American Oversight v. U.S. Department of Justice

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

AMERICAN OVERSIGHT, Plaintiff,

v. Case No. 18-cv-0319 (CRC)

UNITED STATES DEPARTMENT OF JUSTICE,

Defendant.

MEMORANDUM OPINION

In November 2017, American Oversight filed a Freedom of Information Act (“FOIA”)

request that sought records of any guidance Department of Justice (“DOJ”) officials had provided to United States Attorney John W. Huber in his investigation of certain issues related to the 2016 presidential election. Rather than immediately conduct an electronic search for documents, DOJ’s Office of Information Policy (“OIP”) staff reached out to a liaison in the Office of the Attorney General (“OAG”) to communicate the request. This liaison subsequently discussed the FOIA request with senior leadership within OAG and the Office of the Deputy Attorney General (“ODAG”), and eventually with Huber himself. Exactly what was said in these conversations is unknown, but they led to an initial determination that no responsive records existed—largely because the individuals interviewed maintained that the guidance American Oversight sought was never reduced to writing but was instead communicated orally.

And yet, when Huber finally saw for himself the reply brief DOJ was set to file in this case—and presumably the actual text of American Oversight’s request—he readily located a responsive email record. DOJ thereafter conducted a supplemental email search, which did not turn up any additional responsive documents. The agency now claims it has fulfilled its FOIA

2

obligations and moves for summary judgment. American Oversight says the agency’s search was inadequate in several ways. It further contends that the DOJ’s initial person-to-person search was carried out in bad faith, and asks the Court’s permission to conduct limited discovery, unusual as it may be in FOIA cases, to better examine the agency’s search efforts. After careful review of the parties’ submissions, and for the reasons that follow, the Court will grant DOJ’s motion for summary judgment and deny American Oversight’s cross-motion for summary judgment and its request for discovery. I. Background In July 2017, twenty members of the House Judiciary Committee, including Chairman Robert Goodlatte, sent a letter to then-Attorney General Jeff Sessions and then-Deputy Attorney General Rod Rosenstein asking them “to appoint a second special counsel to investigate a plethora of matters connected to the 2016 election and its aftermath.” Declaration of Vanessa R. Brinkmann (“First Brinkmann Decl.”), Ex. A, ECF No. 16-3, at 12. Congressman Goodlatte and his cosigners expressed the view that the then-existing special counsel investigation into the 2016 election focused too narrowly on “Russian influence on the election and connections with the Trump campaign.” Id. at 12. To correct for this purported “[p]olitical gamesmanship,” they asked that a second special counsel be appointed to investigate “actions taken by previously public figures like Attorney General Loretta Lynch, FBI Director James Comey, and former Secretary of State Hillary Clinton.” Id. Congressman Goodlatte and 13 other House Judiciary members followed up with a second, similar letter in September 2017. Id. at 19.

In November 2017, the Assistant Attorney General for the Office of Legislative Affairs, Stephen E. Boyd, responded to Congressman Goodlatte and his colleagues’ request. Id. at 23. Boyd reported that “the Attorney General has directed senior federal prosecutors to evaluate

3

certain issues raised in your letters” and to “make recommendations as to whether any matters not currently under investigation should be opened [and] whether any matters currently under investigation require further resources[.]” Id.

Roughly one year later, American Oversight filed four FOIA requests related to DOJ’s decision to act on Congressman Goodlatte’s invitation. See First Brinkmann Decl., Ex. B, at 30– 71. Three of these requests have been resolved by the parties during the course of this litigation: the “Drafting FOIA” request seeking records “relating to the drafting” of DOJ’s response to the Goodlatte letters, Compl., ECF No. 1 ¶ 18; the “Prosecutors FOIA” request seeking identification of those “senior federal prosecutors” DOJ tasked to “evaluate certain issues” raised in the Goodlatte letters, id. ¶ 10; and the “Recusal FOIA” request seeking records reflecting “any analysis of government or legal-ethics issues or evaluating any recusal obligations” related to former Attorney General Session’s participation in the matter, id. ¶ 7.

American Oversight’s fourth FOIA request presents the sole remaining battleground in this case. See First Brinkmann Decl., Ex. B, at 66–71. In what the parties call the “Guidance FOIA,” American Oversight sought

[a]ll guidance or directives provided to the senior federal prosecutors who have been directed to evaluate certain issues raised in Congressman Robert Goodlatte’s letters, as indicated in the Department of Justice’s November 13, 2017 response signed by Assistant Attorney General Stephen Boyd, attached for your convenience, regarding their performance of that task.

Id. at 67 (internal quotation marks and alteration omitted). The “senior federal prosecutors” turned out to be just one federal prosecutor, namely the United States Attorney for the District of Utah, John W. Huber. First Brinkmann Decl., Ex. B, at 27. The request, therefore, sought all “guidance or directives” that had been given to Huber regarding the Goodlatte investigation.

4

Upon receiving the request, OIP “determined that a direct inquiry to knowledgeable staff in [OAG] would be the most logical and effective search method.” First Brinkmann Decl. ¶ 14. OIP consulted with the Counselor to the Attorney General in OAG to “ascertain . . . what guidance or directives, if any, were issued.” Id. The Counselor to the Attorney General “then conferred with other Department officials with direct knowledge of the subject matter,” id., who reported that “no written guidance or directives were issued to Mr. Huber in connection with” his assignment to investigate the matters raised in Congressman Goodlatte’s letters, id. ¶ 15.

Based on those representations, OIP concluded that “further searches would be unlikely to identify records relevant to [American Oversight’s] request,” id., and moved for summary judgment, see Def’s Mot. Summ. J., ECF No. 16. American Oversight, in turn, filed its own motion for summary judgment, and briefing in the case continued. See Pl’s Cross-Mot. Summ. J., ECF No. 18.

The twist in this saga occurred on the day DOJ’s reply brief was due, February 28, 2019.

Second Declaration of Vanessa R. Brinkmann (“Second Brinkmann Decl.”), ECF No. 26-1, ¶ 15. While reviewing DOJ’s draft reply brief, Huber “brought to the attention of the then-ODAG point of contact that it appeared that drafters of the brief were not aware of the letter to him from Attorney General Sessions at the initiation of his assignment.” Id. The letter was delivered to Huber as an attachment to a November 22, 2017 email from Matthew Whitaker, then serving as Attorney General Sessions’ Chief of Staff. Id. When the email made its way to OIP, “[i]t was immediately apparent that this material was responsive to the Guidance Request”—so DOJ that same day notified American Oversight of its discovery and asked this Court for more time to determine whether additional responsive records existed. Id.

