In re Sealed Case

ORDER

Per Curiam.

In light of the recent indictment of former Secretary of Agriculture, Mike Espy, it is no longer necessary to maintain under seal any portion of the Court’s opinion issued on June 17, 1997. It is, therefore, ORDERED, on the Court’s own motion, that the opinion be amended as follows:

The previously sealed portions are hereby unsealed. The Clerk is directed to file and issue the unredacted opinion.

Footnote 24 of the opinion issued on June 17,1997, shall be deleted. [Editor’s Note: A different footnote 24 is included as part of this published opinion.]

Appeal from the United States District Court for the District of Columbia (No. 95ms00192).

Before: WALD, GINSBURG and ROGERS, Circuit Judges. WALD, Circuit Judge:

This case involves an effort by the Office of the Independent Counsel (“OIC”) to compel performance of a subpoena duces tecum issued by the grand jury investigating former Secretary of Agriculture Alphonso Michael (Mike) Espy (“Espy”) and served on the Counsel to the President (“White House Counsel”). The White House provided several folders of documents to the OIC in response to the subpoena but withheld 84 documents as privileged. After ordering that the withheld documents be produced for in camera review, the district court upheld the White House’s claims of privilege in full. We now vacate the district court’s opinion and remand for the court to conduct a more detailed review of the documents consistent with the principles set out in this opinion.

I. Background

A Factual Background

Allegations that Espy may have improperly accepted gifts from individuals and organizations with business before the U.S. Department of Agriculture (“USDA”) first surfaced publicly in March of 1994. These allegations led to the appointment of an Independent Counsel, on September 9,1994, to investigate whether Espy had unlawfully accepted gifts and related matters and to prosecute any related violations of federal law that the Independent Counsel reasonably believed had occurred. See In re Alphonso (Mike) Espy, No. 94-2 (D.C.Cir. Spec. Div.1994); see also In re Espy, 80 F.3d 501 (C.A.D.C.1996) (per curiam). This investigation into Espy’s actions is still ongoing.

*735The same allegations also led the President of the United States to direct the White House Counsel to investigate Espy’s conduct in order to advise the President on whether he should take executive action against Espy. On October 3, 1994, Espy announced his resignation, effective December 31, 1994. A little over a week later, on October 11, 1994, the White House publicly released a report on Espy produced by the White House Counsel. The report stated that the President had asked the White House Counsel to address two issues: “(1) whether the President should direct that any further action be taken with respect to Secretary Espy’s conduct; and (2) what actions should be taken to ensure that similar incidents are avoided by other Members of the Cabinet.” After detailing several areas in which questions had been raised regarding Espy’s conduct, the report concluded that no further executive action need be taken against Espy since he had announced his resignation, reimbursed the cost of questionable transactions, recused himself from matters involving meat and poultry inspection and undertaken screening measures for his travel. The report also recommended that efforts be undertaken to ensure that all cabinet members and other executive branch officers be given ethics training and be familiarized with applicable ethical standards for executive branch officers.

On October 14, 1994, the grand jury issued the subpoena duces tecum at issue in this case. The subpoena seeks all documents on Espy and other subjects of the OIC’s investigation that were “accumulated for, relating in any way to, or considered in any fashion, by those persons who were consulted and/or contributed directly or indirectly to all drafts and/or versions” of the White House Counsel’s report. Within this broad category of documents relating to the White House Counsel’s report, the subpoena specifically requests notes of any meetings in the White House concerning Espy and of any conversations between Espy or his counsel and White House employees. On October 20, 1994, the White House issued a press statement stating that it had received a subpoena for documents relating to the White House Counsel’s report and would comply with the subpoena. On November 17, 1994, the White House produced several folders of documents for the OIC, which the White House maintained represented all responsive documents except those withheld on the basis of privilege. On December 12,1994, at the OIC’s request, the White House produced a privilege log identifying the date, author, and recipient of each document withheld as well as a general statement of the nature of each document and the basis for the privilege on which the document was withheld. This privilege log indicated that 84 documents were withheld on grounds of the deliberative process privilege, with one document additionally withheld on grounds of attorney-client privilege.1 In a later draft of the privilege log, the White House lists the privilege basis of all 84 documents as being “executive/deliberative privilege.”2

The OIC negotiated with the White House for access to the withheld documents for several months, finally filing a motion to compel production on June 7, 1995. The White House resisted the motion, arguing that the withheld documents came within both the privilege for presidential communications, recognized in United States v. Nixon, 418 U.S. 683, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974) (Nixon), and the deliberative process privilege that protects the deliberations and decisionmaking process of executive officials generally. After a hearing on the motion to compel, the district court ordered the White House to produce the withheld documents for in camera review and the White House complied. Each document produced was accompanied by an ex parte cover *736sheet that explained the purpose of the document. The OIC also made an ex parte submission justifying the grand jury’s need for the documents. On September 30, 1996, the court denied the motion to compel. The memorandum opinion accompanying the denial quoted from Nixon to the effect that the “generalized assertion of privilege [for presidential communications] must yield to the demonstrated, specific need for evidence in a pending criminal trial,” 418 U.S. at 713, 94 S.Ct. at 3110, but then concluded that the White House had properly asserted the claimed privileges in this case. In reaching this conclusion, the court stated that it had carefully reviewed the documents, but did not discuss the documents in any further detail and provided no analysis of the grand jury’s asserted need for the documents.

The OIC appeals from the district court’s decision. The OIC argues that, at a minimum, the district court’s order should be vacated and the matter remanded because the district court failed to provide any account of its reasoning in denying the OIC’s motion to enforce the subpoena. On the merits, the OIC maintains that the district court erred in denying the motion to compel because the White House had waived its claims of privilege by releasing the final White House Counsel’s report, stating it would comply with the subpoena, and unduly delaying in invoking privilege. The OIC further argues that the presidential communications privilege does not apply to the withheld documents because none of the documents was sent to or received from the President; the only document that the President received regarding the Espy investigation was the White House Counsel’s final report, which was publicly released. Alternatively, the OIC claims that even if the withheld documents do enjoy the presidential privilege, the district court should have applied a less restrictive need standard than that articulated in Nixon, because this case involves a grand jury subpoena instead of a criminal trial subpoena, and the grand jury’s need for the documents is sufficient to overcome the claims of executive privilege raised in this case. Although the OIC does not separately discuss the applicability of the deliberative process privilege in any detail, it maintains in passing that the need to obtain evidence that may shed light on governmental misconduct outweighs the deliberative process privilege.

The White House challenges each of these arguments. It insists that it has not waived its claims of privilege and that the withheld documents come under the presidential communications privilege because they were generated in response to the President’s request for advice on whether to retain a cabinet officer, one of the President’s core functions under Article II of the Constitution. The White House also notes that the deliberative privilege would apply to the documents in their entirety because the factual material in the documents is inseparable from the documents’ deliberative portions. The White House contends that the same standard of need applies when the presidential privilege is raised in response to a grand jury subpoena as when a criminal trial subpoena is involved, and the OIC has failed to demonstrate a sufficient need to justify release under either the presidential privilege or the deliberative process privilege. Finally, the White House maintains that, since the district court reviewed the documents in camera, it provided sufficient explanation for its decision to deny the motion to compel even though it did not discuss the documents individually.

B. Legal Background: On Executive Privilege Generally and the Deference Due to the District Court

Since the beginnings of our nation, executive officials have claimed a variety of privileges to resist disclosure of information the confidentiality of which they felt was crucial to fulfillment of the unique role and responsibilities of the executive branch of our government. Courts ruled early that the executive had a right to withhold documents that might reveal military or state secrets. See United States v. Reynolds, 345 U.S. 1, 6-8, 73 S.Ct. 528, 531-32, 97 L.Ed. 727 (1953); Chicago & Southern Air Lines, Inc. v. Waterman Steamship Corp., 333 U.S. 103, 111, 68 S.Ct. 431, 436, 92 L.Ed. 568 (1948); Totten v. United States, 92 U.S. 105, 106-07, 23 L.Ed. 605 (1875). The courts have also granted the executive a right to withhold the *737identity of government informers in some circumstances, see Roviaro v. United States, 353 U.S. 53, 59-61, 77 S.Ct. 623, 627-28, 1 L.Ed.2d 639 (1957), and a qualified right to withhold information related to pending investigations. See Friedman v. Bache Halsey Stuart Shields, Inc., 738 F.2d 1336, 1341-43 (D.C.Cir.1984). Other privileges sanctioned by the Supreme Court include the grant of absolute immunity to the President from civil liability for official acts, see Nixon v. Fitzgerald, 457 U.S. 731, 749, 102 S.Ct. 2690, 2701, 73 L.Ed.2d 349 (1982) (Fitzgerald), and from judicial compulsion to perform a discretionary act. See Franklin v. Massachusetts, 505 U.S. 788, 802-03,112 S.Ct. 2767, 2776-77, 120 L.Ed.2d 636 (1992) (plurality opinion); Swan v. Clinton, 100 F.3d 973, 977-78 (D.C.Cir.1996).3

The most frequent form of executive privilege raised in the judicial arena is the deliberative process privilege; it allows the government to withhold documents and other materials that would reveal “advisory opinions, recommendations and deliberations comprising part of a process by which governmental decisions and policies are formulated.” Carl Zeiss Stiftung v. V.E.B. Carl Zeiss, Jena, 40 F.R.D. 318, 324 (D.D.C.1966), aff'd, 384 F.2d 979 (D.C.Cir.1967); accord NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 151-53, 95 S.Ct. 1504, 1516-18, 44 L.Ed.2d 29 (1975); EPA v. Mink, 410 U.S. 73, 86-93, 93 S.Ct. 827, 835-39, 35 L.Ed.2d 119 (1973). Although this privilege is most commonly encountered in Freedom of Information Act (“FOIA”) litigation, it originated as a common law privilege. See Wolfe v. Department of Health and Human Services, 839 F.2d 768, 773 (D.C.Cir.1988) (en banc); Jordan v. Department of Justice, 591 F.2d 753, 772 (D.C.Cir.1978) (en banc).4 Two requirements are essential to the deliberative process privilege; the material must be predecisional and it must be deliberative. See Army Times Publ’g Co. v. Department of the Air Force, 998 F.2d 1067, 1070 (D.C.Cir.1993); Wolfe, 839 F.2d at 774. Both requirements stem from the privilege’s “ultimate purposed which] ... is to prevent injury to the quality of agency decisions” by allowing government officials freedom to debate alternative approaches in private. Sears, 421 U.S. at 151, 95 S.Ct. at 1516-17. The deliberative process privilege does not shield documents that simply state or explain a decision the government has already made or protect material that is purely factual, unless the material is so inextricably intertwined with the deliberative sections of documents that its disclosure would inevitably reveal the government’s deliberations. See id. at 150-54, 95 S.Ct. at 1516-18; Mink, 410 U.S. at 87-91, 93 S.Ct. at 836-38; Wolfe, 839 F.2d at 774; see generally Russell L. Weaver & James T.R. Jones, The Deliberative Process Privilege, 54 Mo.L.Rev. 279, 290-98 (1989).

The deliberative process privilege is a qualified privilege and can be overcome by a sufficient showing of need.5 This need determination is to be made flexibly on a case-by-case, ad hoc basis. “[E]ach time [the deliberative process privilege] is asserted the district court must undertake a fresh balancing of the competing interests,” taking into account factors such as “the relevance of the *738evidence,” “the availability of other evidence,” “the seriousness of the litigation,” “the role of the government,” and the “possibility of future timidity by government employees.” In re Subpoena Served Upon the Comptroller of the Currency, 967 F.2d 630, 634 (D.C.Cir.1992) (internal quotations omitted) (quoting In re Franklin Nat’l Bank Securities Litig., 478 F.Supp. 577, 583 (E.D.N.Y.1979)); see also Tuite v. Henry, 98 F.3d 1411, 1417 (D.C.Cir.1996) (describing need in the context of the law enforcement investigatory privilege, which involves balancing similar factors, as “an elastic concept”); Developments in the Law — Privileged Communications, 98 Harv. L.Rev. 1450, 1621 (1985) (“courts simply engage in an ad hoc balancing of the evidentiary need against the harm that may result from disclosure”); Larkin, supra, § 5.03 at 5-89 to 5-92 (“need for [privileged materials] may vary considerably, depending on the circumstances’/^. For example, where there is reason to Ibebeve the documents sought may shed light on government misconduct, “the privilege is routinely denied,” on the grounds that shielding internal government deliberations in this context does not serve “the public’s interest in honest, effective government.'’ Texaco Puerto Rico, Inc. v. Departments of Consumer Affairs, 60 F.3d 867, 885 (1st Cir.1995); see also In re Comptroller of the Currency, 967 F.2d at 634 (“the privilege may be overridden where necessary ... to ‘shed light on alleged government malfeasance’ ”) (quoting Franklin Nat’l Bank, 478 F.Supp. at 582); Wetlaufer, supra, at 852 n. 25, 855 (listing cases).

