¶1Opinion for the Court filed by Circuit Judge WILKEY.
¶2This is a suit under the Freedom of Information Act (FOIA) seeking to compel disclosure of “Evaluation of Personnel Management” reports, as well as reports and studies of a similar nature, prepared by the Bureau of Personnel Management of the Civil Service Commission (the “Commission”). After examining a representative sample of nine documents, the District Court granted plaintiff’s motion for summary judgment, and thus ruled disclosable those portions of the reports which the Government claimed were exempt under Exemption 2
¶3For the reasons set out, we affirm the District Court’s disposition.
¶4I. Facts and Procedural Background
¶5Plaintiff, a law professor doing research into the Civil Service Commission, seeks disclosure of reports prepared by the Commission’s Bureau of Personnel Management during the fiscal years 1969 through 1972. These reports, denominated “Evaluations of Personnel Management,” are “the Commission’s evaluation of the way the agencies’ managers and supervisors are carrying out their personnel management responsibilities . .”
¶6Plaintiff’s request for disclosure was denied by the Director of the Bureau of Personnel Management Evaluation on 15 June 1972.
¶7On remand the Government, with plaintiff’s acquiescence, filed nine representative reports from which identifying details such as agency names were deleted. A table accompanied each report *191correlating the page, paragraph, or sentence for which a claim of exemption is being made with the exemption being relied upon. The parties have agreed that the nine sample reports are representative of the documents sought and that any decision as to those documents will be applicable to all of the documents
¶8After examining the sample reports, the' District Court ruled
¶9Those portions of the sample reports found not to be exempt from disclosure cover a wide range of topics: labor-management relations, position classification, equal employment opportunity, the merit promotion program, processing of personnel actions, incentive awards and the employee suggestion program, management’s evaluation of employee performance, employment of Vietnam era veterans, employee training, manpower planning, employment of handicapped individuals, recruitment efforts, and implementation of reductions in force.
¶10Each of the reports concludes with a series of recommendations made by the evaluating team. These recommendations are variously labelled “Action Items,” “Recommendations,” or begin simply with the words “we recommend” or “we suggest.” The District Court held those to be clearly recognizable, easily severable from the narrative portions of the reports, and exempt from disclosure.
¶11Although we follow a slightly different reasoning process from that of the District Court, we affirm the results reached.
¶12II. Exemption 2
¶13Exemption 2 provides that agency records “related solely to the internal personnel rules and practices of an agency” need not be disclosed under the Freedom of Information Act. Despite the seeming clarity of the exemption, the Senate and House Reports recommending passage of FOIA express conflicting views as to the scope of the exemption. The Senate Report stated:
Exemption No. 2 relates only to the internal personnel rules and practices of an agency. Examples of these may be rules as to personnel’s use of park*192ing facilities or regulation of lunch hours, statements of policy as to sick leave, and the like.11
¶14The House Report, on the other hand, declared:
2. Matters related solely to the internal personnel rules and practices of any agency: Operating rules, guidelines, and manuals of procedure for Government investigators or examiners would be exempt from disclosure, but this exemption would not cover all “matters of internal management” such as employee relations and working conditions and routine administrative procedures which are withheld under the present law.12
¶15Virtually all of the courts that have considered the conflict between the two reports have held that the Senate Report most accurately reflects the intent of Congress.
¶16First, it must be remembered that pri- or to the enactment of FOIA, the Administrative Procedure Act contained a public information section, Section 3.
Federal agencies may limit the dissemination of a wide range of information that they deem related “solely to the internal management” of the agency. What are the limitations, if any, that are attached to this provision?18
¶17In a similar vein, Congressman Moss of California, the Chairman of the House Subcommittee that considered the bill, pointed specifically to the phrase “internal management” and declared that the purpose of the bill was to replace the old statute with “workable standards . [containing] specific definitions of information which may be withheld.”
¶18*193Despite the similar wording of Exemp-tion 2, we refuse to believe that it was intended merely to reenact the old exemption. Congress that Exemption 2 be interpreted narrowly and specifically. In our view, the House Report carries the potential of exempting a wide swath of information under the category of “operating rules, guidelines, and manuals of procedure .”
¶19This is a standard, a guide, which an agency and then a court, if need be, can apply with some certainty, consistency and clarity. If we be wrong in finding this line of demarcation drawn by doubtless Congress will correct it&emdash;and hopefully this time, with both Houses in accord.
¶20Reinforcing this interpretation is "the clear legislative intent [of FOIA] to assure public access to all governmental records whose disclosure would not significantly harm specific governmental
¶21The second major consideration favor-ing reliance upon the Senate Report is the fact that it was the only committee report that was before both houses of Congress.
