FEDERAL TRADE COMMISSION, ET AL., PETITIONERS V. GROLIER INCORPORATED No. 82-372 In the Supreme Court of the United States October Term, 1982 On Writ of Certiorari to the United States Court of Appeals for the District of Columbia Circuit Brief for the Petitioners PARTIES TO THE PROCEEDING The parties in the court of appeals, in addition to those named in the caption, were the members of the Federal Trade Commission. At present, the members of the Commission are James C. Miller, III, Chairman, and Michael Pertschuk, David M. Clanton, Patricia P. Bailey, and George W. Douglas, Commissioners. TABLE OF CONTENTS Opinions below Jurisdiction Statute involved Statement Summary of argument Argument The work product of government attorneys is protected from disclosure under Exemption 5 of the Freedom of Information Act after the conclusion of the litigation for which that material was prepared and without regard to the existence or potential existence of related litigation A. The work-product privilege continues despite the termination of the litigation for which it was prepared and even though no related litigation exists or potentially exists B. Even if work product may be disclosed in civil discovery when the case for which it was prepared has ended and no related case exists or potentially exists, the work product of government attorneys is nevertheless exempt from mandatory disclosure under the FOIA Conclusion OPINIONS BELOW The opinion of the court of appeals (Pet. App. 1a-14a) is reported at 671 F.2d 553. The opinion of the district court (Pet. App. 21a-23a) is not reported. JURISDICTION The judgment of the court of appeals was entered on February 5, 1982 (Pet. App. 15a), and a timely petition for rehearing was denied on April 7, 1982 (id. at 16a, 19a). By orders of June 29, 1982, and August 3, 1982, the Chief Justice extended the time within which to file a petition for writ of certiorari to and including September 4, 1982. The petition was filed on September 2, 1982, and was granted on November 8, 1982 (J.A. 52). The jurisdiction of this Court is invoked under 28 U.S.C. 1254(1). STATUTE INVOLVED Subsection (a)(3) of the Freedom of Information Act, 5 U.S.C. 552(a)(3), provides: Except with respect to the records made available under paragraphs (1) and (2) of this subsection, each agency, upon any request for records which (A) reasonably describes such records and (B) is made in accordance with published rules stating the time, place, fees (if any), and procedures to be followed, shall make the records promptly available to any person. Subsection (b)(5) of the Freedom of Information Act, 5 U.S.C. 552(b)(5) provides: (b) This section does not apply to matters that are -- * * * * * (5) 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. QUESTION PRESENTED Whether Exemption 5 of the Freedom of Information Act, 5 U.S.C. 552(b) (5), continues to exempt from mandatory disclosure the work product of government attorneys after the litigation for which the material was prepared has ended and related litigation does not exist or potentially exist. STATEMENT This Freedom of Information Act ("FOIA") suit arose out of a request by respondent Grolier Incorporated for documents of the Federal Trade Commission ("FTC") concerning, among other things, a covert investigation by the FTC of one of Grolier's subsidiaries, the Americana Corporation (Pet. App. 1a). /1/ This investigation had been undertaken in connection with a civil penalty action filed by the Department of Justice in 1972 (United States v. Americana Corp., Civil No. 388-72 (D.N.J.)) (Pet. App. 1a). /2/ The FTC complied in part with the FOIA request, but it withheld part or all of seven documents relating to the investigation of Americana. Grolier then brought suit in the United States District Court for the District of Columbia to compel release of those documents. The FTC contended that certain of the documents, /3/ including the four still at issue (i.e., Numbers 3, 5, 6, and 7), fell within Exemption 5 of the FOIA, 5 U.S.C. 552(b)(5), because they were work product compiled by FTC attorneys in anticipation of the civil penalty action against Americana and therefore would normally be privileged from discovery in civil litigation. Since Exemption 5 "exempt(s) those documents, and only those documents, normally privileged in the civil discovery context." (NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 149 (1975)), the FTC maintained that the documents sought by Grolier were exampt from disclosure under the FOIA. After reviewing the documents in camera, the district court found (Pet. App. 23a) that they "encompass opinions by attorneys regarding the evidentiary needs of the Americana action" and that "(t)hey also discuss specific methods of obtaining evidence in that litigation." The court held (ibid.) that "the documents (fell) within the rubric of 'mental impressions, conclusions, opinions or legal theories'" under Rule 26(b)(3) of the Federal Rules of Civil Procedure and that they therefore fell within Exemption 5 of the FOIA. A divided court of appeals vacated the district court's judgment regarding those four documents. The panel majority conceded that "(t) here is no question that the documents involved were work-product prepared as part of the Americana action" (Pet. App. 2a). However, the majority stated that the applicability of Exemption 5 of the FOIA depended upon "whether these documents continue to be privileged against disclosure several years after the Americana suit was terminated" (ibid.). The court below noted (Pet. App. 3a-4a, quoting In re Murphy, 560 F.2d 326, 334 (8th Cir. 1977) (emphasis added in opinion)) that several courts of appeals and district courts have held that "there is 'a perpetual protection for work product' extending beyond the termination of the litigation for which the documents were prepared." But the court concluded (Pet. App. 5a) that "(e)xtending the work-product protection only to subsequent related cases best comports with the fact that the privilege is qualified, not absolute." The court reasoned (id. at 6a) that effective legal representation would not be adversely affected by disclosure of attorneys' mental impressions, conclusions, opinions, or legal theories "(w)hen litigation has ended and no potential for related actions exists * * *." Accordingly, the court held (id. at 7a; emphasis in original) that "in the context of an FOIA request, attorney work-product from terminated litigation remains exempt from disclosure only when litigation related to the terminated action exists or potentially exists." Applying that test to the present case, the court of appeals rejected the suggestion that Grolier's FOIA suit constituted related litigation (Pet. App. 8a) and stated (id. at 7a) that "there does not appear to be any suit or potential suit related to the original Americana action." The court therefore remanded the case "for reconsideration of the applicability of the work-product privilege in light of the apparent absence of related litigation" (id. at 7a-8a). Judge MacKinnon dissented. In his view, the Commission properly withheld the requested documents under Exemption 5 "because the termination of the litigation to which the work-product relates does not destroy this privilege" (Pet. App. 9a). Moreover, assuming that the work-product privilege were limited by a "related case" rule, Judge MacKinnon concluded that "the present suit could not be more directly related to the Americana litigation" (id. at 13a). The court of appeals subsequently denied the government's suggestion for rehearing en banc by a vote of five to four (Pet. App. 19a-20a). SUMMARY OF ARGUMENT This is a Freedom of Information Act case in which respondent sought disclosure of work product prepared by Federal Trade Commission attorneys for a civil action that ended several years ago. The Commission denied disclosure on the ground that the documents were protected by the attorney work-product privilege and were accordingly exempt from disclosure under the FOIA by virtue of Exemption 5, 5 U.S.C. 552(b)(5). That exemption protects "memorandums or letters which would not be available by law to a party other than an agency in litigation with the agency." It is settled that this exemption was intended to protect the working papers of government attorneys. See FOMC v. Merrill, 443 U.S. 340, 353 (1979); NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 154 (1975). The court of appeals nonetheless ordered the documents disclosed, holding (Pet. App. 7a; emphasis in original) that "in the context of an FOIA request, attorney work-product from terminated litigation remains exempt from disclosure only when litigation related to the terminated action exists or potentially exists." This holding is clearly wrong. 