Federal courts issue hundreds of decisions in FOIA cases every year. Most do not break new legal ground or attract media attention. As 2023 gets under way, the legal eagles of FOIA Advisor -- Allan Blutstein (AB) and Ryan Mulvey (RM) -- look back at 2022 and discuss the top ten decisions that stood out to them (in no particular order).
(1) Barnes v. FBI (D.C. Cir. June 3, 2022) -- affirming district court’s decision that requester’s criminal plea bargain precluded requester from seeking government records pertaining to his case, because government identified legitimate criminal-justice interest in enforcing waiver provision.
AB: In 2017, the D.C. Circuit notoriously became the first court to reject a FOIA waiver contained in a plea bargain, holding that the government was required to demonstrate a “legitimate criminal-justice interest.” The dissent warned that the government would never be able to meet the Circuit’s nebulous test. So, I was pleased as pie to see the government prevail when the Circuit had another opportunity to consider the same issue. Earlier in the year, a district court upheld a plea agreement’s FOIA waiver, noting that plaintiff failed to identify any public policy harms that would override criminal justice interests favoring enforcement of plaintiff’s voluntary waiver. The Circuit’s balancing test is wrong, in my view, but at least it is not necessarily fatal for the government.
RM: The whole FOIA Advisor staff was troubled by the Circuit’s 2017 decision in Price, as I recall. I certainly found Judge Brown’s dissent in Price to be persuasive. So, I agree that Barnes is a welcome development. I would even venture to describe Barnes as a sort of reversal of Price. I was worried that post-Price it would be difficult to sustain any waiver precisely because of the way the Circuit described its apparent balancing test between the interest in enforcement and the public policy interests enshrined in the FOIA—including, importantly, the use of public records requests to seek exculpatory material. But the Barnes court seems to deny there even is a balancing test. Judge Katsas simply wrote that, so long as the government identifies a “legitimate” criminal-justice interest, “[n]othing more is required” to enforce a waiver. Is that really reconcilable with an honest reading of Price?
(2) Behar v. DHS (2d Cir. July 8, 2022) -- reversing district court’s decision and holding that: (1) records obtained by the Secret Service from Donald Trump’s presidential campaign and transition were not “agency records” because agency did not control them; and (2) even if disputed records qualified as agency records, they were protected from disclosure under Exemption 7(C).
RM: The Second Circuit applied the modified-control test to records exchanged between the Secret Service and the Trump presidential campaign and post-election transition team. By itself, that’s a surprising approach. The modified-control test—effectively, the first two Burka factors—is only supposed to be used in cases involving purported presidential or congressional records. So, I think it’s a real question whether the proper test was employed. At the same time, it may not matter. Other courts that have considered access to transition-team records, for example, have likewise avoided Burka and instead looked to so-called “first principles” to find agency control lacking. See, e.g., Democracy Forward Foundation v. GSA, 393 F. Supp. 3d 45 (D.D.C. 2019). All the various “control” tests are ultimately grounded in a functional understanding of “agency record”; they are trying to grapple with what it means for a record to come into an agency’s “possession in the legitimate conduct of its official duties,” to quote Dep’t of Justice v. Tax Analysts. I imagine the Second Circuit didn’t want to open a Pandora’s box of indirect public access to presidential campaign or transition team records through Executive Branch agencies. I’d only add that Mr. Behar has filed a petition for writ of certiorari, which we reported last month. The last petition to ask the Supremes to consider the legitimacy of a post-Tax Analysts control test—in fact, the same modified-control test—was ACLU v. CIA, 823 F.3d 655 (D.C. Cir. 2016), cert. denied, 137 S. Ct. 1837 (Apr. 24, 2017). The unique aspect of Behar, however, is the Circuit reversed despite the agency waiving the control issue below during summary judgment and neither party raising it on appeal. It’s interesting to consider whether the decision violates the party-presentation principle and is otherwise adequately justified by the meager evidentiary record.
