AT&T MOBILITY LLC v. CONCEPCION et ux.

No. 09-893.

Argued November 9, 2010 —

Decided April 27, 2011

*334& alia, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Auto, JJ., joined. Thomas, J, filed a concurring opinion, post, p. 352. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined, post, p. 357.

Andrew J. Pincus argued the cause for petitioner. With him on the briefs were Kenneth S. Getter, Evan M. Tager, *335Archis A. Parasharami, Kevin Ranlett, Donald M. Falk, and Neal Berinhout.

Deepak Gupta argued the cause for respondents. With him on the brief were Scott L. Nelson, Gregory A. Beck, Kirk B. Hulett, Craig M. Nicholas, and Alex M. Tomasevic *

*

Briefs of amici curiae urging reversal were filed for the State of South Carolina et al. by Henry D. McMaster, Attorney General of South Carolina, James Emory Smith, Jr., Assistant Deputy Attorney General, and Mark L. Shurtleff, Attorney General of Utah; for the American Bankers Association et al. by Alan S. Kaplinsky, Jeremy T. Rosenblum, and Mark J. Levin; for the Center for Class Action Fairness by Brian P. Brooks; for the Chamber of Commerce of the United States of America by Roy T. Englert, Jr., Robin S. Conrad, and Amar D. Sarwal; for CTIA — The Wireless Association by Paul D. Clement and Michael F. Altschul; for DIRECTV, Inc., et al. by Jeffrey S. Davidson; for Distinguished Law Professors by Andrew G. McBride; for DRI — The Voice of the Defense Bar by Kevin C. Newsom and John R. Kouris; for the Equal Employment Advisory Council by Rae T. Vann; for the New England Legal Foundation by Benjamin G. Robbins and Martin J. Newhouse; and for the Pacific Legal Foundation by Deborah J. La Fetra.

Briefs of amici curiae urging affirmance were filed for the State of Illinois et al. by Lisa Madigan, Attorney General of Illinois, Michael A. Scodro, Solicitor General, and Jane Elinor Note, Deputy Solicitor General, and by the Attorneys General for their respective jurisdictions as follows: Peter J. Nickles of the District of Columbia, Douglas F. Gansler of Maryland, Lori Swanson of Minnesota, Steve Bullock of Montana, Gary K. King of New Mexico, Robert E. Cooper, Jr., of Tennessee, and William H. Sorrell of Vermont; for the American Antitrust Institute by Richard M. Brunell and Albert A. Foer; for the American Association for Justice by Andre M. Mura and John Vail; for Civil Procedure and Complex Litigation Professors by William B. Rubenstein, Theodore Eisenberg, John Leubsdorf, Arthur R. Miller, and Judith Resnik; for the Constitutional Accountability Center by Douglas T. Kendall and Elizabeth B. Wydra; for Contracts Professors by Peter K. Stris; for Federal Jurisdiction Professors by Stephen I. Vladeck and Michael J. Quirk; for the Lawyers’ Committee for Civil Rights Under Law et al. by Sarah Crawford, Terisa E. Chaw, Catherine Ruckelshaus, Rebecca Hamburg, and Sharyn A Tejani; for the Legal Aid Society of the District of Columbia et al. by Bonnie I. RobinVergeer, Michael D. Donovan, and James C. Sturdevant; for the NAACP Legal Defense & Educational Fund, Inc., by John Payton, Debo P. Adegbile, and Joshua Civin; for the National Academy of Arbitrators by James *336A. Feldman; for the National Workrights Institute by Theodore J. St. Antoine and Lewis Maltby; for Marygrace Coneff et al. by Leslie A. Bailey, Arthur H. Bryant, F. Paul Bland, Jr., and Matthew Wessler; and for Jonathan C. Kaltwasser by Joseph N. Kravec, Jr.

Biro N. Aragaki filed a brief for Arbitration Professors as amici curiae.

*336Justice Scalia

delivered the opinion of the Court.

Section 2 of the Federal Arbitration Act (FAA) makes agreements to arbitrate “valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U. S. C. §2. We consider whether the FAA prohibits States from conditioning the enforceability of certain arbitration agreements on the availability of classwide arbitration procedures.

I

In February 2002, Vincent and Liza Concepcion entered into an agreement for the sale and servicing of cellular telephones with AT&T Mobility LLC (AT&T).1 The contract provided for arbitration of all disputes between the parties, but required that claims be brought in the parties’ “individual capacity, and not as a plaintiff or class member in any purported class or representative proceeding.” App. to Pet. for Cert. 61a.2 The agreement authorized AT&T to make unilateral amendments, which it did to the arbitration provision on several occasions. The version at issue in this case reflects revisions made in December 2006, which the parties agree are controlling.

The revised agreement provides that customers may initiate dispute proceedings by completing a one-page Notice of Dispute form available on AT&T’s Web site. AT&T may *337then offer to settle the claim; if it does not, or if the dispute is not resolved within 30 days, the customer may invoke arbitration by filing a separate Demand for Arbitration, also available on AT&T’s Web site. In the event the parties proceed to arbitration, the agreement specifies that AT&T must pay all costs for nonfrivolous claims; that arbitration must take place in the county in which the customer is billed; that, for claims of $10,000 or less, the customer may choose whether the arbitration proceeds in person, by telephone, or based only on submissions; that either party may bring a claim in small claims court in lieu of arbitration; and that the arbitrator may award any form of individual relief, including injunctions and presumably punitive damages. The agreement, moreover, denies AT&T any ability to seek reimbursement of its attorney’s fees, and, in the event that a customer receives an arbitration award greater than AT&T’s last written settlement offer, requires AT&T to pay a $7,500 minimum recovery and twice the amount of the claimant’s attorney’s fees.3

The Concepcions purchased AT&T service, which was advertised as including the provision of free phones; they were not charged for the phones, but they were charged $30.22 in sales tax based on the phones’ retail value. In March 2006, the Concepcions filed a complaint against AT&T in the United States District Court for the Southern District of California. The complaint was later consolidated with a putative class action alleging, among other things, that AT&T had engaged in false advertising and fraud by charging sales tax on phones it advertised as free.

In March 2008, AT&T moved to compel arbitration under the terms of its contract with the Concepcions. The Concepcions opposed the motion, contending that the arbitration agreement was unconscionable and unlawfully exculpatory *338under California law because it disallowed classwide procedures. The District Court denied AT&T’s motion. It described AT&T’s arbitration agreement favorably, noting, for example, that the informal dispute-resolution process was “quick, easy to use,” and likely to “promp[t] full or . .. even excess payment to the customer without the need to arbitrate or litigate”; that the $7,500 premium functioned as “a substantial inducement for the consumer to pursue the claim in arbitration” if a dispute was not resolved informally; and that consumers who were members of a class would likely be worse off. Laster v. T-Mobile USA, Inc., 2008 WL 5216255, *11-*12 (SD Cal., Aug. 11,2008). Nevertheless, relying on the California Supreme Court’s decision in Discover Bank v. Superior Court, 36 Cal. 4th 148, 113 P. 3d 1100 (2005), the court found that the arbitration provision was unconscionable because AT&T had not shown that bilateral arbitration adequately substituted for the deterrent effects of class actions. Laster, 2008 WL 5216255, *14.

