In Re: Kellogg Brown & Root, Inc.

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued May 7, 2014 Decided June 27, 2014

No. 14-5055

IN RE: KELLOGG BROWN & ROOT, INC., ET AL.,
PETITIONERS

On Petition for Writ of Mandamus
(No. 1:05-cv-1276)

John P. Elwood argued the cause for petitioners. With
him on the petition for writ of mandamus and the reply were
John M. Faust, Craig D. Margolis, Jeremy C. Marwell, and
Joshua S. Johnson.

Rachel L. Brand, Steven P. Lehotsky, Quentin Riegel,
Carl Nichols, Elisebeth C. Cook, Adam I. Klein, Amar
Sarwal, and Wendy E. Ackerman were on the brief for amicus
curiae Chamber of Commerce of the United States of
America, et al. in support of petitioners.

Stephen M. Kohn argued the cause for respondent. With
him on the response to the petition for writ of mandamus were
David K. Colapinto and Michael Kohn.

Before: GRIFFITH, KAVANAUGH, and SRINIVASAN, Circuit
Judges.
 2
Opinion for the Court filed by Circuit Judge
KAVANAUGH.

KAVANAUGH, Circuit Judge: More than three decades
ago, the Supreme Court held that the attorney-client privilege
protects confidential employee communications made during
a business’s internal investigation led by company lawyers.
See Upjohn Co. v. United States, 449 U.S. 383 (1981). In this
case, the District Court denied the protection of the privilege
to a company that had conducted just such an internal
investigation. The District Court’s decision has generated
substantial uncertainty about the scope of the attorney-client
privilege in the business setting. We conclude that the
District Court’s decision is irreconcilable with Upjohn. We
therefore grant KBR’s petition for a writ of mandamus and
vacate the District Court’s March 6 document production
order.

I

Harry Barko worked for KBR, a defense contractor. In
2005, he filed a False Claims Act complaint against KBR and
KBR-related corporate entities, whom we will collectively
refer to as KBR. In essence, Barko alleged that KBR and
certain subcontractors defrauded the U.S. Government by
inflating costs and accepting kickbacks while administering
military contracts in wartime Iraq. During discovery, Barko
sought documents related to KBR’s prior internal
investigation into the alleged fraud. KBR had conducted that
internal investigation pursuant to its Code of Business
Conduct, which is overseen by the company’s Law
Department.

KBR argued that the internal investigation had been
conducted for the purpose of obtaining legal advice and that
 3
the internal investigation documents therefore were protected
by the attorney-client privilege. Barko responded that the
internal investigation documents were unprivileged business
records that he was entitled to discover. See generally Fed. R.
Civ. P. 26(b)(1).

After reviewing the disputed documents in camera, the
District Court determined that the attorney-client privilege
protection did not apply because, among other reasons, KBR
had not shown that “the communication would not have been
made ‘but for’ the fact that legal advice was sought.” United
States ex rel. Barko v. Halliburton Co., No. 05-cv-1276, 2014
WL 1016784, at *2 (D.D.C. Mar. 6, 2014) (quoting United
States v. ISS Marine Services, Inc., 905 F. Supp. 2d 121, 128
(D.D.C. 2012)). KBR’s internal investigation, the court
concluded, was “undertaken pursuant to regulatory law and
corporate policy rather than for the purpose of obtaining legal
advice.” Id. at *3.

KBR vehemently opposed the ruling. The company
asked the District Court to certify the privilege question to
this Court for interlocutory appeal and to stay its order
pending a petition for mandamus in this Court. The District
Court denied those requests and ordered KBR to produce the
disputed documents to Barko within a matter of days. See
United States ex rel. Barko v. Halliburton Co., No. 05-cv-
1276, 2014 WL 929430 (D.D.C. Mar. 11, 2014). KBR
promptly filed a petition for a writ of mandamus in this Court.
A number of business organizations and trade associations
also objected to the District Court’s decision and filed an
amicus brief in support of KBR. We stayed the District
Court’s document production order and held oral argument on
the mandamus petition.
 4
The threshold question is whether the District Court’s
privilege ruling constituted legal error. If not, mandamus is of
course inappropriate. If the District Court’s ruling was
erroneous, the remaining question is whether that error is the
kind that justifies mandamus. See Cheney v. U.S. District
Court for the District of Columbia, 542 U.S. 367, 380-81
(2004). We address those questions in turn.

