Index Newspapers LLC v. United States Marshals Service

FILED FOR PUBLICATION OCT 9 2020 UNITED STATES COURT OF APPEALS MOLLY C. DWYER, CLERK U.S. COURT OF APPEALS

FOR THE NINTH CIRCUIT

INDEX NEWSPAPERS LLC, DBA No. 20-35739 Portland Mercury; DOUG BROWN; BRIAN CONLEY; SAM GEHRKE; D.C. No. 3:20-cv-01035-SI MATHIEU LEWIS-ROLLAND; KAT District of Oregon, MAHONEY; SERGIO OLMOS; JOHN Portland RUDOFF; ALEX MILAN TRACY; TUCK WOODSTOCK; JUSTIN YAU, and those similarly situated, ORDER

Plaintiffs-Appellees, v.

UNITED STATES MARSHALS SERVICE; U.S. DEPARTMENT OF HOMELAND SECURITY,

Defendants-Appellants, and

CITY OF PORTLAND, a municipal corporation; JOHN DOES, 1-60; individual and supervisory officers of Portland Police Bureau and other agencies working in concert,

Defendants. Before: O’SCANNLAIN, RAWLINSON, and CHRISTEN, Circuit Judges. Order by Judges RAWLINSON and CHRISTEN, Dissent by Judge O’SCANNLAIN

On May 25, 2020, George Floyd was killed by a Minneapolis police officer while being arrested. Bystanders on the sidewalk recorded videos of a police officer kneeling on Floyd’s neck for several minutes while Floyd begged for his life. A video showing the last minutes of Floyd’s life was circulated nationwide, and it ignited protests across the country in support of the Black Lives Matter movement.

This case arises out of the protests in Portland, Oregon. Most of the protests have been peaceful, but some have become violent. There have been incidents of vandalism, destruction of property, looting, arson, and assault, particularly late at night. Since the protests began, state and local authorities in Oregon have actively monitored the protests and engaged in crowd control measures. Plaintiffs—a newspaper organization and individual journalists, photojournalists, and legal observers who have attended the protests to serve as reporters and recorders—filed a class-action complaint against the City of Portland on June 28, 2020.

The complaint alleged that the City’s response to the protests violated their rights under the First and Fourth Amendments to the United States Constitution, and Article I, Sections 8 and 26 of the Oregon Constitution. Specifically, plaintiffs asserted that although they had not participated in the protests, the local authorities shot them with less-lethal munitions (pepper balls, impact munitions, paint *3 markers, and tear gas canisters), and pepper sprayed, shoved, and otherwise prevented them from recording and reporting on the protests and on law enforcement’s response to the same. Four days after the complaint was filed, on July 2, the district court entered a temporary restraining order (TRO) against the City regulating the local authorities’ use of crowd-control tactics against journalists and legal observers. On July 16, the City and plaintiffs stipulated to a preliminary injunction that was largely identical to the TRO.

Many of the protests in Portland have centered around the Mark O. Hatfield Federal Courthouse. In response to the threat to federal property, the Department of Homeland Security (DHS) and the United States Marshals Service (USMS) (collectively, the Federal Defendants) deployed federal law enforcement agents to Portland. It appears undisputed that the intensity of the protests escalated after the Federal Defendants arrived.

Plaintiffs filed a second amended complaint on July 17 joining as defendants DHS and USMS. This complaint alleged that the Federal Defendants “intentionally targeted and used physical force and other forms of intimidation against journalists and authorized legal observers for the purpose of preventing or deterring them from observing and reporting on unreasonably aggressive treatment *4 of lawful protestors.” The district court entered a TRO against the Federal Defendants on July 23.

On July 29, 2020, DHS and the State of Oregon reached an agreement regarding their respective crowd control efforts. The agreement is not part of the record, but the district court described it as generally providing that the City would take the lead in responding to the protests. The court’s findings also made clear that the agreement contains numerous caveats and is terminable at any time, without notice. Though the agreement was to take effect on July 29, the district court observed that the record includes video clips that purport to show federal agents firing tear gas and less-lethal munitions at journalists standing on SW Main Street on July 29 and into the morning of July 30. The district court found that “there was no one nearby on the street but numerous federal enforcement officers and six journalists when the munitions were deployed.”

The Federal Defendants assert that the Oregon State Police are no longer enforcing crowd control in Portland, and that the Portland Police are currently filling that role instead. But it is clear that the federal agents have remained in *5 Portland, and Acting Secretary of DHS, Chad Wolf, stated that “no determination of timetables for reduction in protective forces has yet been made.”1

On August 10, plaintiffs filed a motion for a preliminary injunction against the Federal Defendants. After briefing was complete, the parties stipulated that the court could base its decision on the record and the parties’ arguments without holding an evidentiary hearing. The record comprises dozens of declarations, many of which include photographs and links to video files. The district court issued a detailed, sixty-one page order granting plaintiffs’ motion on August 20 and entered a preliminary injunction with terms largely identical to the terms of the July 23 TRO.

The district court’s order began by observing that the Constitution reserves the general police power to the states, and pursuant to the general police power, local officials have the authority to issue general dispersal orders on the public streets and sidewalks. The court noted that the City had separately stipulated that it would not require members of the press or legal observers to disperse, and

1 On July 28, plaintiffs filed a motion for a finding of contempt and imposition of sanctions against the Federal Defendants, alleging several violations of the July 23 TRO. The district court has not yet ruled on the motion, but noted “serious concerns” that the Federal Defendants had not complied with the July 23 TRO, and that some of the alleged misconduct occurred after the Federal Defendants reached the agreement with Governor Brown.

explained that the Federal Defendants did not assert the authority to issue general dispersal orders to clear city streets and that the statutory authority the Federal Defendants relied upon did not so provide. The court’s order recounts the Federal Defendants’ position, which was that federal officers had been dispatched to Portland with the stated mission to protect federal property and personnel. Nevertheless, the district court was confronted with compelling photographic evidence showing that federal officers “routinely have left federal property and engaged in crowd control and other enforcement on the streets, sidewalks and parks of the City of Portland.” The court’s order detailed several of the dozens of declarations, photos, and video clips introduced into evidence to support plaintiffs’ contention that at least some of the federal officers had intentionally targeted journalists and legal observers in retaliation for their news-reporting efforts.

Having explained that local officials had separately stipulated they were not requiring journalists and legal observers to disperse, the preliminary injunction entered to address the Federal Defendants’ conduct states that journalists and legal observers “shall not be subject to arrest for not dispersing following the issuance of an order to disperse.” The order states that journalists and legal observers may not impede, block, or otherwise physically interfere with the lawful activities of the Federal Defendants, and recognizes that the Federal Defendants are free to issue *7 “otherwise lawful crowd-dispersal orders for a variety of lawful reasons;” i.e. crowd-dispersal orders not issued to clear city streets and sidewalks. The preliminary injunction also requires that journalists and observers “must comply with all laws other than general dispersal orders.”

Because the Federal Defendants argued that some protestors had masqueraded as members of the press by wearing press badges or clothing identifying them as members of the press corps, the order provides that it does not protect unlawful conduct and that anyone, even a person who appears to be a journalist, is subject to arrest for engaging in such conduct. Finally, the injunction sets out a number of indicia to assist the Federal Defendants in distinguishing between journalists, legal observers, and protesters. These indicia include visual identifiers such as press passes, people standing off to the side of protests not engaging in protest activities, people not intermixed with protest activities, and people carrying professional-grade photographic equipment. The order requires that the Federal Defendants’ uniforms bear marks allowing federal officers to be identified. The injunction also provides that if a journalist or legal observer is incidentally exposed to crowd-control devices after remaining in the area where such devices are deployed to enforce a lawful dispersal order, the Federal Defendants will not be liable for violating the injunction.

On August 25, the district court denied the Federal Defendants’ motion for a stay of the preliminary injunction pending appeal, principally concluding that the Federal Defendants had not shown a sufficient likelihood that they would suffer irreparable injury absent a stay. On appeal, a divided three-judge motions panel issued a brief, two-page order on August 27 granting the Federal Defendants’ motion for an administrative stay of the injunction pending resolution of their emergency motion for a stay pending appeal.

Having considered the parties’ complete briefing, and after hearing oral argument, we conclude that the Federal Defendants have not shown a strong likelihood of success on the merits. The Federal Defendants also failed to demonstrate they are likely to suffer irreparable injury if the preliminary injunction is not stayed pending appeal. Accordingly, we deny the Federal Defendants’ emergency motion.

I

“A stay is not a matter of right, even if irreparable injury might otherwise result to the appellant.” Virginian Ry. Co. v. United States, 272 U.S. 658, 672 (1926). “The party requesting a stay bears the burden of showing that the circumstances justify an exercise of that discretion.” Nken v. Holder, 556 U.S. 418, 433–434 (2009).

To decide whether to grant the Federal Defendants’ motion for a stay pending appeal, our case law requires that we consider: (1) whether the Federal Defendants have made a strong showing that they are likely to succeed on the merits; (2) whether the Federal Defendants will be irreparably injured absent a stay; (3) whether issuance of the stay will substantially injure the other parties interested in the proceeding; and (4) where the public interest lies. Id. at 426.

To decide whether the Federal Defendants have demonstrated a likelihood that they will succeed on the merits of their claims, we review the district court’s findings of fact for clear error, its legal conclusions de novo, and the injunction’s scope for abuse of discretion. Armstrong v. Brown, 768 F.3d 975, 979 (9th Cir. 2014); Walters v. Reno, 145 F.3d 1032, 1047 (9th Cir. 1998) (“A district court’s factual findings are entitled to deference unless they are clearly erroneous.”).

II

The bar for obtaining a stay of a preliminary injunction is higher than the Winter standard for obtaining injunctive relief. Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). We have explained that the first two Nken factors are the most critical, and that the second two factors are only considered if the first two factors are satisfied. Nken, 556 U.S. at 434–35; Al Otro Lado v. Wolf, 952 F.3d 999, 1007 (9th Cir. 2020). The Federal Defendants must show a strong likelihood *10 of success on the merits. Doe #1 v. Trump, 957 F.3d 1050, 1062 (9th Cir. 2020). And “simply showing some possibility of irreparable injury fails to satisfy the second factor.” Nken, 556 U.S. at 434–35 (internal citations and quotations omitted). The demanding standard applicable here requires that the Federal Defendants show “that irreparable injury is likely to occur during the period before the appeal is decided.” Doe #1, 957 F.3d at 1059.

A

The Federal Defendants argue they are likely to succeed on the merits for three reasons. First, they argue plaintiffs lack standing to pursue injunctive relief on their First Amendment retaliation claim because plaintiffs have not shown a sufficient likelihood that they will be deprived of their constitutional rights if the Federal Defendants’ crowd control measures are not subject to the district court’s preliminary injunction pending appeal. Second, they argue they will succeed on the merits of plaintiffs’ retaliation claim because there is no evidence to support the district court’s conclusion that plaintiffs’ protected activity was a substantial or motivating factor that prompted the Federal Defendants’ actions to disperse them. Third, the Federal Defendants argue they are likely to succeed on plaintiffs’ First Amendment right-of-access claim because the press and legal observers have no First Amendment right to access the streets and sidewalks where the protests are *11 staged if the Federal Defendants order them to disperse. For these three reasons, the Federal Defendants argue they are entitled to a stay of the preliminary injunction pending appeal.2

Three elements make up the “irredicuble constitutional minimum of standing”: (1) injury in fact; (2) a causal connection between the injury and the conduct complained of; and (3) a likelihood that the injury will be redressed by a favorable decision. Lujan v. Defenders of Wildlife, 504 U.S. 555, 560–561 (1992). Here, only the “injury in fact” element is disputed.

