FOR PUBLICATION

UNITED STATES COURT OF APPEALS

FOR THE NINTH CIRCUIT

EAST BAY SANCTUARY COVENANT; Nos. 18-17274

AL OTRO LADO; INNOVATION LAW 18-17436

LAB; CENTRAL AMERICAN

RESOURCE CENTER, D.C. No. Plaintiffs-Appellees, 4:18-cv-06810-

JST

v.

JOSEPH R. BIDEN, President of the ORDER AND United States; MERRICK B. AMENDED GARLAND, Attorney General; JEAN OPINION KING, Acting Director, Executive Office for Immigration Review (EOIR); ALEJANDRO MAYORKAS, Secretary, U.S. Department of Homeland Security; TRACY RENAUD, Senior Official Performing the Duties of the Director, U.S. Citizenship and Immigration Services; TROY MILLER, Senior Official Performing the Duties of the Commissioner, U.S. Customs and Border Protection; TAE D. JOHNSON, Acting Director, U.S. Immigration and Customs Enforcement, Defendants-Appellants. Appeal from the United States District Court for the Northern District of California Jon S. Tigar, District Judge, Presiding

Argued and Submitted October 1, 2019 San Francisco, California

Filed February 28, 2020 Amended March 24, 2021

Before: Ferdinand F. Fernandez, William A. Fletcher, and Richard A. Paez, Circuit Judges.

Order; Opinion by Judge Paez; Concurrence by Judge Fernandez; Concurrence in Denial of Rehearing En Banc by Judge Paez; Dissent from Denial of Rehearing En Banc by Judge Bumatay; Dissent from Denial of Rehearing En Banc by Judge VanDyke

EAST BAY SANCTUARY COVENANT V. BIDEN 3

SUMMARY *

Immigration / Preliminary Injunctions

The panel filed: 1) an order denying on behalf of the court a petition for rehearing en banc; 2) an amended opinion affirming the district court’s grant of a temporary restraining order and a subsequent grant of a preliminary injunction enjoining enforcement of a rule and presidential proclamation that, together, strip asylum eligibility from every migrant who crosses into the United States along the southern border of Mexico between designated ports of entry; and 3) an amended concurrence.

Addressing briefing on the President’s revocation of the proclamation at issue, the panel agreed with the parties that this appeal was not moot, but declined to hold the case in abeyance while the government reviews the interim final rule at issue. The panel noted that the parties may address further developments and whether any such developments render the case moot on remand.

In the amended opinion, the panel explained that the Department of Justice and Department of Homeland Security adopted an interim final rule in November 2018 (“the Rule”) that makes migrants who enter the United States in violation of a “presidential proclamation or other presidential order suspending or limiting the entry of aliens along the southern border with Mexico” categorically ineligible for asylum. The same day, President Trump issued a presidential proclamation (“the Proclamation”) that

* This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. suspended the entry of all migrants along the southern border of the United States for ninety days, except for any migrant who enters at a port of entry and properly presents for inspection.

Legal services organizations representing asylum- seekers (“the Organizations”) sued to prevent enforcement of the Rule. The district court entered a temporary restraining order enjoining the Rule, and the government appealed, seeking a stay in this court of the district court’s order pending appeal. In a published order, a motions panel denied the stay, and the Supreme Court denied a stay as well. The district court issued an injunction barring enforcement of the Rule, the government appealed, and this court consolidated the two appeals.

First, addressing the effect of the motions panel’s order on the present panel’s decision, the panel concluded that a published motions panel order may be binding as precedent for other panels deciding the same issue, but it is not binding here. The panel explained that this is because the issues are different: in deciding whether to stay a preliminary injunction pending appeal, the motions panel was predicting the likelihood of success of the appeal, meaning it was predicting rather than deciding what the merits panel will decide; however, in resolving the merits of a preliminary injunction appeal, the merits panel was deciding the likelihood of success of the actual litigation. The panel explained that such a predictive analysis should not, and does not, forever decide the merits.

The panel noted that there may be circumstances where a motions panel does answer the same legal question that is presented to the merits panel, observing that this court addressed such a circumstance in Lair v. Bullock, 697 F.3d 1200 (9th Cir. 2012), in which the merits panel held that it

EAST BAY SANCTUARY COVENANT V. BIDEN 5

was bound by the motions panel’s published decision on a particular issue, where the motions panel answered precisely the same question that was before the merits panel. Accordingly, the panel noted that, to the extent the issues share predictive similarity here, the motions panel may be persuasive but not binding.

Next, the panel considered the government’s challenge to the court’s jurisdiction. First, the panel concluded that the Organizations had established organizational standing by showing that the Rule perceptibly impaired their ability to perform their services. Second, the panel rejected the government’s argument that the court should avoid interfering with the Rule on the ground that the power to expel or exclude aliens is a fundamental sovereign attribute exercised by the government’s political departments largely immune from judicial control. The panel explained it was responsible for reviewing whether the government has overstepped its delegated authority under the Immigration and Nationality Act (“INA”) and encroached upon Congress’s legislative prerogative. Third, the panel rejected the government’s argument that three statutory provisions, 8 U.S.C. §§ 1252(e)(3), 1252(a)(5), and 1252(b)(9), divested this court of jurisdiction. The panel explained that none of these provisions have any bearing on the Rule because they govern judicial review of removal orders or challenges inextricably linked with actions taken to remove migrants from the country. Finally, the panel concluded that the Organizations fell within the zones of interests of the INA.

The panel next addressed the Organizations’ likelihood of success on the merits of their claims under the Administrative Procedure Act (“APA”). Applying the Chevron framework, the panel held that the Rule conflicts with the INA’s section on asylum, which states that a migrant may apply for asylum when she is “physically present in the United States” or “arrives in the United States (whether or not at a designated port of arrival…)[.]” 8 U.S.C. § 1158(a)(1). Because the Rule requires migrants to enter at ports of entry to preserve their eligibility for asylum, the panel explained that it is effectively a categorical ban on migrants who use a method of entry explicitly authorized by Congress in § 1158(a).

The panel further concluded that, even if the text of § 1158(a) were ambiguous, the Rule fails at the second step of Chevron because it is an arbitrary and capricious interpretation of the statute. The panel explained that the BIA and this court have long recognized that a refugee’s method of entry is a discretionary factor in determining whether the migrant should be granted relief, but that the method of entry should be carefully evaluated in light of the harsh consequences that may result. Thus, the panel concluded that, given the Rule’s effect of conditioning asylum eligibility on a factor that has long been understood as worth little if any weight in adjudicating asylum applications, it is an arbitrary and capricious interpretation of § 1158(a).

The panel also concluded that the Rule is unreasonable in light of the United States’s treaty obligations under the 1951 United Nations Convention Relating to the Status of Refugees (“1951 Convention”) and the 1967 United Nations Protocol Relating to the Status of Refugees.

The panel briefly addressed the procedural arguments raised by the parties regarding whether the Rule was invalid because it was issued without public notice and comment or complying with the thirty-day grace period required by the APA. The panel concluded that the Rule likely does not

EAST BAY SANCTUARY COVENANT V. BIDEN 7

properly fall under the good-cause exception or the foreign- affairs exception to these procedural requirements.

Next, the panel concluded that the Organizations had demonstrated a sufficient likelihood of irreparable injury to warrant injunctive relief, explaining that the Organizations had shown that they will suffer a significant change in their programs and a concomitant loss of funding absent a preliminary injunction. The panel also concluded that the public interest weighs sharply in the Organizations’ favor.

Finally, addressing the scope of the remedy, the panel concluded that the district court did not abuse its discretion in issuing an injunction preventing any action to implement the Rule. The panel noted that the Organizations do not limit their potential clients to refugees who enter only at the Mexican border with California and Arizona, and that the government had not proposed an alternative form of the injunction that accounts for the scope of the harms, but applies only within the Ninth Circuit.

Concurring in the result, Judge Fernandez wrote that he concurred in the majority opinion because, and for the most part only because, he believes that this panel is bound by the motions panel’s published decision in this case. Judge Fernandez wrote that the panel is bound by the law of the circuit as well as the law of the case doctrine. Judge Fernandez noted that the majority had now taken steps, in its amended opinion, to obscure its attacks on the doctrines of law of the case and law of the circuit, which were set forth in its original opinion. Writing that his concurrence nevertheless remained the same and was based on the same reasoning, Judge Fernandez stated that the majority’s attempt to pull an invisibility cloak over the mainspring of its attacks cannot hide damage wrought by them. Concurring in the denial of rehearing en banc, Judge Paez, joined by Judge W.Fletcher, wrote to respond to several of the arguments in Judge Bumatay’s and Judge VanDyke’s dissents. Responding to Judge Bumatay’s contention that the amended majority opinion’s application of precedent relating to organizational standing deviated from John Marshall’s and James Madison’s respective visions of Article III, Judge Paez wrote that the majority opinion’s organizational standing holding was consistent with Supreme Court precedent, Article III, and decisions of this and other courts.

Judge Paez also responded to Judge VanDyke’s attacks on the integrity of the majority opinion. First, Judge Paez wrote that Judge VanDyke accused the majority of impropriety in addressing the effect of the motions panel’s published opinion on the merits panel’s subsequent consideration of the appeal. Noting that this potentially dispositive issue was expressly raised by the parties’ briefs, Judge Paez concluded that it was clearly appropriate for the majority to address the issue. Second, Judge Paez wrote that Judge VanDyke complained that the majority engaged in “mischief” by amending its opinion during the en banc process. Judge Paez described the court’s en banc process. Judge Paez wrote that, consistent with this process, the panel majority here considered the debate about law-of-the-case where a motions panel has published an order, the panel majority was persuaded to revise its opinion by the discussions in the memorandum exchange, and the revisions, which were proposed to the court during the memorandum exchange period, were reflected in the amended majority opinion.

Dissenting from the denial of rehearing en banc, Judge Bumatay, joined by Judges Ikuta, Bennett, R. Nelson, Lee,

EAST BAY SANCTUARY COVENANT V. BIDEN 9

and VanDyke, wrote that judges are not “Platonic Guardians” of the nation’s public policies; judges have no business standing athwart the choices of the political branches no matter how misguided the judges believe them to be. That fundamental limitation, Judge Bumatay wrote, is even more pronounced in the immigration context.

Judge Bumatay concluded that the majority opinion ignored constitutional limits on jurisdiction by stretching organizational standing doctrine beyond Article III’s reach, inaugurating a new standard of organizational standing— one that allows an organization to challenge a disfavored government policy by merely asserting that the change could result in “one less client” or cause it to shift resources to better support its mission.

Judge Bumatay also wrote that the majority re-wrote the asylum statute to add a prohibition on the Executive’s authority not found anywhere in the legislative text. Judge Bumatay concluded that the Rule easily fit within the asylum scheme enacted by Congress, explaining that the Rule does nothing to limit the ability of aliens to apply for asylum, as it only restricts the parameters for a successful asylum petition. Judge Bumatay wrote that the panel’s reading essentially conflates applying for and receiving asylum, and the result is that the court has substituted its version of the asylum law for the one actually passed by Congress.

Dissenting from the denial of rehearing en banc, Judge VanDyke wrote separately to emphasize it was never necessary in this case for the panel majority to wade into the issue of the binding effect of motions panels’ published opinions and purport to overrule Lair v. Bullock. The panel majority only did so because of Innovation Law Lab v. Wolf, the other case it decided the same day it decided this case. That was the panel majority’s “mischief”—not the panel amending its opinion during the en banc process.

Further, Judge VanDyke wrote that while it is true that the amended rationale in the East Bay majority’s opinion does not represent the ham-fisted upheaval its original opinion did, it merely trades one evil for another. Instead of rudely casting aside precedent, it distinguishes—rather than overrules—Lair, creating a new standardless standard that will allow any merits panel to disregard a motions panel’s published decision resolving virtually identical claims.

EAST BAY SANCTUARY COVENANT V. BIDEN 11

COUNSEL

Scott Grant Stewart (argued) and Sarah Harrington, Deputy Assistant Attorneys General; Francesca Genova and T. Benton York, Trial Attorneys; Erez Reuveni, Assistant Director; William C. Peachey, Director; August E. Flentje, Special Counsel; Joseph H. Hunt, Assistant Attorney General; Office of Immigration Litigation, United States Department of Justice, Washington, D.C.; for Defendants- Appellants.

Lee P. Gelernt (argued), Judy Rabinovitz, Omar C. Jadwat, Anand Balakrishnan, Daniel Galindo, and Celso Perez, American Civil Liberties Union Foundation, Immigrants’ Rights Project, New York, New York; Katrine Eiland, Jennifer Chang Newell, Cody Wofsy, and Spencer Amdur, American Civil Liberties Union Foundation, Immigrants’ Rights Project, San Francisco, California; Melissa Crow, Southern Poverty Law Center, Washington, D.C.; Mary Bauer, Southern Poverty Law Center, Charlottesville, Virginia; Gracie Willis, Southern Poverty Law Center, Decatur, Georgia; Baher Azmy, Angelo Guisado, and Ghita Schwarz, Center for Constitutional Rights, New York, New York; Christine P. Sun and Vasudha Talla, American Civil Liberties Union Foundation of Northern California Inc., San Francisco, California; for Plaintiffs-Appellees.

Lawrence J. Joseph, Washington, D.C.; Christopher J. Hajec, Director of Litigation, Immigration Reform Law Institute, Washington, D.C.; for Amicus Curiae Immigration Reform Law Institute.

Patrick W. Pearsall, Karthik P. Reddy, and Vaishalee V. Yeldandi, Jenner & Block LLP, Washington, D.C.; Alice Farmer, Office of the United Nations High Commissioner for Refugees, Washington, D.C.; for Amicus Curiae Office of the United Nations High Commissioner for Refugees.

Richard D. Bernstein, Washington, D.C.; Richard Mancino, Willkie Farr & Gallagher LLP, New York, New York; for Amici Curiae Peter Keisler, Stuart Gerson, Carter Phillips, John Bellinger III, Samuel Witten, Ray Lahood, Brackett Denniston, Stanley Twardy, and Richard Bernstein.

Alan E. Schoenfeld, Wilmer Cutler Pickering Hale and Dorr LLP, New York, New York; Alex Gazikas, Wilmer Cutler Pickering Hale and Dorr LLP, Washington, D.C.; Peter S. Margulies, Bristol, Rhode Island; Shoba Sivaprasad Wadhia, University Park, Pennsylvania; for Amici Curiae Professors of Immigration Law.

Margaret L. Carter, Dmitiri D. Portnoi, and Daniel R. Suvor, O’Melveny & Myers LLP, Los Angeles, California; Barbara J. Parker, City Attorney; Maria Bee, Erin Bernstein, Malia McPherson, Zarah Rahman, and Suzanne Dershowitz; Office of the City Attorney Oakland, California; Edward N. Siskel, Corporation Counsel, City of Chicago Department of Law, Chicago, Illinois; Zachary W. Carter, Corporation Counsel, New York City Law Department, New York, New York; for Amici Curiae 21 Counties, Cities, and Local Officials.

Xavier Becerra, Attorney General; Michael L. Newman, Senior Assistant Attorney General; Christine Chuang, Supervising Deputy Attorney General; Shubhra Shivpuri and James F. Zahradka II, Deputy Attorneys General; Office of the Attorney General, Oakland, California; Philip J. Weiser, Attorney General, Denver, Colorado; William Tong, Attorney General, Hartford, Connecticut; Kathleen Jennings, Attorney General, Wilmington, Delaware; Karl A.

EAST BAY SANCTUARY COVENANT V. BIDEN 13

Racine, Attorney General, Washington, D.C.; Clare E. Connors, Attorney General, Honolulu, Hawaii; Kwame Raoul, Attorney General, Chicago, Illinois; Tom Miller, Attorney General, Des Moines, Iowa; Brian E. Frosh, Attorney General, Baltimore, Maryland; Maura Healey, Attorney General, Boston, Massachusetts; Dana Nessel, Attorney General, Lansing, Michigan; Keith Ellison, Attorney General, St. Paul, Minnesota; Aaron D. Ford, Attorney General, Carson City, Nevada; Gurbir S. Grewal, Attorney General, Trenton, New Jersey; Hector Balderas, Attorney General, Santa Fe, New Mexico; Letitia James, Attorney General, New York, New York; Ellen F. Rosenblum, Attorney General, Salem, Oregon; Josh Shapiro, Attorney General, Harrisburg, Pennsylvania; Peter F. Neronha, Attorney General, Providence, Rhode Island; Thomas J. Donovan, Jr., Attorney General, Montpelier, Vermont; Mark R. Herring, Attorney General, Richmond, Virginia; Robert W. Ferguson, Attorney General, Olympia, Washington; for Amici Curiae California, Colorado, Connecticut, Delaware, District of Columbia, Hawaii, Illinois, Iowa, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Jersey, New Mexico, New York, Oregon, Pennsylvania, Rhode Island, Vermont, Virginia, and Washington. 1. The full court was advised of the petition for rehearing en banc. A judge of the court called for rehearing en banc. The matter failed to receive a majority of the votes of the non-recused active judges in favor of en banc consideration. Fed. R. App. P. 35. Rehearing en banc is

DENIED.

2. Attached are Judge Paez’s concurrence to and Judge Bumatay’s and Judge VanDyke’s dissents from the denial of rehearing en banc.

3. The opinion filed on February 28, 2020 is amended as follows:

On page 7, strike the sentence starting, “How strictly the order binds this court,” up and through the sentence on page 13 starting, “We discuss the merits of a stay request[.]”

Replace it with:

The Organizations contend we must affirm the preliminary injunction because the published motions panel order denying a stay of the injunction pending appeal controls our decision on the merits of the preliminary injunction appeal. They argue that reversal of the injunction would amount to overruling the motions panel order, which we cannot do. See Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003) (holding that a published opinion may only be overruled when it is clearly irreconcilable with an intervening higher authority). The published motions panel order may be binding as precedent for other

EAST BAY SANCTUARY COVENANT V. BIDEN 15

panels deciding the same issue, but it is not binding here. This is because the issues are different. In deciding whether the court should stay the grant or denial of a preliminary injunction pending appeal, the motions panel is predicting the likelihood of success of the appeal. That is, the motions panel is predicting rather than deciding what our merits panel will decide. In resolving the merits of a preliminary injunction appeal, our merits panel is deciding the likelihood of success of the actual litigation.

The Supreme Court has recognized this distinction and made clear that it leads to different analyses of the equities. See Trump v. Int’l Refugee Assistance Project, 137 S. Ct. 2080, 2087 (2017). There is more judicial discretion with respect to a stay. See id.; see also Doe #1 v. Trump, 957 F.3d 1050, 1058 (9th Cir. 2020) (“A request for a stay pending appeal is committed to the exercise of judicial discretion.”); 11 WRIGHT & MILLER,

FEDERAL PRACTICE & PROCEDURE

§ 2904, at 702–03 (3d ed. 2012).

In Nken v. Holder the Supreme Court emphasized that “[a]n injunction and a stay have typically been understood to serve different purposes” and a stay halting “the conduct or progress of litigation before the court, ordinarily is not considered an injunction.” 556 U.S. 418, 428, 430 (2009) (internal quotations omitted). Indeed, our court has relied on Nken for the proposition that “there are important differences between a preliminary injunction and a stay pending review…” Leiva-Perez v. Holder, 640 F.3d 962, 966 (9th Cir. 2011) (citing Nken, 129 S. Ct. at 1756–59). A stay “operates upon the judicial proceeding itself,” while a preliminary injunction “direct[s] an actor’s conduct.” Nken, 556 U.S. at 428, 429.

In the government’s appeal, we are charged with determining whether the district court abused its discretion in granting the preliminary injunction, see All. for the Wild Rockies v. Cottrell, 632 F.3d 1127, 1131 (9th Cir. 2011); the motions panel, instead, considered whether the government raised serious questions relating to the propriety of the district court’s preliminary injunction and whether the government would likely prevail on appeal, see Leiva-Perez, 640 F.3d at 965– 66. The question presented to the motions panel is an additional step removed from the underlying merits of the district court’s preliminary injunction. We discuss the merits in “likelihood terms” and exercise restraint in assessing the merits of either question, see Sierra Club, 929 F.3d at 688, but particularly so when considering the “extraordinary request” to stay a preliminary injunction granted by a district court. Barr v. E. Bay Sanctuary Covenant, No. 19A230, 2019 WL 4292781, at *1 (Sept. 11, 2019) (Sotomayor, J., dissenting from grant of a stay).”

EAST BAY SANCTUARY COVENANT V. BIDEN 17

On page 14, insert the following footnote after the citation to Leiva-Perez v. Holder, 640 F.3d 962, 967 (9th Cir. 2011):

There may be circumstances where a motions panel does answer the same legal question that is presented to the merits panel. See Lair v. Bullock, 798 F.3d 736, 747 (9th Cir. 2015) (“Lair II”). In Lair I, the motions panel, while conducting a probabilistic and discretionary analysis of the Nken factors, addressed a pure question of law—whether the Supreme Court had abrogated a relevant circuit precedent. Lair v. Bullock, 697 F.3d 1200, 1206 (9th Cir. 2012) (“Lair I”). The motions panel held that the very question at issue had already been dispositively answered in prior circuit precedent. Id. The motions panel’s specific holding on this pure question of law was neither based on an assessment of probability nor an exercise of discretion—it was necessarily compelled by preexisting binding precedent.