5

Recognizing that the discovery of the email “raised questions about the adequacy” of the agency’s initial search, DOJ decided to search “the emails of the senior leadership officials overseeing Mr. Huber’s evaluation,” including Sessions, two of Sessions’ personal assistants, Whitaker, Rosenstein, and former Principal Associate Deputy Attorney General Robert Hur. Id. ¶ 17. In this search, DOJ used keywords to identify any correspondence that was transmitted to Huber, or on which Huber was copied, in light of the agency’s interpretation of the Guidance FOIA request as seeking “[a]ll guidance or directives [actually] provided” to Huber. Id. ¶ 20.1 The email search did not turn up any responsive documents beyond the attachment Huber had already located and which had already been produced to American Oversight. Id. ¶ 26. Believing it had discharged its duties under FOIA, DOJ reasserted its call for summary judgment. See Def.’s Reply in Supp. of Def.’s MSJ (“Def.’s MSJ Reply”), ECF No. 26, at 10.

American Oversight did not wait to learn of the fruits of this supplemental search.

Instead, seizing on DOJ’s error, it promptly moved to stay summary judgment briefing and demanded discovery. See Pl’s Mot. Stay Summ. J. Briefing & Mot. Discovery (“Pl’s Mot. Discovery”), ECF No. 25-1. In asking for discovery, American Oversight intones that DOJ’s late-in-the-game discovery of the responsive letter “raised grave questions regarding both the thoroughness of [DOJ’s] search and whether [DOJ] has complied in good faith with its duties under FOIA.” Id. at 7. It insists that discovery is necessary to determine “whether the misrepresentations contained in the initial declaration resulted from negligence or a lack of good faith” by DOJ in its response to the Guidance FOIA. Decl. of Counsel in Supp. Mot. Discovery, ECF No. 25-2, ¶ 40. In addition to its discovery demand, American Oversight maintains that the

1 Further details about this email search—and about the agency’s overall search effort—

are discussed in Part IV of this opinion.

6

agency’s search was inadequate for a variety of reasons and argues that it, not the agency, is entitled to summary judgment. See Pl’s Reply in Supp. of Cross-Mot. for Summ. J (“Pl’s MSJ Reply”), ECF No. 29.

With briefing now complete, the motions are ripe for the Court’s resolution.

II. Legal Standards FOIA cases are typically resolved on summary judgment. See Brayton v. Office of U.S.

Trade Rep., 641 F.3d 521, 527 (D.C. Cir. 2011). Summary judgment is warranted if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a).

When faced with a challenge to the adequacy of its FOIA search, the agency must demonstrate “beyond material doubt that its search was reasonably calculated to uncover all relevant documents.” Ancient Coin Collectors Guild v. U.S. Dep’t of State, 641 F.3d 504, 514 (D.C. Cir. 2011) (internal quotation marks omitted) (quoting Valencia-Lucena v. U.S. Coast Guard, 180 F.3d 321, 325 (D.C. Cir. 1999)). The adequacy of an agency’s search depends on the reasonableness of its methods, not the quantity or quality of documents it unearths. See CREW v. U.S. Gen. Servs. Admin., No. 18-cv-377, 2018 WL 6605862, at *2–3 (D.D.C. Dec. 17, 2018). An agency must show that it made “a good faith effort to conduct a search for requested records, using methods which can be reasonably expected to produce the information requested.” Oglesby v. U.S. Dep’t of Army, 920 F.2d 57, 68 (D.C. Cir. 1990). That showing can be made through declarations that detail “what records were searched, by whom, and through what process.” Steinberg v. U.S. Dep’t of Justice, 23 F.3d 548, 552 (D.C. Cir. 1994). Agency declarations are “accorded a presumption of good faith” and “cannot be rebutted by purely

7

speculative claims about the existence and discoverability of other documents.” SafeCard Servs., Inc. v. SEC, 926 F.2d 1197, 1200 (D.C. Cir. 1991) (internal quotation marks omitted).

“Discovery in FOIA is rare and should be denied where an agency’s declarations are reasonably detailed, submitted in good faith and the court is satisfied that no factual dispute remains.” Schrecker v. U.S. Dep’t of Justice, 217 F. Supp. 2d 29, 35 (D.D.C. 2002), aff’d, 349 F.3d 657 (D.C. Cir. 2003). Because an “agency may establish the adequacy of its search by submitting reasonably detailed, nonconclusory affidavits describing its efforts,” Baker & Hostetler LLP v. U.S. Dep’t of Commerce, 473 F.3d 312, 318 (D.C. Cir. 2006), district courts “ha[ve] discretion to forego discovery and award summary judgment on the basis of affidavits,” Freedom Watch, Inc. v. Nat. Sec. Agency, 783 F.3d 1340, 1345 (D.C. Cir. 2015) (quoting Goland v. CIA, 607 F.2d 339, 352 (D.C. Cir. 1978)) (internal quotation marks omitted). “Even if an agency’s affidavits regarding its search are deficient, courts generally do not grant discovery but instead direct the agency to supplement its affidavits.” Flowers v. IRS, 307 F. Supp. 2d 60, 68 (D.D.C. 2004) (citing Judicial Watch, Inc. v. Dep’t of Justice, 185 F. Supp. 2d 54, 65 (D.D.C. 2002)). III. Analysis American Oversight contends that DOJ’s efforts to process the Guidance FOIA request was technically flawed and carried out in bad faith. It therefore argues that the Court should not only order a supplemental search, but should also grant American Oversight the right to conduct limited discovery to probe DOJ’s compliance with its FOIA obligations. While American Oversight’s bad-faith argument directly undergirds its request for discovery, it also colors its attack on the adequacy of the agency’s search. Because that issue pervades the entire case, the Court begins its analysis there. Beyond the bad-faith question, American Oversight identifies

8

four discrete deficiencies in DOJ’s search. First, it alleges that DOJ failed to identify, consult, or search all relevant custodians. Second, it claims that DOJ did not search all locations or systems of records where responsive files might exist. Third, it argues that DOJ adopted an interpretation of the Guidance FOIA that is “improperly cramped” and inconsistent with the general rule that FOIA requests are to be liberally construed. Finally, it says that DOJ cabined its search to an artificially truncated timeframe. Pl’s MSJ Reply at 1. The Court will turn to those arguments second.

A. Whether DOJ’s Declarations Warrant Deference To support the adequacy of its searches in response to the Guidance FOIA, DOJ provides two declarations from Vanessa R. Brinkmann, Senior Counsel in OIP. The first declaration sets forth the rationale for, and mechanics of, the initial person-to-person oral inquiry. The second declaration, meanwhile, explains how DOJ’s eleventh-hour discovery of the Whitaker email spurred it to conduct a broader email search, and then details the specifics of that search. Because American Oversight alleges the agency responded to its FOIA request in bad faith, the Court will recount DOJ’s process in more detail than is usual in the typical FOIA case. Some of what follows was set forth in Part I of this opinion, but bears repeating here.

Start with the initial search, which the agency conducted before it filed its motion for summary judgment. OIP determined the most effective response to the Guidance FOIA was a “direct inquiry to knowledgeable staff in the Office of the Attorney General (OAG)” to, first, identify whom the Attorney General directed to evaluate the issues raised in Congressman Goodlatte’s letters and, second, find out “what guidance or directives, if any, were issued.” First Brinkmann Decl. ¶ 14. After relaying the request to the Counselor to the Attorney General, who in turn conferred with “other Department officials with direct knowledge of the subject matter,”

9

OIP quickly learned the answer to the first question: Huber was the only federal prosecutor assigned to the Goodlatte matter. Id.