II

We first consider whether the District Court’s privilege
ruling was legally erroneous. We conclude that it was.

Federal Rule of Evidence 501 provides that claims of
privilege in federal courts are governed by the “common law
– as interpreted by United States courts in the light of reason
and experience.” Fed. R. Evid. 501. The attorney-client
privilege is the “oldest of the privileges for confidential
communications known to the common law.” Upjohn Co. v.
United States, 449 U.S. 383, 389 (1981). As relevant here,
the privilege applies to a confidential communication between
attorney and client if that communication was made for the
purpose of obtaining or providing legal advice to the client.
See 1 RESTATEMENT (THIRD) OF THE LAW GOVERNING
LAWYERS §§ 68-72 (2000); In re Grand Jury, 475 F.3d 1299,
1304 (D.C. Cir. 2007); In re Lindsey, 158 F.3d 1263, 1270
(D.C. Cir. 1998); In re Sealed Case, 737 F.2d 94, 98-99 (D.C.
Cir. 1984); see also Fisher v. United States, 425 U.S. 391,
403 (1976) (“Confidential disclosures by a client to an
attorney made in order to obtain legal assistance are
privileged.”).

In Upjohn, the Supreme Court held that the attorney-
client privilege applies to corporations. The Court explained
that the attorney-client privilege for business organizations
 5
was essential in light of “the vast and complicated array of
regulatory legislation confronting the modern corporation,”
which required corporations to “constantly go to lawyers to
find out how to obey the law, . . . particularly since
compliance with the law in this area is hardly an instinctive
matter.” 449 U.S. at 392 (internal quotation marks and
citation omitted). The Court stated, moreover, that the
attorney-client privilege “exists to protect not only the giving
of professional advice to those who can act on it but also the
giving of information to the lawyer to enable him to give
sound and informed advice.” Id. at 390. That is so, the Court
said, because the “first step in the resolution of any legal
problem is ascertaining the factual background and sifting
through the facts with an eye to the legally relevant.” Id. at
390-91. In Upjohn, the communications were made by
company employees to company attorneys during an attorney-
led internal investigation that was undertaken to ensure the
company’s “compliance with the law.” Id. at 392; see id. at
394. The Court ruled that the privilege applied to the internal
investigation and covered the communications between
company employees and company attorneys.

KBR’s assertion of the privilege in this case is materially
indistinguishable from Upjohn’s assertion of the privilege in
that case. As in Upjohn, KBR initiated an internal
investigation to gather facts and ensure compliance with the
law after being informed of potential misconduct. And as in
Upjohn, KBR’s investigation was conducted under the
auspices of KBR’s in-house legal department, acting in its
legal capacity. The same considerations that led the Court in
Upjohn to uphold the corporation’s privilege claims apply
here.
 6
The District Court in this case initially distinguished
Upjohn on a variety of grounds. But none of those purported
distinctions takes this case out from under Upjohn’s umbrella.

First, the District Court stated that in Upjohn the internal
investigation began after in-house counsel conferred with
outside counsel, whereas here the investigation was conducted
in-house without consultation with outside lawyers. But
Upjohn does not hold or imply that the involvement of outside
counsel is a necessary predicate for the privilege to apply. On
the contrary, the general rule, which this Court has adopted, is
that a lawyer’s status as in-house counsel “does not dilute the
privilege.” In re Sealed Case, 737 F.2d at 99. As the
Restatement’s commentary points out, “Inside legal counsel
to a corporation or similar organization . . . is fully
empowered to engage in privileged communications.” 1
RESTATEMENT § 72, cmt. c, at 551.

