Opinion for the Court filed by Circuit Judge WALD.
This case raises issues concerning the scope of Exemptions 5 and 7 to the general disclosure requirements of the Freedom of Information Act (FOIA), 5 U.S.C. § 552 (1976). In 1975 and 1976, plaintiff Coastal States Gas Corporation (Coastal States) filed Freedom of Information requests with the defendant,1 seeking copies of agency interpretations of its regulations which had not been made public. The plaintiff’s requests were never processed, but after suit was filed some documents were released. *276The issue in this appeal is focused on memoranda from regional counsel to auditors working in DOE’s field offices, issued in response to requests for interpretations of regulations within the context of particular facts encountered while conducting an audit of a firm. The plaintiff contends that these memoranda constituted a body of “secret law” which the agency was using in its dealings with the public and which must be disclosed, while DOE responds that the documents were properly withheld under Exemption 5,2 as documents which would not be subject to disclosure during discovery,3 and in a few cases, under Exemption 7 4 as documents within an investigatory file. The district court ordered, with a few specific exceptions, that the documents must be released under the FOIA. It rejected each of DOE’s general claims of exemption, finding either that the rationale of the particular exemption did not apply to these documents, or that the agency had failed to demonstrate the prerequisites to proper invocation of the exemption. We agree, and affirm the decision of the district court in all respects.
I. The Facts: In order to determine whether the agency’s claim that the documents were properly withheld is valid, an understanding of the function the documents serve within the agency is crucial. NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 95 S.Ct. 1504, 44 L.Ed.2d 29 (1975). In our explanation of the facts, we draw upon the district court’s findings, which, of course, we are bound to accept unless they are clearly erroneous. We will note, however, those places at which DOE contends the district court has misconstrued the internal functioning of the agency.
After the 1973 oil embargo, a compliance program was established to assure the observance of petroleum pricing and allocation regulations. Ten regional offices were established within which regional counsel were located. Each regional office also employed auditors and other investigative personnel, whose job was auditing individual firms to assure compliance with the regulations. These audits were not “investigations;” at that point, no charge had been made nor was a violation necessarily suspected. According to the regional director of Region VI in Dallas, the auditor “begins the audit without any preconceived notion that there’s a violation at that firm. He is auditing for compliance.” Deposition of Larry White, Regional Director for Compliance, Region VI, at 22 (hereinafter White Dep.).
While the regional counsel has many responsibilities, the particular task relevant to this case is that of providing interpretations of the pertinent regulations to the auditors at this early stage of compliance review. If the auditors should encounter a problem of regulatory interpretation, a request for ad*277vice would be sent to the regional counsel, couched in a specific factual context, either real or hypothetical.5 The response would be a legal memorandum, interpreting any applicable regulations in light of those facts, and often pointing out additional factors which might make a difference in the application of the regulation. We set out, as an example, one of the fourteen documents submitted by the agency as “typical” of the memoranda at issue in this case.6
The agency points out that these were not “formal” interpretations of the regulations, emphasizing that there is a published procedure for issuing such interpretations. Also, the agency insists that the interpretations were not “binding” on the audit staff; it contends that the agency staff “is free to reject the memorandum.”7 The district, *278court found, however, that in fact the advice was regularly and consistently followed by the non-legal staff, a conclusion which we find to be fully supported by the evidence.8 There is evidence in the record that agency staff failed to follow a regional counsel opinion only if it could be distinguished on the facts, or if the matter were referred to a higher authority within the agency.9 Furthermore, in some of the offices the documents were indexed by subject matter and used as precedent in later cases;10 they were circulated among the area offices and supplied to new personnel; 11 they were at times “amended” or “rescinded,” which would hardly be necessary if the documents contained merely informal suggestions to staff which could be disregarded;12 and on at least one occasion a regional counsel memorandum involving the audit of a different firm was cited to a member of the public as binding prece-. dent.13 The fourteen documents which are a part of the record are brief memoranda which explain the meaning of a particular regulation when applied to certain facts.
II. The District Court Opinion: Although the district court rejected each of DOE’s general claims of exemption, the court did find that some documents were properly withheld. Those documents labeled as “drafts, proposals and recommendations” in the agency’s index of documents were found to be deliberative documents within the scope of Exemption 5. A few documents were found to be protected by the attorney work-product privilege because the index revealed they were drafted at a time in the audit when litigation was likely because specific potential violations had been revealed. However, the agency’s claim of attorney-client privilege was rejected as to all the documents because the agency had failed to establish that these documents had been treated with any measure of confidentiality within the agency.