Therefore, the merits panel in Lair II held that it was bound by the motions panel’s published decision in Lair I on this particular issue where the motions panel answered precisely the same question that was before the merits panel. 798 F.3d at 747 (“The motions panel in Lair I explicitly held that Randall did not contain a majority opinion capable of abrogating Eddleman.”) (internal citations omitted). Lair II’s statement in this context that “a motions panel’s published opinion binds future panels the same as does a merits panel’s published opinion,” id., simply confirms our circuit’s stare decisis principle that a question already answered in binding precedent will be controlled by that answer when the same question is presented in the future. See also Miller v. Gammie, 335 F.3d 889, 893 (9th Cir. 2003) (en banc).

But here, unlike in Lair II, none of the questions answered by the motions panel were pure questions of law for which preexisting binding authority necessarily compelled the answer.

On page 14, strike the sentence starting, “The question before us now,” up and through the sentence on page 15 starting with, “Given the preliminary stage of the appellate process[.]”

Replace it with:

The inquiry with respect to the stay differs from the inquiry as to the preliminary injunction. To the extent the issues share predictive similarity, the motions panel may be persuasive but not binding.

On page 15, strike “re-evaluate,” and replace it with “consider.”

On page 16, strike “, as it did previously before the district court and before the motions panel.”

On page 16, strike “renews,” and replace it with “makes.”

EAST BAY SANCTUARY COVENANT V. BIDEN 19

On page 16, strike “before this court.”

On page 18, strike “We agree with the motions panel and the district court,” and replace it with “We conclude.”

On page 27, strike “again” in the sentence before the beginning of subsection B.

On page 28, strike “continue to” from the last full sentence of the page.

On page 34, strike “again” from the sentence starting with, “The government again suggests that the existence of these eligibility bars[.]”

On page 37, strike “the motions panel” from the sentence starting with, “The Attorney General’s interpretation,” and the citation to EBSC II, 932 F.3d at 772–773, that appears after the sentence.

On page 50, strike “we agree with the motions panel that.”

4. An amended opinion is filed concurrently with this order. An amended opinion concurring in the result, which adds a new footnote 1 on page 1 and renumbers the ensuing footnotes, is also filed concurrently with this order. No further petitions for rehearing or rehearing en banc may be filed.

5. In light of Appellants’ February 16, 2021 letter informing the court that the President signed Executive Order 14010 revoking “Proclamation 9880 of May 9, 2019,” Dkt. No. 87, the panel directed the parties to simultaneously file letter briefs addressing whether all or any aspect of this appeal has been rendered moot. We have considered the parties’ responses and agree with them that this appeal is not moot. We decline to accept the parties’ suggestion to hold the case in abeyance while Appellants review the interim final rule at issue. The parties may address future developments related to Appellants’ review of the interim final rule and whether any such developments render the case moot in the district court on remand.

OPINION

PAEZ, Circuit Judge:

Forty years ago, Congress recognized that refugees fleeing imminent persecution do not have the luxury of choosing their escape route into the United States. It mandated equity in its treatment of all refugees, however they arrived. 1

This principle is embedded in the Refugee Act of 1980, which established an asylum procedure available to any migrant, “irrespective of such alien’s status,” and irrespective of whether the migrant arrived “at a land border or port of entry.” Pub. L. No. 96-212, § 208(a), 94 Stat. 102, 105 (1980). Today’s Immigration and Nationality Act (“INA”) preserves that principle. It states that a migrant who arrives in the United States—“whether or not at a designated port of arrival”—may apply for asylum. See 8 U.S.C. § 1158(a).

1 See 125 Cong. Rec. 35,813–14 (1979) (statement of Rep. Holtzman).

EAST BAY SANCTUARY COVENANT V. BIDEN 21

In November 2018, the Departments of Justice and Homeland Security jointly adopted an interim final rule (“the Rule”) which, coupled with a presidential proclamation issued the same day (“the Proclamation”), strips asylum eligibility from every migrant who crosses into the United States between designated ports of entry. In this appeal, we consider whether, among other matters, the Rule unlawfully conflicts with the text and congressional purpose of the INA. We conclude that it does.

I.

The Rule announces a new bar to asylum eligibility. It makes migrants who enter the United States in violation of “a presidential proclamation or other presidential order suspending or limiting the entry of aliens along the southern border with Mexico” categorically ineligible for asylum. See Aliens Subject to a Bar on Entry Under Certain Presidential Proclamations; Procedures for Protection Claims, 83 Fed. Reg. 55,934, 55,952 (Nov. 9, 2018) (codified at 8 C.F.R. §§ 208.13, 208.30). Migrants who are ineligible for asylum under the Rule will also automatically receive negative credible-fear determinations in expedited- removal proceedings. See id. at 55,935, 55,952. Typically, a migrant in expedited-removal proceedings who demonstrates a “credible fear” of persecution must be allowed to present her asylum claim before an immigration judge. See 8 U.S.C. § 1225(b)(1)(A)(ii), (B)(v). A migrant who enters the United States in contravention of a proclamation will instead need to demonstrate a “reasonable fear” of persecution or torture—which is more difficult than establishing a credible fear of persecution—to obtain other forms of relief. See 83 Fed. Reg. at 55,936, 55,952; see also 8 C.F.R. § 208.31(c); 8 U.S.C. § 1225(b)(1)(B)(v). The same day the Departments of Justice (“DHS”) and Homeland Security (“DHS”) adopted the Rule, President Trump issued the Proclamation. The Proclamation suspends the entry of all migrants along the southern border of the United States for ninety days, except for any migrant who “enters the United States at a port of entry and properly presents for inspection.” See Presidential Proclamation No. 9,822, Addressing Mass Migration Through the Southern Border of the United States, 83 Fed. Reg. 57,661, 57,663 (Nov. 9, 2018).

Individually, the Rule and Proclamation have little effect. The Proclamation does not have the force of law, and the Rule only effectuates proclamations. But together, the Rule and Proclamation make asylum entirely unavailable to migrants who enter the country between ports of entry. The magnitude of the Rule’s effect is staggering: its most direct consequence falls on “the more than approximately 70,000 aliens a year (as of FY 2018) estimated to enter between the ports of entry [who] then assert a credible fear in expedited- removal proceedings.” 83 Fed. Reg. at 55,948. These migrants would typically proceed to an asylum hearing before an immigration judge but will now be unable to do so because they have entered the country at a place other than a port of entry.

The day the Proclamation and Rule issued, four legal services organizations that represent current and future asylum-seekers sued to prevent enforcement of the Rule. East Bay Sanctuary Covenant, Al Otro Lado, Innovation Law Lab, and Central American Resource Center of Los Angeles (collectively, “the Organizations”) argued that the Rule was likely unlawful because it was issued without public notice and comment or complying with the thirty-day grace period required by the Administrative Procedure Act

EAST BAY SANCTUARY COVENANT V. BIDEN 23

(“APA”), see 5 U.S.C. § 553(b)–(d). The Organizations also argued that the Rule conflicts with the plain text of the INA and is arbitrary and capricious because it constitutes a severe departure from the Board of Immigration Appeals’s and the Ninth Circuit’s interpretation of asylum practices in the United States.

The district court agreed that the Rule “irreconcilably conflicts with the INA and the expressed intent of Congress” and entered a temporary restraining order enjoining the Rule’s enforcement and ordering the government “to return to the pre-Rule practices for processing asylum applications.” See E. Bay Sanctuary Covenant v. Trump (EBSC I), 349 F. Supp. 3d 838, 844, 868–69 (N.D. Cal. 2018). Eight days after the court’s order, the government filed an appeal and an emergency motion in the district court to stay the temporary restraining order pending appeal. The court denied the stay motion three days later.

The following day, the government sought an immediate stay in our court of the district court’s order pending appeal. In a lengthy published order, a motions panel of this court denied the government’s request to stay enforcement of the court’s order. See E. Bay Sanctuary Covenant v. Trump (EBSC II), 932 F.3d 742, 755, 762 (9th Cir. 2018). Although temporary restraining orders are typically not appealable, the panel concluded that appellate jurisdiction existed under 28 U.S.C. § 1292(a)(1) because the temporary restraining order was effective for thirty days, well beyond the fourteen- day limit imposed by Federal Rule of Civil Procedure 65(b). Id. at 762–63. The government’s application for a stay from the Supreme Court was also denied. See Trump v. E. Bay Sanctuary Covenant, 139 S. Ct. 782 (2018).

While the government’s stay application was pending before the Supreme Court, the Organizations filed a motion for a preliminary injunction in the district court. The arguments presented during the second round of litigation were “nearly identical” to those made during the first. See E. Bay Sanctuary Covenant v. Trump (EBSC III), 354 F. Supp. 3d 1094, 1102 (N.D. Cal. 2018). Relying heavily on the motions panel’s published order, the district court again issued an injunction barring enforcement of the Rule. See id. at 1121.

The government again appeals, arguing that the district court erred when it entered the injunction or that the injunction should at least be narrowed. We consolidated the government’s appeal from the temporary restraining order with the appeal from the preliminary injunction. 2 For the reasons explained below, we agree with the district court that the Rule is inconsistent with the INA, and we affirm the district court’s orders granting preliminary injunctive relief.

II.

We first consider the effect of the motion panel’s order on the present panel’s decision. The Organizations contend we must affirm the preliminary injunction because the published motions panel order denying a stay of the injunction pending appeal controls our decision on the merits

2 Although the Proclamation expired by its terms in February 2019, the President issued a new Proclamation, which did not substantially change the terms of the original Proclamation and extended its effect for an additional ninety days. When that Proclamation expired in May, the President again re-issued it and extended the effect of the initial Proclamation “for an additional 90 days beyond the date when the United States obtains relief from the preliminary injunction of the interim final rule[.]” See Presidential Proclamation No. 9,880, Addressing Mass Migration Through the Southern Border of the United States, 84 Fed. Reg. 21,229, 21,229 (May 8, 2019).

EAST BAY SANCTUARY COVENANT V. BIDEN 25

of the preliminary injunction appeal. They argue that reversal of the injunction would amount to overruling the motions panel order, which we cannot do. See Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003) (holding that a published opinion may only be overruled when it is clearly irreconcilable with an intervening higher authority).

The published motions panel order may be binding as precedent for other panels deciding the same issue, but it is not binding here. This is because the issues are different. In deciding whether the court should stay the grant or denial of a preliminary injunction pending appeal, the motions panel is predicting the likelihood of success of the appeal. That is, the motions panel is predicting rather than deciding what our merits panel will decide. In resolving the merits of a preliminary injunction appeal, our merits panel is deciding the likelihood of success of the actual litigation.

The Supreme Court has recognized this distinction and made clear that it leads to different analyses of the equities. See Trump v. Int’l Refugee Assistance Project, 137 S. Ct. 2080, 2087 (2017). There is more judicial discretion with respect to a stay. See id.; see also Doe #1 v. Trump, 957 F.3d 1050, 1058 (9th Cir. 2020) (“A request for a stay pending appeal is committed to the exercise of judicial discretion.”); 11 WRIGHT & MILLER, FEDERAL PRACTICE & PROCEDURE § 2904, at 702–03 (3d ed. 2012).

In Nken v. Holder the Supreme Court emphasized that “[a]n injunction and a stay have typically been understood to serve different purposes” and a stay halting “the conduct or progress of litigation before the court, ordinarily is not considered an injunction.” 556 U.S. 418, 428, 430 (2009) (internal quotations omitted). Indeed, our court has relied on Nken for the proposition that “there are important differences between a preliminary injunction and a stay pending review…” Leiva-Perez v. Holder, 640 F.3d 962, 966 (9th Cir. 2011) (citing Nken, 129 S. Ct. at 1756–59). A stay “operates upon the judicial proceeding itself,” while a preliminary injunction “direct[s] an actor’s conduct.” Nken, 556 U.S. at 428, 429.

In the government’s appeal, we are charged with determining whether the district court abused its discretion in granting the preliminary injunction, see All. for the Wild Rockies v. Cottrell, 632 F.3d 1127, 1131 (9th Cir. 2011); the motions panel, instead, considered whether the government raised serious questions relating to the propriety of the district court’s preliminary injunction and whether the government would likely prevail on appeal, see Leiva-Perez, 640 F.3d at 965–66. The question presented to the motions panel is an additional step removed from the underlying merits of the district court’s preliminary injunction. We discuss the merits in “likelihood terms” and exercise restraint in assessing the merits of either question, see Sierra Club, 929 F.3d at 688, but particularly so when considering the “extraordinary request” to stay a preliminary injunction granted by a district court. Barr v. E. Bay Sanctuary Covenant, No. 19A230, 2019 WL 4292781, at *1 (Sept. 11, 2019) (Sotomayor, J., dissenting from grant of a stay). Such a predictive analysis should not, and does not, forever decide the merits of the parties’ claims. This sort of “pre- adjudication adjudication would defeat the purpose of a stay, which is to give the reviewing court the time to ‘act responsibly,’ rather than doling out ‘justice on the fly.’” Leiva-Perez v. Holder, 640 F.3d 962, 967 (9th Cir. 2011) (quoting Nken v. Holder, 556 U.S. 418, 427 (2009)). 3

3 There may be circumstances where a motions panel does answer the same legal question that is presented to the merits panel. See Lair v.

EAST BAY SANCTUARY COVENANT V. BIDEN 27

Notably, when acting on the government’s stay motion in this case, the motions panel acknowledged the preliminary nature of the stay proceedings. The panel issued a lengthy opinion with detailed analysis, but repeatedly “stress[ed]” that the case was still “at a very preliminary stage of the proceedings,” and expected that “[f]urther development of the record as the case progresses may alter [their] conclusions.” EBSC II, 932 F.3d at 780. The panel also left open various mixed questions of law and fact for a later court—pointing out, for example, that if “facts develop in the district court that cast doubt on the Organizations’

Bullock, 798 F.3d 736, 747 (9th Cir. 2015) (“Lair II”). In Lair I, the motions panel, while conducting a probabilistic and discretionary analysis of the Nken factors, addressed a pure question of law—whether the Supreme Court had abrogated a relevant circuit precedent. Lair v. Bullock, 697 F.3d 1200, 1206 (9th Cir. 2012) (“Lair I”). The motions panel held that the very question at issue had already been dispositively answered in prior circuit precedent. Id. The motions panel’s specific holding on this pure question of law was neither based on an assessment of probability nor an exercise of discretion—it was necessarily compelled by preexisting binding precedent.

Therefore, the merits panel in Lair II held that it was bound by the motions panel’s published decision in Lair I on this particular issue where the motions panel answered precisely the same question that was before the merits panel. 798 F.3d at 747 (“The motions panel in Lair I explicitly held that Randall did not contain a majority opinion capable of abrogating Eddleman.”) (internal citations omitted). Lair II’s statement in this context that “a motions panel’s published opinion binds future panels the same as does a merits panel’s published opinion,” id., simply confirms our circuit’s stare decisis principle that a question already answered in binding precedent will be controlled by that answer when the same question is presented in the future. See also Miller v. Gammie, 335 F.3d 889, 893 (9th Cir. 2003) (en banc).

But here, unlike in Lair II, none of the questions answered by the motions panel were pure questions of law for which preexisting binding authority necessarily compelled the answer. standing, the district court is, of course, free to revisit this question,” id. at 763 n.6, and reiterating that its conclusions were reached “at [the current] stage of the proceedings,” see id. at 763, 767, 778, 779.

The inquiry with respect to the stay differs from the inquiry as to the preliminary injunction. To the extent the issues share predictive similarity, the motions panel may be persuasive but not binding.

III.

We next consider the government’s challenge to our jurisdiction. The government argues that the Organizations lack Article III standing because they have not suffered a cognizable injury and are outside the zone of interests protected by the INA. The government also makes three arguments before this court: (1) the Organizations lack a “legally protected interest in maintaining their current organizational structure or in the [R]ule’s application to third parties,” Op. Br. of Gov’t at 29, 4 (2) the “immigration context” of the Rule counsels against judicial intrusion, and (3) various portions of the INA divest this court of jurisdiction to entertain this appeal. We address each argument in turn.

A.

The Article III standing inquiry serves a single purpose: to maintain the limited role of courts by ensuring they protect against only concrete, non-speculative injuries. See Lujan v. Defenders of Wildlife, 504 U.S. 555, 583 (1992). Parties

4 We refer to the government’s opening brief as “Op. Br. of Gov’t,” and to the government’s reply brief as “Reply Br. of Gov’t.”

EAST BAY SANCTUARY COVENANT V. BIDEN 29

must have a “personal stake in the outcome” sufficient to ensure the court that, absent judicial review, they will suffer or have suffered some direct injury. See id.

Organizations can assert standing on behalf of their own members, see Friends of the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 528 U.S. 167, 181 (2000), or in their own right, Havens Realty Corp. v. Coleman, 455 U.S. 363, 378– 79 (1982). To determine whether organizational standing requirements have been satisfied, we “conduct the same inquiry as in the case of an individual: Has the plaintiff ‘alleged such a personal stake in the outcome of the controversy as to warrant his invocation of federal-court jurisdiction?’” Havens, 455 U.S. at 378–79. The Organizations therefore have the burden of demonstrating that (1) they have suffered an injury-in-fact, meaning an injury that is “concrete and particularized” and “actual and imminent,” (2) the alleged injury is “fairly traceable” to the defendants’ conduct, and (3) it is “more than speculative” that the injury is judicially redressable. Lujan, 504 U.S. at 560–61.

In Havens, the Supreme Court held that a fair housing organization had standing under the Fair Housing Act where the defendants’ allegedly racial steering practices had frustrated the organization’s ability to assist equal access to housing, and it had to devote “significant resources” to identify and counteract those practices. 455 U.S. at 379. Because the defendants’ practices had “perceptibly impaired” the organization’s ability to provide its services, the Court explained, “there can be no question that the organization has suffered injury in fact.” Id.

We have read Havens to hold that an organization has direct standing to sue where it establishes that the defendant’s behavior has frustrated its mission and caused it to divert resources in response to that frustration of purpose. See Fair Hous. of Marin v. Combs, 285 F.3d 899, 905 (9th Cir. 2002). Of course, organizations cannot “manufacture the injury by incurring litigation costs or simply choosing to spend money fixing a problem that otherwise would not affect the organization at all,” but they can show they “would have suffered some other injury” had they “not diverted resources to counteracting the problem.” La Asociacion de Trabajadores de Lake Forest v. Lake Forest, 624 F.3d 1083, 1088 (9th Cir. 2010); see also, e.g., El Rescate Legal Servs., Inc. v. Exec. Office of Immigration Review, 959 F.2d 742, 745, 748 (9th Cir. 1991).

We conclude that the Organizations have established that the Rule has “perceptibly impaired” their ability to perform the services they were formed to provide. EBSC II, 932 F.3d at 765. This is sufficient for organizational standing. See Combs, 285 F.3d at 904–05.

The Organizations share the same mission of assisting migrants seeking asylum. “[B]ecause the Rule significantly discourages a large number of [asylum-seekers] from seeking asylum given their ineligibility,” the Rule frustrates their mission. EBSC II, 932 F.3d at 766. The Rule has also caused the Organizations to divert their already limited resources in response to the collateral obstacles it introduces for asylum-seekers. East Bay Sanctuary Covenant (“EBSC”) and Innovation Law Lab (“ILL”), for example, are located near Berkeley, California, and in Oregon, respectively, and because most asylum-seekers who enter at a designated port of entry will “remain detained in detention facilities near the border hundreds of miles away,” EBSC III, 354 F. Supp. 3d at 1109 (internal quotation marks omitted), those organizations “cannot represent asylum seekers.” Decl. of Michael Smith at ¶ 6. Unaccompanied minors are

EAST BAY SANCTUARY COVENANT V. BIDEN 31

now often unable to seek asylum alone, and “[s]ince the new rule was announced, Al Otro Lado [(“AOL”)] has been overwhelmed with children who traveled to the southern border of the United States to apply for asylum but now cannot do so.” Supp. Decl. of Erika Pinheiro at ¶¶ 4, 15. Caring for the often nonlegal needs of these unaccompanied children is not part of AOL’s core mission and is “causing a near complete diversion of [AOL’s] resources.” Id. ¶ 16. It has “expended significant resources to send staff to the border as it attempts to shift its programs.” EBSC III, 354 F. Supp. 3d at 1109.

The funding on which the Organizations critically depend is also jeopardized by the Rule. EBSC only “rarely” represents people in removal proceedings. Decl. of Michael Smith at ¶ 8. Because 80 percent of its clients have entered without stopping at a port of entry in the past, EBSC stands to “lose a significant amount of business and suffer a concomitant loss of funding” if these individuals are deemed categorically ineligible for asylum. EBSC III, 354 F. Supp. 3d at 1109 (citing EBSC II, 932 F.3d at 767). AOL and CARECEN explain that the Rule decreases the funding they stand to receive from the California Department of Social Services. AOL often represents detained immigrants in their bond proceedings, and “[s]ince the [R]ule went into effect,” AOL has “not received a single referral for a bond case, as persons who enter without inspection are ostensibly being put into ‘Withholding-only’ proceedings and no longer initially eligible for bond.” Supp. Decl. of Erika Pinheiro at ¶ 22. CARECEN receives from the Department a flat amount of funding per client it assists, and because more of its clients are being put into more time- and resource- intensive withholding proceedings, it will assist less clients and receive less funding. Decl. of Daniel Sharp at ¶ 7. Each organization would have lost clients seeking refuge in the United States had it not diverted resources toward counteracting the effect of the Rule. La Asociacion de Trabajadores de Lake Forest, 624 F.3d at 1088. The Organizations are not required to demonstrate some threshold magnitude of their injuries; 5 one less client that they may have had but-for the Rule’s issuance is enough. In other words, plaintiffs who suffer concrete, redressable harms that amount to pennies are still entitled to relief.

