This Essay examines the roles that federal administrative agencies have begun to play in response to the rise of private arbitration, particularly in the consumer and employment contexts. Such agency actions have included enforcement strategies designed to mimic the effects of private litigation when such litigation may not be possible due to the presence of arbitration agreements. And, in some cases, they have involved regulatory responses, including direct regulation of the terms governing private arbitration in particular areas.
The Essay explores how agencies can partially address some of the concerns that scholars of regulation and civil procedure have noted regarding the rise of arbitration. It also sketches the benefits of area-specific regulation of arbitration by administrative agencies compared to the current system of regulation through courts. Finally, it discusses some of the potential downsides of agency involvement and how to address them.


    1. History of the FAA and Its Interpretation
    2. Concerns
    1. Substitution
      1. Mimicking Class Actions
      2. Precedent Setting
    2. Regulation
      1. Information Gathering
      2. Information Forcing
      3. Regulation Pursuant to an Express Delegation
      4. Regulation Pursuant to an Implied Power
      5. Conditions
    1. The Case for Agency Involvement
      1. The Reassertion of the Public in Arbitration
      2. Agency Action May Help Partially Address Concerns Regarding Deterrence, Compensation, and the Rule of Law
      3. Congress Should Consider Further Targeted Delegations
      4. Courts Should Rethink How They Approach Implied Agency Authority
      5. Summing Up and Stepping Back: Just How Much Authority Does Current Law Provide?
    2. Some Worries (and How to Mitigate Them)
      1. The Rise of Class-Arbitration Bans Undermines Part of the Case Against Agency Disgorgement Actions
      2. Duplication
      3. Agency Tunnel Vision



