Florida House of Representatives, Honorable T.K. Wetherell, Speaker, Cross-Appellant v. United States Department of Commerce, Cross-Appellee
961 F.2d 941
11th Cir.
1992
Check Treatment
COX, Circuit Judge:

On July 19, 1991, the Florida House of Representatives (“Florida”) filed a Freedom of Information Act (“FOIA” or “Act”) request seeking release of the United States Department of Commerce’s (“Department” or “Secretary”) adjusted block level census data which the Department had developed in the course of its 1990 census duties. The Secretary denied the request on the ground that the data was exempt from disclosure under Exemption 5 of the FOIA. After its request was denied, Florida sued the Secretary, seeking disclosure under the FOIA. Finding that the adjusted block level data did not fall within the purview of Exemption 5, the district court granted summary judgment in favor of Florida.

The Secretary appeals the district court’s order granting summary judgment in favor of Florida, arguing that the district court erred when it found Exemption 5 inapplicable to the adjusted census data. Florida cross-appeals, contending that the district court erred in finding that the Secretary did not waive his privilege under Exemption 5, assuming that he had the privilege in the first place.

Because we find both that (1) the data at issue are covered by Exemption 5 and (2) the Secretary did not waive this privilege, we reverse.

I. FACTS AND PROCEDURAL HISTORY

The Constitution requires that a census be taken every ten years. U.S. Const, art. 1, § 2. By statute, Congress requires the Secretary of Commerce to conduct the census “in such form and content as he may determine.” 13 U.S.C. § 141(a). For the 1990 census,-the Secretary initially undertook an “actual” census headcount. 1 Basically, the actual headcount involved an enormous mailing of questionnaires by the Department, as well as a great number of census takers knocking on a great number of doors.

After the actual headcount was completed, the Secretary considered a Census Bureau proposal, accompanied with adjusted census data, to statistically adjust the actual headcount. To arrive at the adjusted census data, the Census Bureau first conducted a post-census, follow-up, sample survey referred to as a “post-enumeration survey” (“PES”). The PES re-surveyed approximately 5,000 block clusters from across the nation. The blocks 2 were selected to ensure that they included representatives of 1392 different population subgroups known as “post-strata.” These subgroups theoretically reflect the various demographic components of the entire United States population. The information gathered from the PES sample blocks was then compared with the information for those same blocks that was obtained during the actual headcount. By comparing the results of both surveys, the Bureau developed an adjustment factor which reflected the extent to which the PES suggested an over or undercount in the census.

The Bureau then multiplied the number of people from each post-stratum repre *944 sented in each populated block (as counted by the actual census) by the appropriate adjustment factor for each post-stratum. This process was completed for all of the inhabited blocks in the United States.

On July 15, 1991, the Secretary issued a final decision determining that the actual census headcount should not be adjusted based on the statistical methodology proposed. 56 Fed.Reg. 33,583 (July 22, 1991). The Secretary’s final decision outlines the adjustment methodology considered and includes statistically adjusted calculations for the state, county, and city levels; it does not include the adjusted block level data at issue.

Florida points out that the Department did in fact disclose the block level data at issue under a protective order in City of New York v. United States Dep’t of Commerce, 713 F.Supp. 48 (E.D.N.Y.1989). The protective order limited the information under pain of contempt to experts and other identified parties in City of New York. In addition, the Department disclosed half of the adjusted block level data to Congress; data which Florida eventually obtained.

II. ISSUES ON APPEAL

1. Whether the adjusted block level census estimates at issue in this case are exempt from disclosure under Exemption 5 of the FOIA.

2. Assuming that the information may be withheld by the Department under Exemption 5, whether the Department waived this privilege.

III. DISCUSSION

A. Background

The Freedom of Information Act, 5 U.S.C. § 552(a), requires federal agencies to disclose to the public a wide range of information unless the information at issue falls within one of the nine enumerated exemptions listed in § 522(b). NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 136, 95 S.Ct. 1504, 1509, 44 L.Ed.2d 29 (1975) (noting that “[a]s the Act is structured, virtually every document generated by an agency is available to the public in one form or another, unless it falls within one of the Act’s nine exemptions.”).