AB: You’ve given this decision a lot more consideration than me, perhaps because your employers have gamely litigated several “agency record” decisions before the D.C. Circuit. Be that as it may, I had no qualms with the Second’s Circuit approach here. Presidential campaign and transition records warrant the same special treatment that the court afforded to White House’s visitor logs two years earlier in Doyle. The alleged procedural defects set forth in the petition for certiorari are striking, but I still wouldn’t bet on SCOTUS taking this case. If I am wrong, I’ll make sure the case is at the top of next year’s top ten list.
(3) Seife v. FDA (2nd Cir. Aug. 5, 2022) -- affirming district court’s decision that Exemption 4 protected portions of pharmaceutical company's successful application for accelerated approval of a drug. Of note, concluding that government met statute’s foreseeable harm provision, which the Court held, in the Exemption 4 context, required the government to address “the submitter's commercial or financial interests.”
RM: This is an important decision for the requester community, even though the government prevailed. It is the first Circuit decision to engage with the intersection of Exemption 4 and the foreseeable-harm standard. We mentioned the district court opinion on our list of top cases from the second half of 2020, but we expected the appeals court to delve into the Trade Secrets Act and the question of whether Exemption 4 is a “discretionary” privilege. Instead we got an opinion that considered which precise interests are protected by Exemption 4. The Second Circuit rejected the notion that Exemption 4 protected confidentiality as such, which it thought would have rendered the foreseeable-harm standard surplusage. Instead, the government must show how disclosure would reasonably be foreseen to harm the commercial or financial interests of a submitter. In a sense, the Seife decision brings back some flavor of the old National Parks test rejected by the Supreme Court in Argus Leader. That’s great for requesters. And the Circuit isn’t alone in its reading of the statute; Seife is consistent with how the case law is developing elsewhere.
AB: Bah, humbug. I agree this is good news for requesters . . . in the Second Circuit, at least. DOJ might (and should) decline to follow it elsewhere. This decision might take a little bit of steam out of any congressional efforts to undo the Supreme Court’s decision in Argus Leader, which did not address the foreseeable harm provision. It behooves me to note that competitive harm is not the be-all and end-all of the National Parks test. The D.C. Circuit—and subsequently, many other courts—also recognized that the government itself could be harmed if businesses declined to voluntarily provide information that the government needed. This became known as the “impairment prong” of the National Parks test. In Seife, the disputed information almost certainly was required to be submitted to the FDA. Thus, the parties and the court had no reason to consider “impairment” as a foreseeable harm.
RM: Fair point. But whether we’re talking about the government-impairment prong of the National Parks test, or Critical Mass’s “customary” standard for voluntary submissions, I’m not sure it makes a huge difference in the end vis-à-vis foreseeable harm. It seems to me an agency’s explanation of “impairment” would need to be substantive and context-bound. The well-developed caselaw on foreseeable harm and the deliberative-process privilege could be instructive. I also tend to think, at the theoretical level, the government-impairment prong is duplicative of the “competitive harm” prong, as explained in an amicus brief I filed in Argus Leader. But I digress…
(4) Newman v. BOP (D.D.C. May 13, 2022) -- determining that government performed “thorough” searches for records concerning former Cuban exile and inmate Antonio Veciana, whose incarceration and supervision ended in the early 1980s. Of note, the court remarked that this case was “yet another example of the “mismatched incentives” that FOIA creates” because “nonprofit FOIA requestees like [plaintiff] pay little to nothing for their FOIA requests . . . [s]o they do not internalize the costs of a wild goose chase like this one. This case has tasked multiple attorneys at three agencies (including the U.S. Attorney's Office) and several FOIA specialists in the search for decades-old inmate records that by regulation should have been transferred or destroyed years ago. Unsurprisingly, they were. But the cost of this predictably fruitless search is borne by the agencies, and ultimately, American taxpayers.”