The government advances three additional justiciability arguments. First, the government argues that the Organizations have “no legally protected interest in maintaining their current organizational structure or in the Rule’s application to third parties, which the motions panel did not consider in its analysis.” Op. Br. of Gov’t at 28. This position misunderstands the injury-in-fact inquiry and conflates organizational standing with third-party standing,

5 The government notes that “East Bay Sanctuary Covenant has only ‘around 35 clients who have entered without inspection and [who] expect to file for affirmative asylum in the upcoming months,’” while, “[b]y comparison, the ‘current backlog of asylum cases exceeds 200,000’ and more than 200,000 inadmissible aliens present themselves for inspection at ports of entry annually (even without the additional incentive to do so that the Rule will create).” Op. Br. of Gov’t at 27 n.4.

The comparative magnitude of the harms alleged by the parties, however, is not relevant for standing purposes; “a loss of even a small amount of money is ordinarily an ‘injury.’” Czyzewski v. Jevic Holding Corp., 137 S. Ct. 973, 983 (2017); see also Wallace v. ConAgra Foods, Inc., 747 F.3d 1025, 1029 (8th Cir. 2014) (“The consumers’ alleged economic harm—even if only a few pennies each—is a concrete, non- speculative injury.”); Carpenters Indus. Council v. Zinke, 854 F.3d 1, 5 (D.C. Cir. 2017) (“A dollar of economic harm is still an injury-in-fact for standing purposes.”).

EAST BAY SANCTUARY COVENANT V. BIDEN 33

which the Organizations have conceded is not at issue. 6 An injury-in-fact is “an invasion of a legally protected interest,” see Lujan, 504 U.S. at 560, but this means an interest that is only concrete and particularized and actual or imminent— not an interest protected by statute. This distinction prevents Article III standing requirements from collapsing into the merits of a plaintiff’s claim; “a petitioner’s ‘legally protected interest’ need not be a statutorily created interest,” Ass’n of Pub. Agency Customers v. Bonneville Power Admin., 733 F.3d 939, 950 (9th Cir. 2013), and a plaintiff can have standing despite losing on the merits. See also In re Special Grand Jury 89-2, 450 F.3d 1159, 1172 (10th Cir. 2006) (citing Warth v. Seldin, 422 U.S. 490, 500 (1975) (“[S]tanding in no way depends on the merits of the plaintiff’s contention that particular conduct is illegal….”)).

More recent Supreme Court opinions have described injury-in-fact as “a judicially cognizable interest”— implying that “an interest can support standing even if it is not protected by law…so long as it is the sort of interest that courts think to be of sufficient moment to justify judicial intervention.” In re Special Grand Jury 89-2, 450 F.3d at 1172 (citing Bennett v. Spear, 520 U.S. 154, 167 (1997));

6 Many of the cases cited by the government in support of this proposition do not concern organizational standing under Article III. In O’Bannon v. Town Court Nursing Center, 447 U.S. 773 (1980), the Court addressed whether nursing home residents have a right to an administrative hearing before a state or federal agency hearing before the agency revokes the home’s authority to provide them with nursing care at government expense. Linda R.S. v. Richard D., 410 U.S. 614 (1973), DaimlerChrysler Corp. v. Cuno, 547 U.S. 332 (2006), Arpaio v. Obama, 797 F.3d 11 (D.C. Cir. 2015), and Sure-Tan, Inc. v. N.L.R.B., 467 U.S. 883 (1984) all describe limitations on third-party, not organizational, standing. see also, e.g., Hollingsworth v. Perry, 570 U.S. 693, 707 (2013). Whether the Organizations have a sufficient statutory or otherwise legal basis for their claims is irrelevant at this threshold stage.

The government next argues that we should avoid interfering with DOJ’s and DHS’s decision to adopt the Rule because “[t]he Supreme Court has ‘long recognized the power to expel or exclude aliens as a fundamental sovereign attribute exercised by the Government’s political departments largely immune from judicial control.’” See Op. Br. of Gov’t at 30 (quoting Fiallo v. Bell, 430 U.S. 787, 792 (1977)).

We do not conduct independent policy analyses of executive decisions. But we do “police the separation of powers in litigation involving the executive[.]” In re Cheney, 334 F.3d 1096, 1106 (D.C. Cir. 2003), vacated and remanded on other grounds, 542 U.S. 367 (2004). For this reason, there is a strong presumption favoring judicial review of administrative action, see Bowen v. Mich. Acad. of Family Phys., 476 U.S. 667, 670 (1986); non-reviewability is an exception that must be clearly evidenced in the statute, see Barlow v. Collins, 397 U.S. 159, 166–67 (1970). Without such review, “statutes would in effect be blank checks drawn to the credit of some administrative officer or board.” Bowen, 476 U.S. at 671 (citing S. Rep. No. 752, 79th Cong., 1st Sess., 26 (1945)). Efficient agency administration always requires some authority and responsibility to resolve questions left unanswered by Congress. It does not include the “power to revise clear

EAST BAY SANCTUARY COVENANT V. BIDEN 35

statutory terms.” 7 Utility Air Reg. Grp. v. E.P.A., 573 U.S. 302, 327 (2014).

We are therefore responsible for reviewing whether the government has overstepped its delegated authority under the INA and encroached upon Congress’s legislative prerogative. See 5 U.S.C. § 706(2)(A).

Finally, the government argues that three provisions of the Illegal Immigration Reform and Immigration Responsibility Act (“IIRIRA”), 8 U.S.C. §§ 1252(e)(3), 1252(a)(5), and 1252(b)(9), divest this court of jurisdiction to entertain this appeal. These statutes, in the government’s view, require the Organizations to bring their claims in individual-removal proceedings or in the District Court for the District of Columbia.

Section 1252(e)(3) authorizes a limited court review of expedited-removal proceedings. The statute requires that judicial review of such administrative decisions be initiated in the District Court for the District of Columbia, and limits review to “determinations of (i) whether such section, or any regulation issued to implement such section, is constitutional; or (ii) whether such a regulation…is not consistent with applicable provisions of this subchapter or is

7 The irony of the government’s position is that section 1158(b)(2)(C)—the INA rule-making delegation upon which it relies—is based on a congressional mandate that was intended, at least in part, to curtail “unfettered executive discretion” and assure Congress’s “proper and substantial role in refugee admissions, given [its] plenary power over immigration.” 125 Cong. Rec. 35,814–15 (1979) (statement of Rep. Holtzman) (emphasis added). “[I]f there is a separation-of- powers concern here, it is between the President and Congress.” EBSC II, 932 F.3d at 774. otherwise in violation of law.” 8 Section 1252(e)(3), in short, limits jurisdiction over challenges to regulations implementing expedited-removal orders. See Barajas- Alvarado, 655 F.3d at 1086 n.10.

Section 1252(a)(5) operates in conjunction with section 1252(e). It limits review of expedited-removal orders to habeas review under 1252(e) and further restricts any appellate habeas review to considering only whether the migrant is lawfully in the country. See id. at 1082; 8 U.S.C. § 1252(e)(2). Section 1252(b)(9) also applies only to removal orders, but instead channels “[j]udicial review of all questions of law and fact…arising from any action taken or proceeding brought to remove an alien from the United States[,]” to the courts of appeals. 8 U.S.C. § 1252(b)(9); see also I.N.S. v. St. Cyr, 533 U.S. 289, 313 (2001).

In the APA context, these provisions prohibit “a claim by an alien, however it is framed, [that] challenges the procedure and substance of an agency determination that is ‘inextricably linked’ to the order of removal[.]” Martinez v. Napolitano, 704 F.3d 620, 623 (9th Cir. 2012). “[C]laims that are independent of or collateral to the removal process” are not actions taken to “remove an alien from the United States.” J.E.F.M. v. Lynch, 837 F.3d 1026, 1032 (9th Cir. 2016); 8 U.S.C. § 1252(b)(9). The purpose of these claim- channeling provisions is to “limit all aliens to one bite of the

8 Migrants can be placed in expedited removal proceedings when they arrive at ports of entry without documents, misrepresent their identities, or present fraudulent documents. See United States v. Barajas-Alvarado, 655 F.3d 1077, 1081 (9th Cir. 2011). Undocumented migrants who receive removal orders but indicate an intention to apply for asylum or a fear of persecution may still be considered for asylum. See id. (citing 8 U.S.C. § 1225(b)(1)(A)(i)).

EAST BAY SANCTUARY COVENANT V. BIDEN 37

apple with regard to challenging an order of removal.” Martinez, 704 F.3d at 622.

None of these provisions have any bearing on the Rule. Sections 1252(a)(5), (b)(9), and (e)(3) govern judicial review of removal orders or challenges inextricably linked with actions taken to remove migrants from the country. The Rule “governs eligibility for asylum and screening procedures for aliens subject to a presidential proclamation or order restricting entry[.]” 83 Fed. Reg. at 55,934 (emphasis added). Bars to asylum eligibility may eventually be relevant to removal proceedings, but they are not “regulation[s]…to implement [removal orders]” or otherwise entirely linked with removal orders. 9 8 U.S.C. § 1252(e)(3); see also Martinez, 704 F.3d at 623; O.A. v. Trump, 404 F. Supp. 3d 109, 141 (D.D.C. 2019) (“§ 1252(e)(3) is about challenges to expedited removal orders and the implementation of the expedited removal provisions that Congress enacted in IIRIRA.”). This is consistent with the purposes of these jurisdictional limitations: allowing collateral APA challenges to an

9 In another, strikingly similar context, the government appears to agree with this interpretation of section 1252(e)(3). Before the District Court for the District of Columbia, the government argued that section 1252(e)(3) divested the court of jurisdiction to hear an APA challenge to an immigration decision issued by the Attorney General. See Grace v. Whitaker, 344 F. Supp. 3d 96, 105 (D.D.C. 2018). The Attorney General’s decision, and a policy memorandum that adopted the standards in the decision, invoked the expedited-removal statute and required that “claims based on membership in a putative particular social group defined by the members’ vulnerability to harm…will not establish the basis for asylum, refugee status, or a credible or reasonable fear of persecution.” Id. at 110. The government there argued that the policy memorandum and the Attorney General’s decision did not “implement” section 1225(b) because it “was a decision about petitions for asylum under section 1158.” Id. at 115–16 (emphasis added). asylum-eligibility rule does not undermine Congress’s desire to “limit all aliens to one bite of the apple with regard to challenging” their removal orders. See Martinez, 704 F.3d at 622.

At best, the law governing asylum is collateral to the process of removal. Migrants in the country who file affirmatively for asylum, or who are otherwise lawfully in the country—such as those who have a valid visa, maintain Temporary Protected Status, or are given parole, for example—can apply and be eligible for asylum and never encounter any of the statutory provisions governing removal. See 8 C.F.R. § 208.4(a)(5)(iv). Other subsections of the INA explicitly grant this court jurisdiction to review denials of individual asylum applications, further reinforcing that the jurisdiction-stripping provisions cited by the government were not intended to apply at all to challenges to asylum eligibility rules. See 8 U.S.C. §§ 1252(a)(2)(B)(ii), (a)(2)(D); see also Morales v. Gonzales, 478 F.3d 972, 978–79 (9th Cir. 2007).

We hold that the Organizations’ claims are justiciable and they have otherwise satisfied the Article III standing requirements.

B.

We generally also require that plaintiffs fall within the “zone of interests” protected by the statute in question to bring their claims in federal court. Lexmark Int’l, Inc., v. Static Control Components, Inc., 572 U.S. 118, 129 (2014). The breadth of the zone-of-interests test varies, depending on the provisions of law at issue. Id. Under the APA, the test is not “especially demanding.” Id. at 130 (quotations and citations omitted). The zone-of-interests analysis forecloses suit “only when a plaintiff’s interests are so

EAST BAY SANCTUARY COVENANT V. BIDEN 39

marginally related to or inconsistent with the purposes implicit in the statute that it cannot reasonably be assumed that Congress authorized that plaintiff to sue.” Id. (quotations and citations omitted).

The Organizations bring their claims under the APA, but because the APA provides a cause of action only to those “suffering legal wrong because of agency action…within the meaning of a relevant statute,” 5 U.S.C. § 702, the relevant zone of interest is that of the INA. EBSC II, 932 F.3d at 767–68. And the relevant purpose is not that of the entire INA; it is “by reference to the particular provision of law upon on which the plaintiff relies.” Bennett, 520 U.S. at 175–76.

In our review, we are “not limited to considering the [specific] statute under which [plaintiffs] sued, but may consider any provision that helps us to understand Congress’ overall purposes” in enacting the statute. Clarke v. Sec. Indus. Ass’n, 479 U.S. 388, 401 (1987); see also EBSC II, 932 F.3d at 768. This inquiry is intended only to help clarify the act’s scope—not determine whether Congress intended a cause of action to arise for the plaintiff in question. See Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians v. Patchak, 567 U.S. 209, 225 (2012) (“We do not require any indication of congressional purpose to benefit the would-be plaintiff.”) (internal quotation marks omitted).

The Organizations’ claims fall within the zone of interests of the INA and of the regulatory amendments implemented by the Rule. The Rule, much like the scope of section 1158(b) of the INA, shapes asylum eligibility requirements for migrants. The Organizations’ purpose is to help individuals apply for and obtain asylum, provide low- cost immigration services, and carry out community education programs with respect to those services. EBSC III, 354 F. Supp. 3d at 1108–10. This is sufficient for the Court’s lenient APA test: at the very least, the Organizations’ interests are “marginally related to” and “arguably within” the scope of the statute. See Patchak, 567 U.S. at 224, 225.

IV.

We turn to the merits of the preliminary injunction 10 entered by the district court. A plaintiff seeking a preliminary injunction must establish that (1) he is likely to succeed on the merits, (2) he is likely to suffer irreparable harm in the absence of preliminary relief, (3) the balance of equities tips in his favor, and (4) an injunction is in the public interest. All. for the Wild Rockies, 632 F.3d at 1131 (citing Winter v. National Res. Def. Council, 555 U.S. 7, 20 (2008)). When the government is a party, the last two factors (equities and public interest) merge. Nken, 556 U.S. at 435. These factors are evaluated on a sliding scale. All. for the Wild Rockies, 632 F.3d. at 1131–34.

We review for abuse of discretion the district court’s grant of a preliminary injunction. Arc of Cal. v. Douglas, 757 F.3d 975, 983 (9th Cir. 2014). District courts abuse their discretion when they rely on an erroneous legal standard or clearly erroneous finding of fact. Id. (internal quotations omitted).

10 The terms of the temporary restraining order entered by the district court in EBSC I technically differ from the terms of the preliminary injunction entered by the court in EBSC III, but the difference has no practical effect: both injunctions prevent enforcement of the Rule and are identical in scope. Therefore, we review them together.

EAST BAY SANCTUARY COVENANT V. BIDEN 41

A.

The likelihood of the Organizations’ success on the merits depends on the substantive and procedural validity of the Rule. See EBSC III, 354 F. Supp. 3d at 1111–12. They must establish a likelihood that the Rule is either substantively or procedurally invalid. See EBSC II, 932 F.3d at 770. Because the record on appeal is now “fully developed,” and the substantive validity of the Rule “rest[s] primarily on interpretations of law, not the resolution of factual issues, we may consider the merits of the case and enter a final judgment to the extent appropriate.” Beno v. Shalala, 30 F.3d 1057, 1063 (9th Cir. 1994) (internal quotation marks omitted).

1.

The APA requires that we “hold unlawful and set aside agency action, findings, and conclusions found to be…an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). Presidential action is not ordinarily “agency action,” and is typically unreviewable under the APA. Franklin v. Massachusetts, 505 U.S. 788, 796 (1992). But the Proclamation and Rule together create an “operative rule of decision” for asylum eligibility that is reviewable by this court. EBSC II, 932 F.3d at 770; see also City of Carmel-by-the-Sea v. U.S. Dep’t of Transp., 123 F.3d 1142, 1166 (9th Cir. 1997) (holding that executive orders with “specific statutory foundation[s]” that do not expressly preclude judicial review are treated as agency action and reviewed under the APA); Public Citizen v. U.S. Trade Rep., 5 F.3d 549, 552 (D.C. Cir. 1993) (“Franklin is limited to those cases in which the President has final constitutional or statutory responsibility for the final step necessary for the agency action directly to affect the parties.”). To determine whether the Rule is “not in accordance with law,” we apply the framework established in Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 842 (1984). Under Chevron, we first consider “whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter.” Campos-Hernandez v. Sessions, 889 F.3d 564, 568 (9th Cir. 2018) (quoting Chevron, 467 U.S. at 842). Federal courts are “the final authority on issues of statutory construction and must reject administrative constructions which are contrary to clear congressional intent.” Chevron, 467 U.S. at 843 n.9.

a.

We consider, then, whether the Rule conflicts with Congress’s intent. The only section of the INA implicated by the Rule is section 1158 (“Asylum”). That section begins by stating that an undocumented migrant may apply for asylum when she is “physically present in the United States” or “arrives in the United States (whether or not at a designated port of arrival…)[.]” 8 U.S.C. § 1158(a)(1). DOJ and DHS adopted the Rule under section 1158(b)(2)(C)’s grant of authority to the Attorney General to “establish additional limitations and conditions, consistent with this section, under which an alien shall be ineligible for asylum[.]” 11

11 A separate subsection of section 1158, 1158(d)(5)(B), grants the Attorney General authority to impose “conditions or limitations on the consideration of an application for asylum not inconsistent with this chapter.” As the motions panel observed, had the Rule explicitly conditioned applications for asylum (instead of eligibility for asylum) on arriving at a designated point of entry, the Rule would be “quite obviously, ‘not in accordance with law,’” EBSC II, 932 F.3d at 770

EAST BAY SANCTUARY COVENANT V. BIDEN 43

We agree with the district court that the Rule is “not in accordance with law.” 5 U.S.C. § 706(2)(A). Section 1158(a) provides that migrants arriving anywhere along the United States’s borders may apply for asylum. The Rule requires migrants to enter the United States at ports of entry to preserve their eligibility for asylum. It is effectively a categorical ban on migrants who use a method of entry explicitly authorized by Congress in section 1158(a). As the district court stated, “[i]t would be hard to imagine a more direct conflict” than the one presented here. EBSC III, 354 F. Supp. 3d at 1112.

The government argues that the structure of section 1158 mandates a different result. Critical to the government’s argument is that section 1158 splits asylum applications (§ 1158(a)) and eligibility (§ 1158(b)) into two different subsections; therefore, the government explains, Congress intended to allow DOJ to promulgate limitations on asylum eligibility without regard to the procedures and authorizations governing asylum applications. The text in section 1158(a) requires only that migrants arriving between ports of entry be permitted to “apply for asylum,” and the Rule does not prevent migrants from submitting futile asylum applications. (emphasis added).

This argument is unconvincing. We avoid absurd results when interpreting statutes. Rowland v. Cal. Men’s Colony, Unit II Men’s Adv. Council, 506 U.S. 194, 200–01 (1993). Explicitly authorizing a refugee to file an asylum application because he arrived between ports of entry and then summarily denying the application for the same reason borders on absurdity. The consequences of denial at the

(quoting 5 U.S.C. § 706(2)(A)), because 1158(a) directs migrants to “apply for asylum” in accordance with section 1158. application or eligibility stage are, to a refugee, the same. See EBSC II, 932 F.3d at 771. Had Congress intended to allow DOJ and DHS to override this provision, it could have said so in its delegation of authority to the Attorney General or in the statutory provisions governing asylum applications. And Congress signaled its desire that any eligibility limitations be consistent with application requirements; limitations promulgated under the eligibility subsection of the statute must be “consistent with this section”—meaning the entirety of section 1158—not just consistent with this subsection.

The other categorical bars to asylum in section 1158(b) of the INA do not meaningfully inform our reading of the statute and the Rule. See EBSC II, 932 F.3d at 771 n.12. The INA contains various provisions making ineligible asylum applicants who committed a serious, nonpolitical crime outside the United States prior to arrival (8 U.S.C. § 1158(b)(2)(A)(iii)), assisted or otherwise participated in the persecution of another person (8 U.S.C. § 1158(b)(2)(A(i)), or were firmly resettled in another country prior to arriving in the United States (8 U.S.C. § 1158(b)(2)(A)(vi)), among other things. The government suggests that the existence of these eligibility bars in the INA demonstrates that Congress intended certain categories of migrants to be permitted to apply for asylum even though they are categorically ineligible. A migrant who was firmly resettled in another country, for example, is still free to complete an asylum application, even though she will be barred from seeking asylum under section 1158(b)(2)(A)(vi).

But—unlike the eligibility bar effected by the Rule—the statutory asylum bars in the INA do not separately conflict with explicit text in section 1158(a). There is no provision

EAST BAY SANCTUARY COVENANT V. BIDEN 45

in section 1158(a), for example, that affirmatively requires that migrants who were firmly resettled in another country be permitted to apply for asylum. The Rule creates the only bar to eligibility under section 1158(b) that directly conflicts with language in section 1158(a). The statutory eligibility bars noted above do not suggest Congress intended that migrants who are subject to them be permitted to apply for asylum. See also EBSC II, 932 F.3d at 772 (“‘[t]o say that one may apply for something that one has no right to receive is to render the right to apply a dead letter.’”) (quoting EBSC I, 349 F. Supp. 3d at 857). The district court correctly concluded that the Rule is substantively invalid because it conflicts with the plain congressional intent instilled in 8 U.S.C. § 1158(a), and is therefore “not in accordance with law,” 5 U.S.C. § 706(2)(A).

b.