In recent years, scholars have identified mandatory arbitration clauses as a looming threat to the civil justice system. Such clauses—which require resolution of disputes by private arbitrators and often re­strict consumers, employees, and others from using class procedures—may frustrate the ability of injured persons to seek compensation. 1 See, e.g., Myriam Gilles, Opting Out of Liability: The Forthcoming, Near-Total Demise of the Modern Class Action, 104 Mich. L. Rev. 373, 430 (2005) (“[I]n most cases, class members ‘would not likely have received any monetary compensation absent a class action.’” (quoting Deborah R. Hensler et al., Class Action Dilemmas: Pursuing Public Goals for Private Gain 467 (2000))); David Horton & Andrea Cann Chandrasekher, After the Revolution: An Empirical Study of Consumer Arbitration, 104 Geo. L.J. 57, 58–59 (2015) (noting and discussing objections to the predominance of compelled arbitration); Judith Resnik, Fairness in Numbers: A Comment on AT&T v. Concepcion, Wal-Mart v. Dukes, and Turner v. Rogers, 125 Harv. L. Rev. 78, 111–12 (2011) [hereinafter Resnik, Fairness in Numbers] (describing how, absent class procedures, small individual recoveries and high costs dissuade attorneys from taking cases); Jean R. Sternlight & Elizabeth J. Jensen, Using Arbitration to Eliminate Consumer Class Actions: Efficient Business Practice or Unconscionable Abuse?, Law & Contemp. Probs., Winter/Spring 2004, at 75, 85–87 (not­ing that bringing small individual claims, absent class procedures, is typically irrational). Moreover, academics and civil practitioners have argued that mandatory arbitration, paired with class-arbitration bans, threatens the underlying substantive law itself. 2 Class arbitration refers to class procedures used in arbitration and typically gov­erned by procedural rules, such as those put out by the American Arbitration Association. See generally Am. ‌Arbitration ‌Ass’n, ‌Supplementary Rules‌ for Class ‌Arbitrations (‌2003),
/mda0/~edisp/adrstg_004129.pdf []. Scholars have laid out many concerns with these rules and mandatory arbitration clauses. See Margaret Jane Radin, Boilerplate: The Fine Print, Vanishing Rights, and the Rule of Law 16–17 (2013) (describing the ability of mandatory arbitration clauses to supersede legal rights); Einer Elhauge, How Italian Colors Guts Private Antitrust Enforcement by Replacing It with Ineffective Forms of Arbitration, 38 Fordham Int’l L.J. 771, 776–77 (2015) [hereinafter Elhauge, How Italian Colors Guts]; Catherine Fisk & Erwin Chemerinsky, The Failing Faith in Class Actions: Wal-Mart v. Dukes and AT&T Mobility v. Concepcion, 7 Duke J. Const. L. & Pub. Pol’y 73, 85–86 (2011) (noting that mandatory arbitration clauses can effectively shield a firm from liability); J. Maria Glover, Disappearing Claims and the Erosion of Substantive Law, 124 Yale L.J. 3052, 3076–83 (2015) [hereinafter Glover, Disappearing Claims] (arguing Supreme Court jurisprudence has permitted firms to skirt substantive legal obligations via mandatory arbitration clauses); David Horton, Arbitration as Delegation, 86 N.Y.U. L. Rev. 437, 483–85 (2011) (describing how the Supreme Court’s interpretation of the Federal Arbitration Act (FAA) allows firms to “change the law”); David S. Schwartz, Enforcing Small Print to Protect Big Business: Employee and Consumer Rights Claims in an Age of Compelled Arbitration, 1997 Wis. L. Rev. 33, 59–60 (arguing mandatory arbitration clauses may remove an entire industry’s disputes from courts).
In areas in which individual litigation is unlikely due to transaction costs, whole categories of harm may go unremedied (and, in the long term, undeterred). Thus, scholars have argued that courts, and the U.S. Supreme Court in particular, should stem the tide and refuse to enforce arbitration agreements, at least in certain circum­stances. 3 See, e.g., Horton, supra note 2, at 493–98 (urging the Supreme Court to apply the nondelegation doctrine to limit the scope of the Federal Arbitration Act); David L. Noll, Rethinking Anti-Aggregation Doctrine, 88 Notre Dame L. Rev. 649, 681–86 (2012) (en­couraging courts to adopt the “outcomes approach,” which considers companies’ ex post behavior if prohibitions on class procedures are per­mitted); see also Christopher R. Leslie, The Arbitration Bootstrap, 94 Tex. L. Rev. 265, 322–29 (2015) (proposing ways lower courts can strike clauses in arbitration contracts, notwithstanding Supreme Court precedent). To date, however, their calls have fallen on deaf ears, at least when it comes to the Supreme Court. 4 See Am. Express Co. v. Italian Colors Rest., 133 S. Ct. 2304, 2312 (2013) (holding contractual waivers of class arbitration are enforceable under the FAA even when the plaintiff’s cost of individually arbitrating a federal statutory claim exceeds the potential recovery); AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 344 (2011) (“Requiring the availability of classwide arbitration interferes with fundamental attributes of arbitration and thus creates a scheme inconsistent with the FAA.”); Stolt-Nielsen S.A. v. AnimalFeeds Int’l Corp., 559 U.S. 662, 687 (2010) (holding that an arbitration agreement that is con­ceded to be “silent” regarding the availability of class arbitration cannot be read to author­ize it); see also infra Part I (discussing cases in greater depth). But see Oxford Health Plans LLC v. Sutter, 133 S. Ct. 2064, 2071 (2013) (finding an arbitrator did not exceed his powers when interpreting an agreement to allow for class arbitration and limiting Stolt-Nielsen).

