Garcia Ramirez v. U.S. Immigration and Custom Enforcement

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA WILMER GARCIA RAMIREZ, et al., : : Plaintiffs, : Civil Action No.: 18-508 (RC) : v. : : U.S. IMMIGRATION AND CUSTOMS : ENFORCEMENT, et al., : : Defendants. : FINDINGS OF FACT AND CONCLUSIONS OF LAW CONCERNING LIABILITY This case concerns alleged violations of the Administrative Procedure Act (“APA”) by the U.S. Immigration and Customs Enforcement (“ICE” or “the Agency”) in connection with ICE’s processing of eighteen-year-olds who came to the United States as unaccompanied alien children (“UACs”). Plaintiffs—immigrant teenagers who entered the United States as UACs— bring this class action against ICE, the Acting Director of ICE, the Department of Homeland Security (“DHS”), and the Secretary of Homeland Security (collectively “Defendants” or “the Government”). When minors lacking immigration status arrive in the United States without parents or other guardians, they are placed in the custody of the Department of Health and Human Services, Office of Refugee Resettlement (“HHS” and “ORR”). If they are still in custody on their eighteenth birthday, the now-adult immigrants “age out” of HHS and ORR custody and are transferred to ICE custody. Immigrants who undergo this transfer from HHS to ORR are referred to by the parties as “age-outs” and a subset of these age-outs make up the plaintiff class in this case. A provision of a 2013 statute amending the Trafficking Victims Protection Reauthorization Act (“TVPRA”) 1 codified at 8 U.S.C. § 1232 (c)(2)(B), requires that when ICE receives custody of an age-out it must “consider placement in the least restrictive setting available after taking into account the alien’s danger to self, danger to the community, and risk of flight.” 8 U.S.C. § 1232 (c)(2)(B). In short, the plaintiffs argue that a significant number of ICE field offices and officers automatically place many age-outs in adult detention settings without giving less-restrictive settings the consideration required. In their Amended Complaint, plaintiffs allege that ICE’s failure to consider the least restrictive placement violates the APA. In Count One, the Plaintiffs argue that it is arbitrary, capricious, or an abuse of discretion for ICE not to consider placement in the least restrictive setting. In Count Two, they argue that ICE is unlawfully withholding consideration of placement in the least restrictive setting and they ask the Court to compel the Agency to undertake such consideration. The Counts, though legally distinct, challenge the same conduct and both are focused on what ICE is actually doing with each age-out it encounters. Neither is about what official policy ICE might have written down. In both Counts, the Plaintiffs are alleging that ICE is not fulfilling its obligation to actually consider placement in the least restrictive setting for each age-out that comes into its custody. The Court conducted a bench trial over the Course of eighteen days between December 2, 2019 and January 15, 2020. Following the trial, the parties submitted proposed findings of fact and conclusions of law, as well as subsequent briefs in opposition. The Court now makes its 1 These amendments to the TVPRA were made as part of a broader set of amendments to the Violence Against Women Act (“VAWA”). See Violence Against Women Reauthorization Act of 2013, Pub. L. No. 113-4, § 1261 , 127 Stat. 54 , 56 (codified at 8 U.S.C. § 1232 (c)(2)(B)). As a result, the statute governing ICE’s handling of age-outs is sometimes referred to as the TVPRA and sometimes referred to as VAWA. 2 Findings of Fact and Conclusions of law, as required by Rule 52(a)(1) of the Federal Rules of Civil Procedure. For the reasons discussed in detail below, the Court will enter judgment in favor of Plaintiffs. The statute requires that ICE field officers take into account the statutory risk factors of danger to self, danger to community, and risk of flight, and that they consider placing age-outs in the least restrictive setting available. To consider placing an age-out in the least restrictive setting available, ICE officers must be able to identify what available setting would be least restrictive. This requires making an inquiry into available placements for age-outs that ICE officers throughout the country frequently do not undertake. ICE headquarters does not train officers on proper decisionmaking, and in fact gives guidance that is contrary to the statute in several ways. ICE does not require any number of practices that could facilitate compliance with the statute, but leaves it up to field officers’ discretion whether, for example, to review age-outs’ files in detail, to contact group homes and shelters, or to respond to communications from ageouts’ attorneys suggesting settings for them that are less restrictive than adult detention. Many officers choose not to take these steps, with the result that in many of ICE’s largest field offices, age-outs are detained nearly automatically. In the most extreme cases, this means that ICE field officers refuse to release age-outs to organizational sponsors who have said they would be happy to take them in, see infra ¶ 187, or to eighteen-year-olds’ own parents living in the United States, infra ¶ 200, when nothing in the age-outs’ records indicated they posed a flight risk or a danger to themselves or to the public. These are not the decisionmaking processes that Congress required. By failing to make decisions in the way Congress dictated, and based on the factors Congress identified as relevant, ICE fails to fulfill its obligations under the statute and therefore violates the APA. 3 I. BACKGROUND For clarity in following the Court’s specific findings of fact and conclusions of law, the Court first provides a brief overview of the facts underlying this case and a summary of the case’s procedural history and current posture. A. Factual Overview Most immigration enforcement functions in the United States are carried out by the DHS, in which ICE is housed. See 6 U.S.C. §§ 111 , 251, 291. Congress established a different legal framework, however, for the care and custody of “unaccompanied alien children”— defined as children under age eighteen, who have no lawful immigration status in the United States and no parent or legal guardian in the United States available to provide care and physical custody. 6 U.S.C. § 279 (g)(2). Except in certain exceptional circumstances, unaccompanied minors apprehended by immigration officials are transferred to the custody of HHS. 8 U.S.C. § 1232 (b)(3). ORR, an office within a division of HHS, is thereafter responsible for, among other things, “coordinating and implementing the care and placement” of such children. 6 U.S.C. § 279 (a)–(b)(1)(A). Congress has established that these children “shall be promptly placed in the least restrictive setting that is in the best interest of the child” and that “[i]n making such placements, the Secretary [of HHS] may consider danger to self, danger to the community, and risk of flight.” 8 U.S.C. § 1232 (c)(2)(A). HHS only has authority over the care and custody of immigrant children, however. See 6 U.S.C. § 279 . And, of course, children do not stay children forever. Congress accounted for that fact of life, extending certain protections to newly adult immigrants who were formerly in the care and custody of HHS. Pursuant to 8 U.S.C. § 1232 (c)(2)(B): If [an unaccompanied alien child in the custody of the Secretary of HHS] reaches 18 years of age and is transferred to the custody of the Secretary of Homeland 4 Security, the Secretary [of DHS] shall consider placement in the lease restrictive setting available after taking into account the alien’s danger to self, danger to the community, and risk of flight. Such aliens shall be eligible to participate in alternative to detention programs, utilizing a continuum of alternatives based on the alien’s need for supervision, which may include placement of the alien with an individual or an organizational sponsor, or in a supervised group home. Under this provision, DHS—that is to say, ICE—must “tak[e] into account” specified statutory factors and must “consider” placement in the least restrictive setting available for those who aged out of HHS’s jurisdiction. See id. But, unlike unaccompanied minors, these young adults are not promised placement in the least restrictive setting. Compare 8 U.S.C. § 1232 (c)(2)(A) (“shall promptly be placed”), with id. § 1232(c)(2)(B) (“shall consider placement”). B. Procedural History Plaintiffs filed this lawsuit on March 5, 2018. Compl., ECF No. 1. The plaintiffs were three immigrant teenagers who had previously been held in ORR custody as unaccompanied alien children. See Mem. Op. Denying Defs.’ Mot. Dismiss and Granting Pls.’ Mot. for Class Cert. (“MTD and Class Cert. Op.”), ECF No. 50 (citing First Am. Compl. (“Am. Compl.”) ¶¶ 1, 33, 46, 61, ECF No. 21). Each turned eighteen, was transferred to ICE custody, and was placed in an adult detention facility, purportedly without receiving the statutorily mandated consideration of less restrictive placement options. Id. (citing Am. Compl. ¶¶ 1, 4, 13–15). Plaintiffs assert two causes of action, both under the APA. First, pursuant to 5 U.S.C. § 706 (2), Plaintiffs argue that Defendants have acted in a manner that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” by failing to consider placements for each Plaintiff in the “least restrictive setting available after taking into account [their] danger to self, danger to community, and risk of flight” and by failing to make available to Plaintiffs any alternatives to detention. Am. Compl. ¶¶ 98–106. Second, Plaintiffs contend that Defendants 5 have ‘‘failed to take a discrete agency action that [the agency] is required to take,’’ in violation of Section 706(1) of the APA. Id. ¶¶ 107–11; 5 U.S.C. § 706 (1). Plaintiffs moved for class certification, ECF No. 6, and shortly thereafter filed an Amended Complaint, ECF No. 21. The Amended Complaint, which remains the operative complaint today, alleges that ICE’s handling of age-outs violates the APA in two ways. Count One alleges a violation of Section 706(2), which says that a reviewing court shall “hold unlawful and set aside agency action, findings, and conclusions found to be” among other things “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law” or “in excess of statutory jurisdiction, authority, or limitations, or short of statutory right.” 5 U.S.C. § 706 (2). Count Two is similar. It alleges a violation of Section 706(1) which says that a reviewing court shall “compel agency action unlawfully withheld or unreasonably delayed.” Id. § 706(1). The two counts focus on the same conduct. Count One alleges that ICE’s lack of compliance with Section 1232(c)(2)(B) makes its detention of age-outs arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law, while Count Two alleges that ICE’s failure to take required action under Section 1232(c)(2)(B) amounts to the unlawful withholding of required agency actions. See Am. Compl. ¶¶ 98–111. Defendants moved to dismiss arguing that the plaintiffs lacked standing to bring their claims, that their claims were moot, that ICE’s relevant actions were not reviewable under the APA, and that Plaintiffs failed to state claims upon which relief could be granted. MTD and Class Cert. Op. at 18. The Court disagreed and denied the motion to dismiss. Id. Among the Court’s conclusions in that opinion was that “Section 1232(c)(2)(B) does not limit ‘consider[ation]’ or ‘eligib[ility] to participate in alternative to detention programs’ to those who DHS has determined pose no risk of flight.” Id. at 40. Instead, DHS and ICE must account for 6 an age-out’s flight risk, danger to self, and danger to community and then “consider placement in the least restrictive setting available” no matter what the level of flight risk or danger has been assessed to be. Id. “[T]he statute calls for an individualized assessment of the proper placement for each former unaccompanied minor in light of DHS’s assessment of his or her danger to self, danger to the community, and risk of flight. It includes no provision permitting DHS to shortcircuit this inquiry . . . .” Id. In the same opinion that denied Defendants’ motion to dismiss, the Court granted Plaintiffs’ motion for class certification, allowing Plaintiffs to proceed on behalf of a class defined as: All former unaccompanied alien children who are detained or will be detained by ICE after being transferred by ORR because they have turned 18 years of age and as to whom ICE did not consider placement in the least restrictive setting available, including alternatives to detention programs, as required by 8 U.S.C. § 1232 (c)(2)(B). MTD and Class Cert Op. at 56–57. The case then proceeded to discovery. After the close of discovery, the Government moved for partial summary judgment. Mot. Partial Summ. J. (“MSJ”), ECF No. 209. The Court denied partial summary judgment because the same dispute of fact that the Government conceded precluded summary judgment on Count Two was equally central to the Plaintiffs’ case on Count One. Order Denying Mot. for Partial Summ. J. (“MSJ Order”), 2019 WL 7370368 at *1–2, ECF No. 240. Specifically, as the Government recognized, there was a genuine dispute concerning “whether ICE Officers in the field are indeed considering the least restrictive setting available.” Id. at *1 , ECF No. 240 (citing MSJ at 8 n.3). This issue would have to be resolved at trial. 7 C. Bench Trial The Court conducted a bench trial over the Course of eighteen days between December 2, 2019 and January 15, 2020. Transcript of Bench Trial (“Trial Tr.”), ECF Nos. 280–313. 2 Plaintiffs presented testimony from twenty-three witnesses at trial, calling some to the stand for live testimony and reading deposition testimony into the record for others. They also provided the Court with designated deposition testimony from a number of additional witnesses. Defendants called fourteen witnesses, some of whom had also been presented, in one form or another, by Plaintiffs. Both parties introduced a substantial volume of exhibits and designated significant deposition testimony for the Court’s consideration. Following the close of trial, each party submitted proposed findings of fact and conclusions of law to the Court. At the Court’s instruction, they have divided their arguments across two sets of briefs, with liability and remedies addressed separately. Defs.’ Proposed Findings of Fact and Conclusions of Law Concerning Remedies 3 (“Defs.’ Remedies Br.”), ECF No. 267; Defs.’ Proposed Findings of Fact & Conclusions of Law [Concerning Liability] (“Defs.’ Br.”), ECF No 268; Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Liability (“Pls.’ Br.”), ECF No. 271; Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Remedies (“Pls.’ Remedies Br.”), ECF No. 272; Pls.’ Mem. Concerning Remedies (“Pls.’ Remedies Mem.”), ECF No. 273. Plaintiffs also filed a series of demonstrative exhibits that they had referred to during trial. Notice of Filing Demonstrative Exs. (“Pls.’ Demonstratives”), ECF No. 275. The parties then filed responses to each other’s 2 The Trial Transcript has been consecutively paginated. Accordingly, the Court will only reference page and line numbers when citing to the transcript, rather than identifying the particular date of the cited testimony or the Electronic Case Filing Number assigned to that portion of the testimony. 3 The Court has supplied this title because Defendants did not title this document. 8 proposed findings of fact and conclusions of law. Pls.’ Resp. to Defs.’ Br. (“Pls.’ Resp.”), ECF No. 317; Pls.’ Resp. to Defs.’ Remedies Br. (“Pls.’ Remedies Resp.”), ECF No. 318; Defs.’ Reply to Pls.’ Remedies Mem. (“Defs.’ Remedies Mem. Resp.”), ECF No. 319; Defs.’ Resps. to Pls.’ Br. (“Defs.’ Resp.”), ECF Nos. 321, 323 4; Defs.’ Resp. To Pls.’ Remedies Br. (“Defs.’ Remedies Resp.”), ECF No. 322. Both parties redacted certain information in their briefing and each has filed unredacted versions of these documents under seal. 5 See Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Liability (“Pls.’ Sealed Br.”), ECF No. 270-1; Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Remedies (“Pls.’ Sealed Remedies Br.”), ECF No. 270-2; Defs.’ Proposed Findings of Fact & Conclusions of Law [Concerning Liability] (“Defs.’ Sealed Br.”), ECF No. 276; Pls.’ Resp. to Defs.’ Br. (“Pls.’ Sealed Resp.”), ECF No. 316-1; Defs.’ Resps. to Pls.’ Br. (“Defs.’ Sealed Resp.”), ECF No. 320-1; Defs.’ Resps. To Pls.’ Remedies Br. (“Defs.’ Sealed Remedies Resp.”), ECF No. 320-2. The Court will cite primarily to the publiclyfiled versions of these documents. Here, the Court only reaches findings of fact and conclusions of law concerning liability. The Court will address remedies in the near future. 4 Defendants responded to Plaintiffs’ findings of fact and conclusions of law in two separate filings. ECF No. 323 contains paragraphs 218 through 257 of Plaintiffs’ responses to Defendants’ proposed findings of fact concerning liability. Responses to all of Defendants’ other proposed findings of fact and conclusions of law concerning liability are found in ECF No. 321. 5 The parties each filed motions seeking the Court’s leave to make sealed filings pursuant to the Protective Order governing this case, ECF No. 63. Pls.’ Mot. to File Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Liability and Remedies Under Seal, ECF No. 270; Pls.’ Mot. To File Resp. to Defs.’ Br. Under Seal, ECF No. 316; Defs.’ Unopposed Mot. For Leave to File Under Seal, ECF No. 320. All three motions will be granted and the attached sealed briefings will be deemed filed. 9 II. LEGAL STANDARDS A. Rule 52(a) of the Federal Rules of Civil Procedure Pursuant to Rule 52(a) of the Federal Rules of Civil Procedure, in an action tried without a jury, the Court “must find the facts specifically and state its conclusions of law separately.” Fed. R. Civ. P. 52(a)(1). The Court’s “[f]indings and conclusions may be incorporated in any opinion or memorandum of decision the court may file.” Defenders of Wildlife, Inc. v. Endangered Species Scientific Auth., 659 F.2d 168 , 176 (D.C. Cir. 1981); see also Fed. R. Civ. P. 52(a)(1) (“The findings and conclusions . . . may appear in an opinion or memorandum of decision filed by the court.”). The Court’s findings must be “sufficient to indicate the factual basis for the ultimate conclusion.” Kelley v. Everglades Drainage Dist., 319 U.S. 415, 422 (1943); see also Lyles v. United States, 759 F.2d 941 , 943–44 (D.C. Cir. 1985) (“One of [Rule 52(a)’s] chief purposes is to aid the appellate court by affording it a clear understanding of the ground or basis of the decision of the trial court.” (internal quotation omitted)). “But the judge need only make brief, definite, pertinent findings and conclusions upon the contested matters; there is no necessity for over-elaboration of detail or particularization of facts.” Fed. R. Civ. P. 52(a) advisory committee’s note to 1946 amendment; accord Paleteria La Michoacana, Inc. v. Productos Lacteos Tocumbo S.A. DE C.V., 188 F. Supp. 3d 22, 34 (D.D.C. 2016); Wise v. United States, 145 F. Supp. 3d 53, 57 (D.D.C. 2015); Moore v. Hartman, 102 F. Supp. 3d 35, 65 (D.D.C. 2015). The Court is neither “require[d]” nor “encourage[d]” “to assert the negative of each rejected contention as well as the affirmative of those which they find to be correct.” Schilling v. Schwitzer-Cummins Co., 142 F.2d 82, 84 (D.C. Cir. 1944); Paleteria La Michoacana, 188 F. 10 Supp. 3d at 34. Similarly, “the court need not ‘address every factual contention and argumentative detail raised by the parties,’ or ‘discuss all evidence presented at trial.’” Moore, 102 F. Supp. 3d at 65 (quoting Mayagüez v. Corporación Para El Desarrollo Del Oeste, 824 F. Supp. 2d 289, 295 (D.P.R. 2011); Wachovia Bank N.A., Nat’l Ass’n v. Tien, 598 F. App’x 613 , 617–18 (11th Cir. 2014)) (internal citation omitted); see also Wise, 145 F. Supp. 3d at 57 (“[T]he Court need not address all the evidence presented at trial, and must simply make findings sufficient to allow the appellate court to conduct a meaningful review.”); Fasolino Foods Co. v. Banca Nazionale del Lavoro, 961 F.2d 1052, 1058 (2d Cir. 1992) (“All that is required by Rule 52(a) is that the trial court provide findings that are adequate to allow a clear understanding of its ruling.”); 9C Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 2579 (3d ed. 2008). In accordance with these standards, the Court has cited portions of the record, including trial testimony, trial exhibits, and designated deposition testimony, to support its findings of fact. The Court has not, however, exhaustively identified every portion of the record upon which it has relied to make its findings of fact, and the omission of a citation to a particular portion of the record does not necessarily mean that the Court did not rely on that portion to make its findings of fact. Rather the Court has considered the record in its entirety, including the Court’s own determination of witnesses’ credibility, and its citations to portions of the record are primarily intended to provide helpful references for the basis of its findings. The Court has not noted every instance in which a finding it adopts was purportedly disputed by a party but in which that disputing party failed to make a substantive argument against the finding. The Court has addressed those substantive arguments that present a material dispute of fact. The Court therefore has not addressed most instances in which a party purported 11 to dispute a proposed finding but only raised legal arguments concerning the legal consequences or implications of the proposed finding, without actually challenging the validity of the fact. Legal arguments are addressed in the Court’s conclusions of law section. Finally, the Court notes that Rule 52 provides that the Court’s “[f]indings of fact, whether based on oral or other evidence, must not be set aside unless clearly erroneous, and the reviewing court must give due regard to the trial court’s opportunity to judge the witnesses’ credibility.” Fed. R. Civ. P. 52(a)(6); see also Anderson v. City of Bessemer City, N.C., 470 U.S. 564 , 573–76 (1985) (discussing the “clearly erroneous” standard). B. Administrative Procedure Act Under 5 U.S.C. § 706 (2)(A), a reviewing court must set aside agency actions that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706 (2)(A). The touchstone of arbitrary-and-capricious review is reasoned decisionmaking. Harry T. Edwards, Linda A. Elliott & Marin K. Levy, The Requirement of Reasoned Decisionmaking: Arbitrary and Capricious Review Under the APA, Federal Standards of Review Ch. XV (Apr. 2013) (“Federal Standards of Review Ch. XV”) (citing Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)). Thus, a court cannot set aside an agency action that is “rational, based on consideration of the relevant factors[,] and within the scope of the authority delegated to the agency by the statute.” State Farm, 463 U.S. at 42–43. Although this is a deferential standard, it still requires a reviewing court to take a “hard look” at an agency’s reasoning. See Federal Standards of Review Ch. XV (quoting Nat’l Lime Ass’n v. EPA, 627 F.2d 416 , 451 n.126 (D.C. Cir. 1980)). The agency must provide at least a “‘brief statement’” of its reasoning that “explains ‘why it chose to do what it did,’” so that “the court can ‘reasonably discern[ ]’ the agency’s path.” Coe v. McHugh, 968 F. Supp. 2d 237 , 240 12 (D.D.C. 2013) (quoting Tourus Records, Inc. v. DEA, 259 F.3d 731, 737 (D.C. Cir. 2001); Pub. Citizen, Inc. v. FAA, 988 F.2d 186, 197 (D.C. Cir. 1993)) (alteration in original). The agency’s decision should be upheld “so long as the agency ‘engaged in reasoned decision making and its decision is adequately explained and supported by the record.’” Clark County v. FAA, 522 F.3d 437, 441 (D.C. Cir. 2008) (quoting N.Y. Cross Harbor R.R. v. Surface Transp. Bd., 374 F.3d 1177, 1181 (D.C. Cir. 2004)). The APA also authorizes courts to “compel agency action unlawfully withheld or unreasonably delayed[.]” 5 U.S.C. § 706 (1). “Unlike the provision that instructs courts to ‘set aside’ unlawful agency action, id. § 706(2), the [Section] 706(1) provision ‘provides relief for a failure to act[.]’” Ctr. for Biological Diversity v. Zinke, 260 F. Supp. 3d 11 , 19–20 (D.D.C. 2017) (citing Norton v. S. Utah Wilderness Alliance (“SUWA”), 542 U.S. 55, 62 (2004)). However, “a claim under [Section] 706(1) can proceed only where a plaintiff asserts that an agency failed to take a discrete agency action that it is required to take.” SUWA, 542 U.S. at 64 . A plaintiff who asks a court to “compel agency action . . . unreasonably delayed” under § 706(1) must pinpoint an agency’s failure to take an action that is both discrete and mandatory. See SUWA, 542 U.S. at 61, 64 . “[Section] 706(1) empowers a court only to compel an agency ‘to perform a ministerial or non-discretionary act,’ or ‘to take action upon a matter, without directing how it shall act.’” Id. (quoting Attorney General's Manual on the Administrative Procedure Act 108 (1947) (emphasis added)). The Court discusses the applicability of both provisions of the APA in more detail in its conclusions of law. 13 C. Burden of Proof In the typical civil trial, plaintiffs bear the burden of proving their case “by a preponderance of the evidence.” Escamilla v. Nuyen, 227 F. Supp. 3d 37, 47 (D.D.C. 2017) (citing Ascom Hasler Mailing Sys., Inc. v. U.S. Postal Serv., 885 F. Supp 2d 156, 181 (D.D.C. 2012)); see also, e.g., Woodby v. Immigration & Naturalization Serv., 385 U.S. 276, 285 (1966) (referencing the “burden of proof . . . generally imposed in civil cases and administrative proceedings” as being “the duty of prevailing by a mere preponderance of the evidence”). The parties have not discussed the burden of proof in their briefing in any detail, and Defendants do not appear to dispute that the preponderance of the evidence standard applies here. “[T]he preponderance of the evidence standard ‘simply requires the trier of fact to believe that the existence of a fact is more probable than its nonexistence before [the trier of fact] may find in favor of the party who has the burden.’” Barhoumi v. Obama, 609 F.3d 416, 424 (D.C. Cir. 2010) (quoting Concrete Pipe & Prods. of Cal., Inc. v. Constr. Laborers Pension Trust for S. Cal., 508 U.S. 602, 622 (1993)). “In the simplest of terms, preponderance of the evidence means more likely than not[.]” La Botz v. Fed. Election Comm’n, 889 F. Supp. 2d 51, 59 (D.D.C. 2012). The standard “demands only 51% certainty.” Nat’l Lime Ass’n, 627 F.2d at 453 n.139. For this bench trial, the preponderance of the evidence standard dictates that the Plaintiffs were required to convince the trier of fact—the Court—that it was, at a minimum, slightly more likely than not that ICE failed to comply with Section 1232(c)(2)(B). Plaintiffs are not required to disprove compliance, or to present evidence to counter every alternative theory Defendants may suggest to account for some fault or other in their compliance. See, e.g., infra ¶¶ 290 & n.37, 298 (discussing unfounded theories presented by Defendants relating to the parties’ expert statistical analysis). If the trier of fact is sufficiently convinced that it is more likely than not that 14 a violation occurred, the party with the burden of proof is under no obligation to present evidence—or even argumentation—to counter unpersuasive theories and speculation from the other side. See Dennis v. Ford Motor Co., 332 F. Supp. 901, 903 (W.D. Pa. 1971) (observing that a party bearing the burden of proof by a preponderance of the evidence “is not required to disprove every possibility conceived by [opposing] counsel’s inventive mind”), aff’d, 471 F.2d 733 (3d Cir. 1973); Miles v. Sec’y of Health & Human Servs., 142 Fed. Cl. 136 , 146 (2018), aff’d, 769 F. App’x 925 (Fed. Cir. 2019) (“Respondent is not required to disprove a theory of causation that the Special Master has already determined to be insufficient.”). III. FINDINGS OF FACT A. Plaintiffs 1. Named Plaintiffs 6 a. Sulma Hernandez Alfaro 1. Sulma Hernandez Alfaro was born on January 19, 2000 in Honduras. Trial Tr. at 689:7–9 (Alfaro). Her eighteenth birthday was January 19, 2018. Id. at 697:12–18 (Alfaro). 2. Ms. Alfaro left Honduras for the United States late in 2016, and arrived in the United States in December, 2016. Id. at 694:7–10; 694:22–695:5 (Alfaro). 3. Ms. Alfaro turned herself in to immigration officials, was identified as an unaccompanied alien child, and was placed in the custody of ORR, in a shelter in San Benito, Texas. Id. at 695:6–19, 696:20–697:1 (Alfaro); see also DX-7 at 2–3, 17. 4. At some time before her eighteenth birthday, Ms. Alfaro filed an application for asylum. Trial Tr. at 706:11–13 (Alfaro). 6 In April 2019, the parties stipulated to the dismissal of a third named Plaintiff, who had been identified as Ana P. She took no further part in this case. See Stipulation to Dismiss Plaintiff Ana P. Under Federal Rule of Civil Procedure 41(a)(1)(A), ECF No. 143. 15 5. On or before January 17, 2018, La Posada Providencia (“La Posada”), “a transitional shelter for people in crisis . . . who are seeking legal refuge in [the United States]” agreed to shelter Ms. Alfaro when she turned eighteen. Id. at 706:14–19 (Alfaro); DX-7 at 23. 6. On her eighteenth birthday, Ms. Alfaro was taken into ICE custody. She was picked up by ICE and taken from the ORR shelter where she was being housed to an ICE office for processing, and was then taken to the Port Isabel ICE Detention Center in Harlingen, Texas. Trial Tr. at 697:23–698:17 (Alfaro); DX-7 at 2, 17. 7. No immigration officer ever spoke with Ms. Alfaro about release to La Posada or to any other sponsor, or any other alternative to detention (i.e. release on bond or with an ankle monitor). Id. at 706:20–707:18 (Alfaro). She did not bring up La Posada with the ICE officers processing her because she was not asked. Id. at 730:24–31:1 (Alfaro). 8. On February 2, 2018, Ms. Alfaro’s immigration counsel contacted ICE and unsuccessfully requested that Ms. Alfaro be released to La Posada Providencia. Ltr. from ProBAR to ICE (Feb. 2, 2018), ECF No. 2-9; see also Mem. Op. Granting Prelim. Injunction (“P.I. Op.”) at 9, ECF No. 28. 9. On March 21, 2018, an Immigration Judge granted Ms. Alfaro a bond in the amount of $10,000, but she and her family were unable to pay this amount so she remained in ICE detention. Trial Tr. at 707:19–708:2 (Alfaro); DX-8. 10. On April 18, 2018, this Court granted Plaintiffs’ Motion for a Preliminary Injunction and ordered ICE to comply with § 1232(c)(2)(B) in placing Ms. Alfaro. Order, ECF No. 27. 11. On April 24, 2018, Ms. Alfaro was granted Asylum and she was released from ICE custody. Trial Tr. at 690:1–2, 708:3–4 (Alfaro). 16 b. Wilmer Garcia Ramirez 12. Wilmer Garcia Ramirez was born on September 23, 1999 in Guatemala. Trial Tr. at 739:23–740:13 (Ramirez). His eighteenth birthday was September 23, 2017. Id. at 749:1–3 (Ramirez). 13. Mr. Ramirez left Guatemala for the United States in early 2017, and arrived in the United States on March 26, 2017. Id. at 743:20–21, 744:9–11 (Ramirez). 14. Mr. Ramirez was apprehended by immigration officials, was identified as an unaccompanied alien child, was placed in the custody of ORR, and was taken to ORR’s Southwest Key Glendale shelter, located in Glendale, Arizona. Id. at 744:14–745:2 (Ramirez); see DX-11. 15. On September 19, 2017, a state court in Arizona granted a petition declaring Mr. Ramirez a dependent of the state due to his parents’ neglect of him in Guatemala and found that it was not in Mr. Ramirez’s best interest to be returned to Guatemala. Order Regarding Child’s Eligibility for Special Immigrant Juvenile Status at 3, ECF No. 2-4. Based on this, he then filed a petition for Special Immigrant Juvenile Status (“SIJS”), seeking lawful permanent residency in the United States. Petition, ECF no. 2-3; see also P.I. Op. at 4. 16. On September 22, 2017—the day before Mr. Ramirez’s eighteenth birthday—Mr. Ramirez’s immigration counsel wrote to ICE explaining that Mr. Ramirez’s removal proceedings had been administratively closed, and that he had a petition for SIJS pending. His counsel requested that Mr. Ramirez be released on his own recognizance to stay with a family friend in Pennsylvania. DX-11 at 1. The request was denied, and ICE explained to Mr. Ramirez’s counsel that it intended to reopen removal proceedings, though it had not yet done so. Id. at 3–5. 17. On his eighteenth birthday, Mr. Ramirez “had understood that” he would be going to live with his family friend in Pennsylvania. He was instead picked up from his ORR facility by 17 ICE and taken to an ICE adult detention facility in Eloy, Arizona. Trial Tr. at 749:1–14 (Ramirez). 18. ICE did not explain why it rejected the possibility of his living with the family friend, nor did it discuss any alternative sponsors or other alternatives to detention with Mr. Ramirez. Id. at 751: 13–25 (Ramirez). 19. On January 5, 2018, Mr. Ramirez’s counsel again requested that Mr. Ramirez be released from custody “to the least restrictive setting available.” DX-17 at 1. Albert Carter, ICE Deputy Field Office Director for the Phoenix Field Office, denied the request, indicating, “I find the totality of circumstances do not support a favorable exercise of discretionary authority in this case.” DX-18. 20. On April 18, 2018, this Court granted Plaintiffs’ Motion for a Preliminary Injunction and ordered ICE to comply with § 1232(c)(2)(B) in placing Mr. Ramirez. Order, ECF No. 27. 21. Shortly after that Order was issued, in April 2018, Mr. Ramirez was released from ICE detention with instructions to appear periodically for hearings. He went to live with an uncle in Bessemer, Alabama. Trial Tr. at 740:18–23, 753:8–23 (Ramirez); see also P.I. Op., at 35–36 (indicating Mr. Ramirez was still in custody at the time of the Court’s Order). 2. Class 22. The Court certified a class under Federal Rule of Civil Procedure 23(b)(2) and defined it as consisting of: All former unaccompanied alien children who are detained or will be detained by ICE after being transferred by ORR because they have turned 18 years of age and as to whom ICE did not consider placement in the least restrictive setting available, including alternatives to detention programs, as required by 8 U.S.C. § 1232 (c)(2)(B). MTD and Class Cert. Op. at 69. 18 3. Harm to Class Members 23. Plaintiffs’ expert Dr. Julie Linton presented unrebutted expert testimony concerning the forms of harm that eighteen-year-olds may suffer as a result of detention. Dr. Linton is a pediatrician, a professor and assistant dean for admission as the University of South Carolina School of Medicine in Greenville, South Carolina, and medical director of PASOs, a program that “connects Latino or Hispanic families to community-based resources.” Trial Tr. at 2390:10– 19 (Linton). She has published papers on pediatrics and on care for immigrant children specifically. Id. at 2396:12–24. The Court found her to be well-qualified as an expert, and Defendants did not object at trial to her qualifications as an expert. Id. at 2404:6–9. She reached and testified to “three core opinions” in this case. Id. at 2404:23 . 24. Dr. Linton’s first expert opinion was that “because youth from ages 15 to 24 are still undergoing critical brain development in response to their environment and their experiences . . . the experience of detention for 18-year-old immigrant youth would be particularly detrimental to this continuing development. Id. at 2404:24–2405:3; see id. at 2405:12–2419:6 (elaborating on this opinion in more detail and explaining Dr. Linton’s basis for reaching it). 25. Dr. Linton also provided the expert opinion that “detention, particularly in adult settings, is physically and psychologically harmful to children and youth, and threatens both their short- and long-term health and well-being.” Id. at 2405:4–7; see id. at 2419:19–2422:7, 2425:15–2426:13 (elaborating on this opinion in more detail and explaining Dr. Linton’s basis for reaching it). 26. Dr. Linton’s third expert opinion was that “alternatives to detention, including placement in community settings . . . while awaiting court cases are markedly better for the 19 health and well-being of youth.” Id. at 2405:8–11; see id. at 2426:14–2428:23 (elaborating on this opinion in more detail and explaining Dr. Linton’s basis for reaching it). 27. This testimony was not rebutted by Defendants. Defendants briefly cross-examined Dr. Linton, id. at 2429:2–2437:7. During Dr. Linton’s direct testimony, the Court observed that “there’s case law that says deprivation of liberty is irreparable [harm] in and of itself,” and lead counsel for Defendants agreed, commenting that “this is not going to be a contested issue” and that Defendants were “not contesting that as a legal position.” Id. at 2423:6–2424:9. B. Defendants 1. ICE Structure 28. ICE has 24 field offices nationwide and many more sub-field offices. Trial Tr. at 4311:15–21 (Harper). 29. Mellissa Harper has been the unit chief of ICE’s Juvenile and Family Residential and Management Unit (“JFRMU”) since the fall of 2017. Id. at 46:6–17 (Harper). 30. JFRMU is an ICE headquarters unit within Enforcement and Removal Operations (“ERO”) that—among other duties—oversees implementation of § 1232(c)(2)(B). Id. at 46:6– 11, 279:11–12 (Harper); see also DX-4 at 12 (“Introduction to JFRMU”). This involves providing advice, training, and guidance to ICE field officers regarding age-out custody determinations, collecting data, responding to inquiries from the field, and dealing with related issues. Trial Tr. at 46:24–48:2 (Harper). 31. Beyond Chief Harper, other JFRMU personnel who testified in this case were National Juvenile Coordinators Ana Sanchez-Zimak and Eric Ravenell, and Section Chief Dawn Helland. Sanchez Dep. Desig. at 9:18–10:1; Ravenell Dep. Desig. at 9:15–20; Trial Tr. at 4004:4–6 (Helland). 20 32. The term “age-out” is used to refer to a UAC who reaches the age of 18 while in ORR custody and is then transferred to ICE custody. E.g. Trial Tr. at 46:18–23 (Harper). 33. JFRMU’s instructions to ICE field officers come in the form of “broadcasts” or “guidance,” which are understood by field officers to constitute orders from ICE headquarters. Id. at 4388:9–4389:2 (Harper). These are sometimes transmitted via e-mail. E.g. DX-6. 34. Field Office Juvenile Coordinators (“FOJCs”) are ICE officers who work in the Juvenile Units of ICE field offices. See Galvez Dep. Tr. at 7:19–8:6; DX-19 at 2. FOJCs serve as local points of contact for all juvenile-related issues in ICE field offices and Areas of Responsibility (“AORs”). This includes requesting ORR placements for UACs, screening apprehended UACs, processing age-outs, and conducting site visits. DX-23 at 4. 35. FOJCs are responsible for making “recommendations” of custody determinations for age-outs, and those recommendations are approved by their supervisors, called SDDOs (Supervisory Deportation and Detention Officers). Trial Tr. at 54:12–54:24 (Harper); see DX-4 at 9 (list of acronyms). JFRMU and Chief Harper do not to make custody determinations in ordinary cases. Trial Tr. at 54:25–55:9 (Harper). Chief Harper testified that “[g]enerally speaking,” it is true that “JFRMU defers to the field office with regards to detention decisions.” Id. at 56:13–15 (Harper). 36. There has been “significant rotation” of field officers into and out of the FOJC position, with FOJCs in many field offices rotating out of the role every one or two years. Trial Tr. at 149:19–150:8 (Harper). 37. The following forms of training are available to new FOJCs: annual in-person training held by JFRMU once a year, monthly calls, access to online resources, and on-the-job training from an outgoing FOJC. Id. at 150:11–151:2 (Harper); see id. at 489:7–10 (Officer 21 Ravenell agreeing that “the only training that officers receive when they become FOJCs is on the job training”). 2. Age-Out Placement Alternatives 38. ORR caseworkers sometimes prepare post-18 plans for UACs in ORR custody who are approaching their eighteenth birthday. Id. at 72:14–73:17 (Harper). There is no set format for these and they can range from a single sentence to a longer document. Id. 72:23–24 (Harper). They may include various types of information, including the age-out’s history, including medical and mental health history. Id. 72:20–23 (Harper). Most importantly for Section 1232(c)(2)(B), post-18 plans sometimes include names and contact information for potential sponsors, including potential sponsors that ORR has contacted but did not place an age-out with (if a UAC were placed with a sponsor by ORR, they would not become an age-out). Id. 73:3–5, 73:21–74:2 (Harper). 39. Placement with a sponsor is one of several options ICE FOJCs have for placing an age-out. Section 1232(c)(2)(B) explicitly references two options for placement. It says that ageouts “shall be eligible to participate in alternative to detention programs, utilizing a continuum of alternatives based on the [age-out]’s need for supervision, which may include placement of the [age-out] with an individual or an organizational sponsor, or in a supervised group home.” 8 U.S.C. § 1232 (c)(2)(B). 40. Releasing an age-out to live with a sponsor is one option available to FOJCs processing age-outs. This entails releasing the age-out to reside at the address provided by a sponsor where the sponsor will also live and provide support to the age-out, and where the ageout will be able to receive immigration court notices and similar communications. See, e.g., Trial Tr. 434:24–435:7 (Harper); Munguia Dep. Desig. at 193:20–195:11. 22 41. Release to an organizational sponsor or supervised group home is another option available. See PX-1 at 8 (age-out review worksheet asking “[a]re any individual or organizational sponsors or supervised group home available to this age-out for placement?”). Like individual sponsors, organizational sponsors also provide a support for age-outs and help ensure their appearance at any future immigration proceedings. Id. at 4339:19–25 (Harper) (reading into the record ERO guidance from DX-108 at 2); see also id. at 844:20–25 (Enriquez) (testimony from a third-party legal services provider that “an institutional sponsor might be a homeless shelter or an organization that provides housing services to youth”). 42. Organizational sponsors did not need to be in an office’s AOR, as some shelters accepted age-outs from across the country. See Trial Tr. at 845:20–847:14 (Enriquez) (testifying that Catholic Charities has always been able to identify a potential sponsor organization for ageouts by first reaching out to local shelters and then expanding the search to organizations around the country); id. at 1317:14–25 (Lehner) (Miami-area legal services provider testifying that she was able to find placement for age-outs from the Miami AOR with shelters in Chicago and Texas); see also DX-88 (AORW for a Miami office age-out form whom Officer Reardon recommended placement in a Texas shelter). 43. Age-outs can be released under an Order of Supervision (“OSUP”) or on their own recognizance (“OREC”). See, e.g. PX-1 at 8 (age-out review worksheet listing these as among the “Discretionary Release Action(s)” available). In practical terms these are the same—the age out is released—but release on an Order of Supervision is appropriate when the age-out is under an existing order of removal and is being supervised by an immigration court, whereas release on the age-out’s own recognizance is appropriate when no order of removal has been issued. See Trial Tr. at 942:7–943:2, 1107:1–22. 23 44. ICE’s Alternatives to Detention (“ATD”) program provides ICE officers with a number of other means for tracking age-outs who are not placed in detention. Carbonneau Dep. Des. at 7:5–8:5. 7 These include tracking by ankle bracelet monitor or by smartphone application, telephonic reporting, or ICE’s Intensive Supervision Appearance Program, a program through which outside service providers conduct office visits, home visits, residence visits, orientation enrollment and other forms of tracking. Id. at 10:5–10, 11:4–12:7, 56:10–57:5; see also Trial Tr. at 146:13–147:1 (Harper). ICE field officers can release age-outs on ATD anywhere in the country, with the exception of certain rural or sparsely populated areas. Carbonneau Dep. Desig. at 80:12–21; Trial Tr. at 3697:23–3698:11 (Irizzary). 45. ICE’s ATD Unit Chief Eric Carbonneau confirmed that age-outs are eligible to participate in the ATD program. Carbonneau Dep. Desig. at 115:13–16; see also id. at 75:20– 76:14, 77:8–10. He also explained that the ATD program is “an effective flight risk mitigation tool,” and that it “has demonstrated great success in improving compliance rates for those aliens assigned to the program.” Carbonneau Dep. Des. at 27:11–13; see also PX-21 (memo from Carbonneau to field officers making the same claims about the ATD program’s effectiveness). 46. ICE officers can also release age-outs on an ICE bond. Similar to a criminal bond, this entails the age-out “pay[ing] a bond as collateral for their release from ICE custody.” Trial Tr. at 159:12–20 (Harper). The minimum amount for an ICE bond is $1,500. Trial Tr. at 3143:12–16 (Hyde). 7 When referring to ICE’s ATD program, the Court uses the acronym ATD, or capitalizes the term (“Alternatives to Detention”). Section 1232(c)(2)(B) uses the term “alternatives to detention” in lowercase, more generically. When referring to placements other than detention, the court sometimes uses this phrase without capitalizing or, to avoid confusion, refers to “detention alternatives.” 