But even if the text of section 1158(a) were ambiguous, the Rule fails at the second step of Chevron because it is an arbitrary and capricious interpretation of that statutory provision. If the statute is “silent or ambiguous with respect to the specific issue, the question for the court is whether the agency’s answer is based on a permissible construction of the statute.” Campos-Hernandez, 889 F.3d at 568 (quoting Chevron, 467 U.S. at 843). Under this standard, we must give effect to an agency’s reasonable interpretation of a statute, unless the interpretation is inconsistent with clearly expressed congressional intent. See United States v. Fulton, 475 U.S. 657, 666–67 (1986).

The Board of Immigration Appeals (“BIA”) and this court have long recognized that a refugee’s method of entering the country is a discretionary factor in determining whether the migrant should be granted humanitarian relief. EBSC II, 932 F.3d at 772. More than thirty years ago, the BIA stated that “an alien’s manner of entry or attempted entry is a proper and relevant discretionary factor” to adjudicating asylum applications under section 1158(a), but “it should not be considered in such a way that the practical effect is to deny relief in virtually all cases.” 12 Matter of Pula, 19 I. & N. Dec. 467, 473 (B.I.A. 1987), superseded in part by statute on other grounds as stated in Andriasian v. I.N.S., 180 F.3d 1033, 1043–44 (9th Cir. 1999); see also EBSC II, 932 F.3d at 772. The court explained that it would instead evaluate “the totality of the circumstances and actions of an alien in his flight from the country where he fears persecution,” rather than deny asylum outright because of a single procedural flaw in the migrant’s application. Matter of Pula, 19 I. & N. Dec. at 473–74.

Especially where a migrant may be eligible only for asylum and cannot establish the more stringent criteria for withholding-of-removal, the discretionary factors— including method of entry—should be “carefully evaluated in light of the unusually harsh consequences which may befall an alien[.]” Id. at 474. Indeed, “the danger of persecution should generally outweigh all but the most egregious of adverse factors.” Id.

We have supported the BIA’s understanding of section 1158(a). The most vulnerable refugees are perhaps those fleeing across the border through the point physically closest to them. That a refugee crosses a land border instead of a

12 The BIA in Pula interpreted section 1158(a) before it was amended to include the particular phrase at issue (“whether or not at a designated port of arrival”). At the time, the relevant sentence stated “The Attorney General shall establish a procedure for an alien physically present in the United States or at a land border or port of entry, irrespective of such alien’s status, to apply for asylum[.]” 8 U.S.C. § 1158(a) (1980).

EAST BAY SANCTUARY COVENANT V. BIDEN 47

port-of-entry says little about the ultimate merits of her asylum application; “if illegal manner of flight and entry were enough independently to support a denial of asylum,…virtually no persecuted refugee would obtain asylum.” EBSC II, 932 F.3d at 773 (quoting Huang v. I.N.S., 436 F.3d 89, 100 (2d Cir. 2006)). Given the Rule’s effect of conditioning asylum eligibility on a factor that has long been understood as “worth little if any weight,” see Mamouzian v. Ashcroft, 390 F.3d 1129, 1138 (9th Cir. 2004), in adjudicating whether a migrant should be granted asylum, it is an arbitrary and capricious interpretation of section 1158(a).

The Attorney General’s interpretation of section 1158(a) is also unreasonable, as the district court discussed, in light of the United States’s treaty obligations. See EBSC III, 354 F. Supp. 3d at 1112–13. The United States agreed to comply with Articles 2 through 34 of the 1951 United Nations Convention Relating to the Status of Refugees (“1951 Convention”) and the 1967 United Nations Protocol Relating to the Status of Refugees (“1967 Protocol”) in 1968. H.R. Rep. 96-781 (Conf. Rep.), at 19–20 (1980), as reprinted in 1980 U.S.C.C.A.N. 160, 160–62; see also I.N.S. v. Cardoza-Fonseca, 480 U.S. 421, 429, 436–37 (1987). To streamline the United States’s refugee procedures and implement the country’s new treaty commitments, Congress passed the Refugee Act of 1980, which amended the INA and created the country’s first codified rules governing asylum. S. Rep. No. 96-256, at 1 (1979), as reprinted in 1980 U.S.C.C.A.N. 141, 141–42, 144; H.R. Doc. No. 96- 608, at 17–18 (1979); see also Negusie v. Holder, 555 U.S. 511, 535–36 (2009). As the United Nations High Commissioner of Refugees (“UNHCR”) explains, 13 the Rule runs afoul of three of these codified rules: the right to seek asylum, the prohibition against penalties for irregular entry, and the principle of non- refoulement embodied in Article 31(1) of the 1951 Convention. Neither the 1967 Protocol nor the 1951 Convention require countries to accept refugees, but they do ensure that refugees at each signatory’s borders have legal and political rights and protections. See Cong. Research Serv. S522-10, Review of U.S. Refugee Resettlement Programs and Policies 15–16 (1980).

The definition of “refugee” used in the 1951 Convention is “virtually identical” to the one adopted by Congress in the INA. Cardoza-Fonseca, 480 U.S. at 437. Under both the INA and the 1951 Convention, refugees are all individuals who—because of a “well-founded fear of being persecuted for reasons of race, religion, nationality, membership in a particular social group or political opinion”—are “unable,” or, because of such fear, “unwilling to return” to their home countries. See 8 U.S.C. § 1101(a)(42); 1951 Convention, Art. 1(A)(2). Once individuals meet the statutory definition of a “refugee,” they may be granted asylum under the INA. See 8 U.S.C. § 1158(b)(1)(A).

13 The arguments presented by the United Nations in its amicus brief on how the 1951 Convention and 1967 Protocol should be construed are not binding on this court. See Cardoza-Fonseca, 480 U.S. at 439 n.22. But they do “provide[] significant guidance in construing the [1967] Protocol, to which Congress sought to conform[,]” and are “useful in giving content to the obligations that the Protocol establishes.” Id; see also Miguel-Miguel v. Gonzales, 500 F.3d 941, 949 (9th Cir. 2007) (“We view the UNHCR Handbook as persuasive authority in interpreting the scope of refugee status under domestic asylum law.”) (internal quotation marks and citation omitted).

EAST BAY SANCTUARY COVENANT V. BIDEN 49

Both the INA and the 1951 Convention acknowledge that individuals may be stripped of their refugee status even when they meet the other eligibility criteria for asylum. The refugee provisions of the 1951 Convention “shall not apply” to “any person with respect to whom there are serious reasons for considering” that such a person has committed a crime against peace, a war crime, a crime against humanity, a non-political crime outside of the country of refuge, prior to their admission as a refugee, or has been “guilty of acts contrary to the purposes and principles of the United Nations.” 1951 Convention, Art. 1(F)(a)–(c). The statutory bars for eligibility in the INA are similarly severe. Individuals who are otherwise refugees may not apply for asylum if the Attorney General determines that they “ordered, incited, assisted, or otherwise participated” in the persecution of another, based on a trait protected by the INA; “constitute[] a danger to the community of the United States”; committed a “serious nonpolitical crime” outside the country; are a “danger to the security” of the country; have engaged in terrorist activities; or were “firmly resettled in another country prior to arriving in the United States.” 8 U.S.C. § 1158(b)(2)(A)(i)–(vi).

The exceptions listed in the 1951 Convention “require individualized assessments and ‘must be [interpreted] restrictive[ly].” Br. for UNHCR as Amicus Curiae at 14 n.6 (quoting Office of the United Nations High Commissioner for Refugees, Handbook on Procedures and Criteria for Determining Refugee Status ¶ 149 (Geneva, 1979)). So too the categorical bars on eligibility in the INA are interpreted with lenience toward migrants to avoid infringing on the commitments set forth in the 1951 Convention and 1967 Protocol. See, e.g., Ali v. Ashcroft, 394 F.3d 780, 790 (9th Cir. 2005) (A “narrow interpretation of the firm resettlement bar would limit asylum to refugees from nations contiguous to the United States or to those wealthy enough to afford to fly here in search of refuge. The international obligation our nation agreed to share when we enacted the Refugee Convention into law knows no such limits.”); Cardoza- Fonseca, 480 U.S. at 449.

The asylum bars in the INA and in the 1951 Convention appear to serve either the safety of those already in the United States or, in the case of the firm-resettlement bar, the safety of refugees. The Rule ensures neither. Even a broad interpretation of these eligibility bars does not naturally encompass a refugee’s method of entry. Illegal entry is not ordinarily considered a “serious crime.” See Pena- Cabanillas v. United States, 394 F.2d 785, 788 (9th Cir. 1968) (stating that the statute criminalizing entry into the United States “is not based on any common law crime, but is a regulatory statute enacted to assist in the control of unlawful immigration by aliens” and “is a typical mala prohibita offense”). Nor does a migrant’s method of entry per se create a danger to the United States, serve as a useful proxy for terrorist activity, or suggest the persecution of another.

And the Rule surely does not suggest that the migrant has received protection in a third country. Many migrants enter between ports of entry out of necessity: they “cannot satisfy regular exit and entry requirements and have no choice but to cross into a safe country irregularly prior to making an asylum claim.” Br. for UNHCR as Amicus Curiae at 15 (citing Memorandum by the Secretary-General, Ad Hoc Comm. on Statelessness, Status of Refugees & Stateless Persons, at Annex Art. 24, cmt. ¶ 2, U.N. Doc. E/AC.32/2 (Jan. 3, 1950); UNHCR Executive Committee Conclusion No. 58 (XL) ¶ (i) (Oct. 13, 1989)). This was well recognized when the Refugee Act of 1980 was drafted. See Pub. L. No.

EAST BAY SANCTUARY COVENANT V. BIDEN 51

96-212, § 208(a), 94 Stat. 102, 105 (1980). Prior to the passage of the Act, migrants who arrived at a port of entry were “given an opportunity to have their [asylum] applications heard in a hearing before an immigration judge,” but refugees arriving “at a land border of the United States [we]re not given this right.” Refugee Act of 1979: Hearing on H.R. 2816 before the Subcomm. on Immigration, Refugees, and Int’l Law of the Comm. on the Judiciary, 96th Cong. 190 (1979) (testimony of David Carliner, American Civil Liberties Union). In its attempt to streamline the country’s refugee and asylum laws, Congress was urged to consider that “persons who seek any benefits under [the INA] should be entitled to a uniform procedure.” Id. Congress heeded this consideration during the drafting of the Refugee Act, eventually describing it as “establish[ing] a more uniform basis for the provision of assistance to refugees, and [] other purposes.” Refugee Act of 1980, Pub. L. No. 96-212, 94 Stat. 102 (1980). The Rule defies this desire for uniformity and denies refuge to those crossing a land border. The effects of the Rule contravene the United States’s commitments in the 1951 Convention.

Article 31(1) of the 1951 Convention also explains that signatories “shall not impose penalties” on account of refugees’ “illegal entry or presence,” 1951 Convention Art 31(1). Notwithstanding the government’s interpretations otherwise, “deportation is an integral part— indeed, sometimes the most important part—of the penalty that may be imposed” on migrants who are found guilty of specified crimes, or for other reasons are barred from seeking asylum. 14 See Padilla v. Kentucky, 559 U.S. 356,

14 The UNHCR’s view is that “penalties” in Article 31(1) “encompasses civil or administrative penalties as well as criminal ones.” Br. for UNHCR as Amicus Curiae at 20. 364 (2010) (footnote omitted). The Rule imposes an additional penalty on refugees because of their “illegal entry” by risking the deportation of migrants who enter the country at a land border. 1951 Convention Art. 31(1).

And by categorically denying refugees an opportunity to seek asylum only because of their method of entry, the Rule is also in tension with the United States’s commitment to avoid refouling individuals to countries where their lives are threatened. Article 31(1) of the 1951 Convention prohibits signatories from “expel[ling] or return[ing] (‘refouler’) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened[.]” The INA’s withholding-of-removal, 8 U.S.C. § 1231(b)(1), and Convention Against Torture (“CAT”) protections, 8 C.F.R. § 1208.16–18, are not as great as those conferred by the INA’s asylum provisions. The evidentiary standard that applicants must meet for either withholding-of-removal or CAT relief is higher than the evidentiary standard for asylum. See, e.g., Ling Huang v. Holder, 744 F.3d 1149, 1152 (9th Cir. 2014). Applicants for withholding-of- removal and CAT relief must establish a “clear probability” that they would be persecuted or tortured, respectively, if they were removed to their home countries. See Korablina v. I.N.S., 158 F.3d 1038, 1045–46 (9th Cir. 1998); Wakkary v. Holder, 558 F.3d 1049, 1053 (9th Cir. 2009). A “clear probability” of persecution or torture means that it is “more likely than not” that applicants will be persecuted upon their removal. I.N.S. v. Stevic, 467 U.S. 407, 424, 429–30 (1984).

Applicants for asylum instead must demonstrate only that they are “unable or unwilling” to return to their home countries “because of persecution or a well-founded fear of persecution[.]” 8 U.S.C. § 1101(a)(42)(A). “One can certainly have a well-founded fear of an event happening

EAST BAY SANCTUARY COVENANT V. BIDEN 53

when there is less than a 50% chance of the occurrence taking place”; it would only be “too apparent,” for example, for a refugee to have a “well-founded fear of being persecuted” where “every tenth adult male person is either put to death or sent to some remote labor camp” in the applicant’s home country. Cardoza-Fonseca, 480 U.S. at 431 (citing 1 A. Grahl-Madsen, The Status of Refugees in International Law 180 (1966)). The Rule, then, risks the removal of individuals with meritorious asylum claims who cannot petition for withholding of removal or CAT relief. By doing so, it is inconsistent with our treaty commitment to non-refoulement.

The Rule is “arbitrary, capricious, or manifestly contrary to the statute,” Chevron, 467 U.S. at 844, both because it is contrary to plain congressional intent, and because it is an arbitrary and capricious interpretation of section 1158(a). Even if we agreed that the text of section 1158(a) is ambiguous, the Rule flouts this court’s and the BIA’s discretionary, individualized treatment of refugees’ methods of entry, and infringes upon treaty commitments we have stood by for over fifty years.

2.

Because we conclude that the Rule is substantively invalid, we only briefly address the procedural arguments raised by the parties. The APA requires public notice and comment and a thirty-day grace period before a proposed rule takes effect. 5 U.S.C. § 553(b)–(d). The notice-and- comment requirements are exempted when “there is involved a military or foreign affairs function of the United States[,]” id. § 553(a), or when “the agency for good cause finds…that notice and public procedure thereon are impracticable, unnecessary or contrary to the public interest.” Id. § 553(b)(B). The thirty-day lag in publication can be waived where “good cause [is] found.” 15 Id. § 553(d)(3).

The Rule was issued without notice and comment or the grace period. The government argues that the Rule was properly issued because it falls under either the good-cause or the foreign-affairs exceptions to these procedural requirements.

a.

Proper invocation of the good-cause exception is “sensitive to the totality of the factors at play.” United States v. Valverde, 628 F.3d 1159, 1164 (9th Cir. 2010). The exception is a “high bar” because it is “essentially an emergency procedure.” Id. at 1164, 1165. The government must make a sufficient showing that “‘delay would do real harm’ to life, property, or public safety,” EBSC II, 932 F.3d at 777 (quoting Valverde, 628 F.3d at 1164–65), or that “some exigency” interferes with its ability to carry out its mission. Nat. Res. Def. Council, Inc. v. Evans, 316 F.3d 904, 911 (9th Cir. 2003).

In support of its reliance on the exception, the government now cites a Washington Post article indicating that when the United States stopped its policy of separating migrant parents from their children, smugglers told asylum-

15 “Different policies” underlie the good-cause exception for the thirty-day grace period and the good-cause exception for the notice-and- comment requirement, and “they can be invoked for different reasons.” Riverbend Farms, Inc. v. Madigan, 958 F.2d 1479, 1485 (9th Cir. 1992). Notice-and-comment requirements are intended to ensure public participation in rulemaking, and the thirty-day waiting period is “intended to give affected parties time to adjust their behavior before the final rule takes effect.” Id.

EAST BAY SANCTUARY COVENANT V. BIDEN 55

seekers that “the Americans do not jail parents who bring children—and to hurry up before they might start doing so again.” See Nick Miroff and Carolyn Van Houten, The Border is Tougher to Cross Than Ever. But There’s Still One Way into America, Wash. Post (Oct. 24, 2018). The district court concluded that the article “at least supports the inference” that the Rule might result in similar changes in immigration policy, and held that the government had “identified a ‘rational connection between the facts found and the choice made’ to promulgate the interim Rule on an emergency basis.” EBSC III, 354 F. Supp. 3d at 1115 (quoting Valverde, 628 F.3d at 1168).

A citation to this single article is not sufficient to demonstrate that the delay caused by notice-and-comment or the grace period might do harm to life, property, or public safety. See EBSC II, 932 F.3d at 777. The government’s reasoning continues to be largely speculative, see id. at 778; no evidence has been offered to suggest that any of its predictions are rationally likely to be true. The article does not directly relate to the Rule, the consequences of the Rule, or anything related to asylum eligibility.

Even if it did, that “the very announcement of [the] proposed rule itself can be expected to precipitate activity by affected parties that would harm the public welfare,” Reply Br. of Gov’t at 21, is likely often, or even always true. The lag period before any regulation, statute, or proposed piece of legislation allows parties to change their behavior in response. If we were to agree with the government’s assertion that notice-and-comment procedures increase the potential harm the Rule is intended to regulate, these procedures would often cede to the good-cause exception. Because the government has failed to demonstrate the existence of an exigency justifying good cause, we hold that the Rule likely does not properly fall under the good cause exception.

b.

For the foreign affairs exception to apply, “the public rulemaking provisions should provoke definitely undesirable international consequences.” Yassini v. Crosland, 618 F.2d 1356, 1360 n.4 (9th Cir. 1980). Otherwise, the exception “would become distended if applied to INS actions generally, even though immigration matters typically implicate foreign affairs.” Id. Use of the exception is generally permissible where the international consequences of the rule-making requirements are obvious or thoroughly explained. We have rejected its use where the government has failed to substantiate its reliance on the exception or explain the detrimental effects of compliance with the APA’s requirements. See EBSC II, 932 F.3d at 776– 77.

The government cites to four documents in support of its renewed argument that the foreign-affairs exception is justified: a Memorandum of Understanding (“MOU”) between DHS and the Mexican government, the Washington Post article, credible-fear origin data published by the Executive Office of Immigration Review (“EOIR”), and a speech by President Trump. The four documents appear to demonstrate that the Rule and Proclamation are related to ongoing changes in the national immigration landscape, but still fail to establish that adhering to notice and comment and a thirty-day grace period will “provoke definitely undesirable international consequences.” Yassini, 618 F.3d at 1360 n.4.

We agree with the government that the cited MOU does broadly “show[] that [immigration] negotiations have

EAST BAY SANCTUARY COVENANT V. BIDEN 57

happened in the past,” Op. Br. of Gov’t at 49, but this is insufficient to demonstrate that notice and comment will provoke undesirable international consequences. Indeed, the MOU’s substance seems to undermine the “broader diplomatic program involving sensitive and ongoing negotiations with Mexico.” Op. Br. of Gov’t at 47 (internal quotations omitted). Article 3 of the MOU states that “[l]ocal repatriation agreements should conform to mutually established criteria and principles for the repatriation of Mexican nationals being repatriated from the United States to Mexico.” The unilateral repatriation of Mexican nationals set forth by the Rule—without requesting public participation—undermines these terms.

The cited Washington Post page discusses an increase in the proportion of families that seek asylum and the EOIR data lists the country of origin of credible-fear cases and summarizes the number of people that attempt to enter the United States with an asylum application, the number of cases completed in 2018, and the outcome of credible fear cases. It is unclear how these data “reflect[] motivations for crossing the border illegally,” Op. Br. of Gov’t at 49, and even less clear how they demonstrate the consequences of requesting public notice-and-comment on foreign policy. And the speech by President Trump, as the district court noted, discusses the domestic consequences of foreign immigration, not the foreign policy consequences of immigration into the United States. See EBSC III, 354 F. Supp. 3d at 1114. The speech—like the MOU, the article, and the EOIR data—does not suggest that the APA’s rulemaking provisions might trigger or even shape immediate consequences in foreign affairs.

The evidence relied on by the government here is largely the same as the evidence previously before the motions panel and the district court. While we remain “sensitive to the fact that the President has access to information not available to the public, and…[are] cautious about demanding confidential information,” the connection between negotiations with Mexico and the immediate implementation of the Rule is still “not apparent.” EBSC II, 932 F.3d at 776. Broadly citing to the Rule’s immigration context is insufficient to invoke the foreign-affairs exception. See Yassini, 618 F.2d at 1360 n.4. The government has not made a “sufficient showing” that “the public rulemaking provisions should provoke definitely undesirable international consequences.” Id.; see also Evans, 316 F.3d at 912.

In sum, the government has not established that DOJ and DHS properly invoked the foreign-affairs exception to the notice-and-comment requirement and thirty-day grace period.

B.