“A plaintiff threatened with future injury has standing to sue ‘if the threatened injury is certainly impending, or there is a substantial risk the harm will occur.’” In re Zappos.com, Inc., 888 F.3d 1020, 1024 (9th Cir. 2018) (quoting Susan B. Anthony List v. Driehaus, 573 U.S. 149, 158 (2014)). A plaintiff may not rely “on mere conjecture about possible governmental actions” to demonstrate injury, and must instead present “concrete evidence to substantiate their fears.”

2 Our case law has frequently observed the importance of the press as surrogates for the public, but we have not considered whether legal observers serve the same function. Neither the parties nor the district court focused on whether the legal observers’ right of access differs from the one enjoyed by the press. Because we do not need to decide this question in order to rule on the emergency motion for a stay, we leave it for another day.

Clapper v. Amnesty Int’l USA, 568 U.S. 398, 420 (2013). “Past exposure to illegal conduct does not in itself show a present case or controversy regarding injunctive relief . . . if unaccompanied by any continuing, present adverse effects.” City of Los Angeles v. Lyons, 461 U.S. 95, 102 (1983) (quoting O’Shea v. Littleton, 414 U.S. 488, 495–96 (1974)).

The Federal Defendants’ standing argument relies primarily on Lyons, a case involving a claim for injunctive relief asserted by a man who had been subjected to a chokehold by police officers. Id. at 102. In Lyons, the Court explained that to establish standing, the plaintiff was required to “credibly allege that he faced a realistic threat from the future application of the City’s [chokehold] policy.” Id. at 106 n.7. Because Lyons had not been subjected to a second chokehold in the time before he filed his federal complaint, the Supreme Court concluded that his assertion that he might face such abuse in the future was premised on a speculative sequence of events. Id. at 105–06. The Supreme Court explained that Lyons did not have standing to pursue equitable relief barring the use of chokeholds because “[p]ast exposure to illegal conduct does not in itself show a present case or controversy regarding injunctive relief . . . if unaccompanied by any continuing, present adverse effects.” Id. at 102 (emphasis added) (citation omitted).

Here, plaintiffs’ injuries are different for several reasons. First, their risk of future injury is not speculative. Plaintiffs introduced powerful evidence of the Federal Defendants’ ongoing, sustained pattern of conduct that resulted in numerous injuries to members of the press between the date the complaint was filed and the date the district court entered its preliminary injunction. The district court’s preliminary injunction included twelve pages solely dedicated to factual findings that describe in detail dozens of instances in which the Federal Defendants beat plaintiffs with batons, shot them with impact munitions, and pepper sprayed them. The court’s findings were supported by nineteen declarations and video and photographic evidence. The Federal Defendants do not argue that any of the district court’s findings are clearly erroneous, and we conclude the findings are amply supported.

As of the time the preliminary injunction was entered, the district court found that the Federal Defendants had engaged in a pattern of conduct that had persisted for weeks and was ongoing. After reviewing plaintiffs’ declarations, photos, and video clips, the district court found that many victims had been standing on public streets, sidewalks, and parks, well away from protestors, and were not engaged in unlawful activity when they were shot, tear gassed, shoved, or pepper sprayed by the Federal Defendants. Unlike Lyons, the district court found *14 that some journalists and legal observers monitoring the protests had been injured by the Federal Defendants more than once. The district court’s findings are compelling because “the possibility of recurring injury ceases to be speculative when actual repeated incidents are documented.” Thomas v. Cnty. of Los Angeles, 978 F.3d 504, 507 (9th Cir. 1992) (internal quotation marks omitted).

The nature of plaintiffs’ injuries also sharply differs from the substantive due process injury asserted in Lyons. Plaintiffs allege that the Federal Defendants’ crowd-control measures have “chilled” the exercise of their First Amendment rights, and that this First Amendment injury is ongoing. A chilling of First Amendment rights can constitute a cognizable injury, so long as the chilling effect is not “based on a fear of future injury that itself [is] too speculative to confer standing.” Munns v. Kerry, 782 F.3d 402, 410 (9th Cir. 2015) (citing Clapper, 568 U.S. at 417–18); Libertarian Party of L.A. Cty. v. Bowen, 709 F.3d 867, 870 (9th Cir. 2013) (“[A]s the Supreme Court has recognized, a chilling of the exercise of First Amendment rights is, itself, a constitutionally sufficient injury.”).

The district court agreed that the Federal Defendants’ targeting of the plaintiffs chilled their First Amendment rights, and after analyzing the factors prescribed by Furgatch, the court concluded that the Federal Defendants’ conduct *15 was likely to continue.3 Fed. Election Comm’n v. Furgatch, 869 F.2d 1256, 1263 n.5 (9th Cir. 1989). The district court issued a lengthy and detailed order and the Federal Defendants do not challenge its factual findings. On this record, we conclude the Federal Defendants have not made a strong showing that their standing argument is likely to succeed, and have not shown that the district court abused its discretion by entering a preliminary injunction. This cuts against the emergency motion for a stay pending appeal.

We also conclude the Federal Defendants have not made the strong showing required by Nken that they are likely to succeed on the merits of plaintiffs’ First Amendment retaliation claim. For this claim, plaintiffs were required to show that they were engaged in a constitutionally protected activity, the Federal Defendants’ actions would chill a person of ordinary firmness from continuing to engage in the protected activity, and the protected activity was a substantial or motivating factor

3 Furgatch instructs courts to consider five factors when determining whether conduct is likely to occur in the future: (1) the degree of scienter involved; (2) the isolated or recurrent nature of the infraction; (3) the defendant’s recognition of the wrongful nature of his conduct; (4) the extent to which the defendant’s professional and personal characteristics might enable or tempt him to commit future violations; and (5) the sincerity of any assurances against future violations. Fed. Election Comm’n v. Furgatch, 869 F.2d 1256, 1263 n.5 (9th Cir. 1989). The Federal Defendants do not argue that the district court misapplied any of these factors, and we see no error.

in the Federal Defendants’ conduct. Pinard v. Clatskanie Sch. Dist. 6J, 467 F.3d 755, 770 (9th Cir. 2006). The Federal Defendants do not contest the first or second elements of the retaliation claim, nor does there appear to be a good faith basis for doing so.4

The Federal Defendants only argue that they are likely to succeed on the merits of plaintiffs’ retaliation claim because “plaintiffs have not shown their First Amendment activity was a ‘substantial or motivating factor’ in the government’s conduct.” This element of a First Amendment retaliation claim may be met with either direct or circumstantial evidence, and we have said that it involves questions of fact that normally should be left for trial. Ulrich v. City & Cty. of San

4 As to the first element, plaintiffs were clearly observing and recording law enforcement activity in public, as the district court found. Fordyce v. City of Seattle, 55 F.3d 436, 439 (9th Cir. 1995) (recognizing plaintiff was exercising his “First Amendment right to film matters of public interest” when filming activities of police officers during a public protest march). The First, Third, Fifth, Seventh and Eleventh Circuits have all recognized the public’s First Amendment right to observe and film police activities in public. See Fields v. City of Philadelphia, 862 F.3d 353, 359–60 (3d Cir. 2017); Turner v. Lieutenant Driver, 848 F.3d 678, 688 (5th Cir. 2017); Gericke v. Begin, 753 F.3d 1, 7 (1st Cir. 2014); ACLU of Illinois v. Alvarez, 679 F.3d 583, 600 (7th Cir. 2012); Smith v. City of Cumming, 212 F.3d 1332, 1333 (11th Cir. 2000). As to the second element of the retaliation claim, the Federal Defendants do not challenge the district court’s finding that being shot with less-lethal munitions like pepper balls, tear gas, and paint-marking munitions, being pepper sprayed at close range, or being shoved by a law enforcement officer would chill a person of ordinary firmness from continuing to exercise their First Amendment rights.

Francisco, 308 F.3d 968, 979 (9th Cir. 2002). The district court’s extensive and thorough factual findings provide robust support for its conclusion that plaintiffs’ exercise of their First Amendment rights was a substantial or motivating factor in the Federal Defendants’ conduct. To highlight just four of the district court’s findings:

• • On July 29, plaintiff Brian Conley was wearing a photographer’s vest marked “PRESS,” a helmet marked “PRESS,” and was carrying a large camera with an attached LED light and telephoto lens. After reviewing video footage submitted by plaintiffs, the district court found that Conley was filming a line of federal officers moving down the street pepper spraying peaceful protesters—including spraying a woman in the face at point blank range who was on her knees in the middle of the street with her hands up—when, without warning, a federal officer pepper sprayed Conley at point blank range.
• • On the night of July 19, Jungho Kim, a photojournalist, was wearing a neon yellow vest marked “PRESS” and a white helmet marked “PRESS” on the front and rear. The district court found that Kim was standing alone, about 30 feet from federal agents, taking photographs, when suddenly and without warning, Kim was shot in the chest, just below his heart with a less-lethal munition. A photograph submitted with Kim’s declaration shows that he was shot where the word “PRESS” was printed on his vest.
• • On the night of July 26, Daniel Hollis, a videographer, was wearing a press pass and a helmet marked “PRESS” in bright orange tape, and carrying a large, professional video-recording camera. Hollis was filming a group of federal agents massed outside the federal courthouse. “Almost immediately,” the federal agents shot at him, striking him just left of his groin. He turned and began to run away, but was shot again in the lower back.

• • On July 27, Amy Katz, a photojournalist, was wearing a hat and tank top marked “PRESS” and carrying a camera with a telephoto lens while covering the protests. Katz was photographing a federal agent who pushed a man down a flight of stairs while arresting him. Another federal agent physically blocked Katz and tried to stop her from photographing the arrest. Katz stepped to the side to continue photographing the arrest, and the federal agent physically shoved her away.

Plaintiffs’ expert witness, Gil Kerlikowske, provided a declaration supporting the district court’s conclusion that these incidents were retaliatory in nature and did not reflect appropriate crowd-control tactics.5 Kerlikowske opined that defending the federal courthouse in Portland mainly involves establishing a perimeter around the building, and that there is no need to target or disperse journalists. According to Kerlikowske, in crowd-control situations it is inappropriate to shoot non-lethal munitions at a person’s head, chest, or back. Kerlikowske also opined that pepper balls and tear gas canisters should not be aimed at people at all, as those munitions are intended to be shot at the ground

5 The district court found Kerlikowske to be a “qualified, credible, and persuasive expert witness.” Kerlikowske is a former Commissioner of U.S. Customs and Border Protection, served as the Chief of Police in Seattle, Washington for 10 years, and as the Police Commissioner in Buffalo, New York. The district court recognized Kerlikowske’s “substantial training and experience with crowd control and civil unrest in the context of protests [and] use of force in that context,” and observed that Kerlikowske has “led and orchestrated the policing of hundreds of large and potentially volatile protests, many of which were considerably larger than the recent protests in Portland.”

where they explode and release their contents into the air. In his view, virtually all of the journalists’ injuries were caused by the improper use of force, including shooting people who were not engaged in threatening acts, and the Federal Defendants’ misuse of crowd-control munitions.

All told, the district court’s findings describe at least forty-five instances similar to the four highlighted here, and all of them occurred between July 15 and July 30 while plaintiffs were observing and recording the Black Lives Matter protests in downtown Portland. The forty-five instances were “just several examples selected” by the district court “from the extensive evidence provided by Plaintiffs.” The court was clear that “[t]here are more.” Plaintiffs submitted a total of nineteen declarations with their motions for a temporary restraining order and preliminary injunction. Many of the events described by the declarations were corroborated by accompanying photographs and video clips.

