United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued May 4, 2017 Decided July 21, 2017
No. 16-3102
UNITED STATES OF AMERICA,
APPELLEE
v.
KHAN MOHAMMED,
APPELLANT
Appeal from the United States District Court
for the District of Columbia
(No. 1:06-cr-00357-1)
Nathaniel H. Nesbitt argued the cause for appellant. With
him on the briefs was Peter S. Spivack.
Vijay Shanker, Attorney, U.S. Department of Justice,
argued the cause for appellee. With him on the brief were
Trevor N. McFadden, Deputy Assistant Attorney General, and
Matthew Robert Stiglitz, Attorney. Elizabeth Trosman,
Assistant U.S. Attorney, entered an appearance.
Before: ROGERS, MILLETT and PILLARD, Circuit Judges.
Opinion for the Court filed by Circuit Judge PILLARD.
2
PILLARD, Circuit Judge: Appellant Khan Mohammed was
extradited from Afghanistan to the United States and
convicted, following a 2008 jury trial, of international drug
trafficking and narcoterrorism, i.e., using drug proceeds to fund
terrorists or terrorism. This case comes to us on direct appeal
a second time, after this court affirmed Mohammed’s sentence
and conviction but remanded for an evidentiary hearing on his
claim of ineffective assistance of counsel. See United States v.
Mohammed, 693 F.3d 192 (D.C. Cir. 2012). We conclude that
the performance of Mohammed’s trial counsel was
constitutionally deficient. Counsel failed to investigate the
possibility of impeaching the government’s central witness as
biased against Mohammed, despite ample indication that he
should and could do so. Our review persuades us, however,
that counsel’s deficient performance did not prejudice
Mohammed as to the drug trafficking charge. As to the
narcoterrorism charge, we cannot on this record confidently
assess prejudice, and therefore remand to the district court for
further proceedings on that issue.
I.
We assume familiarity with our prior decision and the
district court’s opinion on remand. See United States v.
Mohammed, No. 06-357, 2016 WL 3982447 (D.D.C. July 22,
2016). We recite here only those facts relevant to our
disposition.
At a status conference more than two months before trial,
Mohammed’s trial counsel advised the district court that he
intended to seek witnesses in Afghanistan on Mohammed’s
behalf, but noted that there were “very difficult obstacles in
terms of finding witnesses, locating them, and then somehow
bringing them to the United States under some type of parole
visas.” 2/25/08 Status Hearing Tr. at 9. Months later, on the
eve of trial, counsel confirmed that he “did look into … how
do you even get [to Afghanistan], and who [from the office]
3
was I going to take,” but there were “no volunteers.” 5/2/08 Ex
Parte Hearing Tr. at 38. Ultimately, as counsel conceded, he
failed to “follow through” on his previously expressed intent to
contact witnesses in Afghanistan. Id. Although the
government provided a contact list from Mohammed’s phone
book in discovery, counsel never attempted to call any potential
witnesses in Afghanistan and, in fact, mistakenly represented
that he “wasn’t given telephone numbers.” Id. at 37.
The district court also asked Mohammed himself whether
he had told counsel to contact any potential witnesses in
Afghanistan. 5/2/08 Ex Parte Hearing Tr. at 21. Mohammed
responded that he had “asked [counsel] to bring my witnesses,”
and specifically identified four witnesses who would say that
he was not associated with the Taliban. Id. at 21-22. Counsel
confirmed Mohammed had given him those four names but
apparently thought the only way he could have spoken with the
potential witnesses would have been to travel to Afghanistan,
which he concluded posed “insurmountable” difficulties. Id. at
36.
One of the key issues on which witnesses in Afghanistan
might have shed light was the credibility of the government’s
confidential informant and central witness at trial, a man
known by the pseudonym Jaweed. Mohammed’s trial counsel
was aware of the possibility that Jaweed had a criminal history
that could be useful in undermining his credibility, as shown
by his request that the government perform a Lewis check—
pursuant to Lewis v. United States, 393 A.2d 109 (D.C.
