United States v. Maher Obagi

FOR PUBLICATION
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
UNITED STATES OF AMERICA, No.18-50170
Plaintiff-Appellee,
D.C. No.
v. 8:13-cr-00001-
AG-2
MAHER OBAGI, AKA Maher Abaji,
AKA Mahir Abaji, AKA Mike Abaji,
Defendant-Appellant.
UNITED STATES OF AMERICA, No. 18-50171
Plaintiff-Appellee,
D.C. No.
v. 8:13-cr-00001-
AG-6
MOHAMED SALAH, AKA Mohamed
Ismail, AKA Mohamed Morales,
AKA Ahmed Salah, AKA Mohamed OPINION
Ahmed Salah,
Defendant-Appellant.
Appeal from the United States District Court
for the Central District of California
Andrew J. Guilford, District Judge, Presiding
Argued and Submitted June 4, 2020
Pasadena, California
2 UNITED STATES V. OBAGI
Filed July 17, 2020
Before: John B. Owens and Patrick J. Bumatay, Circuit
Judges, and Donald W. Molloy, * District Judge.
Opinion by Judge Owens;
Dissent by Judge Bumatay
SUMMARY **
Criminal Law
The panel reversed Maher Obagi’s and Mohamed
Salah’s convictions for federal mortgage fraud, and
remanded for further proceedings, in a case in which the
government disclosed after the close of evidence
information impeaching a government witness in violation
of Brady v. Maryland.
The panel wrote that had the information impeaching
Halime “Holly” Saad been disclosed prior to the close of
evidence, the presumption that juries are presumed to follow
their instructions and the normal rules concerning curative
instructions would govern, but in this case the genie was out
of the bottle. The panel noted that (1) the government’s
closing argument theme had been cast—the jury could trust
witness Jacqueline Burchell, who had pled guilty in this
*
The Honorable Donald W. Molloy, United States District Judge
for the District of Montana, sitting by designation.
**
This summary constitutes no part of the opinion of the court. It
has been prepared by court staff for the convenience of the reader.
UNITED STATES V. OBAGI 3
investigation and perjured herself in civil deposition, and
other cooperators because Saad was an “independent
witness” who reliably corroborated Burchell; (2) Obagi’s
counsel had completed closing argument without the benefit
of being able to attack Saad’s credibility; (3) Salah could
have used evidence that Saad was not a reliable independent
witness—and that the prosecution team had failed to
discover Saad’s severe credibility problems until closing
arguments—to attack the thoroughness and even the good
faith of the investigation; and (4) one could not expect Salah
at the last minute to reframe his defense to incorporate this
impeachment.
Rejecting the government’s argument that the failure to
disclose was not material, the panel wrote that Saad’s
impeachment substantially weakened the credibility of the
government’s cooperating witnesses and the strength of its
case.
Given the difficulty the jury faced in reaching a verdict,
the panel could not say with confidence that the undisclosed
impeachment did not affect the jury’s judgment.
The panel likewise could not conclude that the district
court’s instruction fully cured the prejudice that resulted
from the Brady violation. Noting that it may well be that no
instruction (or judge) could have corrected the government’s
significant error, the panel wrote that it does not criticize the
district court for how it handled this fluid and very tricky
complication.
Judge Bumatay dissented. Fearing that the panel is
unnecessarily curtailing the discretion afforded district
courts in responding to Brady violations, he would affirm the
convictions because the district court ably tailored a remedy
4 UNITED STATES V. OBAGI
that sufficiently abated the prejudice of the government’s
late disclosure of evidence.
COUNSEL
Craig Wilke (argued), Fullerton, California; Davina T. Chen,
Glendale, California; for Defendant-Appellant Maher
Obagi.
H. Dean Steward (argued), San Clemente, California, for
Defendant-Appellant Mohamed Salah.
Kerry L. Quinn (argued), Assistant United States Attorney;
L. Ashley Aull, Chief, Criminal Appeals Section; United
States Attorney’s Office, Los Angeles, California; for
Plaintiff-Appellee.
