Brennan Center for Justice at New York University School of Law v. United States Department of Justice

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

THE BRENNAN CENTER FOR JUSTICE AT NEW YORK UNIVERSITY SCHOOL OF LAW et al.,

Plaintiffs,

Civil Action No. 18-1860 (RDM)

v.

UNITED STATES DEPARTMENT OF JUSTICE,

Defendant.

MEMORANDUM OPINION AND ORDER Plaintiffs brought this Freedom of Information Act (“FOIA”), 5 U.S.C. § 552, action seeking records to illuminate “how the government is prosecuting cases involving domestic terrorism.” Dkt. 1-1 at 6. As Plaintiffs explained, the “available data” indicated that the Department of Justice (“Department”) had brought “4,496 terrorism-related prosecutions” since 2001 and had obtained “3,772 convictions or guilty pleas.” Dkt. 16-1 at 5 (Kurzman Decl. ¶ 13). But Plaintiffs could not tell, based on that publicly available data, which “activities the Department of Justice considers terrorism,” particularly “domestic terrorism,” or whether international terrorism and domestic terrorism cases are treated differently. Dkt. 1-1 at 3–4.

To better understand how the Department categorizes and treats “terrorism” cases, Plaintiffs the Brennan Center for Justice and Professor Charles Kurzman sent a FOIA request to the Department’s Executive Office for United States Attorneys (“EOUSA”), seeking all records in the Department’s Legal Information Office Network System (“LIONS”) database relating to terrorism cases, including the docket numbers associated with the court proceedings in each case.

Id. at 4. After the Department withheld the docket numbers for each case pursuant to FOIA Exemptions 6 and 7(C), Plaintiffs brought this suit, which the Court resolved last year on the parties’ cross-motions for summary judgment. Applying the binary standard that the D.C. Circuit set forth in two cases raising similar issues—Am. C.L. Union v. U.S. Dep’t of Just., 655 F.3d 1 (D.C. Cir. 2011) (“ACLU I”), and Am. C.L. Union v. U.S. Dep’t of Just., 750 F.3d 927 (D.C. Cir. 2014) (“ACLU II”)—the Court held that the Department permissibly withheld the docket numbers in cases that resulted in acquittals or that were dismissed but that the Department impermissibly withheld the docket numbers in cases resulting in convictions. Brennan Ctr. for Just. at N.Y. Univ. Sch. of Law v. U.S. Dep’t of Just., No. 18-cv-1860, 2020 WL 1189091 (D.D.C. Mar. 12, 2020) (“Brennan I”). In reaching that conclusion, the Court weighed the “minimal” privacy interest threatened by releasing information relating to a public, criminal conviction and the more “substantial” privacy interest threatened by releasing information relating to a case resulting in an acquittal or dismissal of charges—and balanced those privacy interests against the public interest in disclosure. Id. at *5–11. Consistent with ACLU I and ACLU II, that balance tipped in favor of disclosure with respect to cases resulting in convictions and in favor of withholding in cases resulting in acquittals or dismissals. Id.

Now pending before the Court is the Department’s motion for reconsideration under Rule 59(e). Dkt. 32. In its motion, the Department argues that the privacy interests at stake are far greater than it had previously indicated and, correspondingly, far greater than the Court’s decision in Brennan I recognized. In particular, the Department now explains that prosecutors categorize each case within LIONS at an early stage of an investigation and that those categorizations are rarely updated as a case develops. Dkt. 32-1 at 5. Consequently, “it is not uncommon” for those categorizations to no longer be germane by the time the cases reach

3

fruition. Id. In addition, the Department acknowledges that at least some cases are categorized as terrorism-related in error or for unknown reasons. See Dkt. 42-1. As such, a case that is initially categorized (or miscategorized) as terrorism-related within LIONS may later result in charges and a conviction that bear no obvious connection to terrorism. In that subset of cases, the Department contends, disclosure of the docket numbers in LIONS could reveal for the first time that the Department at some point believed the crime might be related to terrorism—or, worse, disclosure could suggest that the Department considered the case related to terrorism when in fact the categorization was a mistake. In the Department’s view, even if a criminal defendant has only a minimal privacy interest in the fact of his public conviction, he holds a much stronger privacy interest in the undisclosed fact that the Department suspected him of involvement in terrorism, mislabeled his case as terrorism-related, or ultimately charged him with a crime that the Department might internally regard as terrorism-related but that the public would not ordinarily associate with terrorism. Based on the difficulty in ascertaining which cases are correctly labeled as terrorism-related convictions or guilty pleas, the Department argues that the Court should now grant summary judgment in its favor and relieve it of the obligation of releasing any of the docket numbers corresponding to terrorism-related convictions or guilty pleas.

As the history of this litigation and the Department’s recent filings demonstrate, the interests on both sides of the relevant balance are weighty—and far weightier than the Court previously appreciated. On the one hand, disclosure of docket numbers in “terrorism-related” cases that resulted in convictions will often reveal information about the investigation or the Department’s internal characterization of the case that was not previously public. The relevant privacy interests thus go beyond the risk of reminding the public about events that previously

4

occurred in the light of day. In some cases, for example, the disclosure might reveal that the defendant was the subject of a terrorism investigation that never resulted in any terrorism-related charges. In others, it might suggest that a defendant was connected in some way to terrorism when the Department’s categorization was simply a mistake. And in still others, it might show that the Department regarded a conviction as terrorism-related, even though it never publicly disclosed that characterization. Disclosing that information could cause grave harm to the subjects of the investigations and prosecutions and would do so in a manner that would provide little opportunity for the subjects to rebut the Department’s suspicions or internal characterizations.

On the other side of the balance, however, Plaintiffs’ interest—and the public interest—in better understanding how the Department characterizes “terrorism” cases is also more compelling than the Court previously recognized, given that the Department’s submissions in support of its motion for reconsideration call into serious question the reliability of its case categorizations and, by extension, its public reports regarding “terrorism-related” prosecutions. Understanding what the Department means when it describes its efforts to fight terrorism—what it means, for example, when it announces that it has obtained “3,772 convictions or guilty pleas” in terrorism-related cases since 2001, Dkt. 16-1 at 5 (Kurzman Decl. ¶ 13) —is of intense public interest. Similarly, the public has a considerable interest in assessing the accuracy of the Department’s reports regarding terrorism prosecutions. In short, this is a case with anvils on both sides of the scale.

The Court previously denied the Department’s motion for reconsideration in part, with respect to any “conviction or plea for an international terrorism offense that has a clear public connection to terrorism.” See Brennan Ctr. for Just. at N.Y. Univ. Sch. of Law v. U.S. Dep’t of

5

Just., No. 18-cv-1860, 2020 WL 7685612, at *1 (D.D.C. Aug. 19, 2020) (“Brennan II”). For the following reasons, the Court will now GRANT in part and DENY in part the remaining portion of the Department’s motion.

I. BACKGROUND

As the Court explained in its prior opinions, the Department uses the LIONS database “to categorize and to track information about cases” in which the ninety-four U.S. Attorneys’ Offices are involved. Brennan I, 2020 WL 1189091, at *2. Whenever a federal prosecutor opens an investigation, she creates a record in the database, which includes such information as the name of the person under investigation, the investigating agencies, and the nature of the possible charges. Id. Once the government files charges, the prosecutor adds additional information, including the docket number associated with the case in court and, later, the outcome of the case. Id. Among the categories to which a case can be assigned in the LIONS database are six related to terrorism: “International Terrorism Incidents Which Impact U.S.,” “Domestic Terrorism,” “Terrorism Related Hoaxes,” “Terrorist Financing,” “Export Enforcement Terrorism-Related,” and “Critical Infrastructure Protection.” Id. Although the Department provides public access to its case management system, the public version of the records redacts personally identifying information, including the docket numbers, which could be used to access public court records that identify the defendant in each prosecution. Id.