AB: This is a fairly routine “adequacy of search” case, but it made my year to read Judge McFadden’s dicta quoted above. He’s absolutely right that American taxpayers should not bear the full costs of FOIA fishing expeditions. Congress should require all requesters, including journalists, to pay fees for all search time beyond the first two hours. Instead, Congress has encouraged requesters to submit far-reaching requests by expanding the meaning of a “representative of the media” and prohibiting agencies from collecting fees when they miss the statute’s short response deadlines. Perhaps Judge McFadden should not have singled out non-profit groups, however; they are certainly not alone in submitting burdensome requests.
RM: Judge McFadden has been on a roll publishing dicta that singles out sections of the requester community and blames them for everything wrong with FOIA . It started in American Center for Law & Justice v. Department of Homeland Security, and it has continued with Newman and Harrington v. Food & Drug Administration. Personally, I find it inappropriate, especially since Judge McFadden’s assumptions are open to challenge. Professor Bernard Bell, for example, has published a short and thought-provoking response to ACLJ and Newman at the Yale Journal on Regulation’s Notice & Comment blog. He makes the point that Judge McFadden ignores the significant costs borne by non-profit requesters when they review agency records post-production, or when they spend time fighting with agencies over ostensibly broad requests pre-litigation. Judge McFadden also fails to appreciate that agencies aren’t always innocent. Although they may face underfunding or understaffing, agencies can be abusive in their rejection of requests as imperfect/unreasonably described. They can be reluctant to provide requesters with information to help with narrowing, and sometimes requesters’ broad requests (and their rush to the courthouse) are entirely warranted. I’m also skeptical the FOIA’s attorney’s fee provision provides as strong an incentive as Judge McFadden suggests. Prevailing in a lawsuit isn’t enough, and courts hardly rubberstamp fee awards; they do a good job considering whether a requester is both eligible and entitled to recover. As for preferential fee categories, call me skeptical that this has any wide-reaching impact. Some agencies choose not to collect, even when they aren’t statutorily barred. (I agree the wisdom of the 2016 FOIA Amendments, in this respect, is open to debate!) And, as Professor Margaret Kwoka’s research demonstrates, the majority of requests aren’t coming from the media or non-profit “watchdogs”—they’re coming from commercial requesters and first-party requesters, neither of which get special fee treatment.
AB: You’ve been dying to get that off your chest for a while, haven’t you? Perfectly understandable given your dutiful service to several non-profit groups with laudable FOIA practices. I confess to being surprised by your remark about preferential fee categories. If you think they have no impact, let’s get rid of them and see what happens. Shall we submit a joint recommendation to the Federal FOIA Advisory Committee asking Congress to enact a three-year pilot program? Judge McFadden has not complained—as far as I know—about the overall number of FOIA requests or lawsuits, but rather that requesters unburdened by fees are more likely to submit requests for voluminous records and/or requests requiring burdensome searches. Does any of Professor Kwoka’s research suggest otherwise? I doubt it.
RM: Professor Kwoka, as far as I’m aware, hasn’t examined whether the fee categories create an incentive for overly broad requests. Happy to be corrected there. But she has acknowledged that, even when fees are charged and collected, they don’t provide a sustainable funding model, if only because many expenses aren’t captured under “search,” “review,” or “duplication.” She’s probably right. And FOIA does tend to be more expensive than anyone ever expects. But all that being said, my fundamental point is that preferential requesters aren’t the problem. As OIP reported for FY 2021, agencies collected fees amounting to “less than 0.4% of the total costs related to the government’s FOIA activities.” There’s simply no way this is happening because the agencies are treating everyone as, say, representatives of the news media, let alone liberally granting public-interest fee waivers. It probably isn’t even because of the fee limitations in the 2016 amendments. Maybe a better pilot program would be to have OIP collect hard data on who is requesting (i.e., how are requesters being categorized for fee purposes), and the specifics of why fees aren’t being collected.
