United States v. Tanisha Banks

In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 19-3245
UNITED STATES OF AMERICA,
Plaintiff-Appellee,
v.
TANISHA A. BANKS,
Defendant-Appellant.
____________________
Appeal from the United States District Court for the
Northern District of Indiana, Hammond Division.
No. 2:18CR61-3 — Theresa L. Springmann, Judge.
____________________
ARGUED SEPTEMBER 22, 2020 — DECIDED DECEMBER 18, 2020
____________________
Before SYKES, Chief Judge, and FLAUM and ROVNER, Circuit
Judges.
SYKES, Chief Judge. Tanisha Banks was indicted on charges
of conspiracy and aiding and abetting a robbery of the
United States Post Office in Gary, Indiana, where she
worked as a mail clerk. See 18 U.S.C. §§ 371, 2114(a). After a
five-day trial and four hours of deliberation that stretched to
about 8:45 p.m., the jury returned a verdict of guilty on both
counts.
2 No. 19-3245
At the request of Banks’s counsel, the judge polled the
jurors. See FED. R. CRIM. P. 31(d). The first four affirmed the
verdict. The fifth did not. When asked whether the guilty
verdict was in fact his verdict, Juror 32 responded, “Forced
into.” The judge repeated the question. Juror 32 responded
that he needed more time. The judge continued the poll, and
the remaining jurors affirmed the verdict, singling out
Juror 32 as the lone dissenter. The judge then instructed the
jurors to continue deliberating and sent them back to the
jury room at 9:06 p.m. Twenty-nine minutes later, the jury
again returned a guilty verdict. This time the poll confirmed
a unanimous decision.
Banks raises several issues on appeal, but her main ar-
gument concerns the circumstances surrounding the jury
poll, which she contends exerted impermissible pressure on
the wavering juror. We agree. The totality of the circum-
stances—most notably, the dissenting juror’s troubling
responses to the poll questions, the judge’s decision to
complete the poll notwithstanding the juror’s dissent, the
lateness of the hour, and the extreme brevity of the jury’s
renewed deliberations—were unacceptably coercive. We
vacate the judgment and remand for a new trial.
I. Background
On August 3, 2017, a masked man with a gun robbed the
United States Post Office in Gary’s Tolleston neighborhood,
absconding with almost $6,000 in cash. According to the
evidence presented at trial, the robbery was the culmination
of a scheme hatched a few days earlier by Banks, who
worked as a mail clerk at the Tolleston branch, and James
Caffey, her boyfriend. They desperately needed money—
they had less than $700 between them and were delinquent
No. 19-3245 3
on rent and car payments—so they recruited their friend
Leeroy Beck and set the robbery plan in motion. Banks
provided information about the layout of the post office and
its closing procedures. Caffey supplied a mask and a
9mm handgun and was the getaway driver. Beck committed
the robbery using Caffey’s gun.
A grand jury indicted Banks, Caffey, and Beck on charges
of conspiracy to rob a post office, 18 U.S.C. § 371, and rob-
bing or aiding and abetting the robbery of mail, money, or
property of the United States, id. § 2114(a). The indictment
also charged Beck and Caffey with brandishing and aiding
and abetting the brandishing of a firearm during a crime of
violence. Id. § 924(c)(1)(A). Beck pleaded guilty to the fire-
arm count and agreed to testify for the government.
The case against Banks and Caffey proceeded to trial,
which spanned five days and featured testimony from Beck
and the government’s lead investigator, among other evi-
dence. The investigator’s testimony included statements
from several witness interviews, prompting hearsay objec-
tions from the defense. The objections were overruled and
Banks challenges those rulings on appeal. We can omit the
details because the juror-coercion issue resolves the appeal
in Banks’s favor, so there’s no need to address the hearsay
objections. Banks also challenges Beck’s testimony as incred-
ible as a matter of law, but again we have no need to address
that issue. We focus our attention on the circumstances
surrounding the jury’s deliberations, the jury poll, and the
aftermath of Juror 32’s dissent.
Before submitting the case to the jury, the judge gave this
instruction regarding the requirement of a unanimous
verdict:
4 No. 19-3245
The verdict must represent the considered
judgment of each juror. Your verdict, whether
it is guilty or not guilty, must be unanimous.
You should make every reasonable effort to
reach a verdict. In doing so, you should consult
with each other, express your own views and
listen to your fellow jurors’ opinions[.]
[D]iscuss your differences with an open
mind[.] [D]o not hesitate to re-examine your
own view and change your opinion if you
come to believe it is wrong. But you should not
surrender your honest beliefs about the weight
or effect of evidence just because [of] the opin-
ions of your fellow jurors or just so that there
could be a unanimous verdict.
The jurors retired to deliberate at 4:45 p.m. Four hours later
they returned to the courtroom and announced a verdict
finding both defendants guilty of conspiracy and aiding and
abetting the robbery, and acquitting Caffey on the firearm
charge.