“[T]he mandate of the FOIA calls for broad disclosure of Government records, and for this reason [the Supreme Court has] consistently stated that FOIA exemptions are to be narrowly construed.” United States Dep’t of Justice v. Julian, 486 U.S. 1, 8, 108 S.Ct. 1606, 1611, 100 L.Ed.2d 1 (1988) (internal citations and quotations omitted).

The Secretary contends that the adjusted block level data may be withheld based on Exemption 5 of the FOIA. Exemption 5 permits agencies to withhold “inter-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). The Supreme Court has construed this provision to exempt from disclosure those documents normally or routinely privileged under the relevant statutory and case law in the pretrial discovery context. Julian, 486 U.S. at 11, 108 S.Ct. at 1613; Sears, 421 U.S. at 149, 95 S.Ct. at 1515.

In EPA v. Mink, 410 U.S. 73, 93 S.Ct. 827, 35 L.Ed.2d 119 (1973), the Supreme Court held that Exemption 5 incorporated the rule of discovery protecting inter-agency predecisional “advisory opinions.” 3 Today, as was the case when Mink was decided, predecisional deliberative documents are not routinely and normally disclosed to a party in litigation with the government. See, e.g., FTC v. Grolier, Inc., 462 U.S. 19, 27, 103 S.Ct. 2209, 2214, 76 L.Ed.2d 387 (1983); Centifanti v. Nix, 865 F.2d 1422 (3rd Cir.1989); Litton Indus. v. Lehman Bros. Kuhn Loeb Inc., 125 *945 F.R.D. 51 (S.D.N.Y.1989); N.O. v. Callahan, 110 F.R.D. 637 (D.Mass.1986); J.R. Norton Co., v. Arizmendi, 108 F.R.D. 647 (S.D.Cal.1985). This privilege against discovery is commonly referred to as the “deliberative process” privilege.

Before the Government can withhold inter-agency documents under the deliberative process privilege, it must show that the document is (1) predecisional, i.e., “prepared in order to assist an agency decisionmaker in arriving at his decision,” Renegotiation Bd. v. Grumman Aircraft Eng’g Corp., 421 U.S. 168, 184, 95 S.Ct. 1491, 1500, 44 L.Ed.2d 57 (1975), and (2) deliberative, i.e., “a direct part of the deliberative process in that it makes recommendations or expresses opinions on legal or policy matters.” Nadler v. United States Dep’t of Justice, 955 F.2d 1479, 1491 (11th Cir.1992) (quoting Vaughn v. Rosen, 523 F.2d 1136, 1144 (D.C.Cir.1975)). Florida does not argue that the adjusted block level date is not predecisional. 4 Therefore, *946 our attention will focus on whether the material is deliberative. Prior to that discussion, however, we will address the question of waiver. 5

B. Waiver

Florida has cross-appealed, arguing that even if the data are privileged under Exemption 5, the district court erred in finding that the Secretary did not waive his right to assert the privilege. Florida relies on three discrete disclosures by the Department to support its waiver argument: (1) disclosure of the data in City of New York; (2) disclosure of half of the data to Congress; and (3) disclosure of other data and reasoning in the Secretary’s final decision.

Waiver is the voluntary relinquishment of a known right or privilege. In Shell Oil Co. v. IRS, 772 F.Supp. 202 (D.Del.1991), the District Court of Delaware arrived at a definition of waiver for purposes of Exemption 5: “Where an authorized disclosure is voluntarily made to a non-federal party, the government waives any claim that the information is exempt from disclosure under the deliberative process privilege.” Id. at 211 (emphasis supplied).

The Department claims that it released the information in City of New York for discovery purposes only because the court, over the Department’s protests, ordered it disclosed. These protests, not surprisingly, were claims that the information was protected by the deliberative process privilege. Florida does not dispute these allegations. We find it difficult to characterize the court-ordered disclosure of the data in City of New York as a voluntary waiver of the Department’s privilege in light of the court order, especially considering that the Department attempted to exercise the very privilege it is supposed to have waived. If documents are exempt from disclosure under the FOIA, the fact that they were involuntarily disclosed by means other than the FOIA should not lead to a finding of waiver.