The Ninth Circuit affirmed, also finding the provision unconscionable under California law as announced in Discover Bank. Laster v. AT&T Mobility LLC, 584 P. 3d 849, 855 (2009). It also held that the Discover Bank rule was not pre-empted by the PAA because that rule was simply “a refinement of the unconscionability analysis applicable to contracts generally in California.” 584 F. 3d, at 857 (internal quotation marks omitted). In response to AT&T’s argument that the Concepcions’ interpretation of California law discriminated against arbitration, the Ninth Circuit rejected the contention that “ ‘class proceedings will reduce the efficiency and expeditiousness of arbitration’” and noted that ‘“Discover Bank placed arbitration agreements with class action waivers on the exact same footing as contracts that bar class action litigation outside the context of arbitration.’ ” Id., at 858 (quoting Shroyer v. New Cingular Wireless Services, Inc., 498 F. 3d 976, 990 (CA9 2007)).

We granted certiorari, 560 U. S. 923 (2010).

*339h-i HH

The FAA was enacted in 1925 in response to widespread judicial hostility to arbitration agreements. See Hall Street Associates, L. L. C. v. Mattel, Inc., 552 U. S. 576, 581 (2008). Section 2, the “primary substantive provision of the Act,” Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U. S. 1, 24 (1983), provides, in relevant part, as follows:

“A written provision in any maritime transaction or a contract evidencing a transaction involving commerce to settle by arbitration a controversy thereafter arising out of such contract or transaction ... shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U.S. C. §2.

We have described this provision as reflecting both a “liberal federal policy favoring arbitration,” Moses H. Cone, supra, at 24, and the “fundamental principle that arbitration is a matter of contract,” Rent-A-Center, West, Inc. v. Jackson, 561 U. S. 63, 67 (2010). In line with these principles, courts must place arbitration agreements on an equal footing with other contracts, Buckeye Check Cashing, Inc. v. Cardegna, 546 U. S. 440, 443 (2006), and enforce them according to their terms, Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 478 (1989).

The final phrase of § 2, however, permits arbitration agreements to be declared unenforceable “upon such grounds as exist at law or in equity for the revocation of any contract.” This saving clause permits agreements to arbitrate to be invalidated by “generally applicable contract defenses, such as fraud, duress, or unconseionability,” but not by defenses that apply only to arbitration or that derive their meaning from the fact that an agreement to arbitrate is at issue. Doctor’s Associates, Inc. v. Casarotto, 517 U. S. 681, 687 (1996); see also Perry v. Thomas, 482 U. S. 483, 492-493, n. 9 (1987). *340The question in this case is whether § 2 pre-empts California’s rule classifying most collective-arbitration waivers in consumer contracts as unconscionable. We refer to this rule as the Discover Bank rule.

Under California law, courts may refuse to enforce any contract found “to have been unconscionable at the time it was made,” or may “limit the application of any unconscionable clause.” Cal. Civ. Code Ann. § 1670.5(a) (West 1985). A finding of unconscionability requires “a ‘procedural’ and a ‘substantive’ element, the former focusing on ‘oppression’ or ‘surprise’ due to unequal bargaining power, the latter on ‘overly harsh’ or ‘one-sided’ results.” Armendariz v. Foundation Health Psychcare Servs., Inc., 24 Cal. 4th 83, 114, 6 P. 3d 669, 690 (2000); accord, Discover Bank, 36 Cal. 4th, at 159-161, 113 P. 3d, at 1108.

In Discover Bank, the California Supreme Court applied this framework to class-action waivers in arbitration agreements and held as follows:

“[W]hen the waiver is found in a consumer contract of adhesion in a setting in which disputes between the contracting parties predictably involve small amounts of damages, and when it is alleged that the party with the superior bargaining power has carried out a scheme to deliberately cheat large numbers of consumers out of individually small sums of money, then . . . the waiver becomes in practice the exemption of the party ‘from responsibility for [its] own fraud, or willful injury to the person or property of another.’ Under these circumstances, such waivers are unconscionable under California law and should not be enforced.” Id., at 162-163, 113 P. 3d, at 1110 (quoting Cal. Civ. Code Ann. § 1668).

California courts have frequently applied this rule to find arbitration agreements unconscionable. See, e. g., Cohen v. DIRECTV, Inc., 142 Cal. App. 4th 1442, 1451-1453, 48 Cal. Rptr. 3d 813, 819-821 (2006); Klussman v. Cross Country *341Bank, 134 Cal. App. 4th 1283, 1297, 36 Cal Rptr. 3d 728, 738-739 (2005); Aral v. EarthLink, Inc., 134 Cal. App. 4th 544, 556-557, 36 Cal. Rptr. 3d 229, 237-239 (2005).

Ill

A

The Concepcions argue that the Discover Bank rule, given its origins in California’s unconscionability doctrine and California’s policy against exculpation, is a ground that “exist[s] at law or in equity for the revocation of any contract” under FAA §2. Moreover, they argue that even if we construe the Discover Bank rule as a prohibition on collective-action waivers rather than simply an application of unconscionability, the rule would still be applicable to all dispute-resolution contracts, since California prohibits waivers of class litigation as well. See America Online, Inc. v. Superior Court, 90 Cal. App. 4th 1, 17-18, 108 Cal. Rptr. 2d 699, 711-713 (2001).

When state law prohibits outright the arbitration of a particular type of claim, the analysis is straightforward: The conflicting rule is displaced by the FAA. Preston v. Ferrer, 552 U. S. 346, 353 (2008). But the inquiry becomes more complex when a doctrine normally thought to be generally applicable, such as duress or, as relevant here, unconscionability, is alleged to have been applied in a fashion that disfavors arbitration. In Perry v. Thomas, 482 U. S. 483 (1987), for example, we noted that the FAA’s pre-emptive effect might extend even to grounds traditionally thought to exist “‘at law or in equity for the revocation of any contract.’” Id., at 492, n. 9 (emphasis deleted). We said that a court may not “rely on the uniqueness of an agreement to arbitrate as a basis for a state-law holding that enforcement would be unconscionable, for this would enable the court to effect what .. . the state legislature cannot.” Id., at 493, n. 9.

An obvious illustration of this point would be a ease finding unconscionable or unenforceable as against public policy *342consumer arbitration agreements that fail to provide for judicially monitored discovery. The rationalizations for such a holding are neither difficult to imagine nor different in kind from those articulated in Discover Bank. A court might reason that no consumer would knowingly waive his right to full discovery, as this would enable companies to hide their wrongdoing. Or the court might simply say that such agreements are exculpatory — restricting discovery would be of greater benefit to the company than the consumer, since the former is more likely to be sued than to sue. See Discover Bank, supra, at 161, 113 P. 3d, at 1108-1109 (arguing that class waivers are similarly one sided). And, the reasoning would continue, because such a rule applies the general principle of unconscionability or public-policy disapproval of exculpatory agreements, it is applicable to “any” contract and thus preserved by §2 of the FAA. In practice, of course, the rule would have a disproportionate impact on arbitration agreements; but it would presumably apply to contracts purporting to restrict discovery in litigation as well.

Other examples are easy to imagine. The same argument might apply to a rule classifying as unconscionable arbitration agreements that fail to abide by the Federal Rules of Evidence, or that disallow an ultimate disposition by a jury (perhaps termed “a panel of twelve lay arbitrators” to help avoid pre-emption). Such examples are not fanciful, since the judicial hostility towards arbitration that prompted the FAA had manifested itself in “a great variety” of “devices and formulas” declaring arbitration against public policy. Robert Lawrence Co. v. Devonshire Fabrics, Inc., 271 F. 2d 402, 406 (CA2 1959). And although these statistics are not definitive, it is worth noting that California’s courts have been more likely to hold contracts to arbitrate unconscionable than other contracts. Broome, An Unconscionable Application of the Unconscionability Doctrine: How the California Courts Are Circumventing the Federal Arbitration Act, 3 Hastings Bus. L. J. 39, 54, 66 (2006); Randall, Judicial *343Attitudes Toward Arbitration and the Resurgence of Unconscionability, 52 Buffalo L. Rev. 185,186-187 (2004).

