658 F. 2d 957 - Radowich v. United States Attorney District of Maryland

The plaintiff/appellee sought the release under the Freedom of Information Act (FOIA)1 of certain information given the office of the United States Attorney in confidence by one Gudelsky in the course of a criminal investigation of political corruption conducted by that office. The Government objects to release asserting that such information was exempt from disclosure under work product,2 personal privacy,3 and confidential information4 exemptions. Other than for the excision under the work product exemption of the purely deliberative statements in the requested information made by the Assistant United States Attorney himself, the district court rejected all the exemption claims of the Government and ordered disclosure.5

The Government has appealed, limiting its claims to those under the work product and confidential information exemptions. Since we find, contrary to the ruling of the district court, that the requested information is not subject to compulsory disclosure under exemption (b)(7)(D) (confidential information), it is unnecessary to consider the work product claim of the Government.

Exemption (b)(7)(D) protects from compulsory disclosure two specific types of confidential information, as stated in two separate clauses of the exemption, i. e., (1) any information disclosing the identity of a "confidential source" in any law enforcement investigation, whether civil or criminal; and (2) all information furnished by a confidential source if compiled in the course of a criminal investigation.6 The distinction between the two types is that the first covers "(p)ersonnel, regulatory, and civil enforcement investigations" as well as criminal investigations if they reveal the identity of a confidential source, whereas the second embraces all information furnished by a confidential source but only in the course of a criminal investigation.7

The district court found the first clause of this exemption to grant protection "only to the identity of a confidential source and (held that, since) the identity of Mr. Gudelsky's attorney is neither confidential nor at issue," the clause offered no protection from disclosure in this case.8 It dismissed the second clause as inapplicable since "there (had) been no showing that these notes (which were the subject of the request for disclosure) contain 'confidential information furnished only by the confidential source.' "9 We shall address initially the claim of the appellants under the first clause of the exemption. In so doing, we conclude the district court erred in its construction of the exemption and reverse its order of disclosure. Since the district court, however, found disclosure also not foreclosed under the second clause of 7(D), we must address that point too. We find the district court's ruling that this second clause offers no protection in this case is likewise erroneous.

It is manifest that, in dismissing the claim of exemption under the first clause of 7(D), the district court construed the term "confidential" as "secret" and, having so defined the term, concluded that, because Gudelsky and his attorney were at the time of the disclosure request known and thus were not "secret" sources, their information, even though suggestive of their identity, was not protected from disclosure. Under its construction of the term "confidential," the fact that the "source" had talked to the Assistant United States Attorney under the conceded pledge or promise of confidentiality was "irrelevant" on the issue of disclosure under 7(D) after the "identity" of the confidential source became known. The flaw in this reasoning, though, is that "confidential," as used in the exemption, is not to be construed as "secret" but as "given in confidence" or "in trust." This was made clear in the Joint Explanatory Statement of the Committee of Conference, Conference Report No. 93-1200, 1974, U.S.Code Cong. & Admin.News at 6291:

"The substitution of the term 'confidential source' in section 552(b)(7)(D) is to make clear that the identity of a person other than a paid informer may be protected if the person provided information under an express assurance of confidentiality or in circumstances from which such an assurance could be reasonably inferred. Under this category, in every case where the investigatory records sought were compiled for law enforcement purposes either civil or criminal in nature the agency can withhold the names, addresses, and other information that would reveal the identity of a confidential source who furnished the information."

And this construction of the exemption in the Joint Conference Report was followed in Keeney v. Federal Bureau of Investigation, 630 F.2d 114, 119 n. 2 (2d Cir. 1980), where the Court, after a careful review of the legislative history of the exemption, declared:

"We also reject plaintiff's argument that the exemption is inapplicable because the local law enforcement agencies in the present case have now been identified, and thus cannot constitute 'confidential' sources. As we have discussed above, the term, 'confidential' is used in exemption 7(D) in the sense of reposing confidence or trust, rather than in the sense of 'secret.' "

This definition of "confidential source," as stated in the Joint Conference Report and in Keeney is substantially the same as we announced earlier in Deering Milliken, Inc. v. Irving, 548 F.2d 1131, 1138 (4th Cir. 1977), where we held that to establish "the confidentiality of the source," which is, according to Lesar v. United States Dept. of Justice, 636 F.2d 472 (D.C.Cir.1980), the condition placed on the applicability of the exemption, it is only necessary to show that the information was given under an express assurance of confidentiality or in circumstances where such an assurance could reasonably be inferred. See also Maroscia v. Levi, 569 F.2d 1000, 1002 (7th Cir. 1977).

