Supreme Court: Ambiguous Contract, Like Silent One, Insufficient to Support Class Arbitration — Contra Proferentem Is Public Policy Rule, Not Helpful in Determining Parties’ Intent Nor Basis for Imposing Classwide Arbitration
Lamps Plus v. Varelam 2019 U.S. LEXIS 2943 (U.S. April 24, 2019):
The Federal Arbitration Act requires courts to enforce covered arbitration agreements according to their terms. See 9 U. S. C. §2. In Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662 (2010), we held that a court may not compel arbitration on a classwide basis when an agreement is “silent” on the availability of such arbitration. Because class arbitration fundamentally changes the nature of the “traditional individualized arbitration” envisioned by the FAA, Epic Systems Corp. v. Lewis, 584 U. S. ___, ___ (2018) (slip op., at 8), “a party may not be compelled under the FAA to submit to class arbitration unless there is a contractual basis for concluding that the party agreed to do so,” Stolt-Nielsen, 559 U. S., at 684 (emphasis in original). We now consider whether the FAA similarly bars an order requiring class arbitration when an agreement is not silent, but rather “ambiguous” about the availability of such arbitration.
Petitioner Lamps Plus is a company that sells light fixtures and related products. In 2016, a hacker impersonating a company official [*6] tricked a Lamps Plus employee into disclosing the tax information of approximately 1,300 other employees. Soon after, a fraudulent federal income tax return was filed in the name of Frank Varela, a Lamps Plus employee and respondent here.
Like most Lamps Plus employees, Varela had signed an arbitration agreement when he started work at the company. But after the data breach, he sued Lamps Plus in Federal District Court in California, bringing state and federal claims on behalf of a putative class of employees whose tax information had been compromised. Lamps Plus moved to compel arbitration on an individual rather than classwide basis, and to dismiss the lawsuit. In a single order, the District Court granted the motion to compel arbitration and dismissed Varela’s claims without prejudice. But the court rejected Lamps Plus’s request for individual arbitration, instead authorizing arbitration on a classwide basis. Lamps Plus appealed the order, arguing that the court erred by compelling class arbitration.
The Ninth Circuit affirmed. 701 Fed. Appx. 670 (2017). The court acknowledged that Stolt-Nielsen prohibits forcing a party “to submit to class arbitration unless there is a contractual basis for concluding that [*7] the party agreed to do so” and that Varela’s agreement “include[d] no express mention of class proceedings.” 701 Fed. Appx., at 672. But that did not end the inquiry, the court reasoned, because the fact that the agreement “does not expressly refer to class arbitration is not the ‘silence’ contemplated in Stolt-Nielsen.” Ibid. In Stolt-Nielsen, the parties had stipulated that their agreement was silent about class arbitration. Because there was no such stipulation here, the court concluded that Stolt-Nielsen was not controlling.
The Ninth Circuit then determined that the agreement was ambiguous on the issue of class arbitration. On the one hand, as Lamps Plus argued, certain phrases in the agreement seemed to contemplate “purely binary claims.” Ibid. At the same time, as Varela asserted, other phrases were capacious enough to include class arbitration, such as one stating that “arbitration shall be in lieu of any and all lawsuits or other civil legal proceedings relating to my employment.” Ibid. The Ninth Circuit followed California law to construe the ambiguity against the drafter, a rule that “applies with peculiar force in the case of a contract of adhesion” such as this. Ibid. (quoting Sandquist v. Lebo Auto., Inc., 1 Cal. 5th 233, 248, 376 P. 3d 506, 514 (2016)). Because Lamps [*8] Plus had drafted the agreement, the court adopted Varela’s interpretation authorizing class arbitration. Judge Fernandez dissented. In his view, the agreement was not ambiguous, and the majority’s holding was a “palpable evasion of Stolt-Nielsen.” 701 Fed. Appx., at 673.
Lamps Plus petitioned for a writ of certiorari, arguing that the Ninth Circuit’s decision contravened Stolt-Nielsen and created a conflict among the Courts of Appeals. In opposition, Varela not only disputed those contentions but also argued for the first time that the Ninth Circuit lacked jurisdiction over the appeal, and that this Court therefore lacked jurisdiction in turn. We granted certiorari. 584 U. S. ___ (2018).
We begin with jurisdiction. Section 16 of the FAA governs appellate review of arbitration orders. 9 U. S. C. §16. Varela contends that the Ninth Circuit lacked statutory jurisdiction because section 16 permits appeal from orders denying motions to compel arbitration, §16(a)(1)(B), but not orders granting such motions, §16(b)(2). Brief for Respondent 9-12; see also post, at 3 (Breyer, J., dissenting). This argument is beside the point, however, because Lamps Plus relies for jurisdiction on a different provision of section 16, section 16(a)(3).
Section 16(a)(3) provides that an appeal may be taken from “a final decision with respect to an arbitration [*9] that is subject to this title.” We construed that provision in Green Tree Financial Corp.-Ala. v. Randolph, 531 U. S. 79 (2000), a case where, as here, the District Court had issued an order both compelling arbitration and dismissing the underlying claims. We held that such an order directing “the parties to proceed to arbitration, and dismiss[ing] all the claims before [the court], . . . is ‘final’ within the meaning of §16(a)(3), and therefore appealable.” Id., at 89. 1
Varela attempts to distinguish Randolph on the ground that the appeal here was taken by the party who sought an order to dismiss the claim and compel arbitration, Lamps Plus. He claims the company “lacked standing to appeal the dismissal,” because the District Court’s order “provided precisely the relief Lamps Plus sought.” Brief for Respondent 13, 15.
But Lamps Plus did not secure the relief it requested. It sought an order compelling individual arbitration. What it got was an order rejecting that relief and instead compelling arbitration on a classwide basis. We have explained—and will elaborate further below—that shifting from individual to class arbitration is a “fundamental” change, Stolt-Nielsen, 559 U. S., at 686, that “sacrifices the principal [*10] advantage of arbitration” and “greatly increases risks to defendants,” AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 348, 350 (2011). Lamps Plus’s interest in avoiding those consequences gives it the “necessary personal stake in the appeal” required by our precedent. Camreta v. Greene, 563 U. S. 692, 702 (2011). 2
The Ninth Circuit applied California contract law to conclude that the parties’ agreement was ambiguous on the availability of class arbitration. In California, an agreement is ambiguous “when it is capable of two or more constructions, both of which are reasonable.” 701 Fed. Appx., at 672 (quoting Powerine Oil Co. v. Superior Ct., 37 Cal. 4th 377, 390, 118 P. 3d 589, 598 (2005)). Following our normal practice, we defer to the Ninth Circuit’s interpretation and application of state law and thus accept that the agreement should be regarded as ambiguous. See, e.g., Expressions Hair Design v. Schneiderman, 581 U. S. ___, ___ (2017) (slip op., at 7). 3
We therefore face the question whether, consistent with the FAA, an ambiguous agreement can provide the necessary “contractual basis” for compelling class arbitration. Stolt-Nielsen, 559 U. S., at 684. We hold that it cannot—a conclusion that follows directly from our decision in Stolt-Nielsen. Class arbitration is not only markedly different from the “traditional individualized arbitration” contemplated by the FAA, it also undermines the most important benefits of that familiar form of arbitration. Epic Systems, 584 U. S., at ___ (slip op., at 8); see Stolt-Nielsen, 559 U. S., at 686-687. The statute therefore [*11] requires more than ambiguity to ensure that the parties actually agreed to arbitrate on a classwide basis.
The FAA requires courts to “enforce arbitration agreements according to their terms.” Epic Systems, 584 U. S., at ___ (slip op., at 5) (quoting American Express Co. v. Italian Colors Restaurant, 570 U. S. 228, 233 (2013)). Although courts may ordinarily accomplish that end by relying on state contract principles, First Options of Chicago, Inc. v. Kaplan, 514 U. S. 938, 944 (1995), state law is preempted to the extent it “stands as an obstacle to the accomplishment and execution of the full purposes and objectives” of the FAA, Concepcion, 563 U. S., at 352 (internal quotation marks omitted). At issue in this case is the interaction between a state contract principle for addressing ambiguity and a “rule[ ] of fundamental importance” under the FAA, namely, that arbitration “is a matter of consent, not coercion.” Stolt-Nielsen, 559 U. S., at 681 (internal quotation marks omitted).
“[T]he first principle that underscores all of our arbitration decisions” is that “[a]rbitration is strictly a matter of consent.” Granite Rock Co. v. Teamsters, 561 U. S. 287, 299 (2010) (internal quotation marks omitted). We have emphasized that “foundational FAA principle” many times. Stolt-Nielsen, 559 U. S., at 684; see also, e.g., Howsam v. Dean Witter Reynolds, Inc., 537 U. S. 79, 83 (2002); First Options, 514 U. S., at 943; Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U. S. 52, 57 (1995); Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 479 (1989); Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U. S. 614, 626 (1985).
Consent is essential under the FAA because arbitrators wield only the authority they are given. That is, they derive their “powers from the parties’ agreement to forgo the legal process and submit [*12] their disputes to private dispute resolution.” Stolt-Nielsen, 559 U. S., at 682. Parties may generally shape such agreements to their liking by specifying with whom they will arbitrate, the issues subject to arbitration, the rules by which they will arbitrate, and the arbitrators who will resolve their disputes. Id., at 683-684. Whatever they settle on, the task for courts and arbitrators at bottom remains the same: “to give effect to the intent of the parties.” Id., at 684.