24 47. Finally, ICE can also take an age-out into ICE custody. When ICE does this, the eighteen-year-old age-out is placed in an adult detention facility. Id. at 83:17–84:15. ICE has adult detention facilities with several different levels of security. Id.at 83:21–23 (“We have low, low-medium, medium, medium high.”). 3. Documentation of Age-Out Custody Determinations 48. In discovery in this case ICE produced evidence of age-out custody determinations between April 2016 and May 2019. See, e.g., id. at 18:12–13, 311:13–15 (Plaintiffs’ counsel representing that this is “the period for which ICE has provided data”). Plaintiffs summarized these records in PX-334, as well as in other exhibits. See id. at 1657:16–1665:24 (introducing PX-334 and explaining its contents), 2120:16–2121:25 (describing revisions from PX-334 to PX334A); see also id. at 2169:8–2171:2 (counsel discussing admissibility of PX-334A and explaining the data sources provided by ICE from which Plaintiffs’ counsel constructed summary charts); id. at 4223:12–4228:12 (testimony concerning ICE’s responses to Freedom of Information Act Requests, see PX-505, reflecting that the earliest data ICE could compile concerning age-out custody determinations was from April 2016). 49. For the time period from April 2016 through October 2018, ICE tracked age-out custody determinations through an informal monthly age-out report in the form of a spreadsheet. Id. at 284:2–13; Harper 12/6/18 Dep. Desig. at 27:19–21. 50. Monthly age-out reports tracked the name and A-number (alien number) of each ageout, the field office and/or AOR in which they were processed, their date, and country of birth. Id. at 284:18–285:13. From April 2016 through approximately January 2018 the report also included a column tracking the “Reason for Detention,” but around January 2018, at Chief Harper’s instruction, the spreadsheet was changed to instead ask, “If not detained, reason why?” 25 Id. at 285:18–286:21. This subtle change shifted the administrative burden on the field office decisionmaker from justifying the reason an age-out was detained to subsequently justifying the reason an age-out was released. 51. On October 17, 2018, after this court had granted Plaintiffs’ motions for a preliminary injunction and to certify the class, JFRMU / ERO announced to field officers that they would be using a new system to document age-out custody determinations. The new system involved uploading documentation to a SharePoint database and filling out an Age-Out Review Worksheet (“AORW”) for each age-out that was processed by ICE. PX-1; Trial Tr. at 250:20–251:6 (identifying PX-1 as a memorandum from Chief Harper dated October 17, 2018 “announcing . . . the new worksheet and SharePoint documentation system”); id. at 524:22–525:22 (same). 52. The memorandum announcing the new system went to all field office directors and deputy directors and announced that “[t]he process consists of a standardized worksheet that the FOJCs will use to document their considerations and a SharePoint site to track the entire population of age-outs.” PX-1 at 1. 53. The SharePoint database required FOJCs to fill out the same basic information regarding each age-out that they had entered into the monthly age-out reports under the earlier system, including the revised entry for “Non-Detention Reason . . . If not detained reason why[.]” It also allowed for uploading of completed AORWs. See id. at 3–4. The database would also “serve as the repository for the completed worksheets and other documentation supporting ICE’s custody determinations for individual age-outs.” Id. at 1 . 54. The AORW / SharePoint documentation system was designed only to document decision processes that were already in place. It was not intended to change anything about the decisionmaking processes. The JFRMU / ERO guidance emphasized that “The new worksheet 26 and tracking system are critical to ensure that ICE’s compliance with the law is properly captured. It is imperative that FOJCs ensure that all age-outs, and the considerations in [FOJCs’] custody determinations for each age-out, are documented in the SharePoint site within 24 hours of the transfer or release from ORR custody.” PX-1 at 1. Defendants apparently dispute this at one point in their response. They argue that the new documentation system was “not intended to change how officers made custody determinations so long as they were doing it properly,” implicitly suggesting that it was intended to change or correct improper determination practices. Defs.’ Resp. ¶ 47. The Court finds otherwise because this new suggestion, in support of which Defendants cite no testimony or evidence, is at odds with Chief Harper’s testimony at trial and with the text of the guidance introducing the worksheet. Trial Tr. at 251:18–252:13 (Harper), 526:1–22 (Harper) (“[T]he process still remained the same, it just added documentation . . . .”); see also PX-1 at 1 (“The new process merely helps document the work that officers already do regarding custody determinations.”). FOJCs understood this as well. Munguia Dep. Desig. at 36:22–37:3; Trial Tr. at 2771:21–2772:2 (Munguia); Hyde Dep. Desig. at 102:13–103:5; Barnes Dep. Desig. at 56:16– 57:20; Reardon Dep. Desig. at 36:1–37:16; Galvez Dep. Desig. at 149:21–150:16; Venegas Dep. Desig. at 50:8–51:10; Black Dep. Desig. at 73:23–74:17; Trial Tr. at 2646:15–2647:14 (Black); Zanello Dep. Desig. at 168:24–169:9. 8 Moreover, despite apparently disputing Plaintiffs’ proposed finding of fact on this point, Defendants then do not dispute Plaintiffs’ subsequent proposed findings that “the evidence is undisputed that the AORW and SharePoint “documentation system” were not intended to change how officers made custody determinations, 8 Officer Nancy Zanello’s name was Nancy Sullivan at the time of her deposition, but the Court uses her current last name for consistency with her testimony at trial. See Trial Tr. at 3770:24–3771:2 (Zanello). 27 or to ensure that those decisions were made in compliance with the statute” or that “[t]he officers in the field implementing the new documentation system understood the system in the same way.” Defs.’ Resp. ¶¶ 48, 49. 55. The new documentation system was intended to capture and record all the information an FOJC considered in making a custody recommendation. PX-1 at 1 (“[I]t is imperative that FOJCs ensure that all age-outs, and the considerations in their custody determinations for each age-out, are documented . . . .”); Trial Tr. at 4367:13–16 (Harper) (testifying that any sponsor considered by an FOJC is “supposed to” be documented on an AORW); see also Defs.’ Resp. ¶ 54 (not disputing this). 56. The October 2018 AORW asked for basic information concerning an age-out (AOR, age-out’s name, date of birth, alien number, time in custody, gender) and for the FOJC’s name, and then included 16 items, which were as follows, see PX-1 at 6–8: a. Item 1 asked whether the age-out’s identity and citizenship had been sufficiently determined and provided boxes which could be checked Yes or No. b. Item 2 asked about the “Health and Notoriety” of the age-out: whether the age-out was “a public figure,” whether there was “Congressional or media interest” in the age-out, and whether the age-out had “[k]nown or reported medical issues.” Again these were yes-or-no questions. c. Item 3 was headed “Database Queries” and asked whether there were positive or negative results to a search for the age-out in each of a variety of databases, including “Benefits Pending” and “Asylum Pending.” This item also included a space for remarks from the FOJC explaining the results. 28 d. Item 4 was headed “Current ORR Facility Type” and asked whether the age-out was coming from a shelter, foster care, or a staff secure or secure facility. e. Item 5 asked whether the age-out had certain types of underlying charges. f. Item 6 asked whether any Significant Incident Reports 9 were provided by ORR and included boxes to be checked for Yes, No, and Not Applicable. g. Item 7 asked whether behavioral information was provided, and provided boxes for Yes, No, and Not Applicable, along with space for the FOJC to “briefly describe” the behavioral information. h. Item 8 asked first “Is this age-out a flight risk, danger to the community, or danger to themselves?” and provided boxes for Yes or No responses. Below that question, Item 8 read “NOTE: If Yes, check and explain the reason(s) why in the remarks.” Below this were five possible reasons: Terrorist Affiliation or Ties; Criminal/Delinquency History; Gang Affiliation; Escapes/Non-compliance History; and Other Credible Threat. Each of these was followed by Yes and No check boxes, and there was space for remarks below. i. Item 9 asked whether the age-out had an attorney and provided space to record the attorney’s information. j. Item 10 asked whether a post-18 plan has been provided and provides space for the names and relationships of any potential sponsors. k. Item 11 asked whether certain types of post-18 plan documentation was provided. 9 Item 6 reads “SIR(s) Provided?” SIR is an acronym for Significant Incident Report and these are provided by ORR. The fact that a UAC has an SIR does not necessarily mean the UAC did something wrong. Trial Tr. at 563:19–25. 29 l. Item 12 asked whether “any individual or organizational sponsors or supervised group homes [are] available to this age-out for placement” and then provides space for remarks with the instruction “if No, briefly explain in the Remarks which of these options were evaluated and why they were not used.” m. Item 13 asks for the Risk Classification Assessment (“RCA”) Recommendation for the age-out. 10 n. Item 14 is headed “Discretionary Release Action(s).” It instructs “Check all of the options that were evaluated. If any options were not evaluated, or were evaluated and not used, briefly explain why in the Remarks.” It asks about four types of discretionary release actions: OSUP, OREC, see supra ¶ 43, Bond, and ATD. There is also space for remarks. o. Item 15 reads “Detention Decision, Considering Least Restrictive Setting Available” and has boxes to be checked for Detain or Not Detain. It also asks for a custody classification, if the age-out will be detained. The response to item 15 reflects the FOJC’s custody recommendation. Trial Tr. at 518:12–21. p. Item 16 provides space for the FOJC’s supervising officer (SDDO) to indicate approval of the FOJC’s recommendation. See also Trial Tr. at 518:18–21. It asks whether the SDDO approves or disagrees with the recommendation, and provides space for remarks along with instructions to “Briefly justify the recommendation in the Remarks, to include alternatives if supervisor disagrees.” 10 The RCA tool is a program used by field officers that makes recommendations as to whether to release or detain an individual and, if the recommendation is to detain, as to what appropriate level of ICE custody detention (e.g. low, medium, high) would be for that individual. See Ravenell Dep. Desig. at 84:3–23; Harper 12/6/18 Dep. Desig. at 168:3–169:4; see also infra ¶¶ 125–130. 30 57. At an April 2019 JFRMU training in San Antonio, FOJCs were trained on how to fill out an April 2019 version of the AORW that was different from the October 2018 AORW that had been in use for almost six months. Trial Tr. 416:20–419:2 (Harper); see also PX-197 (completed April 2019 version of the AORW); PX-203 at 4–5 (illegible April 2019 version of the AORW appearing in a PowerPoint presentation from the April 2019 training); infra ¶¶ 83– 100 (discussing the April 2019 training). 58. In August 2019 JFRMU/ERO issued another guidance to field officers announcing the implementation of another updated version of the AORW to be used effective September 1. DX-6; see Trial Tr. at 242:3–9, 253:12–19 (Harper). As in the October 2018 guidance, ERO/JFRMU emphasized that the system was “critical to ensure that ICE’s compliance with the law is properly captured,” and stated that it was “imperative” that materials be documented and uploaded “within 24 hours of the age-out’s transfer or release from ORR custody.” DX-6 at 2 (emphasis in original); see also Trial Tr. at 256:16–258:6 (Harper) (comparing similar language in the October 2018 Guidance (PX-1) and the August 2019 Guidance (DX-6). 59. The April 2019 version of the AORW that had been used at the April 2019 JFRMU training in San Antonio was never used in the field. Trial Tr. at 419:24–420:3 (Harper); id. at 546:16–547:2 (Sanchez). 60. The new August 2019 version of the AORW was changed in a number of substantive ways. Compare PX-1 at 6–8, with DX-6 at 3–6. Among the changes were the following: a. Item 8 from the earlier AORW was revised and broken up into three items, numbered 8a, 8b, and 8c. b. Item 8a asks “Is this age-out a danger to themselves?” and provides Yes and No boxes. Below this there are additional options to which the FOJC can answer Yes 31 or No: Medical/Mental Health Issues, Suicidal Ideation, History of Self-Harm, and Other Reason. Each has its own check boxes. Below these, the FOJC is instructed to “Write here a detailed explanation for why this person is or is not a danger to himself or herself.” c. Item 8b asks “Is this age-out a danger to the community?” and provides Yes and No boxes. Again there are additional options to which the FOJC can answer Yes or No: Gang Affiliation; Criminal/Delinquency History; Terrorist Affiliation or Ties; ORR Reported Behavioral Issues; and Other Reason. Below these, the FOJC is instructed to “Write here a detailed explanation for why this person is or is not a danger to the community, providing details of any question answered ‘Yes’ above.” d. Item 8c asks “Is this age-out a flight risk?” Again there are additional options to which the FOJC can answer Yes or No: Family and Community Ties; Escapes/Non-Compliance History; Prior Immigration History; Lacks Fixed Address in the United States; Other Reason; and Final Order. Next to “Final Order” is a space for the FOJC to enter the date of the Final Order if one exists. Finally there are instructions to “Write here a detailed explanation for why this person is or is not a flight risk.” e. Item 9, asking whether the age-out has an attorney and for that attorney’s contact information, is now Item 9a. A new Item 9b asks “Does this person have an advocate who is not an attorney, such as a social worker or staff member at a shelter?” If the answer is Yes, the FOJC is instructed to provide a name and contact information. 32 f. Item 12 asks whether there is a potential sponsor or organizational sponsor (with the question phrased in the singular (“sponsor”) rather than the plural (“sponsors”) on the earlier form. The space for remarks explaining which options were evaluated and not used has been eliminated, and instead there is only space for the sponsor’s contact information. g. Items 13 through 16 are significantly changed and consolidated into an Item 13 and an Item 14 only. h. Item 13 asks for “FOJC Recommendation, Considering Least Restrictive Setting Available” and instructs the FOJC to choose either Detained or Non-Detained. It then asks “Were all options/settings considered when making this recommendation?” (Yes or No) and “Are any alternatives to detention available for this age out?” (Yes or No). There is space for the FOJC to “detail alternatives” and a separate space for the FOJC to “provide supporting factors for the FOJC recommendation.” i. Item 13 continues instructing the FOJC to choose one of three possible “RCA Custody Recommendation[s]” (High, Medium, or Low) if the recommendation is to detain. Alternatively, if the recommendation is not to detain, the FOJC is instructed to choose one of the following: OREC, OSUP, OREC/ATD, OSUP/ATD, Parole, or Bond. There is then a space for the FOJC’s signature and the date. j. Item 14 has the heading “ICE Supervisor Custody Determination.” It asks whether the supervisor “Concur[s] with FOJC Recommendation” and allows for a 33 Yes or No answer. Space is provided for comments, and for the supervisor’s name and signature and the date. 61. The revised AORW did not change the fact that the AORW and SharePoint system was only designed to document the decisionmaking process, not to change it. The revisions did not change any decisionmaking procedures for FOJCs, but only to help document the processes they were already carrying out. Trial Tr. at 258:7–259:5 (Chief Harper testimony indicating there was no “new or additional guidance” in the new documentation system); DX-6 at 1 (“These procedures help ICE document the work its officers do every day . . . .”); see also PX-53 at 12 (training document stating that “FOJCs should use the Age-Out Review Worksheet to assist in documenting the Age-Out custody determination process”); Trial Tr. at 179:9–19 (Chief Harper reviewing PX-53 at 12). a. AORW Reliability Issues 62. A significant portion of the AORWs produced in this litigation were not completed by the FOJC who made the custody recommendation at or around the time that the age-out was processed, as the JFRMU/ERO guidance suggests is proper procedure. See supra ¶¶ 54, 58. In their April 2019 Motion to De-Certify the Certified Class for Lack of Numerosity and Commonality (“Mot. to De-Certify”), Defendants had told the Court that “since October 17, 2018, ICE [FOJCs] routinely document their custody determinations . . . [and] have been required to upload, for each and every age-out, an [AORW].” Mot. to De-Certify at 2, ECF No. 140. The Court was informed that “[individual] mistakes are, as they should be, infrequent outliers.” Id. Chief Harper filed a declaration stating that “[AORWs] are now being completed on a nationwide basis” and that “ICE ERO [was] complying with the JFRMU October 17, 2018 guidance.” Decl. of Mellissa B. Harper (“Harper 3/28/19 Decl.”) ¶¶ 25–26, ECF No. 140-3. These representations were incorrect in several respects. 34 63. A number of AORWs were created or updated shortly before depositions were taken in connection with this litigation. Officer Ana Sanchez, a National Juvenile Coordinator (“NJC”) in JFRMU, testified that in the days before Chief Harper’s December 6, 2018 deposition she and another NJC, Dawn Helland, were tasked with identifying age-outs processed since October 18, 2018 for whom no AORW had been uploaded to the SharePoint database. Trial Tr. 631:8–633:17; see also PX-267; Trial Tr. at 351:4–352:2 (Chief Harper’s testimony). Officer Sanchez then contacted field officers and asked them to complete and upload the missing AORWs. Trial Tr. 635:8–636:8, 638:3–8. E-mails show Officer Sanchez telling Chief Harper, “[w]e reviewed all the cases. 17 cases are missing the worksheet. . . . We sent emails to the FOJCs and the SDDOs,” PX-286 at 2, and later reporting that “All the missing [AORWs] were completed and uploaded to SharePoint,” id. at 1. 64. The day after her deposition, Chief Harper tasked Officer Sanchez with working with Chicago FOJC Geoffrey Pepple to “review” AORWs from the Chicago field office, to “update[]” the worksheets, and to “get more detail on the forms that need it.” PX-290 at 1; see also Trial Tr. at 649:24–651:24. Officer Sanchez then e-mailed Chicago FOJC Shawn Holmes, asking him to “review” certain AORWs and to “make sure that the reason for detention or release [was] stated” on them. PX-275; Trial Tr. 652:17–659:12 (Sanchez Testimony regarding PX-275 and PX-276); see also PX-276 (similar email to Officer Holmes). Officer Sanchez next e-mailed SDDO Linda Hyde, in the New York City field office, requesting similar review of and revisions to AORWs. Trial Tr. 660:7–25, 662:21–663:14; PX-282; PX-284; see also PX-285 (subsequent email from Sanchez to Hyde asking if Hyde had any additional changes for the worksheets). She also reached out to the Phoenix, El Paso, and Los Angeles offices regarding revisions to their AORWs. See PX-335.09849–50 (Phoenix); PX-289 (El Paso); PX-501 (Los Angeles); see also 35 Trial Tr. 3717:21–3719:17 (Los Angeles FOJC Oscar Irizarry identifying PX-501 as a thread of e-mails in which, at Officer Sanchez’s request, he altered responses on an AORW six weeks after the age-out was detained). 65. After Officer Sanchez secured the revisions she requested, Defendants then produced the revised or altered versions of the AORWs to the Plaintiffs in this case. Compare, e.g., PX– 335.09849–.09850, with PX–335.09835–.09837; PX–335.18899–.18901; PX–335.19082– .19084; PX–335.19197–.19199 (e-mail from Officer Sanchez to Phoenix FOJC Zanello requesting changes and produced AORWs reflecting those changes). 66. In addition, more than one field officer testified that his or her name was on AORWs that he or she did not fill out, or testified to filling out or adjusting AORWs with others’ names on them. See Trial Tr. at 3029:5–3034:3 (Mason) (SDDO testifying that he believed someone else may have put his name on an AORW that he did not fill out because “in a rush to finish these [AORWs], he put my name on every single one of them,” id. at 3033:23–25); id. at 3855:12–23 (FOJC Nancy Zanello of Phoenix confirming that she “made . . . changes to [a] worksheet according to Ms. Sanchez’s email . . . even though the decision had been made about a month earlier,” id. at 3855:12–15, by a different FOJC and that she did not believe she was involved in that decision); Hormiga Dep. Desig. at 291:8–292:12 (Houston FOJC Richard Hormiga identifying PX-335.22446–80 as an AORW filled out by someone else even though he had processed the age-out it covers). Officer Zanello testified that, to complete one worksheet, she copied and pasted information from another form, probably did not speak to the age-out, and did not necessarily even speak with the colleague who actually processed the age-out. Trial Tr. at 3848:22–3849:1, 3850:1–3 (Zanello). The revisions that Sanchez requested and Zanello 36 carried out included changes as substantive as change the answer for “flight risk” from “no” to “yes.” See id. at 3854:2–15 (Zanello). 67. Finally, in July 2019, ICE informed the Court that it had been failing to document a significant portion of age-outs for at least several years. See PX-351 (Aug. 29, 2019 Decl. of Mellissa B. Harper (“Harper 8/29/19 Decl.”), ECF No. 204-1); see also Mot. to Hold Defs.’ Mot. to Decertify the Class in Abeyance, ECF No. 192. 68. Between October 2018 and July 2019 ICE and ICE officials represented to the Court that all age-out custody determinations were being memorialized in AORWs that were being uploaded to the SharePoint database. See, e.g., Harper 3/28/19 Decl. (“Following JFRMU’s October 17, 2018 guidance, the Age-Out Review Worksheets are now being completed on a nationwide basis.”); Defs.’ Mot. To De-Certified the Certified Class at 1, ECF No. 140 (“ICE implemented a process that assures proper, contemporaneous documentation of its compliance with 8 U.S.C. § 1232 (c)(2)(B). Under this procedure, ICE comprehensively documents its process . . . . The implementation of this nationwide process forecloses the introduction of new members into the class, absent some sort of discrete error or mistake by an individual ICE officer.” (footnote omitted)). 69. Between April 1, 2016 and March 31, 2019, ICE had failed to make records for 1,473 age-outs, including 228 age-outs processed between October 17, 2018 and March 31, 2019 while the AORW/SharePoint system was in place. PX-351 ¶ 2. Additionally, 88 individuals had aged out between March 31, 2019 and May 24, 2019 without being recorded in the SharePoint site. Id. This was a significant upward revision. See Trial Tr. at 347:17–18; 348:3–4 (Chief Harper characterizing the discrepancy as “significant”). To take as an example one date range Chief Harper addressed in two declarations, between October 17, 2018 and January 30, 2019, the Court 37 had previously been told that 581 individuals aged out, and that 154 of these had been detained. PX-351 ¶ 6 (citing Harper 3/28/19 Decl. ¶ 27). In fact, 742 age-outs had been processed and 263 of these had been detained—a 27.7% increase in the number of age-outs, and an increase in detention rate from 26.5% to 35.4%. Id. C. Policy, Guidance, and Procedures at the National Level 70. In many, but not all, field offices, deportation officers rotate in and out of the position of FOJC with some frequency—every year or two. See Trial Tr. at 149:24–150:4 (Harper) (stating that “[g]enerally speaking it’s every two years” but that “it can be” every year in some circumstances); see also, e.g., Trial Tr. at 3032:4–8 (Mason), 3623:6–11 (Pepple); Hormiga Dep. Desig. at 26:10–29:12 (describing recent turnover of the entire Houston FOJC team); Venegas Dep. Desig. at 5:25–6:11 (“We rotate[d] the units on a yearly basis”). In some other offices, officers can remain FOJCs for long periods of time. See, e.g., Pepple Dep. Desig. at 8:5–6; Trial Tr. at 3667:22–25 (Irizarry). 71. FOJCs are trained on the job by their predecessors as FOJCs, are trained at the annual JFRMU training, and have access to the FOJC handbook and other online materials from JFRMU. Harper 12/6/18 Dep. Desig. at 220:2–221:2; see also, e.g., Trial Tr. at 2562:22–2563:3 (Black); 2673:19–2674:17 (Kaskanlian); Trial Tr. 2757:18–2758:1 (Munguia); Trial Tr. 3133:15–3134:1 (Mason). There is also no formal headquarters or other training specifically designed for onboarding new FOJCs. Harper 12/6/18 Dep. Desig. at 220:2–221:2; see Defs.’ Resp. ¶ 190 (not disputing that there is no formal or other training specifically designed for onboarding new FOJCs). 38 1. Training and Guidance on Age-Out Custody Determinations a. Written Guidance 72. Between the 2013 amendment of the TVPRA that included Section 1232(c)(2)(B) and March 2018, ICE headquarters issued guidance to the field regarding Section 1232(c)(2)(B) three times, in May 2013, January 2016, and November 2016, 11 and that ICE also addressed the statute in a 2017 FOJC handbook. PX-338 at 2–5 (3/27/2018 Harper Decl.); see also Defs.’ Resp. ¶¶ 160–65 (identifying no additional guidance from this time period and disputing only Plaintiffs’ characterization of this guidance as improper). 73. The May 2013 guidance says that the language of the amended TVPRA/VAWA “comports with existing ICE practices in that custody determinations are made after considering the circumstances of each individual case” and that age-outs “may also be eligible to participate in [ATD] programs based on the alien’s need for supervision.” PX-338 at 7; see also Ravenell Dep. Desig. at 99:21–100:16 (identifying the document). It also explains that ICE will begin tracking age-out custody determinations by means of the monthly age-out reports described above. PX-338 at 7. It does not quote the statute or reference the statutory factors: danger to self, danger to community, and risk of flight. 74. The January 2016 guidance is a multi-page outline of guidance for apprehending and processing UACs, most of which was devoted to the initial apprehension of UACs and transfer of UACs to ORR custody. Id. at 11–15. The final page addresses age-outs. It paraphrases the statutory language, including mention of the requirement to “consider placing [age-outs] in the 11 Chief Harper’s March 2018 declaration gave the wrong date for this third guidance document. PX-338 at 3. A version produced by ICE in the context of an email thread clarifies the proper date. PX-133. 39 least restrictive setting available, taking into account the UAC’s danger to self, danger to community, and flight risk.” Id. at 15 . It also referenced DHS Enforcement Priorities. Id. 75 . The November 2016 guidance states that the statute “specifically requires ICE to consider placement in the least restrictive setting based on the circumstances of each individual case.” PX-338 at 9. It further states: “A default to detention is never acceptable in these cases; instead, all custody determinations must include an assessment of the individual’s danger to self, the community, and risk of flight, as well as recommendations submitted by ORR in ‘Post-18 Plans.’ Following this assessment, if detention is not justified, release through one of ICE’s alternatives to detention mechanisms should be considered.” Id. 76 . The FOJC Handbook, PX-338 at 24–94, references Section 1232(c)(2)(B) at three points. See Defs.’ Resp. ¶¶ 164–65 (not challenging this). Page 18 of the Handbook says that age-outs “shall be considered for placement in the least restrictive setting” and that “[s]uch placements should be considered after evaluating the individual’s danger to self and the community, and risk of flight.” PX-338 at 44. Pages 43 to 44 of the Handbook discuss the statute again, stating that it “requires ICE consider least restrictive settings when making custody determinations” for age-outs, that “[d]eterminations should be made considering the circumstances of each individual case,” and that ATD programs are available for some age-outs. PX-338 at 71. The third reference comes in a discussion of post-18 plans. See id. at 71–72. After quoting the statute, the Handbook says that “[b]ased on the totality of all the information . . . a custody determination should be made prior to transfer back to ICE custody.” Id. at 72 . 40 77. An updated FOJC Handbook introduced in April 2018 did not make any substantive changes to any material referencing Section 1232(c)(2)(B). PX-346 at 24–25, 54–56; see Defs.’ Resp. ¶ 166 (not disputing this); see also DX-4 (2018 FOJC Handbook). 78. Other than materials used in conjunction with ICE’s April 2019 annual training, the only other written policy guidance from ICE concerning Section 1232(c)(2)(B) has been a “one pager job aid” sent by Chief Harper to all FOJCs on December 21, 2018. PX-135; see Ravenell Dep. Desig. at 153:24–154:8 (NJC Ravenell recalling no additional written guidance); Defs.’ Resp. ¶ 167 (identifying no additional written guidance beyond training materials). The “onepager” consists of a flowchart “Guidance on Age-Outs.” PX-135 at 2. It lays out that ORR should notify an FOJC when a UAC is approaching their 18th birthday and that “ORR provides FOJC with a Post-18 Plan 1-2 weeks prior” to that birthday. Id. Next, according to the flowchart, the “FOJCs assess [the age-out’s] danger to self, community, and flight risk,” and then “FOJCs make a Custody Determination based on the totality of the circumstances” and should consider “[p]ublic safety,” “flight risk,” “Final Order Status,” “Post-18 Plan,” “Communication from ORR,” and “External stakeholder input.” Id. The next box says that “[i]f appropriate, FOJCs consider alternatives to detention programs based on an [age-out’s] need for supervision,” which can include placement with a sponsor, OREC, OSUP, or a program through the ATD unit. Id. 79 . The same “one-pager” flowchart guidance that Chief Harper circulated on December 21, 2018 also appeared in training materials used at the national ICE training in April 2019. PX203 at 14. See Trial Tr. at 224:22–225:3; Defs.’ Resp. ¶ 216. 80. The “one-pager” is the only written guidance Chief Harper has issued concerning how age-out custody determinations are to be made. Any additional guidance she has given on 41 this topic has been during monthly phone calls, and any additional written guidance she has issued has concerned only documentation of age-out custody determinations. See Defs.’ Resp. ¶ 167 (identifying no additional written guidance and citing only training materials as providing instruction beyond the one-pager). b. Annual In-Person Training 81. Until the fall of 2019, the only in-person training that field officers received with respect to age-out processing and custody determinations was an annual training conducted by JFRMU. Trial Tr. at 535:24–537–11. Trainings were held in 2016 and 2018, but not 2017 due to funding constraints. Harper 12/6/18 Dep. Desig. at 44:17–22. 82. In each of the 2016 and 2018 trainings, JFRMU presented only one slide addressing offering instruction on age-out custody determinations. See PX-338 at 4–5, 18, 21 (Harper Decl. and exhibits); PX-137 (2016 presentation excerpt); PX-139 (2018 presentation excerpt); see also Defs.’ Resp. ¶ 168 (failing to challenge this contention or to point to any other materials presented at the 2016 or 2018 training). Both of these years, FOJCs were told to “[s]eek [an] ORR Post-18 Plan,” to run the RCA tool, to make a “case-specific custody determination,” and to “[c]onsider VAWA.” Id. at 18, 21 . In 2016, though not in 2018, the slide instructed FOJCs to “[c]onsider least restrictive setting,” but in neither year did it mention the statutory factors mentioned in Section 1232(c)(2)(B)—danger to self, danger to the community, and risk of flight. Id. 83 . JFRMU’s spring 2019 training—the first training after implementation of the AORW and SharePoint system—addressed age-out processing during two of approximately thirty training sessions. One of these sessions was titled “Current litigation: Garcia Ramirez v. ICE” and the other focused on “Documentation Activity” for age-out custody determinations. Trial Tr. at 539:11–14 (Sanchez), 541:6–542:22 (Sanchez); PX-198 (training schedule); see also 42 Defs.’ Resp. ¶ 169 (failing to identify any additional coverage of age-out processing during the 2019 training). 84. The session reviewing this litigation, Garcia Ramirez v. ICE, was led by Officer Sanchez. PX-203 (PowerPoint presentation); Trial Tr. at 542:23–25 (Sanchez). The goals of the session included “[d]iscuss[ing] the custody determination process” and “[d]escrib[ing] the importance of documenting all the articulable facts related to the custody decision.” PX-203 at 3; Trial Tr. at 543:11–18. 85. Part of the Garcia Ramirez v. ICE presentation involved Officer Sanchez describing the AORW “as an example of the available job aids” and emphasizing “the importance of documenting all the articulable facts related to the custody decision.” PX-203 at 4–5. The version of the AORW that Officer Sanchez displayed and discussed was the interim April 2019 version that was never implemented for use by officers in the field. Trial Tr. at 546:16–547:2 (Sanchez); supra ¶¶ 57, 59. 86. Officer Sanchez next displayed a slide addressing Section 1232(c)(2)(B) and summarized it as requiring that “[a]ny minor transferred from ORR to ICE custody at the age of 18 shall be considered for placement in the least restrictive setting.” PX-203 at 6. 87. Another slide summarized the “Age-Out Process.” It explained that FOJCs making custody determinations “must consider the totality of the circumstances and all available information” including not only flight risk, public safety, and danger to self, but also “Post-18 Plan or lack thereof,” “Recommendation letter from sponsor, attorney . . . , advocates, and/or NGOs,” “Behavioral issues,” and “Significant incident reports.” Id. at 8. Officer Sanchez testified that although the phrase “totality of the circumstances” does not appear in the statute, it 43 was her understanding that FOJCs not considering “every piece of information available” would “not be doing their job.” PX-203 at 8; Trial Tr. at 548:16–24 (Sanchez). 88. The next slide showed a chart reflecting a sequence of steps for FOJCs to take when processing age outs: first, “[p]erform a Risk Classification Assessment (RCA)”; second, “[c]onsider VAWA in all Age-Out custody determinations”; third, “[m]ake a new custody recommendation to the [SDDO]”; and fourth, serve the necessary form (I-286) reflecting the custody determination. PX-203 at 10; Trial Tr. at 548:25–549:22 (Sanchez). At the second step—considering VAWA—there are additional notes on the chart: “Detention as a default is NOT acceptable” and “Consider the least restrictive setting, ORR facility type, and other available resources.” PX-203 at 10. 89. The next slide emphasized that “Good documentation is key,” and listed forms and documents, including the AORW, that could help memorialize decisionmaking. PX-203 at 12. Officer Sanchez’s presenter notes reflect that she told the FOJCs that a detailed entry in ICE’s ENFORCE Alien Removal Module (“EARM”) 12 was “paramount” and that “[a] comment such as [‘]aged out transferred to ICE adult detention[’] is not acceptable” because it fails to “provide the reason why the detain decision was made.” Id. 90. The next slide displayed the flowchart guidance document that Chief Harper had circulated in December 2018. PX-203 at 14; PX-135; see supra ¶¶ 78–132 (explaining that the flowchart indicates that FOJCs should first make a custody determination and then, “[i]f appropriate,” consider alternatives to detention). 91. The remainder of the presentation—aside from some concluding and review slides— involved review of three hypothetical age-out scenarios. PX-203 at 15–33. In the first, the 12 EARM is a case management system used by ICE. Trial Tr. at 599:3–7. 44 hypothetical age-out was considered a flight risk because he lacked family or community ties in the United States, but could properly be released to a program in the local community identified by ORR in a post-18 plan because the program had a “good reputation” for “providing young adults with housing, schooling, and medical and legal assistance.” See id. at 15–18. 92. In the second hypothetical, the age-out was a young mother who had been placed in ORR custody along with her one-year-old child. PX-203 at 21. ORR had learned that the only potential sponsor they had been able to identify was a suspected smuggler, and had explained this to ICE but had not provided a post-18 plan. Id. ORR then requested that ICE take custody of the aging-out mother while leaving the one-year-old in ORR custody. Id. The presentation suggests that the correct analysis would be to regard the age-out as a flight risk and a danger to herself because the smuggler/sponsor posed a safety concern. Id. at 23. The fact that the age-out had an infant child would mean that ORR would not release the infant to ICE if the mother was not going to be released from ICE custody. Id. The presentation concluded that “ICE detention was the least restrictive / safer setting for this Age-Out because at the time of custody determination she has no viable sponsor except the smuggler.” Id. at 25–26. 93. Chief Harper testified that this second hypothetical was based on a real case from ICE’s New York office but that she was not aware of any cases presenting similar fact patterns. Trial Tr. at 186:3–25 (Harper). She could not identify another case in which ORR suggested that a mother be separated from a one-year-old child. Id. at 187:10–13 (Harper). She further testified that she believed the presentation’s conclusion was the correct one and that an officer faced with this situation was not obligated to consider or try to identify any alternatives other than what ORR recommends. Id. at 187:23–190:11. 45 94. In the third hypothetical, ORR had given ICE a post-18 plan stating that an age-out with no family or community ties in the United States had been diagnosed with PTSD and had expressed suicidal ideations. PX-203 at 28. ORR was recommending release to an emergency shelter. Id. The presentation suggested that the age-out was a flight risk because of his lack of ties and that he posed a risk to himself. Id. at 30. The presentation concluded that “ICE detention was the most appropriate setting” because “ICE needs to consider the danger to self” and “ICE has the responsibility to place this subject where his medical needs can be met.” Id. at 32. 95. Chief Harper testified that under this third scenario, an ICE officer would not be under any obligation to try to identify alternative options for placing this age-out, and that she agreed detention was the right outcome in this scenario. Trial Tr. at 192:13–193:4 (Harper). 96. The documentation activity session was also led by Officer Sanchez, along with Officer Ravenell. PX-199 at 2. The second half of this presentation involved having attendees review two completed AORWs. See id. at 17–27; see also PX-197 (training worksheet); PX-198 (same). 97. “Key takeaways” identified in the presentation included: k. “[T]here is no right/wrong answer, but you should properly document your decision[.]” PX-199 at 22. l. “Ensure EARM case comments match the [AORW]; Worksheet should also match SharePoint[.]” Id. m. “The age-out was in ORR custody for 86 days, and ORR was not able to identify an individual sponsor, a Post 18 program, or a supervised group home for [the age-out.]” Id. at 23. 46 n. “He has no equity, family, or community ties in the U.S., thus posing a significant flight risk [.]” Id. 98. The AORW used in the documentation activity session was the interim April 2019 version that was never used in the field. PX-199 at 21; Trial Tr. at 559:19–21 (Sanchez). 99. Regarding Section 1232(c)(2)(B), the presentation explained that “ICE considered VAWA in the custody determination,” but that “ICE was not able to consider any release options because the age-out does not currently have a viable sponsor.” Id. at 24. 100. The presentation listed the RCA recommendation as among the “Documents to be Completed” as part of the custody determination, id. at 25, and identified failure to process an RCA recommendation as a deficiency because “RCA should be used as an aid to complete the custody redetermination process,” id. at 27. c. Additional Instruction 101. In 2018, JFRMU instituted an “e-learning module,” covering UAC processing . This was an online training program lasting around thirty minutes, based on the FOJCs handbook and covering the role of the FOJC and relevant legal authorities. PX-140; see Ravenell Dep. Desig. at 166:18–168:24. 102. One slide on the UAC processing e-learning module mentions VAWA. PX-140 at 56–57. 13 It explains that the statute “states that [age-outs] are eligible to participate in programs such as Alternatives to Detention (ATD), using a continuum of alternatives based on the need for supervision.” Id. at 57. It does not include the language of the statute, does not mention the requirement to consider the least restrictive setting, and does not refer to the statutory factors of danger to self, danger to community, and risk of flight. Id. 13 The same slide also mentions provisions of the TVPRA other than Section 1232(c)(2)(B). 47 103. Chief Harper also holds monthly phone calls with FOJCs in which she “cover[s] the most current topics, anything that comes to [JFRMU’s] attention, [and] answer[s] any question the field may have.” Trial Tr. at 151:3–12 (Harper). These calls last thirty to sixty minutes and cover all of JFRMU’s areas of focus, not only age-outs. Id. at 151:3–25; Harper 12/6/18 Dep. Desig. at 51:11–14. No minutes are prepared and no other records of what is discussed on the monthly calls is kept. Harper 12/6/18 Dep. Desig. at 53:17–20. Plaintiffs do not dispute that “JFRMU holds monthly phone calls with FOJCs to discuss issues and disseminate information,” but they dispute whether JFRMU’s monthly calls constitute or provide training. Pls.’ Resp. FF ¶ 88. Whether the calls can be said to constitute or provide training of one form or another is immaterial, as “training” and “guidance” are not terms that need to be defined in order for this case to be resolved. To the Court, it seems that the calls are best described as a form of guidance or, at most, informal training, considering that no materials are prepared for distribution or to memorialize the topics covered. This is in contrast with the formal training held each year and with the guidance materials distributed from time to time by JFRMU/ERO. It is appropriate to describe the phone calls as informal because, whatever training they contained, it was less rigorous, less memorable, and more difficult for FOJCs to refer back to as compared with the other forms of training which produced written records. This is underscored by the fact that no witness, including Chief Harper, ever pointed to a specific monthly phone call at which a particular issue was discussed. To the extent the Court finds, below, that FOJCs did not receive formal training on a given topic, it does not mean to suggest that this topic was never mentioned on a monthly call. A finding that there was no formal training on a given topic means that that topic was not addressed 48 in annual training or guidance documents and thus that it was only addressed, if at all, in the monthly phone calls. 104. In the fall of 2019, Chief Harper and Officer Helland visited the Miami field office for additional in-person training by JFRMU for all field officers—not just FOJCs—covering all UAC-related issues and practices. Trial Tr. at 4324:23–4325:13 (Harper). 105. Among other reasons for visiting the Miami office, Chief Harper and her staff were concerned that FOJCs in Miami were automatically detaining age-outs without making a VAWA-compliant decision and, instead, leaving it up to ICE’s detention unit to seek out alternatives to detention and to consider less restrictive settings. Trial Tr. at 4327:6–4328:12 (Harper); id. at 4074:23–4075:3 (Helland). 106. The training given during the Miami visit was “not actually geared towards the FOJCs” but JFRMU officials met with FOJCs subsequent to the formal training. Id. at 4112:23– 4413:18 (Helland). This included explaining to the FOJCs that there was no legal basis for the Miami FOJCs’ practice of rejecting proposed sponsors for age-outs solely because they were not lawful permanent residents or United States citizens. Id. at 4151:7–4152:20 (Helland); see also id. at 3925:10–24, 3926:9–17 (Reardon). 107. JFRMU officials conveyed to officers in the Miami office that a decision not to release an age-out to a particular sponsor needed to be based on danger or flight risk, not simply on the fact that the sponsor was not a U.S. citizen or lawful permanent resident. Id. at 4151:7– 4152:20 (Helland). 