The disclosure to Congress similarly does not sustain a finding of waiver. The record again reveals that this disclosure was involuntary. A December 10, 1991 letter from Under Secretary Michael R. Darby to the House Subcommittee on Census and Population makes it clear that the Department was dead set against releasing this information to Congress. R. 2-37. It was only under the threat of Congress’s power of subpoena that they reluctantly released half the data. The Subcommittee Chairman’s reply letter only confirms the forced nature of the disclosure. 6 The Secretary may not have brought up the privilege to the Subcommittee, but that is easily understandable since Exemption 5 may not be exercised against Congress. See 5 U.S.C. § 552(c) (“This section [which includes Exemption 5] is not authority to withhold information from Congress.”). The disclosure to Congress was not voluntary and therefore does not support a finding of waiver. 7

*947 Lastly, Florida argues that the Department waived its privilege when it included in its notice of final decision adjusted census estimates for the state, county and city levels, as well as the reasoning behind the estimates. As the Ninth Circuit stated in Mobil Oil Corp. v. EPA, 879 F.2d 698 (9th Cir.1989), when “[t]here is no evidence in the record that the documents that are the subject of [the] appeal were ever revealed ..., [i]mplying a waiver-of exemption for these documents based on the release of related documents ... would be contrary both to the case law on waiver and to the policies underlying FOIA and its exemptions.” Id. at 700.

The reasoning behind Mobil’s holding is persuasive in this case:

Implying such a waiver [for one document, in a situation where the agency only disclosed related documents,] could tend to inhibit agencies from making any disclosures other than those explicitly required by law because voluntary release of documents exempt from disclosure requirements would expose other documents ... [to] disclosure. An agency would have an incentive to refuse to release all exempt documents if it wished to retain an exemption for any documents.

Mobil, 879 F.2d at 701.

Florida does not contend that the Secretary released the adjusted block level data in the final opinion; rather, it merely argues that the Secretary disclosed to the public related material. The district court did not err in finding no waiver.

C. Exemption 5

1. Deliberative

As stated above, the Government may withhold inter-agency documents under the deliberative process privilege only if that material is both predecisional and deliberative. For Exemption 5 purposes, the term “deliberative” began its definitional journey in EPA v. Mink, 410 U.S. 73, 93 S.Ct. 827, 35 L.Ed.2d 119 (1973).' There, the Court acknowledged, and indeed explicitly adopted for Exemption 5 purposes, the distinction between “factual” and “opinion” data already recognized by trial courts in ordinary discovery situations. “Exemption 5 ... requires different treatment for material reflecting deliberative or policy-making processes on the one hand, and purely factual, investigative matters on the other.” Id. at 89, 93 S.Ct. at 837. The Court issued this mandate despite its acknowledgment that the task of drawing a line between what is fact and what is opinion can at times be frustrating and' perplexing. This is partly borne out by its detailed discussion of what a court should do in situations in which a memo contains both fact and opinion or when a memo consists of factual material that is intertwined with policy making processes. Id. at 90-91, 93 S.Ct. at 837-38.

While it was delineating the contours of the privilege in NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 95 S.Ct. 1504, 44 L.Ed.2d 29 (1975), the Supreme Court implicitly reaffirmed this fact/opinion distinction by noting that the focus of the deliberative process privilege was “on documents reflecting advisory opinions, recommendations and deliberations....’” Id. at 150, 95 S.Ct. at 1516. Later, in Federal Open Mkt. Comm. v. Merrill, 443 U.S. 340, 99 S.Ct. 2800, 61 L.Ed.2d 587 (1979), the Court explained that the purpose of the privilege for predecisional deliberations is to ensure that a decisionmaker will receive the unimpeded advice of his associates.

It seems safe to say then that the Supreme Court has never backed away from the fact/opinion distinction first announced in Mink, although some courts, including the district court in the present case, have questioned its continuing vitality. See, e.g., National Wildlife Fed’n v. United States Forest Serv., 861 F.2d 1114, 1118-20 (9th Cir.1988) (disposing of the factual/opinion semantics debate and adopting a “process-oriented” test 8 ); Wolfe v. Department of Health and Human Services, 839 F.2d *948 768, 774 (D.C.Cir.1988) (“[T]his court cannot mechanically apply the fact/opinion test.”).