Second, the District Court noted that in Upjohn the
interviews were conducted by attorneys, whereas here many
of the interviews in KBR’s investigation were conducted by
non-attorneys. But the investigation here was conducted at
the direction of the attorneys in KBR’s Law Department.
And communications made by and to non-attorneys serving as
agents of attorneys in internal investigations are routinely
protected by the attorney-client privilege. See FTC v. TRW,
Inc., 628 F.2d 207, 212 (D.C. Cir. 1980); see also 1 PAUL R.
RICE, ATTORNEY-CLIENT PRIVILEGE IN THE UNITED STATES
§ 7:18, at 1230-31 (2013) (“If internal investigations are
conducted by agents of the client at the behest of the attorney,
they are protected by the attorney-client privilege to the same
extent as they would be had they been conducted by the
attorney who was consulted.”). So that fact, too, is not a basis
on which to distinguish Upjohn.
 7
Third, the District Court pointed out that in Upjohn the
interviewed employees were expressly informed that the
purpose of the interview was to assist the company in
obtaining legal advice, whereas here they were not. The
District Court further stated that the confidentiality
agreements signed by KBR employees did not mention that
the purpose of KBR’s investigation was to obtain legal
advice. Yet nothing in Upjohn requires a company to use
magic words to its employees in order to gain the benefit of
the privilege for an internal investigation. And in any event,
here as in Upjohn employees knew that the company’s legal
department was conducting an investigation of a sensitive
nature and that the information they disclosed would be
protected. Cf. Upjohn, 449 U.S. at 387 (Upjohn’s managers
were “instructed to treat the investigation as ‘highly
confidential’”). KBR employees were also told not to discuss
their interviews “without the specific advance authorization of
KBR General Counsel.” United States ex rel. Barko v.
Halliburton Co., No. 05-cv-1276, 2014 WL 1016784, at *3
n.33 (D.D.C. Mar. 6, 2014).

In short, none of those three distinctions of Upjohn holds
water as a basis for denying KBR’s privilege claim.

More broadly and more importantly, the District Court
also distinguished Upjohn on the ground that KBR’s internal
investigation was undertaken to comply with Department of
Defense regulations that require defense contractors such as
KBR to maintain compliance programs and conduct internal
investigations into allegations of potential wrongdoing. The
District Court therefore concluded that the purpose of KBR’s
internal investigation was to comply with those regulatory
requirements rather than to obtain or provide legal advice. In
our view, the District Court’s analysis rested on a false
dichotomy. So long as obtaining or providing legal advice
 8
was one of the significant purposes of the internal
investigation, the attorney-client privilege applies, even if
there were also other purposes for the investigation and even
if the investigation was mandated by regulation rather than
simply an exercise of company discretion.

The District Court began its analysis by reciting the
“primary purpose” test, which many courts (including this
one) have used to resolve privilege disputes when attorney-
client communications may have had both legal and business
purposes. See id. at *2; see also In re Sealed Case, 737 F.2d
at 98-99. But in a key move, the District Court then said that
the primary purpose of a communication is to obtain or
provide legal advice only if the communication would not
have been made “but for” the fact that legal advice was
sought. 2014 WL 1016784, at *2. In other words, if there
was any other purpose behind the communication, the
attorney-client privilege apparently does not apply. The
District Court went on to conclude that KBR’s internal
investigation was “undertaken pursuant to regulatory law and
corporate policy rather than for the purpose of obtaining legal
advice.” Id. at *3; see id. at *3 n.28 (citing federal
contracting regulations). Therefore, in the District Court’s
view, “the primary purpose of” the internal investigation “was
to comply with federal defense contractor regulations, not to
secure legal advice.” United States ex rel. Barko v.
Halliburton Co., No. 05-cv-1276, 2014 WL 929430, at *2
(D.D.C. Mar. 11, 2014); see id. (“Nothing suggests the reports
were prepared to obtain legal advice. Instead, the reports
were prepared to try to comply with KBR’s obligation to
report improper conduct to the Department of Defense.”).