Banks’s counsel asked the judge to poll the jury. The
judge asked each juror individually, “[I]s this your verdict as
to Ms. Banks?” The first four jurors responded, “Yes.” When
the next juror was polled, the following colloquy ensued:
THE COURT: Juror No. 32, is this your ver-
dict as to Ms. Banks?
JUROR NO. 32: Forced into.
THE COURT: Is this your verdict?
JUROR NO. 32: I suppose so.
No. 19-3245 5
THE COURT: Is it your verdict that she is
guilty on both Counts One
and Two?
JUROR NO. 32: I don’t know how to answer
that.
THE COURT: I’m asking you to answer that
at this time.
JUROR NO. 32: I feel like I need more time.
THE COURT: Let me go finish the poll, and
then I’ll come back to you.
The judge polled the remaining jurors, and each affirmed the
verdict.
The judge then called a sidebar with counsel and noted
that the verdict did not sound unanimous. Defense counsel
agreed. The government added, “I think we have to send
them back.” Caffey’s counsel asked the judge to poll the jury
regarding the verdict on the counts against Caffey. The
judge did so. All jurors affirmed the verdict as to Caffey, and
the court accepted it.
The judge then returned to the unanimity problem in the
verdict against Banks and sent the jurors back to the jury
room to continue deliberating with this instruction:
At this time, ladies and gentlemen of the jury,
because the verdict that you have returned …
with regard to Ms. Tanisha Banks does not ap-
pear to be unanimous given the polling of the
jury, the [c]ourt is going to direct that you re-
turn to your deliberations with regard to
Ms. Banks at this time—it’s about five minutes
after 9:00—and to continue your best, good
6 No. 19-3245
faith efforts in doing so to attempt to come to a
unanimous verdict with regard to Ms. Banks.
The judge asked if counsel had anything to add regarding
this instruction and counsel declined.
Deliberations resumed at 9:06 p.m. Just 29 minutes later,
the jury returned a verdict finding Banks guilty on both
counts. The judge again polled the jury, and this time all
jurors confirmed that the verdict was unanimous.
II. Discussion
“Any criminal defendant … being tried by a jury is enti-
tled to the uncoerced verdict of that body.” Lowenfield v.
Phelps, 484 U.S. 231, 241 (1988). Impermissible coercion
occurs “when jurors surrender their honest opinions for the
mere purpose of returning a verdict.” United States v.
Williams, 819 F.3d 1026, 1030 (7th Cir. 2016) (quotation marks
omitted). We assess the risk of juror coercion based on the
totality of the circumstances from the juror’s perspective. Id.The inquiry is objective and focuses on “the situation facing
the juror”; the subjective intent of the judge and the juror are
irrelevant. Id.; accord United States v. Blitch, 622 F.3d 658, 668
(7th Cir. 2010).
Banks did not object to the judge’s approach to Juror 32,
so we review the claim of impermissible coercion for plain
error. FED. R. CRIM. P. 52(b). Under the plain-error standard,
the defendant ordinarily must establish that an “obvious”
error occurred that affected his substantial rights and seri-
ously affected the fairness, integrity, or public reputation of
the judicial proceedings. United States v. Triggs, 963 F.3d 710,
714 (7th Cir. 2020). But juror coercion, if it occurs, is a grave
error. Jenkins v. United States, 380 U.S. 445, 446 (1965) (per
No. 19-3245 7
curiam) (“[T]he principle that jurors may not be coerced into
surrendering views conscientiously held is so clear as to
require no elaboration.” (quotation marks omitted)). Accord-
ingly, if the totality of the circumstances presents a clear
impermissible risk of juror coercion, we presume that the
error prejudiced the defendant and seriously affected the
fairness of the proceedings. Williams, 819 F.3d at 1031.
We begin with the circumstance most indicative of coer-
cion: the dialogue between the judge and Juror 32. Juror 32’s
initial response to the judge’s polling question was startling.
When the judge asked if the guilty verdict against Banks
was, in fact, his verdict, he replied with two striking words:
“Forced into.” This was far more than a simple “no.” Indeed,
the words “forced into” amount to Juror 32 telling the court
that he was coerced into the verdict.
To be clear, we do not hold that this response alone ne-
cessitated a mistrial. The rules of procedure give the trial
judge the discretion to either mistry the case or order further
deliberation when the jury poll reveals a lack of unanimity.
See FED. R. CRIM. P. 31(d). But Juror 32’s response was alarm-
ing enough that an immediate sidebar with counsel was
warranted.
Instead, the judge asked several follow-up questions that
only pressed Juror 32 further. Those questions and answers
did little to lessen the pressure at the time, and they do
nothing to assuage our concerns now. After Juror 32 said,
“Forced into,” the judge again asked him, “Is this your
verdict?” And he responded, “I suppose so.” The govern-
ment characterizes this response as suggesting that Juror 32
agreed with the guilty finding. That hardly seems the case.
