SEIU LOCAL 121RN, Plaintiff-Appellee, v. LOS ROBLES REGIONAL MEDICAL CENTER, DBA Los Robles Hospital and Medical Center, Defendant-Appellant.
This case presents the question of whether the arbitrability of an issue is itself arbitrable, where the relevant agreement includes a broad arbitration clause but is otherwise silent on the question. Over a quarter-century ago, the Supreme Court established that a court must make that determination when the agreement does not specifically address the “who” question. See First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938, 944–45, 115 S.Ct. 1920, 131 L.Ed.2d 985 (1995). That instruction notwithstanding, this court subsequently stated that labor cases are different, and in those cases, an arbitrator should decide arbitrability as long as the agreement includes a broad arbitration clause. See United Bhd. of Carpenters & Joiners of Am., Local No. 1780 v. Desert Palace, Inc., 94 F.3d 1308, 1309 (9th Cir. 1996). The Desert Palace court reached its conclusion by distinguishing collective bargaining 1 disputes (at issue in Desert Palace) from commercial arbitration disputes (at issue in First Options) on policy grounds, and thus opted not to apply First Options. Id. at 1310–12.
After Desert Palace, the Supreme Court expressly rejected the notion that labor arbitration disputes should be analyzed differently than commercial arbitration disputes. See Granite Rock Co. v. Int'l Bhd. of Teamsters, 561 U.S. 287, 300–01, 130 S.Ct. 2847, 177 L.Ed.2d 567 (2010). The Court disagreed that “courts may use policy considerations as a substitute for party agreement.” Id. at 303, 130 S.Ct. 2847. Since then, our court has not meaningfully considered whether Desert Palace’s rationale holds after Granite Rock. We do so now. Because the rationale in Desert Palace is “clearly irreconcilable with the reasoning or theory of intervening higher authority” (i.e., Granite Rock), we are not bound by Desert Palace. Miller v. Gammie, 335 F.3d 889, 893 (9th Cir. 2003) (en banc). We therefore reverse and remand for further proceedings.
Los Robles Regional Medical Center (the “Hospital”) entered into a collective bargaining agreement (“CBA”) with SEIU Local 121RN (“SEIU”), who represented registered nurses working at the Hospital. The CBA established the terms and conditions of employment for these nurses and was effective from September 16, 2014 through September 15, 2017, subject to certain renewal provisions. Article 38 of the CBA provided for a three-step procedure to address grievances, with the final step resulting in arbitration. A “grievance” subject to Article 38 was defined as “a dispute or disagreement involving the interpretation, application or compliance with specific provisions of this Agreement (including Article and Section) or a dispute or disagreement concerning whether or not discipline including discharge was for just cause.” But an arbitrator was not allowed to “add to or subtract from or to modify the terms of [the CBA] or to arbitrate any matter after [the CBA] has expired[,] other than matters which arose prior to the time of the expiration of [the CBA], or to hear any dispute on any claim that has been asserted in state or federal court or other governmental adjudicatory forum.” Other articles of the CBA exempted certain disputes from the grievance procedures, including health and safety issues under Article 22 and certain staffing and workload issues under Article 25.
On September 7, 2017, SEIU filed a grievance asserting that the Hospital placed certain types of patients with nurses who did not have the appropriate training to care for those patients. The grievance also accused the Hospital of violating nurse-to-patient ratios established by state law. SEIU claimed that these practices violated Articles 22 (“Safety”), 39 (“Job Descriptions”), and 64 (“In-Service Education”) of the CBA, along with other provisions of the CBA and state and federal laws. The Hospital and SEIU were unable to resolve the grievance among themselves, so SEIU notified the Hospital on November 13, 2017 that it was pursuing arbitration. The Hospital responded that the grievance appeared to be a staffing issue covered under Article 25 of the CBA and thus was not arbitrable. After a series of further communications between the parties regarding the arbitrability issue, the Hospital sent an email “confirm[ing] we continue to disagree that this matter is substantively arbitrable AND disagree that an arbitrator has the authority to decide that issue.”
SEIU filed a complaint in the district court on May 10, 2018 and subsequently filed a motion to compel arbitration, which the district court granted on January 15, 2019. The court identified two main questions raised by the motion—first, whether the parties were bound by the arbitration provision in the CBA, and second, whether the grievance fell within the scope of that provision. Acknowledging no argument to the contrary, the district court determined that the parties were bound by the arbitration provision, and thus it needed to address only the second question. But as part of that second inquiry, the district court explained that it had to first determine whether the arbitration provision gave an arbitrator—rather than the court—the authority to decide its own jurisdiction (i.e., to decide if the grievance was subject to the arbitration provision).
The district court engaged in a thorough analysis of Supreme Court and Ninth Circuit precedent on the issue and, although it “believe[d] that the Ninth Circuit's reasoning [in Desert Palace] no longer controls in light of more recent Supreme Court decisions,” the district court concluded that it could not overrule Desert Palace. Applying Desert Palace, the district court found that the arbitration provision in the CBA was broad enough to authorize the arbitrator—rather than the court—to determine whether the grievance was arbitrable and therefore granted SEIU's motion to compel arbitration without reaching the question of whether the grievance was in fact arbitrable. The Hospital filed a timely notice of appeal. We have jurisdiction under 28 U.S.C. § 1291.
We review de novo the district court's decision to compel arbitration. Bushley v. Credit Suisse First Boston, 360 F.3d 1149, 1152 (9th Cir. 2004); Chiron Corp. v. Ortho Diagnostic Sys., 207 F.3d 1126, 1130 (9th Cir. 2000).