1978)—into Jaweed’s criminal history in Afghanistan. That
check produced no governmental records on Jaweed and
counsel did not further pursue the issue.
Counsel briefly raised the possibility of impeaching
Jaweed as biased against Mohammed. The two men, hailing
4
from the same village, had known each other for some time;
indeed, Jaweed’s mother once asked Mohammed if he would
consider allowing Jaweed to marry Mohammed’s sister, an
inquiry that evidently did not yield the desired result for
Jaweed. See Mohammed, 2016 WL 3982447, at *14. At a
status conference a couple of weeks before trial, counsel noted
that Jaweed had “been previously accused with a lawsuit.”
4/22/08 Status Hearing Tr. at 6. But counsel represented
(erroneously, it is now claimed) that the lawsuit did not involve
Mohammed, and the district court advised that, barring some
connection to Mohammed, evidence relating to the lawsuit
would not be admissible to show bias. Id. at 5-8. The district
court stated, however, that counsel would be free to explore
possible bias arising from Mohammed’s victory in a local
election over Jaweed’s cousin. See id. at 7-8. Counsel took no
steps to investigate the election, Mohammed’s potential
connection to the lawsuit, or other possible sources of bias.
At trial, the government’s case rested on two pillars: the
recordings of Jaweed’s undercover conversations with
Mohammed, and Jaweed’s own testimony, which addressed
the meaning of those conversations and, more broadly, the two
men’s interactions. Jaweed’s testimony was, as trial counsel
later recalled, “the bread and butter of the case.” 12/3/15
Evidentiary Hearing Tr. at 25. The government repeatedly
asked Jaweed to clarify the meaning of exchanges between him
and Mohammed. Mohammed identifies 118 such
clarifications, but counsel objected only once, about halfway
through. See Appellant Br. 7. On that occasion, the
government asked Jaweed to clarify the meaning of
Mohammed’s statement, “I will, God willing, speak with them,
this will be done and that, too, and the work will be ready.”
5/12/08 A.M. Trial Tr. at 25. Counsel objected that the
question called for speculation and lacked foundation. Id. at
25-26. The district court overruled the objection, explaining
5
that, in its view, Jaweed was “entitled to say what he
understands Khan Mohammed to be saying.” Id. at 27. Jaweed
then testified that Mohammed “was trying to say that he would
see or speak with [Taliban commanders, and] the work of
missiles and the opium would be done.” Id.
Trial counsel’s cross-examination of Jaweed focused on
the possibility that Jaweed had monetary incentives—he was
paid $8,000 for his work as an informant, see 5/13/08 A.M.
Trial Tr. at 86-87—to testify against Mohammed. He did not
probe any preexisting bias on Jaweed’s part against
Mohammed. Trial counsel called no defense witnesses.
Mohammed appealed on several grounds, including
ineffective assistance of counsel. This court affirmed the
conviction and sentence and remanded for the district court to
hold an evidentiary hearing on the ineffectiveness claim. See
Mohammed, 693 F.3d at 195, 204. On the initial appeal, we
noted that “the thrust of Mohammed’s argument” was that
“counsel should have made some effort to learn if the
[potential] witnesses could undermine Jaweed’s credibility in
general, by testifying, for example, that he had a reputation for
dishonesty or that he harbored a grudge against Mohammed.”
Id. at 203. The court determined that, because “Jaweed’s
testimony arguably shaped how the jury understood
Mohammed’s words,” counsel’s failure to challenge Jaweed’s
credibility could have been prejudicial and a remand was
necessary for the district court to develop evidence and decide
the issue in the first instance. Id. at 204.