UNITED STATES V. OBAGI 5
OPINION
OWENS, Circuit Judge:
Maher Obagi and Mohamed Salah appeal from their
convictions for federal mortgage fraud. We have
jurisdiction under 28 U.S.C. § 1291, and we reverse. 1
I. BACKGROUND
A. The Excel Mortgage Fraud Scheme
Obagi and Salah worked for Excel Investments
(“Excel”), a mortgage broker that engaged in widespread
fraud. To make a long and complicated story short, Excel
approached condominium developers and offered to
purchase unsold inventory in exchange for kickbacks. Excel
would then use part of these kickbacks to make the initial
down payments on properties, pocketing the rest. To cover
up its activities, Excel created false “marketing agreements”
that purportedly reflected Excel’s work in advertising the
units, when in fact Excel did nothing to promote these sales.
To make this scheme work, Excel also recruited
individuals to pose as buyers on the mortgage applications.
These “straw” buyers required extensive false paperwork,
including bogus employment, income, and asset
documentation. At times, Excel staff impersonated the straw
buyers (and their supposed employers) when lenders
attempted to telephonically verify information listed in the
mortgage applications.
1
We resolve the companion appeal, United States v. Abaji, No. 18-
50241, in a concurrently filed memorandum disposition.
6 UNITED STATES V. OBAGI
B. Obagi’s and Salah’s Alleged Roles in the Excel
Scheme
Obagi began working for Excel in late 2007 and,
according to the prosecution, quickly took a leading role in
the operation. He supervised the creation of fraudulent loan
documents, communicated with banks, and recruited straw
buyers. He also supervised other Excel employees who
impersonated the buyers in phone calls, and he pocketed
large sums from the scheme.
Salah played a smaller role at Excel and, unlike Obagi,
was not part of the management team. But, as the
government alleged, he assisted in forging tax records and
similar documents to ensure that banks would approve the
mortgage applications. He also helped facilitate the bogus
straw buyer phone calls by Excel employees and concealed
payments to straw buyers by converting Excel checks into
cashier’s checks.
C. The Indictment, Trial, and Post-Trial
Proceedings
In January 2013, an indictment charged Obagi and Salah
with conspiracy to commit wire and bank fraud, and Obagi
with an additional six substantive counts of wire fraud. A
number of other individuals involved with Excel were
charged as well, several of whom pled guilty and agreed to
cooperate with the government. Others testified in exchange
for immunity.
At trial, the government introduced the fraudulent
records that Obagi and Salah allegedly created for Excel.
The government also called several cooperating witnesses
with significant credibility problems, including escrow
officer Jacqueline Burchell. Burchell testified that Obagi
UNITED STATES V. OBAGI 7
directed her to conceal the kickback payments, among other
things. In addition, Burchell told the jury that she overheard
a conversation between Obagi and Salah explaining why
they had developers send kickbacks to Excel instead of
reducing the prices of properties. But Burchell pled guilty
in this investigation, participated in a separate mortgage
fraud scheme, and perjured herself in a civil deposition. So
to bolster her credibility—as well as that of other
cooperating witnesses—the government also presented three
Excel witnesses who purportedly never had cut a deal to
avoid prosecution. One such witness was Halime “Holly”
Saad, another Excel escrow officer. Saad, like Burchell,
testified that Obagi instructed her to conceal the kickback
payments. The prosecution later described Saad’s testimony
as going “right to the heart of what was happening at Excel
during all the events set forth in the Indictment.”
After the close of evidence and during the opening
portion of its closing argument, the government highlighted
the testimony of Burchell, the cooperating escrow officer.
And to blunt the upcoming defense attack on Burchell’s
credibility, the government relied heavily on Saad. The
prosecution acknowledged that the jury should “treat
cooperating witness testimony more skeptically than you
would the average witness,” but emphasized that “on all
major points, these witnesses’ testimonies are corroborated
from independent witnesses,” as well as by documents and
the defendants’ own statements. “In particular, you heard
from Holly Saad. She said she dealt primarily with Maher
Obagi as the guy who controlled the escrow process for those
properties in Florida.” The prosecutor reminded the jury that
Saad had “no agreement[] wherein there’s any
representation about . . . [her] sentence[] being reduced or
non-prosecution, et cetera.” Indeed, Saad testified that she
8 UNITED STATES V. OBAGI
had not “entered into any agreement” with the prosecutor’s
office or received immunity in exchange for her cooperation.