¶22Whether this entire scenario is true or not, it suggests the reason why we as a court viewing the legislative history must. be wary of relying upon the House Report, or even the statements of House sponsors, where their views differ from those expressed in the Senate. As Professor Davis said: "The basic principle is quite elementary: The content of the law must depend upon the intent of both Houses, not of just one."
¶23*194concur) to the interpretation written into the House Report (or voiced in floor colloquy). This being the case, we choose to rely upon the Senate Report.
¶24Interpreted in light of the Senate Report, we hold that the Personnel Management Evaluations here under consideration are not exempt from disclosure by virtue of Exemption 2. They relate not to such “house-keeping” matters such as parking facilities, lunchrooms, sick leave, and the like. The reports deal with the compliance of federal agencies with policies set down by statute, Executive order, and Commission regulations. They deal with such programs as equal employment opportunity, labor-management relations, and employment of Vietnam era veterans. These programs are now or have been the focus of legitimate public interest and attention.
¶25III. Exemption 5
¶26Exemption 5 was intended to exempt from the operation of the FOIA “those documents, and only those documents, normally privileged in the civil discovery context.”
¶27Appellant disputes this finding of the District Court and argues that the evaluative portions of the reports constitute an integral part of an ongoing, pre-decisional deliberative process because (1) they play a consultative role by which the agency evaluates and changes its personnel policies, rules, regulations, and standards, and (2) because they act as a basis for the exercise of the Commission’s power to withdraw an agency’s authority over its personnel affairs.
¶28We can easily reject the second prong of appellant’s position. Appellant made no such argument to the District Court in its cross-motions for summary judgment. Nowhere in the affidavits offered by the Government in the District Court nor in the nine sample reports is there any mention of the fact that the reports are to be used as part of a deliberative process within the Commission.
¶29Turning to appellant’s first argument, we note initially that it is not enough to assert, in the context of Exemption 5, that a document is used by a decisionmaker in the determination of policy. Unevaluated factual reports or summaries of past administrative determinations are frequently used by decisionmakers in coming to a determination, and yet it is beyond dispute that such documents would not be exempt from disclosure.
¶30Of considerable importance to our determination here is that we have been along this road once before. We remanded to the trial court with instructions as to the Government’s burden of showing specifically the character of the documents in issue. On the record we now have from the District Court, the Government attempted to make no clear distinction between the facts obtained as a result of the surveys made by the Commission, to enable the facts to be distinguished from the evaluative, interpretative, or final conclusions of the Commission. Rather, the Government sought to lump all of the reports written by the Commission—facts, interpretation, evaluation, and recommendations— into one mass to be protected. Perhaps there are no true distinctions to be made. Perhaps the “facts” as observed and recorded by the Commission’s investigative staff differ little from the “interpretative” or “evaluative” material. If this is the situation, then the Government has sought to protect too much; it has failed in its obligation to classify and differentiate meaningfully, for certainly not all of this can be characterized as part of the deliberative process.
¶31The Government has thus failed to carry its burden of proof on the issue. Two of the affidavits, those of Mr. Rosen and Mr. Lafferty, merely asserted in conclusory terms that Exemption 5 was applicable to evaluative portions of the reports, e. g., “It is my belief that the *196information contained in this section is exempt from disclosure under exemptions two and five . . . Mr. Schulkind’s affidavit at one point makes a plaintive plea for confidentiality because, in his view, confidentiality has resulted in agencies being forthright and candid with the Commission (not the other way around). However, at no point does the affidavit discuss the role which the evaluative portions play in agency deliberations. We thus reaffirm what we said in our prior opinion: “This affidavit . . . [merely] set forth in conclusory terms the Director’s opinion that the evaluations were not subject to disclosure under the FOIA.”
¶32Looking at the evaluative portions of the sample reports themselves, we note nothing in them to suggest that they are anything other than “final objective analyses of agency performance under existing policy.” While the Commission’s evaluating team probably hopes that its analyses will have a salutary effect on agency personnel practices, the evaluative reports appear to be informational in nature. They provide the raw data upon which decisions can be made; they are not themselves a part of the decisional process.
¶33 The Government appears to argue that this entire process of management appraisal, evaluation, and recommendations for improvement is a seamless whole, that it is in its entirety a deliberative process, and that it is this process which the Government seeks to protect as an ongoing continuous affair. On this view, starting when the Commission’s staff initiates a survey of an agency, when it notes down significant facts of its operation, compares these with personnel and other management directives, evaluates what has been observed, until it formulates these evaluations into conclusions which are usually put as recommendations for improvement to the surveyed agency—all would be a part of a deliberative process. As such it would be protected under Exemption 5; to make any of it subject to disclosure is to do violence to the protection written in Exemption 5.