Since I’m on a roll, I’d like to mention one more adequacy-of-search decision from Judge McFadden that not only cites ACLJ but raises a few other red flags: Center for Immigration Services v. USCIS. In this case, Judge McFadden held a FOIA request to be unreasonably described even though the agency had already conducted a search in part because the agency cited “overly burdensome post-search efforts.” Frankly, I see no textual basis in the FOIA to reject voluminous (as opposed to unreasonably described) requests. OIP has long advised agencies that “[t]he sheer size or burdensomeness of a FOIA request, in and of itself, does not entitle an agency to deny that request on the ground that it does not ‘reasonably describe’ records within the meaning of 5 U.S.C. § 552(a)(3)(A).” And district courts have explained the only solution for an agency faced with a voluminous request, at least in litigation, is to move for an Open America stay. Most importantly, I find the CIS decision insidious because Judge McFadden balanced the burden of processing against his perception of whether the requested records were relevant to (and would advance) the requester’s “purpose” as a non-profit organization. As a rule, the identity and motive of a requester are absolutely irrelevant to an agency’s disclosure obligations. There are exceptions with certain fee issues. But motive—let alone the perceived relevance of responsive records—should have no bearing on the reasonable-description requirement.
(5) Transgender Law Ctr. v. ICE (9th Cir. Aug. 19, 2022) -- holding that: (1) agencies must prove adequacy of search “beyond material doubt” and that district court erred in finding that DHS adequately searched or records pertaining to asylum-seeker’s death from HIV in federal custody; (2) district court should not have “essentially treated all drafts as necessarily covered by the deliberative process privilege”; (3) district court erred in permitting government to withhold email domain addresses under Exemptions 6 and 7(C); (4) government’s use of Exemption 7(E) was overbroad and district court neglected to analyze whether withheld records were techniques and procedures, and not guidelines; and (5) district court failed to ensure that: (a) the government’s Vaughn Index entries were non-conclusory, (b) the government explained why documents were not segregable, and (c) the government properly designated certain documents as non-responsive or duplicates.
RM: I’m not sure this decision breaks any new ground so much as it’s an important reminder to agencies not to try and be creative with their withholdings, and a warning to district judges not to be lackluster in their review of supporting declarations and Vaughn indices. Of particular note is the panel’s reiteration that not all draft documents are covered by the deliberative-process privilege, as well as the passing comment that records reflecting deliberations over press statements may not implicate “the type of policy decisions the privilege covers.” That latter question has been answered in the D.C. and Second Circuits, but seems open still in the 9th. Finally, I’m amazed there even needed to be a fight over the redaction of government agency domains in official email addresses. The court was right to explain how such information cannot constitute a “similar file.”
AB: I share your puzzlement about the agency’s refusal to release domain email addresses, as well as the defense of Exemption 6 by counsel and the district court. Hello attorney’s fees. The answer to whether public messaging records fall within Exemption 5 in the Ninth Circuit seems to be much closer to “no” than “yes.” The court here directed the release of the agency’s draft press statements instead of giving the agency an opportunity to offer additional evidence to the district court. What else can I say other than the D.C. and Second Circuit courts are right and the Ninth Circuit is wrong.
(6) Harrington v. FDA (D.D.C. Jan. 2022) -- denying plaintiff’s motion seeking “immediate” production of agency’s pet-food-related records and finding that FDA’s proposed production schedule was reasonable; taking into account that agency’s relevant FOIA office had backlog of 336 requests, due “in no small part” to plaintiff’s “staggering 2220+ requests to FDA since 2018”; further noting that plaintiff never sought expedited processing at administrative level and did not meet the statutory test.
AB: Of all the unsuccessful FOIA plaintiffs since 1967, few have deserved to lose more than the plaintiff in this case. As the court pointed out, the agency generously offered to process the requested records in a mere 60 days if plaintiff would allow the agency to pause its work on one of plaintiff’s earlier requests involving hundreds of thousands of pages. Because plaintiff rejected that offer, Judge McFadden was oh-so-right to state that plaintiff had only himself to blame for the agency’s alternative production timeline of one year. Bonus points to the court for inviting Congress to address vexatious FOIA requesters like this this one. Amazingly, plaintiff’s attorney actually considered appealing this case.