Unfortunately, one of prosecution’s key tenets during
closing—that the jurors could trust the culpable cooperators
because they could trust Saad as an independent
corroborating witness—was false. During a break between
Obagi’s and Salah’s defense closings, a different prosecutor
from the U.S. Attorney’s Office who just happened to watch
the closing arguments recognized that Saad had in fact
received immunity in a separate mortgage fraud
investigation and alerted the trial prosecutors to the
enormous oversight. The prosecution then notified the court
and defense counsel about Saad’s June 2014 immunity
agreement, immediately disclosing both the agreement and
two investigative reports.
The introduction of this new information in the middle
of closings placed defense counsel and the trial judge in an
impossible position. The parties discussed various options,
including granting a mistrial, reopening the case to recall
Saad to the witness stand, or instructing the jury about
Saad’s undisclosed immunity agreement.
After extensive discussion, the trial court decided to
proceed with the remaining closing arguments, and to
instruct the jury as follows:
It has come to my attention that the United
States Attorney’s Office and the FBI failed to
disclose evidence that the witness, Halime
Saad, previously received immunity from the
United States Attorney in a separate case, and
thereafter, she may have knowingly made
false statements to the FBI. This information
UNITED STATES V. OBAGI 9
is relevant to the witness’s credibility in this
case.
The failure to disclose this information
implicates defendants’ right to due process of
law. For these reasons, you should disregard
the testimony of witness, Halime Saad, and
not consider it for any purpose, nor should
you consider any arguments made by the
government concerning Ms. Saad.
Immediately after crafting the instruction, the trial court told
counsel: “That’s it. I think you’ve all made your record, and
I think the Ninth Circuit should carefully look at that record
and see if I’ve drawn the line in the right place.”
In addition to the three documents disclosed during
closings, the government disclosed 3,750 pages of materials
the next morning and another 1,000 pages in the month after
trial. The defense did not have the opportunity to review any
of these materials before the case was submitted to the jury.
These documents revealed that Saad was a culpable
participant in the separate mortgage fraud scheme, had taken
a bribe to falsify escrow documents, lied to law enforcement
about her participation in the scheme, admitted to her actions
only after receiving limited use immunity from the U.S.
Attorney, and remained under investigation by a state
agency in connection with her fraudulent activities.
The jury deliberated for three days, then returned guilty
verdicts on the conspiracy charge for both Obagi and Salah,
convicted Obagi on three substantive counts, and hung on
the three remaining substantive counts.
Obagi and Salah moved for a new trial. After extensive
briefing and two days of post-trial hearings, the district court
10 UNITED STATES V. OBAGI
denied the motion. The court found no evidence of
intentional misconduct 2 and concluded that the
government’s failure to disclose did not deprive Obagi or
Salah of a fair trial. The court sentenced Obagi to
78 months’ imprisonment with $10,042,638 in restitution,
and sentenced Salah to 57 months’ imprisonment with
$7,487,163 in restitution.
The district court recognized the difficulties caused by
the prosecution’s failure to disclose Saad’s immunity deal
and related discovery. Urging Obagi and Salah to preserve
their objection for appeal, the court stated, “The Ninth
Circuit is going to have a look at this.” After sentencing, the
court allowed Obagi and Salah to remain on release status
pending appeal because of the disclosure issue.
II. DISCUSSION
A. STANDARD OF REVIEW
We review de novo whether the government violated its
discovery obligations. United States v. Stinson, 647 F.3d
1196, 1208 (9th Cir. 2011). Where, as in this case, the
district court has sanctioned the government for its discovery
violations, we review the choice of sanctions for abuse of
discretion. United States v. Garrison, 888 F.3d 1057, 1064
(9th Cir. 2018).
2
We agree with the district court that the failure to timely disclose
Saad’s separate immunity agreement was not due to any malfeasance by
the prosecutors in this case, but due to an apparent failure to
communicate between investigating agents who worked on the various
mortgage fraud investigations.
UNITED STATES V. OBAGI 11
B. BRADY AND ITS APPLICATION HERE
The government violates its discovery obligations under
Brady v. Maryland, 373 U.S. 83 (1963), if it suppresses
material evidence that is favorable to the accused. Strickler
v. Green, 527 U.S. 263, 281–82 (1999). Evidence favorable
to the accused “includes evidence that would help the
defendant impeach a witness.” Sanders v. Cullen, 873 F.3d
778, 802 (9th Cir. 2017) (citing Giglio v. United States,
405 U.S. 150, 154–55 (1972)); see also U.S. Dep’t of
Justice, Justice Manual § 9-5.001 (2020) (imposing
disclosure obligations on the government beyond that
provided for by either the Federal Rules or the Constitution).