¶34We cannot accept this. If we consider this entire continuous ongoing process of management appraisal, beginning with the action of the Commission’s staff inquiries through the final recommendations to the subject agency and its final action thereon, as a deliberative process, then surely we would be interpreting Exemption 5 to protect too much.
¶35*197In another respect the proposed Government interpretation seeks to protect too much, i. e., it assumes that the Commission recommendations, which it seeks to protect as part of a continuous deliberative process, eventually always result in final decisions and actions by the surveyed agency which can be made public. After the Commission transmits its evaluations and recommendations to the agency surveyed, there is no legally enforceable obligation on the subject agency to take any action at all. Those with some knowledge of the daily functioning of the bureaucracy may surmise that an unknown number of these recommendations simply go into the files and rest peacefully there, with no action which the Government here would define as final and subject to disclosure ever being taken at all.
¶36Thus, the Government’s characterization of this mass of material it seeks to protect as the “deliberative process” would result in a huge mass of material being forever screened from public view because the administrative bureaucracy had never reached a “final” decision on the management matters involved. The public has an interest in decisions deferred, avoided, or simply not taken for whatever reason, equal to its interest in decisions made, which from their very nature may more easily come to public attention than those never made.
¶37We are not saying that a “final decision” is necessary for there to be a “deliberative process” which is protected by Exemption 5. Rather, we cite the absence of any assured final decision as indicative of the amorphous nature of the mass of information the Government seeks to protect, i. e., the failure of the affidavits relied on to come to grips with and define what it is out of this mass of documents that the Government considers the “deliberative process” and thus entitled to protection.
¶38We consider Exemption 5 as basically a codification of the common sense-common law privilege, i. e., the recognition that the Government cannot operate in a fish bowl. By “common sense-common law privilege” we mean what is usually referred to as “executive privilege,” shorn of any constitutional overtones of separation of powers. As Attorney General Levi has wisely ' observed, “[T]he term ‘executive privilege’ fails to express the nature of the interests at issue; its emotive value presently exceeds and consumes what cognitive value it might have possessed. The need for confidentiality is old, common to all governments, essential to ours since its formation.”
¶39Exemption 5 is designed to protect subordinates’ advice to superiors, it is designed to protect a true deliberative process usually leading up to final decisions. If we construed Exemption 5 as broadly as the Government seeks to do here, we would go a long way toward undercutting the entire Freedom of Information Act. There is a huge quantity of amorphous management improvement activity in every agency which would be protected by an equivalent rationale, u we held that the evaluation reports of the Commission and the mass of facts behind them in this case were so protected.
¶40The affidavits relied on by the Government not only fail because they are conclusory; the affidavits fail to carry the Government’s burden of proof here because at no place do they define, explain, or limit the “deliberative process” which the Government seeks to protect.
¶41We decline to hold on this record that there is a deliberative process, beginning with the Commission’s staff inspection of the subject agency and continuing to a definitely assured “final” decision by the other agency, which can invoke the pro*198tection of Exemption 5 throughout until the final agency decision is reached.
¶42IV. Conclusion
¶43The present case has been pending in administrative channels and before the courts for over three years. After this long period of time we are in a position to affirm finally a disposition that grants plaintiff most of the records he requested of the Commission.
¶44Congress has ordered that actions brought under the FOIA be “expedited in every way.”
¶45Affirmed.
¶46. “[T]his section does not apply to matters that are related solely to the internal personnel rules and practices of an agency . . . ." 5 U.S.C. § 552(b)(2) (1970).
¶47. “[T]his section does not apply to matters that are . . . inter-agency and intra-agency memorandums or letters which would not be available by law to a party other than an agency in litigation with the agency.” 5 U.S.C. § 552(b)(5) (1970).
¶48. Affidavit of Gilbert A. Schulkind, Director, Bureau of Personnel Management Evaluation, United States Civil Service Commission (hereinafter Schulkind Affidavit), ¶] 2; Appendix (App.) 23.
¶49. App. 16-17.
¶50. App. 20-21.
¶51. 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).
¶52.All told, 2,448 reports are covered by plaintiffs request. These reports fill 17 standard-size, five-drawer filing cabinets. Affidavit of John J. Lafferty, Deputy Director, Bureau of Personnel Management Evaluation (Lafferty Affidavit), ilfl 8, 9a; App. 45, 46.
¶53. Vaughn v. Rosen, 383 F.Supp. 1049 (D.D.C.1974).
¶54. Brief for Plaintiff-Appellee, pp. 8-9.
¶56. S.Rep.No. 813, 89th Cong., 1st Sess. (1965), p. 8.
¶57. H.Rep.No. 1497, 89th Cong., 2d Sess. (1966), p. 10, U.S.Code Cong. & Admin.News, pp. 2418, 2427.