As for the Government’s Exemption 7 claim, the district court found that while it was asserted as to fifty-three documents, no attempt was made to indicate the present status of any investigation involving any of the documents. Since Exemption 7 only applies to concrete prospective or presently active investigations,14 the district court ruled that the agency had failed to establish its entitlement to the exemption, but nevertheless — as a precaution — permitted very *279recent documents as to which Exemption 7 was claimed to be withheld.
On appeal, DOE contends that the district court was in error as to all documents which were ordered to be released. All of the documents, it argues, were properly withheld because they were “pre-decisional” or deliberative intra-agency memoranda, protected by the attorney-client or the attorney work-product privileges, or constitute part of an investigative file, and their disclosure would cause harm to compliance investigations. Coastal States does not challenge the district court decision to permit the continued withholding of the few documents which the district court found were protected by the exemptions.
III. The Vaughn Index: In lieu of in camera inspection, DOE submitted an index of the withheld documents, along with affidavits from regional counsel in support of its decision not to release the memoranda. The parties have referred to these materials as the Government’s “Vaughn Index,” but we wish to make clear that this index is not what we had in mind in our decision in Vaughn v. Rosen, 157 U.S.App.D.C. 340, 484 F.2d 820 (1973), cert. denied, 415 U.S. 977, 94 S.Ct. 1564, 39 L.Ed.2d 873 (1974) (Vaughn I), in which we set out suggested procedures to allow the courts to determine the validity of the Government’s claims without physically examining each document. We repeat, once again, that conclusory assertions of privilege will not suffice to carry the Government’s burden of proof in defending FOIA cases. A typical line from the index supplied in this case identifies who wrote the memorandum, to whom it was addressed, its date, and a brief description of the memorandum such as “Advice on audit of reseller whether product costs can include imported freight charges, discounts, or rental fees. Sections 212.93 and 212.92.” DOE claimed this document was “PD” (predecisional), “ATWP” (attorney work-product) and that “some” of it was in an investigatory file. That is all we are told, save for the affidavits submitted by the regional counsel which repeat in conclusory terms that all the documents withheld fall within one or another of the exemptions.
Such an index is patently inadequate to permit a court to decide whether the exemption was properly claimed, as will be discussed more fully in the course of this opinion. Contrast the index submitted by the agency and described in Mead Data Central, Inc. v. U. S. Dep’t of the Air Force, 184 U.S.App.D.C. 350, 566 F.2d 242 (1977), which clearly describes the characteristics of the documents which the agency felt brought them within the exemption claimed, and which was still inadequate to permit the court to determine whether all elements of the privileges were present in each document. After describing what it expected from a Vaughn index in that opinion, the court acknowledged that the provision of adequate justification for withholding could be a substantial burden on an agency:
Certainly these procedures add significantly to the resource costs an agency must bear if it chooses not to disclosure [sic] material it has in good faith decided is exempt. Those burdens may be avoided at the option of the agency, however, by immediate disclosure. Congress has encouraged the agencies to disclose exempt material for which there is no compelling reason for withholding, and an agency’s own balancing of the resource costs of justifying nondisclosure against the value of secrecy may provide a rough estimate of how compelling is its reason for withholding.
Id., 184 U.S.App.D.C. at 369, 566 F.2d at 261 (footnote omitted).
At several points in the course of this opinion we will rely on a conclusion not that the documents are not exempt as a matter of law, but that the agency has failed to supply us with even the minimal information necessary to make a determination. We remind the agencies, once again, that the burden is on them to establish their right to withhold information from the public and they must supply the courts with sufficient information to allow us to make a reasoned determination that they were correct.
*280IV. Exemption 5 to the FOIA: The clear purpose of the FOIA is to assure that the public has access to all government documents, subject to only nine specific limitations, to be narrowly interpreted, which Congress decided were necessary to protect our national interests and permit the efficient operation of the government. Only two of these limitations are relevant here; Exemption 5, which permits the withholding of documents which would be protected in the course of discovery, and Exemption 7(A), which permits the temporary protection of materials in investigatory files, to prevent the premature disclosure of the government’s case and possible disruption of adversary proceedings.
