HUNTON & WILLIAMS, Plaintiff-Appellant, v. UNITED STATES DEPARTMENT OF JUSTICE, Defendant-Appellee, 590 F.3d 272

Affirmed in part, vacated and remanded in part by published opinion. Judge WILKINSON wrote the majority opinion, in which Judge KEELEY joined. Judge MICHAEL wrote a dissenting opinion. OPINION WILKINSON, Circuit Judge: This appeal centers on communications between the U.S. Department of Justice (“DOJ”) and a telecommunications company, in which the company allegedly lobbied DOJ to take its side in litigation with a client of law firm Hunton and Williams, LLC (“Hunton”).

The district court upheld DOJ’s decision to deny Hunton’s request under the Freedom of Information Act, 5 U.S.C. § 552 (2006), (“FOIA”) for records of those communications. Hunton contends that it is entitled to the records, regardless of whether they satisfied the requirements of the so-called common interest doctrine, which enables parties with a shared legal interest to pursue a joint legal strategy.

DOJ argues not only that common interest communications are exempt from FOIA, but that we should defer to the agency’s invocation of the common interest doctrine without demanding any serious inquiry into the validity of its common interest claims. Both sides have a point, though only a partial one. DOJ argues persuasively that FOIA does not strip the government of its civil discovery privileges or its valuable right to partner with other parties in litigation or in anticipation of the same.

At the same time, however, Hunton correctly contends that common interest assertions by government agencies must be carefully scrutinized. For the doctrine to apply, an agency must show that it had agreed to help another party prevail on its legal claims at the time of the communications at issue because doing so was in the public interest. It is not enough that the agency was simply considering whether to become involved. I. A.

This FOIA action grows out of an earlier patent suit brought by New Technology Products, Inc., (“NTP”), a client of Hun-ton’s, against Research in Motion, Ltd., (“RIM”), manufacturer of BlackBerry communications devices. In August 2003, following an earlier jury finding that RFM had infringed various patents held by NTP, the district court in the BlackBerry litigation entered an order enjoining RIM’s use of the patented technology. Enforcement of the injunction was stayed, however, pending RIM’s appeal, which ultimately resulted in partial affirmation and a remand to the district court. NTP, Inc. v.

Research In Motion, Ltd., 418 F.3d 1282, 1326 (Fed.Cir.2005). District court proceedings did not resume until October 2005. Shortly after the jury reached its verdict, proceedings to reexamine NTP’s patents were initiated before the United States Patent and Trademark Office (PTO). While RFM’s appeal in the BlackBerry litigation was pending, RIM began contacting officials from various executive branch departments to express its concern about the injunction.

On March 10, 2005, several DOJ attorneys, including John Fargo, Director of the Intellectual Property Staff in the Commercial Litigation Branch of DOJ’s Civil Division, met with RIM attorney Herbert Fenster. At their meeting, Fenster expressed his opinion that RIM and the federal government had a mutual interest in opposing the BlackBerry injunction because the injunction would interfere with the federal government’s BlackBerry use.

The United States government is the largest single user of BlackBerry devices, and as a matter of law, it cannot be subject to injunctive relief against the use of patented technology. 28 U.S.C. § 1498(a) (2006); Trojan, Inc. v. Shat-R-Shield, Inc., 885 F.2d 854, 856-57 (Fed.Cir.1989). Fenster offered to furnish DOJ with information and drafts of affidavits RTM was then in the process of obtaining as part of its efforts to overturn the injunction. DOJ contends that, immediately after the meeting, Fenster and Fargo orally agreed to exchange documents on a confidential “common interest” basis.

Fenster continued to discuss the BlackBerry litigation with Fargo in the months that followed, supplying information, documents, and declarations for DOJ’s use. The first time the phrase “common interest” appeared in any written communication between RTM and DOJ was October 6, 2005, when Fargo added the disclaimer “protected by joint and common interest privilege” to an email reply he sent to Fenster. On November 8, 2005, two weeks after proceedings in the BlackBerry patent litigation resumed in district court, DOJ filed a Statement of Interest and requested that the matter be stayed for 90 days.

It argued that the injunction contemplated by the district court could operate as a de facto injunction against the government’s BlackBerry use and that the government needed more time to consider the issue. According to Fargo, the decision to file the Statement was not made until shortly before the actual filing. On November 10, 2005, two days after the Statement of Interest was filed, Fargo and Fenster signed a written common interest agreement on behalf of DOJ and RTM, which stated that their common interest relationship had come into being on February 4, 2005.

On February 1, 2006, DOJ filed a motion to intervene in the district court proceedings, which was granted. The litigation settled the next month. At some point prior to DOJ’s intervention, NTP became concerned about communications between RIM and the PTO in connection with the patent reexamination proceedings, and in January 2006, counsel for NTP filed a FOIA request with the PTO and its parent agency, the Department of Commerce, to obtain any such communications. See Rein v. U.S. Patent & Trademark Office, 553 F.3d 353, 362 (4th Cir.2009).

In the wake of that request, Hunton learned of the common interest agreement between DOJ and RIM. Shortly after the BlackBerry litigation settled, Hunton filed a second FOIA request, this time with DOJ, seeking records of communications between RIM and DOJ, as well as related communications between DOJ and other agencies such as the PTO. DOJ withheld roughly half of the documents Hunton requested, and Hunton challenged the withholding of those documents that DOJ claimed were protected from disclosure by Exemption 5 of FOIA. 5 U.S.C. § 552(b)(5).