More than three years ago, on January 17, 2018, Plaintiffs submitted a FOIA request to the EOUSA, the office within the Department responsible for providing administrative support to U.S. Attorneys’ Offices. Id. Plaintiffs sought “[a]ll records in the [LIONS] database involving public charges that are marked with at least one of” the six categories used to track terrorism cases. Id. (quoting Dkt. 1-1 at 4) (alternations in original). The request specified that Plaintiffs

6

wanted to obtain the docket numbers associated with each record “to understand and to analyze how the Department characterizes conduct as ‘terrorism’ and how it prosecutes ‘terrorism’ cases.” Id. The Department denied Plaintiffs’ request for the docket numbers on the ground that they are shielded from disclosure by FOIA Exemptions 6 and 7(C), both of which protect privacy interests. Id.

Plaintiffs filed this FOIA action to challenge the Department’s withholdings, Dkt. 1, and the parties filed cross-motions for summary judgment, Dkt. 13; Dkt. 16. The Department’s motion argued that Exemptions 6 and 7(C) justify the challenged withholdings. Exemption 6 applies to any “personnel and medical files and similar files the disclosure of which would constitute a clearly unwarranted invasion of personal privacy.” 5 U.S.C. § 552(b)(6). Exemption 7(C) covers a narrower set of records—only those that are “compiled for law enforcement purposes.” Id. § 552(b)(7). But when Exemption 7(C) does apply, it covers a broader range of privacy interests. See Nat’l Archives & Records Admin. v. Favish, 541 U.S. 157, 165–66 (2004); Dep’t of Just. v. Reporters Comm. for Freedom of Press, 489 U.S. 749, 756 (1989). While Exemption 6 applies to only those disclosures that “would constitute a clearly unwarranted invasion of personal privacy,” 5 U.S.C. § 552(b)(6) (emphasis added), Exemption 7(C) applies to all disclosures that “could reasonably be expected to constitute an unwarranted invasion of personal privacy,” id. § 552(b)(7)(C) (emphasis added). Within its scope of coverage, Exemption 7(C) therefore “establishes a lower bar for withholding” than Exemption 6. ACLU I, 655 F.3d at 6. So long as the records in question are “compiled for law enforcement purposes,” a Court need consider only whether Exemption 7(C) applies, because any law enforcement record not protected by Exemption 7(C) would necessarily fall outside the scope of Exemption 6.

7

Although the Department bore the burden of demonstrating that Exemption 7(C) applied to the records at issue, see Mobley v. CIA, 806 F.3d 568, 580 (D.C. Cir. 2015), the original declaration that the Department submitted in support of its motion for summary judgment failed to show that these records were compiled for law enforcement purposes. See Minute Entry (Nov. 25, 2019). But given the potentially significant privacy interests at stake, the Court offered the Department an opportunity to supplement the record to establish the applicability of Exemption 7(C) by describing how it uses and compiles information in the database. Id. The Department’s supplemental declaration, however, was still “largely silent” on how the “requested document numbers” related to records compiled for law enforcement purposes. See Minute Order (Jan. 28, 2020). But, because of the “privacy interests at play,” the Court gave the Department a third chance to justify its withholdings. Id.

On March 12, 2020, the Court issued an opinion granting in part and denying in part the parties’ cross-motions for summary judgment. Brennan I, 2020 WL 1189091, at *11. As a threshold matter, the Court held that, although docket numbers are created by courts, they are compiled in the LIONS database for law enforcement purposes, such that the challenged withholdings were subject to the more lenient standard of Exemption 7(C). Id. at *3–5. After determining that the records in question were compiled for law enforcement purposes, the Court turned to the question whether disclosure of the docket numbers “could reasonably be expected to constitute an unwarranted invasion of personal privacy.” 5 U.S.C. § 552(b)(7)(C). To answer that question, the Court “balance[d] the public interest in disclosure against the [privacy] interest[s] Congress intended the Exemption to protect.” Brennan I, 2020 WL 118909, at *5 (second and third alterations in original) (quoting Reporters Comm., 489 U.S. at 776).

8

The Court began by assessing the asserted privacy interest. Under D.C. Circuit precedent, the privacy interest at stake in the disclosure of public docket numbers varies depending on the outcome of the underlying case: “defendants whose prosecutions ended in acquittal or dismissal have a much stronger privacy interest in controlling information concerning those prosecutions than defendants who were ultimately convicted.” ACLU II, 750 F.3d at 933. Although “an agency’s disclosure of public docket numbers for cases that resulted in convictions implicates a cognizable privacy interest,” that interest is “‘at the lower end of the privacy spectrum.’” Brennan I, 2020 WL 1189091, at *6 (quoting ACLU I, 655 F.3d at 7). Because the conviction or plea is already public, the privacy interest at issue is limited to preventing disclosure of records that could draw renewed attention to the defendant’s crime. Thus, “even though the disclosure of public docket numbers ‘compromise[s] more than a de minimis privacy interest,’ it does ‘not compromise much more.’” Id. (alteration in original) (quoting ACLU I, 655 F.3d at 12).

The Department argued that, even where a case ended in a conviction or plea, the public docket numbers categorized as terrorism-related in the LIONS database implicate heightened privacy interests because association with terrorism is “stigmatic.” Id. (quoting Dkt. 13 at 14). The Court was “persuaded that the stigma of a terrorism conviction is likely substantial and that disclosure of the docket numbers of cases that the Department has characterized for its internal purposes as terrorism-related risks inviting unwanted attention on the subjects of those prosecutions.” Id. at *6. But, as in ACLU I, this asserted privacy interest was “still far weaker than the core interests protect by Exemption 7(C),” which is “principally concerned with protecting individuals’ ‘strong interest in not being associated unwarrantedly with alleged criminal activity.’” Id. at *7 (quoting Stern v. FBI, 737 F.2d 84, 91–92 (D.C. Cir. 1984))

9

(emphasis in original). If anything, the privacy interest at stake in the disclosure of a docket number associated with a terrorism-related conviction might be weaker than in other classes of cases, because “terrorism-related cases are likely to generate more press coverage and greater public attention than the average case.” Id. at *6. Significantly for present purposes, the Court emphasized that “the Department cannot plausibly argue (nor does it attempt to argue) that the public ‘will hear of’ the terror-related nature of the cases ‘for the first time merely because the Justice Department releases a list of docket numbers.’” Id. (quoting ACLU I, 655 F.3d at 10). The disclosure of the contested docket numbers would thus reveal “only information that has already been the subject of a public proceeding.” ACLU I, 655 F.3d at 8.

Expanding on this point, the Court observed in a footnote:

One can imagine an argument that the Department’s characterization of an offense as “terrorism-related” is not information that is already public or that could be easily found by reviewing the public docket. Someone’s neighbors, for example, might know that a subject had pleaded guilty to a financial crime of some sort but might not know that, in the Department’s view, that crime was connected to domestic or international terrorism. The Department, however, has not made this argument, nor can the Court discern from the limited information that is before it whether this hypothetical harm is fanciful or real.