Because the district court’s findings include so many instances in which plaintiffs were standing nowhere near protesters while photographing and observing the Federal Defendants’ actions, they provide exceptionally strong evidentiary support for the district court’s finding that some of the Federal Defendants were motivated to target journalists in retaliation for plaintiffs’ exercise of their First Amendment rights. Indeed, in response to this shocking pattern of *20 misconduct, the dissent contemplates that plaintiffs’ allegations may well support Bivens actions and claims of excessive force against individual federal agents.6

The evidence that at least some of the Federal Defendants’ conduct was retaliatory supports the district court’s conclusion that the plaintiffs are likely to succeed on the merits of their retaliation claim. On this record, we do not hesitate to conclude that the Federal Defendants have not made the required strong showing that they are likely to prevail on the merits of the claim. This evidence of retaliatory conduct also cuts against the emergency motion for a stay pending appeal.7

The Federal Defendants have not shown that they are likely to succeed on the merits of plaintiffs’ First Amendment right-of-access claim. To begin, the Federal Defendants reframe the issue and mischaracterize the preliminary injunction as recognizing a special, across-the-board exemption for members of the press and legal observers. But the threshold issue presented is whether plaintiffs

6 A Bivens claim requires a showing of purposeful misconduct. See Ashcroft v. Iqbal, 556 U.S. 662, 676–77 (2009). 7 The dissent argues that the retaliation claim does not justify enjoining the Federal Defendants from issuing dispersal orders because the dispersal orders themselves are not retaliatory. This argument overlooks that the preliminary injunction expressly states the Federal Defendants are not precluded from issuing lawful crowd-dispersal orders for a variety of reasons.

have a constitutionally protected right to access the public forum where the protests are staged, and as the district court observed, the preliminary injunction does not afford plaintiffs any special rights beyond those enjoyed by the general public pursuant to the First Amendment.

In Press-Enterprise II, the Supreme Court articulated a two-part test to determine whether a member of the public has a First Amendment right to access a particular place and process. Press-Ent. Co. v. Superior Court of Cal., 478 U.S. 1 (1986). First, a court must ask “whether the place and process has historically been open to the press and general public” and “whether public access plays a significant positive role in the functioning of the particular process in question.” Id. at 8. If a qualified right of access exists, the government can overcome that right and bar the public by showing that it has “an overriding interest based on findings that closure is essential to preserve higher values and is narrowly tailored to serve that interest.” Id. at 9.

The Federal Defendants argue that the press is not entitled to any special First Amendment right of access to observe and record the protests taking place on Portland’s streets and sidewalks. But the Press-Enterprise II test is not dependent upon plaintiffs’ occupation, and plaintiffs do not argue that it affords them a special right of access not shared by the general public. We agree with plaintiffs *22 that the press is entitled to a right of access at least coextensive with the right enjoyed by the public at large; the press is certainly not disfavored. See Pell v. Procunier, 417 U.S. 817, 833–34 (1974). Indeed, the Supreme Court has repeatedly observed that excluding the media from public fora can have particularly deleterious effects on the public interest, given journalists’ role as “surrogates for the public,” Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 572–73 (1980); Cox Broad. Corp. v. Cohn, 420 U.S. 469, 492 (1975) (“Without the information provided by the press most of us and many of our representatives would be unable to vote intelligently or to register opinions on the administration of government generally.”). Recognizing the outsized effect of denying access to the press, we have observed that the Supreme Court’s Press-Enterprise II test “balances the vital public interest in preserving the media’s ability to monitor government activities against the government’s need to impose restrictions if *23 necessary for safety or other legitimate reasons.” Leigh v. Salazar, 677 F.3d 892, 900 (9th Cir. 2012).8

The Federal Defendants do not contest that the place—Portland’s streets and sidewalks—and the process—public protests and law enforcement’s response to them—have historically been open to the public. See Hague v. Comm. for Indus. Org., 307 U.S. 496, 515 (1939) (“Wherever the title of streets and parks may rest, they have immemorially been held in trust for the use of the public and, time out of mind, have been used for purposes of assembly, communicating thoughts between citizens, and discussing public questions.”).

Public demonstrations and protests are clearly protected by the First Amendment, and a protest not open to the press and general public is not a public demonstration. See, e.g., Snyder v. Phelps, 562 U.S. 443, (2011) (reiterating that “speech on matters of public concern . . . is at the heart of the First Amendment’s protection” (internal quotation marks omitted)); City of Houston v. Hill, 482 U.S.

8 The Press-Enterprise II test emerged from a line of cases involving access to criminal judicial proceedings, but by its terms the test is not limited to any particular type of plaintiff or any particular type of forum. The Ninth Circuit and several other courts have applied Press-Enterprise II’s analytical framework to other settings, including planning commission meetings, student disciplinary records, state environmental agency records, settlement records, transcripts of state utility commission meetings, resumes of candidates for school superintendents, and legislator’s telephone records, among others. See Leigh, 677 F.3d at 899 and n.5 (collecting cases).

451, 472 (1987) (“[T]he First Amendment recognizes, wisely we think, that a certain amount of expressive disorder not only is inevitable in a society committed to individual freedom, but must itself be protected if that freedom would survive.”); NAACP v. Claiborne Hardware Co., 458 U.S. 886, 907–12 (1982) (holding that a boycott of local businesses “clearly involved constitutionally protected activity” including “speech, assembly, association, and petition”); Collins v. Jordan, 110 F.3d 1363, 1371 (9th Cir. 1996) (“Activities such as demonstrations, protest marches, and picketing are clearly protected by the First Amendment.”).

Nor do the Federal Defendants deny that public access plays a significant positive role in the functioning of our democracy. Just as streets and sidewalks historically have been recognized as being open to the public, the press has long been understood to play a vitally important role in holding the government accountable.9 Indeed, the public became aware of the circumstances surrounding George Floyd’s death because citizens standing on a sidewalk exercised their First

9 Leigh, 677 F.3d at 897 (“A popular Government, without popular information, or the means of acquiring it, is but a Prologue to a Farce or a Tragedy; or, perhaps both.” (quoting 9 WRITINGS OF JAMES MADISON 103 (G. Hunt ed. 1910))).

Amendment rights and filmed a police officer kneeling on Floyd’s neck until he died.

“The free press is the guardian of the public interest,” and “[o]pen government has been a hallmark of our democracy since our nation’s founding.” Leigh, 677 F.3d at 897, 900. “In a society in which each individual has but limited time and resources with which to observe at first hand the operations of his government, he relies necessarily upon the press to bring to him in convenient form the facts of those operations.” Cox Broad. Corp., 420 U.S. at 490–91. Transparency assures that the government’s response is carried out “fairly to all concerned,” and public access discourages “misconduct of participants, and decisions based on secret bias or partiality.” Richmond Newspapers, 448 U.S. at 569. Given our deeply entrenched recognition of the public’s right to access city streets and sidewalks, circuit precedent establishing the right to film public police activity, and the broadly accepted principle that the public’s interest is served by the role the press plays, the district court had strong support for its conclusion that plaintiffs demonstrated a likelihood of success on the merits of their First Amendment right-of-access claim.

We are mindful that the Federal Defendants could have overcome plaintiffs’ right of access by demonstrating “an overriding interest based on findings that *26 closure is essential to preserve higher values and is narrowly tailored to serve that interest.” Press-Enterprise II, 478 U.S. at 9. There is no question the Federal Defendants have a strong interest in protecting federal property and persons on federal property, and we do not doubt the district court’s findings related to the difficult and dangerous situation posed by protesters who engaged in violent and criminal conduct. But Federal Defendants argue that dispersing the press, regardless of whether they are on federal property, is essential to protecting the government’s interests. They further argue that their dispersal orders cannot be tailored in any way and that the district court erred by granting a special exemption from crowd-control measures to members of the press and legal observers. We disagree.

First, the district court did not grant a special exemption to the press; it found that dispersing the press was not essential to protecting the government’s interests. The district court was faced with a mountain of evidence that the Federal Defendants routinely left federal property to engage in crowd control. The injunction recognizes that the Federal Defendants did not claim the authority to issue general dispersal orders on Portland’s streets and sidewalks, that local law enforcement retains that authority pursuant to the general police power, and that Portland’s law enforcement agreed not to require journalists and legal observers to *27 disperse. The preliminary injunction does nothing to hinder Federal Defendants from arresting individuals engaged in violent or criminal acts.

The Federal Defendants’ argument that the injunction grants a broad special exemption to the plaintiffs hinges on the implied assumption that they have the authority to take action to disperse members of the public who are neither on federal property nor threatening it. At oral argument before our court, the Federal Defendants declined to provide their view of the scope of their authority to take such action, but the district court’s order makes clear that, in the district court, the Federal Defendants did not argue they have “the legal authority to declare a riot and order persons to disperse from the city streets in Portland.” We need not precisely define the limits of the Federal Defendants’ authority in order to resolve their emergency motion, but it cannot be debated that the United States Constitution reserves the general police power to the states, U.S. CONST. amend. X; United States v. Morrison, 529 U.S. 598, 618 (2000), and the district court found that the Federal Defendants “routinely have left federal property and engaged in crowd control and other enforcement on the streets, sidewalks, and parks of the City of Portland.”

The district court did not question that the provision relied upon by the Federal Defendants, 40 U.S.C. § 1315, grants them the authority to protect federal *28 property, including issuing and enforcing dispersal orders against people on or threatening federal property. Paragraph six of the injunction expressly recognizes that the Federal Defendants may issue “lawful crowd-dispersal orders for a variety of lawful reasons.” In footnoting that the authority provided by § 1315 does not allow the Federal Defendants to declare an unlawful assembly on the city’s streets or to disperse people from city streets, the court carefully distinguished the Federal Defendants’ ability to disperse people from federal property and described their authority outside the property as limited to performing authorized duties “to the extent necessary to protect the property and persons on the property.” 40 U.S.C. § 1315(b)(1) (emphasis added). But the Federal Defendants’ suggestion that § 1315 confers authority to take action to disperse members of the public who *29 are neither on nor threatening federal property is dubious.10 See United States v. Baldwin, 745 F.3d 1027, 1029 (10th Cir. 2014) (Gorsuch, J.) (discussing § 1315 and its implementing regulations as they relate to “[p]ersons in and on [Federal] property” (alterations in original)). On remand, the district court may have occasion to more precisely define the scope of the Federal Defendants’ authority if the Federal Defendants indicate that they intend to issue dispersal orders outside of federal property for lawful purposes.

The district court was not persuaded that the Federal Defendants’ response to the plaintiffs was essential or narrowly tailored to serve the government’s interests. Press-Enterprise II, 478 U.S. at 9. The district court’s conclusions are well supported and the Federal Defendants have not established that they will likely prevail in their efforts to show that the dispersal of press was essential. Nor

10 Pursuant to § 1315, the Secretary of Homeland Security “shall protect the buildings, grounds, and property that are owned, occupied, or secured by the Federal Government . . . and the persons on the property.” 40 U.S.C. § 1315(a). Relevant here, the governing regulations: (1) prohibit disorderly conduct “in or on Federal property,” 41 C.F.R. § 102–74.390; (2) prohibit people “entering in or on Federal property” from improperly disposing of rubbish on property, willfully damaging property, stealing property, creating a hazard on property, throwing articles at a building, or climbing on a building, 41 C.F.R. § 102–74.380; and (3) require people “in and on property” to obey the “lawful direction of federal police officers and other authorized individuals,” 41 C.F.R. § 102–74.385; United States v. Baldwin, 745 F.3d 1027, 1029 (10th Cir. 2014) (Gorsuch, J.) (construing 41 C.F.R. § 102–74.385 as being applicable to people “in and on [Federal] property” (alteration in original)).

did the Federal Defendants show that the need to defend federal property made it impossible to tailor their dispersal orders.