¶58. Stokes v. Brennan, 476 F.2d 699 (5th Cir. 1973); Hawkes v. Internal Revenue Service, 467 F.2d 787 (6th Cir. 1972); Stern v. Richardson, 367 F.Supp. 1316, 1319 (D.D.C.1973); Consumers Union of United States, Inc. v. Veterans Administration, 301 F.Supp. 796, 801 (S.D.N.Y.1969), appeal dismissed as moot, 436 F.2d 1363 (2d Cir. 1971); Benson v. General Services Administration, 289 F.Supp. 590, 595 (W.D.Wash.1968), affd on other grounds, 415 F.2d 878 (9th Cir. 1969); see Getman v. NLRB, 146 U.S.App.D.C. 209, 212 n. 8, 450 F.2d 670, 673 n. 8 (1971). But see Cuneo v. Laird, 338 F.Supp. 504 (D.D.C.1972), rev’d on other grounds, sub nom. Cuneo v. Schiesinger, 157 U.S.App.D.C. 368, 484 F.2d 1086 (1973), cert. denied, sub nom., Vaughn v. Rosen, 415 U.S. 977, 94 S.ct. 1564, 39 L.Ed.2d 873 (1974); City of Concord v. Ambrose, 333 F.Supp. 958 (N.D. Cal.1971).
¶59. K. Davis, Administrative Law Treatise, § 3A.17, pp. 144, 145 (1970 Supp.).
¶60. 5 U.S.C. § 1002 (1964).
¶61. Environmental Protection Agency (EPA) v. Mink, 410 U.S. 73, 79, 93 S.Ct. 827, 832, 35 L.Ed.2d 119 (1973).
¶62. Ibid.
¶63. 112 Cong.Rec. 13644—45 (20 June 1966).
¶65. H.Rep.No. 1497, 89th Cong., 2d Sess. (1966), p. 10, U.S.Code Cong. & Admin.News, p. 2427.
¶66. Ibid.
¶67. Soucie v. David, 145 U.S.App.D.C. 144, 157, 448 F.2d 1067, 1080 (1971).
¶68. Ibid.; Vaughn v. Rosen, 157 U.S.App.D.C. 340, 343, 484 F.2d 820, 823, cert. denied, 415 U.S. 977, 94 S.Ct. 1564, 39 L.Ed.2d 873 (1974).
¶69. See generally, K. Davis, Administrative Law Treatise, § 3A.31 (1970 Supp.).
¶70. Ibid.; Hearings Before the Subcommittee on Administrative Practice and Procedure and Separation of Powers of the Senate Comm, on the Judiciary and the Subcommittee on Inter-governmental Relations of the Senate Comm, on Government Operations on S. 1142, et al. (Vol. 2), 93d Cong., 1st Sess. 122-26 (1973).
¶71. See generally, K. Davis, Administrative Law Treatise, § 3A.31 (1970 Supp.) at 175.
¶72.In Rose v. Department of Air Force, 495 F.2d 261 (2d Cir. 1974), cert. granted, 420 U.S. 923, 95 S.ct. 1115, 43 L.Ed.2d 392 (1975), the Second Circuit considered the “difference of approach between the House and Senate Reports” but came to no resolution because the difference was not critical to the case before it. The court did note that if it were to adopt the Senate construction of Exemption 2, “legitimate public interest” would be a factor in determining the scope of the exemption. 495 F.2d at 265.
¶73. National Labor Relations Board v. Sears, Roebuck & Co., 421 U.S. 132, 149, 95 S.Ct. 1504, 1515, 44 L.Ed.2d 29, 43 U.S.L.W. 4491, 4496 (1975).
¶74. Ibid., quoting from EPA v. Mink, 410 U.S. at 86-87, 93 S.Ct. 827.
¶75. 383 F.Supp. 1053-54.
¶76. Brief for Defendant-Appellant, pp. 24-26.
¶77. EPA v. Mink, 410 U.S. at 87-88, 93 S.Ct. 827.
¶78. Perhaps the court’s disposition here can be better appreciated if the flavor of the material the Government seeks to protect is sampled. In arguing that the “evaluative” sections of the Reports were in many instances indistinguishable from the “recommendations,” although the trial court ordered the first disclosed but the second protected, the Government cited the following from “Report No. 6, a representative Nationwide Review of Personnel Management”:
a. Manpower Planning
Evaluation: The evaluative section observes that improvements need to be made in planning, work organization, and position management. No generally accepted rule of thumb for predicting manpower needs is said to exist. (Representative Reports, hereinafter “RR”, pp. 133-34.