Brennan I, 2020 WL 1189091, at *7 n.2 (emphasis added). In the absence of such an argument, the Court concluded that, as to the cases resulting in convictions or guilty pleas, the privacy interest implicated by disclosure of the public docket numbers was relatively weak. Id. at *7. In short, “where someone has been convicted of terrorism-related charges, disclosure of the information contained in the LIONS database—including the docket number of the case—would risk little additional public opprobrium.” Id. at *8.

Turning to the public interest in disclosure, the Court held that it was strong. The Court explained that the relevant public interest to weigh was “‘the extent to which disclosure advances the basic purpose of the [FOIA] to open agency action [to] the light of public scrutiny, . . .

10

thereby furthering the citizens’ right to be informed about what their government is up to.’” Id. at *7 (first and third alteration in original) (quoting ACLU I, 655 F.3d at 6). Plaintiffs had argued persuasively “that the public has a significant interest in understanding the Department’s ‘current strategies and success rates with regard to terrorism-related prosecutions.’” Id. (quoting Dkt. 16 at 24). “Terrorism prosecutions—and which cases the Department characterizes as ‘terrorism’ cases—moreover, touch on an array of issues of public attention, ranging from immigration policy, national security, and the allocation of prosecutorial resources.” Id. (noting that the Department spent $5.6 billion on its counterterrorism efforts in fiscal year 2019). Even more to the point, Plaintiffs had established that “the information they seek would advance the public’s interest in knowing how the Department categorizes ‘terrorism’ cases, how it allocates its resources in prosecuting domestic and international terrorism cases, and how the threat of terrorism has evolved in recent years.” Id. Indeed, “the only way to conduct a comprehensive study of the prosecutions [that the] Department categorizes in this manner is through the disclosure of the requested docket numbers.” Id. The Court thus concluded that the public interest in disclosure outweighed the countervailing privacy interest and granted summary judgment to Plaintiffs with respect to docket numbers for cases that resulted in convictions or pleas. Id. at *9, *11.

The balance was different for the 724 cases that resulted in acquittals or dismissals. Id. at *9. Following the D.C. Circuit’s reasoning in ACLU II, the Court explained that “‘defendants whose prosecutions ended in acquittal or dismissal have a much stronger privacy interest in controlling information concerning those prosecutions than defendants who were ultimately convicted.’” Id. (quoting ACLU II, 750 F.3d at 932–33). This is because “‘[t]he presumption of innocence stands as one of the most fundamental principles of our system of criminal justice,’

11

and, ‘[u]nfortunately, public perceptions’ do not always hold the government to its burden of proof.” Id. (second alteration in original) (quoting ACLU II, 750 F.3d at 933). Although the public interest in disclosure remained strong, the Court granted summary judgment in favor of the Department with respect to docket numbers for cases ending in acquittal or dismissal. Id. at *10–11.

Having resolved the parties’ cross-motions for summary judgment, the Court entered final judgment on March 12, 2020. Dkt. 31. Twenty-eight days later, the Department moved for partial reconsideration pursuant to Federal Rule of Civil Procedure 59(e). Dkt. 32; see Fed. R. Civ. P. 59(e) (providing that a “motion to alter or amend a judgment must be filed no later than 28 days after the entry of the judgment”). In its motion, the Department seizes upon the Court’s footnote observing that the Department had failed to argue that its “characterization of an offense as ‘terrorism-related’ is not information that is already public or that could be easily found by reviewing the public docket.” Brennan I, 2020 WL 1189091, at *7 n.2. Now, the Department contends that the Court’s “concern is, in fact, quite real and affects a significant number of cases” at issue. Dkt. 32-1 at 5.

To explain why, the Department offers the declaration of Kevin Krebs, Assistant Director of the EOUSA’s FOIA staff. Dkt. 32-2 (Krebs. Decl.). Krebs attests that categorizations in the LIONS database are “entered at a very early stage of an investigation” and then are “generally not changed to reflect the ultimate charges.” Id. at 2 (Krebs. Decl. ¶ 6). Even when further investigation does “not uncover sufficient evidence to charge terrorism offenses” or when prosecutors “discover that the case does not involve terrorism at all,” the terrorism-related designation in LIONS might remain in place. Id. at 3 (Krebs Decl. ¶ 7). The categorization is thus “often not an accurate indicator of what [a] case . . . actually involves,” and “it is not

12

uncommon for a case that includes a terrorism-related initial program categorization in LIONS . . . to result only in non-terrorism-related charges or convictions and for the public record in the case to not reflect any connection to terrorism.” Id. at 2–3 (Krebs. Decl. ¶ 6).

Based on these representations, the Department argues that, “if the docket numbers of convictions identified with an internal terrorism program category in LIONS are disclosed, many people who had previously been associated publicly only with crimes not related to terrorism will for the first time be publicly, and in some instances mistakenly, associated with terrorism or terrorist activity.” Dkt. 32-1 at 6. Such disclosure would, in the Department’s view, implicate greater third-party privacy interests than the Court recognized (or than the Department itself urged) in Brennan I. Id. at 16. The Department therefore requests that the Court “reconsider its decision . . . [and] hold that the Department has properly withheld the docket numbers in its database in order to protect the privacy of individuals who have not been linked publicly to terrorism.” Id. at 6.

On July 17, 2020, the Court held a telephonic status conference to address the third-party privacy interests implicated by the Department’s motion for reconsideration. See Minute Entry (July 17, 2020); see also Brennan II, 2020 WL 7685612, at *1. At the status conference, the Court expressed concern about the privacy interests of third parties whose cases are categorized as terrorism-related in the LIONS database “even though they were convicted of crimes bearing no self-evident connection to terrorism and even though the Department never made any public statement tying their cases to terrorism.” Id. Of particular concern was the possibility that the database included “some matters that were initially opened as terrorism investigations but that ultimately led to charges and convictions unrelated to terrorism.” Id. Although the public has a significant interest in learning about the possibility that the database over-reports the number of

13

terrorism convictions by including cases that resulted in convictions for crimes unrelated to terrorism, “the Court also recognized that the defendants in those cases may have a substantial privacy interest in preventing public disclosure of the fact that the Department even suspected their involvement in terrorism.” Id.

At the same time, however, the Court observed that the supplemental declaration from Krebs was “wholly conclusory and of little assistance to the Court.” Id. Although the Department maintained that it was common for cases to be categorized as terrorism in the database even when the investigation did not lead to terrorism-related charges or any other public link to terrorism, the Department “could not offer even a rough guess of how many such cases might exist.” Id. The Court therefore ordered the parties to file a joint status report “setting forth their proposed methodologies for identifying the number of cases that raise legitimate third-party privacy concerns and for segregating those cases from the ones for which the connection to terrorism is either evident from the nature of the charges or has previously been publicly disclosed.” Id.

In the joint status report, the parties were unable to reach consensus on a path forward.

