United States v. Cabrera

19-3363-cr
United States v. Cabrera

United States Court of Appeals
for the Second Circuit
AUGUST TERM 2020
No. 19-3363-cr

UNITED STATES OF AMERICA,
Appellee,

v.

JOHN E. CABRERA,
Defendant-Appellant.

ARGUED: SEPTEMBER 15, 2020
DECIDED: SEPTEMBER 8, 2021

Before: JACOBS, LYNCH, SULLIVAN, Circuit Judges.

John Cabrera appeals from the judgment of the United States District

Court for the Southern District of New York (Failla, J.) convicting him of four

counts of distributing and possessing with intent to distribute fentanyl. On

appeal, Cabrera argues that the jury instruction misstated the burden on the

inducement element of his entrapment defense, and that the district court abused
its discretion by admitting a special agent’s opinion that Cabrera was an

experienced drug dealer. The effect of these two errors was reciprocal.

We VACATE and REMAND for a new trial.

JUDGE SULLIVAN dissents in the Court’s opinion, and files a dissenting

opinion.

____________________

DANIEL HABIB, Federal Defenders of New York, New
York, NY, for Defendant-Appellant John Cabrera.

DANIELLE R. SASSOON, Assistant United States
Attorney (Dominic Gentile, Rebekah Donaleski, Thomas
McKay, Assistant United States Attorneys, on the brief),
United States Attorney’s Office for the Southern District
of New York, New York, NY, for Appellee.

DENNIS JACOBS, Circuit Judge:

John Cabrera engaged in four drug transactions with his barber, who was

a government informant. Cabrera’s sole defense was entrapment, which (as the

district court acknowledged) was a close call as to the element of inducement.

He appeals chiefly on the grounds that: the charge misstated his burden by

requiring the defendant to establish that the government initiated the crime; and

that testimony from a special agent, who opined that Cabrera was an

2
experienced drug dealer, was inadmissible as lay opinion under Federal Rule of

Evidence 701.

Cabrera and his barber gave opposite accounts of who first proposed

partnering in the drug trade. It was therefore crucial that the charge accurately

state Cabrera’s burden: the slight burden of adducing “some credible” evidence

that the government initiated the crime. The charge overstated that burden,

effectively requiring that the jury weigh the evidence and definitively accept

Cabrera’s account as a precondition to considering predisposition.

Compounding the prejudice to Cabrera’s defense, the special agent’s

testimony that Cabrera was an “experienced” drug dealer was inadmissible as

lay opinion. And it undercut Cabrera’s account of how the transactions with his

barber originated, as well as his lack of predisposition to deal.

We vacate Cabrera’s conviction and remand for a new trial.

I

Cabrera is a legal permanent resident who came to New York from the

Dominican Republic in 2013, when he was 20. After arriving, Cabrera held

3
several minimum-wage jobs before becoming a carpenter. Around 2014, he met

a barber and fellow Dominican immigrant named Marcos. Cabrera’s apartment

was located near the barbershop where Marcos worked, and Cabrera began

visiting him weekly for a shave and haircut.

Marcos had immigrated to the United States in 1992 when he was 17; but

in 2001 he was deported after serving a sentence on a drug conviction. He

reentered illegally that same year. In 2016 Marcos became a paid informant for

the Drug Enforcement Administration (“DEA”). He received cash payments and

deportation deferrals renewed annually so long as he remained an informant.

(Put another way, Marcos was compensated and deferred so long as he was

useful, that is, so long as he had people on whom he could inform.)

Over a two month period in late 2017, Cabrera and Marcos partnered to

sell drugs. Cabrera delivered pills containing fentanyl, and Marcos, under the

DEA’s direction, paid Cabrera and pretended to resell the pills to customers in

North Carolina. There were five transactions. On September 7, Cabrera gave

Marcos a small free sample. Six days later, Cabrera sold Marcos 200 pills for

$3,000; a week later, 198 pills for $3,000; and another six days later, 397 pills for

4
$3,000 up front and $3,000 in two days. Following a month-long gap, they met

again on October 27 to exchange 1,000 pills for $15,000, and agents arrested

Cabrera; he had 1,100 pills on him.

The government charged Cabrera in a four-count indictment. Counts I

and II were for distributing and possessing with intent to distribute fentanyl on

September 13 and 21 in violation of 21 U.S.C. §§ 812, and 841(a)(1) and (b)(1)(C).

Counts III and IV were for distributing and possessing with intent to distribute

40 grams or more of fentanyl on September 27 and October 27 in violation of 21

U.S.C. §§ 812, and 841(a)(1) and (b)(1)(B). At trial, Cabrera and Marcos gave

sharply divergent testimony about how their partnership began.

Cabrera--conceding he sold the pills to Marcos--claimed he was entrapped.

He testified as follows. Marcos asked him several times during barbershop visits

to supply drugs; Cabrera refused, telling Marcos he already made sufficient

money as a carpenter. But Marcos renewed his invitation approximately five or

six times until, in early 2017, Cabrera relented, having become desperate after

losing his job, girlfriend, and apartment--and confiding his problems to Marcos.

Cabrera began searching for a supplier; after six months, he found one at a

5
nightclub, and told Marcos that he was ready: Cabrera would serve as the

middleman, earning $2 from the supplier for each pill that he sold to Marcos,

who would then resell to (fictitious) customers in North Carolina.

Marcos’s version of events, as follows, was different in every material

respect. Marcos first learned in 2016 that Cabrera dealt drugs when Cabrera told

him that his supplier had unfortunately been arrested. At that point, Cabrera

and Marcos had known each other for eight months. Cabrera then disappeared

for a year, during which time Marcos became an informant. When Cabrera

returned to the barbershop in September 2017, he told Marcos that he was back

in business. Cabrera was looking to sell oxycodone pills and asked Marcos if he

knew any buyers. When Marcos said that he knew some in North Carolina,

Cabrera proposed that the two do business together. Marcos promptly contacted

his handlers at the DEA.