In carrying out that responsibility, it is important to recognize the “fundamental” difference between class arbitration and the individualized form of arbitration envisioned by the FAA. Epic Systems, 584 U. S., at ___ (slip op., at 8); see also Concepcion, 563 U. S., at 349, 351; Stolt-Nielsen, 559 U. S., at 686-687. In individual arbitration, “parties forgo the procedural rigor and appellate review of the courts in order to realize the benefits of private dispute resolution: lower costs, greater efficiency and speed, and the ability to choose expert adjudicators to resolve specialized disputes.” Id., at 685. Class arbitration lacks those benefits. It “sacrifices the principal advantage of arbitration—its informality—and makes the process slower, more costly, and more likely to generate procedural morass than final judgment.” Concepcion, 563 U. S., at 348. Indeed, we recognized just last Term that with class arbitration “the [*13] virtues Congress originally saw in arbitration, its speed and simplicity and inexpensiveness, would be shorn away and arbitration would wind up looking like the litigation it was meant to displace.” Epic Systems, 584 U. S., at ___ (slip op., at 8). Class arbitration not only “introduce[s] new risks and costs for both sides,” ibid., it also raises serious due process concerns by adjudicating the rights of absent members of the plaintiff class—again, with only limited judicial review. See Concepcion, 563 U. S., 349; see also Stolt-Nielsen, 559 U. S., at 686 (citing Ortiz v. Fibreboard Corp., 527 U. S. 815, 846 (1999)).
Because of these “crucial differences” between individual and class arbitration, Stolt-Nielsen explained that there is “reason to doubt the parties’ mutual consent to resolve disputes through classwide arbitration.” 559 U. S., at 687, 685-686. And for that reason, we held that courts may not infer consent to participate in class arbitration absent an affirmative “contractual basis for concluding that the party agreed to do so.” Id., at 684. Silence is not enough; the “FAA requires more.” Id., at 687.
Our reasoning in Stolt-Nielsen controls the question we face today. Like silence, ambiguity does not provide a sufficient basis to conclude that parties to an arbitration agreement agreed to “sacrifice[ ] the principal advantage of arbitration.” Concepcion, 563 U. S., at 348.
This conclusion aligns with our [*14] refusal to infer consent when it comes to other fundamental arbitration questions. For example, we presume that parties have not authorized arbitrators to resolve certain “gateway” questions, such as “whether the parties have a valid arbitration agreement at all or whether a concededly binding arbitration clause applies to a certain type of controversy.” Green Tree Financial Corp. v. Bazzle, 539 U. S. 444, 452 (2003) (plurality opinion). Although parties are free to authorize arbitrators to resolve such questions, we will not conclude that they have done so based on “silence or ambiguity” in their agreement, because “doing so might too often force unwilling parties to arbitrate a matter they reasonably would have thought a judge, not an arbitrator, would decide.” First Options, 514 U. S., at 945 (emphasis added); see also Howsam, 537 U. S., at 83-84. We relied on that same reasoning in Stolt-Nielsen, 559 U. S., at 686-687, and it applies with equal force here. Neither silence nor ambiguity provides a sufficient basis for concluding that parties to an arbitration agreement agreed to undermine the central benefits of arbitration itself. 4
The Ninth Circuit reached a contrary conclusion based on California’s rule that ambiguity in a contract should be construed against the drafter, a doctrine known as contra proferentem. The rule applies “only [*15] as a last resort” when the meaning of a provision remains ambiguous after exhausting the ordinary methods of interpretation. 3 A. Corbin, Contracts §559, pp. 268-270 (1960). At that point, contra proferentem resolves the ambiguity against the drafter based on public policy factors, primarily equitable considerations about the parties’ relative bargaining strength. See 2 E. Farnsworth, Contracts §7.11, pp. 300-304 (3d ed. 2004); see also 11 R. Lord, Williston on Contracts §32:12, pp. 788-792 (4th ed. 2012) (stating that application of the rule may vary based on “the degree of sophistication of the contracting parties or the degree to which the contract was negotiated”); Restatement (Second) of Contracts §206, pp. 80-81, 105-107 (1979) (classifying contra proferentem under “Considerations of Fairness and the Public Interest” rather than with rules for interpreting “The Meaning of Agreements”); 3 Corbin, Contracts §559, at 270 (noting that contra proferentem is “chiefly a rule of public policy”). Although the rule enjoys a place in every hornbook and treatise on contracts, we noted in a recent FAA case that “the reach of the canon construing contract language against the drafter must have limits, no matter who the drafter was.” DIRECTV, Inc. v. Imburgia, 577 U. S. ___, ___ (2015) (slip op., at 10). This case [*16] brings those limits into focus.
Unlike contract rules that help to interpret the meaning of a term, and thereby uncover the intent of the parties, contra proferentem is by definition triggered only after a court determines that it cannot discern the intent of the parties. When a contract is ambiguous, contra proferentem provides a default rule based on public policy considerations; “it can scarcely be said to be designed to ascertain the meanings attached by the parties.” 2 Farnsworth, Contracts §7.11, at 303. Like the contract rule preferring interpretations that favor the public interest, see id., at 304, contra proferentem seeks ends other than the intent of the parties.
“[C]lass arbitration, to the extent it is manufactured by [state law] rather than consen[t], is inconsistent with the FAA.” Concepcion, 563 U. S., at 348. We recently reiterated that courts may not rely on state contract principles to “reshape traditional individualized arbitration by mandating classwide arbitration procedures without the parties’ consent.” Epic Systems, 584 U. S., at ___ (slip op., at 8). But that is precisely what the court below did, requiring class arbitration on the basis of a doctrine that “does not help to determine the meaning that the two parties gave to the words, or even the meaning [*17] that a reasonable person would have given to the language used.” 3 Corbin, Contracts §559, at 269-270. Such an approach is flatly inconsistent with “the foundational FAA principle that arbitration is a matter of consent.” Stolt-Nielsen, 559 U. S., at 684.
Varela and Justice Kagan defend application of the rule on the basis that it is nondiscriminatory. It does not conflict with the FAA, they argue, because it is a neutral rule that gives equal treatment to arbitration agreements and other contracts alike. See Brief for Respondent 18, 25-26; post, at 6-9 (Kagan, J., dissenting). We have explained, however, that such an equal treatment principle cannot save from preemption general rules “that target arbitration either by name or by more subtle methods, such as by ‘interfer[ing] with fundamental attributes of arbitration.’” Epic Systems, 584 U. S., at ___ (slip op., at 7) (quoting Concepcion, 563 U. S., at 344).
That was the case in Concepcion. There, the Court considered the general contract defense of unconscionability, which had been interpreted by the state court to bar class action waivers in consumer contracts, whether in the litigation or arbitration context. See id., at 341-344. The general applicability of the rule did not save it from preemption under the FAA with respect to arbitration agreements, because [*18] it had the consequence of allowing any party to a consumer arbitration agreement to demand class proceedings “without the parties’ consent.” Epic Systems, 584 U. S., at ___ (slip op., at 8) (describing the “essential insight” of Concepcion). That, for the reasons we have explained, “interferes with fundamental attributes of arbitration and thus creates a scheme inconsistent with the FAA.” Concepcion, 563 U. S., at 344; see Epic Systems, 584 U. S., at ___-___ (slip op., at 8-9). The same reasoning applies here: The general contra proferentem rule cannot be applied to impose class arbitration in the absence of the parties’ consent. 5
Our opinion today is far from the watershed Justice Kagan claims it to be. Rather, it is consistent with a long line of cases holding that the FAA provides the default rule for resolving certain ambiguities in arbitration agreements. For example, we have repeatedly held that ambiguities about the scope of an arbitration agreement must be resolved in favor of arbitration. See, e.g., Mitsubishi Motors Corp., 473 U. S., at 626; Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U. S. 1, 24-25 (1983). In those cases, we did not seek to resolve the ambiguity by asking who drafted the agreement. Instead, we held that the FAA itself provided the rule. As in those cases, the FAA provides the default rule for resolving ambiguity [*19] here.
Courts may not infer from an ambiguous agreement that parties have consented to arbitrate on a classwide basis. The doctrine of contra proferentem cannot substitute for the requisite affirmative “contractual basis for concluding that the part[ies] agreed to [class arbitration].” Stolt-Nielsen, 559 U. S., at 684.
We reverse the judgment of the Court of Appeals for the Ninth Circuit and remand the case for further proceedings consistent with this opinion.
It is so ordered.
Concur by: THOMAS
Justice Thomas, concurring.