We next consider whether the Organizations have established that, in the absence of a preliminary injunction, they are likely to suffer irreparable harm. See Arizona Dream Act Coalition v. Brewer, 757 F.3d 1053, 1068 (9th Cir. 2014). Irreparable harm is “harm for which there is no adequate legal remedy, such as an award for damages.” Id. For this reason, economic harm is not generally considered irreparable. But where parties cannot typically recover monetary damages flowing from their injury—as is often the case in APA cases—economic harm can be considered irreparable. See California v. Azar, 911 F.3d 558, 581 (9th Cir. 2018). Intangible injuries may also qualify as irreparable harm, because such injuries “generally lack an adequate legal remedy.” Brewer, 757 F.3d at 1068.

EAST BAY SANCTUARY COVENANT V. BIDEN 59

We agree with the district court that the Organizations have established that they will suffer a significant change in their programs and a concomitant loss of funding absent a preliminary injunction enjoining enforcement of the Rule. EBSC II, 932 F.3d at 767. Both constitute irreparable injuries: the first is an intangible injury, and the second is economic harm for which the Organizations have no vehicle for recovery.

The Rule has already prompted the Organizations to change their core missions. Since the Rule issued, ILL has placed programmatic expansions on hold and has “had to lessen its caseload[.]” Supp. Decl. of Stephen W. Manning at ¶ 14. CARECEN notes that it will “divert significant resources,” including “staff time and organizational resources” to respond to the Rule. Decl. of Daniel Sharp at ¶¶ 11–13. EBSC has had to “divert resources away from its core programs to address the new policy.” Decl. of Michael Smith at ¶ 15. And, as discussed in Part III, supra, the Organizations each stand to lose funding because of their core changes in mission.

Importantly, the Organizations also filed suit the same day that the Rule and the first proclamation issued; while not dispositive, this suggests urgency and impending irreparable harm. See Oakland Tribune, Inc. v. Chronicle Pub. Co., Inc., 762 F.2d 1374, 1377 (9th Cir. 1985). We agree with the district court that the Organizations have demonstrated a sufficient likelihood of irreparable injury to warrant injunctive relief. EBSC III, 354 F. Supp. 3d at 1116.

C.

The government next argues that the harms it will suffer because of the preliminary injunction—namely, the harm caused by the injunction “undermin[ing] the Executive Branch’s constitutional and statutory authority to secure the Nation’s borders,” and the “entry of illegal aliens”— outweigh the benefit to the public and the Organizations conferred by the injunction. Op. Br. of Gov’t at 51–52. Relevant equitable factors include the value of complying with the APA, the public interest in preventing the deaths and wrongful removal of asylum-seekers, preserving congressional intent, and promoting the efficient administration of our immigration laws at the border.

First, “[t]he public interest is served by compliance with the APA.” Azar, 911 F.3d at 581. Indeed, it “does not matter that notice and comment could have changed the substantive result; the public interest is served from the proper process itself.” Id. at 581–82. The Organizations and various Amici informed the district court that they would have submitted comments explaining why the Rule disrupts their organizational missions and fails to meet its intended purpose, had they had the opportunity. The APA’s requirements reflect “a judgment by Congress that the public interest is served by a careful and open review of proposed administrative rules and regulations.” Alcaraz v. Block, 746 F.2d 593, 610 (9th Cir. 1984) (citing Phil. Citizens in Action v. Schweiker, 669 F.2d 877, 881 (9th Cir. 1982)). The government’s failure to comply with the APA—particularly given the strength of the Organizations’ procedural attack on the Rule—weighs in favor of granting injunctive relief.

Second, the public has an interest in “ensuring that we do not deliver aliens into the hands of their persecutors,” Leiva-Perez, 640 F.3d at 971, and “preventing aliens from being wrongfully removed, particularly to countries where they are likely to face substantial harm,” Nken, 556 U.S. at 436. The Rule will likely result in some migrants being wrongfully denied refugee status in this country. For

EAST BAY SANCTUARY COVENANT V. BIDEN 61

migrants affected by the Rule, withholding of removal and CAT protection are the only forms of relief available. As discussed, these forms of relief demand a higher burden of proof than an asylum claim. At the initial screening interview with an asylum officer, an applicant seeking asylum need only present a “credible fear” of persecution, while an applicant seeking withholding of removal of CAT protection must demonstrate the higher “reasonable fear” of persecution or torture.

The government’s opening brief notes that 17 percent of the 34,158 migrants whose cases were completed in 2018 received asylum. See Op. Br. of Gov’t at 52. Assuming the number of migrants remains constant, if even just 25 percent of asylum-seekers with meritorious claims are denied asylum because of their method of entry, over 1,000 people will either be returned to home countries where they face “persecution based on ‘race, religion, nationality, membership in a political social group, or political opinion,’” EBSC III, 354 F. Supp. 3d at 1117 n.15 (quoting 8 U.S.C. §§ 1101(a)(42), 1158(b)(1)), or forced to proceed on limited-relief claims that demand more stringent showings. If the rate of migration and the rate of migrants claiming fear during the expedited removal process continues to increase, see 84 Fed. Reg. at 21,229, the scale of this wrongful removal will only worsen.

Third, the public has an interest in ensuring that the “statutes enacted by [their] representatives are not imperiled by executive fiat.” EBSC II, 932 F.3d at 779 (internal quotation marks omitted). The INA, and the United States’s signatory status to the 1951 Convention, “reflect the balance Congress struck between the public interests in rendering aliens who enter illegally inadmissible and subject to criminal and civil penalties, and…preserving their ability to seek asylum.” EBSC III, 354 F. Supp. 3d at 1117–18 (citations omitted). The Rule and Proclamation disrupt that balance by overriding plain congressional intent.

Finally, the government and the public have an interest in the “efficient administration of the immigration laws at the border.” EBSC II, 932 F.3d at 779 (internal quotation marks omitted). This interest is “weighty.” Landon v. Plasencia, 459 U.S. 21, 34 (1982). “[C]ontrol over matters of immigration is a sovereign prerogative, largely within the control of the executive and the legislature.” Id. The government has a compelling interest in ensuring that injunctions—such as the one granted here—do not undermine separation of powers by blocking the Executive’s lawful ability to regulate immigration and rely on its rulemaking to aid diplomacy.

The role of the judiciary in reviewing such policies is narrow. It is merely to ensure that executive procedures do not violate principles of due process or “displace congressional choices of policy.” Id. at 35. This executive deference, then, is closely linked with our determination on the substantive validity of the Rule. Essentially, the weight we ascribe to this factor depends on the extent to which we agree that the Rule overrides plain congressional intent. Because the Organizations have established that the Rule is invalid, we do not place much weight on this factor. As the motions panel noted: “[t]here surely are enforcement measures that the President and the Attorney General can take to ameliorate the [immigration] crisis, but continued inaction by Congress is not a sufficient basis under our Constitution for the Executive to rewrite our immigration laws.” EBSC II, 932 F.3d at 774.

In sum, we agree with the district court that there is a significant basis for concluding that the public interest

EAST BAY SANCTUARY COVENANT V. BIDEN 63

weighs “sharply” in the Organizations’ favor. See EBSC III, 354 F. Supp. 3d at 1111.

V.

Finally, we turn to the remedy entered by the district court: an injunction preventing enforcement of the Rule. The injunction enjoins the part of the Rule that removes asylum eligibility from migrants who fail to follow a presidential proclamation. EBSC III, 354 F. Supp. 3d at 1121. It does not enjoin the credible-fear amendments, but “they have no independent effect,” so they are effectively enjoined as well. Id. at 1121 n.22. We conclude that the district court did not abuse its discretion in enjoining enforcement of the Rule.

Injunctive relief should be “no more burdensome to the defendant than necessary to provide complete relief to the plaintiffs before the court.” Univ. of Cal. v. U.S. Dep’t of Homeland Sec., 908 F.3d 476, 511 (9th Cir. 2018) (internal quotations omitted). “Where relief can be structured on an individual basis, it must be narrowly tailored to remedy the specific harm shown,” but there is “no general requirement that an injunction affect only the parties in the suit.” Bresgal v. Brock, 843 F.2d 1163, 1169–1170 (9th Cir. 1987). The equitable relief granted by the district court is acceptable where it is “necessary to give prevailing parties the relief to which they are entitled.” Id. at 1170–71. District courts have “considerable discretion” in crafting suitable equitable relief; correspondingly, appellate review is “narrow.” Lamb-Weston, Inc. v. McCain Foods, Ltd., 941 F.2d 970, 974 (9th Cir. 1991).

As discussed, the harms caused to the Organizations as a result of the Rule include a (1) loss of funding and (2) disruption of organizational purpose. Adequate equitable relief must remedy both harms. Bresgal, 843 F.2d at 1170–71. Both harms are due, in part, to the Rule’s likely consequence of preventing asylum-seekers with meritorious claims from entering the country along our southern border and successfully obtaining asylum. The stymied flow of refugees will result in less funding for the Organizations, and a shift (sometimes wholesale) in their organizational missions.

The Organizations do not limit their potential clients to refugees that enter the United States only at the California- Mexico or Arizona-Mexico border; they represent “asylum seekers” broadly. Unlike the plaintiffs in California v. Azar—individual states seeking affirmance of an injunction that applied past their borders—the Organizations here “do not operate in a fashion that permits neat geographic boundaries.” EBSC III, 354 F. Supp. 3d at 1120–21; see also Califano v. Yamasaki, 442 U.S. 682, 702 (1979) (The “scope of injunctive relief is dictated by the extent of the violation established, not by the geographical extent of the plaintiff class.”). An injunction that, for example, limits the application of the Rule to California, would not address the harm that one of the Organizations suffers from losing clients entering through the Texas-Mexico border. One fewer asylum client, regardless of where the client entered the United States, results in a frustration of purpose (by preventing the organization from continuing to aid asylum applicants who seek relief), and a loss of funding (by decreasing the money it receives for completed cases).

The government suggests that plaintiffs “identify actual aliens in the United States who would otherwise be subject to the Rule,” Op. Br. of Gov’t at 57, but this suggestion fails to redress the scope of the Organizations’ harms. Part of the harm the Organizations have alleged is the difficulty posed

EAST BAY SANCTUARY COVENANT V. BIDEN 65

by the Rule in helping them reach migrants who will cross the border; their missions are not limited to helping individuals currently present in the United States. Even if their missions were so limited, asking the Organizations to seek and list every person in the country they might help in the coming months is infeasible and impracticable. The “Government has not proposed a workable alternative form of the [injunction] that accounts” for the harm at issue but “nevertheless appl[ies] only within the [] borders” of the Ninth Circuit. Washington v. Trump, 847 F.3d 1151, 1167 (9th Cir. 2017); see also EBSC II, 932 F.3d at 779; EBSC III, 354 F. Supp. 3d at 1121.

Two other factors support the district court’s decision to enjoin Defendants from taking any action to implement the Rule. First, “[w]hen a reviewing court determines that agency regulations are unlawful, the ordinary result is that the rules are vacated—not that their application to the individual petitioners is proscribed.” Univ. of Cal., 908 F.3d at 511 (internal quotation marks omitted). Singular equitable relief is “commonplace” in APA cases, and is often “necessary to provide the plaintiffs” with “complete redress.” Id. at 512. Our “typical response is to vacate the rule and remand to the agency”; we “ordinarily do not attempt, even with the assistance of agency counsel, to fashion a valid regulation from the remnants of the old rule.” Harmon v. Thornburgh, 878 F.2d 484, 494 (D.C. Cir. 1989). Because of the broad equitable relief available in APA challenges, a successful APA claim by a single individual can affect an “entire” regulatory program. Lujan v. Nat’l Wildlife Fed’n, 497 U.S. 871, 890 n.2 (1990).

Second, as the district court noted, there is an important “need for uniformity in immigration policy.” Id. at 511; see also EBSC III, 354 F. Supp. 3d at 1120–21. We previously have recognized that the “Constitution requires a uniform Rule of Naturalization; Congress has instructed that the immigration laws of the United States should be enforced vigorously and uniformly; and the Supreme Court has described immigration policy as a comprehensive and unified system.” Univ. of Cal., 908 F.3d at 511 (quoting United States v. Texas, 809 F.3d 134, 187–88 (5th Cir. 2014) (emphases in original)). The INA itself “was designed to implement a uniform federal policy, and the meaning of concepts important to its application are not to be determined according to the law of the forum, but rather require[] a uniform federal definition.” Kahn v. I.N.S., 36 F.3d 1412, 1414 (9th Cir. 1994) (internal quotation marks omitted). Different interpretations of executive policy across circuit or state lines will needlessly complicate agency and individual action in response to the United States’s changing immigration requirements. For these reasons, in immigration cases, we “consistently recognize[] the authority of district courts to enjoin unlawful policies on a universal basis.” EBSC II, 932 F.3d at 779 (citing Univ. of Cal., 908 F.3d at 511).

The government again “raises no grounds on which to distinguish this case from our uncontroverted line of precedent.” Id. Given the context of this case and the harm the district court sought to address, we find no error or abuse of discretion in the terms or scope of the preliminary injunction.

VI.

For the reasons discussed, the district court’s orders granting preliminary injunctions are AFFIRMED.

EAST BAY SANCTUARY COVENANT V. BIDEN 67

FERNANDEZ, Circuit Judge, concurring in the result:

I concur in the majority opinion because, and for the most part only because, I believe that we are bound by the published decision in East Bay Sanctuary Covenant v. Trump (East Bay I), 932 F.3d 742 (9th Cir. 2018).

More specifically, we are bound by both the law of the circuit and the law of the case. 1 Of course, the rules that animate the former doctrine are not the same as those that animate the latter. See Gonzalez v. Arizona, 677 F.3d 383, 389 n.4 (9th Cir. 2012) (en banc).

As we have said: “Circuit law…binds all courts within a particular circuit, including the court of appeals itself. Thus, the first panel to consider an issue sets the law not only for all the inferior courts in the circuit, but also future panels of the court of appeals.” Hart v. Massanari, 266 F.3d 1155, 1171 (9th Cir. 2001). Moreover: “Once a panel resolves an issue in a precedential opinion, the matter is deemed resolved, unless overruled by the court itself sitting en banc, or by the Supreme Court.” Id. (footnote omitted). Published opinions are precedential. See id. at 1177; see also Gonzalez, 667 F.3d at 389 n.4. That remains true, even if some later panel is satisfied that “arguments have been characterized differently or more persuasively by a new

1 The majority has now taken steps to obscure its attacks on our doctrines of law of the case and law of the circuit, which were set forth in its original opinion, East Bay Sanctuary Covenant v. Trump, 950 F.3d 1242, 1261–65 (9th Cir. 2020). Nevertheless, my concurrence remains the same and is based upon the same reasoning. See id. at 1284–86 (Fernandez, J., concurring in the result). In short, the majority’s attempt to pull an invisibility cloak over the mainspring of its attacks cannot hide damage wrought by them. litigant,” 2 or even if a later panel is convinced that the earlier decision was “incorrectly decided” and “needs reexamination.” 3 And those rules are not mere formalities to be nodded to and avoided. Rather, “[i]nsofar as there may be factual differences between the current case and the earlier one, the court must determine whether those differences are material to the application of the rule or allow the precedent to be distinguished on a principled basis.” Hart, 266 F.3d at 1172. In this case, there are no material differences—in fact, the situation before this panel is in every material way the same as that before the motions panel. Furthermore, there is no doubt that motions panels can publish their opinions, 4 even though they do not generally do so. 5 Once published, there is no difference between motions panel opinions and other opinions; all are entitled to be considered with the same principles of deference by ensuing panels. Thus, any hesitation about whether they should be precedential must necessarily come before the panel decides to publish, not after. As we held in Lair v. Bullock, 798 F.3d 736 (9th Cir. 2015):

Lair contended at oral argument that a motions panel’s decision cannot bind a merits panel, and as a result we are not bound by the motions panel’s analysis in this case. Not so. We have held that motions panels can issue published decisions.…[W]e are bound by a

2 United States v. Ramos-Medina, 706 F.3d 932, 939 (9th Cir. 2013). 3 Naruto v. Slater, 888 F.3d 418, 425 n.7 (9th Cir. 2018). 4 See 9th Cir. Gen. Order 6.3(g)(3)(ii); see also id. at 6.4(b). 5 See Haggard v. Curry, 631 F.3d 931, 933 n.1 (9th Cir. 2010) (per curiam).

EAST BAY SANCTUARY COVENANT V. BIDEN 69

prior three-judge panel’s published opinions, and a motions panel’s published opinion binds future panels the same as does a merits panel’s published opinion.

Id. at 747 (citations omitted). 6 Therefore, the legal determinations in East Bay I are the law of the circuit.

We have explained the law of the case doctrine as “a jurisprudential doctrine under which an appellate court does not reconsider matters resolved on a prior appeal.” Jeffries v. Wood, 114 F.3d 1484, 1488–89 (9th Cir. 1997) (en banc), overruled on other grounds by Gonzalez, 677 F.3d at 389 n.4. While we do have discretion to decline application of the doctrine, “[t]he prior decision should be followed unless: (1) the decision is clearly erroneous and its enforcement would work a manifest injustice, (2) intervening controlling authority makes reconsideration appropriate, or (3) substantially different evidence was adduced at a subsequent trial.” Id. at 1489 (internal quotation marks and footnote omitted). 7 We have also indicated that, in general,

6 The majority opines that in this respect Lair’s holding is dicta. Not so. The court’s first basis for rejecting Lair’s contention was the basis just quoted. Its second basis was then set forth. Id. It gave both of those alternatives weight and attention. See Woods v. Interstate Realty Co., 337 U.S. 535, 537, 69 S. Ct. 1235, 1237, 93 L. Ed. 1524 (1949) (holding “where a decision rests on two or more grounds, none can be relegated to the category of obiter dictum.”); see also United States v. Vidal- Mendoza, 705 F.3d 1012, 1016 n.5 (9th Cir. 2013); Guadalupe-Cruz v. INS, 240 F.3d 1209, 1211 & n.5 (9th Cir.), corrected, 250 F.3d 1271 (9th Cir. 2001). 7 The majority seems to add a fourth exception, that is, motions panel decisions never constitute the law of the case. That would be strange if those decisions can constitute the law of the circuit, which they can. Moreover, the case primarily cited for that proposition did not “our decisions at the preliminary injunction phase do not constitute the law of the case,” 8 but that is principally because the matter is at the preliminary injunction stage and a further development of the factual record as the case progresses to its conclusion may well require a change in the result. 9 Even so, decisions “on pure issues of law…are binding.” Ranchers Cattlemen, 499 F.3d at 1114. Of course, the case at hand has not progressed beyond the preliminary injunction stage. It is still at that stage, and the factual record has not significantly changed between the record at the time of the decision regarding the stay motion and the current record. Therefore, as I see it, absent one of the listed exceptions, which I do not perceive to be involved here, the law of the case doctrine would also direct that we are bound by much of the motions panel’s decision in East Bay I.

Applying those doctrines:

(1) The Organizations have standing. East Bay I, 932 F.3d at 765–69.

(2) The Organizations are likely to succeed on the substantive merits. See id. at 770–74. As to procedural validity regarding adoption of the regulation, the motions

indicate it was dealing with a published motions panel decision or one that set forth its reasoning. See United States v. Lopez-Armenta, 400 F.3d 1173, 1175 (9th Cir. 2005). It also dealt with the unique area of jurisdiction. See id. 8 Ranchers Cattlemen Action Legal Fund United Stockgrowers of Am. v. U.S. Dep’t of Agric., 499 F.3d 1108, 1114 (9th Cir. 2007); see also Stormans, Inc. v. Wiesman, 794 F.3d 1064, 1074, 1076 n.5 (9th Cir. 2015); Ctr. for Biological Diversity v. Salazar, 706 F.3d 1085, 1090 (9th Cir. 2013). 9 See Ctr. for Biological Diversity, 706 F.3d at 1090.

EAST BAY SANCTUARY COVENANT V. BIDEN 71

panel decision that the foreign affairs exception to the notice and comment procedures does not apply is binding. Id. at 775–77. In addition, while the motions panel decision regarding the good cause exceptions is not fully binding, what it did determine was that the information then brought to the attention of the panel and the district court did not suffice. Id. at 777–78. In light of that, I agree with the majority that merely adding the twenty-five-word sentence from a Washington Post article was insufficient to justify changing the motions panel result.

(3) The decisions made by the motions panel regarding harm to the Organizations and balance of hardships are also binding decisions regarding the propriety of the preliminary injunction. Id. at 767, 778–79.

(4) The scope of the injunction is not overly broad. Id. at 779–80.

Thus, I respectfully concur in the result of the majority opinion.

PAEZ, Circuit Judge, concurring in the denial of rehearing en banc, joined by FLETCHER, Circuit Judge.

We concur fully in the court’s denial of rehearing en banc. We respond to several of the arguments in Judge Bumatay’s and Judge VanDyke’s dissents.

I.

In his dissent from denial of rehearing en banc, Judge Bumatay takes issue with the entire doctrinal category of “organizational standing.” He contends that the amended majority opinion’s 1 application of Supreme Court and circuit precedent, concluding that the plaintiff Organizations established standing, deviates from John Marshall’s and James Madison’s respective visions of Article III. PJB Dissent 83–84, 91. The majority opinion, however, does not revise or deviate from organizational standing doctrine. Instead, it applies well-settled, binding authority. Its organizational standing holding fits squarely within the Supreme Court’s analysis in Havens Realty Corp. v. Coleman, 455 U.S. 363 (1982) and related cases in our court. Although the majority opinion thoroughly addresses the Organizations’ standing, we respond briefly to several of the points raised by Judge Bumatay.