Trial evidence included government recordings of meetings and phone

calls between Cabrera and Marcos, all of which post-date the agreement to

partner. Cabrera boasted of his experience selling drugs, telling Marcos, for

example, that “with me there will always be many good things,” and “I’m only

6
24 . . . but I’m not new at this.” App’x 80–81. Cabrera and Marcos occasionally

pushed each other to do bigger deals. At their second meeting (their first sale),

Marcos voiced frustration at being unable to buy pills in greater bulk; and soon

after, over the phone, Cabrera expressed disappointment about how long it was

taking to plan their next deal. On a call following their third meeting, Cabrera

urged Marcos to visit North Carolina more frequently; when Marcos demurred,

Cabrera offered to give him more pills on credit. Later, Marcos asked Cabrera to

locate a pure form of heroine called China White, but this time it was Cabrera

who declined.

Cabrera went silent after their September 29 meeting. He testified that he

wanted to cut ties with Marcos because he regretted breaking the law and feared

he was under DEA surveillance. Marcos left multiple voicemails throughout

October, pushing Cabrera to resume deals. At the DEA’s direction, Marcos

showed up at Cabrera’s workplace to ask where he had been (Marcos does not

recall being wired on that occasion). On October 25 at the barbershop, they

planned the fifth deal in an unrecorded meeting; according to Marcos, Cabrera

7
was scared he had been followed and insisted on increasing the deal to 1,000

pills.

Special Agent Daniel Son, who had surveilled Cabrera at the September 21

and 27 deals, also testified. Over Cabrera’s objection, Agent Son opined on

rebuttal that Cabrera, unlike the “average drug dealer,” appeared to be

“experienced” because he had employed countersurveillance driving techniques

(which consisted of really bad driving). App’x 649. For support, Agent Son cited

his experience conducting narcotics investigations.

After a six-day trial, the jury convicted Cabrera on all counts. Cabrera was

sentenced to concurrent terms of 48 months’ imprisonment on each count.

II

The first issue is whether the jury instruction on Cabrera’s entrapment

defense contained error, specifically as to the element of inducement. “We

review a jury instruction challenge de novo.” United States v. Coppola, 671 F.3d

220, 247 (2d Cir. 2012) (citation omitted). “Instructions are erroneous if they

mislead the jury as to the correct legal standard or do not adequately inform the

8
jury of the law.” Hudson v. New York City, 271 F.3d 62, 67 (2d Cir. 2001)

(quotation omitted).

A

The affirmative defense of entrapment consists of “two related elements:

government inducement of the crime, and a lack of predisposition on the part of

the defendant to engage in the criminal conduct.” Mathews v. United States, 485

U.S. 58, 63 (1988) (citations omitted). “[W]hen a defendant has presented

credible evidence of inducement by a government agent, the government has the

burden of proving beyond a reasonable doubt that the defendant was

predisposed to commit the crime.” United States v. Flores, 945 F.3d 687, 717 (2d

Cir. 2019) (citing Jacobson v. United States, 503 U.S. 540, 548–49 (1992)).

That approach balances two considerations. The government may not

manufacture crime where there would be none by “implant[ing] in the mind of

an innocent person the disposition to commit the alleged offense and induc[ing]

its commission . . . .” Jacobson, 503 U.S. at 553 (quoting Sorrells v. United States,

287 U.S. 435, 442 (1932)). At the same time, “stealth and strategy are necessary

9
weapons in the arsenal of the police officer.” Sherman v. United States, 356 U.S.

369, 372 (1958). The entrapment defense thus seeks to protect the “unwary

innocent” while leaving room for investigative techniques that catch the

“unwary criminal who readily availed himself of the opportunity to perpetrate

the crime.” Mathews, 485 U.S. at 63 (quotation omitted).

The first element--inducement--is relatively straightforward. It happens

when the government has “initiated the crime.” United States v. Brand, 467 F.3d

179, 190 (2d Cir. 2006) (quoting United States v. Mayo, 705 F.2d 62, 67 (2d Cir.

1983)). More broadly, inducement covers “soliciting, proposing, initiating,

broaching or suggesting the commission of the offence charged.” United States

v. Sherman, 200 F.2d 880, 883 (2d Cir. 1952) (Hand, J.). The degree of pressure

exerted, and the type, are matters that bear mainly on the element of

predisposition. United States v. Dunn, 779 F.2d 157, 158 (2d Cir. 1985); United

States v. Pugliese, 346 F.2d 861, 863–64 (2d Cir. 1965); see also United States v.

Mayfield, 771 F.3d 417, 437 (7th Cir. 2014) (“The nature of the government

inducement is significant chiefly as evidence bearing on predisposition.”

(quotation omitted)).

10
 We have long held that the jury instruction on inducement should not

specify a burden of proof; it should require only “some” or “credible” evidence

the government initiated the crime. United States v. Braver, 450 F.2d 799, 805 (2d

Cir. 1971) (Feinberg, J.); United States v. Valencia, 645 F.2d 1158, 1166 (2d Cir.

1980); United States v. Groob, 451 F.2d 1210, 1210–11 (2d Cir. 1971).

At the same time, we have previously characterized the defendant’s

burden to establish inducement as a burden of proof by a preponderance.

United States v. Williams, 23 F.3d 629, 635 (2d Cir. 1994); see also Braver, 450

F.2d at 802. We now recognize that this “preponderance” burden is inconsistent

with the jury instruction we have endorsed. As our sister circuits recognize, a

“some evidence” instruction on inducement communicates a burden of

production, not one of persuasion. See, e.g., Mayfield, 771 F.3d at 440; United

States v. Isnadin, 742 F.3d 1278, 1297 (11th Cir. 2014); United States v. Gurolla,

333 F.3d 944, 955 (9th Cir. 2003). And in this Circuit, “some evidence” describes

a burden of production in the context of burden shifting. United States v.

Archer, 671 F.3d 149, 173–74 (2d Cir. 2011). “Some evidence” is evidence that is

11
detected or recognized--without being weighed, as would be needed to find a

thing by a preponderance.

Similarly, we have sometimes conflated the defendant’s burden to obtain

an entrapment charge with the defendant’s burden at trial. Brand characterized

the burden at trial as a preponderance of the evidence, but a burden that is

nonetheless “relatively slight.” 467 F.3d at 190 (quoting Mayo, 705 F.2d at 67).