Id. (citing Dkt. 37). In light of their dueling proposals, the Court entered an order on August 19, 2020, requiring the Department to take four steps intended to clarify both how many cases in the LIONS database are categorized as terrorism-related without ever having been publicly linked to terrorism and, relatedly, how difficult it would be to segregate those cases from ones whose connection to terrorism is more apparent. First, to ascertain the universe of records at issue, the Court ordered the Department to “identify the exact number of cases in the LIONS database since January 1, 2006, that include one of the six terrorism-related classifications and that resulted in a conviction or guilty plea.” Id. Second, the Court ordered the Department to

14

identify the subset of those cases “involving ‘a conviction or plea for an international terrorism offense that has a clear public connection to terrorism.’” Id. (quoting Dkt. 37 at 4). With respect to those docket numbers, which implicated minimal privacy interests and which were easily identifiable (and thus segregable), the Court denied the Department’s motion for reconsideration and ordered the docket numbers released “without delay.” Id. at *2. Third, turning to the remaining docket numbers, the Court ordered the Department to review a random sampling of 100 cases to determine whether they had been publicly linked to terrorism in press releases or public court records. Id. Fourth and finally, the Court ordered the Department to create a Vaughn index for the 100 cases in the sample detailing whether each case had been publicly linked to terrorism and, for those that had not, “a brief explanation of why the case appears to be miscategorized in the database.” Id.

The Department filed the results of its survey on October 23, 2020, in the form of a second declaration from Krebs and an accompanying Vaughn index. Dkt. 40-1 (2d Krebs Decl.). Krebs reported that the LIONS database includes 3,843 cases entered since January 1, 2006, that are tagged with one of the six terrorism-related classifications and that ended in a conviction or guilty plea. Dkt. 40-1 at 2 (2d Krebs Decl. ¶ 3). The Department separated out the 427 cases that ended in a conviction for an international terrorism offense with a clear connection to terrorism and released the docket numbers for those cases to Plaintiffs, consistent with the Court’s order in Brennan II. 1 Id. (2d Krebs Decl. ¶ 4). From the remaining 3,416 cases, which

1 In identifying which cases ended in a conviction for an international terrorism offense, the Department included, as directed by the Court, any cases “charged under any provision of Chapter 113B, 113C, or 118 of Title 18 or brought for conspiracy, attempt, or aiding and abetting a violation of Chapter 113B, 113C, or 118.” Brennan II, 2020 WL 7685612, at *1. Sweeping more broadly, the Department also included any cases in which the defendant was convicted or pleaded guilty under any of “18 U.S.C. §§ 32, 42, 112, 878, 1116, 1201(a)(4), 175, 175b, 229,

15

primarily involve offenses committed domestically, the Department selected a random sample of 100 cases, following the procedure mandated by the Court, and created a Vaughn index indicating the crime of conviction for each case and whether each case had been publicly associated with terrorism through a press release or in court records, such as the charging documents or sentencing memoranda. Id. at 2–3 (2d Krebs Decl. ¶¶ 5–8). In light of the Department’s prior assertion that it “ha[d] not identified any widespread inaccuracies with the LIONS data,” Dkt. 35 at 14, the sampling produced surprising results. The Vaughn index shows that of the 100 sampled cases, only eleven cases had ever been publicly linked to terrorism, at least as the Department defines a public link. 2 Id. at 3 (2d Krebs Decl. ¶ 9). With substantial overlap, those eleven cases included five references to terrorism in press releases, nine references in charging documents, six references in sentencing memoranda, and just one case with a terrorism-related sentencing enhancement. Id.

As to the 89 convictions that the Department contends were not publicly linked to terrorism, the Court had directed the Department to include “a brief explanation of why the case appears to be miscategorized in the database.” Brennan II, 2020 WL 7685612, at *2. In response, the Department “maintains that a case is not necessarily ‘miscategorized’ as terrorism- related merely because the publicly filed documents lack an express link to terrorism.” Dkt. 40-1

831, 2332a, 175c, 832, 956, 1203, 1993, 2332, 2332b, 2332f, 2332g, 2332h, 2339, 2339A, 2339B, 2339C, 2339D; 21 U.S.C. § 1010A; 42 U.S.C. § 2884; 49 U.S.C. § 46502; [or] 50 U.S.C. § 1705(b).” Dkt. 40-1 at 2 (Krebs Decl. ¶ 4). The Court is thus satisfied that the Department made a good-faith effort to release all docket numbers that are associated with international terrorism offenses as to which the connection to terrorism is clear from the nature of the offense. 2 Unlike with respect to international terrorism offenses, the Department in considering which cases had been publicly linked to domestic terrorism did not identify any statutes under which a conviction would necessarily bear a self-evident connection to domestic terrorism (even where records in a particular case did not reference “terrorism”).

16

at 4 (2d Krebs Decl. ¶ 11). According to the Department, “many such cases are categorized in the LIONS database with a terrorism-related categorization because some aspect of the underlying investigation may have related to terrorism.” Id. And Krebs attests that only five cases out of the sample of 100 “included a terrorism-related classification that was confirmed to have been entered in error.” 3 Id. The Vaughn index, however, tells a more complicated story. It does indicate that in 47 cases, the investigation suggested a possible link to terrorism, although what that means is not explained. Id. at 6–14. But in an additional 32 cases, the reason for the terrorism-related categorization is listed as “unknown.” Id. It thus seems at least possible that the error rate is far higher than five percent.

On November 10, 2020, the Department filed an updated version of the Vaughn index, including internal U.S. Attorneys’ Office case ID numbers, as requested by Plaintiffs. Dkt. 42. On November 23, 2020, Plaintiffs filed a response to the Department’s supplemental filing. Dkt. 43. The portion of the motion for reconsideration not already resolved is now ripe for decision.

II. ANALYSIS

Rule 1 of the Federal Rules of Civil Procedure admonishes courts and parties to “construe[], administer[], and employ[]” the rules to secure the “just, speedy, and inexpensive determination of every action.” Fed. R. Civ. P. 1. In most cases, those interests are disserved by permitting a disappointed party to relitigate a case after the Court has entered final judgment. In light of the interest of finality, the orderly administration of justice, and conservation of scarce judicial resources, motions for reconsideration under Rule 59(e) are “disfavored and should only be granted in extraordinary circumstances.” SEC v. Bilzerian, 729 F. Supp. 2d 9, 13 (D.D.C.

3 In his declaration, Krebs confesses error with respect to six of the 100 cases, but only five cases in the Vaughn index are tagged with “[e]ntered in error” in the explanation column.

17

2010). The standard for a court’s consideration of a Rule 59(e) motion is discretionary, and the motion “need not be granted unless the district court finds that there is an intervening change of controlling law, the availability of new evidence, or the need to correct a clear error or prevent manifest injustice.” Ciralsky v. CIA, 355 F.3d 661, 671 (D.C. Cir. 2004) (internal quotation marks and citation omitted). “Rule 59(e) permits a court to alter or amend a judgment, but it ‘may not be used to relitigate old matters, or to raise arguments or present evidence that could have been raised prior to the entry of judgment.’” Exxon Shipping Co. v. Baker, 554 U.S. 471, 485 n.5 (2008) (quoting 11 Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 2810.1 (2d ed. 1995)). For these reasons, in general, “courts will not address new arguments or evidence that the moving party could have raised before the decision issued.” Banister v. Davis, 140 S. Ct. 1698, 1703 (2020).

“[T]he interests of judicial finality and economy” apply with “special force” in FOIA cases, moreover, “because the statutory goals—efficient, prompt, and full disclosure of information—can be frustrated by agency actions that operate to delay the ultimate resolution of the disclosure request.” Senate of Puerto Rico ex rel. Judiciary Comm. v. U.S. Dep’t of Just., 823 F.2d 574, 580 (D.C. Cir. 1987) (quotations marks and citations omitted). “[F]airness to parties seeking disclosure ordinarily requires that they be accorded a full and concentrated opportunity to challenge and test comprehensively the agency’s evidence regarding all claimed exemptions.” Id. Correspondingly, an agency that has full access to the contested records in the first instance will generally be hard pressed to explain how a late-asserted rationale for withholding records is compelling but was nonetheless overlooked by the agency until after the court entered judgment.

