UNITED STATES OF AMERICA v. RENE SANCHEZ GOMEZ

Reset A A Font size: Print

United States Court of Appeals, Ninth Circuit.

UNITED STATES OF AMERICA, Plaintiff-Appellee, v. RENE SANCHEZ-GOMEZ, Defendant-Appellant.

UNITED STATES OF AMERICA, Plaintiff-Appellee, v. MOISES PATRICIO-GUZMAN, Defendant-Appellant.

UNITED STATES OF AMERICA, Plaintiff-Appellee, v. JASMIN ISABEL MORALES, AKA Jasmin Morales, Defendant-Appellant.

UNITED STATES OF AMERICA, Plaintiff-Appellee, v. MARK WILLIAM RING, Defendant-Appellant.

No. 13-50561, No. 13-50562, No. 13-50566, No. 13-50571

    Decided: May 31, 2017

Before: Sidney R. Thomas, Chief Judge, and Mary M. Schroeder, Stephen Reinhardt, Alex Kozinski, Diarmuid F. O'Scannlain, Barry G. Silverman, Susan P. Graber, Richard A. Paez, Marsha S. Berzon, Consuelo M. Callahan and Sandra S. Ikuta, Circuit Judges. COUNSEL Reuben Camper Cahn (argued), Shereen J. Charlick, and Ellis M. Johnston III, Federal Defenders of San Diego, Inc., San Diego, California, for Defendants-Appellants. Daniel E. Zipp (argued) and Kyle Hoffman, Assistant United States Attorneys; Bruce R. Castetter, Chief, Appellate Section, Criminal Division; United States Attorney's Office, San Diego, California; for Plaintiff-Appellee.

OPINION

SUMMARY *

Criminal Law

The en banc court denied mandamus relief regarding the United States District Court for the Southern District of California's policy of routinely shackling all pretrial detainees in the courtroom.

The en banc court construed four defendants' appeals challenging the district-wide policy as petitions for writs of mandamus and found that it had jurisdiction to consider them under the court's supervisory authority.

Applying the analysis of Gerstein v. Pugh, 420 U.S. 103 (1975), a class action, the en banc court held that even though the named defendants' cases had ended and the challenged policy was no longer in effect, the supervisory mandamus case was not moot because the capable-of-repetition-yet-evading-review mootness exception applied.

The en banc court clarified that the Fifth Amendment right to be free of unwarranted restraints applies whether the proceeding is pretrial, trial, or sentencing, with a jury or without. Before a presumptively innocent defendant may be shackled, the court must make an individualized decision that a compelling government purpose would be served and that shackles are the least restrictive means for maintaining security and order in the courtroom. Courts cannot delegate this constitutional question to those who provide security, such as the U.S. Marshals Service. Nor can courts institute routine shackling policies reflecting a presumption that shackles are necessary in every case. The en banc court wrote that the right to be free of unwarranted restraints has deep roots in the common law, which did not draw a bright line between trial and arraignment. The en banc court rejected the government's contention that individualized determinations are required only before shackles are used in the jury's presence, and that otherwise the right is sufficiently protected by considering generally applicable security concerns, deferring to the Marshals Service and leaving the rest to individual judges' discretion.

Even though the en banc court held the district court's shackling policy to be unconstitutional, it withheld the issuance of a formal writ of mandamus because the shackling policy was not then in effect.

Concurring, Judge Schroeder wrote that she fully concurred in the majority's opinion. In addition to her disagreement with the dissent's interpretation of common law and Supreme Court authority, Judge Schroeder observed that the dissent lacked sensitivity to the dignity with which court proceedings should be conducted and to the proper role of the judges as opposed to the Marshals Service in determining how a courtroom should be run.

Dissenting, Judge Ikuta, joined by Judges O'Scannlain, Silverman, Graber, and Callahan, wrote that the case was moot and no exception to mootness applied. Judge Ikuta wrote that the majority's theory of a “functional class action” exception was inconsistent with Supreme Court precedent and incompatible with Article III's case-or-controversy requirement. She wrote that the defendants' appeals also did not meet the requirements for granting a writ of supervisory mandamus, an extraordinary remedy justified only by exceptional circumstances amounting to a judicial usurpation of power. On the merits, Judge Ikuta wrote that the majority announced a new rule of constitutional criminal procedure that was contrary to Supreme Court precedent, created a split with the Second and Eleventh Circuits, and put trial courts throughout the Ninth Circuit at risk.

We consider whether a district court's policy of routinely shackling all pretrial detainees in the courtroom is constitutional.

BACKGROUND

In 2013, the judges of the Southern District of California acceded to the U.S. Marshals Service's request for “a district-wide policy of allowing the Marshals Service to produce all in-custody defendants in full restraints for most non-jury proceedings.” “Full restraints” means that a defendant's hands are closely handcuffed together, these handcuffs are connected by chain to another chain running around the defendant's waist, and the defendant's feet are shackled and chained together.

After seeking input from the U.S. Attorney's Office, the Federal Defenders of San Diego and a Criminal Justice Act panel representative, the judges adopted the policy 1 of deferring to the Marshals' shackling decisions, with a few minor exceptions. The judges retained discretion to “direct the Marshals to produce an in-custody defendant without restraints.” And the district judges, but not the magistrates, directed the Marshals to “remove arm and hand restraints during guilty pleas and sentencing hearings before them unless the Marshals [were] aware of information that the particular defendant need[ed] to be fully restrained.” Additionally, “defendants in individual cases may ask the judge to direct that the restraints be removed in whole or in part,” at which point the judge would “weigh all appropriate factors, including all of the concerns” expressed by the Marshals in justifying the routine use of full restraints. Only one district judge, Judge Marilyn Huff, opted out of the policy altogether. For the rest of the Southern District's judges, the Marshals shackled all in-custody defendants at pretrial proceedings.

Starting on the first day of the policy's implementation, the Federal Defenders of San Diego objected to the routine use of shackles and requested that each defendant's shackles be removed. The judges routinely denied the requests, relying on the Marshals Service's general security concerns as well as concerns particular to the Southern District. They pointed to increasing security threats from what they viewed as changing demographics and increasing case loads in their district.2 After ruling on a few individual objections, the judges indicated that they didn't “want to go through it a bunch of times.” “For the record,” one judge helpfully noted, “every defendant that has come out is in th[e] exact same shackling; so [counsel doesn't] have to repeat that every time.”

The shackling was the same regardless of a defendant's individual characteristics. One defendant had a fractured wrist but appeared in court wearing full restraints. The judge denied her motion “for all of the reasons previously stated.” Another defendant was vision-impaired. One of his hands was free of restraint so he could use his cane, but his other hand was shackled and secured to a chain around his waist and his legs were shackled together. His objection was “denied for all the reasons previously stated.” And another defendant was shackled despite being brought into court in a wheelchair due to her “dire and deteriorating” health. The court “noted” her objection to the shackles and “appreciate[d] [counsel] not taking anymore time” with it.

The four defendants here, Rene Sanchez-Gomez, Moises Patricio-Guzman, Jasmin Isabel Morales and Mark Ring, all appeared in shackles and objected to their use. The magistrate judges overruled the objections in each instance. Defendants appealed these denials to the district court and also filed “emergency motions” challenging the constitutionality of the district-wide policy. The district courts denied all relief. All four cases are now consolidated before us.3

ANALYSIS

A. Appellate Jurisdiction

1. In United States v. Howard, we considered shackling claims similar to the ones raised here. 480 F.3d 1005, 1008 (9th Cir. 2007). The Central District of California had adopted a routine shackling policy in consultation with the U.S. Marshals Service. Id. The policy required defendants to be shackled in leg restraints at their initial appearances. Id. The public defenders objected, claiming that the use of leg restraints on individual defendants violated the defendants' liberty interests under the Fifth Amendment. Id. at 1009, 1013. They appealed the district court's denial of the unshackling motions without waiting for the defendants' criminal cases to conclude. Id.

We held that we had jurisdiction to review the district's shackling decisions as immediately appealable collateral orders. Id. at 1011. Such orders “(1) conclusively determine[ ] the disputed question, (2) resolve[ ] an important issue completely separate from the merits of the action, and (3) [are] effectively unreviewable on appeal from a final judgment.” Sell v. United States, 539 U.S. 166, 176 (2003) (internal quotation marks, brackets and citation omitted). The government urges us to reconsider Howard, arguing that shackling decisions don't satisfy the requirements for immediately appealable collateral orders.

Presented for our review in this appeal are individual shackling decisions as well as district-wide challenges to the shackling policy. The main dispute in this case, however, is the district-wide shackling policy. Because we do not review the individual defendants' shackling decisions, we see no reason to revisit Howard's appellate jurisdiction analysis as it applies to those appeals.

The district-wide challenges introduce a wrinkle in this case that Howard didn't address. Defendants challenge the Southern District's policy of routinely shackling in-custody defendants without an individualized determination that they pose a material risk of flight or violence. Defendants seek relief not merely for themselves, but for all in-custody defendants in the district. Thus, defendants are making class-like claims and asking for class-like relief.

Such claims are sometimes brought as civil class actions.4 See, e.g., De Abadia-Peixoto v. U.S. Dep't of Homeland Sec., 277 F.R.D. 572, 574 (N.D. Cal. 2011) (using a civil class action to challenge an Immigration and Customs Enforcement policy of shackling all detainees in San Francisco's immigration court). But we can also construe such claims as petitions for writs of mandamus when we lack appellate jurisdiction and mandamus relief is otherwise appropriate. See Miller v. Gammie, 335 F.3d 889, 895 (9th Cir. 2003) (en banc). We “treat the notice of appeal as a petition for a writ of mandamus and consider the issues under the factors set forth in Bauman.” Id. (citation omitted).

2. “The common-law writ of mandamus against a lower court is codified at 28 U.S.C. § 1651(a): ‘The Supreme Court and all courts established by Act of Congress may issue all writs necessary or appropriate in aid of their respective jurisdictions and agreeable to the usages and principles of law.’ ” Cheney v. U.S. Dist. Court for the Dist. of Columbia, 542 U.S. 367, 380 (2004). “Historically, a writ of mandamus was an order compelling a court or officer to act.” In re United States, 791 F.3d 945, 953 (9th Cir. 2015).

Another use of the writ is to exercise our “supervisory” or “advisory” authority. Supervisory and advisory writs are appropriate in cases “involving questions of law of major importance to the administration of the district courts.” In re Cement Antitrust Litig. (MDL No. 296), 688 F.2d 1297, 1307 (9th Cir. 1982); see also La Buy v. Howes Leather Co., 352 U.S. 249, 259–60 (1957) (“We believe that supervisory control of the District Courts by the Courts of Appeals is necessary to proper judicial administration in the federal system.”). This authority allows courts to provide broader relief than merely ordering that the respondent act or refrain from acting, which promotes the writ's “vital corrective and didactic function.” Will v. United States, 389 U.S. 90, 107 (1967); see also 16 Charles Alan Wright et al., Federal Practice and Procedure §§ 3934, 3934.1 (3d ed. 2016) (describing the history and modern usage of this authority).

The Supreme Court has announced three conditions for issuing the writ: First, to ensure that the writ doesn't replace the regular appeals process, there must be “no other adequate means to attain the relief”; second, the petitioner must have a “clear and indisputable” right to the writ; and, lastly, the court, in its discretion, must be “satisfied that the writ is appropriate under the circumstances.” Cheney, 542 U.S. at 380–81 (internal quotation marks and citations omitted). These conditions are consistent with the five factors our circuit has used since Bauman v. U.S. Dist. Court, 557 F.2d 650 (9th Cir. 1977), to determine whether mandamus relief is appropriate:

(1) whether the petitioner has no other means, such as a direct appeal, to obtain the desired relief; (2) whether the petitioner will be damaged or prejudiced in any way not correctable on appeal; (3) whether the district court's order is clearly erroneous as a matter of law; (4) whether the district court's order is an oft repeated error or manifests a persistent disregard of the federal rules; and (5) whether the district court's order raises new and important problems or issues of first impression.

Hernandez v. Tanninen, 604 F.3d 1095, 1099 (9th Cir. 2010) (internal quotation marks omitted) (quoting Perry v. Schwarzenegger, 591 F.3d 1147, 1156 (9th Cir. 2009)); see also Bauman, 557 F.2d at 654–55.