As for the second question, OAG reported that “no written guidance or directives were issued to Mr. Huber in connection with this directive, either by the Attorney General, or by other senior leadership office staff.” Id. ¶ 15. Instead, OAG “advised that details of Mr. Huber’s direction were addressed orally, in meetings and discussions among a small group of Department officials, including the Attorney General, the Deputy Attorney General, the OAG Chief of Staff, the Principal Associate Deputy Attorney General,” and Huber himself. Id. “[T]he lack of written guidance or directives was confirmed by OAG, pursuant to internal OAG discussions as well as discussions with Mr. Huber himself.” Id. Notwithstanding these representations, OIP, “as a supplementary measure . . . took the additional step of conferring with the Office of the Deputy Attorney General (ODAG) regarding the information provided to [it] by OAG” and “concluded that the OAG information regarding the lack of written guidance or directives to Mr. Huber was adequate and that further searches would be unlikely to identify” responsive records. Id.

According to OIP, other clues lent credence to the claim that any guidance given to Huber was oral rather than written. See id. ¶ 16. For one, the belief was “confirmed over the course of several discussions between OIP and OAG.” Id. For another, “nothing in the records reviewed by OIP”—apparently while fulfilling other American Oversight FOIA requests— “indicated the existence of written guidance or directives issued to Mr. Huber.” Id. Thus, Brinkmann declared that, “[i]n light of the clear comprehensive, and conclusive information [she] received directly from OAG, and the lack of any indication that other records responsive to

10

Plaintiff’s [other requests], [she] determined that no additional searching was necessary in this instance.” Id.

DOJ held firm to that representation—until it didn’t. The day its reply brief in support of its motion for summary judgment was due, Huber reviewed a draft of the brief and “brought to the attention of the then-ODAG point of contact that it appeared that drafters of the brief were not aware of the letter to him from Attorney General Sessions at the initiation of his assignment.” Second Brinkmann Decl. ¶ 15. The letter was attached to an email sent to Huber by Sessions’ then-Chief of Staff Whitaker. Id. Huber “promptly forwarded this letter to the ODAG point of contact, who forwarded it to [Brinkmann].” Id. Once Brinkmann determined that it was responsive to the Guidance FOIA request, she “immediately contacted counsel of record to notify him of the discovery of this responsive material.” Id. DOJ then “promptly notified [American Oversight] that evening of the discovery” and petitioned the Court for additional time to determine whether other responsive records existed and to revise its reply brief. Id.

Because the letter belied OIP’s initial conclusion that Huber had received no written guidance, it decided that a further search was in order. Id. ¶ 17. Brinkmann “determined that the most appropriate response would be to conduct a complete electronic search for records responsive to [American Oversight’s] request[.]” Id. The search had three components. First, DOJ “search[ed] the emails of the senior leadership officials overseeing Mr. Huber’s evaluation,” namely Sessions (and his two personal assistants), Whitaker, Rosenstein, and Hur. Id.; see also id. ¶ 20. To do so, OIP canvassed the “unclassified email of the designated OAG and ODAG custodians,” id. ¶ 18, for “any reference to ‘Huber’ in the ‘To,’ ‘From,’ ‘cc,’ or ‘bcc’ email message fields,” id. ¶ 20. The timeframe of the email search was from July 27, 2017 (the date of Congressman Goodlatte’s first letter to Sessions) to January 19, 2018, the date OIP began

11

its search for American Oversight’s Guidance FOIA request. Id. OIP reasoned that this email search was appropriately calibrated to American Oversight’s request “because the Guidance FOIA sought all guidance or directives provided to the senior federal prosecutors concerning the matters raised in Chairman Goodlatte’s letters, and any guidance or directives provided to Mr. Huber via email would reasonably include Mr. Huber as a party to the communication.” Id. (alteration and internal quotation marks omitted).

Second, OIP searched the Departmental Executive Secretariat (“DES”), which is the “electronic database of the official records repository of OAG and ODAG.” Id. ¶ 18. The DES “is used to track internal Department correspondence sent through formal channels, as well as certain external correspondence including Departmental correspondence with Congress.” Id. ¶ 22. OIP conducted the DES search “using several combinations of search parameters: (1) correspondence where Mr. Huber was listed as ‘contact’ (either sender or recipient); (2) correspondence where Attorney General Sessions was listed as a ‘contact’ and contained the keyword ‘Huber,’ and (3) a general keyword search for correspondence containing the keywords ‘Sessions’ and ‘Huber.’” Id. ¶ 23. The DES search timeframe was the same as the email search—July 2017 to January 2018. Id.

Third, and lastly, OIP conducted a “manual search of Mr. Huber’s mailbox.” Id. ¶ 18.

The parameters for this search were identical to the unclassified email search of the OAG and ODAG custodians—i.e., seeking correspondence between Mr. Huber and any of the DOJ officials who supervised him from July 2017 to January 2018. Id. ¶ 25. “An official in Mr. Huber’s office [ ] searched the active archive of Mr. Huber’s email account, consisting of the inbox, sent, and deleted items folders[.]” Id.

12

These searches turned up nothing new. Id. ¶ 26. DOJ therefore determined that the email attachment Huber had provided to DOJ after reviewing the draft reply brief, and which DOJ had promptly provided to American Oversight, was the lone record responsive to the Guidance FOIA request. Id.

American Oversight insists this sequence of events should dislodge the presumption of good faith courts normally accord to agency declarations in FOIA cases. See Pl’s Discovery Mot. at 17–18; Pl’s Discovery Reply at 3–6. While much of this argument comes in service of American Oversight’s request to stay summary judgment proceedings and conduct limited discovery, it also bears on whether—the discovery issue aside—the Court can sanction DOJ’s search. For if the agency’s declarations cannot be trusted, the Court would have little reason to think the agency conducted a search reasonably calculated to uncover all responsive records.

Central to American Oversight’s bad-faith argument is the discovery of the November 22 appointment letter after the first DOJ declaration averred that all guidance was given orally. See Pl’s Discovery Reply at 5. Because the senior DOJ officials who represented that Huber received only oral guidance had been “personally involved in crafting and sending” the appointment letter, American Oversight says “it is difficult to credit the possibility that these officials failed to recall” the letter “when questioned a few months later.” Id. The more likely explanation, American Oversight suggests, is that the officials were not cooperating in good faith with the FOIA request. Id. To further support its suspicion, American Oversight notes that the “delay in producing [the letter] prevented Attorney General Sessions and Acting Attorney General Whitaker from having to answer questions about the Attorney General Directive while in office”—since the letter was released to American Oversight four days after Whitaker left his

13

post as Acting AG. Id. at 6. Taking it all together, American Oversight submits this behavior warrants a finding of bad faith. The Court is unable to reach that conclusion.