Recommendation: “Place a higher priority on the development of general guidelines and standard for planning manpower needs on a coordinated basis. . . . ”
(RR, p. 135.)
b. Position Classification
Evaluation: There are shortcomings in the classification of positions in one category of employment, due to failure to conform to the agency’s “cyclic audit process to review and revise positions.” (RR, p. 136.)
Recommendation: “Ensure that all inaccurate position descriptions are corrected through full and effective implementation of the cyclic audit process.” (RR, p. 137.)
c. Employee Development and Training
Evaluation: “We found that numerous training programs are developed and administered without sufficient knowledge of, and coordination by the central training staff.” (RR, p. 144.)
Recommendation: “Develop a coordinated system to identify and meet training needs on an agency-wide basis.” (RR, p. 145.)
Evaluation: “Personnel Officers generally exhibit little or no involvement in identifying critical training needs. ...” (RR, p. 145.)
Recommendation: “Substantially increase the involvement of personnel officers in identifying training needs. . . . ” (RR, p. 145.)
d. Promotion Program
Evaluation: Too few promotions are made through competitive procedures, because supervisors don’t understand this program. (RR, p. 146.)
Recommendation: “Competitive procedures must be applied when filling a position with known promotion potential. . ” (RR, p. 149.) Supervisors of the program must be given formal training. (Id.)
¶80. The same management appraisals, which are frequently done in each agency’s own internal self-improvement process, would likewise seem to be protected.
¶81. On an alternative interpretation, the Government would fare differently, but not mu,ch better. We might consider that the Civil Service Commission has completed its work when it makes its evaluation and recommendations to the surveyed agency. On this view, the recommendations to the agency and perhaps the evaluation also would be the final decision of the Commission itself. Under accepted rules this decision would then be publicly disclosable as the final product of the Commission’s work. However, the factual basis of the Commission’s work, i. e., the staffs gathering of facts as to the operation of the subject agency surveyed, would also be disclosable.
¶82We prefer to view the Commission’s recommendations as the seed for deliberations by the subject agency itself, and thus protected by Exemption 5, as the trial court so found.
¶83. Confidentiality and Democratic Government, “The Record of the Association of the Bar of the City of New York,” Vol. 30, p. 323 (May-June 1975). See Nixon v. Sirica, 159 U.S. App.D.C. 58, 121-23, 487 F.2d 700, 763-65 (1973) (Wilkey, J., dissenting).
¶84. The Government relies at this point principally on three recent cases: Renegotiation Board v. Grumman Aircraft Engineering Corp., 421 U.S. 168, 95 S.Ct. 1491, 44 L.Ed.2d 57 (1975); Washington Research Project, Inc. v. Department of Health, Education, and Welfare, 164 U.S.App.D.C. 169, 504 F.2d 238 (1974), cert. denied, 421 U.S. 963, 95 S.Ct. 1951, 44 L.Ed.2d 450, 43 U.S.L.W. 3601 (1975); Montrose Chemical Corporation v. Train, 160 U.S.App.D.C. 270, 491 F.2d 63 (1974). None are applicable here.
¶85In Grumman it was clear that neither the Regional Board decisions nor the decisions of the panel of the National Board had any effect whatsoever until they were considered by the full Board, and the full Board of five members then took action. So, in Grumman we find a deliberative process well defined from the start, with a final decision by the full Board obligatory after one of the Regional Board or panel decisions. This is not true here. As indicated in the text, no final action by the surveyed agency is guaranteed. While the recommendations of the Civil Service Commission may serve as a seed for a deliberative process resulting in such final decision which would be public, there is no guarantee that this necessarily will take place. The prior work of the Commission is to accumulate the factual data, on which the Commission does make its recommendations, which we and the trial court hold protected, but we also hold that the facts accumulated are not within Exemption 5.
¶86In Washington Research a definite decision was called for in each individual case as to whether a person or institution applying for a grant actually received a grant. A decision, grant or denial, was made in each case. So, the evaluative reports there were part of a single deliberative process assuredly leading to a decision which would be made public.
¶87The same can be said of Montrose Chemical, which is even more clearly distinguishable. There the part of the deliberative process we held to be protected was the evaluative summary of the whole vast record, made on the specific direction of the Administrator by his aides in order to assist him in making a decision which he was obligated to make, and did make.
¶88Furthermore, in both Montrose and Washington Research the basic factual data was available to the public. Only the evaluative interpretative summaries were held by us to be protected. Here none of the factual data, this mass in 17 five-drawer file cabinets accumulated by the Civil Service Commission in its survey, would be available to the public in any other form if the Government prevailed.
¶89. 5 U.S.C. § 552(a)(3) (1970).