All of the Bauman factors need not be present to justify the writ. See In re Cement Antitrust Litig., 688 F.2d at 1301, 1304 (noting that the fourth and fifth factors are rarely present in the same case). “Except for supervisory mandamus cases, the absence of factor three—clear error as a matter of law—will always defeat a petition for mandamus.” Calderon v. U.S. Dist. Court for the Cent. Dist. of Cal., 163 F.3d 530, 534 (9th Cir. 1998) (en banc), abrogated on other grounds by Woodford v. Garceau, 538 U.S. 202 (2003). “In the final analysis, the decision of whether to issue the writ lies within our discretion.” In re Van Dusen, 654 F.3d 838, 841 (9th Cir. 2011) (citation omitted).

The Bauman and Cheney factors favor our review. There is no danger that the writ will supplant the normal appeals process because the district-wide shackling claims aren't connected to defendants' individual criminal cases.5 The policy doesn't apply to jury trials; thus, it causes no prejudice that would justify reversal of a conviction in a direct appeal. This case also raises new and important constitutional issues that haven't been fully considered by this court. See United States v. Brandau, 578 F.3d 1064, 1065 (9th Cir. 2009). And a survey of our circuit's district courts shows that some form of routine shackling has become a common practice and thus is an oft-repeated error.6

Accordingly, we construe defendants' appeals as petitions for writs of mandamus under our supervisory authority and find that we have jurisdiction to consider them.

B. Mootness

Article III's “case-or-controversy limitation” on federal court jurisdiction requires a live controversy between two adversaries. Friends of the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 528 U.S. 167, 180 (2000). Supervisory mandamus cases require live controversies even when we don't order a lower court to take or refrain from a specific action. See In re United States, 791 F.3d at 952. Neither party claims that this case is moot, but the court “must assure itself of its own jurisdiction.” Terenkian v. Republic of Iraq, 694 F.3d 1122, 1137 (9th Cir. 2012). There are two circumstances in this case that raise the possibility of mootness: (1) the named defendants' cases have ended, so they're no longer subject to the complained-of policy, and (2) the challenged policy is no longer in effect.

1. “In cases where intervening events have rendered the writ an ineffective or superfluous remedy, but where the controversy nonetheless remains live, we have occasionally reviewed the district court's decision for error while withholding a formal writ.” In re United States, 791 F.3d at 953 (citing Phoenix Newspapers, Inc. v. U.S. Dist. Court for the Dist. of Ariz., 156 F.3d 940, 952 (1998); United States v. Brooklier, 685 F.2d 1162, 1173 (9th Cir. 1982)). We do so when it would have been appropriate to issue the writ at the time the petition was filed. Id. at 954. This allows us to review “important issues that would otherwise escape review, while [e]nsuring that such review is limited to truly extraordinary circumstances.” Id.

Two of the defendants, Rene Sanchez-Gomez and Jasmin Isabel Morales, were not yet convicted and so were still subject to the pretrial shackling policy when they filed their notices of appeal. Construing their notices of appeal as petitions for writs of mandamus, they had a direct stake in the resolution of the controversy at the time their petitions were filed.

Named plaintiffs—or, in the mandamus setting, petitioners—must also have a continuing personal interest in the outcome of the case throughout the litigation. See Campbell-Ewald Co. v. Gomez, 136 S. Ct. 663, 669 (2016). Because they are no longer subject to the policy, defendants' personal interests in the outcome of this case have expired.

We faced the same issue in Howard. The defendants' criminal cases ended before their shackling appeals could be heard. 480 F.3d at 1009–10. We held that the case wasn't moot because it fell into the capable-of-repetition-yet-evading-review exception. Id. This exception requires repetition as to the particular complainants, and we cannot presume that defendants will be subject to criminal proceedings in the future. Id. But some criminal defendants would have been subject to the challenged policy during the litigation and would personally benefit from resolving the case. Thus, we employed the capable-of-repetition-yet-evading-review mootness exception that applied to the class action in Gerstein v. Pugh, 420 U.S. 103 (1975). Though Howard wasn't a class action, the case served the same functional purpose—it was a functional class action. See 480 F.3d at 1009–10.

The Supreme Court in Gerstein applied the capable-of-repetition-yet-evading-review mootness exception even though the named plaintiff was no longer subject to the challenged practice. 420 U.S. at 110 n.11. In that case, the class was composed of defendants held in pretrial detention without a probable cause hearing. Id. at 105–06. It wasn't clear that any representative plaintiff would remain in pretrial custody long enough for the judge to certify the class, much less decide the case. Id. at 110 n.11. But the class would continually fill with new in-custody defendants who had a live interest in the case. Id. The attorney representing the class was a public defender who would continue to represent at least some of those new defendants and class members. Id. Under those circumstances, the Court held that the case wasn't moot because the harm was capable of repetition yet evading review as to some member of the class throughout the litigation. Id.

We have applied Gerstein's analysis to functional class actions with inherently transitory claims. See Howard, 480 F.3d at 1009–10; Or. Advocacy Ctr. v. Mink, 322 F.3d 1101, 1117–18 (9th Cir. 2003).7 These cases involve circumstances “analogous to those found in class action cases where, because of the inherently transitory nature of the claims,” an individual's interests would expire before litigation could be completed. Or. Advocacy Ctr., 322 F.3d at 1117. Functional class actions share the same three features that animated the Supreme Court in Gerstein: They challenge not merely individual violations, but also broader policies or practices. See id. at 1118. Thus, they consist of continually changing groups of injured individuals who would benefit from any relief the court renders. And they have common representation, thereby guaranteeing that the cases will be zealously advocated even though the named individuals no longer have live interests in the case. See id. at 1117.

The dissent disputes this application of Gerstein. According to the dissent, Gerstein and related cases require “the existence of a procedural mechanism, such as [Federal Rule of Civil Procedure] 23,” for their mootness exceptions to apply. Dissent at 46–47. But the rule in Gerstein doesn't turn on the presence of a procedural device like Rule 23. 420 U.S. at 110 n.11. Rather, Gerstein's rule resolves the problem of inherently transitory claims while ensuring there is a live controversy for which the court can provide relief. Id.

The Supreme Court itself has indicated that Gerstein's broadening of the capable-of-repetition-yet-evading-review mootness exception could apply to cases sufficiently similar to class actions. The Court discussed Gerstein's factors in a case brought under the Fair Labor Standards Act (FLSA), Genesis Healthcare Corp. v. Symczyk, 133 S. Ct. 1523 (2013). Unlike the class action procedures in Rule 23, the FLSA's “ ‘conditional certification’ does not produce a class with an independent legal status.” Id. at 1530. The Court nonetheless considered whether, under Gerstein, the plaintiff's injury might be capable of repetition yet evading review. Id. at 1531; see also id. at 1530 (recognizing that the Court's holdings in Sosna v. Iowa, 419 U.S. 393 (1975), and U.S. Parole Commission v. Geraghty, 445 U.S. 388 (1980), depended on the “independent legal status” of class actions while making no such claim about Gerstein's holding).

The dissent claims that Genesis Healthcare still requires “the existence of a procedural mechanism ․ to aggregate the claims” as a “necessary prerequisite” for Gerstein's analysis to apply. Dissent at 46–47. But the Court did not say so. Instead, the Court noted that its application of Gerstein has “invariably focused on the fleeting nature of the challenged conduct giving rise to the claim.” 133 S. Ct. at 1531. The dissent's excursus on mootness also ignores that this is a supervisory mandamus case. See dissent at 39–53. In its supervisory mandamus role, a court of appeals properly addresses the harm of a district court policy affecting a huge class of persons who aren't parties to the mandamus petition. See, e.g., Will, 389 U.S. at 95, 104–06; Schlagenhauf v. Holder, 379 U.S. 104, 110–12 (1964); La Buy, 352 U.S. at 257–60. Unlike the dissent, see dissent at 50 n.5, the Supreme Court hasn't found a constitutional infirmity with such cases. Thus, the dissent's concerns about the lack of formal joinder and whether the decision binds other defendants, see id. at 46–49, are misplaced.

All of the Court's considerations in Gerstein are present here, and the harm—unconstitutional pretrial shackling—is inherently ephemeral, just like the pretrial detention challenges in Gerstein. We are faced with an ever-refilling but short-lived class of in-custody defendants who are subject to the challenged pretrial shackling policy. At least some members of this functional class continue to suffer the complained-of injury. Most of the defendants are represented by the Federal Defenders of San Diego. And even if we must withhold a formal writ, we can provide district-wide relief by exercising our supervisory mandamus authority, thus demonstrating that there is a live controversy here. See Knox v. Serv. Emps. Int'l Union, Local 1000, 132 S. Ct. 2277, 2287 (2012) (“A case becomes moot only when it is impossible for a court to grant any effectual relief whatever to the prevailing party.” (internal quotation marks and citations omitted)); see also In re United States, 791 F.3d at 954 (“[W]e are not categorically precluded from opining on the merits of a mandamus petition when issuance of the writ would no longer be effective.”).

2. Shortly after the original panel decision in this case, the Southern District of California changed its shackling policy in response to additional litigation about its continued use of five-point restraints. But the district court's decision to change the policy was only a voluntary cessation. See Friends of the Earth, 528 U.S. at 189. The appealed policy could be reinstated at any time. In fact, the government has indicated that it will seek to reinstate the policy unless we hold it unconstitutional. Thus, there is still a live controversy over the shackling policy.

C. The Fundamental Right to be Free of Unwarranted Restraints

At the heart of our criminal justice system is the well-worn phrase, innocent until proven guilty. See Taylor v. Kentucky, 436 U.S. 478, 483 (1978). And while the phrase may be well-worn, it must also be worn well: We must guard against any gradual erosion of the principle it represents, whether in practice or appearance. This principle safeguards our most basic constitutional liberties, including the right to be free from unwarranted restraints. See Deck v. Missouri, 544 U.S. 622, 629–30 (2005).

1. Under the Fifth Amendment, no person shall be “deprived of life, liberty, or property, without due process of law.” U.S. Const. amend. V. The Supreme Court has said time and again that “[l]iberty from bodily restraint always has been recognized as the core of the liberty protected by the Due Process Clause from arbitrary governmental action.” Youngberg v. Romeo, 457 U.S. 307, 316 (1982) (alteration in original) (quoting Greenholtz v. Neb. Penal Inmates, 442 U.S. 1, 18 (1979) (Powell, J., concurring in part and dissenting in part)). Liberty from bodily restraint includes the right to be free from shackles in the courtroom. See Deck, 544 U.S. at 629.

The Supreme Court held in Deck v. Missouri that “the Constitution forbids the use of visible shackles during the penalty phase, as it forbids their use during the guilt phase, unless that use is ‘justified by an essential state interest’—such as the interest in courtroom security—specific to the defendant on trial.” Id. at 624 (quoting Holbrook v. Flynn, 475 U.S. 560, 568–69 (1986)). In evaluating the government's justification, a court may “take into account the factors that courts have traditionally relied on in gauging potential security problems and the risk of escape at trial.” Id. at 629. While the decision whether to shackle is entrusted to the court's discretion, routine shackling isn't permitted. Id. at 629, 633. Instead, courts must make specific determinations of necessity in individual cases. Id. at 633.

The Supreme Court identified three constitutional anchors for the right: (1) the presumption that a defendant is innocent until proven guilty; (2) the Sixth Amendment right to counsel and participation in one's own defense; and (3) the dignity and decorum of the judicial process, including “the respectful treatment of defendants.” Id. at 630–31. In jury proceedings, an additional concern is that the sight of a defendant in shackles would prejudice the jury against him. Because prejudice is difficult to discern from a cold record, shackles visible to the jury are considered “inherently prejudicial.” Id. at 635 (quoting Holbrook, 475 U.S. at 568). But when security needs outweigh these other concerns, even visible restraints may be used. Id. at 632.

Consistent with Deck, we have held that criminal defendants have a “constitutional right to be free of shackles and handcuffs in the presence of the jury absent an essential state interest that justifies the physical restraints.” Williams v. Woodford, 384 F.3d 567, 591 (9th Cir. 2004) (citations omitted). We require lower courts to consider concerns similar to those articulated by the Court in Deck, such as whether shackles would prejudice the jury, diminish the presumption of innocence, impair the defendant's mental capabilities, interfere with the defendant's ability to communicate with counsel, detract from the dignity and decorum of the courtroom or cause physical pain. See Spain v. Rushen, 883 F.2d 712, 721 (9th Cir. 1989). “ ‘In all [ ] cases in which shackling has been approved,’ we have noted, there has been ‘evidence of disruptive courtroom behavior, attempts to escape from custody, assaults or attempted assaults while in custody, or a pattern of defiant behavior toward corrections officials and judicial authorities.’ ” Gonzalez v. Pliler, 341 F.3d 897, 900 (9th Cir. 2003) (alteration in original) (quoting Duckett v. Godinez, 67 F.3d 734, 749 (9th Cir. 1995)).