The language of Exemption 5 is cast in terms of discovery law; the agencies need turn over no documents “which would not be available by law to a private party in litigation with the agency.” This discovery standard can only serve as a “rough guide” to the courts, EPA v. Mink, 410 U.S. 73, 86, 93 S.Ct. 827, 835, 35 L.Ed.2d 119 (1973), since decisions as to discovery are usually based on a balancing of the relative need of the parties, and standards vary according to the kind of litigation involved. Furthermore, the most fundamental discovery and evidentiary principle, relevance to the issues being litigated, plays no part in FOIA cases. It is clear, however, that Congress intended that agencies should not lose the protection traditionally afforded through the evidentiary privileges simply because of the passage of the FOIA. The courts have recognized that Exemption 5 protects, as a general rule, materials which would be protected under the attorney-client privilege, Mead Data Central, 184 U.S.App.D.C. at 360-363, 566 F.2d at 252-255; the attorney work-product privilege, NLRB v. Sears, 421 U.S. at 154, 95 S.Ct. at 1518, Bristol-Myers Co. v. FTC, 194 U.S.App.D.C. 99, 598 F.2d 18 (1978); or the executive “deliberative process” privilege, EPA v. Mink, 410 U.S. at 85-90, 93 S.Ct. at 835-837, Vaughn v. Rosen, 173 U.S.App.D.C. 187, 523 F.2d 1136 (1975) (Vaughn II). DOE argues that all three are applicable in this case.
a. The Attorney-Client Privilege : The familiar attorney-client privilege is the oldest of the evidentiary privileges. 8 J. Wigmore, Evidence § 2290 at 542 (McNaughton rev. 1961). Its purpose is to assure that a client’s confidences to his or her attorney will be protected, and therefore encourage clients to be as open and honest as possible with attorneys. Uninhibited confidence in the inviolability of the relationship is viewed as essential to the protection of a client’s legal rights, and to the proper functioning of the adversary process. Like other privileges which protect communications within a particular relationship, the privilege reflects society’s judgment that promotion of trust and honesty within the relationship is more important than the burden placed on the discovery of truth.15 The privilege is not limited to communications made in the context of litigation or even a specific dispute, but extends to all situations in which an attorney’s counsel is sought on a legal matter. While its purpose is to protect a client's disclosures to an attorney, the federal courts extend the privilege also to an attorney’s written communications to a client, to ensure against inadvertent disclosure, either directly or by implication, of information which the client has previously confided to the attorney’s trust.16
Like all privileges, however, the attorney-client privilege is narrowly construed and is limited to those situations in which its purposes will be served. The Supreme Court has stated that the privilege “protects only those disclosures — necessary to obtain informed legal advice — which might not have been made absent the privilege.” Fisher v. United States, 425 U.S. *281391, 403, 96 S.Ct. 1569, 1577, 48 L.Ed.2d 39 (1976). We have difficulty in perceiving any purpose which would be served by applying the attorney-client privilege in this case. While it is clear that an agency can be a “client” and agency lawyers can function as “attorneys” within the relationship contemplated by the privilege, this does not seem to be such a case. It is hard to imagine the “confidential information” which an auditor might have communicated to the regional counsel. The factual situations the auditor communicates to the attorneys are encountered in the course of auditing third parties, the companies. They do not contain private information concerning the agency. Rather than “counseling,” intended to assist the agency in protecting its interests, the memoranda here seem to be neutral, objective analyses of agency regulations. They resemble, in fact, question and answer guidelines which might be found in an agency manual. In sharp contrast are the documents and memoranda in issue in Mead Data Central, 184 U.S.App. D.C. 350, 566 F.2d 242, in which disclosure was sought of material generated in the course of negotiating a contract between the Air Force and a private company. In such a ease, the Government is dealing with its attorneys as would any private party seeking advice to protect personal interests, and needs the same assurance of confidentiality so it will not be deterred from full and frank communications with its counselors. This case bears little resemblance to that situation.