Suppression occurs whenever the government fails to
disclose evidence, regardless of the government’s good or
bad faith. Wearry v. Cain, 136 S. Ct. 1002, 1006 (2016).
Evidence is material—and therefore requires reversal—
when there is “any reasonable likelihood that it could have
affected the judgment of the jury.” Id. (internal quotation
marks and citation omitted). Thus, a defendant can prevail
under Brady even if “the undisclosed information may not
have affected the jury’s verdict.” Id. at 1006 n.6; see also
Browning v. Baker, 875 F.3d 444, 470 (9th Cir. 2017)
(“Even if the jury—armed with all of this new evidence—
could have voted to convict [Browning], we have no
confidence that it would have done so.” (alteration in
original) (quoting Wearry, 136 S. Ct. at 1007)).
The district court made a noble effort in light of the
circumstances to craft a remedy that would protect the rights
of the defendants without requiring a mistrial. Ordinarily,
“juries are presumed to follow their instructions.”
Richardson v. Marsh, 481 U.S. 200, 211 (1987). Had the
information impeaching Saad been disclosed prior to the
12 UNITED STATES V. OBAGI
close of evidence, this presumption and the normal rules
concerning curative instructions likely would govern here.
But in this case, it was too late—the genie was out of the
bottle. Not only had the government’s closing argument
theme been cast—the jury could trust Burchell and the other
cooperators because Saad was an “independent witness”
who reliably corroborated Burchell—but Obagi’s own
counsel had completed closing argument without the benefit
of being able to attack Saad’s credibility. Asking defense
counsel to reframe his theory of the case—both in terms of
examining witnesses and arguing to the jury—after he had
spoken to the jury for the last time was simply too much.
Saad’s testimony did not incriminate Salah to the same
extent as Obagi, but the late-disclosed impeachment
evidence still could have played an important role in Salah’s
theory of the case. Salah could have used evidence that Saad
was not a reliable independent witness—and still worse, that
the prosecution team had failed to discover her severe
credibility problems until closing arguments—to attack “the
thoroughness and even the good faith of the investigation.”
Kyles v. Whitley, 514 U.S. 419, 445 (1995). The
investigating agents in this case failed to uncover Saad’s
credibility problems before trial, and they could have
similarly failed to vet the prosecution’s other witnesses. As
with Obagi, one could not expect Salah at the last minute to
reframe his defense to incorporate this impeachment.
The government argues that the failure to disclose was
not material because Saad’s testimony was “duplicative of
other testimony and evidence.” This was not how the
government viewed Saad’s testimony at trial. Saad’s
testimony mattered not only because she directly
incriminated Obagi, but because she corroborated testimony
from the government’s cooperating witnesses. Indeed, the
UNITED STATES V. OBAGI 13
government’s closing mentioned Saad by name six times
because her corroboration was so important. The
government told jurors that they could believe Burchell—a
confessed fraudster who testified to receive a favorable
plea—because Saad was trustworthy and “independent.”
The government used Saad to bolster the otherwise dubious
credibility of its cooperating witnesses. Saad’s
impeachment substantially weakened the credibility of those
witnesses and the strength of the government’s case.
Prejudice is especially likely here because the case was
so close. The jury deliberated for three days and still
delivered a split verdict on the charges against Obagi.
United States v. Leal-Del Carmen, 697 F.3d 964, 976 (9th
Cir. 2012) (finding prejudice where “jury deliberations
spanned two days and ended in a split verdict”); United
States v. Velarde-Gomez, 269 F.3d 1023, 1036 (9th Cir.
2001) (en banc) (“Longer jury deliberations weigh against a
finding of harmless error because lengthy deliberations
suggest a difficult case.” (internal quotation marks, citations,
and alterations omitted)). Given the difficulty the jury faced
in reaching a verdict, we cannot say with confidence that the
undisclosed impeachment did not affect the jury’s judgment.