Rare is the case that compels a court to cast a skeptical eye on an agency’s FOIA declarations—and this is not one of them. DOJ did an about face, to be sure. But the fact that an agency discovers an error in its earlier representations, and thereafter changes course, does not alone displace the good-faith presumption courts accord its declarations. In Military Audit Project v. Casey, 656 F.2d 724, 754 (D.C. Cir. 1981), a FOIA plaintiff argued that an agency’s admission that “it was initially in error” demonstrated that “the agency is fallible” and thus “its affidavits [were] suspect.” Id. As a result, the plaintiff contended, summary judgment could not be granted “on the basis of those affidavits alone.” Id. The D.C. Circuit “emphatically reject[ed] this line of argument.” Id. “If accepted,” the court reasoned, “it would work mischief in the future by creating a disincentive for an agency to reappraise its position, and when appropriate, release documents previously withheld. It would be unwise for us to punish flexibility, lest we provide the motivation for intransigence.” Id. Indeed, if anything, the D.C. Circuit has suggested the opposite reaction—bolstering the good-faith presumption—should follow from an agency change-in-position. The contrary view “is based on the perverse theory that a forthcoming agency is less to be trusted in its allegations than an unyielding agency.” Id.

Taking their cue from Military Audit Project, case after case in this jurisdiction has held that the focus should be on whether the agency admits and corrects error, not on the fact that it made a mistake in the first place. For example, in Meeropol v. Meese, 790 F.2d 942, 953 (D.C. Cir. 1986), the Circuit said that “what is expected of a law-abiding agency is that it admit and correct error when error is revealed.” And because the FBI did so in that case, the court found that its “additional releases suggest[ed] ‘a stronger, rather than a weaker, basis’ for accepting the

14

integrity of the search,” and rejected the plaintiffs’ bad-faith argument. Id. (quoting Military Audit Project, 656 F.2d at 754). Likewise, in People for the Ethical Treatment of Animals, Inc. v. Bureau of Indian Affairs, 800 F. Supp. 2d 173, 179 (D.D.C. 2011), a fellow district court rejected “plaintiff’s attempt to render defendant’s post-litigation efforts as immaterial (or as proof of bad faith),” noting the well-established principle “that subsequent production cannot serve as proof that the agency conducted an unreasonable search initially or acted in bad faith, for such a rule would punish those agencies that attempted to correct past inadequate searches.” There are plenty more cases where those came from. See, e.g., Tushnet v. U.S. Immigration & Customs Enf’t, 246 F. Supp. 3d 422, 432 (D.D.C. 2017) (Cooper, J.) (holding that “a series of errors and inconsistencies found across [the agency’s] multiple declarations” did not reflect bad faith in light of search’s complexity and fact “that the agency promptly rectified its mistake”); Competitive Enter. Inst. v. Office of Sci. & Tech. Policy, 161 F. Supp. 3d 120, 136 (D.D.C. 2016) (“Although the initial Leonard Declaration contained a significant error—asserting that none of the drafts were shared outside the Executive Branch—mistakes alone do not imply bad faith.”); Leopold v. Dep’t of Justice, 130 F. Supp. 3d 32, 42 (D.D.C. 2015) (“Although the Brinkmann Declaration contained errors, those errors are not indicative of bad faith. Mistakes alone do not imply bad faith.”); Fischer v. U.S. Dep’t of Justice, 723 F. Supp. 2d 104, 109 (D.D.C. 2010) (“To be sure, defendant has not performed its duties under FOIA perfectly, but error-free performance is not required.”)

In view of this authority, the Court concludes that DOJ remains entitled to the presumption of good faith. At bottom, American Oversight has predicated its bad-faith argument on a single error in the first Brinkmann declaration. The error was no doubt significant. And it may even raise some doubts about the sincerity of Mr. Whitaker’s apparent belief that no

15

guidance or directives had been reduced to record form, especially because Whitaker himself sent the email attaching the responsive document. But the error does not seriously impugn the credibility of the other DOJ officials or Huber himself—who all represented that any guidance or directives had been communicated orally—nor does it call into question OIP’s efforts to track down all responsive documents. Brinkmann’s supplemental declaration makes clear that OIP was aware of the existence of the Sessions appointment letter when it conducted its initial search and took steps to find it. See Second Brinkmann Decl. ¶¶ 9–13. But because those efforts uncovered only a draft of the letter that was not provided to Huber, OIP determined it was not responsive. See id. That OIP did not locate the final version of the appointment letter does not bespeak bad faith anymore than it does imperfect record-keeping practices. Id. ¶ 24 (explaining why letter was not stored in DES, which OIP searched). Moreover, if DOJ wished to conceal any guidance or directives given to Huber, it was a strange choice to ask Huber himself about any such records and to have him review its submissions in this case, which led to the discovery and production of the appointment letter. Thus, although the agency made a mistake, it has been forthright in acknowledging and correcting it, and other considerations undermine the suggestion that it has proceeded in bad faith. At the end of the day, as DOJ contends, “[a]n error alone, and one that the agency has acknowledged, explained, and corrected . . . is not sufficient to vitiate the presumption of good faith[.]” Defendant’s Opposition to Plaintiff’s Motion for Summary Judgment and Opposition to Plaintiff’s Motion for Discovery (“Def’s Opp.”), ECF No. 26, at 14.

American Oversight resists this conclusion by attempting to distinguish the cases discussed above. Pl’s Discovery Reply at 7–9. It asserts that the courts in those cases “declined to find questions about an agency’s good faith” because “the agencies conducted very complex searches and made typical, unremarkable errors in describing or conducting searches of such size

16

and complexity[.]” Id. at 7. Here, by contrast, the search called for was not similarly complex and DOJ’s mistake was neither minor nor clerical, so those precedents cannot bear on this case. But even assuming that American Oversight is correct that the decisions on which DOJ relies involved more onerous requests and searches than the one here, that does not take its argument very far. The reasoning of those cases did not turn primarily on the complexity of the searches required but rather on whether the government acknowledged its error and took appropriate steps to fix it. See, e.g., Meeropol, 790 F.2d at 953 (“[W]hat is expected of a law-abiding agency is that it admit and correct error when error is revealed.”); Tushnet, 246 F. Supp. 3d at 432 (emphasizing “that the agency promptly rectified its mistake” and “described [subsequent efforts] in a supplemental affidavit”); Fischer, 723 F. Supp. 2d at 109 (“[T]he agency’s cooperative behavior of notifying the Court and plaintiff that it had discovered a mistake, if anything, shows good faith.”).

Remember, the very reason Military Audit Project held that an initial mistake could not doom an agency’s declarations was because the Circuit wanted to encourage agencies to reflect and do better, not retreat into recalcitrance. 656 F.2d at 754 (“It would be unwise for us to punish flexibility, lest we provide the motivation for intransigence.”). A myopic focus on the reasonableness of the initial failures, divorced from the agency’s remedial efforts to cure them, would erode that principle. The upshot is that the rule applied in Military Audit Project, Meeropol, and Tushnet (and Fischer and Leopold and so on) is not limited to cases, as American Oversight appears to suggest, where the agency’s initial mistakes could be explained by having to perform particularly complex searches. To the contrary, an agency’s declarations will almost always remain entitled to a presumption of good faith so long as the agency acknowledges error and corrects for it—which is what DOJ did in this case.