The Concepcions suggest that all this is just a parade of horribles, and no genuine worry. “Rules aimed at destroying arbitration” or “demanding procedures incompatible with arbitration,” they concede, “would be preempted by the FA A because they cannot sensibly be reconciled with Section 2.” Brief for Respondents 32. The “grounds” available under § 2’s saving clause, they admit, “should not be construed to include a State’s mere preference for procedures that are incompatible with arbitration and ‘would wholly eviscerate arbitration agreements.’” Id., at 33 (quoting Carter v. SSC Odin Operating Co., LLC, 237 Ill. 2d 30, 50, 927 N. E. 2d 1207, 1220 (2010)).4

We largely agree. Although § 2’s saving clause preserves generally applicable contract defenses, nothing in it suggests an intent to preserve state-law rules that stand as an obstacle to the accomplishment of the FAA’s objectives. Cf. Geier v. American Honda Motor Co., 529 U. S. 861, 872 (2000); Crosby v. National Foreign Trade Council, 530 U. S. 363, 372-373 (2000). As we have said, a federal statute’s saving clause “ ‘cannot in reason be construed as [allowing] a common law right, the continued existence of which would be absolutely inconsistent with the provisions of the act. In other words, the act cannot be held to destroy itself.’” American Telephone & Telegraph Co. v. Central Office Telephone, Inc., 524 U. S. 214, 227-228 (1998) (quoting Texas & Pacific R. Co. v. Abilene Cotton Oil Co., 204 U. S. 426, 446 (1907)).

*344We differ with the Concepcions only in the application of this analysis to the matter before us. We do not agree that rules requiring judicially monitored discovery or adherence to the Federal Rules of Evidence are “a far cry from this case.” Brief for Respondents 32. The overarching purpose of the FAA, evident in the text of §§2, 3, and 4, is to ensure the enforcement of arbitration agreements according to their terms so as to facilitate streamlined proceedings. Requiring the availability of classwide arbitration interferes with fundamental attributes of arbitration and thus creates a scheme inconsistent with the FAA.

B

The “principal purpose” of the FAA is to “ensur[e] that private arbitration agreements are enforced according to their terms.” Volt, 489 U. S., at 478; see also Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 681-682 (2010). This purpose is readily apparent from the FAA’s text. Section 2 makes arbitration agreements “valid, irrevocable, and enforceable” as written (subject, of course, to the saving clause); § 3 requires courts to stay litigation of arbitral claims pending arbitration of those claims “in accordance with the terms of the agreement”; and §4 requires courts to compel arbitration “in accordance with the terms of the agreement” upon the motion of either party to the agreement (assuming that the “making of the arbitration agreement or the failure ... to perform the same” is not at issue). In light of these provisions, we have held that parties may agree to limit the issues subject to arbitration, Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U. S. 614, 628 (1985), to arbitrate according to specific rules, Volt, supra, at 479, and to limit with whom a party will arbitrate its disputes, Stolt-Nielsen, supra, at 683.

The point of affording parties discretion in designing arbitration processes is to allow for efficient, streamlined procedures tailored to the type of dispute. It can be speci*345fled, for example, that the decisionmaker be a specialist in the relevant field, or that proceedings be kept confidential to protect trade secrets. And the informality of arbitral proceedings is itself desirable, reducing the cost and increasing the speed of dispute resolution. 14 Penn Plaza LLC v. Pyett, 556 U. S. 247, 269 (2009); Mitsubishi Motors Corp., supra, at 628.

The dissent quotes Dean Witter Reynolds Inc. v. Byrd, 470 U. S. 213, 219 (1985), as ‘“rejecting] the suggestion that the overriding goal of the Arbitration Act was to promote the expeditious resolution of claims.’” Post, at 360 (opinion of Breyer, J.). That is greatly misleading. After saying (accurately enough) that “the overriding goal of the Arbitration Act was [not] to promote the expeditious resolution of claims,” but to “ensure judicial enforcement of privately made agreements to arbitrate,” 470 U. S., at 219, Dean Witter went on to explain: “This is not to say that Congress was blind to the potential benefit of the legislation for expedited resolution of disputes. Far from it . . . .” Id., at 220. It then quotes a House Report saying that “the costliness and delays of litigation . . . can be largely eliminated by agreements for arbitration.” Ibid, (quoting H. R. Rep. No. 96, 68th Cong., 1st Sess., 2 (1924)). The concluding paragraph of this part of its discussion begins as follows:

“We therefore are not persuaded by the argument that the conflict between two goals of the Arbitration Act — enforcement of private agreements and encouragement of efficient and speedy dispute resolution — must be resolved in favor of the latter in order to realize the intent of the drafters.” 470 U. S., at 221.

In the present case, of course, those “two goals” do not conflict — and it is the dissent’s view that would frustrate both of them.

Contrary to the dissent’s view, our cases place it beyond dispute that the FAA was designed to promote arbitration. *346They have repeatedly described the Act as “embod[ying] [a] national policy favoring arbitration,” Buckeye Check Cashing, 546 U. S., at 443, and “a liberal federal policy favoring arbitration agreements, notwithstanding any state substantive or procedural policies to the contrary,” Moses H. Cone, 460 U. S., at 24; see also Hall Street Assocs., 552 U. S., at 581. Thus, in Preston v. Ferrer, holding pre-empted a state-law rule requiring exhaustion of administrative remedies before arbitration, we said: “A prime objective of an agreement to arbitrate is to achieve 'streamlined proceedings and expeditious results/ ” which objective would be “frustrated” by requiring a dispute to be heard by an agency first. 552 U. S., at 357-358. That rule, we said, would, “at the least, hinder speedy resolution of the controversy.” Id., at 358.5

California’s Discover Bank rule similarly interferes with arbitration. Although the rule does not require classwide arbitration, it allows any party to a consumer contract to demand it ex post. The rule is limited to adhesion contracts, Discover Bank, 36 Cal. 4th, at 162-163, 113 P. 3d, at 1110, but the times in which consumer contracts were anything *347other than adhesive are long past.6 Carbajal v. H&R Block Tax Servs., Inc., 372 F. 3d 903, 906 (CA7 2004); see also Hill v. Gateway 2000, Inc., 105 F. 3d 1147, 1149 (CA7 1997). The rule also requires that damages be predictably small, and that the consumer allege a scheme to cheat consumers. Discover Bank, supra, at 162-163, 113 P. 3d, at 1110. The former requirement, however, is toothless and malleable (the Ninth Circuit has held that damages of $4,000 are sufficiently small, see Oestreicher v. Alienware Corp., 322 Fed. Appx. 489, 492 (2009) (unpublished)), and the latter has no limiting effect, as all that is required is an allegation. Consumers remain free to bring and resolve their disputes on a bilateral basis under Discover Bank, and some may well do so; but there is little incentive for lawyers to arbitrate on behalf of individuals when they may do so for a class and reap far higher fees in the process. And faced with inevitable class arbitration, companies would have less incentive to continue resolving potentially duplicative claims on an individual basis.