As our precedents make clear and the Court acknowledges, the Federal Arbitration Act (FAA) requires federal courts to enforce arbitration agreements “just as they would ordinary contracts: in accordance with their terms.” Howsam v. Dean Witter Reynolds, Inc., 537 U. S. 79, 87 (2002) (Thomas, J., concurring in judgment). Federal courts must therefore apply “background principles of state contract law” when evaluating arbitration agreements. Arthur Andersen LLP v. Carlisle, 556 U. S. 624, 630 (2009); Perry v. Thomas, 482 U. S. 483, 492, n. 9 (1987). “In this endeavor, ‘as with any other contract, the parties’ intentions control.’” Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 682 (2010) (quoting Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U. S. 614, 626 (1985)). Thus, where an agreement is silent as to class arbitration, a court may not infer from that silence that the parties agreed to arbitrate on a class basis. 559 U. S., at 687.
Here, the arbitration agreement between [*20] Varela and Lamps Plus is silent as to class arbitration. If anything, the agreement suggests that the parties contemplated only bilateral arbitration. * App. to Pet. for Cert. 24a (waiving “any right I may have to file a lawsuit or other civil action or proceeding relating to my employment with the Company” (emphasis added)); ibid. (“The Company and I mutually consent to the resolution by arbitration of all claims . . . that I may have against the Company” (emphasis added)); id., at 24a-25a (“Specifically, the Company and I mutually consent to the resolution by arbitration of all claims that may hereafter arise in connection with my employment” (emphasis added)). This agreement provides no “contractual basis” for concluding that the parties agreed to class arbitration, Stolt-Nielsen, supra, at 684, and I would therefore reverse on that basis.
The Court instead evaluates whether California’s contra proferentem rule, as applied here, “‘stands as an obstacle to the accomplishment and execution of the full purposes and objectives’ of the FAA.” Ante, at 6 (quoting AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 352 (2011)). I remain skeptical of this Court’s implied pre-emption precedents, see Wyeth v. Levine, 555 U. S. 555, 582-604 (2009) (opinion concurring in judgment), but I join the opinion of the Court because it correctly [*21] applies our FAA precedents, see Epic Systems Corp. v. Lewis, 584 U. S. ___ (2018); Concepcion, supra.
Dissent by: GINSBURG; BREYER; SOTOMAYOR; KAGAN
Justice Ginsburg, with whom Justice Breyer and Justice Sotomayor join, dissenting.
Joining Justice Kagan’s dissenting opinion in full, I write separately to emphasize once again how treacherously the Court has strayed from the principle that “arbitration is a matter of consent, not coercion.” Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 681 (2010) (internal quotation marks omitted).
Congress enacted the Federal Arbitration Act (FAA) in 1925 “to enable merchants of roughly equal bargaining power to enter into binding agreements to arbitrate commercial disputes.” Epic Systems Corp. v. Lewis, 584 U. S. ___, ___ (2018) (Ginsburg, J., dissenting) (slip op., at 19) (emphasis in original). The Act was not designed to govern contracts “in which one of the parties characteristically has little bargaining power.” Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U. S. 395, 403, n. 9 (1967); see Gilmer v. Interstate/Johnson Lane Corp., 500 U. S. 20, 42 (1991) (Stevens, J., dissenting) (“I doubt that any legislator who voted for [the FAA] expected it to apply . . . to form contracts between parties of unequal bargaining power, or to the arbitration of disputes arising out of the employment relationship.”); Miller, Simplified Pleading, Meaningful Days in Court, and Trials on the Merits: Reflections on the Deformation of Federal Procedure, 88 N. Y. U. L. Rev. 286, 323 (2013) (The FAA [*22] was “enacted in 1925 with the seemingly limited purpose of overcoming the then-existing ‘judicial hostility’ to the arbitration of contract disputes between businesses.”).
The Court has relied on the FAA, not simply to overcome once-prevalent judicial resistance to enforcement of arbitration disputes between businesses. In relatively recent years, it has routinely deployed the law to deny to employees and consumers “effective relief against powerful economic entities.” DIRECTV, Inc. v. Imburgia, 577 U. S. ___, ___ (2015) (Ginsburg, J., dissenting) (slip op., at 9). Arbitration clauses, the Court has decreed, may preclude judicial remedies even when submission to arbitration is made a take-it-or-leave-it condition of employment or is imposed on a consumer given no genuine choice in the matter. See Epic, 584 U. S., at ___-___ (Ginsburg, J., dissenting) (slip op., at 21-22) (surveying “court decisions expansively interpreting” the FAA); Circuit City Stores, Inc. v. Adams, 532 U. S. 105, 132 (2001) (Stevens, J., dissenting) (“There is little doubt that the Court’s interpretation of the [FAA] has given it a scope far beyond the expectations of the Congress that enacted it.”); Miller, supra, at 324 (describing as “extraordinary” “judicial extension of the [FAA] to a vast array of consumer contracts . . . characterized by their adhesive nature and by the individual’s complete lack [*23] of bargaining power”). Propelled by the Court’s decisions, mandatory arbitration clauses in employment and consumer contracts have proliferated. See, e.g., Economic Policy Institute, A. Colvin, The Growing Use of Mandatory Arbitration 2, 4‒6 (Apr. 6, 2018) (mandatory arbitration imposed by private-sector employers on nonunionized employees notably increased between 1995 and 2017), online at https://www.epi.org/files/pdf/144131.pdf (all Internet materials as last visited Apr. 22, 2019); Consumer Financial Protection Bureau, Arbitration Study §1.4.1 (Mar. 2015) (“Tens of millions of consumers use consumer financial products or services that are subject to . . . arbitration clauses.”), online at https://files.consumerfinance.gov/f/ 201503_cfpb_arbitration-study-report-to-congress-2015.pdf.
Piling Pelion on Ossa, the Court has hobbled the capacity of employees and consumers to band together in a judicial or arbitral forum. See Epic, 584 U. S., at ___, n. 12 (Ginsburg, J., dissenting) (slip op., at 22, n. 12) (noting Court decisions enforcing class-action waivers imposed by the party in command, who wants no collective proceedings). The Court has pursued this course even though “neither the history nor present practice suggests that class arbitration [*24] is fundamentally incompatible with arbitration itself.” AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 362 (2011) (Breyer, J., dissenting).
Employees and consumers forced to arbitrate solo face severe impediments to the “vindication of their rights.” Stolt-Nielsen, 559 U. S., at 699 (Ginsburg, J., dissenting). “Expenses entailed in mounting individual claims will often far outweigh potential recoveries.” Epic, 584 U. S., at ___ (Ginsburg, J., dissenting) (slip op., at 27); see American Express Co. v. Italian Colors Restaurant, 570 U. S. 228, 246 (2013) (Kagan, J., dissenting) (“[The defendant] has put [the plaintiff] to this choice: Spend way, way, way more money than your claim is worth, or relinquish your . . . rights.”); Concepcion, 563 U. S., at 365 (Breyer, J., dissenting) (“What rational lawyer would have signed on to represent the [plaintiffs] for the possibility of fees stemming from a $30.22 [individual] claim?”); Resnik, Revising Our “Common Intellectual Heritage”: Federal and State Courts in Our Federal System, 91 Notre Dame L. Rev. 1831, 1888 (2016) (“Few individuals can afford to pursue small value claims; mandating single-file arbitration serves as a means of erasing rights, rather than enabling their ‘effective vindication.’”).
Today’s decision underscores the irony of invoking “the first principle” that “arbitration is strictly a matter of consent,” ante, at 7 (internal quotation marks and alterations omitted), to justify imposing individual [*25] arbitration on employees who surely would not choose to proceed solo. Respondent Frank Varela sought redress for negligence by his employer leading to a data breach affecting 1,300 employees. See Complaint in No. 5:16‒cv‒00577 (CD Cal.), Doc. 1, ¶¶1, 59. The widely experienced neglect he identified cries out for collective treatment. Blocking Varela’s path to concerted action, the Court aims to ensure the authenticity of consent to class procedures in arbitration. Ante, at 7‒8. Shut from the Court’s sight is the “Hobson’s choice” employees face: “accept arbitration on their employer’s terms or give up their jobs.” Epic, 584 U. S., at ___, n. 2 (Ginsburg, J., dissenting) (slip op., at 7, n. 2); see Circuit City, 532 U. S., at 139 (Souter, J., dissenting) (employees often “lack the bargaining power to resist an arbitration clause if their prospective employers insist on one”).
Recent developments outside the judicial arena ameliorate some of the harm this Court’s decisions have occasioned. Some companies have ceased requiring employees to arbitrate sexual harassment claims, see McGregor, Firms May Follow Tech Giants on Forced Arbitration, Washington Post, Nov. 13, 2018, p. A15, col. 1, or have extended their no-forced-arbitration policy to a broader range of claims, [*26] see Wakabayashi, Google Scraps Forced Arbitration Policy, N. Y. Times, Feb. 22, 2019, p. B5, col. 4. And some States have endeavored to safeguard employees’ opportunities to bring sexual harassment suits in court. See, e.g., N. Y. Civ. Prac. Law Ann. §7515 (West 2019) (rendering unenforceable certain mandatory arbitration clauses covering sexual harassment claims). These developments are sanguine, for “[p]lainly, it would not comport with the congressional objectives behind a statute seeking to enforce civil rights . . . to allow the very forces that had practiced discrimination to contract away the right to enforce civil rights in the courts.” Barrentine v. Arkansas-Best Freight System, Inc., 450 U. S. 728, 750 (1981) (Burger, C. J., dissenting).