108. Additional field office visits were planned but, as of the conclusion of trial in January 2020, none had taken place due to lack of funding. Trial Tr. at 4330:21–4331:2 (Harper). 49 2. Substance of JFRMU Training a. Totality of the Circumstances 109. In training materials, ICE emphasizes to FOJCs that they should apply a “totality of the circumstances” approach when making age-out custody determinations. PX-203 at 8 (April 2019 training presentation instructing that “[w]hen making an Age-Out custody determination, the FOJC must consider the totality of the circumstances”), id. at 14 (“one-pager” flowchart instructing that “FOJCS make a custody determination based on the totality of the circumstances”); PX-203 at 2 (same); Trial Tr. at 64:14–17 (Harper); Defs.’ Resp. ¶ 95 (“Defendants admit that ICE JFRMU trains officers to make custody determinations based on the ‘totality of the circumstances.’”); see also Harper 5/2/19 Dep. Desig. at 9:21–10:13 (agreeing that totality of the circumstances is the standard by which “all custody decisions are made regardless of who . . . the determination is being made for”); Ravenell Dep. Desig. at 142:24– 143:24 (agreeing “[a] hundred percent” with Chief Harper’s testimony); Trial Tr. at 61:12–25, 62:20–63:1 (Harper). These materials stress the “totality of the circumstances” over and above the three statutory factors identified in Section 1232(c)(2)(B). 110. In Officer Sanchez’s presentation on this litigation from the April 2019 JFRMU training bold text reading “FOJC must consider the totality of the circumstances” appears in a paragraph above a box listing bullet points identifying types of information that might matter. These are: “Post-18 Plan or lack thereof”; “Recommendation letter from sponsor, attorney of record, advocates and/or NGOs”; “Flight risk”; “Public safety”; “Behavioral issues during custody with ORR and any other agencies”; “Significant incident reports”; “Danger to self”. PX-203 at 8. The three statutory factors all appear on this sheet, but they are not emphasized over other sources of information that are not in the statute, and all of these are presented under the umbrella of “totality of the circumstances.” Id. The preceding slide of the same presentation 50 identifies the “Relevant Legal Authority” as “VAWA Section 1232(c)(2)(B)” and summarizes it thus: “Any minor transferred from ORR to ICE custody at the age of 18 shall be considered for placement in the least restrictive setting”—no mention is made of the three statutory factors. Id. at 6. The statutory factors are treated separately at one point in Officer Sanchez’s presenter’s notes, see id. at 9, though even there they are not connected to the statute. And nowhere in the slides that were displayed to FOJCs are the three statutory criteria singled out or highlighted. An FOJC reviewing the presentation’s instruction to “Consider VAWA in all Age-Out custody determinations,” id. at 10, would not be able to tell from the presentation that VAWA also obligates her to “tak[e] into account the alien’s danger to self, danger to the community, and risk of flight,” 8 U.S.C. § 1232 (c)(2)(B). 111. The “one-pager” job aid that was circulated by Chief Harper in 2018 and which appears again in this same presentation, downplays the statute even further. Id. at 14 ; PX-135 at 2. It instructs that the custody determination is “based on the totality of the circumstances” and then instructs FOJCs making custody determinations to consider the following: “Public safety and flight risk factors”; “Final Order Status”; “Post-18 Plan”; Communication from ORR, advocates and others; and “External stakeholder input.” PX-203 at 14; PX-135 at 2. Danger to self is not mentioned and, again, statutory factors are listed alongside other inputs and under the heading of “totality of the circumstances.” Id. The FOJC Handbook does list the three statutory factors at two points but does not elaborate on them or provide any instruction on how to apply them. See PX-338 at 44, 71–72; see also supra ¶¶ 76–77 (discussing the FOJC handbook in more detail). The handbook also emphasizes that custody determinations should be “[b]ased on the totality of all the information.” PX-338 at 72. 51 112. ICE does not provide and Chief Harper has not seen any guidance or standards defining “totality of the circumstances.” Trial Tr. at 70:12–71:18 (Harper); Ravenell Dep. Desig. at 144:1–4. Chief Harper explained her understanding that considering the “totality of the circumstances” means “that we consider every possible factor that we’re aware of at the time,” Trial Tr. at 62:4–6 (Harper), and that “totality of the circumstances” is “not a standard, it’s a phrase meaning you consider all the information that you have at the time to make a decision based on the factors that apply at that given time,” id. at 69:14–17. See also id. at 578:9–14 (Sanchez) (“[F]or the most part, if we’re looking at the totality of the circumstances, we are looking at everything that is there . . . .”). b. Flight Risk and Sponsors 113. JFRMU and ICE train FOJCs to consider an age-out’s flight risk based on factors that are present for nearly all age-outs. The training materials used at the April 2019 JFRMU training provide, as examples of reasons an age-out could be a flight risk, “lack of ties, no viable sponsor, no-Post 18 Plan, and long time in ORR custody.” PX-203 at 9; see also id. at 17 (providing “the subject has no family or community ties” as an example of a flight risk); id. at 23 (“the subject does not have equity, family, and/or community ties”); PX-54 at 23 (“He has no equity, family, or community ties in the U.S., thus posing a significant flight risk . . . .”); id. at 26 (“The subject does not have family or community ties in the U.S., thus posing a significant flight risk[.]”). Most age-outs lack community ties in the United States. See Galvez Dep. Desig. 225:11–17; Pepple Dep. Desig. at 175:13–176:5 (“The majority of our cases are what we call recent entrants so they don’t have a lot of community ties here, and as recent entrants they would be considered a flight risk.”); Black Dep. Des. at 24:10–17 (agreeing that her unit, and ICE generally, consider most age-outs to be flight risk.); PX–335.09849–.09850 (email from Sanchez to Zanello 52 requesting edits to AORWs to indicate for each of three age-outs that “the subject is a flight risk as he does not have a viable sponsor”). 114. JFRMU also instructs FOJCs to view a long period of time in ORR custody as indicating that an age-out is a flight risk. See, e.g., Trial Tr. (Harper) at 303:12–304:13 (discussing PX-55 at 1, an email from an FOJC referencing a February 2018 monthly FOJC call on which Chief Harper expressed that “if UACs had been in ORR custody for an extended period of time . . . they may be considered a flight risk”); id. at 390:10–19 (Harper) (testifying, while reviewing training materials, that “155 days a is a pretty lengthy stay to be in ORR custody” and that this would be significant because it showed ORR “hadn’t found a sponsor for that person”); PX-155 at 4 (email from Chief Harper stating “One big factor [regarding detention] is the amount of time that ORR has a minor and hasn’t been able to find anyone to release them to. That in and of itself indicates a lack of ties and flight risk.”). This guidance is, in certain contexts, wrong. First, it does not take into account that certain organizational sponsors have age limits and that some do not accept children. See, e.g., Barnes Dep. Desig. at 51:24–52:9 (discussing Casa Juan Diego, a Houston-area facility that only accepts adults); see also Trial Tr. at 619:19–25 (Sanchez) (reviewing a letter from an organizational sponsor that only accepts “young men ages 17 to 21”). Second, the opposite is true for children who, while in ORR custody, had been placed in foster care and were thus free to attend school and other activities. See, e.g., PX-335.32027 (post-18 plan for a UAC in foster care indicating that the age-out “continues to show effort in school, with no prompting obtains tutoring . . . is respectful in the home, and participates [in] extracurricular activities”). If children with this level of freedom have not taken advantage of it and attempted to abscond, their length of time in ORR custody suggests a lack of flight risk. See Trial Tr. at 2617:12–25 (Black) 53 (explaining that she viewed an age-out’s long period of time in foster care as indicating she was not a flight risk); id. at 2730:10–2731:1 (indicating that it was a “very straightforward” decision to release on OREC an age-out who “had been living successfully in the community in a foster care program”). 115. ICE instructs its FOJCs that a risk of flight can be mitigated, including by release to sponsors. In one age-out scenario reviewed at the April 2019 JFRMU training, a hypothetical age-out was said to be a flight risk because of a lack of family ties, but could be released to a “Post-18 program” because “[t]he flight risk concern will be mitigated by the fact that the program has an excellent record of compliance for previous cases.” PX-203 at 17; see id. at 15 (describing the program as having “a good reputation of providing young adults with housing, schooling, and medical and legal assistance”). 116. An age-out’s risk of flight could also be mitigated by a pending application or court case that could yield immigration benefits, like asylum, Special Immigrant Juvenile Status, or family-based immigration visas. Andrea Saenz, a legal services provider who works with people who are in immigration detention or facing deportation in the New York area, testified that in her experience with immigrant clients, “somebody who wants to come back to court, who has defenses to removal, who has counsel, who has an incentive to come back is going to come back, so they’re not going to run away.” Trial Tr. at 1057:13–21; id. at 1029:18–25; see also Tafur Dep. Desig. at 50:14–20 (San Antonio–area legal services provider describing a case in which, after an age-out client was released, “she was very great at communicating . . . and making all diligent efforts to comply with her requirement to [go] to court.”). Officer Black agreed that a strong case for immigration benefits would disincentivize an age-out to abscond. See Trial Tr. at 54 2619:21–2621:3 (Black) (“I think most people would assume that would be good reason to stick around and see their case through . . . .” Id. at 2620:7–8.). 14 117. Following the JFRMU visit to the Miami field office, Officer Helland sent an email to Miami SDDO Lee-Fatt—with Chief Harper’s approval—saying, among other things, that “ERO 14 Defendants ostensibly dispute that pending immigration proceedings can disincentivize flight, but point to no evidence in doing so other than the testimony of Plaintiffs’ expert retired immigration Judge Robert Vinikoor. Defs.’ Resp. ¶ 111. Judge Vinikoor did testify, when asked, that a flight risk remains unmitigated when no sponsor is available, see Trial Tr. at 1872:22–25, but he also said that “you consider other factors,” id. at 1872:18, and his statement about unmitigated flight risk came in response to Defendants’ counsel’s instruction to disregard those other factors, id. at 1872:19–20. When asked a clearer question by Plaintiffs’ counsel, Judge Vinikoor agreed that “eligibility for relief to remain in the United States is a very strong motivational factor that oftentimes shows or compels [age-outs] to return to court.” Id. at 1800:7–10. Defendants challenge Judge Vinikoor’s testimony on a variety of grounds, including that he lacks relevant expertise because he never encountered Section 1232(c)(2)(B) and may never have held a bond redetermination hearing for an adult who arrived in the United States as a UAC. See Defs.’ Resp. ¶ 112. The Court declines to parse the Defendants’ numerous critiques of Judge Vinikoor’s testimony because the essentially unrebutted testimony of Ms. Saenz, Ms. Tarfur, and Officer Black serves to establish well enough the fact that pending immigration proceedings might incentivize age-outs not to abscond. Defendants also accuse Judge Vinikoor of having given “misleading testimony” regarding the fees he charged for his services in this case. Defs.’ Resp. ¶ 267. They suggest that Judge Vinikoor “claimed that [his work on the case] was in part ‘pro bono’—or ‘for free’—while in reality he charged hundreds of dollars per hour” and that he “even felt the need to apologize to the Court for his misleading testimony on this point.” Id. The Court feels compelled to set the record straight. During his direct examination Judge Vinikoor said he was accepting “a $200 an hour pro bono fee” for his work on this case, in contrast to his usual fee of $450 an hour. Trial Tr. at 1778:23–25. On cross-examination, Defendants’ counsel confirmed this pro bono rate, and asked Judge Vinikoor if he knew “what pro bono means” in Latin. Trial Tr. at 1885:10. He responded “For free,” and counsel agreed and asked, “Is $200 an hour for free?” Id. at 1884:12– 13. Judge Vinikoor then said “I apologize.” Id. at 1884:14. Judge Vinikoor had nothing to apologize for, because he had never claimed to have been working “pro bono” independent of quoting his $200 an hour pro bono rate. He never suggested his services were provided for free. Further, both Judge Vinikoor and counsel mistranslate pro bono, which translates from Latin as “for good” or “for the public good” and is commonly understood to include work performed at reduced rates. See, e.g. Sally Kane, What is Pro Bono?, The Balance Careers (Jun. 27, 2020), https://www.thebalancecareers.com/what-does-pro-bono-mean-2164411; Carla Tardi, What does Pro Bono Really Mean?, Investopedia (Feb. 22, 2019), https://www.investopedia.com/ask/ answers/08/pro-bono.asp. It is Defendants’ argument, not Judge Vinikoor’s testimony, that is misleading. 55 may . . . evaluate whether release to [a] proposed undocumented sponsor sufficiently mitigates concerns of flight risk.” DX-108 at 2; see generally Trial Tr. at 4335:22–23 (Harper) (approval of the email). But, as discussed below, ICE does not encourage its officers to make any effort to identify, verify, or investigate sponsors for age-outs, infra ¶¶ 146–150, and officers tend not to do so, infra ¶¶ 157–163. 118. The ATD program is another way to mitigate risk of flight. Carbonneau Dep. Desig. at 23:12–53:2 (ICE 30(b)(6) witness on the ATD program agreeing “that the ATD program is ‘an important flight mitigation tool,’” as he had stated in a 2015 guidance document, PX-22 at 1). 119. An April 2018 email shows that on at least one phone call, in February 2018, Chief Harper told FOJCs that releasing an age-out on ATD was another way that flight risk could be mitigated. See PX-55 at 1 (“During the February FOJC teleconference, it was suggested that if UACs had been in ORR custody for an extended period they could be placed on ATD as they may be considered a flight risk.”); see also supra ¶ 114 (discussing the implication that a long period in ORR custody indicates a flight risk). However, the same email, from an FOJC, makes it clear that the February 2018 call was the first time the FOJC had heard that flight risk could be mitigated in this way. PX-55 at 1 (indicating that the age-out in question, who had been released to a sponsor “aged out before that [February] teleconference”). In fact, the testimony from a number of FOJCs across the country indicated that broadly speaking, they did not release age-outs on ATD, and many were unsure whether age-outs even qualified for the program. See infra ¶ 139 (finding that, of those who testified, five FOJCs said they were not trained on how to place age-outs on ATD, and that three more did not believe ageouts were eligible for the ATD program). 56 120. Across field offices, then, FOJCs view age-outs without verified individual sponsors as lacking community ties and therefore treat them as flight risks. E.g. Trial Tr. at 3394:7–12 (Hyde) (agreeing, in reference to an age-out, that “[t]he fact that he didn’t have a sponsor meant that he was a danger to himself and a significant flight risk”); Hormiga Dep. Des. 139:6–11 (agreeing that an age-out’s “flight risk . . . determination [was] just based on a lack of sponsorship”); id. at 220:5–12 (testifying that an age-out was a flight risk based on “no ties to the community . . . no ways to earn income, and no way to verify where they were going to stay”); Galvez Dep. Desig. 241:12–242:16 (“If they don’t have a sponsor or a place to go they would be a flight risk.”); id. at 243:16–244:15 (testifying that lack of a sponsor can be determinative of whether an age-out is regarded as a flight risk); see also Reardon Dep. Desig. at 139:10–16 (testifying that he viewed an age-out as a flight risk until the detained unit could vet a proposed sponsor); Trial Tr. at 567:7–16 (Sanchez) (“If you don’t have a sponsor, you don’t have an address. I don’t see how they could release you without you having a location to go to.”); id. at 851:4–14 (Enriquez) (third party testifying that his organization “provide[s] [sponsor] information to ICE because that’s a signal that the child has the strong community ties, which we know from experience is a signal that the child is not a flight risk . . . .”); PX-54 at 24 (age-out documentation exercise concluding an activity with the “key takeaway[]” that “ICE was not able to consider any release options because the age-out does not currently have a viable sponsor”). 121. An age-out’s lack of a verified sponsor can also cause that age-out to be perceived by ICE as a danger to himself or herself because in at least two field offices—New York and El Paso—FOJCs have interpreted danger to self to include the risk that an age-out will not have a place to live. Trial Tr. at 3328:3–11 (Hyde) (“[I]n my interpretation, if someone has no place to go and [is] placed out in the elements, it could be a danger to self, yes.”); Galvez Dep. Desig. at 57 300:19-301:3 (agreeing that “an age-out that doesn’t have a fixed address or a sponsor could be a danger to themselves because they might become homeless or otherwise suffer adverse consequences”). This argument is largely a red herring. Regardless of the fact that this logic does not appear to be supported by the plain language of the statute, it is not supported by the facts presented at trial. With respect to the danger of exposing age-outs to the elements, this runs counter to the data indicating that two of the field offices with the highest detention rates are Houston and Miami, where the purported risk of exposure to the elements does not hold up. See infra ¶¶ 176, 191. With respect to offices like Chicago and New York where the elements could pose a hazard to age-outs with nowhere to go, the data presented at trial indicated that Chicago was able to place the vast majority of its age-outs when it attempted to find a place for them, see infra ¶¶ 233–240, and although New York detained large percentages of its age-outs this was not due to a risk of homelessness because there was “a rich array of organizational sponsors” in the area to take them, see infra ¶ 211 (quoting Trial Tr. at 846:3–846:21 (Enriquez)). 122. Officer Hyde, SDDO in ICE’s New York City office, additionally suggested that lack of a sponsor could make an age-out a danger to the community. Trial Tr. at 3139:4–7 (Hyde) (explaining that she understands “harm to the community” to mean “someone, [with] no place to go that could potential[ly] be a harm or risk to the community at large”). No evidence was presented supporting this suggestion. 123. JFRMU condones and permits interpretations of the statute that view lack of a sponsor as indicative of at least some of the three statutory factors, despite recognizing that it is not the obvious meaning of the statute. Officer Sanchez, who is responsible for reviewing documentation of age-out custody determinations in JFRMU, testified that while “danger to self 58 usually is some kind of mental illness issues,” it “could” also include lack of a sponsor because an officer “could easily claim the fact that you had [no] address and you [are] going to be a homeless person that could constitute a danger to self because you have no place to stay and it’s cold outside.” Trial Tr. at 664:12–23 (Sanchez). When reviewing AORWs in advance of depositions in 2018, however, Officer Sanchez e-mailed Officer Hyde in New York to say that “[t]he factors for danger to self are self-harming or any other mental health issues” and to request revision of an AORW on which an FOJC had indicated that an age-out was a danger to himself because he had no sponsor. PX-284; see Trial Tr. at 662:20–666:11 (discussing PX-284). New York City SDDO Linda Hyde testified at one point that JFRMU “told [the New York office] that not having an address is not necessarily harm to self.” Trial Tr. at 3329:13–14. The Court does not find the assertion that JFRMU articulated a general policy along these lines to be credible. Officer Hyde gave this testimony owhile she was reviewing an email, PX-285, in which Officer Sanchez asked that revisions be made to a particular AORW. See Trial Tr. at 3322:2–3331:22 (Hyde). From context, the Court found it sufficiently clear that Officer Hyde did not mean to refer to this email when she said JFRMU had “told” the office not to consider lack of an address or sponsor as signifying danger or harm to self because she said she did not remember when this instruction came from JFRMU. Id. at 3329:18–20 (Hyde). In general, Officer Hyde’s credibility was undermined by her evasive and combative testimony. See Trial Tr. at 3252:16–20 (Hyde) (“The Court: . . . You need to answer his questions.”), 3496:8–9 (Pan) (“The Court: . . . Officer Hyde was a very combative witness this morning.”); see also, e.g., id. at 3238:6–3329:11 (Hyde) (refusing to respond to the question 59 asked even after the Court’s instruction to do so). 15 On the particular issue of when JFRMU instructed her office not to view lack of a fixed address as indicative of harm to self, she claimed both that she had already taken corrective actions and that she was closely monitoring the situation going forward, id. at 3330:2–3331:1, but she was not able to say when the message from JFRMU was received, id. at 3329:18–22, or when the corrective actions were taken, id. at 3331:2–3331:10. In light of Officer Sanchez’s testimony indicating that JFRMU did not have a problem with viewing lack of a fixed address or sponsor as indicating a danger to self, the Court thinks it unlikely that JFRMU communicated such instructions to the New York office, otherwise than in Officer Sanchez’s email regarding the AORW revisions. 124. Because lack of a sponsor can be viewed by ICE as signifying that an age-out is either a flight risk or, in some offices, a danger, an age-out’s lack of a sponsor carries determinative weight in many age-out custody determinations. See, e.g., Trial Tr. at 3230:2–3231:17 (Hyde), 3394:7–12 (Hyde) (agreeing, in reference to an age-out, that “[t]he fact that he didn’t have a sponsor meant that he was a danger to himself and a significant flight risk”); Galvez Dep. Desig. at 291:12–292:20 (Officer Galvez would designate an age-out a flight risk based solely on the lack of a sponsor); Pepple Dep. Desig. at 175:22–176:10 (testifying that the “vast majority” of age-outs are recent entrants to the country who have limited community ties and are therefore flight risks); see also Harper 5/2/19 Dep. Desig. at 28:20–29:8 (Chief Harper was not personally aware of any case where ICE released an age-out without a sponsor); PX-213 at 2 (AORW indicating “ORR was unable to provide a sponsor making subject a flight risk”). 15 Plaintiffs attacked Officer Hyde’s credibility in a footnote in their briefing, identifying a number of points on which she was impeached. Pls.’ Br. at 180 n. 5. Not only did Defendants not dispute anything in this footnote, but they did not even respond to it, Defs.’ Resp. at pp.333– 334, despite having responded to similar footnotes attacking the credibility of other witnesses, see id. at 63–66 (Chief Harper), 392–94 (Officer Zanello). 60 c. Risk Classification Assessment 125. FOJCs were required to run the RCA tool on each age-out they processed. Officer Ravenell attested to this, as did many FOJCs, and Defendants do not dispute it. Ravenell Dep. Desig. at 85:1–8; see, e.g., Hyde Dep. Desig. at 239:5–17; Barnes Dep. Desig. at 75:24–76:13; Pepple Dep. Desig. at 28:12–29:25; Reardon Dep. Desig. at 77:15–17; Galvez Dep. Desig. at 88:1–6; Love Dep. Desig. at 34:8–35:5, 36:11–22; see also Defs.’ Resp. ¶¶ 99–106 (“ICE policy is to require that the RCA tool be used for every adult custody decision, including age-outs.”). 126. In August 2017, the possible outputs given by the RCA tool regarding whether an age-out should be detained were adjusted: whereas before it had recommended detention or release, after the adjustment it would only either recommend detention or “say[] . . . something like see your supervisor” or “discuss with supervisor.” Trial Tr. at 176:13–177:5 (Harper); see also Gibney Dep. Desig. at 55:22–60:07 (stating that the tool was changed to recommend only “detain” or “supervisor to determine”); id. at 62:6–7 (stating that the change came from ERO leadership); see also Defs.’ Resp. ¶ 100 (failing to dispute that the tool was adjusted in this way). Thus, after the adjustment, the RCA tool would never recommend that an age-out be released. 127. JFRMU provided no guidance to the field when the release option was eliminated, and in fact continued to tell field officers to run the RCA on every age-out they process. At the April 2019 training FOJCs were instructed they must run the RCA as the first step of age-out processing. PX 203 at 10; Trial Tr. at 548:25–549:8 (Sanchez). See also PX-203 at 34 (same graphic repeated toward the end of the presentation under the heading “Key Take-Aways”); PX199 at 27 (slide identifying failure to process an RCA recommendation as a potential “case deficienc[y]”). Officer Ravenell, one of the presenters at the April 2019 training, did not recall when he learned about the change to the RCA outputs when asked at his deposition in May 2019, but said that it might have been in 2018. Ravenell Dep. Desig. at 86:5–17. Officer Sanchez, 61 another presenter, testified that she did not recall anyone telling the FOJCs in April 2019 that the RCA no longer recommended release. Trial Tr. at 684:13–685:4 (Sanchez). Chief Harper testified that when she learned of the change to the RCA tool’s outputs JFRMU “immediately started telling people, the FOJCs in the field, not to rely on the RCA.” Id. at 342:9–13 (Harper). If Chief Harper meant to suggest that she began telling FOJCs not to rely on the RCA at all, or to alert them to the change in the RCA tool’s outputs, the Court does not find this assertion credible. 16 Defendants have not produced or identified any evidence of any such communication from Chief Harper or JFRMU to the field. Defendants identify testimony from officers who said they understood that the RCA output was only a recommendation to be considered and testimony that the advisory nature of the RCA output was discussed on monthly JFRMU calls. See Defs.’ Resp. ¶ 101. But communications stating that the RCA was only a recommendation or only one among many factors to consider is not the same as communications explaining to the field that the RCA would longer suggest release in any case. Communications of this latter sort did not take place. At least two FOJCs learned about the changes to the RCA outputs for the first time at their depositions. See, e.g., Barnes Dep. Desig. at 75:11–76:22; Galvez Dep. Desig. 90:1–16. And Chief Harper in fact later testified that there was never any official communication from ICE headquarters announcing the change to the RCA outputs and that she only learned of the change from speaking with officers who run the RCA tool. Trial Tr. at 3363:7–22 (Harper); see also Trial Tr. at 3363:20–22 (Hyde) (“The Court: Was there ever any communication from headquarters saying, hey, we made this change? The Witness: No, sir.”). 16 If Chief Harper meant only that she began telling or reminding FOJCs not to rely solely on the RCA—the Court makes no finding one way or the other on that. 62 128. Chief Harper also testified at one point that the RCA tool was not used with age-outs because it no longer recommends release. Trial Tr. at 177:6–8 (Harper) (“Q: So now it says either detain or something like see your supervisor, correct? A: Yes, which is why we don’t use it with age-outs.”). This cannot be the case, because, even if the RCA recommendation was not relied on, the tool was undoubtedly “used with” age-outs, as Defendants admit in their briefing. Defs.’ Resp. ¶¶ 99–106. Chief Harper also testified otherwise, explaining near the end of trial that the tool is run in every instance as a matter of policy. Trial Tr. at 4320:14–20 (Harper). 129. ICE continued to cite RCA recommendations in monthly reports to Congress as a reason why age-outs were detained. In the first of ICE’s monthly reports to Congress summarizing age-out custody decisions, “RCA recommends detention” was the reason for detention for 46 of 93 age-outs, and no mention was made of the fact that the RCA tool, by that point, no longer recommended release. PX-153 (first monthly report to Congress); Trial Tr. at 340:7–342:2 (discussing the reporting requirements). 130. The Court finds that ICE FOJCs were never told by JFRMU that the RCA tool would no longer recommend releasing age-outs, and that FOJCs were—and remain—required to run the tool on each age-out and to take its recommendation under consideration, though not to rely only on the RCA tool in explaining a custody decision. d. Sequential decision making 131. The “one-pager” flowchart—JFRMU’s only written guidance concerning Section 1232(c)(2)(B) outside of annual training—suggests that ICE officers should first determine whether an age-out can be released or is a flight risk, danger to the community, or danger to himself who must be detained, and second, consider the full range of alternatives to detention only if the answer is the age-out can be released. The flowchart instructs in one box that “FOJCs make a Custody Determination based on the totality of the circumstances” and in the next that 63 “[i]f appropriate, FOJCs consider alternatives to detention programs based on Aged-Out UAC’s need for supervision.” PX-135 at 2 (emphasis added); see also PX-203 at 14 (same flowchart in 2019 Training presentation). The flowchart thus clearly implies that in at least some circumstances, it might not be appropriate to consider alternatives. Plaintiffs refer to this way of structuring the age-out custody determination as “sequential decision-making,” and Defendants do not appear to take issue with the terminology. Defs.’ Resp. ¶ 107. 132. Both Chief Harper and Officer Ravenell—another National Juvenile Coordinator in JFRMU—understood the flowchart as accurately reflecting the sequential decisionmaking process that FOJCs should follow when processing age-outs. Trial Tr. at 180:22–183:18; Ravenell Dep. Desig. at 149:5–152:24. Specifically, they both testified that, as reflected in the flowchart, FOJCs should first consider whether an age out is a danger to self, danger to the community, and that considering alternatives to detention “[i]f appropriate” meant considering them if the age-out was not determined to be a flight risk, danger to self, or danger to others at that previous step. Trial Tr. at 181:7–183:18; Ravenell Dep. Desig. at 151:13–152:24. 133. Relatedly, as Chief Harper testified, the ATD program (which covers a narrower range of options than the broader “alternatives to detention” mentioned in the statute, see supra ¶ 44 n.7) is “used as [a] mitigation factor,” and generally only considered as an option for ageouts whom ICE has already decided to release. Trial Tr. at 148:7–149:9. 3. Lack of Guidance 134. In a number of areas relating to age-out custody determinations and documentation, JFRMU (a) provides no training to FOJCs and SDDOs; (b) has no policy; and / or (c) has a policy that the issue is left to the unbridled discretion of individual FOJCs and SDDOs. 64 a. Age-Out Review Worksheets 135. FOJCs never received any formalized training on the October 2018 version of the AORW, which was in use between October 2018 and August 2019. Trial Tr. at 420:5–7 (Harper); see supra ¶¶ 51, 58. FOJCs testified that they received no guidance or training aside from the memo introducing the AORW and some discussion on monthly phone calls. Munguia Dep. Desig. at 97:22–98:2; Barnes Dep. Desig. at 69:2–71:12; Black Dep. Desig. at 52:6–17; Trial Tr. at 2650:11–12 (Black); Trial Tr. at 2829:24–2830:1 (Munguia) (“[T]hey told us to review it; and that if we had any questions, we were supposed to call them.”). Defendants have not identified any guidance concerning the October 2018 AORW other than the email introducing the AORW and whatever was conveyed to the field during monthly JFRMU phone calls. Trial Tr. at 224:6–9 (Harper) (“I don’t know if I issued official policy guidance, but I know we began to discuss the age-out process and the documentation and VAWA on the monthly phone calls”); see Defs.’ Resp. ¶ 183. 136. The April 2019 version of the AORW, which FOJCs trained on during the April 2019 training in San Antonio, was never used in the field. Supra ¶ 59 (citing Trial Tr. at 419:24–420:3 (Harper); id. at 546:16–547:2 (Sanchez)). 137. As of the time of trial, the only additional written guidance FOJCs received concerning the August 2019 revised AORW was the two-page email introducing the revised worksheet, see DX-6, and guidance that might have been given during monthly phone calls or on a case-by-case basis in response to an FOJC affirmatively reaching out to JFRMU with questions. Trial Tr. at 534:10–535:9 (Sanchez); see also Trial Tr. at 2370:12–17 (Hormiga) (“I don’t believe we ever had any training on any worksheets.”). 138. Defendants argue that the August 2019 guidance introducing the revised AORW “was discussed on several JFRMU monthly phone calls,” but they cite only to testimony by 65 Chief Harper stating in general terms what the monthly calls entail. Defs.’ Resp. ¶ 186 (citing Trial Tr. at 151:3–14). Defendants also argue more generally that during monthly phone calls “JFRMU provided training as to what individual items on the worksheets mean or how they should be completed.” Defs.’ Resp. ¶ 187. The testimony they cite again provides no detail about any particular instructions given or topics covered during the monthly phone calls, and instead consists only of general assertions about what kind of guidance is available during these phone calls. See id. (citing Trial Tr. at 46:24–47:13 (Harper), 387:4–25 (Harper), 420:8–13 (Harper), 534:10–13 (Sanchez)). Thus, there was no evidence presented that ICE ever provided guidance on critical issues relating to the AORWs, including, for example, what constituted an “other credible threat” in Item 8; whether one of the potential reasons in Item 8 had to be checked whenever an age-out was designated a flight risk or danger; what it meant for a sponsor to be “available”; what it meant to “evaluate” a discretionary release action like ATD or Bond; or what kind of justification an SDDO ought to provide for the ultimate decision. b. Detention Alternatives 139. FOJCs are not formally trained on placing age-outs in ICE’s ATD program. Defendants dispute this finding, arguing that Plaintiffs highlight the testimony of only three FOJCs and that some FOJCs did place age-outs on ATD. See Defs.’ Resp. ¶ 188 (citing testimony by Chief Harper). By the Court’s count, the numbers are slightly worse for the Defendants. At least five FOJCs testified that they did not receive such training. See Hyde Dep. Desig. at 32:20–22; Barnes Dep. Desig. at 40:6–8; Hormiga Dep. Desig. at 83:11–18; Trial Tr. at 3715:14–17 (Irizarry); see also Trial Tr. at 2944:10–13 (Venegas) (indicating no written guidance on how to place an age-out on ankle bracelet monitoring). Three more testified that they understood the ATD program not to be an option for age-outs, at least in their AOR. Zanello Dep. Desig. at 117:6–9; Pepple Dep. Desig. at 178:3–4 (“ATD is not additionally used for age66 out cases in this AOR [Chicago].”); Barnes Dep. Desig. at 40:16–22; see also Black Dep. Desig. at 180:23–181:9 (unaware that ankle bracelet monitoring, specifically, was an option); but see Trial Tr. 2515:13–15 (Black) (testifying at trial that she had released an age-out on an ankle bracelet, presumably at some point between her deposition and the trial). The fact that some FOJCs managed to place age-outs in the ATD program (in many cases only after this litigation was ongoing for a year or more) does not establish that JFRMU trained them on how to do so. While ATD is mentioned as an option in certain training materials, including the “one-pager” flowchart, PX-135, and the FOJC handbook, DX-4 at 48, Defendants have identified no written evidence or testimony suggesting that JFRMU ever provided any formal or informal guidance on how FOJCs could actually go about putting an age-out on ATD, or on what circumstances might make doing so appropriate. To the contrary, as the abovereferenced testimony demonstrates, some FOJCs were confused about the use of the ATD program with age-outs. Chief Harper agreed that “field officers consider age-outs for participation in ICE’s alternative detention program only rarely, if at all.” Trial Tr. at 149:6–9 (Harper). Further, Defendants do not dispute that “several FOJCs were not even aware the ATD program was available to Age-Outs” or “ICE does not promote or require consideration of ATD when making custody determinations for Age-Outs.” See Defs.’ Resp. ¶¶ 152–53; see also Trial Tr. at 149:18–22 (Harper) (“[T]here has also been significant rotation of people into the FOJC position, and [the availability of ATD for age-outs] may not have been relayed to them.”). 140. To the extent FOJCs ever recommended placing age-outs in the ATD program—and this rarely happened, see infra ¶ 164—they only did so as part of the flawed sequential decisionmaking process that misunderstands how Section 1232(c)(2)(B) ought to be applied. Rather than treat the ATD program as an alternative to detention (lowercase), they treated it as 67 an alternative to releasing an age-out on his or her own recognizance. Chief Harper testified that “the decision to put somebody on ATD is made after you release somebody” and agreed that “[FOJCs] make the decision whether you’re going to detain first.” Trial Tr. at 148:12–25; id.at 149:1–5 (“Q: And only if you decide you’re not going to detain and you’re going to release do you get to the question of looking at alternatives to detention such as release with ankle-bracelet monitoring, right? A: That is how the FOJCs do it.”). Officer Ravenell agreed that this decisionmaking structure, reflected in the “one-pager” flowchart, was the process FOJCs should follow. Ravenell Dep. Desig. at 149:5–152:24. This view of the program turns VAWA’s “continuum of alternatives” on its head. 8 U.S.C. § 1232 (c)(2)(B). 141. FOJCs are not formally trained on releasing age-outs subject to an ICE Bond. Defendants “do not dispute that there has not been any written guidance on bond,” Defs.’ Resp. ¶ 189, but suggest that there has been some training during monthly phone calls. Chief Harper testified that ICE bonds had “been discussed at length,” Trial Tr. at 160:9, but she did not say when or where, and she was unable to identify any document other than the AORW itself in which “field officers were advised . . . concerning the use of ICE bonds with respect to ageouts.” Trial Tr. at 160:10–23 (Harper); see also id. at 160:8–9 (Harper) (“I don’t know if I’ve provided written guidance to them.”). In disputing this fact, Defendants cite testimony from Chief Harper and Officer Sanchez in which they described the wide breadth of topics discussed on JFRMU’s monthly phone calls to the field, but none of this testimony makes any reference to ICE Bonds, see Defs.’ Resp. ¶ 189 (citing Trial Tr. at 46:24–47:13 (Harper), 387:4–25 (Harper), 420:8–13 (Harper), 534:10–13 (Sanchez)), and Chief Harper did not mention the monthly calls when asked about training on ICE bonds, Trial Tr. at 159:12–160:14. Even if the Court credits Defendants’ weakly supported assertion that ICE bonds were discussed on monthly JFRMU calls 68 at some point, the fact remains that no formal training on the topic took place, and the evidence indicates that ICE bonds are also rarely used with age-outs, see infra ¶ 168. 142. In one of the few instances where ICE Bonds were issued, they were issued by the Miami field office not as part of an alternative to detention intended to mitigate risk of flight, but instead $5,000 bonds were set as a prerequisite for release from ICE detention. Trial Tr. at 1313:22–1315:15 (Lehner); id. at 3935:10–16 (Reardon) (reviewing an AORW reflecting that Officer Reardon set a $5,000 ICE bond); id. at 3966:15–3967:6 (same, for another AORW). This practice is, again, the opposite of what the statute requires as it treats what could be a riskmitigating alternative to detention as only available if the age-out is already otherwise on track for release. 143. FOJCs are not trained or given guidance on releasing age-outs to shelters outside their AOR. See Trial Tr. at 3723:20–25 (Irizarry), 3867:14–3868:12 (Zanello). c. Statutory Risk Factors 144. JFRMU does not train FOJCs on the factors that Section 1232(c)(2)(B) requires ICE to take into account when making age-out custody determinations. Put another way, FOJCs are not trained on what makes an age-out a flight risk, a danger to themselves, or a danger to the community. See, e.g., Trial Tr. at 424:13–425:1 (Harper) (no formal JFRMU training on any of these three factors); Id. at 666:16–22 (Sanchez) (“Q: So whatever the consideration is, just so [long as] you put it on the paper and tie it to one of the statutory factors, that’s sufficient? A: No, I’m not saying it’s sufficient, but if you could articulate in a clear and conscious way that when you read it you can understand what you are talking about, then you can go ahead and do it.”); Hyde Dep. Desig. at 124:1–9 (flight risk, danger to community), 124:17–20 (danger to self); Trial Tr. at 3363:23–3364:19 (Hyde) (all three factors); Black Dep. Desig. at 115:10–17 (flight risk), 115:18–116:5 (danger to self or community); Trial Tr. at 2653:11–2654:16 (Black) (same); 69 Hormiga Dep. Desig. at 191:15–192:5 (flight risk, danger to community). Defendants argue that FOJCs are law enforcement officers who do not need specialized training on the meaning of these terms, but they do not dispute that FOJCs do not receive any training as to their meaning. Defs.’ Resp. ¶ 196. d. Policies of Discretion 145. JFRMU defers to the discretion of FOJCs and SDDOs with regard to the outcomes of age-out custody determinations. Trial Tr. (Harper) at 54:2–55:2, 56:13–15, 64:3–7; Sanchez Dep. Des. at 18:1–16 (“[JFRMU] only look[s] at the documentation because the decision is solely up to the field.”). Chief Harper did not believe she had ever asked a field office or field officer that a custody determination needed to be changed, was incorrect, or not compliant with Section 1232(c)(2)(B). Trial Tr. (Harper) at 55:14–56:12; see also Trial Tr. at 513:1–14 (Sanchez) (agreeing that JFRMU reviews only the documentation, not the substance of custody determinations). At the 2019 annual training, FOJCs were instructed that “there is no right/wrong answer, but you should properly document your decision.” PX-54 at 22. Defendants explain this as a matter of ICE chain of command and hierarchy, but do not dispute that the substance of age-out custody determinations is not reviewed by JFRMU. See Defs.’ Resp. ¶¶ 208–09. 146. It is up to the discretion of the field officer whether to attempt to contact an ORR shelter in advance of an age-out’s 18th birthday to ask for a post-18 plan. Trial Tr. (Harper) at 86:14–23, 92:1–19; see also Ravenell Dep. Desig. at 41:12–18. 17 17 The parties dispute whether Chief Harper’s testimony on this point was inconsistent. See Defs.’ Resp. at pp. 64–65. At trial she was asked whether “ICE encourage[s] its FOJCs to affirmatively reach out” to ORR for a post-18 plan, and initially responded affirmatively. Trial Tr. (Harper) at 85:21–24. Plaintiffs’ counsel then confronted her with testimony from her deposition in which she was asked “Should the FOJC contact the shelter in advance to request -- 70 147. JFRMU has no written guidance advising FOJCs to reach out to ORR about potential sponsors to whom ORR did not release the age-out, Ravenell Dep. Desig. at 35:5–12, and no written guidance concerning whether FOJCs should attempt to identify or verify sponsors on their own, id. at 41:12–42:22, including by reviewing information available on the age-out’s form I-213 18, id. at 49:11–50:14. 148. ICE leaves it entirely up to the discretion of the field officer whether to interview or speak with an age-out on or before the age-out’s eighteenth birthday, either in order to ask about or attempt to identify potential sponsors or for any other reason. Trial Tr. at 93:22–94:15 (Harper); Ravenell Dep. Des. at 46:18–47:1. 