Following an in camera inspection of a substantial portion of the documents DOJ withheld, the district court in Hunton’s DOJ FOIA suit granted DOJ’s motion for

summary judgment for all but three of the documents at issue. B. The question of whether a district court properly granted the government

summary judgment in a FOIA action is one of law, which we review de novo. Ethyl Corp. v. U.S. E.P.A., 25 F.3d 1241, 1246 (4th Cir.1994). Whether a document fits within one of FOIA’s prescribed exemptions is also a matter of law, unless the legal conclusion is based upon factual findings, which we review for clear error. Id. FOIA provides that federal agencies shall “upon request for records which reasonably describes such records ... make the records promptly available to any person.” 5 U.S.C. § 552(a)(3)(A). The Act specifies nine exemptions from its regime of disclosure, however. 5 U.S.C. § 552(b).

In general, these exemptions are to be narrowly construed. Bowers v. U.S. Dept. of Justice, 930 F.2d 350, 354 (4th Cir.1991). The burden of demonstrating that a requested document falls under an exemption rests on the government. 5 U.S.C. § 552(a)(4)(B); City of Virginia Beach, Va. v. U.S. Dep’t of Commerce, 995 F.2d 1247, 1252 (4th Cir.1993).

Exemption 5 provides that FOIA disclosure rules do not apply to “inter-agency or intra-agency memorandums or letters which would not be available by law to a party other than an agency in litigation with the agency.” 5 U.S.C. § 552(b)(5). “To qualify, a document must thus satisfy two conditions: its source must be a Government agency, and it must fall within the ambit of a privilege against discovery under judicial standards that would govern litigation against the agency that holds it,” such as the attorney-client, deliberative process, or attorney work product privileges. Dep’t of the Interior v.

Klamath Water Users Protective Ass’n, 532 U.S. 1, 8, 121 S.Ct. 1060, 149 L.Ed.2d 87 (2001). The district court’s finding that the documents now at issue are of a character qualifying them for the privileges asserted by DOJ is not challenged in this appeal. The only question we address is whether the district court erred in finding that the materials sought by Hunton qualify as “inter-agency or intra-agency memorandums or letters.” II.

For more than 40 years, the Freedom of Information Act has provided a way for the American people to keep a watchful eye on their government. “The basic purpose of FOIA is to ensure an informed citizenry, vital to the functioning of a democratic society, needed to check against corruption and to hold the governors accountable to the governed.” N.L.R.B. v. Robbins Tire & Rubber Co., 437 U.S. 214, 242, 98 S.Ct. 2311, 57 L.Ed.2d 159 (1978). The Act discourages agencies from keeping in the dark actions that might not withstand the light of day.

For that reason, “disclosure, not secrecy, is the dominant purpose of the Act.” Klamath, 532 U.S. at 8,121 S.Ct. 1060. FOIA was the first in a series of laws adopted between 1966 and 1976 that were designed to bring the activities of federal agencies out of the shadows of official secrecy. FOIA and its sister statutes — the Federal Advisory Committee Act, Pub.L.

No. 92-463, 86 Stat. 770 (1972) (codified as amended at 5 U.S.C. app. at 455 (2006)), and the Government in the Sunshine Act, 5 U.S.C. § 552b (2006), — reflected a concern about the effects of self-interested private parties on the administration of federal law. See H.R.Rep. No. 89-1497, at 26 (1966), U.S.Code Cong. & Admin.News 1966, pp. 2418, 2426 (federal agency’s refusal to disclose bids on a multimillion dollar project concealed the fact that the lowest bidder had not been chosen and demonstrated the need for FOIA).

Cozy public-private relationships, selective disclosures to favored parties, and ex parte lobbying of agencies by special interests were among the practices that FOIA sought to subject to public scrutiny. See N. Dakota ex rel. Olson v. Andrus, 581 F.2d 177, 182 (8th Cir.1978) (“Preferential treatment of persons or interest groups fosters precisely the distrust of government that the FOIA was intended to obviate.”).

The danger that agencies will be captured by the beneficiaries of public largesse or the subjects of federal regulation has not diminished with time, and FOIA’s disclosure mandate continues to play a vital

part in the supervision of federal agencies by their ultimate principals, the American people. Although FOIA establishes a broad policy of transparency, its commitment to that policy is not unlimited. The Act acknowledges that “public disclosure is not always in the public interest.” Baldrige v. Shapiro, 455 U.S. 345, 352, 102 S.Ct. 1103, 71 L.Ed.2d 199 (1982).

FOIA’s nine specified exemptions reflect a wide array of concerns, in this case Exemption 5’s recognition that any agency’s ability to discharge its duties effectively depends on the agency’s not having to operate “in a fishbowl.” Virginia Beach, 995 F.2d at 1252 (citation omitted). In situations where the government consults with an attorney, Exemption 5 expresses Congress’s view that the public interest is not served by stripping government agencies of the privileges otherwise available to them in litigation.

FOIA was meant to foster political accountability, not to force the United States into a uniquely disadvantaged litigation posture. Indeed, the Supreme Court has noted that to allow a party to “obtain through the FOIA material that is normally privileged would create an anomaly in that the FOIA could be used to supplement civil discovery. We have consistently rejected such a construction of the FOIA.” United States v. Weber Aircraft Corp., 465 U.S. 792, 801, 104 S.Ct. 1488, 79 L.Ed.2d 814 (1984) (citations omitted). III.