18

Here, the Department seeks reconsideration on the ground that its withholding of docket numbers is justified under Exemption 7(C) “in order to protect the privacy of individuals who have” been convicted of a crime but have “not been linked publicly to terrorism.” Dkt. 32-1 at 6. Nothing prevented the Department from raising this argument in its motion for summary judgment; the Department simply failed to do so. Indeed, before deciding the parties’ cross- motions for summary judgment, the Court permitted the Department to file two supplemental declarations in support of its withholdings, yet at no point in any of its submissions did the Department raise the argument that it currently presses.

The Department does not argue—nor could it—that Rule 59(e) provides an open door for litigants to raise new arguments that were previously available but not previously made. Principles of finality, efficiency, and fairness, as well as the need to promote confidence in the courts’ judgments, all preclude relitigating fully briefed and adjudicated motions based on nothing more than disagreement with the Court’s ruling or the hope that an alternative argument might fare better than the arguments the litigant previously raised. In the Department’s view, however, those principles do not apply here because the parties’ prior briefs focused only on “cases with publicly identified connections to terrorism-related offenses.” Id. at 12. But saying that the parties did not focus on an issue is a far cry from showing that the parties could not have raised that issue. The Department moved for summary judgment before Plaintiffs, Dkt. 13, and it was free to raise whatever arguments and to marshal whatever evidence it pleased. It was the Department that decided how to frame the arguments that were before the Court, and, accordingly, it is no answer to say that “Plaintiffs and the Court” did not previously consider whether the Department’s initial categorization of a case as terrorism-related during the investigation phase might be disconnected from the nature of the later charges and conviction.

19

Dkt. 32-1 at 12. Unlike the Department, Plaintiffs and the Court had no way of knowing that the categorizations in the LIONS database are not updated as a case progresses and are not always accurate; indeed, it appears that the Department relies on the LIONS designations in publicly reporting the number of “terrorism/national security critical infrastructure” prosecutions brought each year. See, e.g., United States Attorneys’ Annual Statistical Report for Fiscal Year 2019 at 14, https://www.justice.gov/usao/page/file/1285951/download (reporting 225 “terrorism/national security critical infrastructure” cases filed against 287 defendants, with 181 convictions, 1 acquittal, and 38 dismissals).

If the pending motion implicated only the interests of the parties, the Court’s analysis might end here. But the motion also implicates (and, in fact, most directly implicates) the interests of third parties, “whose identity and information are [arguably] protected by FOIA,” Computer Pros. for Social Resp. v. U.S. Secret Serv., 72 F.3d 897, 903 (D.C. Cir. 1996), and who cannot be blamed for the Department’s delay in identifying any disconnect between the characterization of the case included in the LIONS database and the actual crime of conviction. In these circumstances, the Court must assess whether reconsideration is necessary to “prevent [a] manifest injustice to the[se] innocent third parties.” Changzou Laosan Grp. v. U.S. Customs and Border Prot. Bureau, 374 F. Supp. 2d 129, 132 (D.D.C. 2005); see also Delta Ltd. v. U.S. Customs and Border Prot. Bureau, 393 F. Supp. 2d 15, 17 (D.D.C. 2005) (“[I]t seems clear that injury to innocent third parties would fall beneath the ‘manifest injustice’ umbrella”); see also 11 Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 2810.1 n.18 (3d ed. April 2021 update). Although Plaintiffs present a substantial argument that the Department’s motion is “[i]mproper” and “[s]hould [n]ot [b]e [e]ntertained,” Dkt. 33 at 15, the Court must decline Plaintiffs’ invitation to reject the Department’s motion at the threshold.

20

But the Department still faces a heavy burden. The Court will consider the motion under the only available prong of the governing standard, which asks whether relief is necessary “to correct a clear error or [to] prevent manifest injustice.” Ciralsky, 355 F.3d at 671 (quotation marks and citation omitted). When analyzing the Department’s withholdings under Exemption 7(C), the Court must, as in Brennan I, “‘balance the public interest in disclosure against the [privacy] interest[s] Congress intended the Exemption to protect.’” Brennan I, 2020 WL 1189091, at *5 (alternations in original) (quoting Reporters Comm., 489 U.S. at 776). Putting the Rule 59(e) standard together with the Exemption 7(C) standard, the Court may grant the motion for reconsideration only if it concludes that its prior balancing of the public interest in disclosure against the privacy interests protected by Exemption 7(C) constituted a “clear error” or, if allowed to stand, would result in a “manifest injustice.” Ciralsky, 355 F.3d at 671. Upon consideration of the Department’s motion for reconsideration and its supplemental declaration and Vaughn index, as well as Plaintiffs’ responses, the Court concludes that release of at least some of the docket numbers at issue would constitute a manifest injustice. The Court will, accordingly, grant the motion for reconsideration in part. The Court cannot, however, accept the most sweeping version of the Department’s motion, which would relieve the Department of the obligation to segregate and to disclose even those docket numbers associated with cases that the Department has publicly associated with terrorism.

A.

For ease of analysis, the Court begins with the public interest in disclosure. As the Court explained in Brennan I, even before the Department’s recent revelations, that interest was substantial. The Court wrote:

[T]he public interest weighs heavily in favor of disclosure. Understanding how and when the Department categorizes cases as terrorism cases and following

21

trends relating to these prosecutions would shed light on the workings of government and “would inform . . . ongoing public policy discussion[s]” on [a]

range of issues, ACLU I, 655 F.3d at 13, from immigration, to national security, to prosecutorial priorities, see Reporters Comm., 489 U.S. at 766 n.18; Steinberg v. Dep’t of Just., 179 F.R.D. 366, 370 (D.D.C. 1998) (finding a “significant”

public interest in the disclosure of documents relating to the “criminal investigation of alleged counter-terrorist activities”).

2020 WL 1189091, at *8. But, as the Department’s more recent filings show, that analysis just scratched the surface of the public interest in disclosure. These most recent filings raise questions not only about how the Department characterizes terrorism-related cases, but also about whether the Department has long overstated the number of terrorism-related cases that it prosecutes. As Plaintiffs note, the entire purpose of their FOIA request is to help the public understand “which cases the Department characterizes as ‘terrorism’ cases,” Dkt. 33 at 26 (emphasis omitted), and a review of the Department’s filings in this case suggests that the Department itself lacks a grasp on that important question.

In its supplemental declaration and Vaughn index, as Plaintiffs emphasize, the Department “suggests that 89 out of the 100 cases it reviewed that are categorized as ‘terrorism- related’ in its own database have no public link to terrorism.” Dkt. 43 at 2 (emphasis omitted). The Court agrees with Plaintiffs that “the claimed disconnect between [the Department’s] internal categorization of cases as ‘terrorism-related’ and [its] charging decisions in, and public disclosures regarding, those cases only enhances the public interest in lock-step” with the asserted privacy interests. Id. Plaintiffs argue that the Department’s representation that most cases classified as terrorism-related in the LIONS database do not ultimately lead to charges that are publicly linked to terrorism suggests that the Department’s reporting could be both under- and over-inclusive. Id. To the extent that the database includes mistakes, cases that may have started as terrorism-related investigations but that did not lead to the filing of terrorism-related

22

charges, and cases that bear an indeterminate relationship to terrorism, the Department may be significantly overstating how many terrorism prosecutions it ultimately brings. Id. At the same time, if the Department has reason to believe that a large number of cases that it never publicly identified as having ties to terrorism were actually related to terrorism in some way, the public is not receiving an accurate picture of the true scope of domestic terrorism threats. Id. And to the extent the designation of a case as terrorism-related is made by individual Assistant United States Attorneys, without meaningful guidance or oversight, like cases might not receive like classifications.