Notwithstanding recent steps to counter the Court’s current jurisprudence, mandatory individual arbitration continues to thwart “effective access to justice” for those encountering diverse violations of their legal rights. DIRECTV, 577 U. S., at ___ (Ginsburg, J., dissenting) (slip op., at 1). The Court, paradoxically reciting the mantra that “[c]onsent is essential,” ante, at 7, has facilitated companies’ efforts to deny employees and consumers the “important right” to sue in court, and to do so collectively, by inserting solo-arbitration-only clauses that [*27] parties lacking bargaining clout cannot remove. CompuCredit Corp. v. Greenwood, 565 U. S. 95, 115 (2012) (Ginsburg, J., dissenting). When companies can “muffl[e] grievance[s] in the cloakroom of arbitration,” Merrill Lynch, Pierce, Fenner & Smith, Inc. v. Ware, 414 U. S. 117, 136 (1973), the result is inevitable: curtailed enforcement of laws “designed to advance the well-being of [the] vulnerable.” Epic, 584 U. S., at ___ (Ginsburg, J., dissenting) (slip op., at 26). “Congressional correction of the Court’s elevation of the FAA over” the rights of employees and consumers “to act in concert” remains “urgently in order.” Id., at ___ (slip op., at 2).
Justice Breyer, dissenting.
Although I join Justice Ginsburg’s and Justice Kagan’s dissents in full, I also dissent for another reason. In my view, the Court of Appeals lacked jurisdiction to hear this case. Consequently, we lack jurisdiction as well. See 28 U. S. C. §1254. My reason for reaching this conclusion is the following. The Federal Arbitration Act, at §4, says that a “court,” upon being satisfied that the parties have agreed to arbitrate a claim, “shall make an order directing the parties to proceed to arbitration in accordance with the terms of the agreement.” 9 U. S. C. §4. Section 16 of the Act then says that “an appeal may not be taken from an interlocutory order . . . directing arbitration to proceed under section 4 of this title.” §16(b)(2) (emphasis added). And directing arbitration [*28] to proceed is just what the District Court did here. App. to Pet. for Cert. 23a.
These statutory provisions reflect a congressional effort (in respect to a specific subject matter) to help resolve a more general problem. Too few interlocutory appeals will too often impose upon parties delay and expense that an interlocutory appeal, by quickly correcting a lower court error, might have spared them. But too many interlocutory appeals will too often unnecessarily delay proceedings while a party appeals and loses. And delays can clog the appellate system, thereby slowing down the workings, and adding to the costs, of the judicial system seen as a whole. Congress’ jurisdictional statutes consequently compromise, providing, for example, for interlocutory appeals in some instances, such as cases involving injunctive orders, see, e.g., 28 U. S. C. §1292(a)(1), or where important separable legal questions are at issue, see, e.g., Ashcroft v. Iqbal, 556 U. S. 662, 671 (2009), or where a district court certifies an open legal question to a court of appeals for determination, see, e.g., 28 U. S. C. §1292(b). But often statutes and rules require the parties to proceed to the end of a trial before obtaining appellate review. See, e.g., §1291.
The statutory provisions before us are a local [*29] species of this jurisdictional genus. In them, Congress limited interlocutory review of orders concerning arbitration in a way that favors arbitration. Consequently, §16(a) of the FAA will normally allow an immediate appeal where arbitration is denied, but §16(b) will normally require parties to wait until the end of the arbitration in order to bring legal questions about that proceeding to a court of appeals.
A couple of examples illustrate the point. Take first §4 of the FAA. Section 4 provides that a “court,” upon being satisfied that the parties have agreed to arbitrate a claim, “shall make an order directing the parties to proceed to arbitration in accordance with the terms of the agreement.” 9 U. S. C. §4. Section 16(a) of the FAA provides that a party may immediately appeal a district court order refusing to compel arbitration under §4, while §16(b) provides that a party generally may not immediately appeal a district court order compelling arbitration under §4. Compare §16(a)(1)(B) (“An appeal may be taken from” an order “denying a petition under section 4 of this title”) with §16(b)(2) (“[A]n appeal may not be taken from an interlocutory order . . . directing arbitration to proceed under section 4 of this title”).
Section 3 of the FAA provides another good example. Where a suit contains several claims, and the district [*30] court has determined that the parties agreed to arbitrate only a subset of those claims, §3 of the FAA provides that the district court must stay the litigation at the request of either party. See §3 (providing that a court, when referring claims for arbitration, “shall on application of one of the parties stay” the case “until such arbitration has been had”). The stay relieves the parties of the burden and distraction of continuing to litigate any remaining claims while the arbitration is ongoing. And true to the FAA’s proarbitration appellate scheme, §16(a) permits immediate appeals of district court orders refusing to enter a stay, while §16(b) generally prohibits immediate appeals of district court orders granting a stay. Compare §16(a)(1)(A) (“An appeal may be taken from” an order “refusing a stay of any action under section 3 of this title”) with §16(b)(1) (“[A]n appeal may not be taken from an interlocutory order . . . granting a stay of any action under section 3 of this title”).
I could go on. Section 16(a) of the FAA permits immediate appeal of an interlocutory order granting an injunction against arbitration, while §16(b) generally prohibits immediate appeal of an order refusing to enjoin an arbitration. Compare §16(a)(2) with §16(b)(4). Section 16(a) of the FAA permits immediate appeal of an order denying [*31] an application to compel arbitration pursuant to §206, while §16(b) generally prohibits immediate appeal of an order compelling arbitration pursuant to §206. Compare §16(a)(1)(C) with §16(b)(3). Et cetera.
The point, however, is that the appellate scheme of the FAA reflects Congress’ policy decision that, if a district court determines that arbitration of a claim is called for, there should be no appellate interference with the arbitral process unless and until that process has run its course.
With §16’s structure, and Congress’ policy in mind, we can turn to the facts of this case.
Respondent Frank Varela is an employee of petitioner Lamps Plus, Inc. At the outset of their employment relationship, Varela and Lamps Plus agreed to arbitrate employment-related claims. Varela later filed suit against Lamps Plus on behalf of himself and a class of Lamps Plus’ employees. Lamps Plus asked the District Court to compel arbitration. And the District Court granted Lamps Plus’ request. Despite having won the relief that it requested, Lamps Plus appealed the District Court’s order because Lamps Plus objected to the District Court’s conclusion that the parties’ agreement permitted arbitration on a classwide basis. The Court of Appeals [*32] affirmed the District Court’s judgment. And we granted Lamps Plus’ petition for certiorari to consider whether the Court of Appeals erred in so ruling.
But on those facts, I think that the Court lacks jurisdiction over Lamps Plus’ petition. When Lamps Plus responded to Varela’s lawsuit by seeking a motion to compel arbitration, and the District Court granted that motion, this case fell neatly into §16(b)’s description of unappealable district court orders under the FAA. The parties were obligated by the FAA to arbitrate their dispute without the expense and delay of further litigation. If, after arbitration, the parties were dissatisfied with the award or with the District Court’s arbitration related decisions, §16(a) of the FAA provides for an appeal at that later date. See §§16(a)(1)(D)-(E) (permitting appeals of orders confirming, modifying, or vacating an award); see also §16(a)(3) (permitting appeal of “a final decision with respect to an arbitration”). But, in the interim, §16(b) deprived the Court of Appeals of jurisdiction to hear any such complaint. See §§16(b)(1)-(4). I recognize that Lamps Plus is dissatisfied with the arbitration that the District Court ordered here. But the District Court’s order nonetheless granted the motion compelling [*33] arbitration, leaving Lamps Plus to bring its claim to an appellate court only after the arbitration is completed. See §16(b)(2). I believe we should enforce the statutory provisions that lead to this conclusion.
Lamps Plus offers three arguments in response. First, Lamps Plus suggests the Court of Appeals had jurisdiction over Lamps Plus’ appeal because the District Court order at issue here not only granted Lamps Plus’s motion to compel arbitration, but also granted Lamps Plus’ motion to dismiss the case. See Brief for Petitioners 29. Lamps Plus points out that §16(a) permits the appeal of “a final decision with respect to an arbitration.” 9 U. S. C. §16(a)(3). Lamps Plus reasons that, so long as a decision is final, it is appealable under the FAA.
I disagree because I do not believe that the District Court had the discretion to dismiss the case immediately after granting Lamps Plus’ motion to compel arbitration. Section 4 of the FAA permits a district court to compel the parties to arbitrate their claim, and §16(b)(2) explains that “an appeal may not be taken from an interlocutory order . . . directing arbitration to proceed under section 4 of this title.” Thus, the District Court order compelling arbitration was interlocutory and generally unappealable. [*34] As I have just explained, to read the statute any other way would contravene §16’s proarbitration appeal scheme by turning an interlocutory order that would have been unappealable under §16(b) of the Act into a dismissal order that is appealable under §16(a).
And because the order granting Lamps Plus’ motion to compel was interlocutory, the District Court’s dismissal of the case—in the very same order, see App. to Pet. for Cert. 23a—did not give the Court of Appeals jurisdiction over Lamps Plus’ appeal. An improper dismissal cannot create appellate jurisdiction to review an interlocutory order.