The District Court erred because it employed the wrong
legal test. The but-for test articulated by the District Court is
not appropriate for attorney-client privilege analysis. Under
 9
the District Court’s approach, the attorney-client privilege
apparently would not apply unless the sole purpose of the
communication was to obtain or provide legal advice. That is
not the law. We are aware of no Supreme Court or court of
appeals decision that has adopted a test of this kind in this
context. The District Court’s novel approach to the attorney-
client privilege would eliminate the attorney-client privilege
for numerous communications that are made for both legal
and business purposes and that heretofore have been covered
by the attorney-client privilege. And the District Court’s
novel approach would eradicate the attorney-client privilege
for internal investigations conducted by businesses that are
required by law to maintain compliance programs, which is
now the case in a significant swath of American industry. In
turn, businesses would be less likely to disclose facts to their
attorneys and to seek legal advice, which would “limit the
valuable efforts of corporate counsel to ensure their client’s
compliance with the law.” Upjohn, 449 U.S. at 392. We
reject the District Court’s but-for test as inconsistent with the
principle of Upjohn and longstanding attorney-client privilege
law.

Given the evident confusion in some cases, we also think
it important to underscore that the primary purpose test,
sensibly and properly applied, cannot and does not draw a
rigid distinction between a legal purpose on the one hand and
a business purpose on the other. After all, trying to find the
one primary purpose for a communication motivated by two
sometimes overlapping purposes (one legal and one business,
for example) can be an inherently impossible task. It is often
not useful or even feasible to try to determine whether the
purpose was A or B when the purpose was A and B. It is thus
not correct for a court to presume that a communication can
have only one primary purpose. It is likewise not correct for a
court to try to find the one primary purpose in cases where a
 10
given communication plainly has multiple purposes. Rather,
it is clearer, more precise, and more predictable to articulate
the test as follows: Was obtaining or providing legal advice a
primary purpose of the communication, meaning one of the
significant purposes of the communication? As the
Reporter’s Note to the Restatement says, “In general,
American decisions agree that the privilege applies if one of
the significant purposes of a client in communicating with a
lawyer is that of obtaining legal assistance.” 1 RESTATEMENT
§ 72, Reporter’s Note, at 554. We agree with and adopt that
formulation – “one of the significant purposes” – as an
accurate and appropriate description of the primary purpose
test. Sensibly and properly applied, the test boils down to
whether obtaining or providing legal advice was one of the
significant purposes of the attorney-client communication.

In the context of an organization’s internal investigation,
if one of the significant purposes of the internal investigation
was to obtain or provide legal advice, the privilege will apply.
That is true regardless of whether an internal investigation
was conducted pursuant to a company compliance program
required by statute or regulation, or was otherwise conducted
pursuant to company policy. Cf. Andy Liu et al., How To
Protect Internal Investigation Materials from Disclosure, 56
GOVERNMENT CONTRACTOR ¶ 108 (Apr. 9, 2014) (“Helping a
corporation comply with a statute or regulation – although
required by law – does not transform quintessentially legal
advice into business advice.”).

In this case, there can be no serious dispute that one of
the significant purposes of the KBR internal investigation was
to obtain or provide legal advice. In denying KBR’s privilege
claim on the ground that the internal investigation was
conducted in order to comply with regulatory requirements
and corporate policy and not just to obtain or provide legal
 11
advice, the District Court applied the wrong legal test and
clearly erred.

III

Having concluded that the District Court’s privilege
ruling constituted error, we still must decide whether that
error justifies a writ of mandamus. See 28 U.S.C. § 1651.
Mandamus is a “drastic and extraordinary” remedy “reserved
for really extraordinary causes.” Cheney v. U.S. District
Court for the District of Columbia, 542 U.S. 367, 380 (2004)
(quoting Ex parte Fahey, 332 U.S. 258, 259-60 (1947)). In
keeping with that high standard, the Supreme Court in Cheney
stated that three conditions must be satisfied before a court
grants a writ of mandamus: (1) the mandamus petitioner must
have “no other adequate means to attain the relief he desires,”
(2) the mandamus petitioner must show that his right to the
issuance of the writ is “clear and indisputable,” and (3) the
court, “in the exercise of its discretion, must be satisfied that
the writ is appropriate under the circumstances.” Id. at 380-
81 (quoting and citing Kerr v. United States District Court for
the Northern District of California, 426 U.S. 394, 403
(1976)). We conclude that all three conditions are satisfied in
this case.