1. The court of appeals' interpretation of the duration of the work-product privilege is not supported by precedent or logic. Certainly nothing in this Court's decisions or Fed. R. Civ. P. 26(b)(3), which codifies some aspects of the privilege, supports the view that the privilege ends upon the conclusion of the litigation for which it was prepared unless related litigation "exists or potentially exists" (Pet. App. 7a). In addition, the interpretation of the privilege adopted by the court below has been rejected by three other courts of appeals that have considered the issue. In re Murphy, 560 F.2d 326, 335 (8th Cir. 1977); Duplan Corp. v. Deering Milliken, Inc., 540 F.2d 1215, 1219 (4th Cir. 1976); United States v. Leggett & Platt, Inc., 542 F.2d 655, 660 (6th Cir. 1976), cert. denied, 430 U.S. 945 (1977). The decision below also is fundamentally inconsistent with the underlying purposes of the work-product privilege. As this Court has recognized, the disclosure of attorneys' working papers may seriously undermine the proper functioning of the adversary process. See Upjohn Co. v United States, 449 U.S. 383, 398 (1981); United States v. Nobles, 422 U.S. 225, 236-240 (1975); Hickman v. Taylor, 329 U.S. 495 (1947). In preparing for litigation, an attorney must "assemble information, sift what he considers to be the relevant from the irrelevant facts, prepare his legal theories and plan his strategy without undue and needless interference" (Hickman v. Taylor, supra, 329 U.S. at 511). "This work is reflected * * * in interviews, statements, memoranda, correspondence, briefs, mental impressions, personal beliefs, and countless other tangible and intangible ways * * *" (ibid.). If such materials were subject to disclosure, "much of what is now put down in writing would remain unwritten," and both "the interests of the clients and the cause of justice would be poorly served" (ibid.). Subjecting the work-product privilege to the limitations imposed by the court of appeals would produce many of the deleterious consequences that the privilege was designed to prevent. The concepts of "related" litigation and litigation that "potentially exists" (Pet. App. 7a) are so vague that attorneys would be unable to predict how they would be applied in particular cases. Cases may be "related" in innumerable ways. For example, they may involve similar facts, legal theories, parties, or other personnel. There are also infinite degrees of relatedness. At the very least, much litigation would be required to determine what is meant in this context by "related" cases, and we frankly doubt whether it would be possible to develop clear standards that would permit attorneys to predict when and under what circumstances their working papers might be ordered disclosed. The concept of related litigation that "potentially exists" is of even less practical utility. Attorneys and judges would have a difficult time indeed attempting to ascertain what future lawsuits may involve. As a result of these ambiguities, cautious attorneys would be forced to assume that any working papers kept after the termination of the litigation for which they were prepared might be disclosed to adversaries and others. In many instances, attorneys would refrain from preparing or collecting work product that might prove damaging if released. In addition, many valuable files would likely be destroyed upon the conclusion of the litigation for which they were assembled. This would impair the effective representation of clients, make legal services more expensive, and interfere with the functioning of the legal system. Applying the concepts of related and potential litigation would be particularly difficult in FOIA and other government cases. The volume of government litigation (and potential litigation) that would have to be surveyed is enormous, and most cases involving the same agency are related in many important respects. For example, they may involve the same statutory provisions, similar factual situations, similar investigative strategies or practices, the same or similar adversaries, and the same agency personnel. Assuming arguendo that attorneys had a rough idea of what was meant by "related" cases and cases that "potentially exist," attorneys' trial preparation and representation of clients might well be adversely affected by the knowledge that damaging work product would be discoverable in subsequent, unrelated cases. Certain work-product materials may be equally damaging no matter in what context they are disclosed. These include documents discussing materials that could prove harmful or embarrassing to a client, as well as documents that reflect unfavorably upon an attorney's performance. Anxious to prevent disclosure of such materials at any time, attorneys would probably refrain from preparing, collecting, or preserving many documents if they were discoverable in subsequent, "unrelated" cases. 2. Even if the court of appeals' decision correctly reflected the duration of the work-product privilege in civil discovery, the working papers of government attorneys would nevertheless be protected from disclosure under Exemption 5 of the FOIA. First, Exemption 5 exemps those documents "normally privileged in the civil discovery context" (NLRB v. Sears, Roebuck & Co., supra, 421 U.S. at 149). Documents may be obtained despite that exemption only if they would "'routinely be disclosed' in private litigation" (id. at 149 n.16, quoting H.R. Rep. No. 1497, 89th Cong., 2d Sess. 10 (1966)). Under the court of appeals' decision, however, work product from terminated litigation would be discoverable only if (a) it was sought in an "unrelated" case and there was no potential for related litigation or (b) the party seeking discovery was able to show a sufficient need for the materials. Accordingly, since work-product materials from terminated cases would not "routinely be disclosed" in civil discovery but would be discoverable only under certain circumstances, those materials would be shielded by Exemption 5. Second, as Judge MacKinnon remarked in dissent (see Pet. App. 13a), an FOIA suit seeking work product from an earlier case is clearly "related" to the prior case. This is especially so where, as here, the party making the FOIA request was a party (or is the parent corporation of a party) in the previous action. ARGUMENT THE WORK PRODUCT OF GOVERNMENT ATTORNEYS IS PROTECTED FROM DISCLOSURE UNDER EXEMPTION 5 OF THE FREEDOM OF INFORMATION ACT AFTER THE CONCLUSION OF THE LITIGATION FOR WHICH THAT MATERIAL WAS PREPARED AND WITHOUT REGARD TO THE EXISTENCE OR POTENTIAL EXISTENCE OF RELATED LITIGATION Exemption 5 of the Freedom of Information Act, 5 U.S.C. 552(b)(5), protects from disclosure "memorandums or letters which would not be available by law to a party other than an agency in litigation with the agency." This exemption was clearly intended to protect the working papers of government attorneys from disclosure. The Senate Report states (S. Rep. No. 813, 89th Cong., 1st Sess. 2 (1965)) that the material protected by Exemption 5 "include(s) the working papers of the agency attorney and documents which would come within the attorney-client privilege if applied to private parties." /4/ This Court has also noted that it is "clear that Congress had the attorney's work-product privilege specifically in mind when it adopted Exemption 5." NLRB v. Sears, Roebuck & co., supra, 421 U.S. at 154. See FOMC v. Merrill, 443 U.S. 340, 353, 357 (1979) (House and Senate Reports provide "unequivocal support for an Exemption 5 privilege for * * * attorney work product privileges"). Although Exemption 5 was meant to protect the work product of government attorneys from mandatory disclosure to the public, the court of appeals held that such materials lose their privileged status upon the termination of the litigation for which they were prepared unless the government is able to demonstrate that related litigation exists or potentially exists. This holding is erroneous for at least two reasons. First, the work-product privilege is not subject to such limitations, in either the civil discovery or the FOIA context. Second, even if such limitations exist in civil discovery, disclosure in a subsequent FOIA suit would not be warranted. Exemption 5 applies unless covered materials would "'routinely be disclosed' in private litigation" (NLRB v. Sears, Roebuck & Co., supra, 421 U.S. at 149 n.16, quoting H.R. Rep. No. 1497, 89th Cong., 2d Sess. 10 (1966) (emphasis added)), and even under the court of appeals' interpretation of the privilege, the working papers of government attorneys would not "routinely be disclosed" in civil discovery but could be obtained only in limited circumstances. A. The work-product privilege continues despite the termination of the litigation for which it was prepared and even though no related litigation exists or potentially exists As noted, the court of appeals held that the documents at issue in the present case, although concededly the work product of government attorneys (see Pet. App. 2a), were no longer protected by the work-product privilege because the litigation for which they had been prepared had ended and related litigation did not exist or "potentially" exist. There is no basis for these unprecedented limitations upon the work-product doctrine. 