It is also clear that, as the Court declared in Lesar v. United States Dept. of Justice, supra, 636 F.2d at 491, the proscription of involuntary disclosure provided by the exemption does not disappear if the "identity" of the "confidential source" later becomes known but continues until the beneficiary of the promise of confidentiality waives disclosure. See also, to the same effect: Volz v. United States Dept. of Justice, 619 F.2d 49, 50 (10th Cir. 1980), cert. denied, --- U.S. ----, 101 S.Ct. 397, 66 L.Ed.2d 244 (1980); Scherer v. Kelley, 584 F.2d 170, 176, n. 7 (7th Cir. 1978), cert. denied, 440 U.S. 964, 99 S.Ct. 1511, 59 L.Ed.2d 778 (1979).

It follows that the district court erred in holding that the first clause of the exemption was inapplicable in this case because the identities of the confidential sources were known. This is true because it is not suggested that the "confidential" sources in this case who concededly had talked to the Assistant United States Attorney only because of a pledge of confidentiality, had waived the assurance of confidentiality given them by the Assistant United States Attorney. The United States Attorney accordingly was perfectly right in denying disclosure of information that might disclose the identity of his "confidential" sources and the district court should have sustained his refusal under the first clause of 7(D).10

Turning now to the second clause of 7(D), which protects "information furnished by a confidential source," the district court premised its conclusion that such exemption was inapplicable in this case on its construction of the phrase "confidential information furnished only by the confidential source" in that exemption. In its view, this language exempted only such information given by a confidential source that was not available from another source, and it would impose on the law enforcement agency the obligation of establishing that the information was not available to the agency from another source, as a condition to the right to the exemption.

Admittedly, this clause of the statute, in this particular at least, is awkwardly phrased. It is perhaps possible to give the language the construction adopted by the district court. On the other hand, the language can be construed as intended to limit the protection to information furnished by a confidential source only and not to information furnished by others. Under this construction any information furnished by the confidential source would be absolutely protected but disclosure could be compelled for information furnished by other than a confidential source. To repeat, under this construction, this exemption is limited to information furnished by a confidential source and only that information. And this construction seems more in keeping with the exemption's manifest purpose as clearly mirrored in the legislative history.11 It would be unreasonable to assume that Congress could have intended to protect all information reasonably likely to disclose the identity of a confidential source, even though known, in connection with "investigatory records, compiled for law enforcement purposes," either civil or criminal, but not to protect all information furnished by a confidential source, even though available from some other source, "in the course of a criminal investigation, or ... a lawful national security intelligence investigation." This appears to be a classic example of a situation where the Court, in construing statutory language, should look to the clearly expressed intention as expressed without dissent in the legislative history.

If the phrase in question is to be read in line with the repeated expressions of legislative purpose in its enactment, the construction placed on the phrase by the district court violates the principle followed in Chrysler Corp. v. Brown, 441 U.S. 281, 313, 99 S.Ct. 1705, 1723, 60 L.Ed.2d 208 (1979), in looking to "legislative intent," as evidenced in the legislative history, in construing the meaning of phrases in other exemptions in the FOIA, and the principle as stated in Passenger Corp. v. Passengers Assn., 414 U.S. 453, 458, 94 S.Ct. 690, 693, 38 L.Ed.2d 646 (1974) that "even the most basic general principles of statutory construction must yield to clear contrary evidence of legislative intent."12 The proper construction of the phrase, if legislative intent is respected, was accurately stated by the Court in Lesar, supra, 636 F.2d 472. In that case, the appellant, who sought disclosure against a claim of exemption under 7(D), asserted that the records held by law enforcement agencies were not "confidential" because, between court proceedings, news leaks, and FOIA requests, much of the information sought was in the public domain. The Court, however, dismissed the argument and held that the information was protected under 7(D), saying at p. 491:"Appellant next argues that even if nonfederal state and local law enforcement agencies fit within the definition of the word 'source,' the bulk of the records furnished by the Memphis and Atlanta police departments are not 'confidential.' As a result of court proceedings, news leaks, and other FOIA requests, appellant asserts, much of the information in dispute is within the public domain, and the district court therefore erred in granting summary judgment to the Department on this issue. We disagree.

