pulled from Rule 33) will

be apparent later.

From that decision, the government now appeals. After

setting out the guiding legal principles, we turn directly to the

issues that confront us — adding additional details necessary to

put matters into workable perspective. When all is said and done,

we affirm.

Guiding Legal Principles

Judges can grant a new trial if required in "the interest

of justice," see Fed. R. Crim. P. 33(a) — though they should grant

these motions only "sparingly" and to prevent "a perceived

miscarriage of justice," see United States v. Veloz,

(1st Cir. 2020) (quoting United States v. Gramins,

(2d Cir. 2019)). Applying abuse-of-discretion review,

United States v. Gonzalez,

(1st Cir.), cert.

denied,

(2020), we can affirm a judge's new-trial

decision even if "there was sufficient evidence to convict," United

States v. Rothrock,

(1st Cir. 1986). This review

standard is multifaceted, requiring us to inspect "fact findings

for clear error, legal issues de novo (in nonlegalese, with fresh

eyes), and judgment calls with some deference." United States v.

McCullock,

(1st Cir. 2021). Showing an abuse of

discretion is especially difficult when, "as here, the judge who

- 6 - hear[d]" the new-trial motions "is the same judge who presided

over the trial," because in that scenario, "substantial deference

is due to the judge's perceptions." See Gonzalez,

(emphasis added). And we ultimately will reverse "only when left

with a definite conviction that 'no reasonable person could agree

with the judge's decision,'" see McCullock,

(quoting United States v. Cruz-Ramos,

(1st Cir.

2021)) — a rule that stops us from switching our discretion for

the judge's, see Rothrock,

. But at the same

time (and as the government is quick to note), a material error of

law is never discretionary and so always is an abuse of discretion.

See Gonzalez,

.

To grant a new trial on an ineffective-assistance claim,

a judge must find that counsel performed objectively unreasonably

and that prejudice followed. See, e.g., United States v. Silvia,

(1st Cir. 2020). Deficient performance "requires

showing that counsel made errors so serious that [he] was not

functioning as the 'counsel' guaranteed the defendant by the Sixth

Amendment." Strickland v. Washington,

(1984).

And deficient performance prejudices the defense when it is

reasonably probable "that, but for counsel's unprofessional

errors, the result of the proceeding would have been different" —

i.e., "a probability sufficient to undermine confidence" in the

- 7 - result.

. The probability "of a different result must

be substantial, not just conceivable." Harrington v. Richter,

(2011). But that does not require a showing that

counsel's actions "more likely than not altered the outcome."

Strickland,

(emphasis added); see also Williams v.

Taylor,

(2000) (stating that this standard is

a lesser showing than a preponderance of the evidence). And when

assessing prejudice, a judge "must consider the totality of the

evidence before the . . . jury," because "a verdict . . . only

weakly supported by the record is more likely to have been affected

by errors than one with overwhelming record support," Strickland,

-96 — though we judges must never forget that the

purpose of the prejudice prong is to ensure a defendant has not

suffered a fundamentally unfair or unreliable outcome, see

.

Arguments and Analysis

The government does not quarrel with the judge's

deficient-performance findings — findings premised on a long list

of failures on Baptiste's lawyer's part (we hit the highlights,

quoting from the judge's decision):

• He could not "open discovery produced by the [g]overnment."

• He "did not provide copies of documents or audio and video

recordings to . . . Baptiste, nor did they ever sit down

- 8 - together to review all of the materials that the [g]overnment

had provided."

• He did not "'thoroughly review' certain documents."

• "[H]e [did] not investigate[]" the case "sufficiently to

understand the import" of the government's evidence or to

craft an appropriate response.

• He did not get English translations of Haitian-Creole

recordings, even after learning about "potential errors" in

one of the government's translations.

• He "did not subpoena any witness" or "formulate his own list

of potential witnesses in support of . . . Baptiste's

potential defenses."

• "[H]e did not . . . identify or contact any expert witnesses

that could have provided evidence on Haitian law or business

practices."

• He "continued to pursue an entrapment defense," even though

"others had previously told him that the defense was not

available to . . . Baptiste on the facts of the case" — a

mistake that essentially put Baptiste in the thick of the

conspiracy. 3

3Consider this snippet from Baptiste's lawyer's opening statement: - 9 - [T]he FBI undercover agents . . . were predisposed, predisposed to thinking that the only way to invest successfully in Haiti was to commit a bribe. The FBI agents fixated on the concept of pay-to-play [and] sought to manufacture a crime and criminals out of men who had impeccably, impeccably clean records. Remember, the FBI agents solicited . . . Baptiste and . . . Boncy to travel to Boston, and they came here on their own dime. They solicited them to travel to Boston to receive the FBI's pitch for them to engage in criminal activity. In short, the FBI did not care about benefiting Haiti or having a profitable investment. What the FBI hoped for, what they had hoped for was for corruption to occur. Conversely, . . . Baptiste and . . . Boncy sought to operate legally while bringing foreign direct investment into Haiti that would benefit the people of Haiti. What you have going on in all of this fake investment talk is the undercover agents repeatedly asking whether Haiti is pay-to- play, a pay-to-play country. . . . While you are carefully listening to the government's evidence, ask yourself how many times does the government need to be told by an American citizen that he will not commit a crime before the government leaves him alone and goes away but continues to attempt to manufacture a new criminal and a new crime. Testifying at the hearing on the new-trial motions, Baptiste's attorney admitted raising the entrapment theory even though other "people" told him he "didn't have the elements to establish that defense. - 10 - • He "only cross-examined two of the [g]overnment's six

witnesses, none of whom [he] had contacted or sought to

interview prior to trial."

• He "elicited damaging testimony" from the two he did cross.

• And he deferred to Boncy's lawyer on the "cross-examinations

of the remaining witnesses," even though Boncy's "trial

strategy was to portray . . . Baptiste as the primary driver

of the alleged conspiracy" — a conspiracy that Boncy's

attorney insisted Boncy was not a part of.

But while the government makes no argument against the deficient-

performance finding, the arguments it does present do not persuade

us to reverse.

A.

The government starts off by insisting that "[t]he

evidence of Baptiste's and Boncy's guilt was overwhelming,"

claiming that "[i]n call after recorded call" they "agreed to bribe

Haitian officials" to grease the skids for the project. And

according to the government, the evidence underpinning "Baptiste's

convictions for violating the Travel Act and conspiring to commit

money laundering was no less compelling." Yet the judge never

"addressed the weight of this evidence" — or so the government

- 11 - continues, with a nod to Strickland — and thus "erred as a matter

of law." We see several problems with this argument, however.

First off, the government's thesis suggests that

sufficiently strong evidence can sink any ineffective-assistance

claim. No one denies that the strength of the prosecution's case

is a factor in the prejudice analysis. See Turner v. United

States,

(1st Cir. 2012). But it is not the be-

all and end-all, for (after all) the chief "focus" remains "on the

'fundamental fairness of the proceeding.'" See Dugas v. Coplan,

(1st Cir. 2007) (emphasis added and quoting

Strickland,

). To put it in slightly different

terms, the prejudice probe is not designed to be applied

"mechanical[ly]" — because "when a court is evaluating an

ineffective-assistance claim, the ultimate inquiry must

concentrate on 'the fundamental fairness of the proceeding.'"

Weaver v. Massachusetts,

(2017) (quoting

Strickland,

). And while the government is

convinced of Baptiste's and Boncy's guilt based on its own —

basically unchallenged — evidence, "we have never intimated that

the right to counsel is conditioned upon actual innocence," because

actually "[t]he constitutional rights of criminal defendants are

granted to the innocent and the guilty alike." See Kimmelman v.

Morrison,

(1986). "The prejudice essential to

- 12 - a violation of the Sixth Amendment right to effective assistance

of counsel is not being convicted though one is innocent, although

that is the worst kind," we explained recently, echoing a sibling

circuit — "it is being convicted when one would have been

acquitted, or at least would have had a good shot at acquittal,

had one been competently represented." United States v. Mercedes-

De La Cruz,

, 67 n.6 (1st Cir. 2015) (quotations omitted

and quoting Owens v. United States,

(7th Cir.

2004) (per Posner, J.)). And in the context of a new-trial motion,

courts recognize that Rule 33's standard "is the interest of

justice" — a standard that "comprehends the interests of the law-

abiding as well as those of possibly guilty defendants." See

United States v. Morales,

(7th Cir. 1990) (per

Posner, J.).

Quoting Strickland, the government's papers below also

spent a lot of time emphasizing how the judge had to — but did not

— "consider the 'totality of the evidence' at trial because errors

are less likely to create prejudice when a verdict has

'overwhelming record support.'" But with the government's having

put this all front and center before the judge, we think it fair

to infer that she considered and rejected the government's points

— rather than ignoring them. See generally United States v.

Jimenez-Beltre,

(1st Cir. 2006) (en banc) (noting

- 13 - that we can conclude that the judge analyzed a defendant's claim

"by comparing what was argued by the parties . . . with what the

judge did"). And as for the government's "overwhelming evidence"

refrain, we need only point out again that a judge can grant a new

trial under Rule 33 to prevent "a miscarriage of justice" — even

if "there [is] sufficient evidence to convict." See Rothrock,

.

On top of all that, we "presume" — per Supreme Court

directive — that judges "know" and correctly "apply" the law "in

making their decisions." Walton v. Arizona,

(1990), overruled on other grounds by Ring v. Arizona,

(2002). So we presume the judge considered the totality of

the evidence as required, particularly since the government offers

no persuasive reason not to apply that presumption here.

B.

The government next faults the judge for not

"consider[ing]" Baptiste's counsel's failings "in light of Boncy's

complementary defense." Boncy's lawyer, the government writes,

"took the lead on cross-examining witnesses and, through his

skilled questioning, advanced an argument" that no "conspiratorial

agreement" ever existed — which matters because Baptiste could not

- 14 - be convicted of any conspiracy without a coconspirator. The

government, however, cannot win under this theory either.

Hurting the government here is that other supposed

coconspirators existed, albeit unindicted ones (as even the

government concedes). Putting that point aside does the government

no good, however. Even granting that Boncy's lawyer tried on

cross-examination to play up the no-conspiratorial-agreement

angle, the government is still stuck with the judge's finding —

based mostly on Baptiste's counsel's evidentiary-hearing testimony

— that the two defense teams "did not 'coordinate' . . . or see

'eye to eye'" on "their defense strategy." And the judge so found

because Boncy's attorney had "'his direction'" and Baptiste's

attorney had his own "'in terms of how [they] were proceeding with

the trial'" (the internal quotations are from Baptiste's counsel's

testimony at the new-trial motion hearing, by the way).

This finding does not rest on thin air either — for the

record shows that Boncy's counsel spent much energy trying make

Baptiste the real culprit in this conspiracy while distancing Boncy

from Baptiste's deception. Boncy's lawyer, for example,

emphasized recordings in which Baptiste — not Boncy — was "[t]he

person who's talking about . . . how" bribery is "done in Haiti."

And Boncy's attorney suggested that Baptiste — not Boncy —"raised"

the bribery idea by demanding that "something had to be put on the

- 15 - table" for bribes (that "something" according to the government,

being payola). More, Boncy's counsel accused Baptiste of "chiding"

Boncy about Boncy's "unwillingness to put money on the table" for

bribes. More still, Boncy's lawyer elicited agent testimony that

Baptiste and Boncy said "two different things" about the "pay-

offs" to Haitian officials — Baptiste, according to the agent, was

talking about "pay-offs" while Boncy was not. Boncy's counsel did

all this while also painting Boncy (himself a lawyer) as "very

strict" on compliance issues compared to Baptiste. And more still,

Boncy's lawyer alternatively suggested that Baptiste was trying to

"hustle[]" some money from the feds — that Baptiste wanted "5

percent" built into the project cost (a key part of the

government's conspiracy theory; recall our "overview" remarks at

the opinion's beginning) simply to enrich himself rather than to

buy off Haitian politicos. Boncy's attorney returned to this theme

during closing arguments, calling Baptiste a liar who told the

undercover agents what they "wanted to hear . . . to keep the money

flowing" and adding that if Baptiste "was scamming the FBI, it's

reasonable to conclude that he was scamming . . . other so-called

co-conspirators," including "Boncy."

The bottom line is that despite what the government

thinks, Boncy's lawyer was hardly defending Baptiste. So this

- 16 - facet of the government's reversal argument does not assist its

cause.

C.

The government does not like how the judge used the

cumulative-error doctrine. To hear it tell it, that doctrine

simply provides "that there need not be any singularly dispositive

deficiency." But it believes that the judge failed to find

"concrete ways" that any of "counsel's deficiencies . . .

contribute[d] to a substantial likelihood of a different result,"

and because "none would, the cumulative error doctrine" is not in

play. We, however, see things differently.

The cumulative-error doctrine holds that errors not

individually reversible can become so cumulatively. See, e.g.,

United States v. Sepulveda,

(1st Cir. 1993).

That is because "[i]ndividual errors, insufficient in themselves

to necessitate a new trial, may in the aggregate have a more

debilitating effect" and thus add up to prejudice. See

(double

emphasis added). And we have long held that the prejudice inquiry

under Strickland can be a cumulative one as to the effect of all

of counsel's slipups that satisfy the deficient-performance prong

— meaning that a defendant need show it more likely than not that

the several blunders, even if not prejudicial on their own,

prejudiced him when taken together. See Dugas v. Coplan, 428 F.3d

- 17 - 317, 335 (1st Cir. 2005) (proclaiming that "Strickland clearly

allows the court to consider the cumulative effect of counsel's

errors in determining whether a defendant was prejudiced" (quoting

Kubat v. Thieret,

(7th Cir. 1989))).

The government makes much of language in United States

v. Sampson that "the cumulative error doctrine finds no foothold"

if no individual error "worked any cognizable harm to [the

defendant's] rights." See

(1st Cir. 2007). But

we have read that passage as saying that if the combined errors

resulted in an unfair trial, the cumulative-error doctrine

"requires the vacation of a defendant's conviction even though the

same compendium of errors, considered one by one, would not justify

such relief." See United States v. Padilla-Galarza,

(1st Cir. 2021) (double emphasis added, citing Sampson,

).

So what matters is the cumulative effect of counsel's

errors (even if no error in isolation suffices to establish

qualifying prejudice) — i.e., the focus must be on the collective

impact of counsel's deficiencies, some of which we have mentioned:

• His raising an entrapment theory, despite being told it had

no legs — a blooper that essentially put Baptiste in the

middle of a bribery scheme.

- 18 - • His not challenging a photo of cash sent from Baptiste's phone

(a picture counsel first saw at trial, despite getting it

from prosecutors in discovery), which the government argued

showed money Baptiste used for bribes — even though Baptiste

told counsel during the trial that the money represented

Christmas bonuses for Baptiste's staff, not bribes for

Haitian functionaries.

• His not effectively contesting agent testimony labeling

certain words by Baptiste (captured on tape) code for bribes

(e.g., "tips," "tak[ing] care of . . . people on the ground,"

"leeway," "unforeseen expenses," "something," "anything,"

"stuff on the table," "social programs," and "Christmas

bonuses") — a failure that tacitly endorsed the idea that the

agents were not simply speculating about what the words meant.

• And his not calling any witnesses, including a 20-year

acquaintance of Baptiste and a Haitian legal expert — the

first of whom (according to an affidavit filed with Baptiste's

new-trial motion) would have testified that Baptiste had an

excellent reputation with Haitian officials and so had no

need to bribe his way into their circle, and the second of

whom (according to another affidavit) would have testified

that certain actions (making political contributions,

- 19 - lobbying officials by promising them to fund social programs

to help local communities, etc.) were legal in Haiti.

The government believes the judge should have found counsel's

inadequacies "harmless." But the government attacks each gaffe

item by item, without providing any convincing basis to question

the judge's conclusion that the defects as a group cast a shadow

on the verdict's integrity — which dooms its argument.

In the same section of its opening brief, the government

calls the judge's prejudice finding too "hypothetical" because, at

least in its view, the record lacks a sufficient "description" of

the tack the defense might take at a retrial and so prevents us

from doing a prejudice analysis on appeal. But our discussion of

the deficiencies bulleted throughout the opinion (in the beginning

of the "arguments and analysis" section and in preceding paragraph)

sheds sufficient light on what the defense could do differently at

a do-over trial.

D.

Last but not least is the government's claim that the

judge reversibly erred by finding that Baptiste's lawyer's

ineffectiveness prejudiced Boncy. This claim has three subparts.

The first is the government's contention "that the Sixth Amendment

right to counsel is personal and cannot be asserted vicariously"

by Boncy (which is the government's "principal argument"). The

- 20 - second is the government's insistence that "[e]ven if" we "construe

Boncy's claim as an argument that he was prejudiced by joinder of

the trials due to the presentation of conflicting defense

strategies," his claim would still "fail" because the judge "found

no such prejudice here" when she ordered them retried together.

And the third is the government's suggestion that "the prejudice

the [judge] did find" has no "factual basis." Nothing the

government says, however, moves the needle in its favor.

Before getting into the nitty-gritty, it is worth going

over some of the new-trial basics again — while also amplifying

some of our earlier points.

Rule 33 lets a defendant file a new-trial motion "based

on newly discovered evidence" (Rule 33(b)(1)) or "on any reason

other than newly discovered evidence" (Rule 33(b)(2), emphasis

ours). Rule 33 also says a judge "may" grant a new trial in "the

interest of justice" (Rule 33(a), emphasis again ours).

The "interest of justice" catchphrase is undefined. But

according to a leading treatise on criminal procedure, Rule 33's

words vest a judge with "broad powers to grant a new trial" if she

"concludes for any reason that the trial resulted in a miscarriage

of justice" (with the caveat that a new-trial motion is "to be

granted with caution"). 3 Sarah N. Welling, Federal Practice and

Procedure § 581 (4th ed. 2021) [hereinafter Federal Practice and

- 21 - Procedure] (emphasis added). See generally Veloz,

(using the "miscarriage of justice" idiom). We ourselves have

stressed the "broad discretion" Rule 33 gives a judge — contrasting

(in a way that provides some insight into what this means) the

spacious "interest[] of justice" lingo in Rule 33 with a federal

habeas statute that lets a judge grant postconviction relief if

(for example) she "finds . . . that there has been . . . a denial

. . . of constitutional rights." Trenkler v. United States,

(1st Cir. 2001) (quotations omitted). Or to steal a

line from the same leading treatise, "[t]he grounds for relief are

broader under Rule 33 than under [the habeas statute]," because

"[t]he Rule allows a new trial whenever the interest of justice

requires it, while the statutory [habeas] remedy . . . for the

most part reaches only constitutional defects in the proceedings."

3 Federal Practice and Procedure § 591. Which is why Rule 33 is

"likely more enticing to a" defendant than the statute. Trenkler,

(quotations omitted). And a somewhat related line

of cases elsewhere — not challenged by the government here —

recognizes that the "interest of justice" empowers a judge to grant

a new trial based on perceived unfairness of something not

amounting to reversible error. See United States v. Vicaria,

(11th Cir. 1994) (holding that the district judge

had the discretion to grant a new trial after "conclud[ing]" that

- 22 - he "should have given [a jury] instruction [he] had declined [to

give]" — even though the judge did not legally err in not giving

that instruction); Morales,

(noting that if a

district judge believes "there is a serious danger that a

miscarriage of justice has occurred," she can grant a new trial —

"even if [she] does not think that [she] made any erroneous rulings

at the trial" (citing our Rothrock opinion,

)).

A judge's handling of a Rule 33 motion "is ordinarily a

'judgment call.'" United States v. Connolly,

(1st Cir. 2007) (quoting United States v. Maldonado-Rivera,

(1st Cir. 2007)). So when the judge deciding the

motion is the judge who ran the trial, we owe "an appreciable

measure of respect . . . to '[her] sense of the ebb and flow of

the recently concluded'" proceedings.

(quoting United States

v. Natanel,

(1st Cir. 1991)); see also United

States v. Paniagua-Ramos,

(1st Cir. 1998)

(commenting that our review "is deferential to the district court"

because "the court was present at the trial and had the opportunity

first hand to observe the evidence, the witnesses, and the jury").

What this means here is that the government must meet the high

threshold of showing the judge abused her discretion. See

Connolly,

; see also United States v. Rodríguez-

- 23 - Soler,

(1st Cir. 2014) (calling abuse of

discretion a "difficult standard[]").

With all that in mind — and knowing too that our job is

to focus on the arguments that an appellant like the government

actually makes, not to think up other ones besides, see Cruz-

Ramos,

; Rivera-Corraliza v. Morales,

(1st Cir. 2015) — we turn to our analysis of the government's

multidimensional thesis.

Contrary to the government's position, Boncy did not

pursue a Sixth Amendment claim keyed to Baptiste's lawyer's

ineffective assistance. Instead, he pursued a Fifth Amendment

claim tied to the denial of his own due-process rights — to quote

again from his motion (emphasis ours, though), he argued that

"deficiencies in the conduct of Baptiste's trial counsel

necessarily affected the jury's view of both defendants in this

joint trial, thereby depriving Boncy of a fair trial in violation

of his due process right to a fair trial." So we need not worry

about the government's Sixth Amendment-based first sub-argument.

Nor need we spend time on the government's second sub-

argument. And that is because we accept (without further ado) the

government's claim that the judge's prejudice analysis did not

turn on Boncy's being tried jointly with Baptiste, given how she

(as the government notes) found severance unnecessary.

- 24 - Which leaves us with the government's final sub-

argument, i.e., that the judge's prejudice finding lacks

evidentiary support — this, despite the fact that the judge cited

to the record lots of times. Anyway, the government's initial

brief does not tie its attack to the controlling clearly-erroneous

standard — a hard-to-satisfy test, seeing how a "challenger must

show that the contested finding stinks like 'a 5 week old,

unrefrigerated, dead fish.'" See United States v. Rivera-

Carrasquillo,

(1st Cir. 2019) (quoting Toye v.

O'Donnell (In re O'Donnell),

(1st Cir. 2013)),

cert. denied,

(2020); see also United States v.

Oliveira,

(1st Cir. 2018) (stating that the

challenger's arguments must cause us to form "a strong, unyielding

belief that a mistake has been made" (quotation omitted)); United

States v. Cates,

(1st Cir. 2018) (highlighting

how clear error's "heights are difficult to scale"). And it should

now go without saying (though we say it anyway) that an appellant

risks waiver if its opening brief does not properly develop

arguments showing how its claims can succeed under the proper

review standard. See Cruz-Ramos,

(collecting

authority); see also Rivera-Carrasquillo,

.

But even brushing waiver away, we cannot stamp the

challenged findings clearly erroneous. As noted several pages

- 25 - ago, the judge found (again, with specific citations to the record)

that Boncy's lawyer wanted to expose Baptiste to the jury "as the

primary driver" of the charged conspiracy to bribe foreign

officials — a strategy, to quote the government's lower-court

papers, that "Boncy's counsel" pursued as he basically "led the

defense side of the case," almost from the start. And "[g]iven

that . . . Boncy played a different role in the charged conspiracy

than . . . Baptiste" — to quote some more from the judge's order

— "his counsel might well have decided to say and do considerably

less during trial to emphasize [Boncy's] more minor participation

in the alleged conspiracy had he not had to carry the defense so

much on his own." Looking to deflect the force of the judge's

finding, the government says that Boncy's lawyer also sometimes

portrayed Baptiste as running an uncharged conspiracy to rip-off

the FBI (as we intimated above in section B, Boncy's counsel

occasionally theorized that "the 5 percent" was Baptiste's self-

serving scam to line his pockets — and his alone). Still, no one

can seriously doubt that Boncy's counsel worked to show that the

charged conspiracy did not exist. Yet, Baptiste's lawyer basically

conceded the charged conspiracy's existence by hyping an

entrapment theory (and to be perfectly clear: by charged

conspiracy, we mean the conspiracy involving Baptiste and Boncy).

Also conflicting with his bid to show his client had no part in

- 26 - the charged conspiracy, Boncy's counsel — as the judge supportably

found — had "to play an outsized role at trial rather than pursue

his preferred defense strategy for his own client." Which explains

why the government's arguments do not provoke a strong, unyielding

belief that the judge botched the prejudice finding. And to the

extent the government thinks a plausible view of the facts supports

its no-prejudice position, we need only say that a judge's choice

among supportable alternatives cannot be clearly erroneous. See

Cooper v. Harris,

(2017) (emphasizing that

"[a] finding that is 'plausible' in light of the full record —

even if another is equally or more so — must govern" (emphasis

added and citation omitted)).

The district judge (not we) actually heard the

witnesses, saw how each counsel performed, and watched the jurors

as the proceedings unfolded. That put her (not us) in the best

spot to decide if the interest of justice demanded a new trial.

See Connolly,

. For, to borrow the words of Judge

Posner, "[t]he trial judge will always be in a better position

than the appellate judges to assess the probable reactions of

jurors in a case over which [she] has presided," because she can

understand, "as we cannot, . . . the atmosphere of the trial —

that congeries of intangibles that no stenographic transcript can

convey." See United States v. Bruscino,

(7th

- 27 - Cir. 1982) (en banc) (per Posner, J.). Perhaps if we had been

there we would have decided matters differently. Or perhaps not.

See generally Glossip v. Gross,

(2015)

(instructing us not to "overturn a finding 'simply because [we

are] convinced that [we] would have decided the case differently'"

(alteration in original and quoting Anderson v. City of Bessemer

City,

(1985))). But because our job is not to

play district judge, we cannot substitute our judgment for her

discretion and ring-side insights — unless those insights sunk to

an abuse of discretion. See Veloz,

. And as with

its other claims, this multilayered argument does not satisfy the

government's heavy burden of showing an abuse of discretion.

We end this segment by again emphasizing a fundamental

point. "The premise of our adversarial system is that appellate

courts do not sit as self-directed boards of legal inquiry and

research, but essentially as arbiters of legal questions presented

and argued by the parties before them." See NASA v. Nelson,

, 147 n.10 (2011) (quoting Carducci v. Regan,

(D.C. Cir. 1983) (per Scalia, J.)); accord Greenlaw v.

United States,

(2008) (stating that "we rely on

the parties to frame the issues for decision and assign to courts

the role of neutral arbiter of matters the parties present"). So

in doing what judges are paid to do, we have taken — and, it turns

- 28 - out, rejected — the government's arguments as we have found them.

Maybe there is a better argument for why a new trial would not

serve "the interest of justice" — e.g., maybe there is some other

limiting principle to that legal standard's broad reach. But if

a better argument does exist, we need not deal with it today: it

is enough for us to hold (as we do) that the arguments the

government does make do not convince us that the reach of this

broad standard should be circumscribed in this instance.

Wrap Up

No one doubts (at least, no one should doubt) that "[t]he better

the lawyers at a trial are, provided they are evenly matched, the more likely

is the trier of fact to find the truth." See Morales,

. So

while both sides here face the expense of retrial, "[t]he result will be a .

. . proceeding much more likely to render a verdict in which the legal system

and the public can have confidence."4 See

And thus, for the reasons recorded above, we affirm the judge's

grant of new trials.

4 Given our holding, we need not consider Baptiste's and Boncy's alternative grounds for affirmance — e.g., Baptiste's argument that a judge can grant a new trial even if counsel's performance did not fall into the category of a Sixth Amendment violation; or Boncy's argument that he should get a new trial based on the government's destruction of certain evidence. See generally PDK Labs. Inc. v. DEA,

(D.C. Cir. 2004) (Roberts, J., concurring in part and concurring in the judgment) (declaring that "if it is not necessary to decide more, it is necessary not to decide more"). - 29 -