17

That fact distinguishes the cases cited by American Oversight, in which courts granted some limited discovery. See Def’s Opp. at 16–18 (discussing cases cited by American Oversight). In CREW v. U.S. Dep’t of Veterans Affairs, 828 F. Supp. 2d 325 (D.D.C. 2011), for example, the agency acknowledged that a declarant “inadvertently had made inaccurate statements in his first two declarations,” id. at 321, but failed to “correct his prior inaccurate statements—or even suggest that they were inaccurate in any way,” id. at 333–34. The agency indicated that “another, unnamed VA employee would provide a declaration further clarifying” those inaccuracies, but that declaration came long after it would have been useful in a deposition of the first declarant who made the initial mistakes, and even that declaration contained deficiencies that “had to be supplemented” later on. Id. at 334. These “litigation tactics,” the court concluded, “rendered [the first declarant’s] deposition at best incomplete, and perhaps useless” and provided a ground for further discovery. Id. In another case on which American Oversight relies, Pulliam v. U.S. Envtl. Prot. Agency, 292 F. Supp. 3d 255, 260 (D.D.C. 2018), the court had ordered the Department of Defense (“DOD”) to search all of its electronic records after the agency indicated that it conducted a search for email records only. A DOD declarant thereafter averred that his original search request had in fact sought “all electronic files, not just emails,” but DOD “never once sought to supplement or correct [the initial] declaration” nor sought “reconsideration of the Court’s ruling on the grounds of mistake.” Id. at 261. Because the agency never acknowledged the apparent mistake and attempted to correct it, the court concluded that a discrepancy remained which “g[ave] rise to a question of fact that is best resolved by discovery.” Id. In contrast to both CREW and Pulliam, here the agency—without any prodding by either its adversary or the Court—acknowledged its error in the plainest of terms and took steps to correct it.

18

For a similar reason, Bangoura v. U.S. Dep’t of the Army, No. CIV.A. 05-0311, 2006 WL 3734164, at *6 (D.D.C. Dec. 8, 2006), does not help American Oversight. In that case, as here, the defendant found documents during second and third searches that it had not located in its first search. Id. at *5. But the similarity ends there. Although the agency in Bangoura produced the additional records, its declarations “d[id] not describe in reasonable detail the methods used in conducting its search, and fail[ed] to explain why more documents were found after the Plaintiff was told twice there were no more documents to be found.” Id. at *6. The court ordered discovery because the defendant “d[id] not provide any rationale for the delay in finding the requested documents and d[id] not explain why the search procedures were not adequate enough to find the documents during the initial search.” Id. Here, by contrast, DOJ did not simply provide the appointment letter bereft of any explanation for its late discovery; it instead explained in detail the basis for its initial search, why that search turned out to be inadequate, why it opted to conduct an electronic records search, and how that search was conducted. See Second Brinkmann Decl. ¶¶ 16–26. The DOJ erred, no doubt, but “[n]ot every error bespeaks a knave.” LaCedra v. Exec. Office for U.S. Attorneys, 317 F.3d 345, 348 (D.C. Cir. 2003).

In light of the foregoing analysis, the Court concludes that the Brinkmann declarations remain entitled to a presumption of good faith. This finding means different things for American Oversight’s discovery demand and its challenge to the adequacy of DOJ’s search. As for the discovery request, the Court now must deny it. Discovery is unusual in FOIA cases, see Schrecker, 217 F. Supp. 2d at 35, especially when the plaintiff has not established that the defendant has proceeded in bad faith, see Baker & Hostetler LLP, 473 F.3d at 318 (denying

19

discovery because plaintiff “offered no evidence of bad faith to justify additional discovery”). Because American Oversight has not done so here, it is not entitled to discovery.

That DOJ’s declarations remain entitled to a presumption of good faith does not, however, automatically mean that they describe an adequate search. The Court therefore now turns to the four shortcomings that American Oversight spies in DOJ’s searches.

B. American Oversight’s Specific Attacks on the Search 1. Search Custodians

American Oversight first contends that DOJ failed to “search[ ] or even consult[ ] all the custodians likely to be in possession of, or know about potentially responsive records.” Pl’s MSJ Reply at 5. Recall that, after Huber brought the November 22 appointment letter to DOJ’s attention, the agency decided to conduct a supplemental search of the email accounts of the senior DOJ officials who supervised Huber’s assignment to the Goodlatte matters. Specifically, DOJ searched the accounts of former Attorney General Sessions, former OAG Chief of Staff Matthew Whitaker, former Deputy Attorney General Rod Rosenstein, former Principal Associate Deputy Attorney General Robert Hur, and two of Sessions’ personal assistants, on the theory that they may have corresponded with Huber on Sessions’ behalf. See Second Brinkman Decl. ¶¶ 17, 20. DOJ also searched the DES, which serves as the official records repository for OAG and ODAG and maintains records of formal, controlled, unclassified correspondence sent to or from those offices. Id. ¶¶ 22–24.

American Oversight says this is not enough. It maintains that DOJ has “failed to even identify, much less consult or search, the set of officials most likely to know about the existence and location of potentially responsive records.” Pl’s MSJ Reply at 5. That is because DOJ did not identify every official who attended the “meetings and discussions” where Huber’s

20

assignment was discussed, id. (quoting First Brinkmann Decl. ¶ 15), including “first assistants and counselors to the Attorney General . . . who are most likely to possess the documentation surrounding their principals’ activities,” id. at 6. American Oversight says that it is “common sense, and commonplace knowledge to anyone familiar with how things operate at senior levels of major departments” that the staff of the senior leadership officials, not the officials themselves, would possess the records targeted by the Guidance FOIA request. Id. at 5.

When faced with a FOIA request, an agency must search all locations likely to contain responsive records. See Oglesby, 920 F.2d at 68. Generally, of course, “the agency is in the best position to determine custodians most likely to have relevant records.” Judicial Watch, Inc. v. U.S. Dep’t of Justice, 373 F. Supp. 3d 120, 127 (D.D.C. 2019) (alteration and internal quotation marks omitted). “But that truism—and the deference that accompanies it—does not insulate a search that overlooks locations where responsive materials are likely to be found.” Id. In Judicial Watch, for example, the plaintiff sought “any and all” correspondence related to former FBI Agent Peter Strzok’s assignment to Special Counsel Robert Mueller’s investigation and his subsequent reassignment. Id. at 127. This Court rejected as unreasonable the FBI’s decision to search only Agent Strzok’s email account “because one would reasonably expect that others besides Strzok exchanged emails on the topic.” Id. The Bureau’s declarant tried to justify its limited search by arguing that Strzok was the “named subject” of the request—but that was no answer because the request sought communication between high-level FBI personnel that was about Strzok’s assignment, not simply communication involving Strzok. Id. The Court therefore ordered that the FBI’s next search “shall include the email accounts of any of Agent Strzok’s superiors or other Bureau officials who were involved in the decision” to assign or reassign Strzok. Id.