This case implicates three types of disputes: (1) the “Merits Question”—a dispute between the parties regarding the merits of an issue (e.g., whether the Hospital's conduct, as set out in the grievance, violates the CBA); (2) the “Arbitrability Question”—a dispute regarding whether the parties agreed to arbitrate the Merits Question (e.g., whether the arbitration provision in the CBA requires the Hospital and SEIU to arbitrate the grievance); and (3) the “Delegation Question”—a dispute regarding whether an arbitrator or a court is tasked with deciding the Arbitrability Question.2 Resolution of the main issue in this appeal—the Delegation Question—largely depends on the interplay of three different cases: the Supreme Court's decision in First Options, the Ninth Circuit's subsequent decision in Desert Palace, and the Supreme Court's even later decision in Granite Rock. We will briefly describe each in turn.
In First Options, the Supreme Court considered whether, under the Federal Arbitration Act, the arbitrator or a court “has the primary power to decide arbitrability”—i.e., the Delegation Question. 514 U.S. at 940, 943, 115 S.Ct. 1920. The Court reasoned that “[j]ust as the arbitrability of the merits of a dispute [i.e., the Arbitrability Question] depends upon whether the parties agreed to arbitrate that dispute, so the [Delegation Question] ․ turns upon what the parties agreed about that matter.” Id. at 943, 115 S.Ct. 1920 (internal citations omitted). This is because “arbitration is simply a matter of contract between the parties; it is a way to resolve those disputes—but only those disputes—that the parties have agreed to submit to arbitration.” Id.
The Court in First Options explained that, when answering the Delegation Question, courts “should not assume that the parties agreed to arbitrate arbitrability unless there is clear and unmistakable evidence that they did so.” Id. at 944, 115 S.Ct. 1920 (emphasis added) (internal citations, quotation marks, and alteration marks omitted). In fact, the presumption in favor of arbitration, which applies when a valid arbitration agreement is silent or ambiguous as to the arbitrability of a particular dispute, does not apply to “silence or ambiguity” with respect to the Delegation Question. Id. at 944–45, 115 S.Ct. 1920. Instead, “the law reverses the presumption.” Id. at 945, 115 S.Ct. 1920. The Court noted that this difference in presumptions follows from the fact that the Delegation Question is “rather arcane” and might not be a focus of negotiation and, thus, courts should not “force unwilling parties to arbitrate a matter they reasonably would have thought a judge, not an arbitrator, would decide.” Id. The Court ultimately concluded that the court was presumptively the appropriate entity to review the question of arbitrability because the parties “did not clearly agree to submit the question of arbitrability to arbitration.” Id. at 947, 115 S.Ct. 1920.
Just one year later, the Ninth Circuit considered the Delegation Question in the context of a collective bargaining agreement between a union and an employer. Desert Palace, 94 F.3d at 1309. The Desert Palace court initially acknowledged that courts, not arbitrators, should determine the arbitrability of an issue unless “the parties clearly and unmistakably provide otherwise.” Id. at 1310 (citation omitted). But it went on to cite a series of pre-First Options decisions from this circuit for the proposition that construing a broad arbitration clause as conferring authority on the arbitrator to decide arbitrability “is an act of legitimate contract interpretation.” Id. (citation and quotation marks omitted). The court therefore concluded that the broad arbitration clause at issue in Desert Palace, which covered matters of interpretation and application of the agreement and did not explicitly exclude arbitrability or jurisdictional issues, evinced the parties’ agreement to have an arbitrator decide arbitrability. Id.
The Desert Palace court asserted that its “conclusion [wa]s not affected by the Supreme Court's recent decision in First Options” because the policy concerns and other considerations driving the analysis in First Options were specific to the commercial arbitration context and did not apply to the collective bargaining context at issue in Desert Palace. Id. at 1310–11 (“These policy concerns [described in First Options], however, do not come into play in the collective bargaining context” and “unlike parties in the commercial arbitration context, parties entering into a collective bargaining agreement know they are granting the arbitrator tremendous power ․.” (emphases added)).3 Relying on the policy differences between commercial and labor arbitration, the Desert Palace court determined that First Options did not specifically speak to “whether a broad arbitration clause in a collective bargaining agreement grants the arbitrator the power to decide his or her own jurisdiction” and thus “decline[d] to extend First Options to the collective bargaining context.” Id. at 1311 (emphasis added).
Nearly fifteen years later, the Supreme Court in Granite Rock discussed arbitration in both the commercial and labor contexts when it “reemphasize[d] the proper framework for deciding when disputes are arbitrable under [its] precedents.” 561 U.S. at 297, 130 S.Ct. 2847. Granite Rock involved a dispute between an employer and a union regarding when their collective bargaining agreement was ratified and whether the court or an arbitrator had the authority to resolve that dispute. Id. at 291–92, 130 S.Ct. 2847. The Court began by explaining that the Delegation Question presumption applies in both the commercial and labor contexts: “It is well settled in both commercial and labor cases that whether parties have agreed to submit a particular dispute to arbitration is typically an issue for judicial determination.” Id. at 296, 130 S.Ct. 2847 (citations, quotation marks, and alteration marks omitted). It emphasized—repeatedly—that the analysis for determining whether a dispute is arbitrable in the labor context is the exact same analysis applied in the commercial context. See id. at 301, 130 S.Ct. 2847 (explaining “[o]ur cases invoking the federal policy favoring arbitration of commercial and labor disputes apply the same framework” (emphasis added) (quotation marks omitted)).4
The Court explicitly rejected the argument that its precedents required that labor disputes be arbitrated “based on policy grounds even where evidence of the parties’ agreement to arbitrate the dispute in question [was] lacking.” Id. at 300, 130 S.Ct. 2847. The Court responded: “we have never held that this policy [favoring arbitration] overrides the principle that a court may submit to arbitration only those disputes that the parties have agreed to submit. Nor have we held that courts may use policy considerations as a substitute for party agreement.” Id. at 302–03, 130 S.Ct. 2847 (emphasis added) (internal citations, quotation marks, and alteration marks omitted).
The Supreme Court's emphasis in Granite Rock on applying the same arbitrability framework to both commercial and labor disputes—combined with its conclusion that the policy favoring arbitration of labor disputes cannot substitute for party agreement—is clearly irreconcilable with the reasons that this court in Desert Palace relied on to distinguish First Options. SEIU contends, however, that Granite Rock answered a completely different question than the question addressed in Desert Palace (i.e., the Delegation Question). The Court in Granite Rock specifically did not reach the Delegation Question, argues SEIU, based on a footnote stating that “[b]ecause neither party argues that the arbitrator should decide [whether the ratification dispute is arbitrable], there is no need to apply the rule requiring ‘clear and unmistakable’ evidence of an agreement to arbitrate arbitrability.” Granite Rock, 561 U.S. at 297 n.5, 130 S.Ct. 2847 (quoting First Options, 514 U.S. at 944, 115 S.Ct. 1920). To that end, SEIU asserts that Desert Palace is not inconsistent with Granite Rock.
But SEIU misses the problem that Granite Rock creates for Desert Palace. “[T]he issues decided by the higher court need not be identical in order to be controlling” over a prior circuit decision; rather, the Supreme Court “must have undercut the theory or reasoning underlying the prior circuit precedent in such a way that the cases are clearly irreconcilable.” Miller, 335 F.3d at 900 (emphases added). Even though Granite Rock dealt with a formation-related Arbitrability Question and Desert Palace dealt with a Delegation Question, these questions ultimately derive from the same inquiry—what did the parties agree to have the arbitrator decide?5 Because of this shared genesis, the Supreme Court in Granite Rock and the Ninth Circuit in Desert Palace unsurprisingly identified the same initial step for their analyses. The Court in Granite Rock explained that:
[Our precedents] recognize that, except where the parties clearly and unmistakably provide otherwise, it is the court's duty to interpret the agreement and to determine whether the parties intended to arbitrate grievances concerning a particular matter.
561 U.S. at 301, 130 S.Ct. 2847 (emphasis added) (internal citations and quotation marks omitted). The Desert Palace court likewise asserted that:
Generally, the question of whether a dispute is arbitrable is decided by the courts. However, where the parties clearly and unmistakably provide otherwise, the courts will be divested of their authority and an arbitrator will decide in the first instance whether a dispute is arbitrable.
94 F.3d at 1310 (emphasis added) (internal citations and quotation marks omitted). But what precedents was the Supreme Court referring to in Granite Rock in its statement quoted above? The Court tells us: “[o]ur cases [involving] ․ commercial and labor disputes.” Granite Rock, 561 U.S. at 301, 130 S.Ct. 2847 (emphasis added). Because the Court interpreted both its commercial and labor arbitration cases as requiring the same framework to determine whether a dispute is arbitrable, and the initial step of such framework is the same standard as that applied to Delegation Questions, it follows that a Delegation Question must be analyzed the same way in both commercial and labor contexts. That is not what Desert Palace said. Indeed, Desert Palace reached the opposite conclusion, distinguishing First Options on the basis that the “policy concerns” underlying the Supreme Court's decision in First Options “do not come into play in the collective bargaining context.” Desert Palace, 94 F.3d at 1311. Granite Rock forcefully (and repeatedly) put the lie to any such difference.6
The footnote addressing the Delegation Question in Granite Rock further undermines SEIU's attempt to distance that case. That footnote reinforces that “clear and unmistakable” evidence would be required to authorize an arbitrator to decide if a dispute is arbitrable, even in the collective bargaining context at issue in Granite Rock. See Granite Rock, 561 U.S. at 297 n.5, 130 S.Ct. 2847. Importantly, the source cited for this proposition is none other than First Options. Id. This citation indicates the Court's view that First Options, despite involving a commercial dispute, applies in the collective bargaining context. Again, this logic cannot be reconciled with this court's attempt in Desert Palace to differentiate a labor dispute from the commercial dispute in First Options.
SEIU separately contends that Desert Palace is distinguishable from First Options because the two courts faced different types of “silence or ambiguity” with respect to the Delegation Question. The “silence or ambiguity” at issue in First Options arose from the fact that the parties were not governed by the language in an arbitration clause—instead, the party seeking arbitration argued that the resisting party agreed to arbitrate by filing objections at the arbitration proceedings. First Options, 514 U.S. at 941, 945–46, 115 S.Ct. 1920. The parties in Desert Palace, by contrast, were governed by an arbitration clause, and the dispute in that case related to whether the clause covered the Delegation Question. Desert Palace, 94 F.3d at 1310. SEIU argues that this difference—in particular, the fact that the Supreme Court had no arbitration clause to interpret in First Options—was the reason for the Desert Palace court's observation that “First Options did not address the question whether a broad arbitration clause in a collective bargaining agreement grants the arbitrator the power to decide his or her own jurisdiction.”
But this argument, based on one sentence plucked from its context, also founders on the text of Desert Palace when read as a whole. In the discussion immediately following the statement quoted above, the Desert Palace court did not point to factual differences between the two cases; it instead pointed to the Supreme Court's acknowledgment in First Options that there was “no strong arbitration-related policy favoring First Options,” and contrasted that with the “collective bargaining context, where there is a strong federal policy favoring arbitration of labor disputes.” Desert Palace, 94 F.3d at 1311–12 (emphasis in original). The entirety of the Desert Palace opinion turns on one repeated argument: that policy differences justify different application of the arbitrability framework in commercial versus labor disputes—precisely the argument that Granite Rock rejected.7
Since Granite Rock, none of our decisions have squarely grappled with the fundamental inconsistencies between Desert Palace and Granite Rock. Only one subsequent Ninth Circuit majority opinion has even cited to Desert Palace. See Kramer v. Toyota Motor Corp., 705 F.3d 1122, 1127 (9th Cir. 2013). But Kramer merely cited Desert Palace for the general proposition that “clear and unmistakable evidence” of the parties’ agreement is required to “arbitrate arbitrability.” Id. Kramer says nothing about Desert Palace’s conclusion that, notwithstanding First Options, because of unique policy considerations in the labor context, a broad arbitration agreement in that context can itself constitute “clear and unmistakable evidence” of the parties’ agreement to arbitrate arbitrability. Because Kramer involved a commercial arbitration dispute, not a labor dispute, it is not surprising that its passing citation to Desert Palace had nothing to do with whether Desert Palace’s cordoning off of labor cases survives Granite Rock. See Head v. Wilkie, 936 F.3d 1007, 1012–13 (9th Cir. 2019) (acknowledging Supreme Court precedent overruled two prior Ninth Circuit cases despite continued citation of the overruled cases because none of the subsequent citations “expressly addressed the[ir] continuing viability” (internal citations omitted)); cf. In re Magnacom Wireless, LLC, 503 F.3d 984, 993–94 (9th Cir. 2007) (“In our circuit, statements made in passing, without analysis, are not binding precedent.”); Estate of Magnin v. C.I.R., 184 F.3d 1074, 1077 (9th Cir. 1999) (“When a case assumes a point without discussion, the case does not bind future panels.”).
The only meaningful discussion from this court of the interplay between Granite Rock and Desert Palace appears in Judge Ikuta's dissent in ASARCO LLC v. United Steel, Paper & Forestry, Rubber, Manufacturing, Energy, Allied Industrial & Service Workers International Union, where the majority opinion never addressed the issue. 910 F.3d 485, 496–98 (9th Cir. 2018) (Ikuta, J., dissenting). Judge Ikuta determined that Desert Palace distinguished itself from First Options on the basis that “First Option’s holding applied only in the commercial context, not in the collective bargaining context, where there is a strong federal policy favoring arbitration of labor disputes.” Id. at 498 (citations and quotation marks omitted). She concluded, as the court now does, that “Granite Rock superseded Desert Palace.” Id.
SEIU counters that the Ninth Circuit applied the “broad arbitration clause” rule post-Granite Rock in International Alliance of Theatrical Stage Employee & Moving Picture Technicians, Artists, & Allied Crafts v. InSync Show Productions, Inc., 801 F.3d 1033 (9th Cir. 2015). Because, according to SEIU, the current case is “indistinguishable” from InSync, concluding that a court should determine arbitrability in this case would improperly overturn InSync without intervening Supreme Court precedent.
It is unclear whether SEIU argues that the continued application of the “broad arbitration clause” rule indicates the continuing vitality of Desert Palace, or that SEIU's case is separately governed by InSync even if Desert Palace is no longer good law. With respect to the first argument, InSync does not discuss Desert Palace. It does not cite to the case even once. InSync is also completely silent as to the alleged differences between commercial and labor arbitration. The court in InSync does indicate that a broad arbitration clause can constitute “clear and unmistakable” evidence of an arbitrator's authority to determine arbitrability, but it makes such acknowledgment in a conclusory statement in a footnote. InSync, 801 F.3d at 1043 n.7 (stating “[b]ecause the scope of the arbitration provision broadly covers claims or allegations of employer violations of the agreement, this case is not one in which the arbitration provision's scope is too narrow to ‘clearly and unmistakably’ leave questions of arbitrability for determination by the arbitrator” (internal citations and quotation marks omitted)). Again, concepts only referenced in passing, without analysis, in a prior opinion do not have precedential value. In re Magnacom Wireless, LLC, 503 F.3d at 993–94.
SEIU also could be arguing that, regardless of the fate of Desert Palace, this case should be governed by InSync’s holding that a broad arbitration clause was sufficient to authorize the arbitrator to determine the parties’ dispute over the termination clause in their collective bargaining agreement. But the arbitrability of a termination clause is obviously different than the arbitrability of the arbitration requirement itself. See, e.g., First Options, 514 U.S. at 944–45, 115 S.Ct. 1920 (“[T]he law treats ․ the question ‘who (primarily) should decide arbitrability’ differently from the way it treats ․ the question ‘whether a particular merits-related dispute is arbitrable ․’—for in respect to this latter question the law reverses the presumption.”). Although the InSync court initially recited general principles of arbitration, it determined that Brotherhood of Teamsters & Auto Truck Drivers Local No. 70 v. Interstate Distributor Co., 832 F.2d 507 (9th Cir. 1987), governed its analysis because “Interstate Distributor explained how to approach a case in which the parties to a collective bargaining agreement disagree about the proper meaning or interpretation of a termination clause in their agreement, and whether an arbitrator should decide that question.” InSync, 801 F.3d at 1041–42 (emphasis added) (internal citations, quotation marks, and alteration marks omitted). The InSync court repeatedly stated that it was applying principles specific to addressing the arbitrability of expiration and termination disputes, characterizing such disputes as “one step removed from the issue of substantive arbitrability.” Id. at 1042 (describing the proper inquiry set out in Interstate Distributor for cases, like InSync, where “a dispute exists over whether a contract with an arbitration clause has expired or been terminated” (emphasis added)).8 The InSync court explicitly distinguished its case—a case resolving a termination dispute—from cases that address disputes involving contract formation or the scope of an existing arbitration provision, such as in Granite Rock. Id. at 1044 n.9 (“Because this appeal involves the parties’ dispute relating to expiration or termination of their CBA, cases involving contract formation and disputes regarding the scope of an arbitration provision are distinct.”). The case before us does not involve a dispute over the termination or expiration of the CBA, but rather whether the arbitration provision delegates the arbitrability question to the arbitrator. InSync therefore cannot control.
This court has, in fact, already recognized the incompatibility of the rationale in Desert Palace with that in Granite Rock, albeit indirectly so. In Pacesetter Construction Co. v. Carpenters 46 Northern California Counties Conference Board, this court addressed the question of whether an arbitration panel had the authority to determine the arbitrability of a dispute between an employer and a union. 116 F.3d 436, 437 (9th Cir. 1997), cert. denied, 522 U.S. 1014, 118 S.Ct. 599, 139 L.Ed.2d 488 (1997), abrogated by Granite Rock, 561 U.S. 287, 130 S.Ct. 2847, 177 L.Ed.2d 567 (2010), as recognized in Local Joint Exec. Bd. v. Mirage Casino-Hotel, Inc., 911 F.3d 588 (9th Cir. 2018). The underlying dispute in Pacesetter related to whether the agreement between the parties had been repudiated. Id. This court concluded that submitting the repudiation dispute to the arbitration panel, without also challenging the panel's jurisdiction, constituted implicit consent to having the panel arbitrate the latter issue. Id. at 438–39. But the Supreme Court vacated and remanded that decision for reconsideration in light of First Options. Id. at 438.
On remand, the employer argued that First Options controlled the case and invalidated the court's prior holding, but the Pacesetter panel rejected this contention, citing Desert Palace for the assertion that “[o]ur court has recently held that it is not appropriate to transfer wholesale the rationale of First Options to the collective bargaining context.” Id. at 439. As the court in Pacesetter explained, the court in Desert Palace declined to follow First Options due to the “fundamentally different policy concerns” that are implicated in commercial, as opposed to labor, arbitration agreements. Id. The court concluded that “all of Desert Palace’s policy reasons for arbitrating disputes between the parties to labor disputes” were applicable in Pacesetter also and thus that case was not controlled by First Options, but rather a prior Ninth Circuit opinion in the labor context. Id. at 439–40.
This court has since recognized that the analysis in Granite Rock undermined Pacesetter and its reliance on the differences between commercial and labor arbitration cases. Specifically, in Local Joint Executive Board v. Mirage Casino-Hotel, Inc., the court explained:
It is true that in Pacesetter Construction Co. v. Carpenters 46 Northern California Counties Conference Board, 116 F.3d 436, 439 (9th Cir. 1997), this court declined to adopt “wholesale” the Court's reasoning in First Options—a commercial arbitration case—and held that [prior Ninth Circuit precedent] remained good law in the labor context. But since Pacesetter, the Supreme Court has made clear that the “same framework” applies in both labor and commercial arbitration disputes. Indeed, several other circuits have applied First Options in the labor context.
911 F.3d 588, 598 (9th Cir. 2018) (citations omitted). We now reach the same conclusion with respect to Desert Palace that the court in Local Joint Executive Board reached with respect to Pacesetter—the Supreme Court's reasoning in Granite Rock is clearly irreconcilable with Desert Palace, and thus Desert Palace has been abrogated. See Miller, 335 F.3d at 893 (“We hold that in circumstances ․ where the reasoning or theory of our prior circuit authority is clearly irreconcilable with the reasoning or theory of intervening higher authority, a three-judge panel should consider itself bound by the later and controlling authority, and should reject the prior circuit opinion ․.”).9
Having concluded that Desert Palace is no longer good law, we must determine whether there is “clear and unmistakable” evidence of the parties’ intent to have an arbitrator—rather than the court—decide whether the grievance is arbitrable. Under Article 38 of the CBA, an arbitrator is tasked with resolving any “dispute or disagreement involving the interpretation, application or compliance with specific provisions of [the CBA].” The CBA is otherwise silent as to the arbitrator's authority to determine its own jurisdiction, and SEIU does not provide any other evidence of an agreement to have an arbitrator decide that issue.
As First Options instructs us, “[c]ourts should not assume that the parties agreed to arbitrate arbitrability” in the face of “silence or ambiguity” with respect to the Delegation Question. 514 U.S. at 944–45, 115 S.Ct. 1920. Without “clear and unmistakable” evidence, then, we rely on the well-settled principle “that whether parties have agreed to submit a particular dispute to arbitration is typically an issue for judicial determination.” Granite Rock, 561 U.S. at 296, 130 S.Ct. 2847 (citations, quotation marks, and alteration marks omitted). We therefore hold that the district court is responsible for determining whether the grievance filed by SEIU is arbitrable.
The district court's order granting SEIU's motion to compel arbitration is reversed and the case is remanded for further proceedings consistent with this opinion.
REVERSED and REMANDED.
I agree with much of the majority's analysis. But I am not convinced that Granite Rock Co. v. International Brotherhood of Teamsters, 561 U.S. 287, 130 S.Ct. 2847, 177 L.Ed.2d 567 (2010) has effectively overruled United Brotherhood of Carpenters & Joiners of America, Local No. 1780 v. Desert Palace, Inc., 94 F.3d 1308 (9th Cir. 1996) because they address two related—but distinct—issues.
To be clear, the foundation for Desert Palace appears flawed. The Supreme Court in First Options of Chicago, Inc. v. Kaplan went to great lengths to specify that “[c]ourts should not assume that the parties agreed to arbitrate arbitrability unless there is clear and unmistakable evidence that they do so.” 514 U.S. 938, 944, 115 S.Ct. 1920, 131 L.Ed.2d 985 (1995) (internal quotation marks and alterations omitted) (emphasis added). This court in Desert Palace, however, distinguished First Options by limiting it to the commercial arbitration context only. We held that, in contrast, a broad arbitration agreement in the collective bargaining context reflects “clear and unmistakable evidence” that the parties agreed to arbitrate arbitrability—even though the contract is silent on that issue. Desert Palace, 94 F.3d at 1311–12.
But it is counterintuitive to construe silence as “clear and unmistakable evidence” of an agreement on an issue that neither party has addressed in the contract. To be sure, courts sometimes infer silence as an implicit agreement. But silence certainly does not indicate “clear and unmistakable evidence” of an agreement. Cf. First Options, 514 U.S. at 945, 115 S.Ct. 1920 (“And, given the principle that a party can be forced to arbitrate only those issues it specifically has agreed to submit to arbitration, one can understand why courts might hesitate to interpret silence or ambiguity on the ‘who should decide arbitrability’ point as giving the arbitrators that power.”); 1 Corbin on Contracts § 3.18 (2020) (explaining that it is “not a rule of law” that “silence gives consent,” but sometimes “silence coupled with conduct or with expectations engendered by a prior relationship can reasonably be understood by the offeror as an acceptance”). Our holding in Desert Palace has thus created a strained distinction between commercial arbitration and labor arbitration on who decides arbitrability.
Nonetheless, I do not believe that Desert Palace is clearly irreconcilable with Granite Rock because it dealt with a related but different question: Desert Palace addressed who decides arbitrability (i.e., the Delegation Question), while Granite Rock dealt with whether a specific issue is arbitrable (i.e., the Arbitrability Question). Compare Granite Rock, 561 U.S. at 297–98, 130 S.Ct. 2847 (“The parties agree that it was proper for the District Court to decide whether their ratification dispute was arbitrable [i.e., Delegation Question]. They disagree about whether the District Court answered the question correctly [i.e., Arbitrability Question].” (footnote omitted)) with Desert Palace, 94 F.3d at 1311 (“As our precedents hold, a broad arbitration clause—even one that does not specifically mention who decides arbitrability—is sufficient to grant the arbitrator authority to decide his or her own jurisdiction.”).1
As the majority notes, the Supreme Court in Granite Rock did emphasize courts should apply the same arbitration framework in both commercial and labor arbitration contexts. But the Court in Granite Rock was addressing the need to ensure that courts “compel[ ] arbitration of a dispute only after [a court is] persuaded that the parties’ arbitration agreement was validly formed and that it covered the dispute in question and was legally enforceable.” Granite Rock, 561 U.S. at 300, 130 S.Ct. 2847. It also explained that courts discharge their duty to apply the “settled framework for deciding arbitrability” by “(1) applying the presumption of arbitrability only where a validly formed and enforceable arbitration agreement is ambiguous about whether it covers the dispute at hand; and (2) adhering to the presumption and ordering arbitration only where the presumption is not rebutted.” Id. at 301–02, 130 S.Ct. 2847. The Supreme Court in Granite Rock appeared to implicitly recognize that this does not necessarily implicate the Delegation Question: It explained that the “duty to interpret the agreement and to determine whether the parties intended to arbitrate grievances concerning a particular matter” applies “except where the parties clearly and unmistakably provide otherwise.” Id. at 301, 130 S.Ct. 2847 (emphasis added, internal quotation marks omitted).
Moreover, Granite Rock’s rejection of using policy consideration as a substitute for party agreement does not necessarily undermine Desert Palace. Desert Palace did not purport to rely on policy considerations to overcome a lack of agreement on the Delegation Question. Rather, Desert Palace stated that it was relying on “an act of legitimate contract interpretation” to interpret a broad arbitration clause in the collective bargaining context as delegating the Arbitrability Question to the arbitrator. 94 F.3d at 1310 (emphasis added, internal quotation marks omitted). The Desert Palace court explained that “[i]n the collective bargaining context, the issue of who gets to decide arbitrability is not an ‘arcane’ question the parties do not think about,” so “[b]y entering into a collective bargaining agreement, the parties know they will get an arbitrator rather than the courts to decide any labor dispute that arises under their collective bargaining agreement, including a dispute over arbitrability.” Id. at 1311. While one may question the analysis in Desert Palace, we need to accept the opinion's assertion that it relied on contract principles, not policy considerations.
In sum, I do not believe that Granite Rock “undercut the theory or reasoning underlying the prior circuit precedent in such a way that the cases are clearly irreconcilable.” Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003). So I believe that Desert Palace should still stand, even if it teeters on a flawed foundation. See Federal Trade Commission v. Consumer Defense, LLC, 926 F.3d 1208, 1213 (9th Cir. 2019) (explaining that “if we can apply our precedent consistently with that of the higher authority, we must do so”).
I respectfully dissent.
1. This opinion refers to “collective bargaining” and “labor” arbitration interchangeably.
2. The concept of classifying the “Merits Question,” the “Arbitrability Question,” and the “Delegation Question” originated from Judge Ikuta's dissent in ASARCO LLC v. United Steel, Paper & Forestry, Rubber, Manufacturing, Energy, Allied Industrial & Service Workers International Union. 910 F.3d 485, 496–97 (9th Cir. 2018) (Ikuta, J., dissenting). The “Delegation Question” received its name “[b]ecause this third species of dispute raises the question whether the parties delegated the arbitrability decision to the arbitrator.” Id. at 497 (emphasis added). Because these labels aid in differentiating between the types of disputes at issue, we adopt the use of them here.
3. See also Desert Palace, 94 F.3d at 1312 (opining that the Supreme Court's observation in First Options that “there is no strong federal policy favoring arbitration of commercial disputes, does not apply in the collective bargaining context”).
4. See also Granite Rock, 561 U.S. at 302–03, 130 S.Ct. 2847 (citing a commercial arbitration case for the assertion that “[t]he FAA's proarbitration policy does not operate without regard to the wishes of the contracting parties” and, in the same string cite, citing a labor case as “applying the same rule to the presumption of arbitrability for labor disputes” (emphases added) (internal citations and quotation marks omitted)); id. at 303, 130 S.Ct. 2847 (“We have applied the presumption favoring arbitration, in FAA and in labor cases, only where it reflects ․ that arbitration of a particular dispute is what the parties intended ․.” (emphases added)).
5. Compare Granite Rock, 561 U.S. at 297, 299, 130 S.Ct. 2847 (explaining, in response to the Arbitrability Question, that a dispute should be decided by an arbitrator “only where the court is satisfied that the parties agreed to arbitrate that dispute” because “[a]rbitration is strictly a matter of consent, and thus is a way to resolve those disputes—but only those disputes—that the parties have agreed to submit to arbitration” (internal citations and quotation marks omitted)), with First Options, 514 U.S. at 943, 115 S.Ct. 1920 (“Just as the arbitrability of the merits of a dispute depends upon whether the parties agreed to arbitrate that dispute, so the question ‘who has the primary power to decide arbitrability’ [i.e., the Delegation Question] turns upon what the parties agreed about that matter.” (internal citations omitted)).
6. The dissent's conclusion that, although Desert Palace is wrong, it isn't necessarily irreconcilable with Granite Rock, relies on essentially the same argument made by SEIU—that Desert Palace and Granite Rock are reconcilable because the former dealt with the Delegation Question and the latter dealt with the Arbitrability Question. The dissent elaborates that, although the Supreme Court in Granite Rock applied the same arbitration framework with respect to both commercial and labor disputes, the “arbitration framework” at issue was only in the context of Arbitrability Questions. Relying on certain text in Granite Rock, the dissent asserts that the Court “appeared to implicitly recognize that this [uniform framework in commercial and labor cases] does not necessarily implicate the Delegation Question.”As an initial matter, the dissent's conclusion that the Supreme Court's broad statements made in the context of Arbitrability Questions do not “necessarily implicate” Delegation Questions ignores that a Delegation Question is, really, just a very specific type of Arbitrability Question. The Arbitrability Question asks whether the parties agreed to have an arbitrator decide “a particular matter.” Granite Rock, 561 U.S. at 301, 130 S.Ct. 2847. The Delegation Question asks whether the parties agreed to have an arbitrator decide the specific matter of arbitrability. Given this close relationship between the two questions, if the Supreme Court has generally declared—as it has in Granite Rock—that commercial and labor disputes are subject to the same analysis in the context of Arbitrability Questions, then presumably such a declaration applies to Delegation Questions.Indeed, we don't actually need to presume that the two related questions share the same starting point for the “who decides” issue—the Supreme Court, as far back as First Options, made clear they do when it said: “Just as the arbitrability of the merits of a dispute depends upon whether the parties agreed to arbitrate that dispute, so the question ‘who has the primary power to decide arbitrability’ turns upon what the parties agreed about that matter.” 514 U.S. at 943, 115 S.Ct. 1920 (first emphasis added and citation omitted). The only difference between the two questions is that, unlike the Arbitrability Question, for the Delegation Question, courts apply both a presumption and a heightened evidentiary requirement against arbitrability. Id. at 944–45, 115 S.Ct. 1920 (for the Delegation Question, courts “reverse[ ] the presumption” and require “clear and unmistakable evidence” of the parties’ intent to arbitrate (citation, quotation marks, and alteration marks omitted)). Since Delegation Questions are generally treated similarly to Arbitrability Questions—except arbitrability is harder to establish for Delegation Questions—the dissent's speculation that the Granite Rock Court may have intended to implicitly cordon off what our court has subsequently labeled “Delegation Questions” from the Court's discussion about arbitrability seems a stretch. Why would the Granite Rock Court's emphasis that labor policy concerns cannot substitute for evidence of party agreement not apply a fortiori to the subset of Arbitrability Questions that we have labeled Delegation Questions?Lastly, the dissent's rationale—that the Granite Rock Court might have carved out the Delegation Question from its broad statement that labor and commercial cases should be treated no differently—is belied by the context of the very Granite Rock text the dissent references. The full text from Granite Rock that the dissent quotes from reads: “Our cases invoking the federal ‘policy favoring arbitration’ of commercial and labor disputes apply the same framework. They recognize that, except where ‘the parties clearly and unmistakably provide otherwise,’ it is ‘the court's duty to interpret the agreement and to determine whether the parties intended to arbitrate grievances concerning’ a particular matter.” 561 U.S. at 301, 130 S.Ct. 2847 (citations omitted). The “clearly and unmistakably” language quoted in the second sentence comes from AT&T Technologies, Inc. v. Communications Workers of America, 475 U.S. 643, 649, 106 S.Ct. 1415, 89 L.Ed.2d 648 (1986), where the Court was discussing the Delegation Question. It would be quite odd for the Court to have tacitly excluded the Delegation Question from its broad statement when it relied on a Delegation Question case to support it.
7. The dissent contends that “Desert Palace did not purport to rely on policy considerations,” but instead relied on a principle of contract interpretation. It purported to rely on both. After stating that the rule it was applying was “ ‘an act of legitimate contract interpretation’ ․ under this circuit's precedents,” Desert Palace, 94 F.3d at 1310, it proceeded to distinguish “the Supreme Court's recent decision in First Options” on the sole basis that the Court's expressed “policy concerns, however, do not come into play in the collective bargaining context.” Id. at 1310–11; see also id. at 1312 (asserting “there is no strong federal policy favoring arbitration of commercial disputes,” but there is “a strong federal policy favoring arbitration of labor disputes” (first and third emphasis added)); id. (stating “federal labor policy strongly favors the resolution of labor disputes through arbitration” (emphasis added) (quoting United Food & Commercial Workers Int'l Union, Local 588 v. Foster Poultry Farms, 74 F.3d 169, 173 (9th Cir. 1995))).But even the “act of legitimate contract interpretation” relied on by Desert Palace was itself a result of a labor-context policy decision. Any time a court says that it will assume the parties agreed to something they didn't actually expressly agree to, that rule—whether labeled a “contract interpretation” rule or something else—is a result of a policy decision. Here, the case cited in Desert Palace for the broad arbitration clause rule as “an act of legitimate contract interpretation” does not offer any rationale for the rule other than labor policy considerations. See Frederick Meiswinkel, Inc. v. Laborer's Union Local 261, 744 F.2d 1374, 1376 (9th Cir. 1984) (asserting that “[f]ederal policy generally favors arbitration of labor disputes” (citation omitted)). Indeed, other decisions from this court have explained in more detail the labor policy considerations at play in Desert Palace, as Desert Palace itself emphasized. Only a few paragraphs after relying on “this circuit's precedents,” it block-quotes from one of those “precedents” that fleshes out in more detail its strong reliance on labor policy considerations. See Desert Palace, 94 F.3d at 1311 (“The labor arbitrator does not merely provide an alternative means of resolving disputes ․. Unlike the commercial contract, ․ [t]he labor arbitrator is the person the parties designate to fill in the gaps; ․ the arbitrator will state the parties’ bargain. He is ‘the parties’ officially designated “reader” of the contract ․ their joint alter ego ․’ to handle matters omitted from the agreement.” (underlining added and alteration marks omitted) (quoting Stead Motors of Walnut Creek v. Auto. Machinists Lodge No. 1173, 886 F.2d 1200, 1205 (9th Cir. 1989) (en banc))). So ultimately, it's policy considerations all the way down.The Desert Palace court had to rely on these policy differences so that its “conclusion [wa]s not affected by the Supreme Court's recent decision in First Options.” Id. at 1310. Absent its policy-based distinction, the Desert Palace court itself recognized First Options mandated a different outcome.
8. See also InSync, 801 F.3d at 1042 (discussing the application of “Interstate Distributor’s rule that when the collective bargaining agreement contains a customary arbitration clause acts of repudiation and other acts of termination must be submitted to arbitration” (emphasis added) (citations and quotation marks omitted)); id. at 1043 (setting out the principle from Interstate Distributor’s progeny that “an agreement to arbitrate any differences that may arise regarding the meaning and enforcement of this Agreement, or any other broad arbitration clause, ․ ordinarily requires us to hold that the parties have provided for arbitration of disputes regarding termination” (emphasis added) (citations, internal quotation marks, and emphasis omitted)); id. at 1043–44 (distinguishing a Supreme Court case on the basis that the parties in InSync disputed “the issue of whether the agreement has expired or been terminated” (emphasis added)); id. at 1044 (acknowledging that “once it is found that a contract did exist at some time, the questions of whether that contract has expired, or has been terminated or repudiated, may well present arbitrable issues” (emphasis added)).
9. See also Head, 936 F.3d at 1014 (concluding that Ninth Circuit precedent was irreconcilable with a subsequent Supreme Court opinion and overruling such precedent without taking the case en banc); Palmer v. Sanderson, 9 F.3d 1433, 1437 n.5 (9th Cir. 1993) (noting “a panel can reexamine the earlier decision of a three-judge panel if that earlier decision has been undermined by later overriding precedent” (quoting United States v. Magana, 797 F.2d 777, 779 (9th Cir. 1986))).
1. Admittedly, Desert Palace remains vague on whether it actually applied the clear and unmistakable evidence standard. After articulating that standard in the second sentence of the discussion section, the opinion curiously never mentions it again. 94 F.3d at 1310. But cases applying Desert Palace have construed it as applying the clear and unmistakable standard. See, e.g., IATSE Local 720 v. InSync Show Productions, Inc., 801 F.3d 1033, 1043 n.7 (9th Cir. 2015) (“Because the scope of the arbitration provision broadly covers claims or allegations of employer violations of the agreement, this case is not one in which the arbitration provision's scope is too narrow to ‘clearly and unmistakably leave questions of arbitrability for determination by the arbitrator’ ” (internal quotation marks omitted)).
VANDYKE, Circuit Judge:
Dissent by Judge Lee