Assuming, however, that the purposes of the attorney-client privilege might be served by extending its protection to the situation here, we agree with the district court that DOE has failed to demonstrate a fundamental prerequisite to assertion of the privilege: confidentiality both at the time of the communication and maintained since. The burden is on the agency to demonstrate that confidentiality was expected in the handling of these communications, and that it was reasonably careful to keep this confidential information protected from general disclosure. Not only has the DOE failed to affirmatively establish confidentiality, but the evidence shows no attempt whatsoever to protect these memoranda within the agency. The agency has admitted that it does not know who has had access to the documents,17 and there is undisputed testimony that at least in some regions, copies of the memoranda were circulated to all area offices,18 filed and indexed for future use,19 relied on as precedent and used as training materials for new personnel.20
DOE argues that circulation limited to the confines of the agency of a document otherwise entitled to protection under the attorney-client privilege should not defeat the privilege, but that would be far too broad a grant of privilege. When the client is by nature a group, as is true of both the government and corporations, the courts have agreed that the privilege should not be defeated by some limited circulation beyond the attorney and the person within the group who requested the advice. The test, as this court held in Mead Data Central, is whether the agency is able to demonstrate that the documents, and therefore the confidential information contained therein, were circulated no further than among those members “of the organization who are authorized to speak or act for the organization in relation to the subject matter of the communication.” 184 U.S.App.D.C. at 361 n. 24, 566 F.2d at 253 n. 24. The purpose of the privilege is limited to protection of confidential facts. If facts have been made known to persons other than those who need to know them, there is nothing on which to base a conclusion that they are confidential.
If DOE were able to establish that some attempt had been made to limit disclosure of the documents to the agency personnel *282responsible for the audit under discussion in the memorandum, we would have a different case. One significant measure of confidentiality is the degree of care exhibited in the handling of the documents; we can find nothing in this record to indicate that the person requesting advice from the regional counsel had any expectation of confidentiality whatsoever.
We must therefore agree with the district court that DOE has failed to carry its burden of establishing that these documents should be granted protection under the attorney-client privilege.
b. Attorney Work-Product: Another traditional area of privilege which has been recognized under Exemption 5 is attorney work-product. This doctrine stands in contrast to the attorney-client privilege; rather than protecting confidential communications from the client, it provides a working attorney with a “zone of privacy” within which to think, plan, weigh facts and evidence, candidly evaluate a client’s case, and prepare legal theories. There is one significant limitation of the doctrine, however, which defeats the agency’s claim of privilege here; it has uniformly been held to be limited to documents prepared in contemplation of litigation. See, Jordan v. U. S. Dep’t of Justice, 192 U.S.App.D.C. 144, 591 F.2d 753 (1978) (en banc); Fed.R.Civ.P. 26(b)(3). Beyond the few documents which the trial court held could be properly withheld pursuant to this privilege, the DOE has failed to establish that any of the remaining documents were prepared in apprehension of litigation.
The principles of the work-product privilege have been developed in federal courts from the landmark decision in Hickman v. Taylor, 329 U.S. 495, 510-11, 67 S.Ct. 385, 393, 91 L.Ed. 451 (1947), which held that “it is essential that a lawyer work with a certain degree of privacy, free from unnecessary intrusion by opposing parties and their counsel.” The purpose of the privilege, however, is not to protect any interest of the attorney, who is no more entitled to privacy or protection than any other person, but to protect the adversary trial process itself. It is believed that the integrity of our system would suffer if adversaries were entitled to probe each other’s thoughts and plans concerning the case. Certainly less work-product would be committed to paper, which might harm the quality of trial preparation. In any case, the scope of the protection is limited, as Judge Wilkey made clear in Jordan :
The work-product rule does not extend to every written document generated by an attorney; it does not shield from disclosure everything that a lawyer does. Its purpose is more narrow, its reach more modest. . . . [T]he 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. In other words, the privilege focuses on the integrity of the adversary trial process itself . . . . This focus on the integrity of the trial process is reflected in the specific limitation of the privilege to materials “prepared in anticipation of litigation or for trial.”
192 U.S.App.D.C. at 166, 591 F.2d at 775 (emphasis added). The same limitation is reflected in the discovery rule in Fed.R. Civ.P. 26(b)(3), which provides:
Trial Preparation: Materials.
[A] party may obtain discovery of documents and tangible things . . . prepared in anticipation of litigation or for trial by or for another party or by or for that other party’s representative (including his attorney, consultant, surety, indemnitor, insurer, or agent) 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. In ordering discovery of such materials when the required showing has been made, the court-shall protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation.
*283(emphasis added). There is still some dispute among the courts as to the limits of the privilege, for instance, whether the protection afforded by the privilege lapses once the litigation has ended or the prospects of litigation have faded, but it is firmly established that there is no privilege at all unless the document was initially prepared in contemplation of litigation, or in the course of preparing for trial. Jordan, 192 U.S.App.D.C. at 166, 591 F.2d at 775. Of course, the courts will not penalize litigants for doing initial preparation before filing a complaint, but we agree with the district court that at the very least some articulable claim, likely to lead to litigation, must have arisen. To the extent the Government provided some indication in its index that a specific claim had taken shape in the course of an audit, so that the attorney’s work could fairly, if generously, be characterized as “in contemplation of litigation,” the district court permitted these documents to be withheld.21 Beyond that,' the DOE has failed to carry its burden of establishing that litigation was fairly foreseeable at the time the memoranda were prepared, and thus is not entitled to invoke the exception.