These concerns, at least with respect to the possibility of over-reporting, are far from academic. When issuing its annual reports, the Department states the total number of terrorism- related cases as one number, without differentiating between those that led to publicly identified terrorism-related charges and those where the only connection to terrorism was a lead that might not have been pursued or might not have panned out in the early stages of the investigation. That is, in its public reporting and even in its earlier briefing in this case, the Department at least implies that all of these “terrorism-related” cases led to terrorism-related prosecutions. See United States Attorneys’ Annual Statistical Report for Fiscal Year 2019 at 14, https://www.justice.gov/usao/page/file/1285951/download (reporting 225 cases categorized as “Terrorism/National Security Critical Infrastructure” filed against 287 defendants resulting in 181 guilty verdicts); see also Dkt. 13 at 21 (referring to LIONS as containing “information on terrorism prosecutions”). The Department’s supplemental filings, however, suggest that only a small portion of the reported “terrorism-related” cases in fact led to charges that anyone publicly associates with terrorism. In its public reports, moreover, the government warrants that its data is accurate because U.S. Attorneys’ Offices “routinely examine current and historical data sets,

23

as well as look for trends, to confirm that the data [is] accurate and reliable.” U.S. Dep’t of Just., Fiscal Year 2019 Annual Performance Report and Fiscal Year 2021 Annual Performance Plan at 38, https://www.justice.gov/doj/page/file/1249306/download. But that review does not seem to encompass updating the categorization of cases in the LIONS database as those cases develop, catching all errors, or confirming the basis for each designation, all of which cast doubt upon the Department’s public statements based on LIONS data.

This lack of certainty about the data may warp other information that the Department provides to the public as well. In 2019, for instance, the Department reported that 96 percent of “counterterrorism” cases were favorably resolved, but it is unclear whether that data is limited to defendants who were convicted of terrorism-related offenses or also includes all cases categorized as terrorism-related in the LIONS database. Id. at 30. The Department also uses data on the number of terrorism cases that it prosecutes in its annual budget submissions to Congress, for instance seeking more than $43 million in Fiscal Year 2021 for the Criminal Division’s work “disrupting and defeating terrorist operations.” See U.S. Dep’t of Just., Criminal Division, Performance Budget FY 2021 Congressional Submission at 16, https://www.justice.gov/doj/page/file/1246356/download. The Department’s position in its motion for reconsideration that many cases that it categorizes as terrorism-related either lead to charges not related to terrorism or are categorized as terrorism for unknown reasons raises questions of substantial public importance about these statements. Disclosure of the contested records would thus “pierce the veil of administrative secrecy” and help the public understand “what their government is up to.” Brennan I, 2020 WL 1189091, at *7–8.

The public’s interest in how the Department pursues domestic terrorism threats is especially acute in the context of the current moment. In October 2020, the Department of

24

Homeland Security reported that “Domestic Violent Extremists present[] the most persistent and lethal” terrorist threat in the United States. See U.S. Dep’t of Homeland Security, Homeland Threat Assessment at 17 (Oct. 2020), https://www.dhs.gov/sites/default/files/publications/2020_10_06_homeland-threat- assessment.pdf. And the director of the Federal Bureau of Investigation recently testified to the Senate that domestic terrorism is “metastasizing across the country.” See Adam Goldman, Domestic Terrorism Threat Is ‘Metastasizing’ in U.S., F.B.I. Director Says, N.Y. Times (Mar. 2, 2021), https://www.nytimes.com/2021/03/02/us/politics/wray-domestic-terrorism-capitol.html. Which types of cases the Department thinks of as contributing to that threat, and how the Department goes about meeting that threat, are matters of the utmost public interest.

B.

Many, although not all, of the privacy interests at stake in this case are also weighty.

These third-party interests fall into several categories and raise distinct considerations:

1.

To start, Krebs now attests that the LIONS database includes convictions in cases that began as terrorism investigations but failed to “uncover sufficient evidence to charge terrorism offenses” or that found no evidence of “terrorism at all.” Dkt. 32-2 at 2–3 (Krebs Decl. ¶¶ 6–7). As a result, “it is not uncommon for a case that includes a terrorism-related initial program categorization in LIONS . . . to result only in non-terrorism-related charges or convictions and for the public record in the case to not reflect any connection to terrorism.” Id. at 3 (Krebs. Decl. ¶ 6). These representations, which are confirmed by the Vaughn index, are at odds with the Court’s understanding in Brennan I that releasing the docket numbers categorized as terrorism- related in the LIONS database would disclose only “‘information that has already been the

25

subject of a public proceeding.’” Brennan I, 2020 WL 1189091, at *7 (quoting ACLU I, 655 F.3d at 8). As such, the Court understated the privacy interest at stake when it observed that “the Department cannot plausibly argue (nor does it attempt to argue) that the public ‘will hear of’ the terror-related nature of the cases ‘for the first time merely because the Justice Department releases a list of docket numbers.’” Id. at *6 (quoting ACLU I, 655 F.3d at 10).

The asserted third-party privacy interest is compelling in those cases where the initial investigation suggested some possible connection to terrorism—and the case was categorized as terrorism-related on that basis—but the ultimate charges in the case were unrelated to terrorism. In such a case, the never-before-disclosed terrorism investigation and the publicly known non- terrorism charges could fairly be viewed as two distinct law enforcement matters. And disclosing the terrorism-related LIONS categorization in those cases would reveal information about the initial investigation (as opposed to about the prosecution) that had never been made public—namely, that the subject was at one time suspected of terrorism. When the records at issue in this case are analyzed through that lens, the rule from ACLU I that a criminal defendant possesses only a relatively weak privacy interest in information pertaining to his plea or conviction would apply to the previously disclosed non-terrorism prosecution but would have little or no bearing on the separate, undisclosed terrorism investigation.

The D.C. Circuit’s decisions in SafeCard Services, Inc. v. SEC, 926 F.2d 1197 (D.C. Cir.

1991), and Citizens for Resp. & Ethics in Wash. v. U.S. Dep’t of Just., 854 F.3d 675 (D.C. Cir. 2017) (“CREW”), bear on the question whether the privacy interests of those who were subject to these abandoned terrorism-related investigations outweigh the public interest in disclosure. In SafeCard, the D.C. Circuit adopted a categorical rule, permitting agencies to withhold “names and addresses of private individuals appearing in files within the ambit of Exemption 7(C),”

26

unless disclosure is “necessary in order to confirm or refute compelling evidence that the agency is engaged in illegal activity.” 926 F.2d at 1206. The court refined that categorical rule in CREW, holding that the “categorical rule of non-disclosure announced in SafeCard does not apply” to individuals who have already been publicly identified “as having been charged, convicted[,] or otherwise implicated in connection with” a crime.” 854 F.3d at 682. But such individuals, even those convicted of a crime, retain a privacy interest in “new information about [the] person’s conduct, going beyond the facts in the public record related to that person’s conviction and sentencing.” Id. In any event, as the CREW court further explained, “[t]here is little [a court] can conclude” about this privacy interest “in the abstract.” Id. at 683. The Court must, accordingly, engage in “a particularized weighing of the public and privacy interests that would be implicated by the disclosure.” Id.