At best, “I suppose so” suggests that Juror 32 is unsure of his
8 No. 19-3245
verdict. At worst, it implies he is giving up—“surrender[ing
his] honest opinions for the mere purpose of returning a
verdict.” Blitch, 622 F.3d at 668 (quotation marks omitted).
Either way, “I suppose so” certainly does not convey affir-
mance of the verdict. When asked a third time, Juror 32 said,
“I don’t know how to answer that.” We do not need to
speculate about the implications of this answer. The purpose
of a jury poll is to confirm unanimity, Williams, 819 F.3d at
1031, and this response decidedly did not do that.
More concerning is the judge’s next request: “I’m asking
you to answer that at this time.” The government argues that
this was a neutral inquiry intended only to clear up the
uncertainty in Juror 32’s responses. We see it differently. By
this time there was no uncertainty about Juror 32’s position.
He had already said he was “forced into” the verdict, and he
did not retreat from that position when asked two more
times. Whether the judge intended to merely clear up the
uncertainty is irrelevant. Our inquiry focuses on the juror’s
perspective. Id. at 1030. Viewed through that lens, continu-
ing to press Juror 32 for a different answer was unnecessari-
ly coercive.
The way in which the judge conducted the jury poll also
informs our analysis. Although criminal defendants are
entitled to poll the jury, the judge must conduct the poll in a
manner that minimizes its coercive effect. Id. at 1031. Banks
argues that the judge’s decision to complete the poll after
Juror 32 rejected the verdict unduly increased the pressure
on him by revealing him as the lone dissenter.
Continuing to poll the jury after one juror disagrees with
the verdict does not automatically amount to reversible
coercion. United States v. Carraway, 108 F.3d 745, 751 (7th Cir.
No. 19-3245 9
1997); see also Williams, 819 F.3d at 1032 n.2 (collecting out-of-
circuit cases). After all, the purpose of polling is to “ferret
out” dissenting jurors, Lyell v. Renico, 470 F.3d 1177, 1184
(6th Cir. 2006), and polling always risks revealing a lone
holdout because the last juror to be polled could reject the
verdict after the first 11 affirm it, Williams, 819 F.3d at 1031.
Revealing the jury’s numerical division through polling
therefore does not equate to inquiring into the numerical
division of a deadlocked jury—a per se reversible error. See
Brasfield v. United States, 272 U.S. 448, 450 (1926); Lyell,
470 F.3d at 1183–84 (emphasizing the differences between a
jury poll and an inquiry into the jury’s division).
But our decision in Williams also recognized that contin-
ued polling in this situation is relevant to the coercion
inquiry because it undoubtedly puts pressure on the holdout
juror. 819 F.3d at 1032–33. In most cases there is “little point
to continuing to poll … because one holdout suffices to send
the jury back to deliberate.” Lyell, 470 F.3d at 1183. That is
why we have endorsed the “wise” approach of terminating
the poll as soon as a lack of unanimity is revealed. Williams,
819 F.3d at 1032 & n.3. Doing so avoids revealing the num-
ber and identity of any dissenting jurors, reducing the poll’s
coercive effect.
The judge here did not follow that advice. Had she done
so, no one (save for the other jurors) would know whether
Juror 32 was alone in his disagreement or one of several
dissenters. We do not mean to suggest that the judge intend-
ed to pressure Juror 32; quite the contrary, we are confident
that she did not. But her intent is immaterial. What matters
is whether the manner in which the judge conducted the poll
10 No. 19-3245
unnecessarily risked coercion. Id. at 1033. We conclude that
it did.
The government argues that the risk of coercion attribut-
able to the jury poll is less here than it was in Williams.
There, a juror answered “no” when asked if the guilty
verdict was indeed her verdict. Id. at 1028. But the judge
apparently did not hear the juror’s response and simply
continued with the poll—revealing the dissenter as the lone
holdout—and then dismissed the jury, believing the verdict
was unanimous. Id. When counsel brought the issue to the
judge’s attention, he polled the jury a second time. We held
that these “additional coercive actions” rendered the contin-
ued polling an obvious error. Id. at 1032–33 (emphasis
omitted).
By contrast, here the judge polled the jury only once and
did not inadvertently dismiss the jury after the first poll. The
government emphasizes this distinction, but it’s not disposi-
tive. We assess juror coercion based on the totality of the
circumstances, and the jury poll is only one factor in that
inquiry.
The use of a supplemental instruction after a jury poll re-
veals division can help guard against an impermissibly
coercive atmosphere. We have not mandated a specific jury
instruction in this situation, id. at 1034, and we do not do so
today. But a robust cautionary instruction can lessen the
pressure on a dissenting juror. Williams emphasizes the
importance of reminding jurors “not to surrender their
honest beliefs” just to reach a unanimous verdict. Id.