"The affidavits submitted by the Department aver that the Memphis and Atlanta police records were provided to the Task Force with the explicit understanding that these documents would remain confidential. Assuming arguendo that some of the information in these records has found its way into the public domain by one means or another, that does not alter the fact that this information originally was obtained in confidence. No waiver of confidentiality has occurred. The cooperating agencies, to which the right to claim the privilege belongs, have not only supplied the information in confidence originally, but, by affidavits have continued to object to the disclosure of these records." (Emphasis added)

Again in Pratt v. Webster, 508 F.Supp. 751, 760 (D.C.D.C.1981), the Court, relying on Lesar, reached the same result:

"In Lesar, at 492, the court noted that the focus of its analysis was on determining whether the information was supplied by a confidential source. 'Once that question is answered in the affirmative, all such information obtained from the confidential source receives protection.' The affidavits properly index the material for which the exemption was invoked. In all five categories for the claimed exemption the defendant has established the confidentiality of the source."

Lesar and Pratt, in their discussion, are fully supported by the legislative record of 7(D). The author of the final version of 7(D) was Senator Hart. It was prepared by him, while the 1974 Amendments to the FOIA were in conference. He was one of the Senate conferees and he submitted the language that appears in 7(D) as a response to the objection voiced by President Ford to the original language in the section. In connection with the submission of the Amendments as revised in conference, including the final language in 7(D), to the Senate, Senator Hart offered the following explanation of the meaning of such exemption and particularly that of the very language in issue here:

"The major change in conference was the provision which permits law enforcement agencies to withhold 'confidential information furnished only by a confidential source'. In other words, the agency not only can withhold information which would disclose the identity of a confidential source but also can provide blanket protection for any information supplied by a confidential source. The President is therefore mistaken in his statement that the FBI must prove that disclosure would reveal an informer's identity; all the FBI has to do is to state that the information was furnished by a confidential source and it is exempt. In fact, this protection was introduced by the conferees in response to the specific request of the President in a letter to Senator Kennedy during the conference. All of the conferees endorsed the Hart amendment as modified."13

(Emphasis added)

Beyond this Senator Kennedy, as Chairman of the Senate Conferees, and Congressman Moorhead, as Chairman of the House Conferees, answering the letter of President Ford suggesting changes in the language of 7(D), among other things, wrote the President after the Conference Committee had agreed on Senator Hart's language and stated the meaning of the present language of 7(D). In this joint letter of September 23, 1974, they said:

"After reviewing the points made in your letter on this point, (i. e., 7(D)) the conference committee also agreed to adopt language offered by Senator Hruska to permit the withholding of the information provided by a confidential source to a criminal law enforcement authority during the course of a criminal or 'lawful national security intelligence investigation.' The Federal agency may, in addition, withhold the identification of the confidential source in all law enforcement investigations civil as well as criminal."14 (Emphasis in text)

Despite these firm assertions made by the author of the exemption in question and by the respective chairmen of the two Conference Committees responsible for the presentation of the revised exemption to the Senate and House, the President vetoed the Amendments, finding, inter alia, the language of 7(D) too narrow in its definition of protected information of criminal investigative records. In urging the Senate to override the veto and in attempting to assuage any fears that the members of the Congress might have that 7(D) would permit disclosure of any confidential material furnished by a confidential source, in the course of a criminal investigation, Senator Kennedy, who was in charge of the Amendments in the debate, assured the Senate as follows:

"Then we also provided (in 7(D)) that there be no requirement to reveal not only the identity of a confidential source, but also any information obtained from him in a criminal investigation. The only source information that would be available would be that compiled in civil investigations." (Emphasis added)15

As the debate in the Senate on the question of overriding the President's veto was concluding, Senator Byrd, the majority leader, gave the Senate his considered construction of the language of 7(D) in assuring them that in enacting the exemption as drafted by Senator Hart, information provided by a confidential source in connection with a criminal investigation would be protected and that the fears to the contrary expressed by President Ford in his veto message were groundless. He said:

"The Senate-passed version of the bill contained an amendment which would have required disclosure of information from a law enforcement agency unless certain information was specifically exempted by the act. What particularly disturbed me was that while the identity of an informer would be protected, the confidential information which he had given the agency would not have been protected from disclosure. Another matter that disturbed me was the use of the word 'informer', since that could be construed to mean that only the identity of a paid 'informer' was to be protected and not the identity of an unpaid confidential source. I was deeply concerned that without such protection, law enforcement agencies would be faced with a 'drying-up' of their sources of information and their criminal investigative work would be seriously impaired.