That caveat, however, derived from United States v. Henry, in which the issue

was whether the defendant was entitled to an entrapment charge at all. 417 F.2d

267, 269–70 (2d Cir. 1969). This conflation arose here: the government argued to

the district court that “some evidence is really, goes more to whether they meet

their threshold for getting the jury instruction.” App’x 688.

In light of this confusing – and inconsistent – case law describing the

defendant’s burden to establish inducement, we now reconsider the burden that

a defendant bears at trial and the proper jury instruction that should accompany

it.

We hold that the defendant has the burden to produce “some credible”

evidence--but need not prove by a preponderance of the evidence--that the

12
government induced him to commit the crime. This formulation best aligns with

much of our recent precedent and eliminates any conflict with the language we

have previously endorsed in jury instructions. See Flores, 945 F.3d at 717 (citing

Jacobson, 503 U.S. at 548–49); United States v. Kopstein, 759 F.3d 168, 174 (2d Cir.

2014) (citing United States v. Bala, 236 F.3d 87, 94 (2d Cir. 2000)); see also United

States v. Salerno, 66 F.3d 544, 547 (2d Cir. 1995). Compared to a “some evidence”

instruction, the phrase “some credible evidence” makes explicit what is implicit--

that a jury need not consider evidence it finds unworthy of credit or belief. 1

By definition, “some credible” evidence suggests a burden of production.

And, as a matter of administration, requiring a jury to apply two different

burdens of proof to a single defense would “tend[] to distract the jury from the

real issue and may result in the imposition of too heavy a burden on the

defendant.” Dunn, 779 F.2d at 160. “[T]he ultimate question basic to all claims

of entrapment” is whether the defendant was “ready and willing to commit the

1 This opinion has been circulated to all the judges of the Court prior to filing.
See Jon O. Newman, In Banc Practice in the Second Circuit, 1984-1988, Brook. L.
Rev. 355, 367–68 (1989) (“On occasion . . . a panel opinion is circulated prior to
filing when the panel deems it important for the full court to be aware of what
the panel proposes to say.”).

13
offense if given an opportunity to do so.” United States v. Martinez-Carcano, 557

F.2d 966, 970 (2d Cir. 1977). Predisposition--not inducement--is the “principal

element” of entrapment. Mathews, 485 U.S. at 63 (quoting United States v.

Russell, 411 U.S. 423, 433 (1973)). Inducement is merely the threshold inquiry for

whether “the defense of entrapment is at issue.” Jacobson, 503 U.S. at 549.

Traditionally, the defendant’s burden on an affirmative defense, when the

government has the ultimate burden of persuasion, is to produce evidence

creating an issue of fact. See Archer, 671 F.3d at 173 (collecting examples). There

is no reason to depart from that principle here. A defendant’s prima facie case of

inducement raises an issue of fact: whether the defendant “likely would have

never run afoul of the law” but for the hand of the government. Jacobson, 503

U.S. at 549; see also Sherman, 356 U.S. at 376. The government must then justify

its conduct, see Henry, 417 F.2d at 270, and undertake its proper burden to prove

“beyond reasonable doubt that the defendant was disposed to commit the

criminal act prior to first being approached by Government agents,” Jacobson,

503 U.S. at 549.

14
 B

The charge on Cabrera’s entrapment defense implied that the jury could

not consider predisposition unless it made a finding that the government “did

initiate” the crime. That was legal error. In full and relevant part, the instruction

referenced the proper standard (“any evidence”) but deviated:

You should first consider whether there is any evidence that the
government, acting through the confidential informant, induced Mr.
Cabrera to commit the offense you are considering by taking the first
step that led to the criminal act. Inducement may include soliciting,
proposing, initiating, broaching, or suggesting the submission of the
offense. If you find that there is no evidence that the government
induced Mr. Cabrera to commit the offense you are considering,
then you should find that there was no entrapment and you need
not consider this defense any further. If, on the other hand, you find
that the government did initiate the offense you are considering,
then you must decide whether the government has proven beyond a
reasonable doubt that Mr. Cabrera was already predisposed to
commit the offense.

App’x 807.

This instruction suggested that Cabrera had to satisfy a burden of proof.

No matter what standard of proof the jury applied or intuited--whether it was a

preponderance, beyond a reasonable doubt, or a standard from the jury’s

imagination--there was error. Cabrera was obliged to produce no more than

15
“some credible” evidence of inducement. The error was considerable. Only one

standard was referenced: the government’s burden to prove predisposition

beyond a reasonable doubt. The jury “would naturally infer . . . that was the

standard they were to apply, not only to the government but also to [Cabrera].”

Pugliese, 346 F.2d at 863 (quotation omitted).

The government effectively concedes this error, focusing instead on other

passages in the charge that purportedly conveyed the gist of defendant’s “slight”

burden. The government contends that the verb “did initiate” was “necessarily

defined by the preceding [instruction] to consider whether there is ‘any evidence’

of inducement and to terminate the entrapment inquiry only if there is ‘no

evidence’ of inducement.” Appellee Br. 17.

However, the final two sentences left the jury with a dichotomy that

reinforced the charge’s basic error: the jury could either (1) reject the entrapment

defense if “you find . . . no evidence” of inducement, or (2) move on to

predisposition if, “on the other hand,” “you find that the government did

initiate” the offense. That dichotomy has the natural tendency to confuse and

skew the jury’s decision-making.

16
 The government argues in the alternative that Cabrera forfeited his

objection by not explicitly requesting the “some evidence” language. But

Cabrera submitted a proposed instruction with that precise wording, and he later

objected to the district court’s proposal by asking the court to replace “find that”

with “find evidence that.” That is all counsel was required to do to “inform the

court of the specific objection and the grounds for the objection before the jury

retires to deliberate.” Fed. R. Crim. P. 30(d). Counsel need not pester the judge

to preserve the objection.

III

We also agree with Cabrera that the district court abused its discretion by

admitting Special Agent Son’s testimony as lay opinion. Agent Son testified that

he believed Cabrera was an “experienced” drug dealer. His testimony was

expert opinion, in violation of Federal Rule of Evidence 701.