We now clarify the scope of the right and hold that it applies whether the proceeding is pretrial, trial, or sentencing, with a jury or without.8 Before a presumptively innocent defendant may be shackled, the court must make an individualized decision that a compelling government purpose would be served and that shackles are the least restrictive means for maintaining security and order in the courtroom.9 See, e.g., Gonzalez, 341 F.3d at 900; Duckett, 67 F.3d at 748; Spain, 883 F.2d at 721, 728. Courts cannot delegate this constitutional question to those who provide security, such as the U.S. Marshals Service. Nor can courts institute routine shackling policies reflecting a presumption that shackles are necessary in every case.10

This right to be free from unwarranted shackles no matter the proceeding respects our foundational principle that defendants are innocent until proven guilty. The principle isn't limited to juries or trial proceedings. It includes the perception of any person who may walk into a public courtroom, as well as those of the jury, the judge and court personnel. A presumptively innocent defendant has the right to be treated with respect and dignity in a public courtroom, not like a bear on a chain. See Zuber, 118 F.3d at 106 (Cardamone, J., concurring) (“The fact that the proceeding is non-jury does not diminish the degradation a prisoner suffers when needlessly paraded about a courtroom, like a dancing bear on a lead, wearing belly chains and manacles.”).

And it's not just about the defendant. The right also maintains courtroom decorum and dignity:

The courtroom's formal dignity, which includes the respectful treatment of defendants, reflects the importance of the matter at issue, guilt or innocence, and the gravity with which Americans consider any deprivation of an individual's liberty through criminal punishment. And it reflects a seriousness of purpose that helps to explain the judicial system's power to inspire the confidence and to affect the behavior of a general public whose demands for justice our courts seek to serve.

Deck, 544 U.S. at 631. The most visible and public manifestation of our criminal justice system is the courtroom. Courtrooms are palaces of justice, imbued with a majesty that reflects the gravity of proceedings designed to deprive a person of liberty or even life. A member of the public who wanders into a criminal courtroom must immediately perceive that it is a place where justice is administered with due regard to individuals whom the law presumes to be innocent. That perception cannot prevail if defendants are marched in like convicts on a chain gang. Both the defendant and the public have the right to a dignified, inspiring and open court process. Thus, innocent defendants may not be shackled at any point in the courtroom unless there is an individualized showing of need.

2. This right “has deep roots in the common law.” Deck, 544 U.S. at 626. The Supreme Court has “regularly observed that the Due Process Clause specially protects those fundamental rights and liberties which are, objectively, deeply rooted in this Nation's history and tradition, and implicit in the concept of ordered liberty, such that neither liberty nor justice would exist if they were sacrificed.” Washington v. Glucksberg, 521 U.S. 702, 720–21 (1997) (internal quotation marks and citations omitted).

One traditional justification for the right was allowing defendants to try their cases without the distraction of shackles and any attendant physical pain. See Deck, 544 U.S. at 626; see also id. at 638–39 (Thomas, J., dissenting).11 An early commentator noted that defendants should approach the court free of shackles “so that their pain shall not take away any manner of reason, nor them constrain to answer, but at their free will.” Id. at 626 (quoting 3 Edward Coke, Institutes of the Laws of England 34 (1797)). But the right was also motivated by the desire to protect defendants' dignity:

[E]very person at the time of his arraignment, ought to be used with all the humanity and gentleness which is consistent with the nature of the thing, and under no other terror or uneasiness than what proceeds from a sense of his guilt, and the misfortune of his present circumstances; and therefore ought not to be brought to the bar in a contumelious manner; as with his hands tied together, or any other mark of ignominy and reproach; nor even with fetters on his feet, unless there be some danger of a rescous or escape.

2 William Hawkins, A Treatise of the Pleas of the Crown 434 (John Curwood, 8th ed. 1824). Still, there were certain situations when the need for security overcame the right to be free of shackles: “[A] defendant ‘must be brought to the bar without irons, or any manner of shackles or bonds; unless there be evident danger of an escape.’ ” Deck, 544 U.S. at 626 (quoting 4 William Blackstone, Commentaries on the Laws of England 317 (1769)).

The Supreme Court in Deck found that the common law drew a distinction between trial and pretrial proceedings when applying the right because “Blackstone and other English authorities recognized that the rule did not apply at ‘the time of arraignment,’ or like proceedings before the judge.” Id. (quoting 4 Blackstone, Commentaries on the Laws of England 317) (citing Trial of Christopher Layer, 16 How. St. Tr. 94, 99 (K.B. 1722)). This statement on pretrial proceedings is undoubtedly dictum in a case about shackling at capital sentencing. Persuasive Supreme Court dicta are usually heeded by lower courts. See United States v. Montero-Camargo, 208 F.3d 1122, 1132 n.17 (9th Cir. 2000) (en banc). But dicta “ought not to control the judgment in a subsequent suit, when the very point is presented for decision.” Humphrey's Ex'r v. United States, 295 U.S. 602, 627 (1935) (quoting Cohens v. Virginia, 19 U.S. 264, 399 (1821) (Marshall, C.J.)). The Supreme Court's dictum on pretrial proceedings in Deck doesn't control this case because it's contradicted by the very sources on which the Supreme Court relied.12

The early commentators didn't draw the bright line between trial and arraignment that the Deck Court seemed to believe they did. Coke's discussion of shackling noted that “[i]t is an abuse that prisoners be charged with irons, or put to any pain before they be attainted.” 3 Coke, Institutes of the Laws of England 34. And Blackstone did not recognize that the rule against shackles didn't apply at the time of arraignment or proceedings before a judge. Instead, the language the Court cited and partially quoted said the opposite: Shackles at arraignment and pretrial proceedings are acceptable only in situations of escape or danger.

The prisoner is to be called to the bar by his name; and it is laid down in our an[c]ient books, that, though under an indictment of the highest nature, he must be brought to the bar without irons, or any manner of shackles or bonds; unless there be evident danger of an escape, and then he may be secured with irons. But yet in Layer's case, A.D. 1722[,] a difference was taken between the time of arraignment, and the time of trial; and accordingly the prisoner stood at the bar in chains during the time of his arraignment.

4 Blackstone, Commentaries on the Laws of England 317. Shackles at arraignment and trial are different, as Blackstone noted, but only because shackles are more easily justified at the former, which was demonstrated by Layer's case.

Layer's case, relied on by both Blackstone and the Supreme Court, began with Layer's appeal to be unshackled at his arraignment. The Trial of Christopher Layer, esq; at the King's-Bench for High-Treason, Nov. 21. 1722, in 6 A Complete Collection of State-Trials, and Proceedings Upon High-Treason 229–32 (2d ed. 1730). The government justified the shackles on the ground that Layer had previously attempted to escape. Id. Layer's lawyer objected strongly, explaining that “by Law he ought not to be called upon, even to plead, till his Fetters are off.” Id. at 231. He argued that shackles not only caused physical and mental “uneasiness,” but also that they besmirched the decorum of the court:

[S]omething of the Dignity of the Court might be considered in this Matter, for a Court of Justice, the highest in the Kingdom for criminal Matters, where the King himself is supposed to be personally present, to have a Man plead for his Life before them in Chains, seems to be very unsuitable. He is now before the same awful and just Tribunal which he will be before when he is tried, and why not therefore without Chains as well now as then ․ ?

Id. While Layer was ultimately unsuccessful, his argument demonstrates that shackling at arraignment was not a standard practice, or even permissible, absent a demonstrated need.

The dissent struggles manfully against the plain language of Layer's case and Blackstone. See dissent at 58–62. It claims to “follow the Supreme Court's interpretation” of Layer's case by pointing to Deck, id. at 61–62 n.13, but nowhere does the Deck majority analyze the case. We merely repeat what Blackstone and Layer's case provide—that shackling at arraignment was allowed after a showing of need. Layer's case applied the exception to Blackstone's basic rule: A prisoner “must be brought to the bar without irons, or any manner of shackles or bonds; unless there be evident danger of an escape.” 4 Blackstone, Commentaries on the Laws of England 317. There's nothing to indicate that shackles were used at arraignments more generally without a particular reason; Layer's case suggests the contrary.

Early American courts “traditionally followed Blackstone's ‘ancient’ English rule.” Deck, 544 U.S. at 626–27 (collecting cases). Blair v. Commonwealth, relying on a legal encyclopedia, explained that courts followed “the common–law rule” that “shackling defendant[s] during arraignment, during the calling and examination of the jurors, or at any time during the trial, except in extreme cases to prevent escape or to protect the bystanders from the danger of defendant's attack, [was] reversible error.” 188 S.W. 390, 393 (Ky. Ct. App. 1916) (internal quotation marks omitted) (quoting 12 William Mack, Cyclopedia of Law and Procedure 529 (1904)). Likewise, Rainey v. State quoted Bishop's authoritative treatise to note that “ ‘the rule [against shackling] at arraignment where only a plea is required is less strict’ ” than the rule at trial. 20 Tex. App. 455, 472 (1886) (quoting 1 Joel Prentiss Bishop, Criminal Procedure § 955 (3d ed. 1880)). Contrary to the dissent's belief, that the rule “is less strict” doesn't mean it didn't exist at all.13 Bishop understood the common law rule just as we do: “[I]f a keeper deems it necessary,” then the general rule that the defendant “should not be in irons” at arraignment could be relaxed. 1 Bishop, Criminal Procedure § 731; see also Parker v. Territory, 52 P. 361, 363 (Ariz. 1898) (“ ‘A person charged with a public offense shall not before conviction be subjected to any more restraint than is necessary for his detention to answer the charge,’—which is but the common-law and constitutional right of a prisoner embodied in the statute.” (citation omitted)). Thus, we have a tradition dating from time out of mind that defendants will appear in court prior to their conviction as free men with their heads held high.

3. The government contends that individualized determinations are required only before shackles are used in the jury's presence. Otherwise, it argues, the right is sufficiently protected by considering generally applicable security concerns, deferring to the U.S. Marshals Service and leaving the rest to individual judges' discretion. The government also asks us to analyze this case under Bell v. Wolfish, 441 U.S. 520 (1979).

But Bell dealt with pretrial detention facilities, not courtrooms.14 Those facilities are meant to restrain and keep order, not dispense justice. They are a mere step away from detention in prison. We emphatically reject the idea that courtrooms are (or should be) perceived as places of restraint and punishment, or that courtrooms should be governed exclusively by the type of safety considerations that justify detention facility policies. We must make every reasonable effort to avoid the appearance that courts are merely the frontispiece of prisons.

We have a long tradition of giving correctional officials a wide berth in maintaining security within their own facilities.15 See id. at 540 n.23. But we don't have a tradition of deferring to correctional or law enforcement officers as to the treatment of individuals appearing in public courtrooms. In the courtroom, law enforcement officers have no business proposing policies for the treatment of parties as a class. Insofar as they have information pertaining to particular defendants, they may, of course, bring it to the court's attention. But a blanket policy applied to all defendants infuses the courtroom with a prison atmosphere. The Marshals Service should not have proposed it and the judges should not have paid heed.

We must take seriously how we treat individuals who come into contact with our criminal justice system—from how our police interact with them on the street to how they appear in the courtroom. How the justice system treats people in these public settings matters for the public's perception, including that of the defendant. Practices like routine shackling and “perp walks” are inconsistent with our constitutional presumption that people who have not been convicted of a crime are innocent until proven otherwise. That's why we must examine these practices more skeptically than those deployed in an institutional setting like Bell. See, e.g., Deck, 544 U.S. at 634 (holding that a defendant's Fifth Amendment rights were violated by visible shackling before a jury at capital sentencing proceedings); Lauro v. Charles, 219 F.3d 202, 212–13 (2d Cir. 2000) (holding that a defendant's Fourth Amendment rights were violated by a staged and filmed perp walk done without a legitimate law enforcement reason). We must treat people with respect and dignity even though they are suspected of a crime.