This Essay shifts the focus from the courts. In recent years, a differ­ent set of actors has begun to step in and address the issues raised by ar­bitration: federal administrative agencies. The greater involvement of agencies confronts us with a set of institutional questions that stands apart from the particular content of the current set of rules governing arbitra­tion. Because of agencies’ sensitivity to enforcement needs in particular areas and their ability to amass information, they have a potentially valu­able role to play in comparison to the inherited system of regulation by litigants and courts. But agency action also triggers a set of well-known concerns, which may play out differently in different regulatory envi­ronments. For its part, the scholarly literature, while nibbling around the edges of the issue, 5 See Mark A. Lemley & Christopher R. Leslie, Antitrust Arbitration and Merger Approval, 110 Nw. U. L. Rev. 1, 55 (2015) (arguing that the Federal Trade Commission (FTC) and Department of Justice Antitrust Division should address the rise of mandatory bilateral arbitration by “[c]onditioning merger approval on not imposing mandatory arbi­tration clauses in their contracts”); Maureen A. Weston, The Clash: Squaring Mandatory Arbitration with Administrative Agency and Representative Recourse, 89 S. Cal. L. Rev. 103, 105 (2015) (exploring doctrinal issues regarding whether “mandatory predispute arbitration agreements displace a party’s right to access state and federal administrative agency procedures, or laws authorizing representative actions”); see also Myriam Gilles & Gary Friedman, After Class: Aggregate Litigation in the Wake of AT&T Mobility v Concepcion, 79 U. Chi. L. Rev. 623, 654–58 (2012) (noting the delegation of authority to the Consumer Financial Protection Bureau (CFPB) to regulate arbitration). Professor Myriam Gilles and Gary Friedman also explore the role of state attorneys general in poten­tially pursuing parens patriae actions “to fill the void left by class actions.” Id. at 660–75. This Essay will largely sidestep actions taken at the state level or under state law, which raise a separate set of issues. There is also a small but growing literature on agency actions mimicking class actions, largely in the securities context. See Barbara Black, Should the SEC Be a Collection Agency for Defrauded Investors?, 63 Bus. Law. 317, 318 (2008); Urska Velikonja, Public Compensation for Private Harm: Evidence from the SEC’s Fair Fund Distributions, 67 Stan. L. Rev. 331, 359–91 (2015) (assessing the effectiveness and effi­ciency of fair funds in compensating investors for their losses and in transferring wealth from shareholders to investors); Verity Winship, Fair Funds and the SEC’s Compensation of Injured Investors, 60 Fla. L. Rev. 1103, 1110–23 (2008) (discussing the emergence of SEC “mechanisms for compensating injured investors: distribution of disgorged funds and . . . distribution [under] the Fair Fund provision”). For an article that goes beyond the securities laws, see Adam S. Zimmerman, Distributing Justice, 86 N.Y.U. L. Rev. 500 (2011) (discussing other agencies, including the FTC and Food and Drug Administration (FDA), which distribute awards to consumers). These articles, however, do not discuss the implications of mandatory arbitration in the context of agency actions that mimic class actions, an issue this Essay takes up in section III.B. has yet to fully grapple with the role agencies have played, and might yet play, in the context of arbitration. 6 One recent and notable exception is David L. Noll, Regulating Arbitration, 105 Calif. L. Rev. (forthcoming 2017) [hereinafter Noll, Regulating Arbitration], (on file with the Columbia Law Review). Professor David Noll’s article provides a useful framework for thinking about whether and how policymakers—particularly Congress and agencies—should regulate arbitration in different areas. But Professor Noll is less concerned with the institutional-allocation questions at the heart of this Essay, such as whether agencies in particular should step into a greater regulatory role vis-à-vis courts and Congress. Professor Noll’s article also does not discuss agency enforce­ment authority in the context of arbitration. The time is therefore ripe for a reassessment not only of the particular legal rules governing arbitration but also of the institutional frameworks in which those rules operate.

This Essay advances two broad ways that agencies have involved themselves in the issues brought about by mandatory bilateral arbitra­tion. 7 This Essay will refer to mandatory arbitration paired with a class waiver as “man­datory bilateral arbitration.” First, drawing on the insight that private enforcement and public action often work in tandem to produce regulatory outcomes, the Essay argues that administrative agencies can make compensating adjustments to their enforcement practices in order to partially offset the ill effects of class-arbitration bans. For example, agencies can increase their reliance on disgorgement-type remedies and agency-administered recovery funds in areas in which private recovery is unlikely because, for example, the claims are too low value to support individual resolution. Such actions are in many cases already legally available to agencies and—because it is the agency and not the affected individual initiating the suit—are unaf­fected by the presence of arbitration agreements. Agencies can also use these and other types of suits to establish precedent that is then available to private parties in litigation or arbitration and that contributes to the continuing development of the substantive law.