The district court cited plaintiffs’ expert, Kerlikowske, who opined that “[d]efending the federal courthouse in Portland mainly involves establishing a perimeter around the building, and there is no reason to target or disperse journalists from that position.” The district court further relied on Kerlikowske’s opinion that “trained and experienced law enforcement personnel are able to protect public safety without dispersing journalists and legal observers and can differentiate press from protesters, even in the heat of crowd control.” The district court found this expert qualified, credible, and persuasive. Rather than deferring to the court’s findings, the dissent examines the record anew, decides that Kerlikowske did not adequately address crowd control, and questions the district court’s tailoring analysis. But the Federal Defendants conceded that they made no effort to tailor their response, and on the record at this preliminary stage they have not made the strong showing required by Nken that dispersing the press was essential or that their response was narrowly tailored to serve the government’s interest in protecting federal property.

We also agree with the district court that the City’s ability to comply with a similarly worded injunction strongly undercuts the Federal Defendants’ argument. *31 The City has not required journalists and authorized legal observers to disperse when it has issued crowd control orders to the protesters. After the district court entered the first temporary restraining order against the City on July 2, the district court “specifically invited the City to move for amendment or modification if the original TRO was not working or to address any problems at the preliminary injunction phase.” But the City did not seek modification. Instead, on July 16 the City stipulated to entry of a preliminary injunction that was “nearly identical to the original TRO.” The City’s willingness to tailor the dispersal orders it issues pursuant to its general police power is strong evidence that the Federal Defendants’ dispersal of journalists and legal observers is not essential to defend federal property, and that it is possible for the Federal Defendants to tailor their methods more narrowly.

By its terms, the preliminary injunction the district court entered against the Federal Defendants addresses each of the reasons the Federal Defendants advanced to argue that it was impossible to tailor their dispersal orders. As to the contention that journalists or legal observers might interfere with federal law enforcement if not required to disperse, the preliminary injunction expressly prohibits journalists and legal observers from impeding, blocking, or otherwise interfering with the lawful conduct of the Federal Defendants. The preliminary injunction leaves the *32 Federal Defendants free to make arrests if there is probable cause to believe a crime has been committed, even if the perpetrator is dressed as a journalist or legal observer. The preliminary injunction also provides that the Federal Defendants will not be liable for violating the injunction if journalists or legal observers remain in the area after a dispersal order is issued, and are incidentally exposed to crowd-control devices. Finally, though the Federal Defendants argued that large and unique identifying markings on their uniforms could inhibit their ability to carry out their duties, the district court concluded they did not support this claim. Indeed, the district court went to great lengths to make sure the terms of the injunction do not impede the federal defendants’ ability to safely achieve their mission.11

The dissent faults us for deferring to the district court’s findings, but that is precisely what our precedent requires. Walters, 145 F.3d at 1047. It is not our role

11 Plaintiffs’ expert Kerlikowske opined that identifiable markings on law enforcement officers’ uniforms increase accountability, act as a check and deterrent against misconduct, and will not interfere with federal officers’ ability to perform their duties. This term of the injunction was added after the Federal Defendants were unable to identify their own officers in videos submitted in support of plaintiffs’ still-pending motion for sanctions and contempt of the July 23 TRO. The Federal Defendants contend the district court overreached, but requiring the officers’ uniforms to bear unique identifiable markings is a common- sense method to ensure that non-compliance with the court’s order may be addressed.

to second-guess the district court’s factual findings; we review the district court’s findings for clear error, and we do not see any. The dissent is not so constrained. It reviews the facts de novo, reframes all of the protests as riots, and concludes the Federal Defendants must be permitted to issue dispersal orders without limit. Yet the majority of the protests have been peaceful, and the record is replete with instances in which members of the press were targeted when they were not mixed with, or even proximate to, protesters. Even the Federal Defendants recognize that the general police power is reserved to the states, and the response to protesters on the public streets of Portland is being handled by the state and local police. As for the Federal Defendants’ actions on federal property, the injunction expressly recognizes that the Federal Defendants are free to issue dispersal orders for a variety of lawful reasons. Their authority to issue dispersal orders to protect federal property has not been questioned.

But on the record before us, the Federal Defendants have not shown the general dispersal orders they issued were lawful, much less essential or narrowly tailored. Press-Enterprise II, 478 U.S. at 9. We do not condone any form of violence, nor did the district court, but the court found no evidence that any of the named plaintiffs engaged in unlawful conduct. The many peaceful protesters, journalists, and members of the general public cannot be punished for the violent *34 acts of others. “[T]he proper response to potential and actual violence is for the government to ensure an adequate police presence . . . and to arrest those who actually engage in such conduct, rather than to suppress legitimate First Amendment conduct as a prophylactic measure.” Collins v. Jordan, 110 F.3d 1363, 1373 (9th Cir. 1996) (internal citations omitted). Accordingly, we conclude the Federal Defendants have not made a strong showing that they are likely to succeed on the merits of plaintiffs’ First Amendment right-of-access claim, nor that this argument supports their emergency motion for a stay pending appeal.

B

We turn next to the second Nken factor: whether the Federal Defendants have shown a likelihood they will suffer irreparable injury if the district court’s preliminary injunction is not stayed pending appeal. Nken, 556 U.S. at 426. The Federal Defendants contend the district court abused its discretion because the scope of the injunction is unworkable. Specifically, they argue the injunction will force federal officers to make snap judgments to distinguish journalists and legal observers from protesters. They argue federal officers will face irreparable injury absent a stay pending appeal because the preliminary injunction will hinder their ability to safely protect federal property and people on federal property, and will *35 generally place them in the untenable position of having to choose between risking their safety and violating the preliminary injunction.

The district court was not persuaded, and for purposes of their emergency motion for a stay pending appeal, the Federal Defendants have not shown that the court likely erred. First, as we have explained, the preliminary injunction entered against the Federal Defendants is one of two preliminary injunctions the district court entered. In a separate preliminary injunction, the City stipulated that it would not require journalists and legal observers to disperse from Portland’s streets and sidewalks after it issues general dispersal orders. In the lengthy preliminary injunction the court issued to address the Federal Defendants’ conduct, the court took pains to explain that the general police power is reserved to the states, and that the Federal Defendants had not taken the position that they had the authority to issue general dispersal orders on Portland’s streets and sidewalks.

Second, it is clear the district court has worked tirelessly to respond to a tense and sometimes chaotic situation. In order to provide clear direction, the district court required the Federal Defendants to broadly disseminate, to the federal agents responding to the protesters, the three pages of its opinion and order that enumerate the terms of the injunction. The Federal Defendants read one sentence from the three-page excerpt in isolation and argue that the preliminary injunction *36 provides a special, citywide exemption to dispersal orders for journalists and legal observers. In fact, it is apparent the district court was actually providing the Federal Defendants with an unambiguous statement of actions they may and may not take in the field, including the requirement that the Federal Defendants mark their uniforms in some way to allow officers to be identified, thereby incentivizing compliance with the court’s orders. Read as a whole, the preliminary injunction does not provide a special exemption for journalists and legal observers. Rather, the terms of the injunction account for the City’s stipulation that journalists and legal observers will not be required to disperse from Portland’s streets and sidewalks. The injunction also accounts for the district court’s finding that the Federal Defendants, at least at this preliminary stage, have not shown that it is essential to disperse press to protect federal property, nor that their response was narrowly tailored.

Third, the preliminary injunction unambiguously provides that the Federal Defendants will not be held liable for violating the preliminary injunction by incidentally exposing journalists or legal observers to otherwise lawful crowd-control measures. The Federal Defendants’ argument that they may be irreparably harmed if individuals disguise themselves as journalists or legal observers in order to commit crimes or interfere with law enforcement is similarly unpersuasive *37 because the order explicitly allows the Federal Defendants to arrest anyone if they have probable cause to believe a crime is being committed—regardless of whether that person is, or appears to be, a journalist or legal observer. The preliminary injunction expressly prohibits journalists and legal observers from impeding, blocking, or otherwise physically interfering with the lawful activities of the Federal Defendants.

The district court recognized that Federal Defendants have sustained injuries over the course of the summer, but found no evidence that any of the named plaintiffs engaged in any of the unlawful conduct that caused their injuries, and the Federal Defendants point to no evidence that the injuries they sustained were more severe or more frequent during the time they were operating under the substantially similar terms of the July 23 TRO, or that the alleged confusion in distinguishing between protestors and plaintiffs resulted in any injury.

The district court was heavily influenced by the City’s agreement to enter into a stipulated preliminary injunction that largely mirrors the preliminary injunction entered against the Federal Defendants, and observed “[t]he City did not contend that the terms of the stipulated preliminary injunction were intrusive, unworkable, or vague.” In fact, the City supported entry of the instant preliminary injunction against the Federal Defendants, arguing “[t]he actions of [F]ederal *38 [D]efendants are escalating violence, inflaming tensions in [Portland], and harming Portlanders who seek to engage in nonviolent protests in support of racial justice.”

Plaintiffs’ expert, Gil Kerlikowske, also seriously undermined the Federal Defendants’ argument that they faced irreparable injury. Relying on Kerlikowske’s expert opinion, the district court concluded that the Federal Defendants’ concerns regarding the workability of the injunction were exaggerated. The district court noted Kerlikowske’s statement that “during his tenure in Seattle, law enforcement did not target or disperse journalists and there were no adverse consequences.” Kerlikowske opined that the prohibitions contained in the July 23 temporary restraining order, which the district court incorporated into the preliminary injunction, were both safe and workable for law enforcement. Kerlikowske stated that dispersing press and legal observers is not necessary to protect public safety, and further explained that trained and *39 experienced law enforcement personnel can differentiate press from protesters in the heat of crowd control.12

On the present record, despite the Federal Defendants’ assertion that all of their officers and agents are adequately trained, the district court found numerous instances in which Federal Defendants shot munitions directly at journalists’ and legal observers’ chests, arms, backs, and heads while they were standing entirely apart from the protesters. These methods directly conflict with Kerlikowske’s opinion that crowd-control munitions are not appropriately aimed at the upper body, and that pepper balls and tear gas canisters should not be aimed at people at all. We review the court’s findings for clear error, and for purposes of the Federal Defendants’ emergency motion, the Federal Defendants have not shown that they will likely establish the district court’s findings are clearly erroneous.

We also conclude the Federal Defendants’ have not made the required showing that they will suffer irreparable harm if the preliminary injunction is not

12 Plaintiffs’ briefing repeatedly asserts that the Federal Defendants lack crowd control training, and the Federal Defendants repeatedly respond that they are trained in the appropriate use of force. At this preliminary stage, the record did not allow the district court to determine whether the Federal Defendants differentiate between crowd control training and training in the proper use of force. Nor does the record make clear whether the training provided to U.S. Marshals differs from the training provided to personnel from the Department of Homeland Security. Those questions may be resolved at a later stage in the proceedings.

stayed pending a decision on the merits of their appeal. The district court took care to address the Federal Defendants’ concerns regarding the workability of the injunction. The terms of the injunction itself adequately address their concerns, and the Federal Defendants’ continued objection that the injunction is unworkable is undermined by the City’s agreement to operate pursuant to a substantially similar order. Kerlikowske’s opinions, which the court found persuasive and credible, further support the district court’s finding that the terms of the preliminary injunction are safe and workable.