A

First, a mandamus petitioner must have “no other
adequate means to attain the relief he desires.” Cheney, 542
U.S. at 380. That initial requirement will often be met in
cases where a petitioner claims that a district court
erroneously ordered disclosure of attorney-client privileged
documents. That is because (i) an interlocutory appeal is not
available in attorney-client privilege cases (absent district
court certification) and (ii) appeal after final judgment will
 12
come too late because the privileged communications will
already have been disclosed pursuant to the district court’s
order.

The Supreme Court has ruled that an interlocutory appeal
under the collateral order doctrine is not available in attorney-
client privilege cases. See Mohawk Industries, Inc. v.
Carpenter, 558 U.S. 100, 106-13 (2009); see also 28 U.S.C.
§ 1291. To be sure, a party in KBR’s position may ask the
district court to certify the privilege question for interlocutory
appeal. See 28 U.S.C. § 1292(b). But that avenue is available
only at the discretion of the district court. And here, the
District Court denied KBR’s request for certification. See
United States ex rel. Barko v. Halliburton Co., No. 05-cv-
1276, 2014 WL 929430, at *1-3 (D.D.C. Mar. 11, 2014). It is
also true that a party in KBR’s position may defy the district
court’s ruling and appeal if the district court imposes
contempt sanctions for non-disclosure. But as this Court has
explained, forcing a party to go into contempt is not an
“adequate” means of relief in these circumstances. See In re
Sealed Case, 151 F.3d 1059, 1064-65 (D.C. Cir. 1998); see
also In re City of New York, 607 F.3d 923, 934 (2d Cir. 2010)
(same).

On the other hand, appeal after final judgment will often
come too late because the privileged materials will already
have been released. In other words, “the cat is out of the
bag.” In re Papandreou, 139 F.3d 247, 251 (D.C. Cir. 1998).
As this Court and others have explained, post-release review
of a ruling that documents are unprivileged is often
inadequate to vindicate a privilege the very purpose of which
is to prevent the release of those confidential documents. See
id.; see also In re Sims, 534 F.3d 117, 129 (2d Cir. 2008) (“a
remedy after final judgment cannot unsay the confidential
 13
information that has been revealed”) (quoting In re von
Bulow, 828 F.2d 94, 99 (2d Cir. 1987)).

For those reasons, the first condition for mandamus – no
other adequate means to obtain relief – will often be satisfied
in attorney-client privilege cases. Barko responds that the
Supreme Court in Mohawk, although addressing only the
availability of interlocutory appeal under the collateral order
doctrine, in effect also barred the use of mandamus in
attorney-client privilege cases. According to Barko, Mohawk
means that the first prong of the mandamus test cannot be met
in attorney-client privilege cases because of the availability of
post-judgment appeal. That is incorrect. It is true that
Mohawk held that attorney-client privilege rulings are not
appealable under the collateral order doctrine because
“postjudgment appeals generally suffice to protect the rights
of litigants and ensure the vitality of the attorney-client
privilege.” 558 U.S. at 109. But at the same time, the Court
repeatedly and expressly reaffirmed that mandamus – as
opposed to the collateral order doctrine – remains a “useful
safety valve” in some cases of clear error to correct “some of
the more consequential attorney-client privilege rulings.” Id.
at 110-12 (internal quotation marks and alteration omitted). It
would make little sense to read Mohawk to implicitly preclude
mandamus review in all cases given that Mohawk explicitly
preserved mandamus review in some cases. Other appellate
courts that have considered this question have agreed. See
Hernandez v. Tanninen, 604 F.3d 1095, 1101 (9th Cir. 2010);
In re Whirlpool Corp., 597 F.3d 858, 860 (7th Cir. 2010); see
also In re Perez, 749 F.3d 849 (9th Cir. 2014) (granting
mandamus after Mohawk on informants privilege ruling); City
of New York, 607 F.3d at 933 (same on law enforcement
privilege ruling).
 14
B