Here, the Court is persuaded that disclosure of information about previously unknown terrorism investigations, which did not result in terrorism-related charges, raise concerns similar to those that animated the SafeCard rule and that underlie the D.C. Circuit’s decision in ACLU II. As the D.C. Circuit explained in ACLU II, “[t]he presumption of innocence stands as one of the most fundamental principles of our system of criminal justice: defendants are considered innocent unless and until the prosecution proves their guild beyond a reasonable doubt.” 750 F.3d at 933. But, as the court further explained, “[u]nfortunately, public perceptions can be quite different,” id., and many may unjustifiably assume that the subjects of terrorism-related investigations are, in fact, terrorists. This unfairness is particularly acute, moreover, when a publicly disclosed investigation never resulted in charges or a trial, leaving the subject of the investigation without a forum in which to clear his name. Indeed, the desire to forestall guilty verdicts in the court of public opinion while a case is still under investigation is one of the main

27

reasons that grand jury proceedings are subject to strict rules of secrecy. See Fed. R. Crim. P. 6(e); see also United States v. Sells Eng’g, Inc., 463 U.S. 418, 424 (1983) (“[B]y preserving the secrecy of the proceedings, we assure that persons who are accused but exonerated by the grand jury will not be held up to public ridicule.”). Disclosure of LIONS terrorism-related cases, moreover, would not serve the public interest in understanding which prosecutions the Department of Justice classifies as terrorism-related and how the Department prosecutes those cases. To be sure, the public might be interested in how the Department handles terrorism- related investigations. But specific criminal investigations, as opposed to prosecutions, are not generally subject to public disclosure—and for good reason.

Against this background, the Court concludes that the Department has properly invoked Exemption 7(C) to protect the identities of those individuals who were subject to terrorism- related investigations but were never charged with or convicted of a terrorism-related charge. The Court will, accordingly, grant the Department’s motion for reconsideration with respect to this class of docket numbers.

As explained below, however, the Court cannot tell whether the Department’s sample Vaughn index has appropriately limited the “investigation-suggested-possible-terrorism- connection” label to only cases in which the ultimate charges of conviction were truly unrelated to terrorism. It appears that the Department may have applied this same tag to cases in which the ultimate conviction was in fact connected to the terrorism-related investigation, but where that link to terrorism had not been made public. The Court’s holding in this section of the opinion does not apply to that class of cases, regardless of how they are labeled in the Vaughn index; instead, the Court deals separately with those cases below. See infra Section II.B.3.

28

2.

The Department’s sample Vaughn index also shows that some of the terrorism-related designations were entered in error, while other terrorism-related designations were entered for reasons that the Department, after investigation, could not ascertain (meaning, presumably, that the Department could not rule out that those designations were also entered in error). Dkt. 42-1 at 1–12. With respect to those designations entered in error, the Court is persuaded that the subjects’ privacy interests outweigh the public interest in disclosure. In many respects, these cases are, like the cases discussed above, most analogous to the SafeCard line of cases: the individuals’ names appear in law enforcement records and those records bear no relationship to any prior public disclosures or to the individuals’ criminal convictions. To be sure, the public has an interest in knowing that the LIONS database contains inaccuracies and that those inaccuracies may have resulted in erroneous public reporting from the Department. And the sting of public disclosure might be mitigated by a contemporaneous explanation that the terrorism-related designation was wrong. But just as members of the public might assume that someone who is subject to criminal investigation has done something wrong, they might give insufficient weight to the Department’s confession of error. Whether accurate or not, for most people it would be unwelcome news—and an unwelcome public disclosure—that they appear in the LIONS database under a terrorism-related categorization. The public interest in learning that the LIONS database is, at times, inaccurate, moreover, can be served through publication of a Vaughn index, like the sample one that the Department has already produced, which would explain that particular entries are being withheld because they are inaccurate.

The docket numbers for cases that are categorized as terrorism-related for “unknown”

reasons present a more difficult question. As to those cases, the Court concludes that the

29

Department needs to work harder to reach the best conclusion it can. If, after careful review, the Department cannot find any reason to conclude that the matter was, in fact, related to terrorism, it may make that representation in its final Vaughn index and may treat the designation as erroneous. But if it cannot reach that conclusion, the Department will need to decide whether the case should be included in one of the other buckets discussed in this opinion. 4 3.

The next category is the easiest to address: those cases that resulted in convictions and where the Department has already characterized the case as terrorism-related, in a press release, charging document, or sentencing memorandum. See Dkt. 40-1 at 3 (2d Krebs Decl. ¶ 9). As to those cases, the Court’s prior reasoning in Brennan I continues to apply, and there is no reason to revisit the Court’s conclusion requiring disclosure.

The Department does not disagree with any of that, but it nonetheless argues that the Court should relieve it of the duty to release this information because the process of sorting through 3,416 cases to identify those cases publicly linked to terrorism would be unduly burdensome. Specifically, Krebs explains that “LIONS does not provide . . . the functional capability to separate cases where the defendants were publicly connected to terrorism from those where there was no such public connection.” Id. at 4 (2d Krebs Decl. ¶ 13). Instead,

4 In addition to five cases designated in error and thirty-two cases designated for unknown reasons, the Vaughn index also lists five cases that were designated as terrorism-related in LIONS because the cases were “[r]elated to case with possible terrorism.” See Dkt. 42-1. In the absence of any further explanation, it is unclear to the Court exactly what that means. If the cases in question were classified as related to each other because they involved the same defendants or the same transactions, then a public disclosure that one case involved terrorism may be enough to show that the related case was also publicly identified as being connected to terrorism. With respect to these cases, as with the cases designated for “unknown” reasons, the Department must do more legwork to determine whether the cases were related to terrorism in a manner that the public could ascertain or whether the terrorism connection was purely internal to the Department.

30

prosecutors must conduct a “manual review of thousands of pages of records for a single case to determine whether there was any public reference to terrorism.” Id. at 5 (2d Krebs Decl. ¶ 15). But this issue is before the Court on the Department’s post-judgment motion for reconsideration, and Krebs suggests only that segregating the docket numbers would be “burdensome,” id., not impossible. In these circumstances, the Court is unpersuaded that requiring the Department to do the work required to identify the docket numbers that it can lawfully withhold would itself constitute a “manifest injustice.”

As discussed in greater detail below, although the Department has identified various international terrorism offenses that obviously relate to terrorism on their face, the Department has not identified any offenses that by their nature typically relate to domestic terrorism. If the Department can categorically identify certain offenses that are inherently related to domestic terrorism, that will lighten its burden of reviewing cases one by one. If the Department wants to withhold some of the docket numbers that the Court previously ordered released, it simply must do the necessary legwork to demonstrate that each docket number may be properly withheld.

4.

The preceding analysis resolves the motion for reconsideration with respect to every class of docket numbers identified in the sample Vaughn index. As noted above, of the 100 cases in the sample, the Department submits that eleven cases had been publicly identified as related to terrorism, forty-seven cases involved terrorism-related investigations but not prosecutions, thirty- two cases were categorized as terrorism-related for unknown reasons, five were related to other terrorism-related cases, and five were entered in error. The Court suspects, however, that some cases labeled as either “unknown” or “investigation suggested possible terrorism connection” actually fall into an additional category—those prosecutions that the Department believes were

31

terrorism-related but that it has never previously identified as such in any public filing, statement, or press release. Such cases, to the extent they exist, would present an especially difficult question.