The dissent decides that the Federal Defendants are likely to suffer irreparable harm absent a stay pending appeal because the preliminary injunction does not explain how arresting individual suspects is as feasible or safe as using general crowd control tactics during a riot. But the district court found that the protests have been largely peaceful, and the preliminary injunction does not prevent the Federal Defendants from issuing lawful dispersal orders to protect federal property if and when it is threatened by violent protests. We conclude the Federal Defendants have not shown that they will suffer irreparable injury if the district court’s preliminary injunction is not stayed.

C

The Federal Defendants have not satisfied the first two Nken factors, but we briefly note that the final two factors also strongly suggest the Federal Defendants’ motion must be denied. See Nken, 556 U.S. at 435; Al Otro Lado, 952 F.3d at 1006.

The third Nken factor asks whether the other parties to the litigation will be substantially injured if the district court’s preliminary injunction is stayed pending appeal. Nken, 556 U.S. at 426.

The City supported the imposition of the preliminary injunction against the Federal Defendants. As explained, the City asserted that the Federal Defendants’ presence in Portland escalated violence and inflamed tensions. Although the Federal Defendants have entered into some type of agreement with Governor Brown, the district court voiced “serious concerns that the Federal Defendants have not fully complied with the Court’s original TRO.” The district court also highlighted evidence in the record suggesting intentional targeting of journalists or legal observers after the imposition of the TRO. Further, the district court found that the day after the Federal Defendants reached the agreement with the Governor, federal agents fired tear gas at journalists standing nowhere near protesters. In light of this evidence, and the Federal Defendants’ stated intention to remain in *42 Portland to continue to protect the federal buildings should they deem local authorities’ efforts unsatisfactory, the likelihood that the City will suffer substantial injury supports denial of the emergency motion for a stay pending appeal.

Plaintiffs also face substantial injury if the Federal Defendants’ motion is granted because the district court found that the Federal Defendants’ conduct chilled the exercise of their First Amendment rights. The district court made this finding after reviewing plaintiffs’ vivid descriptions and photographic evidence of injuries they sustained as bystanders. “It is well established that the deprivation of constitutional rights ‘unquestionably constitutes irreparable injury.” Melendres v. Arpaio, 695 F.3d 990, 1002 (9th Cir. 2012) (quoting Elrod v. Burns, 427 U.S. 347, 373 (1976)); see also, e.g., Assoc. Press v. Otter, 682 F.3d 821, 826 (9th Cir. 2012) (“The loss of First Amendment freedoms, even for minimal periods of time, unquestionably constitutes irreparable injury.”). In sum, the Federal Defendants have failed to show that the other parties to the litigation will not be substantially injured if the district court’s preliminary injunction is stayed pending appeal.

The fourth Nken factor requires courts to determine where the public interest lies. Nken, 556 U.S. at 426. When the government is a party, the irreparable injury and public interest factors merge, id. at 435, but the Federal Defendants are *43 incorrect to suggest that a showing of harm to the government commands the conclusion that the public interest weighs entirely in favor of whichever outcome the government seeks. Our court has consistently balanced the public interest on the side of the plaintiffs against the public interest on the side of the government to determine where the public interest lies. See, e.g., Padilla v. Immigration & Customs Enforcement, 953 F.3d 1134, 1147–48 (9th Cir. 2020) (determining the “balance of the equities and public interest favors plaintiffs”).

Here, the public interest on the Federal Defendants’ side is the uncontested interest in protecting federal agents and property. The harms the Federal Defendants assert relate to the potential challenges the preliminary injunction poses to their ability to safely and effectively protect federal property and personnel. On the other hand, plaintiffs also assert a strong public interest: “It is always in the public interest to prevent the violation of a party’s constitutional rights.” Padilla, 953 F.3d at 1147–48 (internal quotation marks omitted). When weighing public interests, courts have “consistently recognized the significant public interest in upholding First Amendment principles.” Assoc. Press, 682 F.3d at 826 (quoting Sammartano v. First Judicial Dist. Court, 303 F.3d 959, 974 (9th Cir. 2002), abrogated on other grounds by Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 22 (2008)). The Federal Defendants assert a very important public interest, but the *44 record fully supports the district court’s conclusion that the Federal Defendants’ interest does not require dispersing plaintiffs. They have not threatened federal property, and the journalists, in particular, provide a vitally important service to the public. Accordingly, the final Nken factor does not weigh in favor of a stay.

The Federal Defendants have not made a strong showing that they are likely to succeed on the merits of plaintiffs’ claims. Nor have they shown that they are likely to suffer irreparable injury as a result of the district court’s preliminary injunction. Further, a stay of the district court’s injunction would substantially injure both the City and the plaintiffs. For these reasons, we cannot say at this juncture that the Federal Defendants are entitled to a stay of the preliminary injunction pending appeal. The Federal Defendants’ emergency motion for a stay pending appeal is DENIED, and the administrative stay entered August 27, 2020 is lifted.

Index Newspapers v. U.S. Marshals Serv., No. 20-35739 FILED OCT 9 2020 O’SCANNLAIN, J., dissenting: MOLLY C. DWYER, CLERK U.S. COURT OF APPEALS In the words of the majority—and I agree—“the district court has worked

tirelessly to respond to a tense and sometimes chaotic situation” 1 arising from peaceful urban protest events that have degenerated into riots and destructive mob violence, resulting, inevitably, in crowd dispersal actions by law enforcement. Unfortunately, because the constitutional interests of the parties are misaligned in the provisions of the injunction before us, I must, respectfully, dissent from the order. Since the government is likely to prevail on the merits and the other requisite factors are met, I would grant the motion for stay pending appeal.

With its decision today, the majority of this motions panel validates the transformation of the First Amendment-based “right of public access” to governmental proceedings into a special privilege for self-proclaimed journalists and “legal observers” to disregard crowd dispersal orders issued by federal law enforcement officers. The district court’s injunction erroneously curtails an important law enforcement tool for responding to protest events that threaten federal property and personnel, thereby limiting options available for federal officers precisely when they are most needed. While well-meaning, the district court’s decision constitutes a significant and unwarranted departure from the

1 Majority Opinion at 35.

traditional, qualified “right of public access” to criminal judicial proceedings that has been carefully delineated by the Supreme Court. In short, the majority’s decision approves the mutation of a very limited historical right reinforced by a millennium of legal tradition into a broad, amorphous entitlement that finds support nowhere in our precedents or in the historical sources of the First Amendment.

Similarly, the majority’s decision to uphold the injunction before us ostensibly rests on the deference that it accords to the district court’s factual findings with respect to plaintiffs’ “retaliation” claim, which, indeed, reveal quite a disturbing pattern of apparent misconduct by certain federal officers. But even these unfortunate facts cannot justify granting journalists and “legal observers” a unique exemption from lawful dispersal orders—orders that were neither found, nor alleged, to be retaliatory.

I

Because the facts set forth in the majority opinion do not adequately reveal the full picture, I respectfully restate them as found in the record.

A

In the early morning of July 3, 2020, the recent and ongoing political protests in downtown Portland, Oregon took a violent and destructive turn. Rioters smashed the glass entryway doors of the Mark O. Hatfield Federal Courthouse and *3 attempted to set fire to the building. They threw balloons containing an accelerant into the lobby and fired powerful commercial fireworks toward the accelerant, which ignited a fire in the lobby. Vandalism, destruction of property, and assault on federal law enforcement officers securing the building continued throughout the Fourth of July holiday weekend, and federal agents made multiple arrests.

Before July 3rd, federal law enforcement officers at the Hatfield Courthouse had been stationed in a defensive posture, intended to de-escalate tensions with protesters by remaining inside and responding only to breach attempts on the building and assaults on personnel or to other serious crimes. With limited support from the Portland Police Bureau (“PPB”), however, federal agents struggled to contain protests that often focused on the Courthouse and frequently devolved into violence in the late evenings and early mornings.

When this pattern of violent unrest culminated in the July 3rd attack, the Department of Homeland Security (“DHS”) changed its tactics and authorized federal agents to take additional action to protect the Courthouse, and to identify and to arrest serious offenders. After federal officers adopted this more assertive posture, the protests became larger and more intense. These protest events were chaotic and dynamic, and federal officers had frequent confrontations with rioters. According to DHS’s Gabriel Russell, the law enforcement officer leading the federal response in Portland, 120 federal officers experienced injuries, including *4 broken bones, hearing damage, eye damage, a dislocated shoulder, sprains, strains, and contusions. Conflict between federal officers and rioters continued until the early morning of July 30th, after which incidents diminished as a result of DHS reaching an agreement with the Governor of Oregon for the Oregon State Police to provide security in the areas adjacent to the Hatfield Courthouse.

During the period of unrest, journalists and “legal observers” ostensibly reporting on law enforcement’s response to the riots were frequently interspersed with protesters when events degenerated into violence. Some of these individuals even participated in violent and unlawful conduct, including assaults on federal officers and destruction of federal property. For example, a person with a helmet marked “press” used a grinder to attempt to breach the fence surrounding the Hatfield Courthouse. Another person with a “press” helmet entered Courthouse property and encouraged others to join, yelling to the crowd that “they can’t arrest us all!” A man wearing a vest labeled “press” was seen throwing a hard object toward police. In yet another incident, a Courthouse staff member reported being kicked by someone wearing clothing marked “press.”

B

Plaintiffs are a newspaper organization and individual journalists and “legal observers,” some of whom are affiliated with the National Lawyers Guild (“NLG”) and the American Civil Liberties Union (“ACLU”). They allege that federal law *5 enforcement officers with DHS and the U.S. Marshals Service (“USMS”) operating in Portland during the month of July (1) infringed their First Amendment “right of access” to public streets and sidewalks to observe and to document law enforcement’s response to the riots near the Hatfield Courthouse; and, (2) deliberately and unlawfully “retaliated” against them for exercising their putative First Amendment right to report on those events by targeting them with tear gas, less-lethal munitions, and pepper spray.

Plaintiffs initially filed suit against the City of Portland, and unnamed individual PPB officers, in federal district court, alleging similar constitutional violations arising out of the PPB’s response to the protest events. For example, Plaintiffs alleged a “broader pattern of the Portland police repeatedly and intentionally shooting, gassing, and beating journalists and [legal] observers.” Among other incidents, Plaintiffs alleged that the PPB slammed a reporter from The Oregonian in the back with a truncheon, even as she was displaying her press pass, and shoved a reporter from the Portland Tribune into a wall, after he had identified himself as media, when he initially refused to comply with an order to disperse. Plaintiffs further alleged that the PPB had publicly announced that it would use force to disperse reporters unless they had been previously selected to embed with officers. Plaintiffs obtained a temporary restraining order (“TRO”) against the PPB, without the City of Portland’s consent, on July 2nd, with terms *6 similar to those contained in the instant preliminary injunction. In its order granting the TRO, the district court concluded that Plaintiffs had demonstrated “serious questions going to the merits” with respect to their claim of a First Amendment-based “right of public access” to observe law enforcement’s response to protest events. The TRO specified that press and “legal observers” were exempt from any orders to disperse issued by the PPB.

After alleged retaliation by a federal law enforcement agent on July 12th, plaintiffs filed an emergency motion seeking the district court’s leave to file an amended complaint describing such incident and also adding DHS and USMS as defendants in the case. The City of Portland filed an objection, arguing, inter alia, that plaintiffs’ claims against the City of Portland and those against DHS and USMS raised no common questions of law or fact. The City maintained that PPB operates under fundamentally different conditions than federal law enforcement agencies, including different directives governing the use of force, different limitations on the use of force, and a separate command structure.