Although we have had little occasion to examine classwide arbitration, our decision in Stolt-Nielsen is instructive. In that case we held that an arbitration panel exceeded its power under § 10(a)(4) of the FAA 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. 559 U. S., at 684-687. We then held that the agreement at issue, which was silent on the question of class procedures, could not be interpreted to allow them because the “changes brought about by the shift from bilateral arbitration to class-action arbitration” are “fundamental.” Id., at 686. This is obvious as a *348structural matter: Classwide arbitration includes absent parties, necessitating additional and different procedures and involving higher stakes. Confidentiality becomes more difficult. And while it is theoretically possible to select an arbitrator with some expertise relevant to the class-certification question, arbitrators are not generally knowledgeable in the often-dominant procedural aspects of certification, such as the protection of absent parties. The conclusion follows that class arbitration, to the extent it is manufactured by Discover Bank rather than consensual, is inconsistent with the FAA.

First, the switch from bilateral to class arbitration 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. “In bilateral 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.” 559 U. S., at 685. But before an arbitrator may decide the merits of a claim in classwide procedures, he must first decide, for example, whether the class itself may be certified, whether the named parties are sufficiently representative and typical, and how discovery for the class should be conducted. A cursory comparison of bilateral and class arbitration illustrates the difference. According to the American Arbitration Association (AAA), the average consumer arbitration between January and August 2007 resulted in a disposition on the merits in six months, four months if the arbitration was conducted by documents only. AAA, Analysis of the AAA’s Consumer Arbitration Caseload, online at http://www.adr.org/si.asp7idr: 5027 (all Internet materials as visited Apr. 25, 2011, and available in Clerk of Court’s case file). As of September 2009, the AAA had opened 283 class arbitrations. Of those, 121 remained active, and 162 had been settled, withdrawn, *349or dismissed. Not a single one, however, had resulted in a final award on the merits. Brief for AAA as Amicus Curiae in Stolt-Nielsen, O. T. 2009, No. 08-1198, pp. 22-24. For those cases that were no longer active, the median time from filing to settlement, withdrawal, or dismissal — not judgment on the merits — was 583 days, and the mean was 630 days. Id., at 24.7

Second, class arbitration requires procedural formality. The AAAs rules governing class arbitrations mimic the Federal Rules of Civil Procedure for class litigation. Compare AAA, Supplementary Rules for Class Arbitrations (effective Oct. 8, 2003), online at http://www.adr.org/sp.asp?id=21936, with Fed. Rule Civ. Proc. 23. And while parties can alter those procedures by contract, an alternative is not obvious. If procedures are too informal, absent class members would not be bound by the arbitration. For a class-action money judgment to bind absentees in litigation, class representatives must at all times adequately represent absent class members, and absent members must be afforded notice, an opportunity to be heard, and a right to opt out of the class. Phillips Petroleum Co. v. Shutts, 472 U. S. 797, 811-812 (1985). At least this amount of process would presumably be required for absent parties to be bound by the results of arbitration.

We find it unlikely that in passing the FAA Congress meant to leave the disposition of these procedural requirements to an arbitrator. Indeed, class arbitration was not even envisioned by Congress when it passed the FAA in 1925; as the California Supreme Court admitted in Discover Bank, class arbitration is a “relatively recent development.” 36 Cal. 4th, at 163, 113 P. 3d, at 1110. And it is at the very *350least odd to think that an arbitrator would be entrusted with ensuring that third parties’ due process rights are satisfied.

Third, class arbitration greatly increases risks to defendants. Informal procedures do of course have a cost: The absence of multilayered review makes it more likely that errors will go uncorrected. Defendants are willing to accept the costs of these errors in arbitration, since their impact is limited to the size of individual disputes, and presumably outweighed by savings from avoiding the courts. But when damages allegedly owed to tens of thousands of potential claimants are aggregated and decided at once, the risk of an error will often become unacceptable. Faced with even a small chance of a devastating loss, defendants will be pressured into settling questionable claims. Other courts have noted the risk of “in terrorem” settlements that class actions entail, see, e. g., Kohen v. Pacific Inv. Management Co. LLC, 571 F. 3d 672, 677-678 (CA7 2009), and class arbitration would be no different.

Arbitration is poorly suited to the higher stakes of class litigation. In litigation, a defendant may appeal a certification decision on an interlocutory basis and, if unsuccessful, may appeal from a final judgment as well. Questions of law are reviewed de novo and questions of fact for clear error. In contrast, 9 U. S. C. § 10 allows a court to vacate an arbitral award only where the award “was procured by corruption, fraud, or undue means”; “there was evident partiality or corruption in the arbitrators”; “the arbitrators were guilty of misconduct in refusing to postpone the hearing ... or in refusing to hear evidence pertinent and material to the controversy^] or of any other misbehavior by which the rights of any party have been prejudiced”; or if the “arbitrators exceeded their powers, or so imperfectly executed them that a mutual, final, and definite award ... was not made.” The AAA rules do authorize judicial review of certification decisions, but this review is unlikely to have much effect given these limitations; review under § 10 focuses on misconduct *351rather than mistake. And parties may not contractually expand the grounds or nature of judicial review. Hall Street Assocs., 552 U. S., at 578. We find it hard to believe that defendants would bet the company with no effective means of review, and even harder to believe that Congress would have intended to allow state courts to force such a decision.8

The Concepcions contend that because parties may and sometimes do agree to aggregation, class procedures are not necessarily incompatible with arbitration. But the same could be said about procedures that the Concepcions admit States may not superimpose on arbitration: Parties could agree to arbitrate pursuant to the Federal Rules of Civil Procedure, or pursuant to a discovery process rivaling that in litigation. Arbitration is a matter of contract, and the FAA requires courts to honor parties’ expectations. Rent-A-Center, West, 561 U. S., at 67-69. But what the parties in the aforementioned examples would have agreed to is not arbitration as envisioned by the FAA, lacks its benefits, and therefore may not be required by state law.

The dissent claims that class proceedings are necessary to prosecute small-dollar claims that might otherwise slip through the legal system. See post, at 365. But States cannot require a procedure that is inconsistent with the FAA, even if it is desirable for unrelated reasons. Moreover, the claim here was most unlikely to go unresolved. As noted earlier, the arbitration agreement provides that AT&T will *352pay claimants a minimum of $7,500 and twice their attorney’s fees if they obtain an arbitration award greater than AT&T’s last settlement offer. The District Court found this scheme sufficient to provide incentive for the individual prosecution of meritorious claims that are not immediately settled, and the Ninth Circuit admitted that aggrieved customers who filed claims would be “essentially guarantee^]” to be made whole, 584 F. 3d, at 856, n. 9. Indeed, the District Court concluded that the Concepcions were better off under their arbitration agreement with AT&T than they would have been as participants in a class action, which “could take months, if not years, and which may merely yield an opportunity to submit a claim for recovery of a small percentage of a few dollars.” Laster, 2008 WL 5216255, *12.

* * *

Because it “stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress,” Hines v. Davidowitz, 312 U. S. 52, 67 (1941), California’s Discover Bank rule is pre-empted by the FAA. The judgment of the Ninth Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion.

It is so ordered.

1

The Concepcions’ original contract was with Cingular Wireless. AT&T acquired Cingular in 2005 and renamed the company AT&T Mobility in 2007. Laster v. AT&T Mobility LLC, 584 F. 3d 849, 852, n. 1 (CA9 2009).

2

That provision further states that “the arbitrator may not consolidate more than one person’s claims, and may not otherwise preside over any form of a representative or class proceeding.” App. to Pet. for Cert. 61a.

3

The guaranteed minimum recovery was increased in 2009 to $10,000. Brief for Petitioner 7.