Second, apart from their role as enforcers, agencies have in some cases taken on a role in regulating arbitration agreements, including by banning them in certain circumstances or limiting the use of class waiv­ers. 8 See infra notes 157–175 and accompanying text. One insight of this Essay is that the current system for regulating arbitration is neither nonregulation nor regulation by Congress but a system in which the courts—led by the Supreme Court—set the rules, often by employing contestable assumptions about the role of litigation, including class litigation, in various regulatory regimes. The result, largely brought about by a five-Justice majority on the Court, is a system in which legal challenges to arbitration have mostly been cut off—that is, in which arbitration agreements are enforced according to their terms in the vast majority of cases and across vastly different regulatory regimes.

From a purely institutional perspective, there is reason to doubt that the one-size-fits-all approach developed by the courts—the Supreme Court in particular—is the best model. For reasons well known to the ad­ministrative law literature, agencies’ ability to amass information about particular regulatory areas will often make them better area-specific regu­lators than either the courts or Congress. 9 See infra notes 227–240 and accompanying text (describing the expertise and knowledge agencies require in order to decide when and how to regulate arbitration). There is no reason in principle why that should not also be the case on contested issues regarding whether mandatory arbitration should be allowed for certain kinds of disputes and on what terms. This suggests that agencies should be given a wider regulatory role vis-à-vis other institutional actors than they are in the current system. 10 In making this suggestion, this Essay is of a piece with other recent works that have suggested moving some areas traditionally under the primary control of courts to a more administrative model. See, e.g., James R. Hines Jr. & Kyle D. Logue, Delegating Tax, 114 Mich. L. Rev. 235, 253–68 (2015) (arguing for greater delegation of tax policy to agencies); Rafael I. Pardo & Kathryn A. Watts, The Structural Exceptionalism of Bankruptcy Administration, 60 UCLA L. Rev. 384, 452–60 (2012) (making a similar argu­ment with respect to bankruptcy policy). Moreover, as explained below, there are good rea­sons to think that Congress is simply more likely to bestow area-specific reg­ulatory authority on an administrative agency than it is to write definitive rules itself—for example, by amending the Federal Arbitration Act (FAA) to overturn recent court decisions or by banning predispute binding arbitration entirely in certain areas, as would the repeatedly in­troduced (but still unenacted) Arbitration Fairness Act. 11 See infra notes 240–246 and accompanying text.

This Essay unfolds as follows. Part I describes the problem: Over the past two decades, businesses have increasingly inserted mandatory pre­dispute arbitration provisions into their contracts with consumers, em­ployees, and others. This practice has caused concern regarding victim com­pensation, deterrence, and, ultimately, the underlying substantive law itself.

Part II details the recent actions that agencies have undertaken in response to the rise of mandatory bilateral arbitration. There are two broad methods agencies have employed in response to the effect arbitra­tion has on the regulatory system as a whole. I call these methods “substi­tution” and “regulation.” Section II.A focuses on substitution. When adopting substitution strategies, agencies seek to mimic civil litigation that may have occurred between private parties but for an arbitration agreement. Section II.B focuses on regulation. Regulation strategies in­volve agencies acting not as surrogates for private parties but more classi­cally in their role as public regulators. Agency regulatory efforts have ranged from “soft” methods, such as information gathering or reporting requirements, to “harder” means, such as imposing conditions on the use of arbitration agreements or even barring companies from inserting certain terms into those agreements.

Part III provides a normative evaluation of agencies’ involvement in arbitration. Section III.A suggests that agencies have potential roles to play in addressing the issues raised by the rise of arbitration. For similar reasons that have led some scholars to suggest that agencies should have greater power to create private rights of action for the statutes they en­force, 12 See, e.g., Matthew C. Stephenson, Public Regulation of Private Enforcement: The Case for Expanding the Role of Administrative Agencies, 91 Va. L. Rev. 93, 121–43 (2005) [hereinafter Stephenson, Public Regulation] (arguing that an agency “is likely to be both especially good at ascertaining whether private enforcement would aid the pursuit of stat­utory objectives and especially sensitive to the risks” of private enforcement). agencies are in a good position to evaluate whether arbitration is helping or hindering congressional purposes and take action according­ly. Agencies likely already have better information than either Congress or the courts with which to make the relevant decisions, and they can gather information more efficiently. Indeed, one of the first and primary roles agencies should perform is simply to gather information about the oftentimes obscure facts regarding arbitration as it occurs on the ground.