Our decision in Microsoft Corp. v. Baker, 582 U. S. ___ (2017), holds as much. The plaintiffs in Microsoft sought to appeal a district court order denying certification of a class. Under Federal Rule of Appellate Procedure 23(f), plaintiffs can ordinarily bring such an appeal only with the court of appeals’ permission. But the plaintiffs in Baker, who had been denied permission to appeal, tried to circumvent that denial by stipulating to a voluntary dismissal of their claims. The voluntary dismissal, they claimed, was an appealable “final decisio[n]” under 28 U. S. C. §1291. And in their appeal of the dismissal, they would be free to also seek review of the order denying class certification. We disagreed. As we [*35] explained there, to permit plaintiffs to “transform a tentative interlocutory order into a final judgment . . . simply by dismissing their claims with prejudice” would be to “undermine §1291’s firm finality principle, designed to guard against piecemeal appeals, and subvert the balanced solution Rule 23(f) put in place for immediate review of class-action orders.” Microsoft, supra, at ___, ___ (slip op., at 2, 16) (citation omitted).
The same reasoning applies here. Section 16(a)(3) of the FAA, like 28 U. S. C. §1291, creates appellate jurisdiction only over “final decisions.” Despite that jurisdictional limit, Lamps Plus, like the plaintiffs in Microsoft, seeks review of an interlocutory order. Like the plaintiffs in Microsoft, Lamps Plus attempts to obtain appellate review by “transform[ing]” an interlocutory order into a final decision. 582 U. S., at ___ (slip op., at 16). Like the plaintiffs in Microsoft, Lamps Plus has done so based on an order “purporting to end the litigation”—an order that Lamps Plus itself “persuade[d] a district court to issue.” Ibid. And like the plaintiffs in Microsoft, Lamps Plus does not “complain of the ‘final’ order that dismissed [the] case,” but instead seeks “review of only the inherently interlocutory order” compelling arbitration. Ibid. (alterations omitted). Therefore, like the Court [*36] in Microsoft, I would hold that Lamps Plus cannot, by securing an unlawful dismissal, find a way around the appellate jurisdiction scheme that Congress wrote into the FAA.
Second, Lamps Plus suggests that this Court has already decided that a district court order compelling arbitration and dismissing a plaintiff’s complaint creates no jurisdictional problem. Brief for Petitioners 29-30. Lamps Plus cites Green Tree Financial Corp.-Ala. v. Randolph, 531 U. S. 79 (2000), in support of that argument. And according to Lamps Plus, “this Court held in Randolph” that “when a district court orders arbitration and dismisses the plaintiff’s claims,” the order is “final” and therefore appealable under §16 of the FAA. Brief for Petitioners 29-30.
But Randolph does not control the jurisdictional aspect of this case. The Randolph Court explicitly reserved the question that we face now, stating: “Had the District Court entered a stay instead of a dismissal in this case, that order would not be appealable. 9 U. S. C. §16(b)(1). The question whether the District Court should have taken that course is not before us, and we do not address it.” Randolph, supra, at 87, n. 2 (emphasis added). Thus, although the Randolph Court stated that §16(a)(3) of the FAA permits appeals of final orders entered under the FAA, the Court did not decide whether a district court could convert [*37] an interlocutory, unappealable order under §16(b) into an appealable order under §16(a) by entering a dismissal instead of a stay. For that reason, Randolph does not answer the jurisdictional question here.
Third, and finally, Lamps Plus suggests that the Court of Appeals had jurisdiction because the District Court “effectively denied Lamps Plus’s motion to compel arbitration” when the District Court interpreted the arbitration agreement to permit class arbitration. Brief for Petitioners 31 (emphasis deleted). Leaning heavily on dicta from Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662 (2010), Lamps Plus argues that class arbitration is so “fundamental[ly]” different from individual arbitration that the fact that “the district court purported to grant Lamps Plus’s motion is not controlling.” Brief for Petitioners 31.
But Stolt-Nielsen cannot bear the weight Lamps Plus would place on it. We held in Stolt-Nielsen that a party may not be compelled to “submit to class arbitration unless there is a contractual basis for concluding that the party agreed to do so.” 559 U. S., at 684. We did not hold that class arbitration is not arbitration at all. And because class arbitration is arbitration, the District Court’s interpretation of [*38] Lamps Plus and Varela’s arbitration agreement to permit class arbitration could not create appellate jurisdiction over the District Court order compelling the parties to arbitrate their dispute. See 9 U. S. C. §16(b)(2) (prohibiting interlocutory appeals of district court orders “directing arbitration to proceed”).
Nor did we hold in Stolt-Nielsen (or anywhere else) that §16 of the FAA permits appeals of interlocutory orders directing arbitration to proceed, so long as the order incorporates some ruling that one party dislikes. If that were the rule, then §16’s limitations on appellate jurisdiction would be near meaningless. Consequently, the courts of appeals have—rightly, I believe—long recognized that they lack jurisdiction over appeals from orders that compel arbitration, “albeit not in the ‘first-choice’” manner of the party that moved to compel. Al Rushaid v. National Oilwell Varco, Inc., 814 F. 3d 300, 304 (CA5 2016). See also, e.g., Blue Cross Blue Shield of Mass., Inc. v. BCS Ins. Co., 671 F. 3d 635, 638 (CA7 2011) (concluding that the court of appeals lacked jurisdiction over an order compelling arbitration but denying a motion to direct arbitrators to “hold separate rather than consolidated proceedings”); Bushley v. Credit Suisse First Boston, 360 F. 3d 1149, 1154 (CA9 2004) (similar holding with respect to a request that arbitration take place before a different forum); Augustea Impb Et Salvataggi v. Mitsubishi Corp., 126 F. 3d 95, 98 (CA2 1997) (similar holding with respect to a request that the parties [*39] arbitrate in a different location). As one of those courts explained, “[p]ursuant to the plain meaning of th[e] statute . . . a party cannot appeal a district court’s order unless, at the end of the day, the parties are forced to settle their dispute other than by arbitration.” Id., at 99. And Lamps Plus’ characterization of the District Court’s order compelling arbitration as an “effectiv[e] den[ial]” of Lamps Plus’ motion “does not make it so.” Blue Cross Blue Shield, supra, at 637.
Consequently, I would hold that we lack jurisdiction over this case. But because the Court accepts jurisdiction and decides the substantive legal question before us, I shall do the same. And in respect to that question I agree with Justice Ginsburg and Justice Kagan, and I join their dissents.
Justice Sotomayor, dissenting.
I join Justice Ginsburg’s dissent in full and Part II of Justice Kagan’s dissent. 1 This Court went wrong years ago in concluding that a “shift from bilateral arbitration to class-action arbitration” imposes such “fundamental changes,” Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 686 (2010), that class-action arbitration “is not arbitration as envisioned by the” Federal Arbitration Act (FAA), AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 351 (2011). See, e.g., id., at 362-365 (Breyer, J., dissenting). A class action [*40] is simply “a procedural device” that allows multiple plaintiffs to aggregate their claims, 1 W. Rubenstein, Newberg on Class Actions § 1:1 (5th ed. 2011), “[f]or convenience . . . and to prevent a failure of justice,” Supreme Tribe of Ben-Hur v. Cauble, 255 U. S. 356, 363 (1921). Where, as here, an employment agreement provides for arbitration as a forum for all disputes relating to a person’s employment and the rules of that forum allow for class actions, an employee who signs an arbitration agreement should not be expected to realize that she is giving up access to that procedural device.
In any event, as Justice Kagan explains, the employment contract that Frank Varela signed went further. It states that “‘any and all disputes, claims or controversies arising out of or relating to[ ] the employment relationship between the parties[ ] shall be resolved by final and binding arbitration.’” Post, at 2 (quoting App. to Pet. for Cert. 24a). It adds that Varela and Lamps Plus “consent to the resolution by arbitration of all claims that may hereafter arise in connection with [Varela’s] employment.” Id., at 24a-25a. And it provides for arbitration “‘in accordance with’” the rules of the arbitral forum, which in turn allow for class arbitration. Post, at 3 (opinion of Kagan, J.) (citing [*41] App. to Pet. for Cert. 25a-26a). That is enough to persuade me that the contract was at least ambiguous as to whether Varela in fact agreed that no class-action procedures would be available in arbitration if he and his co-workers all suffered the same harm “relating to” and “in connection with” their “employment.” See id., at 24a-25a. And the court below was correct to turn to state law to resolve the ambiguity.
The Court today reads the FAA to pre-empt the neutral principle of state contract law on which the court below relied. I cannot agree. I also note that the majority reaches its holding without actually agreeing that the contract is ambiguous. See ante, at 5 (“[W]e defer to the Ninth Circuit’s interpretation and application of state law”). The concurrence, meanwhile, offers reasons to conclude that the contract unambiguously precludes class arbitration, see ante, at 1-2, and n. (opinion of Thomas, J.), which would avoid the need to displace state law at all. 2 This Court normally acts with great solicitude when it comes to the possible pre-emption of state law, see, e.g., Medtronic, Inc. v. Lohr, 518 U. S. 470, 485 (1996), but the majority today invades California contract law without pausing to address whether its incursion is necessary. [*42] Such haste is as ill advised as the new federal common law of arbitration contracts it has begotten.