Second, a mandamus petitioner must show that his right
to the issuance of the writ is “clear and indisputable.”
Cheney, 542 U.S. at 381. Although the first mandamus
requirement is often met in attorney-client privilege cases,
this second requirement is rarely met. An erroneous district
court ruling on an attorney-client privilege issue by itself does
not justify mandamus. The error has to be clear. As a result,
appellate courts will often deny interlocutory mandamus
petitions advancing claims of error by the district court on
attorney-client privilege matters. In this case, for the reasons
explained at length in Part II, we conclude that the District
Court’s privilege ruling constitutes a clear legal error. The
second prong of the mandamus test is therefore satisfied in
this case.

C

Third, before granting mandamus, we must be “satisfied
that the writ is appropriate under the circumstances.” Cheney,
542 U.S. at 381. As its phrasing suggests, that is a relatively
broad and amorphous totality of the circumstances
consideration. The upshot of the third factor is this: Even in
cases of clear district court error on an attorney-client
privilege matter, the circumstances may not always justify
mandamus.

In this case, considering all of the circumstances, we are
convinced that mandamus is appropriate. The District Court’s
privilege ruling would have potentially far-reaching
consequences. In distinguishing Upjohn, the District Court
relied on a number of factors that threaten to vastly diminish
the attorney-client privilege in the business setting. Perhaps
most importantly, the District Court’s distinction of Upjohn
 15
on the ground that the internal investigation here was
conducted pursuant to a compliance program mandated by
federal regulations would potentially upend certain settled
understandings and practices. Because defense contractors
are subject to regulatory requirements of the sort cited by the
District Court, the logic of the ruling would seemingly
prevent any defense contractor from invoking the attorney-
client privilege to protect internal investigations undertaken as
part of a mandatory compliance program. See 48 C.F.R.
§ 52.203-13 (2010). And because a variety of other federal
laws require similar internal controls or compliance programs,
many other companies likewise would not be able to assert
the privilege to protect the records of their internal
investigations. See, e.g., 15 U.S.C. §§ 78m(b)(2), 7262; 41
U.S.C. § 8703. As KBR explained, the District Court’s
decision “would disable most public companies from
undertaking confidential internal investigations.” KBR Pet.
19. As amici added, the District Court’s novel approach has
the potential to “work a sea change in the well-settled rules
governing internal corporate investigations.” Br. of Chamber
of Commerce et al. as Amici Curaie 1; see KBR Reply Br. 1
n.1 (citing commentary to same effect); Andy Liu et al., How
To Protect Internal Investigation Materials from Disclosure,
56 GOVERNMENT CONTRACTOR ¶ 108 (Apr. 9, 2014)
(assessing broad impact of ruling on government contractors).

To be sure, there are limits to the impact of a single
district court ruling because it is not binding on any other
court or judge. But prudent counsel monitor court decisions
closely and adapt their practices in response. The amicus
brief in this case, which was joined by numerous business and
trade associations, convincingly demonstrates that many
organizations are well aware of and deeply concerned about
the uncertainty generated by the novelty and breadth of the
District Court’s reasoning. That uncertainty matters in the
 16
privilege context, for the Supreme Court has told us that an
“uncertain privilege, or one which purports to be certain but
results in widely varying applications by the courts, is little
better than no privilege at all.” Upjohn Co. v. United States,
449 U.S. 383, 393 (1981). More generally, this Court has
long recognized that mandamus can be appropriate to
“forestall future error in trial courts” and “eliminate
uncertainty” in important areas of law. Colonial Times, Inc.
v. Gasch, 509 F.2d 517, 524 (D.C. Cir. 1975). Other courts
have granted mandamus based on similar considerations. See
In re Sims, 534 F.3d 117, 129 (2d Cir. 2008) (granting
mandamus where “immediate resolution will avoid the
development of discovery practices or doctrine undermining
the privilege”) (quotation omitted); In re Seagate Technology,
LLC, 497 F.3d 1360, 1367 (Fed. Cir. 2007) (en banc) (same).
The novelty of the District Court’s privilege ruling, combined
with its potentially broad and destabilizing effects in an
important area of law, convinces us that granting the writ is
“appropriate under the circumstances.” Cheney, 542 U.S. at
381. In saying that, we do not mean to imply that all of the
circumstances present in this case are necessary to meet the
third prong of the mandamus test. But they are sufficient to
do so here. We therefore grant KBR’s petition for a writ of
mandamus.