149. It is likewise entirely up to the field officer’s discretion whether to contact a potential sponsor identified in a post-18 plan, Trial Tr. (Harper) at 88:7–19, 93:4–10, to try to identify a sponsor if none is provided or if a proposed sponsor is deemed unsuitable, id. at 93:13–16, 190:12–18, or to acknowledge, respond to, or follow up on communications received from in advance of the child’s 18th birthday to request the post-18 plan” and responded “I wouldn’t say they should do that. The burden for the post-18 plans falls on ORR.” Id. at 86:2–13 (citing Harper 5/2/2019 Dep. Desig. at 12:14–18). Asked again whether “ICE encourage[s] its FOJCS to affirmatively reach out, Chief Harper responded in the negative. Id. at 86:14–17. Defendants note, correctly, that the question Chief Harper was asked at trial was not the same as the question Plaintiffs’ counsel read from her deposition testimony. Defs.’ Resp. at p. 65. At the same deposition, however, the question immediately following the one Plaintiffs’ counsel confronted Chief Harper with at trial was the same question from trial regarding whether “ICE encourage[s] its FOJCs to affirmatively reach out,” and Chief Harper responded “ICE doesn’t make a recommendation either way.” Harper 5/2/2019 Dep. Desig. at 12:19–13:1. So, while Chief Harper was not confronted with the most precise excerpt from her deposition, she did change her testimony at trial to become more consistent with her deposition testimony—and more consistent with Officer Ravenell’s testimony. 18 A form I-213 records a narrative of ICE’s experience with a UAC or other alien, and begins by explaining the circumstances in which they were encountered, whether by ICE or by Customs and Border Patrol. It includes information such as where a UAC was headed and who, if anyone, the UAC called on the phone when he or she was apprehended. See Ravenell Dep. Desig. at 48:15–49:23; Trial Tr. at 969:21–970:5 (Hyde); Pepple Dep. Desig. at 33:8–20. 71 attorneys or advocates for age-outs that request release or propose sponsors, id. at 88:20–91:25 19; Ravenell Dep. Desig. at 37:24–38:6; see also id. at 36:15–19 (agreeing that ICE has no requirements “with respect to what an FOJC should do when someone provides him or her with the name of a potential sponsor”). 150. ICE generally requires that a proposed sponsor be deemed “viable” before ICE will release an age-out to that sponsor. Trial Tr. at 130:6–11 (Harper); see also PX-334A (summary of AORWs indicating that “no viable sponsor” was identified as the reason for detention for over 100 age-outs between October 2018 and May 2019). 151. JFRMU and ICE have no “set or written requirements” for what makes a proposed sponsor viable or suitable; viability is left up to the discretion of individual FOCJs and SDDOs. Trial Tr. at 130:23–131:11 (Harper); Harper 5/2/2019 Dep. Desig. at 22:1–23:3; see also Ravenell Dep. Desig. at 44:17–45:4 (confirming the lack of written guidance on steps FOJCs should take to identify or vet sponsors, and that this topic was not addressed at the April 2019 training); id. at 57:11–14 (“ICE is not responsible for finding sponsors, so . . . we wouldn’t have any policy, procedures, in order to determine whether [a sponsor] is viable or not). After eighteen days of trial, it is still unclear to the Court how ICE defines a “viable sponsor.” 152. It is “up to the discretion of individual officers and field officers whether to use ATD,” and JFRMU does not monitor the extent to which they do so. Trial Tr. at 157:22–158:2 (Harper). 19 During questioning on this third point, Chief Harper was confronted with testimony from her deposition and Plaintiffs argue that she changed her testimony at trial in response. See Pls.’ Br at 35 n.2. Defendants argue the testimony was not inconsistent. Defs.’ Resp. at p. 65. The Court agrees with Defendants with regard to this particular line of questioning. While Chief Harper’s answers at trial were more nuanced than those she gave at her deposition, they were not so different as to damage her credibility on this issue. 72 D. Outcomes at the Field Office Level 1. Exercises of Discretion a. Sponsors 153. Officers in JFRMU take the position that ORR, not ICE, is responsible for identifying potential sponsors. Trial Tr. at 74:3–20, 76:15–77:3 (Harper); Harper 5/2/19 Dep. Desig. at 43:7–17 (agreeing that “the officer’s only obligation is just to consider what somebody else provides” and that “ORR is responsible for identifying potential sponsors”); Trial Tr. at 573:7–9 (Sanchez); Ravenell Dep. Desig. at 36:15–37:9; see also PX-270 at 1 (Chief Harper email to Miami field office stating “VAWA . . . doesn’t convey an affirmative responsibility onto ICE to find group homes or sponsors”). 154. ORR has different and higher standards that it applies when considering potential sponsors for UAC than ICE has when it considers sponsors for adults. Before HHS and ORR can release a UAC to a proposed sponsor they are required by Section 1232(c)(3)(A) to “make[] a determination that the proposed custodian is capable of providing for the child’s physical and mental well-being.” 8 U.S.C. § 1232 (c)(3)(A). This inquiry must “at a minimum, include verification of the custodian’s identity and relationship to the child, if any, as well as an independent finding that the individual has not engaged in any activity that would indicate a potential risk to the child.” Id.; see also id. § 1232(c)(3)(B) (requiring more extensive home studies for proposed sponsors of children with certain particular vulnerabilities or traumatic histories). Nothing in the TVPRA or VAWA sets similar requirements for ICE. Officer Ravenell explained that this is because immigrants who have aged out of ORR custody are, at that point, adults. Ravenell Dep. Desig. at 58:10–18 (“Q: And to your knowledge, ICE doesn’t have requirements with respect to the suitability of sponsors for adults, correct? A: If you’re an adult, you’re an adult . . . [W]e don’t go out looking for sponsors, we don’t go out checking to 73 see what type of individuals they are. We don’t have to do suitability checks for sponsors.”); id. at 58:22–24 (“If you’re an adult, you’re 18, it’s not ICE’s responsibility to do suitability checks for where you’re going to live.”). 155. As discussed, JFRMU takes the position that it is up to individual field officers whether to communicate with ORR regarding potential sponsors for age-outs, supra ¶¶ 146–47, and that it is up to individual officers whether to seek out or attempt to identify potential sponsors, supra ¶¶ 148–49, or to request a post-18 plan, supra ¶ 146. 156. Defendants have intimated, if not directly argued, that differences in detention rates between field offices may be attributable at least in part to whether post-18 plans are generally received by FOJCs in those field offices that detain at higher rates. Compare, e.g., Defs.’ Br. FF ¶¶ 263–65 (explaining that the San Antonio field office generally received post-18 plans a sufficient amount of time in advance of the age-out’s eighteenth birthday), with id. ¶¶ 321–23 (explaining that this is not the case for the Houston office); id. ¶ 456 (explaining that the New York City office “would not always receive Post-18 Plans”); and id. ¶ 535 (explaining that the Miami office did not reliably receive post-18 plans before a certain point). Plaintiffs dispute Defendants’ proposed findings regarding the Houston and Miami offices. See Defs.’ Resp. ¶¶ 321–23, 535; see also id. ¶ 456 (conceding that Officer Hyde so testified but not conceding the truth of the matter). Even assuming, however, that offices detaining at higher rates were those that did not receive post-18 plans, this would only show a correlation, not a causal effect. Cf. Trial Tr. at 2048:23–2049:25 (Lenzo) (Defendants’ counsel making the point that correlation does not equal causation). Further, even if there were a causal relation between high detention rates and a lack of post-18 plans, it is far from clear which would be more likely to cause the other. For example, 74 during a time period when El Paso was detaining most of its age-outs, the SDDO in that office had limited what kinds of materials ICE would accept from legal services and required hard copy delivery of these materials rather than submission via email. See infra ¶¶ 248, 250. At least one legal services provider found fulfilling these onerous requirements “pretty time-intensive” and “ultimately . . . made a decision that [they] just couldn’t afford the time to submit packets anymore” given that even the strongest cases for release “continued to get denied.” Trial Tr. at 1509:25–1510:10 (Escontrias). Officer Pepple from the Chicago office testified that he began receiving more post-18 plans when he started releasing more age-outs and reaching out to ORR and ORR contractors for names of potential sponsors. See infra ¶ 238. It stands to reason that ORR and third parties may have seen no reason to waste time submitting post-18 plans to offices that would detain all their age-outs regardless. If a causal relationship exists between submission of post-18 plans and detention rates, it strikes the Court that this causal relationship may plausibly run in either direction. 157. In many instances, ICE does not release an age-out to a proposed sponsor because the sponsor fails to meet criteria set by ORR for release of a UAC—as opposed to an age-out—from ICE custody to a sponsor. For instance, at one time, ORR required that sponsors and members of sponsors’ households be fingerprinted before an age-out could be released to them. Trial Tr. at 143:18–146:4 (Harper) (describing a memorandum of understanding between ICE and ORR which, from roughly April 2018 through December 2018, had required fingerprinting all members of a sponsor’s household before a UAC could be released); id. at 399:10–23 (Harper) (describing types of people that ORR no longer requires fingerprints from, including “[p]eople that claim to be adult siblings or grandparents” and “other adult household members”); see also id. at 878:3–879:23 (Enriquez) (third-party legal services provider describing his experience with 75 ORR fingerprinting requirements); id. at 2629:9–25 (Black) (testifying to her awareness that ORR sometimes requires fingerprints but that ICE does not have a fingerprinting requirement); id. at 3266:13–23 (Hyde). A number of age-outs were detained by ICE because the age-out did not have a “viable sponsor” when ORR did not manage to complete its fingerprinting check on a potential sponsor, despite ICE not having such a requirement. See, e.g., Reardon Dep. Desig. at 216:10–217:24; PX-211 at 8 (age-out memo from HHS to ICE indicating the case manager “was unable to have [a] new household member fingerprinted on time before [UAC] aged out”); Trial Tr. at 3791:17– 20 (Zanello) (“[I]f ORR had already vetted a potential sponsor, we didn’t need the specifics, we would just need to know yes, that sponsor works, no, that sponsor doesn’t, so we can move on.”); Zanello Dep. Desig. at 212:21–214:1 (discussing PX-211); Hyde Dep. Desig. at 287:1–288:21 (discussing PX-74); id. at 289:8–290:9 (same); Galvez Dep. Desig. at 314:15–315:8, 318:6– 320:1 (referencing PX-40); see also, e.g., Black Dep. Desig. at 159:7–161:21 (reviewing PX-16 at 55, 60); Reardon Dep. Desig. at 111:24–112:1 (“Q: So the mother was the proposed sponsor and the only holdup was fingerprinting of her? A: Yes.”); Trial Tr. at 3966:11–14 (Reardon) (“Q: So your understanding is that for a sponsor to be available within the meaning of item 12 [on the AORW], ORR has to have completed their vetting or approved that sponsor, correct? A: Yes.”); PX-335.16133–16135 (AORW from Los Angeles field officer recommending detention and noting that “shelter staff was unable to locate any potential sponsors”); id. at .14767–810 (AORW from Los Angeles field officer recommending detention and noting that a potential sponsor did not participate in ORR reunification efforts). 158. Another way in which ICE applies ORR standards for release of UACs to adult ageouts is by viewing an age-out as a flight risk based on his or her having spent an extended period 76 of time in ORR custody as a child. See supra ¶ 114 (citing, among other evidence, PX-155 at 4 (email from Chief Harper stating “One big factor [regarding detention] is the amount of time that ORR has a minor and hasn’t been able to find anyone to release them to. That in and of itself indicates a lack of ties and flight risk.”)). In Houston, for example, Officer Mason approved detention of an age-out who had been in ORR custody for 157 days and for whom ORR had prepared a post-18 plan recommending placement at a group home. See PX-4; Trial Tr. at 3058:23–3063:22 (Mason) (discussing PX-4). Officer Mason explained that, notwithstanding ORR’s recommendation, “[o]ne big concern that [ICE] had with this case is if they do have a sponsor that has already said they would take him then why didn’t ORR release him[?]” Trial Tr. at 3061:18–20 (Mason); see also id. at 3190:16–3191:14 (Hyde) (“That’s an indicator that they were not able . . . to find someone or someplace for the potential age-out.”); id. at 3221:6– 11 (Hyde) (“It’s telling me that they have been in there for an extremely long time, and ORR has not been able to find someone for that length of time for sponsorship.”); Hormiga Dep. Desig. at 171:15–172:8, 176:18–20 (testifying that he considered an age-out to be a flight risk because “she had been in custody for 151 days, and her uncle [a proposed sponsor] had the opportunity to come forward and go through the process and have her released, and apparently, he didn’t,” id. at 171:17–20, but also that he did not recall contacting ORR about the uncle); Trial Tr. at 2878:5– 16 (Venegas) (“When you see an age-out pass a certain time frame, say 30 to 40 days [in ORR custody], . . . that’s one of the factors that I would take into account in determining to detain or not detain.”). 159. The Miami field office went even further and developed a “local ICE requirement” that, in order to be deemed valid, sponsors not only needed to show proof of U.S. citizenship or lawful status but also needed “proof that they had ties to the community, like through a utility 77 bill [or] an affidavit of support.” Trial Tr. at 3925:10–24 (Reardon). When Chief Harper and JFRMU staff visited the Miami office they informed that office that there was “no legal basis” for these requirements. Id. at 3961:9–17 (Reardon). 160. Chief Harper also believed that, in situations where ORR was aware of a sponsor but had not released an age-out to them, that sponsor was presumably not viable. In April 2018, she sent an email asking her JFRMU staff to find out more information concerning the Washington D.C. field office’s decisions to release four out of five age-outs in February 2018. See generally PX-55; see also Trial Tr. at 291:24–296:9 (Harper) (beginning discussion of PX-55). For three of the age-outs, ICE had released to sponsors that had been known to ORR for some time, but to whom ORR did not release the age-outs. PX-55 at 4; Harper 5/2/19 Dep. Desig. at 308:8–310:7. Chief Harper “wanted to know what information [the FOJC] had that ORR didn’t have.” Harper 5/2/19 Dep. Desig. at 309:10–11. Chief Harper’s question demonstrates that she at least did presume that a sponsor to whom ORR had not released an age-out had at least uncertain viability, unless some facts on the ground, unknown to ORR, had changed. Chief Harper agreed at her second deposition that the fact that ORR does not place a UAC with a sponsor does not mean there is no viable sponsor at the time that the UAC becomes an age-out. Harper 5/2/2019 Dep. Desig. at 62:2–7. She answered the same question evasively at trial and was impeached with her prior testimony. Trial Tr. at 141:22–142:18 (Harper). Given her inconsistent testimony, the Court finds that the position suggested in her April 2018 e-mail was more likely her actual stance on the question. 161. Officer Galvez took the same view. He testified that a sponsor recommended by ORR but not vetted and approved by ORR would still have to be vetted in some way. Galvez Dep. Desig. at 312:15–314:13. 78 162. With no obligation to identify sponsors on their own and ample discretion regarding whether to communicate with the office—ORR—that is expected to identify sponsors, many FOJCs do not release age-outs to sponsors unless ORR has already gone through the work of verifying or approving placement with the sponsor. See, e.g., Barnes Dep. Desig. at 44:11–23 (testifying that he was unaware of any situation where an age-out was released to an individual or organizational sponsor “identified and/or arranged for by ICE”); Galvez Dep. Desig. at 68:2– 18 (testifying that in his experience, only the ORR case manager arranges for a sponsor); Hyde Dep. Desig. at 55:20–56:6 (testifying that ICE does not “attempt to identify places where th[e] age-out . . . could live”); Hormiga Dep. Desig. at 80:10–17 (testifying that calling organizational sponsors is “not what we do” and “just not in our job description.”); id. at 87:19–88:5; Trial Tr. at 3023:16–20 (Mason) (agreeing it is “not the FOJC’s job to look for an organizational sponsor” if the age-out does not have one on his 18th birthday); Reardon Dep. Desig. at 59:2–13; Trial Tr. at 3232:7–10 (Hyde); see also Sanchez Dep. Desig. at 324:8–12 (testifying she was not aware of “an instance where the FOJC found a sponsor for an age-out that had not been identified by ORR or an advocate”); Pls.’ Resp. FF ¶ 127 (confirming that this is consistent with Defendants’ understanding of ICE’s jurisdiction); see also Trial Tr. at 3865:11–19 (Zanello) (testifying she has only placed age-outs with a particular organizational sponsor when attorneys have provided her with a letter from that sponsor, never on her own initiative). 163. In a number of offices, ICE’s refusal to take any action by way of investigating, vetting, or verifying sponsors 20 means that age-outs are essentially never released to sponsors when they turn 18 unless ORR has done all, or nearly all, the vetting it normally requires and has 20 With no official policies in this area, the language used to describe what needed to be done with regard to sponsors tended to vary from office to office and officer to officer. 79 made any necessary arrangements. At these offices, if ORR has not released the UAC to a sponsor by the time they turn eighteen, the age-outs will, with near-certainty, be placed in adult detention on their eighteenth birthday when they are transferred to ICE custody, even if sponsors have been identified. See infra ¶¶ 182–184 (Houston), 196 (Miami), 210–211 (New York City), 249 (El Paso). b. Detention Alternatives 164. As attested by ICE’s 30(b)(6) witness on the ATD program, Eric Carbonneau, the ATD program can mitigate a risk of flight, and age-outs are eligible for the program (unless they intend to live in certain rural areas). Carbonneau Dep. Desig. at 23:12–53:2, 80:12–21, 115:13– 116:7. Some officers do place age-outs on ATD. Officer Jones in Los Angeles, for example, has released age-outs on ATD several times and agreed that ATD could mitigate risk of flight and would consider it if risk of flight was the only concern. Trial Tr. at 4271:22–4272:21 (Jones). 165. ICE does not train its field officers on placing age-outs into the ATD program and, to the extent officers nonetheless know how to do so, it is up to their discretion whether they choose to place age-outs in the ATD program. See supra ¶¶ 139, 152. 166. Overall, as Chief Harper agreed, “field officers consider age-outs for participation in ICE’s alternative detention program only rarely, if at all.” Trial Tr. at 149:6–9 (Harper). Defendants do not dispute that “the evidence at trial was overwhelming and undisputed that the ATD program is almost never considered as an alternative to detention for Age-Outs.” See Defs.’ Resp. ¶ 149 (citing, inter alia, Trial Tr. at 2785:7–8 (Munguia) (“Q: Do you ever consider ATD? A. No.”); id. at 2908:24–25 (Venegas) (“Q: And ICE in Harlingen have an ATD program for age-outs? A: Not in the Harlingen office, we do not.”); id. at 2990:6–8 (Mason) (“Q. Has your office ever released an age–out to the ATD program? A. No, we haven’t.”); id. at 3339:12– 16 (Hyde) (testifying she was not aware of a juvenile being released on ATD from the New York 80 office in the past ten years); id. at 3715:11–13 (Irizarry) (testifying he had never released an ageout to the ATD program); id. at 3873:6–8 (Zanello) (same); Hormiga Dep. Desig. at 84:1–4; Pepple Dep. Des. at 23:6–10; Reardon Dep. Des. at 12:6–13); see also Trial Tr. at 3339:7 (Hyde) (“I have never considered ATD, correct.”). And, as noted, supra ¶ 139, several FOJCs were not aware that ATD was an option for age-outs. 167. From the date that the AORW and SharePoint system was instituted—October 17, 2018—through the close of discovery on May 24, 2019, just 34 of 1,360 age-outs processed by ICE were placed on ATD (2.5%). Harper 8/29/19 Decl. ¶ 7, ECF No. 204-1; see also Defs.’ Resp. ¶ 150 (not disputing these figures). 168. ICE bonds are likewise rarely used by field officers making age-out custody determinations. Defendants do not dispute that most field offices do not release age-outs on ICE bonds and that those offices that do often do so only where an age-out or family member has specifically requested one. Defs.’ Resp. ¶ 155; see Trial Tr. at 161:17–163:1 (Harper) (testifying it would not be surprising to learn that field officers from some of the largest field offices testified they had never released an age out on an ICE bond); Munguia Dep. Desig. at 88:15–24; Trial Tr. at 2824:7–10 (Munguia); Hyde Dep. Desig. at 144:6–145:10; Trial Tr. at 3339:24– 3340:12 (Hyde); Barnes Dep. Desig. at 79:9–15; Pepple Dep. Desig. at 45:13–23, 47:2–5; Reardon Dep. Desig. at 69:22–70:1; Galvez Dep. Desig. at 80:15–81:10; Black Dep. Desig. at 46:17–24 (testifying she was not sure when issuing an ICE bond would be appropriate); Trial Tr. at 2909:3–7 (Venegas); id. at 3681:10–11 (Irizarry); id. at 3873:15–20 (Zanello). 169. On 51.8% of the AORWs produced in this litigation, the officer who filled out the form indicated that no consideration was given to releasing on OREC or OSUP, placing the age81 out on ATD, or using an ICE Bond. Trial Tr. at 1654:3–6 (Horton); see supra ¶ 56, subparagraph n (describing Item 14 on the AORW). c. Release After Detention 170. Persons in immigration detention have a right to have their bond determined or redetermined before an immigration judge. Trial Tr. at 1774:2–5 (Vinikoor). Plaintiffs presented expert testimony from retired Immigration Judge Robert Vinikoor concerning these types of proceedings. Id. at 1774:24–1775:5 (Vinikoor). 171. Immigration judges making a bond redetermination “consider[] almost identical factors” to those considered by ICE in making an age-out custody determination. Id. at 1780:6– 16 (Vinikoor). The relevant immigration regulation requires that a detained immigrant seeking release has the burden of “demonstrat[ing] to the satisfaction of the [immigration judge] that such release would not pose a danger to property or persons, and that the alien is likely to appear for any future proceeding.” 8 C.F.R. §§ 236.1 (c)(8). Judge Vinikoor explained that in his expert opinion, because the factors considered are so similar to those in Section 1232(c)(2)(B), “[t]he release of an age-out by an immigration judge on bond is compelling evidence that the age-out was not a danger to the community or significant flight risk at the time ICE placed the individual in custody.” Trial Tr. at 1780:6–9 (Vinikoor). Defendants dispute the fact that the standards overlap only partially and, even then, ineffectively. First they quibble with Judge Vinikoor’s use at trial of the phrase “critical overlap,” id. at 1780:11 (Vinikoor), and claim that the overlap is not “critical.” Defs.’ Resp. ¶ 261 The plain text of the two provisions make clear that two of the three statutory factors are encompassed in the regulation’s standard. This overlap is significant, regardless of whether it is properly described as “critical.” Second, Defendants note that “§ 1232(c)(2)(B) contains an entirely additional factor, ‘harm to self,’” which is not mentioned in the regulation. Id. Leaving 82 aside the fact that the statutory factor is “danger to self” not “harm to self,” § 1232(c)(2)(B), the absence of this factor does not meaningfully undermine the point that an immigration judge evaluates factors very similar to those evaluated by an FOJC. Relatively little testimony at trial focused on the “danger to self” factor, and there is no reason to think that ICE was detaining many age-outs specifically because they were dangers to themselves. 21 In fact, ICE’s SharePoint system, which asked FOCJs to identify a reason for detention of each age-out they detained, did not provide an option for “danger to self,” and it was therefore not identified in the system as a reason for detention for any of the 363 age-outs detained between August 2018 and April 2019 for whom ICE provided SharePoint data. Defs.’ Resp. ¶¶ 596–97 (not disputing this); see also PX-505 (SharePoint data in excel format). This goes to show that ICE did not consider this factor particularly important, and immigration judges’ lack of obligation to consider it does not make the standard they apply meaningfully less comparable to the standard applied by FOJCs. An immigration judge’s evaluation is also different in that they are typically determining or redetermining a bond amount for a detained immigrant. This again does not make the immigration judge’s evaluation incomparable to the FOJC’s because the FOJC also has the option of setting bond and releasing an age-out subject to that bond. 172. Of the 1,342 age-outs detained by ICE between January 1, 2018 and May 24, 2019, sixty-two percent (446 in total) were subsequently released on bond by an immigration judge. Trial Tr. at 1951:10–13 (Lenzo) (discussing Pls.’ Demo Ex 9, which summarizes relevant data from PX-399 and PX-400). Defendants dispute the conclusions Plaintiffs draw from this fact, but do not dispute the numbers. See Defs.’ Resp. ¶ 260. 21 Officers in two field offices suggested that age-outs are dangers to themselves if they have no place to live. The Court has explained why it doubts this was a genuine explanation for these offices’ detention decisions. Supra ¶ 121. 83 173. This means that a high proportion of age-outs were released by immigration judges applying a standard similar to the standard applied by FOJCs, with two notable differences: danger to self was not taken into account, and the burden was placed on the former age-out to demonstrate lack of danger to the community and lack of flight risk. The Court finds this release rate significant, and views it as meaningful evidence that the FOJCs were not properly considering the factors laid out by the statute. Judge Vinikoor agreed with this conclusion. See Trial Tr. at 1796:9–20 (“[T]his is compelling evidence that they weren’t a danger to the community or flight risk at the time they were taken into custody.”). Defendants challenge Judge Vinikoor’s qualifications to offer this opinion because he was not sure he had ever held a bond redetermination hearing for an age-out, because former age-outs could bring new evidence to their bond hearing, because facts may have changed between the initial custody determination and the bond hearing, and because immigration judges can reach different conclusions on the same set of facts. Defs.’ Resp. ¶ 260. The Court finds none of these arguments compelling. First, while the Court found Judge Vinikoor to be a credible and helpful expert witness, his view of these facts only supports conclusions that the Court could have reached on its own. Expert testimony is not required for the Court to compare the standard set in Section 1232(c)(2)(B) with the standard set in the immigration regulations, nor is it required for the Court to review the parties’ summary of the rate at which detained age-outs were later released by immigration judges. It is helpful to know that the high rate of release caused Judge Vinikoor to question the thoroughness of the initial custody determinations, but the Court could have reached that conclusion on its own. As for the possibility of changed circumstances or new evidence, that argument is speculative and does not significantly diminish the persuasiveness of the comparison. 84 Defendants do not support their suggestion with any specific examples. Moreover, it is counterintuitive. It is unclear how former age-outs could reduce their danger to the community or risk of flight while in adult detention—they are unlikely to establish community ties or change their criminal history or gang affiliations while locked up. As for danger to self, no evidence was presented that former age-outs receive better or more effective psychological or psychiatric care while in adult detention than they were able to receive in an ORR facility (and the Court deems that scenario unlikely)—and very few age-outs fall in this category anyway. Finally, the possibility that two different immigration judges exercising discretion could reach different conclusions on the same set of facts, see Trial Tr. at 1828:10–15 (Vinikoor), is obvious and irrelevant. One or two instances of an immigration judge disagreeing with an FOJC’s detention recommendation could be disregarded on this ground, but when immigration judges disagree with FOJCs sixty-two percent of the time, despite placing a higher burden on the immigrant, this scale of systematic disagreement suggests something more than reasonable decision-makers reaching different conclusions. 2. Age-Out Detention Rates by Field Office 174. Historically, ICE field offices have detained age-outs at varying rates, with some detaining all or nearly all age-outs, some releasing all or nearly all, and others with results that fall somewhere in between or swing from one extreme to the other. As will be explained, these differences in detention rates can be seen in the ERO Custody Management Division data produced by Defendants, PX-345, and summarized by Plaintiffs in PX-334 and PX-334A. At trial, Plaintiffs also established the variations in field office detention rates by showing a series of Rule 1006 summary graphs and charts. See Pls.’ Demo. Exs. 21A, 22A, 23A, 24A, 25A, 26A, 27A, 28A, 28B, 28C, 68 (“the Rule 1006 Summaries”); see also Fed. R. Evid. 1006 85 (allowing the use of “a summary, chart, or calculation to prove the content of voluminous writings, recordings, or photographs that cannot be conveniently examined in court.”). Each is a line graph that charts, for each month, the percentage of age-outs encountered at one or more field offices. They summarize the information contained in Defendants’ raw data, PX-345. Defendants objected to the use of these summaries at trial, e.g. Trial Tr. at 1918:19– 1919:16, and have renewed their objection that the Rule 1006 Summaries were not timely produced to them by Plaintiffs and thus should not be admitted, Defs.’ Resp. ¶ 219. The Court has already ruled on the admissibility of the Rule 1006 Summaries in a Minute Order dated January 19, 2020. Min. Entry (Jan. 9, 2020); see also Min. Entry (Feb. 19, 2020) (clarifying prior Minute Order). The Court noted that the Government had not argued that the Rule 1006 Summaries were inaccurate or otherwise not proper summaries of the data. Min. Entry (Jan. 9, 2020). The Court agreed that the Summaries were produced and updated after the deadline for exhibits, but found it equitable to extend the deadline for this limited purpose. Id. Defendants had also updated their own production after the deadline, and admitting Plaintiffs’ demonstratives as Rule 1006 Summaries was not prejudicial to Defendants because Plaintiffs had timely produced nearly identical versions of the Summaries and because the Court did not see how Plaintiffs’ cross-examination at trial would have been different had they known the materials would be admitted as demonstratives rather than as substantive evidence. Id. Defendants do not identify any errors in this analysis and the Court stands by it. Defendants also argue that the Rule 1006 Summaries are visually misleading in a few ways. Pls.’ Resp. FF ¶ 219. They argue the graphs are misleading because they do not show the number of age-outs processed each month, but only the percentage of age-outs detained. This is not a serious concern because the Summaries prominently state that they depict the “Monthly 86 Detention Rate,” (emphasis added), not absolute figures. Defendants also argue that the Plaintiffs have misleadingly drawn a line through “100%” rather than “0%” on the charts’ y-axis for months where a field office processed no age-outs. Id. (citing Pls.’ Demo. Exs. 25A, 26A, 27A). Again, the Court finds the summaries are clear enough, as the charts include a star on the lower boundary of the y-axis for these months. The Court appreciates that in creating a summary of voluminous data, decisions must be made regarding how to represent certain nuances, and finds the Plaintiffs’ design choice reasonable and easy enough to bear in mind when reviewing the charts. The rate for these months is nonexistent, so indicating 0% would have been just as inaccurate as indicating 100%. The summaries qualify under Rule 1006, and Plaintiffs and their expert have explained what they do and do not reflect. The Court has not relied on the Rule 1006 Summaries as substantive evidence, though it finds them to accurately represent the underlying data. 175. Eight of the ICE field offices discussed at trial have rates of age-out detention that the Court found significant in one way or another. Three demonstrated consistently high average rates of age-out detention: Houston, Miami, and New York City. See infra ¶¶ 176–214. One field office demonstrated a consistently low average rate of age-out detention: San Antonio. See infra ¶¶ 215–223. Four demonstrated extreme shifts in age-out detention rate: El Paso, Harlingen, Chicago, and San Francisco. See infra ¶¶ 225–259. a. High Average Detention Rate i. Houston 176. Between April 2016 and May 24, 2019, the ICE field office in Houston detained 515 of 538 age-outs it processed (96%). PX-334 at 43–53; PX–334A at 147–184. In no month did 87 the Houston office detain fewer than 80% of the age-outs it processed. 22 Id. Between October 17, 2018 and May 24, 2019—the portion of the discovery period during which ICE was using the AORW documentation system—the Houston office detained 97% of age-outs it processed. PX334 at 50–53 23; PX-334A at 147–84; see also Pls.’ Demo. Ex. 24A. 24 177. Officer Barnes was the lead FOJC in Houston from 2015 through October 2019, when Officer Hormiga became the lead FOJC. Barnes Dep. Desig. at 8:17–18; Hormiga Dep. Desig. at 27:4–11. Officer Mason was the SDDO in Houston during this entire time period. Barnes Dep. Desig. at 8:18–25; Hormiga Dep. Desig. at 24:5–10. The officers’ testimony described procedures that could be expected to yield a very high detention rate. 178. Officer Barnes could not recall any age-outs he encountered that he did not view as flight risks. Barnes Dep. Desig. at 149:6–12. He never released an age-out on his or her own 22 This statistic alone belies Defendants’ suggestion that the Houston office’s high detention rate was the result of a “surge” or “influx” of UACs in 2018. See Defs.’ Br. FF ¶¶ 313–15; see also Trial Tr. at 2959:18–23 (Mason) (suggesting that the Houston office lacked sufficient FOJCs during an “influx”). The detention rate there was consistently high for years— both before and after any temporary increase. Further, the Houston office has never had fewer than four FOJCs, see Trial Tr. at 2959:5–10 (Mason), and processed an average of fourteen ageouts per month, PX–334A at 147–84. The San Antonio office had the same number of FOJCs, and processed the same average number of age-outs per month. See Trial Tr. at 2754:21–23 (Munguia); PX-334A at 329–66. That office was able to release nearly every age-out they processed, despite having the same ratio of personnel to age-outs. Infra ¶ 215. 23 PX-334 lists age-outs processed by each field office, sorted by their dates of birth. In order to translate this data in to date ranges for age-out processing, the Court has assumed, based on the testimony presented at trial, that each age-out was processed by ICE on his or her eighteenth birthday. 24 Defendants agree that these statistics are “by and large,” accurate. Defs.’ Resp. ¶ 354. In their own proposed findings of fact, they argue that Plaintiffs incorrectly count 137 age-outs handled by the Houston office during the AORW time period, when in fact the number is only 134. See Defs.’ Br. FF ¶ 325 n.1. Plaintiffs disagree. See Pls.’ Resp. FF ¶ 325. Neither party meaningfully disputes the figures its adversary presents for the Houston office. See id.; Defs.’ Resp. ¶ 354. Because a difference of three either way would not change the agreed-upon fact that the Houston office detained nearly all its age-outs, the dispute over these three cases is not material and the Court need not resolve it. 88 recognizance, id. at 28:11–13, and could not remember anyone else in his field office having done so, id. at 43:11–14. He could not identify a single age-out released by the Houston field office who was not a woman with a child or a person with a serious medical issue. Id. at 135:10– 16. 179. In a number of AORWs reviewed at his deposition, Officer Barnes had indicated that the age-out was a flight risk despite checking “no” for each of the specific characteristics in Item 8 that would make an age-out a flight risk (Terrorist Affiliation or Ties; Criminal/Delinquency History; Gang Affiliation; Escapes/Non-compliance History; and Other Credible Threat). See generally id. at 277:13–332:17 (reviewing AORWs). 180. Office Barnes filled in identical answers on a number of AORWs, and testified that he may have been using a completed AORW as a template from which he created additional AORWs for other age-outs. Id. at 138:2–5, 141:17–25. 181. Officer Barnes was unable to identify a single factor that might prevent an age-out from coming into ICE custody. Id. at 247:1–24, 249:15–22. 182. The Houston office also failed to follow ICE’s requirements for documenting age-out custody determinations. For more than thirty age-outs processed by the Houston office between October 17, 2018 and May 24, 2019, no AORW was uploaded to the SharePoint Database. Trial Tr. at 2995:18–21, 2997:10–14. Officer Hormiga acknowledged that six of the twenty-four AORWs with his name on them had been edited by someone else, and that he had not filled out twelve of them in the first place. Hormiga Dep. Desig. at 304:19–305:10. 183. The Houston field office made little to no effort to identify potential sponsors for ageouts. SDDO Mason testified that it is not an FOJC’s job to look for organizational sponsors when a UAC turns eighteen without having been released by ORR. Trial Tr. at 3023:11–20 89 (Mason). Officer Hormiga, the primary FOJC in Houston, and Officer Barnes, his predecessor in that role, both agreed. Hormiga Dep. Desig. at 80:7–17; Barnes Dep. Desig. at 44:11–23. Officer Hormiga views his role with regard to sponsors and alternatives to detention as “being limited to approving or denying alternatives that someone else has given [him].” Hormiga Dep. Desig. at 87:19–88:5. Even when an alternative was presented, though, the Houston office would sometimes reject it, saying it had not been given the information early enough to allow for sufficient investigation of the proposed sponsor. See Trial Tr. at 3048:12–3049:8 (Mason). Because ICE has issued no guidance on what constitutes a “viable sponsor,” it is unclear what investigation was necessary that would have required additional time. 184. Neither Officer Hormiga nor Officer Barnes has ever tried to identify an individual or group sponsor. Hormiga Dep. Desig. at 85:11–86:11; Barnes Dep. Desig. at 45:4–8. Neither was aware of anyone else in the unit ever having made such an attempt. Hormiga Dep. Desig. at 86:12–18; Barnes Dep. Desig. at 45:9–11. Officer Barnes was not aware of any situation in which an age-out was released to a shelter or group home or sponsor that ICE had identified or made arrangements for. Barnes Dep. Desig. at 44:4–45:11. Although Houston FOJCs are not categorically opposed to vetting potential sponsors like FOJCs in some other field offices, they also explained many detention decisions on AORWs by asserting that there was “no viable sponsor.” See PX-334A at 178–184 (reflecting that “no viable sponsor” was identified as a reason for detention for 73 of the 120 age-outs detained by the Houston office between November 2018 and May 2019). Neither officer would speak with age-outs about finding a sponsor. Barnes Dep. Desig. at 151:14–24; Hormiga Dep. Desig. at 56:9–18. 185. Officer Hormiga viewed his window of responsibility for age-outs as only the day of their eighteenth birthday, and did not see himself as having any obligation to prepare in advance 90 of their aging out or to follow up on their case after making a placement decision. Hormiga Dep. Desig. at 60:7–19. He testified that “one day is not enough time to be able to release somebody. But if you have time, then you can possibly release somebody.” Id. at 199:17–19. 186. No detention alternatives are used by the Houston office when processing age-outs. Officer Barnes viewed age-out custody determinations as presenting only two options: release or adult detention. Barnes Dep. Desig. at 42:13–18. The Houston office has not used ICE bonds for age-outs. Id. at 79:9–15, 79:24–80:9; Hormiga Dep. Desig. at 114:22–115:7. Defendants dispute this characterization of Officer Hormiga’s testimony because, they argue, he testified that although ICE bonds are not the office’s preferred means of release, an age-out processed by the Houston office could be given one. Defs.’ Resp. ¶ 374 (citing Hormiga Dep. Desig. at 114:2– 18). This is an accurate account of Officer Hormiga’s testimony but it overlooks the fact that he was then asked, “has that happened in your experience” and responded “[n]o . . . Never has.” Hormiga Dep. Desig. at 114:19–21. The Court’s finding is thus sufficiently specific—the office has not used an ICE bond, even if it could. No age-out has ever been placed into the ATD program from the Houston field office. Barnes Dep. Desig. at 40:23–41:1; Hormiga Dep. Desig. at 84:1–9; Trial Tr. at 2990:6–8 (Mason). Officer Barnes was not aware that ICE’s ATD program was an option for age-outs and had not been trained on it. Barnes Dep. Desig. at 40:6– 8, 40:15–22. Officer Hormiga also was not trained on using ATD. Hormiga Dep. Desig. at 83:11–13. 187. The Houston office’s failure to consider alternatives extended even to extreme cases. Officer Barnes detained one age-out in January 2018 even though the age-out was represented by a Catholic Charities staff attorney who contacted Officer Barnes a week before the age-out’s eighteenth birthday to request that the age-out be placed at an organizational sponsor. See PX-4. 91 The attorney attached a copy of a letter from the organization saying that they were willing to sponsor the age-out. Id. at 10–12. ORR also submitted a post-18 plan recommending placement with the same organization. Id. at 13–15. The age-out had no criminal history and had an application for asylum pending. Id. at 1–3, 10. The closest thing to a reason for detention that Officer Barnes wrote on the AORW was the conclusory sentence, “Subject is a flight risk.” Id. at 2. Officer Mason approved the detention recommendation. Id. at 3. This decision was so unreasoned that ORR officials contacted JFRMU to ask for reconsideration. See PX-269 at 3 (email from Sanchez to Barnes and Mason). When Officer Sanchez emailed the Houston officers to ask about it, Officer Mason gave no explanation except the fact that the age-out “ha[d] no family in the United States and [no] ties to the community,” that he was therefore a flight risk, and that “it is in the best interest of the Government that the subject be transferred to adult detention.” Id. at 2. 188. JFRMU knew that Houston detained nearly all of their age-outs, but did not see any substantive problem in need of correcting. In a follow up email conversation with Chief Harper, Officer Sanchez expressed concerns only with Houston’s paperwork, and with what she anticipated would be Officer Barnes’s anticipated poor performance as a witness in this case. Id. at 1 (“The reason for detention is flight risk, but the officer did not articulate the facts. The officer did a poor job completing the [AORW]. . . . Officer Barnes should be able to articulate the facts for a flight risk. He has been a [FOJC] for years. I have no faith in him for the deposition; I hope that I am wrong.”); see also Trial Tr. at 227:23–228:18 (Harper) (discussing this email). She further wrote “it appears that they [Houston] detain most of the age outs. My concern is that they do a terrible job with the paperwork.” PX-269 at 1. JFRMU evidently did 92 not take issue with the substance of the decision to detain this eighteen year old who could have been placed in a shelter. 189. One age-out processed in Houston aged out of ORR custody when she was six months pregnant and was detained despite ORR having submitted a post-18 plan recommending release to a cousin. PX-335.22449 (AORW); PX-335.22479 (post-18 plan). Officers responsible for placing her in adult detention were not informed of her pregnancy. See PX335.22467 (email from an officer working in the intake unit stating “we were not aware of the pregnancy”); see Hormiga Dep. Desig. at 301:5–302:1. The eighteen year old was sent to the emergency room about a week after being placed in adult detention, suffering from what appeared to be serious medical issues. See PX-335.22468 (“moderate to severe back and abdominal pain, worsening over the past day, leaking yellow fluid she does not believe to be urine; also believes she has not been able to feel the baby move today”). She was released on OREC a week after her eighteenth birthday, and ICE officers filled out an inaccurate AORW indicating that she was released on OREC when she aged out—glossing over her week in adult detention and trip to the emergency room. See PX-335.22449–51; see also Hormiga Dep. Desig. at 303:7–304:13; Defs.’ Resp. ¶ 370 (not disputing the facts of this individual’s experience). 190. Officer Hormiga’s deposition was taken—outside of court—during the trial in December 2019. As of that time, there had been no changes made to the process by which Houston FOJCs make age-out custody determinations. Hormiga Dep. Desig. at 196:10–17. ii. Miami 191. Miami: Between April 2016 and May 24, 2019, the ICE field office in Miami detained 181 of 189 age-outs it processed (96%). PX-334 at 57–60; PX–334A at 218–47. Between October 17, 2018 and May 24, 2019 the Miami office detained over 98% of age-outs it processed. PX-334 at 59–60; PX-334A at 240–47; see also Pls.’ Demo. Ex. 25A. 93 192. Officer Sean Reardon was deposed in January 2019. See Reardon Dep. Desig. At that time he had been an FOJC in the Miami office for about four or five years. Id. at 5:1–4. His deposition testimony thus describes practices in Miami prior to the August 2019 Miami field office visit during which Chief Harper and other officers in JFRMU traveled to the Miami field office for a training session. See supra ¶¶ 104–08. Officer Reardon also testified at trial in December 2019. Trial Tr. at 3895–3990 (Reardon). 193. Prior to the JFRMU visit in August 2019, Miami FOJCs had been given several instructions that JFRMU viewed as incorrect and clarified at the training. SDDO Lee-Fatt had told the Miami FOJCs that they were not allowed to contact potential sponsors. Trial Tr. at 3925:10–3926:8 (Reardon). Miami Officers had also been told that they could not release an age-out to a sponsor unless the sponsor had lawful immigration status (as a lawful permanent resident or U.S. citizen). Trial Tr. at 3925:10–24, 3926:9–18 (Reardon). Following JFRMU’s visit to Miami and additional guidance, Miami FOJCs now understand that they may contact and vet sponsors and that sponsors do not need any particular status. Trial Tr. at 3927:10–20, 3928:11–18, 3929:9–21 (Reardon). 194. The Miami office was not reliably compliant with ICE’s systems for documenting age-out custody determinations. Officer Reardon testified that when ICE compiled a monthly age-out report (before the AORW was introduced), Miami’s reporting was not consistent and that the office probably missed at least some months. Reardon Dep. Desig. at 281:6–25 (“Yeah, we slacked on it. No doubt.”). After the AORW and SharePoint system was adopted in late 2018, Miami field officers sometimes completed forms for one another. Id. at 83:10–14. 195. Officer Reardon’s understanding, at least at the time of his deposition, was that an age-out could be detained even if they were not determined to be a flight risk, danger to the 94 community, or danger to themselves. Reardon Dep. Desig. at 138:5–12; see, e.g., PX-171; PX335.08683–85. He also viewed age-outs without verified sponsors as potential flight risks. See Reardon Dep. Desig. at 139:10–16. 196. The Miami office regularly detained age-outs with proposed sponsors, stating on their AORWs that there was “insufficient . . . time for ICE to verify sponsorship.” See, e.g., Reardon Dep. Desig. at 96:20–98:8 (discussing PX–169); id. at 129:22–131:14, 142:18–143:1 (discussing PX-170); id. at 142:18–144:6, 149:23–150:7 (discussing PX-171). Prior to the August training, having the name of the proposed sponsor earlier would hardly have made a difference, for at least two reasons. 197. As Officer Reardon testified, Miami FOJCs believed they were not allowed to vet sponsors. Id. at 87:15–22. Even at the time of trial, after the August training, Officer Reardon was not able to identify any procedures in place for vetting or approving sponsors. Trial Tr. 3952:19–3953:5 (Reardon) (testifying he was not aware of any procedures for vetting). 198. Second, although FOJCs in the Miami office cited lack of time as a reason for detention, they made no effort to do anything with the time available to them. In the Miami office, a weekly list of UACs approaching their eighteenth birthdays was distributed, but Officer Reardon did not do anything with the list. Reardon Dep. Desig. at 112:8–113:1, 278:22–279:22. Miami FOJCs only ever had one day to work on a given age-out because they did no work before the age-out’s eighteenth birthday and would transfer the age-out to adult detention under the authority of ICE’s detained unit on that day, along with any information on possible sponsors. See Reardon Dep. Desig. at 116:25–117:17, 119:6–22; Trial Tr. at 3952:11–18 (Reardon); id. at 3953:13–3955:14 (Reardon) (testifying that when a post-18 plan was received from ORR, as early as two weeks before the age-out date, it would just be passed along to the detained unit); id. 95 at 3955:15–17 (Reardon) (agreeing that only the detained unit would do anything with the proposed sponsor information). Fifty-seven of seventy-eight worksheets from the Miami office gave “insufficient amount of time” or “lack of sufficient time” as a reason why an age out was detained. See Pls.’ Br. FF ¶ 385 (collecting citations to AORWs); Defs. Resp. FF ¶ 385 (not disputing this). 199. The Miami field office at one point had what Officer Reardon called a “local ICE requirement” that, in order to be deemed valid, sponsors needed to show proof of U.S. Citizenship or lawful status and had to demonstrate ties to the community through evidence like a utility bill or affidavit of support. Trial Tr. at 3925:10–24 (Reardon); supra ¶ 159. JFRMU later informed that office that there was “no legal basis” for these requirements. Id. at 3961:9–17 (Reardon). 200. In nine cases processed while these unlawful local ICE requirements were in place, the Miami office refused to release an age-out to a parent who had been proposed as a sponsor, citing insufficient time to investigate the sponsor and fulfill these local requirements—even though there was no documentation suggesting any question about the identity of the parents or any information suggesting that the age-out was a flight risk, danger to self, or danger to the community. See PX-172; PX–335.15716, PX–335.18946, PX–335.14143, PX–335.14218, PX– 335.09290, PX–335.25874, PX–335.10945, PX–335.26507. 25 201. In at least some instances, EARM comments in the age-outs’ files indicate that ICE’s detained unit was able to release age-outs detained by Officer Reardon within a week of their 25 In eight more cases, ICE denied release to a proposed sponsor who was the age-out’s sibling, aunt, or uncle, again citing no reason to doubt the relationship and no evidence of the statutory risk factors but only saying there was insufficient time to investigate the sponsor. See PX–335.26545, PX–335.09175, PX–335.14414, PX–335.18355, PX–335.19855, PX–335.02434, PX–335.17037, PX–335.23340. 96 detention to sponsors that he claimed he lacked time to verify. See PX-169 at 7; PX-170 at 6; see also PX-171 at 19 (release to sponsor approximately one month after detention). 202. The Miami field office regularly released the age-outs that it detained after habeas petitions were filed on their behalf by legal services providers like Americans for Immigrant Justice. Trial Tr. at 1320:12–1321:2 (Lehner) (attorney for Americans for Immigrant Justice). There were twenty-two such cases between July 2018 and December 2019. Id. at 1320:12–18; see also, e.g., id. at 1307:13–1313:4 (Lehner) (discussing one case in more detail). None of the cases proceeded to a hearing on the merits, meaning that no judge ever ruled on the lawfulness of ICE’s detention of these twenty-two age-outs. Id. at 1313:12–18. ICE released each of them voluntarily before the habeas case reached that point. Id. at 1313:19–21; see Defs.’ Resp. ¶ 401 (not disputing this Plaintiffs’ proposed finding regarding these habeas cases). 26 203. Even following the August 2019 JFRMU site visit, the Miami field office believes that sponsors are only available for age-out placement if they have been approved by ORR. When reviewing the AORW at his deposition in January 2019, Officer Reardon testified that in order to answer affirmatively to Item 12, which asks whether an individual or organizational sponsor is available for placement, it was required that the sponsor be “verified or approved by ORR.” Reardon Dep. Desig. At 67:19–22. At trial, when asked the same question, he first responded “we’ll work with ORR on group homes if we can and on sponsors if we can. So I would verify the group home and the sponsor,” and, when pressed, responded that he wasn’t sure whether the sponsor had to be approved or verified by ORR in order to be available for placement. Trial Tr. at 3965:11–17. 26 Releasing these age-outs without waiting for a judge to rule on the legality of their detention suggests that the ICE office tasked with detaining them was not entirely comfortable with the juvenile unit’s decision to detain them in the first place. 97 Defendants dispute this proposed finding. Defs.’ Resp. ¶¶ 390–93. It is not clear from their response whether they challenge the notion that Officer Reardon and other Miami FOJCs believed sponsors needed to be approved by ICE before the August 2019 site visit, or whether they only challenge the notion that Miami FOJCs still believed this at the time of trial or at present. Either way, the Court disagrees. Before the August 2019 site visit, the evidence is clear—Officer Reardon rejected proposed sponsors and detained age-outs, explaining on the AORW that ORR requirements that ICE did not share had not been met. See Trial Tr. at 3970:19–22 (Reardon) (reviewing PX-169, an AORW and accompanying documentation for an age-out whose proposed sponsor was rejected for failing to complete ORR’s fingerprinting requirements); id. at 3980:11–13 (Reardon) (reviewing PX-335.18946–48, AORW for an ageout whose proposed sponsor was rejected for failing to complete an ORR home study). For the time period after the site visit, and up through the present, it is less obvious what Officer Reardon understood it to mean for a sponsor to be available. The Court finds that his testimony suggests he still thinks ORR approval is a necessary step. Asked “[w]hat exactly do you do now to vet sponsors?” he responded that, though he had not had many cases since the change in policy, “now if ORR is saying they have a sponsor but they don’t have any status, or if ORR is -- needs help in any way to contact a sponsor to call, we’ll try to assist them, anything we can do.” Trial Tr. at 3929:15–21 (Reardon). And, as quoted above, he also testified, “we’ll work with ORR on group homes if we can and on sponsors if we can.” Id. at 3965:11–13 (Reardon). Both these descriptions of Miami’s post–August 2019 procedures suggest that Officer Reardon views his role in the process as helping facilitate ORR approval of sponsors, rather than vetting or investigating proposed sponsors to see if they meet ICE’s own criteria. 98 204. Prior to 2019, the Miami field office did not consider placing age-outs on ATD or releasing them on ICE bonds. Officer Reardon was familiar with the ATD program, as he formerly worked in that unit. Reardon Dep. Desig. at 6:7–7:7. Nonetheless, in his four or five years as an FOJC prior to his deposition he had never referred an age-out to the ATD unit and was not aware of any age-outs that were placed there by others. Id. at 11:11–19, 12:6–13; but see PX-256 (2018 email from ICE ERO writing to the juvenile unit to say that an age-out would have to be placed on ATD because the adult detention facility was full); Defs.’ Resp. ¶ 403 (“Defendants do not dispute that the Miami Field Office did not utilize ICE’s ATD Program . . . . ”). As of his deposition in January 2019, Officer Reardon had never released an age-out on an ICE bond and was not aware of anyone else in the Miami office having done so. Reardon Dep. Desig. at 69:22–70:1. At some point, in March or April 2019, SDDO Lee-Fatt told the Miami FOJCs that they could offer ICE bonds to age-outs. Trial Tr. at 3931:19–3932:12 (Reardon); see also id. at 3932:13–18 (Reardon) (expressing uncertainty about when in 2019 this instruction was given); id. at 1313:22–1314:18 (Lehner) (testifying that ICE began offering bonds in the summer of 2019). 205. The bonds offered by the Miami field office beginning in 2019 have been in the amount of $5,000—more than the $1,500 minimum used by other offices. Id. at 1314:19–1315:1 (Lehner); see DX-59 at 3 (Miami office AORW indicating a bond set for $5,000); see also Trial Tr. at 3143:12–21 (Hyde) (testifying that the New York City office set bonds at the minimum amount of $1,500). No evidence was presented concerning how this amount was set, why it was greater than other offices, or whether it was calibrated based on risk of flight. The families of many age-outs struggled to pay this bond, so many age-outs who had $5,000 bonds set were nonetheless sent to adult detention facilities. See Trial Tr. at 1315:2–1316:8 (Lehner). 99 206. The current policy in the Miami field office is that ICE’s detained unit will no longer release age-outs to sponsors and that if an age-out has been detained, the detained unit will only release them if they pay the $5,000 ICE bond or obtain a lower bond from an immigration judge. Id. at 1317:9–13 (Lehner); see also DX-108 (September 2019 guidance email from JFRMU to Officer Lee-Fatt explaining that after the initial age-out decision is made, Section 1232(c)(2)(B) no longer applies and the age-out should be treated as any other adult in ICE custody); Trial Tr. at 4327:6–4328:12 (Harper) (explaining this guidance and the basis for JFRMU conveying it to the Miami office). iii. New York City 207. New York City: Between April 2016 and May 24, 2019, the ICE field office in New York City detained 121 of 135 age-outs it processed (90%). PX-334 at 61–63; PX-334A at 248– 85. Between October 17, 2018 and May 24, 2019 the New York City office detained 82% of age-outs it processed. PX-334 at 62–63; PX-334A at 275–85; see also Pls.’ Demo. Ex. 27A. 208. The only witness from ICE’s New York City field office from whom the Court heard testimony was Officer Linda Hyde, who has been SDDO in the New York City office since 2015. Trial Tr. at 3102:16–21. 209. Officer Hyde and the New York City office view those age-outslacking fixed addresses as flight risks, dangers to themselves, and dangers to the community. See Hyde Dep. Desig. at 165:11–14 (“Lack of ties is an indicator that they have no ties to the community. They have no one here. And that’s a potential . . . flight risk.”); id. at 169:17–170:4, 246:6–13 (flight risk); id. at 228:16–229:7 (“[A] person with no place to live, no means of -- to support themselves, could be a danger to themselves . . . . I attribute it to being, you have no place to live, you have no place to eat, no place to drink, you could become a danger to yourself.”); Trial Tr. at 3328:3–11 (Hyde) (“[I]n my interpretation, if someone has no place to go and [is] placed 100 out in the elements, it could be a danger to self, yes.”); Hyde Dep. Desig. at 192:24–193:14, 227:13–19 (danger to self); Trial Tr. at 3139:4–7 (Hyde) (explaining that she understands “harm to the community” to mean “someone [with] no place to go that could potential[ly] be a harm or risk to the community at large”). 210. In the vast majority of cases, the New York City office does not attempt to identify, investigate, or vet potential sponsors. Officer Hyde testified that it is not ICE’s responsibility to identify sponsors. Hyde Dep. Desig. at 55:20–56:6, 99:16–21, 128:9–13; Trial Tr. at 3124:14– 18. She also testified that her office does not vet sponsors, but relies on ORR to do that. Trial Tr. at 3268:2–9 (Hyde). With investigation and both vetting off the table under Officer Hyde’s leadership, this leaves nothing for ICE to do with regard to sponsors. Because the New York field office neither identifies nor vets sponsors, a UAC in their AOR that has not already been released to a sponsor by ORR will not be released to a sponsor on their eighteenth birthday. 211. The New York City field office does not maintain a list of potential organizational sponsors in anything approaching a formal or orderly manner. Officer Hyde gave inconsistent and confusing testimony on this issue. E.g. id. at 3341:25–3342:3 (“Q: Do you maintain a list of potential placement folks and organizations who accept age-outs in the New York area? A: No, we have a list in the office.”); id. at 3343:2 (“We looked at the list of sponsors -- entities in New York.”); id. at 3343:17–19 (“Q: Okay. You don’t have your own list of potential sponsors for young adults, 18 year olds? A: No, I get it from the internet.”); id. at 3344:16–18 (“Q: So you do have a list now or you don’t have a list? A: I think we have one somewhere.”). Needless to say, no such list was produced at trial. The Court has explained its concerns regarding Officer Hyde’s credibility generally, see supra ¶ 123 & n.15, and this issue is no different. While there may have been some kind of list created in the New York office at some point in Officer Hyde’s 101 five years as SDDO, it is clear from Officer Hyde’s confused testimony that no formal list was in use. The high detention rate supports the conclusion that no such list was being used even if it existed somewhere. This is despite the fact that there is a “rich array of institutional sponsors” in the New York area. Trial Tr. at 846:3–846:21 (Enriquez). 212. The New York City field office does not consider proposed sponsors or post-18 plans coming from anyone other than ORR. Id. at 3268:24–3269:14 (“[w]e don’t know who's credible or who is not credible. We don't know if we're going to get correct information or incorrect information. So we depend on ORR to do the vetting for us.”); PX-488 at 1 (email from Officer Hyde to her supervisors asking “Should we consider Post-18 plans from advocate[s] and attorneys? (I would stick with Post 18 plan from the shelters)”); see also Trial Tr. at 3151:9–10 (Hyde) (“[S]ince the minors are not 18 as yet, we try to steer them to ORR . . . .); id. at 3297:1–5 (Hyde) (“[L]obbying [on behalf of a UAC] should go to ORR, because ORR still had custody. . . . [A]ll this information that the [legal services provider] had would go to ORR, and that would be included in the post-18 plan that would come to ICE.”); PX-292 (email from Officer Hyde stating “ICE cannot get involve[d] with ORR’s procedures . . . prior to a minor turning 18. When a minor turn[s] 18 or ages out of custody that is when ICE involvement begins.”) Officer Hyde felt strongly enough that she sent an email to Chief Harper expressing incredulity with the suggestion by immigration attorneys in New York that ICE bore some responsibility for identifying sponsors. PX-451 at 1 (email from Officer Hyde to Chief Harper, complaining about immigration attorneys sending emails to ORR shelters citing this Court’s opinion granting a preliminary injunction in this case, ECF No. 28); Trial Tr. at 3321:18–20 (Hyde) (“I thought it was unbelievable that they believe that we should go door-to-door looking for sponsorship.”). 102 213. Like some other field offices, the New York City office takes the position that it only has one day—the age out’s eighteenth birthday—to take any action with regard to an age-out. Trial Tr. at 3155:19–3156:8, 3240:8–16 (Hyde). The New York City field office makes age-out custody determinations on that day. Hyde Dep. Desig. at 83:7–18, 84:16–85:3. 214. The New York City field office rarely released age-outs with ICE bonds, and never places them in ICE’s ATD program. Officer Hyde testified that she has given age-outs ICE bonds in the minimum amount of $1,500, but only rarely and only when an age-out or an ageout’s family member asks for one. Hyde Dep. Desig. at 142:11–25, 144:6–145:8. Officer Hyde has never referred any age-outs to the ATD unit, and is not aware of anyone else in her field office having done so. Id. at 32:17–33:18. 27 She testified that she “do[esn’t] believe in ATD” because she believes requiring age-outs to wear ankle bracelets would be “shaming” them. Id. at 139:3–7. Even if there was a flight risk that needed to be mitigated—as there is in nearly every case as Officer Hyde views it—she would prefer to use any option other than ATD. Id.at 34:9– 19 (“If I believe a person is a flight risk, I would look for other alternatives. But ATD in my estimation is the last resource because he’s going to cut that bracelet regardless. I believe that I’m shaming an individual by putting a bracelet on him. 18-year-old going to school, I don’t think that’s appropriate.”). But this position, again, appears to view ATD as an alternative to ROR rather than as an alternative to detention. Officer Hyde did not explain why being released with an ankle bracelet would be any more shameful than being placed in adult detention. One age-out processed by the New York City office was deemed “not eligible for alternative to 27 Somewhat inconsistently with ICE’s broader emphasis on the need to vet sponsors for the sake of age-outs’ wellbeing, Officer Hyde testified that once a sponsor pays a bond, the sponsor has no ongoing responsibility for taking care of the age-out, and need not even be in this country. See Trial Tr. at 3185:10–3186:6 (Hyde). 103 detention” because—according to ICE—he did not have a fixed address. See id. at 192:17–23 (quoting PX-69 at 2). b. Low Average Detention Rate i. San Antonio 215. San Antonio: Between April 2016 and May 24, 2019, the ICE field office in San Antonio released 527 of 545 28 age-outs it processed (96%). PX-334 at 76–87; PX-334A at 329– 66. Between October 17, 2018 and May 24, 2019 the San Antonio office released 117 of 119 age-outs it processed (98%). PX-334 at 85–87; PX-334A at 359–66; see also Pls.’ Demo. Ex. 21A; Trial Tr. at 2805:23–2806:12 (Munguia) (testifying that the San Antonio office has historically released all or almost all age-outs on their own recognizance except for those cases in which the age-out has a criminal history or serious medical condition.). Each of the two ageouts who were detained by the San Antonio office during the time that the AORW and SharePoint system was in place had a criminal history. PX– 335.07680–.07682; PX–335.27174– .27176. 216. Officer Jose Munguia is the primary FOJC in the San Antonio field office (one of four FOJCs in that office), and has been an FOJC for roughly eleven of the past sixteen years. Trial Tr. at 2754:5–25 (Munguia). In those eleven years he does not recall ever having detained an age-out that he would otherwise have released because he was unable to find a sponsor. Id. at 2822:5–21 (Munguia). 28 These figures are based on the Court’s review of PX-334A. Plaintiffs quote slightly different figures, as they say the San Antonio office released 526 of 543 age-outs in this time period. Pls.’ Br. ¶ 269. Defendants do not dispute Plaintiffs’ figures. Defs.’ Resp. ¶ 269. The small differences between the two sets of figures make for a negligible 0.02% change in overall detention rate. 104 217. Officer Munguia frequently accesses the ORR UAC data portal and compares the UAC information there with ICE’s own records in order to maintain an up-to-date list of UACs in ORR custody whose eighteenth birthdays are approaching. Id. at 2809:23–2810:22 (Munguia). He then researches each upcoming age-out and speaks with their ORR case managers to learn more about the UAC and his or her time in ORR custody, with the aim of determining whether that UAC is a flight risk, danger to self, or danger to the community. See id. at 2813:3–2814:6(Munguia). He does not view how long the UAC has been in ORR custody as a relevant consideration. Id. at 2832:25–2833:12 (Munguia). 218. FOJCs in the San Antonio office consider post-18 plans, either from ORR or from other advocates, that include contact information for proposed sponsors. See id. at 2764:12–20, 2811:4–12 (Munguia). These post-18 plans are typically received three to seven days in advance of the age-out’s eighteenth birthday. Id. at 2764:21–25 (Munguia). 219. Officer Munguia is often able to place age-outs with sponsors even when he only receives a post-18 plan one or two days before the age-out’s eighteenth birthday. Id. at 2812:4– 7. Officer Munguia will reach out to ORR if he has not received a post-18 plan, and can usually get one either from ORR or from a legal services provider working with the age-out. Id. at 2812:8–21 (Munguia); see also id. at 2764:6–11 (Munguia) (describing the San Antonio field office’s “good working relationship” with ORR). Officer Munguia will sometimes also look at a UAC’s Form I-213, which ORR cannot access, and attempt to identify potential sponsors that he can give to ORR. Id. at 2814:7–24 (Munguia). When a proposed sponsor has been identified, Officer Munguia will verify their address and sometimes call the proposed sponsor. Id. at 2818:4–2819:4. 105 Defendants dispute the proposed finding “that Officer Munguia identifies sponsors.” Defs.’ Resp. ¶ 280. The Court finds that Officer Munguia did in fact review the Form I-213 for certain age-outs and that he passed information from that Form along to ORR. Officer Munguia testified consistently to this effect in his deposition and again at trial. See Munguia Dep. Desig. at 82:24–83:18; Trial Tr. at 2814:7–24 (Munguia). Defendants have identified no inconsistent testimony or alternative interpretation of Officer Munguia’s testimony. The Court found him to be a credible witness. (If Defendants only meant to suggest that passing names to ORR does not count as “identif[ying] sponsors,” the Court would find that semantic dispute to be immaterial.). 220. In cases where an individual sponsor cannot be identified, Officer Munguia has been able to place age-outs with organizational sponsors. Id. at 2820:16–20 (Munguia). He is familiar with people working at two adult shelters in the San Antonio AOR and speaks to them on the phone. Id. at 2820:16–2821:25 (Munguia). 221. Officer Munguia speaks with each age-out he processes. Trial Tr. at 2816:6–9 (Munguia). When a proposed sponsor has been identified, Officer Munguia speaks with an ageout to confirm that the age-out is comfortable living with the proposed sponsor. Id. at 2817:9–18 (Munguia). If a sponsor has not yet been identified, he will ask the age-out if he or she knows of any possibilities. Id. at 2817:19–24 (Munguia). 222. Nearly all the age-outs processed by the San Antonio office are placed with sponsors and released on their own recognizance. Id. at 2804:12–2805:2. 223. Other FOJCs in the San Antonio field office follow the same general practices described by Officer Munguia. Id. at 2820:7–15 (Munguia). 106 224. No evidence has been presented suggesting that the liberal release procedures followed in San Antonio result in more age-outs absconding than for those offices utilizing more detention-focused procedures. c. Extreme Shifts in Detention Rate i. Harlingen 225. Between April 2016 and May 24, 2019, the ICE sub-office in Harlingen, Texas processed more age-outs than any other. It detained 674 of 1563 age-outs it processed (43%). PX-334 at 11–42; PX-334A at 78–146. This moderate overall rate is somewhat misleading because it is the product of several fairly dramatic swings in the detention rate. Between April 2016 and October 15, 2016, the Harlingen office detained 88% of the age-outs it processed. PX334 at 11–17. From October 15, 2016 through February 25, 2017, 95% of age-outs were released. Id. at 13–17. Eighty-one percent were detained between February 26, 2017 and December 2, 2017; 94% were released between December 3 and December 14, 2017; then 91% were detained between December 15, 2017 and September 25, 2018. Id. at 17–29. Finally, between September 26, 2018 and May 24, 2019, 99% of age-outs were released. PX-334 at 29– 42; see also PX-334A at 78–146; Pls.’ Demo. Ex. 22A (illustrating these shifts in detention rate). 226. The Court heard testimony concerning the Harlingen office from Officer Stephen Venegas, who worked as the primary FOJC in Harlingen from October 2017 to October 2018, and as a backup FOJC from then until October 2019. Trial Tr. at 2859:2–17 (Venegas). 227. Officer Venegas’s testimony showed that the dramatic shift in detention rate was caused by new guidance given by the Harlingen SDDO. At his deposition, Officer Venegas was unable to provide an explanation for the shift in detention rate. See Venegas Dep. Desig. at 144:20–145:25. At trial he testified that in the fall of 2018, Harlingen’s SDDO, José Cortez, instructed his FOJCs to look at flight risk less “stringently” than they previously had, and to 107 “broaden the spectrum on how [they] looked at flight risk” by “tak[ing] into consideration . . . were they really a flight risk.” Trial Tr. at 2881:6–2882:5 (Venegas). New guidance from JFRMU may also have played a role in the change in detention rate at the Harlingen office. The Harlingen office began releasing nearly all the age-outs it processed starting on September 25, 2018. See PX-334 at 29. That same day, Chief Harper held a call with FOJCs and SDDOs on which she discussed age-outs. See PX-61 at 1 (email from Chief Harper addressed to “FOJC SDDOs” and referencing “the call today”). She sent an email soliciting information from the field offices concerning age-out custody determination policies, guidance, and training, explaining that “we just need to get an overview of any independent guidance or trainings that may have been done in the field.” Id. at 1–2. The email also stated that she “w[ould] be discussing additional documentation that will be needed for age-outs” on the call as well. Id. at 1. This was less than a month before the AORW and SharePoint system was introduced on October 17, 2018, see PX-1, but was subsequent to this Court’s issuance of its orders granting a preliminary injunction and shortly after the Court first certified a class, P.I. Op. (issued April 18, 2018); MTD and Class Cert. Op. (issued August 30, 2018). Given the timing and content of the email and call, the Court thinks it more likely than not that guidance from JFRMU played a role in shifting the outcomes of age-out custody determinations in Harlingen. 228. At his deposition in March of 2019, Officer Venegas testified regarding his process for making age-out custody determinations. This was after Harlingen’s shift to detaining almost no age-outs and Officer Venegas only testified to his then-current practices. His testimony at trial was broadly consistent. 229. Officer Venegas testified at his deposition that Harlingen FOJCs would log onto ICE’s EARM database on a near-daily basis to identify UACs approaching their eighteenth 108 birthdays. Venegas Dep. Desig. at 11:15–12:11. At trial he testified that he only did this rarely. Trial Tr. at 2924:21–2925:12 (Venegas). Considering that the deposition was taken while Officer Venegas was still working as an FOJC while the trial was held two months after he stopped working as one, the Court credits his earlier testimony. Defendants dispute Plaintiffs’ proposed finding that the EARM tool could “identify upcoming Age-Outs,” stating that Officer Venegas needed to look to post-18 plans to know whether “a UAC was really aging out.” Defs.’ Resp. ¶ 301; see id. ¶ 300. This is a semantic distinction worth clarifying. Officer Venegas testified that, with the EARM tool, FOJCs “could see who was going to turn 18, and -- and we could get that information ahead of time.” Trial Tr. at 2873:10–11. The Court takes Defendants’ point that a UAC approaching her eighteenth birthday will not necessarily become an age-out, but Officer Venegas accounted for this possibility, noting that “sometimes they would even be released by ORR before they even aged out.” Id. at 2873:18–19. The EARM tool could not tell the FOJCs whether this would happen, but it could help them identify likely or potential upcoming age-outs—those UACs approaching their eighteenth birthdays. Officer Venegas testified that he would keep track of these upcoming potential age-outs, but would not take much action on them until they received post-18 plans from ORR. Id. at 2873:15–21. 230. Harlingen FOJCs view their responsibilities with age-outs as beginning a day or two before the age-out’s eighteenth birthday and ending on that day. Venegas Dep. Desig. at 14:17– 22. They typically only have one to two days’ advance notice of the age-out, and would receive post-18 plans at that time. Id. at 11:1–7; Trial Tr. at 2920:25–2921:2 (Venegas). 231. Harlingen FOJCs investigated and attempted to identify potential sponsors and would seek out organizational sponsor placements if no individual sponsor was available. They would 109 receive post-18 plans, either from ORR or from legal services providers, in most cases and would investigate proposed sponsors. Venegas Dep. Desig. at 17:6–7 (testifying he receives post-18 plans “on most occasions”); id. at 17:24–18:18 (testifying he would call proposed sponsors, run background checks, verify addresses); id. at 65:6–20 (discussing post-18 plans from legal services provider); id. at 66:7–16 (testifying that the process is the same whether the recommendation comes from the legal services provider or from ORR); see also Trial Tr. at 2874:1–15 (Venegas) (describing his phone calls to potential sponsors). Officer Venegas would speak with the age-out to determine if they had any potential sponsors and, if necessary, would review the age-out’s Form I-213 to see if any potential sponsors were mentioned there. Venegas Dep. Desig. at 81:16–82:25, 89:23–90:5. If no individual sponsor could be found, he would reach out to La Posada, a shelter, to see if they would have space for the age-out. Id. at 18:22– 25; 19:7–12. He would release the age-out on his or her own recognizance to La Posada if they had space, which they always did. Trial Tr. at 2876:14–17, 2928:1–3. 232. Office Venegas identified a sponsor for and release every age-out he processed between October 17, 2018 and May 24, 2019, having determined that they were not flight risks, dangers to the community, or dangers to themselves. Id. at 2930:24–2931:20. During the same time period, the Harlingen office as a whole released 603 age-outs and detained only four. PX334A at 125–46. Again, no evidence was presented that these release-favoring procedures affected the rate at which age-outs later absconded. ii. Chicago 233. Between April 2016 and May 24, 2019, the ICE field office in Chicago detained 128 of 201 age-outs it processed (64%). PX-334 at 1–5; PX-334A at 1–38. Again, this is the product of shifts in the average detention rate. From April 2016 through August 2018, the Chicago office detained 90% of the age-outs it processed. PX-334 at 1–3; PX-334A at 1–29. From September 110 2018 through May 24, 2019, just thirteen of the seventy-four age-outs processed by the Chicago office (17.5%) were detained and 82.5% were released. PX-334 at 3–5; PX 334A at 30–38; see also Pls.’ Demo. Ex. 23A (illustrating this shift in detention rate). 234. The Court heard testimony concerning the Chicago office from Officer Geoffrey Pepple, who has worked as an FOJC in Chicago since 2012. Trial Tr. at 3616:5–7 (Pepple). 235. Shifts in the Chicago field office’s detention rate have been attributable to changes in office leadership and to this litigation. See infra ¶¶ 236–237. 236. Officer Pepple testified that around the time Section 1232(c)(2)(B) was enacted, in 2013, his office released most age-outs. Id. at 3621:10–16. In 2015, a new SDDO “started emphasizing” that age-outs were flight risks and “she wanted them detained when they turned 18.” Id. at 3658:4–23. This shift would have predated the period for which ICE produced data, but is consistent with the high detention rates seen in 2016 at the start of the data. See, e.g., Pls.’ Demo Ex. 23A. 237. In the summer of 2018—around when the Chicago office began releasing more of its age-outs—the office “change[d] that emphasis” on flight risk and detention of age-outs. Trial Tr. at 3658:24–3659:2 (Pepple). Officer Pepple candidly attributed the change in attitude primarily to this lawsuit, Garcia-Ramirez v. ICE, and also to the fact that “[n]ew supervisors have different opinions about how we handle cases.” Id. at 3659:3–7; see also Pepple Dep. Desig. at 203:19– 20, 22–23 (“Q: [W]hat was the reason for that change in your opinion? . . . A: I think it was primarily motivated by your lawsuit.”). 238. Officer Pepple’s practice with regard to age-out processing changed in the last few months of 2018. He began taking more active steps to obtain names of potential sponsors from ORR and ORR contractors, and began receiving post-18 plans more often as a result. Pepple 111 Dep. Desig. at 191:5–192:8; see also Trial Tr. at 3632:19–3633:9 (Pepple). He had not previously reached out to a charitable organization about sponsoring an age-out until he did for the first time at some point between October and December 2018. Pepple Dep. Desig. at 17:1–6, 44:21–45:4. 239. By the time he testified at trial in December 2019, Officer Pepple was using ICE’s EARM module and ORR’s UAC portal to keep track of UACs in ORR custody who were approaching their eighteenth birthdays. Trial Tr. at 3629:7–16, 3630:25–3631:23. He typically reaches out to an age-out’s ORR shelter ten days before the UAC’s birthday to find out information about the UAC and about any potential sponsors, and ORR typically responds. Id. at 3628:14–3629:6, 3629:17–23, 3632:17–3633:9. If no individual sponsor is available, Officer Pepple asks ORR to look into organizational sponsors. Id. at 3634:7–21. On two occasions Officer Pepple has had to do this himself by contacting a charitable organization because ORR did not find one. Id. at 3634:22–3635:20. 240. The Chicago office released the vast majority of its age outs without using detention alternatives other than release to an individual or organizational sponsor. Officer Pepple has never set an ICE bond for an age-out and was not aware of anyone else in his office having done so, though he was aware that it was an option. Pepple Dep. Desig. at 45:13–23, 47:2–5, 49:2–5. He has also never recommended or used the ATD program with respect to an age-out though, again, he was familiar with it as an option. Id. at 22:13–23:10. He has considered ankle monitoring for age-outs but decided against it because it is more labor intensive for the FOJC to set up. Id. at 173:2–174:10. He believes other officers in his unit feel the same way and do not use ATD with age-outs for the same reason. Id. at 174:9–16. Again, no evidence has been 112 presented that Chicago’s more recent release-favoring procedures have negatively impacted ageout rates of absconding. iii. El Paso 241. Between April 2016 and May 24, 2019, the ICE field office in El Paso detained 123 of 218 age-outs it processed (56%). PX-334 at 6–10; PX-334A at 39–77. The office’s detention rate changed dramatically in early 2017. From April 2016 through January 2017, only four of eighty-seven processed age-outs were detained (5 %). PX-334 at 6–7; PX-334A at 39–48. Between February 2017 and May 2019, however, the rate of detention was higher, with 119 of 131 age-outs detained (91%). PX-334 at 7–10; PX-334A at 49–77. Between October 17, 2018 and May 24, 2019, the El Paso office detained 84% of age-outs it processed. PX-334 at 2–10; PX-334A at 68–77; see also Pls.’ Demo. Ex. 26A. 242. Officer Victor Galvez is one of six FOJCs in the El Paso office. Galvez Dep. Desig. at 7:19–8:15. His deposition was taken and designations were read into the record at trial but he did not testify at trial. The Court also heard testimony concerning the El Paso office from Danielle Escontrias, an immigration attorney who works for Catholic Charities in New Mexico and who previously worked for Diocesan Migrant & Refugee Services in El Paso. Trial Tr. at 1494:24–1495:11 (Escontrias). 243. The El Paso office has had three different SDDOs during the time period for which ICE produced data in this lawsuit: Joel Beard, Hector Juarez, and Danielle Hernandez, the current SDDO. See Galvez Dep. Desig. at 11:17–12:7, 12:14–13:13, 107:19–108:1; Trial Tr. at 1499:8–20, 1514:22–1515:1 (Escontrias). 244. The detention rate increased sharply over Officer Juarez’s first three or four months as SDDO, between December 2016 and March 2017. See Pls.’ Demo. 26A (charting El Paso’s monthly detention rate). See Galvez Dep. Desig. at 306:17–307:8, 307:14–19, 308:1–309:3 113 (discussing the shift in rate at Officer Galvez’s February 2019 deposition, but quoting specific figures that were superseded by subsequent productions of and corrections to ICE’s data); see also supra ¶ 241 (summarizing the shift in detention rate visible in the final version of the data). While the El Paso office had released nearly 100% of its age outs through November 2016, it was detaining nearly 100% by the end of February 2017. Id. Plaintiffs suggest that Officer Beard was replaced by Officer Juarez in January 2017. Pls.’ Br. FF ¶ 428. Defendants do not challenge this, Defs.’ Resp. ¶ 428, but the Court does not see where the testimony establishes a changeover date with this specificity. Officer Galvez testified he began working with the juvenile unit in 2016, possibly in October of that year, and that he does not recall having worked under Officer Beard. Galvez Dep. Desig. at 106:1–16, 107:19–108:1. The shift in age-out detention rate began in December of 2016 and escalated in January and February of 2017. See Trial Tr. at 1505:21–1506:17 (Escontrias) (describing a shift in policy from the El Paso office); see also Pls.’ Demo. 26A. Whether Officer Galvez simply forgets that he worked briefly under Officer Beard, or whether Officer Juarez began as El Paso SDDO slightly earlier than Plaintiffs believe is immaterial. Either way, the detention rate increased sharply over Officer Juarez’s first three or four months as SDDO. 245. Officer Galvez could not recall a single age-out who was not detained while Officer Juarez was SDDO. Galvez Dep. Desig. at 116:11–117:10; see also Trial Tr. at 1506:13–23 (Escontrias) (testifying that nearly 100% of the age-outs she worked with were detained under Officer Juarez, including those with serious health problems). 246. At some time before the end of October 2018, Officer Hernandez became SDDO. See PX-335.27398–.27400 (AORW from late October 2018 approved by Officer Hernandez as SDDO). Around this same time, the detention rate in El Paso began to drop somewhat. From 114 February 2017 through September 2018 the office had detained 95.8% of all age-outs it processed. PX-334 at 7–9; PX-334A at 49–67. Between October 2018 and the close of discovery on May 24, 2019, 84.4% of processed age outs were detained. PX-334 at 9–10; PX334A at 68–77. While not as dramatic as the increase in the detention rate when Officer Juarez became SDDO, these figures are consistent with Ms. Escontrias’s testimony that once Officer Hernandez took over as SDDO, her organization had somewhat more success in getting age-outs released. See Trial Tr. at 1514:21–1515:16 (Escontrias). In raw numbers the shift is more obvious: between February 2017 and September 2018 the El Paso office processed seventy age outs and released only three, while between October 2018 and May 2019, it processed only forty-nine but released nine—three times as many. PX-334 at 7–10; PX-334A at 49–77. 247. The Court finds that the shifts in the age-out detention rate in the El Paso field office thus coincided with changes in the position of SDDO. Defendants “dispute that any changes were caused by policy changes that accompanied turnover in the SDDOs.” Defs.’ Resp. ¶¶ 428, 448, 450, 452–53. The Court disagrees. Defendants cite Officer Galvez’s testimony denying that there have been any policy changes. See Galvez Dep. Desig. at 118:3–8 (“Q: Do you recall any change that occurred starting after August of 2018 with respect to the El Paso Juvenile Office’s decisions whether to detain age outs? A: No, there was no change.”); see also id. at 98:18–99:17. The testimony Defendants cite could only relate to changes between Officer Juarez and Officer Hernandez, because of the timeframe indicated by counsel’s question, and because officer Galvez testified that he did not remember working under Officer Beard. So he would not be in a position to say what policy changes officer Juarez put in place when he came in. Further, there was testimony and evidence suggesting that FOJCs in the El Paso office do not necessarily see the post-18 plans that are provided by ORR or by attorneys like Ms. Escontrias. 115 See Galvez Dep. Desig. at 29:1–22 (testifying that post-18 plans “go straight to my supervisor” and that FOJCs “don’t get to see them,” unless they are copied on an email, but that generally the materials are physically dropped off and go straight to the SDDO); see also PX-466 (email from Galvez to a FOJC in another field office in which he writes “we don’t get Post-18s”). So if there were a policy change regarding what kinds of post-18 plans are accepted or considered by the El Paso office—the kind of change Ms. Escontrias’s testimony suggests—it is not at all clear that Officer Galvez would have been in a position to know. Defendants also argue that Ms. Escontrias, who did not work in the field office, was not in a position to know what policy changes were taking place. Even if Ms. Escontrias did not know precisely what policy changes were taking place, she nonetheless testified credibly to certain policy changes implemented by Officers Juarez and Hernandez that she could see from the outside because they related to what materials immigration attorneys could submit to the El Paso office. See Trial Tr. at 1505:21–1506:17 (Escontrias). It is reasonable to assume that if the SDDOs changed outward facing policy, they likely also had changed internal procedures, and the shifts in detention rate are strongly correlated with changes in the SDDO position. Defendants did not present testimony from any witnesses who did work in the El Paso office to refute Ms. Escontrias’s credible testimony. 