The dissatisfaction with this bright line rule evidently arose because of situations where the disclosure of so called “facts” would nevertheless reveal an agency’s deliberative process. But as the Court of-Appeals for the District of Columbia recognizes: The fact/opinion distinction continues to offer a “quick, clear, and predictable rule of decision for most cases.” Quarles v. Department of Navy, 893 F.2d 390, 392 (D.C.Cir.1990).

After noting the uncertain state of the law surrounding the issue of whether a court should consider the policies behind the FOIA when deciding whether or not a document is deliberative, the district- court disregarded the fact/opinion distinction and chose to “go beyond the literal definition” of deliberative to determine “whether [the] granting [of the] privilege [would] fulfill the intent of the privilege.” This approach included the consideration of both the Department’s voluntary and involuntary disclosures. -To support its method of analysis, the district court relied on language culled from Wolfe v. Department of Health and Human Services, 839 F.2d 768 (D.C.Cir.1988), namely, that “courts must examine the information requested in light of the policies and goals that underlie the deliberative process privilege.” Id. at 774.

For a number of reasons, we believe the district court’s approach was incorrect. First, although the court in Wolfe did not “mechanically apply the fact/opinion test,” the opinion implicitly follows the distinction’s dictates and remains true to the ultimate goal of the test, namely, determining “whether [the .information] reflects the give-and-take of the consultative process.” Id. at 774 (quoting Coastal States Gas Corp. v. Department of Energy, 617 F.2d 854, 866 (D.C.Cir.1980). After examining the information within the context of the predecisional process, it became clear to the court in Wolfe that the information was protected partly because it was an opinion:

The information would disclose that proposals have been made, and that these preliminary recommendations have been accepted or rejected.... This court has previously noted that the deliberative process privilege ... serves a. number or purposes among which are the protection of subordinates’ willingness to provide decisionmakers with frank opinions and recommendations....

Wolfe, 839 F.2d at 774-75 (emphasis added).

Second, the district court’s additional inquiry into the underlying policies of the FOIA (which, significantly, included consideration of the Secretary’s prior disclosures) does not follow from the Wolfe decision. Wolfe did not involve a governmental agency which, prior to plaintiff’s request, had voluntarily disclosed what would have otherwise been exempt information. In essence, the district court employed a “pseudo-waiver” approach which would have us distinguish between this case and a hypothetical case in which the Department had chosen not to voluntarily disclose the reasoning behind its rejected recommendation. If we were to distinguish between these two scenarios for this analysis, the very applicability of Exemption 5 would rest on the magnanimity of the governmental agency. This approach seems to us a poor substitute for a well-delineated, narrowly construed waiver analysis.

Third, in determining whether the information was deliberative, the Wolfe court did not venture into a fact-finding thicket to determine whether, in the specific circumstances of its case, the requested disclosure would “chill discussion” within the agency. Nor should it have for that matter. Instead, its analysis examined the documents in the context in which the documents were used, or put another way, within the context of the agency’s predeci-sional approval process. In fact, in the later case of Quarles v. Department of Navy, 893 F.2d 390 (D.C.Cir.1990), the D.C. Court expressed its lingering skepticism of whether a fact-specific “chilling inquiry” is “necessary.to support nondisclosure under Exemption 5.” Id. at 393.

The approach employed by the district court necessarily examines the very same *949 policies recognized by the Supreme Court in Mink; policies that were clearly considered by the Court when it adopted the Exemption 5 fact/opinion distinction in the first place. Undertaking this inquiry for a second time renders the fact/opinion standard useless. Stated another way, the Supreme Court has held that if, looking at the deliberative process as a whole, data can be deemed advice or opinion, it is necessarily deliberative. To engage in such an additional analysis contradicts the very premise that the distinction relies on, namely, that data classified as advice or opinion necessarily involves the deliberative process. See Chilivis v. SEC, 673 F.2d 1205, 1211 (11th Cir.1982) (“The exemption has generally been held to encompass ... staff opinions/recommendations that are part of an agency’s deliberative process.”); Sears, 421 U.S. at 151, 95 S.Ct. at 1516-17 (“The quality of a particular agency decision will clearly be affected by the communications received by the decisionmaker on the subject of the decision prior to the time the decision is made.”).