The majority opinion makes clear that an organization suffers a cognizable Havens injury only if the challenged action “perceptibly impaired,” id. at 379, the organization by frustrating its mission and causing a diversion of resources in response to that frustration, Fair Hous. of Marin v. Combs, 285 F.3d 899, 905 (9th Cir. 2002). The majority opinion summarizes the record evidence demonstrating the various ways that the Rule’s extensive restrictions on asylum frustrated the missions of the Organizations and caused them to divert significant resources away from their mission in response. Amend. Op. 28–38.

Judge Bumatay’s dissent ignores the record evidence indicating the injuries suffered by the Organizations and substitutes his own conclusion that the Organizations simply refuse to see the Rule as a new opportunity to pursue their missions. PJB Dissent 89. Judge Bumatay quizzically asks, “But how does spending money assisting migrants seeking

1 For simplicity, hereinafter we refer to the amended majority opinion as the “majority opinion.”

EAST BAY SANCTUARY COVENANT V. BIDEN 73

asylum ‘frustrate’ the organizations’ mission to help migrants seeking asylum?” Id. The record evidence, as set forth in the Supplemental Excerpts of Record (“SER”), thoroughly documents how the Rule harmed the Organizations: 2

• “Since the new rule was announced, Al Otro Lado has been overwhelmed with children who traveled to the southern border of the United States to apply for asylum but now cannot do so. These children are very vulnerable, as they are unaccompanied and so have no one to look out for them in Tijuana, a city currently seeing record levels of violence…Caring for these children is incredibly time consuming and causing a near complete diversion of Al Otro Lado’s resources away from its core mission, which is providing legal services. In addition to providing these children with legal advice, attending to the many nonlegal needs of Al Otro Lado’s unaccompanied children clients is forcing Al Otro Lado to divert resources away from advising the many other clients who need the organization’s legal help, thus undermining our ability to achieve our core mission. For example, staff members are going with the children to try to get them on the list to cross at the port of entry. Staff have also had to negotiate with Mexican officials so that they will not take these children into custody and instead allow them to stay at a local youth shelter while Al Otro Lado determines the best way to ensure that the children have access to a legal means to seek asylum in the

2 Instead, Judge Bumatay answers with his own evidence-free musing. PJB Dissent 89 (“This sounds more like fulfilling their mission, rather than ‘frustrating’ it.”). United States.” SER 4 (Declaration of Erika Pinheiro).

• “Immediately after the interim final rule was announced, up to 15 children at a time were waiting in our office around the clock while we arranged their accommodations. [Al Otro Lado’s] office often felt more like a daycare center than a legal services organization. Because our office was being used to shelter our unaccompanied minor clients, we could not spend as much time with our other clients.” SER 22 (Declaration of Nicole Ramos).

• “[East Bay Sanctuary Covenant]’s mission is to serve poor individuals who have few resources and cannot afford attorneys…EBSC is neither staffed nor funded to provide representation for asylum applicants in removal proceedings. To do so would involve a fundamental change in our program and mission…[I]f the new policy remains in effect, EBSC stands to lose nearly all of our funding for our affirmative asylum program, because 80% of our clients in that program entered without inspection. But under the new policy, if they enter without inspection, they will no longer be able to access asylum at all.” SER 26–27 (Declaration of Michael Smith).

• “The new Proclamation would also put a financial strain on this organization…[F]or much of its work in removal proceedings, [Central American Resource Center] received a flat fee regardless of the time spent on a particular case. Because withholding and CAT cases are more resource-intensive than asylum cases, precluding certain CARECEN clients

EAST BAY SANCTUARY COVENANT V. BIDEN 75

from applying for asylum would create a greater financial loss in these cases. In other words, we would likely receive the same funding, but have to devote more hours per case.” SER 58 (Declaration of Daniel Sharp).

Judge Bumatay’s dissent ignores the record evidence, glosses over the majority opinion’s analysis of the evidence, and fails to cite a single case holding that the injuries which the Organizations suffered fall outside the Supreme Court’s decision in Havens. 3 His dissent also overlooks circuit precedent that addresses a variety of analogous situations in which organizations demonstrated they were “perceptibly impaired” by being forced to divert “significant resources” to establish standing under Havens. See 455 U.S. at 379; Fair Hous. of Marin v. Combs, 285 F.3d 899, 905 (9th Cir. 2002); La Asociacion de Trabajadores de Lake Forest v. Lake Forest, 624 F.3d 1083, 1088 (9th Cir. 2010); El Rescate Legal Servs., Inc. v. Exec. Office of Immigration Review, 959 F.2d 742, 745, 748 (9th Cir. 1991). Finally, his dissent ignores the fact that the opinion is consistent with the interpretation of organizational standing by other circuits. 4

3 Judge Bumatay’s attempt to reduce the majority’s analysis to, “Any loss of clients caused by the Rule, no matter how minimal, was sufficient to confer Article III standing,” PJB Dissent 90, is misguided and similarly ignores the record evidence. 4 See e.g., Centro de la Comunidad Hispana de Locust Valley v. Town of Oyster Bay, 868 F.3d 104, 109–111 (2d Cir. 2017) (holding that plaintiff organization dedicated to ending discrimination against immigrant workers had standing to challenge enforcement of ordinance that prohibited persons from stopping a vehicle to solicit employment because the organization had committed attention, time, money and personnel to prepare a response to the ordinance and because enforcement of the ordinance might deter the organization from In short, the majority opinion’s organizational standing holding is consistent with Supreme Court precedent, Article III, the decisions of our court and the decisions of other courts.

II.

We respond briefly to Judge VanDyke’s attacks on the integrity of the majority opinion.

First, Judge VanDyke accuses the majority of impropriety in addressing the effect of the motions panel’s published opinion on the merits panel’s subsequent consideration of the appeal. This issue was expressly raised in the parties’ briefs, and the majority opinion does nothing more than address it. The government’s opening brief argued that the merits panel was not bound by the motions panel’s published opinion. The Organizations answered by arguing that we were bound. The government replied to bolster their position that we were not bound.

Under the circumstances here, where the parties presented and fully briefed a potentially dispositive issue, it

counseling clients at a worksite where they congregated); OCA–Greater Houston v. Texas, 867 F.3d 604, 610–612 (5th Cir. 2017) (holding that a plaintiff organization that promoted voting among Asian-Americans had standing to challenge voting laws that restricted reliance on assistants to translate for voters with limited English-language abilities because the organization had redirected resources towards educating members and the public on how to comply with the law); Arcia v. Florida Secretary of State, 772 F.3d 1335, 1341–1342 (11th Cir. 2014) (holding that plaintiff organizations engaged in voter registration efforts had standing to challenge a program designed to remove individuals from voting rolls because the organizations diverted resources to address the program); see also 13A Charles Alan Wright et al., Federal Practice & Procedure § 3531.9.5 n.15 (3d ed.) (collecting cases).

EAST BAY SANCTUARY COVENANT V. BIDEN 77

was clearly appropriate for the majority to address the issue. This is how the process of adjudication is designed to function. The majority opinion abides by the Supreme Court’s unambiguous explanation that “in our adversarial system of adjudication, we follow the principle of party presentation…[I]n both civil and criminal cases, in the first instance and on appeal, we rely on the parties to frame the issues for decision and assign to courts the role of neutral arbiter of matters the parties present.” United States v. Sineneng-Smith, 140 S. Ct. 1575, 1579 (2020) (internal citations and quotation marks omitted). There is nothing remarkable about the majority opinion addressing an issue presented and framed by the parties.

Second, Judge VanDyke complains that the majority engages in “mischief,” LVD Dissent 104, by amending its opinion during the en banc process. The majority did, indeed, revise its opinion as a result of the en banc process. There was, however, no “mischief” in that decision. A brief explanation of our court’s en banc process provides the backdrop for the amended opinion.

After a judge of our court calls for a vote to determine whether a case should be reheard en banc under Fed. R. App. P. 35(f), our General Orders require the calling judge to “forward a memorandum” to all members of the court “setting forth reasons” for the call. G.O. 5.4.c.1. Thereafter, “any judge may circulate memoranda in response to [the] en banc call,” to facilitate robust discussion about the issues. G.O. 5.5.a. The opportunity to exchange views about the arguments in the en banc call includes the judges who authored the majority opinion. Id. The panel majority may propose revisions to address concerns raised by off-panel judges. Id. If, after the close of the memorandum exchange period the en banc call fails to obtain the votes of a majority, the panel resumes control of the case and may amend its opinion. G.O. 5.5.c.

Consistent with this process, the panel majority here considered the debate about law-of-the-case where a motions panel has published an order. The panel majority was persuaded to revise its opinion by the discussions in the memorandum exchange. The revisions, which were proposed to the court during the memorandum exchange period, are reflected in the amended majority opinion. The process by which the panel majority reached this conclusion is nothing new. The role this process plays to collaboratively strengthen the opinions of our court is acknowledged in several opinions. See e.g., Carver v. Lehman, 558 F.3d 869, 878–79 (9th Cir. 2009) (Smith, J.) (“Until the mandate has issued, opinions can be, and regularly are, amended or withdrawn, by the merits panel at the request of the parties pursuant to a petition for panel rehearing, in response to an internal memorandum from another member of the court who believes that some part of the published opinion is in error, or sua sponte by the panel itself…[This] collaborative process strengthens, not weakens, the final quality of those opinions, thereby better enabling them to stand the test of time, and engender the respect of thoughtful citizens for both the opinion, and the court that produced it.”); Perez v. City of Roseville, 926 F.3d 511, 526 (9th Cir. 2019) (Ikuta, J.) (“Like all three-judge panels, we must resolve the case before us to the best of our abilities, which may include reconsidering and revising an opinion that has not yet mandated.”).

In sum, there was nothing improper in the majority’s revision of its opinion to accommodate the views of our colleagues.

EAST BAY SANCTUARY COVENANT V. BIDEN 79

* * *

For all the above reasons, in addition to those in the amended opinion, we concur in the court’s denial of rehearing en banc.

BUMATAY, Circuit Judge, joined by IKUTA, BENNETT,

R. NELSON, LEE, and VANDYKE, Circuit Judges,

dissenting from the denial of rehearing en banc:

We are not “Platonic Guardians” of our nation’s public policies. See L. Hand, The Bill of Rights 73 (1958). As judges, we have no business standing athwart the choices of the political branches no matter how misguided we believe them to be. That fundamental limitation on our role is even more pronounced in the immigration context, where it is long settled that “the admission and exclusion of foreign nationals is a fundamental sovereign attribute exercised by the Government’s political departments largely immune from judicial control.” See Trump v. Hawaii, 138 S. Ct. 2392, 2418 (2018) (simplified). The Supreme Court has repeatedly warned us we overstepped our bounds when we tried to curtail immigration policies in the recent past. See, e.g., id.; Dep’t of Homeland Sec. v. Thuraissigiam, 140 S. Ct. 1959 (2020); Nielsen v. Preap, 139 S. Ct. 954 (2019); Barr v. E. Bay Sanctuary Covenant, 140 S. Ct. 3 (2019); Jennings v. Rodriguez, 138 S. Ct. 830 (2018). Unfortunately, we have not learned from our mistakes. Today, we once again second-guess the Executive’s immigration policies.

This time, we enjoin an immigration regulation temporarily limiting asylum eligibility to those who enter the country at a port of entry, deeming the policy “absurd.” E. Bay Sanctuary Covenant v. Trump (“East Bay”), 950 F.3d 1242, 1272 (9th Cir. 2020). To get there, we disregard two central precepts of the judicial role. First, we ignore constitutional limits on our jurisdiction by stretching organizational standing doctrine beyond Article III’s reach. Second, we re-write the asylum statute to add a prohibition on the Executive’s authority not found anywhere in the legislative text.

Whatever we may think of the prior administration’s immigration policies, we are not “free-wheeling enforcers of the Constitution and laws.” Initiative & Referendum Inst. v. Walker, 450 F.3d 1082, 1087 (10th Cir. 2006) (en banc). We lack any constitutional power to offer meaningful relief in the absence of an injured party. With no injured party here, we had no role to play in this dispute. And even if a proper party were before the court, our role would be a limited one—to determine the law as written by Congress. We have no authority to look beyond the plain text of a statute to manufacture a legislative prohibition against disfavored policies.

The proper forum for policy disputes is the ballot box, not the courthouse. Yet, by failing to revisit our decision here, we again ordain ourselves a super-legislature— presiding over our nation’s immigration policies. Because we donned the cloaks of Platonic Guardians rather than the robes of impartial judges in this case, I respectfully dissent from the denial of rehearing en banc.

I.

In late 2018, the Department of Justice and the Department of Homeland Security jointly adopted an interim final rule restricting eligibility for asylum. See Aliens Subject to a Bar on Entry Under Certain Presidential

EAST BAY SANCTUARY COVENANT V. BIDEN 81

Proclamations; Procedures for Protection Claims, 83 Fed. Reg. 55,934 (Nov. 9, 2018) (“Rule”). The Rule didn’t curb who may apply for asylum; rather, it would have made aliens who enter the United States “contrary to the terms of the proclamation or order” ineligible for asylum. Id. at 55,952. That same day, former President Trump issued a proclamation suspending the entry of non-permanent- resident aliens crossing from Mexico outside of a designated port of entry. See Proclamation No. 9822, Addressing Mass Migration Through the Southern Border of the United States, 83 Fed. Reg. 57,661 §§ 1, 2 (Nov. 9, 2018) (“Proclamation”). The suspension on entry was set to expire after 90 days. Id. at 57,663. Together, the Rule and Proclamation would have had the effect of barring asylum for any alien who illegally enters the United States between ports of entry on the southern border during the 90-day period.

The Rule never went into effect. We enjoined it immediately. East Bay Sanctuary Covenant v. Trump (“East Bay II”), 932 F.3d 742, 755 (9th Cir. 2018). Yet, the Rule was not challenged by any alien subject to the new asylum bar. Instead, the Rule was challenged by four lawyers’ organizations that represent current and future asylum- seekers. Id. On review of the merits, we concluded that the organizations had standing to bring suit and held that the Rule was “unlawful[]” because it conflicted with “the text and congressional purpose” of the asylum statute, 8 U.S.C. § 1158. East Bay, 950 F.3d at 1259. Both conclusions are wrong and should have been reviewed en banc. The Constitution limits our jurisdiction to “Cases” and “Controversies.” U.S. Const. art. III, § 2. This limitation is grounded in the Framers’ “concern about the proper—and properly limited—role of the courts in a democratic society.” Warth v. Seldin, 422 U.S. 490, 498 (1975). Outside of our jurisdiction, we lack the “power to declare the law.” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 94 (1998). To appreciate the jurisdictional bounds of the judicial power, “we must refer directly to the traditional, fundamental limitations upon the powers of common-law courts.” Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1550–51 (2016) (Thomas, J., concurring) (simplified).

At the time of the Founding, “the requirements of public control over public rights and private control over private rights predominated in American law.” Ann Woolhandler & Caleb Nelson, Does History Defeat Standing Doctrine?, 102 Mich. L. Rev. 689, 695 (2004). 1 Indeed, when it came to standing, “eighteenth- and nineteenth-century courts were well aware of the need for proper parties, and they linked

1 “Private rights” are those “belonging to individuals, considered as individuals.” 3 W. Blackstone, Commentaries on the Laws of England *2 (1768). These rights traditionally included the rights of personal security, property rights, and contract rights. Spokeo, 136 S. Ct. at 1550 (Thomas, J., concurring). By contrast, “only the government had the authority to vindicate a harm borne by the public at large[.]” Id. at 1551; see Lujan v. Defs. of Wildlife, 504 U.S. 555, 576 (1992) (“Vindicating the public interest (including the public interest in Government observance of the Constitution and laws) is the function of Congress and the Chief Executive.”).

EAST BAY SANCTUARY COVENANT V. BIDEN 83

that issue to the distinction between public and private rights.” Id. at 691. 2

From the earliest days of our Republic, “federal courts did not…entertain mandamus actions initiated by private relators who lacked private injury.” Id. at 707. When asked to command the Secretary of State to deliver Mr. Marbury’s judicial commission, the Supreme Court in Marbury v. Madison commented on the limits of the judicial power: “The province of the court is, solely, to decide on the rights of individuals, not to enquire how the executive, or executive officers, perform duties in which they have…discretion.” 5 U.S. (1 Cranch) 137, 170 (1803) (emphasis added). Chief Justice Marshall was careful to identify that the Court was adjudicating an individual right—that of Mr. Marbury to his commission—and not a “political” subject. Marbury, 5 U.S. at 166. Political subjects, by contrast, “respect the nation, not individual rights, and being entrusted to the executive, the decision of the executive is conclusive.” Id. 3

2 See, e.g., In re Wellington, 33 Mass. 87, 105 (1834) (“Undoubtedly,…a private individual can apply for a writ of mandamus only in a case where he has some private or particular interest to be subserved, or some particular right to be pursued or protected by the aid of this process, independent of that which he holds in common with the public at large; and it is for the public officers exclusively to apply, where public rights are to be subserved.”). 3 Chief Justice Marshall would later caution that “[i]f the judicial power extended to every question under the laws of the United States[,] [t]he division of power [among the branches of government] could exist no longer, and the other departments would be swallowed up by the judiciary.” Jonathan Adler, God, Gaia, the Taxpayer, and the Lorax: Standing, Justiciability, and Separation of Powers After Massachusetts and Hein, 20 Regent U. L. Rev. 175, 180–81 (2008) (quoting Chief Justice Marshall’s papers). That view of jurisdiction remained constant. Nearly a century and half later, the Supreme Court similarly observed that Article III authorizes federal courts to supervise the proper functioning of the political bodies “only to the extent necessary to protect justiciable individual rights against administrative action fairly beyond the granted powers.” Stark v. Wickard, 321 U.S. 288, 310 (1944).

But in the latter half of the 20th Century, a “sea change…occurred in the judicial attitude towards the doctrine of standing—particularly as it affects judicial intrusion into the operations of the other two branches.” See Antonin Scalia, The Doctrine of Standing as an Essential Element of the Separation of Powers, 17 Suffolk U. L. Rev. 881, 882–83 (1983); see, e.g., Flast v. Cohen, 392 U.S. 83 (1968) (holding federal taxpayers had standing to challenge federal expenditures that would assist religious schools under the Establishment Clause). Courts began venturing beyond protecting individual rights to serve the growing belief that the judicial branch should adjudicate the public law claims of public interest organizations. See, e.g., Scenic Hudson Pres. Conference v. FPC, 354 F.2d 608, 616 (2d Cir. 1965) (“In order to insure that the Federal Power Commission will adequately protect the public interest in the aesthetic, conservational, and recreational aspects of power development, those who by their activities and conduct have exhibited a special interest in such areas, must be held to be included in the class of ‘aggrieved’ parties[.]”).

In the last three decades, however, the Supreme Court has deliberately reined in Article III’s standing analysis. See, e.g., Lujan, 504 U.S. at 562–67 (holding that environmental organizations did not have standing to challenge environmental regulations when they suffered no direct injury); Spokeo, 136 S. Ct. at 1549 (holding that a

EAST BAY SANCTUARY COVENANT V. BIDEN 85

violation of a statutorily granted right does not automatically confer standing); Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 129 n.3 (2014) (calling into question prudential standing doctrine).

Nonetheless, lower courts still at times exhibit a flippant attitude towards Article III’s injury-in-fact requirements. Nowhere is this more evident today than in the muddled doctrine of organizational standing. See, e.g., People for the Ethical Treatment of Animals v. U.S. Dep’t of Agric., 797 F.3d 1087 (D.C. Cir. 2015) (granting standing to public interest groups that voluntarily expended resources to counteract government inaction with respect to environmental regulation); see id. at 1099 (Millett, J., dubitante) (declaring that this “ruling is in grave tension with Article III precedent and principles”).

But the era of judicial adventurism has passed. It is time we returned to our proper province—which “is, solely, to decide on the rights of individuals.” Marbury, 5 U.S. at 170.

B.

Against this backdrop, it is difficult to see how the legal services groups here could bring their suit under the traditional notions of Article III. Arguably the most important of Article III’s jurisdictional requirements is that plaintiffs have standing to bring their case to federal court. For either an individual or an organization to establish Article III standing, the party must show (1) an injury in fact that is concrete, particularized, and actual or imminent (not conjectural or hypothetical); (2) that the injury is fairly traceable to the challenged conduct; and (3) that the injury will be redressed by a favorable decision. Lujan, 504 U.S. at 561. Here, the plaintiff organizations cannot assert a proper injury-in-fact. The four organizations “share the same mission of assisting migrants seeking asylum.” East Bay, 950 F.3d at 1266. But the organizations didn’t bring the suit on behalf of client asylum-seekers harmed by the new Rule—as one might expect. Rather, they sued in their own right on behalf of their own interests. Id. As laudable as the organizations’ work may be, the Rule didn’t regulate their attorneys’ ability to provide legal services in any way; it only conditioned a successful asylum application on an alien’s manner of entry. The Rule therefore didn’t represent a direct harm to the organizations themselves. As the historical record shows, it is “substantially more difficult” to establish standing when a plaintiff asserts an injury arising from the government’s regulation of “someone else.” Lujan, 504 U.S. at 562.

So how, then, were the organizations able to meet Article III standing? First, the panel majority overcame the injury- in-fact hurdle by holding that the organizations were injured because the Rule amorphously “frustrate[d] their mission” by causing them to redirect resources to assist asylum- seekers in different ways. East Bay, 950 F.3d at 1266. Second, the panel majority proclaimed that the lawyers’ organizations were injured because the Rule would diminish their client base—announcing a shockingly broad rule that standing is conferred if a policy change would result in even “one less client” for an organization. Id. at 1267.