1. The attorney work-product privilege was first recognized by this Court in Hickman v. Taylor, 329 U.S. 495 (1947), a Jones Act suit brought by the representative of a crewman drowned in a tugboat accident. After counsel for defendants, the tug owners and underwriters, had interviewed survivors of the accident, opposing counsel sought to obtain copies of any written statements provided by the survivors and also to compel defense counsel to "set forth in detail" any oral statements the survivors had made (id. at 498-499). This Court held that the written and oral statements were privileged and refused to permit their discovery. The decision in Hickman was based upon the recognition that discovery of attorney work product would interfere with sound trial preparation and thereby undermine the adversary process. The Court stated (329 U.S. at 510-511): In performing his various duties * * *, it is essential that a lawyer work with a certain degree of privacy, free from unnecessary intrusion by opposing parties and their counsel. Proper preparation of a client's case demands that he assemble information, sift what he considers to be the relevant from the irrelevant facts, prepare his legal theories and plan his strategy without undue and needless interference. That is the historical and the necessary way in which lawyers act within the framework of our system of jurisprudence to promote justice and to protect their clients' interests. This work is reflected, of course, in interviews, statements, memoranda, correspondence, briefs, mental impressions, personal beliefs and countless other tangible and intangible ways -- aptly though roughly termed by the Circuit Court of Appeals in this case as the "work product of the lawyer." Were such materials open to opposing counsel on mere demand, much of what is now put down in writing would remain unwritten. An attorney's thoughts, heretofore inviolate, would not be his own. Inefficiency, unfairness and sharp practices would inevitably develop in the giving of legal advice and in the preparation of cases for trial. The effect on the legal profession would be demoralizing. And the interests of the clients and the cause of justice would be poorly served. The Court added, however, that the work-product privilege is not absolute. The Court stated (329 U.S. at 511-512) that "written materials obtained or prepared * * * with an eye toward litigation" may be discovered upon a showing of necessity. On the other hand, the Court observed (id. at 512-513), the "mental impressions" of counsel may be discovered, if at all, only in "rare" circumstances. /5/ The considerations that led this Court to recognize the work-product privilege apply with particular force in the case of government attorneys. The working papers of government lawyers may be sought, not only in civil discovery, but under the FOIA as well. Yet, government attorneys are far more likely to be dependent upon the existence of comprehensive written files than most of their counterparts in private practice: government cases tend to be larger and to last longer; government attorneys tend to be more transient; coordination of the positions taken in many varieties of cases is of unusual importance; and maintaining supervision and coordination in an organization as vast and complex as the United States government naturally requires great reliance upon past records. The strong public policy underlying the work-product doctrine has been repeatedly reaffirmed by this Court. See, e.g., Upjohn Co. v. United States, 449 U.S. 383, 398 (1981); United States v. Nobles, 422 U.S. 225, 236-240 (1975). That policy has also been embodied in Federal Rule of Civil Procedure 26(b)(3), which provides in pertinent part: (3) Trial Preparation: Materials. Subject to the provisions of subdivision (b)(4) of this rule, a party may obtain discovery of documents and tangible things otherwise discoverable under subdivision (b)(1) of this rule and prepared in anticipation of litigation or for trial by or for another party or by or for that other party's representative * * * only upon a showing that the party seeking discovery has substantial need of the materials in the preparation of his case and that he is unable without undue hardship to obtain the substantial equivalent of the materials by other means. Nothing in this Court's decisions or in Rule 26(b)(3) suggests that the work-product privilege terminates with the conclusion of the litigation that gave rise to the work product. On the contrary, Rule 26(b)(3) by its terms protects "documents and tangible things * * * prepared in anticipation of litigation or for trial" (emphasis added) and not just those prepared in anticipation of or for the suit then before the court. Nor is there a hint in this Court's decisions or in Rule 26(b)(3) that the privilege survives the termination of the suit for which the work product was prepared only if related litigation "exists or potentially exists" (Pet. App. 7a), whatever that may mean. /6/ With the exception of the court below, all the courts of appeals that have reached the issue have held that work product from terminated litigation remains privileged in subsequent, unrelated cases. In Duplan Corp. v. Moulinage et Retorderie de Chavanoz, 487 F.2d 480, 483-484 (4th Cir. 1973), the court concluded that perpetual protection for work product best comports with the rationale of Hickman and explained (id. at 483): (Hickman's) overriding concern is that the lawyer's morale be protected as he performs his professional functions in planning litigation and preparing his case. This work product immunity is the embodiment of a policy that a lawyer doing a lawyer's work in preparation of a case for trial should not be hampered by the knowledge that he might be called upon at any time to hand over the result of his work to an opponent. The Fourth Circuit specifically rejected the suggestion that "documents prepared for one case are protected in a second case only if the two cases are 'closely related'" (id. at 484 n.15). The court stated (ibid.) that "to dispose of this delicate and important question by such a technical touchstone is incompatible with the essential basis of the Hickman decision." See Duplan Corp. v. Deering Milliken, Inc., 540 F.2d 1215, 1219 (4th Cir. 1976); Duplan Corp. v. Moulinage et Retorderie de Chavanoz, 509 F.2d 730, 735 (4th Cir. 1974), cert. denied, 420 U.S. 997 (1975). The Sixth Circuit followed the Fourth in United States v. Leggett & Platt, Inc., 524 F.2d 655 (6th Cir. 1976), cert. denied, 430 U.S. 945 (1977). In that case, the government initiated a civil antitrust action to require a manufacturing company to divest itself of two other recently acquired firms in the same industry, and the defendant company then sought discovery of work product prepared by government lawyers in anticipation of litigation concerning similar acquisitions by other companies in that industry. Holding that work product retains its protection when sought in subsequent cases, the court stated (id. at 660): Were the work product doctrine an unpenetrable protection against discovery, we would be less willing to apply it to work produced in anticipation of other litigation. But the work product doctrine provides only a qualified protection against discovery, at least of products other than (opinion work product). Finally, in In re Murphy, 560 F.2d 326 (8th Cir. 1977), the Eighth Circuit reached a similar conclusion. Rejecting the view that the work product privilege applies only in the very case for which the materials were prepared or in a "related" case, the court reasoned (id. at 335; footnote omitted): If work product is protected in related, but not unrelated future cases, an attorney would be hesitant to assemble extensive work product materials because of the concern that the materials will not be protected in later, unrelated litigation. The unrelatedness of the subsequent litigation provides an insufficient basis for disregarding the privilege articulated in Hickman and incorporated in Rule 26(b)(3). The mischief engendered by allowing discovery of work product recognized in Hickman would apply with equal vigor to discovery in future, unrelated litigation. In the present case, the court of appeals purported to find support for its decision in the purpose of the work-product privilege (see Pet. App. 6a). It seems clear, however, that the policy underlying the privilege would be poorly served if subjected to the constraints imposed by the decision below. We will first discuss whether there is any justification for holding that the work-product privilege automatically dissolves upon the conclusion of the litigation in connection with which it was originally prepared. We will then address the two additional refinements adopted by the court of appeals -- i.e., the "related" litigation and "potential" related litigation tests. 