Justice Kagan, with whom Justice Ginsburg and Justice Breyer join, and with whom Justice Sotomayor joins as to Part II, dissenting.
The Federal Arbitration Act (FAA or Act) requires courts to enforce arbitration agreements according to their terms. See ante, at 6. But the Act does not federalize basic contract law. Under the FAA, state law governs the interpretation of arbitration agreements, so long as that law treats other types of contracts in the same way. See DIRECTV, Inc. v. Imburgia, 577 U. S. ___, ___ (2015) (slip op., at 6). That well-established principle ought to resolve this case against Lamps Plus’s request for individual arbitration. In my view, the arbitration agreement Lamps Plus wrote is best understood to authorize arbitration on a classwide basis. But even if the Court is right to view the agreement as ambiguous, a plain-vanilla rule of contract interpretation, applied in California as in every other State, requires reading it against the drafter—and so likewise permits a class proceeding here. See Sandquist v. Lebo Auto., Inc., 1 Cal. 5th 233, 247, 376 P. 3d 506, 514 (2016). The majority can reach the opposite conclusion only by insisting that the FAA trumps that neutral state rule whenever its application [*43] would result in class arbitration. That holding has no basis in the Act—or in any of our decisions relating to it (including the heavily relied-on Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 686 (2010)). Today’s opinion is rooted instead in the majority’s belief that class arbitration “undermine[s] the central benefits of arbitration itself.” Ante, at 9. But that policy view—of a piece with the majority’s ideas about class litigation—cannot justify displacing generally applicable state law about how to interpret ambiguous contracts. I respectfully dissent.
From its very beginning, the arbitration agreement between Lamps Plus and Frank Varela announces its comprehensive scope. The first sentence states: “[T]he parties agree that any and all disputes, claims or controversies arising out of or relating to[ ] the employment relationship between the parties[ ] shall be resolved by final and binding arbitration.” App. to Pet. for Cert. 24a. The phrase “any and all disputes, claims, or controversies” encompasses both their individual and their class variants—just as any other general category (e.g., any and all chairs) includes all particular types (e.g., desk and reclining). So Varela’s class action (which [*44] arose out of or related to his employment) was a “dispute, claim or controversy” that belonged in arbitration.
The next paragraph continues in the same vein, by describing what Varela gave up by signing the agreement. “[A]rbitration,” the agreement says, “shall be in lieu of any and all lawsuits or other civil legal proceedings relating to my employment.” Ibid.; see ibid. (similarly waiving the right “to file a lawsuit or other civil action or proceeding”). That is the language of forum selection: Any and all actions (both individual and class) that I could once have brought in court, I am agreeing now to bring in arbitration. The provision carries no hint of consent to surrender altogether—in arbitration as well as court—the ability to bring a class proceeding.
Further on, the remedial and procedural terms of the agreement support reading it to authorize class arbitration. The arbitrator, according to the contract, may “award any remedy allowed by applicable law.” Id., at 26a. That sweeping provision easily encompasses classwide relief when the “any and all disputes” that the contract’s first sentence places in arbitration call for such remedies. 1 And under the agreement, the arbitration shall [*45] be conducted “in accordance with” the rules of either of two designated arbitration providers—both of which furnish rules for arbitrators to conduct class proceedings. Id., at 25a-26a; see, e.g., American Arbitration Assn., Supplementary Rules for Class Arbitrations (2011).
Even the section Lamps Plus cites in arguing that the agreement bars class arbitration instead points to the opposite conclusion. In describing what the agreement covers, one provision states: “The Company and I mutually consent to the resolution by arbitration of all claims or controversies (‘claims’), past, present or future that I may have against the Company.” App. to Pet. for Cert. 24a; see id., at 24a-25a (“Specifically, the Company and I mutually consent to the resolution by arbitration of all claims that may hereafter arise in connection with my employment”). Lamps Plus (along with the concurrence, see ante, at 1-2 (opinion of Thomas, J.)) highlights “th[e] repeated use of singular personal pronouns” there, contending that it is incompatible with a form of arbitration that also involves other parties’ claims. Brief for Petitioners 17. But the use of the first person singular merely reflects that the agreement is bilateral in [*46] nature—between Varela and Lamps Plus. Those pronouns do not resolve whether one of those parties (“I”) can bring to arbitration class disputes, as well as individual disputes, relating to his employment. The part of the quoted section addressing that question is instead the phrase “all claims or controversies.” And that phrase supplies the same answer as the agreement’s other provisions. For it too is broad enough to cover both individual and class actions—the ones Varela brings alone and the ones he shares with co-workers. 2
Suppose, though, you think that my view of the agreement goes too far. Maybe you aren’t sure whether the phrase “any and all disputes, claims or controversies” must be read to include class “disputes, claims or controversies.” Or maybe you wonder whether the surrounding “I” and “my” references limit that phrase’s scope, rather than merely referring to one of the contract’s signatories. In short, you can see reasonable arguments on both sides of the interpretive dispute—for allowing, but also for barring, class arbitration. You are then in the majority’s position, “accept[ing]” the arbitration agreement as “ambiguous.” Ante, at 5. What should follow?
Under California [*47] law (which applies unless preempted) the answer is clear: The agreement must be read to authorize class arbitration. That is because California—like every other State in the country—applies a default rule construing “ambiguities” in contracts “against their drafters.” Sandquist, 1 Cal. 5th, at 247, 376 P. 3d, at 514; see Cal. Civ. Code Ann. §1654 (West 2011); see also Brief for Contract Law Scholars as Amici Curiae 10-12, and n. 4 (listing decisions from all 50 States applying that rule). This anti-drafter canon—which “applies with peculiar force” to form contracts like Lamps Plus’s—promotes clarity in contracting by resolving ambiguities against the party who held the pen. Sandquist, 1 Cal. 5th, at 248, 376 P. 3d, at 514 (quoting Graham v. Scissor-Tail, Inc., 28 Cal. 3d 807, 819, n. 16, 623 P. 2d 165, 172, n. 16 (1981)); see Ayres & Gertner, Filling Gaps in Incomplete Contracts, 99 Yale L. J. 87, 91, 105, n. 80 (1989). And the rule makes quick work of interpreting the arbitration agreement here. Lamps Plus drafted the agreement. It therefore had the opportunity to insert language expressly barring class arbitration if that was what it wanted. It did not do so. It instead (at best) left an ambiguity about the availability of class arbitration. So California law holds that Lamps Plus cannot now claim the benefit of the doubt as to the agreement’s meaning. Even the majority does not dispute that point. See ante, at 5, 9.
And contrary [*48] to the rest of the majority’s opinion, 3 the FAA contemplates that such a state contract rule will control the interpretation of arbitration agreements. Under the FAA, courts must “enforce arbitration agreements according to their terms.” Epic Systems Corp. v. Lewis, 584 U. S. ___, ___ (2018) (internal quotation marks omitted) (slip op., at 5); see 9 U. S. C. §4 (requiring that “arbitration proceed in the manner provided for in such agreement”). But the construction of those contractual terms (save for in limited circumstances, addressed below) is “a question of state law, which this Court does not sit to review.” Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 474 (1989). The Court has made that crucial point many times. Nothing in the FAA (as contrasted to today’s majority opinion) “purports to alter background principles of state contract law regarding” the scope or content of agreements. Arthur Andersen LLP v. Carlisle, 556 U. S. 624, 630 (2009). Or again: When ruling on an arbitration agreement’s meaning, courts “should apply ordinary state-law principles.” First Options of Chicago, Inc. v. Kaplan, 514 U. S. 938, 944 (1995). Or yet again: The interpretation of such an agreement is “a matter of state law to which we defer.” DIRECTV, Inc., 577 U. S., at ___ (slip op., at 6). In short, the FAA does not federalize contract law.
Except when state contract law discriminates against arbitration agreements. As this Court has explained, the FAA came about because courts had shown themselves [*49] “unduly hostile to arbitration.” Epic Systems, 584 U. S., at ___ (slip op., at 5). To remedy that problem, Congress built an “equal-treatment principle” into the Act, requiring courts to “place arbitration agreements on an equal footing with other contracts.” Kindred Nursing Centers L. P. v. Clark, 581 U. S. ___, ___ (2017) (slip op., at 4); AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 339 (2011) (internal quotation marks omitted); see 9 U. S. C. §2 (making arbitration agreements “valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract”). So any state rule treating arbitration agreements worse than other contracts “stand[s] as an obstacle” to achieving the Act’s purposes—and is preempted. Concepcion, 563 U. S., at 343. That means the FAA displaces any state rule discriminating on its face against arbitration. See id., at 341. And the Act likewise preempts any more subtle law “disfavoring contracts that (oh so coincidentally) have the defining features of arbitration agreements.” Kindred Nursing, 581 U. S., at ___ (slip op., at 5). What matters, as this Court reiterated last Term, is whether the state law in question “target[s]” arbitration agreements, blatantly or covertly, for substandard treatment. Epic Systems, 584 U. S., at ___ (slip op., at 7). 4 When the law does so, it cannot operate; when, conversely, it treats arbitration agreements the same as all other contracts, the FAA leaves it alone.