We review the district court’s evidentiary decisions for abuse of discretion.

United States v. Garcia, 413 F.3d 201, 210 (2d Cir. 2005). “A district court abuses

its discretion when it bases its ruling on an erroneous view of the law or on a

17
clearly erroneous assessment of the evidence, or renders a decision that cannot

be located within the range of permissible decisions.” United States v. Vayner,

769 F.3d 125, 129 (2d Cir. 2014) (quoting Porter v. Quarantillo, 722 F.3d 94, 97 (2d

Cir. 2013)).

Rule 701 provides:

If a witness is not testifying as an expert, testimony in the form of an
opinion is limited to one that is:

(a) rationally based on the witness’s perception;

(b) helpful to clearly understanding the witness’s testimony or
to determining a fact in issue; and

(c) not based on scientific, technical, or other specialized
knowledge within the scope of Rule 702.

Fed. R. Evid. 701.

The “specialized knowledge” restriction in Part (c) “prevent[s] a party

from . . . conferring an aura of expertise on a witness without satisfying the

reliability standard for expert testimony set forth in Rule 702 and the pre-trial

disclosure requirements set forth in Fed. R. Crim. P. 16.” Garcia, 413 F.3d at 215

(citing Fed. R. Evid. 701, Advisory Committee Note (2000)). “If the opinion rests

in any way upon scientific, technical, or other specialized knowledge, its

18
admissibility must be determined by reference to Rule 702, not Rule 701.” Id.

(quotation omitted). Accordingly, lay opinion testimony is “limited to opinions

that result from a process of reasoning familiar in everyday life.” Flores, 945 F.3d

at 707 (quoting United States v. Cuti, 720 F.3d 453, 457 (2d Cir. 2013)).

Agent Son testified that when Cabrera was driving to sell drugs to Marcos,

Agent Son saw Cabrera “conducting countersurveillance”; that is, “[e]xcessively

speeding, [making] erratic lane changes, making U turns and then making

another U-turn, those types of maneuvers.” App’x 648–49. Cabrera does not

challenge that (seemingly unobjectionable) testimony.

Cabrera’s challenge is focused on Agent Son’s conclusion: that Cabrera

was “experienced” as compared to the “average drug dealer”:

Q. In your experience was it unusual to see someone engaged in a
drug transaction doing this?

A. Yes. It’s unusual.

Q. And why was that?

A. The average drug dealer does not know that they’re being
followed. They’re [sic] don’t do those type of techniques.

Q. And based on what you saw, those--from those
countersurveillance techniques, what conclusion did you draw?

19
 A. That the defendant was experienced to know that he knows
some of our law enforcement techniques and to deploy those
countersurveillance techniques, to lose us or lose the tail.

App’x 649.

Garcia is closely analogous. We ruled that the district court abused its

discretion in admitting a DEA agent’s opinion that the defendant was a member

of a drug conspiracy: the witness’s “reasoning process was not that of an average

person in everyday life; rather, it was that of a law enforcement officer with

considerable specialized training and experience in narcotics trafficking.” 413

F.3d at 216–17.

Likewise, Agent Son drew upon his specialized knowledge and experience

as a DEA detective to infer that Cabrera was more experienced than your

average drug dealer. That is, Cabrera had done this frequently--testimony that,

not coincidentally, had obvious bearing on Cabrera’s entrapment defense. Agent

Son’s opinion did not relate events, or describe Cabrera’s atypical driving

patterns, or contextualize the relationship between Cabrera and Marcos. See id.

at 213–14. It was therefore an abuse of discretion to admit his testimony as lay

opinion under Rule 701. See also United States v. Figueroa-Lopez, 125 F.3d 1241,

20
1246 (9th Cir. 1997) (finding inadmissible under Rule 701 a law enforcement

officer’s testimony that the defendant’s countersurveillance driving was

consistent with that of an “experienced drug trafficker”).

The government points out that a lay person could have reached the same

conclusion as Agent Son--that Cabrera’s driving was suspicious. But Agent Son

went well beyond that; he inferred from Cabrera’s driving that Cabrera must be

one of those experienced drug-dealers who had mastered the technique of

evading law enforcement. Cf. United States v. Grinage, 390 F.3d 746, 750 (2d Cir.

2004) (explaining that jurors were not “helped” within the meaning of Rule 701

by opinion testimony that, in addition to telling them “what was in the

evidence,” also told them “what inferences to draw from it”).

Although a lay witness may in some circumstances call on special

experience in aid of the witness’s perceptive faculties, that is not the case here.

Flores ruled that the district court was within its discretion in admitting a DEA

informant’s lay opinion that a substance was cocaine. 945 F.3d at 709. The

witness’s testimony was based on his experience working for a drug cartel and

having examined the substance by smell, touch, and appearance. Id. at 708–09.

21
The district court had made clear that the witness was no expert and cabined his

testimony; accordingly, the witness testified to whether he believed the

substance was or was not cocaine based on practical tests of perception, rather

than opining on its purity. Id.

Agent Son, however, reached his opinion through an opaque, intuitive

process grounded in some kind of specialized knowledge as to how your average

drug dealer typically behaves compared to a drug dealer who is experienced. A

lay person is unfamiliar with law enforcement surveillance techniques and

incapable of inferring that a suspect’s driving maneuvers evince (1) experience

with evading those techniques and, consequently, (2) experience dealing drugs.

A juror might as easily ascribe those maneuvers to watching the movies, or to a

paranoia born of inexperience.

IV

We must vacate and remand for a new trial. The error in the jury

instruction and the improperly admitted testimony, considered together,

prejudiced Cabrera’s defense of entrapment--and he had no other.

22
 Because Cabrera objected to the charge at trial and has raised the same

claim on appeal, we review for harmless error. United States v. Botti, 711 F.3d

299, 308 (2d Cir. 2013) (citation omitted). We vacate when “there was a

prejudicial error” in the charge viewed as a whole, assessing whether “it is clear

beyond a reasonable doubt that a rational jury would have found the defendant

guilty absent the error.” United States v. Atilla, 966 F.3d 118, 123–24 (2d Cir.