East Bay inaugurates a new standard of organizational standing—one that allows an organization to challenge a disfavored government policy by merely asserting that the change could result in “one less client” or cause it to shift resources to better support its mission. Id. at 1266–67. Such

EAST BAY SANCTUARY COVENANT V. BIDEN 87

a broad and malleable standard is an end-run around Article

III.

1.

Under traditional standing principles, the organizations could not have established an injury-in-fact. As a general rule, a plaintiff “must allege facts showing that he is himself adversely affected” in order to establish standing. Sierra Club v. Morton, 405 U.S. 727, 740 (1972) (emphasis added). A plaintiff does not suffer injury by voluntarily expending resources to counteract a governmental action that only indirectly affects the plaintiff. Clapper v. Amnesty Int’l USA, 133 S. Ct. 1138, 1148–51 (2013); see Linda R.S. v. Richard D., 410 U.S. 614, 619 (1973) (stating that a party’s interest in having the law properly enforced against others is not a cognizable injury under Article III). While vindicating one’s rights is central to our American legal tradition, it must be the party’s rights being vindicated. In this case, none of the organizations had a right that was violated by the Rule. As such, we had no right to interfere with the political branches.

The Rule imposed no new restriction or obligation on attorneys of asylum seekers. It only sought to add a condition for the successful grant of asylum. While it is easy to see how the Rule would have cut against the organizations’ goal to “assist[] migrants seeking asylum” as a policy matter, East Bay, 950 F.3d at 1266, “a setback to the organization’s abstract social interests” is not a concrete injury, Havens Realty Corp. v. Coleman, 455 U.S. 363, 379 (1982) (holding that injury occurs when a policy “perceptibly impair[s] [the organization’s] ability to provide…services for [its clients]”). “[A] mere interest in a problem, no matter how longstanding the interest and no matter how qualified the organization is in evaluating the problem, is not sufficient by itself to” confer standing. Sierra Club, 405 U.S. at 739 (simplified). Praiseworthy as the organizations’ mission may be, “Article III requires more than a desire to vindicate value interests.” Diamond v. Charles, 476 U.S. 54, 66 (1986). This principle is meant, in part, to ensure that courts do not wade into “political or ideological disputes about the performance of government.” United States v. Richardson, 418 U.S. 166, 192 (1974) (Powell, J., concurring).

To be sure, under our existing precedent, an organization can sue “on its own behalf…when it suffered both a diversion of its resources and a frustration of its mission.” La Asociacion de Trabajadores de Lake Forest v. City of Lake Forest, 624 F.3d 1083, 1088 (9th Cir. 2010) (simplified) (emphasis added). But that’s not enough. It must also “show that it would have suffered some other injury if it had not diverted resources to counteracting the problem.” Sierra Club v. Trump, 963 F.3d 874, 884 (9th Cir. 2020) (simplified). An organization is not injured by choosing, of its own volition, to spend more money or to shift resources from one mission goal to another. To be injured, the organization must expend resources outside of its mission. See Havens Realty Corp., 455 U.S. at 378–79 (recognizing injury when the organization devoted resources to identifying and counteracting racially discriminatory steering practices when its mission was to provide counseling and referral services to low- and moderate- income home-seekers). As recently as 2019, we found that Second Amendment rights groups had no standing to challenge the enforcement of a gun-confiscation law because implementation of the law did not “impede[] their ability to carry out their mission or require[] them to divert substantial resources away from the organizations’ preferred uses.”

EAST BAY SANCTUARY COVENANT V. BIDEN 89

Rodriguez v. City of San Jose, 930 F.3d 1123, 1135 (9th Cir. 2019).

But we lifted these restrictions on organizational expenditures in East Bay. Instead, we held that any change in resources—even if consistent with its mission—is enough to confer organizational injury. See East Bay, 950 F.3d at 1266 (holding the organizations were injured because the Rule “caused the Organizations to divert their already limited resources in response to the collateral obstacles it introduces for asylum-seekers”). For example, we noted how one organization had shifted resources to send staff to the border to be closer to migrant detention facilities. Id. But how does spending money assisting migrants seeking asylum “frustrate” the organizations’ mission to help migrants seeking asylum? This sounds more like fulfilling their mission than “frustrating” it. 4 See Fair Hous. Council of San Fernando Valley v. Roommate.com, LLC, 666 F.3d 1216, 1224 (9th Cir. 2012) (Ikuta, J., concurring in part and dissenting in part) (“How can an organization have a legally protected interest in not spending money to advance its core mission?”). Given that the Rule did not directly “impair[] [the organizations’] ability to” assist asylum-seekers, Havens Realty Corp., 455 U.S. at 379, we should have declined to find jurisdiction here.

2.

East Bay’s second ground of injury is even more problematic than the first. In addition to the diversion of 4 The panel seemingly concedes the deficiency of its opinion by relying now on extra facts from the supplemental record to support its analysis. See Concurrence in Order at 73. But future litigants should not have to pore through the supplemental record in order to reconstruct what the panel actually meant. resources, we recognized an injury to the organizations simply because the Rule would have reduced their client base. See 950 F.3d at 1266 (“Because the Rule significantly discourages a large number of asylum-seekers from seeking asylum given their ineligibility, the Rule frustrates [the organizations’] mission.”) (simplified). Any “los[s of] clients” caused by the Rule, no matter how minimal, was sufficient to confer Article III standing. Id. We even went so far as to announce that “one less client that [the organizations] may have had but-for the Rule’s issuance is enough” for constitutional injury. Id. at 1267.

As a result of our ruling, the distinction between injuries to an organization and the injuries to third parties served by the organization becomes meaningless. Under East Bay, a group has standing to petition our court simply by asserting that a government policy impacts third parties assisted by the organization. Such a low threshold, I predict, will lead to significant unintended consequences.

Under the “one less client” theory of injury, any group could challenge any governmental policy in its own right if the policy would have any de minimis effect on its client base. Law firms, for example, could directly challenge statutes and regulations without bothering to obtain an injured client. Public interest groups could seek to enjoin any regulation that targets its area of interest. We have moved well beyond requiring particularized and concrete injury and have embraced a “general grievance” theory of jurisdiction by construing organizational standing so broadly.

Jurisdiction is not like the Big Bang Theory, with an ever-expanding universe. Yet, East Bay’s sweeping pronouncement marks a new outer bound in our inflation of organizational standing.

EAST BAY SANCTUARY COVENANT V. BIDEN 91

C.

While the expansion of organizational standing doctrine ostensibly redounds to the benefit of public-interest groups, the real issue here is the aggrandizement of our own authority. Article III’s “limitations preserve separation of powers by preventing the judiciary’s entanglement in disputes that are primarily political in nature.” Spokeo, 136 S. Ct. at 1551 (Thomas, J., concurring). By ignoring these fundamental principles, we risk our courts becoming “publicly funded forums for the ventilation of public grievances.” Valley Forge Christian Coll. v. Ams. United for Separation of Church & State, Inc., 454 U.S. 464, 473 (1982).

By loosening organizational standing requirements, we only increase our own authority to adjudicate policy disputes. We no longer need to wait for a rule or regulation to actually injure a party. Now, we can skip ahead and immediately superintend any policy disagreement from the get-go by entertaining the bevy of public interest organizations willing to challenge the disfavored policy du jour. That makes us a super-legislature, not a court. Cf. The Federalist No. 47 (James Madison) (quoting 1 Montesquieu, Spirit of the Laws 181 (1750)) (“Were the power of judging joined with the legislative, the life and liberty of the subject would be exposed to arbitrary control, for the judge would then be the legislator.”). We should have fixed this problem by reviewing this case en banc and articulating a clear organizational standing doctrine grounded in Article III and the standing principles respected by our courts since the Founding. Compounding the error of improperly expanding organizational standing, our court also substantially misreads the asylum law’s text. The prior administration promulgated the Rule under the grant of authority Congress conferred on the Executive Branch to adopt limitations on asylum eligibility. See 8 U.S.C. § 1158(b)(2)(C). But instead of acceding to the express will of Congress and the Executive, we enjoin the Rule by manufacturing prohibitions in the text to avoid what we believe would be an “absurd” policy.

We start, as always, with the text of the statute. United States v. Graf, 610 F.3d 1148, 1166 (9th Cir. 2010). The asylum statute is divided into two relevant parts:

• Section 1158(a) – “Authority to apply for asylum” – provides that “any alien” who arrives in the United States may apply for asylum “whether or not [the alien arrives] at a designated port of arrival.” 8 U.S.C. § 1158(a)(1).

• Section 1158(b) – “Conditions for granting asylum” – permits the grant of asylum to an alien who meets the definition of a “refugee” and authorizes the Attorney General to “establish” by regulation “additional limitations and conditions, consistent with this section, under which an alien shall be ineligible for asylum.” Id. § 1158(b)(2)(C).

By these provisions, Congress endowed the Executive with broad discretion to regulate eligibility for asylum. The plain text and structure of the asylum statute indicate that Congress sought to separate the requirements and procedures for applying for asylum from granting asylum.

EAST BAY SANCTUARY COVENANT V. BIDEN 93

In subsection (a), Congress tells us who can apply for asylum. Subject to some exception, “any alien” has authority to do so, regardless of the alien’s manner of entry into the country. See id. § 1158(a)(1)–(2). In subsection (b), Congress tells us who is eligible to be granted asylum. See id. § 1158(b)(1)–(2). On that front, § 1158(b) broadly authorizes the Attorney General to promulgate regulations establishing conditions on eligibility for asylum. Id. § 1158(b)(2)(C). Accordingly, while allowing “any alien” to apply for asylum with little limitation, Congress also affords the Executive extensive authority to regulate who receives asylum. Importantly, the sole textual limitation on Congress’s grant of authority is that any regulation be “consistent with” the rest of § 1158.

The Rule easily fits within the asylum scheme enacted by Congress. The Rule does nothing to limit the ability of aliens to apply for asylum and only restricts the parameters for a successful asylum petition. Under a straightforward reading of § 1158, the Executive was permitted to do so. Contrary to the panel majority’s decision here, the Rule does not conflict with any provision of § 1158. The bifurcation of the statute indicates that the two authorities—applying for asylum and receiving asylum—are purposefully distinct. The Executive cannot interfere with the former but can expressly alter the latter. And it is perfectly “consistent” to allow an alien to apply for asylum generally and to restrict asylum eligibility to only a subset of those who apply. This is especially so where, as here, the restriction is of limited time and scope. 5 See 83 Fed. Reg. 57,661 §§ 1, 2 (confining

5 Contrary to the panel’s opinion here—and fatal to its analysis— the Rule doesn’t serve as a complete or categorical ban on eligibility for all aliens arriving outside of a port of entry. See East Bay, 950 F.3d at 1272 (describing the Rule as a “categorical ban”). the Rule to a 90-day period along the southern border). And the temporary restriction was promulgated to respond to a specific crisis at the border. See East Bay II, 932 F.3d at 754 (describing the asylum system as “overburdened” and acknowledging that courts were receiving a “staggering increase in asylum applications,” with a backlog “exceed[ing] 200,000,” resulting in “thousands of [asylum] applicants who had been detained by immigration authorities [being] released into the United States”). Accordingly, nothing in the plain language or structure of § 1158 precludes temporarily conditioning eligibility for asylum on the alien’s manner of entry—especially given the Executive’s concern for the situation at the border.

Despite the clear grant of discretionary authority to the Executive, our court strikes down the Rule by inventing a conflict with Congress’s supposed intent. See East Bay, 950 F.3d at 1272–73 (holding the Rule “unlawful” because “it conflicts with the plain congressional intent”). As explained above, the Rule is wholly consistent with § 1158(a); yet the panel professes that it “would be hard to imagine a more direct conflict than the one presented here.” Id. at 1272 (simplified).

The Supreme Court has warned us against going beyond the “plain language” of the statute and “read[ing] into the text additional limitations designed to narrow the scope” of the statute. Millbrook v. United States, 569 U.S. 50, 55 (2013). By grafting § 1158(a)’s broad scope of who can apply for asylum onto § 1158(b)’s necessarily narrower scope of who may (or may not) receive asylum, we do just that.

The panel’s reading essentially conflates applying for and receiving asylum. But just because an alien can apply for asylum does not entitle him or her to receive asylum.

EAST BAY SANCTUARY COVENANT V. BIDEN 95

Instead, whether an alien is ultimately granted asylum is purely a matter of executive grace. See 8 U.S.C. § 1158(b)(1)(A) (establishing that the Executive “may grant asylum”). 6 Indeed, the government is authorized to deny a bid for asylum even if the “alien is a refugee within the meaning of [the statute].” Id. § 1158(b)(1)(A), (b)(2)(A). All the more baffling, the panel even acknowledges that the Executive may consider an alien’s manner of entry when adjudicating an individual asylum case. See East Bay, 950 F.3d at 1273 (“[We] have long recognized that a refugee’s method of entering the country is a discretionary factor in determining whether the migrant should be granted [asylum].”). If the Executive can consider manner of entry in individual cases, it stands to reason it can do so more broadly as well. Cf. Lopez v. Davis, 531 U.S. 230, 243–44 (2001) (rejecting argument that the Bureau of Prisons was required to make “case-by-case assessments” of eligibility for sentence reductions and explaining that an agency “is not required continually to revisit ‘issues that may be established fairly and efficiently in a single rulemaking’”) (simplified).

The panel also justified its departure from the plain text by arguing that this reading would lead to “absurd results.” 950 F.3d at 1272 (“Explicitly authorizing a refugee to file an asylum application because he arrived between ports of entry and then summarily denying the application for the

6 See Moncrieffe v. Holder, 569 U.S. 184, 187 (2013) (describing grants of asylum as a “form[] of discretionary relief”); INS v. Aguirre– Aguirre, 526 U.S. 415, 420 (1999) (explaining that the “decision whether asylum should be granted to an eligible alien is committed to the Attorney General’s [and the Secretary’s] discretion” under the INA); INS v. Cardoza-Fonseca, 480 U.S. 421, 443 (1987) (holding that an alien who satisfies asylum requirements under the INA “does not have a right to remain in the United States; he or she is simply eligible for asylum, if the Attorney General, in his discretion, chooses to grant it”). same reason borders on absurdity.”). The “absurdity canon isn’t a license for us to disregard statutory text where it conflicts with our policy preferences.” Tamm v. UST–U.S. Trustee, Honolulu (In re Hokulani Square, Inc.), 776 F.3d 1083, 1088 (9th Cir. 2015). And frankly, there is no absurdity here at all. The Rule says that everyone who arrives outside of a port of entry is able to apply for asylum, but because of the identified migrant crisis, for the 90-day period at the southern border, applicants must come in through a port of entry to successfully gain asylum. While the panel majority may disagree with that policy decision, there is nothing absurd about it. Indeed, § 1158 sets numerous categorical exclusions from asylum eligibility for aliens who are statutorily authorized to apply for asylum. See 8 U.S.C. § 1158(b)(2).

If Congress wanted to preclude the Executive from basing asylum decisions on the manner of the alien’s arrival, it could have easily said so. Simmons v. Himmelreich, 136 S. Ct. 1843, 1848 (2016) (“[W]e presume Congress says what it means and means what it says.”). But here, we have supplied the policy preference. Although the panel majority purports to divine Congress’s intent, that is beside the point. Based on the text of the statute, the asylum-application provision expressly authorizes aliens who enter between ports of entry to file asylum applications, while the asylum- granting provision contains no such guarantee. Instead of trying to extract the intent of some 535 legislators in enacting § 1158, we should have simply adhered to the plain meaning of the text.

We should have been especially reluctant to exercise judicial power to circumvent the political branches’ implementation of immigration law. Just last term, the Supreme Court explained,

EAST BAY SANCTUARY COVENANT V. BIDEN 97

[T]he power to admit or exclude aliens is a sovereign prerogative[.] [T]he Constitution gives the political department of the government plenary authority to decide which aliens to admit, and a concomitant of that power is the power to set the procedures to be followed in determining whether an alien should be admitted.

Thuraissigiam, 140 S. Ct. at 1982 (simplified).

The Rule, therefore, rests on solid statutory ground. Congress permits the Attorney General to establish by regulation “additional limitations and conditions” on asylum eligibility. 8 U.S.C. § 1158(b)(2)(C). The Attorney General did just that. That should have been the end of the analysis. Nevertheless, we substitute our version of the asylum law for the one actually passed by Congress.

IV.

For these reasons, I respectfully dissent from the denial of rehearing en banc.

VANDYKE, Circuit Judge, dissenting from denial of rehearing en banc:

I agree with Judge Bumatay’s analysis and join his dissent in full. I write separately to address the East Bay panel majority’s amended decision—specifically its past and present treatment of the binding effect of motions panels’ published opinions—and the tortured path that got us here. It was never necessary in this case for the panel majority to wade into this issue at all (more on that later), but both the panel’s original and now-amended analyses are equally unsatisfying and will undoubtedly lead to more mischief on an important issue that we will wrestle with for years to come. Although the revised East Bay panel opinion now no longer purports to overrule Lair v. Bullock, 798 F.3d 736 (9th Cir. 2015)—which, of course, it never had the authority to do in the first place—the replacement rationale it provides is no less troublesome, and can only serve as an ever-ready escape valve for merits panels who wish to disregard a published motions panel decision that decided effectively identical issues.

I.

A.

Some table-setting is necessary. First, the law. Before this case, our law-of-the-circuit rule vis-à-vis a motions panel’s published opinion was well-established and straightforward. In Lair v. Bullock, a panel of our court held as a matter of first impression that “a motions panel’s published opinion binds future panels the same as does a merits panel’s published opinion.” 798 F.3d at 747. There’s not much ambiguity in that statement.

But what about our law-of-the-case jurisprudence? Does that separate doctrine somehow undermine Lair’s clear ruling by allowing a merits panel to avoid being bound by a motions panel’s published opinion in the same case? That question was resolved by our court years before Lair. “No,” we said, en banc. See Gonzalez v. Arizona, 677 F.3d 383, 389 n.4 (9th Cir. 2012) (en banc) (“We now hold that the exceptions to the law of the case doctrine are not exceptions to our general ‘law of the circuit’ rule, i.e., the rule that a published decision of this court constitutes binding authority which ‘must be followed unless and until overruled by a

EAST BAY SANCTUARY COVENANT V. BIDEN 99

body competent to do so.’”) (quoting Hart v. Massanari, 266 F.3d 1155, 1170 (9th Cir. 2001)).

Taken together, Lair and Gonzalez were therefore clear and categorical: a motions panel’s published opinion is the law of the circuit—like any other published panel opinion— for all subsequent panels, even the merits panel in the same appeal. 1

B.

Next, this case. On October 1, 2019, the same merits panel heard and submitted this and another high-profile case, both challenging Trump Administration immigration policies. See E. Bay Sanctuary Covenant v. Trump (“East Bay”), 950 F.3d 1242 (9th Cir. 2020); Innovation Law Lab v. Wolf (“ILL”), 951 F.3d 1073 (9th Cir. 2020), cert. granted, 141 S. Ct. 617 (2020). In both cases, district courts had enjoined the government from carrying out the Administration’s policies, and the government applied to this court for emergency stays of those injunctions pending appeal. In this case (East Bay), a motions panel of our court denied the government’s stay request in a published opinion. See E. Bay Sanctuary Covenant v. Trump, 932 F.3d 742, 755 (9th Cir. 2018). But in ILL, a different motions panel granted the government’s stay request, again in a published opinion. See Innovation Law Lab v. McAleenan (“ILL (motions panel)”), 924 F.3d 503, 510 (9th Cir. 2019).

Both cases proceeded to be heard by the same panel considering the merits of the district courts’ preliminary 1 It’s worth emphasizing up front that I’m somewhat ambivalent as to whether the Lair-Gonzalez rule was the correct rule. But right or wrong, it clearly was the rule in our circuit, and if we wanted to change it, we should have done so en banc. injunctions. That panel published decisions in both cases on the same day: February 28, 2020. In East Bay, a two-judge majority on the merits panel agreed with the motions panel and affirmed the district court’s injunction. In ILL, the same merits panel majority also affirmed the district court’s injunction. 2 But the merits panel in ILL had a Lair problem: the motions panel’s published decision had reached a directly contrary result on at least one major issue in the case. Compare, e.g., ILL (motions panel), 924 F.3d at 509 (“[W]e conclude that DHS is likely to prevail on its contention that § 1225(b)(1)” is consistent with the Administration’s policy.), with ILL, 951 F.3d at 1081 (“We conclude that plaintiffs have shown a likelihood of success on their claim that [the policy] is inconsistent with 8 U.S.C. § 1225(b).”).

So the panel majority resourcefully applied some good old-fashioned judge-jitsu, admirable perhaps for its audacity if not its propriety or subtlety. The majority first turned to East Bay—where Lair was both not at issue (because the motions and merits panels in East Bay reached the same conclusion) and, for that reason, not particularly relevant— to create a new rule effectively toppling Lair and neutralizing its thorny imposition on the panel majority’s desired outcome in ILL. In East Bay, the majority’s opinion announced a new rule: “we treat the motions panel’s decision as persuasive, but not binding.” East Bay, 950 F.3d at 1265. The decision simply asserted that Lair’s holding was dicta, and that even if it were precedential, there were “good policy and practical reasons for departing from

2 Judge Fernandez, the third judge on both merits panels, wrote separately in both cases because, applying Lair, he believed the merits panel in both cases was bound by the earlier published motions panel opinions. East Bay, 950 F.3d at 1284 (Fernandez, J., concurring in the result); ILL, 951 F.3d at 1095 (Fernandez, J., dissenting).