The judge’s supplemental instruction lacked this im-
portant reminder. To be sure, the judge did ask the jury to
No. 19-3245 11
continue deliberating in “good faith.” But the heart of the
coercion inquiry is whether a juror “surrender[s his] honest
opinions for the mere purpose of returning a verdict.” Blitch,
622 F.3d at 668 (quotation marks omitted). That’s why the
so-called Silvern charge—the instruction given to a dead-
locked jury in this circuit—reminds jurors not to “surrender
[their] honest beliefs about the weight or effect of evidence
just because of the opinions of [their] fellow jurors or just so
that there can be a unanimous verdict.” Pattern Criminal Jury
Instructions of the Seventh Circuit No. 7.03 (2020). The likeli-
hood of coercion would have been diminished had the
judge’s supplemental instruction included this important
warning.
The government argues that because the judge gave the
Silvern charge with the general instructions before the jury
commenced deliberations, she did not need to do so again.
This argument misses the point. The supplemental instruc-
tion comes “at a very sensitive moment—immediately after
[a juror is] identified as the lone dissenter.” Williams,
819 F.3d at 1033. Cautioning jurors at that moment that they
should not surrender their honest beliefs for the sake of
reaching a unanimous verdict goes a long way toward
mitigating the risk of impermissible coercion.
The timing of the judge’s instruction to continue deliber-
ating also increased the likelihood of coercion. The jurors
reached a verdict at about 8:45 p.m., after listening to hours
of closing argument at the conclusion of a five-day trial and
deliberating for about four hours. After the first poll re-
vealed a dissenting juror, the judge returned the jurors to the
jury room at 9:06 p.m. with an instruction to continue delib-
erations. Having just been identified as the lone holdout,
12 No. 19-3245
Juror 32 no doubt well understood that he was the only
person preventing his fellow jurors, the attorneys, and the
judge from going home for the night. Under those circum-
stances any dissenting juror would have felt pressure to
surrender his beliefs. See United States v. Fiorilla, 850 F.2d 172,
176–77 (3d Cir. 1988) (noting the judge’s decision to dismiss
the jury for the day after a jury poll revealed a lone dissenter
lessened the risk of coercion by “remov[ing the dissenting
juror] from the immediate ‘attacks’ of his peers”); cf. Blitch,
622 F.3d at 670–71 (holding that the court’s instruction to
continue deliberating given around the time the court had
earlier promised the jury they would be able to leave for the
day was unduly coercive).
We conclude with a few words about the length of the
jury’s renewed deliberations. A verdict returned very quick-
ly after the jury is given a supplemental instruction to con-
tinue deliberations may signal that a juror was coerced.
Lowenfield, 484 U.S. at 240; Williams, 819 F.3d at 1034. While
longer deliberations indicate the jurors took time to “reex-
amine [their] own views” and “consult with one another,”
United States v. Silvern, 484 F.2d 879, 883 (7th Cir. 1973) (en
banc), shorter deliberations may support an inference of a
“coercive effect of the majority running roughshod over the
minority,” United States v. De Stefano, 476 F.2d 324, 337 (7th
Cir. 1973).
The circumstances support the latter inference here. The
jury returned its second verdict against Banks just
29 minutes after being instructed to continue deliberating.
That rapid turnaround suggests that Juror 32 may have felt
pressured into surrendering his views for the sole purpose
of returning a unanimous verdict.
No. 19-3245 13
The government points to cases involving shorter delib-
erations in which the court held that there was no coercion.
See Amos v. United States, 496 F.2d 1269, 1272–73 (8th Cir.
1974) (25 minutes); United States v. Brooks, 420 F.2d 1350,
1354 (D.C. Cir. 1969) (20 minutes). But neither Amos nor
Brooks expressly considered the deliberation time in their
analyses. In contrast, the Supreme Court in Lowenfield explic-
itly recognized that a 30-minute supplemental deliberation
“suggests the possibility of coercion.” 484 U.S. at 235, 240.
We likewise conclude that the 29-minute deliberation sug-
gests impermissible coercion.
III. Conclusion
The totality of the circumstances here created a clear and
obvious risk of juror coercion. Juror 32’s “forced into” re-
sponse to the poll question is powerful evidence of imper-
missible coercion. And the circumstances that followed—the
judge’s repeated pressing for another answer, the incom-
plete cautionary instruction, the late hour, and the brief
duration of the renewed deliberations—only amplify our
concern. Because the risk of juror coercion was clear and
obvious, we presume that the error prejudiced Banks and
seriously affected the fairness of the proceedings. Williams,
819 F.3d at 1031. The judgment must be vacated.
VACATED AND REMANDED