¶90(concurring):
¶91In considering this case, I have devoted particular reflection to exploring the proper scope of exemption 2, which provides that agency records “related solely to the internal personnel rules and practices of an agency” need not be disclosed under the Freedom of Information Act (FOIA). The legislative history is obscure, and previous decisions less than probing. I find there is more scope than the majority contemplates for exemption 2. But I conclude that the Civil Service Commission studies at issue here must be disclosed because they relate predominantly to evaluation of government-wide personnel policy rather than policy implementation internal to an agency.
¶92A. Legislative History. Both the legislative history and the commentary on the history of the (b)(2) exemption seem confused.
¶93There is a marked difference between the Senate Report and the House Report. The Senate Report provides:
*199Exemption No. 2 relates only to the internal personnel rules and practices of an agency. Examples of these may be rules as to personnel’s use of parking facilities or regulation of lunch hours, statements of policy as to sick leave, and the like.1
¶94The House Report provides:
Matters related solely to the internal personnel rules and practices of an agency: Operating rules, guidelines, and manuals of procedure for Government investigators or examiners would be exempt from disclosure, but this exemption would not cover all “matters of internal management” such as employee relations and working conditions and routine administrative procedures which are withheld under the present law.2
¶95A number of decisions sound the theme that the Senate Report is stronger because the Senate passed the bill before the House issued its report, and therefore the “surer indication of congressional intent is to be found in the Senate Report which was available for consideration in both houses.”
¶96The real question is, which report better fits the language of the statute, and purpose of the law as best the courts can discover that purpose. Putting aside the dates of the reports, it may fairly be said that in general the Senate Report places “emphasis on the fullest responsible disclosure.”
¶97As to exemption (2), all the Senate Report says is that certain items are exempt. No one can fairly doubt that these particular items are exempt under that provision. But the Senate Report cites these as illustrative, not definitional or exclusive. Nor is an intelligible principle of construction supplied in the House Report. It says that under exemption 2 there would be an exemption *200for “operating rules, guidelines and manuals of procedure.” But obviously “operating rules” as a category is far broader than “internal personal rules and practices.”
¶98B. Analysis. As demonstrated in the foregoing discussion, neither previous cases nor legislative history defines the core meaning of the (b)(2) exemption. When there are choices to be made in the application of a core principle, it may be helpful to invoke general guidelines, such as that the Information Act’s spirit favors disclosure and that exemptions are to be strictly construed.
¶99The public disclosure concerns of Congress were first crystallized into law in the APA, which in § 3
¶100The FOIA was meant to abolish the old “internal management” exception, which apparently operated as a license *201for bureaucratic secrecy, perhaps by considering instructions concerning administration as exempt tools of “management” even when they were predominantly administrative constructions of statutes and regulations.
¶101However, the legislature did decide to retain some part of the exception—recast in (b)(2) as “matters related solely to the internal personnel rules and practices of an agency.” This apparently signified a determination that the public interest would not be furthered by a requirement of public disclosure of certain “internal” matters. The pertinent policy has a different quality from the policies underlying exceptions such as those for national security secrets, or discussion of policy matters, or pending cases, as to which there is a high requirement of confidence that is closely confined. Personnel rules and practices are better known than that, and not generally truly “secret.” But there is still a legislative inclination that while the public has a right to know all the activities of an agency that bear on its intentions concerning outsiders, whether formal or informal interpretations and instructions, when purely “internal” matters are involved there is a combination of diminished valid interest in the outsiders relative to the administrative burden imposed, plus a recognition that management of government needs some elbow room in developing and revising internal practices, so as to achieve efficiency, without becoming embroiled in continuous public discussion. The problem is to give effect to both of these policy goals without rendering either of them nugatory by a too broadly sweeping construction of the statutory provisions.
¶102The Civil Service reports at issue here have been specifically authorized by either statute, Executive Order or government-wide Civil Service Commission regulation.
¶103In some attenuated sense, virtually everything that goes on in the Federal Government, and much that goes on outside of it, could be said to be “related” through some chain of circumstances to the “internal personnel rules and practices of an agency.” The potentially all-encompassing sweep of a broad exemption of this type undercuts the vitality of any such approach. The legislature added the qualification that limited the exemption to items “relating solely” to internal personnel practices. Various opinions have relied on “solely” as a means of limiting the range of the (b)(2) exemption.
¶104It seems unlikely that the (b)(2) exemption is applicable only to the kind of routine or trivial agency personnel policies and practices itemized in the Senate Report. But even so the exemption is limited to predominantly “internal personnel rules and practices of an agency.” In contrast, the Civil Service Commission’s responsibility for federal personnel policy is government-wide, not oriented internally within an agency. This construction of the exemption is not only literally accurate, but meaningful in a practical and a policy sense. The reports sought here are not evaluations of the Commission’s own personnel practices, but are instead concerned with evaluating the implementation of government-wide personnel policies established by statute, executive order and regulation. Their primary function is the effectuation of the watchdog and oversight duties assigned to the Commission by Congress and the President. This mandate is the substantive on-line policy responsibility of the Commission; shaping and implementation of personnel policy is its primary duty, not a necessary but secondary problem incidental to implementing executive policy in discrete areas of substantive national regulation.