These functional considerations also suggest different postures for the other two branches. Congress, for its part, should consider more tar­geted express delegations of authority such as that contained in Dodd-Frank. And the courts should grant agencies more leeway when agencies interpret their organic statutes in order to regulate arbitration. Thus, at the end of section III.A, the Essay provides a novel doctrinal argument for why agencies should receive Chevron deference when they reasonably interpret ambiguous provisions of statutes they administer in order to regulate arbitration. This is an issue that has vexed the lower courts and that the Supreme Court may take up in an upcoming set of cases. 13 See Jessica Karmasek, SCOTUS to Decide Arbitration Issue; Unclear If Trump Pick Will Be on Bench in Time, Forbes (Jan. 19, 2017, 2:04 pm), (on file with the Columbia Law Review) (reporting the Supreme Court “agreed to review the validity of class action waiver clauses in employer/employee arbitration agreements”); see also infra note 249 (describing the circuit split over the National Labor Relations Board’s authority to bar waivers of class processes under the federal labor laws).

Section III.B considers some possible worries with agencies taking action to offset the effects of arbitration. First, the section argues that the rise of arbitration—paired with restrictions on class processes—may actu­ally undercut some of the concerns associated with agencies pursuing class-like remedies. The section then discusses institutional mechanisms to address the fear that agencies will overcompensate or overregulate arbitration due to tunnel vision or other administrative pathologies.

I. Arbitration and Its Dangers

This Part briefly details the history of the Federal Arbitration Act and issues that arise under it. The history, discussed in section I.A, is one of expansion, with the Supreme Court interpreting section 2 of the Act—which generally makes arbitration agreements valid and enforceable—to cover more and more types of agreements, while simultaneously narrow­ing the various exceptions to the Act and increasing its power vis-à-vis other sources of law. The end result has been to make it increasingly hard to challenge arbitration clauses, including those that restrict class processes and that mandate bilateral arbitration. Section I.B then takes up some problems that scholars have identified with the current regime.

A. History of the FAA and Its Interpretation

Congress passed the FAA (then called the United States Arbitration Act) in 1925. 14 United States Arbitration Act, Pub. L. No. 68-401, 43 Stat. 883 (1925) (codified as amended at 9 U.S.C. §§ 1–16 (2012)); see also Judith Resnik, Diffusing Disputes: The Public in the Private of Arbitration, the Private in Courts, and the Erasure of Rights, 124 Yale L.J. 2804, 2860 (2015) [hereinafter Resnik, Diffusing Disputes] (describing the history of the FAA). The Act was largely designed to allow parties to enforce arbitration awards in federal court and to prevent judges from deploying arbitration-specific grounds to set aside agreements to arbitrate, as some judges had been doing. 15 See Resnik, Diffusing Disputes, supra note 14, at 2860–61; Resnik, Fairness in Numbers, supra note 1, at 113–14; see also Horton, supra note 2, at 444–45 (noting how judges had “invented special rules, such as the ouster and revocability doctrines, to nullify contracts to arbitrate”). Thus, the current version of the Act states that agreements to arbitrate “shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 16 9 U.S.C. § 2. The Act allows a party to petition the district courts for an order to compel arbitration, 17 Id. § 4 (“A party aggrieved by the alleged failure, neglect, or refusal of another to arbitrate under a written agreement for arbitration may petition any United States district court . . . for an order directing that such arbitration proceed in the manner provided for in such agreement.”). and it directs courts to enforce arbitration awards once made. 18 Id. § 9 (indicating that after an arbitration award “any party to the arbitration may apply to the court so specified for an order confirming the award, and thereupon the court must grant such an order”).