* * *

The Constitution enshrines a fundamental right to be free of unwarranted restraints. Thus, we hold that if the government seeks to shackle a defendant, it must first justify the infringement with specific security needs as to that particular defendant. Courts must decide whether the stated need for security outweighs the infringement on a defendant's right. This decision cannot be deferred to security providers or presumptively answered by routine policies. All of these requirements apply regardless of a jury's presence or whether it's a pretrial, trial or sentencing proceeding. Criminal defendants, like any other party appearing in court, are entitled to enter the courtroom with their heads held high.

The policy that defendants challenged here isn't presently in effect. Thus, although we hold that policy to be unconstitutional, we withhold the issuance of a formal writ of mandamus at this time.

DENIED.

I fully concur in Judge Kozinski's opinion with its comprehensive historical analysis. I write separately only to offer a brief comment about Judge Ikuta's lengthy, well written dissent.

In addition to noting my disagreement with the dissent's interpretation of common law and Supreme Court authority, I also observe that the dissent unfortunately lacks sensitivity to two of the most important components of our system of justice. The first is the dignity with which court proceedings should be conducted. The dissent thus ignores the degradation of human beings who stand before a court in chains without having been convicted, or in many instances, without even having been formally charged with any crime. Second, the dissent lacks sensitivity to the proper role of the judges as opposed to the Marshals Service in determining how a courtroom should be run. Thus the dissent accepts the data provided by the Marshals Service even though no district court judge has ever made any finding of fact concerning the data's accuracy or whether it provides a good reason for this unprecedented mass shackling.

Our court today correctly upholds the proper role of the judges, as opposed to the jailors, in the courtroom.

Far removed from the potential dangers of a trial court, the majority holds that criminal defendants whose cases are now moot can use their individual appeals as vehicles to invalidate the prospective application of a federal district court's policy of deferring to the United States Marshals Service on questions of courtroom security. In reaching this conclusion, the majority ignores Article III's limitations on federal judicial power, conjures up an unsupported and unprecedented exception to mootness, chastises district judges for following our case law, brushes aside inconvenient Supreme Court reasoning, creates an unjustifiable circuit split, and discovers a one-size-fits-all courtroom security policy in the Constitution. We should not be hearing this case at all, much less using it to announce a sweeping and unfounded new constitutional rule with potentially grave consequences for state and federal courthouses throughout this circuit. I dissent.

I

In July 2013, the United States Marshals Service, pursuant to its congressional charge “to provide for the security ․ of the United States District Courts,” 28 U.S.C. § 566(a), recommended that the judges of the Southern District of California allow the Marshals Service to produce all in-custody defendants in full restraints for non-jury proceedings. The Marshals Service based this recommendation on several factors. For one, a number of dangerous incidents had recently occurred in the courthouse. In 2013 alone, there were two separate inmate-on-inmate assaults inside courtrooms; an inmate was stabbed in the face as a result of one of those assaults. The Marshals Service also discovered that several detainees had armed themselves with homemade weapons in holding cells, including a detainee with no violent background who attempted to smuggle a razor blade in his shoe.

Second, the Marshals Service determined that it lacked sufficient information to predict which detainees would present a danger. In many cases, detainees with no history of violence, or those who were charged with non-violent offenses, engaged in violent acts while in custody. For instance, in 2013 there were seven detainee-on-staff assaults in the Southern District of California; six of the offenders had been charged with non-violent offenses, and five of those six had no histories of violence. Moreover, the Marshals Service can access only limited criminal background information regarding detainees who are not residents of the United States, and the Southern District of California hears an unusually high number of cases involving such detainees. Accordingly, the Marshals Service concluded that it had little ability to predict which detainees would present a danger.

The Marshals Service also noted logistical concerns that enhanced the potential danger arising from the large number of criminal defendants cycling through the courthouse. In the years leading up to the policy's implementation, the Marshals Service produced approximately 40,000 in-custody defendants for court appearances, with an average of over 200 defendants moving through district cellblocks per day. The high volume of in-custody criminal defendants, the close quarters in the courtrooms used by magistrate judges, the configurations of the courtrooms used by district judges, and budgetary constraints that forced the Marshals Service to reduce the allocation of resources to courtroom protection duties all contributed to heightened security concerns. In short, the Marshals Service's security recommendation arose from a confluence of factors, many of which were specific to the Southern District of California.

After consulting with the United States Attorney's Office, the Federal Defenders of San Diego, and a Criminal Justice Act panel representative, the district court concluded that it should defer to the Marshals Service's recommendation on this courtroom security issue, with two exceptions. First, the district court declined to adopt the Marshals Service's recommendation with respect to guilty plea colloquies and sentencing hearings. Second, the district court reserved the right of any individual judge to opt out of the policy. In deciding to implement the Marshals Service's recommendation, the district court relied on our decision in United States v. Howard, 480 F.3d 1005, 1013 (9th Cir. 2007), and on the Second Circuit's decision in United States v. Zuber, 118 F.3d 101, 104 (2d Cir. 1997), each of which held that deference to the Marshals Service's judgment regarding the use of restraints on detainees during non-jury pretrial proceedings did not violate the detainees' constitutional rights.

Challenges to the new policy came quickly, including from the defendants now before us on appeal. In October 2013, Jasmin Morales made her initial appearance before a magistrate judge in full restraints pursuant to the new security policy. Morales had been charged with felony importation of a controlled substance, in violation of 21 U.S.C. §§ 952 and 960. Her counsel moved to have the restraints removed during the pretrial proceedings, but the magistrate judge denied the motion. While her criminal case was moving forward, Morales filed an emergency motion with the district court challenging the pretrial restraint policy. A district court judge denied that motion, and her counsel filed a notice of appeal in November 2013. A few months later, in April 2014, Morales pleaded guilty. The district court imposed a sentence of eighteen months imprisonment and three years of supervised release and entered a final judgment on June 19, 2014. At that point, Morales's criminal case before the district court was over.1 The other defendants whose appeals are before us—Rene Sanchez-Gomez, Moises Patricio-Guzman, and Mark Ring—have similar stories.2

This case accordingly comes to us in an odd procedural posture: Each of the four defendants' criminal cases came to a close before we heard their appeals, and the four defendants (represented here by the Federal Defenders of San Diego) are before us challenging only the Marshals Service's prospective use of restraints during pretrial proceedings. They do not seek review of the individual decisions to permit the use of restraints in their cases. They do not seek damages for any injury they incurred due to this policy. Nor do they seek to have their convictions or sentences set aside as a result of any prejudicial effect of the restraint policy. Instead, the Federal Defenders of San Diego, allegedly on behalf of the four defendants, seeks prospective relief for all future pretrial detainees who may have pretrial proceedings in the Southern District of California. The defendants seek this relief even though, as the majority concedes, Maj. op. at 16, they are no longer subject to the challenged policy. In fact, none of these defendants has any reason to step foot in a federal courtroom as a pretrial detainee again. Thus, as the majority acknowledges, these defendants “are making class-like claims and asking for class-like relief,” Maj. op. at 11, but are doing so via their individual criminal cases. The threshold question presented in this case is whether, consistent with Article III of the Constitution, they may do so.

II

Because Morales, Sanchez-Gomez, Patricio-Guzman, and Ring have no ongoing interest in the purely prospective relief they seek, see Maj. op. at 16, their appeals are moot unless some exception to the ordinary rules of mootness applies. But neither the Supreme Court nor our precedent has established any applicable exception. The majority implicitly concedes as much by contriving a new exception—the “functional class action,” id. at 16—in order to rescue these appeals from mootness. Because this theory is inconsistent with Supreme Court precedent and incompatible with Article III's case-or-controversy requirement, the majority's creative effort to sidestep mootness should be rejected.

A

The majority treats Article III's case-or-controversy requirement as a mere obstacle in its path to the merits that can be avoided through calculated maneuvering. But our adherence to this requirement “is essential if federal courts are to function within their constitutional sphere of authority.” North Carolina v. Rice, 404 U.S. 244, 246 (1971) (per curiam). The Constitution constrains federal “judicial Power” to nine classes of “Cases” and “Controversies.” U.S. Const. art. III, § 2; Rice, 404 U.S. at 246. A dispute is not a qualifying case or controversy unless we can afford relief to the parties before us, see Rice, 404 U.S. at 246, and the “case-or-controversy requirement subsists through all stages of federal judicial proceedings,” Fed. Election Comm'n v. Wis. Right to Life, Inc., 551 U.S. 449, 461 (2007) (internal quotation marks omitted). Thus, “it is not enough that a dispute was very much alive when suit was filed.” Lewis v. Cont'l Bank Corp., 494 U.S. 472, 477 (1990). Instead, if a party seeking relief loses a “cognizable interest in the outcome” at any stage of the litigation, Murphy v. Hunt, 455 U.S. 478, 481 (1982) (per curiam) (internal quotation marks omitted), then the matter becomes moot and is “no longer a ‘Case’ or ‘Controversy’ for purposes of Article III, ․ [n]o matter how vehemently the parties continue to dispute the lawfulness of the conduct that precipitated the lawsuit,” Already, LLC v. Nike, Inc., 133 S. Ct. 721, 726–27 (2013). This constraint on federal judicial power exists, as the majority acknowledges, whether the parties are before the court on an appeal, a petition for a writ of mandamus, or any other means of obtaining relief. Maj. op. at 16. Here, the defendants' claims that the pretrial restraint policy violates the Constitution are moot “because even a favorable decision” would not entitle the defendants to any relief. Murphy, 455 U.S. at 481. Accordingly, absent some exception to the ordinary rules of mootness, we lack jurisdiction over these consolidated appeals, and they must be dismissed.

The established exceptions to mootness do not give the majority much to work with in its effort to find a live case or controversy. The majority references the exception for cases capable of repetition, yet evading review, Maj. op. at 16–17, but this exception applies only if “there is a reasonable expectation that the same complaining party will be subject to the same action again,” Spencer v. Kemna, 523 U.S. 1, 17 (1998) (brackets omitted) (quoting Lewis, 494 U.S. at 481) (internal alterations omitted). Here, as the majority concedes, “we cannot presume that defendants will be subject to criminal proceedings in the future.” Maj. op. at 16; accord O'Shea v. Littleton, 414 U.S. 488, 497 (1974); Cox v. McCarthy, 829 F.2d 800, 804 n.3 (9th Cir. 1987). Accordingly, the alleged injury is not capable of repetition as to the parties before us, and the “capable of repetition, yet evading review” exception to mootness is inapplicable. Spencer, 523 U.S. at 17.

B

Instead of conceding that this case is beyond our power to decide, the majority invents a new “functional class action” exception to mootness. Maj. op. at 16–20. Relying on Gerstein v. Pugh, 420 U.S. 103 (1975), a case considering mootness in the class action context, the majority reasons that a case is not moot whenever there is “an ever-refilling but short-lived class” of defendants who are subject to a challenged policy, “[a]t least some members of this functional class continue to suffer the complained-of injury,” and most of the members are represented by zealous advocates. Maj. op. at 19. But a group of ever-changing individuals with similar concerns (as the majority envisions) does not constitute the sort of class that can avoid mootness. Even when a plaintiff purports to bring an action on behalf of others, the action will become moot when the plaintiff's own claims become moot, unless the plaintiff has used a procedural mechanism, such as class certification under Rule 23 of the Federal Rules of Civil Procedure, that can produce a class with “an independent legal status” or otherwise effectively joins “additional parties to the action.” Genesis Healthcare Corp. v. Symczyk, 133 S. Ct. 1523, 1530 (2013). Without the creation of such a class pursuant to a statute or rule, a group of interested individuals cannot be a party to the action before the court, and therefore the court may not consider their interests in a particular case for purposes of a mootness inquiry. See id.

To understand why Gerstein is inapposite here, some background is needed to explain how the mootness doctrine applies in the class action context. Rule 23 provides a procedure that allows courts to aggregate the claims of multiple parties. See Deposit Guar. Nat'l Bank v. Roper, 445 U.S. 326, 339 (1980) (stating that a class action is “[t]he aggregation of individual claims in the context of a classwide suit”). Once a class is certified under Rule 23, it “acquires an independent legal status.” Genesis Healthcare, 133 S. Ct. at 1530. The members of a class are parties to the action and are generally bound by the judgment. See Taylor v. Sturgell, 553 U.S. 880, 884 (2008).