¶105The reports are therefore not limited “solely” or “predominantly” to the internal practices of “an agency.” Thus, the Federal Personnel Manual, issued by the Commission for government-wide application, could certainly not be withheld from the public in reliance on exemption 2;
¶106Personnel evaluations undertaken within an agency fall within the precise terms of the (b)(2) exemption. The fact *203that a citizen postulates some connection between such evaluations and some general federal personnel policy would not of itself require disclosure of agency evaluations focusing on the agency’s discretionary implementation of its personnel practices to conform to government policy. Congress can fairly be said, by terms, to have established a presumptive zone of protection for such studies, and disclosure would raise more difficult questions than face us on this appeal.
¶107C. Precedents. Perhaps a word should be said about precedents, although in my view they are for the most part not on target. Cases that have previously analyzed the applicability of exemption 2 have done so largely in the discrete context of deciding disclosability of government manuals. These manuals set guidelines for employees carrying out agency policies applicable to the public. The courts have ordered disclosure of “secret law” as within the disclosure mandate of 5 U.S.C. § 552(a)(2)(c) (“administrative staff manual that affect[s] a member of the public”), while protecting agency techniques that if disclosed would materially lessen the agency’s effectiveness vis-a-vis the public it regulates.
¶108Two cases have dealt with the exempt-ability of specific aspects of personnel policy. Hicks v. Freeman, 397 F.2d 193 (4th Cir. 1968) cert. denied 393 U.S. 1064, 89 S.Ct. 717, 21 L.Ed.2d 707 (1969) relies on the language of (b)(2) and on 44 U.S.C. § 305(a) in connection with the Civil Service Commission’s and Depart*204ment of Agriculture’s reduction in force regulations. However that case did not require reflection on whether items need be made public because the regulations were in fact made public, since they were available in the Civil Service Commission’s Federal Personnel Manual and in the Department of Agriculture regulations. The issue was different— whether they were invalid for lack of publication in the Federal Register. Rose v. Department of Air Force, 495 F.2d 261 (2d Cir. 1974), cert. granted, 420 U.S. 923, 95 S.Ct. 1115, 43 L.Ed.2d 392 (1975) is more directly on point in holding that case abstracts under the Air Force Academy’s Honor and Ethics Code were not within exemption 2 because they “have a substantial potential for public interest outside the Government.” That doctrine seems to me to be debatable. But even assuming, for discussion, that the Supreme Court will hold that the Ethics Code is governed by the exemption for personnel policies and practices developed and applied entirely within one agency, the Civil Service Commission’s focus on government-wide federal personnel policy differs from the internally-generated practices at issue in Rose.
¶109. KNen.No.813. 89th Cong., 1st Sess. (1965) at 8.
¶110. H.R.Rep.No. 1497. 89th Cong., 2d Sess. (1966) at 10, U.S.Code Cong. & Admin.News, p. 2427.
¶111. Benson v. General Services Administration, 289 F.Supp. 590, 595 (W.D.Wash.) aff’d on other grounds 415 F.2d 878 (9th Cir.1969). This language is quoted in K. Davis Administrative Law (1970 Supp.) § 3A.18, p. 145, and has been used in various opinions, e. q., Getman v. NLRB, 146 U.S.App.D.C. 209, 212, n.8, 450 F.2d 670, 673 n.8 (1971); Hawkes v. IRS, 467 F.2d 787, 794 (6th Cir. 1972).
¶112. NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 149, n.16, 95 S.Ct. 1504, 44 L.Ed.2d 29, 43 U.S.L.W. 4491, 4496, n.16.
¶113. S.Rep. at 3.
¶114. See, e. g., Professor Davis, loc. cit. at 145: “ ‘Operating rules’ may be ‘internal personnel rules’ only to the extent that they deal with the relations between an agency and its employees, not to the extent that they deal with the relations between an agency and an outsider or between employees of the agency and an outsider.”
¶115. Ibid.
¶116. Strict construction is also implicit in the command of § 552(c): “This section does not authorize withholding of information or limit the availability of records to the public, except as specifically stated in this section.”
¶117. FAA Administrator v. Robertson, 422 U.S. 255, 95 S.Ct. 2140, 45 L.Ed.2d 164, 43 U.S.L.W. 4833.
¶118. 5 U.S.C. § 1002 (1964).
¶119. Final Report of the Attorney General’s Committee, Administrative Procedure in Government Agencies (1941), p. 29.