Nor can we conclude that the district court’s instruction
fully cured the prejudice that resulted from the government’s
Brady violation. While the instruction informed the jury that
the government had erred and that it should disregard Saad’s
testimony and argument about her, it did not tell the jury that
the government’s powerful closing was premised on a false
narrative—Saad’s reliability. Nor did it explain how defense
counsel had presented the case one way, only to learn
afterwards that the truth was something else. Despite the
trial court’s best efforts, the failure to disclose Saad’s
immunity deal “undermine[s] confidence” in the jury’s
14 UNITED STATES V. OBAGI
verdict. Wearry, 136 S. Ct. at 1007 (quoting Smith v. Cain,
565 U.S. 73, 76 (2012)).
Although we conclude that the instructions given to the
jury did not sufficiently cure the problems that the late
disclosures created, we do not criticize the district court for
how it handled this fluid and very tricky complication—
indeed, under these unique circumstances, it may well be
that no instruction (or judge) could have corrected the
government’s significant error.
III. CONCLUSION
Because there is a reasonable likelihood that the
undisclosed evidence impeaching Saad could have affected
the judgment of the jury, we are compelled to reverse the
convictions and remand the case to the district court for
further proceedings. 3
REVERSED AND REMANDED.
3
Because we reverse on Brady grounds, we do not address the other
issues raised in the defendants’ appeal. The parties or the district court
may revisit these issues on remand.
UNITED STATES V. OBAGI 15
BUMATAY, Circuit Judge, dissenting:
No one questions the seriousness of the government’s
failure here. I agree with the majority that the government
wrongly withheld Brady material in failing to disclose
impeachment material against its witness, Halime “Holly”
Saad, until after closing arguments. The error, even if
inadvertent, undermined Maher Obagi’s and, to a lesser
extent, Mohamed Salah’s right to a fair trial.
But, I’m concerned that we have overcorrected for that
mistake here. As serious as a Brady violation is, that is not
the end of the inquiry for us. Even in cases where the
government’s withholding of Brady material prejudices
defendants, we still look to the district courts to determine if
a remedy is possible short of a mistrial. In this case, a long-
serving district court judge, who presided over the lengthy
trial, exercised that discretion and fashioned a remedy that
adequately cured any prejudice from the government’s trial
shortcomings. In the majority’s view, this cure falls short.
In the process, the majority suggests that any Brady violation
disclosed after closing arguments automatically warrants
reversal—a result found nowhere in our precedent. I
respectfully dissent.
I.
We ordinarily affirm guilty verdicts—even where the
government commits Brady violations—when any prejudice
is outweighed by (1) the district court’s remedy and
(2) extensive evidence of guilt. See United States v.
Garrison, 888 F.3d 1057, 1065 (9th Cir. 2018). Determining
whether the government’s Brady violation prejudiced
defendants, then, requires us to balance the severity of the
injury to defendants from the government’s conduct against
16 UNITED STATES V. OBAGI
the court’s chosen remedy and the record evidence of guilt
as a whole. Id.
In Garrison, for example, the government repeatedly
failed to timely disclose impeachment material regarding its
two key cooperating witnesses. 888 F.3d at 1061–62. In
response, the district court did not strike the witnesses’
testimony, but advised the jury that the government failed to
disclose the evidence and that it could draw adverse
inferences from this failure. Id. at 1064–66. On appeal, we
upheld the conviction, holding that in light of the court’s
instruction and “the extensive evidence against Garrison, we
cannot conclude that any prejudice stemmed from the late
disclosure.” Id. at 1055–56.
Moreover, we give deference to district courts in crafting
remedies for Brady violations. United States v. Struckman,
611 F.3d 560, 577 (9th Cir. 2010) (reviewing the district
court’s Brady remedy for abuse of discretion); see also
United States v. Ubaldo, 859 F.3d 690, 704 (9th Cir. 2017)
(“The district court acted within its discretion when it elected
to issue a curative instruction rather than granting a mistrial
for the purported violation of Rule 16.”). “[R]emedies
should be tailored to the injury suffered from the
constitutional violation and should not unnecessarily
infringe on competing interests.” United States v. Morrison,
449 U.S. 361, 364 (1981). When the district court’s remedy
sufficiently mitigates the harm, any prejudice is abated. See
Struckman, 611 F.3d at 578.