Essentially, the Supreme Court has set up an entire class of data which is per se deliberative. It defies reason, as well as Supreme Court precedent, to then go back and weigh the policies underlying the distinction to decide whether disclosure would in fact discourage frank discussion in some specific case. The only inquiry that should be made in deciding whether something should be denoted opinion, and hence deliberative, is: Does the information reflect the give-and-take of the consultive process? Coastal States, 617 F.2d at 866; see also Sears, 421 U.S. at 150, 95 S.Ct. at 1516 (courts should “focus on documents ‘reflecting advisory opinions, recommendations and deliberations comprising part of a process by which governmental decisions and policies are formulated.’ ”).

Lastly, the Supreme Court has pointed out time and time again that the FOIA exemptions “are plainly intended to set up concrete, workable standards for determining whether particular material' may be withheld or must be disclosed.” Mink, 410 U.S. at 79, 93 S.Ct. at 832. See also Grolier, 462 U.S. at 27, 103 S.Ct. at 2214 (“[B]y establishing a discrete category of exempt information, [we implement] the congressional intent to provide ‘workable’ rules.”); Julian, 486 U.S. at 13, 108 S.Ct. at 1614. The fact/opinion distinction continues to be an efficient and workable standard for separating out what is, and what is not, deliberative.

2. Fact or Opinion

With this in mind, we turn now to the question of whether the adjusted block level data are factual or opinion. 9 The facts of this case are undisputed. Pursuant to congressional mandate, the Secretary was directed to take a census in any form or manner he wished. He had choices. He first ordered that a census be taken in a manner we have referred to as an actual headcount. He then asked others at the Department to devise a statistical methodology to estimate the country's population. This led to the creation of a statistical model which someone other than the ultimate decisionmaker thought would best represent this country’s true population. After conducting another partial headcount, a subordinate plugged these results into a host of equations based on numerous assumptions and out emerged a numerical recommendation. This proposal was presented to the Secretary to replace the initial headcount numbers. The advice was rejected for reasons not relevant to our analysis.

In light of the foregoing, we have no problem characterizing the final adjusted block level data as a proposal or a recom *950 mendation that eventually was rejected by the person in charge. See Pies v. IRS, 668 F.2d 1350, 1353 (D.C.Cir.1981) (holding that draft policy options which are ultimately rejected are protected from disclosure under the deliberative process privilege).

One reason the information in this case is difficult to classify is because the data are numbers. The District of Columbia Court of Appeals best illustrated the amorphous personality of numbers in Quarles v. Department of Navy, 893 F.2d 390 (D.C.Cir.1990):

Numbers have a surface precision that may lead the unsophisticated to think of them as fixed, and of course some are— Waterloo was fought in 1815. But cost estimates such as these are far from fixed, as anyone knows who has had two contractors bid on a home improvement or has compared budget estimates with final costs of a government project. They derive from a, complex set of judgments — projecting needs, studying pri- or endeavors and assessing possible suppliers. They partake of just that elasticity that has persuaded courts to provide shelter for opinions generally.

Id. at 392-93 (emphasis added).

The adjusted numbers in our case were a recommendation — nothing more than one rejected recommendation in a process which ultimately led to the Secretary accepting a different recommendation as his final decision. The advice may have been in the form of numbers but they were no more “facts” than were the numbers in Quarles. Wé therefore conclude that the adjusted census data should be characterized as opinion.

The district court attempted to distinguish 'Quarles in two ways. First, it could not understand how disclosure of the data would chill discussion in light of the Secretary’s candor. As we stated above, the Secretary’s candor is not relevant in determining whether material is deliberative.