DOE relies on several cases to support its claim of privilege, each of which is distinguishable because in each case, at the very least, a specific charge or allegation was under investigation. Even under the broad scope of the attorney work-product privilege described in Kent Corp. v. NLRB, 530 F.2d 612 (5th Cir.), cert. denied, 429 U.S. 920, 97 S.Ct. 316, 50 L.Ed.2d 287 (1976), which held that investigatory reports prepared very early in the course of the agency’s involvement in the case, before there had been any determination that the charges had substance, were protected by the privilege, we note that the Fifth Circuit relied on the fact that “the prospect of litigation [was] identifiable because of specific claims that [had] already arisen.” Id. at 623. The NLRB investigations involved in Kent only occurred after specific charges of unfair labor practices had been filed. We need not here explore the outer boundaries of the privilege; with the exception of the few documents held to be privileged by the district court, there is no indication in the index or affidavits that there was even the dimmest expectation of litigation-when these documents were drafted. The mere fact that many of the memoranda deal with specific factual situations is not sufficient; if an agency were entitled to withhold any document prepared by any person in the Government with a law degree simply because litigation might someday occur, the policies of the FOIA would be largely defeated. To argue that every audit is potentially the subject of litigation is to go too far. While abstractly true, the mere possibility is hardly tangible enough to support so broad a claim of privilege. We need not decide here whether litigation need be consciously contemplated by the attorney; the documents must at least have been prepared with a specific claim supported by concrete facts which would likely lead to litigation in mind, and that has not been demonstrated here.
We note that among the fourteen documents submitted by DOE as “representative” there are two which might, if we accepted the broad Kent test, be characterized as in “contemplation of litigation.” In one document the auditor had already determined that an illegal price had been charged by a company which had recently been sold, and advice was sought as to whether the former shareholders could be liable for the overcharges. In the other, a similar transfer of operations had occurred, and the regional counsel opinion explored the question of whether the agency should seek reimbursement from the former sole proprietor or the new corporation. If these documents had been adequately described *284at an earlier point we would have to face the question whether protection of work produced at such an early stage in the administrative process would sufficiently further the goals of the work-product privilege to justify withholding them from the public. But in pursuing its broader claim that all of these documents represent attorney work-product, DOE neglected to supply the court with sufficient facts, either in its index or its submitted affidavits, to permit a conclusion that in fact specific claims had arisen and were likely to be pursued to the point of litigation by the agency. For instance, DOE indexes the second document described above simply as a “detailed discussion of facts and applicability of § 212.111(a)(2) to a specific case.” At the very least, the agency must establish in its affidavits or indexes the fact that a specific claim had arisen, was disputed by the company, and was being discussed in the memorandum; DOE cannot expect the district court to simply assume the fundamental prerequisites which are its burden to establish.
c. The Deliberative Process Privilege : A privilege unique to the government is one which is variously described as predecisional or deliberative process privilege. The privilege has a number of purposes: it serves to assure that subordinates within an agency will feel free to provide the decisionmaker with their uninhibited opinions and recommendations without fear of later being subject to public ridicule or criticism; to protect against premature disclosure of proposed policies before they have been finally formulated or adopted; and to protect against confusing the issues and misleading the public by dissemination of documents suggesting reasons and rationales for a course of action which were not in fact the ultimate reasons for the agency’s action. See Jordan, 192 U.S.App.D.C. at 163-165, 591 F.2d at 772-774.
DOE argues that these memoranda were “pre-decisional” because the regional counsel did not have final decisionmaking authority over interpretation of the regulations, and they were “deliberative” because they were an early part of the enforcement process, subject to continuing debate within the agency as the investigation continued. Coastal States, on the other hand, argues that the documents display none of the give-and-take of the decisionmaking process, are not recommendatory in nature, and in any case were utilized within the agency as an informal, functioning body of “secret law.”