2020) (quoting United States v. Aina-Marshall, 336 F.3d 167, 170 (2d Cir. 2003)

and Neder v. United States, 527 U.S. 1, 18 (1999)). We similarly review an

evidentiary error for whether it substantially influenced the verdict, Garcia, 413

F.3d at 210 (quotations omitted); but because the errors had reciprocal effect--and

the standards have no discernible difference--we consider whether the two

together were harmless.

Harmlessness turns on whether it is clear Cabrera would have failed on his

entrapment defense absent the errors; i.e., whether it is clear a rational jury

would still have found: (1) that Cabrera presented no credible evidence of

inducement, or (2) that the government proved predisposition beyond a

reasonable doubt.

23
 At trial, inducement was vigorously contested. Marcos and Cabrera gave

dueling accounts of who initiated the venture. Marcos testified that Cabrera

proposed selling drugs in a bid to revive Cabrera’s drug-dealing business after

losing his supplier and disappearing for a year. Cabrera testified that Marcos

nagged him to sell drugs until Cabrera relented under personal and financial

stress.

A properly instructed jury could easily have found that Cabrera put forth

some credible evidence of inducement. It was Marcos’s word against Cabrera’s.

Marcos had compelling needs to feed the government new drug dealers; and

Cabrera impeached Marcos with records of phone calls between them during the

year Marcos claimed Cabrera had disappeared. The district court itself was “in

equipoise” “on the issue of who initiated the transactions.” App’x 1083. A

similarly dubious jury would necessarily have found at least some of Cabrera’s

testimony credible.

The government argues that the difference between the parties’ proposed

instructions on inducement was too fine to matter and that at one point the

instruction referenced “any evidence” of inducement, a correct statement of the

24
law. We rejected this argument in Part II.B; considering the entire instruction,

the error was the non-trivial difference between burdens of production and

persuasion. The unusual feature of this case is that improper testimony

reinforced the error’s prejudicial effect. Agent Son testified with the authority of

a federal agent that he believed Cabrera was an “experienced” drug dealer,

which necessarily invited the inference that Marcos’s account of who initiated

the crime was right, Cabrera’s was wrong, and, further, that Cabrera was

predisposed. The testimony cleared the way for the jury to reject Cabrera’s

defense of entrapment with ease.

Nor is Cabrera’s predisposition clear beyond a reasonable doubt. To prove

predisposition, the government may present evidence of, but not limited to, “(1)

an existing course of criminal conduct similar to the crime for which [the

defendant] is charged, (2) an already formed design on the part of the accused to

commit the crime for which he is charged, or (3) a willingness to commit the

crime for which he is charged as evidenced by the accused’s ready response to

the inducement.” Flores, 945 F.3d at 717 (quoting Salerno, 66 F.3d at 547).

25
 The government first contends that Cabrera was keen to do deals with

Marcos. According to Marcos’s testimony and recorded conversations, Cabrera

gave Marcos a free sample before the first sale, told Marcos that “with me there

will always be many good things,” and gave similar assurances. App’x 80–81,

875, 892, 885. Cabrera also admitted at trial that he was eager to sell Marcos

more pills more frequently after their first deal; eagerness which showed in the

quick succession and increasing size of new deals, and Cabrera’s wariness of

surveillance. App’x 576–77.

But the government’s argument is off target. What matters is Cabrera’s

“state of mind prior to” when they first broached transacting drugs. United

States v. Cromitie, 727 F.3d 194, 208 (2d Cir. 2013) (emphasis in original); see

Jacobson, 503 U.S. at 549 n.2 (“[T]he proposition that the accused must be

predisposed prior to contact with law enforcement officers is . . . firmly

established.”). The government’s evidence of eagerness lacks probative value as

to Cabrera’s state of mind at the time Marcos--in his capacity as informant--and

Cabrera made contact. As the government acknowledged, its evidence “pick[s]

up midstream” in Cabrera and Marcos’s venture. App’x 771. And Cabrera (if he

26
is to be believed) had repeatedly refused to partner in drug deals even before

they struck an agreement. It is therefore far from clear whether Cabrera’s

eagerness was “independent and not the product of the attention that the

Government had directed at [him].” Jacobson, 503 U.S. at 550. The

government’s only evidence of a “prompt response” to an early solicitation was

Marcos’s disputed claim that Cabrera returned to the barbershop and proposed

partnering in the drug business. United States v. Harvey, 991 F.2d 981, 993 (2d

Cir. 1993).

The government also contends that it proved Cabrera’s predisposition

with evidence showing that he was an established drug dealer. The government

cites: Cabrera’s advice to Marcos on how to evade detection by (inter alia)

speaking in code, avoiding police, changing phone numbers, and hiding drugs in

hidden car compartments; Cabrera’s touting of his experience selling drugs (as

heard on the recordings); and Cabrera’s possession of 1,104 pills when arrested

even though Marcos had only agreed to buy 1,000, from which the government

infers that Cabrera must have “had other drug customers to whom he was

selling those pills.” Appellee Br. 26.

27
 We disagree. Most notably, the government could identify no other

customers, even though it had visually surveilled Cabrera, collected historical

cell-cite location and call information, and searched his phone post-arrest.

Moreover, Cabrera’s advice to Marcos was not the counsel of a mastermind; it

could have been given by a novice. And the district court discounted Cabrera’s

boasts to Marcos as “puffery,” explaining that:

[I]t’s been my experience that folks involved in drug trafficking do
not discuss drug trafficking in the way that Mr. Cabrera did. . . . I
can’t say for sure whether he had other customers, but I don’t
necessarily think that the statements in the transcripts would give
me reason to believe that he did, because a lot of them were said
with a tone of voice and with language that suggested he was trying
to puff himself up as a more established drug dealer.

App’x 1083. The district court was similarly unimpressed with the 100 extra pills

Cabrera had at his arrest: “I can’t say whether those were for others or for

additional transactions with Marcos.” App’x 1085.

Given how thin was the government’s case, Agent Son’s testimony was

pivotal. By definition, an experienced drug dealer is predisposed. Without his

testimony, the government’s evidence clearly establishes only that Marcos was

28
Cabrera’s customer, not “that [Cabrera] himself was in the trade [of selling

narcotics].” Sherman, 356 U.S. at 375.