RM: This seems to have been a poorly prosecuted case. As the district court noted, the requester failed to file a reply brief. And it seems the requester tried to make an argument that seriously confused the statutory deadline to provide a response, and the separate (and subsequent) requirement to provide a requester with non-exempt responsive records “promptly.” I won’t comment on the vexatious-requester question except to note that if this is real problem in need of a solution—and I’m not convinced it is—then the court was right to invite Congress to provide a statutory fix.
(7) Rocky Mountain Wild, Inc. v. U.S. Forest Serv. (10th Cir. Dec. 30, 2022) -- affirming district court’s decision and finding that: (1) plaintiff waived argument that Vaughn Index was insufficient; (2) agency performed reasonable search for voluminous records concerning proposed highway project; (3) agency properly withheld records pursuant to Exemption 5 and demonstrated prospective harm, but declining to hold that the foreseeable harm provision imposed a “heightened burden”; and (4) district court properly ordered plaintiff to return two documents that agency accidentally disclosed to plaintiff without redaction, even though another organization subsequently posted the documents online.
RM: This is a surprising decision, in my mind. Although the panel averred it was not deciding whether the foreseeable-harm standard imposes any additional burden on an agency to justify withholding beyond satisfying the technical requirements for applying an exemption in the first place, avoiding that question speaks volumes. At least in other circuits, such as D.C. and the 9th, the government has abandoned its atextual and, frankly, preposterous position that the 2016 FOIA Amendments merely codify the discretionary “presumption of openness” most recently articulated in the March 2009 Holder Memo. Rocky Mountain Wild seems to breath new life into that position, sadly, at least in the Tenth Circuit. I also find the Circuit’s holding that the foreseeable-harm standard would have been met, even if it applied as it does elsewhere, to be equally troubling. The panel seriously confused why Exemption 5 exists and can be used in conjunction with civil discovery privileges, with the distinct question of whether disclosure could be reasonably foreseen to cause harm. Mere recitation of the prerequisite for a privilege, or the broad policy reasons for the existence of a privilege, cannot be enough. The statutory language cannot sustain that approach.
AB: Just as I was about to close the book on 2022, the Tenth Circuit pulled me back in. My first thought when I saw the decision was: “Ryan is going to hate this.” Then I cheerfully started humming John Denver’s signature song Rocky Mountain High, because the foreseeable harm provision is truly horrible. FOIA processors should not be expected to be seers who prophesize about harms likely to occur if particular documents are disclosed. The Senate Report to the FOIA Improvement Act of 2016 claimed that the foreseeable harm standard would provide “clearer guidance regarding when to withhold information covered by a discretionary FOIA exemption.” Seriously? What could be clearer than permitting information to be withheld when it falls within an exemption, as the statute operated for 50 years? FOIA improvement? Hardly.
(8) Eddington v. DOD (D.C. Cir., June 3, 2022) -- affirming district court’s decision that requester failed to present sufficient evidence to overcome agency’s declaration that it did not receive any of his requests emailed to 14 DOD components.
AB: We did not include the district court’s decision in our list of top cases last year, but since plaintiff would not leave well enough alone, I felt compelled to add it now. This case illustrates the obvious risk of submitting FOIA requests by email. Just because you click “send” and your email doesn’t bounce back, that doesn’t necessarily mean that the agency actually received it. A prudent requester would have asked the agencies for acknowledgments of receipt within 6 months. But this frequent requester and trigger-happy litigant mistakenly assumed that all 14 agencies were ignoring his requests for 6 months and he marched into court. As the old saying goes, “assuming makes an ass of u and me.”
RM: I’ll admit this appeal really had me scratching my head. As you say, it would have been quite easy (and more prudent!) for the requester to have asked the agency if it had received his many requests. I wouldn’t be surprised if this lawsuit were set up as some sort of strategic challenge, say, to establish a “mailbox” rule for receipt of FOIA requests. If it were really an effort to get records, then it would have been quicker (as the panel suggested in a footnote) to forego appeal and start over.