"The bill in the form now presented to the Senate has been significantly changed by the conference on these critical issues. The language of section 552(b)(7) has been changed from protecting from disclosure the identity of an 'informer' to protecting the identity of a 'confidential source' to assure that the identity of a person other than a paid informer may be protected. The language has also been broadened substantially to protect from disclosure all of the information furnished by a confidential source to a criminal law enforcement agency if the information was compiled in the course of a criminal investigation. Thus, not only is the identity of a confidential source protected but also protected from disclosure is all the information furnished by that source to a law enforcement agency in the course of a criminal investigation."16

It is plain from this review of the legislative history that the indisputable legislative intent, as expressed in 7(D), was to exempt from involuntary disclosure all information furnished by a confidential source. Congress had included the adverb "only" for the purpose of confining its exemption to that single type of "confidential information" but not for the purpose of weakening the exemption in favor of information "furnished by a confidential source" in connection with a criminal investigation, by fragmenting that information into two types, one subject to involuntary disclosure and one not so disclosable. There is not a single statement in the legislative history that suggests that Congress, by enacting the exemption, did not intend to cast a veil of protection over all information furnished by a confidential source in the course of a criminal investigation, irrespective of whether the information in whole or in part, might have been available from some other source. It is our obligation to respect that Congressional intent in our construction of 7(D). See Steelworkers v. Weber, supra, and Chrysler Corp. v. Brown, supra. Lesar did that and so do we.17

The appellee has advanced the argument that note 16 in Church of Scientology v. U.S. Dept. of Justice, 612 F.2d 417, 428 (9th Cir. 1980) is contrary to the construction of the second clause of 7(D) adopted by us. We do not find the language of the note enlightening on the point with which we are concerned. It simply repeats the phrase "confidential information only furnished by a confidential source" without offering any further explanation of the phrase.18 This phrasing itself can be understood as merely limiting the exemption to "confidential information from a confidential source." If this is its meaning, it is not different from our construction of the exemption. If, however, it is to be given any other meaning, it is contrary to the clearly articulated legislative intent, which we are obligated to follow, and would be contrary to the result in Lesar and Kenney, supra, in construing the exemption.

In summary, we conclude that (1) the term "confidential information" as used in the first clause of 7(D) refers to not "secret" information, but information given "in confidence or trust," protected from compelled disclosure until the assurance of confidentiality has been waived, and (2) that the protection from compelled disclosure given by the second clause of that exemption extends to all information furnished by the confidential source in the course of a criminal investigation.

The judgment of the district court is accordingly reversed and the cause is remanded to the district court for action in conformity with this opinion.

The majority holds that the requested materials are exempt from disclosure under both the first and second clauses of exemption (b)(7)(D) of the Freedom of Information Act, 5 U.S.C. § 552(b)(7)(D). The pertinent portion of that section provides:

(b) This section does not apply to matters that are ...

(7) investigatory records compiled for law enforcement purposes, but only to the extent that the production of such records would ... (D) disclose the identity of a confidential source and, in the case of a record compiled by a criminal law enforcement authority in the course of a criminal investigation, ... confidential information furnished only by the confidential source ....

Keeping in mind that "FOIA exemptions are to be narrowly construed in accordance with the legislative purpose of Congress that disclosure rather than secrecy is the dominant objective of the Act," Nix v. United States, 572 F.2d 998, 1002 (4 Cir. 1978), citing Department of the Air Force v. Rose, 425 U.S. 352, 360-61, 96 S.Ct. 1592, 1599, 48 L.Ed.2d 11 (1976), I am impelled to the conclusion that the majority misconstrues the Act with respect to both clauses of exemption 7(d). I therefore respectfully dissent.

I set forth my views with respect to each of the two clauses:

The district court held that the first clause of 7(D) "disclose the identity of a confidential source" protected only the identity of confidential sources and found the exemption inapplicable because "the identity of Mr. Gudelsky's attorney is neither confidential nor at issue." Relying on legislative history and on language from Keeney v. FBI, 630 F.2d 114, 119 n.2 (2 Cir. 1980), and Lesar v. Dept. of Justice, 636 F.2d 472 (D.C.Cir.1980), the majority concludes that confidential means not "secret" but "reposing confidence," and that the first clause of 7(D) remains applicable despite the fact that the identity of Gudelsky's attorney and the fact that he talked to the government are not "secret."