EAST BAY SANCTUARY COVENANT V. BIDEN 101

Lair[],” including law-of-the-case considerations. See id. at 1262–63.

That alone was remarkable for its chutzpah, considering it casually disregarded one of our court’s foundational rules undergirding the rule of law itself: that three-judge panels cannot overturn prior panel opinions unless they are “clearly irreconcilable with the reasoning or theory of intervening higher authority.” Miller v. Gammie, 335 F.3d 889, 893 (9th Cir. 2003) (en banc); see also United States v. Contreras, 593 F.3d 1135, 1136 (9th Cir. 2010) (en banc). But a full- blown critique of this rationale is unnecessary now because the majority has amended its opinion and thereby withdrawn some more offensive aspects of its prior analysis. To summarize, though, the East Bay majority’s curt analysis in its initial opinion rather gruffly dismissed Lair and merely paid lip service to Gonzalez, which is why it was later forced to amend its opinion.

But the real head-scratcher for the original East Bay decision was why the majority promulgated its new rule at all. If you were looking only at East Bay itself, you would wonder what prompted the majority there to even delve into the Lair discussion. Again, in East Bay, the motions panel and the merits panel were in vigorous agreement that the government was unlikely to succeed on the merits. So why drag Lair into it?

The reason is obvious once you zoom out: because of ILL, the other case decided by the same panel on the same day. In ILL, the same two-judge majority needed to navigate around a published motions panel decision that went the other way. See ILL, 951 F.3d at 1081 (“A preliminary question is whether a merits panel is bound by the analysis of a motions panel on a question of law, performed in the course of deciding an emergency request for a stay pending appeal.”). And how did the ILL majority answer that question? By dutifully following precedent: “On that question, we follow East Bay…in which we held that a motions panel’s legal analysis, performed during the course of deciding an emergency motion for a stay, is not binding on later merits panels.” Id. The only thing missing from the majority’s ILL decision is a solemn citation to Miller v. Gammie so everyone can know how serious the majority was about faithfully adhering to binding precedent—its own, in East Bay, that is. 3

To summarize ILL: the motions panel in a published opinion addressed the same legal questions presented to the merits panel, and under Lair and Gonzalez, the motions panel’s resolution of those questions should have controlled. But the merits panel majority didn’t like the result reached by the motions panel. It needed that published opinion to be

3 The panel majority defends its original East Bay decision rejecting Lair by reaching for the oldest excuse in the Book (literally): “But we were asked to do it!” See Genesis 3:12 (NIV) (“The man [Adam] said, ‘The woman you put here with me—she gave me some fruit from the tree, and I ate it.’”). The obvious response is no less antediluvian, and will be familiar to any parent: just because you were asked to do something doesn’t mean you should have. Indeed, here the panel majority has implicitly acknowledged it shouldn’t have done what it originally did, by amending its opinion to remove the more offensive aspects of its original rationale. So even assuming the parties had asked the panel to do what it did, that wouldn’t excuse it. But, in fact, the panel majority can’t fairly blame the parties. The panel in this case only needed to address Lair if it departed from the motions panel’s conclusions, which it didn’t. The government didn’t even cite Lair, much less argue Lair’s rule was dicta, until a footnote in its reply brief. The panel majority purported to nullify Lair in its original opinion only because it needed Lair’s rule gone for ILL—not because Lair stood in the way of the panel’s conclusion in this case (which the panel’s amended opinion demonstrates, by reaching the same conclusion but without dissing Lair).

EAST BAY SANCTUARY COVENANT V. BIDEN 103

“persuasive,…not binding.” East Bay, 950 F.3d at 1265. So the majority overruled Lair in East Bay—where it was completely unnecessary to do so—so that it could immediately apply East Bay’s ILL-conceived rule in a decision where it could do some real work. 4

Back to East Bay: after the government’s request for en banc rehearing, the majority agreed to amend its opinion; a majority of our court was unwilling to vote in favor of en banc rehearing; and we arrived at our current juncture. The East Bay panel majority’s amended opinion has now nixed its categorical rejection of Lair. It replaces its prior repudiation of Lair with one of the oldest common law axioms: a “published motions panel order may be binding as precedent for other panels deciding the same issue.” It vaguely concedes that “[t]here may be circumstances where a motions panel does answer the same legal question that is presented to the merits panel.” But, as you might have guessed, the majority still concludes that in East Bay itself, the motions panel opinion “is not binding…because the issues are different.” That way, you see, East Bay can still purport to serve as precedent for ILL and subsequent decisions that have involved Trump Administration policies

4 Along with the same policy arguments outlined in East Bay, the ILL majority noted that its disregard of the motions panel’s earlier published decision was especially appropriate in ILL because, in that case, one of the ILL motions panel judges had “expressed the hope that the merits panel, with the benefit of full briefing and argument, would decide the legal questions differently.” ILL, 951 F.3d at 1081. For those not tracking the judicial composition of the various panels, that same motions panel judge was also part of the two-judge majority on the East Bay and ILL merits panels, thus triggering the venerable “I-told-you-so” rationale for disregarding precedent. where the merits panel disagreed with the published motions panel opinion. 5

In just a moment, I’ll discuss the deficiencies of the now- amended East Bay opinion’s seemingly innocuous and banal reasoning, and why we still should have taken this case en banc despite the panel’s retrenchment. But first, it’s important to recap what happened here. Originally, the East Bay majority adopted a clearly indefensible rationale to fabricate a new rule it could apply in a different case, where it needed to circumvent a motions panel opinion that decided the very same issue of law. See ILL, 951 F.3d at 1081 (reaching a directly contrary legal conclusion to that of the motions panel on whether the executive branch’s policy was consistent with the relevant statute). After being called to the carpet by a petition for en banc rehearing, the majority stepped back from that blatantly inappropriate power-grab behind its original East Bay rule. Of course, that’s easy enough to do now. The original East Bay rule has already served its purpose: the majority got to its preferred outcome in ILL by applying the precedent-disregarding “rule” it manufactured in East Bay. Discarding the old East Bay rule now, after the fact, and in a case where the “rule” was complete dicta to begin with, is pretty much costless. Mischief managed.

5 At least one other merits panel of our court followed ILL’s lead when it declined to be bound by a published motions panel opinion. That case involved the Public Charge Rule, where, as discussed further below, the motions panel reached legal conclusions contrary to those of the merits panel majority on virtually every question in the case. Compare generally City & County of San Francisco v. USCIS, 981 F.3d 742 (9th Cir. 2020) (merits panel), with City & County of San Francisco v. USCIS, 944 F.3d 773 (9th Cir. 2019) (motions panel).

EAST BAY SANCTUARY COVENANT V. BIDEN 105

Or maybe not. In the real world, the lingering effects of pulling a stunt like this don’t just vanish. Even putting aside the problems identified in Judge Bumatay’s dissent, and the problems with East Bay’s new replacement rationale discussed in the next section, this looks bad for our court— because it is. Even if none of the other reasons for taking this case en banc had merit (which they do), we should have done so if only to demonstrate to the public that we don’t approve of the barely disguised shenanigans that happened in East Bay and ILL. It would have helped promote public confidence in the integrity and impartiality of our court. 6

II.

It is true that the amended rationale in the East Bay majority’s opinion doesn’t represent the ham-fisted upheaval its original opinion did. But it merely trades one evil for another. Instead of rudely casting aside precedent, it distinguishes—rather than overrules—Lair, creating a new standardless standard that, as a practical matter, will allow any merits panel to disregard a motions panel’s published decision resolving virtually identical claims. In addition to the several compelling reasons given in Judge Bumatay’s dissent as to why we should have taken this case en banc,

6 The panel majority misconstrues what I characterize as its “mischief.” I agree “there was nothing improper in the majority’s revision of its opinion to accommodate the views of our colleagues.” The “mischief” I reference was the panel’s “ILL-conceived” rule from its original opinion—not the panel “amending its opinion during the en banc process.” My point is that while the panel majority (and apparently a majority of our court) may think that the panel has “managed” its original mischief by amending its opinion, it hasn’t—especially when you consider the indiscernible mess left by the panel’s new but still unsatisfying rationale, combined with our other recent decisions that I discuss in detail below. there are at least two more considerations—related to each other—that further counseled in favor of reviewing the revised East Bay opinion.

First, the East Bay majority’s new standard is indeterminate at best, retreating to the truism that two things are different, except when they’re not: “The published motions panel order may be binding as precedent for other panels deciding the same issue, but it is not binding here.” This nebulous standard is completely unhelpful to future panels attempting to apply our precedent. It is only helpful if we want to maximize merits panels’ discretion to do whatever they want with published motions panels’ opinions in the same case. It is terrible for predictability and the rule of law, and should have been considered en banc for this reason alone.

Second, and relatedly, although there might be one logically satisfying way to reconcile East Bay’s new “sometimes-it’s-binding, sometimes-it’s-not” rule with Lair, that specific reconciliatory interpretation of the East Bay majority’s rule obviously cannot be what the East Bay majority actually meant. How do we know? Because of what the judges from the East Bay majority have done in subsequent cases. Their clear actions prevent their mushy words in East Bay from meaning the only thing that might have been meaningful.

So after East Bay we really are left with a truly ambiguous standard giving merits panels maximum discretion and minimal guidance. To explain requires me to first elaborate on how one could try to reconcile Lair and the new East Bay, and then describe how the panel majority’s actions in subsequent cases make clear that that can’t be what East Bay means.

EAST BAY SANCTUARY COVENANT V. BIDEN 107

Here is how an intrepid law professor might try to reconcile Lair and East Bay: The amended East Bay opinion no longer rejects Lair out-of-hand, but nods to Lair while distinguishing it by retaining a lengthy but abstract discussion about the differences between adjudicating stay applications and preliminary injunctions. The opinion engages in a superficial and formalistic analysis of the technical distinctions between stays and preliminary injunctions, but never pins down what those differences might mean here in the real world. Once we dip below 30,000 feet, however, we observe that whether a motions panel grants or denies a stay should have very different implications for a subsequent panel considering the merits of the preliminary injunction itself. And therein lies a theoretical key to reconciling Lair and East Bay.

To see how the stay and preliminary injunction standards interact differently depending on whether a stay is granted or denied requires looking closely at the standards themselves. A party seeking a preliminary injunction before a district court bears the burden of establishing:

[1] that he is likely to succeed on the merits, [2] that he is likely to suffer irreparable harm in the absence of preliminary relief, [3] that the balance of equities tips in his favor, and [4] that an injunction is in the public interest.

Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). By comparison, after a district court issues a preliminary injunction, the losing party seeking a stay of that injunction bears the considerably heavier burden of establishing:

(1) [that] the stay applicant has made a strong showing that he is likely to succeed on the merits; (2) [that] the applicant will be irreparably injured absent a stay; (3) [that] issuance of the stay will [not] substantially injure the other parties interested in the proceeding; and (4) [that] the public interest lies [with the issuance of a stay].

Nken v. Holder, 556 U.S. 418, 426 (2009) (emphases added) (citation omitted). And the stay applicant doesn’t present its arguments in a vacuum, but must overcome a presumptively valid lower court decision that went the other way. See Sw. Voter Registration Educ. Project v. Shelley, 344 F.3d 914, 918 (9th Cir. 2003) (explaining that appellate review of district court preliminary injunctions is “limited[,] and deferential”). On top of that, the stay applicant bears the substantial burden of strongly showing that it is the party who is likely to succeed on the merits. And the applicant must establish it “will be irreparably injured absent a stay.” Nken, 556 U.S. at 434 (emphasis changed). Compare those showings to the less demanding “likely to succeed on the merits” and “likely to suffer irreparable harm” of the preliminary injunction inquiry. Given the higher standards and inverted burdens, it is no surprise that stay applications are (and should be) denied most of the time; they are granted even more infrequently than are preliminary injunctions, which are themselves an “extraordinary remedy.” Winter, 555 U.S. at 22.

Because of that, losing a stay motion (even in a published opinion) 7 logically should not usually foreordain that party’s fate on the merits of the preliminary injunction. A stay applicant unable to meet the onerous stay standard could still

7 Most motions panel decisions aren’t published, and therefore cannot tie the hands of later merits panels. See Haggard v. Curry, 631 F.3d 931, 933 n.1 (9th Cir. 2010) (per curiam); 9th Cir. R. 36-3(a).

EAST BAY SANCTUARY COVENANT V. BIDEN 109

prevail in reversing a district court preliminary injunction, since the burden before the merits panel flips back to the original injunction proponent. See id. at 20. In most cases, like East Bay, the interplay of the shifting stay and injunction standards means that a motions panel’s denial of a stay won’t control the outcome of the later panel just considering the merits of the preliminary injunction itself. So, if it could be limited to that context, the amended East Bay opinion’s observation that published stay opinions are not always binding on later merits panels could make some sense. It also might not be inconsistent with Lair, because unlike in East Bay, Lair involved a situation where a stay was granted by the motions panel.

In a Lair-type circumstance—that is, in the doubly rare case where the motions panel granted a stay of the preliminary injunction in a published opinion—the stay applicant is necessarily very likely to succeed on all the legal factors relevant to the preliminary injunction itself. If a motions panel concludes that an applicant has demonstrated (1) a strong likelihood of success on the merits, (2) that it will be irreparably harmed absent a stay, (3) that other parties won’t be substantially harmed by the stay, and (4) that the public interest is served by a stay, then a fortiori the other party will almost always be prevented from meeting the considerably less burdensome preliminary injunction standard—unless the merits panel is allowed to just disregard the motions panel’s conclusions. Both parties can’t likely succeed on the merits. Both can’t show that the equities and public interest tip in their favor. That’s simply not how the standards logically interact. If, in a published decision, a motions panel concludes that the stay applicant surmounted its Sisyphean burden—despite the odds stacked against it—that’s generally a conclusive legal determination that the applicant should also prevail on the merits, and so a preliminary injunction is unwarranted. 8

One might conclude, therefore, that, because Lair involved precisely this rare situation of a merits panel faced with a published motions panel decision granting a stay, this is the circumstance Lair had in view for its categorical motions-panels-opinions-bind-later-merits-panels rule. And, in contrast, East Bay’s new observation that motions panel decisions don’t always bind later merits panels is simply a reflection of the reality in that case—that the earlier motions panel had denied the stay. Accordingly, one could argue that the East Bay panel majority’s focus on the distinctions between stays and preliminary injunctions was intended to serve as precisely this basis for reconciling East Bay and Lair—i.e., that East Bay applies if the motions panel denied the stay, but Lair applies if the motions panel granted the stay.

But here in the real world, whoever makes this otherwise possibly satisfying academic argument is delusional. The East Bay panel majority can’t have meant to reconcile Lair and East Bay this way because the very same two-judge panel majority immediately cited and relied on East Bay to ignore a motions panel opinion granting a stay in ILL, and has never retreated from that. In fact, one of those two judges did the same thing again months later as part of the

8 The exception might arise in a more traditional case where, after a stay is granted, the parties further develop the record in front of the district court before having the merits of their preliminary injunction considered on appeal. In that case, a merits panel might actually be reviewing a meaningfully different case than the motions panel did. But as the procedural histories of the various cases recounted herein illustrate, the development of our caselaw in this area does not seem to be driven much by “traditional” cases—at least not recently.

EAST BAY SANCTUARY COVENANT V. BIDEN 111

merits panel majority in the Public Charge cases, which also involved a published motions panel opinion granting a stay that was subsequently ignored by the two-judge merits panel majority. So that theoretically reconciliatory reading of Lair and East Bay, as nice as it might be, just isn’t on the table— unless we want to ignore reality.

As if ILL were not bad enough, our merits panel decision in the Public Charge cases highlights just how unstructured our treatment of published motions panel stay opinions has become. In three recent cases argued and decided together by this court, challengers to the Trump Administration’s revamped “Public Charge” rule obtained preliminary injunctions from two district courts in our circuit. The government sought an emergency stay of these injunctions, which a motions panel of our court granted in a lengthy published opinion. According to the motions panel, the government “mustered a strong showing of likelihood of success on the merits.” City & County of San Francisco, 944 F.3d at 807. More specifically, the motions panel concluded:

• The new rule was not contrary to law. See id. at 798 (“Congress has not spoken directly to the interpretation of ‘public charge’ in the INA. Nor did it unambiguously foreclose the interpretation articulated in the Final Rule.”).

• The new rule was not arbitrary and capricious. See id. at 790.

• The government would suffer irreparable harm absent a stay of the injunctions. See id. at 806.

• A stay would serve the public interest. See id. at 807. Contrast these clear holdings with those of the two-judge merits panel majority, which entirely reversed course on the same record and issues. City & County of San Francisco, 981 F.3d at 763. The merits panel majority concluded:

• “[T]he plaintiffs have demonstrated a high likelihood of success in showing that the Rule is inconsistent with any reasonable interpretation of the statut[e].” Id. at 758.

• “We must conclude that the Rule’s promulgation was arbitrary and capricious.” Id. at 762.

• Plaintiffs established that they will likely suffer irreparable harm absent the preliminary injunctions. See id.

• The public interest and equities favor the injunctions. See id.

Same case, same legal issues, no interstitial factual developments—diametrically opposed legal conclusions. 9 The merits panel majority in the Public Charge appeal explained that it could depart from the carefully-reasoned holdings of the motions panel because (a) it had the benefit of more briefing on the same issues (including amicus briefs); and (b) some other circuits had since reached conclusions contrary to those of the motions panel. See id. at 753–54. Because our Public Charge decision never

9 As a result, we now have on-point, circuit precedent in two different published opinions (from the motions panel and the merits panel) that reach opposite conclusions in the Public Charge Rule cases. Which binding precedent should the lower courts, and panels in subsequent cases, follow? Addressing this problem is another reason our court should have taken this case en banc.

EAST BAY SANCTUARY COVENANT V. BIDEN 113

mentioned East Bay or ILL (or Lair or Gonzalez, or, for that matter, any precedent for disregarding the motions panel’s published legal conclusions), it isn’t clear whether the panel majority was implicitly applying the East Bay/ILL rule, 10 or whether it was creating yet another novel rule that motions panels’ opinions are not binding whenever there is a lot more merits briefing—including some really good amicus briefs—and other circuits have reached a different conclusion than did our motions panel. 11

As these cases illustrate, the undeniable reality is that after Gonzalez, Lair, East Bay, ILL, and City & County of San Francisco have all been thrown in the mixer, our circuit doesn’t have anything close to a cognizable rule about how merits panels should treat motions panels’ earlier published stay opinions. Perhaps, sitting en banc, we could adopt the approach described above reconciling Lair and East Bay. Or we could make a rule that motions panels’ decisions granting a stay should always be unpublished, and thus not binding. Or, amending our earlier Gonzalez decision, we could conclude that motions panels’ decisions—published or not—are not binding on later panels in the same appeal (including merits panels), and address how our court and the

10 As already noted, the Public Charge merits panel majority included an overlapping judge from the East Bay and ILL panel majority, so the Public Charge panel majority could hardly have been unaware of the East Bay rule, or Lair. 11 Of course, if other circuits’ intervening contrary decisions were actually a legitimate consideration in deciding whether to apply binding circuit precedent, then so might be the fact that the U.S. Supreme Court had stayed those very same contrary decisions. See Dep’t of Homeland Sec. v. New York, 140 S. Ct. 599, 599 (2020) (staying Second Circuit’s preliminary injunction); Wolf v. Cook County, 140 S. Ct. 681, 681 (2020) (same for Seventh Circuit). district courts in other cases should treat inconsistent published motions and merits opinions from the same case. Our court sitting en banc might come up with any number of good solutions to the problems presented in cases like this. But until we do, we and those who litigate before us must content ourselves with the slop at the intersection of East Bay, ILL, City & County of San Francisco, Lair, Gonzalez— and whatever future panel opinions applying those divergent precedents deliver. It’s a pity that a majority of our court could not be persuaded to take this case en banc to untangle the muddle we’ve made of the Lair–Gonzalez rule.

III.

Recognizing the amended East Bay rationale for what it is—a plastic truism that, without now saying so, still subversively displaces Lair’s rule—it’s hard to see how anyone could think this is a good thing for the state of the law in our circuit. Going forward, the hyper-elasticity of the new Lair–East Bay standard will undoubtedly be useful for merits panels seeking flexibility, but obviously less helpful to litigants seeking clarity and predictability. This problem is not going to go away. Given our circuit’s propensity to be the venue of choice to overrule certain administrations’ policies, we are sure to see this problem manifest again (and again) before too long. Until we fix it, everybody will know that whether a motions panel’s published stay opinion has any binding force just depends on which merits panel you draw—not the law. Cite Lair. Cite East Bay. Cite ILL or City & County of San Francisco. Trust us to pick which one fits best.

That breaks faith with our most fundamental obligation, which is to apply the law, not our preferences. Panel majorities should not be allowed to simply change the rules when they find them inconvenient obstacles to preferred

EAST BAY SANCTUARY COVENANT V. BIDEN 115

substantive outcomes. In the end, this type of overt results- oriented judging can’t help but discredit our court and the rule of law generally.

Any of these considerations should individually have been important enough to correct East Bay en banc. But apparently, even in toto, they were not. I must therefore respectfully express my disappointment.