2. Little discussion is required to demonstrate that the policy underlying the work-product doctrine would be seriously eroded if the privilege immediately evaporated upon the termination of the litigation in connection with which it was prepared. We readily acknowledge that the need to shield work product is usually strongest during the pendency of that litigation, for the more immediate the threat of disclosure the greater the effect upon an attorney's preparation and overall representation is likely to be. But common sense suggests that the possibility of disclosure at some future date may also have a highly deleterious effect upon an attorney's preparation and representation and consequently upon the adversary process. This may be illustrated by considering a slight variation of the situation in Hickman. In that case, five crewmen drowned in a tugboat accident, but representatives of four eventually settled their claims. As previously noted, the representative of the fifth crewman filed suit and sought discovery of the witness statements obtained by defense counsel. This Court refused to allow discovery of those statements, which constituted work product gathered in preparation for the case before the Court. Suppose, however, that the other four crewmen, rather than settling their claims, had decided to wait until after the initial action ended before bringing suit. Under those circumstances, it could hardly be argued that the policy underlying the work product privilege would be adequately served if the privilege terminated upon the completion of the initial suit. If the preparation of defendants' counsel would have been adversely affected by the knowledge that his work product might be revealed during the first suit, the realization that the same materials might be disclosed in the subsequent actions almost certainly would have had an equivalent effect. To take another example, suppose that the complaint in Hickman had been dismissed on grounds that did not bar reinstitution of suit -- for example, for lack of in personam jurisdiction or venue or because suit was brought under the wrong statute or legal theory. Surely the policy underlying the work-product privilege would demand continued protection of work product prepared in connection with the first, aborted proceeding. Even where the likelihood of subsequent related litigation is not as strong as it is in the above examples, an attorney's preparation and representation might be affected significantly and adversely if he knows that his work product may be disclosed upon the conclusion of the suit at hand. For example, sound trial preparation often requires an attorney to probe all possible weaknesses in his client's case. An attorney may consequently find it necessary or helpful to prepare memoranda discussing materials that reflect unfavorably upon his client, such as materials that might be used to cross-examine or impeach his client or other persons whom the attorney is contemplating calling as witnesses. An attorney may believe that only by collecting and evaluating such materials can he intelligently make or advise his client concerning numerous litigation decisions, such as whether to file suit, whether to settle a pending action, whether to assert a particular claim or defense, which witnesses to call at trial, and which items of evidence to introduce. However, if an attorney believed that documents discussing materials damaging to a client might be disclosed upon completion of the case at hand, the attorney's trial preparation could well be altered. Potentially damaging information might not be collected, examines, or reduced to writing. The attorney and his client might therefore be required to make critical litigation decisions and perhaps proceed to trial without the benefit of the preparation that the attorney would otherwise deem essential. Both the interest of the client and the adversary process would suffer as a consequence. If the court of appeals' decision is affirmed, attorneys' trial preparation might be altered not only to protect their clients but also to protect themselves. If work product lost all protection following the conclusion of the case for which it was prepared, attorneys might hesitate to prepare or retain materials revealing or acknowledging their own mistakes. This too would have undesirable consequences, because a frank assessment of the conduct of litigation may benefit both an attorney's client and society's larger interests. A recognition of past errors may permit their correction or may lead to the conclusion that proposed litigation should not be initiated or that pending litigation should be settled or dismissed. Yet many attorneys, both in government and private practice, would probably be much more reluctant to engage in candid analysis of their own performance or that of their colleagues if materials containing such analysis were subject to disclosure at the conclusion of the pending litigation. By the same token, if the attorney work-product privilege dissolved upon the completion of the litigation for which the requested materials were prepared, valuable files might well be destroyed at that time to prevent possible future disclosure. Many irreplaceable documents would be lost, and other documents destroyed at the conclusion of one case would have to be prepared or gathered again when subsequent cases arose. The efficient functioning of the legal process would accordingly be hampered. In sum, it seems obvious that the strong public policy underlying the work-product doctrine would be frustrated if the privilege automatically dissolved upon the completion of the litigation for which the work-product materials were prepared. If the privilege were limited in that fashion, "much of what is now put down in writing would remain unwritten. An attorney's thoughts, heretofore inviolate, would not be his own. Inefficiency, unfairness and sharp practices would inevitably develop in the giving of legal advice and in the preparation of cases for trial. The effect on the legal profession would be demoralizing. And the interests of the clients and the cause of justice would be poorly served." Hickman v. Taylor, supra, 329 U.S. at 511. On the other hand, recognizing that the work-product privilege is not subject to temporal restraints would not create undue hardship for civil litigants because, at least with respect to non-opinion work product, the privilege may be overcome where sufficient need for the materials is shown. 