4

The dissent seeks to fight off even this eminently reasonable concession. It says that to its knowledge "we have not. . . applied the Act to strike down a state statute that treats arbitrations on par with judicial and administrative proceedings,” post, at 366 (opinion of Breyer, J.), and that “we should think more than twice before invalidating a state law that ... puts agreements to arbitrate and agreements to litigate 'upon the same footing,’ ” post, at 361.

5

Relying upon nothing more indicative of congressional understanding than statements of witnesses in committee hearings and a press release of Secretary of Commerce Herbert Hoover, the dissent suggests that Congress “thought that arbitration would be used primarily where merchants sought to resolve disputes of fact. . . [and] possessed roughly equivalent bargaining power.” Post, at 362. Such a limitation appears nowhere in the text of the FAA and has been explicitly rejected by our eases. “Relationships between securities dealers and investors, for example, may involve unequal bargaining power, but we [have] nevertheless held ... that agreements to arbitrate in that context are enforceable.” Gilmer v. Interstate/Johnson Lane Corp., 500 U. S. 20, 33 (1991); see also id., at 32-33 (allowing arbitration of claims arising under the Age Discrimination in Employment Act of 1967 despite allegations of unequal bargaining power between employers and employees). Of course the dissent’s disquisition on legislative history fails to note that it contains nothing — not even the testimony of a stray witness in committee hearings — that contemplates the existence of class arbitration.

6

Of course States remain free to take steps addressing the concerns that attend contracts of adhesion — for example, requiring class-action-waiver provisions in adhesive agreements to be highlighted. Such steps cannot, however, conflict with the FAA or frustrate its purpose to ensure that private arbitration agreements are enforced according to their terms.

7

The dissent claims that class arbitration should be compared to class litigation, not bilateral arbitration. Post, at 363. Whether arbitrating a class is more desirable than litigating one, however, is not relevant. A State cannot defend a rule requiring arbitration-by-jui'y by saying that parties will still prefer it to trial-by-jury.

8

The dissent cites three large arbitration awards (none of which stems from elasswide arbitration) as evidence that parties are willing to submit large claims before an arbitrator. Post, at 364. Those examples might be in point if it could be established that the size of the arbitral dispute was predictable when the arbitration agreement was entered. Otherwise, all the eases prove is that arbitrators can give huge awards — which we have never doubted. The point is that in class-action arbitration huge awards (with limited judicial review) will be entirely predictable, thus rendering arbitration unattractive. It is not reasonably deniable that requiring consumer disputes to be arbitrated on a elasswide basis will have a substantial deterrent effect on incentives to arbitrate.

Justice Thomas,

concurring.

Section 2 of the Federal Arbitration Act (FAA) provides that an arbitration provision “shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U. S. C. § 2. The question here is whether California’s Discover Bank rule, see Discover Bank v. Superior Court, 36 Cal. 4th 148, 113 P. 3d 1100 (2005), is a “groun[d]... for the revocation of any contract.”

It would be absurd to suggest that § 2 requires only that a defense apply to “any contract.” If §2 means anything, it *353is that courts cannot refuse to enforce arbitration agreements because of a state public policy against arbitration, even if the policy nominally applies to “any contract.” There must be some additional limit on the contract defenses permitted by § 2. Cf. ante, at 351 (opinion of the Court) (state law may not require procedures that are “not arbitration as envisioned by the FAA” and “lac[k] its benefits”); post, at 361 (Breyer, J., dissenting) (state law may require only procedures that are “consistent with the use of arbitration”).

I write separately to explain how I would find that limit in the FAA’s text. As I would read it, the FAA requires that an agreement to arbitrate be enforced unless a party successfully challenges the formation of the arbitration agreement, such as by proving fraud or duress. 9 U. S. C. §§ 2, 4. Under this reading, I would reverse the Court of Appeals because a district court cannot follow both the FAA and the Discover Bank rule, which does not relate to defects in the making of an agreement.

This reading of the text, however, has not been fully developed by any party, cf. Brief for Petitioner 41, n. 12, and could benefit from briefing and argument in an appropriate case. Moreover, I think that the Court’s test will often lead to the same outcome as my textual interpretation and that, when possible, it is important in interpreting statutes to give lower courts guidance from a majority of the Court. See US Airways, Inc. v. Barnett, 535 U. S. 391, 411 (2002) (O'Con-nor, J., concurring). Therefore, although I adhere to my views on purposes-and-objectives pre-emption, see Wyeth v. Levine, 555 U. S. 555, 582 (2009) (opinion concurring in judgment), I reluctantly join the Court’s opinion.

I

The FAA generally requires courts to enforce arbitration agreements as written. Section 2 provides that “[a] written provision in ... a contract ... to settle by arbitration a controversy thereafter arising out of such contract. . . shall *354be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” Significantly, the statute does not parallel the words “valid, irrevocable, and enforceable” by referencing the grounds as exist for the “invalidation, revocation, or non-enforcement” of any contract. Nor does the statute use a different word or phrase entirely that might arguably encompass validity, revocability, and enforceability. The use of only “revocation” and the conspicuous omission of “invalidation” and “nonenforcement” suggest that the exception does not include all defenses applicable to any contract but rather some subset of those defenses. See Duncan v. Walker, 533 U. S. 167, 174 (2001) (“It is our duty to give effect, if possible, to every clause and word of a statute” (internal quotation marks omitted)).

Coneededly, the difference between revocability, on the one hand, and validity and enforceability, on the other, is not obvious. The statute does not define the terms, and their ordinary meanings arguably overlap. Indeed, this Court and others have referred to the concepts of revocability, validity, and enforceability interchangeably. But this ambiguity alone cannot justify ignoring Congress’ clear decision in §2 to repeat only one of the three concepts.

To clarify the meaning of §2, it would be natural to look to other portions of the FAA. Statutory interpretation focuses on “the language itself, the specific context in which that language is used, and the broader context of the statute as a whole.” Robinson v. Shell Oil Co., 519 U. S. 337, 341 (1997). “A provision that may seem ambiguous in isolation is often clarified by the remainder of the statutory scheme ... because only one of the permissible meanings produces a substantive effect that is compatible with the rest of the law.” United Sav. Assn. of Tex. v. Timbers of Inwood Forest Associates, Ltd., 484 U. S. 365, 371 (1988).

Examining the broader statutory scheme, § 4 can be read to clarify the scope of § 2’s exception to the enforcement of *355arbitration agreements. When a party seeks to enforce an arbitration agreement in federal court, §4 requires that “upon being satisfied that the making of the agreement for arbitration or the failure to comply therewith is not in issue,” the court must order arbitration “in accordance with the terms of the agreement.”

Reading §§ 2 and 4 harmoniously, the “grounds ... for the revocation” preserved in § 2 would mean grounds related to the making of the agreement. This would require enforcement of an agreement to arbitrate unless a party successfully asserts a defense concerning the formation of the agreement to arbitrate, such as fraud, duress, or mutual mistake. See Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U. S. 395, 403-404 (1967) (interpreting § 4 to permit federal courts to adjudicate claims of “fraud in the inducement of the arbitration clause itself” because such claims “g[o] to the 'making’ of the agreement to arbitrate”). Contract defenses unrelated to the making of the agreement — such as public policy — could not be the basis for declining to enforce an arbitration clause.*

*356II

Under this reading, the question here would be whether California’s Discover Bank rule relates to the making of an agreement. I think it does not.