We must vacate. The errors prejudiced Cabrera’s only defense in this case,

and we cannot say with certainty that a rational jury would have rejected that

defense and convicted him notwithstanding.

CONCLUSION

We have considered the government’s remaining arguments and found

them to be meritless. For the foregoing reasons, we VACATE Cabrera’s

conviction and REMAND for a new trial.

29
RICHARD J. SULLIVAN, Circuit Judge, dissenting:

Although I agree with the majority that this Circuit’s case law concerning

the applicable standard for entrapment defenses has been a source of confusion,

and I have no objection to the new standard announced today concerning the

proper instructions for an entrapment defense, I cannot say that the district court’s

jury instruction here constituted error under even that new standard. Moreover,

while I agree with the majority that Agent Son’s testimony exceeded the

permissible bounds of lay opinion testimony, I am equally convinced that the

introduction of that testimony was harmless in light of the totality of the evidence

presented at trial. For these reasons, I would affirm Cabrera’s conviction in all

respects.

As an initial matter, the majority is correct to point out that our prior cases

have sown confusion regarding the burden of proof that a defendant bears in

establishing inducement when asserting an entrapment defense. As a result, I am

unopposed to the formulation of a new standard that will offer greater clarity to

district courts and parties moving forward. For nearly 70 years, our Court has

recognized that when a defendant asserts an affirmative defense of entrapment,

“two questions of fact arise: (1) did the agent induce the accused to commit the
offence charged in the indictment,” and “(2) if so, was the accused ready and

willing without persuasion . . . to commit the offence.” United States v. Sherman,

200 F.2d 880, 882 (2d Cir. 1952) (Hand, J.). “On the first question the accused has

the burden; on the second the prosecution has it.” Id. at 882–83. In the intervening

seven decades, we have consistently recognized that the defendant’s burden on

the first question involves proving government inducement “by a mere

preponderance of evidence.” United States v. Thomas, 351 F.2d 538, 539 (2d Cir.

1965); see also United States v. Brand, 467 F.3d 179, 190 (2d Cir. 2006) (holding that

“a defendant hoping to assert the entrapment defense bears the burden of

establishing inducement by a preponderance of the evidence”); United States v.

Braver, 450 F.2d 799, 803 (2d Cir. 1971) (determining that the “defendant’s burden

of proof [as to inducement] must be at least the ‘preponderance’ or ‘more-likely-

than-not’ standard”).

Nevertheless, we have also considered the struggle that juries may face in

distinguishing between the defendant’s burden of proving inducement “by a

preponderance of the evidence” and the government’s burden of proving

predisposition “beyond a reasonable doubt.” Braver, 450 F.2d at 803. And,

recognizing that the shifting burdens may indeed be confusing, we have suggested

2
that “it would be preferable for the district courts of this [C]ircuit to use an

entrapment charge that does not give to the jury two ultimate factual issues to

decide on two different burdens of persuasion imposed upon two different

parties.” Id. at 805. In addition, because this Circuit’s broad definition of

inducement “requires so little evidence to satisfy the defendant’s burden of proof,”

and since “production of ‘some evidence’ of government initiation almost always

satisfies it,” we have found “simplification of the charge on [entrapment]

appropriate.” Id. at 805; see Sherman, 200 F.2d at 883 (defining inducement to

include “soliciting, proposing, initiating, broaching or suggesting the commission

of the offence charged”). Accordingly, without replacing the preponderance

standard – indeed, while confirming its continued viability – we have determined

that “it will be enough to tell the jury that if it finds some evidence of government

initiation of the illegal conduct, the [g]overnment has to prove beyond a

reasonable doubt that the defendant was ready and willing to commit the crime.”

Braver, 450 F.2d at 805 (emphasis added); see id. at 804–05 (expressly declining to

hold that an instruction describing the preponderance burden would constitute

prejudicial error).

3
 To focus and clarify the jury’s analysis on entrapment, we now overrule by

“mini en banc” our longstanding precedent requiring defendants to prove

inducement by a preponderance of the evidence. See Doscher v. Sea Port Grp. Sec.,

LLC, 832 F.3d 372, 378 (2d Cir. 2016) (explaining that a three-judge panel may

overrule prior decisions of this Court after “circulat[ing] its opinion among all

active judges and receiv[ing] no objections to its filing”). In replacing that

standard, we draw from past discussions about jury instructions in cases like

Braver and announce that, going forward, a defendant seeking to assert an

affirmative defense of entrapment “has the burden to produce ‘some credible’

evidence – but need not prove by a preponderance of the evidence – that the

government induced him to commit the crime.” Maj. Op. at 12–13. I have no

particular objection to this new standard, and agree that it will likely provide

welcome clarity to district court judges grappling with our conflicting precedents. 1

1 While the rule we adopt today is certainly clearer than the patchwork of cases it replaces, it is
by no means the only way to effectuate a more coherent approach to the defense of entrapment.
Another option might be to adopt one of the more demanding standards endorsed by our Sister
Circuits. See, e.g., United States v. Bradfield, 113 F.3d 515, 522 (5th Cir. 1997) (requiring defendants
to make a “prima facie showing of both . . . lack of predisposition and true inducement by the
government” to be entitled to a jury instruction on entrapment, at which point the burden shifts
to the government to “prove beyond a reasonable doubt that the defendant was [pre]disposed”);
United States v. Mayweather, 991 F.3d 1163, 1176 (11th Cir. 2021) (explaining that trial courts must
first “determine if the defendant has met his initial burden of producing sufficient evidence of
government inducement,” and only if he meets this burden is he “entitled to have his defensive
theory of the case put before the jury with appropriate instructions from the trial judge” and “the

4
 But even under this new – and arguably less burdensome – standard, the

district court’s instruction was not, in my view, improper. As noted by the

majority, the district court instructed the jury on Cabrera’s entrapment defense as

follows:

You should first consider whether there is any evidence that the
government, acting through the confidential informant, induced Mr.
Cabrera to commit the offense you are considering by taking the first
step that led to the criminal act. Inducement may include soliciting,
proposing, initiating, broaching, or suggesting the submission of the
offense. If you find that there is no evidence that the government
induced Mr. Cabrera to commit the offense you are considering, then
you should find that there was no entrapment, and you need not
consider this defense any further. If, on the other hand, you find that
the government did initiate the offense you are considering, then you
must decide whether the government has proven beyond a
reasonable doubt that Mr. Cabrera was already predisposed to
commit the offense.