On July 16th, before the district court had an opportunity to rule on the motion to bring DHS and USMS into the case, plaintiffs and the City jointly filed a “Stipulated Preliminary Injunction” that substantially mirrored the TRO’s terms. The following day, the district court granted plaintiffs’ motion for leave to file the operative Second Amended Complaint.

The Second Amended Complaint sets forth independent causes of action based on the First and Fourth Amendments of the U.S. Constitution and Article I, Sections 8 and 26 of the Oregon Constitution. It seeks both damages and equitable relief. The day it was filed, Plaintiffs immediately moved for a TRO against DHS and USMS, with the request for injunctive relief limited only to their aforementioned First Amendment claims.

On July 22nd, the City filed a brief in support of the entry of the TRO against DHS and USMS. The City accused both agencies of escalating violence, harming non-violent protesters, and effectively kidnapping people off of Portland streets. Notably, on the same day, the Portland City Council passed a resolution prohibiting the PPB from cooperating with federal officers deployed in Portland.

The district court granted the TRO on July 23rd and extended it for an additional 14 days on August 6th. On August 20th, the district court entered the instant preliminary injunction, from which DHS and USMS now seek emergency relief pending appeal.

The preliminary injunction provides, among other things, that journalists and “legal observers” are exempt 2 from general dispersal orders issued by federal

2 The precise language of the district court’s order provided that journalists and “legal observers” “shall not be required to disperse following the issuance of an order to disperse, and such persons shall not be subject to arrest for not dispersing following the issuance of an order to disperse.”

officers. It further requires that federal officers refrain from using force or threatening arrest to compel such persons to disperse after an order to disperse has been issued. It also sets forth a non-exclusive list of indicia by which officers are to determine who qualifies as a journalist or “legal observer.” 3

3 The eight-part injunction entered by the district court is lengthy, not to say labyrinthine, but warrants repetition in full for appreciation of its extraordinary scope:

1. The Federal Defendants, their agents and employees, and all persons acting under their direction are enjoined from arresting, threatening to arrest, or using physical force directed against any person whom they know or reasonably should know is a Journalist or Legal Observer (as explained below), unless the Federal Defendants have probable cause to believe that such individual has committed a crime. For purposes of this Order, such persons shall not be required to disperse following the issuance of an order to disperse, and such persons shall not be subject to arrest for not dispersing following the issuance of an order to disperse. Such persons shall, however, remain bound by all other laws. No Journalist or Legal Observer protected order this Order, however, may impede, block, or otherwise physically interfere with the lawful activities of the Federal Defendants.

2. The Federal Defendants, their agents and employees, and all persons acting under their direction are further enjoined from seizing any photographic equipment, audio- or video- recording equipment, or press passes from any person whom they know or reasonably should know is a Journalist or Legal Observer (as explained below), or ordering such person to stop photographing, recording, or observing a protest, unless the Federal Defendants are also lawfully seizing that person consistent with this Order. Except as expressly provided in Paragraph 3 below, the Federal Defendants must return any seized equipment or press passes immediately upon release of a person from custody.

3. If any Federal Defendant, their agent or employee, or any person acting under their direction seize property from a Journalist or Legal Observer who is lawfully arrested consistent with this Order, such Federal Defendant shall,

as soon thereafter as is reasonably possible, make a written list of things seized and shall provide a copy of that list to the Journalist or Legal Observer. If equipment seized in connection with an arrest of a Journalist or Legal Observer lawfully seized under this Order is needed for evidentiary purposes, the Federal Defendants shall promptly seek a search warrant, subpoena, or other court order for that purpose. If such a search warrant, subpoena, or other court order is denied, or equipment seized in connection with an arrest is not needed for evidentiary purposes, the Federal Defendants shall immediately return it to its rightful possessor.

4. To facilitate the Federal Defendants’ identification of Journalists protected under this Order, the following shall be considered indicia of being a Journalist: visual identification as a member of the press, such as by carrying a professional or authorized press pass, carrying professional gear such as professional photographic equipment, or wearing a professional or authorized press badge or other official press credentials, or distinctive clothing, that identifies the wearer as a member of the press. It also shall be an indicium of being a Journalist under this Order that the person is standing off to the side of a protest, not engaging in protest activities, and not intermixed with persons engaged in protest activities, although these are not requirements. These indicia are not exclusive, and a person need not exhibit every indicium to be considered a Journalist under this Order. The Federal Defendants shall not be liable for unintentional violations of this Order in the case of an individual who does not carry or wear a press pass, badge, or other official press credential, professional gear, or distinctive clothing that identifies the person as a member of the press.

5. To facilitate the Federal Defendants’ identification of Legal Observers protected under this Order, the following shall be considered indicia of being a Legal Observer: wearing a green National Lawyers Guild-issued or authorized Legal Observer hat (typically a green NLG hat) or wearing a blue ACLU-issued or authorized Legal Observer vest. It also shall be an indicium of being a Legal Observer protected under this Order that the person is standing off to the side of a protest, not engaging in protest activities, and not intermixed with persons engaged in protest activities, although these are not requirements.

6. The Federal Defendants are not precluded by the Order from issuing otherwise lawful crowd-dispersal orders for a variety of lawful reasons. The

A prior motions panel of this court entered an administrative stay of the injunction pending the adjudication of the government’s motion for emergency

Federal Defendants shall not be liable for violating this injunction if a Journalist or Legal Observer is incidentally exposed to crowd-control devices after remaining in the area where such devices were deployed after the issuance of an otherwise lawful dispersal order.

7. Plaintiffs and the Federal Defendants shall promptly confer regarding how the Federal Defendants can place unique identifying markings (using numbers and/or letters) on the uniforms and/or helmets of the officers and agents of the Federal Defendants who are specially deployed to Portland so that they can be identified at a reasonable distance and without unreasonably interfering with the needs of these personnel. Based on the Court’s understanding that Deputy U.S. Marshals and Courtroom Security Officers stationed in Portland who are under the direction of the U.S. Marshal for the District of Oregon are not part of the force that has given rise to events at issue in the lawsuit, they are exempt from this requirement. Agents wearing plain clothes and assigned to undercover duties also are exempt from this requirement. If the parties agree on a method of marking, they shall submit the terms of their agreement in writing to the Court, and the Court will then issue a modified preliminary injunction that incorporates the parties’ agreement. If the parties cannot reach agreement within 14 days, each party may submit its own proposal, and each side may respond to any other party’s proposal within seven days thereafter. The Court will resolve any disputes on this issue and modify this preliminary injunction appropriately.

8. To promote compliance with this Preliminary Injunction, the Federal Defendants are ordered to provide copies of the verbatim text of the first seven provisions of this Preliminary Injunction, in either electronic or paper form, within 14 calendar days to: (a) all employees, officers, and agents of the Federal Defendants currently deployed in Portland, Oregon (or who later become deployed in Portland, Oregon while this Preliminary Injunction is in force), including but not limited to all personnel in Portland, Oregon who are part of Operation Diligent Valor, Operation Legend, or any equivalent; and (b) all employees, officers, and agents of the Federal Defendants with any supervisory or command authority over any person in group (a) above.

relief. As the court, in its role as this motions panel, today denies such emergency request for a stay pending appeal, the injunction will go back into effect and this matter will proceed before the district court, pending disposition of the government’s appeal of the preliminary injunction by a merits panel of this court. Plaintiffs’ Fourth Amendment and state constitutional claims did not form part of the request for preliminary relief and remain pending before the district court, as do plaintiffs’ requests for compensatory and punitive damages, attorney’s fees, and costs. As the City’s stipulation to a preliminary injunction resolved only Plaintiffs’ request for equitable relief, Plaintiffs’ remaining claims against the City and individual PPB officers also remain pending in the district court.

II

I agree with the majority that the Nken v. Holder factors must determine our disposition of the government’s request for emergency relief, but I respectfully disagree with how the majority analyzes those factors. 556 U.S. 418, 426 (2009). I address each factor in turn, beginning with the government’s burden to make a strong showing of likelihood of success on the merits.4

4 Upon appeal of a preliminary injunction, the district court’s conclusions of law are reviewed de novo, its underlying factual findings are reviewed for clear error, and the scope of the injunction is reviewed for abuse of discretion. Padilla v. Immigration & Customs Enf’t, 953 F.3d 1134, 1141 (9th Cir. 2020). In addition, “we review First Amendment questions de novo since they present mixed questions of law and fact, requiring us to apply principles of First Amendment

The district court granted injunctive relief on the basis of Plaintiffs’ two First Amendment claims: (1) a “right of public access” to public streets and sidewalks to observe and to document law enforcement officers engaged in riot control and crowd dispersal; and (2) a right to be free from “retaliation” by federal officers for reporting on law enforcement’s response to civil unrest.

A

With respect to the “right of public access” issue, the district court purported to apply the framework articulated in Press-Enterprise Co. v. Superior Court of Cal. (“Press-Enterprise II”) for evaluating “claim[s] of a First Amendment right of access to criminal proceedings[.]” 478 U.S. 1, 8–9 (1986). Pursuant to that doctrine, in evaluating a purported claim of public access to a proceeding, a court must consider: (1) “whether the place and process have historically been open to the press and general public;” and (2) “whether public access plays a significant positive role in the functioning of the particular process in question.” Id. “If the particular proceeding in question passes these tests of experience and logic, a qualified First Amendment right of public access attaches.” Id. “A presumptive right of access to any particular proceeding may be overcome by an overriding

jurisprudence to the specific facts of this case.” Gerritsen v. City of Los Angeles, 994 F.2d 570, 575 (9th Cir. 1993).

government interest based on findings that closure is essential to preserve higher values and is narrowly tailored to serve that interest.” Id.

But the First Amendment-based right of public access and its corresponding framework have never been deemed to apply to riot control and crowd dispersal in a public street. 5 The Supreme Court has discussed only a qualified right of access to certain criminal judicial proceedings and has never recognized a right of public access outside of that context. See Press-Enterprise II, 478 U.S. at 8–9 (right of public access to preliminary hearings in criminal cases); Press-Enterprise Co. v. Superior Court of Cal., 464 U.S. 501, 503, 508 (1984) (right of public access to voir dire hearings in criminal cases); Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 580 (1980) (right of public access to criminal trials).

5 The majority, echoing arguments offered by plaintiffs’ counsel, invokes prior decisions of our court referencing a First Amendment-based right to record law enforcement activity in public. See Reed v. Lieurance, 863 F.3d 1196, 1211 (9th Cir. 2017); Fordyce v. City of Seattle, 55 F.3d 436, 439 (9th Cir. 1995). Those cases are inapposite, however, as they do not address situations where law enforcement is responding to rioting and violent unrest. At most, those cases merely recognize the right of a person to use a recording device in a public forum, before any measures have been taken to restrict access to the forum, such as issuance of a general dispersal order. They certainly do not stand for the extraordinary proposition that an individual is exempt from a dispersal order or other riot control measure merely because he is engaged in the act of recording law enforcement operations. Moreover, as a matter of doctrine, neither case applied right-of-public-access analysis. In fact, Reed applied public forum analysis, which the district court notably chose not to do here. 863 F.3d at 1211. Cf. fn. 9, infra.