In Discover Bank, 36 Cal. 4th 148, 113 P. 3d 1100, the California Supreme Court held that “class action waivers are, under certain circumstances, unconscionable as unlawfully exculpatory.” Id., at 165, 113 P. 3d, at 1112; see also id., at 161, 113 P. 3d, at 1108 (“[C]lass action waivers [may be] substantively unconscionable inasmuch as they may operate effectively as exculpatory contract clauses that are contrary to public policy”). The court concluded that where a class-action waiver is found in an arbitration agreement in certain consumer contracts of adhesion, such waivers “should not be enforced.” Id., at 163, 113 P. 3d, at 1110. In practice, the court explained, such agreements “operate to insulate a party from liability that otherwise would be imposed under California law.” Id., at 161, 113 P. 3d, at 1109. The court did not conclude that a customer would sign such an agreement only if under the influence of fraud, duress, or delusion.

The court’s analysis and conclusion that the arbitration agreement was exculpatory reveals that the Discover Bank rule does not concern the making of the arbitration agreement. Exculpatory contracts are a paradigmatic example of contracts that will not be enforced because of public policy. *35715 G. Giesel, Corbin on Contracts §§85.1, 85.17, 85.18 (rev. ed. 2003). Indeed, the court explained that it would not enforce the agreements because they are “‘against the policy of the law.’” 36 Cal. 4th, at 161, 113 P. 3d, at 1108 (quoting Cal. Civ. Code Ann. § 1668 (West 1985)); see also 36 Cal. 4th, at 166, 113 P. 3d, at 1112 (“Agreements to arbitrate may not be used to harbor terms, conditions and practices that undermine public policy” (internal quotation marks omitted)). Refusal to enforce a contract for public-policy reasons does not concern whether the contract was properly made.

Accordingly, the Discover Bank rule is not a “groun[d]... for the revocation of any contract” as I would read § 2 of the FA A in light of § 4. Under this reading, the FA A dictates that the arbitration agreement here be enforced and the Discover Bank rule is pre-empted.

*

The interpretation I suggest would be consistent with our precedent. Contract formation is based on the consent of the parties, and we have emphasized that “[ajrbitration under the Act is a matter of consent.” Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 479 (1989).

The statement in Perry v. Thomas, 482 U. S. 483 (1987), suggesting that §2 preserves all state-law defenses that “arose to govern issues concerning the validity, revocability, and enforceability of contracts generally,” id., at 493, n. 9, is dicta. This statement is found in a footnote concerning a claim that the Court “decline[d] to address.” Id., at 492, n. 9. Similarly, to the extent that statements in Rent-A-Center, West, Inc. v. Jackson, 561 U. S. 63, 69, n. 1 (2010), can be read to suggest anything about the scope of state-law defenses under § 2, those statements are dicta, as well. This Court has néver addressed the question whether the state-law “grounds” referred to in §2 are narrower than those applicable to any contract.

Moreover, every specific contract defense that the Court has acknowledged is applicable under §2 relates to contract formation. In Doctor’s Associates, Inc. v. Casarotto, 517 U. S. 681, 687 (1996), this Court said *356that fraud, duress, and unconscionability “may be applied to invalidate arbitration agreements without contravening § 2.” All three defenses historically concern the making of an agreement. See Morgan Stanley Capital Group Inc. v. Public Util. Dist. No. 1 of Snohomish Cty., 554 U. S. 527, 547 (2008) (describing fraud and duress as “traditional grounds for the abrogation of [a] contract” that speak to “unfair dealing at the contract formation stage”); Hume v. United States, 132 U. S. 406, 411, 414 (1889) (describing an unconscionable contract as one “such as no man in his senses and not under delusion would make” and suggesting that there may be “contracts so extortionate and unconscionable on their face as to raise the presumption of fraud in their inception” (internal quotation marks omitted)).

Justice Breyer,

with whom Justice Ginsburg, Justice Sotomayor, and Justice Kagan join, dissenting.

The Federal Arbitration Act says that an arbitration agreement “shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U. S. C. §2 (emphasis added). California law sets forth certain circumstances in which “class action waivers” in any contract are unenforceable. In my view, this rule of state law is consistent with the federal Act’s language and primary objective. It does not “stan[d] as an obstacle” to the Act’s “accomplishment and execution.” Hines v. Davidowitz, 312 U. S. 52, 67 (1941). And the Court is wrong to hold that the federal Act pre-empts the rule of state law.

I

The California law in question consists of an authoritative state-court interpretation of two provisions of the California Civil Code. The first provision makes unlawful all contracts “which have for their object, directly or indirectly, to exempt anyone from responsibility for his own .. . violation of law.” *358Cal. Civ. Code Ann. § 1668 (West 1985). The second provision authorizes courts to “limit the application of any unconscionable clause” in a contract so “as to avoid any unconscionable result.” § 1670.5(a).

The specific rule of state law in question consists of the California Supreme Court’s application of these principles to hold that “some” (but not “all”) “class action waivers” in consumer contracts are exculpatory and unconscionable under California “law.” Discover Bank v. Superior Court, 36 Cal. 4th 148, 160, 162, 113 P. 3d 1100, 1108, 1110 (2005). In particular, in Discover Bank the California Supreme Court stated that, when a class-action waiver

“is found in a consumer contract of adhesion in a setting in which disputes between the contracting parties predictably involve small amounts of damages, and when it is alleged that the party with the superior bargaining power has carried out a scheme to deliberately cheat large numbers of consumers out of individually small sums of money, then ... the waiver becomes in practice the exemption of the party ‘from responsibility for [its] own fraud, or willful injury to the person or property of another.’ ” Id., at 162-163, 113 P. 3d, at 1110.

In such a circumstance, the “waivers are unconscionable under California law and should not be enforced.” Id., at 163, 113 R 3d, at 1110.

The Discover Bank rule does not create a “blanket policy in California against class action waivers in the consumer context.” Provencher v. Dell, Inc., 409 F. Supp. 2d 1196, 1201 (CD Cal. 2006). Instead, it represents the “application of a more general [unconscionability] principle.” Gentry v. Superior Court, 42 Cal. 4th 443, 457, 165 P. 3d 556,564 (2007). Courts applying California law have enforced class-action waivers where they satisfy general unconscionability standards. See, e. g., Walnut Producers of Cal. v. Diamond Foods, Inc., 187 Cal. App. 4th 634, 647-650, 114 Cal. Rptr. 3d *359449, 459-462 (2010); Arguelles-Romero v. Superior Court, 184 Cal. App. 4th 825, 843-845, 109 Cal. Rptr. 3d 289, 305-307 (2010); Smith v. Americredit Financial Servs., Inc., No. 09cv1076, 2009 WL 4895280 (SD Cal., Dec. 11, 2009); cf. Provencher, supra, at 1201 (considering Discover Bank in choice-of-law inquiry). And even when they fail, the parties remain free to devise other dispute mechanisms, including informal mechanisms, that, in context, will not prove unconscionable. See Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 479 (1989).

II

A

The Discover Bank rule is consistent with the federal Act’s language. It “applies equally to class action litigation waivers in contracts without arbitration agreements as it does to class arbitration waivers in .contracts with such agreements.” 36 Cal. 4th, at 165-166, 113 P. 3d, at 1112. Linguistically speaking, it falls directly within the scope of the Act’s exception permitting courts to refuse to enforce arbitration agreements on grounds that exist “for the revocation of any contract.” 9 U. S. C. §2 (emphasis added). The majority agrees. Ante, at 343.