Second, the district court relies on the reasoning of Assembly of the State of California v. United States Dep’t of Commerce, No. S-91-990 (E.D.Cal. Feb. 7, 1992), where in similar circumstances, a court distinguished Quarles by noting that the information in Quarles was sought pri- or to the decisionmaker’s final decision. The district court pointed out that in our case Florida requested the data after the Secretary’s final decision was published. Even assuming that the court’s reading of Quarles is correct, the Supreme Court has nevertheless addressed this point:

The purpose of the privilege for prede-cisional deliberations is to insure that a decisionmaker will receive the unimpeded advice of his associates. The theory is that if advice is revealed, associates may be reluctant to be candid and frank. It follows that documents shielded by executive privilege remain privileged even after the decision to which they pertain may have been effected, since disclosure at any time could inhibit the free flow of advice, including analysis, reports, and expression of opinion within the agency.

Open Market, 443 U.S. at 359-60, 99 S.Ct. at 2812 (emphasis added).

IV. CONCLUSION

“Exemption 5, properly construed, calls for ... the withholding of all papers which reflect the agency’s group thinking in the process of working out its policy and determining what its law shall be.” Sears, 421 U.S. at 153, 95 S.Ct. at 1517-18. Because the adjusted census block level data are a subordinate’s opinion and reflect the give- and-take of the deliberative process, we hold that the data are' deliberative, and in turn, within the scope of the deliberative process privilege. We also conclude that the district court did not err in finding that the Department did not waive its privilege. It follows that the data are exempt from disclosure pursuant to 5 U.S.C. § 552(b)(5).

Summary judgment for the Plaintiff is REVERSED and the case is REMANDED to the district court with instructions to enter judgment for the Defendant.

REVERSED and REMANDED.

Notes

1

. In 1987, the Department had determined that no statistical adjustment would be made for the 1990 census. In response to a subsequent lawsuit, the Department stipulated that it would reconsider its determination. The disputed methodology and adjusted figures at issue are the product of that reconsideration.

2

. Census blocks are a relatively small geographical unit compared to a city or county unit.

3

. Other privileges protected by Exemption 5 include: (1) the attorney work product privilege, FTC v. Grolier, Inc., 462 U.S. 19, 103 S.Ct. 2209, 76 L.Ed.2d 387 (1983); (2) the confidential commercial information privilege, Federal Open Mkt. Comm. v. Merrill, 443 U.S. 340, 99 S.Ct. 2800, 61 L.Ed.2d 587 (1979); and (3) the Machin v. Zukert, 114 U.S.App.D.C. 335, 316 F.2d 336 (1963) privilege, United States v. Weber Aircraft Corp., 465 U.S. 792, 104 S.Ct. 1488, 79 L.Ed.2d 814 (1984).

4

. Nevertheless, we feel it necessary to address one of the conclusions reached by the district court that indirectly raises the predecisional issue:

[The] FOIA requires that each federal agency make available for public inspection the general methodology, as well as statements of policy and interpretation, that have been adopted by the agency. In many cases, draft documents may not constitute policy interpretations, but in this case, the defendant published and used many of the adjusted data figures to explain its decision not to adjust the data. Since the figures were used in the Secretary’s explanation not to adjust, these take on a greater significance than mere “rejected data,” and rise to the level of a policy interpretation. Since these, data are not protected by the deliberative process privilege, they are, under FOIA, to be made available to the public.

Apparently, this conclusion was an outgrowth of the district court’s "deliberative analysis.” Implicitly, the conclusion represents a finding by the court that the adjusted data are no longer predecisional, even though the district court expressly noted that it was not reaching that question.

The Supreme Court holds that “if an agency chooses expressly to adopt or incorporate by reference an intra-agency memorandum previously covered by Exemption 5 in what would otherwise be a final opinion, that memorandum may not be withheld under Exemption 5.” Sears, 421 U.S. at 161, 95 S.Ct. at 1521-22; Grumman, 421 U.S. at 184-85, 95 S.Ct. at 1500.