3. The court of appeals introduced two qualifications designed to soften the impact of the rule that work product loses all protection upon the conclusion of the case for which it was prepared. The court concluded that the work-product privilege could survive the termination of the original litigation if (a) related litigation is in existence or (b) there is a potential for such litigation. However, neither of these qualifications is workable or comports with the rationale for the privilege. What the court of appeals meant by "related" litigation is far from clear. Cases may be related in countless ways. For instance, they may involve similar facts; the same or similar investigative techniques, causes of action, or defenses; or the same lay or expert witnesses, investigators, trial attorneys, or supervisory personnel. There are also infinite degrees of relatedness. The very fact that work product prepared for one case is relevant and hence potentially discoverable in another case (see Fed. R. Civ. P. 26(b)(1) suggests that the two cases are related in at least one significant respect. /7/ Application of the concept of related litigation would be particularly difficult in cases involving government agencies. Because of their continuing enforcement responsibilities, many federal agencies are almost constantly engaged in litigation involving the same laws; recurring factual situations; the same or similar opposing parties; the same investigative, trial, and supervisory personnel; analogous enforcement decisions; and the same or similar investigative and litigative strategies or techniques. Thus, cases involving such agencies are typically related in many significant ways, and disclosure of work product from cases long since closed may reveal privileged information of considerable value to regulatees and potential adversaries. In addition, the sheer volume of cases involving government agencies would compound the difficulty of ascertaining whether any are related to the litigation for which the work product at issue was created. For the courts that will ultimately be required to adjudicate work-product disputes, the related litigation test would create the difficult, if not impossible, task of specifying what sort of relationship is required and just how close the relationship between cases must be before they are deemed to be related. At the very least, developing such standards would take many years and much litigation, and we frankly doubt whether it would ever be possible to articulate clear standards capable of producing predictable results. Of course, if attorneys cannot predict what would be deemed "related" litigation for these purposes, they would not know if, when, and under what circumstances their work product may be ordered disclosed. Collecting and preserving potentially damaging or embarrassing work-product materials would accordingly be a risky undertaking from which many cautious attorneys might elect to refrain. As this Court has observed (Upjohn Co. v. United States, supra, 449 U.S. at 393), "(a)n uncertain privilege, or one which purports to be certain but results in widely varying applications by the courts, is little better than no privilege at all." Even if attorneys had a rough idea what might later be deemed "related" litigation, their trial preparation and overall representation might well be adversely affected by the knowledge that disclosure of damaging work product might be ordered in subsequent unrelated cases. Although the court of appeals did not explain the conceptual basis for the related litigation test, in order to comport with the policy underlying the work-product privilege that test would have to rest upon the premise that disclosure of work product in or during the pendency of related litigation is likely to have a greater adverse impact upon attorneys' trial preparation and representation than is disclosure in or during unrelated suits. This could be so for two reasons: (a) it could be argued that attorneys are more likely to foresee (and thus fear) the possibility of disclosure during subsequent related cases; and (b) it could be argued that disclosure during the pendency of such suits is likely to be more damaging to attorneys and their clients. If this is indeed the underlying rationale of the related litigation test, we question the validity of the broad, unsubstantiated generalizations upon which it is based. While there are undoubtedly circumstances in which subsequent related cases may be foreseen, there are also many circumstances in which it is apparent to attorneys that work-product materials gathered for one case may be relevant to and sought in other cases that may be deemed unrelated. In addition, the very fact that such materials are sought by an opposing party in civil discovery and are found to be relevant to that suit is some evidence that their disclosure may not be innocuous. Furthermore, it seems obvious that revelation of certain work-product materials may be equally damaging no matter what the context in which they are disclosed. For example, documents discussing materials that could be used to impeach a particular client in one case might alert an adversary to the possibility of using those materials for the same purpose in a subsequent, unrelated case. Documents containing information that would damage the business or personal reputation of a client if made public might be just as harmful if revealed in an unrelated as opposed to a related suit. And materials that reflect unfavorably upon an attorney's performance would be equally damaging without regard to the nature of the case in which their disclosure is ordered. Finally, the related litigation test is fatally flawed because subsequent unrelated litigation may precede suits intimately related to the case in connection with which the work-product materials were prepared. Disclosure of those materials in the unrelated suit would mean that they will be available to adversaries in the related cases not yet begun. 4. Apparently recognizing that the "related litigation" test is inadequate, the court of appeals added that the work-product privilege continues as long as there is a "potential" for related litigation. In our view, the concept of whether related litigation "potentially exists" is meaningless for all practical purposes. To be sure, there are undoubtedly instances in which the possibility of subsequent related litigation is clear -- for example, where an action is dismissed without prejudice -- but absent such circumstances it is hard to imagine how a court would go about determining whether there is a "potential" for related litigation. /8/ Indeed, in our litigious society, what sort of litigation does not "potentially exist" at any time? Moreover, the concept of litigation that "potentially exists," like the concept of related litigation, would be particularly difficult to apply in government cases. Because of the enormous volume of government litigation, surveying all cases in the offing would be a Herculean chore. And no attorney could accurately predict what even a single agency such as the FTC is likely to investigate in the future, where those investigations are likely to lead, and what sort of litigation may ultimately result. Thus, the rule adopted by the court of appeals in this case would seriously undermine the important public policy underlying the work-product privilege. Attorneys preparing for trial would be unable to predict when or if potentially damaging work-product materials might be ordered disclosed, and they would have little assurance that such materials, if prepared or collected, would not work to the detriment of their clients or themselves. The results would be precisely those this Court sought to avoid in Hickman. /9/ 5. We have already touched upon the reasons why continuing protection for work product is especially important in government litigation and why the concepts of "related" and "potential" litigation would be particularly unworkable in this context. For many of these same reasons, we believe that even if the work product of private attorneys were subject to temporal restraints, continuing protection for the work product of government attorneys would be required. Unlike most private litigants, many government agencies exist in large measure for the purpose of litigating in either administrative or judicial proceedings. In addition, most government agencies have specialized enforcement and litigative responsibilities. Accordingly, they may participate over the years in a large number of cases involving the same laws and similar factual patterns, investigative techniques, evidentiary problems, and trial strategies and decisions. Work product prepared for prior cases is "closed" in only the most technical sense; it may reveal how the agency customarily deals with recurring legal problems and therefore may be valuable to regulatees and actual or potential adversaries for reasons that are seldom applicable with respect to the work product of private attorneys. To take one example, in United States v. Leggett & Platt, Inc., supra, in which the Sixth Circuit held that work product is perpetually protected, the company against which the government had initiated a civil antitrust action sought discovery of work product prepared by government attorneys in anticipation of the same type of litigation with other firms in the same industry. If those materials had been ordered disclosed, the defendant company might have gained much valuable information about the investigative and litigative strategies customarily employed by the government in cases of that sort. By contrast, if the antitrust action had been filed by a private party, it is doubtful whether disclosure of work product from any past case in which that party had been involved would yield similar fruit. And unless the defendant company had previously defended a similar suit, it is