On remand, the district court received declarations from
Mohammed’s appellate counsel and his trial counsel and held
an evidentiary hearing at which trial counsel and Mohammed
testified. Mohammed’s appellate counsel stated that, in 2011,
while working on Mohammed’s first appeal, he managed to
6
interview 28 witnesses in Afghanistan over the phone. See
Declaration of Shardul Desai ¶¶ 9-11 (Feb. 28, 2013). Those
witnesses stated that Jaweed had a bad reputation for “criminal
behavior and gangsterism,” with one witness adding that he
was “a liar and an oath-breaker,” and that Mohammed was a
well-regarded local leader who was not in the Taliban. Id. at
¶¶ 13-14, 20-21. Of particular note, one witness, Malek
Rezwan, stated that he sat on a tribal arbitration council (jirga)
with Mohammed in which Jaweed was found to have stolen
jewelry, opium, and money. Id. at ¶ 16. Rezwan recalled
Jaweed being particularly “vengeful” toward Mohammed for
casting his vote in support of a ruling against Jaweed because
he thought Mohammed should have voted “along village
lines.” Id. Rezwan said that he and others overheard Jaweed
tell Mohammed: “I will not leave you alone even if it takes 20
years.” Id.
The district court rejected Mohammed’s ineffectiveness
claim principally on the ground that trial counsel reasonably
responded to matters that Mohammed timely brought to his
attention for possible investigation. In particular, the court
focused on the four potential witnesses whom, it found,
Mohammed specifically identified for counsel. See 2016 WL
3982447, at *17. The court concluded that counsel reasonably
did not pursue those witnesses given the limited information he
had, and that, in any event, Mohammed failed to establish that
those witnesses would have provided helpful information. See
id. at *23.
II.
To prevail on a claim of ineffective assistance of counsel,
“the defendant must show that (1) his counsel’s performance
‘fell below an objective standard of reasonableness,’ and (2)
‘there is a reasonable probability that, but for counsel’s
unprofessional errors, the result of the proceeding would have
7
been different.’” Payne v. Stansberry, 760 F.3d 10, 13 (D.C.
Cir. 2014) (quoting Strickland v. Washington, 466 U.S. 668,
687-88, 694 (1984)). Reviewing courts “must indulge a strong
presumption that counsel’s conduct falls within the wide range
of reasonable professional assistance,” and the defendant “must
overcome the presumption that, under the circumstances, the
challenged action ‘might be considered sound trial strategy.’”
Strickland, 466 U.S. at 689 (quoting Michel v. Louisiana, 350
U.S. 91, 101 (1955)). On appeal of a district court’s denial of
an ineffectiveness claim, this court reviews “for clear error any
findings of historical fact embedded in the District Court’s
conclusions on deficient performance and prejudice,” and
reviews questions of law de novo. United States v. Nwoye, 824
F.3d 1129, 1135 n.4 (D.C. Cir. 2016); see also Payne, 760 F.3d
at 13.
A.
Mohammed principally contends that his trial counsel
rendered constitutionally deficient performance by failing to
investigate Jaweed’s possible bias. Defense counsel “has a
duty to make reasonable investigations or to make a reasonable
decision that makes particular investigations unnecessary.”
Strickland, 466 U.S. at 691. “Where the case involves a failure
to investigate, the ‘particular decision not to investigate must
be directly assessed for reasonableness in all the circumstances,
applying a heavy measure of deference to counsel’s
judgments.’” United States v. McDade, 699 F.3d 499, 506
(D.C. Cir. 2012) (quoting Strickland, 466 U.S. at 691).
Counsel reasonably may decline to investigate when she or he
determines that any potential information an investigation
might uncover would have limited value or “could be easily
attacked on cross-examination.” Id. at 507.
The complete failure to investigate potential impeachment
witnesses cannot be construed as a strategic decision on the part
of defense counsel. See, e.g., Reynoso v. Giurbino, 462 F.3d
8
1099, 1112 (9th Cir. 2006) (“[I]f counsel’s failure to
investigate possible methods of impeachment is part of the
explanation for counsel’s impeachment strategy (or a lack
thereof), the failure to investigate may itself constitute
ineffective assistance of counsel.”); Tucker v. Ozmint, 350 F.3d
433, 444 (4th Cir. 2003) (“Trial counsel have an obligation to
investigate possible methods for impeaching a prosecution
witness, and failure to do so may constitute ineffective
assistance of counsel.”); cf. United States v. Debango, 780 F.2d
81, 85 (D.C. Cir. 1986) (“The complete failure to investigate
potentially corroborating witnesses . . . can hardly be
considered a tactical decision.”). “It is especially important
that counsel adequately investigate the case,” because “[o]nly
when reasonable investigation has been performed is counsel
in a position to make informed tactical decisions.” United
States v. Barbour, 813 F.2d 1232, 1234 (D.C. Cir. 1987). “In
assessing the reasonableness of an attorney’s investigation …
a court must consider not only the quantum of evidence already
known to counsel, but also whether the known evidence would
lead a reasonable attorney to investigate further.” Wiggins v.