248. Officer Galvez views most age outs as flight risks and dangers to themselves because of their lack of community ties. Officer Galvez agreed that age-outs lack community ties in most cases, and treats lack of a sponsor as indicating a lack of community ties. Galvez Dep. Desig. at 225:11–17, 226:5–227:17. He views age-outs who lack community ties as flight risks. Id. at 241:12–13, 243:16–244:15. He also views them as dangers to themselves if they do not have a place to go. Id. at 299:19–300:7 (“[W]hen you release a kid out on the street by themselves I 116 would say yes [danger to self] because they are in a new land, different language. So if you let them out by themselves, they are a danger to themselves, because red light/green light, if they cross, they could get run over . . . .”); see also id. at 300:19–301:3. 249. Similar to the practice in several other offices, FOJCs in the El Paso field office take little to no action to attempt to identify or vet sponsors for age-outs and leaves this work entirely to ORR. Officer Galvez was not aware of anyone in the Juvenile Unit there ever arranging for or identifying a sponsor for an age out. Galvez Dep. Desig. at 66:15–22. In his experience only an ORR case manager would identify or arrange for a sponsor. Id. at 66:10–22, 68:2–6. Officer Galvez was not aware of any organizational sponsors or group homes—aside from the fact that he recognized the name of one. Id. at 70:1–71:18. Between the institution of the AORW system in October 17, 2018 and the end of January 2019 (two weeks before Officer Galvez’s deposition), for all fourteen age-outs detained by the El Paso office, the explanation on the AORW for the detention was “no viable sponsor.” Id. at 287:13–18, 295:2–22. Officer Galvez summarized his process for age-outs as being that he will “talk to the lead case manager, see if they have any sponsors that [they] will possibly release him [to] before his birthday. And then they will tell [Galvez] yes or no.” Id. at 41:4–7. “If they [are] not released by ORR,” Galvez testified, “they will be adults that will be going to [ICE] custody,” which means “they are taken directly from the shelter to [adult detention].” Id. at 42:1–6. 250. FOJCs in the El Paso field office do not receive post-18 plans. While Officer Beard allowed legal services providers like Diocesan Migrant & Refugee Services to email post-18 plans, Officer Juarez changed the office’s procedures and began requiring attorneys to deliver in person hard copies of the materials they had previously sent via email. Trial Tr. at 1506:4–12 117 (Escontrias). Officer Hernandez continued the policy of requiring hard copy materials. Id. at 1514:25–1515:2 (Escontrias). 251. The El Paso office does not make detention alternatives available for age-outs. Officer Galvez has never released an age-out on an ICE bond or an ankle bracelet monitor and he was unaware of anyone else in his office having released an age-out using either alternative. Galvez Dep. Desig. at 80:15–81:10, 85:22–86:13, 87:6–12. iv. San Francisco 252. The San Francisco office processed a smaller number of age-outs in comparison to some of ICE’s other field offices. Between October 17, 2018 and May 24, 2019 it processed sixteen age-outs and detained only four of them (25%). 29 This was a significant shift from the time period for which data was produced before the AORW system was put in place, as the San Francisco office detained 59 of 77 age-outs it processed (77%) between April 2016 and October 17, 2018. 30 PX-334 at 88–89; PX-334A at 394–400. Overall, then, between April 2016 and May 24, 2019, the ICE field office in San Francisco detained 63 of 93 age-outs it processed (68%). 29 The following sixteen AORWs came from the San Francisco field office: PX– 335.00007, PX–335.01955, PX–335.02270, PX–335.03355, PX–335.10458, PX–335.11798, PX–335.12040, PX–335.14570, PX–335.17801, PX–335.18631, PX–335.22692, PX–335.23359, PX–335.25794, PX–335.29740, PX–335.33224, PX–335.34258. Plaintiffs cite these AORWs directly to establish the 25% detention rate during this time period and Defendants do not dispute it. Defs.’ Resp. ¶ 334. Oddly, the Defendants’ summary charts, PX-334 and PX-334A seem to misstate the contents of these sixteen worksheets, as they show that the San Francisco office detained ten of the sixteen age-outs it processed between October 17, 2018 and May 2019. See PX-334 at 88–89; PX-334A at 394–400. The Court has reviewed the underlying AORWs and confirmed that the figures quoted in the briefing are correct. 30 Based on the Court’s comparison of unredacted selections from PX-334 with ICE’s raw data in PX-345, it does not appear that there are any problems with the San Francisco data summarized in PX-334 and PX-334A from the time period before the AORWs were used. 118 253. Officer Joe Love was the primary FOJC in the San Francisco field office from October 2016 through September 2018. Love Dep. Desig. at 5:19–25. Officer Andrew Kaskanlian replaced him in that role and was the primary—and only—FOJC from October 2018 through August 2019, at which time he was promoted to SDDO. Trial Tr. at 2671:25–2672:6 (Kaskanlian); see also id. at 2673:17–18. Officer Kaskanlian testified at trial and the parties introduced designated testimony from Officer Love’s deposition, which was conducted in October of 2018. 254. The San Francisco office’s shift in September and October of 2018 from detaining most age-outs to releasing most of them thus coincides both with the introduction of the AORW and SharePoint reporting system, see PX-1, and with the changeover from Officer Love to Officer Kaskanlian, supra ¶ 253. 255. The San Francisco field office processes age-outs from one of the highest-security ORR shelters in the country, Yolo County Juvenile Detention Facility. Trial Tr. at 2680:16–24 (Kaskanlian). “[M]inors with more egregious backgrounds are housed” by ORR in this type of facility. Id. at 2680:22 (Kaskanlian). The four age-outs that the San Francisco office detained under Officer Kaskanlian all came from a medium- or high-security ICE shelter. See PX– 335.00007; PX–335.03355; PX–335.22692; PX–335.34258. The Court appreciates that under Officer Love, the San Francisco office may also have processed a disproportionate number of age-outs with “egregious backgrounds” relative to other field offices, but Defendants have not argued that this had any significant impact on the office’s detention rate. 256. If an age-out has no criminal history and has complied with all immigration court obligations, and is not a flight risk, the San Francisco office will typically release them on their 119 own recognizance, either to an individual sponsor or to a shelter or group home. See Trial Tr. at 2743:18–2744:24 (Kaskanlian). 257. ICE officers in San Francisco relied on ORR or on legal services providers to arrange for placements with group home or shelters, they do not arrange for these placements themselves. Love Dep. Desig. at 65:14–66:4, 67:18–20; Trial Tr. at 2743:25–2744:12 (Kaskanlian); see also id. at 2681:2–8, 2738:22–2739:3 (Kaskanlian). 258. Officers in the San Francisco field office use the full range of detention alternatives available to them. Officer Love offered age-outs ICE bonds. Love Dep. Desig. at 18:11–12. He testified that he would first consider whether the age-out was a flight risk or danger to himself or the community, and would generally not offer ICE bonds if the age-out was a danger to himself or others. See id. at 19:3–14. Officer Kaskanlian has both given ICE Bonds and placed age-outs in the ATD program. Trial Tr. at 2687:7–21 (Kaskanlian). d. Intermediate Detention Rates 259. The Court heard testimony concerning four additional ICE field offices for which the age-out detention rate over time did not stand out as particularly high, particularly low, or particularly indicative of fluctuation: Phoenix, Seattle, Los Angeles, and Washington D.C. The Court has relied on the testimony of officers from these offices in cataloging the general training and performance of ICE officers in the field. The custody determination outcomes in these offices are material to the extent that they were incorporated into the experts’ quantitative analyses of ICE’s decisionmaking, discussed below. At the same time, the particular procedures of field officers in these four offices warrant less discussion than those procedures in other offices that led to more uniform results. Accordingly, the Court discusses these four offices only relatively briefly here. 120 i. Phoenix 260. The age-out detention rate in ICE’s Phoenix office fluctuated between April 2016 and May 2019. See Defs.’ Resp. ¶ 521 (not disputing this). Plaintiffs do not suggest precise figures for the age-out detention rate in the Phoenix office over time, possibly because of complications resulting from mid-trial corrections to the Defendants’ Phoenix data. See Defs.’ Resp. in Opp’n to Pls.’ Mem. in Supp. of Admitting their Demonstratives at 3–4, ECF No. 265; see also Pls.’ Mem. in Supp. of Admitting their Demonstratives at 7–9, ECF No. 262. 31 Because the parties do not meaningfully rely on or emphasize the average rate of detention in the Phoenix office, 32 neither does the Court. 261. Officer Nancy Zanello (formerly Nancy Sullivan) was the primary FOJC in the Phoenix office from October 2016 through January 2019. Trial Tr. at 3772:5–12 (Zanello). She testified at trial and was also deposed in February 2019. See Zanello Dep. Desig. 262. In the Phoenix AOR, ORR notifies FOJCs about one day before a UAC would turn eighteen. Trial Tr. at 3834:1–3. Age-outs are picked up and brought to the Phoenix field office on the morning of their eighteenth birthdays. Trial Tr. at 3834:4–11. The office’s policy is that age-outs cannot remain at the office for more than twelve hours, and if they have not been 31 Plaintiffs’ demonstratives 12A, 13A, and 68 were admitted as Rule 1006 summaries of voluminous records. See Fed. R. Evid. 1006. The Court understands these charts to reflect the final version of the Phoenix data, but the Court is unsure whether the most up-to-date Phoenix data has itself been submitted into evidence. The summary information in PX-334A and PX-334 does not appear consistent with the demonstratives, and the raw form Phoenix data in PX-345 and the AORWs in PX-335 are too voluminous for the Court to reconstruct the data from them without clearer guidance from the parties. 32 At two points in their briefing Plaintiffs mention changes in detention rates at the Phoenix office and cite generally to PX-505, a May 9, 2019 email from Officer Helland attaching raw age-out documentation data dating back to April 2016. See Pls.’ Br. ¶¶ 525–26 (citing PX-505). Considering the scope of the data, which covers hundreds of age outs over several years, these general citations to PX-505 with no additional explanation or page reference are not sufficiently specific to support any claims about the Phoenix office’s detention practices. 121 released to a sponsor by that point, they are taken to adult detention. Id. at 3834:12– 3835:12;Zanello Dep. Desig. at 57:24–58:2. Officer Zanello clarified at trial that transfer to adult detention at the end of the twelve-hour period was not necessarily indefinite—the age-out could return to the field office thereafter and FOJCs would continue working on arranging for a sponsor or for transportation. See Trial Tr. at 3834:15–3835:5. Nonetheless, they are sent to adult detention at least temporarily at the end of the twelve-hour period. Id.; see also Defs.’ Resp. ¶ 529 (not disputing this). 263. FOJCs in the Phoenix office make limited use of organizational sponsors and do not attempt to identify them on their own. Officer Zanello was only aware of one possible organizational sponsor in the Phoenix area, Casa Alitas. Trial Tr. at 3864:1–3, 3866:5–7. She had released age-outs there, but only when a local immigration non-profit affirmatively suggested it. Id. at 3865:11–19. She does not consider placing age-outs in out-of-state shelters. Id. at 3867:14–23. 264. FOJCs in the Phoenix office do not consider the full range of detention alternatives available to them. Officer Zanello has never released an age-out on an ICE bond and was not aware whether anyone in the Phoenix office had done so. Id. at 3873:12–20. She had never placed an age-out in the ATD program. Id. at 3873:6–8;Zanello Dep. Desig. at 117:10–12. She believed that ATD was only an option for families and that it could not be used for age-outs. Zanello Dep. Desig. at 243:18–244:14. ii. Seattle 265. Between April 2016 and May 24, 2019, the ICE field office in Seattle detained 38 of 79 age-outs it processed (48%). PX-334 at 90–91; PX-334A at 401–31. The Seattle office processed only eleven age-outs between October 17, 2018 and the close of discovery on May 24, 2019, and detained seven of them. PX-334 at 91; PX-334A at 426–31. 122 266. Officer Lika Black has been an FOJC in the Seattle field office since December of 2017. Trial Tr. at 2499:19–20 (Black). As of her testimony at trial she was detailed to the JFRMU unit in Washington, D.C. Id. at 2499:11–18 267. Officer Black uses data available in the UAC portal to keep track of upcoming likely age-outs and reaches out to UACs’ ORR shelters around one month in advance of their eighteenth birthday to ask whether they anticipate submitting a post-18 plan. Id. at 2518:1–20 268. The Seattle field office does not attempt to identify or arrange for placement with individual or organizational sponsors, and only releases age-outs to sponsors when ORR has proposed one in a post-18 plan. Id. at 2604:12–2605:6 Officer Black does not release age-outs other than to a placement suggested by ORR or an ORR contractor in a post-18 plan. Id. Her age-out custody determination decisions are therefore “more or less” a binary choice between the proposal in the post-18 plan or adult detention. See id. at 2605:15–22 269. FOJCs in the Seattle office do not consider the full range of detention alternatives available to them. Officer Black did not know what “Bond” referred to on the AORW at her deposition. Black Dep. Desig. at 46:17–19. She had never issued an ICE bond, did not know of anyone else in the Seattle office having done so, and did not know when it would be appropriate to do so. Id. at 20–24; Trial Tr. at 2607:4–11. At her deposition, she was unaware that she had the option of placing an age-out into ICE’s ATD program and was not aware of anyone else having done so. Black Dep. Desig. at 47:16–25, 180:23–182:12. By the time she testified at trial in December 2019, Officer Black had since released one age-out on ATD. Trial Tr. at 2607:24– 2608:1. iii. Los Angeles 270. Between April 2016 and May 24, 2019, the ICE field office in Los Angeles detained 60 of 111 age-outs it processed (54%). PX-334 at 54–56; PX-334A at 185–217. Between 123 October 17, 2018 and May 24, 2019 the Los Angeles office detained 66% of age-outs it processed. PX-334 at 55–56; PX-334A at 210–17; see also Pls.’ Demo. Ex. 12A. 271. Officer Oscar Irizarry, who has been an FOJC in the Los Angeles field office since 2013, testified at trial. See Trial Tr. at 3667:20–25. 272. Officer Irizarry recommended detaining eight of the eleven age-outs he processed between October 17, 2018 and May 24, 2019, while the AORW and SharePoint documentation system was in place. Id. at 3731:16–19 (Irizzary). The three age-outs he recommended releasing were all eighteen-year-olds with children of their own. Id. at 3731:20–22. 273. Officer Irizarry has not granted an ICE bond for an age-out because, he said, “[m]ost of the time they don’t have the money to pay” the $1,500 minimum. Id. at 3681:10–22. He did not explain why, even if true, they could not be given the opportunity to try to arrange for payment of the bond rather than face adult detention. He also had not released an age-out by using ICE’s ATD program, and he was not aware of anyone else in his office having done so. Id. at 3715:1–20. iv. Washington, D.C. 274. Between April 2016 and May 24, 2019, the ICE field office in the District of Columbia detained 76 of 156 age-outs it processed (49%). PX-334 at 92–95; PX-334A at 432– 69. 33 From October 17, 2018 through May 24, 2019, the D.C. office processed twenty-four ageouts and detained sixteen of them (66.7%). PX-334 at 95; PX 334A at 450–69. 33 Citing the same pages of PX-334A, Plaintiffs suggest that the Washington office detained only 41% of its age-outs during this time period. Pls.’ Br. ¶ 564. Defendants do not directly challenge Plaintiffs’ math, see Defs.’ Resp. ¶ 564 (disputing the proposed finding on other grounds), but, upon the Court’s calculation, the detention rate comes out to 48.7%. 124 275. Officer Rebecca Jones has been FOJC in the Washington, D.C. field office since June of 2018. Trial Tr. at 4241:22–23 (Jones). She did not attend the April 2019 JFRMU training in San Antonio. Id. at 4245:7–8. 276. Between April 2016 and August 2018, two of the three ORR facilities from which the Washington field office received age-outs for processing were high security ORR shelters. Id. at 4265:15–4267:9 (Jones). Age-outs placed at these kinds of facilities were more likely to have criminal history, to have made escape attempts, or to have behavioral issues. Id. at 873:5–16 (Enriquez); id. at 2715:23–2716:23 (Kaskanlian); see also id. at 4265:11–13 (Jones) (colloquy with the Court in which Officer Jones suggested that age-outs coming from higher-security facilities were more likely to present reasons why detention might be appropriate). 277. Washington, D.C. FOJCs reach out to ORR facilities in advance of a UAC’s eighteenth birthday to request a post-18 plan. Id. at 4242:20–4243:3 (Jones). They also consider sponsors proposed by attorneys or by age-outs themselves. Jones Dep. Desig. at 33:16–34:15. 278. The Washington D.C. FOJCs do not maintain a list of potential shelters or group homes. Id. at 42:17–20. 279. The Washington D.C. field office does not place age-outs on ICE bonds. See Jones Dep. Desig. at 227:18–228:3. 280. Officer Jones has placed age-outs in ICE’s ATD program as a means of mitigating flight risk concerns. Trial Tr. at 4244:23–4245:2, 4272:3–21. 3. Expert Statistical Testimony 281. Plaintiffs presented expert testimony from economist Dr. Justin Lenzo. Dr. Lenzo holds a Ph.D. in economics, has taught courses in business analytics, and has worked as a consultant since 2014 focusing on empirical analysis. See Trial Tr. at 1891:21–1893:23 (Lenzo). 125 The Court found him to be well-qualified to discuss and explain statistical methods and the forms of regression analysis that he performed here, and found him to be a credible witness. 282. Defendants presented rebuttal expert testimony from Dr. Qin Pan. 34 Dr. Pan holds a Master’s degree in statistics and a Ph.D. in biostatistics, and is an associate professor of statistics at George Washington University. Id. at 3409:6–22. The Court found her to be well-qualified to discuss statistical methods and also found her testimony to be credible. 283. Dr. Lenzo and his team created a database in which they collected the contents of AORWs and SharePoint material produced by ICE (“age-out documentation”). They reviewed materials relating to all age-outs processed between October 17, 2018 and May 24, 2019— approximately the first seven months during which ICE was using the AORW and SharePoint documentation system. Id. at 1926:24–1957:10. During this time period, ICE processed 1,005 age-outs, and detained 285 of these. Id. (citing Pls.’ Demo. Ex. 7). To collect data from this varied set of documents, they developed a form, PX-393, that could be filled out by reviewers in order to provide a consistent structure to all the information documented by field officers making detention decisions. See Trial Tr. (Lenzo) at 1897:21– 1899:10. The form, and the resulting database, focused on five primary questions: (1) was there “[a]ny evidence that [the] age-out was a danger to the community”; (2) was there “[a]ny 34 Defendants also presented, and the Court has reviewed, designated testimony from the deposition of Dr. Joseph L. Gastwirth, Dr. Pan’s coauthor on her rebuttal to Dr. Lenzo’s expert report. See generally Gastwirth Dep. Desig. While Dr. Gastwirth appears to be equally qualified, the Court did not find his deposition testimony added anything of substance to the testimony Dr. Pan provided in Court. In particular, the materiality of Dr. Gastwirth’s testimony was lessened by the fact that Dr. Gastwirth never reviewed the full scope of produced AORWs and SharePoint materials that Dr. Lenzo did. See Trial Tr. at 3606:2–3608:13 (Pan). Dr. Pan likewise had not formally updated the co-authored rebuttal report, see id. at 3608:5–111311, but she testified after hearing much of Dr. Lenzo’s testimony at trial. Her testimony is therefore a more reliable source for rebuttal arguments focused on the latest version of Dr. Lenzo’s opinions. 126 evidence that [the] age-out was a danger to self”; (3) were any individual or organizational sponsors available; (4) was there evidence of “flight risk that might not be addressed by release to [a] sponsor”; and (5) were there any errors or omissions in the completed AORW. PX-393. For each question, sub-questions asked more detailed questions—primarily in yes-or-no format, and asked the reviewer to copy over additional information like remarks from the AORW or names of available sponsors. Id. The data that was collected regarding errors or omissions in the AORW was not used in Dr. Lenzo’s final analysis. Trial Tr. at 1910:22–1911:2. 284. In reviewing the age-out documentation in order to construct their database, Dr. Lenzo’s reviewers indicated any evidence that might indicate danger to the community, danger to self, or risk of flight, but did not assess whether such a danger or risk in fact existed or evaluate the degree of that danger or risk. Id. at 1899:15–22 (Lenzo); see also, e.g., id. at 1902:20–1903:2 (“Q: And if the age-out had any criminal history at all did [reviewers] look deeper to find out whether it was a nonviolent misdemeanor and just a charge so it shouldn’t be counted as a danger to the community? A: No, we did not. If there was any indication[] of criminal history, we counted that as evidence of potential danger to the community. We did not further evaluate it for severity or anything else.”). Defendants do not dispute this description of the first stage of Dr. Lenzo’s analysis, but argue that “coding any evidence that might indicate danger to self, danger to the community, and risk of flight, regardless of severity, is a study design weakness” because “[i]n the real world, these variables are not binary.” Defs.’ Resp. ¶ 244 (citing Trial Tr. at 2011:14–2013:17 (Lenzo) (agreeing that these variables are not binary)); see also id. ¶ 252 (arguing Dr. Lenzo’s “oversimplified model . . . d[id] not capture the degree of each risk factor and the interactions between risk factors.”); Trial Tr. at 3467:6–10 (“[U]sing binary indicators for danger to 127 community, danger to self and flight risk or potential sponsor cannot capture the complexity of all possible scenarios that were considered by ICE officer in making the decisions.”). The Court does not see how this particular simplification would do any meaningful damage to the reliability of Dr. Lenzo’s analysis. His research question involved investigating the effect of “the presence or absence of the risk factors” identified in the statute on age-out custody decisions. Trial Tr. at 1934:14–18 (Lenzo). Presence or absence is necessarily binary, even if the underlying substance of the risk factor is not. Further, to the extent this necessary simplification distorts reality, it does so in a direction that is favorable to the Defendants. By coding any evidence that could suggest one of these risk factors equally, Dr. Lenzo, if anything, overvalued evidence of insignificant risks like misdemeanor charges that did not indicate a real danger to the community. Non-serious risk factors are grouped in with serious, obvious ones, so more age-outs are viewed as presenting the risk factors than actually would if they were considered on a case-by-case basis. Dr. Lenzo’s analysis essentially treats every borderline case as breaking in favor of there being a risk factor and from there, presumably, in favor of detention. His conclusion that the presence or absence of risk factors is not the strongest indicator of whether the age-out is detained is bolstered, not weakened, by the fact that he treated risk factors as binary and counted any minor indication of a risk factor as a positive on that binary assessment. This gives the officers the benefit of the doubt by essentially rounding up the degree of risk and treating any evidence of a risk factor (whether minor or serious) as an equally sufficient basis for a detention decision. While reducing the age-out documentation to quantifiable data necessarily required some simplification of the contents of the documents, the Court does not find that he oversimplified the data or that he simplified them in a way that weakens his conclusions. 128 285. Turning to the availability of sponsors, Dr. Lenzo’s reviewers indicated that an individual sponsor was available only so long as the age-out documentation lacked any evidence of negative or derogatory information about a potential sponsor—including criminal history or unwillingness to serve as a sponsor. Id. at 1904:16–1905:11. This likely resulted in at least some undercount of available sponsors, because potential sponsors could be found in Forms I-213, which were only created and available for review for about ten percent of the age-outs. Id. at 1905:12–1906:16; see also Defs.’ Resp. ¶ 246 (acknowledging that a Form I-213 was not always included in the SharePoint materials, but noting that information from a Form I-213 could have been duplicated elsewhere in the SharePoint materials that were available). 286. Lack of ORR approval was not treated as negative or derogatory information. If the age-out documentation contained evidence that a potential sponsor existed but that this potential sponsor had not been approved by ORR, Dr. Lenzo would code that sponsor as available. Id. at 1909:7–11. This makes sense because, as indicated before, ORR applies different criteria for the sponsors of children than what ICE applies for adults, including age-outs. Thus, the fact that ORR may have found a potential sponsor unqualified under ORR sponsorship criteria does not negate the viability of that individual as a sponsor under ICE’s criteria, to the extent any ICE criteria exist. 287. For organizational sponsors or group homes, Dr. Lenzo’s team first consulted a list of such potential sponsors prepared by counsel for Plaintiffs which indicated that shelter or group homes existed in all of ICE’s AORs except for Los Angeles and San Diego. Trial Tr. at 1906:22–1907:12 (Lenzo) (identifying PX-463 as a modified version of the list); see also id. at 1700:1–1701:9 (the Court accepting PX-463 into the record as a demonstrative, but not admitting 129 it as an exhibit). 35 For age-outs from every office except Los Angeles and San Diego, Dr. Lenzo essentially assumed an organizational sponsor or group home was available unless there was evidence to the contrary in the AORW. See id. at 1906:20–1908:7. That is to say, if there was any indication in the AORW or SharePoint material that an ICE officer had considered placing the age-out with an organizational sponsor, but that it was not a viable placement for some reason (bed space, amenability to taking age-outs), that organization would not have been considered an available sponsor. Id. at 1907:24–1908:7. However, because most field offices had more than one potential organizational sponsor or group home in their AOR, Dr. Lenzo would still have considered other organizational placements available for age-outs. See id. at 2024:7–11 (“I believe we did find instances where one of these organizational sponsors was considered and rejected. But there were other organizational sponsors available in the AOR, and so there was still a potential sponsor for that age-out.”). Dr. Lenzo did not recall there being more than a handful of age-outs for whom there was no organizational sponsor deemed available based on ICE’s having considered and rejected every available organizational sponsor or group home in the AOR. Id. at 2023:9–2024:24. 35 Defendants argue that “Plaintiffs elicited no evidence of an established organizational sponsor that operated a shelter in the Miami AOR that could accept age-outs.” Defs.’ Resp. ¶ 229 This is not accurate. The Court admitted PX-485 into evidence. See Trial Tr. at 2321:18–2322:16. This exhibit is a list of shelters created by Lisa Lehner, an attorney at Americans for Immigrant Justice, id. at 1281:12–13, 1318:1–1319:5, 1404:6–10. The list encompassed shelters at which her organization had placed UACs, and she had shared the list with ICE in an effort to get the Miami office to place age-outs at some of these shelters. Id. at 1318:1–13. Two of the shelters on the list were in the Miami AOR. Id. at 1318:16; see also PX485. While some shelters that accepted children may not have accepted eighteen-year-olds, Defendants did not cross-examine Ms. Lehner on that theory. And, regardless, the list included shelters from around the country, some of which did accept adults, so the Miami office had more reason than most to have realized that shelters outside their AOR would accept age-outs. 130 288. Defining sponsor availability in this way resulted in 268 of 285 detained age-outs (94.0%) being coded as having potential individual or organizational sponsors. Pls.’ Demo. Ex. 8. Of these, 153 were coded as having individual sponsors available, id., which means that 115 detained age-outs (40.4%) were coded as having a sponsor available based only on the availability of an organizational sponsor as Dr. Lenzo defined it. 289. Dr. Lenzo did not instruct his team to consider conclusory assertions on AORWs as constituting evidence. For example, Dr. Lenzo would not code a sponsor as unavailable if a potential sponsor was listed on the AORW or was identifiable in the accompanying materials but the AORW said “no viable sponsor” without indicating why the sponsor was not viable. See Trial Tr. at 1908:21–1909:6 (“I would not have coded that as not an available sponsor. . . . [T]hat’s a conclusion, that’s not evidence.”). Likewise if an AORW contained the phrase “flight risk” but no indication as to a reason why an age-out was a flight risk, Dr. Lenzo would not have coded this as evidence of a flight risk that would not be addressed by release to a sponsor. Id. at 1910:3–13. 290. Synthesizing the information from the age-out documentation in this way yielded the following results: 285 age-outs were detained between October 17, 2018 and May 24, 2019. Of these, 229 (80.4%) had individual or organizational sponsors available and ICE had recorded no evidence that they presented a danger to themselves, danger to the community, or risk of flight that could not be mitigated by release to a sponsor. Trial Tr. at 1927:16–1928:12 (Lenzo); see also Pls.’ Demo. Ex. 7. This means that only 19.6% of detained age-outs in this time period had age-out documentation showing evidence of one of the statutory risk factors and had no sponsor (that had not been affirmatively investigated and rejected). Trial Tr. at 1930:3–11 (Lenzo). 131 Defendants and their expert Dr. Pan do not dispute these figures directly but argue that Dr. Lenzo did not take “sponsor viability” into account. See Defs.’ Resp. FF ¶ 249. Dr. Pan challenges Dr. Lenzo’s assumptions about both individual and organizational sponsors. For individual sponsors, they reject Dr. Lenzo’s assumption that an identifiable sponsor was available unless ICE documented evidence indicating that they were not. See Trial Tr. at 3445:1–3446:5 (Pan). Dr. Pan took at face value an FOJC’s response of “no” to question 12 on an AORW (“Are there any individual or organizational sponsors . . . available . . . ?”), id. at 3441:6–3443:9, and further assumed that an organizational sponsor was not available unless there was a specific reference to that sponsor in the AORW or SharePoint materials, id. at 3568:12–17. Under Dr. Pan’s set of assumptions, only around seven percent of detained ageouts had “viable” sponsors, id. at 3445:3–16, and only around three percent of detained ageouts—rather than Dr. Lenzo’s 80.4%—had no statutory risk factors and a viable sponsor, id. at 3447:21–3448:13. More broadly, Defendants argue that Dr. Lenzo’s assumptions surrounding the availability of organizational sponsors and group homes “did not reflect reality.” Defs.’ Resp. ¶ 247. It bears repeating that although ICE generally requires that sponsors be deemed “viable,” and although over one hundred AORWs stated “no viable sponsor” as part of the reason why an age-out was detained, ICE does not have a definition of viability or any articulated standards for assessing it. See supra ¶¶ 150–51; see also Trial Tr. at 3432:11–3433:15 (colloquy between the Court and Dr. Pan regarding her definition of viable, which she agreed was not based on any “ICE guidance” or “formal definition” but which instead “just mean[t] the ones without negative comments or negative evidence”). Though they phrase their dispute in terms of “viability,” 132 Defendants are essentially arguing only for a different set of presumptions and assumptions about which age-outs could potentially have been placed with sponsors. Defendants are correct that Dr. Lenzo’s count is likely not a completely accurate measure of the reality of how many age-outs could have been placed if ICE wanted to place them. But the Court does not understand Dr. Lenzo to have ever claimed as much. This type of after-thefact analysis based on documentation can only approximate the facts on the ground at the moment of the custody determination based on the information that ICE provided. The true and historically accurate count of how many age-outs could have gone to sponsors is not calculable in hindsight, though it likely falls somewhere in between Dr. Lenzo’s count and Dr. Pan’s. The Court must then evaluate which statistician suggests the set of assumptions that better approximates reality, and, more specifically, whether Dr. Lenzo’s assumptions are reasonable and whether they have been undermined by Dr. Pan’s critiques. Despite some argumentation in the parties’ briefing about the difference between “structured data” versus “unstructured data,” see, e.g., Defs.’ Br. FF ¶ 712, the Court does not understand the disagreements about how to count available sponsors as ones that require any kind of advanced statistical analysis or understanding to resolve. Rather, these disagreements require a decision about which default rules and assumptions will make the statistical model more closely resemble the real world. Having heard extensive testimony concerning ICE’s age-out documentation procedures, the Court is well positioned to make factual findings about which set of assumptions are most reasonable. Considering everything the Court has learned concerning ICE’s age-out custody determinations, and the facts it has found, the Court finds that the assumptions about sponsor availability made by Dr. Lenzo are reasonable and that Dr. Pan’s challenges do not call into question his methodology. 133 Regarding individual sponsors, Dr. Pan recommended taking AORW responses “at face value,” id. ¶ 714, and trusting that there was no individual sponsor available so long as Item 12 on the AORW indicated as much, Trial Tr. at 3441:6–3443:9 (Pan); id. at 444:9–11. This is a particularly credulous approach to mining data from the age-out documentation considering the several ways in which Plaintiffs demonstrated their unreliability. As Plaintiffs’ counsel demonstrated repeatedly at trial and in depositions, a statement on an AORW that a sponsor is not “viable” does not necessarily mean that no sponsor was available. Officer Hyde, for example, admitted that “it wasn’t truthful” when an FOJC answered “no” for Item 12 on one particular AORW, id. at 3374:24 (Hyde), and indicated in the remarks section that a particular shelter had not provided a post-18 plan, id. at 3375:13–21; see generally id. at 3365:12–3375:21 (discussing DX-60). On another AORW reviewed with Officer Hyde, ICE had the name and phone number of a family member and potential sponsor, and an FOJC indicated there was no sponsor, without documenting any attempt at contacting this family member. Id. at 3392:8– 3399:21 (Hyde) (reviewing DX-64). ICE generally viewed sponsors with whom ORR failed to place age-outs as not “viable” sponsors, despite a lack of any articulable reasons why ICE ought to follow ORR sponsorship standards in deciding to whom age-outs could be released. See, e.g., supra ¶¶ 157–63. On top of this, significant testimony showed that many AORWs were filled out long after a custody determination was made, or by an officer who did not make the custody recommendation, or under JFRMU’s instruction in advance of a deposition. See, e.g., supra ¶¶ 63–67. Dr. Pan admitted that her method would treat the answer to Item 12 as accurate even when there was substantive evidence in the age-out documentation that contradicted the conclusory “no” response. Trial Tr. at 3601:10–21 (“Q. . . . [Y]ou testified that you relied on a 134 no answer to [Item 12] even where there was contrary evidence? . . . A. Yes, I did.”)); see also id. at 3458:7–12 (“The Court: So if the officer clicked no . . . you took that as a given? [Dr. Pan]: I took that as a no.”). On cross-examination, Plaintiffs’ counsel illustrated the flaws with Dr. Pan’s “face value” approach to age-out documentation in a few analogous contexts: Dr. Pan admitted that certain AORWs indicated that an age-out was a “flight risk, danger to the community, or danger to themselves” and checked the box for “other credible threat,” but with purportedly explanatory remarks that bore no relation to any type of threat or danger. Id. at 3482:20–3483:3. Another group of AORWs indicated that age-outs were flight risks because they had removal orders pending, but a review of the substance of their SharePoint materials revealed that the orders were not entered until after the age-out was placed in custody by ICE. See id. at 3518:7–3528:4. 36 Elsewhere in her testimony, Dr. Pan noted that her model was “much more likely to observe the detention or release outcomes in real life,” meaning that “[her] definition [of sponsorship] is closer to the real factors that were considered by the field officer.” Id. 3453:23– 3454:3; see also id. at 3455:13–17 (“[S]tatistically our model fits better. And the reason is the viable sponsor, it’s closer to the true outcomes, the consideration in real life what they relied on. They didn’t rely on the assumption of everyone has a sponsor . . . .”); id. at 3464:4–8; 3473:19– 36 There were nineteen age-outs coded by Dr. Lenzo as not presenting risk factors whom Dr. Pan did code as having risk factors based on a face value evaluation of Item 8 on the AORW. See Trial Tr. 3474:16–3475:24 (Pan). The testimony cited in this paragraph comes from Plaintiffs’ cross-examination of Dr. Pan regarding these nineteen age-outs. At this stage, Defendants do not appear to dispute Dr. Lenzo’s evaluation of the documentation for these ageouts. See Defs.’ Resp. ¶¶ 239–57 (raising no challenge to these aspects of Dr. Lenzo’s work); see also Trial Tr. at 3453:6–15 (Pan) (“The only difference, in fact, from my regression and Dr. Lenzo’s regression is the definition of the sponsor. Again it’s the sponsor that we differ. Others are all the same variables, the same dataset[.]”). For the same reasons discussed in this paragraph, the Court finds that Dr. Lenzo’s method of counting evidence of the statutory risk factors is reasonable and preferable to taking the AORWs at face value. 135 20 (“[T]he main thing is sponsor availability -- available, viable sponsor.”). The Court does not find this compelling, as it presumes that field officers were following the correct factors and the law. That was the question that the Court set out to decide at trial, and cannot be taken as a given. If, as seems to be the case, ICE officers were stating that there were no viable sponsors in cases where in fact there were viable sponsors, then of course an officer’s statement that there was no viable sponsor would be a good predictor of a detention decision. But it would reveal nothing about the officer’s actual attention to the statutory factors or consideration of placement in the least restrictive setting. Consider, by way of analogy, an employer using a similar form to document hiring or promotion decisions. If that form included a box that could be checked to indicate whether an applicant was qualified, and if that employer were sued for discrimination, no court would take at face value the employer’s forms indicating that white applicants were qualified and nonwhite applicants were not. Courts would look for evidence of the applicants’ qualifications. Any statistical analysis of applicants’ qualifications that took the boxes at face value would be useless. Dr. Pan suggests doing just that here. Her analysis even disregards contrary substantive evidence. Dr. Pan further testified that, to the extent she did evaluate substantive information contained in the age-out documentation that might have affected whether sponsors were viable, evaluations were made by Dr. Pan and her assistant, “just by [their] judgments.” Id. 3433:8– 3434:4. This methodology is, perhaps, a necessary consequence of ICE’s lack of standards for sponsor viability. Nonetheless it is not a methodology that would necessarily approximate the reality of sponsor availability more accurately than Dr. Lenzo’s method of designating sponsors unavailable so long as some substantive derogatory information was found. This method took 136 into account any evidence that ICE had found an individual sponsor was not viable. Id. at 1904:16–1905:11 (Lenzo). Because ICE officers were instructed to document all considerations that went into their custody determination, see PX-1 at 1, any such information should have been recorded on the AORW. Assuming derogatory information when none was presented is speculative. Turning to organizational sponsors, the Court first observes that Dr. Pan and defense counsel mischaracterized Dr. Lenzo’s method as having “assumed that anyone sent to [13 field officers with shelters in their AOR] has a potential sponsor.” Trial Tr. at 3422:17–24 (Pan); id. at 3447:23–3448:1 (question by Ms. Parascandola referencing “Dr. Lenzo’s assumption that a sponsor was available” for these age-outs); id. at 3435:17–18 (same). As Dr. Lenzo testified, he did not make such a blanket assumption. Rather, he assumed that these organizational sponsors in the area of a field office were available only if the age-out’s AORW and SharePoint materials lacked evidence of ICE’s having considered and rejected them. See id. at 1905:25–1907:8, 1907:24–1908:7, 2024:7–11 (Lenzo). The Court considers it self-evident that there were limits on the capacity of shelters and group homes, and that it may not necessarily have been the case that 94% of age-outs could realistically have been placed in one on the precise day they turned eighteen. At the same time, placing a significant number of those age-outs would not necessarily have been impossible. Testimony at trial showed that when FOJCs made efforts to locate organizational sponsors for age-outs who did not have individual sponsors, they were able to do so in almost all instances. See supra ¶¶ 215 (San Antonio); 225 (Harlingen after September 2018); 233 (Chicago after September 2018). Moreover, this analysis is conservative in that it treats as available only organizational sponsors from a field office’s particular AOR. In reality, some shelters accepted 137 age-outs from throughout the country. See supra ¶ 42 (citing Trial Tr. at 1317:14–25 (Lehner) (testifying that age-outs processed by the Miami field office have been placed at Viator House in Chicago and at La Posada and Casa Marinella in Texas)). Dr. Lenzo did not take this into account, see Trial Tr. at 1929:6–25 (Lenzo), so he may have missed some proportion of age-outs who could have been placed with a shelter outside the processing office’s AOR. To the extent that Dr. Lenzo does overestimate the availability of organizational sponsors, he does not seem to have done so to such a degree as to call his broader analysis into question. Dr. Lenzo explained that he considered the possibility that he was assuming too much in regard to organizational sponsor availability and ran a “robustness check” which confirmed that this assumption was not driving the statistical analysis. Trial Tr. at 1942:9–1943:20; see also id. at 2023:16–21. While Defendants challenge the reasonableness of Dr. Lenzo’s assumptions about organizational sponsors, they do not suggest that Dr. Pan identified this assumption as having any kind of outsize influence on Dr. Lenzo’s conclusions. The Court expects that if the assumption about organizational sponsor availability made a significant difference in the conclusions Dr. Lenzo put forward, Dr. Pan would have so testified. It makes sense that this assumption would not have a large impact on the overall analysis because the availability of organizational sponsors is somewhat ancillary to the main conclusions Dr. Lenzo reaches in his analysis, which concern the relative effect on age-out custody determinations of the risk factors identified in Section 1232(c)(2)(B) and the field office making the custody determination. Dr. Lenzo’s analysis would certainly have been more robust if he had been able to account for variables like bed space availability on a given night, but the kind of regression analysis that Dr. Lenzo was attempting to perform necessarily requires some simplification and reduction of the nuances of particular cases, and the Court understands him to have taken 138 reasonable precautions in his analysis to avoid making any assumptions that would overly distort his results. 