Here, California’s anti-drafter rule is as [*50] even-handed as contract rules come. It does not apply only to arbitration contracts. Nor does it apply (as the rule we rejected in Concepcion did) only a tad more broadly to “dispute-resolution contracts,” pertaining to both arbitration and litigation. 563 U. S., at 341 (holding that a ban on collective-action waivers in those contracts worked to “disfavor[ ] arbitration”). Instead, the anti-drafter rule, as even the majority admits, applies to every conceivable type of contract—and treats each identically to all others. See Sandquist, 1 Cal. 5th, at 248, 376 P. 3d, at 514 (“This general principle of contract interpretation applies equally to the construction of arbitration provisions”); ante, at 9-10. And contrary to what the majority is left to insist, the rule does not “target arbitration” by “interfer[ing] with [one of its] fundamental attributes”—i.e., its supposed individualized nature. Ante, at 11 (internal quotation marks omitted); see ante, at 7-9. The anti-drafter rule (again, quite unlike Concepcion’s ban on class-action waivers) takes no side—favors no outcome—as between class and individualized dispute resolution. All the anti-drafter rule asks about is who wrote the contract. So if, for example, Varela had drafted the agreement here, [*51] the rule would have prevented, rather than permitted, class arbitration. 5 Small wonder, then, that this Court has itself used the anti-drafter canon to interpret an arbitration agreement. See Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U. S. 52, 62 (1995) (construing an ambiguous arbitration agreement against the drafter’s interest). In that case (as properly in any other), the rule’s through-and-through neutrality made preemption unthinkable. 6
So this case should come out Varela’s way even if the agreement is ambiguous. To repeat the simple logic applicable here: Under the FAA, state law controls the interpretation of arbitration agreements unless that law discriminates against arbitration; the anti-drafter default rule is subject to no such objection; the rule therefore compels this Court to hold that the agreement here authorizes class arbitration. That the majority thinks the contract, as so read, seriously disadvantages Lamps Plus, see ante, at 7-8, is of no moment (any more than if state law had instead construed the contract to produce adverse consequences for Varela). The FAA was enacted to protect against judicial hostility toward arbitration agreements. See supra, at 6. But the Act provides no warrant for courts to disregard neutral state [*52] law in service of ensuring that those agreements give defendants the best terms possible. Or said otherwise: Nothing in the FAA shields a contracting party, operating against the backdrop of impartial state law, from the consequences of its own drafting decisions. How, then, could the majority go so wrong?
Stolt-Nielsen offers the majority no excuse: Far from “control[ling]” this case, ante, at 8, that decision addressed a different situation—and explicitly reserved decision of the question here. In Stolt-Nielsen, the contracting parties entered into a formal stipulation that “they had not reached any agreement on the issue of class arbitration.” 559 U. S., at 673. The case thus involved not the mere absence of express language about class arbitration, but a joint avowal that the parties had never resolved the issue. Facing that oddity, an arbitral panel compelled class arbitration based solely on its “own conception of sound policy.” Id., at 675; see id., at 676 (“[T]he panel did [nothing] other than impose its own policy preference”). This Court rejected the panel’s decision for that reason, holding that a party need not “submit to class arbitration unless there is a contractual basis for concluding that the party agreed [*53] to do so.” Id., at 684. But the Court went no further. In particular, it did not resolve cases like this one, where a neutral interpretive rule (even if not an express term) enables an adjudicator to determine a contract’s meaning. To the contrary, the Court disclaimed any view on that question. Yes, the Court held, “a contractual basis” was needed for class arbitration. Ibid. (emphasis added). But given the panel’s reliance on policy alone, the Court explained that it had “no occasion to decide what contractual basis” was required. Id., at 687, n. 10 (emphasis added); see Oxford Health Plans LLC v. Sutter, 569 U. S. 564, 571 (2013) (“We overturned the arbitral decision [in Stolt-Nielsen] because it lacked any contractual basis for ordering class procedures,” not because it relied on an inadequate one).
Indeed, parts of Stolt-Nielsen—as well as later decisions—indicate that applying the anti-drafter rule to ambiguous language provides a sufficient contractual basis for class arbitration. In Stolt-Nielsen, we faulted the arbitrators for failing to inquire whether the relevant law “contain[ed] a default rule” that would construe an arbitration clause “as allowing class arbitration in the absence of express consent.” 559 U. S., at 673 (internal quotation marks omitted). We thus implied that [*54] such a default rule—like the anti-drafter canon here—can operate to authorize class arbitration when an agreement’s language is ambiguous. And that is just how Concepcion (the other decision the majority relies on, see ante, at 7-8, 10-12) understood Stolt-Nielsen’s reasoning. Said Concepcion: We held in Stolt-Nielsen “that an arbitration panel exceeded its power [by] imposing class procedures based on policy judgments rather than the arbitration agreement itself or some background principle of contract law that would affect its interpretation.” 563 U. S., at 347 (emphasis added); see Oxford Health, 569 U. S., at 571 (similarly noting that Stolt-Nielsen criticized the arbitrators for failing to consider whether a “default rule” resolved the class arbitration question (internal quotation marks omitted)). The Court has thus (rightly) viewed the use of default rules as a run-of-the-mill aspect of contract interpretation, which (so long as neutrally applied) can support class arbitration.
And nothing particular to the anti-drafter rule justifies a different conclusion, as the majority elsewhere suggests, see ante, at 9-11. 7 That rule, proclaims the majority, reflects “public policy considerations,” rather than “help[ing] to interpret [*55] the meaning of a term” as understood by the parties. Ante, at 10. The majority here notes that some commentators have viewed some equitable factors as supporting the rule, see ante, at 9-10—which is no doubt right. But see 11 R. Lord, Williston on Contracts §30:1, p. 11 (4th ed. 2012) (Williston) (stating that the rule is not justified by public interest considerations). But if the majority means to claim—as it must to prove its point—that the anti-drafter rule has no concern with what “the part[ies] agreed to,” Stolt-Nielsen, 559 U. S., at 684, then the majority is flat-out wrong. From an ex ante perspective, the rule encourages the drafter to set out its intent in clear contractual language, for the other party then to see and agree to. See Ayres & Gertner, 99 Yale L. J., at 91, 105, n. 80 (stating the modern view); 2 W. Blackstone, Commentaries on the Laws of England 380 (1766) (anticipating that view by 200-plus years). And from an ex post perspective, the rule enables an interpreter to resolve any remaining uncertainty in line with the parties’ likely expectations. See 11 Williston §30:1, at 11. Consider this very contract. Lamps Plus, knowing about the anti-drafter rule, still chose not to include a term prohibiting class arbitration. And Varela, [*56] seeing only the language sending “any and all disputes, claims, or controversies” to arbitration, had no reason to think class disputes barred. Cf. ibid. (“[T]he party addressed will understand ambiguous language in the sense most favorable to itself”). The upshot is that the rule (as this Court recognized in another arbitration case) protects against “unintended” consequences. Mastrobuono, 514 U. S., at 63.
And even if that were not so evident, the FAA does not empower a court to halt the operation of such a garden-variety principle of state law. Nothing in the Act’s text requires the displacement of state contract rules, as the majority implicitly concedes. See ante, at 6. Nor do the Act’s purposes, so long as the state rule (as is true here) extends to all contracts alike, without disfavoring arbitration. See supra, at 6-7. The idea that the FAA blocks a state rule satisfying that standard because (a court finds) the rule has too much “public policy” in it comes only from the majority’s collective mind. That approach disrespects the preeminent role of the States in designing and enforcing contract rules. It discards a universally accepted principle of contract interpretation in favor of unsupported assertions [*57] about what the parties must have (or could not possibly have) consented to. It subordinates authoritative state law to (at most) the impalpable emanations of federal policy, impossible to see except in just the right light. 8 For that reason, it would never have graced the pages of the U. S. Reports save that this case involves . . . class proceedings.
The heart of the majority’s opinion lies in its cataloging of class arbitration’s many sins. See ante, at 7-8. In that respect, the opinion comes from the same place as (though goes a step beyond) this Court’s prior arbitration decisions. See, e.g., Concepcion, 563 U. S., at 350 (lamenting that class arbitration “greatly increases risks to defendants” by “aggregat[ing] and decid[ing] at once” the “damages allegedly owed to tens of thousands of potential claimants”); Epic Systems, 584 U. S., at __ (slip op., at 8) (similarly bemoaning the greater costs and complexity of class proceedings). The opinion likewise has more than a little in common with this Court’s efforts to pare back class litigation. See, e.g., Comcast Corp. v. Behrend, 569 U. S. 27 (2013); Wal-Mart Stores, Inc. v. Dukes, 564 U. S. 338, 348-360 (2011). In this case, the result is to disregard the actual contract the parties signed. And to dismiss the neutral and commonplace default rule that would construe that contract against the drafting [*58] party. No matter what either requires, the majority will prohibit class arbitration. Does that approach remind you of anything? It should. Here (again) is Stolt-Nielsen as Concepcion described it: The panel exceeded its authority by “imposing class procedures based on policy judgments rather than the arbitration agreement itself or some background principle of contract law that would affect its interpretation.” 563 U. S., at 347; see supra, at 10. Substitute “foreclosing” for “imposing” and that is what the Court today has done. It should instead—as the FAA contemplates—have left the parties’ agreement, as construed by state law, alone.