IV

We have one final matter to address. At oral argument,
KBR requested that if we grant mandamus, we also reassign
this case to a different district court judge. See Tr. of Oral
Arg. at 17-19; 28 U.S.C. § 2106. KBR grounds its request on
the District Court’s erroneous decisions on the privilege
claim, as well as on a letter sent by the District Court to the
Clerk of this Court in which the District Court arranged to
transfer the record in the case and identified certain
 17
documents as particularly important for this Court’s review.
See KBR Reply Br. App. 142. KBR claims that the letter
violated Federal Rule of Appellate Procedure 21(b)(4), which
provides that in a mandamus proceeding the “trial-court judge
may request permission to address the petition but may not do
so unless invited or ordered to do so by the court of appeals.”

In its mandamus petition, KBR did not request
reassignment. Nor did KBR do so in its reply brief, even
though the company knew by that time of the District Court
letter that it complains about. Ordinarily, we do not consider
a request for relief that a party failed to clearly articulate in its
briefs. To be sure, appellate courts on rare occasions will
reassign a case sua sponte. See Ligon v. City of New York,
736 F.3d 118, 129 & n.31 (2d Cir. 2013) (collecting cases),
vacated in part, 743 F.3d 362 (2d Cir. 2014). But whether
requested to do so or considering the matter sua sponte, we
will reassign a case only in the exceedingly rare circumstance
that a district judge’s conduct is “so extreme as to display
clear inability to render fair judgment.” Liteky v. United
States, 510 U.S. 540, 551 (1994); see also United States v.
Microsoft Corp., 253 F.3d 34, 107 (D.C. Cir. 2001) (en banc).
Nothing in the District Court’s decisions or subsequent letter
reaches that very high standard. Based on the record before
us, we have no reason to doubt that the District Court will
render fair judgment in further proceedings. We will not
reassign the case.

***

In reaching our decision here, we stress, as the Supreme
Court did in Upjohn, that the attorney-client privilege “only
protects disclosure of communications; it does not protect
disclosure of the underlying facts by those who
communicated with the attorney.” Upjohn Co. v. United
 18
States, 449 U.S. 383, 395 (1981). Barko was able to pursue
the facts underlying KBR’s investigation. But he was not
entitled to KBR’s own investigation files. As the Upjohn
Court stated, quoting Justice Jackson, “Discovery was hardly
intended to enable a learned profession to perform its
functions . . . on wits borrowed from the adversary.” Id. at
396 (quoting Hickman v. Taylor, 329 U.S. 495, 515 (1947)
(Jackson, J., concurring)).

Although the attorney-client privilege covers only
communications and not facts, we acknowledge that the
privilege carries costs. The privilege means that potentially
critical evidence may be withheld from the factfinder.
Indeed, as the District Court here noted, that may be the end
result in this case. But our legal system tolerates those costs
because the privilege “is intended to encourage ‘full and frank
communication between attorneys and their clients and
thereby promote broader public interests in the observance of
law and the administration of justice.’” Swidler & Berlin v.
United States, 524 U.S. 399, 403 (1998) (quoting Upjohn, 449
U.S. at 389).

We grant the petition for a writ of mandamus and vacate
the District Court’s March 6 document production order. To
the extent that Barko has timely asserted other arguments for
why these documents are not covered by either the attorney-
client privilege or the work-product protection, the District
Court may consider such arguments.

So ordered.