Smith, 539 U.S. 510, 527 (2003).
This case involves a “complete failure to investigate.”
Debango, 780 F.2d at 85. Counsel failed to take the obligatory
step of calling potential witnesses, starting with the contact list
that the government provided in discovery months before trial
that identified people in Mohammed’s and Jaweed’s home
town in Afghanistan who knew Mohammed and might be able
to provide information relevant to the case or to Jaweed’s and
Mohammed’s past interactions. In fact, counsel evidently
forgot that he had been given the contact list. See 5/2/08 Ex
Parte Hearing Tr. at 37. As counsel himself acknowledged, he
failed to “follow through” on his previously expressed intent to
contact witnesses in Afghanistan. Id. at 38. Indeed, as this
court previously noted, counsel “expressed to the court his fear
9
that he had ‘been ineffective in assisting [Mohammed]’ by
failing to follow up on his lead that witnesses in Afghanistan
would rebut the government’s allegation that he was part of the
Taliban.” Mohammed, 693 F.3d at 196.
Counsel’s failure to place calls or otherwise reach out to
potential witnesses cannot be traced to any strategic decision.
Counsel did not decide against investigating based on any
reasoned anticipation that the evidence unearthed or testimony
of anyone he might reach would be of limited value or would
be vulnerable on cross-examination. Cf. McDade, 699 F.3d at
507. Mohammed gave him no reason to think that
investigation “would be fruitless or even harmful.” Strickland,
466 U.S. at 691. Counsel did not and could not know what he
would find out had he picked up the phone; he was thus in no
“position to make informed tactical decisions.” Barbour, 813
F.2d at 1234.
Notably, counsel was aware that Jaweed’s testimony
would be central to the upcoming trial. And counsel took
preliminary steps to understand Jaweed’s background,
including asking the government to search for any official
criminal records. But he did not investigate the possibility of
Jaweed’s bias, despite having some important clues on that
score. First, counsel knew several months before trial that,
while in custody pending extradition, Mohammed made
statements about his history of conflict with Jaweed and about
Jaweed’s past misconduct “in an attempt to ascribe a motive
for Jaweed to fabricate evidence or lie, or otherwise to impugn
Jaweed’s credibility.” Mohammed, 2016 WL 3982447, at *6.
And, as the district court found, Mohammed told counsel at
least a couple weeks before trial that “Jaweed may have a
vendetta against Mohammed because Mohammed won a local
election” against Jaweed’s preferred candidate. Id. at *20.
Counsel was also aware that Jaweed had once sought
10
unsuccessfully to marry Mohammed’s sister. See id. at *14. In
light of all that information, the possibility that Jaweed might
be biased against Mohammed was clearly worth investigating.
Cf. Wiggins, 539 U.S. at 527-28 (“[C]ounsel chose to abandon
their investigation at an unreasonable juncture, making a fully
informed decision … impossible.”).
We respectfully disagree with the district court that
counsel “was under no duty to investigate whether Jaweed was
biased against Mohammed when the only information he had
about this purported bias was the election.” Mohammed, 2016
WL 3982447, at *22. Any information about potential sources
of bias in a witness as crucial as Jaweed should have led to
further investigation. Yet counsel did not investigate,
apparently because he mistakenly believed he would need to
travel to Afghanistan to do so.
The district court appears to have concluded that counsel
could not be expected to take investigative steps that
Mohammed did not specifically suggest to him. But even when
defendants are “fatalistic or uncooperative,” that “does not
obviate the need for defense counsel to conduct some sort of …
investigation.” Porter v. McCollum, 558 U.S. 30, 40 (2009).