37 Finally, considering as a whole the assumptions that Dr. Lenzo made in synthesizing ICE’s production into a manageable database format, the impact of the Plaintiffs-favoring aspects of his assumptions about both individual and organizational sponsors are arguably mitigated somewhat by his Defendants-favoring assumptions, including (a) that organizational sponsors outside an office’s AOR were not available for placement, see supra ¶ 290; (b) that weak indications of risk factors should be considered on par with strong ones, see supra ¶ 284; and (c) that no individual potential sponsor was available when the SharePoint materials were missing a Form I-213, see supra ¶ 285. These assumptions do not all directly offset Dr. Lenzo’s assumptions about organizational sponsor availability, but they contribute to Dr. Lenzo’s credibility and rigor by demonstrating that he did not attempt to draw every possible assumption in favor of Plaintiffs. 291. Having thus determined the presence or absence of the statutory risk factors and the availability of sponsors—as reflected in the age-out documentation—Dr. Lenzo performed two 37 Defendants did not produce any analysis of available bed space either, and the Court suspects that doing so would not have been feasible for either party. Dr. Lenzo and his firm should not be criticized for not undertaking an analysis that would have increased exponentially the resources his firm put into this case. So long as the analysis performed was adequate—as it is here—an expert need not disprove every possible alternative theory, especially when there is no factual basis. The testimony was clear that FOJCs who attempted to place age-outs with organizational sponsors were able to do so in nearly every case. See supra ¶¶ 210–15 (San Antonio); 225 (Harlingen after September 2018); 233 (Chicago after September 2018). Although Defendants intimate that Dr. Lenzo should have considered bed space availability, they point to no testimony from any ICE officer suggesting that availability of bed space at organizational sponsors was an obstacle to release. See, e.g., Defs.’ Br. FF ¶ 654; Defs.’ Resp. ¶¶ 242–57 (responding to Plaintiffs’ discussion of the expert statistical testimony). 139 sets of statistical analyses seeking to explain the disparities in detention rates across different field offices. Trial Tr. at 1930:12–19. 292. Dr. Lenzo’s first analysis asked “whether the presence or absence of the risk factors among age-outs could explain the disparities” between field office detention rates. Id. at 1934:14–18. This involved looking only at those age-outs for whom there was no evidence of any of the statutory risk factors in their age-out documentation and determining whether the different field offices determined members of this subset of age-outs at rates different enough to be statistically significant. Id. at 1934:19–1935:5. Dr. Lenzo concluded that “[t]he differences in assortment of risk factors among age-outs assigned to offices cannot explain the differences we see in . . . detention rates across those offices.” Id. at 1935:17–22. 293. Dr. Lenzo’s second analysis was designed to control for a variety of variables including field office and the presence or absence of evidence for each risk factor, with the aim of evaluating whether the offices’ different detention rates could be explained by the presence or absence of the risk factors or of any particular risk factor. See id. at 1935:23–1937:21. This analysis was able to compare how much more likely an age-out was to be detained if, for instance, he was processed by a particular office or if his file contained evidence of one of the statutory risk factors. See id. at 1938:11–1940:13 (referencing Pls. Demo. Ex. 16 and explaining the regression analysis). From this, Dr. Lenzo concluded that, holding all else equal, an age-out was more likely to be detained if processed by one of six ICE field offices—Houston, Miami, El Paso, New York, Los Angeles, or Phoenix—than if that age-out’s documentation contained evidence of any of the statutory risk factors—danger to community, danger to self, or risk of flight. See id. at 1940:3–13, 1941:3–10. Dr. Lenzo then ran a further analysis to confirm that the 140 field office “matter[ed] significantly to the probability the age-out is detained, even once controlling for risk factors,” and confirmed that it did. Id. at 1941:11–16. 294. The second analysis confirmed the outsize influence of the field office on age-out custody determinations. Age-outs were determined to be 12.7% more likely to be detained if their documentation contained evidence of any one risk factor, and 28.0% more likely to be detained if their documentation contained evidence of all the risk factors. Id. at 1943:11–1945:2 (discussing Pls.’ Demo. Ex. 19); id. at 1945:8–1946:17(discussing Pls.’ Demo. Ex. 18). Field offices mattered much more: age-outs processed by the Houston office were over 90% more likely to be detained than otherwise identical age-outs processed by the San Antonio office. Id. at 1946:1–17; see also Pls.’ Demo. Exs. 18, 19. The same was true for the Miami office. See id. Age-outs processed in El Paso were over 80% more likely to be detained, New York over 60%, and Los Angeles and Phoenix also showed significant increases. See id. 295. Defendants do not dispute these characterizations of Dr. Lenzo’s methods and opinions but challenge that his model was “oversimplified” or “defective.” SeeDefs.’ Resp. ¶¶ 252–53. The Court has already addressed the argument that the model oversimplified the assessment of risk factors in particular by treating them as binary. See supra ¶ 284. More broadly, Defendants and Dr. Pan challenge the application of a regression model in the first place, charging that it “oversimplifies real world conditions, and any Field Office location effects estimated from such regression models could result from other factors such as the characteristics of the age-outs assigned to a particular Field Office.” Defs.’ Br. FF ¶ 756–57. The Court does not find this critique convincing. Dr. Pan’s critique of using regression analysis as a methodology amounts primarily to the position that more complex forms of analysis might provide better information if they could be 141 brought to bear on the question at hand—but Dr. Pan acknolwedged that this was not possible. Trial Tr. at 3580:15–3581:25. As Dr. Pan explained at trial, “to mimic the process of the decision process of field officers,” and to avoid oversimplifying, “you need something like a random forest or recurrent neural network . . . to reflect the complicated detailed thinking process.” Trial Tr. at 3425:20–25. She admitted, though, that such analyses would be impossible here because there are not nearly as many data points (age-outs) as would be needed to set up algorithms of these sort. Id. at 3426:1–16. Because these more precise methods could not be applied here, Dr. Pan suggested that the best way to understand ICE officers’ decisions would be to review the limited set of age-out documentation to see what reasons they provided. See id. at 3426:17–3427:22. She and her assistant had done this, and “summarized” the officers’ records. Id. Simply looking at the stated reasons on the worksheets, of course, does not require expert statistical analysis. As the Court noted at trial—and as Dr. Pan agreed—hearing “three weeks of testimony of the actual decision makers,” would be a better way of evaluating compliance with the law than simply working from a summary of the documentation produced by Dr. Pan and her assistant. Id. at 3427:23–3428:11 (Dr. Pan colloquy with the Court). This does not amount to a compelling critique of Dr. Lenzo’s regression methodology. The fact that a more complex but practically impossible statistical technique might have revealed more does not show that regression did not reveal what Dr. Lenzo asserts it did. The Court agrees with Dr. Pan that there are limits to regression analysis and that it is useful to review the substance of the worksheets and to “interview each field officer, go through the case and judge is this good practice, is this complying [with] the law.” Id. at 3428:2–5. This too does not mean the regression analysis is improper or unhelpful. Instead it describes the substance of the extensive non-expert testimony and evidence that the Court has taken under consideration and 142 has used to assess the credibility of Dr. Lenzo’s work. This testimony was highly relevant to the Court’s evaluation of Dr. Lenzo’s methodology, as the Court applied its understanding of ICE’s practices and policies in determining that Dr. Lenzo’s assumptions about risk factors and sponsor availability were appropriate. 296. Dr. Pan did not challenge Dr. Lenzo’s execution of the regression methodology. Though she formed the opinion that regression was inappropriate at some point after drafting her report and after testifying at her deposition, her rebuttal report agreed with Dr. Lenzo’s analysis that differences in prevalence of risk factors did not fully explain why El Paso, Houston, Los Angeles, Miami, New York and Phoenix had significantly higher detention rates than other offices. See Trial Tr. at 3578:23–3579:19. This was true even when risk factors were identified in the way Dr. Pan proposed—by taking the AORWs at face value. Id. at 3586:21–3588:6. Dr. Pan and her coauthor Dr. Gastwirth tested a number of other independent variables to attempt to explain the variations in detention rates, and nothing they could come up with—availability of counsel for age-outs, third-party involvement, country of origin, field office caseload, number of days in ORR custody, or ORR bed space—was significantly correlated with detention decisions. Id. at 3559:7–3565:13. 297. Defendants raised a new argument for the first time in their Addendum to Defendants’ Responses to Plaintiffs’ Proposed Findings of Fact and Conclusions of Law Concerning Liability. ECF No. 323. In this document they argue that Dr. Lenzo and Plaintiffs “make the fatal mistake of ignoring community environment, including the availability of viable sponsors.” Defs.’ Resp. ¶ 218. The Court has already addressed Dr. Lenzo’s assumptions about sponsors. To the extent that “community environment” means anything more, this argument is underdeveloped, unexplained, and waived by virtue of having been raised for the first time in a 143 response to Plaintiffs’ Proposed Findings. The phrase “community environment”—which was not uttered once in eighteen days of trial—appears over twenty times in Defendants’ Addendum to their Responses, but is nowhere defined. Defendants say that “community environment” includes “factors such as the availability of a viable sponsor . . . , the level of cooperation provided by the local ORR contractor, and whether the FOJC was provided a Post-18 Plan.” Id. ¶ 224. It is unclear whether Defendants mean for “community environment” to be a new concept, or whether it is simply a grab-bag reframing a collection of arguments made elsewhere. The Court has addressed each of the identified elements of community environment. As explained above, the Court found Dr. Lenzo’s assumptions about sponsors to be reasonable. Regarding Post-18 plans, Dr. Pan considered whether “NGO organization[s] help[ing] in locating potential sponsors or providing post-18 plans” correlated with detention outcomes and found that it did not. Trial Tr. at 3561:23–3562:8. And the Court has explained that a high rate of detention could reasonably lead third parties to choose not to bother submitting post-18 plans. Supra ¶ 156. ORR officials could have reached the same conclusion. Dr. Pan did not test for effects of “community environment” nor did she address it in her testimony. There is no evidence of any understood meaning of the term. To the extent that Defendants meant for “community environment” to refer to something new, they have waited far too long to raise it and have thus waived the argument. See Campbell v. District of Columbia, No. 12-cv-1769, 2016 WL 3023977 , at *8 (D.D.C. May 25, 2016) (“[B]ecause the District waited until after trial to raise it, the District has waived this argument . . . .”) (citing Estate of Gaither ex rel. Gaither v. District of Columbia, 771 F. Supp. 2d 5, 10 (D.D.C. 2011) (explaining that Rule 59(e) motions 144 cannot be used as “as a vehicle for presenting theories or arguments that could have been advanced earlier”)). 298. As the Court has observed, statistical experts are not required to disprove every possible speculative and unsupported theory about what could be driving detention rates. See supra n.37 (discussing shelter bed space availability). As Dr. Lenzo noted, proper statistical analysis should focus on “legally relevant” factors and on those for which there was data. Trial Tr. at 2074:5–14 (responding to Defendants’ counsel’s question whether it would be significant if “the age-out . . . was wearing red shoes” correlated with detention decisions, id. at 2073:16– 21). Where one party’s statistical expert is able to demonstrate a statistically significant correlation, and where the rebuttal expert is unable to make a stronger case for any other plausible factor or to disprove the first expert’s methods or implementation of those methods, this is enough for a court to make a finding of fact in line with the first expert’s conclusions. 299. Considering Dr. Lenzo’s testimony as a whole, which the Court finds corroborative of the evidence presented in exhibits and witness testimony, the Court finds as Dr. Lenzo did, that variations in the statutory risk factors presented by age-outs at different ICE field offices do not explain those offices’ varying propensities to detain the age-outs they process. IV. CONCLUSIONS OF LAW Having established and explained its findings of fact, the Court now provides its conclusions of law. A. Administrative Procedure Act Plaintiffs’ claims are brought under 5 U.S.C. § 706 which permits a court to “compel agency action unlawfully withheld or unreasonably delayed,” and to “hold unlawful and set aside agency action, findings, and conclusions found to be arbitrary, capricious, an abuse of discretion, 145 or otherwise not in accordance with law.” 5 U.S.C. § 706 . The APA provides for judicial review of all “final agency action for which there is no other adequate remedy in a court,” id. § 704, except when “statutes preclude judicial review” or the “agency action is committed to agency discretion by law,” id. § 701(a). The Court has already ruled on the applicability of the APA to Plaintiffs’ claims. See MTD and Class Cert. Op. at 44–55. Here the Court briefly summarizes its prior holdings and addresses a few discrete issues that are still disputed. The APA “applies, according to the provisions thereof, except to the extent that—(1) statutes preclude judicial review; or (2) agency action is committed to agency discretion by law.” 5 U.S.C. § 701 (a). Neither exception applies here. The language of the statutory provision features no terms either expressly or impliedly precluding judicial review. Compare 8 U.S.C. § 1232 (c)(2)(B), with, e.g., 8 U.S.C. § 1226 (e) (stating that “discretionary judgment regarding the application of [Section 1226] shall not be subject to review”). No legislative history suggests otherwise. See MTD and Class Cert. Op. at 47. The agency action required by § 1232(c)(2)(B) is not in one of the “categories of administrative decisions” that Courts have held unreviewable. Sec’y of Labor v. Twentymile Coal Co., 456 F.3d 151, 156 (D.C. Cir. 2006). Nor is the statutory text “drawn in such broad terms that in a given case there is no law to apply.” Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 410 (1971); see also MTD and Class Cert. Op. at 48–49. Defendants do not challenge the application of the APA at this stage, but do correctly emphasize that, as the Court has previously ruled, the Court’s review only extends to the decisionmaking process established by Section 1232(c)(2)(B), and not to the substance of the individual age-out custody determinations. See MTD and Class Cert. Op. at 35 n.8 (“As 146 Plaintiffs appear to recognize, they have no legally protected interest in any particular placement . . . .”); Defs.’ Resp. ¶ 611. 38 Plaintiffs’ challenges to ICE’s conduct also met the APA requirement that challenged conduct be a final agency action. Again, the Court has already so determined, MTD and Class Cert Op. at 50–53, and Defendants do not dispute this at this stage, Defs.’ Resp. ¶ 612. The APA defines “agency action” as an “agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act.” 5 U.S.C. § 551 (13); 5 U.S.C. § 701 (b)(2) (explaining that “agency action” for purposes of the judicial review provisions of the APA carries the same meaning given by 5 U.S.C. § 551 ). With respect to complaints that an agency failed to act, “a claim under § 706(1) can proceed only where a plaintiff asserts that an agency failed to take a discrete agency action that it is required to take.” Norton v. S. Utah Wilderness All. (“SUWA”), 542 U.S. 55, 64 (2004) (emphases added). Even discrete agency actions are reviewable by a court under the APA only if they are final. See 5 U.S.C. § 704 (establishing reviewability of “final agency action”). As the Supreme Court established in Bennett v. Spear, 520 U.S. 154 38 Inconsistent with this correct description of the scope of the Court’s review, Defendants have argued in their responsive briefing on remedies that “[u]nder the APA’s arbitrary and capricious standard, this Court must defer to the agency, and indeed must grant extra deference to agency decisions made under Section 1232(c)(2)(B).” Defs.’ Remedies Mem. Resp. at 5. This argument misunderstands what is at issue in this case. ICE would certainly be entitled to some level of deference regarding the outcomes of age-out custody determinations, and that is why there is no entitlement to any particular placement for any particular age-out. However, ICE is not entitled to any deference when it comes to whether or not it follows the required procedures laid out in Section 1232(c)(2)(B). Sas Inst., Inc. v. Iancu, 138 S. Ct. 1348, 1358 (2018) (“[W]e owe an agency’s interpretation of the law no deference unless . . . we find ourselves unable to discern Congress’s meaning.”); see Chevron, U.S.A. Inc. v. Natural Res. Def. Council., 467 U.S. 837 , 843 n. 9 (“If a court, employing traditional tools of statutory construction, ascertains that Congress had an intention on the precise question at issue, that intention is the law and must be given effect.”). Nor is the agency entitled to deference with regard to whether certain documentation—like an AORW—demonstrates compliance with the statute’s required procedures. To defer on that question would be to defer to the agency on the question of whether they have complied with the statute, and that is indisputably the Court’s role. 147 (1997), a court will find that an agency action is final if two conditions are met: “First, the action must mark the consummation of the agency’s decisionmaking process—it must not be of a merely tentative or interlocutory nature. And, second, the action must be one by which rights or obligations have been determined, or from which legal consequences will flow.” Id. at 177–78 (internal quotation marks and citations omitted). Plaintiffs have challenged final agency action subject to review under the APA. They seek to compel Defendants to conduct the inquiry that the statute mandates they conduct before determining where to place each former unaccompanied minor now in DHS’s care and custody. See MTD and Class Cert Op. at 52–53 (citing P.I. Op. at 18–20). Turning to matters that remain disputed, Defendants contest whether age-outs who are detained without ICE following the requirements of Section 1232(c)(2)(B) suffer irreparable harm. See Defs.’ Resp. ¶ 613 (“Defendants dispute that an irreparable harm exists here.”). The Court has already rejected this argument. As it explained when granting a preliminary injunction with regard to the named Plaintiffs, “Courts in this and other jurisdictions have found that deprivations of physical liberty are the sort of actual and imminent injuries that constitute irreparable harm” and “have likewise recognized that the ‘major hardship posed by needless prolonged detention’ is a form of irreparable harm.” P.I. Op. at 36–37 (quoting R.I.L-R v, Johnson, 80 F. Supp. 3d 164, 191 (D.D.C. 2015) and citing Seretse-Khama v. Ashcroft, 215 F. Supp. 2d 37 , 53 n.20 (D.D.C. 2002)). “[W]here a plaintiff requests injunctive relief mandating that an agency comply with a process that, if completed could secure plaintiff’s freedom or could alleviate harsh conditions of confinement, the harm from detention surely cannot be remediated after the fact.” Id. at 37 (citing R.I.L-R, 80 F. Supp. 3d at 191 ). Defendants argue that any injury “is quite reparable—this Court merely need instruct the agency to perform the step of 148 consideration.” Defs.’ Resp. ¶ 613. As the Court already explained, this kind of re-evaluation would not “lessen the harm associated with each additional day Plaintiffs endure purportedly inappropriate detention.” P.I. Op. at 37. The harm that detention causes to age-outs was established at trial by Plaintiffs’ pediatric expert Dr. Linton, see supra ¶¶ 23–26, and was not contested by Defendants, see supra ¶ 27 (quoting Trial Tr. at 2423:16–17 (Mr. Kisor agreeing with the Court that “this is not going to be a contested issue”). While the procedural injury could be rectified, the lasting harm from detention could not be. Last, Plaintiffs have no alternative adequate remedies at law. 39 See 5 U.S.C. § 704 (exempting from judicial review agency action for which there is an “adequate remedy in a court”). The Court has addressed the lack of other adequate remedies in two prior orders. P.I. Op. at 22–25; MTD and Class Cert. Op. at 53–55. As the Court explained, “[w]hen considering whether an alternative remedy is ‘adequate’ and therefore preclusive of APA review, [courts] look for ‘clear and convincing evidence’ of ‘legislative intent’ to create a special, alternative remedy and thereby bar APA review.” MTD and Class Cert Op. at 53 (quoting Citizens for Responsibility & Ethics in Wash. v. U.S. Dep’t of Justice (“CREW”), 846 F.3d 1235, 1244 (D.C. Cir. 2017)). Defendants previously argued that bond hearings were adequate alternatives, but the Court rejected this contention because bond hearings are too speculative, would not necessarily take place quickly, and would not remedy the injury. See P.I. Op. at 24–25. Since that argument was rejected, Defendants have suggested no other alternatives. See MTD and Class Cert Op. at 39 Defendants appear to dispute this. Defs.’ Resp. ¶¶ 614-15. Their briefing provides no suggestion of what other remedy is available, but instead they assert that “because there is no irreparable injury, it is not possible for Plaintiffs to establish that the remedies at law are lacking.” Defs.’ Resp. ¶ 615. Given that the Court has identified an irreparable injury, it is not clear that Defendants would maintain their dispute on this point. 149 54–55; see also Pls.’ Br. (not raising an argument based on 5 U.S.C. § 705 ). The Court therefore stands by its prior rulings. B. Class Membership As a final preliminary matter before returning to the substance of Section 1232(c)(2)(B), the Court addresses Defendants’ arguments concerning class membership. Defendants have suggested, but not directly argued, that the class may not contain sufficient members to be maintained. See Defs.’ Br. CL ¶ 23 (arguing that “Plaintiffs have the continuing burden of proof by preponderance of the evidence to show that the class contains sufficient members to remain certified”); Defs.’ Resp. ¶ 45 (arguing that age-outs are not class members if they either were not detained or were detained through a process consistent with Section 1232(c)(2)(B), and also asserting that “[i]t is now incumbent upon the Court to determine which former UACs detained by ICE are class members”). As Plaintiffs observe, Defendants do not have a motion to decertify the class pending, Pls.’ Resp. CL ¶ 23, and in fact withdrew a motion to decertify that they filed last year, Defs.’ Mot to Hold Defs.’ Motion to Decertify the Class in Abeyance, ECF No. 192. As the Court explained when ruling on Defendants’ Motion for Summary Judgment, the size of the class in this case depends on the finding of liability. See MSJ Order at 4. “If ICE Officers in the field are not actually considering the least restrictive setting, then there might well be plenty of discernable age-out class members.” Id. Because the Court finds today that violations of 1232(c)(2)(B) are occurring, there is little difficulty in saying that the class is sufficiently numerous. See Richardson v. L’Oreal USA, Inc., 991 F. Supp. 2d 181, 196 (D.D.C. 2013) (“In 150 this district, courts have found that numerosity is satisfied when a proposed class has at least forty members . . . .”). 40 C. Statutory Requirements Under Section 1232(c)(2)(B), “[i]f a minor described in subparagraph (A)”—that is, “an unaccompanied alien child in the custody of [HHS],” § 1232(c)(2)(A)— “reaches 18 years of age and is transferred to the custody of the Secretary of [DHS], the Secretary shall consider placement in the least restrictive setting available after taking into account the alien’s danger to self, danger to the community, and risk of flight.” 8 U.S.C. § 1232 (c)(2)(B) (emphasis added). This first sentence requires two things of the Secretary. First, the Secretary must take into account the three statutory factors—dangers to self, danger to the community, and risk of flight. Second, the Secretary must “consider placement in the least restrictive setting available.” This “consider[ation]” is due no matter the outcome of the danger and flight evaluation because the statute does not place any limits or conditions on it other than the sentence’s first clause. From there, “such aliens”—meaning all age-outs because all age-outs are encompassed by the preceding sentence—“shall be eligible to participate in alternative to detention programs, 40 Between April 2016 and May 24, 2019, the ICE field office in Houston detained 515 age-outs and the Miami field office detained 181 age-outs. See supra ¶ 176 (citing PX-334 at 43–53 and PX–334A at 147–84 for Houston data); supra ¶ 191 (citing PX-334 at 57–60 and PX– 334A at 218–47 for Miami data). Even if only a small fraction of detained age-outs from just these two offices were processed without section 1232(c)(2)(B) being followed, the number of class members would still be significantly more than forty. In its most recent filing, the Government has suggested that the Court must scrutinize, one at a time, the AORW created for each age-out. Resp. to Pls.’ Resp. to Defs.’ Notice of Supp. Authority, ECF No. 332. This is too late to raise such an argument. See Brady Ctr. to Prevent Gun Violence v. U.S. Dep’t of Justice, 410 F. Supp. 3d 225 , 240 n.3 (D.D.C. 2019) (finding that an argument raised “for the first time in [a] Notice of Supplemental Authority” had been forfeited because it was “only belatedly and cursorily raised”); see also U.S. ex rel. Miller v. Bill Harbert Intern. Const., Inc., 608 F.3d 871 , 878 n.1 (D.C. Cir. 2010) (treating arguments raised in a post-argument letter as forfeited). And regardless, as the Court has explained, the AORWs are not a reliable record of agents’ actual decisionmaking processes. See supra ¶¶ 62–66. 151 utilizing a continuum of alternatives based on the alien’s need for supervision, which may include placement of the alien with an individual or an organizational sponsor, or in a supervised group home.” Id. Thus, per the plain language of the provision, for each and every age-out, DHS must (1) take into account the statutory factors, (2) do so as it considers placing the age-out in the least restrictive setting available, and (3) make the age-out eligible for the identified alternative to detention options. As the Court has already noted, the statute is unambiguous in this regard and must be enforced “according to its terms.” MTD and Class Cert. Op. at 31 (quoting Bd. of Cty. Comm’rs of Kay Cty., Okla. v. Fed. Hous. Fin. Agency, 754 F.3d 1025, 1029 (D.C. Cir. 2014) (quoting Jimenez v. Quarterman, 555 U.S. 113, 118 (2009))). The Secretary’s obligation to “consider placement [of each age-out] in the least restrictive setting available” and make each age-out “eligible for participation” in certain programs regardless of the age-out’s danger to self, danger to community, or risk of flight warrants further discussion. The Court addressed this in some detail in a prior Opinion, yet it evidently remains a matter of at least some dispute. See Defs.’ Resp. ¶ 621 at pp.451–53 (citing MTD and Class Cert Op. at 40–41). The Court explained that “Section 1232(c)(2)(B) does not limit “consider[ation]” or “eligib[ility] to participate in alternative to detention programs” to those who DHS has determined pose no risk of flight.” MTD and Class Cert. Op. at 40. The assessment of the three statutory factors “is only the beginning of the Secretary’s inquiry.” Id. This is a simple matter of statutory construction because the requirement to “consider” is not limited in any way and because “such aliens” in the second sentence refers back to all the aliens previously specified— all age-outs. See id. at 32. It is therefore contrary to the statute for ICE to engage in what Plaintiffs have called “sequential decisionmaking,” which entails “first determin[ing] whether the Age-Out should be detained, and then consider[ing] alternatives to detention only if the Age152 Out is released, as a form of release, instead of as an alternative to detention.” Pls.’ Br ¶ 621 at p.240. When the Court observed that the statute requires an “individualized assessment of the proper placement” for each age-out that does not “short-circuit th[e] inquiry,” this is what it meant. MTD and Class Cert. Op. at 40. The “consider[ation]” must be given “after taking into account” the factors—ICE cannot take the factors into account and decide directly from that step that no consideration of alternatives to detention is necessary. Defendants argue that “[p]rovided that officers are considering the relevant statutory factors, this Court lacks the authority to require the agency to follow a specific sequence or structure when making decisions under Section 1232(c)(2)(B).” Defs.’ Resp. ¶ 629. They cite Weyerhaeuser Co. v. Costle, 590 F.2d 1011 (D.C. Cir. 1978), for the proposition that if “Congress did not mandate any particular structure or weight” for an agency’s consideration of a variety of factors, then the agency is left with “discretion to decide how to account for the consideration factors, and how much weight to give each factor.” Id. at 1045 . This is precisely correct, but here it is Congress that has mandated a particular structure. The structure being required is derived from the text of the statute, it is not an invention of the Court. Weyerhaeuser concerned a statute requiring the Environmental Protection Agency to “take into account” certain factors, “without prescribing any structure for EPA’s deliberations.” Id. at 1045–46 (quoting 33 U.S.C. § 1314 (b)(2)(B) (1972)). Here, the statute does prescribe a structure—taking the statutory factors into account and then considering the least restrictive setting available. It implicitly proscribes another—taking the statutory factors into account and detaining automatically based on that accounting. JFRMU’s training materials, therefore, instruct FOJCs to follow decisionmaking procedures that are inconsistent with the statute insofar as they suggest 153 sequential decisionmaking in which not every age-out is at least considered and eligible for detention alternatives. See supra ¶¶ 131–33. The first thing the statute requires of the Secretary is that he or she “take into account” the statutory factors. See 8 U.S.C. § 1232 (c)(2)(B) (“[T]he Secretary shall consider placement . . . after taking into account . . . .” (emphasis added)). This “account[ing]” is the type of factor-weighing exercise that the D.C. Circuit described in Weyerhaeuser. See Weyerhaeuser, 590 F.2d at 1046 (“So long as [the agency] pays some attention to the congressionally specified factors, the section on its face lets [the agency] relate the various factors as it deems necessary.”). Congress identified three factors—danger to self, danger to community, and risk of flight—that are relevant to the decision, but did not dictate how they ought to be weighed in relation to one another or how each should be factored into the final decision. This is a relatively common framework for Congress to use in giving instruction to an administrative agency, and courts are familiar with it. See, e.g., New York v. Reilly, 969 F.2d 1147, 1150 (D.C. Cir. 1992) (“Because Congress did not assign the specific weight the [EPA] Administrator should accord each of these factors, the Administrator is free to exercise his discretion in this area.”); Cent. Vermont Ry., Inc. v. I.C.C., 711 F.2d 331 , 336 (D.C. Cir. 1983) (“when a statute requires an agency to ‘consider’ a factor, the agency must reach an express and considered conclusion about the bearing of the factor, but need not give any specific weight to the factor.” (alterations and quotations omitted)); Accrediting Council for Indep. Colleges & Sch. v. DeVos, 303 F. Supp. 3d 77, 112 (D.D.C. 2018) (“the Court finds no reason to question the reasonableness of the weight that the Secretary assigned to the evidence”); see also Am. Corn Growers Ass’n v. EPA, 291 F.3d 1, 6 (D.C. Cir. 2002) (“Although no weights were assigned, the factors were meant to be considered together by the states.”). 154 The Secretary’s second obligation—to “consider placement in the least restrictive setting available”—adds a separate step to the decisional structure. Defendants suggest that “consider” means “take into account,” Defs.’ Br. CL ¶ 31, and they are correct to note that courts— including this one—have treated the two terms as essentially synonymous. N.M. Navajo Ranchers Ass’n v. I.C.C., 850 F.2d 729, 735 (D.C. Cir. 1988) (noting that a prior ruling “held only that the [agency]’s mandate to consider the public interest required it to take into account their underlying policies”); Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506 , 516 n.17 (D.C. Cir. 1983) (interpreting Weyerhaeuser and observing that “[the Clean Water Act] directs EPA to ‘consider’ certain factors; this requires the agency only to ‘take into account’ those factors”); Delta Air Lines, Inc. v. Export-Import Bank of United States, 85 F. Supp. 3d 387, 407 (D.D.C. 2015) (Contreras, J.) (noting that certain statutes “require [an agency] to consider (or, more accurately, ‘take into account’) the potential adverse effects . . . .”). Nonetheless, it is clear from the language of the statute that the obligations are separate. Plaintiffs argue that “consider” means something more than “take into account,” Pls.’ Resp. CL ¶ 31, and appeal to “the usual rule [of statutory construction] ‘when the legislature uses certain language in one part of the statute and different language in another, the court assumes different meanings were intended.’” Sosa v. Alvaraez-Machain, 542 U.S. 692 , 711 n.9 (2004). While the Court agrees that the “consider” requirement and the “take into account” requirement in Section 1232(c)(2)(B) are distinct, the Court does not reach this conclusion based on the definitions of the terms. Instead, the Court relies on the structure of the decisional process described (“consider placement in the least restrictive setting available after taking into account . . .”). Additionally, there is a meaningful difference in kind between the three risk factors and the setting or settings, all of which the Secretary is obligated to think about in some 155 way. The Secretary cannot “consider placement in the least restrictive setting available” in the same way as he “tak[es] into account” the three risk factors, because the age-out’s placement is a potential output of this decisionmaking process, while the factors are characteristics of the input (the age-out). It is equally clear to the Court that the “consider[ation of] placement in the least restrictive setting available” must bear some relation to the risk factors that have been “take[n] into account.” 41 Rather, the sentence outlines the structure of a single agency decision, with two components to it. As the Court put it previously, “the statute calls for an individualized assessment of the proper placement for each former unaccompanied minor in light of DHS’s assessment of his or her danger to self, danger to the community, and risk of flight.” MTD and Class Cert Op. at 40. The Secretary must treat the risk factors as relevant variables, like the “consideration factors” in Weyerhaeuser, and then must determine a placement, the outcome of the decisionmaking process for each age-out. See Weyerhaeuser, 590 F.2d at 1045 . In reaching that outcome, the Secretary must “consider”—must think about as a possibility—the least restrictive setting available when processing an age-out. The use of “consider” at this point in the statute is best understood in contrast with Section 1232(c)(2)(A), which requires that UACs in the custody of the Secretary of HHS “shall be promptly placed in the least restrictive setting that is in the best interest of the child.” 8 U.S.C. § 1232 (c)(2)(A). When processing age-outs, the secretary of DHS is left greater discretion with regard to what setting is deemed appropriate, 41 Although Congress’s decisionmaking structure places one step “after” the other, Defendants have never suggested that taking the risk factors into account is a freestanding exercise that simply must be completed before the Secretary proceeds to the unrelated process of considering where to place the age-out. 156 but the least restrictive setting must be considered, and the specified risk factors must be “tak[en] into account.” To consider the least restrictive setting available, the Secretary—or the ICE officers actually making the decisions—are obligated to determine what settings are “available.” This, too, is a matter of statutory construction. Defendants renew their argument that courts cannot dictate procedures where the agency was left with discretion to structure its own decisionmaking, Defs.’ Resp. ¶ 624, and also suggest that because “[t]he word ‘identify’ appears nowhere in the statute” their officers cannot be under any obligation to identify placements, Defs.’ Resp. ¶ 621 at p.445. These arguments fail because it is Congress, not this Court, that implied some degree of awareness of alternatives. The “least restrictive setting” is defined in relation to other, more restrictive settings, and the Secretary cannot know which setting is least restrictive without knowing the range of alternatives that are available. Congress singled out a particular alternative for consideration but only described it based on its relation to other alternatives. This makes it the Secretary’s obligation to determine the full range of alternatives and work out which one Congress singled out for consideration. The instructions are similar to those given by a diner asking a waiter for a bottle of the restaurant’s least expensive wine. The diner does not have to be familiar with the full list of wines and their prices in order to give this instruction, but the waiter must be, or must review the list and identify the cheapest bottle in order to carry out the instruction. Congress similarly did not need to use the word “identify” in order to impose an obligation that a range of possible settings and the least of these be identified. The Secretary could not consider placing the age-out in the least restrictive setting available without 157 determining what available setting is least restrictive. And the Secretary cannot determine that without evaluating the range of available settings. 42 Even without the kind of specific statutory directive to consider certain alternatives that Section 1232(c)(2)(B) presents, agencies are often obligated to do so. Courts have held agency actions to be arbitrary and capricious for failure to consider “significant and viable” alternatives. Farmers Union Cent. Exchange, Inc. v. FERC, 734 F.2d 1486 , 1511 n.54 (D.C. Cir. 1984). Agencies are not obligated to consider “every alternative . . . conceivable . . . regardless of how uncommon or unknown,” but “[t]he failure of an agency to consider obvious alternatives” often leads to reversal. City of Brookings Mun. Tele. Co. v. F.C.C., 822 F.2d 1153 , 1169 (D.C. Cir. 1987) (first quoting Farmers Union, 734 F.2d at 1151 n.54; and then quoting Yakima Valley Cablevision, Inc. v. F.C.C., 794 F.2d 737 , 746 n.36 (D.C. Cir. 1986)); see also Nat’l Shooting Sports Found., Inc. v. Jones, 716 F.3d 200, 215 (D.C. Cir. 2013). Failure to consider alternatives is an issue more frequently addressed when courts review agency rulemakings as opposed to adjudications, but there is precedent for applying the same obligation in review of adjudications as well. See Laclede Gas Co. v. FERC, 873 F.2d 1494 , 1498 (D.C. Cir. 1989). The D.C. Circuit has also applied some obligation on agencies to consider “facially reasonable alternatives” presented by a party to an adjudication. Id. (“[T]he agency must either consider those alternatives or give some reason . . . for declining to do so.”). With this statute, which specifically mandates consideration of one particular alternative, and where identifying that 42 Defendants argue the opposite, claiming that “the word ‘available’ modifies and limits the requirement to ‘consider.’” Defs.’ Resp. ¶ 632. Grammatically, this is incorrect, as the adjective “available” modifies the noun “setting.” The descriptor “available” limits which settings must be considered. Compared to some potential alternatives, the use of the word “available” might be said to expand the Secretary’s inquiry obligation. For instance, the statute might have required that the Secretary “consider placement in the least restrictive setting known to the Secretary,” in which case no inquiry would have been required. 158 alternative requires awareness of others, the agency’s obligation to know what alternatives are on the table is even clearer. The Secretary’s third obligation under the statute provides some guidance as to what kinds of settings must be considered. All age-outs “shall be eligible to participate in alternative to detention programs, utilizing a continuum of alternatives based on the alien’s need for supervision, which may include placement of the alien with an individual or an organizational sponsor, or in a supervised group home.” § 1232(c)(2)(B). 43 Any officer tasked with making a custody determination should know to evaluate these programs as possible least restrictive settings. Here, “alternative to detention programs,” in lowercase, does not refer only to the monitoring programs administered by ICE’s ATD unit, but refers instead to anything ICE might do with an age-out other than place them in adult detention. ICE bonds also fit this description, though they are not necessarily a “program.” Defendants argue that because ATD monitoring programs are more restrictive than release of an age-out on his or her own recognizance to a 43 Plaintiffs have suggested that ICE must “make available to each 18 year old ‘alternative to detention programs, utilizing a continuum of alternatives . . . which may include placement … with an individual or an organizational sponsor, or in a supervised group home.’” Pls.’ Br. ¶ 620. Defendants dispute whether ICE is obligated to “make available” the “alternative to detention programs” described in the second sentence of Section 1232(c)(2)(B). Defs.’ Resp. ¶ 620 While this language is referenced in both Counts of the Amended Complaint, Am. Compl. ¶¶ 105, 109, the Court does not find it material to their resolution as they have been argued in this case, and thinks the parties may be talking past one another. The statute explicitly states that age-outs “shall be eligible to participate in alternative to detention programs” that “utilize[e] a continuum of alternatives.” § 1232(c)(2)(B). And ICE has done just that—many age-outs have been released to sponsors, and at least some have been placed in the ATD program or released on ICE bonds. The Plaintiffs may simply be rephrasing ICE’s obligation to make age-outs eligible for placement as an obligation to make placements available to age-outs. To the extent that there is a genuine dispute over what ICE is obligated to offer, it is outside the scope of this litigation. This case has not been focused on whether ICE is obligated to maintain the ATD program or to offer bonds, but only over whether it must consider using these with ageouts. Nor has there been any suggestion that ICE itself must somehow make sponsors available, as opposed to identifying sponsors who are already available. 