Given the tension in FOIA between the public’s interest in transparency, on the one hand, and in effective public administration, on the other, it is ironic both sides in this dispute have adopted rather absolute views of the scope of Exemption 5. Hunton contends that Exemption 5 provides no protection to communications between RTM and DOJ, regardless of their congruent interests in the course of actual litigation. DOJ argues that its common interest agreement should shield all of its communications with RJM, without any serious consideration of the scope and duration of the agreement.

Neither position fully reflects the balance struck in FOIA and the cases that have interpreted it. A. Hunton argues that communications between RIM and DOJ are unprotected by Exemption 5 as a matter of law because RIM is a private party, not a part of DOJ, and those communications therefore are neither “inter-” nor “intra-agency.” The district court rejected this conclusion, reasoning that communications between a government agency and a party possessing common and unitary litigation interests should be understood as “intraagency” for purposes of Exemption 5.

The common interest doctrine permits parties whose legal interests coincide to share privileged materials with one another in order to more effectively prosecute or defend their claims. In re Grand Jury Subpoenas, 89-3 and 89-1, John Doe 89-129, 902 F.2d 244, 248-49 (4th Cir.1990). Under Hunton’s reading, however, the decision of a party, here the government, to partner with others in the conduct of litigation would somehow subject that party to the loss of its most basic civil discovery privileges — namely, the attorney-client and attorney work product privileges.

This is a sweeping view, and its impact on the government’s ability to conduct complex and multi-faceted litigation would be staggering. We have made clear that the government was entitled not to favored treatment under Exemption 5, but simply to a level playing field. “The government has the same right to undisclosed legal advice in anticipation of litigation as any private party. And there is nothing in FOIA that prevents the government from drawing confidential counsel from the private sector.

Allowing disclosure here would impair an agency’s ability to prepare effectively for litigation with private parties and thereby thwart its ability to discharge its functions in the public interest.” Hanson v. United States Agency for Int’l Devt, 372 F.3d 286, 294 (4th Cir.2004). Although Hunton argues that Hanson is inapplicable because the attorney in that case was employed by a private party working under a government contract, Hanson’s, result was based upon the presence of shared and substantial interests, not upon some notion that a government contractor’s attorney is a government attorney. Id. at 292.

Nothing in the “inter-agency or intraagency” language in Exemption 5 demands that the government, alone among all litigants, be stripped of civil discovery privileges when it has done nothing more than communicate with other litigating parties with whom it shares a singular and unitary litigation interest.

It is that convergence of interests that entitles the government to communicate within the terms of the Exemption and to do so in a manner that does not strip it of those deliberative privileges that other litigants enjoy and that are widely recognized as necessary “to encourage full and frank communication between attorneys and their clients and thereby promote broader public interests in the observance of law and administration of justice.” Id. at 291 (citation omitted). What is sometimes termed the common interest doctrine is in this sense simply a matter of evenhandedness.

Congress’s whole purpose in drafting Exemption 5 was thus not to have FOIA, which is of course directed only at the government, place an agency at a serious disadvantage vis-a-vis private parties who are free of any FOIA constraints. To allow adverse private litigants to use FOIA to claim this sort of tactical advantage against the government would run counter to the Exemption’s goal. For its clear thrust is simply to ensure that FOIA does not deprive the government of the work-product and attorney-client protections otherwise available to it in litigation. Nat’l Labor Relations Bd. v.

Sears, Roebuck & Co., 421 U.S. 132, 149, 95 S.Ct. 1504, 44 L.Ed.2d 29 (1975) (Exemption 5 should be read “to exempt those documents, and only those documents, normally privileged in the civil discovery context.”); see also H.R.Rep. No. 89-1497, at 31 (1966), U.S.Code Cong. & Admin.News 1966, pp. 2418, 2428 (Exemption 5 was intended to allow an agency to withhold materials that would not be “routinely disclosed to a private party through the discovery process in litigation with the agency....”). What the Supreme Court has termed the “Delphic” wording of Exemption 5, Department of Justice v.

Julian, 486 U.S. 1, 11, 108 S.Ct. 1606, 100 L.Ed.2d 1 (1988), is thus clear in this respect — one side in litigation cannot play by one set of rules and another side play by a more privileged set of rules. If this were the case, attorneys on one side of litigation could freely communicate, safe in the knowledge that their work product and deliberative processes would be privileged, while the other side would be obliged to turn over communications of the very same nature to its adversary.

Further, in the absence of coordination, the government — or any party whose interests align with the government’s — might find its position strafed inadvertently by “friendly fire.” These are just a few of the problems created when one litigant is denied privileges accorded to all others. FOIA does not require the government to litigate on such distinctly disadvantageous terms. This view is entirely consistent with the Supreme Court’s teaching in Department of the Interior v. Klamath Water Users Protective Ass’n, 532 U.S. 1, 121 S.Ct. 1060, 149 L.Ed.2d 87 (2001).

There the Court held that Exemption 5 did not apply to communications between Indian tribes and the Department of the Interior in connection with two proceedings to determine the allocation of water rights, one in state court and one within the Interior Department itself. The Court held that those communications could not be considered “inter-agency or intra-agency,” since the tribes were communicating with the federal government in order to secure water rights for themselves. Id. at 14, 121 S.Ct. 1060.