In deciding whether a document should be protected by the privilege we look to whether the document is “predecisional” —whether it was generated before the adoption of an agency policy — and whether the document is “deliberative” — whether it reflects the give-and-take of the consultative process. The exemption thus covers recommendations, draft documents, proposals, suggestions, and other subjective documents which reflect the personal opinions of the writer rather than the policy of the agency. Documents which are protected by the privilege are those which would inaccurately reflect or prematurely disclose the views of the agency, suggesting as agency position that which is as yet only a personal position. To test whether disclosure of a document is likely to adversely affect the purposes of the privilege, courts ask themselves whether the document is so candid or personal in nature that public disclosure is likely in the future to stifle honest and frank communication within the agency; “Human experience teaches that those who expect public dissemination of their remarks may well temper candor with a concern for appearances and for their own interests to the detriment of the decision-making process.” United States v. Nixon, 418 U.S. 683, 705, 94 S.Ct. 3090, 3106, 41 L.Ed.2d 1039 (1974). We also ask whether the document is recommendatory in nature or is a draft of what will become a final document, and whether the document is deliberative in nature, weighing the pros and cons of agency adoption of one viewpoint or another. Finally, even if the document is predecisional at the time it is prepared, it can lose that status if it is adopted, formally or informally, as the agency position on an issue or is used by the agency in its dealings with the. public.
*285The cases in this area are of limited help to us, because the deliberative process privilege is so dependent upon the individual document and the role it plays in the administrative process. The Supreme Court has considered the privilege on several occasions. EPA v. Mink, 410 U.S. 73, 93 S.Ct. 827, 35 L.Ed.2d 119, dealt with clearly advisory reports prepared for the President, and established the principle that the privilege applies only to the “opinion” or “re-commendatory” portion of the report, not to factual information which is contained in the document.
In NLRB v. Sears, 421 U.S. 132, 95 S.Ct. 1504, 44 L.Ed.2d 29, the plaintiff sought documents prepared by the regional counsel for the Labor Board which either directed the dismissal of a complaint (and terminated the consideration of the case) or directed the filing of a complaint (which formally commenced litigation). Under NLRB procedure, the former action was an unreviewable final disposition of the case, and thus the Supreme Court held that it could not be considered “predecisional.” The latter, however, merely initiated a formal adjudicative procedure, inevitably resulting in a final disposition which would more accurately reflect the agency’s views. The Court recognized that the distinction between the two documents based on their conclusion was not a “bright line,” but was bolstered in its decision that the memoranda directing the filing of a complaint were properly withheld by the reduced public interest in disclosure, since there would inevitably be a “final” opinion which would set out the agency’s policy and law, and by the fact that the attorney work-product privilege was implicated, since those documents began litigation and were prepared by the same attorneys who would be acting as advocates for the agency position. A crucial distinction between Sears and the present case is that in Sears the documents which were properly withheld initiated litigation before the Board; they were not intended to guide and direct subordinates in analogous cases. That is, because they were not final agency action, they were not viewed as having precedential import and were not intended to have effect upon actions of others in the agency. See 421 U.S. at 156-57, 95 S.Ct. at 1519-20.
Renegotiation Bd. v. Grumman Aircraft, 421 U.S. 168, 95 S.Ct. 1491, 44 L.Ed.2d 57 (1975), also involved documents which would inevitably be superceded by later formal agency action. In that case, regional recommendations were prepared, suggesting a disposition of a particular case. In every case, these recommendations were reviewed by the Board and some final action on the case was taken by the Board. The documents at issue in Grumman were transmitted to superior authority for review and final action; they were merely recommendations with no precedential significance. It was only the later decision of the Board that had decisional significance and would guide and direct future conduct of subordinates in analogous cases. See 421 U.S. at 185-86, 95 S.Ct. at 1500-01. For our present decision it is very instructive to noté that the Grumman Court distinguished the facts before it from the situation of a United States District Court, whose decision has real operative effect absent appeal by a party. See id. at 186-87, 95 S.Ct. at 1501-1502. In the present case, by contrast, the memoranda from regional counsel are analogous to trial court decisions in that the district court here found that they have operative and controlling effect over auditors unless the matter was referred to a higher authority in the Department.