App’x at 806–07.

burden shifts to the government to prove the defendant’s predisposition to commit the crime
beyond a reasonable doubt”). Alternatively, to the extent we wish to maintain this Circuit’s
historically expansive notion of inducement – which “requires so little evidence” that
“production of ‘some evidence’ of government initiation almost always satisfies it,” Braver, 450
F.2d. at 805 – we might consider dispensing with an instruction on inducement altogether and
simply allowing the district court to play a gatekeeping role with respect to the defendant’s initial
burden of production. By expressly assigning this threshold inquiry for entrapment to the district
court alone, we would avoid any potential jury confusion over shifting burdens by “focus[ing]
the jury’s attention” exclusively “on the central issue presented by a claim of entrapment: Was the
defendant ready and willing to commit the offense if given an opportunity to do so?” United
States v. Dunn, 779 F.2d 157, 160 (2d Cir. 1995) (internal quotation marks omitted).

5
 The majority insists that “[n]o matter what standard of proof the jury

applied or intuited – whether it was a preponderance, beyond a reasonable doubt,

or a standard from the jury’s imagination – there was error,” Maj. Op. at 15, since

the court’s instruction “effectively requir[ed]” the jury to “weigh the evidence and

definitively accept Cabrera’s account as a precondition to considering

predisposition,” id. at 3. But after reviewing the district court’s jury instruction in

its totality, I see no reason to presume that the instruction here required more of

Cabrera than the production of “some credible” evidence. Maj. Op. 15; see United

States v. Ford, 435 F.3d 204, 210 (2d Cir. 2006) (“We do not review portions of jury

instructions in isolation, but rather consider them in their entirety to determine

whether, on the whole, they provided the jury with an intelligible and accurate

portrayal of the applicable law.” (internal quotation marks and alterations

omitted)).

For starters, the district court’s instruction did not state that Cabrera was

required to prove inducement by a preponderance of the evidence; in fact, the

instruction did not mention any burden with respect to inducement, as Braver

advised was prudent. See Braver, 450 F.2d at 805; Dunn, 779 F.2d at 160 (explaining

that references to “the defendant’s burden of proof with regard to inducement . . .

6
tend[] to distract the jury from the real issue and may result in the imposition of

too heavy a burden on the defendant”); United States v. Valencia, 645 F.2d 1158,

1166 (2d Cir. 1980), amended, 669 F.2d 37 (2d Cir. 1981) (“There is little risk that the

jury will impose too great a burden on defendants if no burden is mentioned at

all.”). It is therefore incorrect to suggest that omitting a standard of proof as to

inducement would cause the jury to “naturally infer” that it should apply the

beyond a reasonable doubt standard “not only to the government but also to

[Cabrera].” Maj. Op. at 15 (alteration in original). Certainly, the court’s instruction

did not expressly advise the jury that Cabrera had to show inducement “beyond a

reasonable doubt” – the harm that cases like Braver and its progeny were most

eager to avoid. See Dunn, 779 F.2d at 160.

Moreover, while the district court did not use the words “some credible

evidence” in its instruction to the jury, the court actually reduced – if not

eliminated – the risk that the jury would apply a standard higher than “some

credible evidence” by first directing the jury to focus on whether the defendant

put forth any evidence of inducement. App’x at 806. The charge next provided

that if the jury found “no evidence,” then – and, arguably, only then – could it

7
reject the defense without addressing the government’s proof of predisposition,

which clearly had to be established beyond a reasonable doubt. Id. at 807.

The majority’s view that the district court’s charge required the jury to

“definitively accept Cabrera’s account as a precondition to considering

predisposition,” Maj. Op. at 3, is thus far from the only plausible interpretation of

the charge. To the contrary, the more natural reading of the charge would suggest

a lower burden on Cabrera, permitting the jury to eschew a finding on

predisposition only if it found no evidence of inducement. Cf. United States v.

Groob, 451 F.2d 1210, 1210–11 (2d Cir. 1971) (rejecting the argument that a charge

requiring “credible evidence” of inducement might have caused the jury to

assume “the defense had to prove inducement beyond a reasonable doubt”).

Finally, even if it could be argued that the court’s jury instruction on inducement

were somehow erroneous, any error would have been harmless because, as

discussed in detail below, the government provided overwhelming evidence of

Cabrera’s predisposition to commit the crimes at trial.

The majority attempts to bolster its argument that the jury instruction at

issue was not harmless by linking it to Agent Son’s improper expert testimony

inferring the extent of Cabrera’s narcotics experience from his aberrant driving

8
maneuvers. See Maj. Op. at 22 (concluding that “[t]he error in the jury instruction

and the improperly admitted testimony, considered together, prejudiced

Cabrera’s defense of entrapment”). I agree with the majority that it was improper

for Agent Son to offer an opinion on Cabrera’s experience as a drug dealer based

merely on his observation of Cabrera’s driving tactics. I also agree that this Court

should apply the harmless error standard on appeal. But a review of the entire

record demonstrates that the improper testimony – whether or not it is paired with

the district court’s entrapment instruction – “had no substantial influence on the

jury verdict” and that any error was therefore harmless. United States v. Garcia,

413 F.3d 201, 210 (2d Cir. 2005).

Agent Son’s testimony was not the only – or even the strongest – evidence

of Cabrera’s predisposition. At trial, the government focused heavily on Cabrera’s

unfettered access to large quantities of drugs and his own statements made over

the course of his dealings with his barber, Marcos, who was the government’s

confidential source. In recorded conversations, Cabrera bragged that he was “only

24 but . . . not new” to dealing drugs, App’x at 885, and alluded to “old customers”

to whom he had been selling drugs for more than two years, id. at 987.