21

A logical blunder like the Bureau’s in Judicial Watch thus might justify a search of additional custodians, notwithstanding an agency declaration that those custodians do not possess responsive records. But such an obvious mistake is absent here. As an initial matter, contrary to American Oversight’s assertion that the staff of the senior leadership officials (rather than the officials themselves) would possess and communicate records of any guidance given to Huber, the only record so far produced in this case was found in an email from Whitaker to Huber. That belies American Oversight’s claim and supports DOJ’s decision to search for correspondence between Huber and the senior officials, like Whitaker, who supervised his assignment. Moreover, simply claiming that it is “common sense” and “commonplace knowledge” that records would likely exist elsewhere—which, again, at least in this case does not appear to be true—is far from the specific evidence that is usually required to overcome an agency’s representations. See Shrecker, 217 F. Supp. 2d at 35 (rejecting argument that FBI should have searched files of two assistant directors because “[t]he FBI has stated that these files do not exist and Ms. Schrecker has not provided any evidence to refute the FBI’s assertion”); see also Judicial Watch, 373 F. Supp. 3d at 127 (ordering new search for text messages because plaintiff adduced evidence that the subject of the FOIA request sent text messages on topics included in the request).

American Oversight persists that it was unreasonable for DOJ to search the emails of Mr.

Sessions’ assistants but not the assistants of the other high-ranking DOJ officials who supervised Huber. Pl’s MSJ Reply at 6 (“If it is possible that Mr. Sessions’s personal assistants transmitted correspondence for him, it seems equally if not more plausible that other, more substantively involved staff to the ‘senior leadership officials’ . . . would have been tasked with sending Mr. Huber, or perhaps his first assistant, any follow-up emails or other guidance.”). The problem

22

with this argument is three-fold. First, as already explained, the one concrete clue DOJ had that at least some written guidance was given to Huber was found in an email from Whitaker to Huber—not in any communication between their subordinates. This undermines American Oversight’s claim that “more substantively involved staff” would communicate with Huber through their assistants rather than personally. Second, the supplemental email search of senior DOJ officials looked for any emails they had received from Huber, just as the manual search of Huber’s email looked for messages from Huber to his supervisors. See Second Brinkmann Decl. ¶¶ 20, 25. It seems unlikely that Huber would communicate with his superiors by emailing their assistants exclusively rather than emailing only the senior officials or at least including them as an additional recipient. Plus, the agency conducted these supplemental searches “with an eye toward identifying any ‘leads’ which would indicate other potential records custodians”—like the assistants about which American Oversight speculates—“but no such leads were identified.” Id. ¶ 21. That leads to the third problem with American Oversight’s argument: the fact that the agency nevertheless opted for good measure to include two of Sessions’ assistants in the email search should not be fashioned into ammunition against the agency that requires it to search the records of still more assistants. The only sure consequence of accepting American Oversight’s logic here would be to incentivize agencies to err on the side of under-inclusion in their searches. An agency would know that if it includes any one of a category of custodians, then it risks being forced to search all custodians belonging to the same category, regardless whether the agency believes the others possess responsive records.

To be sure, the analysis might be different had the search of Sessions’ assistants turned up any responsive documents. If that happened, DOJ would be on notice that the high-level officials’ assistants handled communications with Huber, and it might then be reasonable to

23

expect the agency to search the email accounts of other assistants. See Coleman v. Drug Enf’t Admin., 134 F. Supp. 3d 294, 301 (D.D.C. 2015) (“[A]n agency cannot ignore ‘clear leads . . . [that] may indicate . . . other offices that should have been searched.’” (quoting Rollins v. U.S. Dep’t of State, 70 F. Supp. 3d 546, 550 (D.D.C. 2014))). But the fact that no responsive records were found in the email accounts of Sessions’ assistants lends further support to DOJ’s decision not to search others.2 In sum, there is no evidence before the Court that would raise substantial doubt about the appropriateness of DOJ’s identified custodians. Accordingly, the Court rejects American Oversight’s challenge to the custodians searched.

2. Type of Records Searched American Oversight takes a similar tack for its next challenge, asserting that DOJ’s search of emails and the DES was too narrow a response to the Guidance FOIA. See Pl’s MSJ at 7. It reasons that DOJ “provides no reliable basis to conclude that responsive guidance or direction may not be contained in other locations,” such as “hard copy records, electronic records on personal or shared drives, text messages, voicemail systems, or messages on electronic messaging systems.” Id. at 7–8. American Oversight also challenges the agency’s decision to

2 Although not necessary to the Court’s conclusion that it was reasonable for DOJ to exclude the assistants to other officials from its email search, it is worth noting that the manual search of Huber’s mailbox may have turned up correspondence from support staff. See Second Brinkmann Decl. ¶ 25 (stating that DOJ search Huber’s mailbox “for any messages that involved the officials identified above”). Emails sent by the assistants to Huber may have been discovered in this search, so long as they “involved” one of Sessions, Whitaker, Rosenstein, or Hur; perhaps they were either copied on the email or discussed in the body of the email. The description of this manual search is unclear on this score, however, so the Court mentions it only as an aside. See id. (also describing search as one for “any correspondence between” Huber and the senior DOJ officials).

24

limit its search to unclassified email given that Congressman Goodlatte’s letter implicated classified information. Id. at 8.

While the Guidance FOIA sought a wide range of records, including “any written, typed, recorded, graphic, printed, or audio material of any kind,” the mere broad phrasing of a FOIA request does not require the agency “turn over every stone” in response to such a request. Freedom Watch, Inc. v. Nat’l Sec. Agency, 220 F. Supp. 3d 40, 44 (D.D.C. 2016). FOIA’s text, after all, requires “records sought [to] be reasonably described.” 5 U.S.C. § 552(a)(3). Therefore, the requester must offer some basis for its request beyond merely listing any and all potential formats of records. See Dale v. IRS., 238 F. Supp. 2d 99, 105 (D.D.C. 2002) (observing FOIA does not provide an avenue for “all-encompassing fishing expeditions”). It is around such reasonable descriptions that the agency must subsequently structure its search.

Where this Court has upheld challenges to agency determinations as to the scope of a search and ordered supplemental searches of a broader category of agency records, such as text messages or personal emails, the requester has provided a reasonable basis for concluding that such systems of records might contain responsive items. See, e.g., Judicial Watch, 373 F. Supp. 3d at 127 (D.D.C. 2019) (ordering new search for text messages after revelation that subject of FOIA request had exchanged text messages related to request topics). This does not mean that requesters must describe any particular record with exacting granularity, but rather is consistent with the general requirement that a requester must advance more than “purely speculative claims about the existence and discoverability of other documents” to rebut an agency’s determination as to the scope of its search. SafeCard Servs., Inc., 926 F.2d at 1200.