1 Justice Breyer repeatedly refers to the order in this case as “interlocutory,” post, at 5-7 (dissenting opinion), but—as the language quoted above makes clear—Randolph expressly held that such an order is “final” under the FAA. Justice Breyer also claims that Randolph “explicitly reserved the [jurisdictional] question that we face now,” post, at 7, but Randolph reserved a different question. In that case, the District Court had denied a motion to stay. We noted that, if the District Court had entered a stay instead of dismissing the case, an appeal would have been barred by 9 U. S. C. §16(b)(1). That said, we expressly refrained from addressing whether the District Court should have granted the stay. See 531 U. S., at 87, n. 2. That is the question we reserved. Justice Breyer would have us take up that question today, post, at 3, 7, but there is no basis for doing so. The FAA provides that a district court “shall on application of one of the parties stay” the case pending the arbitration. 9 U. S. C. §3 (emphasis added). Here, no party sought a stay. Thus, Justice Breyer’s jurisdictional analysis is premised on two events that did not happen—a District Court ruling that was never issued denying a stay request that was never made. In short, Justice Breyer has written an opinion for a case other than the one before us.
2 And contrary to Varela’s contention, Brief for Respondent 14-15, and Justice Breyer’s dissent, post, at 6-7, this is hardly a case like Microsoft Corp. v. Baker, 582 U. S. ___ (2017). There, we held that plaintiffs cannot generate a final appealable order by voluntarily dismissing their claim. Here, Lamps Plus was the defendant, and the District Court compelled class arbitration over the company’s vigorous opposition.
3 Justice Kagan offers her own interpretation of the contract, concludes that it unambiguously authorizes class arbitration, post, at 2-4, and criticizes us for “disregard[ing] the actual contract the parties signed,” post, at 14. Justice Sotomayor, on the other hand, concludes that the contract is ambiguous about class arbitration but criticizes us for treating the contract as . . . ambiguous. Post, at 2-3 (dissenting opinion). Again, we simply follow this Court’s ordinary approach, which “accord[s] great deference” to the courts of appeals in their interpretation of state law. Expressions Hair Design, 581 U. S., at ___, (slip op., at 7) (quoting Pembaur v. Cincinnati, 475 U. S. 469, 484 n. 13 (1986) (collecting cases)).
4 This Court has not decided whether the availability of class arbitration is a so-called “question of arbitrability,” which includes these gateway matters. Oxford Health Plans LLC v. Sutter, 569 U. S. 564, 569, n. 2 (2013). We have no occasion to address that question here because the parties agreed that a court, not an arbitrator, should resolve the question about class arbitration.
5 Varela and Justice Kagan contend that our use of contra proferentem in Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U. S. 52, 57 (1995), establishes that the rule is not preempted by the FAA. Brief for Respondent 33-35; post, at 8 (dissenting opinion). In Mastrobuono, however, we had no occasion to consider a conflict between the FAA and contra proferentem because both rules led to the same result. Our holding was primarily based on the FAA policy favoring arbitration, 514 U. S., at 62, and only after establishing that did we apply contra proferentem, noting that the rule was “well suited to the facts of this case,” id., at 63. See also EEOC v. Waffle House, Inc., 534 U. S. 279, 293, n. 9 (2002) (explaining that Mastrobuono resolved an ambiguous provision by “read[ing] the agreement to favor arbitration under the FAA rules”).
* Two intermediate California courts have held, based on similar language, that an arbitration agreement did not authorize class arbitration. See Nelsen v. Legacy Partners Residential, Inc., 207 Cal. App. 4th 1115, 1129-1131, 144 Cal. Rptr. 3d 198, 210-211 (2012); Kinecta Alternative Financial Solutions, Inc. v. Superior Court of Los Angeles Cty., 205 Cal. App. 4th 506, 517-519, 140 Cal. Rptr. 3d 347, 356-357 (2012), disapproved of on other grounds by Sandquist v. Lebo Automotive, Inc., 1 Cal. 5th 233, 376 P. 3d 506 (2016).
1 I am not persuaded at this point that the Court of Appeals lacked jurisdiction over this case, and for that reason I do not join Justice Breyer’s dissenting opinion. Nevertheless, I believe that Justice Breyer’s opinion raises weighty issues that are worthy of further consideration if raised in the appropriate circumstances in the lower federal courts.
2 The majority notes that I criticize it for not checking for such an off-ramp while being unable to take one myself. See ante, at 6, n. 3. But the majority never suggests that it shares my rationale as to why the contract is ambiguous. In other words, the reasons that I reach the issue that the majority decides say nothing about whether the majority would get there itself, short of deferring to the lower federal court.
1 In discussing another arbitration provision, this Court identically reasoned: “[I]t would seem sensible to interpret the ‘all disputes’ and ‘any remedy or relief’ phrases to indicate, at a minimum, an intention to resolve through arbitration any dispute that would otherwise be settled in a court, and to allow the chosen dispute resolvers to award the same varieties and forms of damages or relief as a court would be empowered to award.” Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U. S. 52, 61-62, n. 7 (1995) (internal quotation marks omitted). Here, that means sending to arbitration (among other things) class disputes seeking class relief.
2 An additional semantic point that Lamps Plus makes essentially concedes my reading of the agreement. At oral argument, Lamps Plus acknowledged that the contract would authorize class arbitration if it provided that Varela could bring to the arbitral forum any “lawsuits,” rather than any “claims,” he had or could have brought against the company. Tr. of Oral Arg. 31-32. The idea is apparently that suits can be classwide while claims must be personal. But even assuming (without accepting) that is so, the agreement never speaks only of “claims.” Even when that word appears alone (rather than alongside “disputes” or “controversies”), it in fact functions as a defined term meaning “claims or controversies.” See App. to Pet. for Cert. 24a (referring to “all claims or controversies (‘claims’)”). And if lawsuits are not necessarily personal (as Lamps Plus admits), then neither are controversies. So by Lamps Plus’s own reasoning, Varela should be able to bring to arbitration all controversies (including classwide ones) he had or could have brought to court.
3 I say “the majority’s,” but although five Justices have joined today’s opinion, only four embrace its reasoning. See n. 8, infra.
4 In its many decades of FAA caselaw, the Court has preempted state law in just one other, “narrow” circumstance: Whatever state law might say, courts must find “clear and unmistakable evidence” before deciding that an agreement authorizes an arbitrator to decide a so-called “question of arbitrability.” Green Tree Financial Corp. v. Bazzle, 539 U. S. 444, 452 (2003) (plurality opinion) (internal quotation marks and alterations omitted); Oxford Health Plans LLC v. Sutter, 569 U. S. 564, 569, n. 2 (2013). As the majority acknowledges, that requirement is not at issue here because Varela and Lamps Plus agreed that a judge should decide the availability of class arbitration (even assuming that question is one of arbitrability). See ante, at 9, n. 4.
5 Similarly, if Lamps Plus, as the agreement’s author, had wanted class arbitration (perhaps because that would resolve many related cases at once) and Varela had resisted it (perhaps because he thought his case better than the others), the anti-drafter rule would have prevented, rather than permitted, class arbitration.
6 Our decision in DIRECTV, Inc. v. Imburgia, 577 U. S. ___ (2015), also assumed that a court may generally apply a State’s anti-drafter rule to arbitration agreements. It was only because the court there applied that rule to an unambiguous contract—in contrast to what the court would have done in a non-arbitration case—that we reversed its decision. See id., at ___, ___ (slip op., at 7, 10).
7 The majority actually sends conflicting signals about the extent to which its holding extends beyond the anti-drafter rule to other background principles that serve to discern the meaning of ambiguous contract language. Many of the majority’s statements indicate that any tool for resolving contractual ambiguity is forbidden if it leads to class arbitration. See, e.g., ante, at 6 (stating flatly that “an ambiguous agreement [cannot] provide the necessary ‘contractual basis’ for compelling class arbitration”). But the part of the opinion focusing on the anti-drafter rule suggests that today’s holding applies to only a subset of contract default rules—to wit, those (supposedly) sounding in “public policy considerations.” See ante, at 9-11. On that theory of the decision, courts and arbitrators will have to work out over time which interpretive principles fall within that category. The majority’s own flawed analysis of the anti-drafter canon, see infra, at 11-12, indicates the perils of that undertaking.
8 Given this extraordinary displacement of state law—which, as I have shown, no precedent commands, see supra, at 9-10—I must admit to not understanding Justice Thomas’s full concurrence in today’s opinion. See ante, at 2 (expressing “skeptic[ism]” about the majority’s reasoning but joining its opinion out of a (misplaced) respect for precedent). I would think the opinion a hard pill to swallow for someone who believes that any implied preemption “leads to the illegitimate—and thus, unconstitutional—invalidation of state laws.” Wyeth v. Levine, 555 U. S. 555, 604 (2009) (Thomas, J., concurring in judgment); see, e.g., Bates v. Dow Agrosciences LLC, 544 U. S. 431, 459 (2005) (Thomas, J., concurring in judgment in part) (“[P]re-emption analysis is not a freewheeling judicial inquiry into whether a state statute is in tension with federal objectives” (internal quotation marks and alteration omitted)).
Share this article: