United States v. Paul Bergrin

PRECEDENTIAL

UNITED STATES COURT OF APPEALS

FOR THE THIRD CIRCUIT

_____________

Nos. 11-4300/4552

_____________

UNITED STATES OF AMERICA,

Appellant

v.

PAUL W. BERGRIN

_______________

On Appeal from the United States District Court

for the District of New Jersey

(D.C. No. 09-cr-369)

District Judge: Hon. William J. Martini

_______________

Argued

March 29, 2012

Before: FUENTES, SMITH, and JORDAN, Circuit Judges.

(Filed: June 15, 2012)

_______________

Mark E. Coyne

Steven G. Sanders [ARGUED]

Office of United States Attorney

970 Broad Street – Rm. 700

Newark, NJ 07102

Counsel for Appellant

Lawrence S. Lustberg [ARGUED]

Gibbons

One Gateway Center

Newark, NJ 07102

Counsel for Appellee

_______________

OPINION OF THE COURT

_______________

JORDAN, Circuit Judge.

Paul Bergrin, a former federal prosecutor and

prominent defense attorney, was indicted in the United States

District Court for the District of New Jersey on numerous

charges, including violations of the Racketeering Influenced

and Corrupt Organizations Act (“RICO”). Reasoning that the

RICO charges were inappropriate in light of “the disparate

nature of the substantive crimes that … serve[d] as the

racketeering predicates,” the District Court dismissed them.

United States v. Bergrin, 707 F. Supp. 2d 503, 511 (D.N.J.

2010). The government appealed that decision and we

reversed, observing that the concerns of the District Court

were “either endemic to RICO prosecutions or involve[d] the

application of irrelevant legal standards.” United States v.

Bergrin, 650 F.3d 257, 274 (3d Cir. 2011).

2

After remand, the government filed a 33-count second

superseding indictment (the “Indictment”) charging Bergrin

with RICO violations, witness tampering, participating in a

cocaine-trafficking conspiracy, and tax evasion. Two of the

Indictment’s witness-tampering counts charge Bergrin for his

role in facilitating the murder of a man named Kemo McCray

(“Kemo”), 1 who was to have been a witness against one of

Bergrin’s clients. 2 The District Court ordered those counts

(the “Kemo Murder Counts”) to be severed and tried first and

separately from the rest of the crimes charged. At the ensuing

trial, the Court precluded the government from introducing

evidence of two other witness-murder plots to prove

Bergrin’s intent to have Kemo murdered, and the jury was

ultimately unable to reach a verdict.

As soon as the jury was dismissed, the government, in

anticipation of a retrial, asked whether the District Court

would adhere to its earlier evidentiary rulings. “Absolutely,”

was the response, though the Court noted that the government

would be permitted to try to “convince [the Court]

otherwise.” (Joint App. at 49.) The government now appeals

those evidentiary rulings and also asks us to review an

1

For ease of reference, we will refer to Mr. McCray by

his first name, intending no undue familiarity or disrespect.

2

Specifically, Count 12 charges Bergrin with

conspiring to murder Kemo to prevent his testimony in

violation of 18 U.S.C. § 1512(k), and Count 13 charges that

Bergrin “knowingly and intentionally … counsel[ed], and

induce[d] others to kill” Kemo with “malice aforethought and

with intent to prevent” his testimony in violation of 18 U.S.C.

§ 1512(a)(1)(A). (Joint App. at 199.)

3

additional severance order that the Court entered. 3 In

addition, the government urges that the case be reassigned to

a new judge, contending that a reasonable person would

conclude that the District Court’s impartiality might

reasonably be questioned.

We will vacate the District Court’s decision with

respect to one of the challenged evidentiary rulings, and,

because we will direct the Chief Judge of the District Court to

reassign this matter, will leave the other issues presented to be

considered afresh by the judge who will take up the case.

I. Factual Background and Procedural History

A. Facts

Centered around RICO counts that are substantially

similar to the ones we held to be validly pleaded the last time

this case was before us, see Bergrin, 650 F.3d at 261-63

(summarizing the RICO charges), the Indictment accuses

Bergrin of misusing his law practice to traffic drugs, facilitate

prostitution, tamper with witnesses, and evade taxes. Three

different instances of witness tampering, all of which are

alleged in the RICO violation charged in Count 1, are relevant

to this appeal. Specifically, Bergrin is charged with

3

After the government took its appeal with respect to

the evidentiary decisions pertaining to the Kemo Murder

Counts, the Court severed the majority of the Indictment’s

remaining substantive counts and ordered that they be tried

before the rest of the charges. The government appealed that

ruling, too, see infra note 20, and we consolidated the

government’s two appeals for disposition.

4

instigating Kemo’s murder, plotting to kill witnesses in

connection with the legal defense of an individual named

Vicente Esteves (the “Esteves Plot”), and plotting to kill a

witness who planned to testify against a client named Richard

Pozo (the “Pozo Plot”). 4 Counts 2 through 4 of the

Indictment also plead RICO violations relating to some or all

of those three instances of witness tampering, 5 while the

Indictment’s remaining counts charge Bergrin with other

substantive or conspiracy offenses that rest on many of the

allegations set forth in the RICO counts.

1. The Kemo Murder

4

Although the three witness-tampering plots are all

alleged in Count 1, only the Kemo murder and the Esteves

Plot are charged as predicate racketeering acts. The Pozo

Plot, by contrast, is listed as one of the “methods and means”

through which Bergrin’s firm engaged in racketeering.

5

Count 2 charges Bergrin with participating in a RICO

conspiracy and alleges that the Kemo murder, the Esteves

Plot, and the Pozo Plot were overt acts in furtherance of the

conspiracy. Counts 3 and 4 charge violent crimes in aid of

racketeering offenses for Bergrin’s involvement in the Kemo

murder and the Esteves Plot, respectively. See 18 U.S.C.

§ 1959(a) (providing for criminal sanction where “[one]

murders, kidnaps, maims, assaults with a dangerous weapon,

commits assault resulting in serious bodily injury upon, or

threatens to commit a crime of violence against any

individual in violation of the laws of any State or the United

States, or attempts or conspires so to do” in connection with a

racketeering activity).

5

The Kemo Murder Counts were the subject of the trial

that ultimately led to the present appeal, and, as charged, they

carry a mandatory life sentence. 6 See 18 U.S.C.

§ 1512(a)(3)(A) (tampering with a witness by killing is

punishable as “provided in sections 1111 and 1112”); id. §

1512(k) (“Whoever conspires to commit any offense under

this section shall be subject to the same penalties as those

prescribed for the offense the commission of which was the

object of the conspiracy.”); id. § 1111(b) (“Whoever is guilty

of murder in the first degree shall be punished by death or by

imprisonment for life.”).

At the trial on those counts, the government introduced

evidence that Kemo’s murder arose out of Bergrin’s

representation of William Baskerville. Baskerville was an

associate in a drug-trafficking organization run by Hakeem

Curry and was arrested on federal drug charges in November

2003 for drug sales he made to Kemo. Baskerville told

Bergrin that he suspected Kemo to be the likely source of the

government’s evidence against him. Bergrin, in turn,

telephoned Curry and told him that Kemo was the

confidential witness against Baskerville.

6

The violent crimes in aid of racketeering offense

pertaining to the Kemo murder, see supra note 5, also carries

a mandatory life sentence, see 18 U.S.C. § 1959(a)(1) (violent

crimes in aid of racketeering that result in murder are

punished “by death or life imprisonment, or a fine …, or

both”); United States v. Carson, 455 F.3d 336, 385 n.44 (D.C.

Cir. 2006) (reaching the “common sense conclusion” that,

despite the language employed, the violent crimes in aid of

racketeering statute “does not permit a fine to be levied in lieu

of imprisonment or death”).

6

Anthony Young, a member of Curry’s organization

and the government’s key witness at the trial of the Kemo

Murder Counts, 7 was with Curry during that conversation and

overheard Bergrin say that “Kamo” was the confidential

witness against Baskerville. Young realized, however, that

Bergrin was referring to Kemo. According to Young, Bergrin

met with him and other Curry organization members

approximately one week after Baskerville’s arrest. At that

meeting, Bergrin told the group that “if Kemo testif[ied]

against [Baskerville], [Baskerville] w[ould] never see the

streets again” (Joint App. at 2528), but that he could “get

[Baskerville] out if Kemo d[id]n’t testify” (id. at 2529).

Bergrin twice reiterated “No Kemo, no case” and emphasized

that the group should not “let that kid testify against

[Baskerville].” (Id.)

Members of Curry’s organization thereafter discussed

how to find and kill Kemo, and, in March of 2004, Young

found Kemo and shot him to death.

7

Young was not the only witness who offered

testimony incriminating Bergrin in Kemo’s murder. Alberto

Castro, a drug dealer, testified that Bergrin offered him

$10,000 to murder Kemo, and two former confidants of

Bergrin’s testified that Bergrin implied his complicity in the

events that led to Kemo’s death. (See Joint App. at 3409

(testimony that Bergrin expressed his worry that “Baskerville

would implicate him in the Kemo case”); id. at 3781

(testimony that Bergrin stated he had “met with Baskerville’s

people at the office,” “told them the name of the [witness],”

and that they had “killed [the witness] three months later”).)

7

2. The Other Murder Plots

The government also sought to prove Kemo’s murder

using evidence of the Pozo Plot and the Esteves Plot, which

the District Court ultimately precluded after considering

evidentiary proffers.

The government’s first effort to rely on those other

murder plots developed pretrial when, after we ruled that the

RICO counts had been wrongly dismissed and remanded the

case, Bergrin filed a motion under Federal Rule of Criminal

Procedure 14 to sever the Kemo Murder Counts from the

Indictment. 8 Bergrin argued that a trial on every offense in

the Indictment would be unfairly prejudicial. The

government disagreed, contending that severing the Kemo

Murder Counts “would be a waste of judicial resources, …

would present increased danger for witnesses, and that

regardless of the severance plan … all or most of the evidence

of the related crimes would be admissible at … [any] of the

severed trials.” (Id. at 57-58.) It proffered, in that regard,

that it would seek to prove the Kemo Murder Counts in part

by relying on evidence of the Pozo Plot and the Esteves Plot

under Federal Rule of Evidence 404(b). 9

8

We refer to the Federal Rules of Criminal Procedure

simply as “Criminal Rules.” Criminal Rule 14(a) provides

that a “court may order separate trials of counts, sever the

defendants’ trials, or provide any other relief that justice

requires” when joinder “appears to prejudice a defendant or

the government.” Fed. R. Crim. P. 14(a).

9

We refer to the Federal Rules of Evidence simply as

“Rules.” Rule 404(b), as we discuss further infra, provides

8

i. The Pozo Plot

Pozo, the government asserted, was a “large scale drug

trafficker who distributed multi-hundred kilogram shipments

of cocaine he received in New Jersey via Texas.” (D.N.J.

ECF no. 09-369, doc. no. 304-1, at 13.) 10 In February 2004,

he was charged in the Western District of Texas for his role in

that drug distribution scheme, and he hired Bergrin to

represent him. Bergrin determined that Pozo’s co-defendant,

Pedro Ramos, was cooperating with the government against

Pozo. He told Pozo that Ramos was an informant, asked him

if he knew where Ramos lived, and told him that, if “we

could get to [Ramos] and take him out, Pozo’s headache (his

drug charges) would go away.” (D.N.J. ECF no. 09-369, doc.

no. 302, at 1 (internal quotation marks omitted).) Pozo

responded, “Are you nuts? I am not involved in murdering

people,” and later retained new counsel. (Id. (internal

quotation marks omitted).)

ii. The Esteves Plot

that although “[e]vidence of a crime, wrong, or other act” is

inadmissible to prove a person acted “in accordance with [his

or her] character,” Fed. R. Evid. 404(b)(1), it may be admitted

for “another purpose, such as proving motive, opportunity,

intent, preparation, plan, knowledge, identity, absence of

mistake, or lack of accident,” Fed. R. Evid. 404(b)(2).

10

Our references to documents on the District Court’s

docket cite to the pagination contained in the ECF-generated

header on each page.

9

Esteves, too, was a former client of Bergrin’s who

“operated a large scale drug trafficking business based in

New Jersey.” (D.N.J. ECF no. 09-369, doc. no. 304-1, at 23.)

He was prepared to testify that, when he met with Bergrin in

May 2008, after being charged in the Superior Court of New

Jersey with drug trafficking, Bergrin told him that “the only

way to beat the case was if [Esteves] took care of the

witnesses” on a list of those Bergrin believed were

cooperating with the government. (Id.) During that

conversation, Bergrin also told Esteves that he “hate[d] rats

and … would kill a rat himself,” that “this was not the first

time he ha[d] done this,” and that, “if there are no witnesses,

there is no case.” (Id.) An informant named Oscar Cordova,

whom Bergrin believed was a hitman, subsequently recorded

Bergrin instructing him to kill a witness on that list. (Id.; see

Joint App. at 225-28 (describing the plot).) In that

conversation, Bergrin stated, “we gotta make it look like a

robbery. It cannot under any circumstances look like a hit. …

We have to make it look like a home invasion robbery.”

(D.N.J. ECF no. 09-369, doc. no. 304-5, at 3.)

B. Procedural History

1. The First Severance

In a September 21, 2011 opinion (the “First Severance

Opinion”) citing those proffers, the District Court decided

that severance was necessary and ordered that the Kemo

Murder Counts be tried first.

The Court did say, however, that it would “likely allow

certain … Rule 404(b) evidence into the separate trial on the

[Kemo Murder] Counts.” (Joint App. at 58.) In particular, it

10

indicated that evidence of the Pozo Plot would be admissible

because that plot occurred “before or around the same time as

the [Kemo] murder conspiracy.” (Id. at 59.) Evidence of the

Esteves Plot, by contrast, troubled the Court. The Esteves

Plot was unlike the “other-crime evidence most typically

admitted under Rule 404(b),” the Court said, because it

pertained to acts that “happened more than four years after

the [Kemo] murder conspiracy” and was therefore evidence

of a “subsequent criminal act.” (Id.) Although the Court

acknowledged that there was no categorical “bar to

subsequent act evidence,” it observed that “evidence of a

subsequent act” is not necessarily “permissible or relevant in

the same way that evidence of a prior bad act may be.” (Id.)

Nevertheless, the Court seemed to take for granted that

the government would be permitted to use Bergrin’s own

admissions to Esteves in proving the Kemo Murder Counts.

(See id. at 62 (suggesting that certain evidence pertaining to

the Esteves Plot would “likely be admissible to provide the

requisite background information to support” the testimony of

the witnesses, including Esteves, who would testify to

Bergrin’s admissions).) Aside from that, however, the Court

made it clear that most of the proffered evidence pertaining to

the Esteves Plot would be inadmissible in a trial on the Kemo

Murder Counts. The Court was particularly concerned about

the tape recording of Bergrin’s conversation with Cordova,

evidently believing that the tape’s probative value was likely

to be substantially outweighed by the danger of unfair

prejudice:

The Government proffers that it will introduce

evidence, including audio recordings, showing

that in 2008 Bergrin had conversations with a

11

confidential informant – dubbed by the

Government as “the Hitman” – during which

Bergrin explicitly discussed killing [a witness]

and instructed the Hitman to make the murder

look like a home invasion robbery. By contrast,

the Government’s proffered evidence regarding

the [Kemo] murder is much more

circumstantial. The Government intends to

prove that Bergrin said the words “no Kemo, no

case” to certain other persons and that by

uttering these words Bergrin specifically

intended to cause those individuals to murder

[Kemo] to keep him from testifying. And

although the Government has a variety of

evidence specifically probative of the [Kemo

Murder] Counts it intends to introduce, the

evidence will likely be nowhere near as

overwhelming as the evidence relating to the

[Esteves Plot].

….

[I]n considering Bergrin’s guilt for the

[Kemo Murder] Counts, any limiting

instructions would likely be insufficient. It

would be perhaps unavoidable – and merely

human – for the jury to use the direct, explicit

evidence from the [Esteves Plot] murder

conspiracy case to infer Bergrin’s guilt of the

[Kemo Murder] Counts regardless of any

limiting instruction.

12

(Id. at 56.) Because Bergrin faced a life sentence on the

Kemo Murder Counts, the Court found that risk to be

particularly unacceptable. (See id. at 57 (“[A]lthough he is

charged with a variety of crimes, the stakes on the [Kemo

Murder] Counts are especially high for Bergrin: if a jury finds

him guilty on those counts, he faces a mandatory life

sentence.”).)

Thus, based in part on its view that evidence of the

other witness-murder plots would not, despite the

government’s argument, necessarily be admissible in a trial

on the Kemo Murder Counts, the Court severed those counts

from the Indictment and ordered them to be tried first.

2. The Government’s Motion to Admit Rule

404(b) Evidence and the District Court’s

Initial Ruling

On September 29, 2011, the government moved to

admit much of the Rule 404(b) evidence it had set forth in its

prior proffer, asking the Court to make “preliminary, pretrial

rulings on the admissibility of [the] other acts evidence” that

the Court’s First Severance Opinion had suggested would be

admissible in a trial on the Kemo Murder Counts. (D.N.J.

ECF no. 09-369, doc. no. 304-2, at 3.) Among other things,

the government sought admission of Pozo’s testimony about

the Pozo Plot, and Esteves’s testimony as to Bergrin’s

statements during the Esteves Plot. 11 At an October 7, 2011

11

Although the government implied that it was not

asking to introduce Bergrin’s recorded statement to Cordova

because of the Court’s ruling in the First Severance Opinion,

the government noted that “Bergrin’s defense strategy

13

hearing four days before jury selection was scheduled to

begin, the government followed up on the status of its Rule

404(b) motion, “requesting that the Court make at least some

preliminary rulings … certainly before the jury is sworn.” 12

(Joint App. at 584-85.) The Court did not do so, however,

and a jury was empaneled on October 13, 2011.

The next day, the Court announced its ruling on the

government’s motion which was memorialized in an undated

and unfiled opinion “handed to the parties the following

week.” 13 (D.N.J. ECF no. 09-369, doc. no. 304, at 7.)

Highlighting the factual similarities between the Pozo Plot

and the Kemo murder, the Court ruled that the government

would be permitted to introduce Pozo’s testimony under Rule

404(b):

[would] likely … open the door to additional Rule 404(b)

evidence.” (D.N.J. ECF no. 09-369, doc. no. 304-2, at 3.)

12

As the government explained to the Court, such a

ruling would permit it to “properly prepare an opening

statement” and “properly prepare [witnesses] so that they

don’t say something that’s inadmissible.” (Joint App. at 584-

85.) What was unsaid but perhaps implicit was that the

swearing in of a jury would cut off the government’s right

under 18 U.S.C. § 3731 to an immediate appeal of an adverse

evidentiary ruling.

13

Although it was attached as an exhibit to a motion

for reconsideration the government subsequently filed, the

Court’s Rule 404(b) opinion remains unfiled on the District

Court’s docket.

14

[E]vidence of the [Pozo Plot] … is admissible

under Rule 404(b). The Government seeks to

admit evidence that around February 2004,

while Bergrin was acting as [Pozo’s] lawyer in

a drug-trafficking case in federal court, Bergrin

provided [Pozo] with the identity of a

government witness against him, and counseled

[Pozo] that if the witnesses were killed, Bergrin

would win [Pozo’s] case. The factual

similarities of this case are so striking, and it is

so close in time – occurring contemporaneously

with the [Kemo] murder conspiracy – that this

evidence is highly probative of Bergrin’s intent

with respect to the charged conduct. And while

it carries a risk of undue prejudice, that

prejudice is insufficient to substantially

outweigh its high probative value. And the

Court will, again, mitigate the risk of prejudice

by providing a proper limiting instruction.

(Joint App. at 10 (internal citations omitted).)

The Court, however, retreated from its previous

suggestion that it would allow the government to introduce

some of the evidence pertaining to the Esteves Plot. It ruled

instead that no such evidence – including the “admissions

themselves” – would be allowed “under Rule 404(b) because

the potential for prejudice far outweigh[ed that evidence’s]

minimal probative value.” (Id. at 13.) As the Court

explained it, the admissions were minimally probative

because they were made in connection with a subsequent, as

opposed to a prior, crime and were therefore too attenuated

from the Kemo murder:

15

In its [First Severance Opinion], this Court

expressed at length its concerns regarding the

minimal probative value of – and the undue risk

of prejudice pose[d] by – this subsequent crime

evidence. And while the Court previously

indicated its willingness to consider allowing a

limited amount of evidence to provide the

necessary context as to these admissions, this

no longer seems appropriate now that the Court

has a better understanding of those admissions.

The admissions that Bergrin allegedly made are

too vague to be of great probative value –

indeed, Bergrin does not mention the [Kemo]

murder specifically, but alludes in general terms

to some past act of indeterminate nature. And

they, like the other evidence of the [Esteves

Plot] … are potentially unduly prejudicial. If

the admissions were admitted, the Government

would also be entitled to introduce additional

evidence regarding the [Esteves Plot], thereby

compounding the risk of prejudice. And, as

discussed previously, the potential prejudice of

evidence regarding the murder conspiracy with

Estevez [sic] is so great that it threatens to

prevent the jury from making a proper

determination of Bergrin’s guilt for the [Kemo]

murder – an untenable result, in light of this

Court’s previous rulings.

(Id. (internal citation omitted).)

3. Bergrin’s Opening Statement

16

Opening statements began on October 17, 2011.

Proceeding pro se with standby counsel, Bergrin told the jury

that the evidence would prove he “never wanted, … never

expected, …never believed … that one hair on Kemo’s head

would be hurt.” (Id. at 648.) Instead, as he explained to the

jury, he had simply acted as a zealous advocate on

Baskerville’s behalf: “[W]hen I represented – was called …

to represent William Baskerville, who was accused of a

criminal offense, the Sixth Amendment of the United States

Constitution said that I had to represent him, that he deserved

to be represented effectively. And that’s all I ever did in this

case.” (Id.) Thus, although Bergrin acknowledged calling

Curry and informing him that Kemo was the confidential

witness, he characterized that call as part of his legal duty to

represent Baskerville and denied any malicious motives. 14

Bergrin spoke similarly in explaining the Pozo Plot to

the jury, stating:

Let me tell you about the facts of

Richard Pozo which will come out in this case.

Richard Pozo was dealing cocaine. He sent a

car with a bunch of cocaine in it from Elizabeth,

… where he was living, to Texas. The car

began to be investigated. The car was dropped

off in the driveway of somebody’s house.

While the car is being investigated, Richard

14

Correspondingly, Bergrin claimed during the course

of trial that he never participated in a meeting with Curry-

organization members in which he allegedly implied that

Kemo should be killed by saying, among other things, “No

Kemo, no case.”

17

Pozo comes to see me and says: I think I have a

problem. I believe they detected cocaine in a

car that I had sent to Texas. Will you represent

me?

There is no informant involved. We

have absolutely … no idea whatsoever who any

informants are. The name Pedro Ramon

doesn’t even fit into the equation. We have no

clue who the informant is, he has no clue who

any informant is. And I question him in front of

Peter Willis and another outstanding attorney by

the name of John Whipple in Texas, and that’s

borne out here in this particular case. I never

say to [Pozo]: Let’s get rid of the informant.

Because what does it matter? It doesn’t matter.

I would never say that because it has no impact,

has no effect and I would never say that to this

type of individual.

(Id. at 691-92.)

Believing that Bergrin had made “various door-

opening assertions during his opening statement” the

government filed a letter-motion the next day, asking the

District Court to reconsider its evidentiary ruling excluding

the Esteves Plot evidence. 15 (D.N.J. ECF no. 09-369, doc. no.

263, at 1.) The Court declined.

15

More specifically, the government argued that

Bergrin “exploited [the] Court’s [evidentiary] rulings, and

abused his status as a pro se litigant, by testifying in his

opening statement.” (D.N.J. ECF no. 09-369, doc. no. 263, at

18

4. The District Court’s Decision to Exclude

Evidence of the Pozo Plot

Worse yet for the government, on November 8, 2011,

the Court reversed course on the admissibility of Pozo’s

testimony. Acknowledging that it had previously “indicate[d]

that that testimony would be admissible under [Rule] 404(b),”

the Court said it had changed its mind, “after hearing the case

and the context in which [the testimony was] now being

offered.” (Joint App. at 19.) The Court described a three-

step process for considering whether to admit evidence under

Rule 404(b): first, to “decide whether there is sufficient

evidence that the other act in question actually occurred”;

second, to assess “whether the evidence of other acts is

probative of the material issue other than character”; and

third, to consider “whether the probative value of the

evidence is substantially outweighed by its potential

prejudicial effect.” (Id.)

Pozo’s testimony, the Court said, was problematic

under the first step of that procedure, because there was no

independent documentation corroborating the substance of his

intended testimony:

2.) And because it believed that Bergrin had brought his

intent into question, the government asked the Court to allow

it to introduce, among other things, Esteves’s testimony so as

to rebut Bergrin’s “blanket, self-serving assertion” that he

would never say “[l]et’s get rid of the informant … to … a

client facing charges because of a cooperating witness.” (Id.

at 5 (alteration in original) (internal quotation marks

omitted).)

19

The first step is very rarely even an

issue. … [It] is almost typically a prior

conviction. It will be evidence of even a prior

arrest which has some independen[t]

corroboration because police make a prior arrest

and then they seek to offer that type of

evidence. It might even be a wiretap.

….

One of the concerns I have, and I’ve had,

is that we’re talking about conversations which

allegedly occurred many years ago, and we’re

talking about people’s best recollections of that

conversation without it having been recorded,

without it having been documented

immediately.

….

We have that in this case already. We

have this case, one of the biggest contentions in

this case is if the statement “No Kemo, no case”

was made, what exactly does that mean.

And the conversation with Mr. Pozo, I

know the Government will say that’s Mr.

Pozo’s best recollection. But there’s nothing to

document – when we’re talking about parsing

such important words, there’s nothing to

document what actually was spoken at that time

in those few little sentences that the

Government contends would show that Mr.

20

Bergrin was attempting [to] … you know, to

murder the witness.

(Id. at 19-21.)

The Court suggested that its concern about whether

Pozo’s testimony was truthful also played a role in assessing

whether, under the third step, the probative value of the

evidence was substantially outweighed by “its potential

prejudicial effect.” (Id. at 19.) And the Court declared

Pozo’s testimony would be “cumulative,” “collateral,” and

“confusing.” It explained:

I have no sense of confidence that this evidence

would be so reliable that its probative value

would outweigh its prejudicial effect. And I

think, you know, there’s a concern that it would

be considered by this jury as propensity versus

really going to intent.

Now, in that context, let me also say, one

of the considerations is, is there other evidence

of intent in this case?

And, you know, you have other

evidence, so this would be cumulative and very

collateral and very confusing, in this Court’s

opinion.

….

[Y]ou have evidence of intent, you have,

if the jury believes Mr. Young, you have the

21

conversation that Mr. Young testified to …,

which is a very specific conversation that he

says he recalls Mr. Bergrin making at that time

back in 2004, shortly after … Mr. Baskerville’s

arrest. He testified at … some length about that

conversation. So you have evidence of what

“No Kemo, no case” means.

You also have the evidence that you

brought forth about Mr. Castro. You brought

forth evidence that Mr. Bergrin went to …

another motivated witness, … which the jury

will have to consider in which he says, Mr.

Bergrin went to him at some point and said, you

know: I’ll give you $10,000 if you would, you

know, kill this guy.

Mr. Pozo would be another witness, a

drug dealer who is claiming at some point some

conversation occurred. It’s not documented.

And in weighing the factors that I need to weigh

as far as, you know, the minimum degree it will

have with respect to intent, because the jury

would have to parse those words, whatever they

finally conclude were the words, first of all,

because there’s nothing to document other than

Mr. Pozo saying what he remembers, and then

on cross it may come out to … be something

else, they’d have to document those – they’d

have to parse those words along with the “No

Kemo, no case.” And I think their challenge as

far as dealing with “No Kemo, no case” is

enough.

22

(Id. at 23-26.)

That evening, the government filed a motion asking

the Court to reconsider its decision to exclude evidence of the

Pozo Plot and the Esteves Plot. As the government argued

the next morning in support of its motion, one might perceive

“an inherent tension” between the ruling that the Pozo Plot

was based on “insufficient proof … because we don’t have a

tape” and the ruling that the Esteves Plot was too prejudicial

“because we do have a tape.” (Id. at 36, 37.) The District

Court was not persuaded, however, and reaffirmed its

rulings. 16

5. Closing Arguments and the Jury’s

Verdict

At a subsequent conference about jury instructions,

Bergrin successfully requested that the jury be told it “is a

defense to the charges in the Indictment that the defendant’s

acts constituted lawful and legitimate legal representation of a

client.” (D.N.J. ECF doc. no. 09-369, no. 327, at 46; see Joint

App. at 4024-25 (Bergrin’s request).) Then, in summation, he

16

As to the Esteves Plot, the Court stated that if “there

was a conviction [in Bergrin’s case], I would believe … that

that conviction was the result of the Esteves evidence,

because I don’t see how [the jury] could humanly put that out

of their mind.” (Joint App. at 38.) As to the Pozo Plot, the

Court again laid out its fear that “the jury would … have to

parse out what exactly did Mr. Bergrin say … according to

Mr. Pozo’s recollections eight months after the incident”

given that “we’re talking about a drug dealer, and hearing

words that he thought.” (Id. at 39.)

23

echoed his opening statement’s assertion that he was being

“accused for doing [his] job,” to “defend the Constitution [by]

mak[ing] sure that [Baskerville] ha[d] effective

representation.” (Joint App. at 4188.) Indeed, while Bergrin

again acknowledged that he had discussed Kemo’s name with

Baskerville and disclosed it to Curry over the phone, he

attributed his behavior to legitimate representation, and

implored the jury not to conclude “under any circumstance,

under any leap of bound and faith that [he] ever intended for

one hair to be hurt on poor Kemo’s head.” (Id. at 4277;

accord id. at 4194 (“I, under no circumstances, ever intended,

ever wanted, ever told, ever warned, ever advised, ever

informed anyone to ever harm a hair on the head of Kemo

McCray. I never had that intent.”).)

After six days of deliberation, the jury was unable to

reach a verdict. As a result, the Court declared a mistrial on

November 23, 2011, and scheduled a retrial on the Kemo

Murder Counts for January 2012.

6. The Government’s Appeal and Efforts to

Determine Which Counts to Try Next

Shortly thereafter, the government inquired “about

rulings that [the Court] made excluding evidence,” asking the

Court to clarify if it was “going to adhere to those; Pozo and

Esteves and the things that were contained in the … 404(b)

ruling.” (Id. at 49.) The Court responded as follows:

“Absolutely. I don’t see – unless you can convince me

otherwise, as to why those rulings – I know you feel

otherwise – but on reflection I feel strongly that those rulings

were appropriate. So I don’t expect I would be changing

those rulings.” (Id. at 49-50.) On November 30, 2011, the

24

government filed a notice of appeal challenging those

evidentiary rulings. 17

The next day, the government moved to try the

remainder of the counts in the Indictment at the January 2012

retrial, though it stated it would be willing to sever the tax

evasion counts upon Bergrin’s request. The Court held a

hearing on December 8, 2011 to consider which counts to try

next. Bergrin appeared at the hearing and asked the Court to

stay proceedings pending our disposition of the government’s

appeal of the evidentiary rulings. After consulting with the

government, the Court suggested a second severance in which

the drug-trafficking counts and the witness-tampering counts

pertaining to the Esteves Plot would be severed and tried

before the rest of the Indictment. 18 The government declined

the Court’s suggestion, however, prompting Bergrin to file a

severance motion.

At an ensuing hearing on December 14, 2011, the

government again requested that it be permitted to try the

17

As discussed infra in Part II.A, the government

invoked 18 U.S.C. § 3731 as the basis for appellate

jurisdiction.

18

Under that proposal, Count 5 of the Indictment –

which charges a drug-trafficking conspiracy and lists the

Kemo murder, the Pozo Plot, and the Esteves Plot as part of

the conspiracy’s “manner and means” – would have been

altered to delete allegations relating to the Kemo murder so

that “the Government [would be precluded] from introducing

any such evidence.” (D.N.J. ECF no. 09-369, doc. no. 352, at

1.)

25

entirety of its case against Bergrin, because “[t]he Kemo

murder and the Esteves thing [were] not [disconnected]

bookends” but rather were charged as “a racketeering RICO

violation because” that was what they were. (Id. at 4436.)

The Court, however, made clear that it would not accept the

government’s request to “go forward with the Kemo

allegations … in the” RICO counts:

The concern I always had and the reason I

severed out [the Kemo Murder Counts] was

because of what I believed, and still do believe

– and I think, frankly, the result of the jury

being hung reflects what I had a concern about

– is that charge, standing alone, for the reasons I

stated in the severance, I always was concerned

about the prejudice there would have been if [it]

would have been tried with Esteves and all of

the drug evidence that occurred subsequently.

And I still feel the jury wouldn’t have been able

to separate that out and decide the Kemo case

just based on that case and the prior crime

evidence that this Court didn’t let in.

(Id. at 4433.) Trying the RICO counts next, the Court said,

would unfairly expose Bergrin to a potential life sentence:

The Court: … [I]n my opinion it would

have been inherently unfair to have him

convicted under a RICO – the way that was

framed for the murder case facing a life in

prison sentence tried that way. That’s how I felt

and I still feel that way. And yet, you still feel

insistent on that’s a fair trial, he should be

facing that kind of penalty on the Kemo part of

26

the case when you already now saw a jury come

back and couldn’t reach a verdict on that.

Sure, if you get all your other evidence

in he’ll get convicted on the Kemo murder part

of the case and, you know, that’s what you

want.

[Government’s Counsel]: Well, Judge –

The Court: And that’s the way you want

to do it, and that’s what I have a real difference

of opinion with.

[Government’s Counsel]: I understand.

The Court: And the Government, you

know, they can charge a ham sandwich. I know

that; you know that.

So if you charge a RICO case on its face

on the indictment, it doesn’t take a whole lot to

charge a RICO case.

(Id. at 4463-64.) Given that, in the Court’s view, the

government’s case on the charges other than the Kemo

Murder Counts and the related RICO counts was “very

strong,” 19 that it could be proven without the witnesses who

19

The Court opined that, aside from the Kemo Murder

Counts, the government had a “very clean, strong case,” with

witnesses “more credible than Anthony Young and Castro

type witnesses.” (Joint App. at 4461.)

27

had testified in the trial on the Kemo Murder Counts, and that

it would warrant a “sentence that would reflect the severity of

[those other charges],” the Court suggested that the

government should not “spend the taxpayers’ money to come

in here, put on [the Kemo Murder Counts] evidence again,

[and] stand behind those kinds of witnesses again when [the

government did not] have to do it.” (Id. at 4460.)

In response to those concerns, the government invited

the Court to dismiss the RICO counts if it believed “that Mr.

Bergrin [could not] get a fair trial … as presently

constituted.” (Id. at 4458.) The Court, however, refused to

dismiss the Indictment’s RICO counts, stating that it had

already “[done] that once … because at the time I still was

concerned about the RICO allegations, quite frankly, mostly

for the same reason.” (Id.)

7. The Second Severance

Instead, on December 27, 2011, the Court severed the

substantive counts charging Bergrin with drug trafficking and

participating in the Esteves Plot from the rest of the counts in

the Indictment, and ordered that they be tried in January 2012

(the “Second Severance Order”). The Court explained that its

“original premise [was] that trying Bergrin for his alleged

involvement in the [Kemo] murder conspiracy with extensive

evidence from the [Esteves Plot] … would be fundamentally

unfair and improper” (id. at 67), and it went on to say that the

concerns memorialized in its First Severance Opinion

required an additional severance, because the government’s

appeal with respect to the Kemo Murder Counts made it

“impossible” to pursue the “most logical solution” of simply

retrying those counts (id. at 69).

28

Severing the Indictment’s drug-trafficking and Esteves

Plot counts was the next best solution, the Court said, since

such a severance would

avoid[] undue prejudice because Bergrin faces

no exposure for his alleged involvement in the

[Kemo] murder conspiracy, and so the jury

cannot find him guilty of those charges based

on improper spillover evidence. It also

incorporates as many of the remaining counts as

may properly be joined, and, if Bergrin is

convicted, carries a substantial penalty which

should satisfy the Government’s desire for

justice.

(Id. at 73.) The Court also ruled that it was necessary to

ensure that those counts were tried before the RICO counts in

which the Kemo murder and the Esteves Plot were intrinsic,

rejecting the government’s statement that it should be

permitted to proceed on its RICO charges first, and

characterizing that position as a “thinly veiled attempt to

either circumvent [the Court’s] prior decision or discourage

the Court from taking further actions required by justice.”

(Id.)

That same day, the government filed a second notice

of appeal, this time challenging the Second Severance Order.

II. Discussion

The government argues that the District Court abused

its discretion by precluding the introduction of evidence of

29

the Pozo Plot and the Esteves Plot in the retrial on the Kemo

Murder Counts, and in ordering the drug-trafficking and

Esteves Plot counts to be severed. It also contends that this

case should be reassigned to another district judge. Bergrin

of course disagrees, but spends the bulk of his efforts arguing

that we lack jurisdiction to entertain any of the government’s

arguments.

We begin by addressing the jurisdictional issue.

A. Jurisdiction

The District Court had jurisdiction over this case under

18 U.S.C. § 3231. We, in turn, have appellate jurisdiction to

consider challenges to “decision[s] or order[s] of a district

court suppressing or excluding evidence …, not made after

the defendant has been put in jeopardy,” so long as the

“United States attorney certifies to the district court that the

appeal is not taken for purpose of delay and that the evidence

is a substantial proof of a fact material in the proceeding.” 18

U.S.C. § 3731. The government’s appeal from the District

Court’s ruling excluding evidence of the Pozo Plot and the

Esteves Plot invokes that jurisdiction on the ground that the

District Court’s verbal statement that it would “[a]bsolutely”

adhere to its prior rulings on retrial (Joint App. at 49) was an

appealable “decision or order” excluding evidence.

Given § 3731’s express mandate that its provisions

“shall be liberally construed to effectuate its purposes,” 18

U.S.C. § 3731, there is wide agreement that oral decisions

dealing with subjects within the statute’s scope are

appealable. See United States v. Farnsworth, 456 F.3d 394,

398-99 (3d Cir. 2006) (presuming an “oral ruling” is

30

appealable under § 3731, but holding the ruling at issue was

not appealable because it was not, as the government

contended, a dismissal); United States v. Janati, 374 F.3d

263, 269 (4th Cir. 2004) (exercising appellate jurisdiction

over an “oral ruling”); United States v. Presser, 844 F.2d

1275, 1280 (6th Cir. 1988) (“[W]e view the district court’s

oral statement as evidencing an intent to exclude government

evidence … and consequently, its statement qualifies as an

appealable order … .”); United States v. Flores, 538 F.2d 939,

942-43 (2d Cir. 1976) (oral ruling “exclud[ing] evidence of

prior acts and statements” was appealable under § 3731).

Bergrin argues, however, that the District Court’s statement

was not sufficiently definite to constitute an appealable

decision or order, because the District Court was not

unequivocal in saying it would exclude evidence of the Pozo

Plot and the Esteves Plot at Bergrin’s retrial. Cf. United

States v. Brooks, 145 F.3d 446, 453-54 (1st Cir. 1998) (stating

that orders lacking a requisite “degree of finality … may not

qualify as … order[s] excluding evidence under section

3731”). The record belies that claim.

Although the Court’s colloquy did include some

qualifying language, the first thing it said was that it would

“[a]bsolutely” exclude that evidence from Bergrin’s retrial.

(Joint App. at 49.) And it further stated that it “fe[lt] strongly

that [its] rulings were appropriate.” (Id.) The Court’s rulings

over the course of Bergrin’s trial on the Kemo Murder Counts

reflect similarly strong convictions, even amidst repeated

requests by the government to introduce the Pozo Plot and the

Esteves Plot evidence after Bergrin denied any intent to harm

Kemo. Moreover, the Court confirmed its resolve to keep out

the questioned evidence when, at a hearing after the

government’s first appeal was filed, it reiterated that

31

excluding the evidence was “the right decision” (id. at 4446),

and subsequently ordered a second severance based on its

belief “that trying Bergrin for his alleged involvement in the

[Kemo] murder conspiracy with extensive evidence from the

[Esteves Plot] … would be fundamentally unfair and

improper” (id. at 67).

The District Court did, to be sure, leave open the

possibility that it would reconsider its evidentiary

determinations, and it is possible, as Bergrin points out, that

circumstances may change in the future. But the chance of

change is inherent in virtually every pretrial evidentiary

ruling and treating such rulings as unreviewable “would

insulate [them] from appellate review, thus frustrating … the

purposes of § 3731.” United States v. Siegel, 536 F.3d 306,

315 (4th Cir. 2008). Indeed, even a district court’s explicit

suggestion that a ruling is “preliminary and could change”

does not make it an unappealable one under § 3731. Id. at

314; cf. United States v. Horwitz, 622 F.2d 1101, 1104 (2d

Cir. 1980) (stating that even “conditional … ruling[s], which

raise[] the remote prospect that suppression will not be

ordered, [do not] necessarily deprive[] [an appellate] court of

jurisdiction under section 3731”). Thus, while there was

perhaps some “ambiguity about the district court’s future

actions,” its statement clearly “evidenc[ed] an intent to

exclude government evidence” at Bergrin’s retrial, Presser,

844 F.2d at 1280, and thereby laid the foundation for our

jurisdiction under § 3731.

32

We turn, then, to consider the government’s challenge

to the merits of those evidentiary rulings. 20

20

Although we can undoubtedly review the District

Court’s second severance to determine whether it warrants

mandamus relief, since the government has alternatively

petitioned for that writ, see United States v. Santtini, 963 F.2d

585, 590 (3d Cir. 1992) (“[P]arties are free to proceed

alternatively on application for a writ or by appeal, with the

court determining which, if any, procedure is more

appropriate.”), the jurisdictional question presented by the

appeal of the Second Severance Order is more difficult. The

government argues that we have pendent appellate

jurisdiction to consider that appeal based on our § 3731

jurisdiction to consider the District Court’s evidentiary

rulings. We have “recognized ‘a discretionary, though

‘narrow,’ doctrine of pendent appellate jurisdiction,’” E.I.

DuPont de Nemours & Co. v. Rhone Poulenc Fiber & Resin

Intermediates, S.A.S., 269 F.3d 187, 204 (3d Cir. 2001)

(citation omitted), but there is a split in authority as to

whether that doctrine applies in criminal cases and we have

not expressly employed it in that context, compare, e.g.,

United States v. Ferguson, 246 F.3d 129, 138 (2d Cir. 2001)

(“[T]here is no pendent appellate jurisdiction in criminal

cases.” (citing Abney v. United States, 431 U.S. 651, 662-63

(1977))), with United States v. Lopez-Lukis, 102 F.3d 1164,

1167 & n.10 (11th Cir. 1997) (exercising pendent appellate

jurisdiction over an order striking a count from an indictment

where the government appealed an order suppressing

evidence under § 3731), and United States v. Maker, 751 F.2d

614, 626 (3d Cir. 1984) (reviewing a severance order for an

abuse of discretion without explicitly relying on, or referring

to, pendent appellate jurisdiction). Because we will require

33

B. The District Court’s Exclusion of Pozo’s

Testimony 21

Before trial, the District Court had ruled that the

government would be permitted under Rule 404(b) to

introduce Pozo’s testimony that Bergrin counseled him to

murder a witness. As the Court noted at that time, “[t]he

factual similarities” between that incident and the Kemo

murder are “striking,” and the “evidence is highly probative

of Bergrin’s intent with respect to [the Kemo murder].”

(Joint App. at 10.) Although the Court thought the admission

of that testimony “carrie[d] a risk of undue prejudice,” it

concluded “that [the] prejudice [was] insufficient to

substantially outweigh its high probative value” and noted

that it would “mitigate the risk of prejudice by providing a

proper limiting instruction.” (Id.) At trial, however, even

after Bergrin told the jury in his opening statement that he

would not have made the statements to which Pozo would

testify and declared that he had been acting legitimately as an

attorney in representing Baskerville, the Court turned about

and ruled that the evidence was inadmissible. We agree with

the government that the reasons given for that change reflect

an abuse of discretion.

this case to be reassigned and will ask that the severance

rulings be revisited, see infra Part II.C, we need not determine

the propriety of the Second Severance Order and therefore do

not decide whether we have appellate jurisdiction to review it.

21

We review a “district court’s decision regarding the

admissibility of evidence … for abuse of discretion.” United

States v. Higdon, 638 F.3d 233, 238 (3d Cir. 2011).

34

Rule 404(b), as we have noted, provides that

“[e]vidence of a crime, wrong, or other act is not admissible

to prove a person’s character in order to show that … the

person acted in accordance with the character,” Fed. R. Evid.

404(b)(1), but the Rule permits such evidence “for another

purpose, such as proving motive, opportunity, intent,

preparation, plan, knowledge, identity, absence of mistake, or

lack of accident,” Fed. R. Evid. 404(b)(2). To be admissible

under Rule 404(b), then, evidence of uncharged crimes or

wrongs must have a proper evidentiary purpose. “A proper

purpose is one that is ‘probative of a material issue other than

character.’” United States v. Green, 617 F.3d 233, 250 (3d

Cir. 2010) (quoting Huddleston v. United States, 485 U.S.

681, 686 (1988)); see United States v. Johnson, 199 F.3d 123,

128 (3d Cir. 1999) (stating that Rule 404(b) evidence is

proper “if relevant for any other purpose than to show a mere

propensity or disposition on the part of the defendant to

commit the crime” (citation and internal quotation marks

omitted)). As long as evidence offered under Rule 404(b)

satisfies that criterion, we favor its admission. Johnson, 199

F.3d at 128. Of course, such evidence may be excluded if “its

probative value is substantially outweighed by a danger of …

unfair prejudice, confusing the issues, misleading the jury,

undue delay, wasting time, or needlessly presenting

cumulative evidence.” Fed. R. Evid. 403. We have also

emphasized that limiting instructions may be appropriate

when admitting Rule 404(b) evidence. See Green, 617 F.3d

at 249 (noting that a limiting instruction should be given

where requested).

1. The Decision That There Was

Insufficient Evidence to Establish That

Pozo’s Testimony Was Truthful

35

As the District Court correctly explained, one step in

evaluating whether to admit Rule 404(b) evidence is to

determine whether there is sufficient evidence to conclude

that the crime, wrong, or other act in question actually

occurred, because “similar act evidence is relevant only if …

the act occurred and … the defendant was the actor.”

Huddleston, 485 U.S. at 689. Applying that inquiry to Pozo’s

proffered testimony, however, the Court made its own

credibility assessment, saying that there was “nothing to

document what actually was spoken at that time in those few

little sentences that the Government contends would show

that Mr. Bergrin was attempting [to] … you know, to murder

the witness.” (Joint App. at 21.) Owing to the lack of

independent corroboration, the Court decided that “Mr.

Pozo’s best recollection” would not suffice. (Id.) That was

an error of law.

In Huddleston v. United States, the Supreme Court

expressly rejected the “level of judicial oversight” that the

District Court applied here in excluding Pozo’s testimony.

485 U.S. at 688. It said, rather, that Rule 404(b) evidence

need only be supported by sufficient evidence for a jury to be

able to “reasonably conclude that the act occurred and that the

defendant was the actor.” Id. at 689. A court’s task in that

regard is simply to decide, in accordance with Rule 104(b),

“whether the jury could reasonably find th[ose] facts … by a

preponderance of the evidence.” 22 Id. at 690; see Fed. R.

22

Thus, although the standard of proof in criminal

cases requires a greater showing than a preponderance of

proof, evidence of a contested fact may be admissible towards

that greater burden when the “evidence in the case” permits a

36

Evid. 104(b) (“When the relevance of evidence depends on

whether a fact exists, proof must be introduced sufficient to

support a finding that the fact does exist.”). Importantly, the

Supreme Court instructed that, in making that determination,

trial courts must not “weigh[] credibility” or “make[] a

finding.” Huddleston, 485 U.S. at 690.

By discounting Pozo’s testimony based on a lack of

corroboration and questions about credibility, the Court

usurped the jury’s role. See United States v. Dillon, 532 F.3d

379, 391 (5th Cir. 2008) (“Rule 104(b) does not require

corroboration. It only requires that the district court consider

the witness’s testimony and determine that a reasonable jury

could [make the required] find[ing] by a preponderance of the

evidence … .” (internal footnote omitted)); Siegel, 536 F.3d

at 319 (“Evidence is [sufficiently] reliable for purposes of

Rule 404(b) ‘unless it is so preposterous that it could not be

believed by a rational and properly instructed juror.’”

(citation omitted)); cf. United States v. Haut, 107 F.3d 213,

220 (3d Cir. 1997) (noting that it “is a basic tenet of the jury

system that it is improper for a district court to substitute[ ]

[its] judgment of the facts and the credibility of the witnesses

for that of the jury” (alteration in original) (citation and

internal quotation marks omitted)). While Pozo’s credibility

and motivation for testifying may be open to question, his

testimony itself was sufficient to permit a jury to reasonably

conclude, by a preponderance of the evidence, that Bergrin

did the things that Pozo said he did. See United States v.

Bailey, 990 F.2d 119, 123 (4th Cir. 1993) (explaining that a

witness’s testimony should not be precluded “simply because

jury to “reasonably find the … fact … by a preponderance of

the evidence.” Huddleston, 485 U.S. at 690.

37

it is in conflict with or contradicted by other testimony” or is

offered by a “witness [who] has an unsavory past,” as those

“are merely circumstances for the jury to consider”).

Consequently, the Court was obliged to permit a jury

to consider that testimony, provided it was otherwise

admissible under the Federal Rules of Evidence.

2. The Finding That Pozo’s Testimony Was

Substantially More Prejudicial Than

Probative

The District Court did not believe that Pozo’s

testimony was otherwise admissible, because, under Rule

403, the Court determined that the testimony was cumulative

and confusing, and that the prejudice from it substantially

outweighed any probative value. All of those conclusions are

problematic.

To begin with, it is not clear that the Court applied the

proper test under Rule 403, because, at times, it spoke simply

in terms of “prejudice” to Bergrin. (Joint App. at 23.) It must

always be remembered that unfair prejudice is what Rule 403

is meant to guard against, that is, prejudice “based on

something other than [the evidence’s] persuasive weight.”

United States v. Cruz-Garcia, 344 F.3d 951, 956 (9th Cir.

2003); see United States v. Starnes, 583 F.3d 196, 215 (3d

Cir. 2009) (“[U]nfair prejudice does not simply mean damage

to the opponent’s cause.” (citation and internal quotation

marks omitted)). Assuming the Court was using the term

“prejudice” as shorthand for “unfair prejudice,” we are

examining the kind of balancing decision to which we would

ordinarily accord great deference. See United States v.

38

Kellogg, 510 F.3d 188, 197 (3d Cir. 2007) (noting that if

“judicial self-restraint is ever desirable, it is when a Rule 403

analysis of a trial court is reviewed by an appellate tribunal”

(citation and internal quotation marks omitted)). In general, a

Rule 403 decision will not be reversed unless the “analysis

[undertaken] and resulting conclusion” is “arbitrary or

irrational.” Id. (citation and internal quotation marks

omitted). Here, unfortunately, the District Court’s Rule 403

analysis was arbitrary, in that it was based on the same legally

flawed credibility determination that led the Court to

conclude that Pozo’s testimony was inadmissible without

independent corroboration.

Pozo, as the District Court saw it, “would be another

witness, a drug dealer who is claiming at some point some

conversation occurred.” (Joint App. at 25.) Assessing his

proffered testimony in that light, the Court characterized it as

having a “minimum degree [of persuasiveness] … with

respect to intent.” (Id.) An assumption about how the jury

would view Pozo’s credibility was, however, an improper

basis for discounting his testimony’s probative value. See

United States v. Welsh, 774 F.2d 670, 672 (4th Cir. 1986)

(“[A]s a general rule, the credibility of a witness has nothing

to do with whether or not his testimony is probative with

respect to the fact which it seeks to prove.”); 22 Charles Alan

Wright & Kenneth W. Graham, Jr., Federal Practice and

Procedure § 5214 (4th ed. 1996) (“[I]t seems relatively clear

that in the weighing process under Rule 403 the judge cannot

consider the credibility of witnesses.”). And that errant

starting point likewise tainted the Court’s conclusion that the

jury would be confused by Pozo’s testimony indicating that

Bergrin told him to “take out” a cooperating witness. The

39

only possible confusion, if it can be called that, would arise

from discrediting the source of the testimony. 23

Stripped of improper credibility assessments, Pozo’s

proffered testimony is – as the District Court initially

observed when saying it would be admissible – highly

probative of Bergrin’s guilt, because the factual similarities

between the Pozo Plot and the Kemo murder truly are

“striking.” (Joint App. at 10.) Pozo was a drug dealer

represented by Bergrin around the same time as the Kemo

murder, and he was prepared to testify that Bergrin suggested

that he kill a witness. Pozo’s testimony is, therefore,

powerfully suggestive of Bergrin’s intent in passing Kemo’s

identity on from Baskerville to Curry. It is likewise relevant

to deciding whether Bergrin uttered the words “No Kemo, no

case,” and, if he did, what he meant. 24

23

Nor was the testimony, as the District Court

suggested, cumulative. After all, the Court itself recognized

that the credibility of the primary witnesses against Bergrin

on the Kemo Murder Counts is open to question. See supra

note 19. Pozo’s testimony would therefore have added much

“to the probative force of the other evidence in the case,” and

“contribut[ed] to the determination of truth,” United States v.

Williams, 81 F.3d 1434, 1443 (7th Cir. 1996), and so it cannot

properly be said to be “cumulative,” United States v. Brown,

597 F.3d 399, 406 (D.C. Cir. 2010) (stating evidence should

only be deemed “cumulative” when the “evidence on one side

is so full that no jury that rejected it would be likely to change

its mind because of the introduction of the proffered

evidence” (citation omitted)).

24

The District Court itself recognized that intent was a

key issue in the case in its colloquy excluding Pozo’s

40

In sum, we conclude that the District Court’s ruling

excluding Pozo’s testimony cannot “be reconciled with a

sound exercise of discretion,” United States v. Gatto, 924

F.2d 491, 501 (3d Cir. 1991), and, accordingly, must be

vacated. 25

testimony under Rule 403. (See Joint App. at 20-21 (“[O]ne

of the biggest contentions in this case is if the statement ‘No

Kemo, no case’ was made, what exactly does that mean.”).)

But Bergrin argues that his intent is not at issue with respect

to the Kemo Murder Counts, because his primary defense is

that he never attended the meeting in which he allegedly said

“No Kemo, no case.” We disagree. Bergrin’s insistence that

he did not say those words does not mean the jury will not

have to consider them. It is for the jury to decide whether he

said them. Moreover, as we have just noted, the question of

Bergrin’s intent is not only relevant to determining what “No

Kemo, no case” may mean, but also to ascertaining Bergrin’s

purpose in telling Curry who the witness against Baskerville

was.

25

Pozo’s proffered testimony was proper Rule 404(b)

evidence, and, as we have made plain, our review of the

record thus far reveals no sound basis upon which it should

have been precluded from the government’s case on the

Kemo Murder Counts under Rule 403. We nevertheless leave

it to the new judge to whom this case will be assigned to

conduct his or her own balancing under Rule 403 if the

government again seeks to prove the Kemo Murder Counts

using evidence of the Pozo Plot.

With respect to the Esteves Plot, we agree with the

government that the District Court observed an unwarranted

41

analytical distinction between a “prior bad act” and a

“subsequent bad act,” reasoning that the latter “looks more

like evidence that is being offered to show that the accused is

a ‘bad guy,’ someone with the propensity to commit criminal

acts.” (Joint App. at 60.) Rule 404(b) refers to evidence of

crimes, wrongs, or other acts, saying nothing about whether

the act in question is a “prior” or “subsequent” act. That

makes sense because light can be shed on motive, intent, and

the other issues listed in Rule 404(b)(2) as much by a

subsequent course of behavior as it can by a prior one. Cf.

Huddleston, 485 U.S. at 686 (referring to “similar acts

evidence under Rule 404(b)” (emphasis added)). So although

we once questioned, in dicta “[t]he logic of showing prior

intent or knowledge by proof of subsequent activity,” United

States v. Boyd, 595 F.2d 120, 126 (3d Cir. 1978), the District

Court erred to the extent it dismissed the probative value of

subsequent act evidence. See United States v. McGilberry,

620 F.3d 880, 886 (8th Cir. 2010) (“Rule 404(b) draw[s] no

distinction between prior and subsequent acts that would

support different analyses … .” (citation and internal

quotation marks omitted)); United States v. Rutkoske, 506

F.3d 170, 177 (2d Cir. 2007) (“The courts of appeals mostly

agree that the admission of subsequent acts under Rule 404(b)

is governed by the same four-part test as prior acts … .”);

United States v. Mohr, 318 F.3d 613, 617 (4th Cir. 2003)

(“Rule 404(b) … covers evidence of both prior and

subsequent acts.”); United States v. Jernigan, 341 F.3d 1273,

1283 (11th Cir. 2003) (“[T]he standard for evaluating the

admissibility of a subsequent bad act under Rule 404(b) is

identical to that for determining whether a prior bad act

should be admitted under this Rule.”); United States v.

Echeverri, 854 F.2d 638, 645 (3d Cir. 1988) (“We do not

42

dispute that there may be cases in which evidence of

subsequent wrongful acts may properly be admitted under

Rule 404(b)”); United States v. Alker, 260 F.2d 135, 157 (3d

Cir. 1958) (stating that “prior and subsequent acts …

substantially similar to the subject matter forming the basis of

the indictment [that] are probative to negate the inference that

the crucial conduct was … innocent”).

Unlike the District Court’s ruling with respect to the

Pozo Plot, however, the District Court’s decision to exclude

evidence of the Esteves Plot was not clearly rooted in a

flawed premise. Indeed, the Court spoke at length about its

concerns regarding the nature of the Esteves Plot evidence,

(see, e.g., Joint App. at 38 (explaining that if “there was a

conviction, I would believe … that that conviction was the

result of the Esteves evidence, because I don’t see how they

could humanly put that out of their mind and the purposes of

the cautionary instruction would be and then weigh the rest of

this case accordingly”)), and we cannot glean whether or not

its Rule 403 balancing was tainted by the mistaken distinction

it drew between subsequent and prior acts. Thus, it is

difficult to tell whether or not the Court’s judgment is entitled

to the deference ordinarily accorded a Rule 403 decision. See

Kellogg, 510 F.3d at 197 (stating the general maxim that

“judicial self-restraint” is desirable “when a Rule 403 analysis

of a trial court is reviewed” (citation and internal quotation

marks omitted)). However, because we will be reassigning

this case and directing the new district judge to determine

afresh the admissibility of the Esteves Plot evidence, see infra

Part II.C, we need not tackle the issue at this juncture. It

suffices to say that, in considering that issue on remand, the

judge should bear in mind that subsequent act evidence may

be properly admitted under Rule 404(b), although Rule 403

43

C. Reassignment

The government also asks that this case be given to

another district judge, and we agree, reluctantly, that

reassignment is appropriate. Our authority to direct the

reassignment of a case on remand is based on 28 U.S.C. §

455(a) and 28 U.S.C. § 2106. United States v. Bertoli, 40

F.3d 1384, 1411 (3d Cir. 1994). Under § 455(a), a judge

should no longer preside over a case when “a reasonable

person, with knowledge of all the facts, would conclude that

the judge’s impartiality might reasonably be questioned.” 26

United States v. Wecht, 484 F.3d 194, 213 (3d Cir. 2007)

(citation and internal quotation marks omitted). To warrant

reassignment under § 455(a), a case generally must involve

apparent bias deriving from an extrajudicial source, meaning

permits exclusion when the probative value of such evidence

is “substantially outweighed by a danger of … unfair

prejudice,” see Fed. R. Evid. 403, which, again, refers to

prejudice “based on something other than [the evidence’s]

persuasive weight.” Cruz-Garcia, 344 F.3d at 956.

All of this, of course, becomes essentially moot if the

new judge disagrees with the approach to severance that had

been followed here, though a limiting instruction might still

be warranted with respect to the jury’s consideration of the

Pozo Plot and the Esteves Plot in connection with the Kemo

Murder Counts.

26

“[T]he hypothetical reasonable person … must be

someone outside the judicial system because judicial insiders

… may regard asserted conflicts to be more innocuous than

an outsider would.” In re Kensington Int’l Ltd., 368 F.3d

289, 303 (3d Cir. 2004).

44

something above and beyond judicial rulings or opinions

formed in presiding over the case. See Liteky v. United

States, 510 U.S. 540, 555 (1994) (noting that “judicial rulings

alone almost never constitute a valid basis for a bias or

partiality motion [under § 455(a)]” since they rarely

“evidence the degree of favoritism or antagonism required …

when no extrajudicial source is involved”). Our supervisory

powers under § 2106, however, also permit reassignment and

are not necessarily constrained by that limitation. See id. at

554 (noting that “[f]ederal appellate courts’ ability to assign a

case to a different judge on remand rests not on the recusal

statutes alone, but on the appellate courts’ statutory power”

under § 2106 which “may permit a different standard” than

that applicable to § 455(a)). Notwithstanding the potential

differences between the standards for reassignment under

§ 455(a) and § 2106, we have typically reviewed requests for

reassignment under § 2106 “under an ‘appearance of

impartiality’ standard” like that applicable in the § 455(a)

context. Bertoli, 40 F.3d at 1414 (citation omitted); see Gov’t

of the V.I. v. Walker, 261 F.3d 370, 376 (3d Cir. 2001)

(exercising the supervisory power to reassign a case because

the “conduct and comments of the trial judge … ma[d]e it

exceedingly difficult to resurrect an appearance of

impartiality”).

Although reassignment is an extraordinary remedy that

should seldom be employed, see United States v. Higdon, 638

F.3d 233, 248 (3d Cir. 2011) (recognizing that reassignment

should “be considered seriously and made only rarely”

(citation and internal quotation marks omitted)), we conclude

that it is appropriate in this case despite our sincere respect

for the District Judge who has presided to this point. Key to

our decision is the District Court’s repeated expressions of

45

discomfort with the manner in which the Indictment pulls the

various criminal acts, including the witness-tampering plots,

together under the umbrella of RICO charges. That

discomfort manifested itself when the Court entered its first

dismissal of the RICO counts. While the Court pointed to

what may be called “equitable or logistical concerns,”

Bergrin, 650 F.3d at 274, in examining “the sufficiency of the

[then-existing indictment’s] allegations” under Criminal Rule

12, Bergrin, 707 F. Supp. 2d at 509, it was not explicit about

how prominent a role those concerns played in its decision. It

now appears that the Court ruled as it did, at least in part,

because it believes that it is impossible for Bergrin to get a

fair trial on the RICO counts due to the very nature of RICO,

allowing, as it does, for multiple criminal acts to be charged

as a pattern of racketeering activity. 27 (See Joint App. at 4458

(the Court’s answer, in response to the government’s

invitation to dismiss the RICO counts if the Court believed

they could not be fairly tried, that “I did that once … because

at the time I still was concerned about the RICO allegations,

quite frankly, mostly for the same reason”).)

The Court expressed that same fear when, after the

government appealed the evidentiary rulings relating to the

Kemo Murder Counts, it balked at the government’s request

to try the RICO counts. In suggesting that a trial of those

27

We recognize that the District Court’s primary

concern here was that Kemo’s murder, for which Bergrin

faces a potential life sentence, is an integral component of the

Indictment’s RICO counts. See supra note 6. We do not

intend to suggest that, in different circumstances, there would

be hostility to trying RICO counts simply because they allow

the government to address multiple criminal acts in one

charge.

46

counts would be a fundamentally unfair and inefficient use of

prosecutorial resources, the Court said:

And now you’re sitting here saying:

Judge, we want to do it that way. We’re going

to bring back these guys and we’re going to

spend all this taxpayers’ money, all these people

in witness protection, they’re going to come

flying in, coming in, we’re going to go through

all of this, when you have an option. You have

an option of a five to seven-week trial, clean,

probably a conviction if the evidence is what I

see it is. I mean, you know, and yet you’re

insisting on trying to prove an enterprise, a

pattern, all these predicate acts, confusing a

jury, bringing in these guys again, and he’ll be

cross-examining them again. For what?

(Id. at 4461-62.)

To mitigate that perceived inequity, and in an apparent

effort to dissuade the government from seeking to try the

RICO counts, the Court tried to assure the government that

“there would be a sentence that would reflect the severity of”

the Indictment’s other charges if it secured a conviction on

those charges. (Joint App. at 4460.) In that same colloquy,

the Court did not dispute the government’s assertion that the

Court had “all but accused [the prosecution of] having

wrapped [the Kemo murder and the Esteves Plot] in the

Indictment in order to prevent [Bergrin] from getting a fair

trial” (id. at 4450), confirming instead that, in the Court’s

view, it would indeed “have been inherently unfair to have

[Bergrin] convicted under … RICO” (id. at 4463). Most

47

recently, in ordering a second severance, the Court made clear

its view that “trying Bergrin for his alleged involvement in

the [Kemo] murder conspiracy with extensive evidence from

the [Esteves Plot] … would be fundamentally unfair and

improper.” (Id. at 67.)

The problem with that view is that presenting the

witness-tampering allegations as part of a related pattern of

racketeering activity is exactly what the Indictment and RICO

allow. The Indictment contains valid RICO charges which

allege the Kemo murder along with the Esteves Plot and the

Pozo Plot, and, if the government ever brings its RICO

charges in this case to trial, it will necessarily introduce

evidence of those murder plots to meet its burden of proof.

We do not doubt the depth of the District Court’s

commitment to ensuring a fair trial for all parties, and the

Court’s concern for the rights of a criminal defendant is

commendable. But, as we have already held, Congress

validly paved a path for prosecutions like the one charged in

the Indictment. See Bergrin, 650 F.3d at 276 (reversing the

dismissal of substantially similar RICO counts). It is not a

court’s prerogative to construct a detour around RICO simply

because the court is uncomfortable with how that statute may

“significantly alter[] the way trials are conducted in cases that

involve racketeering acts committed by members of an

enterprise.” 28 Id. at 275; see United States v. Vitillo, 490 F.3d

28

We do not mean to imply that a district court is

powerless in a RICO case to consider severance orders. On

the contrary, as we said the first time we had this case, the

District Court could appropriately “discuss[] joinder and

severance under Rules 8 and 14 of the Federal Rules of

Criminal Procedure” when presented with the former iteration

48

314, 320 (3d Cir. 2007) (“It is well-established that ‘[a]n

indictment returned by a legally constituted and unbiased

grand jury … if valid on its face, is enough to call for trial of

the charge on the merits.” (alterations and emphasis in

original) (quoting Costello v. United States, 350 U.S. 359,

363 (1956))).

Ultimately, in light of the District Court’s statements –

both before and after the earlier appeal in this case – about a

perceived unfairness in trying the various witness-tampering

counts together, we believe that the Court’s “impartiality

might reasonably be questioned,” Wecht, 484 F.3d at 226

(citation and internal quotation marks omitted), and will

therefore order that this case be reassigned under § 2106, see

Bertoli, 40 F.3d at 1414 (noting that supervisory power

reassignment has typically been applied “under an

‘appearance of impartiality’ standard” (citation omitted)).

Because the Court’s discomfort with the Indictment may well

have prompted its evidentiary and case management rulings

(see Joint App. at 4458 (the Court’s statement that it initially

dismissed the RICO counts “because at the time I still was

of the Indictment. Bergrin, 650 F.3d at 276. That authority,

of course, is not unyielding or unbounded, see United States

v. Giannattasio, 979 F.2d 98, 100 (7th Cir. 1992) (Posner, J.)

(rejecting a court’s direction to a prosecutor to “select five …

counts for prosecution” of a fifteen count indictment and

proceed to trial on those counts alone, and noting that “[a]

judge in our system does not have the authority to tell

prosecutors which crimes to prosecute or when to prosecute

them”), but, as we have already noted, we need not and do not

attempt to delineate its contours as applied to the severance

orders entered in this case. See supra note 20.

49

concerned about the RICO allegations, quite frankly, mostly

for the same reason”)), we direct the judge to whom this case

is reassigned to consider anew whether the Indictment should

be severed in any respect and, as necessary, the extent to

which evidence of the Esteves Plot and the Pozo Plot can

properly be used to prove the government’s case against

Bergrin on the Kemo Murder Counts. 29 See 28 U.S.C. § 2106

(affording courts of appeals the authority to “require such

further proceedings to be had as may be just under the

circumstances”); cf. Rimbert v. Eli Lilly & Co., 647 F.3d

1247, 1252 (10th Cir. 2011) (stating that the “law of the case

doctrine has no bearing on the revisiting of interlocutory

orders, even when a case has been reassigned from one judge

to another”); Perez-Ruiz v. Crespo-Guillen, 25 F.3d 40, 42

(1st Cir. 1994) (“Interlocutory orders … remain open to trial

court reconsideration, and do not constitute the law of the

case.”).

III. Conclusion

For the foregoing reasons, we will vacate the District

Court’s decision to exclude evidence of the Pozo Plot and

will direct the Chief Judge of the District Court to reassign

this matter.

29

Although we have vacated the District Court’s

decision to exclude evidence of the Pozo Plot from Bergrin’s

retrial, we note, again, that, depending on what is offered in

evidence, the new judge may well be asked to determine the

admissibility of the Pozo Plot evidence with respect to the

Kemo Murder Counts and will, in that event, need to conduct

an appropriate Rule 403 balancing. See supra note 25.

50