Permitting an agency to withhold such information raises the specter of "secret agency law” — law which Congress finds repugnant to the ideals of an open society. Sears, 421 U.S. at 153, 95 S.Ct. at 1518 (The FOIA "represents a strong congressional aversion to 'secret agency law.’ ”). In this vein, the Act’s purpose is to require disclosure of information which has "the force and effect of law.” Id. (citation omitted).

In light of the above, data "expressly adopt[ed] or incorporate^] by reference” means predecisional deliberative material which is adopted and approved by the agency as its "effective law and policy." Id. 95 S.Ct. at 1517. By expressly adopting the reasoning of her subordinate, the decisionmaker has in effect converted a rejected proposal into the rationale for the agency’s working law. As a consequence, the documents are no longer considered predecisional for they now support and explain the agency’s position in the same manner a postdecisional document explains an agency decision.

In applying this principle to the adjusted figures, one discovers that the circumstances of this case are quite unusual. In one respect, the adjusted data could be viewed as explaining the Department's final decision not to adjust the actual census figures. Indeed, the Department chose to explain its subordinate’s methodology, and went as far as revealing some of the adjusted figures themselves.

But as stated above, this case is unusual. Agencies do not generally issue final opinions solely for the purpose of rejecting a subordinate's proposal. Generally speaking, the adoption of a proposed policy implicitly rejects all others. Yet, no one would contend that every time an agency adopts a policy, it must reveal all the information relating to rejected proposals, since it is that very information that is protected by the predecisional deliberative privilege.

It is therefore necessary to go past the form in this case and ask the following question: What is the Department's final policy determination on the census? The answer, of course, is the adoption of the reasoning of the actual census headcount, and the numbers derived therefrom. The rejected adjusted figures and the methodology used to calculate them, therefore, have never been adopted, even implicitly, by the Department as its rationale for adopting the actual census headcount as its final census. See Access Reports v. Department of Justice, 926 F.2d 1192, 1197 (D.C.Cir.1991) (“The Court has refused to equate reference to a report’s conclusions with adoption of its reasoning, and it is the latter that destroys the privilege."); Taxation with Repr esentation Fund v. IRS, 646 F.2d 666, 678 (D.C.Cir.1981) (holding that “a document that is predecisional at the time of preparation may lose exempt status if 'adopted, formally or informally, as the agency position on an issue....’”) *946 (quoting Coastal States Gas Corp. v. Department of Energy, 617 F.2d 854, 866 (D.C.Cir.1980); Shermco Indus. v. Secretary of the Air Force, 613 F.2d 1314, 1319 (5th Cir.1980) (noting that a predecisional communication is no longer exempt from disclosure once the agency adopts its employee’s advice as its own).

Since the adjusted block level data have never been adopted by the agency, the policy of banning secret agency law is not implicated. The Department simply is not yet using the adjusted numbers in a way that can be described as having the force and effect of law. Summing up, we do not find that the Department “expressly adopted" the reasoning and conclusions associated with the adjusted census figures as its own, as that expression is used in Sears and Grumman.

5

. Because our "deliberative" discussion, infra, touches upon issues of waiver, we feel it better to first address the issue of waiver.

6

. “It is unfortunate that you chose to withhold even a useful amount of the data until the subcommittee was forced to take the extraordinary step of issuing a subpoena.... I also conclude ... that the Department now recognizes the authority of Congress to use the information in any manner it deems appropriate.” R. 2-45.

7

. The Department also maintains that a disclosure to Congress never waives the Exemption 5 privilege. In light of the forced nature of the disclosure, we see no need to address this additional argument.

8

. Ultimately, the court in Wildlife concluded that, "documents containing nonbinding recommendations on law or policy would continue to remain exempt from disclosure" even under its newly adopted functional test. Wildlife, 861 F.2d at 1119 (emphasis added).

9

. In contrast to facts, there is good reason to protect opinions and advice.- By its nature, a subordinate's imprimatur is attached to his or her advice, and hence puts the subordinate out on a limb. We realize that a subordinate’s factual compilation also carries with it his seal of approval. Nevertheless, we believe that there is a significant difference in the type and degree of risk associated, on the one hand, with providing an opinion to a superior, and on the other, supplying factual data to that same person. Clearly, it was the former that Congress ultimately wanted to protect.