B

The Discover Bank rule is also consistent with the basic “purpose behind” the Act. Dean Witter Reynolds Inc. v. Byrd, 470 U. S. 213, 219 (1985). We have described that purpose as one of “ensuring] judicial enforcement” of arbitration agreements. Ibid.; see also Marine Transit Corp. v. Dreyfus, 284 U. S. 263, 274, n. 2 (1932) (“ ‘The purpose of this bill is to make valid and enforcible agreements for arbitration’ ” (quoting H. R. Rep. No. 96, 68th Cong, 1st Sess., 1 (1924); emphasis added)); 65 Cong. Rec. 1931 (1924) (“It creates no new legislation, grants no new rights, except a remedy to enforce an agreement in commercial contracts and in *360admiralty contracts”). As is well known, prior to the federal Act, many courts expressed hostility to arbitration, for example, by refusing to order specific performance of agreements to arbitrate. See S. Rep. No. 536, 68th Cong., 1st Sess., 2 (1924). The Act sought to eliminate that hostility by placing agreements to arbitrate “ ‘upon the same footing as other contracts.’” Scherk v. Alberto-Culver Co., 417 U. S. 506, 511 (1974) (quoting H. R. Rep. No. 96, at 2; emphasis added).

Congress was fully aware that arbitration could provide procedural and cost advantages. The House Report emphasized the “appropriate[ness]” of making arbitration agreements enforceable “at this time when there is so much agitation against the costliness and delays of litigation.” Id., at 2. And this Court has acknowledged that parties may enter into arbitration agreements in order to expedite the resolution of disputes. See Preston v. Ferrer, 552 U. S. 346, 357 (2008) (discussing “prime objective of an agreement to arbitrate”). See also Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U. S. 614, 628 (1985).

But we have also cautioned against thinking that Congress’ primary objective was to guarantee these particular procedural advantages. Rather, that primary objective was to secure the “enforcement” of agreements to arbitrate. Dean Witter, 470 U. S., at 221. See also id., at 219 (we “reject the suggestion that the overriding goal of the Arbitration Act was to promote the expeditious resolution of claims”); id., at 219, 217 (“[T]he intent of Congress” requires us to apply the terms of the Act without regard to whether the result would be “possibly inefficient”); cf. id., at 220 (acknowledging that “expedited resolution of disputes” might lead parties to prefer arbitration). The relevant Senate Report points to the Act’s basic purpose when it says that “[t]he purpose of the [Act] is clearly set forth in section 2,” S. Rep. No. 536, at 2 (emphasis added), namely, the section that says that an arbitration agreement “shall be valid, ir*361revocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract,” 9U.S. C. §2.

Thus, insofar as we seek to implement Congress’ intent, we should think more than twice before invalidating a state law that does just what §2 requires, namely, puts agreements to arbitrate and agreements to litigate “upon the same footing.”

Ill

The majority’s contrary view (that Discover Bank stands as an “obstacle” to the accomplishment of the federal law’s objective, ante, at 344-352) rests primarily upon its claims that the Discover Bank rule increases the complexity of arbitration procedures, thereby discouraging parties from entering into arbitration agreements, and to that extent discriminating in practice against arbitration. These claims are not well founded.

For one thing, a state rule of law that would sometimes set aside as unconscionable a contract term that forbids class arbitration is not (as the majority claims) like a rule that would require “ultimate disposition by a jury” or “judicially monitored discovery” or use of “the Federal Rules of Evidence.” Ante, at 342, 344. Unlike the majority’s examples, class arbitration is consistent with the use of arbitration. It is a form of arbitration that is well known in California and followed elsewhere. See, e. g., Keating v. Superior Court, 167 Cal. Rptr. 481, 492 (App. 1980) (officially depublished); American Arbitration Association (AAA), Supplementary Rules for Class Arbitrations (2003), http://www.adr.org/ sp.asp?id=21936 (as visited Apr. 25, 2011, and available in Clerk of Court’s case file); JAMS, The Resolution Experts, Class Action Procedures (2009). Indeed, the AAA has told us that it has found class arbitration to be “a fair, balanced, and efficient means of resolving class disputes.” Brief for AAA as Amicus Curiae in Stolt-Nielsen S. A. v. Animal-Feeds Int’l Corp., O. T. 2009, No. 08-1198, p. 25 (hereinafter *362AAA Amicus Brief). And unlike the majority’s examples, the Discover Bank rule imposes equivalent limitations on litigation; hence it cannot fairly be characterized as a targeted attack on arbitration.

Where does the majority get its contrary idea — that individual, rather than class, arbitration is a “fundamental attribute]” of arbitration? Ante, at 344. The majority does not explain. And it is unlikely to be able to trace its present view to the history of the arbitration statute itself.

When Congress enacted the Act, arbitration procedures had not yet been fully developed. Insofar as Congress considered detailed forms of arbitration at all, it may well have thought that arbitration would be used primarily where merchants sought to resolve disputes of fact, not law, under the customs of their industries, where the parties possessed roughly equivalent bargaining power. See Mitsubishi Motors, supra, at 646 (Stevens, J., dissenting); Joint Hearings on S. 1005 and H. R. 646 before the Subcommittees of the Committees on the Judiciary, 68th Cong., 1st Sess., 15 (1924); Hearing on S. 4213 and S. 4214 before a Subcommittee of the Senate Committee on the Judiciary, 67th Cong., 4th Sess., 9-10 (1923); Dept, of Commerce, Secretary Hoover Favors Arbitration — Press Release (Dec. 28, 1925), Herbert Hoover Papers, Articles, Addresses, and Public Statements File, No. 536, p. 2 (Herbert Hoover Presidential Library); Cohen & Dayton, The New Federal Arbitration Law, 12 Va. L. Rev. 265, 281 (1926); AAA, Year Book on Commercial Arbitration in the United States (1927). This last mentioned feature of the history — roughly equivalent bargaining power — suggests, if anything, that California’s statute is consistent with, and indeed may help to further, the objectives that Congress had in mind.

Regardless, if neither the history nor present practice suggests that class arbitration is fundamentally incompatible with arbitration itself, then on what basis can the majority hold California’s law pre-empted?

*363For another thing, the majority’s argument that the Discover Bank rule will discourage arbitration rests critically upon the wrong comparison. The majority compares the complexity of class arbitration with that of bilateral arbitration. See ante, at 348-349. And it finds the former more complex. See ibid. But, if incentives are at issue, the relevant comparison is not “arbitration with arbitration” but a comparison between class arbitration and judicial class actions. After all, in respect to the relevant set of contracts, the Discover Bank rule similarly and equally sets aside clauses that forbid class procedures — whether arbitration procedures or ordinary judicial procedures are at issue.

Why would a typical defendant (say, a business) prefer a judicial class action to class arbitration? AAA statistics “suggest that class arbitration proceedings take more time than the average commercial arbitration, but may take less time than the average class action in court.” AAA Amicus Brief 24 (emphasis added). Data from California courts confirm that class arbitrations can take considerably less time than in-court proceedings in which class certification is sought. Compare ante, at 348-349 (providing statistics for class arbitration), with Judicial Council of California, Administrative Office of the Courts, Class Certification in California: Second Interim Report From the Study of California Class Action Litigation 18 (2010) (providing statistics for class-action litigation in California courts). And a single class proceeding is surely more efficient than thousands of separate proceedings for identical claims. Thus, if speedy resolution of disputes were all that mattered, then the Discover Bank rule would reinforce, not obstruct, that objective of the Act.