Because a class is comprised of multiple parties to the legal action, a court's mootness inquiry in a class action lawsuit is broader than in traditional litigation on an individual's own behalf. See, e.g., Franks v. Bowman Transp. Co., 424 U.S. 747, 755–56 (1976) (holding that the interests of “unnamed members of the class” who are entitled to relief may satisfy the case-or-controversy requirement). The named representative of a class must generally have standing at the commencement of an action and when the district court rules on a motion for class certification. Sosna v. Iowa, 419 U.S. 393, 402 (1975). But even if the named representative's case becomes moot after the district court has ruled on a motion for class certification, the case itself is not moot so long as at least one member of the putative class has a live interest. See, e.g., id. (class action not moot when named representative's claim becomes moot after class certification); U.S. Parole Comm'n v. Geraghty, 445 U.S. 388, 404 (1980) (class action not moot when named representative's claim becomes moot after denial of class certification, provided that the class is subsequently certified). This makes sense: A certified class contains parties with ongoing live claims who have an entitlement to relief regardless whether the named representative's case becomes moot after the complaint is filed. See Franks, 424 U.S. at 755–57. Of course, if it turns out that the putative class was never actually eligible for certification, then the entire action dies as moot along with the class representative's claim. Geraghty, 445 U.S. at 404.

Even if a named representative's claims become moot before the district court has ruled on a class certification motion, a class claim may escape mootness under certain circumstances. This is the Gerstein rule. As the Supreme Court has explained, Gerstein “recognized ․ that ‘some claims are so inherently transitory that the trial court will not have even enough time to rule on a motion for class certification before the proposed representative's individual interest expires.’ ” County of Riverside v. McLaughlin, 500 U.S. 44, 52 (1991) (brackets omitted) (quoting Geraghty, 445 U.S. at 399). Under these circumstances, a judicial decision to certify a class after the named representative's individual claim is moot may relate back to the time the named representative filed the class-action complaint, and the action will not be moot so long as members of the class continue to have a live controversy. See Gerstein, 420 U.S. at 110 n.11 (citing Sosna, 419 U.S. at 402 n.11).

In Gerstein, two pretrial detainees sued assorted county officials on behalf of a class of pretrial detainees under 42 U.S.C. § 1983 in order to challenge Florida's practice of not providing detainees with a timely probable cause hearing. Id. at 106–07. By the time the case reached the Supreme Court, the named representatives had been convicted, and it was not clear whether their individual claims had become moot before or after the district court certified the class. Id. at 110 n.11. Given the transitory nature of pretrial custody, the clear existence of a class, and the class's representation by counsel with similarly situated clients, the Supreme Court held that the class action was not moot even if the named representatives' claims expired before certification. Id. “In such cases, the ‘relation back’ doctrine is properly invoked to preserve the merits of the case for judicial resolution.” McLaughlin, 500 U.S. at 52. As Gerstein illustrates, the “relation back” doctrine serves a very particular purpose. Specifically, in inherently transitory situations, the Court deems class certification to have occurred at the time the named representative filed the complaint with class allegations, at which time the named representative's claims were live. See 1 William B. Rubenstein et al., Newberg on Class Actions § 2:13 at 123 (5th ed. 2011). Because the named representative's claims therefore constructively became moot after the class's certification, the rule that a class action does not become moot in such circumstances applies. See id. at 123–24.

The Supreme Court later explained the limits of the Rule 23 mootness doctrine in considering its applicability to a collective action under the Fair Labor Standards Act (FLSA). The FLSA allows employees to bring a collective action on behalf of “other employees similarly situated,” but employees do not become parties to the action unless they elect to opt into it. 29 U.S.C. § 216(b). In Genesis Healthcare, the named plaintiff's case became moot before the district court had “conditionally certified” the action,3 so no other employees had yet opted into the collective action. 133 S. Ct. at 1530. The district court therefore dismissed the lawsuit as moot. Id. at 1527. The Third Circuit reversed, holding that the collective action was not moot because, if the employee were subsequently successful in obtaining conditional certification, the district court should “relate the certification motion back to the date on which respondent filed her complaint.” Id. at 1528.

The Supreme Court disagreed. Under § 216(b), approval of a plaintiff's conditional certification motion “does not produce a class with an independent legal status, or join additional parties to the action,” unlike class action certification under Rule 23. Id. at 1530. Rather, “[t]he sole consequence of conditional certification is the sending of court-approved written notice to employees, who in turn become parties to a collective action only by filing written consent with the court.” Id. (citation omitted). The Court concluded that, even if the original plaintiff “were to secure a conditional certification ruling on remand, nothing in that ruling would preserve her suit from mootness.” Id. In other words, because no claimants had opted into the collective action, a court could not consider their interests in determining whether the plaintiff's suit was moot.

Relying on Gerstein, the plaintiff in Genesis Healthcare argued that in “inherently transitory” cases, a court could give the plaintiff an opportunity to complete the § 216(b) collective action process. Id. at 1530–31. If the court granted the conditional certification and employees subsequently joined the collective action, the plaintiff argued, the court should then “relate back” this successful creation of a collective action to the date the original plaintiff filed the complaint. Id. The Court did not rule on this suggestion, however, because the plaintiff's action in that case was not transitory in nature. Id. at 1531.4 But even this argument rested on the proposition that, at some point in time, multiple plaintiffs with live cases and controversies would be parties to the action before the court, which would overcome the mootness of the original plaintiff's claim.

As the Supreme Court's cases make clear, a necessary prerequisite to applying the mootness doctrine applicable to Rule 23 class actions is the existence of a procedural mechanism, such as Rule 23 or perhaps § 216(b), that allows a court to aggregate the claims of multiple potential claimants and make them parties to the legal action. See id. at 1530 (stating that the “essential” aspect of Sosna and its progeny “was the fact that a putative class action acquires an independent legal status once it is certified”). Contrary to the majority's “functional class action” theory, it is not enough for a party to assert an inherently transitory claim on behalf of others; there must be a statutory or procedural mechanism that aggregates their claims before the court. See id. (characterizing the “line of cases” of which Gerstein is a part as applying to “an ‘inherently transitory’ class-action claim,” not to all inherently transitory claims). Indeed, the Supreme Court has never suggested that “unjoined claimants” could prevent the named plaintiff's case from becoming moot. See id. at 1531. For example, even though a collective action under the FLSA shares certain features of a class action, the class action mootness rules cannot apply unless and until the collective action includes the interests of other employees who have joined the action. See id. at 1530.

The majority's “functional class action” theory cannot create a class that has an independent legal status, whether under Rule 23 or otherwise. Nor does it have the effect of joining any additional criminal defendants as parties to this action. Accordingly, there are no parties before the court with a live case or controversy who could prevent the action from becoming moot. Gerstein merely allows a court that certifies a class to relate the existence of the class back to an earlier point in time, when a named party had a live claim. See id. at 1530–31. Because there is no class action counterpart in the Federal Rules of Criminal Procedure, nor an analogous means of aggregating multiple criminal defendants for class-wide resolution of common claims in the context of federal prosecutions, there is nothing a court can “relate back” after a criminal defendant's individual claim becomes moot. Accordingly, we must apply “the usual rule that litigation is conducted by and on behalf of the individual named parties only.” Califano v. Yamasaki, 442 U.S. 682, 700–01 (1979). The majority's reliance on Gerstein is therefore to no avail. Although criminal defendants could bring civil actions as a class under Rule 23, cf. Gerstein, 420 U.S. at 106–07, a defendant in a criminal prosecution cannot, through his or her individual case, represent and bind other criminal defendants.

The majority argues that “the rule in Gerstein doesn't turn on the presence of a procedural device like Rule 23,” but instead is a free-floating means of “resolv[ing] the problem of inherently transitory claims while ensuring there is a live controversy to which the court can provide relief.” Maj. op. at 18. To the extent the majority means that a federal court can decide a moot claim merely because it is transitory, the majority's theory is clearly contrary to the Constitution's case-or-controversy requirement. See Murphy, 455 U.S. at 481–82 (holding, in a post-Gerstein case, that a pretrial detainee's individual transitory claim became moot “once he was convicted”). Rather, the Supreme Court has been careful to require a live case or controversy pending before the court through a class action that aggregates the claims of multiple parties. See, e.g., Genesis Healthcare, 133 S. Ct. at 1530–31 (characterizing the “line of cases” of which Gerstein is a part as applying to “class-action claim[s]”); McLaughlin, 500 U.S. at 51–52 (applying Gerstein and holding “that by obtaining class certification, plaintiffs preserved the merits of the controversy for our review” (emphasis added)); cf. Murphy, 455 U.S. at 481–84 (implicitly rejecting the view expressed by the dissenting justice that, under Gerstein, “the formalities of class certification are unnecessary,” id. at 486 n.3). A class action's aggregation of claims solves the problem of inherently transitory claims because, as long as at least one member of the class has a live claim, a federal court will have jurisdiction to resolve it.

Accordingly, contrary to the majority's contentions, the rules for mootness in the class action context do not apply to the separate actions brought by Morales, Sanchez-Gomez, Patricio-Guzman, and Ring. Indeed, the majority's reasoning on this point is even weaker than the plaintiff's arguments in Genesis Healthcare. In that case the plaintiff at least made allegations pursuant to a federal statute that allowed collective action. Genesis Healthcare, 133 S. Ct. at 1527. Here, the criminal defendants did not seek any class or collective status, nor did the defendants even raise such an issue before the district court or to us. At best, one might suggest (as the majority does) that the presence of the Federal Defenders of San Diego as counsel binds the parties together. Maj. op. at 19. But not only do the federal public defenders lack the capacity to aggregate their clients' claims into an independent class, Congress has also declined to allow federal public defenders to bring civil rights claims on behalf of criminal defendants. See 18 U.S.C. § 3006A(a) (limiting the scope of representation); id. § 3006A(g)(2)(A) (restricting federal public defenders from “engag[ing] in the private practice of law”). Beyond constituting a misapplication of Supreme Court precedent, the so-called “functional class action” devised by the majority allows the federal public defenders to make an end-run around this statutory limitation by bringing the functional equivalent of a civil rights class action under the guise of a criminal appeal, without ever meeting (or even attempting to meet) the requirements of Rule 23. Thus, the majority errs on all fronts: it contravenes the Constitution, a relevant federal statute, and federal procedural rules.5

In addition to its misplaced reliance on Gerstein to support its “functional class action” theory, the majority's reliance on Oregon Advocacy Center v. Mink, 322 F.3d 1101 (9th Cir. 2003), and United States v. Howard, 480 F.3d 1005 (9th Cir. 2007), Maj. op. at 17, is equally erroneous. Oregon Advocacy Center stands for the unremarkable proposition that a federally authorized organization established to represent the rights of people with disabilities has associational standing to bring a challenge on behalf of mentally incapacitated defendants. 322 F.3d at 1116. That case was not moot because the organization was challenging an ongoing policy causing ongoing harm to the organization's constituents. Id. at 1118. There is no such organization in our case; rather, the only organization involved in this appeal, the federal public defenders, is precluded from bringing civil actions, see 18 U.S.C. § 3006A(a), (g)(2)(A), a stark contrast to the organization in Oregon Advocacy Center, which Congress had authorized to bring such lawsuits. 322 F.3d at 1113.

Nor does Howard provide support. Howard erroneously relied on Oregon Advocacy Center for the proposition that a case is not moot under the “capable of repetition, yet evading review” doctrine “when the defendants are challenging an ongoing government policy.”6 480 F.3d at 1010. As explained above, this is an erroneous reading of the case, which involved associational standing. Moreover, Howard's ruling is contrary to Supreme Court precedent, which limits the “capable of repetition, yet evading review” exception to cases in which there is “a reasonable expectation that the same complaining party will be subject to the same action again.”7 Spencer, 523 U.S. at 17 (internal quotation marks and brackets omitted). The majority correctly acknowledges that this requirement is not met here. Maj. op. at 16. That acknowledgment should have ended this case, not invited the majority's “functional class action” theory.

In short, the criminal defendants here lack a legally cognizable interest in this appeal, and there is no reasonable expectation that they will be subject to the district court's restraint policy again. Nor have these defendants brought a class action under Rule 23 that could be certified, or any equivalent action that produces “a class with an independent legal status” or “join[s] additional parties to the action.” Genesis Healthcare, 133 S. Ct. at 1530. We cannot create jurisdiction where none exists, but that is precisely what the majority has attempted to do with its novel and unfounded “functional class action” theory. Because there is no pretrial detainee with a live case who is a party to this appeal, this case must be dismissed as moot.