25

Here, American Oversight does not point to particular evidence that indicates records must exist elsewhere. Instead, it attacks the basis for DOJ’s conclusion that any additional responsive records would have been captured by the email and DES searches. It complains that

the only explanation that DOJ provides for not searching record locations or systems of records apart from email and DES is the representations that guidance and direction was only conveyed to Mr. Huber orally, representations we know beyond a doubt are false. The fact that the belatedly disclosed records were conveyed by email certainly demonstrates that DOJ’s failure to search email as part of its initial search was misguided, and that its reliance on representations by OAG officials was misplaced.

Pl’s MSJ Reply at 9. American Oversight appears to believe that, because the representation by DOJ officials that guidance or directives were conveyed orally was proven false in at least one instance (the November 22 appointment letter from Sessions), that representation must be cast aside altogether. See Pl’s Discovery Mot. at 2 (“The responsive record that was belatedly produced shows that all of these representations were flatly untrue. Indeed, it shows that the very officials who made these misrepresentations were directly and personally involved in the provision of formal written guidance.”). To American Oversight, then, the existence of one written record means that the Court should now disregard entirely representations by officials that most of their communications with Huber regarding his investigative focus would have occurred orally. The Court declines to take that leap.

That the agency officials’ representations that they spoke with Huber rather than issued written directives to him turned out to be definitively incorrect on one occasion does not mean that those representations were entirely inaccurate. And it does not mean, as already explained, that the presumption of good faith this Court affords those representations has been totally and irrevocably displaced. Indeed, the fact that the agency’s supplemental email search turned up no

26

documents beyond the appointment letter that sparked the supplemental search fortifies, rather than displaces, that presumption.

Stripped of this veneer, American Oversight’s argument about the types of records DOJ should have searched for is revealed as the type of speculation that cannot justify ordering a supplemental search. See SafeCard Servs., Inc., 926 F.2d at 1200 (stating that agency declarations are “accorded a presumption of good faith” and “cannot be rebutted by purely speculative claims about the existence and discoverability of other documents”). For instance, American Oversight hypothesizes about various types of records that might reflect guidance given to Huber. See Pl’s Discovery Reply at 11 (“DOJ provides no reliable basis to conclude that responsive guidance or direction may not be contained in other locations, such as relevant custodians’ hard copy records, electronic records on personal or shared drives, test messages, voicemail systems, or messages on electronic messaging systems such as Lync or Slack.”); id. at 12 (“[C]orrespondence may have been typed and mailed to Mr. Huber, and the record contains no reliable basis to exclude that possibility.”). But that “no reliable basis” contention is what the Court has just rejected; the senior DOJ officials who would have been in position to give Huber guidance or directives said they did so only through oral communication; the discovery of a single email does not render those representations totally unreliable. And Huber, who volunteered the existence of the Sessions’ appointment letter, did not indicate that he received any other written guidance. American Oversight’s argument therefore amounts to speculation that other records must exist, see SafeCard Servs., Inc., 26 F.2d at 1200, which is not the sort of concrete evidence required to support an inference that such records do in fact exist.

In short, DOJ’s declarations set forth in reasonable detail “the search terms and the type of search performed, and aver[] that all files likely to contain responsive materials . . . were

27

searched.” Valencia-Lucena, 180 F.3d at 326. Without evidence that DOJ overlooked some type of records likely to contain responsive information, the Court rejects American Oversight’s challenge on this score.

3. Interpretation of Guidance FOIA Request American Oversight next claims that the agency “adopted an overly cramped and limited view of what records might be responsive to [the Guidance FOIA].” Pl’s MSJ Reply at 11. It contends that its request for “all guidance or directives provided to [Huber]” includes “any record that contains guidance or directives provided to Mr. Huber . . . regardless of whether the document itself was conveyed to Mr. Huber.” Id. Such records, according to American Oversight, could include meeting notes taken by Huber, talking points used by the Attorney General or other officials in directing Huber’s evaluation, Slack or Lync messages directing Huber to include or exclude topics in his review, or notes from other DOJ officials that record the content of meetings conducted with Huber. Id. at 12. But DOJ, according to American Oversight, improperly read the request to cover only written guidance or direction actually provided to Huber. Id.

DOJ counters that its search was adequately responsive to the Guidance FOIA request, distinguishing “provided to,” which DOJ characterizes as transmitted to Huber, from broader terms such as “reflecting,” which DOJ characterizes as containing records of the discussions had with Huber. See Def.’s Opp. at 21–23. DOJ adds that, if American Oversight actually wanted any records that reflected guidance or directives that had been given to Mr. Huber, it knew exactly what words would do the trick. DOJ points out that, in American Oversight’s Recusal FOIA request, it sought “[a]ll records reflecting any analysis of government or legal-ethics issues.” Id. at 22 (quoting Compl. ¶ 7) (emphasis added). In its Drafting FOIA request,

28

meanwhile, American Oversight sought “[a]ll records relating to the drafting of the November 13, 2017 letter” sent by Boyd. Id. at 23 (quoting Compl. ¶ 18) (emphasis added). Here, then, DOJ says American Oversight “could have asked for records ‘reflecting directives or guidance provided to’ Mr. Huber, but it did not.” Id. Although this is somewhat of a close call, DOJ has the better of the argument.

An agency must liberally construe a FOIA request, Nation Magazine, Wash. Bureau v.

U.S. Customs Serv., 71 F.3d 885, 890 (D.C. Cir. 1995), but it “is not obligated to rewrite the request to ask for more than the requester did,” Canning v. U.S. Dep’t of State, 134 F. Supp. 3d 490, 517 (D.D.C. 2015). In construing a request, agencies “are not obligated to search beyond the four corners of the request, nor are they required to divine a requester’s intent.” Center for Biological Diversity v. U.S. Envtl. Prot. Agency, 279 F. Supp. 3d 121, 139 (D.D.C. 2017) (internal quotation marks and citation omitted). Litigation provides the parties with a forum to debate the adequacy of an agency’s response to a FOIA request, which may allow the parties to work together to clarify the scope of the request. Litigation does not, however, give a requester the opportunity to rewrite the request. See Miller v. Casey, 730 F.2d 773, 777 (D.C. Cir. 1984) (observing requests not broadly drawn, but rather narrowly constructed, should be “read [] as drafted, not as either agency officials or [the requester] might wish it was drafted.”).

Applying those principles here, the best reading of the FOIA Guidance request is that it sought written records actually provided to Huber, rather than any record that reflected guidance or directives given to Huber through some other means. The Court grants that the words of the request, in a vacuum, are “reasonably susceptible to the broader reading,” and it acknowledges the principle that a tie between two plausible readings should usually break in favor of disclosure. LaCedra, 317 F.3d at 348. But the Court is persuaded that American Oversight, in

29

light of its extensive success as a FOIA requester and the language used in its related requests, knew how to ask for the broader universe of records and simply failed to do so.