doubtful as well that the government could have obtained information concerning the defendant's investigative or litigative techniques by seeking discovery of work product from cases in which it was previously involved. The need to prevent disclosure of materials reflecting intra-agency advisory opinions, recommendations, and deliberations also finds expression in the government's deliberative process privilege, which is not subject to temporal restraints. See NLRB v. Sears, Roebuck & Co., supra, 421 U.S. at 149-151. This privilege, which is embodied in Exemption 5, of course may not be claimed by private parties. "(T)he ultimate purpose of this long-recognized privilege is to prevent injury to the quality of agency decisions," since such decisions "will clearly be affected by the communications received by the decisionmaker on the subject of the decision prior to the time the decision is made" (id. at 151). This same objective demands continuing protection for much government work product prepared for past cases. Since materials reflecting intra-agency deliberations concerning recurring litigative problems would be perpetually privileged, it would be anomolous indeed if a government attorney's file memoranda concerning the same issues lacked similar protection or if confidential materials reflecting how such problems were approached in practice were freely discoverable following the conclusion of the litigation for which they were prepared and all "related" litigation and potential litigation. See Department of State v. Washington Post Co., No. 81-535 (May 17, 1982), slip op. 6; FBI v. Abramson, No. 80-1735 (May 24, 1982), slip op. 12-13. In sum, the work product of government attorneys is fundamentally different from that of private practitioners and demands continued protection even if private attorneys' work product is subject to temporal limitations. 6. We have thus far addressed the application of the court of appeals' test in the civil discovery context. In a Freedom of Information Act case, like the present suit, that test makes even less sense. Applying its test in the FOIA context, the majority of the court of appeals held that a FOIA suit seeking work product from a terminated case may not be regarded as "related" litigation and that the applicability of the work-product privilege in the FOIA suit depends upon the fortuity of whether other litigation related to the original action is then pending or "potentially exists." We find this approach puzzling. First, we do not understand why a FOIA suit seeking documents from a terminated case should not be regarded as related litigation. See Pet. App. 13a (MacKinnon, J., dissenting). Since the documents sought in the FOIA suit would not have been created or assembled were it not for the first action, the two cases would appear to be closely related in the usual sense of that term. This conclusion also seems consistent with the apparent rationale for the related litigation test. As previously noted, if there is any basis for that test, it is that disclosure of work product in a related case is more likely to be foreseen and to cause damage to an attorney and his client than is disclosure in an unrelated suit. Judged by these criteria, FOIA suits seeking work product from terminated cases fully qualify as related litigation. The filing of a FOIA request -- a simple, inexpensive procedure -- is much more likely and foreseeable in most circumstances than the filing of a civil complaint and the pursuit of work product through civil discovery. In any event, even if not all FOIA suits seeking work product from terminated cases may be regarded as "related" litigation, there is surely no justification for holding that a FOIA suit may never be a "related" case (see Pet. App. 8a). For example, where, as here, documents are sought by the opposing party in the prior suit or its corporate parent, it seems plain that the two actions are closely related. 7. The court of appeals attempted to bolster its interpretation of the work-product privilege by relying upon what it characterized as a "substantial body of case law (that) supports the conclusion that the work-product privilege extends to subsequent cases only when they are related" (Pet. App. 4a; emphasis added). However, none of the cases cited by the court of appeals reached that question. In each of those cases, discovery of attorney work product in a related case was denied. See Republic Gear Co. v. Borg-Warner Corp., 381 F.2d 551, 557 (2d Cir. 1967); Hercules Inc. v. Exxon Corp., 434 F. Supp. 136, 153 (D. Del. 1977); Midland Investment Co. v. Van Alstyne, Noel & Co., 59 F.R.D. 134, 138 (S.D.N.Y. 1973); Philadelphia Electric Co. v. Anaconda American Brass Co., 275 F. Supp. 146, 148 (E.D. Pa. 1967). Indeed, none of those cases even contains dictum embracing the court of appeals' position. /10/ And none of those cases or any other case of which we are aware adopted or approved the "potential" litigation test. /11/ The remainder of the court of appeals' analysis is equally unpersuasive. The court stated (Pet. App. 5a) that "(e)xtending the work-product protection only to subsequently related cases best comports with the fact that the privilege is qualified, not absolute." But simply because the privilege is qualified in one respect -- i.e., some privileged materials may be obtained in civil discovery where sufficient need is shown (see Fed. R. Civ. P. 26(b)(3)) -- it does not follow that the privilege must be qualified in every other respect as well. Merely because a particular litigant's need for privileged materials may justify disclosure, it does not follow that the interests served by the privilege lose all force "(w)hen litigation has ended and no potential for related actions exists" (Pet. App. 6a). The court next stated (Pet. App. 6a, quoting Jordan v. U.S. Department of Justice, 591 F.2d 753, 775 (D.C. Cir. 1978); citations omitted; emphasis in original): The purpose of the privilege * * * "is to encourage effective legal representation within the framework of the adversary system by removing counsel's fears that his thoughts and information will be invaded by his adversary * * *" Therefore, in order to fall within the scope of the privilege a document "must 'relate to the conduct of either ongoing or prospective trials * * *.'" /12/ This is faulty logic. As we have argued, a lawyer's fears about the possibility that his files may be disclosed after all related proceedings have terminated might discourage him from amassing materials in preparation for trial or preserving those files after the case has ended. As a result, "effective legal representation within the framework of the adversary system" would be undermined. Finally, the court stated (Pet. App. 6a; emphasis in original): Moreover, we deal in this case, not with the civil discovery situation, but rather with a Freedom of Information Act request. Here, the presumption in favor of disclosure is at its zenith. "(D)isclosure, not secrecy, is the dominant objective of the Act." Dep't of Air Force v. Rose, 425 U.S. 352, 361 * * * (1976). This reasoning is also fallacious. Just because "disclosure, not secrecy, is the dominant objective of (FOIA)," it does not follow, as respondent maintains (see Br. in Opp. 10) and the court below suggested, that greater disclosure should be available in a FOIA suit than in civil discovery. On the contrary, as we will show, such an interpretation of the FOIA would contravene both the plain language and legislative history of Exemption 5 (see NLRB v. Sears, Roebuck & Co., supra, 421 U.S. at 149); FOMC v. Merrill, supra, 443 U.S. at 353) and would produce absurd results. B. Even if work product may be disclosed in civil discovery when the case for which it was prepared has ended and no related case exists or potentially exists, the work product of government attorneys is nevertheless exempt from mandatory disclosure under the FOIA As previously noted, the work-product privilege is qualified in civil discovery. Facts in an attorney's file that are "essential to the preparation of (an adversary's) case" may be discovered upon a showing of necessity. Fed. R. Civ. P. 26(b)(3); Hickman v. Taylor, supra, 329 U.S. at 511. However, "(i)n ordering discovery of such materials when the required showing has been made," courts must "protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney * * * concerning the litigation" (Fed. R. Civ. P. 26(b)(3); see also Notes of Advisory Committee on 1970 Amendment to Rules, 28 U.S.C. App. at 442; Upjohn Co. v. United States, supra, 449 U.S. at 399-401). Because the work-product privilege is qualified, the particular circumstances of the party requesting discovery -- for example, the party's need for the material in preparing his case -- are an important factor in determining whether the privilege will be honored. Moreover, in civil discovery, a court may enter a protective order permitting disclosure of work product only to particular persons and "only on specified terms and conditions" (Fed. R. Civ. P. 26(c)(2)). Under the FOIA, by contrast, such an individualized approach is not possible. The FOIA does not "by its terms, permit inquiry into particularized needs of the individual seeking the information, although such an inquiry would ordinarily be made of a private litigant." EPA v. Mink, 410 U.S. 73, 86 (1973). Instead, the FOIA requires information subject to disclosure to be made "available to any person" (5 U.S.C. 552(a)(3)), and a large percentage of FOIA requests are made by firms and individuals acting in a representative capacity precisely so that the identity of the person, company, or organization seeking the information will not be known. For these reasons, "at best, the discovery rules can only be applied under Exemption 5 by way of rough analogies." EPA v. Mink, supra, 410 U.S. at 86. Accordingly, in NLRB v. Sears, Roebuck & Co., supra. 