Although this appeal should be dismissed, the district court's policy is not insulated entirely from judicial review. For example, we likely would have jurisdiction over a class action brought by pretrial detainees under Bivens v. Six Unknown Named Agents, 403 U.S. 388 (1971), to recover damages from the individuals implementing the restraint policy,8 or to seek to enjoin the United States Marshal for the district from carrying out the policy. Cf., e.g., Armstrong v. Exceptional Child Ctr., Inc., 135 S. Ct. 1378, 1384 (2015) (noting that federal courts have long had the equitable power to enjoin unlawful conduct by federal officers); Shields v. Utah Idaho Cent. R.R. Co., 305 U.S. 177, 183 (1938) (similar); see also 5 U.S.C. § 702 (waiving sovereign immunity for such claims). Alternatively, the Ninth Circuit could exercise its supervisory powers by issuing appropriate guidance through the judicial council of the circuit. See 28 U.S.C. § 332(d)(1). It is unfortunate that the majority does not deem these procedurally sound avenues of redress even worthy of mention.

III

Because each of the defendants' appeals is moot, it is irrelevant whether their appeals are treated as petitions for a writ of mandamus, as the majority does, Maj. op. at 11–14, or as appeals of collateral orders. In either case, we lack jurisdiction under Article III to consider their claims.

Nevertheless, even if this case were not moot, the defendants' appeals do not meet the requirements for granting a writ of supervisory mandamus, as the majority claims. Maj. op. at 13. Even when “the underlying proceeding is a criminal prosecution,” the writ may issue only when a district court has engaged in “willful disobedience of the rules laid down by” the Supreme Court, or “adopted a deliberate policy in open defiance of the federal rules.” Will v. United States, 389 U.S. 90, 96, 100, 102 (1967). Only such “exceptional circumstances amounting to a judicial ‘usurpation of power’ will justify the invocation of this extraordinary remedy.” Id. at 95.9

In this case, the district court has not defied a higher court or the federal rules of procedure. Rather, the district court complied with our last word on the matter, Howard, 480 F.3d at 1012–14, in which we held that restraining pretrial detainees in proceedings before a judge did not violate due process. The majority therefore oddly equates a good faith effort to follow our case law with “a persistent disregard” for our rulings. Will, 389 U.S. at 96. The majority attempts to distinguish Howard on the ground that the restraints in that case were not as intrusive as the restraints employed under the district court's policy now under review. Maj. op. at 14 n.6. No doubt the majority has detected a factual distinction between Howard and this case, but the district court's failure to anticipate such a distinction (which in any event does not appear to be constitutionally material) is a far cry from “willful disobedience” or “open defiance.” Will, 389 U.S. at 100, 102. As in Will, “the most that can be claimed on this record is that [the district court] may have erred in ruling on matters within [its] jurisdiction.” Id. at 103–04. The record here “simply fails to demonstrate the necessity for the drastic remedy employed by” the majority. Id. at 104.

IV

Because the individual appeals brought by Morales, Sanchez-Gomez, Patricio-Guzman, and Ring are moot, we should not rule on the merits of their arguments that pretrial detainees have a due process right to be free of bodily restraints in pretrial hearings before only a judge. Nevertheless, after proceeding to address the merits, the majority announces a new rule of constitutional criminal procedure that is contrary to Supreme Court precedent, creates a split with the Second and Eleventh Circuits, and puts trial courts throughout this circuit at risk. These errors warrant brief mention.10

A

The question presented on the merits is whether the Constitution precludes placing restraints on detainees during pretrial proceedings before a judge in the absence of a special need. The majority analyzes this question under Deck v. Missouri, in which the Supreme Court considered “whether shackling a convicted offender during the penalty phase of a capital case violates the Federal Constitution.” 544 U.S. 622, 624 (2005). Deck determined that a rule precluding the “routine use of visible shackles during the guilt phase” had “deep roots in the common law.” Id. at 626. In reaching this conclusion, Deck considered treatises on the common law, 18th century English cases, state and federal court opinions adhering to the common law rule, and the Court's own prior cases. Id. at 626–29. From these authorities, Deck concluded that the rule against using visible shackles before a jury was “a principle deeply embedded in the law” and enshrined in the protections of the Fifth and Fourteenth Amendments. Id. at 629. Ultimately, Deck held that “[t]he considerations that militate against the routine use of visible shackles during the guilt phase of a criminal trial apply with like force to penalty proceedings in capital cases.” Id. at 632. In light of a defendant's right to secure a meaningful defense, the need to maintain dignified proceedings, and the concern that visible restraints had the potential to prejudice the jury, the Court concluded that “courts cannot routinely place defendants in shackles or other physical restraints visible to the jury during the penalty phase of a capital proceeding.” Id. at 632–33.

If we apply Deck to the merits question here, we should begin by asking whether the common law rule identified in Deck extends to placing restraints on detainees during pretrial proceedings where there is no jury. Deck itself answers that question: “Blackstone and other English authorities recognized that the rule did not apply at ‘the time of arraignment,’ or like proceedings before the judge.” Id. at 626 (quoting 4 W. Blackstone, Commentaries on the Laws of England 317 (1769) and citing Trial of Christopher Layer, 16 How. St. Tr. 94, 99 (K.B. 1722) (Layer's Case)). Instead, Deck explained that the rule “was meant to protect defendants appearing at trial before a jury.” Id. (citing King v. Waite, 1 Leach 28, 36, 168 Eng. Rep. 117, 120 (K.B. 1743)). In other words, there is no rule regarding restraints on pretrial detainees in non-jury proceedings that has “deep roots in the common law.” Id.

The majority dismisses this conclusion as “undoubtedly dictum” and “contradicted by the very sources on which the Supreme Court relied.” Maj. op. at 27. These rationalizations do not hold water.

First, Deck's statement that the common law rule regulating shackling did not apply at arraignments is not mere dictum, as it responds to arguments raised by the dissent about the rule's scope and purpose. Justice Thomas's dissent argued that the purpose of the English common law rule against leaving a criminal defendant in irons for trial was to ensure that the defendant “was not so distracted by physical pain during his trial that he could not defend himself,” and accordingly modern restraints (which do not cause pain) “do not violate the principle animating the common-law rule.” Deck, 544 U.S. at 638, 640 (Thomas, J., dissenting). To support this point, Justice Thomas noted that because a defendant was not required to “play the main role in defending himself” at the arraignment, courts were not concerned about a defendant's being distracted by pain. Id. at 639–40 (Thomas, J., dissenting). Therefore, “the rule against shackling did not extend to arraignment.” Id. at 639 (Thomas, J., dissenting). In its analysis, the Deck majority conceded the dissent's historical point regarding shackling at arraignments, id. at 626 (majority opinion), but responded that although “[j]udicial hostility to shackling may once primarily have reflected concern for the suffering,” current opinions “have not stressed the need to prevent physical suffering,” but have looked at other legal principles, id. at 630. In light of this implicit give-and-take between the Deck majority and dissent, it is apparent that Deck's conclusion regarding shackling during arraignments is a considered concession of the dissent's historical point. Contrary to the majority, Maj. op. at 26 n.11, Deck's responsive historical analysis is part of its holding, as it bears on Deck's delineation of the scope of the common law rule that constitutes due process under the Constitution. But even if Deck's guidance were dicta, the majority's rejection of the Supreme Court's clear conclusion is contrary to our long established precedent that “Supreme Court dicta have a weight that is greater than ordinary judicial dicta” and therefore “we do not blandly shrug them off.” United States v. Montero-Camargo, 208 F.3d 1122, 1132 n.17 (9th Cir. 2000) (en banc) (internal quotation marks omitted).

Second, Deck's determination on this issue is not contradicted by the historical sources, as the majority seems to believe. Maj. op. at 26–31. In reaching its conclusion that the common law rule applied when the defendant was in the presence of the jury, but not “at ‘the time of arraignment,’ or like proceedings before the judge,” Deck undertook an in-depth historical analysis, considering Blackstone's Commentaries on the Laws of England, original sources setting forth the rule, see Layer's Case, 16 How. St. Tr. at 99; Waite, 1 Leach at 36, and state court cases recognizing the distinction that Blackstone drew, see Parker v. Territory, 5 Ariz. 283 (1898); People v. Harrington, 42 Cal. 165 (1871). Deck, 544 U.S. at 626–27. The majority claims that “Blackstone did not recognize that the rule against shackles didn't apply at the time of arraignment or proceedings before a judge,” but that “[s]hackles at arraignment and pretrial proceedings are acceptable only in situations of escape or danger.” Maj. op. at 28 (emphasis omitted). This is incorrect: Blackstone acknowledged a distinction between arraignment and trial made in Layer's Case. While Blackstone stated the general rule that a prisoner “must be brought to the bar without irons, or any manner of shackles or bonds,” he then observed that “in Layer's Case, A.D. 1722, a difference was taken between the time of arraignment, and the time of trial; and accordingly the prisoner stood at the bar in chains during the time of his arraignment.”11 5 William Blackstone & St. George Tucker, Blackstone's Commentaries with Notes of Reference 322 (1803).

Moreover, the text of Layer's Case better supports Blackstone's analysis. When announcing his decision to keep Layer fettered during his arraignment, the Lord Chief Justice first rejected Layer's reliance on Cranburne's Case for the proposition that restraints were not permitted at arraignments. Layer, 16 How. St. Tr. at 100. Instead, the Lord Chief Justice ruled that Cranburne's Case governed only those cases “when the party was called upon to plead, and was tried at the same time.” Id. The Lord Chief Justice then reasoned that the defendant should be free from chains when he comes to trial so he “should have the use of his reason, and all advantages to clear his innocence.” Id. In pretrial proceedings, however, “he is only called upon to plead by advice of his counsel” and is not to be tried, so there was no reason for “his chains to be taken off this minute, and to be put on again the next,” when he is returned to confinement. Id. at 100–01. This passage supports Blackstone's analysis, as well as that of the Deck majority and dissent; the concern was not with escape, but with the practicalities of removing restraints for a hearing of limited purpose and duration. See Deck, 544 U.S. at 626; id. at 639 n.2 (Thomas, J., dissenting) (“When arraignment and trial occurred on separate occasions, the defendant could be brought to his arraignment in irons.”).

After the decision in Layer's Case, the same rule was stated in King v. Waite, in which “[t]he prisoner, at the time of his arraignment, desired that his irons might be taken off.” 1 Leach 28, 36 (K.B. 1743). The court informed him, however, that it “had no authority for that purpose until the Jury were charged to try him.” Id. So the prisoner pleaded guilty, “and being put upon his trial, the Court immediately ordered his fetters to be knocked off.” Id.

As the common law developed in this country, state courts and treatises interpreted Layer's Case and other common law sources as Deck did, namely as distinguishing the use of restraints during an arraignment from their use during trial. In Lee v. State, for example, the Mississippi Supreme Court noted that Layer's Case and Waite's Case both distinguished between arraignment (where shackles were generally allowed) and trial (where shackles were not allowed except for good cause). 51 Miss. 566, 571 (1875). Lee interpreted the Lord Chief Justice's references to Layer's possible escape as relevant only to his decision to reject Layer's motion to have his restraints removed while in confinement. According to Lee, the Lord Chief Justice was concerned that granting such a motion “might be an excuse to his keeper if he (the prisoner) should escape.” Id. And Lee concluded that the Lord Chief Justice permitted shackling at arraignment because “it would be to no purpose to insist on [unfettering] for so little a time as the prisoner now had to stand at the bar.” Id. Other state courts similarly recognized the distinction between arraignment and trial. See, e.g., State v. Temple, 92 S.W. 869, 872 (Mo. 1906) (noting that in Layer's Case, “it was held that the prisoner might be brought ironed to the bar for arraignment, but that his shackles must be stricken off at the trial,” without reference to concerns regarding escape during proceedings); Rainey v. State, 20 Tex. App. 455, 472 (1886) (citing a treatise for the proposition that prisoners may not be shackled during trial, except in unusual cases, “[t]hough the rule at arraignment where only a plea is required is less strict”). Indeed, some state courts have interpreted Layer's Case as establishing a new common law rule, in contradistinction to a prior common law rule that defendants were generally not shackled at arraignment. See, e.g., Harrington, 42 Cal. at 167 (“[P]rior to 1722, when a prisoner was arraigned, or appeared at the bar of a Court to plead, he was presented without manacles or bonds, unless there was evident danger of his escape.”); Parker, 5 Ariz. at 287 (same).12

Rather than follow Deck, Blackstone, and these early state decisions, the majority provides its own interpretation of Layer's Case, arguing that the Lord Chief Justice held Layer in chains only because Layer had previously attempted to escape. Maj. op. at 28–29. As explained above, this is not a persuasive reading of the case.13 But even if the majority's interpretation of Layer's Case were also plausible, a reasonable difference in interpretations supports (at most) a conclusion that the case is ambiguous, and we should not ignore the Supreme Court's resolution of an ambiguous issue. Even less should we reprimand a district court through mandamus for failing to anticipate that we would do so.