I agree with the majority that "confidential" in 7(D) does not mean "secret," but I do not believe that conclusion makes any difference in this case. Notwithstanding that his identity was known, I would agree that Gudelsky's attorney remains a "confidential source" within the meaning of 7(D), but that conclusion does not alter the fact that the first clause of 7(D) protects only his identity, and not the information he provided.1 This conclusion is compelled by the clear language of the statute and the legislative history. The first clause protects against production of records that would "disclose the identity " of a confidential source. The joint explanatory statement of the House and Senate Conferees on the 1974 amendments to FOIA states that under the first clause:

the agency can withhold the names, addresses, and other information that would reveal the identity of a confidential source....

Conf.Rep.No.1200, 93d Cong., 2d Sess., reprinted in 1974 U.S.Code Cong. & Admin.News 6285, 6291 (emphasis added). I cannot imagine how the statute or legislative history could make it any clearer that the first clause of 7(D) is addressed only to the protection of identities. Because Gudelsky himself disclosed the fact that his attorney spoke with federal prosecutors,2 the district court was quite correct in concluding that clause 1 of 7(D) simply is not "at issue."

As I stated earlier, the fact that Gudelsky's attorney's identity is no longer "secret" does not mean that he is no longer a "confidential source" within the meaning of 7(D). The importance of this conclusion is that, while clause 1 can no longer protect what is already disclosed, the attorney remains a "confidential source" whose information remains protected under clause 2 of 7(D). This is all that Keeney, supra, stands for. There the plaintiff sought information provided in confidence to the federal government by local law enforcement agencies. In footnote 2, the court held that the agencies remained "confidential sources" under 7(D) even though their identities were known. The court stated:

(W)hile there will obviously be circumstances in which (local) agencies will not wish the fact of their own investigations to be made public and hence will not wish their identity disclosed, it is equally conceivable for an agency that is known to be investigating a particular matter to seek confidential treatment for fruits of that investigation.

630 F.2d at 119 n.2 (emphasis added).

Lesar, supra, also relied upon by the majority, is not on point at all. There certain information supplied to federal authorities had already found its way into the public domain. The court held simply that, because some portions of documents given to the government in confidence had been publicized, the documents did not thereby fall without the definition of "confidential information" in 7(D). The protections of clause 1 of 7(D), dealing with identities of sources, were not even at issue in Lesar.

The best argument for extending the protection of clause 1 in the manner the majority seeks, may be made by analogy to Volz v. United States Dept. of Justice, 619 F.2d 49 (10th Cir. 1980), a Privacy Act case. There the court dealt with exemption (k)(5) of the Privacy Act which exempts from disclosure "information ... that ... would reveal the identity of a (confidential) source," in a case where the identity of the informant was known but the particular information supplied was not. The court concluded:

Subsection (k)(5) protects the confidentiality of any substantive information provided by (a confidential source) insofar as disclosure would reveal that he was the agency's source for that information.

619 F.2d at 50 (emphasis added). Thus the court extended (k)(5)'s protection of identity to include the information supplied by a confidential source. But the structure of exemption 7(D) of FOIA dooms any attempt to carry over the Volz reasoning from the Privacy Act context. Unlike (k)(5), 7(D) includes a second clause which specifically defines the protections accorded to information supplied by confidential sources. To read a broad protection of information into the first clause of 7(D), as the Volz court could do with (k) (5), would be to render meaningless the second clause of 7(D). As the majority appears to recognize, the two clauses of 7(D) serve distinctly different functions:

(T)he first covers '(p)ersonnel, regulatory, and civil enforcement investigations' as well as criminal investigations if they reveal the identity of a confidential source, whereas the second embraces all information furnished3 by a confidential source but only in the course of a criminal investigation.

At 959 (emphasis in original). What the opinion fails to recognize is that Gudelsky deprived himself of the protection of the first clause when he voluntarily disclosed the fact of his attorney's talks with prosecutors. He did not thereby forfeit the protection of the information he supplied, but he must meet the requirements of clause 2 in order to receive that protection.Clause 2

Clause 2 exempts from disclosure "confidential information furnished only by the confidential source." The government expressly disavowed any reliance on clause 2 because it failed to show below that the information was supplied only by confidential sources. Br. of Govt. at 20 n.10.4 Nevertheless, the majority resurrects clause 2 and finds that it justifies nondisclosure simply by disregarding the word "only" in the statute. It does so on the basis of portions of the legislative history which really do not consider the issue it decides. More significantly, it does so in the face of Fourth Circuit precedent directly to the contrary.