This court has also considered the deliberative process privilege on many occasions. A strong theme of our opinions has been that an agency will not be permitted to develop a body of “secret law,” used by it in the discharge of its regulatory duties and in its dealings with the public, but hidden behind a veil of privilege because it is not designated as “formal,” “binding,” or “final.” The theme was sounded as early as 1971 when the court emphatically stated that agencies would be required to disclose “orders and interpretations which it actually applies to cases before it,” in order to prevent the development of “secret law.” Sterling Drug, Inc. v. FTC, 146 U.S.App.*286D.C. 237, 247, 450 F.2d 698, 708 (1971). The refrain was repeated in Schwartz v. IRS, 167 U.S.App.D.C. 301, 303, 511 F.2d 1303, 1305 (1975), and Ash Grove Cement Co. v. FTC, 167 U.S.App.D.C. 249, 252, 511 F.2d 815, 818 (1975). The court in Sterling Drug explained the reasoning which underlies this line of cases:
[T]he policy of promoting the free flow of ideas within the agency does not apply here, for private transmittals of binding agency opinions and interpretations should not be encouraged. These are not the ideas and theories which go into the making of the law, they are the law itself, and as such should be made available to the public. Thus, to prevent the development of secret law within the Commission, we must require it to disclose orders and interpretations which it actually applies in cases before it.
146 U.S.App.D.C. at 247, 450 F.2d at 708.
It is also clear that the agency has the burden of establishing what deliberative process is involved, and the role played by the documents in issue in the course of that process. Vaughn II, 173 U.S.App.D.C. at 197, 523 F.2d at 1146. In that case, Judge Wilkey pointed out that if documents are not a part of a clear “process” leading to a final decision on the issue, as they were in both the Sears and the Grumman cases, they are less likely to be properly characterized as predecisional; in such a case there is an additional burden on the agency to substantiate its claim of privilege. The identity of the parties to the memorandum is important; a document from a subordinate to a superior official is more likely to be predecisional, while a document moving in the opposite direction is more likely to contain instructions to staff explaining the reasons for a decision already made. For instance, this court recently identified as “a classic case of the deliberative process at work” a series of memoranda to the Assistant Secretary of the Army from the General Counsel in his department, recommending legal strategy in light of a particular controversy. Murphy v. Dep’t of the Army, 198 U.S.App.D.C. 418 at 421, 613 F.2d 1151 at 1154 (1979).
Applying these principles and precedents to this case, it is readily apparent that the memoranda in issue bear little resemblance to the types of documents intended to be protected under the deliberative process privilege. The documents were not suggestions or recommendations as to what agency policy should be. Unlike the documents in EPA v. Mink and Murphy v. Dep't of the Army, the memoranda are not advice to a superior, nor are they suggested dispositions of a case, as in Grumman. They are not one step of an established adjudicatory process, which would result in a formal opinion, as were the documents held exempt in NLRB v. Sears. There is nothing subjective or personal about the memoranda; they are simply straightforward explanations of agency regulations in specific factual situations. They are more akin to a “resource” opinion about the applicability of existing policy to a certain state of facts, like examples in a manual, to be contrasted to a factual or strategic advice giving opinion. Nor do they reflect “agency give-and-take — of the deliberative process — by which the decision itself is made.” Vaughn II, 173 U.S.App.D.C. at 195, 523 F.2d at 1144. Characterizing these documents as “predecisional” simply because they play into an ongoing audit process would be a serious warping of the meaning of the word. No “decision” is being made or “policy” being considered; rather the documents discuss established policies and decisions — the agency regulations — in the light of a specific, and often hypothetical, fact pattern.
We reemphasize the narrow scope of Exemption 5 and the strong policy of the FOIA that the public is entitled to know what its government is doing and why. The exemption is to be applied “as narrowly as consistent with efficient Government operation.” S.Rep.No. 813, 89th Cong., 1st Sess. 9 (1965).' We do not believe that public knowledge of the contents of these memoranda would affect either “efficient Government operation” or any one of the various policies to be served by the Exemp*287tion. The documents do not contain subjective, personal thoughts on a subject, so public knowledge of the documents will not subject the writer either to ridicule or criticism. Nor do they discuss the wisdom or merits of a particular agency policy, or recommend new agency policy, raising the possibility that their disclosure would mislead the public; rather, they simply explain and apply established policy. DOE asserts that its attorneys will be less “candid” in the future if these memoranda are disclosed, but we are unable to find in any of the fourteen documents any statement which could be described as “candid.” We can see no possibility whatsoever that an attorney performing this job would be less “frank” or “honest” if he or she knew that the document might be made known to the public; there is little to be frank or honest about when explaining on what date' a transaction occurs under 10 C.F.R. § 212.31 or whether 10 C.F.R. § 212.10 permits a buyer and seller to agree to a price higher than that set by the agency. Nor does the general description of the remaining documents in the agency’s index suggest that any “candor” is likely to be found in these legal interpretations.