Besides being ill-considered, the majority's decision to ignore Supreme Court direction also creates a circuit split, again contrary to our precedent. See United States v. Gwaltney, 790 F.2d 1378, 1388 n.4 (9th Cir. 1986) (“Unnecessary conflicts among the circuits are to be avoided.”); see also United States v. Alexander, 287 F.3d 811, 820 (9th Cir. 2002) (“[A]bsent a strong reason to do so, we will not create a direct conflict with other circuits.” (quoting United States v. Chavez-Vernaza, 844 F.2d 1368, 1374 (9th Cir. 1988))). In Zuber, the Second Circuit held that because juror bias “constitutes the paramount concern” in a physical restraint case, and because judges are assumed not to be prejudiced “by impermissible factors,” 118 F.3d at 104, it did not violate due process “for a trial judge (in the absence of the jury) to defer to the judgment of the U.S. Marshals Service without comment or extended colloquy” on the issue of restraints, id. at 103 n.2. Thus, the Second Circuit concluded that “the rule that courts may not permit a party to a jury trial to appear in court in physical restraints without first conducting an independent evaluation of the need for these restraints does not apply in the context of a non-jury sentencing hearing.” Id. at 102. Reaching a similar conclusion, the Eleventh Circuit, after reviewing Deck, Blackstone, and Layer's Case, held that “the rule against shackling pertains only to a jury trial” and “does not apply to a sentencing hearing before a district judge.” United States v. LaFond, 783 F.3d 1216, 1225 (11th Cir. 2015), cert. denied, 136 S. Ct. 213 (2015). The logic of both Zuber and LaFond is, as the majority recognizes, directly contrary to the rule announced today. Maj. op at 22–23 n.8.

Were we empowered to decide this case, we should join our sister circuits in following Deck's reading of the common law, rather than inventing a new right out of whole cloth. Deck establishes that there is no common law rule against the use of restraints during pretrial proceedings. 544 U.S. at 626. Moreover, as indicated in Zuber, there is no danger that the presumption of innocence or the dignity of the courtroom is undermined in the eyes of the jury when pretrial detainees appear in restraints before a judge. 118 F.3d at 103 n.2. Nor have the defendants here indicated that the restraints used in their cases interfered with their ability to communicate with their lawyers or participate in their own defenses. Deck, 544 U.S. at 631. The rule sought by the defendants has no pedigree, nor does it protect a well-established right. Accordingly, it cannot be “objectively, ‘deeply rooted in this Nation's history and tradition,’ ” Washington v. Glucksburg, 521 U.S. 702, 720–21 (1997) (quoting Moore v. City of East Cleveland, 431 U.S. 494, 503 (1977) (plurality opinion)), such that the Due Process Clause requires it, contra Maj. op. at 25. The majority's contrary conclusion grows not from the “deep roots” of the common law, Deck, 544 U.S. at 626, but from the majority's own hothouse.

B

Putting aside the majority's mistreatment of Deck, the appropriate framework for resolving this claim is provided by Bell v. Wolfish. In Bell, pretrial detainees brought a class action to challenge the conditions of their confinement at a federal pretrial detention center. 441 U.S. 520, 523 (1979). The district court granted sweeping relief, which the Second Circuit affirmed in large part. Id. at 523–24. In reviewing this relief, the Supreme Court set up the framework for analyzing constitutional claims by pretrial detainees challenging their conditions of confinement. Because Deck by its terms does not apply to the situation presented here, Deck, 544 U.S. at 626, we ought to apply the general framework for pretrial detention claims that Bell establishes.

Three of Bell's principles bear mentioning in this case. First, Bell teaches us that although “the presumption of innocence plays an important role in our criminal justice system[,] ․ it has no application to a determination of the rights of a pretrial detainee during confinement before his trial has even begun.” 441 U.S. at 533. Second, Bell instructs that pretrial detainment policies “are peculiarly within the province and professional expertise of corrections officials, and, in the absence of substantial evidence in the record to indicate that the officials have exaggerated their response to these considerations, courts should ordinarily defer to their expert judgment in such matters.” Id. at 548 (quoting Pell v. Procunier, 417 U.S. 817, 827 (1974)). This is so even where the officials are “ ‘experts' only by Act of Congress,” because pretrial detainment policies are “peculiarly the province of the Legislative and Executive Branches of our Government, not the Judicial.” Id. Finally, Bell holds that “[i]n evaluating the constitutionality of conditions or restrictions of pretrial detention that implicate only the protection against deprivation of liberty without due process of law, ․the proper inquiry is whether those conditions amount to punishment of the detainee.” Id. at 535. Because the government's authority to detain pending trial extends to its ability “to employ devices that are calculated to effectuate this detention,” id. at 537, when confronted with a particular challenged condition, the question for a court is “whether the disability is imposed for the purpose of punishment or whether it is but an incident of some other legitimate governmental purpose,” id. at 538. In the absence of an intent to punish, a pretrial condition of confinement is not a “punishment” if it is “reasonably related to a legitimate governmental objective.” Id. at 539. By contrast, where a condition is “arbitrary or purposeless,” a court may infer that the true purpose of the condition is to punish. Id.

The majority dismisses Bell as inapplicable because “Bell dealt with pretrial detention facilities, not courtrooms,” and detention facilities “are meant to restrain and keep order, not dispense justice.” Maj. op. at 31. The majority acknowledges that Bell may apply beyond the detention facility walls, see Maj. op. at 32 n.15, but draws a hard line at the courtroom door, see Maj. op. at 31 n.14. Certainly under Deck, a pretrial detainee has additional due process rights when appearing before a jury. But pretrial detainees enjoy no heightened interests when they appear in court outside of the presence of a jury. Cf. Zuber, 118 F.3d at 103–04 & n.2. The government's interest in securing their presence at trial and maintaining order and security, however, remains the same regardless of the location. Thus, as in Bell, the question is whether these interests justify the government's restriction on the liberty of pretrial detainees.

As explained in Bell, the government may restrain detainees to ensure they will be available for trial, 441 U.S. at 539, and may take certain steps necessary to “maintain security and order,” id. at 540. Bell's central lesson is that the reasonable pursuit of these objectives through restrictions on detainees' liberty interests, without more, does not rise to a constitutional violation. Id. at 539. This logic applies beyond the detention facility itself. For example, the government must often ensure that detainees appear at pretrial proceedings. See Fed. R. Crim. P. 10 (providing that a defendant must be physically present at arraignment absent an express waiver of his or her right to appear or express consent to video teleconferencing). But even when detainees are outside the walls of a particular detention facility, they are still subject to detention, and the government maintains a compelling interest in securing their ultimate presence for trial. Cf. Brothers v. Klevenhagen, 28 F.3d 452, 457 (5th Cir. 1994) (holding that pretrial detainee status “never reverts back” to a greater degree of protection “[u]ntil the detainee is released from custody”). Thus, pretrial detainees remain detained while they are in a vehicle transporting them to and from the courthouse, in a holding cell in the courthouse, in any outdoor areas, and even in the courtroom itself. Cf. Beaulieu v. Ludeman, 690 F.3d 1017, 1031–33 (8th Cir. 2012) (upholding under Bell a policy of placing civilly committed detainees in full restraints whenever being transported). In each area, the detainee is subject to reasonable government control aimed at securing his or her presence at trial and his or her orderly and safe interaction with other detainees.

Viewed in this light, the merits of this case would not be difficult, were we empowered to reach it. Because the pretrial detainees are outside the presence of a jury, the majority's rhetoric about the presumption of innocence, Maj. op. at 22, has no place in the analysis. Bell, 441 U.S. at 533. Moreover, because there is no allegation that the restraint policy is intended as a punishment, the question is simply whether requiring detainees to wear restraints while attending their pretrial hearings “is reasonably related to a legitimate governmental objective.” Id. at 539. Here, it clearly is. To the extent the restraints reduce the likelihood of an escape, they further the government's interest in ensuring that detainees will appear at trial. See id. Similarly, given the history of detainee-related assaults and weapons smuggling in the Southern District of California, the restraints are reasonably related to the government's interest in maintaining order and safety among its detainees. Cf. id. at 540 (“[T]he Government must be able to take steps to maintain security and order at the institution and make certain no weapons or illicit drugs reach detainees.”). Requiring detainees to appear at pretrial hearings in restraints is therefore reasonably related to the government's valid interests, and the policy is accordingly a constitutionally permissible condition of pretrial confinement. See id.

Making this case even simpler, the district court's deference to the Marshals Service, the entity that Congress statutorily charged with providing courtroom security, 28 U.S.C. § 566(a), is consistent with the Marshals Service's role as an expert entity charged with securing courtrooms and managing pretrial detainees during their court appearances. As the expert on courtroom security, the Marshals Service is due “wide-ranging deference” absent “substantial evidence in the record to indicate that the officials have exaggerated their response” to the problems they seek to solve. Bell, 441 U.S. at 547, 548. Because there is no substantial evidence on this record that the Marshals Service is punishing detainees by restraining them or otherwise imposing conditions of confinement unrelated to the government's legitimate interests, the challenged policy is not an unconstitutional condition of detention. Accordingly, the district court's deference to the Marshals Service's recommendation does not violate the pretrial detainees' constitutional rights.

To be sure, “district courts have the inherent authority to manage their ․ courtrooms,” Dietz v. Bouldin, 136 S. Ct. 1885, 1892 (2016), and some may choose not to defer to the Marshals Service's recommendation after a careful balancing of the need for safety and security of the courtrooms with the interests of the detainees. These are decisions, however, to be made by the district courts themselves, taking into account facts specific to their situations, including such factors as the adequacy of staffing by security professionals, the configurations of the courtrooms, and prior experiences. They are not decisions that should be made by appellate jurists far removed from the day-to-day administration of criminal justice.

By creating a blanket constitutional rule in this moot case, the majority not only puts federal district courts at risk, but also restricts the choices that states in this circuit can make to secure detainees without inviting a lawsuit under § 1983.14 The ramifications of the majority's holding will reach into courthouses of every size and capacity, yet the majority never once pauses to consider the consequences of its one-size-fits-all security decree. Indeed, the majority fails even to consider the evidence on this particular record that the Marshals Service is unable to make well-founded individual judgments about what threat, if any, a pretrial detainee poses. Instead, the majority lays down the rule that the Marshals Service can either do the impossible (predict risks based on a dearth of predictive information) or sit idly by and suffer an identifiable, compelling harm (violence in the courtroom). The majority's rule therefore fails not only as a matter of law, but also as a matter of common sense.

V

The majority's analysis is wrong at every turn. It contradicts the Supreme Court's rulings on mootness, mandamus, and the merits, and it substitutes the supposed wisdom of the ivory tower for the expertise of the United States Marshals Service and the district courts themselves. Because the four defendants whose criminal appeals are before us have now long since passed through the federal criminal justice system, we should dismiss these appeals as moot, rather than use them as improper vehicles to make a constitutional ruling as sweeping as it is erroneous. I dissent.

FOOTNOTES

FOOTNOTE.   This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.

1.   Several district judges avoid the term “policy” and instead claim it's just a practice. We don't see the difference.

2.   Evidence presented in a mandamus proceeding that transpired during the course of this appeal indicates that the Southern District's case load was increasing up to the year preceding the adoption of the routine shackling policy. But after 2012, case loads decreased and, as of 2015, had reached their lowest level in years.

3.   Defendants also appealed discovery and recusal decisions. We don't reach these issues.

4.   We noted in Howard that indigent defendants have little ability to bring civil class actions as a practical matter. 480 F.3d at 1010. They aren't guaranteed counsel to pursue civil rights claims, cf. Fed. R. Crim. P. 44, and defender organizations—like the Federal Defenders of San Diego—ordinarily have limited mandates that do not include filing class actions on behalf of their clients. See 18 U.S.C. § 3006A(g)(2).