Courts before this one have considered both the plaintiff’s experience with FOIA and the language of its other requests in determining how to best construe ambiguous requests. In Canning, for example, Judge Moss noted that the plaintiff “ha[d] some experience filing FOIA requests” and drafted his FOIA request to seek three specific categories of information concerning a memorandum written by the President, “not for all information relating in any way to [the memorandum].” 134 F. Supp. 3d at 517. While the plaintiff “could have drafted the request broadly to encompass other documents related to [the memorandum], [he] chose not to do so.” Id. Judge Boasberg reasoned in a similar fashion in American Chemistry Council, Inc. v. U.S. Dep’t of Health & Human Servs., 922 F. Supp. 2d 56, 62 (D.D.C. 2013). There, the plaintiff argued that its request for “‘records’ reasonably described a request for data,” but the court disagreed. Id. That was because “when Plaintiff wanted to ask for data, it did so” explicitly on two other occasions. Id. Relatedly, Judge Boasberg also questioned (without deciding) whether a request for records “related to” data should be read to include the data itself, given that elsewhere the requester explicitly sought the latter. See id. at 63 (“Asking for records ‘related to’ data is not the same as asking for the data.”).

The logic of Canning and American Chemistry Council is instructive here. American Oversight is one of the more sophisticated FOIA practitioners in this district, and the comparatively broader language it deployed in its Recusal and Drafting requests—using “reflecting” and “relating to,” respectively—suggests it understands how to use different words to capture different categories of records. Although courts are to construe FOIA requests liberally, they must also give effect to an apparently deliberate decision to use different words for

30

different requests. See Miller, 730 F.2d at 777 (“The agency was bound to read [the FOIA request] as drafted, not as either agency officials or Miller might wish it was drafted.”); Canning, 134 F. Supp. 3d at 517 (“[T]he agency is not obligated to rewrite the request to ask for more than the requester did.”). Finally, something must be said for the agency’s interests in these cases. If courts do not hold FOIA requesters to their word—especially sophisticated ones who have demonstrated an ability to manipulate language to either narrow or broaden a request—agencies will always have to err on the side of the broadest plausible interpretation, forcing them to do more work than even the requesters themselves may have wanted. Cf. American Chemistry Council, 922 F. Supp. 2d at 63 (“The Court sympathizes with Defendants[’] worry that, absent greater specificity, the Agency would be forced to forward vague, lengthy requests [to third parties involved in search.]”).

American Oversight asked for “all guidance or directives provided to” Huber—not for records that reflect guidance or directives communicated to Huber outside of the record itself. The distinction may seem specious in isolation, but in light of American Oversight’s other, more broadly phrased requests, it is perfectly sensible. Thus, while DOJ’s interpretation of the Guidance FOIA may be the more narrow one, it is also the most reasonable one under the circumstances. Accordingly, the Court will not order DOJ to conduct a search based on American Oversight’s preferred reading of its request.

4. Timeframe

The final flaw in DOJ’s search, says American Oversight, is its end-date. DOJ’s supplemental search looked for records from July 27, 2017 (the date of the first Goodlatte-to- Sessions letter) to January 19, 2018 (the date OIP “initiated its search efforts for Plaintiff’s FOIA requests”). Second Brinkmann Decl. ¶ 20. DOJ does not offer much additional explanation for

31

the end-date, other than to cite DOJ FOIA regulation 28 C.F.R. § 16.4(a), which provides that “[i]n determining which records are responsive to a request, a component ordinarily will include only records in its possession as of the date that it begins its search.”

American Oversight acknowledges that regulation but doubts that DOJ actually began its search on January 19, 2018. See Pl’s MSJ Reply at 13. It points to the fact that one of the records it produced in response to the Prosecutors FOIA was dated March 29, 2018. Id. If, as DOJ contends, it began its search on January 19 and used the same date as the search cut-off date, how could the agency have discovered a responsive record that was created more than two months later? Id. at 13–14. More globally, American Oversight contends that DOJ’s declarant did not adequately substantiate its “bald and conclusory assertion” that it began the search on January 19, 2018. Id. at 14.

This argument fails. To start, the fact that DOJ chose to produce a document that post-

dated the beginning of its search efforts does not mean that DOJ was legally required to search for records that originated after the search began. The key question, according to DOJ regulations, is when the agency began its search. See 28 C.F.R. § 16.4(a) (“[A] component ordinarily will include only records in its possession as of the date that it begins its search.” (emphasis added)). DOJ, as American Oversight charges, appeared to conduct “iterative searches” for records responsive to the Prosecutors and Guidance FOIA requests. Id. But that the process was ongoing and evolving does not change when that process began. Consider the analogous situation of a court ordering an agency to conduct a supplemental search to cure frailties in its first one. In that situation, the cut-off date almost always remains the date of the initial search, not the court-ordered supplemental one. See, e.g., McClanahan v. U.S. Dep’t of Justice, 204 F. Supp. 3d 30, 47 (D.D.C. 2016) (“Under the weight of this authority resting

32

on Public Citizen v. United States Department of State, 276 F.3d 634, 642 (D.C. Cir. 2002), this Court, too, finds that a date-of-search cut-off date is reasonable, even if subsequent searches are conducted, to avoid an endless cycle of judicially mandated reprocessing.” (cleaned up)).

As for American Oversight’s charge that DOJ has not provided enough evidence that it actually began its search efforts on January 19, 2018, this simply re-hashes the argument that the Court should discredit the agency’s second declaration because of the mistake contained in the first. See Pl’s MSJ at 14 (arguing that “DOJ’s false representations on other issues have shorn away any presumption of regularity to which it might otherwise be entitled”). Because the Court has rejected that argument, American Oversight must “produce countervailing evidence” to overcome the agency’s representations, Center for Biological Diversity, 279 F. Supp. 3d at 140 (citation omitted), and that it has not done.

The Court therefore concludes that DOJ selected a reasonable end-date for its supplemental searches.

* * *

The Court appreciates American Oversight’s frustration with DOJ’s initial oversight and the accompanying delay in fully processing its FOIA requests. But the record does not reflect sufficient reason to indulge American Oversight’s grim account of the agency’s missteps. “Hazy allegations of administrative malfeasance may sound incriminating, but the Court requires concrete, specific challenges to the sufficiency of [an agency’s] search in order to deny the agency summary judgment on this point.” Competitive Enter. Inst. v. U.S. Envtl. Prot. Agency, 12 F. Supp. 3d 100, 111 (D.D.C. 2014). Finding each of the specific challenges American Oversight advances unconvincing, the Court must deny its motion for summary judgment and grant summary judgment in favor of DOJ on the adequacy of its search.

33

IV. Conclusion For the foregoing reasons, the Court will grant DOJ’s motion for summary judgment and deny Plaintiff’s cross-motion for summary judgment and motion to stay the case and conduct limited discovery.

________________________

CHRISTOPHER R. COOPER

United States District Judge

Date: July 30, 2019