The majority quotes various statements out of the legislative history which suggest that all information provided by confidential sources is exempt from disclosure. I find none of the passages at all convincing because none even purports to deal with an instance where one source provides information in confidence while others are willing to provide the same information with no assurances of confidence. None of the statements make any attempt to explain the limitation intended by the word "only" in the language of the statute. The word is hard to dismiss as an oversight, however, since it is included as well in the joint explanatory statement of the conferees:

However, where the records are compiled by a criminal law enforcement authority, all of the information furnished only by a confidential source may be withheld....

Conf.Rep.No.1200, 93d Cong., 2d Sess., reprinted in 1974 U.S.Code Cong. & Admin.News 6285, 6291 (emphasis in original).

We have previously addressed the meaning of the word "only" in clause 2, and have explicitly rejected the reading the majority gives it:

The (7)(D) exemption is plainly designed to remove impediments to investigation by assuring confidential sources that not only will their identities be retained in confidence save for the proper exercise of the power of subpoena but so, too, will the information itself when obtained only from confidential sources. We find that the exemption is intended to protect all such confidential information when furnished only by a confidential source, whether one or more.

This suggests in turn that such information is not protected under the second phase of subsection (7)(D) unless it is furnished "only" by confidential sources. It is apparent from a review of the record in this case that the information released by the FBI, especially the FBI's interviews with Nix and inmate Isenock, contains some of the same revelations as are to be found in the inmate letters and FBI interviews with other inmates, guards and the prison supervisory official. To that extent, the information which the FBI declines to disclose stems from non-confidential sources, namely, Nix and Isenock.

Nix v. United States, 572 F.2d at 1004. In Nix, we went on to find that, even though clause 2 of 7(D) was inapplicable, the requested information was exempted because its release might divulge the identities and threaten the security of certain unnamed informants. Nevertheless, our interpretation of clause 2 of 7(D) is clear.

I am not at all uncomfortable that release of the information in this case poses any threat to the government's ability to coax informants to speak in the future. All that we need say in denying a 7(D) exemption in this case is that information will be released where an informant has already made known the fact of his communication with the government and where the information he supplies is something that the government has gotten from nonconfidential sources.

Because I would conclude that the requested materials are not exempt from disclosure under either the first or second clauses of exemption (b)(7)(D), I am brought to the government's claim of work product exemption. It is based upon exemption 5, 5 U.S.C. § 552(b)(5), which exempts from production "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."

The district court found that those portions containing the attorney's "mental impressions, opinions and legal theories" were exempt from production. However, it found certain factual materials to be "reasonably segregable" from privileged material and ordered production of the segregable portions of the documents. The government raises two objections to this conclusion. First, it argues that even factual matter in an attorney's notes of a witness's oral statements inevitably discloses the attorney's mental impression regarding the significance or relevance of particular facts. Second, it contends that these materials should be exempt from disclosure because their release would have a significant adverse impact on future investigations.

While NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 95 S.Ct. 1504, 44 L.Ed.2d 29 (1975), and Upjohn Co. v. United States, --- U.S. ----, 101 S.Ct. 677, 66 L.Ed.2d 584 (January 13, 1981), on which the government relies, arguably buttress its first argument, we rejected that argument in Deering Milliken, Inc. v. Irving, 548 F.2d 1131 (4 Cir. 1977). There we directed the district court to order disclosure of "factual" materials in attorneys' notes after deleting portions containing "deliberative" materials, and we explicitly rejected the Board's argument that the FOIA work product privilege "embraces all factual material assembled for trial by a litigant's attorney or other representative." Id. at 1138. Deering Milliken controls here, and I would reject the government's first argument.

The government's second argument does not persuade me. It seems to me that, with regard to law enforcement agencies, exemption 7, 5 U.S.C. § 552(b)(7), provides an exclusive list of circumstances where considerations of confidentiality justify nondisclosure. I cannot accept that exemption 5 provides a more expansive bar to disclosure whenever the government asserts that a breach of confidentiality threatens its fact-gathering ability. Since, as I have shown, disclosure is not protected by exemption 7, I do not think that it is protected by exemption 5.

I would affirm the judgment of the district court.