The majority’s related claim that the Discover Bank rule will discourage the use of arbitration because “ [arbitration is poorly suited to ... higher stakes” lacks empirical support. Ante, at 350. Indeed, the majority provides no convincing reason to believe that parties are unwilling to submit high-*364stake disputes to arbitration. And there are numerous counterexamples. Loftus, Rivals Resolve Dispute Over Drug, Wall Street Journal, Apr. 16, 2011, p. B2 (discussing $500 million settlement in dispute submitted to arbitration); Ziobro, Kraft Seeks Arbitration in Fight With Starbucks Over Distribution, Wall Street Journal, Nov. 30, 2010, p. B10 (describing initiation of an arbitration in which the payout “could be higher” than $1.5 billion); Markoff, Software Arbitration Ruling Gives I.B.M. $833 Million From Fujitsu, N. Y. Times, Nov. 30, 1988, p. A1 (describing both companies as “pleased with the ruling” resolving a licensing dispute).

Further, even though contract defenses, e. g., duress and unconscionability, slow down the dispute resolution process, federal arbitration law normally leaves such matters to the States. Rent-A-Center, West, Inc. v. Jackson, 561 U. S. 63, 68 (2010) (arbitration agreements '“may be invalidated by ‘generally applicable contract defenses’” (quoting Doctor’s Associates, Inc. v. Casarotto, 517 U. S. 681, 687 (1996))). A provision in a contract of adhesion (for example, requiring a consumer to decide very quickly whether to pursue a claim) might increase the speed and efficiency of arbitrating a dispute, but the State can forbid it. See, e. g., Hayes v. Oakridge Home, 122 Ohio St. 3d 63, 67, 2009-0hio-2054, ¶ 19, 908 N. E. 2d 408, 412 (“Unconscionability is a ground for revocation of an arbitration agreement”); In re Poly-America, L. P, 262 S. W. 3d 337, 348 (Tex. 2008) (“Unconscionable contracts, however — whether relating to arbitration or not — are unenforceable under Texas law”). The Discover Bank rule amounts to a variation on this theme. California is free to define unconscionability as it sees fit, and its common law is of no federal concern so long as the State does not adopt a special rule that disfavors arbitration. Cf. Doctor’s Associates, supra, at 687. See also ante, at 355-356, n. (Thomas, J., concurring) (suggesting that, under certain circumstances, California might remain free to apply its unconscionability doctrine).

*365Because California applies the same legal principles to address the unconscionability of class arbitration waivers as it does to address the unconscionability of any other contractual provision, the merits of class proceedings should not factor into our decision. If California had applied its law of duress to void an arbitration agreement, would it matter if the procedures in the coerced agreement were efficient?

Regardless, the majority highlights the disadvantages of class arbitrations, as it sees them. See ante, at 350 (referring to the “greatly increase[d] risks to defendants”; the “chance of a devastating loss” pressuring defendants “into settling questionable claims”). But class proceedings have countervailing advantages. In general agreements that forbid the consolidation of claims can lead small-dollar claimants to abandon their claims rather than to litigate. I suspect that it is true even here, for as the Court of Appeals recognized, AT&T can avoid the $7,500 payout (the payout that supposedly makes the Concepcions’ arbitration worthwhile) simply by paying the claim’s face value, such that “the maximum gain to a customer for the hassle of arbitrating a $30.22 dispute is still just $30.22.” Laster v. AT&T Mobility LLC, 584 F. 3d 849, 855, 856 (CA9 2009).

What rational lawyer would have signed on to represent the Concepcions in litigation for the possibility of fees stemming from a $30.22 claim? See, e. g., Carnegie v. Household Int’l, Inc., 376 F. 3d 656, 661 (CA7 2004) (“The realistic alternative to a class action is not 17 million individual suits, but zero individual suits, as only a lunatic or a fanatic sues for $30”). In California’s perfectly rational view, nonclass arbitration over such sums will also sometimes have the effect of depriving claimants of their claims (say, for example, where claiming the $30.22 were to involve filling out many forms that require technical legal knowledge or waiting at great length while a call is placed on hold). Discover Bank sets forth circumstances in which the California courts believe that the terms of consumer contracts can be manipulated to *366insulate an agreement’s author from liability for its own frauds by “deliberately cheating] large numbers of consumers out of individually small sums of money.” 36 Cal. 4th, at 162-163, 113 P. 3d, at 1110. Why is this kind of decision— weighing the pros and cons of all class proceedings alike— not California’s to make?

Finally, the majority can find no meaningful support for its views in this Court’s precedent. The federal Act has been in force for nearly a century. We have decided dozens of cases about its requirements. We have reached results that authorize complex arbitration procedures. E. g., Mitsubishi Motors, 473 U. S., at 629 (antitrust claims arising in international transaction are arbitrable). We have upheld nondiscriminatory state laws that slow down arbitration proceedings. E. g., Volt Information Sciences, 489 U. S., at 477-479 (California law staying arbitration proceedings until completion of related litigation is not pre-empted). But we have not, to my knowledge, applied the Act to strike down a state statute that treats arbitrations on par with judicial and administrative proceedings. Cf. Preston, 552 U. S., at 355-356 (Act pre-empts state law that vests primary jurisdiction in state administrative board).

At the same time, we have repeatedly referred to the Act’s basic objective as ensuring that courts treat arbitration agreements “like all other contracts.” Buckeye Check Cashing, Inc. v. Cardegna, 546 U. S. 440, 447 (2006). See also, e. g., Vaden v. Discover Bank, 556 U. S. 49, 64 (2009); Doctor’s Associates, supra, at 687; Allied-Bruce Terminix Cos. v. Dobson, 513 U. S. 265, 281 (1995); Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U. S. 477, 483-484 (1989); Perry v. Thomas, 482 U. S. 483, 492-493, n. 9 (1987); Mitsubishi Motors, supra, at 627. And we have recognized that “[t]o immunize an arbitration agreement from judicial challenge” on grounds applicable to all other contracts “would be to elevate it over other forms of contract.” Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U. S. *367395, 404, n. 12 (1967); see also Marchant v. Mead-Morrison Mfg. Co., 252 N. Y. 284, 299, 169 N. E. 386, 391 (1929) (Cardozo, C. J.) (“Courts are not at liberty to shirk the process of [contractual] construction under the empire of a belief that arbitration is beneficent any more than they may shirk it if their belief happens to be the contrary”); Cohen & Dayton, 12 Va. L. Rev., at 276 (the Act “is no infringement upon the right of each State to decide for itself what contracts shall or shall not exist under its laws”).

These cases do not concern the merits and demerits of class actions; they concern equal treatment of arbitration contracts and other contracts. Since it is the latter question that is at issue here, I am not surprised that the majority can find no meaningful precedent supporting its decision.

IV

By using the words “save upon such grounds as exist at law or in equity for the revocation of any contract,” Congress retained for the States an important role incident to agreements to arbitrate. 9 U. S. C. § 2. Through those words Congress reiterated a basic federal idea that has long informed the nature of this Nation’s laws. We have often expressed this idea in opinions that set forth presumptions. See, e. g., Medtronic, Inc. v. Lohr, 518 U. S. 470, 485 (1996) (“[B]ecause the States are independent sovereigns in our federal system, we have long presumed that Congress does not cavalierly pre-empt state-law causes of action”). But federalism is as much a question of deeds as words. It often takes the form of a concrete decision by this Court that respects the legitimacy of a State’s action in an individual case. Here, recognition of that federalist ideal, embodied in specific language in this particular statute, should lead us to uphold California’s law, not to strike it down. We do not honor federalist principles in their breach.

With respect, I dissent.