That holding did not impact the very different situation here — namely, one where the two parties share a unitary interest in achieving a litigative outcome and result. Indeed, the Court’s opinion in Klamath never discussed the common interest doctrine. Further, the common interest doctrine could not have applied in Klamath because the government had not decided what its interests even were or embarked upon a definite course of action. It was communicating with the tribes to make its own assessment of their interests and of how those interests compared with the interests of the public.

Id. at 5, 14, 121 S.Ct. 1060. Finally, in Klamath, the government sought to advance the tribes’ private interests because it had a fiduciary obligation to do so, not because it thought that doing so was in the public interest. Id. at. 14. Klamath addressed the case of self-interested parties attempting to persuade the government to adopt a particular policy, but those concerns are no longer in play once the government is actually persuaded that the policy is in the public interest, as was the case in Hanson.

It would eviscerate the meaning of Exemption 5 if we were to read it to exclude communications between federal agencies and their litigation partners where those communications advance an interest that is both common and, in the government’s considered view, critical to the public’s interest. In Klamath, the Supreme Court recognized the general view that it is “textually possible and ... in accord with the purpose of the provision” to regard as “intra-agency” materials prepared by outside parties with whom a federal agency consults. Klamath, 532 U.S. at 9-10, 121 S.Ct. 1060 (citation omitted).

Numerous Courts of Appeals have done just that. See Stewart v. U.S. Dept. of Interior, 554 F.3d 1236, 1245 (10th Cir.2009); Tigue v. U.S. Dep’t of Justice, 312 F.3d 70, 78 & n. 2 (2d Cir.2002); Hoover v. U.S. Dep’t of Interior, 611 F.2d 1132, 1138 (5th Cir.1980); Brockway v. Department of the Air Force, 518 F.2d 1184, 1191 (8th Cir.1975); Soucie v. David, 448 F.2d 1067, 1078 n. 44 (D.C.Cir.1971); see also Gov’t Land Bank v. G.S.A., 671 F.2d 663, 665 (1st Cir.1982).

Both before and after Klamath, the District of Columbia Circuit has applied this “consultant corollary” doctrine to unpaid as well as paid consultants. Klamath expressly declined to overrule two of the cases in which it did so. Klamath, 532 U.S. at 12 n. 4, 121 S.Ct. 1060 (citing Public Citizen, Inc. v. Department of Justice, 111 F.3d 168 (D.C.Cir.1997); Ryan v. Department of Justice, 617 F.2d 781 (D.C.Cir.1980)).

After all, there is no reason why the advice of a consultant with no conflict of interest should be protected from disclosure if the consultant is paid for the advice but unprotected if it is offered for free. As the D.C. Circuit recently explained in a case which treated as “intraagency” the opinions of outside legal experts who provided free advice concerning the establishment of military tribunals, the consultants’ contributions were “no less valuable or confidential for the lack of compensation or formal contract.” Natl Inst. of Military Justice v.

United States Dep’t of Defense, 512 F.3d 677, 684 (D.C.Cir.2008), cert. denied, - U.S. -, 129 S.Ct. 775, 172 L.Ed.2d 754 (2008). These cases make good sense. Such communications can be considered “intraagency” for the simple reason that the outside attorney or consultant is collaborating with an agency in the agency’s pursuit of the public interest. This same rationale readily reaches cases in which the government partners with another party in litigation.

Because the common interest doctrine requires the agency to determine that the public interest and the litigation partner’s interest have converged, communications between the agency and its partner can be understood as “intra-agency” for purposes of Exemption 5. By cooperating with the agency in pursuit of the agency’s own litigation aims, the litigation partner in a limited sense becomes a part of the enterprise that the agency is carrying out. See Ryan, 617 F.2d at 789.

The underlying idea is essentially the same principle that allows us to refer to certain litigants as “private attorneys general.” The point is not that the outsider attorney general stands to gain from assisting the government — that is equally true of paid consultants, agency employees, and for that matter, of agencies themselves. The point is that there is no conflict of interest when it comes to advancing the public’s interest because the outsider stands to gain personally only if the public’s interest is vindicated.

Hunton concedes that the Exemption 5 was not meant to destroy the common interest doctrine, but argues that the doctrine applies only to the issue of whether a record is privileged. Since it is difficult to imagine a situation in which the common interest doctrine would do any work once a document qualified as “inter-” or “intraagency,” however, it seems clear that the doctrine is relevant to the question of whether the document qualifies as “inter-” or “intra-agency,” not to the question of whether it is privileged. The common interest doctrine, it bears adding, is not a privilege in its own right.

In re Grand Jury Subpoenas, 902 F.2d at 249. Merely satisfying the requirements of the common interest doctrine without also satisfying the requirements of a discovery privilege does not protect documents from disclosure any more than does merely satisfying the “inter-agency or intra-agency” requirement. Courts interpreting the various terms used in Exemption 5 have recognized the need to take a functional approach to ascertaining their meanings.

Thus, although the prefixes “inter-” and “intra-” may suggest that a document be transmitted from one person to another, “common sense” dictates that Exemption 5 reaches handwritten notes kept in a file. Conoco Inc. v. United States Dept. of Justice, 687 F.2d 724, 728 (3rd Cir.1982). Likewise, although Exemption 5 addresses itself only to “letters and memorandums,” the privileges Congress sought to preserve would be gutted if FOIA could be used to reach items like draft pleadings, litigation exhibits, and data on government computers. See Chilivis v.