5.   While these are criminal cases, they aren't subject to special criminal mandamus petition rules. The dissent discusses Will v. United States as though it narrowed the availability of the writ of mandamus in criminal cases. See dissent at 53–54. It didn't. The Supreme Court in Will explained that courts of appeals may resolve erroneous district court practices through mandamus petitions, even in criminal cases. 389 U.S. at 104–05 (discussing La Buy, 352 U.S. at 258). That's exactly what we do here. The Court also cautioned that mandamus petitions brought by the government in criminal cases raise concerns about speedy trials and double jeopardy. Id. at 96–98. None of those concerns are applicable here.

6.   The dissent faults us for “equat[ing] a good faith effort to follow our case law” with a clear and repeated error. Dissent at 54. According to the dissent, “the district court complied with our last word on the matter, Howard.” Id. at 54. The dissent errs. We explicitly noted in Howard that the policy we were addressing was “less restrictive than the previous policy requiring full restraints.” 480 F.3d at 1014. Nothing in Howard endorsed the routine use of full restraints.

7.   Contrary to the dissent's claim, dissent at 50–51, Oregon Advocacy Center was not just about associational standing. After determining that the plaintiffs had standing to bring the suit, we then turned to whether the case was moot—a distinct issue. Compare 322 F.3d at 1108–16 (discussing standing), with id. at 1116–18 (discussing mootness).

8.   The Second Circuit in a pre-Deck case, United States v. Zuber, did not recognize a right to individualized shackling determinations before a sentencing judge. 118 F.3d 101, 104 (2d Cir. 1997). But the court didn't hold that no liberty interest was at issue in nonjury courtroom shackling. Its analysis was limited to whether there would be inherent prejudice in the mind of the sentencing judge seeing the defendant in shackles as there would be in front of a guilt-phase jury. Id. at 103–04.Likewise, the Eleventh Circuit in United States v. LaFond held that a defendant wasn't entitled to an individualized shackling determination before a sentencing judge. 783 F.3d 1216, 1225 (11th Cir. 2015). The court in LaFond went further than Zuber, saying that “the rule against shackling pertains only to a jury trial.” Id. In reaching this conclusion, the Eleventh Circuit disregarded the common law rule embodied in our Constitution that protects an individual from unwarranted shackles in the courtroom, regardless of the presence of a jury. See infra pp. 25–31. Moreover, it failed to consider the three essential interests that Deck identified for deciding shackling cases.

9.   An individual determination cannot resemble what the Southern District judges did here. Courts may not incorporate by reference previous justifications in a general fashion, nor may they refuse to allow defendants to make objections or create evidentiary records. And they cannot flip the presumption against shackling by requiring that the defendant come up with reasons to be unshackled.The Southern District's reliance on postdeprivation process is unconstitutional not only because it often results in no opportunity to be heard at all, but also because many judges failed to exercise discretion when faced with inappropriate shackling. These judges shackled a blind man, a woman in a wheelchair with “dire and deteriorating” health and a woman with a broken wrist. And despite the policy providing that shackles wouldn't be used at sentencing hearings without specific security information showing an individualized need, the defendant in the wheelchair was also shackled at her sentencing hearing. See supra p. 7. The hearing transcript indicates that no evidence of such specific security information was introduced. Routine shackling subject to postdeprivation review is plainly insufficient to protect this fundamental constitutional right.

10.   We therefore overrule Howard to the extent it held that a routine shackling policy largely justified by deference to the U.S. Marshals Service was constitutional.

11.   The dissent relies heavily on the give and take between Justice Thomas and the majority on a matter not central to the majority's holding. See dissent at 57–58. But the Court has recognized that such byplay is not binding if it does not concern the majority's holding. Kirtsaeng v. John Wiley & Sons, Inc., 133 S. Ct. 1351, 1368 (2013) (dismissing as dictum a contrary statement of law in a previous opinion, explaining that it was merely “contained in a rebuttal to a counterargument”).

12.   “Is the Court having once written dicta calling a tomato a vegetable bound to deny that it is a fruit forever after?” Kirtsaeng, 133 S. Ct. at 1368.

13.   The dissent fails to engage with these cases and cites no secondary sources with the view of shackling at arraignment that it espouses. Authoritative secondary sources such as Bishop's treatise and Mack's encyclopedia provide us with a panorama of the law as it was generally understood and applied by a majority of courts at the time.

14.   The dissent expands the scope of Bell to the courtroom by claiming that “[t]he government's interest in securing [pretrial detainees'] presence at trial and maintaining order and security ․ remains the same regardless of the location.” Dissent at 65. Location matters, however. The courtroom is not a pretrial detention facility.

15.   We need not consider the application of Bell to holding cells or transportation between detention centers and the courtroom, which are beyond the scope of this case.

1.   While Morales could appear in federal court again on a supervised release violation, she would not appear as a pretrial detainee. Rather, she is now “[a] criminal defendant proved guilty” who “does not have the same liberty interests as a free man.” Dist. Atty's Office v. Osborne, 557 U.S. 52, 68 (2009).

2.   Sanchez-Gomez was charged with felony misuse of a passport in violation of 18 U.S.C. § 1544. He filed an emergency motion (identical to Morales's motion) challenging the restraint policy. The district court denied the motion, and Sanchez-Gomez filed a notice of appeal on November 22, 2013. By December 2013, Sanchez-Gomez had pleaded guilty to the charge; the district court entered a final judgment and imposed five years of probation. Patricio-Guzman filed an identical emergency motion challenging the restraint policy. It was also denied. He pleaded guilty to misdemeanor illegal entry into the United States, 8 U.S.C. § 1325, and was sentenced to thirty days of imprisonment. Final judgment was entered in his case weeks before he filed a notice of appeal regarding the denial of his motion to have restraints removed during pretrial proceedings. Ring was charged with making an interstate threat in violation of 18 U.S.C. § 875(c). His challenge to the use of pretrial restraints was also denied in November 2013, and he filed an appeal a week later. The district court dismissed the charges against Ring with prejudice on the government's motion in October 2014. We consolidated the appeals brought by each of the defendants.

3.   Genesis Healthcare explained that courts adopted class action terminology, such as “conditional certification,” in the FLSA context “to describe the process of joining co-plaintiffs under 29 U.S.C. § 216(b).” 133 S. Ct. at 1527 n.1.

4.   Because Genesis Healthcare rejected the plaintiff's argument on this ground, it provides no support for the majority's extension of the relation back doctrine to a criminal defendant's claim within his or her criminal case. Contrary to the majority's suggestion that the Court's silence equals permission, Maj. op. at 18–19, Genesis Healthcare consistently refers to the relation back doctrine as applying only to class actions. 133 S. Ct. at 1530–31.

5.   Contrary to the majority's argument, our “supervisory mandamus role” does not give us any authority to address moot claims. Maj. op. at 19. As the majority itself acknowledges, “[s]upervisory mandamus cases require live controversies even when we don't order a lower court to take or refrain from a specific action.” Maj. op. at 15 (citing In re United States, 791 F.3d 945, 952 (9th Cir. 2015)). Indeed, each supervisory mandamus case that the majority cites involved a live Article III case or controversy. Will v. United States, 389 U.S. 90, 91 (1967) (ongoing criminal prosecution for tax evasion); Schlagenhauf v. Holder, 379 U.S. 104, 106–09 (1964) (ongoing diversity personal injury suit); La Buy v. Howes Leather Co., 352 U.S. 249, 251–54 (1957) (ongoing antitrust suit). It is true that the Supreme Court's decisions in these cases constituted binding precedent and therefore affected persons who were not before the court; this is the nature of our federal judicial system. But an opinion that “addresses the harm of a district court policy affecting a huge class of persons that aren't parties to the mandamus petition” in a moot case is just an advisory opinion, Maj. op. at 19, and “[t]he federal courts established pursuant to Article III of the Constitution do not render advisory opinions,” Golden v. Zwickler, 394 U.S. 103, 108 (1969) (quoting United Public Workers of Am. (C.I.O.) v. Mitchell, 330 U.S. 75, 89 (1947)).

6.   Howard's reliance on a case from the D.C. Circuit for the same proposition is equally mistaken. See Howard, 480 F.3d at 1010 (citing Ukrainian-Am. Bar Ass'n v. Baker, 893 F.2d 1374, 1377 (D.C. Cir. 1990)). Like Oregon Advocacy Center, Ukrainian-American Bar Ass'n was a case in which an organization was suing in its own name to challenge an on-going government policy. 893 F.2d at 1376–77. For that reason, Ukrainian-American Bar Ass'n was inapposite in Howard, as it is here.

7.   Howard's treatment of the capable of repetition, yet evading review exception has been rightly criticized elsewhere. See Milwaukee Police Ass'n v. Bd. of Fire & Police Comm'rs, 708 F.3d 921, 932 (7th Cir. 2013) (noting that Howard and cases following it “shoehorned ongoing policy challenges” into the capable of repetition exception even though “the parties would not otherwise qualify for the exception as articulated doctrinally”). The en banc court should have used this case as a vehicle to overrule Howard's error, not to entrench it further.

8.   This should not be read to suggest, however, either that a Bivens remedy would ultimately be appropriate, or that the government defendants would be unable to avail themselves of qualified immunity or other defenses.

9.   The majority suggests that because Will raised “concerns about speedy trials and double jeopardy” that are not present here, the mandamus principles discussed in Will are not applicable. Maj. op. at 14 n.5. But Will merely applied the Supreme Court's general mandamus principles that are applicable in civil and criminal cases alike. See, e.g., La Buy, 352 U.S. at 257–58 (holding that supervisory mandamus “should be resorted to only in extreme cases” such as where “the Court of Appeals has for years admonished the trial judges” not to engage in a particular practice).

10.   Judge Schroeder's concurrence faults the analysis that follows as “lack[ing] sensitivity to two of the most important components of our system of justice,” the dignity of court proceedings and the proper role of judges in managing their courtrooms. Concurrence at 34. It is the majority, however, that “lacks sensitivity to the proper role of ․ judges” in our constitutional system, id., by contravening the “minimum constitutional mandate” that “[t]he Art. III judicial power exists only to redress or otherwise to protect against injury to the complaining party, even though the court's judgment may benefit others collaterally,” Warth v. Seldin, 422 U.S. 490, 499 (1975).

11.   William Hawkins noted this same distinction, also in reliance on Layer's Case. 2 William Hawkins, A Treatise on the Pleas of the Crown 437 (1787).

12.   The majority makes the strange accusation that this analysis of state court cases is flawed because it “cites no secondary sources.” Maj. op. at 30 n.13. The primary sources cited here, however—actual judicial opinions—read Layer's Case as Blackstone and Deck read them. If secondary sources have derived a different rule, this again suggests, at most, that the common law is ambiguous. It is precisely because of this ambiguity that we should follow the Supreme Court's interpretation in Deck, rather than adopt a contrary view that the Court has rejected.

13.   The majority contends that my interpretation of Layer's Case “struggles manfully against the plain language of Layer's case and Blackstone.” Maj. op. at 29. Rather than struggling—manfully or otherwise—with Layer's Case, I would merely follow the Supreme Court's interpretation of Layer's Case, which is well supported by the text and relevant primary and secondary sources. As noted above, the Court relied on Layer's Case for the proposition that “Blackstone and other English authorities recognized that the rule [against shackling] did not apply at ‘the time of arraignment,’ or like proceedings before the judge.” Deck, 544 U.S. at 626. It is the majority that struggles to bypass the Supreme Court's considered statement.

14.   State courtrooms may face even greater dangers than federal courthouses. “Federal judges are protected by a dedicated law enforcement agency, the U.S. Marshals Service,” but “[m]ost state and local judges are protected by all-purpose local sheriff or police departments.” Chuck Weller, What Judges Should Know about Court-Related Violence, 53 Judges' J. 28, 30 (2014). Therefore, “[f]ew state and local judges will ever have the level of protection afforded to their federal counterparts.” Id. Indeed, a mere matter of months ago, a pretrial detainee in Michigan who was handcuffed, but not secured with a belt apparatus that the majority maligns, Maj. op. at 7, managed to disarm a sheriff's deputy, kill two bailiffs, shoot a bystander in the arm, and take hostages. See Associated Press, Sheriff: Inmate who killed 2 at Michigan courthouse was handcuffed, Chicago Tribune (July 12, 2016), available at http://www.chicagotribune.com/news/nationworld/midwest/ct-michigan-courthouse-shooting-20160712-story.html.

KOZINSKI, Circuit Judge:

Opinion by Judge Kozinski; Concurrence by Judge Schroeder; Dissent by Judge Ikuta

FindLaw Career Center

    Select a Job Title


      Post a Job  |  Careers Home

    View More