The FAA also has its limits, however. For one, under the FAA, courts retain the power to set aside arbitration agreements on generally appli­cable (as opposed to arbitration-specific) grounds, such as that the con­tract is unconscionable. 19 Id. §§ 2, 4 (stating that “if the making of the arbitration agreement” is invalid for “such grounds as exist at law or in equity for the revocation of any contract” then the court may not enforce the agreement). And section 4 of the Act provides that issues regarding the “making of the arbitration agreement” itself—for exam­ple, whether there was such an agreement in the first place—are for the courts. 20 Id. § 4. Moreover, there is a general scholarly consensus that the enact­ing Congress understood the Act to be limited in other important ways as well. For example, Congress very likely understood the Act to apply only in federal court and not to displace contrary state law. 21 See Horton, supra note 2, at 445–47 (identifying the lack of an express preemp­tion clause, limitation of enforcement provisions to federal courts exclusively, and legisla­tive history as evidence of Congress’s intent); Margaret L. Moses, Statutory Misconstruction: How the Supreme Court Created a Federal Arbitration Law Never Enacted by Congress, 34 Fla. St. U. L. Rev. 99, 127–31 (2006) (noting the “speculation that Congress had a broader purpose than just creating a remedy in federal courts is repeatedly shown to be unwarranted in the numerous representations that the statute would not ap­ply to the states”); David S. Schwartz, Correcting Federalism Mistakes in Statutory Interpretation: The Supreme Court and the Federal Arbitration Act, Law & Contemp. Probs., Winter/Spring 2004, at 5, 23–25 (“Nothing in the legislative history suggests that states would be bound to apply section 2, or any section, of the FAA.”). Congress also “likely did not intend the FAA to cover employment agreements.” 22 Horton, supra note 2, at 446. In­deed, “[i]n passing the FAA, Congress intended to allow arbitration for only a narrow set of legal claims: inter-merchant contract disputes sounding in breach and maritime claims.” 23 Leslie, supra note 3, at 307. Finally, Congress largely envisioned the Act as applying to agreements between parties of roughly equal bargaining power, and not otherwise. 24 Horton, supra note 2, at 447 (arguing “even a cursory review of the FAA’s legisla­tive history reveals that Congress did not want the statute to apply to contracts between parties with unequal bargaining power”).

Beginning in the midtwentieth century, the Supreme Court decided a number of cases that began to reshape the legal landscape surrounding arbitration. Those cases originally fell into three distinct areas, though the Supreme Court has tended to merge them in more recent decades. First, in the area of labor-grievance arbitration, 25 The rules of labor arbitration were first developed under the federal labor laws and not the FAA. See Stephen L. Hayford, Unification of the Law of Labor Arbitration and Commercial Arbitration: An Idea Whose Time Has Come, 52 Baylor L. Rev. 781, 786–87 (2000) (describing how the Supreme Court’s interpretation of federal labor statutes led to the enforcement of collective bargaining agreements). the Supreme Court de­cided a number of cases in which the Court severely limited the grounds on which courts could set aside arbitrator decisions made pursuant to valid collective bargaining agreements. 26 The most famous of these cases are referred to as the Steelworkers Trilogy. See gen­er­ally United Steelworkers v. Enter. Wheel & Car Corp., 363 U.S. 593 (1960); United Steelworkers v. Warrior & Gulf Navigation Co., 363 U.S. 574 (1960); United Steelworkers v. Am. Mfg. Co., 363 U.S. 564 (1960). Arbitration under such agree­ments has continued to serve an important role in the resolution of workplace disputes. 27 As Professor Jean Sternlight has pointed out, the relative lack of controversy sur­rounding labor arbitration depends on “distinguishing between labor arbitration, which is essentially used in lieu of the labor strike as a mode of governing the workplace, and indi­vidual or commercial arbitration, which is used in lieu of litigation.” Jean R. Sternlight, Panacea or Corporate Tool?: Debunking the Supreme Court’s Preference for Binding Arbitration, 74 Wash. U. L.Q. 637, 654 (1996); see also United Bhd. of Carpenters, Local No. 1780 v. Desert Palace, Inc., 94 F.3d 1308, 1311 (9th Cir. 1996) (distinguishing between labor arbitration and other forms on the ground that “unlike parties in the commercial arbitration context, parties entering into a collective bargaining agreement know they are granting the arbitrator tremendous power to define and ‘fill in the gaps’ of their agreement”).