S.E.C., 673 F.2d 1205, 1212 n. 15 (11th Cir.1982) (“In adopting exemption 5, Congress clearly intended to exempt any document connected with the agency’s deliberative process, not just memoranda and letters.”). To adopt Hun-ton’s view in its entirety would send us in an opposite, doctrinaire direction. We would be required to hold that never, under any circumstances, could the United States claim the most basic civil discovery privileges when partnering with another party in litigation. To say that view invites havoc is to understate the matter.

We think the more sensible course is to ask whether a common and singular public-private interest is present, not to invalidate the proper assertion of civil discovery privileges wholesale. B. 1. The foregoing discussion demonstrates the applicability of Exemption 5 to DOJ’s filing of its Statement of Interest with the district court and its signing of the formal common interest agreement with RJM, both of which occurred in early November 2005.

The written agreement explains the parties’ shared interest in limiting the scope of any injunction in the BlackBerry litigation and clearly manifests an agreement to work together toward that end.

The agreement stated that the United States government is the single largest user of BlackBerry technology, and that “[t]he continued ability to employ those BlackBerry systems and devices is considered of sufficient importance to the U.S. as to make necessary the pursuit of this subject through a common interest agreement.” The agreement defined the scope of the parties’ common interests clearly, providing, for example, a means to terminate the agreement if the parties came to believe that their interests had diverged.

DOJ’s decision to intervene in the litigation in February 2006 provides further evidence that it had committed to taking RIM’s side. The Statement of Interest confirms the veracity of the government’s interest in the BlackBerry litigation and explains that interest in greater detail.

In its Statement, DOJ asserted that the injunction originally entered by the district court “would amount to a de facto injunction against the government’s use of Blackberries” because it would “literally prohibit REVI from providing the services that would be essential for the federal government, as well as state and local governments, to continue their use of the BlackBerry devices.” Even if the government’s use were not directly enjoined, there was a serious risk that the government’s internal communications, as well as its communications with those whom it needed to bé able to contact via BlackBerry, would be disrupted by an injunction against RIM.

Nor could the problem be fixed merely by pointing out the need to protect governmental BlackBerry use. The Statement explained that “there does not appear to be a simple manner” by which to segregate government BlackBerry users from other users. Over several pages, DOJ set forth the technical complexities that fashioning appropriate injunctive relief against RIM would involve. For one thing, it simply would not be possible to distinguish separate communications on an email-by-email basis.

The best that could be hoped for was to create a list of approved BlackBerry users who would be given an across-the-board exemption from any injunction, regardless of whether their communications were being made in their official capacities.

Even then, the government would be required to carry out “a time-consuming inventory of every agency within the federal government” and to create and maintain a database of BlackBerry information that would only become “more complicated and prone to error” when due allowance was made for outside users— such as state and local governments, government contractors, and nongovernmental organizations like the American Red Cross — whose ability to communicate with the federal government was essential to the federal government’s BlackBerry use.

Ensuring that “the federal government’s communications, which include a raft of time-sensitive official communications, are not disrupted by any injunction” would be difficult at best. The task would be Herculean, if not impossible.

In short, the Statement makes it abundantly clear that, in partnering with RIM, DOJ was asserting a genuine public interest in seeking to ensure that “the federal government’s right to continue its use of BlackBerry devices is not rendered a nullity.” Moreover, the record in this case suggests that concern about a BlackBerry injunction against RIM and its impact in turn on government operations was widespread.

As Fargo noted in his sworn statement, the government’s “ability to send messages to, and receive messages from[,] public and private recipients in emergency or crisis situations” was a source of serious concern. Every member of Congress had been issued a BlackBerry, for instance, because BlackBerry technology permitted emergency communications when other communications systems were inoperative. The Department of Homeland Security had numerous contacts with Fargo, including originally alerting Fargo of Fenster’s concerns and making the request that prompted DOJ to file the Statement of Interest.

Federal officials also received information specifically pertaining to BlackBerry use by the Defense Department, and in response to an interrogatory from Hunton, the government disclosed that personnel from the Departments of Defense and Homeland Security may have been informed of the existence of a “recognized common interest” between RIM and DOJ.

Further, Fargo testified that a number of agencies had contacted DOJ in order to find out what the effect of an injunction would be on their BlackBerry use, and he discussed his communications with representatives of the Departments of Defense, Energy, Health and Human Services, Homeland Security, and State, the Executive Office of the President and the Department of Commerce’s Bureau of Industry and Security.

The government also perceived that it might have a “broad general interest in avoiding any disruption to first responders, state and local governments, health care, and the economy generally that might accompany an injunction.” Thus, as of November 2005, there can be no doubt that the government had determined that it was in the public interest for RIM to succeed in its efforts to constrain the scope of injunctive relief entered in the BlackBerry litigation.

The federal government is BlackBerry’s biggest customer, and it had an obvious interest in avoiding the serious disruption to its operations that an overly broad BlackBerry injunction would entail. And the agreement between RIM and DOJ in November 2005 makes it clear that RIM and DOJ had committed to working together to achieve that goal. As a result, Exemption 5 properly applies to communications made pursuant to that agreement.

It does not matter that RIM was motivated by the commercial benefit that would accrue to it if it succeeded in opposing the BlackBerry injunction while the government was motivated by concern for the public interest. What matters is that there was a unity of interest in preserving a non-disruptive pattern of governmental BlackBerry use, and RIM and DOJ could rely on one another’s advice, secure in the knowledge that privileged communications would remain just that. 2.