159 sponsor, ICE officers are not obligated to consider these more restrictive alternatives for age-outs they release on OREC. See Defs.’ Resp. ¶ 632. In one sense this is correct—ATD need not be considered as “the least restrictive setting available” if OREC is available, but it must always be evaluated in order to determine what available setting will be least restrictive for a given age-out. ATD (and ICE bonds) are always on the table and always less restrictive than adult detention. They are not to be considered as an alternative to OSUP or OREC, as many FOJCs seem to think, but as an alternative to detention. Placement with individual or organizational sponsors or in group homes are also expressly contemplated by Section 1232(c)(2)(B). An officer making a custody determination who does not evaluate these alternatives and does not inquire about them is unlikely to be able to identify the least restrictive setting available and to consider placing the age-out there. Again, this is going to require some inquiry or research, even if neither word is used in the statute. If ORR, an age-out’s attorney, or a third party comes forward with a potential sponsor, the inquiry or research obligation may be minimal and may simply entail confirming that there is nothing wrong with the proposed placement. The statute cannot reasonably be interpreted to require any officer to make an indefinite search that goes beyond the requirements of reasonable diligence. See Natural Res. Def. Council, Inc. v. Morton, 458 F.2d 827, 837 (D.C. Cir. 1972) (observing, when discussing an agency’s obligation to “study, develop, and describe appropriate alternatives” under the National Environmental Protection Act, 42 U.S.C. § 4332 (E), that “[t]he statute must be construed in the light of reason” and “not to demand what is, fairly speaking, not meaningfully possible, given the obvious, that the resources of energy and research—and time— available . . . are not infinite.”). In many cases, once the officer identifies a sponsor to whom an age-out can be released on OSUP or OREC, there will be no need to continue searching, as no 160 other form of release will be less restrictive. If, after a reasonable search involving review of—at the very least—proposed sponsors, material in the age-out’s file and on the Form I-213, and known shelters, no placement is “available” for a particular age-out, then sponsorship may not be an option and the officer may be left to evaluate only whatever available placements do not require a sponsor. The obligation to consider alternatives is placed on the Secretary. The Secretary, through the Department’s many field offices, is already aware of a multitude of available placement options that officers have used in the past and that advocacy groups have encouraged the agency to use. Focusing on whether a particular FOJC received materials from ORR or had time to research available placement options distracts from the fact that the obligation is agency-wide. The Secretary, acting through his or her subordinates, knows what options are available and fails to fulfill his or her statutory obligations when there is a failure to make sure those subordinates consider those known alternatives. This means that organizational sponsors outside a field office’s AOR should, in many cases, be treated as available. The Secretary already knows that La Posada and Casa Marinella in Texas and Viator House in Chicago, among others, accept age-outs from around the country, because the Secretary—acting through ICE—has released Miami age-outs to those shelters. See Trial Tr. at 1317:14–25 (Lehner); see also id. at 845:20–847:14 (Enriquez) (identifying organizational sponsors around the country that Catholic Charities has suggested to ORR); id.. Nothing in the text of the statute puts any geographic limit on what counts as “available.” Geography does not appear to be a relevant consideration at any other stage of the process, as there appears to be no rationale (other than space) for why age-outs are placed in any particular ORR shelter (and thus processed by the closest ICE field office). And even if geographic limits were reasonable, there 161 is little reason behind any AOR-based rule, as opposed to a rule based on distance, or even on state lines. 44 The statute does not place any time limits on ICE’s duties under the statute, either on the front end or the back end. Section 1232(c)(2)(B) provides that “[i]f a [UAC] reaches 18 years of age,” the Secretary’s obligations are triggered. It does not say that the Secretary—through ICE—cannot prepare to carry out those obligations until the UAC turns eighteen. A number of ICE field offices view their obligations with regard to age-outs as limited to only the day the ageout turns eighteen. See supra ¶¶ 185 (Houston), 198 (Miami), 213 (New York). The statute does not require this. This is not to say that FOJCs are necessarily obligated to begin working on ageouts’ cases earlier than they do. There may be sound reasons, based in other statutes, regulations, or policies, that require FOJCs to structure their time in certain ways. On the back end, ICE is correct that the statute does not require “periodic re-assessment” of a decision once it has been made. E.g., Defs.’ Resp. ¶ 119. But ICE is not relieved of its obligation to follow VAWA at the end of the day on the age-out’s eighteenth birthday. If an age-out is detained without proper consideration of alternatives, that age-out is owed a compliant decision. This is not re-assessment, but simply the proper assessment that was required in the first place. Officer Zanello’s testimony describing how the Phoenix office would bring age-outs back from detention the day after their birthday shows that there are no legal or practical barriers to an FOJC doing 44 For instance, the Los Angeles-based Southern California AOR lacked identified organizational sponsors. (The Court suspects one or more could probably be located somewhere in that highly populated area of the country.) But organizational sponsors did exist in San Francisco. See PX-463 (list of identified organizational sponsors). The San Francisco office’s AOR included Hawaii and Guam. Trial Tr. at 2673:8–15. If the statute only allowed placement of an age-out with an organizational sponsor in the same AOR this would mean that an age-out in Guam could be placed with an organization in San Francisco, but that an age-out in Los Angeles could not be. 162 what is necessary to make a decision that complies with the statute. 45 Trial Tr: at 3834:15– 3835:5 (Zanello). In short, Section 1232(c)(2)(B) has nothing to say on when exactly FOJCs must start or stop working. Section 1232(c)(2)(B) only requires that ICE begin early enough to fulfill its obligations under the statute, and that it continue working it fulfills those obligations. Testimony from the San Antonio office suggests that the time period required to place an age-out with a sponsor is not too great. Id. at 2764:21–25 (Munguia) (testifying that three days to a week is sufficient time to research sponsors and make placements). So, a wise allocation of resources may not require starting the process too far in advance. But the self-imposed 24-hour limit on the process followed by some offices has no basis in the statute and is unreasonable if the decisionmaking process required has not been completed in that artificial time frame. In short, the statute explicitly requires three things and implicitly requires a fourth from the Secretary and thus from the ICE officers executing his instructions. First, they must take into account the age-out’s statutory risk factors—danger to self, danger to community, and risk of flight. With these in mind, they must second consider placing each and every age-out in the least restrictive setting available for that individual. This implicitly requires that officers be aware of what settings are available, and that they take affirmative steps to identify available settings when they are not self-evident, in order to determine which will be the least restrictive for that age-out. Third, and more broadly, DHS and ICE must make a variety of detention alternatives available for age-outs. The Court now turns to evaluate whether Defendants have fulfilled these obligations. 45 The Phoenix process described by Officer Zanello is far from perfect, of course, because it entails overnight placement of the age-out in adult detention while the process of attempting to comply with the statute is ongoing. That overnight detention could violate the statute. Nonetheless, Phoenix’s continued efforts are preferable to simply giving up at the end of the day as many other field offices do. 163 D. Merits The Court now turns to the merits. Plaintiffs’ two claims challenge the same conduct under two distinct provisions of the APA. Count I alleges that ICE is violating 5 U.S.C. § 706 (2) by acting in a manner that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” by failing to consider placements for each Plaintiff in the “least restrictive setting available after taking into account [their] danger to self, danger to community, and risk of flight” and by failing to make available to Plaintiffs any alternatives to detention. Am. Compl. ¶¶ 98–106. Count II alleges that ICE has ‘‘failed to take a discrete agency action that [the agency] is required to take,’’ in violation of § 706(1) of the APA. Id. ¶¶ 107–111. The Court finds that ICE has violated both provisions of the APA. An agency acts arbitrarily and capriciously in violation of the APA when it makes a decision that is not “based on a consideration of the relevant factors” or fails to “follow[] the necessary procedural requirements.” Overton Park, 401 U.S. at 416–17; see Judulang v. Holder, 565 U.S. 42, 55 (2011) (“[A]gency action must be based on non-arbitrary, ‘relevant factors.’”) (quoting Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)); see id. at 53 (“When reviewing an agency action, [courts] must assess, among other matters, ‘whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.’” (quoting State Farm, 463 U.S. at 43 )). Courts “do not defer to [an] agency’s conclusory or unsupported suppositions.” McDonnell Douglas Corp. v. U.S. Dep’t of the Air Force, 375 F.3d 1182, 1187 (D.C. Cir. 2004) (citing State Farm, 463 U.S. at 43 ). In an adjudication like the decisions at issue here, an agency must “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” NetCoalition v. SEC, 615 F.3d 525 , 532 (D.C. Cir. 2010) 164 (quoting State Farm, 463 U.S. at 43 ). “[A]n agency’s refusal to consider evidence bearing on the issue before it constitutes arbitrary agency action within the meaning of § 706.” Butte Cnty. v. Hogen, 613 F.3d 190, 194 (D.C. Cir. 2010) (citing State Farm, 463 U.S. at 43 ; Comcast Corp. v. FCC, 579 F.3d 1, 8 (D.C. Cir. 2009)). 46 As the Court has already held, there is sufficient “law to apply” here for the Court to review the agency’s decisions. See MTD and Class Cert. Op. at 48–49. While the statute is permissive in that it does not require any particular placement for any particular age-out or group of age-outs, a permissive statute is not a blank check for an agency. On the question of whether 46 Defendants misunderstand the proper standard for review of an agency action. Ostensibly quoting Department of Commerce v. New York, 139 S. Ct. 2551, 2579 (2019), they argue that “Plaintiffs must show that the agency ‘entirely failed to consider an important aspect of the problem.’” Defs.’ Br. CL ¶ 22 (emphasis removed). But this quotation does not appear in Department of Commerce. Justice Breyer’s concurrence in that case comes closest, observing that an agency “ordinarily fails to meet [the arbitrary and capricious] standard if it has ‘failed to consider an important aspect of the problem.’” Dep’t of Comm., 139 S. Ct. at 2585 (quoting State Farm, 463 U.S. at 43 ) (Breyer, J., concurring in part and dissenting in part). The full phrase as Defendants quote it appears in State Farm, but there “entirely fail[ing] to consider an important aspect of the problem” is only one of several ways in which an agency might fail arbitrary and capricious review. State Farm, 463 U.S. at 43 . Public Citizen v. Fed. Motor Carrier Safety ,Admin., 374 F.3d 1209 (D.C. Cir. 2004), which Defendants also cite, Defs.’ Br. CL ¶ 22, makes the same point. See Public Citizen, 374 F.3d at 1216 . The agency in that case failed arbitrary and capricious review because it had “wholly failed to comply with [a] specific statutory requirement””,”” but the D.C. Circuit never suggested that this was the only way an agency could fall short of that standard. Defendants incorrectly suggest that a sufficient condition is a necessary one. Elsewhere in their briefing, Defendants invoke the correct but irrelevant rule that “[t]he ‘arbitrary and capricious’ standard is particularly deferential in matters implicating predictive judgments” concerning the future. Defs.’ Br. CL ¶ 53 (quoting Rural Cellular Ass’n v. F.C.C., 588 F.3d 1095, 1105 (D.C. Cir. 2009)). In the case Defendants cite, “the FCC [was required to] make predictive judgments about the effects of increasing subsidies,” leading the Circuit to note that “certainty is impossible.” Rural Cellular Ass’n, 588 F.3d at 1105 . None of the relevant factors here are forward-looking or predictive in the same way. The Secretary is instructed to evaluate risk factors in the moment, arguably with an eye toward future risk, but not in any sense beyond the everyday sense that any agency decision will have effects in the future. The normal—still reasonably high—level of deference afforded to an agency under the arbitrary and capricious standard is thus sufficient and appropriate here. 165 a statute allows for judicial review, Courts have said that “[l]anguage allowing for discretion does not create unlimited discretion.” Am. Petroleum Tankers Parent, LLC v. United States, 943 F. Supp. 2d 59, 67 (D.D.C. 2013) (quoting Inova Alexandria Hosp. v. Shalala, 244 F.3d 342, 348 (4th Cir. 2001)). A similar principle is at play when, as here, a permissive statute does not dictate much by way of outcomes but only requires consideration of certain factors. Judicial review is, if anything, even more important in situations like these if the agency has declined to issue regulations or even provide guidance to its personnel that might cabin the exercise of their considerable discretion. See Heckler v. Chaney, 470 U.S. 821, 853 (1985) (Marshall, J., concurring) (“[An] agency may well narrow its own enforcement discretion through historical practice, from which it should arguably not depart in the absence of explanation, or through regulations and informal action.”). 47 If agency leadership takes no steps to ensure that discretion is exercised with the thoughtfulness the statute requires, then it falls to courts to scrutinize more closely the decisions being made further down the agency chain-of-command. Defendants argue that an agency’s “obligation to ‘consider’ is a minimal one.” Defs.’ Br. CL ¶ 25; see id. at 26. They do not explain what exactly they mean by minimal, but the cases they cite do not establish that it is necessarily easy for an agency to show it considered something. Defendants argue, for instance, that “[w]hen Congress has in fact intended to affirmatively constrain [an] agency . . . it has done so in no uncertain terms,” by requiring more than simply consideration. Id. (quoting Nat’l Endowment for the Arts v. Finley, 524 U.S. 569 , 47 “[A] large number of agency oversight and review structures are agency created, ranging from informal rules governing what decisions an employee must run past her manager to formal complaint mechanisms to express requirements of sign-off by different subunits or highlevel approval for certain decisions. These internal administrative structures . . . are central to controlling the actions of agency personnel and determining how the agency operates.” Gillian E. Metzger & Kevin M. Stack, Internal Administrative Law, 115 Mich. L. Rev. 1239 , 1253 (2017). 166 581 (1998); see also id. (citing United States v. Bruce, 285 F.3d 69, 74 (D.C. Cir. 2002) (“To ‘consider’ means to ‘reflect on,’ ‘think about,’ ‘deliberate,’ ‘ponder’ or ‘study’ . . . . It does not mean to ‘adhere to,’ ‘be bound by’ or ‘follow.’” (quoting Webster’s Third New International Dictionary, Unabridged 483 (1993))). These and other cases contrast consideration with other requirements Congress could have imposed. The fact that different or more restrictive instructions might have been given does not suggest that the obligation to “consider” or “take into account” factors is, on its own terms minimal or unimportant. Under arbitrary and capricious review the agency is “required to identify the standard and explain its relationship to the underlying regulatory concerns.” WorldCom, Inc. v. FCC, 238 F.3d 449, 462 (D.C. Cir. 2001). Whether this obligation is “minimal” depends on the agency and the decision at issue. 48 For an age-out custody determination, the relevant factors are the age-out’s danger to self, danger to community, and risk of flight, along with availability of certain settings for a particular age-out’s placement. Each of these is identified in and the Secretary is instructed by Congress to “tak[e] into account” the three risk factors. The Court cannot “substitute its judgment for that of the agency,” when evaluating how these factors were weighed in the final custody determination. State Farm, 463 U.S. at 43 . “Agencies . . . have expertise and experience in administering their statutes that no court can properly ignore.” Judulang, 565 U.S. at 53 . Nonetheless, “courts retain a role, an important one, in ensuring that agencies have engaged in reasoned decisionmaking.” Id. The Court therefore “must assess, among other 48 In addition, other definitions could be identified that suggest more of an obligation. Defendants themselves, for example, cite a 1990 edition of Black’s Law Dictionary that defines “consider” as “careful examination” and “[t]o deliberate and ponder over.” Defs.’ Br. CL ¶ 30 (citing Consider, Black’s Law Dictionary 306 (6th ed. 1990)). That definition suggests a higher burden, and one that would not have been carried by the minimal work ICE suggests it needs to do. 167 matters, ‘whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.’” Id. (quoting State Farm, 463 U.S. at 43 ). To evaluate whether Defendants comply with the statute, the Court must evaluate not only the results of the decisionmaking processes, but the process itself. Defendants take issue with this sort of analysis. At many points in Defendants’ briefing, they respond to Plaintiffs’ observation that ICE does not take some particular action by noting that the statute does not require it or that the terms used do not appear in the statute. See, e.g. Defs.’ Resp. at ¶ 621 (arguing, among other things, that the statute “contains no requirement to ‘identify’ the least restrictive setting,” id. at p.445, “does not require Defendants to take any specific actions with regard to sponsors,” id. at p.446, and does not require Defendants “‘to do ‘diligence,’ to ‘vet,’ or to ‘identify’ sponsors,” id. at p.448). In many instances, Defendants are correct that the statute does not require the agency to take the particular steps described. The statute does require something of the agency, though, and what it requires—considering certain alternatives and taking certain factors into account—is not directly observable. Observing what steps officers do or do not take is highly probative when it comes to determining whether the largely mental decisional processes the statute dictates are being followed. For example, while the statute does not require that ICE accept proposed sponsors suggested by third parties, a policy stating that such suggested sponsors will be evaluated would be evidence that ICE is making an effort to identify the least restrictive alternative in every case. Defendants’ own arguments reveal that they understand the concept that procedures can reveal compliance, as they argue that the filling out of an AORW demonstrates compliance with the statute, see Defs.’ Br. CL ¶ 62, even though the statute does not mention AORWs or require that they be filled out, see Defs.’ Resp. ¶ 621 at p.456 (“The statute does not require any of the specific documentation aids that DHS has chosen 168 to use to help confirm that its officers comply with Section 1232(c)(2)(B).”). It is with this in mind that the Court has reviewed, in such detail, the instructions ICE gives its officers and the processes they follow. Defendants argue that their documentation of age-out custody determinations shows they complied with the statute, see Defs.’ Br. CL ¶ 62, but the AORWs do not prove ICE’s compliance with the statute. Both versions of the AORW ask the FOJC to make a custody recommendation while “Considering [the] Least Restrictive Setting Available.” Supra ¶¶ 56, 60. Each also asked whether certain detention alternatives either were or were not considered. Id. The AORWs fail to establish that proper consideration was given for a number of reasons. As the Court has observed, the AORWs are not a reliable record of ICE’s decisionmaking process because so many were produced and edited after the fact by officers other than the ones who actually made the decisions. Supra ¶¶ 62–66. 49 This is to say nothing of the over three hundred age-outs who were processed without having AORWs filled out at all during the time period when ICE was attempting to document its decisions. See supra ¶¶ 67–69. Even an AORW filled out in a timely manner by the correct FOJC would not show that the officer had actually considered the least restrictive setting available rather than simply 49 Defendants argue, in a Notice of Supplemental Authority, ECF No. 330, that a recent case from the District of New Mexico supports their contention that a “properly completed [AORW] satisfies the requirement in section 1232(c)(2)(B) to consider the least restrictive setting available.” Id. at 2 (citing Rodas Godinez v. ICE, No. 2:20-cv-446 KWR/SMV, 2020 WL 3402059 (D.N.M. June 19, 2020)). The Court that decided that case was looking at only one AORW, reflecting the petitioner’s custody determination, and only held that this one particular AORW satisfied section 1232(c)(2)(B)—not that AORWs do so as a general matter. See Rodas Godinez, 2020 WL 3402059 , at *7 (“[T]he worksheet shows that [ICE] considered placing Petitioner in the least restrictive setting . . . .” (emphasis added)). Further, that court was ruling on a Motion for a Temporary Restraining Order, while this Court has heard weeks of testimony and developed a broader sense of the processes by which ICE officers fill out AORWs. 169 engaged in a superficial box-checking exercise. The form does not ask the officer to identify what the least restrictive setting available is, but only to make a recommendation while “[c]onsidering” it. 50 Nor does it ask the officer to take the statutory risk factors into account while considering placements—the form does not connect the detention recommendation with the yes-or-no boxes addressing the elements Congress required the Secretary. Checked boxes indicating that these factors were taken into account do not demonstrate that they were substantively evaluated or thought about in any meaningful way. Checked boxes are, at best, the kinds of “conclusory or unsupported” statements that do not fulfill an agency’s obligation to provide the reasons for its decisions. McDonnel Douglas Corp., 375 F.3d at 1187 ; cf. Delaware Riverkeeper Network v. FERC, 753 F.3d 1304, 1313 (D.C. Cir. 2018) (“[S]imple, conclusory statements of ‘no impact’ are not enough to fulfill an agency's duty under NEPA.” (quoting Found. on Econ. Trends v. Heckler, 756 F.2d 143 , 154 (D.C. Cir. 1985))). The AORW does not require, for instance, elaboration on what level of danger or flight risk an age-out posed or why. And even if merely checked boxes somehow could indicate sufficient consideration of alternatives, more than half of the AORWs reviewed indicated that alternatives to detention were not considered. Supra ¶ 169. 51 Although “[courts] will . . . uphold a decision of less than ideal 50 The FOJC’s ultimate recommendation regarding where the age-out should be placed is not an identification of the least restrictive setting available because, as this Court has held and as Defendants have repeatedly noted, age-outs are not guaranteed placement in the least restrictive setting available, but only consideration of that setting. This is in contrast to 8 U.SC. § 1232(c)(2)(A) which dictates that a UAC “shall be promptly placed in the least restrictive setting that is in the best interest of the child.” Id. If there were a form documenting UAC placement decisions, it would make sense to assume that the recommended setting was the least restrictive. The Court cannot make the same assumption here because the FOJC’s recommended setting need not be the least restrictive that was considered. 51 Defendants argue not only that completed AORWs are sufficient evidence that ICE followed the statute, but that even partially completed ones demonstrate compliance. Defs.’ Br. CL ¶ 62 (citing State Farm, 463 U.S. at 43 ). Because the Court does not think a completed 170 clarity if the agency’s path may reasonably be discerned,” the Court is unable to find that the proper procedures were followed in making these custody determinations based on the AORWs because they do not show that “the decision was based on a consideration of the relevant factors.” State Farm, 463 U.S. at 43 (quotations omitted). The testimony of ICE officials bolsters the Court’s view of the AORWs and also stands on its own as further evidence that the least restrictive setting available was regularly not considered, that risk factors were not taken into account, and that inquiries were often not made to identify available settings that would have been less restrictive than the default of adult detention. To begin at the top, JFRMU and ICE headquarters make no efforts to monitor whether field officers are complying with the statute. As Defendants acknowledge, no review of the substance of age-out custody determinations is ever conducted. Supra ¶ 145. While officers properly have discretion in how they weigh the statutory factors and in whether they ultimately place an age-out in the least restrictive setting available, they do not have discretion to decline to take those factors into account, or to decline to consider the least restrictive setting available. See P.I. Op. at 32 (“[C]ompliance with 8 U.S.C. § 1232 (c)(2)(B) is mandated by statute and therefore not discretionary.”). JFRMU could review the substance of age-out custody determinations for compliance with the statute even while affording officers a reasonable degree of discretion, but instead ICE headquarters allows field officers to decide whether or not they will follow the statute. One reason that many field officers may choose not to follow the statute’s directives is that JFRMU training materials do not emphasize compliance with the statute, and in fact AORW can demonstrate compliance, it should come as no surprise that an incomplete one also falls short for the same reasons. 171 encourage procedures that are contrary to it. Training materials emphasize that officers should consider the “totality of the circumstances”—a term JFRMU does not define—much more than they emphasize the three statutory risk factors Congress has directed the Secretary to take into account. Supra ¶¶ 109–12. Despite emphasizing this standard, which Chief Harper professed means that ICE “contemplate[s] every possible factor that we’re aware of at the time,” Trial Tr. at 62:4–6, ICE allows its FOJCs to ignore certain obvious facts, for instance, FOJCs tend to overlook the fact that nearly all age-outs lack community ties in the United States, supra ¶ 113; or that a long period of time in ORR may mean an age-out is unlikely to flee, supra ¶ 114; the fact that ATD and ICE bonds are available options for age-outs, supra ¶¶ 139–41; or even the simple fact that officers in certain offices like San Antonio manage to place almost all age-outs with sponsors without incident. ICE also permits FOJCs to disregard potential sponsors recommended by advocates, potential sponsors identifiable in an age-out’s form I-213, and organizational sponsors used in prior instances by their colleagues. Supra ¶ 147, 149, 183, 211, 249. Officers are also instructed to follow the kind of sequential decisionmaking process that the Court has concluded violates the statute by denying age-outs consideration of the least restrictive setting available. Supra ¶¶ 131–33. This is compounded by training that comes close to treating detention as a default. Officers are instructed to run a Risk Classification Assessment tool that, unbeknownst to many officers, never recommends release. Supra ¶¶ 125–30. They are instructed to view age-outs as flight risks based on characteristics that are present for nearly all age-outs. Supra ¶ 113–14. Although the statute does not entitle any particular age-out to any particular placement, giving these sorts of detention-favoring instructions alongside an emphasis 172 on sequential decisionmaking that would skip over the required consideration serves to increase the likelihood that officers will not fulfill their statutory obligations. JFRMU’s training concerning sponsors—or rather its lack thereof—increases the importance of sponsors to age-out custody determinations while simultaneously failing to encourage FOJCs to arrange for sponsors or even to accept proposed ones. JFRMU training suggests that nearly all age-outs are flight risks but also states that this risk can be mitigated by placement with a sponsor. Supra ¶¶ 113–15. FOJCs are unlikely to recommend release if they cannot identify a way of mitigating the flight risk they are trained to see in nearly every case. This means that the availability of sponsors is often critical to an age-out’s receiving consideration of the least restrictive setting available for their placement. Nonetheless, JFRMU does not give field officers any instruction on how to investigate what sponsors are available. Nor does it encourage them to do so. Although JFRMU training materials indicate that sponsors must be “viable,” and although the lack of a viable sponsor is often cited by FOJCs as a reason for detention, JFRMU has never developed any standards for what makes a sponsor viable. Supra ¶¶ 150–51. JFRMU does not require, encourage, or take a position on any number of steps an FOJC might take in order to identify the least restrictive setting available including reaching out to ORR, speaking with an age-out, contacting proposed sponsors, or responding to communications from third parties that have suggested possible sponsors. Supra ¶¶ 146–49. JFRMU also does not instruct ICE field officers on how to place or consider placing age-outs in the ATD program or on ICE bonds. Supra ¶¶ 139–41. The result of JFRMU’s lack of instruction concerning sponsor viability or investigation is that, in many cases, ICE essentially leaves it to ORR—a subcomponent of a different agency serving a different population—to fulfill the statutory obligation to consider placement in the 173 least restrictive setting available. JFRMU officials and many field officers view identifying sponsors as ORR’s job, not ICE’s. Supra ¶¶ 153, 157–58, 160–63, ; see also supra ¶¶ 183–84, 197, 210. As the Court has explained, the statute requires the Secretary to identify potential placements for age-outs. When ICE officers decline to identify sponsors, Defendants thus fail to fulfill these statutory obligations. Even if DHS could somehow delegate to ORR its investigative obligations under Section 1232(c)(2)(B), the kind of vetting and investigation that ORR performs, because it deals with children, is materially different than the investigation DHS needs to do with the now-adult ageout. The effect of pushing the duty to investigate onto ORR is that sponsorship standards designed with a different population in mind are applied to age-outs. The unaccompanied alien children in ORR’s custody are just that—children. Age-outs are eighteen years old—legal adults. It is not at all clear that placement with a sponsor should not be treated as “available” to an eighteen-year-old simply because ORR would not or could not place the same individual with that sponsor when they were a child. As the Court has said before, it stands to reason that ORR should and does have more onerous sponsorship requirements than ICE. P.I. Op. at 31 (citing Office of Refugee Resettlement, Sponsor Handbook, (listing duties for sponsors of unaccompanied alien children, including that they must ensure that the child has access to education and must protect the child’s physical and emotional well-being), https://www.acf.hhs.gov/sites/default/files/orr/frp_8_ sponsor_handbook_english_070716.pdf; see also 8 U.S.C. § 1232 (special statutory protections for unaccompanied minors)). The Court appreciates that ORR may have considered and rejected the sponsor only a few days before a UAC became an age-out, and that little may have changed in the intervening time, but Congress has drawn a bright line at eighteen years of age, and 174 identified in Section 1232(c)(2)(B) what protections are provided for age-outs who cross this line and transfer to ICE’s custody. Sponsorship standards designed for children in ORR care are not among the protections afforded to eighteen-year-olds, and applying these higher standards takes away from the protections Congress did give to age-outs by limiting available placements. Because children need more care and supervision than adults, ORR’s standards are different and higher than what is necessary for adult age-outs. Allowing these heightened standards to dictate which placements are considered for age-outs denies those age-outs of consideration of the least restrictive settings that might be available to them. Regarding the FOJC’s training, Defendants argue that “[t]he FOJCs uniformly testified that they understood that they were to consider the least restrictive setting available in making a custody decision as to an age-out.” E.g., Defs.’ Resp. ¶ 182. The officers’ ability to recite the proper standard at trial after preparation by counsel does not show that they were adequately trained by ICE on how to implement it or that they understood what Congress meant by those words. Even granting that ICE officers did know what they needed to do in order to follow the statute, this would not resolve the key question before the Court, which was what they are actually doing in the field. It would not foreclose the possibility that ICE officers knew how to comply with the statute but knowingly chose not to, or found it impossible to follow appropriate and necessary procedures in an actual work environment with limited resources. Testimony from the FOJCs that they knew what they were obligated to do therefore does not preclude a violation of the statute. Turning from JFRMU’s training to the officers in the field making age-out custody determinations, it is clear that the considerable discretion afforded to these officers is exercised in many field offices in ways that do not follow the statute’s directives. As the Court has 175 explained, FOJCs in a number of ICE’s largest field offices do not make attempts to identify sponsors. See, e.g., supra ¶¶ 184, 198–203, 210–11, 263. This leaves many settings unconsidered, any number of which may be the least restrictive setting available—certainly less restrictive, at least, than the adult detention facilities in which these field offices tend to place their age-outs. An even broader set of FOJCs do not consider ICE’s ATD program or ICE bonds as possible detention alternatives to mitigate flight risk. Supra ¶¶ 169, 186, 204, 214, 240, 251, 258, 264, 269, 273, 279. Again, this leads ICE not to consider certain settings that may be less restrictive than those in which the age-outs are regularly placed. This failure to consider ICE’s ATD program or ICE bonds also runs afoul of the Secretary’s obligation to make age-outs “eligible to participate in alternative to detention programs.” 8 U.S.C. § 1232 (c)(2)(B). They are not meaningfully eligible for these programs if few FOJCs know how to properly use them in the context of age-outs. Defendants argue throughout their briefing “that changes in detention and release rates [do not] correspond to compliance or non-compliance with 8 U.S.C. § 1232 (c)(2)(B).” E.g., Defs.’ Resp. ¶¶ 316. Defendants are correct that there is no direct correspondence, but detention and release rates are nonetheless probative evidence of compliance with the statute. If the statutory factors were being taken into account, and if the least restrictive setting was being considered in every case, it would be expected that changes and differences in detention rates— especially large changes and differences—would correspond to changes and differences in risk factors and sponsor availability. The fact that expert statisticians were not able to identify such a correspondence goes to show that the risk factors are not driving detention decisions. This is not direct evidence of a violation, but it corroborates the other evidence of a violation, because if differences in detention rates cannot be explained by differences in risk factors, it is less likely 176 that risk factors are being taken into account. Of course, no age-out has a right under the statute to any particular placement, so it remains possible in theory that ICE officers are properly taking the risk factors into account and considering the least restrictive setting, but simply choosing not to place age-outs in those least restrictive settings. However, considering the full scope of the testimony, evidence, and expert analysis presented at trial, the Court finds the likelihood of this to be vanishingly small. The evidence at trial was overwhelming that ICE officers making ordinary efforts to identify placements for age-outs tend to be able to find them, and tend to release more of their age-outs. See supra ¶¶ 215–23 (discussing the San Antonio Office), ¶¶ 231–232 (discussing changed procedures at the Harlingen office), ¶¶ 237–240 (same for the Chicago office). And no evidence was presented that engaging in these ordinary efforts and releasing more age-outs than are detained would result in the types of harm that the statute considers relevant—higher rates of absconding, more harm to the age-outs, or to the public. The completed AORWs, to the extent that they are reliable, further show the importance of investigating alternatives. Defendants’ expert explained that taking all the AORW responses at face value suggests that the lack of a viable sponsor was the main predictor of whether an ageout is detained or released. Supra ¶ 290. The Court doubts that taking the responses at face value is a reliable means of analyzing ICE’s decisionmaking, but it is significant that if the Court takes Defendants at their word, sponsorship and not risk factors, is the most determinative piece of information captured on the form. Plaintiffs’ expert analysis, which the Court found more reliable than Defendants’ rebuttal analysis, demonstrates that the statutory risk factors are not driving ICE’s detention decisions. Different field offices detain age-outs at vastly different rates. If age-outs’ danger to self, danger to community, and risk of flight were being taken into account, it could be expected that variations in identifiable evidence of these risk factors between 177 age-outs would correspond to whether or not they were detained. Plaintiffs’ expert did not find such a correlation. Instead, he concluded, and the Court found, that variations in the statutory risk factors do not explain the varying rates of age-out detention between field offices. Supra ¶¶ 291–93, 299. The field office matters much more than the age-out’s history and characteristics, and some offices—Houston, Miami, and New York—detain nearly all the ageouts they encounter. And this despite no evidence that these offices’ age-out populations differed in any material way from their sister offices with low detention rates. Needless to say, the location of the ICE field office where the age-out is processed is not a proper consideration in the agency’s decisionmaking. “That fact is as extraneous to the merits of the [age-out’s] case as a coin flip would be.” Judalang v. Holder, 565 U.S. 42, 56 (2011) (holding that the Board of Immigration Appeals violated the APA’s arbitrary and capricious standard when it “hing[ed] a deportable alien’s eligibility for discretionary relief on the chance correspondence between statutory categories—a matter irrelevant to the alien’s fitness to reside in this country,” id. at 53 ); see also id. at 58–59 (“Judge Learned Hand wrote in [an] early immigration case that deportation decisions cannot be made a ‘sport of chance.’” (quoting Di Pasquale v. Karnuth, 158 F.2d 878, 879 (2d. Cir. 1947)). ICE violates the APA by making the outcomes of its detention decisions depend not on the factors Congress has identified as relevant but instead on the field office making the decision. The testimony of field officers offered some explanation for why different offices detain at different rates. The explanations appear to have little to do with the statutory risk factors or sponsor availability and everything to do with the amount of effort made by ICE field officers exercising their considerable discretion under ICE policies. Offices with the highest detention rates are those where officers make among the least efforts to identify sponsors: Houston, Miami, 178 and New York City. Supra ¶¶ 184, 198–203, 210–11. In San Antonio, the detention rate has been consistently low, officers do try to identify sponsors, and they are successful in doing so. Supra ¶¶ 219–20. At offices where the detention rate has fluctuated, fluctuations can be explained by external factors—changes in personnel or responses to litigation—not by changes in the population of age-outs being processed. Supra ¶¶ 227, 237, 247, 254. If field officers at any of these locations were taking the statutory risk factors into account, the detention rate would not be expected to swing so significantly at certain offices. In short, when ICE officers attempt to identify settings for age-outs that are less restrictive than adult detention, they are able to do so almost without fail. The Court heard no testimony from any ICE officer describing a field office that was taking steps to identify detention alternatives that was not able to find such alternatives for the majority of age-outs it processed. And, as set forth above, there was no evidence presented that these efforts resulted in the types of harms of which the statute required ICE to take into account. When ICE officers decline to investigate these kinds of options, they are treating as unavailable certain less restrictive settings that ICE is obligated to consider under the statute. *** The statute requires the Secretary of DHS—acting through ICE officers in the field—to take into account the statutory risk factors as they consider placing each age-out in the least restrictive setting available. Doing this necessarily requires making an inquiry aimed at determining what settings are available and which of these is the least restrictive. As was clear from the testimony presented at trial, ICE officers are consistently failing to take either of these steps. Their training does not emphasize the proper considerations or decisionmaking processes and, in fact, gives instructions that are contrary to the statute. Field officers are left with nearly 179 unbridled discretion to make age-out custody determinations however they would like, and this discretion is exercised in ways that does not comply with the agency’s statutory obligations. To put it in terms of the Administrative Procedure Act, ICE has acted in a manner that is “arbitrary, capricious, an abuse of discretion” and—most clearly—“otherwise not in accordance with law” by failing to follow procedures made necessary by Section 1232(c)(2)(B) and to take into account the factors that the statute requires. 5 U.S.C. § 706 (2). In so doing, ICE has also “unlawfully withheld or unreasonably delayed” required consideration of placement in the least restrictive setting available to many members of the certified class. 5 U.S.C. § 706 (1). V. CONCLUSION For the foregoing reasons, the Court finds in favor of Plaintiffs on Counts I and II. The Court will also GRANT each of the parties’ motions to file unredacted versions of portions of their briefing under seal. Pls.’ Mot. to File Pls.’ Proposed Findings of Fact and Conclusions of Law Concerning Liability and Remedies Under Seal, ECF No. 270; Pls.’ Mot. To File Resp. to Defs.’ Br. Under Seal, ECF No. 316; Defs.’ Unopposed Mot. For Leave to File Under Seal, ECF No. 320. An order concerning the remedies the Court finds appropriate to remedy the violations of law found herein is to follow in the near future. Dated: July 2, 2020 RUDOLPH CONTRERAS United States District Judge 180