The district court’s chosen remedies here fell well within
its discretion. First, the court directed the jury to disregard
the testimony of Saad in its entirety, as well as any
arguments based on it. We’ve counseled before that striking
a witness’s testimony in its entirety is a “drastic remed[y],”
and should be used only after consideration of less severe
UNITED STATES V. OBAGI 17
sanctions. United States v. Polizzi, 500 F.2d 856, 893 (9th
Cir. 1974). Under the presumption that juries generally
follow court instructions, Richardson v. Marsh, 481 U.S.
200, 211 (1987), this directive effectively removed Saad’s
evidence from the proceedings.
The court also instructed the jury that the government
“failed to disclose evidence that [Saad] . . . may have
knowingly made false statements to the FBI,” and that “[t]he
failure to disclose this information implicate[d] defendants’
rights to due process of law.” The district court’s censure of
the prosecution was the last thing the jury heard before
deliberations, bolstering its leveling force. The district court
also afforded defense counsel additional remedies, offering
to allow counsel to recall Saad or the government’s agents
or to re-argue closing. Counsel rejected the court’s offer.1
Given the difficult hand the district court was dealt in the late
stage of trial, these instructions were not unreasonable in
light of the significant evidence marshaled against the
defendants and the discretion granted to district courts in this
area. Struckman, 611 F.3d at 577.
Finally, after the heat of trial, the district court conducted
an extensive evidentiary hearing into the government’s
actions and assured itself of the proprietary of its chosen
remedy. It determined that the government’s error was
inadvertent, that it attempted to correct the mistake at the
first possible moment, and that it had taken steps to avoid
similar mistakes in the future. The district court also found
that the government did not engage in any “flagrant
misbehavior” requiring dismissal. Given the benefit of all
1
This fact makes defendants’ claim on appeal that the court didn’t
do enough hard to swallow.
18 UNITED STATES V. OBAGI
the facts, the district court determined no mistrial was
warranted.
II.
The district court’s actions make even more sense in
light of the extensive evidence against the defendants, even
excluding Saad’s testimony. On the documentary front, the
government presented ample evidence that Obagi submitted
false mortgage documents in his own name, and received
and laundered money pursuant to the scheme. Several other
witnesses testified that Obagi directed participants to omit
kickbacks from statements submitted to banks and convert
personal checks to cashier’s checks, participated in
management-level meetings, and helped shred documents
right before the fraud shut down. Witnesses further testified
that both Obagi and Salah impersonated borrowers and
employers using cell phones in furtherance of the fraud. And
to top things off, investigating agents testified that Obagi
confessed to knowing his company’s actions were “illegal.”
Likewise, Salah admitted on tape that he forged fake W-2s
and other documents.
Given the court’s harsh curative instruction and the
extensive evidence against defendants, it’s hard to see how
Saad’s testimony makes or breaks the government’s case. 2
Saad was not a “star witness.” Cf. United States v. Kohring,
637 F.3d 895, 905 (9th Cir. 2011) (reversing on a Brady
violation related to the prosecution’s “star witness.”)
(simplified). As the district court rightly observed, she “only
testified for about an hour and a half, during the course of a
2
Saad’s unimportance to the government’s case is particularly clear
in relation to Mohamed Salah, who wasn’t even mentioned in her
testimony.
UNITED STATES V. OBAGI 19
three-week long trial.” Moreover, Saad’s testimony
overlapped with that of several other witnesses and covered
events pre-dating the defendants’ substantive criminal
offenses. Cf. Sivak v. Hardison, 658 F.3d 898, 914 (9th Cir.
2011) (finding no prejudice where a witness’s substantive
testimony was “duplicative to the other evidence presented
at trial”). In other contexts, we’ve affirmed guilty verdicts
even when the district court erroneously admitted a
defendant’s confession. See, e.g., Padilla v. Terhune,
309 F.3d 614, 621 (9th Cir. 2002). Saad’s limited testimony
is not nearly as damning as a defendant’s confession.
Because the district court ably tailored a remedy that
sufficiently abated the prejudice of the government’s late
disclosure of evidence, I would affirm the convictions.
While I understand why the majority’s concerns over the
government’s error lead it to reverse these two convictions,
I fear we are unnecessarily curtailing the discretion afforded
district courts in responding to Brady violations here. For
these reasons, I respectfully dissent.