AB: Uh, sorry, I don’t buy for a split second that plaintiff’s lawsuit was an elaborate strategic challenge from the get-go. A simpler explanation makes more sense: plaintiff, a frequent litigator, was unwilling to swallow the agency’s response and he did not want to risk mooting his lawsuit by submitting new requests. Since you referenced it, the court utterly destroyed plaintiff’s mailbox rule argument. For anyone who’s seen the the movie Top Gun: Maverick, picture the massive blast of the enemy’s uranium enrichment plant, aka “Miracle Number 2.”
(9) Citizens for Responsibility & Ethics in Wash v. DOJ (D.C. Cir., Aug. 19, 2022) -- affirming district court’s decision that DOJ failed to adequately explain how an agency memorandum to Attorney General Barr concerning Mueller Report fell within Exemption 5’s deliberative process privilege; declining to decide whether “a purely hypothetical, academic discussion among agency personnel could qualify under the . . . privilege”; refusing to allow DOJ to present a new argument that memorandum was drafted “for the purpose of determining the content of a possible public statement regarding the report.”
AB: The district court’s 2021 opinion drew considerable attention for calling out DOJ’s “misleading” and “incomplete'“ declarations. Democrats on the U.S. Senate Judiciary Committee wasted little time in urging DOJ to forgo an appeal, claiming it was necessary in order to “help rebuild nation’s trust in DOJ’s independence after four years of turmoil.” But the government’s appeal was well taken. The D.C. Circuit all but conceded that DOJ’s new deliberative process privilege argument was likely meritorious. Following precedent, however, the Circuit was constrained to allow the government to belatedly raise a new legal theory in the absence of extraordinary circumstances. In sum, kudos to the DOJ’s appellate staff for yeoman’s work, but all the king’s horses and all the king’s men could not rescue this botched case.
RM: Well, I’m not sure I would describe the government’s appeal as “well taken.” As the CREW opinion explains, DOJ ended up conceding that two of the three grounds for applying the deliberative-process privilege were non-meritorious, and hardly due to the niceties of existing precedent. I would also quibble with whether the panel “all but conceded” the third ground was “likely meritorious.” I read the Circuit as saying the arguments and supporting declarations in the district court before the government’s motion for stay pending appeal did not “make—or even suggest—[a] connection” between the records and deliberations over the preparation of a public statement. The winning argument might have been meritorious in the abstract, but it would have been highly questionable for the agency, which had already been given multiple bites at the apple in the district court, to revise its declarations to correct what it dubiously described as a factual “misimpression.”
(10) NBC 7 San Diego v. DHS (D.D.C. Dec. 20, 2022) -- concluding that DHS and two components failed to provide sufficient explanations as to why they deemed requested records concerning a secret tracking database to be non-responsive, noting that agencies improperly considered “context” of requests instead of identified search terms.
RM: This case grapples with an interesting development in how FOIA requesters draft their requests, and how an agency’s “responsiveness” review occasionally collapses into its search. The court here rejected the agencies’ attempts to conduct a subject-matter responsiveness review about completing a search for records containing identified search terms. Because the underlying requests set the conditions of responsiveness (e.g., “records containing terms x, y, or z”), it was improper for the government to evaluate, as a further matter, whether returned records were relevant to the “substance” or general subject-matter of the request. This approach amounts to an “unreasonable refram[ing].” I’d note, however, that at least one other recent decision has taken a different approach—albeit one that I think is incorrect. Judge Friedrich in Project on Government Oversight v. Department of the Treasury decided to rule in favor of an agency that had used language from the requester’s fee-waiver request to apply a sort of “relevance” review when culling potentially responsive records. The POGO court reasoned that ruling otherwise would have violated the principle that requesters “cannot dictate the search terms for his or her FOIA request.”
AB: I agree with the court’s decision in NBC 7 San Diego. If I had made the same request, I might have authorized the agency to cull the documents by a general subject matter depending on how many records had been located. But many requesters do not trust agencies to do this, and sometimes a requester does not have an umbrella subject. If I were still at Treasury, I would have been reluctant to bless the process used by the agency in the POGO case —that is, not until the agency had reached out to the requester to seek clarification.
* * *