9
 Other evidence introduced at trial bore this out. Cabrera displayed

knowledge of tactics to avoid surveillance throughout his dealings with Marcos,

instructing him on how to set up discrete meetings, hide drugs, change phones,

and use codes in communications about drug deals. Cabrera also admitted during

his testimony at trial that he had independently sought out a drug supplier who

sold him wholesale quantities of fentanyl pills, which enabled him to offer Marcos

“as many pills as he wanted.” And Cabrera had access to other drugs, as

evidenced by the fact that he provided Marcos with a sample of heroin and

discussed a potential transaction involving a strong form of heroin that he claimed

to have sold before.

The ease with which Cabrera located a supplier with unlimited amounts of

fentanyl clearly supported a finding of predisposition, as does the fact that Cabrera

readily took advantage of this apparently endless supply when he offered Marcos

free samples of fentanyl and heroin in early transactions and proposed sales of

hundreds or thousands of pills at more frequent intervals. Notably, at the time of

his arrest, Cabrera had over 1,100 pills in his possession. All of this evidence

showed that Cabrera was already experienced in selling drugs when he began

transacting with Marcos and that Cabrera obviously hoped to sell greater

10
quantities over time, thoroughly demonstrating his predisposition to sell drugs.

See United States v. Salerno, 66 F.3d 544, 547 (2d Cir. 1995) (explaining that

“predisposition may be shown by evidence of . . . an existing course of criminal

conduct similar to the crime for which the defendant is charged,” or “a willingness

to commit the crime for which he is charged as evidenced by the accused’s ready

response to the inducement” (internal quotation marks and alterations omitted)).

In assessing harmlessness, we consider “the overall strength” of the

government’s case, the importance of the improperly admitted testimony, “the

prosecutor’s conduct with respect to” such testimony, and whether that evidence

“was cumulative of other properly admitted evidence.” Garcia, 413 F.3d at 217. In

light of the substantial evidence supporting the government’s case against

Cabrera, Agent Son’s inadmissible opinion borders on the trivial. Indeed, while

the government cited other parts of Son’s testimony in its summation, it barely

discussed his description of Cabrera’s countersurveillance driving and the

inferences to be drawn therefrom. 2 The government relied much more heavily on

evidence of Cabrera’s prior drug-dealing experience that was entirely

2When transcribed, the government’s closing argument spanned approximately 43 pages and
over 1,000 lines of text. [A712–41, 765–779.] Less than 10 lines of that entire summation discussed
Agent Son’s testimony about Cabrera’s countersurveillance techniques. [A732, 768.]

11
independent from Son’s opinion testimony, including Cabrera’s own statements

and actions reflecting his familiarity with dealing drugs and evading law

enforcement and his easy access to a large supply of drugs.

Moreover, although Agent Son should not have been permitted to offer

expert opinion testimony based on his observations of Cabrera’s driving tactics, the

jury was certainly free to draw its own inferences when assessing Agent Son’s

properly admitted observational testimony. Specifically, Agent Son testified that

he personally observed Cabrera “excessively speeding, [making] erratic lane

changes, making U[-]turns and then making another U-turn.” App’x at 649. The

majority’s attempt to dismiss such maneuvers as nothing more than “really bad

driving,” Maj. Op. at 8, is entirely unpersuasive; indeed, we have long recognized

that such evasive driving techniques are highly probative of criminal knowledge

and intent, see United States v. Vasquez, 634 F.2d 41, 42–43 (2d Cir. 1980) (finding a

“reasonabl[y] objective basis” to believe that defendants were engaged in illegal

conduct where they engaged in “peculiar” conduct including circling the block,

signaling to turn in one direction and then abruptly turning in the other, and

glancing in the rearview mirror); see also United States v. Terry, 718 F. Supp. 1181,

1183–85 (S.D.N.Y. 1989), aff’d, 927 F.2d 593 (2d Cir. 1991) (noting that defendants’

12
“erratic and evasive driving away from a building under police surveillance for

drug activity,” making U-turns, driving “somewhat fast,” and changing lanes,

supported a “ reasonable suspicion” of criminal activity).

So while it was not proper for Son to offer opinion testimony concerning the

extent of Cabrera’s drug-dealing experience based on his observations of Cabrera’s

driving, the jury was certainly free to draw common-sense inferences about

Cabrera’s intent and predisposition on its own. See Hygh v. Jacobs, 961 F.2d 359,

364–65 (2d Cir. 1992) (finding harmless error where an expert witness’s

“impermissible testimony was expressed within a larger body of otherwise

unobjectionable testimony concerning police procedures involving violent

arrestees from which the jury could easily have drawn the same conclusions that

[the witness] did”); see also United States v. Duncan, 42 F.3d 97, 103 (2d Cir. 1994)

(similar). Indeed, it was the obviousness of the inference that made the expert

testimony improper in the first place. See United States v. Boissoneault, 926 F.2d 230,

233 (2d Cir. 1991) (noting that under the Federal Rules of Evidence, “a conclusion

that the jury could just as easily have drawn for itself based on its own knowledge

13
or experience is subject to exclusion”). We have rarely tossed a guilty verdict over

stray testimony of this sort, and I see no reason to do so here. 3

* * *

In sum, I am convinced that (1) the district court’s jury charge was wholly

consistent with this Circuit’s prior precedent and the new standard for entrapment

announced today; (2) any conceivable error in the charge would have been

harmless in any event, given the formidable evidence of Cabrera’s predisposition

to sell drugs; and (3) Agent Son’s opinion testimony, while improper, was likewise

harmless in light of that overwhelming evidence introduced at trial. I would

therefore affirm the district court’s judgment of conviction.

3 See, e.g., United States v. Winick, 792 F. App’x 91, 95 (2d Cir. 2019) (concluding that any error was
harmless where the government elicited improper witness testimony and referred to that
testimony at summation, but “offered substantial admissible evidence” at trial and “relied
heavily on admissible evidence” in summation); United States v. Londono-Tabarez, 121 F. App’x
882, 884–85 (2d Cir. 2005) (determining that the district court erred in admitting expert testimony
from a DEA agent who “de-code[d] certain statements” about drug transactions, but deciding
upon examination of the “whole record” – including the defendant’s non-credible testimony –
that the error was harmless (quotation marks omitted)).

14