The Supreme Court’s extension of the FAA’s protections for arbitra­tion in two other contexts has proved much more controversial. In the area of commercial arbitration—governing disputes over business con­tracts—the Court both made it more difficult to challenge arbitration agreements and enforced those agreements in the context of adhesive contracts between companies and their customers. 28 For a more fulsome discussion, see Horton, supra note 2, at 449–56 (detailing how the Supreme Court’s expanding arbitration jurisprudence led to a greater likelihood of arbitrating disputes); Leslie, supra note 3, at 307–15 (arguing that expansive judicial interpretations of the FAA have greatly broadened the Act’s scope, often in spite of con­trary indications of legislative intent); Resnik, Diffusing Disputes, supra note 14, at 2863–71 (discussing expansion of the FAA over time). The Court held that while courts could hear challenges seeking to void an arbitration provi­sion contained within a broader contract, arbitrators must decide chal­lenges seeking to set aside the contract as a whole. 29 Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395, 403–04 (1967). Thus, under the “separability doctrine,” whether a contract containing an arbitration pro­vision was procured by fraud, for example, is for the arbitrator to decide. Horton, supra note 2, at 450. More broadly, the Court announced, in language that has proven to have large sway, that “questions of arbitrability must be addressed with a healthy regard for the federal policy favoring arbitration.” 30 Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983). The Court also held, contrary to precedent at the time, that federal statutory claims, such as those aris­ing under the antitrust laws, could be arbitrated. 31 See Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 625–26 (1985) (explicating the Court’s reasoning in Moses H. Cone Memorial Hospital); see also Horton, supra note 2, at 451–52 (describing earlier cases); Leslie, supra note 3, at 308 (arguing that the Supreme Court’s expansion of the FAA was contrary to legislative intent and that “[a]rbitration was not intended for complex legal issues, such as those involving statutory claims”). And it brushed aside arguments that arbitration should not be ordered when imposed in a contract of adhesion. 32 See Allied-Bruce Terminix Cos. v. Dobson, 513 U.S. 265, 268 (1995) (construing the FAA broadly in spite of the allegedly adhesive nature of the contract); Rodriguez de Quijas v. Shearson/Am. Express, Inc., 490 U.S. 477, 484 (1989) (finding no indication that the agreement to arbitrate was adhesive in nature); Horton, supra note 2, at 455 (noting the Supreme Court “neglect[ed] the colorable argument that Congress never intended the FAA to apply to adhesion contracts”). Finally, in perhaps the most sweeping departure from the FAA as originally conceived, the Court held in Southland Corp. v. Keating that the Act preempts state law rules targeting arbitration. 33 465 U.S. 1, 16 (1984); see also Leslie, supra note 3, at 314 (“Although courts hold that Congress intended the FAA to prevent states from enacting laws to protect their citi­zens, Congress did not intend the FAA to preempt state law.”). It should be stated that the Court did not speak with one voice in all these cases. Justice O’Connor’s dissent in Southland, for example, broadly critiqued the evolution of the Court’s FAA jurisprudence before concluding that “[a]lthough arbitration is a worthy alternative to litigation, today’s exercise in judicial revisionism goes too far.” Southland, 465 U.S. at 36 (O’Connor, J., dissenting). In a subsequent opinion, she wrote that “over the past decade, the Court has abandoned all pretense of ascertaining congressional intent with respect to the Federal Arbitration Act, building instead, case by case, an edifice of its own creation.” Allied-Bruce Terminix Cos., 513 U.S. at 283 (O’Connor, J., concurring).