As an initial matter, the Court concludes that, within this category, the Department may only seek to protect docket numbers associated with charges (and resulting convictions) that are not, on their face, terrorism-related. To take one example, the Department’s sample Vaughn index indicates that a conviction under 18 U.S.C. § 844(f)(1) had not been publicly linked to terrorism and was categorized as terrorism-related for “unknown” reasons. That statute criminalizes “maliciously damag[ing] or destroy[ing]” government buildings, vehicles, or other property “by means of fire or an explosive.” Id. Even if the Department did not issue a press release using the word “terrorism” in connection with that prosecution, it seems likely that the association between an offense committed under that statute and domestic terrorism would be self-evident—or at least as self-evident as with some of the prosecutions that the Department concedes bear an obvious nexus to international terrorism. In identifying those cases with “a clear connection to [international] terrorism,” the Department included “any cases where the individual was convicted or pleaded guilty under at least one” of thirty statutes. Dkt. 40-1 at 2 (2d Krebs Decl. ¶4). Some of those statutes refer to “terrorism,” see, e.g., 18 U.S.C. § 2332b (acts of terrorism transcending national boundaries), but many others do not, see, e.g., 18 U.S.C. § 42 (importation of injurious animals); id. § 112 (protection of foreign officials, official guests, and internationally protected persons); id. § 956 (conspiracy to kill, kidnap, maim, or injure persons or damage property in a foreign country). At least some of those statutes, moreover, are similar in key respects to statutes identified in the Department’s sample Vaughn index as not bearing a public connection to domestic terrorism. Compare id. § 112(b)(1) (prohibiting the

32

intimidation, coercion, or threating of foreign officials), with id. § § 115(a)(1)(B) (prohibiting threats to assault, kidnap, or murder federal officials). It would thus appear that the Department could prepare a list of statutes that bear a self-evident connection to domestic terrorism—and release the docket numbers for cases in the LIONS database that led to convictions under those statutes—just as the Department did for statutes bearing a self-evident connection to international terrorism. Thus, as a first step, the Court will require the Department to determine which, if any, of the statutes associated with domestic terrorism convictions in the LIONS database establish “a clear public connection to terrorism.” As to those cases, the Department will need to disclose the associated docket numbers, as required for international terrorism convictions in the Court’s prior decision.

The Court recognizes, however, that the LIONS database also includes convictions categorized as terrorism-related for offenses that may or may not bear a connection to terrorism, such as making a false statement, 18 U.S.C. § 1001, or unlawful possession or sale of a firearm by a person convicted of a crime or unlawfully present in the United States, id. § 924(d) & (g). As noted above, public disclosure of docket numbers in cases that fall in this category—that is, cases that the Department believes are, in fact, terrorism-related and has included in its various reports relating to terrorism prosecutions, but that the Department has never publicly identified as terrorism-related prosecutions, would serve the public interest by shedding light on how the Department defines “terrorism-related” crimes and how it prosecutes those cases. Those are topics or immense public importance, but, to the extent the Department draws these lines internally without ever disclosing publicly even what types of cases it has in mind when it reports to Congress and the public about its terrorism-related prosecutions, public debate and

33

scrutiny are both distorted and limited. The Court, accordingly, concludes that the public interest in disclosure is substantial—and far more substantial than the Court previously recognized.

But, at the same time, the corresponding privacy interests are also substantial—and, again, far more substantial than the Court previously recognized. It is one thing for the public to know that you have made a false statement in violation of federal law or have unlawfully possessed a firearm. It is another thing entirely for the public to know that, in the Department’s view, your unlawful conduct bore a relationship to “terrorism,” particularly when the Department never conveyed that belief to the Court or to the public while the relevant proceedings were pending, at a time when you could have challenged that characterization.

Although this issue presents a close question, the Court is persuaded that the public disclosure of docket numbers corresponding to cases that the Department has only internally characterized as “terrorism-related” would “constitute an unwarranted invasion of personal privacy,” 5 U.S.C. § 552(b)(7)(C), and that sustaining the Court’s prior judgment would constitute “a manifest injustice,” Ciralsky, 355 F.3d at 671. Disclosing this information would constitute a greater invasion of privacy than the potential disclosure at issue in ACLU II, which the D.C. Circuit nonetheless found sufficient to overcome a substantial public interest in disclosure. See ACLU I, 655 F.3d at 13 (finding a “considerable public interest” in disclosure); ACLU II, 750 F.3d at 935 (assuming “that the public interest in the disclosure . . . equals that in ACLU I”). Here, the public interest in disclosure is as substantial—if not more substantial—than the public interest at stake in ACLU I and ACLU II. But the privacy interests are also far more substantial than those at issue in ACLU I and are greater even than those at issue in ACLU II. In those cases, the D.C. Circuit’s analysis of the relevant privacy interests centered on the disclosure of information that was already public, and the court’s concern related to refocusing

34

public attention on past cases, whether they ended in conviction or acquittal. Here, in contrast, disclosure of docket numbers corresponding to cases that the Department has only internally characterized as “terrorism-related” would risk not only renewed attention on cases closed years ago, but would cast new and potentially damaging aspersions on those convicted of crimes that previously bore no public connection to terrorism. If the crime of conviction was one that, on its face, bore a connection to terrorism, public disclosure is unlikely to cause serious or unjust harm. But if the crime of conviction bore no obvious connection to terrorism, public disclosure of the Department’s internal characterization would not merely open an old wound but would risk inflicting a new one. In the Court’s view, such a disclosure is just a (short) step shy of the disclosure of an uncharged crime or inchoate investigation, and the rationale applied in SafeCard and its progeny to protect that information applies here as well. It follows that with respect to cases that the Department has identified as terrorism-related only internally, the balance tips in favor of withholding.

C.

In summary, the Department’s motion is granted in part, but the Department has more work to do. The Department must first identify those cases in the database involving a conviction or plea for an offense that bears a self-evident public connection to domestic terrorism and disclose those docket numbers. The Department must then sort through the remaining cases and disclose the docket numbers for any cases that have been publicly identified as relating to terrorism in a press release, charging document, sentencing memorandum, or the like. The Department may withhold the remaining docket numbers, to the extent the Department can attest in a Vaughn index that those cases were designated as terrorism-related by accident, involved terrorism-related investigations but not terrorism-related prosecutions, or were categorized as

35

terrorism-related only for the Department’s internal purposes, with no public reference or nexus to terrorism.

CONCLUSION

For the foregoing reasons, it is hereby ORDERED that the Department’s partial motion for reconsideration, to the extent it has not already been denied, is GRANTED in part and DENIED in part. The motion is DENIED with respect to docket numbers associated with (1) offenses that bear a self-evident connection to domestic terrorism or (2) cases that the Department has identified in a public statement or filing as connected to terrorism. Those docket numbers must be disclosed. The motion is GRANTED with respect to docket numbers for cases categorized as terrorism-related based on (1) an error, (2) a connection to terrorism in the investigation but not the prosecution, or (3) the Department’s internal designation. It is further ORDERED that the parties shall meet and confer and file a joint status report on or before July 15, 2021, setting forth their respective positions on (1) the Department’s processing rate for the remaining docket numbers and (2) whether the Court should again enter final judgment in the case or wait to enter a final order until after the records are processed.

SO ORDERED.

/s/ Randolph D. Moss

RANDOLPH D. MOSS

United States District Judge

Date: July 1, 2021