Indeed, when defendants are “actively obstructive,” they
remain entitled to effective counsel. Rompilla v. Beard, 545
U.S. 374, 381 (2005). “[C]ounsel has a duty to investigate,
even if his or her client does not divulge relevant information,”
Vega v. Ryan, 757 F.3d 960, 969 (9th Cir. 2014), and counsel
may not “ignore[] pertinent avenues for investigation of which
he should have been aware,” Porter, 558 U.S. at 40. We
therefore conclude that Mohammed’s trial counsel gave
constitutionally deficient assistance in failing to investigate
Jaweed’s possible bias.
Mohammed also claims that counsel rendered ineffective
assistance by failing to timely object—and to lodge a standing
11
objection—to Jaweed’s frequent “interpretations” of his
conversations with Mohammed. We agree that a proficient
attorney would have promptly objected and, once it became
clear the government would repeatedly solicit Jaweed’s
testimony about Mohammed’s state of mind, lodged a standing
objection. See, e.g., 5/9/08 P.M. Trial Tr. at 47-53 (government
counsel on direct examination repeatedly asking Jaweed to
explain Mohammed’s meaning, rather than Jaweed’s
understanding of what Mohammed was telling him). Had
Jaweed’s entire testimony been confined to what Jaweed
understood, not what Mohammed meant—as the district court
ultimately suggested it should be, see 5/12/08 A.M. Trial Tr. at
27—the jury likely would have better appreciated the
limitations of that testimony. Had Mohammed’s counsel also
independently sown doubt as to Jaweed’s credibility, the subtle
distinction between Jaweed’s actual testimony (about
Mohammed’s state of mind) and testimony to which Jaweed
should have been confined (about what he understood
Mohammed to mean) could have been significant. Thus, the
failure to object to Jaweed’s interpretations should be
considered part and parcel of counsel’s failure generally to
undermine Jaweed’s testimony.
B.
In assessing prejudice, the ultimate question is whether
Mohammed has shown a reasonable probability that adequate
investigation would have enabled trial counsel to sow sufficient
doubt about Jaweed’s credibility to sway even “one juror.”
Hope v. Cartledge, 857 F.3d 518, 524 (4th Cir. 2017) (quoting
Wiggins, 539 U.S. at 537); Vega, 757 F.3d at 974 (same). We
conclude as to the drug trafficking charge that, as the
government put it, “the recordings spoke for themselves and
directly implicated Mohammed through his own unequivocal
words and conduct.” Appellee Br. 44. Even if Jaweed had
been effectively impeached and shown to be biased, the jury
12
would have convicted based on Mohammed’s own words. For
instance, when informed by Jaweed that the drugs he was
selling were bound for America, Mohammed enthused that
“the Jihad would be performed since they send it to America.”
Gov’t Ex. 2I at 1. Mohammed himself, speaking on the
wiretap, disclosed the requisite knowledge or intent that the
drugs he was trafficking would be imported to the United
States. See also Gov’t Ex. 2H at 2-3, 8.
On the existing record, however, we cannot exclude the
possibility of prejudice as to the narcoterrorism conviction.
Jaweed’s testimony was the only evidence that linked
Mohammed to the Taliban. It thus provided critical support for
the narcoterrorism charge. See 5/9/08 P.M. Trial Tr. at 47-53;
5/12/08 A.M. Trial Tr. at 6-8, 28, 32-35, 63, 66. At trial, the
government pursued two distinct theories for convicting
Mohammed of narcoterrorism—that he had personally
engaged in terrorist activity, and that he was “acting as a
conduit of the Taliban,” a terrorist organization. 5/15/08 Trial
Tr. at 50. It is not clear which theory the jury believed. Given
the government’s own emphasis on the Taliban’s role in
Afghan drug trafficking and its use of drug proceeds to fund
terrorism, see id. at 50-51, it is reasonably probable that the jury
convicted Mohammed not of being a terrorist who retained
drug proceeds for himself, but of using drug sales to fund the
Taliban.