421 U.S. at 149, this Court held that "it is reasonable to construe Exemption 5 to exempt those documents, and only those documents, normally privileged in the civil discovery context." The Court carefully emphasized (id. at 149 n.16) that "it is not sensible * * * to require disclosure of any document which would be disclosed in the hypothetical litigation in which the private party's claim is the most compelling." Rather, the question is whether the information sought would "'routinely be disclosed' in private litigation" (ibid., quoting H.R. Rep. No. 1497, 89th Cong., 2d Sess. 10 (1966)). The reason for this construction of Exemption 5 is obvious. If, for example, a person making an FOIA request for work-product materials were deemed to stand in the shoes of a hypothetical civil litigant capable of demonstrating a compelling need for disclosure, the qualified work-product privilege recognized in civil discovery would be rendered meaningless in government litigation. Any civil litigant unable or unwilling to make the showing required for discovery of work product under Fed. R. Civ. P. 26(b)(3) could simply file an FOIA request, and he would automatically be treated as if such a showing had been made. In order to prevent such incongruous results, persons filing FOIA requests must be treated like civil litigants with the least compelling claim for disclosure, i.e., litigants incapable of showing any special need for the documents sought. Such persons are accordingly entitled to only those materials that would be available in civil discovery based upon a simple showing of relevance (Fed. R. Civ. P. 26(b)(1)). In other words, they are entitled to only those documents that would "'routinely be disclosed' in private litigation" (NLRB v. Sears, Roebuck & Co., supra, 421 U.S. at 149 n.16, quoting H.R. Rep. No. 1497, supra, at 10). Applying this teaching to the question presented by the instant case yields a result precisely opposite to that reached by the court of appeals. Of course, if the work-product privilege is perpetual in the context of civil discovery, then work product from terminated litigation would not be available under the FOIA regardless of whether there was any potential for related litigation. But even if, as the court below suggested (Pet. App. 4a), "the work-product privilege extends to subsequent cases only when they are related," it would not follow that "in the context of an FOIA request, attorney work-product from terminated litigation remains exempt from disclosure only when litigation related to the terminated action exists or potentially exists" (id. at 7a; emphasis in original). On the contrary, work product from terminated litigation would not be "routinely" available in civil discovery because it would be available only (a) in unrelated cases where there was also no potential for subsequent related cases or (b) where sufficient need could be shown. As noted, in order to preserve the limitations imposed upon the disclosure of work product in civil discovery, a person seeking such material pursuant to an FOIA request must be deemed to stand in the shoes of a hypothetical civil litigant with the least compelling claim for disclosure -- in this instance, a litigant in a related case who is unable to demonstrate special need for the documents. Consequently, even if the work-product privilege is subject to the limitations suggested by the court of appeals, the work product of government attorneys would be protected under Exemption 5 of the FOIA. CONCLUSION The judgment of the court of appeals should be reversed. Respectfully submitted. REX E. LEE Solicitor General J. PAUL McGRATH Assistant Attorney General KENNETH S. GELLER Deputy Solicitor General SAMUEL A. ALITO, JR. Assistant to the Solicitor General LEONARD SCHAITMAN Attorney JANUARY 1983 /1/ In addition to records relating to the Americana investigation ("Category A" records), Grolier also requested records relating to covert investigations of any of its 14 subsidiaries ("Category B" records) and records relating to covert investigations of any person, company, or entity ("Category C" records). However, Grolier's district court complaint was confined to Category A and B records, and Grolier later withdrew any claim for further disclosure of Category B documents. Thus, Grolier's appeal involved only Category A records. See Appellant's Opening Brief at 2 n.1. /2/ The civil penalty action was based upon an alleged violation of a 1949 cease and desist order prohibiting false advertising and misrepresentations in door-to-door sales. The suit was dismissed with prejudice on Nov. 17, 1976, after the Commission declined to comply with a district court order directing it to turn over certain documents to the defendant (Pet. App. 1a-2a). /3/ The FTC withheld two other documents -- Numbers 1 and 2 -- on the grounds that they were work product and predecisional. Court of Appeals Joint Appendix ("C.A. App.") 60-61. The district court upheld the withholding of these documents under the work-product rationale (J.A. 48, 51). Grolier subsequently withdrew its claim for disclosure of document 2 (Pet. App. 2a n.3), and the court of appeals affirmed the district court's judgment regarding document 1 under the pre-decisional rationale (id. at 8a). The Commission also withheld document 4 as an attorney-client communication (J.A. 36-37). Both the district court and the court of appeals upheld the withholding of this document on that basis (J.A. 49, 51; Pet. App. 2a n.3). /4/ See also 110 Cong. Rec. 17667 (1976) (remarks of Sen. Long) ("(T)here is nothing in this bill which would override normal privileges dealing with the work product * * *.") /5/ Some courts have held that this aspect of the privilege is absolute. See, e.g., In re Grand Jury Proceedings, 473 F.2d 840, 848 (8th Cir. 1973). See also Fed. R. Civ. P. 26(b)(3). /6/ Indeed, the perpetual nature of the work-product privilege is an implicit premise of the Court's analysis in NLRB v. Sears, Roebuck & Co., supra. In that case, Sears filed a FOIA request seeking, among other things, intra-agency memoranda of the NLRB directing the filing of a complaint. Many such memoranda pertained to "cases which had been closed * * * because litigation before the Board had been completed" (421 U.S. at 145). However, the Court did not distinguish between open cases and closed cases but rather concluded (id. at 159-160) that all such memoranda had been "prepared in contemplation of the upcoming litigation (and therefore fell) squarely within Exemption 5's protection of an attorney's work product." /7/ For an illustration of some of the ways in which cases may be related, see Rule 13(b)(4) of the Rules of the United States Court of Appeals for the Ninth Circuit, which requires parties to state whether any "related" cases are pending before the Court. Under this rule, cases may be deemed related if, among other things, they "involve the same or closely related issues," "involve the same basic transaction or event," or "have any other similarities of which you think the Court should be aware." /8/ In the present case, the court of appeals concluded (Pet. App. 7a) that "there does not appear to be any suit or potential suit related to the original Americana action." However, the court reached that conclusion by misconstruing and taking out of context a statement made by the FTC in papers supporting its motion for summary judgment. Pursuant to a provision of the local rules and