In the decades since Press-Enterprise II, the courts of appeals have expanded the right-of-public-access doctrine considerably beyond its initial, paradigmatic application to criminal proceedings—including, in our court, to a variety of non-criminal, non-adjudicative, governmental proceedings, such as a horse gather on federal land, Leigh v. Salazar, 677 F.3d 892, 894 (9th Cir. 2012), and a referendum on a regulatory order conducted by the U.S. Department of Agriculture, Cal-Almond, Inc. v. U.S. Dept. of Agriculture, 960 F.2d 105, 109 (9th Cir. 1992)—but the doctrine is not without limit. Rather, the Press-Enterprise II framework has been confined to claims of access to specific governmental proceedings and has never been applied to public spaces in general or to private events therein. Cf. Leigh, 677 F.3d at 894 (evaluating access to horse gather, not to federal lands); Whiteland Woods, L.P. v. Township of W. Whiteland, 193 F.3d 177, 181 (3d Cir. 1999) (evaluating access to town planning commission meeting, not to town hall). Here, protests in a public street are privately sponsored and organized events, and when they degenerate into riots, the crowd control measures taken by law enforcement are spontaneous and temporary responses to ongoing criminal activity. Protests and resulting riots are simply not governmental proceedings to which a right of public access may be claimed.6

6 Curiously, the complaint might be better viewed as claiming a “right of exclusion” from crowd dispersal actions by federal law enforcement. Plaintiffs seek access to a putative proceeding, the necessary impact of which they actually

Similarly, even where the Press-Enterprise II framework applies, it requires a court to evaluate a claim of access by first determining whether “the place and the process” have historically been open to the public, and whether the public’s presence plays a critical role in the specific proceeding at issue. 478 U.S. at 8–9 (emphasis added). Here, the district court noted that streets, sidewalks, and parks constitute traditional public fora, which have been open to speech and expression from “time out of mind,” Hague v. Comm. for Indus. Org., 307 U.S. 496, 515 (1939), but it failed to evaluate any history of public access to law enforcement operations responding to ongoing criminal activity, including violent civil unrest that threatens federal property and personnel. In the absence of historical analysis regarding the proceeding, as distinguished from the place, a presumptive right of public access simply does not attach. Cf. Leigh, 677 F.3d at 894 (calling for inquiry into history of public access to horse gathers, not to federal lands).

The district court’s reasoning here is reflective of an emerging pattern of lower courts expanding the right-of-public-access doctrine well beyond its original scope, with little consideration of a limiting principle. Cf., e.g., N.Y. Civil Liberties Union v. N.Y. City Transit Auth., 684 F.3d 286, 298 (2d Cir. 2012) (noting, with approval, that “there is no principle that limits the First Amendment right of

wish to avoid. This contradiction highlights the discrepancy between plaintiffs’ claims and traditional right-of-public-access case law.

[public] access to any one particular type of government process”). When the Supreme Court first articulated the First Amendment right of public access in Richmond Newspapers, Inc. v. Virginia, it drew on an extensive historical record of public access to criminal trials in the Anglo-American legal tradition, dating back to “the days before the Norman Conquest.” 448 U.S. at 580. After canvassing more than a thousand years of “unbroken, uncontradicted” history, the Court felt justified in concluding that the right to attend criminal trials is “implicit in the guarantees of the First Amendment.” Id. In Press-Enterprise II, the Court limited its inquiry to post-Bill of Rights history, but nonetheless identified a “near uniform” “tradition of accessibility” to preliminary hearings in criminal cases dating back to the “celebrated trial of Aaron Burr” in 1807. 478 U.S. at 10–11.

Lower courts, by contrast, including ours, have extended the right of public access largely without extensive historical backing and without further guidance from the Supreme Court regarding the specific contours of the doctrine. If the majority’s reasoning here is any indication, the doctrine is growing haphazardly, like a weed in an untended garden, presaging conflict with more established legal rights and powers. This doctrinal disorder warrants further review.

Even if right-of-public-access analysis were appropriate under these circumstances, any right to access the proceeding in question must apply equally to *17 the press and the public. See Cal. First Amendment Coal. v. Woodford, 299 F.3d 868, 873 n.2 (9th Cir. 2002) (“As members of the press, plaintiffs’ First Amendment right of access to governmental proceedings is coextensive with the general public’s right of access.” (citing Houchins v. KQED, Inc., 438 U.S. 1, 15– 16 (1978)). Indeed, it is a long-established and fundamental principle of constitutional law that “the First Amendment does not guarantee the press a constitutional right of special access to information not available to the public generally.” Branzburg v. Hayes, 408 U.S. 665, 684 (1972). Cf. Erwin Chemerinsky, Protect the Press: A First Amendment Standard for Safeguarding Aggressive Newsgathering, 33 U. Rich. L. Rev. 1143, 1145 (2000) (“[The Supreme Court’s] rulings, without exception, have failed to provide any First Amendment protection for newsgathering. Indeed, the Court has declared that there is no exemption for the press from general laws. In other words, while engaged in newsgathering, the press is not exempt from tort liability or criminal laws, no matter how compelling the need for reporting to protect the public’s health and safety.”)

But here, the district court’s injunction, by its own terms, grants selfidentified journalists and “legal observers” a special privilege to disregard dispersal orders with which the general public must comply, which has no legal *18 basis. The injunction is thus at odds with a core First Amendment principle and a common-sense rule of thumb: the media have the same rights as the rest of us.7

The majority opinion here rejects this characterization of the injunction and insists that it creates no special rights. According to the majority, the injunction merely prevents federal agents from seeking to disperse the press from local streets and sidewalks when the City’s current policy is that press may remain there, even during riots, but does not seek to regulate crowd dispersal on federal property. On this view, the injunction is a wholesome exercise in federalism!

But the majority’s analysis is inconsistent with the plain text of the district court’s order and misapplies principles of constitutional structure. The injunction, by its own terms, appears to extend to dispersal orders issued on federal property,

7 Even if journalists had some special claim to enhanced Constitutional protection when reporting on law enforcement activities, grounded in the First Amendment’s “freedom of the press” clause, “legal observers” have never been accorded any special recognition under our law. Cf. Wise v. City of Portland, No. 3:20-CV- 01193-IM, 2020 WL 5231486, at *7 (D. Or. Sept. 2, 2020) (declining to recognize special status for “protest medics” in similar Portland protests) (“[T]his Court has found no legal authority for affording protest medics, as defined by Plaintiffs, unique recognition under the First Amendment beyond that afforded any individual who attends a protest. . . . They simply have no unique status under the First Amendment that allows them to disregard lawful [dispersal] orders.”). That the district court’s injunction appears to empower the ACLU and NLG to bestow immunity from lawful dispersal orders is particularly dubious given the status of these organizations as perennial litigation adversaries of law enforcement agencies. In sum, like “protest medics,” there is no cognizable basis for “legal observers” to receive “special dispensation” to disregard lawful dispersal orders. Wise, 2020 WL 5231486 at *2.

and is certainly not geographically limited in any explicit way. The injunction thus allows the press, but not others, to disregard dispersal orders that are clearly lawful. That can only be understood as a special dispensation that is not consistent with the First Amendment.

In any event, even if federal agents are located on City property when they issue, or seek to enforce, an order to disperse, principles of federalism do not justify carving out a special exemption for the press from such orders simply because City police would typically allow for one. The Federal Government is indeed acknowledged by all to be one of limited and enumerated powers, see Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519, 534 (2012), and it is not entitled to exercise general or residual powers, see United States v. Comstock, 560 U.S. 126, 153 (2010) (Kennedy, J., concurring) (“Residual power, sometimes referred to (perhaps imperfectly) as the police power, belongs to the States and the States alone”), such as the prevention and punishment of crime and disorder on local streets, sidewalks, and parks, see United States v. Morrison, 529 U.S. 598, 618 (2000) ([W]e can think of no better example of the police power, which the Founders denied the National Government and reposed in the States, than the suppression of violent crime and vindication of its victims.”).

It is an inversion of our constitutional structure, however, to require federal officers to abide by municipal policies regarding crowd dispersal when carrying *20 out their statutory prerogative to protect federal property and personnel. Federal officials are prohibited, of course, from “commandeering” state and local law enforcement officers to help secure federal property and must instead rely on voluntary cooperation with state and local officials for this purpose. See Printz v. United States, 521 U.S. 898, 935 (1997). Where such cooperation is inadequate, the federal government must deploy its own agents. In these circumstances, the agency’s lawful directives regarding crowd dispersal, i.e., those adopted pursuant to a constitutionally enacted federal statute or rule, take precedence over state and local ones, not the other way around. Such an arrangement does not violate principles of federalism or dual sovereignty but is rather required by them. See, e.g., Alden v. Maine, 527 U.S. 706, 732 (1999) (federal government sets the supreme law of the land when acting within its enumerated powers).

The majority opinion relies heavily on the district court’s conclusion, with which it agrees, that it is, in fact, unlawful for federal agents to issue orders to disperse if they are situated beyond federal property. According to the majority, DHS and USMS have never claimed to have such authority, and the federal statute upon which they principally rely, 40 U.S.C. § 1315, does not provide for it.

The suggestion that the government has simply conceded this question is overstated. Although this issue was not adequately briefed by either party, the government has consistently articulated the position, both before the district court *21 and on appeal, that federal law enforcement officers may issue dispersal orders on federal property, and in several circumstances, may effectuate those orders beyond federal property, such as by establishing a secure perimeter. In particular, the government has invoked § 1315(b)(1), which provides that the Secretary of Homeland Security may designate DHS agents to protect federal property, including designating agents for duty in “areas outside the property to the extent necessary to protect the property and persons on the property.”

I am inclined to agree with the government’s general understanding of its statutory authority. As the government has pointed out, it would be unreasonable to require that federal officers charged with securing federal buildings wait until violent opportunists have breached the property line or entered the building before taking any protective measures. There is very likely a statutory basis for at least some crowd dispersal activity adjacent to a federal courthouse faced with violent unrest and the other challenging circumstances at issue here.

I also agree with the majority, however, that a determination of the precise scope of DHS’s and USMS’s statutory authority is not required for resolution of this emergency motion. Indeed, the statutory question muddles the First Amendment analysis upon which the district court’s injunction is ultimately grounded. Presumably, if federal officers have no statutory basis for dispersals beyond federal property, then any such dispersals are ultra vires, and the inquiry is *22 at an end. There is no reason to proceed to an evaluation of the constitutional rights of persons subject to such purportedly unlawful measures, let alone to construct a complex injunction that distinguishes the rights of press and “legal observers” from the rights of other participants in a protest. Ultimately, a lack of federal statutory authority for off-property dispersals, as such, cannot serve as the sole, or even primary, basis upon which this particular injunction is upheld, given its reliance on a painstaking analysis of purported constitutional violations with respect to specific persons. Thus, even if I were to accept the majority’s view that the injunction’s aim is simply to prohibit off-property dispersals by federal officers, which I do not, the injunction’s terms would be woefully underinclusive.

Even if a presumptive right of access for press and “legal observers” to witness law enforcement’s response to a riot could be said to exist, the inquiry does not end there. Under Press-Enterprise II, a presumptive right of public access to any particular proceeding may be overcome by an overriding government interest based on findings that closure is essential to preserve higher values and is narrowly tailored to serve that interest. 478 U.S. at 8–9.