We are unable to conclude, either from the DOE’s index or the fourteen documents in the record, that these regional counsel opinions possess any of the characteristics normally displayed by memoranda protected by the deliberative process privilege. We have already pointed out that the opinions are not “predecisional” simply because they are generated early in the audit period. DOE’s contention that these documents are not “final opinions,” absolutely binding on the auditors, misses the point. The evidence strongly supports the district court’s conclusion that, in fact, these opinions were routinely used by agency staff as guidance in conducting their audits, and were retained and referred to as precedent. If this occurs, the agency has promulgated a body of secret law which it is actually applying in its dealings with the public but which it is attempting to protect behind a label. This we will not permit the agency to do. Tentative opinions are not relied on as precedent; they are considered further by the decisionmaker. Suggestions which could be freely disregarded would not need to be rescinded, amended, or referred to a higher authority. These documents, whatever the formal powers of regional counsel to issue binding interpretations of the regulations, in practice represent interpretations of established policy on which the agency relies in discharging its regulatory responsibilities; withholding them would serve no legitimate policy interest of the government.22
V. Exemption 7 to the FOIA : Exemption 7 affords protection to investigatory files to prevent “harm [to] the govern*288ment’s ease in court.” NLRB v. Robbins Tire, 437 U.S. at 224, 98 S.Ct. at 2318, quoting from S.Rep.No. 813, 89th Cong., 1st Sess. (1965). In 1974, the scope of the privilege was sharply narrowed when Congress, dissatisfied with the broad scope given to Exemption 7 by the courts, amended the exemption to make it clear that the Government must establish not only that the document was prepared in the course of an “investigation,” but that disclosure of the document would “interfere with enforcement proceedings.” 5 U.S.C. § 552(b)(7)(A). The Government asserts this privilege as to fifty-three documents.
We need not decide whether any of these memoranda are of a sort which would be protected if the Government had demonstrated — or even conclusorily asserted— that there are presently active investigations underway or contemplated in each of these fifty-three cases. Robbins Tire, 437 U.S. at 230-232, 98 S.Ct. at 2321-2322, points out that a major reason for the 1974 amendments was to eliminate the exemption when there is no enforcement proceeding then pending or contemplated. As Senator Hart, the sponsor of the Exemption 7 amendment, explained:
Let me clarify the instances in which nondisclosure would obtain: first, where the production of a record would interfere with enforcement procedures. This would apply whenever the Government’s case in court — a concrete prospective law enforcement proceeding — would be harmed by the premature release of evidence or information not in the possession of known or potential defendants. This would apply also where the agency could show that the disclosure of the information would substantially harm such proceedings by impeding any necessary investigation before the proceedings. In determining whether or not the information to be released will interfere with a law enforcement proceeding it is only relevant to make such determination in the context of the particular enforcement proceeding.
120 Cong.Rec. 17033 (1974) (emphasis added). There is no reason to protect yellowing documents contained in long-closed files. DOE made no effort whatsoever in the district court to demonstrate that any of these cases are still under investigation or being actively pursued. The district court was correct in concluding that DOE had failed generally to meet its burden of establishing the prerequisites to invocation of Exemption 7.
In any case, it appears highly unlikely that these documents would “tip the Government’s hand” as to the Government’s evidence or approach in an investigation, since the regulatory scheme at DOE is one of discussion and consultation with the company in the course of an audit in an attempt to achieve voluntary compliance or a consent agreement; only if no agreement can be reached does a Notice of Probable Violation issue.23 Since the regional counsel opinions were typically issued in the course of compliance review,24 it is unlikely they would contain anything which was not already known to and discussed with the company being audited. We note this merely to point out that DOE has doubly failed to meet its burden in order to invoke the “investigatory file” exemption. The courts will not speculate as to whether Exemption 7 might, under some possible congruence of circumstances — not proven or even asserted — be properly applied to these documents, nor will we assume that all the necessary conditions are met merely because the agency invokes an exemption.
VI. Conclusion : We agree with the district court that the defendant DOE has failed to carry its burden of establishing that the documents involved in this appeal were properly withheld pursuant to Exemptions 5 or 7 of the FOIA. The decision of the district court ordering release of the documents is therefore affirmed. In view *289of our disposition of this case, the stay previously ordered by this court is vacated.