Although executive privilege in general is no stranger to the courtroom, one form of the executive privilege is invoked only rarely and that is the privilege to preserve the confidentiality of presidential communications. Hints of a presidential communications privilege made an early appearance in Marbury v. Madison where Chief Justice Marshall suggested that for a court to intrude “into the secrets of the cabinet” would give the appearance of “intermeddling] with the prerogatives of the executive.” 5 U.S. (1 Cranch) 137, 170, 2 L.Ed. 60 (1803). Four years later, in 1807, Marshall again addressed the presidential privilege during the trial of Aaron Burr on charges of treason. President Jefferson asserted the privilege in an effort to avoid producing a letter that he had received from General Wilkinson, one of Burr’s main accusers. Marshall, sitting on circuit, issued a subpoena for the letter, ruling that “[i]f [the letter] does contain any matter which it would be imprudent to disclose, which it is not the wish of the executive to disclose, such matter, if it be not immediately and essentially applicable to the point, will, of course, be suppressed.” United States v. Burr, 25 F. Cas. 30, 37 (C.C.Va. 1807) (No. 14,692d). Although Burr was acquitted in his treason trial before there were further proceedings on his subpoena, he was immediately put on trial again on misdemeanor charges and as a result sought production of another letter Wilkinson had sent to Jefferson. See Paul A. Freund, The Supreme Court, 1973 Term — Foreword: On Presidential Privilege, 88 Harv. L. Rev. 13, 22-31 (1974). In neither instance, however, was Marshall forced to definitively decide whether such a presidential privilege existed and if so, in what form. In Marbury, Marshall found that the question of whether a commission as justice of the peace had been issued was a matter of legal and public record, not a confidential cabinet matter, setting the stage for the Court’s pronouncement there that “[i]t is, emphatically, the province and duty of the judicial department, to say what the law is.” 5 U.S. (1 Cranch) at 177.6 In the Burr misdemeanor trial, Jefferson responded to the subpoena by sending Wilkinson’s letter to George Hay, the U.S. Attorney prosecuting Burr, with instructions that the U.S. Attorney should determine what portions should *739be withheld. This delegation induced Marshall to order that the letter be provided to Burr in its entirety, because “[t]he propriety of withholding [the letter] must be decided by [the President] himself.” United States v. Burr, 25 F. Cas. 187, 192 (C.C.Va.1807) (No. 14,694).7

The presidential communications privilege did not resurface in court for over a hundred and fifty years.8 Presidential claims of a right to preserve the confidentiality of information and documents figured more prominently in executive-congressional relations, but these claims too were most often essentially assertions of the deliberative process privilege.9 Moreover, given the restrictions on congressional standing and the courts’ reluctance to interfere in political battles, few executive-congressional disputes over access to information have ended up in the courts.10 As a result, it was not until the 1970s and Watergate-related lawsuits seeking access to President Nixon’s tapes as well as other materials that the existence of the presidential privilege was definitively established as a necessary derivation from the President’s *740constitutional status in a separation of powers regime.

In this case, the White House is asserting both the deliberative process privilege and the presidential communications privilege.11 Our review of the withheld documents indicates that several documents are either wholly factual or contain segregatable factual sections that would not come under the deliberative process privilege. Consequently, we must decide whether the White House properly asserted the presidential communications privilege over the documents.

As a preliminary matter we must first explain the standard under which we should review the district court’s ruling that the presidential privilege applied to the withheld documents. Ordinarily, this court will review a district court’s ruling on a subpoena for the production of documentary evidence only for arbitrariness or abuse of discretion. See In re Comptroller of the Currency, 967 F.2d at 633; In re Sealed Case, 877 F.2d 976, 981-82 (D.C.Cir.1989). No deference is given, however, if the ruling “rests upon a misapprehension of the relevant legal standard or is unsupported by the record.” In re Subpoena on Comptroller of Currency, 967 F.2d at 633. In order to defer we also need to have some articulation of the district court’s reasons for its ruling. See In re Sealed Case (Government Records), 950 F.2d 736, 738 (D.C.Cir.1991) (appeals court cannot apply deferential standard when district court did not provide reasons for denying subpoena or did not review documents in camera).

Here, the district court provided no explanation of its denial of the motion to compel. The denial took the form of a blanket ruling, with no individualized discussion of the documents. Since the district court reviewed the withheld documents in camera before denying the OIC’s motion to compel, the absence of detailed findings would not, on its own, preclude us from according our usual deference to the district court’s opinion. However, the court also failed to provide any explanation of its legal reasoning. It did not address the OIC’s claim that the White House had waived its privileges or analyze whether the presidential communications privilege applies to documents not seen by the President. Moreover, while the court quoted Nixon’s statement to the effect that the presidential privilege must yield to a specific demonstration of need, it never discussed why Nixon applies to grand jury subpoenas as well as trial subpoenas nor indicated why the OIC’s demonstration of need was deficient. Because the district court not only failed to make factual findings but also failed to provide any explanation of its legal reasoning, we believe that no deference to the district court’s denial of the OIC’s motion to compel is appropriate.

II. Waiver

We turn first to the OIC’s contention that the White House has waived its privilege claims; if we find that waiver has occurred, we need not proceed further. In support of its waiver argument, the OIC notes that the White House publicly released the White House Counsel’s report, issued a press statement indicating it would comply with the OIC’s subpoena, and did not formally invoke privilege until after the OIC filed a motion to compel. Only after the briefs in this appeal were submitted did the White House inform us that it had provided Espy’s counsel with a document nearly identical to one of the withheld documents, document 63, the only difference being that document 63 contained certain handwritten notations that the released version lacked. The OIC argues that the release of document 63 is further evidence of a privilege waiver.

We do not credit the OIC’s arguments for waiver. The White House press statement did not explicitly declare that the White House would forego any and all claims of privilege that might apply to the documents. Instead, it described the documents sought in the subpoena and noted “[t]he subpoena requires that documents be produced on November 10, 1994. The White House will comply.” The OIC agreed to extend the return date of the subpoena to November 17, and on that date the White House did in fact produce several folders of documents. *741“Since executive privilege exists to aid the governmental decisionmaking process, a waiver should not be lightly inferred.” SCM Corp. v. United, States, 82 Cust.Ct. 351, 473 F.Supp. 791, 796 (1979); see also Nixon v. Sirica (Sirica), 487 F.2d 700, 717 (D.C.Cir.1973) (explicit statement by President Nixon that “[executive privilege will not be invoked” considered one factor in assessing need to preserve confidentiality of subpoenaed materials, but not held to constitute a waiver). The press statement was not an official response to the subpoena, and it is clear from the record that the OIC was well aware the White House would be asserting privileges in regard to certain documents. Shortly after the statement was issued the White House Counsel informed the OIC that it believed some of the material was privileged, provoking lengthy negotiations between the two over the status of the withheld documents. There is nought to indicate that the press statement misled the OIC.

Nor did the White House have an obligation to formally invoke its privileges in advance of the motion to compel. In its response to the subpoena, the White House informed the OIC that it believed the withheld documents were privileged, thus satisfying Rule 45(c)(2)(B) and Rule 45(d)(2) of the Federal Rules of Civil Procedure, which together require that “a party objecting to a subpoena on the basis of privilege must both (1) object to the subpoena and (2) state the claim of privilege within [the stipulated period] of service.” Twite, 98 F.3d at 1416; see also In re Sealed Case, 856 F.2d 268, 272 n. 3 (D.C.Cir.1988) (where government’s claim of privilege is well taken, remedy for any delay is not waiver but fees and sanctions). The motion to compel was the first event which could have forced disclosure of the documents. Cf. 3 Weinstein’s Federal Evidence § 503.09[4] at 503-44 (failure to assert attorney-client privilege at a hearing at which privileged information is sought may result in waiver of the privilege). Since the OIC was clearly aware in advance of the motion to compel that the White House likely would be asserting privilege, it was not prejudiced by any alleged delay in the White House’s formally invoking its privileges.

The White House’s release of the White House Counsel’s final report also does not constitute waiver of any privileges attaching to the documents generated in the course of producing the report. It is true that voluntary disclosure of privileged material subject to the attorney-client privilege to unnecessary third parties in the attorney-client privilege context “waives the privilege, not only as to the specific communication disclosed but often as to all other communications relating to the same subject matter.” In re Sealed Case, 676 F.2d 793, 809 (D.C.Cir.1982); accord In re Sealed Case, 29 F.3d 715, 719-20 (D.C.Cir.1994); see generally 3 Weinstein’s Federal Evidence § 511. But this all-or-nothing approach has not been adopted with regard to executive privileges generally, or to the deliberative process privilege in particular. Instead, courts have said that release of a document only waives these privileges for the document or information specifically released, and not for related materials. See Mobil Oil Corp. v. United States EPA, 879 F.2d 698, 700-02, 703 (9th Cir.1989); Mehl v. United States EPA, 797 F.Supp. 43, 47-48 (D.D.C.1992); Larkin, su pra, § 5.05 at 5-114.7 to 5-114.14; see also Russell v. Department of the Air Force, 682 F.2d 1045, 1048-49 (D.C.Cir.1982) (although not addressing waiver directly, holding that deliberative process privilege applies to early drafts of Air Force report on use of herbicides in Vietnam despite public release of the final report). This limited approach to waiver in the executive privilege context is designed to ensure that agencies do not forego voluntarily disclosing some privileged material out of the fear that by doing so they are exposing other, more sensitive documents. See Assembly of the State of California v. United States Department of Commerce, 968 F.2d 916, 922 n. 5 (9th Cir.1992); Mobil Oil Corp., 879 F.2d at 701; Mehl, 797 F.Supp. at 47-48.

On that basis, we find that the White House’s release of the final report does not waive the privilege in regard to the documents the White House generated in producing the ultimate version. However, the White House has waived its claims of privilege in regard to the specific documents that *742it voluntarily revealed to third parties outside the White House, namely the final report itself and the typewritten text of document 63, which was sent to Espy’s Counsel. Our review reveals that none of the withheld documents is identical to the final White House Counsel report, that no other withheld document is identical to document 63 and that document 63 has handwritten notations that the White House claims were not on the document sent to Espy’s counsel. Thus, although the White House has waived its privileges regarding the typed text of document 63, the handwritten notations remain subject to our privilege analysis, and if found privileged can be redacted from document 63 before it is released to the grand jury.

In sum, with the exception of document 63 we find that the White House has not waived its privileges as to the withheld documents. We therefore proceed to determine the merits of the White House’s claims of privilege.

III. The Presidential Communications Privilege

Judicial discussion of the presidential communications privilege exploded in the early to mid-1970s when the investigation into the Watergate break-in uncovered the fact that President Nixon had made, and still had in his possession, tape recordings of his conversations in the Oval Office and other locales. This revelation led the Watergate Special Prosecutor to subpoena the tapes for use in the criminal investigation of the break-in. President Nixon asserted the presidential communications privilege in response, and also in several subsequent lawsuits that sought access' to the tapes and other presidential materials generated by his administration. These lawsuits, referred to generically as the Nixon cases, remain a quarter century later the leading — if not the only— decisions on the scope of the presidential communications privilege. We begin our analysis of the White House’s assertion of the presidential privilege in this case by examining in detail the precedent in the Nixon cases. We will then address two specific issues regarding the scope and operation of the privilege presented by this case that are not expressly answered by the earlier decisions: how far down the line of command from the President does the presidential privilege extend, and what kind of demon'stration of need must be shown to justify release to a grand jury of materials that qualify for such a privilege.

A. The Nixon Cases and the General Contours of the Presidential Communications Privilege

We first addressed President Nixon’s assertion of the presidential privilege over the Watergate tapes in Sirica. Sirica involved a subpoena for nine tapes issued by the grand jury investigating the Watergate break-in. The district' court had ordered President Nixon to produce the tapes for in camera review, and on appeal we affirmed that decision, stating that “application of Executive privilege depends on a weighing of the public interest protected by the privilege against the public interests that would be served by disclosure in a particular .case.” 487 F.2d at 716. We initially recognized a “great public interest” in preserving “the confidentiality of conversations that take place in the President’s performance of his official duties” because such confidentiality is needed to protect “the effectiveness of the executive decision-making process,” as a result, we said, presidential conversations “are presumptively privileged.” Id. at 717. But we further held that this privilege could be overcome by a sufficient showing of need by a grand jury, and ruled that President Nixon’s assertion of privilege “must fail in the face of the uniquely powerful showing made by the Special Prosecutor in this case.” Id. We ordered that the tapes be turned over to the court for in camera review, however, rather than given to the grand jury directly, to ensure that only material relevant to the Watergate inquiry was released. Id. at 719-22.

President Nixon did not appeal our decision in Sirica, and thus it was not until a year later, in Nixon, that the question of whether an executive privilege of confidentiality for presidential communications existed reached the Supreme Court. Nixon concerned a subpoena issued by the Watergate Special Prosecutor for additional tapes, this *743time for use in the pending trial of seven individuals indicted by the Watergate grand jury. In a unanimous opinion, the Court agreed that there was “a presumptive privilege for Presidential communications,” 418 U.S. at 708, 94 S.Ct. at 3107, founded on “a President’s generalized interest in confidentiality.” Id. at 711, 94 S.Ct. at 3109. It found such a privilege necessary to guarantee the candor of presidential advisers and to provide “[a] President and those who assist him ... [with] free[dom] to explore alternatives in the process of shaping policies and making decisions and to do so in a way many would be unwilling to express except privately.” 418 U.S. at 708, 94 S.Ct. at 3107. Although not expressly provided for in the Constitution, the privilege nonetheless has constitutional origins; it is “inextricably rooted in the separation of powers under the Constitution,” id., and also “flow[s] from the nature of enumerated powers” of the President. Id at 705 & n. 16, 94 S.Ct. at 3106 & n. 16. But, the Court insisted, “neither the doctrine of separation of powers, nor the need for confidentiality of high-level communications, without more, can sustain an absolute, unqualified Presidential privilege of immunity from judicial process under all circumstances.” Id. at 706, 94 S.Ct. at 3106. Turning to the precise issue at hand, the Court held that an assertion of executive privilege “based only on the generalized interest in confidentiality .... must yield to the demonstrated, specific need for evidence in a pending criminal trial.” Id. at 713, 94 S.Ct. at 3110.12 The Court remanded for the district court to perform an in camera review in which relevant and admissible evidence in the tapes would be isolated for release to the Special Prosecutor; the confidentiality of non-relevant material on the tapes was to be preserved. On remand, the President was also to be given an opportunity to raise more particularized claims of privilege. Id. at 714-15 & n. 21, 94 S.Ct. at 3110-11 & n. 21.

The Nixon Court explicitly limited its ruling to demands for presidential materials relevant to a criminal trial, stating “[w]e are not here concerned with the balance between the President’s generalized interest in confidentiality and the need for relevant evidence in civil litigation, nor with that between the confidentiality interest and congressional demands for information.” Id. at 712 n. 19, 94 S.Ct. at 3109 n. 19. It fell to the remaining Nixon eases to address the scope of the presidential communications privilege in other contexts.13 In Senate Committee, a deei*744sion that pre-dated Nixon, this court refused to enforce a subpoena for tapes issued by the Senate Committee investigating illegal activities connected to the 1972 election, on the grounds that the Senate Committee had not demonstrated that the tapes were “demonstrably critical to the responsible fulfilment of the Committee’s functions.” 498 F.2d at 731. Subsequently, the Court of Claims held that the presidential communications privilege could be overcome by the evidentiary demands of a civil trial, see Sun Oil Co. v. United States, 206 CT.C1. 742, 514 F.2d 1020, 1024 (Ct.C1.1975), and in Dellums v. Powell this court agreed, holding that an adequate showing of need in a civil trial would also defeat the privilege “at least where, as here, the action is tantamount to a charge of civil conspiracy among high officers of government to deny a class of citizens their constitutional rights and where there has been sufficient evidentiary substantiation to avoid the inference that the demand reflects mere harassment.” 561 F.2d 242, 247 (D.C.Cir.1977); see also Dellums v. Powell, 642 F.2d 1351 (D.C.Cir.1980) (remanding to give President Nixon further opportunity to assert more particularized claims of privilege).

The Supreme Court had its next encounter with the presidential communications privilege in Nixon v. Administrator of General Services (GSA), which concerned the operation of the privilege in the context of congressional legislation.14 Congress enacted the Presidential Recordings and Materials Preservation Act (“PRMPA”), which transferred custody of the Nixon tapes along with a vast number of other presidential documents from the Nixon administration to the custody of the General Services Administrator. President Nixon challenged PRMPA as unconstitutional, in part because it infringed on the presidential privilege. The Court first held that a former President could assert the privilege on his own, but his claim would be given less weight than that of an incumbent President. 433 U.S. 425, 449, 97 S.Ct. 2777, 2793, 53 L.Ed.2d 867 (1977). Moreover, it said the privilege was “limited to communications ‘in performance of [a President’s] responsibilities,’ ‘of his office,’ and made ‘in the process of shaping policies and making decisions.’ ” Id. at 449, 97 S.Ct. at 2793 (quoting Nixon) (citations omitted). The Court then noted that the only intrusion into the confidentiality of presidential communications in the case was the screening of the materials by archivists, since the statute provided that the Administrator would promulgate regulations which allowed claims of privilege to be raised before public access occurred. This screening by government archivists who had performed the same task for past Presidents without any apparent interference with presidential confidentiality was viewed by the Court as “a very limited intrusion,” and also as justified in light of the substantial public interests served by the Act. Id. at 450-55, 97 S.Ct. at 2793.15

The Nixon cases establish the contours of the presidential communications privilege^ The President can invoke the privilege when asked to produce documents or other materials that reflect presidential decisionmaking and deliberations and that the President believes should remain eonfidentiáT) If the President does so, the documents become presumptively privileged. 16 *745However, the privilege is qualified, not absolute, and can be overcome by an adequate showing of need. If a court believes that an adequate showing of need has been demonstrated, it should then proceed to review the documents in camera to excise non-relevant material. The remaining relevant material should be released. Further, the President should be given an opportunity to raise more particularized claims of privilege if a court rules that the presidential communications privilege alone is not a sufficient basis on which to withhold the document.

While the presidential communications privilege and the deliberative process privilege are closely affiliated, the two privileges are distinct and have different scopes. Both are executive privileges designed to protect executive branch decisionmaking, but one applies to decisionmaking of executive officials generally, the other specifically to decisionmaking of the President. The presidential privilege is rooted in constitutional separation of powers principles and the President’s unique constitutional role; the deliberative process privilege is primarily a common law privilege. See Fitzgerald, 457 U.S. at 753 & n. 35, 102 S.Ct. at 2703 & n. 35. Consequently, congressional or judicial negation of the presidential communications privilege is subject to greater scrutiny than denial of the deliberative privilege. See 26A Charles Alan Wright & Kenneth W. Graham, Jr., Federal Practice and Procedure § 5673, at 37; contra Freund, supra, at 20 (commenting that question of whether presidential privilege is rooted in the common law or the Constitution is not “very meaningful,” but not discussing effect different derivation has on congressional power).

In addition, unlike the deliberative process privilege, the presidential communications privilege applies to documents in their entirety, and covers final and postdecisional materials as well as pre-deliberative ones. Even though the presidential privilege is based on the need to preserve the President’s access to candid advice, none of the cases suggest that it encompasses only the deliberative or advice portions of documents. Indeed, Nixon argued that the presidential privilege must be qualified to ensure full access to facts in judicial proceedings, thereby assuming that factual material comes under the privilege. 418 U.S. at 709, 94 S.Ct. at 3108; but see Larkin, supra, § 6.01 at 6-1 (asserting, without explanation, that the presidential privilege does not “protect purely factual material”). There is no indication either that the presidential privilege is restricted to pre-decisional materials. GSA cautioned that the privilege only applies to communications made in the process of arriving at presidential decisions, but by this we believe the Court meant that the privilege was limited to materials connected to presidential decisionmaking, as opposed to other executive branch decisionmaking, and not that only pre-decisional materials were covered. 433 U.S. at 449, 97 S.Ct. at 2793. Nor would exclusion of final or post-decisional materials make sense, given the Nixon cases’ concern that the President be given sufficient room to operate effectively. These materials often will be revelatory of the President’s deliberations — as, for example, when the President decides to pursue a particular course of action, but asks his advisers to submit follow-up reports so that he can monitor whether this course of action is likely to *746be successful. The release of final and postdecisional materials would also limit the President’s ability to communicate his decisions privately, thereby interfering with his ability to exercise control over the executive branch.17

Finally, while both the deliberative process privilege and the presidential privilege are qualified privileges, the Nixon cases suggest that the presidential communications privilege is more difficult to surmount. In regard to both, courts must balance the public interests at stake in determining whether the privilege should yield in a particular case, and must specifically consider the need of the party seeking privileged evidence. But this balancing is more ad hoc in the context of the deliberative process privilege, and includes consideration of additional factors such as whether the government is a party to the litigation. Moreover, the privilege disappears altogether when there is any reason to believe government misconduct occurred. On the other hand, a party seeking to overcome the presidential privilege seemingly must always provide a focused demonstration of need, even when there are allegations of misconduct by high-level officials.18 In holding that the Watergate Special Prosecutor had provided a sufficient showing of evidentiary need to obtain tapes of President Nixon’s conversations, the Supreme Court made no mention of the fact that the tapes were sought for use in a trial of former presidential assistants charged with engaging in a criminal conspiracy while in office. Accord Senate Committee, 498 F.2d at 731 (noting .that presidential privilege is not intended to shield governmental misconduct but arguing that showing of need turns on extent to which subpoenaed evidence is necessary for government institution to fulfill its responsibilities, not on type of conduct evidence may reveal); contra 26A Wright & Graham, supra, § 5673, at 53-54 (quoting Senate Committee’s not-a-shield language and arguing that allegations of misconduct qualify the privilege, but not addressing Senate Committee’s comment that need showing turns on function for which evidence is sought and not on conduct revealed by evidence).

These differences between the presidential communications privilege and the deliberative privilege demonstrate that the presidential privilege affords greater protection against disclosure. Consequently, should we conclude as to any document that the presidential privilege applies but that the OIC has demonstrated a sufficient showing of need, there is no reason to examine whether the documents also come under the deliberative process privilege. A fortiori, if release is required under the presidential privilege, it will certainly be required under the deliberative process privilege. Hence, we would need to address application of the deliberative process privilege as to any document only if we determine that the withheld document is not subject to the presidential privilege.

B. How Far Down the Line Does the Presidential Communications Privilege Go?

The withheld documents in this case include materials used in the investigation and formulation of several earlier drafts of the White House Counsel’s report, notes of meetings among White House advisers, and draft press briefings. It is undisputed that none of these documents was actually viewed by the President. As a result, the key issue in this case is whether any, and if so which, of these documents come under the presidential communications privilege. Does the privilege only extend to direct communications with the President, or does it extend further to include communications that in*747volve his chief advisers? And if the privilege does extend past the President, how far down into his circle of advisers does it extend?

Most of the Nixon cases involved subpoenas for tapes of conversations in which President Nixon was a participant, and did not call upon the courts to determine whether the presidential privilege also covered communications in which the President did not directly participate.19 The language used to describe the scope of the privilege in the opinions vacillates between broad and narrow depictions of the privilege. In Nixon the Court referred to “[a] President’s acknowledged need for confidentiality in the communications of his office,” 418 U.S. at 712-13, 94 S.Ct. at 3110 (emphasis added) and elaborated that “[a] president and those who assist him must be free to explore alternatives in the process of shaping policies and making decisions,” id. at 708, 94 S.Ct. at 3107 (emphasis added), suggesting that actual presidential involvement in the communication is not a prerequisite to privilege. See also id. at 705, 94 S.Ct. at 3106 (privilege grounded in the need to protect “communications between high Government officials and those who advise and assist them in the performance of their manifold duties”). But Nixon also uses language that appears to tie the privilege to the President; the opinion repeatedly refers to the privilege as a “privilege of confidentiality of Presidential communications,” id. at 705, 94 S.Ct. at 3106 (emphasis added), and as rooted in “[t]he expectation of a President to the confidentiality of his conversations and correspondence.” Id. at 708, 94 S.Ct. at 3107 (emphasis added). Similar variation can be found in Sirica, which describes the privilege interchangeably as designed to “protect the effectiveness of the executive decision-making process” and as intended to “maintain[ ] the confidentiality of conversations that take place in the President’s performance of his official duties.” 487 F.2d at 717 (emphasis added); see also Dellums, 561 F.2d at 246, 247 (describing the privilege at one point as covering “confidential communications with the President” and at another as “attaching] to the communications, submissions and deliberations essential to the conduct of the office of the [President”).

The scope of the presidential communications privilege did arise in GSA and in Sun Oil, but was not decided in either opinion. Many of the documents which PRMPA gave over to GSA custody had never been seen by the President. After remarking that President Nixon could “legitimately assert the Presidential privilege, of course, only as to those materials whose contents fall within the scope of the privilege,” the Court noted that “[o]f the estimated 42 million pages of documents and 880 tape recordings whose custody is at stake, the District Court concluded that the appellant’s claim of Presidential privilege could apply at most to the 200,000 items with which the appellant was personally familiar.” 433 U.S. at 449, 97 S.Ct. at 2793 (emphasis added); see also id. at 454, 97 S.Ct. at 2795-96 (only a “small fraction of the materials ... implicate Presidential confidentiality”). Since, however, the Court found that the public interests served by PRMPA were sufficient to overcome the presidential communications privilege, it never had to decide which materials came under the privilege. The three-member district court that upheld the statute had explicitly commented that it need not consider “whether the privilege that attaches to presidential communications extends to communications never directly received by the President but rather channelled in a variety of ways to him or his advisers,” because it believed the statute would be constitutional “even if a large proportion of the materials falling within the Act were thought protected.” Nixon v. Administrator of General Servs., 408 F.Supp. 321, 345 n. 29 (D.D.C.1976). The same situation occurred in Sun Oil, which involved a •claim of presidential communications privilege over memoranda that circulated between two presidential aides. The Court of Claims never discussed whether the memoranda actually came under the privilege, but rather assumed the privilege applied and *748held that even so the memoranda should be released because the plaintiffs had made out a sufficient showing of need. 514 F.2d at 1022, 1024.

A case that did directly touch on the question of how far down the line the presidential communications privilege extends was Association of American Physicians and Surgeons v. Clinton (AAPS). AAPS involved an effort to enjoin President Clinton’s Task Force on National Health Care Reform and its subgroups from meeting unless they complied with the Federal Advisory Committee Act (FACA). In holding that FACA’s exemption for advisory groups composed solely of officers or employees of the government applied to the Task Force even though it was chaired by the President’s wife, Hillary Rod-ham Clinton, this court commented that an interpretation of FACA as covering a Task Force that reports directly to the President might well represent an unconstitutional intrusion on the presidential communications privilege. This privilege, we argued, “attaches not only to direct communications with the President, but also to discussions between his senior advisers[, who] ... must be able to hold confidential meetings to discuss advice they secretly will render to the President.” 997 F.2d 898, 909 (D.C.Cir.1993). But in AAPS this court did not actually rule on the scope of the privilege, or determine whether the public interests underlying FACA justified interference with the privilege, since it found that “a strong argument” could be made for exempting the Task Force based on the statutory text. Id. at 905.20

There are acknowledgedly strong arguments in favor of holding that the presidential communications privilege applies to only those communications that directly involve the President. This approach comports with the principle that “the President’s unique status under the Constitution distinguishes him from other executive officials,” Fitzgerald, 457 U.S. at 750, 102 S.Ct. at 2701, particularly in separation of powers analysis. See Wetlaufer, supra, at 901-02. The Constitution after all vests the executive power not in the executive branch, but in the President; it is the President who, as “the chief constitutional officer of the Executive branch, [is] entrusted with supervisory and policy responsibilities of the utmost discretion and sensitivity.” Fitzgerald, 457 U.S. at 750, 102 S.Ct. at 2701. Nixon identified the President’s Article II powers and responsibilities as the constitutional basis of the presidential communications privilege. 418 U.S. at 705 & n. 16, 94 S.Ct. at 3106 & n. 16. Since the Constitution assigns these responsibilities to the President alone, arguably the privilege of confidentiality that derives from them also should be the President’s alone. The uniqueness of the President has frequently led courts to recognize that the President enjoys more extensive privileges than other executive branch officers. For example, the President is absolutely immune from damages liability for official acts, but presidential aides receive only qualified immunity. Compare Fitzgerald, 457 U.S. at 749-54, 102 S.Ct. at 2701-03, with Harlow v. Fitzgerald, 457 U.S. 800, 808-13, 102 S.Ct. 2727, 2732-36, 73 L.Ed.2d 396 (1982); see also Mitchell v. Forsyth, 472 U.S. 511, 520-24, 105 S.Ct. 2806, 2812-14, 86 L.Ed.2d 411 (1985) (holding whether an executive official receive absolute immunity depends on the function the official was performing when she engaged in the actions being challenged). In Franklin the Court emphasized that the separation of powers concerns that arise when the President is personally subjected to judicial process are not implicated when a court exercises jurisdiction over other executive branch officials. 505 U.S. at 801-02, 112 S.Ct. at 2775-76. And in In re Kessler, this court recently rejected the claim that because the President is allowed to appeal a discovery order without being held in contempt the Commissioner of the Food and Drug Admin*749istration should be able to do so as well, noting that “for purposes of separation of powers, the President stands in an entirely different position than other members of the executive branch.” 100 F.3d 1015, 1017 (D.C.Cir.1996).

An additional reason to restrict the presidential communications privilege to direct communications with the President is the general rule, underscored by the Supreme Court in Nixon, that privileges should be narrowly construed: “exceptions to the demand for every man’s evidence are not lightly created nor expansively construed, for they are in derogation of the search for truth.” 418 U.S. at 710, 94 S.Ct. at 3108; accord Jaffee v. Redmond, — U.S. -, -, —-, 116 S.Ct. 1923, 1932, 1933, 135 L.Ed.2d 337 (1996) (Scalia, J., dissenting); Trammel v. United States, 445 U.S. 40, 50, 100 S.Ct. 906, 912, 63 L.Ed.2d 186 (1980); In Re Grand Jury Subpoena Duces Tecum, 112 F.3d 910, 918 (8th Cir.1997). The argument for a narrow construction is particularly strong in cases like this one where the public’s ability to know how its government is being conducted is at stake. In performing his constitutional duties the President may obtain advice and assistance from a broad array of executive officials — cabinet officers, employees in the Executive Office of the President, and agency staff with special expertise, as well as individuals whose sole function in the White House is to provide the President with advice and assistance. See, e.g., Meyer v. Bush, 981 F.2d 1288, 1293-94 (D.C.Cir.1993) (holding President’s Task Force on Regulatory Relief was intended only to advise and assist the President and was not subject to FOIA, even though the Task Force included cabinet officers as members). Indeed, it has been publicly noted that the parts of the executive branch which “directly report[] to the President ha[ve] grown dramatically in the past few decades,” Peter M. Shane, Legal Disagreement and Negotiation in a Government of Laws: The Case of Executive Privilege Claims Against Congress, 71 Minn. L. Rev. 461, 463 (1987); see also Thomas E. Cronin, The State of the Presidency 243-47 (2d ed.1980) (discussing growth of White House staff and its effects).

Extending presidential privilege to the communications of presidential advisers not directly involving the President inevitably creates the risk that a broad array of materials in many areas of the executive branch will become “sequesterfed]” from public view. Wolfe, 815 F.2d at 1533. President Nixon’s attempt to invoke presidential privilege to prevent release of evidence indicating that high level executive officers engaged in illegal acts is perhaps the starkest example of potential for abuse of the privilege. And openness in government has always been thought crucial to ensuring that the people remain in control of their government. According to James Madison,

[a] popular Government, without popular information, or the means of acquiring it, is but a Prologue to a Farce or a Tragedy; or perhaps both. Knowledge will forever govern ignorance: And a people who mean to be their own Governors, must arm themselves with the power which knowledge gives.

Letter from James Madison to W.T. Barry (Aug. 4,1822), in 9 Writings of James Madison 103 (GaiJlard Hunt, ed.1910); see also Soucie, 448 F.2d at 1080 (In enacting FOIA, “Congress recognized that the public cannot make intelligent decisions without [adequate] information, and that governmental institutions become unresponsive to public needs if knowledge of their activities is denied to the people and their representatives”). The very reason that presidential communications deserve special protection, namely the President’s unique powers and profound responsibilities, is simultaneously the very reason why securing as much public knowledge of presidential actions as is consistent with the needs of governing is of paramount importance.

But a very powerful case can also be made for extending the presidential communications privilege beyond those materials with which the President is “personally familiar,” and at the end of the day we find the arguments for a limited extension of the privilege beyond the President to his immediate advisers more convincing. Nixon does not specifically establish how far down the chain of *750command the presidential communication privilege extends, but it does make absolutely clear that the privilege itself is rooted in the need for confidentiality to ensure that presidential decisionmaking is of the highest caliber, informed by honest advice and full knowledge. Confidentiality is what ensures the expression of “candid, objective, and even blunt or harsh opinions” and the comprehensive exploration of all policy alternatives before a presidential course of action is selected. See Nixon, 418 U.S. at 708, 94 S.Ct. at 3107-08; see also GSA, 433 U.S. at 449, 97 S.Ct. at 2793. Several commentators have argued that presidential advisers may not be as likely to “temper candor with a concern for appearances and for their own interests to the detriment of the decisionmaking process,” Nixon, 418 U.S. at 705, 94 S.Ct. at 3106, as the Supreme Court feared. See, e.g., Wetlaufer, supra, at 886-90; 26A Wright & Miller, supra, § 5673 at 38-39. But — even if we were free to ignore Nixon, which we are not — we are not so sanguine that presidential advisers will never be dissuaded from expressing unpopular but correct opinions out of a fear of disclosure, or that able individuals will not shrink from assuming a position as presidential adviser in the first place if by doing so they step unprotected into the limelight. And the critical role that confidentiality plays in ensuring an adequate exploration of alternatives cannot be gainsaid. If presidential advisers must assume they will be held to account publicly for all approaches that were advanced, considered but ultimately rejected, they will almost inevitably be inclined to avoid serious consideration of novel or controversial approaches to presidential problems.

Presidential advisers do not explore alternatives only in conversations with the President or pull their final advice to him out of thin air — if they do, their advice is not likely to be worth much. Rather, the most valuable advisers will investigate the factual context of a problem in detail, obtain input from all others with significant expertise in the area, and perform detailed analyses of several different policy options before coming to closure on a recommendation for the Chief Executive. The President himself must make decisions relying substantially, if not entirely, on the information and analysis supplied by advisers. “Even the most sensitive issues of national security must be brought to the point of presidential decision by staff, who assemble data and views, and then winnow and shape them for the President.” Peter L. Strauss, The Place of Agencies in Government: Separation of Powers and the Fourth Branch, 84 Colum. L. Rev. 573, 661 (1984). In the vast majority of cases, few if any of the documents advisers generate in the course of their own preparation for rendering advice to the President, other than documents embodying their final recommendations, will ever enter the Oval Office. Yet these pre-decisional documents are usually highly revealing as to the evolution of advisers’ positions and as to the different policy options considered along the way. If these materials are not protected by the presidential privilege, the President’s access to candid and informed advice could well be significantly circumscribed.

The protection offered by the more general deliberative process privilege will often be inadequate to ensure that presidential advisers provide knowledgeable and candid advice, primarily because the deliberative process privilege does not extend to purely factual material. As we remarked in AAPS, preservation of the President’s confidentiality requires that a “[gjroup directly reporting and advising the President must have confidentiality at each stage in the formulation of advice to him.” 997 F.2d at 910. In many instances, potential exposure of the information in the possession of an adviser can be as inhibiting as exposure of the actual advice she gave to the President. Without protection for her sources of information, an adviser may be tempted to forego obtaining comprehensive briefings or initiating deep and intense probing for fear of losing deniability. Exposure of the factual portions of presidential advisers’ communications also represents a substantial threat to the confidentiality of the President’s own deliberations. Knowledge of factual information gathered by presidential advisers can quickly reveal the nature and substance of the issues before the President, since “[i]f you know what informa*751tion people seek, you can usually determine why they seek it.” Id.

The greater ease with which the deliberative process privilege can be overcome is another reason to doubt its efficacy in ensuring candid presidential advice. In Nixon the Supreme Court recognized that some possibility of disclosure is unlikely to affect the advice the President receives, stating “we cannot conclude that advisers will be moved to temper the candor of their remarks by the infrequent occasions of disclosure [that might occur if their] ... conversations will be called for in the context of a criminal prosecution.” 418 U.S. at 712, 94 S.Ct. at 3109. The risk of a chill increases, however, as the possibility of disclosure rises, especially if there are situations in which the privilege may virtually disappear, such as when government misconduct is alleged. Nor does it suffice to respond that the public interest in honest and accountable government is stymied if presidential advisers are allowed even a qualified privilege when government misconduct is charged. The President’s supervisory control over executive branch officials is an important means of ensuring that abuse of office is uncovered and swiftly addressed, and the President needs access to candid and informed advice if he is to exercise this control effectively. In this regard it is worth emphasizing that the presidential communications privilege is, at all times, a qualified one, so that an expansion to cover communications of presidential advisers which do not directly involve the President does not mean that these communications will become permanently shielded; they will remain available upon a sufficient showing of need.

Of course, the risk that release of factual information may reveal a pohcymaking official’s area of focus is true at all levels of government. But the President does not represent simply one level of executive branch, but rather the ultimate level of decisionmaking in the executive branch, and intrusion into presidential deliberations is therefore more serious. In ruling on whether General Wilkinson’s letter should be released Chief Justice Marshall remarked that “[i]n no ease of this kind would a court be required to proceed against the president as against an ordinary individual.” Burr, 25 Fed. Cas., at 192. Neither should a court be required to proceed against the President as against any other executive branch official. See Clinton v. Jones, — U.S. at-n. 39, 117 S.Ct. at 1649 n. 39 (quoting Burr and noting “[s]pecial caution is appropriate if the materials or testimony sought by the court relate to a President’s official activities”). Indeed, if the President’s immediate advisers were only covered by the deliberative process privilege, courts might feel compelled to extend the deliberative privilege to cover factual material in order to ensure that the President had sufficient freedom from public review to operate effectively. This result might make the deliberative process privilege better able to meet the particular needs of presidential decisionmaking, but it would hardly advance the goal of open government since it would mean that more factual information was shielded at all levels of the executive branch.

The ultimate question is whether restricting the presidential communications privilege to communications that directly involve the President will “impede the President’s ability to perform his constitutional duty.” Morrison v. Olson, 487 U.S. 654, 691, 108 S.Ct. 2597, 2619, 101 L.Ed.2d 569 (1988); see also Loving v. United States, — U.S.-, --, 116 S.Ct. 1737, 1743, 135 L.Ed.2d 36 (1996) (“[e]ven when a branch does not arrogate power to itself, ... the separation-of-powers doctrine requires that a branch not impair another in the performance of its constitutional duties”). If it does, the constitutional separation of powers will be violated. In Nixon the Court recognized that the President’s access to honest and informed advice and his ability to explore possible policy options privately are critical elements in presidential decisionmaking. Given the President’s dependence on presidential advisers and the inability of the deliberative process privilege to provide advisers with adequate freedom from the public spotlight, we conclude that limiting the privilege in this fashion would indeed impede effective functioning of the presidency.

We believe therefore that the public interest is best served by holding that *752communications made by presidential advisers in the course of preparing advice for the President come under the presidential communications privilege, even when these communications are not made directly to the President. Given the need to provide sufficient elbow room for advisers to obtain information from all knowledgeable sources, the privilege must apply both to communications which these advisers solicited and received from others as well as those they authored themselves. The privilege must also extend to communications authored or received in response to a solicitation by members of a presidential adviser’s staff, since in many instances advisers must rely on their staff to investigate an issue and formulate the advice to be given to the President. We are aware that such an extension, unless carefully circumscribed to accomplish the purposes of the privilege, could pose a significant risk of expanding to a large swath of the executive branch a privilege that is bottomed on a recognition of the unique role of the President.21 In order to limit this risk, the presidential communications privilege should be construed as narrowly as is consistent with ensuring that the confidentiality of the President’s decisionmaking process is adequately protected. Not every person who plays a role in the development of presidential advice, no matter how remote and removed from the President, can qualify for the privilege. In particular, the privilege should not extend to staff outside the White House in executive branch agencies. Instead, the privilege should apply only to communications authored or solicited and received by those members of an immediate White House adviser’s staff who have broad and significant responsibility for investigating and formulating the advice to be given the President on the particular matter to which the communications relate. Only communications at that level are close enough to the President to be revelatory of his deliberations or to pose a risk to the candor of his advisers. See AAPS, 997 F.2d at 910 (it is “operational proximity” to the President that matters in determining whether “[t]he President’s confidentiality interest” is implicated) (emphasis omitted).

Of course, the privilege only applies to communications that these advisers and their staff author or solicit and receive in the course of performing their function of advising the President on official government matters. This restriction is particularly important in regard to those officials who exercise substantial independent authority or perform other functions in addition to advising the President, and thus are subject to FOIA and other open government statutes. See Armstrong v. Executive Office of the President, 90 F.3d 553, 558 (D.C.Cir.1996), cert. denied, — U.S. -•, 117 S.Ct. 1842, 137 L.Ed.2d 1046 (1997). The presidential communications privilege should never serve as a means of shielding information regarding governmental operations that do not call ultimately for direct decisionmaking by the President. If the government seeks to assert the presidential communications privilege in regard to particular communications of these “dual hat” presidential advisers, the government bears the burden of proving that the communications occurred in conjunction with the process of advising the President.

In this case the documents in question were generated in the course of advising the President in the exercise of his appointment and removal power, a quintessential and nondelegable Presidential power.22 In many instances, presidential powers *753and responsibilities, for example the duty to take care that the laws are faithfully executed, can be exercised or performed without the President’s direct involvement, pursuant to a presidential delegation of power or statutory framework. Cf. Morrison, 487 U.S. at 691-92, 108 S.Ct. at 2619-20 (requirement that Independent Counsel can be removed only for good cause is not an unconstitutional restriction on the President’s powers). But the President himself must directly exercise the presidential power of appointment or removal. As a result, in this ease there is assurance that even if the President were not a party to the communications over which the government is asserting presidential privilege, these communications nonetheless are intimately connected to his presidential decisionmaking. In addition, confidentiality is particularly critical in the appointment and removal context; without it, accurate assessments of candidates and information on official misconduct may not be forthcoming. See, e.g., Wash. Legal Foundation v. Department of Justice, 691 F.Supp. 483, 495 (D.D.C.1988), aff'd sub nom. Public Citizen v. Department of Justice, 491 U.S. 440, 109 S.Ct. 2558, 105 L.Ed.2d 377 (1989) (underscoring the “unique need for confidentiality” in the President’s appointment of federal judges).

Finally, we underscore our opinion should not be read as in any way affecting the scope of the privilege in the congressional-executive context, the arena where conflict over the privilege of confidentiality arises most frequently. The President’s ability to withhold information from Congress implicates different constitutional considerations than the President’s ability to withhold evidence in judicial proceedings. See, e.g., Rozell, supra, at 142-57; Norman Dorsen & John H.F. Shattuck, Executive Privilege, the Congress and the Courts, 35 Ohio St. L.J. 1, 16-22, 24-33 (1974). Our determination of how far down into the executive branch the presidential communications privilege goes is limited to the context before us, namely where information generated by close presidential advisers is sought for use in a judicial proceeding, and we take no position on how the institutional needs of Congress and the President should be balanced.

C. Standard of Need

The question of whether the presidential communications privilege applies to communications that do not involve the President is only the first issue we must resolve before turning to an application of the privilege here. We must also determine what type of showing of need the OIC must make in defense of the grand jury subpoena in order to overcome the privilege.

Nixon, GSA, Sirica, and the other Nixon eases all employed a balancing methodology in analyzing whether, and in what circumstances, the presidential communications privilege can be overcome. Under this methodology, these opinions balanced the public interests served by protecting the President’s confidentiality in a particular context with those furthered by requiring disclosure. Since Nixon and Sirica clearly establish that the presidential communications privilege can be overcome by a sufficient showing that subpoenaed evidence is needed for a criminal judicial proceeding, our task is not to weigh anew the public interest in preserving confidentiality against the public interest in assuring fair trials and enforcing the law. Rather, our task is to determine precisely what guidance these cases provide on what counts as a sufficient showing of need in our situation, and more specifically to clarify whether there is any difference between the need standard this court established in Sirica in regard to a grand jury subpoena and the standard articulated by the Supreme Court one year later in Nixon for a criminal trial subpoena.

At the end of its discussion of the presidential communications privilege in Nixon, the Supreme Court stated that the privilege “must yield to the demonstrated, specific need for evidence in a pending criminal trial.” 418 U.S. at 713, 94 S.Ct. at 3110. What the Court meant by a “demonstrated, specific need” is debatable. Compare Cox, supra, at 1414-15 (“[t]he critical test [under Nixon] is *754probably relevance and admissibility”) with Freund, supra, at 31 (Nixon appears to require “a stronger showing of need” than just relevancy). After setting forth this need standard, the Court tersely commented that “[o]n the basis of our examination of the record we are unable to conclude that the District Court erred” in finding that the Watergate Special Prosecutor had made a sufficient showing of need to overcome the presidential privilege; it never explained what parts of the record led it to this conclusion. Id. at 714, 94 S.Ct. at 3110-11. The only occasion where the Court discusses in any detail the showing of need that the Special Prosecutor actually made comes in its analysis of whether the subpoena satisfied Federal Rule of Criminal Procedure 17(e), which governs all subpoenas for documents and materials made in criminal proceedings. The Court concluded that the subpoena met Rule 17(c)’s tripartite requirement of relevancy, admissibility and specificity; the Special Prosecutor’s supporting materials, which listed the date, time and participants in the conversations sought and provided testimony regarding the content of some conversations, established “a sufficient likelihood that each of the tapes contains conversations relevant to the offenses charged in the indictment” and that these conversations would be admissible. Id. at 700, 94 S.Ct. at 3103. The Nixon Court’s failure to elaborate on the demonstrated, specific need standard or provide any further analysis of the Special Prosecutor’s showing led one judge to comment that to overcome presidential privilege “the Court does not appear to have meant anything more than the showing that satisfied Rule 17(c).” North, 910 F.2d at 952 (Silberman, J., concurring in part and dissenting in part). Further, the Court offered varying characterizations of when the presidential communications privilege would be overcome, at one juncture suggesting the privileged material must be “ ‘essential to the justice of the [pending criminal] ease,”’ Nixon, 418 U.S. at 713, 94 S.Ct. at 3110 (quoting Burr, 25 Fed. Cas. at 192), and at others simply that the material must be “preliminarily shown to have some bearing on the pending criminal cases.” Nixon, 418 U.S. at 713, 94 S.Ct. at 3109; see also id. at 712 n. 19, 94 S.Ct. at 3109 n. 19 (referring to the “constitutional need for relevant evidence”) (emphasis added).

It would be strange indeed if Nixon required nothing more to overcome presidential privilege than the initial showing of relevancy, admissibility and specificity necessary to satisfy Rule 17(c) in all cases, even in cases where no claim of privilege is raised. If this were true, the privilege would have no practical benefit. That the Nixon Court believed overcoming the presidential privilege required something more than the ordinary Rule 17(c) showing is apparent from its statement, made at the outset of the discussion of presidential privilege, that “[h]aving determined that the requirements of Rule 17(c) were satisfied, we turn to the claim that the subpoena should be quashed because it demands confidential conversations between a President and his close advisers.” 418 U.S. at 703, 94 S.Ct. at 3105 (internal quotations omitted); see also id. at 713-14, 94 S.Ct. at 3110-11 (distinguishing between inquiry into whether a subpoena was properly issued and review of claim of privilege raised on return of a properly issued subpoena). However, the opinion also cannot be read as demanding that the information sought must be shown to be critical to an accurate judicial determination; such a view simply is incompatible with the Court’s repeated emphasis on the importance of access to relevant evidence in a criminal proceeding.

We conclude that Nixon’s demonstrated, specific need standard has two components. A party seeking to overcome a claim of presidential privilege must demonstrate: first, that each discrete group of the subpoenaed materials likely contains important evidence; and second, that this evidence is not available with due diligence elsewhere. The first component, likelihood of containing important evidence, means that the evidence sought must be directly relevant to issues that are expected to be central to the trial. In practice, this component can be expected to have limited impact, since Rule 17(c) precludes use of a trial subpoena to obtain evidence that is not relevant to the charges being prosecuted or where the claim that subpoenaed materials will contain such evi*755denee represents mere speculation. See, e.g., Nixon, 418 U.S. at 699-700, 94 S.Ct. at 3103-04; United States v. Arditti 955 F.2d 331, 345-46 (5th Cir.1992); Ehrlichman, 559 F.2d at 75-76. But to the extent that Rule 17(c) allows a defendant to subpoena evidence that would be only tangentially relevant or would relate to side issues, the first component of the need standard would come into play. See, e.g., Nixon, 418 U.S. at 701, 94 S.Ct. at 3104 (“Generally, the need for evidence to impeach witnesses is insufficient to require its production in advance of trial.”); Bowman Dairy Co. v. United States, 341 U.S. 214, 219, 71 S.Ct. 675, 678, 95 L.Ed. 879 (1951) (materials can be reached under Rule 17(c) “as long as they are evidentiary”); In re Martin Marietta Corp., 856 F.2d 619, 622 (4th Cir.1988) (upholding subpoena on grounds that materials were “clearly of evidentiary value”). The second component, unavailability, reflects Nixon’s insistence that privileged presidential communications should not be treated as just another source of information. See North, 910 F.2d at 952 n. 29 (Silberman, J., concurring in part and dissenting in part) (acknowledging that one possible difference between the showing necessary to satisfy Rule 17(c) and Nixon’s need standard is that the latter “would also require a showing that the evidence is unavailable from any source other than the President”). Efforts should first be made to determine whether sufficient evidence can be obtained elsewhere, and the subpoena’s proponent should be prepared to detail these efforts and explain why evidence covered by the presidential privilege is still needed. Of course, there will be instances where such privileged evidence will be particularly useful, as when, unlike the situation here, an immediate White House advisor is being investigated for criminal behavior. In such situations, the subpoena proponent will be able easily to explain why there is no equivalent to evidence likely contained in the subpoenaed materials. Finally, while our view of the Nixon need standard is derived from the opinion’s language and a commonsense understanding of “need,” it is worth noting that the factors of importance and unavailability are also used by courts in determining whether a sufficient showing of need has been demonstrated to overcome other qualified executive privileges, such as the deliberative process privilege or the law-enforcement investigatory privilege. See In re Comptroller of the Currency, 967 F.2d at 634; Friedman, 738 F.2d at 1342.

Nixon, however, involved a trial subpoena; what we have here is a grand jury subpoena. In a post-Nixon decision, United States v. R. Enterprises, Inc., the Court emphasized that the unique function of the grand jury fundamentally differentiates its subpoenas from trial subpoenas. “The function of the grand jury is to inquire into all information that might possibly bear on its investigation, ... [and a]s a necessary consequence of its investigatory function, the grand jury paints with a broad brush.” 498 U.S. 292, 297, 111 S.Ct. 722, 726, 112 L.Ed.2d 795 (1991); accord Branzburg v. Hayes, 408 U.S. 665, 688, 92 S.Ct. 2646, 2660, 33 L.Ed.2d 626 (1972). Requiring grand jury subpoenas to comply with the same requirements of relevancy, admissibility, and specificity under Rule 17(c) as applies to trial subpoenas would impose an impossible burden on the grand jury, create untoward delays, and threaten the secrecy of grand jury proceedings. R. Enters., 498 U.S. at 299, 111 S.Ct. at 727. As a result, the Court concluded that a grand jury subpoena is presumed to be reasonable and the burden is on the subpoena’s opponent to disprove this presumption. Where “a subpoena is challenged on relevancy grounds, the motion to quash must be denied unless the district court determines that there is no reasonable possibility that the category of materials the Government seeks will produce information relevant to the general subject of the grand jury’s investigation.” Id. at 301, 111 S.Ct. at 728.

But then again, R. Enterprises concerned a challenge to a grand jury subpoena only on grounds of relevance; it does not govern a ease, such as this, where the grand jury subpoena is being resisted on grounds of privilege. Instead, the case most directly on point in this respect is Sirica, where this court was specifically confronted with a claim of presidential communications privilege raised against a grand jury subpoena. The OIC does not appear to dispute that Sirica is the governing case here; instead, the OIC *756reads Sirica as establishing a significantly less demanding need standard than Nixon, and argues that this differential is justified in light of R. Enterprises’ insistence that a grand jury subpoena is not held to the same standards as a trial subpoena. According to the OIC, Sirica merely requires that the grand jury demonstrate the evidence it seeks is directly relevant to its investigation in order to overcome the President’s claim of privilege.

The OIC’s position represents too selective a reading of Sirica. To be sure, at times in that opinion we used language suggesting the required demonstration was only that the materials sought were “directly relevant” to the grand jury’s inquiry. For example, we commented that “[t]he exception that we have delineated to the President’s confidentiality privilege depends entirely on the grand jury’s showing that the evidence is directly relevant to its decisions.” 487 F.2d at 719 (emphasis added); see also id. at 705-06. But admittedly we also used language on other occasions indicating that a far more substantial showing was required. We stated that the President’s claim of privilege “must fail in face of the uniquely powerful showing made by the Special Prosecutor ... that the subpoenaed tapes contain evidence peculiarly necessary to the carrying out of [the grand jury’s] vital function — evidence for which no effective substitute is available,” 487 F.2d at 717 (emphasis added), and at another point characterized the Special Prosecutor’s showing as being that “the subpoenaed recordings contain evidence critical to the grand jury’s decisions.” Id. at 706 (emphasis added). We echoed this latter characterization in Senate Committee, where we described Sirica as requiring a demonstration that “the subpoenaed evidence is demonstrably critical to the responsible fulfillment of the [grand jury’s] functions.” 498 F.2d at 731 (emphasis added).

In this instance, we agree with the White House that the Sirica need standard which governs grand jury subpoenas is no more lenient than the need standard enunciated for trial subpoenas in Nixon. In both situations, to overcome the presidential privilege it is necessary to demonstrate with specificity why it is likely that the subpoenaed materials contain important evidence and why this evidence, or equivalent evidence, is not practically available from another source. See In re Grand Jury Subpoena, 112 F.3d at 927, 937 (Kopf, J., dissenting) (arguing that Nixon standard applies to grand jury subpoenas as well as trial subpoenas). On the one hand, to the extent that some of this court’s comments in Sirica suggest that a more substantial showing of need must be made when presidential privilege is raised against a grand jury subpoena than the Supreme Court required in regard to a criminal trial subpoena, we do conclude that these comments have been effectively overruled by R. Enterprises. But R. Enterprises’ emphasis on the special leeway given to grand jury subpoenas as opposed to criminal trial subpoenas absent a claim of privilege does not preclude us from finding that the same need standard applies when the presidential communications privilege is asserted. The necessary breadth of the grand jury’s inquiries in fact supports applying a strict standard of need to overcome presidential privilege, because it means that grand jury subpoenas may well represent a much more frequent threat to presidential confidentiality. The Supreme Court has recognized that “the longstanding principle that the public has a right to every man’s evidence” is limited by valid claims of privilege in grand jury proceedings as elsewhere, even as it held that this principle “is particularly applicable to grand jury proceedings.” Branzburg, 408 U.S. at 688, 92 S.Ct. at 2660 (ellipsis omitted); see also United States v. Calandra, 414 U.S. 338, 344, 346, 94 S.Ct. 613, 618, 619, 38 L.Ed.2d 561 (1974) (while grand jury is “accorded wide latitude,” “the grand jury’s subpoena power is not unlimited” and “[j]udicial supervision is properly exercised” to protect claims of privilege).

Nor do we believe the Nixon/Sirica need standard imposes too heavy a burden on grand jury investigation. In practice, the primary effect of this standard will be to require a grand jury to delay subpoenaing evidence covered by presidential privilege until it has assured itself that the evidence sought from the President or his advisers is both important to its investigation and prac*757tieally unavailable elsewhere. As was true in Sirica, a grand jury will often be able to specify its need for withheld evidence in reasonable detail based on information obtained from other sources. And if it has difficulty in obtaining evidence from other sources, this fact in and of itself will go far toward satisfying the need requirement. Although any showing of need has the potential of undercutting the secrecy of grand jury proceedings, a district court can ensure that such secrecy is protected by provisions for sealed, or when necessary ex parte, filings.

We agree with the OIC in one regard, however. R. Enterprises makes clear that a grand jury subpoena is not subject to the same Rule 17(c) requirements of “relevancy, admissibility and specificity” as a criminal trial subpoena. Since to meet the need standard the grand jury will have to make a specific showing of the importance of the evidence it seeks, its exemption from the relevancy and specificity constraints of Rule 17(c) will not be significant. But the same is not true of the grand jury’s freedom from the requirement of admissibility, and in R. Enterprises the Court underscored that a grand jury is often allowed to consider evidence that would be deemed inadmissible in a criminal trial. 498 U.S. at 298, 111 S.Ct. at 726-27. As a result, the fact that evidence covered by the presidential communications privilege may be inadmissible should not affect a court’s determination of the grand jury’s need for the material.

* * *

Based on our review of the Nixon eases and the purpose of the presidential communications privilege, we conclude that this privilege extends to cover communications which do not themselves directly engage the President, provided the communications are either authored or received in response to a solicitation by presidential advisers in the course of gathering information and preparing recommendations on official matters for presentation to the President. The privilege also extends to communications authored or solicited and received by those members of an immediate White House advisor’s staff who have broad and significant responsibility for investigating and formulating the advice to be given to the President on a particular matter. We also hold that in order to overcome a claim of presidential privilege raised against a grand jury subpoena, it is necessary to specifically demonstrate why it is likely that evidence contained in presidential communications is important to the ongoing grand jury investigation and why this evidence is not available from another source.

IV. Examination of the White House’s Claims of Privilege

Our final task is to apply the principles we have heretofore laid out to the documents withheld in this case. We have concluded that although all of the documents come under the presidential communications privilege, the OIC has demonstrated a sufficient showing of need to obtain certain information in some of the documents. Because we believe that the determination of exactly what evidence should be released is one that the district court should make in the first instance, we do not identify any specific portions of the documents to be released. However, we are supplementing our opinion with a sealed appendix to assist the district court with its in camera review of each document on remand.

A. The Presidential Privilege Applies

The withheld documents consist primarily of outlines of issues and questions that needed to be investigated and drafts of the White House Counsel’s report on the Espy investigation. There are also notes of meetings and phone conversations, lists of information on Espy, and press briefings on Espy. Most of the documents were authored by two associate White House Counsel, a few were authored by top presidential advisers, specifically the White House Counsel, Deputy White House Counsel, Chief of Staff and Press Secretary. A few documents were authored by a legal extern in the White House Counsel’s office, and there are also three documents for which no author is listed. According to the White House privilege log, as well as the headings of the documents themselves, it appears that most of the documents circulated only within the White *758House Counsel’s office. Many of the documents were sent to the White House Counsel or Deputy White House Counsel, or represent notes taken from meetings at which these top advisers and others were present. A sizeable number, however, were either authored by the two associate White House Counsel and not disseminated or sent only to them by others. All of the documents relate to the investigation of Espy that the President asked the White House Counsel to undertake.

The documents that were authored by the White House Counsel, Deputy White House Counsel, Chief of Staff and Press Secretary were communications connected to an official matter on which they were directly advising the President, and thus under the principles laid out in this opinion these documents are clearly covered by the privilege. The same is true of notes taken of meetings on the Espy investigation at which these advisers were present, since these notes reflect these advisers’ communications, and of documents that they solicited and received. As established above, the presidential privilege applies to communications made by a member of an immediate White House adviser’s staff when the staff member has broad and significant responsibility for investigating and formulating the advice to be given the President on the particular matter to which the communications relate. It is clear from a review of the documents that the two associate White House Counsel exercised broad and significant responsibility for gathering information on Espy’s actions and authoring initial drafts of the White House Counsel’s report. Consequently, documents they authored or they solicited and received from others also come under the privilege.

The only question regarding application of the presidential communications privilege here concerns the remaining withheld documents, which consist of those documents authored by the legal extern in the White House Counsel’s office and three documents for which no author is listed. It is apparent that the legal extern did not exercise broad and significant responsibility for the Espy investigation, and therefore the documents authored by the legal extern do not, on their own, qualify for the presidential privilege. However, all of the withheld documents authored by the extern were clearly created at the request of the two associate White House Counsel with broad and significant responsibility for the Espy investigation and were received by them. Therefore, the privilege also applies to these documents. The status of the three no-author documents is more difficult to resolve. Two of these documents were received by the Deputy White House Counsel, and the other by one of the associate White House Counsel with broad and significant responsibility for the Espy investigation. These documents relate to operational details of the Espy investigation. Clearly, if these documents were solicited by the Deputy White House Counsel and the associate White House Counsel, they would be also covered by the privilege. The current description of these documents provided by the White House, however, does not specifically indicate whether these documents were in fact solicited. Ordinarily, the White House would be expected to demonstrate that they had been, but we do not believe a remand for that showing is necessary here because our review of the documents themselves demonstrates that from the nature of their contents and the persons to whom they were directed there can be little question that they had been solicited. As we are setting forth for the first time the principles by which we will determine whether the privilege applies to communications of presidential advisers that do not directly involve the President, we believe it would be unrealistic to expect the White House to have foreseen the need to specifically demonstrate that the documents had been solicited.

In sum, we conclude that all of the documents withheld by the White House here are subject to the presidential communications privilege. As a result, we need not determine whether the documents would qualify for the deliberative process privilege.23

*759B. The OIC’s Demonstration of Need

A preliminary question that must be addressed before we turn to an examination of the OIC’s demonstration of need is whether we should be reviewing this demonstration at all. The procedure envisioned by the Nixon cases, as outlined earlier, is that upon a sufficient showing of need, the President must turn over privileged materials for in camera review, whereupon the court reviews the materials and determines what should be released. This case comes to us in a significantly different posture than Nixon and Sirica. In both of those cases, President Nixon was challenging district court orders that instructed him to submit the subpoenaed tapes for in camera review. In this case, the White House has already turned over the subpoenaed materials for in camera review pursuant to the district court’s order, and did not appeal from that order. Instead, we have before us the OIC’s appeal of the district court’s denial of the OIC’s motion to compel. Thus, we are presented with the question of whether we should forego determining whether or not the OIC made a sufficient showing of need to obtain in camera review, and instead simply instruct the district court to review the withheld documents and determine what evidence should be released.

How we resolve this question could have a significant impact on what materials are disclosed to the grand jury, because the standard applied to determine if the OIC has made a sufficient showing of need to obtain in camera review is much more difficult to satisfy than the standard applied during in camera review to determine exactly what evidence should be released. As we explained in the preceding section, the showing required to obtain in camera review is governed by the Nixon/Sirica need standard and entails demonstrating with specificity that the subpoenaed materials likely contain important evidence and that this evidence, or equivalent evidence, is not practically available from another source. The purpose of this initial showing is to protect the confidentiality of presidential communications; it operates on the presumption that these communications are privileged and requires the subpoena proponent to meet a certain threshold of need before a court will consider releasing any of the communications sought.

The district court’s in camera review also aims to ensure that presidential confidentiality is not unnecessarily breached, but it operates on the presumption that some privileged materials will probably be released. The court’s task during its in camera review is simply to ensure that privileged materials that would not be of use to the subpoena proponent are not released. Nixon, 418 U.S. at 714, 94 S.Ct. at 3110; Sirica, 487 F.2d at 719-21. Nixon makes clear that the court determined what evidence could be of use to the subpoena proponent by isolating all evidence that satisfies the applicable Rule 17(c) requirements of admissibility and relevance. This evidence is then released, while the remaining materials are returned to the President. 418 U.S. at 714-15, 94 S.Ct. at 3110-11. As mentioned above, Rule 17(c) does not impose an admissibility requirement on grand jury subpoenas, and requires release of evidence unless there is no reasonable possibility that subpoenaed evidence will be relevant to grand jury proceedings. See R. Enterprises, 498 U.S. at 298, 301, 111 S.Ct. at 726-27, 728. Thus, once a grand jury has provided an adequate demonstration of need to obtain in camera review of materials covered by the presidential privilege, the court should review the subpoenaed material and release any evidence that might reasonably be relevant to the grand jury’s investigation. The question of what evidence might reasonably be relevant to the grand jury’s investigation should be answered by reference to the reasons the grand jury gave in explaining its need for the subpoenaed materials.

We believe that the appropriate course for us is to determine whether the OIC made out a sufficient showing of need to *760obtain in camera review of the documents. Although Nixon established that a President is allowed to immediately appeal an order requiring production of subpoenaed materials for in camera review, the general rule is that an order requiring production of evidence for in camera review “is not final and hence not appealable.” Nixon, 418 U.S. at 691, 94 S.Ct. at 3099; accord Church of Scientology v. United States, 506 U.S. 9, 18 n. 11, 113 S.Ct. 447, 452 n. 11, 121 L.Ed.2d 313 (1992); Kessler, 100 F.3d at 1016-17. Since the provision for immediate appeal by a President is an exception created because “[t]o require a President of the United States to place himself in the posture of disobeying an order of a court merely to trigger the procedural mechanism for review of the ruling would be unseemly,” Nixon, 418 U.S. at 691-92, 94 S.Ct. at 3099, we believe that the White House should not be penalized because it waited until the district court issued its final ruling on the OIC’s motion to compel. To rule otherwise would foster a proliferation of piecemeal appeals in cases implicating the presidential communications privilege. Moreover, both the OIC and the White House have directed their arguments to the question of whether the OIC made a sufficient demonstration of need for the withheld documents, and neither party — nor, it appears, the district court — differentiated between the standard that applies to the OIC’s showing of need to obtain in camera review and the standard the district court subsequently applies during in camera review to determine what material should be released.

The OIC provides two arguments as to why the grand jury needs the documents. One is the general claim that as the White House investigated the same subject matter as the grand jury, namely whether Espy accepted improper gifts or otherwise abused his position, the White House documents will clearly be relevant to the grand jury’s investigation. The other is that the grand jury is investigating whether Espy made false statements to the White House during its investigation, and evidence of statements made by Espy or his lawyers to the White House are key in determining whether false statements were in fact made. The OIC has submitted an ex parte affidavit and other materials in support of these arguments.

We find the OIC’s first justification of the grand jury’s need for the documents, that the withheld documents were generated by the White House Counsel’s office in preparing its report on the same allegations regarding Espy that the grand jury is investigating, insufficient, at this stage, to constitute an adequate showing of need under the Nixon/Sirica standard. It is true, as the OIC contends, that the withheld documents likely will contain evidence that is directly relevant to the grand jury’s investigation of Espy. But the OIC has not yet made a sufficient demonstration of its inability to obtain this information from alternative sources or an explanation of why it particularly needs to know what evidence is in the White House files. Here, unlike in che Nixon cases, the actions of White House officers do not appear to be under investigation.

We recognize the difficulty that the OIC faces in demonstrating that it has not been able to obtain the information contained in the White House Counsel’s documents when it does not know what this information is. This difficulty has been worsened by the extremely sketchy descriptions of the withheld documents that the White House provided in its privilege log. We also realize that in order to preserve the secrecy of the grand jury’s investigation, the OIC is understandably reluctant to detail the witnesses it has interviewed so far or the areas on which the investigation is focusing. But the OIC has not even attempted this task. For example, during their negotiations over the withheld documents, the White House Counsel’s office informed the OIC that the documents contained notes from interviews with two USDA attorneys. Yet the OIC has not indicated whether it interviewed attorneys at USDA and if so whether any one of them admitted to having conversations with the White House Counsel’s office. Again, while the OIC notes in its brief that the withheld documents could contain statements from witnesses who are no longer cooperating with the grand jury’s investigation, it provides no basis on which we could conclude that this is in fact the case. We also note that the *761subpoena for the documents generated in compiling the White House Counsel’s report was issued just three days after the report was released and five weeks after the OIC was appointed. In the face of this timing, it is hard to conclude that the OIC issued its subpoena to the White House as a last resort.

Nonetheless, it is possible that the OIC might be able to provide a sufficient justification for obtaining factual information in the White House files that it might not already possess. The White House has conceded that there is some factual information in the withheld documents that is not also contained in the documents that the White House released, and our own review of the documents has identified a sizeable number of such items of information, though many of them appear to be of minimal consequence. Moreover, the grand jury investigation into Espy’s actions has now lasted over two years, so that if and when the OIC provides some account of the information the grand jury has been unable to obtain, it will be fair to conclude that this information is not obtainable elsewhere. The OIC may also be able to demonstrate a need for information that it currently possesses, but which it has been unable to confirm or disprove.

Consequently, on remand the OIC should be given an opportunity to supplement its showing of need for the information contained in the withheld documents. If the district court determines that the OIC’s demonstration of need satisfies the Nixon/Sirica standard, the court should review the documents in camera and release any information that might reasonably be relevant in light of this demonstration of need. Two caveats should be noted. First, since the grand jury is investigating Espy’s actions, not those of the White House Counsel’s office, the purely deliberative portions of the documents should not be released. Second, only information that is not contained in the documents that the White House earlier released should be provided to the grand jury, since any new release of previously disclosed information would be purely cumulative. See Senate Committee, 498 F.2d at 732.

The OIC’s second, more narrow argument ' as to why the grand jury needs the withheld documents is much more powerful. The OIC argues that these documents may contain evidence of statements made by Espy or his lawyers to the White House, and that the grand jury needs this evidence in order to determine whether Espy made false statements to the White House.24 The White House released copies of all the letters that Espy’s counsel sent to the White House. But among the withheld documents are notes from meetings and conversations with Espy’s counsel, and other evidence indicating statements that Espy’s counsel made. It is likely that these documents contain evidence of statements by Espy’s counsel, and evidence of what Espy or his counsel actually said to the White House may be critical to the grand jury as it seeks to determine whether in fact Espy made any false statements.25 As we commented in Senate Committee, if one of the crimes that the grand jury is investigating relates to “the content of certain conversations, the grand jury’s need for the most precise evidence, the exact text of oral statements ... is undeniable.” 498 F.2d at 732. Obviously, this evidence is not available elsewhere; even if Espy’s counsel offered to provide the grand jury with every statement that was made to the White House, the grand jury would need to review the evidence in the White House files to confirm that no statements were omitted.

The OIC’s second argument of need for evidence in the subpoenaed documents is sufficient to obtain in camera review; the OIC has demonstrated that it is likely the subpoenaed documents contain important evidence that is not available elsewhere. On in camera review, the district court should isolate and release all evidence that might rea*762sonably be relevant to the question of whether Espy made false statements to the White House. Based on our review, we believe there are additional statements made by Espy’s counsel in the withheld documents— although again, several of these statements appear to be of little evidentiary value.

We therefore hold that the OIC has demonstrated -sufficient need in order to overcome the presidential communications privilege in regard to evidence of statements made by Espy or his counsel and contained in the withheld documents, and that the OIC should be given an opportunity to make out a sufficient showing of need in regard to other evidence more generally. On remand, the district court should identify and release specific items of evidence that might reasonably be relevant to the grand jury’s investigation into the potential false statements charge. If the court deems any additional showing of need presented by the OIC to be sufficient, it should also identify any new items of information that merit release. We are submitting a sealed appendix to assist the district court with its review.

Y. Conclusion

This case forces us to engage in the difficult business of delineating the scope and operation of the presidential communications privilege. In holding that the privilege extends to communications authored by or solicited and received by presidential advisers and that a specified demonstration of need must be made even in regard to a grand jury subpoena, we are ever mindful of the dangers involved in cloaking governmental operations in secrecy and in placing obstacles in the path of the grand jury in its investigatory mission. There is a powerful counterweight to these concerns, however, namely the public and constitutional interest in preserving the efficacy and quality of presidential decisionmaking. We believe that the principles we have outlined in this opinion achieve a delicate and appropriate balance between openness and informed presidential deliberation.

The decision of the district court is vacated and the ease is remanded for further proceedings consistent with this opinion.

So ordered.

. Another document was initially withheld on grounds of attorney work product privilege, but has since been released.

. It is clear from the briefs and oral argument in this case, as well as the district court’s opinion, that by "executive privilege" the White House is referring to the privilege that attaches to confidential presidential communications. However, as we discuss below, see infra Part I.B, "executive privilege” is generally used to refer to a wide variety of evidentiary and substantive privileges that courts accord the executive branch. Consequently, we refer to the privileges asserted by the White House more specifically as the presidential communications privilege, or presidential privilege, and the deliberative process privilege.

. For a listing of the different forms of executive privilege sanctioned by courts, see Gerald Wetlaufer, Justifying Secrecy: An Objection to the General Deliberative Privilege, 65 Ind. L.J. 845, 845 n.3 (1990); see generally Murl A. Larkin, Federal Testimonial Privileges §§ 5 to 7 (1996); 3 Weinstein’s Federal Evidence §§ 509-10 (Joseph M. McLaughlin, ed., 2d ed.1997).

. Some aspects of the privilege, for example the protection accorded the mental processes of agency officials, see United States v. Morgan, 313 U.S. 409, 421-22, 61 S.Ct. 999, 1004-05, 85 L.Ed. 1429 (1941), have roots in the constitutional separation of powers. See 3 Weinstein’s Federal Evidence! 509.21 [3] at 509-16.

.This characteristic of the deliberative process privilege is not an issue in FOIA cases because the courts have held that the particular purpose for which a FOIA plaintiff seeks information is not relevant in determining whether FOIA requires disclosure. See Sears, 421 U.S. at 149 n. 16, 95 S.Ct. at 1516 n. 16; Mink, 410 U.S. at 86, 93 S.Ct. at 835; see also Department of Justice v. Reporters Committee for Freedom of the Press, 489 U.S. 749, 771-72, 109 S.Ct. 1468, 1480-81, 103 L.Ed.2d 774 (1989) (determination of whether disclosure of information constitutes an unwarranted invasion of privacy under FOIA’s exemption 7(c) turns on nature of document and what document reveals about operation of government and not on identity or purpose of requestor).

. Marshall's conclusion was presaged by the argument before the Court, where then-Atlorney General and former Secretary of Stale Levi Lincoln had resisted testifying about the whereabouts of Marbury’s commission on the grounds that such information was an official secret he had learned in his position as Secretary of State. The Court had responded that "[t]here was nothing confidential to be disclosed. If there had been he was not obliged to answer it ... but that the fact whether such commissions had been in the office or not, could not be a confidential fact.” Marbury, 5 U.S. (1 Cranch) at 144.

. Jefferson then proceeded to transmit a copy of the letter identifying portions he believed should be deleted to Hay. But since Burr was again acquitted, he did not seek production of the letter until a third set of proceedings, these on the issue of whether he should be committed to custody for trial in Ohio on other charges. Ruling from the bench, Marshall denied Burr's request for the letter, stating “[a]fter such a certificate from the president of the United States as has been received, I cannot direct the production of those parts of the letter, without sufficient evidence of their being relevant to the present prosecution.” Freund, supra, at 29. Marshall instead held that the deleted portions could be inferred to support Burr. Id., 25 F.Cas. 30. Although Marshall never definitively ruled on the President's claims of privilege, his decision to issue the subpoena against President Jefferson has had lasting significance in establishing that "the President is subject to judicial process in appropriate circumstances.” Clinton v. Jones, — U.S.-,-& n. 38, 117 S.Ct. 1636, 1649 & n. 38, 137 L.Ed.2d 945 (1997).

. Two cases, Mink and Soucie v. David, 448 F.2d 1067 (D.C.Cir.1971), involved reports that were prepared pursuant to a presidential request and reviewed by the President, but in both cases the courts viewed the privilege claim at issue as being simply an assertion of the general deliberative process privilege, embodied in exemption five of the Freedom of Information Act, rather than a distinct privilege for presidential communications. See Mink, 410 U.S. at 91-93, 93 S.Ct. at 837-39; Soucie, 448 F.2d at 1071-72, 1075-78.

. See, e.g., Robert Kramer & Herman Marcuse, Executive Privilege — A Study of the Period 1953-1960: Part I, 29 Geo. Wash L. Rev. 623, 682-87, 692-93 (1961) (describing President Eisenhower's refusal to allow any executive branch officers to reveal to Congress internal deliberations on official matters). Although scholars dispute how often Presidents have actually refused to provide Congress with information on grounds of executive privilege, debate over the President’s ability to withhold confidential information from Congress has occurred since the early years of our nation, when President George Washington discussed with his cabinet in 1792 how to respond to a congressional inquiry into the military misfortunes that beset General St. Clair's expedition. See Archibald Cox, Executive Privilege, 122 U. Pa L. Rev. 1383, 1395-1405 (1974); see generally Raoul Berger, Executive Privilege- A Constitutional Myth (1974); Adam C. Breckenridge, The Executive Privilege- Presidential Control Over Information (1974); Daniel N. Hoffman, Governmental Secrecy and the Founding Fathers: A Study in Constitutional Controls (1981); Mark J. Rozell, Executive Privilege The Dilemma of Secrecy and Democratic Accountability (1994). Interestingly, it appears that Congress has at times accepted executive officers’ refusal to testify about conversations they had with the President, even as it was insisting on access to other executive branch documents and materials. See, e.g., Rozell, supra, at 44; Robert Kramer & Herman Marcuse, Executive Privilege — A Study of the Period 1953-1960: Part II, 29 Geo. Wash L. Rev. 827, 872-73 (1961). A very early instance of such a refusal by an executive officer came in the course of the House's investigation into why Alexander Hamilton had deposited into the Bank of the United States certain funds intended to pay off foreign debt. The House sought to know Hamilton’s authority for this act, to which Hamilton replied that he would not provide any instructions President Washington had given him, because "[t]hat question must, then, be a matter purely between the President and the agent, not examinable by the Legislature." Hoffman, supra, at 122. However, the House rejected the claim of privilege, and Hamilton eventually provided the material sought. Id. at 118-24.

.It appears that the courts have been drawn into executive-congressional disputes over access to information on only three recent occasions. These were: United States v. AT&T, 551 F.2d 384 (D.C.Cir.1976), appeal after remand, 567 F.2d 121 (D.C.Cir.1977); Senate Select Comm'ee on Presidential Campaign Activities v. Nixon (Senate Committee), 498 F.2d 725 (D.C.Cir.1974); United States v. House of Representatives, 556 F.Supp. 150 (D.D.C.1983).

. See supra note 2.

. The Court implied, however, that particularized claims of privilege for military and state secrets would be close to absolute, and expressly held only that the presidential communications privilege, which is based only on a generalized interest in confidentiality, can be overcome by an adequate showing of need. See Nixon, 418 U.S. at 710-11, 713, 94 S.Ct. at3108-09, 3110.

. The operation of the presidential communications privilege was addressed in a few other criminal cases. In United States v. Haldeman, 559 F.2d 31 (D.C.Cir.1976), and United States v. Ehrlichman, 546 F.2d 910 (D.C.Cir.1976), John Ehrlichman, an assistant to President Nixon, challenged his convictions stemming from the Watergate investigation on the grounds that the district court had improperly denied requests for information in White House files. However, in neither case is there any significant discussion of the privilege, because Ehrlichman had failed "to argue with specificity the materiality and reasonableness of his discovery request” and thus would have not been entitled to access to this evidence under Rule 17(c) of the Federal Rules of Criminal Procedure even if it were not presumptively privileged. Ehrlichman, 546 F.2d at 931-32; see also Haldeman, 559 F.2d at 76-77. In United States v. Poindexter, 727 F.Supp. 1501 (D.D.C.1989) and United States v. North, 713 F.Supp. 1448 (D.D.C.1989), two prosecutions arising out of the Iran-Contra investigation, former National Security Advisor John Poindexter and Lieutenant Colonel Oliver North subpoenaed President Reagan to testify about conversations; Poindexter also subpoenaed President Reagan’s diaries. Although in both cases the courts noted that the subpoenas implicated the presidential communications privilege, they only addressed the question of whether the subpoenas satisfied Rule 17(c). Poindexter's initial conviction was reversed by this court on other grounds, United States v. Poindexter, 951 F.2d 369 (D.C.Cir.1991), and President Bush subsequently pardoned Poindexter, thus forestalling further appellate review of the district court’s order in his case. This court held that any error in the district court’s refusal to subpoena President Reagan to testify at North's trial was harmless because there was no indication he would have provided evidence that was material or favorable to North. As a result, the issue of presidential privilege was only addressed by Judge Silberman in dissent. United States v. North, 910 F.2d 843, 888-92 & n. 25 *744(D.C.Cir.), vacated, in part, 920 F.2d 940 (D.C.Cir. 1990); id. at 932, 950-54 (Silberman, J., concurring in part and dissenting in part).

.The presidential communications privilege also surfaced in the district court’s opinion in Wayte v. United. States, which later was appealed to the Supreme Court Wayte alleged that the government's enforcement policy on military draft, registration requirements was unconstitutional, and sought discovery of presidential documents and testimony regarding the policy from the White House Counsel. The Court, however, decided the case on other grounds, and the only discussion of the presidential privilege is found in Justice Marshall's dissent. 470 U.S. 598, 614, 615-23, 105 S.Ct. 1524, 1534-39, 84 L.Ed.2d 547 (1985).

. This court subsequently upheld the regulations promulgated by GSA to govern access to the Nixon materials. See Nixon v. Freeman, 670 F.2d 346 (D.C.Cir.1982); see also Nixon v. United States, 978 F.2d 1269 (D.C.Cir.1992) (holding that PRMPA acted as a taking of President Nixon’s materials so as to require just compensation).

. In Nixon, Sirica and GSA, President Nixon personally asserted the presidential communications privilege, and thus these cases do not establish whether the privilege must be invoked by the *745President as opposed to a member of his staff. In discussing the military and state secrets privilege in Reynolds the Supreme Court stated that "[t]here must be a formal claim of privilege, lodged by the head of the department which has control over the matter,” 345 U.S. at 7-8, 73 S.Ct. at 532, which might suggest that the President must assert the presidential communications privilege personally. See Center on Corp. Responsibility, Inc. v. Shultz, 368 F.Supp. 863, 872-73 (D.D.C.1973) (White House Counsel’s affidavit indicating that he is authorized to say that the White House was invoking executive privilege over tapes and documents in White House files is insufficient to invoke the privilege); see also Burr, 25 F. Cas. at 192 (ruling that President Jefferson had to personally identify the passages he deemed confidential and could not leave this determination to the U.S. Attorney). We need not decide whether the privilege must be invoked by the President personally, since the record indicates that President Clinton has done so here; in his affidavit former White House Counsel Abner J. Mikva stated “the President ... has specifically directed me to invoke formally the applicable privileges over those documents.” Moreover, although the OIC challenged the adequacy of the White House's invocation of privilege before the district court, the OIC did not pursue this issue on appeal.

. In some cases, the White House’s ex parte contacts with outside agencies may be subject to disclosure by statute, see, e.g., Portland Audubon Soc'y v. Endangered Species Committee, 984 F.2d 1534, 1543-48 (9th Cir.1993), but this court has refused to require disclosure of conversations between an agency and the President or White House staff, at least where the proceeding was not adjudicatory and the statute did not specifically require disclosure, because of the President's need to oversee executive agencies. See Sierra Club v. Costle, 657 F.2d 298, 404-08 (D.C.Cir.1981).

. The elements of this showing of need are discussed in greater detail infra in Part III.C.

. Commentators have noted that the Nixon opinion did not address this question of who qualifies for the privilege. See Raoul Berger, The Incarnation of Executive Privilege, 22 UCLA L.Rev. 4, 22-26 (1974) (hereinafter Berger, Incarnation).

. In Wolfe v. HHS, 815 F.2d 1527 (D.C.Cir.1987), a panel of this court held that the privilege did not protect communications of the Office of Management and Budget that did not involve the President, stating that such an "extension of the presidential privilege ... is unprecedented and unwarranted .... [and] would create an unnecessary sequestering of massive quantities of information from the public eye.” Id. at 1533. However, the opinion was later vacated by the court en banc, and the government abandoned its presidential privilege claims before the full court. See Wolfe, 839 F.2d at 773 n. 5.

. For example, Professor Berger commented on the Nixon decision: "The real problem is not posed by confidentiality between the President and his immediate advisers, members of his cabinet and the like; it arises from the fact that the claim for executive privilege has sprawled far beyond presidential precincts.” Berger, Incarnation, supra, at 23.

. The Constitution does not explicitly grant the President the power to remove executive branch officials, but it is well established that this power, at least in regard to some officials, can be inferred from the President's other enumerated powers and responsibilities. See Morrison, 487 U.S. at 689-90, 108 S.Ct. at 2618-19; Myers v. United States, 272 U.S. 52, 117, 163-64, 47 S.Ct. 21, 25, 41, 71 L.Ed. 160 (1926). While the President’s removal power over some executive branch officials is limited, the President has unqualified power to appoint and remove cabinet officers. See Myers, 272 U.S. at 134, 47 S.Ct. at 31 ("[The President's] cabinet officers must do his will.... The moment he loses confidence in *753the intelligence, ability, judgment, or loyalty of any one of them, he must have the power to remove him without delay.”).

. The White House has also claimed attorney-client privilege in regard to Document 19. We do not need to examine this claim because it is clear, based on our review of this document, that *759it should not be released. The document comes under the presidential communications privilege as it was authored by the President’s Chief of Staff and was sent to the individual acting as White House Counsel, and contains no information or evidence that could be relevant to the grand jury's inquiry.

. The White House has not suggested that an investigation into whether Espy made false statements is outside the OIC’s mandate, and thus we consider any such argument to be waived.

. We express no opinion as to whether Espy’s counsel's statements are attributable to Espy. See United States v. Curran, 20 F.3d 560, 567-68 (3d Cir.1994); United States v. Fairchild, 990 F.2d 1139, 1141 (9th Cir.1993); United States v. Blazewicz, 459 F.2d 442, 443 (6th Cir.1972).