The district court’s narrow tailoring analysis failed to take proper account of the government’s interests in defense of federal personnel and property, which justify use of general dispersal orders during riot control situations that threaten *23 federal resources, even in a public forum. 8 Here, considering the chaotic and dynamic situation during Portland’s recent protest events, which have frequently devolved into riots, along with the nefarious actions by certain individuals falsely purporting to be press or “legal observers,” closure of the forum through general dispersal orders is essential to the defense of federal personnel and property. Indeed, the closure of governmental proceedings has been deemed proper in several instances where the government’s interest was arguably less immediate and the restriction on access was equally broad. Cf., e.g., Dhiab v. Trump, 852 F.3d 1087, 1095 (D.C. Cir. 2017) (government’s interest in preventing future threats to military operations would justify closure of habeas proceedings); U.S. v. Index Newspapers LLC, 766 F.3d 1072, 1087 (9th Cir. 2014) (government’s interest in

8 In addition, the district court’s narrow tailoring analysis was conceptually flawed because the closure evaluated in the Press-Enterprise II framework should be that of a specific governmental proceeding, not of a public forum generally. Utilizing “right-of-public-access” analysis to evaluate the closure of a “traditional public forum,” such as a public street, is unsettling because government restrictions on First Amendment activity in such locations are usually evaluated under “public forum analysis,” which has been more extensively developed in the case law and provides more guidance regarding the policing of protest events. See, e.g., Perry Educ. Ass’n v. Perry Local Educators’ Ass’n, 460 U.S. 37, 45 (1983); Int’l Action Ctr. v. City of New York, 587 F.3d 521, 527 (2d Cir. 2009); Coal. to Protest Democratic Nat’l Convention v. City of Boston, 327 F. Supp. 2d 61, 69–70 (D. Mass.), aff’d sub nom. Bl(a)ck Tea Soc’y v. City of Boston, 378 F.3d 8 (1st Cir. 2004).

secrecy justified closure of certain grand jury proceedings); ACLU v. Holder, 673 F.3d 245, 252 (4th Cir. 2011) (government’s interest in integrity of ongoing fraud investigation justified sealing of complaints filed in False Claims Act actions).

Moreover, if the categories of “journalist” and “legal observer” in fact include all members of the public engaged in observation, as distinguished from speech or protest—as the majority seems to suggest—then the government’s interests in full closure of the “proceeding” are even more compelling. Otherwise, in the event of a riot in a public forum that threatens federal property, federal officers could disperse only members of the public that are speaking, assembling, and protesting, but not members of the public that are observing or documenting. Peaceful protesters caught up in the riot would have to obey the dispersal order, but peaceful observers would not. This differential treatment is groundless and, in any event, would render federal dispersal orders a dead letter, even in the face of an undeniable threat to federal property and personnel. Federal law enforcement agents simply would not be allowed to clear the street. Such a prohibition is not only inconsistent with the government’s overriding interest in security in cases of violent unrest that threatens federal property and personnel, it is also contrary to established law in other First Amendment settings, which permits general dispersal orders in similar circumstances. Cf., e.g., Bible Believers v. Wayne Cty., Mich., 805 F.3d 228, 252 (6th Cir. 2015) (“The police may go against the hecklers, *25 cordon off the speakers, or attempt to disperse the entire crowd if that becomes necessary.”). Carr v. D.C., 587 F.3d 401, 409–10 (D.C. Cir. 2009) (“[W]hen police face an unruly crowd they may give a dispersal order and then arrest those who, after reasonable opportunity to comply, fail to do so. We continue to acknowledge that this tactic will be invaluable to police in certain circumstances. A dispersal order might well be necessary in a situation in which a crowd is substantially infected with violence or otherwise threatening public safety.” (internal citations and quotation marks omitted)); Wise, 2020 WL 5231486 at *2 (recognizing propriety of general dispersal orders in response to Portland riots).

The only way the majority arrives at a different conclusion is by according deference to the district court’s factual findings, which placed heavy emphasis on the City of Portland’s consent to abide by an injunction with nearly identical terms and a declaration submitted by former DHS official Gil Kerlikowske stating that law enforcement officers may respond effectively to riots without dispersing journalists and “legal observers.” Evaluating whether a government measure is narrowly tailored is not simply a matter of ordinary fact-finding, however. Narrow tailoring is viewed as a mixed question of fact and law that requires a delicate balancing of legal principles as applied to specific circumstances. See Gilbrook v. City of Westminster, 177 F.3d 839, 861 (9th Cir. 1999); Gerritsen v. City of Los Angeles, 994 F.2d 570, 575 (9th Cir. 1993) (“[W]e review First Amendment *26 questions de novo since they present mixed questions of law and fact, requiring us to apply principles of First Amendment jurisprudence to the specific facts of this case.” (internal quotation marks omitted)); see also Mastrovincenzo v. City of New York, 435 F.3d 78, 100 (2d Cir. 2006) (“Our narrow-tailoring inquiry requires us to apply principles of First Amendment jurisprudence to the specific facts of this case, and therefore we treat this issue as a mixed question of law and fact that we may resolve on appeal.” (internal quotations marks omitted)); Casey v. City of Newport, R.I., 308 F.3d 106, 116 (1st Cir. 2002) (“Inescapably, the application of the narrow tailoring test entails a delicate balancing judgment.” (citations omitted)). Accordingly, I would revisit the district court’s narrow tailoring inquiry, which I believe did not correctly balance the interests at stake.

The City’s stipulation does not have the import that the district court, and the majority, ascribe to it. That the City ultimately agreed to the terms of the injunction does not show that it complied with them, let alone that it did so and managed to protect property and personnel. In any event, the City’s agreeableness should not be overstated here. The PPB is still alleged to have followed until recently a policy of dispersing press and “legal observers,” the TRO was entered *27 without the City’s consent, and, after the City agreed to a preliminary injunction, it suggested that modifications would be required.9

Moreover, as already discussed, holding DHS and USMS to the City’s policies and practices reflects a misunderstanding of the relationship between federal and local law enforcement, each of which operates under a separate command structure and is typically entitled to set different enforcement priorities and to follow different directives regarding lawful crowd control tactics, including general dispersal orders. In this case, the City not only sought to distinguish the PPB from federal law enforcement, it has been explicitly adverse to the presence of federal officers in Portland, leveling serious allegations of unlawful conduct against them, and even going so far as to prohibit the PPB from cooperating with federal agents to provide security for the Hatfield Courthouse. The City’s actions, and its filings in the district court, suggest that it has a divergent assessment of the severity of the threat posed to federal personnel and property by protest events that degenerate into riots, and of the proper manner of dealing with that threat. The

9 The City also resisted a very similar request for injunctive relief brought by so- called “protest medics.” Wise, 2020 WL 5231486 at *2. The City apparently argued, and the district court agreed, that an injunction exempting “protest medics” from dispersal orders would be unworkable for the PPB. Why the City expects the PPB to identify and to exempt “legal observers,” but not “protest medics,” is difficult to understand. Cf. fn. 7, supra.

City is entitled, of course, to utilize different crowd control tactics, but the City’s choices obviously do not bind federal law enforcement agencies.

Similarly, Kerlikowske’s testimony does not adequately address crowd control under the specific circumstances faced by federal officers in Portland. For example, he deals in a conclusory manner with the evidence placed in the record regarding the involvement of putative journalists and “legal observers” in criminal acts, stating that federal officers “were not fooled” by the “press” labels and that trained officers are capable of dealing with such incidents on an individualized basis. But effectuating an arrest may not be feasible or safe in the chaotic and dynamic environment of a riot that threatens federal property and personnel, which is why dispersal orders—and related crowd control tactics, such as deployment of tear gas—are understood to be legitimate law enforcement tools in the first place. Cf. Wise, 2020 WL 5231486 at *2 (recognizing propriety of general dispersal orders in responding to Portland riots); Don't Shoot Portland v. City of Portland, No. 3:20-CV-00917-HZ, 2020 WL 3078329, at *4 (D. Or. June 9, 2020) (allowing use of tear gas in situations where safety of public or police is at risk). Given the conclusory nature of Kerlikowske’s testimony on this point, it is hardly definitive.

Accordingly, the government has made a strong showing that it is likely to succeed in demonstrating that the First Amendment-based right of public access does not support the district court’s injunction.

B

With respect to the “retaliation” claim, the district court also concluded that plaintiffs were likely to succeed, largely based on its detailed factual findings indicating a disturbing pattern of unwarranted force by federal agents. The majority opinion here discusses the “retaliation” claim extensively and ultimately defers to these factual findings.

Even if plaintiffs’ retaliation claim were viable, however, that claim alone cannot justify this injunction. The district court’s factual findings regarding retaliation, while apparently based on a meticulous examination of the record, bear no relation to the injunctive relief actually entered. General dispersal orders were not among the acts alleged to be retaliatory, nor did the district court make any findings to support such a conclusion. An injunction that exempts plaintiffs—not to mention, journalists and “legal observers” more generally—from dispersal orders is thus far broader than necessary to provide relief for the injuries alleged, and documented, as a result of retaliation.10 Indeed, Judge Immergut, of the very same district court, relied on such reasoning in denying a similar request for injunctive relief based on First Amendment “retaliation” just two weeks after the

10 Remarkably, some of the allegations in the complaint regarding “retaliation” may well support Bivens actions and claims of excessive force against individual officers, but that is not what is before us today. Bivens v. Six Unknown Fed. Narcotics Agents, 403 U.S. 388, 395 (1971).

instant preliminary injunction was entered. Wise, 2020 WL 5231486 at *8 (injunction not warranted where instances of alleged targeting appeared to occur when “protest medics” refused to follow dispersal orders).

Accordingly, I would hold that, regardless of whether plaintiffs’ have stated a valid First Amendment “retaliation” claim, an injunction that exempts them from non-retaliatory dispersal orders is overbroad and an abuse of discretion. See Cal. v. Azar, 911 F.3d 558, 584 (9th Cir. 2018) (“The scope of the remedy must be no broader and no narrower than necessary to redress the injury shown by the [plaintiff].”)

I conclude that DHS and USMS have made a strong showing that they are likely to succeed in demonstrating that the district court’s extraordinary injunction was issued without an adequate legal basis. This critical Nken factor favors grant of the government’s emergency motion for stay pending appeal.

III

The remaining Nken factors also favor a stay pending appeal here. First, while a closer question, the government has shown that it is likely to suffer irreparable harm during the pendency of the appeal if the injunction is not stayed, because it is unworkable for federal officers to distinguish journalists and “legal observers” in the midst of a riot that threatens federal property and personnel based *31 on the nebulous criteria established by the district court, particularly in light of the incidents of press and “legal observer” involvement in violent unrest.

The majority rejects the government’s showing on this factor, stating that the injunction is carefully drawn to avoid undue interference with DHS’s and USMS’s defense of federal resources, that the PPB has been operating safely and effectively under nearly identical terms, and that Kerlikowske’s declaration indicates that general dispersal orders are unnecessary for crowd control. The majority’s characterization of the order as carefully drawn is misleading because the order merely restates existing legal rules, such as an officer’s power to make an arrest based on probable cause. And the order does not explain how effectuating arrest of individual suspects is as feasible or safe as utilizing general crowd control tactics during a riot that threatens federal property and personnel. Similarly, the City’s stipulation and Kerlikowske’s declaration do not warrant the treatment they receive, for the reasons discussed above.

Second, the harms to the government are serious because the injunction’s curtailment of general dispersal orders will compromise the security of federal personnel and property, whereas, if there is no right of public access, as I have argued, then any harm to plaintiffs from a stay is minimal because they do not have a right to remain in the street after they have been ordered to disperse, and the injunction does not protect them from retaliation. Third, for similar reasons, the *32 public interest in maintenance of order and public safety also favors stay of an overbroad injunction that unduly interferes with law enforcement operations, while offering little, if any, protection for plaintiffs’ actual constitutional rights. This combination of showings justifies a stay pending appeal. See Leiva-Perez v. Holder, 640 F.3d 962, 970 (9th Cir. 2011) (stay warranted where irreparable harm is probable, there is a strong likelihood of success on the merits, and the public interest does not weigh heavily against a stay).

IV

Because the government has made a strong showing that it is likely to succeed in demonstrating that the injunction lacks an adequate legal basis, and the other Nken factors also weigh in favor of a stay, I respectfully dissent and would grant the emergency motion for stay pending appeal.