Hunton next argues that even if a shared interest of RFM and the government in narrowing the scope of a BlackBerry injunction existed once the written common interest agreement was executed, common interest principles should not apply to their efforts to limit a BlackBerry injunction because RIM and DOJ did not face an adverse party on that issue. Hun-ton contends that NTP had no interest in an injunction that was broader than what the law allowed and consistently attempted to work with DOJ to protect the government’s BlackBerry use. Citing language used by the Third Circuit in Haines v.

Liggett Group, Inc., 975 F.2d 81, 90 (3d Cir.1992), Hunton asserts that an adverse party is a prerequisite for invoking the common interest doctrine. We need not address the issue here, however. In this case, there was ample evidence to support the district court’s conclusion that NTP was, in fact, adverse to DOJ and RIM when it came to the scope of injunctive relief in the BlackBerry litigation.

NTP in fact opposed DOJ’s request for an evidentiary hearing regarding the feasibility of a BlackBerry injunction at the time DOJ filed its Statement of Interest, for instance, and Fargo testified in his deposition that he “did regard [NTP] as an adverse party.” NTP sent a series of six requests for information to DOJ, which Fargo viewed as informal requests for discovery, and NTP disagreed with DOJ about the length of time that would be needed to phase in an injunction.

Finally, DOJ successfully argued that it should be allowed to intervene in the BlackBerry litigation because although NTP had conceded that an injunction should not infringe the U.S. government’s BlackBerry rights, NTP had “only offered the court unsupported statements asserting the ease with which” a workable injunction could be crafted, without specifying how this could be done. The foregoing serves to confirm what should be obvious: DOJ became involved in the BlackBerry litigation because it was concerned that its interests might be infringed by an injunctive award in NTP’s favor.

NTP stood to benefit from a broad injunction, while DOJ would likely be harmed by one. There is every reason to suppose that their interests were in conflict, and the district court’s decision on this point cannot be said to be clear error. IV. A. Although the applicability of common interest principles to communications between RIM and DOJ from November 2005 onward has been firmly established, DOJ’s decision to withhold materials generated between March and November 2005 presents a much closer question.

DOJ urges a broad and deferential approach to determining whether those communications are shielded by Exemption 5. The Supreme Court’s decision in Department of the Interior v. Klamath Water Users Protective Ass’n, 532 U.S. 1, 121 S.Ct. 1060, 149 L.Ed.2d 87 (2001), makes clear, however, that any attempt to invoke the common interest doctrine in order to avoid disclosure under FOIA must be more carefully scrutinized. In Klamath, the Court took pains to emphasize that the words “inter-agency or intra-agency” cannot be “drain[ed] ... of independent vitality.” Id. at 12,121 S.Ct. 1060.

The government, not Hunton, bears the burden of demonstrating that the common interest doctrine applies to materials it has withheld, and the elements of the common interest doctrine must therefore be analyzed with precision. Klamath reaffirmed that contacts between self-interested parties and federal agencies are not to be casually exempted from FOIA’s scheme of disclosure. FOIA gives the public the right to know about the government’s communications with an entity that is “pressing its own view of its own interest” on the government. Id. at 14, 121 S.Ct. 1060.

The “dispositive point” in Klamath was that the “apparent object of the Tribe’s communications is a decision by an agency of the Government to support a claim by the Tribe that is necessarily adverse to the interests of competitors.” Id. To be sure, Klamath did not foreclose all confidential communications between federal agencies and private parties, even self-interested ones. Other exemptions may remain available. Indeed, in this case, Hunton has not challenged DOJ’s invocation of Exemption 4, which protects the exchange of confidential trade secret information. See 5 U.S.C. § 552(b)(4).

But Klamath requires the government to show that communications with a party involve something other than self-interested lobbying before the agency can withhold them under the Act. As the Court put it, “the intra-agency condition excludes, at the least, communications to or from an interested party seeking a Government benefit at the expense of other applicants.” Klamath, 532 U.S. at 1, 121 S.Ct. 1060. DOJ argues, however, that Klamath “should be confined to its facts and holding.” Appellee’s Br. at 27.

The department asserts that, unlike Klamath, this case does not entail competition for government benefits at the expense of others. This assertion is too sweeping. It is undeniable that support from DOJ in litigation is a very valuable benefit to confer upon a litigant, and it is equally self-evident that such a benefit entails a corresponding detriment to the litigant’s opponent. In Hanson, the only question was whether a common interest existed at all, not, as in this case, whether it extended as far as the government contended. Hanson, 372 F.3d. at 292.

In this regard, the government argues we should not overturn the district court’s conclusions concerning the scope of the common interest doctrine unless the district court committed clear error. But the appropriateness of the standard used by the district court to make its determinations is a question of law, which we review de novo. Ethyl Corp., 25 F.3d at 1246. The legal standard applied here was not the searching one that Klamath requires. In the proceedings below, the district court ruled that the common interest doctrine applied to exempt from FOIA communications between REVI and DOJ from March 2005 onward.

It explained that, as of March 2005, the record contained “indicia of joint strategy” between RIM and DOJ, insofar as DOJ and REVI “agreed to exchange declarations, other proposed pleadings, and their views on issues relating to the effect of any injunction on the government and the public interest.” This analysis was flawed in two respects.

First, although a common interest agreement can be inferred where two parties are clearly collaborating in advance of litigation, mere “indicia” of joint strategy as of a particular point in time are insufficient to demonstrate that a common interest agreement has been formed. Second, it is not clear that the particular “indicia” identified by the district court pointed to an actual common interest agreement, as opposed to a mere confidentiality agreement.

Hunton contends that the parties exchanged declarations, proposed pleadings, and views as part of the process of persuading DOJ to become involved in the BlackBerry litigation. That contention is entirely plausible, particularly to the extent that the communications were initiated by RIM or flowed primarily from RIM to DOJ. The district court gave no indication that there was anything in the documents themselves that showed them to be part of a joint litigation effort, rather than an attempt by REVI to push its arguments on the government.

Documents exchanged before a common interest agreement is established are not protected from disclosure. In re Grand Jury Subpoena: Under Seal, 415 F.3d 333, 341 (4th Cir.2005). Thus, a proper assessment of the applicability of the common interest doctrine in this case requires a determination of the point in time when DOJ decided that it was in the public interest for RIM to prevail in its litigation with NTP and agreed to partner with RIM in doing so. The danger in this area is once again that mere lobbying efforts, as opposed to joint litigation strategy, will be removed from FOIA’s reach. B.

We thus turn to the record and consider those materials exchanged between RIM and DOJ between March 2005 and November 2005. In his affidavit, Fargo, representing DOJ, stated that he agreed to exchange documents with Fenster, representing RIM, on a confidential basis at the conclusion of their March 10, 2005, meeting. Hunton questions whether any agreement was reached, pointing to the absence of any record of it in Fargo’s contemporaneous notes.

The district court appears not to have determined whether the parties actually formed an agreement as of that date, noting only that there were “indicia of joint strategy” from that point onward. While agreement need not assume a particular form, an agreement there must be. If RIM was simply approaching DOJ over the prospect that there might one day be a joint litigation effort, such contacts and discussions seem too preliminary to remove from disclosure under Exemption 5. The record raises a number of questions as to whether a common interest agreement existed as early as March 2005.

First, there is Fargo’s description of the interest DOJ shared with RIM. In his deposition, Fargo stated that DOJ and RIM shared an interest in communicating about the possible effects that an injunction entered against RIM would have. According to Fargo, RIM and DOJ agreed that “we would exchange views. Certainly, they would provide us with views, potential sources of information, potentially some draft declarations.” This characterization, however, comes close to an agreement to exchange information in order to make an assessment, rather than an agreement to undertake a joint legal strategy.

Second, Fargo stated in his deposition testimony that DOJ did not decide to file a Statement of Interest until shortly before the Statement was filed on November 8, 2005. As late as October 26, 2005, it appears that Fenster was communicating with Fargo regarding the “preparation of [a] declaration” addressing the government’s and the public’s interest. Thus, DOJ did not decide to become involved in the BlackBerry litigation until some time after the March 10, 2005, meeting, and there is reason to believe that Fenster was still pressing the government to take that step as late as October 2005.

Third, the parties failed to create a written common interest agreement until November 2005. This was the time DOJ filed its Statement of Interest, further suggesting that the government may not have made up its mind in the preceding period, or at least, not as early as March 2005. Fargo was an experienced attorney who knew how to create such an agreement if it served the public interest to do so. As DOJ itself notes, Fargo “routinely” creates common interest agreements and reviews those created by his staff.

Yet neither party made any kind of “common interest” notation on their written communications until October 2005. In addition, while Fargo told his supervisor that he was entering a common interest agreement in November 2005, he told his supervisor only that he and Fenster were exchanging materials “on a confidential basis” prior to that time. Finally, the record in this case presents the sort of features that underscore the need for watchfulness where the common interest doctrine is sought as a means to avoid FOIA’s disclosure requirements.

In this regard, it is significant if communications were initiated by the private party, if the bulk of the communications came from a private party, and if there are sparse indications that the government had come to terms with the public interest at stake in the case. The purposes and initiatives of the private party are especially important. In the proceedings below, Hunton suggested that RFM used DOJ in part as a conduit through which it could present its views to the PTO.

Because PTO reexamination proceedings were underway, communications between the PTO and RIM were in some instances required to be on-the-record and in others were prohibited outright. The record in this case reveals that DOJ forwarded draft pleadings supplied by RIM to the PTO. DOJ maintains that it had an interest in finding out the status of the reexamination proceedings, but it is not immediately clear why such an inquiry would require DOJ to pass litigation materials along from RIM to the PTO.

Athough the district court, after examining the DOJ-PTO communications, found that the PTO had not disclosed details of the patent reexaminations for DOJ to relay to RIM, that finding does not speak to Hunton’s suggestion that RIM was seeking to influence the PTO. Hunton’s claim raises precisely the sorts of concerns about self-interested advocacy that were so central in Klamath — particularly since there was a similar risk in Klamath that one agency representing private interests in litigation would communicate with a sister agency that was simultaneously adjudicating closely related claims.

In short, there are a number of items in the record suggesting that DOJ may not have decided to partner with RIM in the BlackBerry litigation much before November 2005. The fact that DOJ later concluded it shared RIM’s interest does not protect communications between the two before that decision was made. An agreement to hear what RIM had to say and to keep what it heard confidential must not be confused with a conclusion that the public interest required taking RIM’s side. We express no opinion on the ultimate merits of Hunton’s objections to DOJ’s withholding of these materials.

Athough we have highlighted aspects of the record that call DOJ’s claims into some doubt, the record does not foreclose the possibility that the parties did indeed reach a common interest understanding at some point before November 2005, which the written agreement served only to memorialize. The common interest doctrine requires a meeting of the minds, but it does not require that the agreement be reduced to writing or that litigation actually have commenced. What FOIA requires is not that the government invariably lose in disclosure disputes, but that its claims be carefully evaluated.

The analysis we set forth combines a recognition of the government’s right to partner with those who share its legal interests with the judicial skepticism that FOIA demands. It will be the district court’s job on remand to address any agreement and submissions of the parties under the standards we have identified. C. We finally thank our colleague for his thoughtful dissent, to which a brief response is in order.

The dissent asserts that DOJ cannot invoke Exemption 5 because, notwithstanding the joint RI1WDOJ effort to defeat the injunction, RIM was “acting in its own interest.” But that principle would eviscerate the government’s discovery privileges, since anyone who joins forces with the government in litigation will benefit if the partnership is successful. What the dissent misses is that where the government decides to work with a party to achieve a legal result beneficial to both, the danger of “self-advoca[cy] at the expense of others” subsides. Klamath, 532 U.S. at 12, 121 S.Ct. 1060.

Within the scope of their shared interest, the time for advocacy has by then passed. In attempting to realize their shared goal, each side advises the other with an eye toward “truth and its sense of what good judgment calls for.” Id. at 10, 121 S.Ct. 1060. Our remand in this case requires district courts to be attentive both to when a common interest is formed and to what communications lie within the scope of that interest. By contrast, our dissenting colleague recognizes virtually no right on the part of the government to share common interests at all.

Recognizing the sweeping nature of its approach, the dissent intimates that the result might be different if the government were a party in litigation and “a communication, privileged under the common interest doctrine, between an agency and a non-agency co-party” were at issue. But why? The government’s co-party would still from the dissent’s perspective be “acting in its own interest.” The communications would still, under the dissent’s view, not be inter- or intra-agency. The criteria announced by the dissent have no logical stopping point short of crippling the government’s discovery privileges.

And beyond that, how is the co-party rule to be applied, to take just a few examples, to parallel litigation, to indemnitors who lack formal party status, to amici, to movants whose request for leave to intervene is under advisement? And further, although privileges may be asserted in litigation, they often protect earlier attorney-client communications and attorney work product because litigation is not something that springs forth instantaneously like Athena from Zeus’s head.

The question simply cannot be resolved by drawing formal or arbitrary lines that overlook the actual point at which a common interest redounding to the public benefit is formed. Words take their meaning only in context. Exemption 5 is brief and its phrasing is, to the Supreme Court at least, “Delphic.” Julian, 486 U.S. at 11, 108 S.Ct. 1606. This much, however, is true: the Court has “consistently” held that FOIA is not to be used to “supplement civil discovery” and obtain “material that is normally privileged.” Weber Aircraft, 465 U.S. at 801, 104 S.Ct. 1488.

In short, the government is not to be denied the right, available to every private litigant, to seek and obtain undiscoverable legal advice, information, and counsel, including from a common interest partner. Yet the dissent, through an unfortunate parsing of Exemption 5’s opaque wording, would hobble the government alone in just this fashion and deny Exemption 5 the only meaning it could have. V. FOIA and its exemptions frequently present a tension.

The Act reflects both a commitment to transparency and to administrative efficacy, to the public’s right to know and to agencies’ duty to fulfill their stated missions. Here those interests are obvious. Congress did not intend to shield from the public’s eye the efforts of private parties to lobby an undecided agency to expend public resources and shape agency policy in ways beneficial to themselves. Neither did it intend, however, to force federal agencies to surrender critical civil discovery protections and litigate with one hand behind their backs.

It is the job of the courts, bearing in mind the government’s burden, to give effect to both of the values that FOIA seeks to advance. For the foregoing reasons, we affirm the district court’s conclusion that privileged communications between DOJ and RIM subsequent to their November 2005 common interest agreement are protected from disclosure by FOIA Exemption 5. We vacate its conclusion that a common interest relationship existed between RFM and DOJ from March 2005 to November 2005.

On remand, the district court should determine the point in time when DOJ decided that the public’s interest converged with RIM’s interest in opposing broad injunctive relief, that it wanted RIM'to prevail in its litigation, and that it would assist RIM in doing so. The judgment is hereby AFFIRMED IN PART, VACATED AND REMANDED IN PART. . Hunton also argues that DOJ did not share RIM’s interest in preventing the entry of a BlackBerry injunction altogether, as opposed to an interest in ensuring that a BlackBerry injunction was properly narrowed. A fair

interpretation of a common interest agreement, however, must leave room for the parties to debate the means by which they will secure their common end. Since defeating a BlackBerry injunction outright was one way — and may realistically have been the only way — to protect governmental BlackBerry use, discussion of this topic fell within the scope of RIM’s and DOJ’s common interests. .

We agree with the dissent that the intraagency prong of Exemption 5 cannot be collapsed into the privilege prong, see Klamath, 532 U.S. at 12, 121 S.Ct. 1060, but as we have already noted, the common interest doctrine is not itself a privilege. It plays the same role in litigation that the intra-agency requirement does under FOIA: it identifies those with whom materials of a privileged nature can be shared.