customary local practice, petitioners stated (J.A. 29; emphasis added): Plaintiff Grolier, is also plaintiff in a proceeding to review an order of the Commission, Grolier Incorporated v. Federal Trade Commission, now pending in the U.S. Court of Appeals for the Ninth Circuit (No. 78-2159). The action in the Ninth Circuit is not directly related to this action. The obvious reason for requiring such a statement in papers supporting a summary judgment motion is to allow the court to ascertain whether a consolidation of actions is appropriate, whether any of the claims asserted are controlled by res judicata or collateral estoppel, and whether the decision in the case at hand will affect any other pending litigation. Petitioners' statement obviously was not made with the court of appeals' "related litigation" test in mind. By seizing upon petitioners' statement, which as noted was made in an entirely different context, the court of appeals masked the difficulty of determining whether pending cases, not to mention potential cases, are related within the meaning of the court's test. /9/ It is generally recognized that the policies and needs underlying the attorney work-product privilege are closely allied to those of the attorney-client privilege. McCormick, Evidence Section 96 (2d ed. 1972). It is therefore significant that the attorney-client privilege survives the termination of the attorney-client relation that it is intended to serve (see, e.g., United States v. Foster, 309 F.2d 8, 15 (4th Cir. 1962)) and even the death of the client (see, e.g., In re Busse's Estate, 332 Ill. App. 258, 266, 75 N.E.2d 36, 39 (1947)). /10/ In Republic Gear Co., the Second Circuit declined to permit discovery of work product prepared for related cases then pending on appeal. The court distinguished the case before it from district court cases involving work product from completed litigation (381 F.2d at 557 & n.5), but the court did not state that it would have allowed discovery if the previous litigation had been completed or was unrelated to the case at hand. The three district court cases cited by the court of appeals held that work product from previous related cases is not discoverable. Those cases did not state that work product from previous unrelated cases is discoverable. As the court of appeals observed (see Pet. App. 3a), some district courts have "concluded that the work-product privilege applies only if the materials were prepared in anticipation of the very suit before the court." As previously discussed, however (see pages 18-21, supra), such a rule would lead to absurd results. Moreover, as the five examples cited by the court illustrate (see Pet. App. 3a), the cases adopting such a rule are devoid of analysis and rely uncritically upon prior district court authority. The most recent district court case cited in the opinion below, United States v. IBM, 66 F.R.D. 154, 178 (S.D.N.Y. 1974), relies solely upon a quotation from Honeywell, Inc. v. Piper Aircraft Corp., 50 F.R.D. 117, 119 (M.D. Pa. 1970), and the latter case contains no analysis but merely cites two of the other cases noted by the court of appeals, i.e., Hanover Shoe, Inc. v. United Shoe Machinery Corp., 207 F. Supp. 407, 410 (M.D. Pa. 1962), and Tobacco and Allied Stocks, Inc. v. Transamerica Corp., 16 F.R.D. 534, 537 (D. Del. 1954). Hanover Shoe relies (207 F. Supp. at 410) in turn solely upon Tobacco and Allied Stocks in which the court reached its conclusion (16 F.R.D. at 537) by incorrectly reading Rediker v. Warfield, 11 F.R.D. 125 (S.D.N.Y. 1951). Rediker concerned material that had not been prepared in anticipation of any litigation and thus was not work product at all (id. at 128). The remaining district court decision cited by the court of appeals, Gulf Construction Co. v. St. Joe Paper Co., 24 F.R.D. 411, 415 (S.D. Tex. 1959), relies solely upon the fact that "Hickman v. Taylor involved the discovery and production of statements of witnesses acquired by counsel in anticipation of the litigation then before the court." /11/ The commentators cited by the court of appeals (see Pet. App. 4a) also provide no analytical support for the court's decision. The court relied upon 8 C. Wright & A. Miller, Federal Practice and Procedure Section 2024 at 201 (1970), but that passage merely states without elaboration that the "sounder view" is that work product does not lose all protection at the termination of the litigation for which it was prepared but instead remains privileged in subsequent cases "at least if the two cases are closely related" (emphasis added). In 4 J. Moore, Federal Practice Paragraph 26.64(2) at 26-415 (1982), it is stated that "(p)rior to the 1970 amendments (to Fed. R. Civ. P. 26(b)(3)), the question as to whether the papers in one action were to be considered 'work product' for purposes of a subsequent one, turned on whether the first action was complete and upon the relationship between the first and second actions." This statement was merely an attempt to summarize the holdings of various lower court decisions prior to 1970 and did not take into consideration the decisions of three courts of appeals after 1970 holding that work product is perpetually privileged in civil discovery. Moreover, for reasons previously explained (see pages 31-32 & n.10, supra), the accuracy of that summary is questionable. The final commentator cited by the court of appeals (Cooper, Work Product of the Rulesmakers, 53 Minn. L. Rev. 1269, 1299 n.100 (1969)) merely states in a footnote and without any supporting analysis that the work-product privilege should terminate "only when there is no danger of disclosure to others pursuing claims related to the claims involved in the litigation giving rise to the one-time work product materials." As we have argued, there is almost always some danger that disclosure of the work product of government attorneys will assist potential adversaries or regulatees. The question involved in this case has attracted surprisingly little additional academic comment. One student commentator concluded that cases recognizing a perpetual work-product privilege that may be overcome in cases of demonstrated need maintain "the delicate balance between reasonable inquiry through discovery and the privacy and professional integrity of the adversary system." Comment, Civil Procedures -- Discovery -- Work-Product Privilege Extends to Subsequent Litigation, 27 Vand. L. Rev. 826, 833 (1974). Another student note on the issue (Note, Discovery of an Attorney's Work Product in Subsequent Litigation, 1974 Duke L. J. 799) proposes the following test: "Under the circumstances existing at the time of the preparation of the initial suit, would there have been in the mind of a reasonable attorney a belief that there was a substantial probability of significant subsequent litigation to which his present work product would be relevant" (id. at 820; emphasis in original). This test, however, is subject to many of the same objections as the "related" and "potential" litigation tests. As the student commentator recognized (id. at 821), no attorney could predict with confidence what a court might view as "significant" litigation or a "substantial probability." Moreover, an attorney's trial preparation and representation might well be adversely affected by the knowledge that there was a chance (although something less than a "substantial probability") that such materials might be ordered disclosed. In Note, Developments in the Law -- Discovery, 74 Harv. L. Rev. 940, 1044 (1961), it is stated without analysis or elaboration that the work-product privilege should continue "so long as a likelihood of litigation remains." However, since the disclosure of work product, whether through civil discovery or the FOIA, may only be ordered in subsequent litigation, the import of this statement is unclear. /12/ The quotation from Jordan that work product must "relate to the conduct of either ongoing or prospective trials" (591 F.2d at 775-776) is taken out of context in the panel opinion, and its meaning is distorted. The point the Jordan court was making was that certain documents were not work product because they were not "prepared in anticipation of a particular trial" or "of trials in general" but "were promulgated as general standards to guide the Government lawyers" (id. at 775). The quoted sentence meant that a document cannot be work product unless, when made or compiled, it related to "ongoing or prospective trials" (id. at 775-776). There is no question that the documents at issue here were prepared in contemplation of specific litigation and accordingly were covered by the work-product privilege at least initially.