Because Jaweed’s testimony provided the only
unambiguous link between Mohammed and the Taliban,
however, it is reasonably probable that the jury would not have
convicted Mohammed if Jaweed’s testimony could have been
effectively undermined. The government objects that the
evidence also showed that Mohammed “provided commissions
to himself, and that he had engaged and planned to engage in
terrorist activity,” so that the jury could have convicted without
relying on Jaweed to link Mohammed to the Taliban. Appellee
13
Br. 47; cf. Griffin v. United States, 502 U.S. 46, 60 (1991)
(refusal to instruct the jury not to consider an “alternative basis
of liability that does not have adequate evidentiary support …
does not provide an independent basis for reversing an
otherwise valid conviction”). But due to counsel’s inadequacy,
the jury had no way to factor in the missing impeachment
evidence. We accordingly cannot safely assume that counsel’s
performance did not affect the verdict.
We cannot evaluate on the record as it stands, however, the
probability that an adequate investigation would have resulted
in effective impeachment of Jaweed’s testimony. On this
point, the district court “focus[ed] its analysis” on the four
potential witnesses whom, it found, Mohammed had
specifically suggested to counsel before trial. Mohammed,
2016 WL 3982447, at *17. The district court concluded that
Mohammed had not demonstrated prejudice because he never
“established that these witnesses would have been able to
testify that any animosity existed between Mohammed and
Jaweed.” Id. at *23; see also id. at *4, *24.
The district court’s exclusive focus on the four potential
witnesses Mohammed named was too limited. As this court
previously suggested, it is relevant not just “what Mohammed
told his attorney” but also “what his attorney [might] have
uncovered himself during trial preparation,” starting with the
contact list the government provided in discovery. 693 F.3d at
203. As appellate counsel’s inquiries illustrated, an adequate
investigation would involve following up with witnesses
besides the four Mohammed specifically identified in an effort
to identify anyone able to testify about what appears likely to
be a potent source of impeachment evidence: Mohammed’s
role in Jaweed’s jirga and Jaweed’s enraged response. See
Mohammed, 2016 WL 3982447, at *22 (district court
acknowledging, even as parties agreed not to postpone trial,
14
that investigation “might have uncovered some evidence of
Jaweed’s purported bias towards Mohammed”).
On the current record, and without additional district court
findings, we cannot confidently assess what a reasonable
investigation in 2008 could have found. Cf. United States ex
rel. Cross v. DeRobertis, 811 F.2d 1008, 1016-17 (7th Cir.
1987) (remanding for further proceedings where “the record
was not sufficiently developed to allow a comprehensive
analysis of the ‘prejudice’ component of the Strickland test”).
That incompleteness, in our view, cannot “fairly be
characterized as a failure of proof” on Mohammed’s part. Id.
at 1017. It does not appear that Mohammed failed to “show to
the extent possible precisely what information would have been
discovered through further investigation.” United States v.
Gwyn, 481 F.3d 849, 855 (D.C. Cir. 2007) (quoting United
States v. Askew, 88 F.3d 1065, 1073 (D.C. Cir. 1996)). Rather,
the proceedings in the district court did not reconstruct what an
adequate investigation in 2008 could have uncovered and how
counsel could have used that information at trial (as fodder for
cross-examination as well as direct testimony). Cf. United
States v. Cordova, 806 F.3d 1085, 1091 (D.C. Cir. 2015)
(finding no prejudice based in part on appellants’ failure to
identify how their cross-examination might have benefitted
from additional information).
Reconstruction of what a reasonable investigation could
have uncovered was made unnecessary by the district court’s
conclusions on the first Strickland prong, but because we hold
that counsel’s performance was deficient, this step must be
taken on remand. We leave to the district court in the first
instance to decide whether more evidence should be
introduced—such as testimony from Mohammed’s appellate
counsel about the investigation he performed, or even a proffer
of testimony in some form from the witnesses he contacted.
15
***
For these reasons, we affirm in part and vacate in part the
decision below and remand to the district court for further
proceedings consistent with this opinion.
So ordered.