¶120. Ibid., pp. 27-28.
¶121. 5 U.S.C. § 1002 (1964).
¶122. Department of Justice, Attorney General’s Manual on the Administrative Procedure Act (1947), p. 18.
¶123. Cf. United States v. Hayes, 325 F.2d 307, 309 (4th Cir. 1963).
¶124. Labor-management relations, Exec. Order No. 11491, 3 C.F.R. 254 (1974). Classification of positions, 5 U.S.C. §§ 5101-5115 (1970). Equal Employment Opportunity: Exec. Order No. 11478, 3 C.F.R. 207 (1974). Recruitment and merit promotion, 5 U.S.C. § 3301. Incentive awards, 5 U.S.C. §§ 4501—4506. Performance evaluation, 5 U.S.C. §§ 4301-08. Em-payment of Vietnam era veterans, Exec. Order No. 11521, 3 C.F.R. 276 (1974). Employee training, 5 U.S.C. §§ 4101-18, Exec. Order No. 11348, 3 C.F.R. 188 (1974). Reductions in force, 5 U.S.C. §§ 3501-04.
¶125.See e. g., Stokes v. Brennan, 476 F.2d 699, 703 (5th Cir. 1973); Benson v. General Services Administration, 289 F.Supp. 590, 595 (W.D.Wash.1968), aff’d on other grounds, 415 F.2d 878 (9th Cir. 1969).
¶126.The Attorney General’s Memorandum on the Public Information Section of the Administrative Procedure Act (1967), asserts that public disclosure of the Federal Personnel Manual is a matter of discretion, and is not required under the Act. P. 31. Of course, agencies may and' do disclose information technically exempted by the FOIA when no purpose is served in keeping the material secret (see Davis, § 3A.5, pp. 122-23). In my view, however, disclosure of the Manual could be compelled were it not made available voluntarily.
¶127. 5 U.S.C. § 553(a)(2) (1970).
¶128. Of course, the fact that information need not be disclosed under the FOIA does not prohibit its disclosure.
¶129. For example, a union might seek an internal agency study of its labor-management relations to gain an edge in collective bargaining. Disclosure of such a study would not be required under the Act.
¶130. In Hawkes v. IRS, 467 F.2d 787, 797 (6th Cir. 1972), the court found it “unlikely” that the IRS manual in question dealt with the “employee-employer type concerns” exempted by the internal practices and policies language of (b)(2). The court did not have occasion to define the scope of the employer-employee concerns that (b)(2) exempts from disclosure. See also Long v. United States Internal Revenue Service, 349 F.Supp. 871, 874 (W.D.Wash. 1972) (IRS Manual “effects members of the public and is not solely related to internal personnel functions of the IRS”). In Stokes v. Brennan, 476 F.2d 699, 703 (5th Cir. 1973) the court affirmed disclosure of a Labor Department training manual “containpng] the substance of what the statute [Occupational Safety and Health Act 29 USC § 651 et seq. (1970)] commands” even though “relatively immaterial parts of the manual, such as the introduction and welcome to the course, could be classified as internal personnel rules and practices.” The court did not justify its failure to require excision of the manual portions falling within the description of exempted personnel rules and practices. In Benson v. General Services Administration, 289 F.Supp. 590, 595 (W.D.Wash.), aff’d on other grounds, 415 F.2d 878 (9th Cir. 1969) the (b)(2) exemption was held to be totally inapplicable to GSA documents dealing with negotiations and sale of real estate. Such documents were not “rules and practices of general application relating to such matters as employee use of the employer’s plant and equipment, and the amount of time in each working day which is to be devoted to the employer’s business and such activity.” See also Cuneo v. Schlesinger, 157 U.S. App.D.C. 368, 372, 484 F.2d 1086, 1090 (1973) cert. denied, Rosen v. Vaughn, 415 U.S. 977, 94 S.Ct. 1564, 39 L.Ed.2d 873 (1974) (“secret law” in defense contract audit agency manual discloseable); Consumers Union v. Veterans Administration, 301 F.Supp. 796 (S.D.N.Y. 1969) appeal dismissed as moot 436 F.2d 1363 (2nd Cir. 1971). But cf. Concord v. Ambrose, 333 F.Supp. 958, 960 (N.D.Cal.1971) (dictum: “ ‘personnel rules’ can be so construed to cover instructions to law enforcement personnel on the tactics by which they should effect arrests”); Polymers, Inc. v. NLRB, 414 F.2d 999 (2d Cir. 1969), cert. denied, 396 U.S. 1010, 90 S.Ct. 570, 24 L.Ed.2d 502 (1970) (follows House Report in finding that NLRB document, “A Guide to the Conduct of Elections” was exempted by (b)(2) and (b)(5)).