The final context in which the Court expanded the FAA’s pro­tections regards employment contracts. Two decisions merit mention in particular. First, the Court held that employees could be required to arbitrate claims of discrimination under the Age Discrimination in Employment Act (ADEA). 34 Gilmer v. Interstate/Johnson Lane Corp., 500 U.S. 20, 27 (1991). Building on prior cases in the antitrust and securities contexts, the Court found little issue with requiring arbitration of federal statutory claims, and it held that the ADEA in particular evinced no intent to preclude arbitration. 35 See id. at 26–27. Second, in Circuit City Stores, Inc. v. Adams, the Court held that the FAA’s exclusion of “‘contracts of employment of seamen, railroad employees, or any other class of workers engaged in foreign or interstate commerce’” pertained only to “trans­portation workers.” 36 532 U.S. 105, 109 (2001) (quoting 9 U.S.C. § 1 (2000)). Arbitration agreements with any other category of worker were afforded the protections of the FAA. As in the commercial context, these protections extended notwithstanding the fact that the agreement to arbitrate might be contained in a contract of adhesion, in­cluding sometimes the employment application itself. 37 The arbitration agreement at issue in Circuit City was contained in such an application. Id. at 109–10. The respondent, Saint Clair Adams, himself alleged he had been subject to on-the-job harassment due to his sexual orientation. Brief for Respondent at 1, Circuit City Stores, Inc., 532 U.S. 105 (No. 99-1379), 2000 WL 1369473.

The judicial expansion of the FAA provided opportunity for compa­nies, particularly (at least initially) in the financial-services industry, to insert arbitration clauses into their contracts with consumers and em­ployees, with little fear that those clauses would be invalidated. 38 See, e.g., Horton, supra note 2, at 456–59 (highlighting how “[t]he Court’s trans­formation of the FAA left . . . [various companies] scrambling to add compulsory arbitra­tion clauses to their contracts”). Companies that did so also started experimenting with provisions restrict­ing procedural opportunities in the arbitral forum. Thus, companies inserted clauses that, among other things, restricted discovery, named par­ticular fora, and shortened statutes of limitations. 39 See id. at 460.

The procedural restrictions that have received the most attention, especially since the early 2000s, are those limiting the availability of class processes. 40 Of course, bans on class arbitration are not the only type of arbitration clause companies have inserted. As Professor Christopher Leslie has pointed out, “firms insert terms into their arbitration clauses to shorten statutes of limitations, to reduce damages, or to prevent injunctive relief.” Leslie, supra note 3, at 266. The Court first indicated its receptivity to restrictions on par­ties’ rights to bring class claims in the context of contracts that said noth­ing at all about the availability of class processes. In Stolt-Nielsen S.A. v. AnimalFeeds International Corp., the Court considered a case in which the parties to an arbitration agreement “stipulated” that their contract was “silent” regarding class arbitration, meaning that “no agreement” had been reached on the issue. 41 559 U.S. 662, 668–69 (2010). Although the parties in Stolt-Nielsen had not entered into a formal “stipulation” regarding their intent, by the time the case reached the Court, the plaintiffs had acknowledged that “the arbitration clause is ‘silent,’ in the sense that it neith­er specifically authorizes nor specifically forbids class arbitration.” Brief for Respondent at 26, Stolt-Nielsen, 559 U.S. 662 (No. 08-1198), 2008 WL 3404244, at *15. A fractured Supreme Court had left open how to treat arbitration agreements that did not mention class arbitration in Green Tree Financial Corp. v. Bazzle, 539 U.S. 444 (2003) (Breyer, J.) (plurality opinion) (failing to reach agreement among a majority of the Court on the issue). The Court held, in a decision penned by Justice Alito, that not only had the arbitrators in Stolt-Nielsen erred by allowing class arbitration to proceed in such circumstances but the FAA imposed “only one possible outcome”: Class arbitration was not permitted. 42 Stolt-Nielsen, 559 U.S. at 677.

The specific result in Stolt-Nielsen is less important than some of the broad statements the Court made concerning class arbitration in gen­eral. 43 Indeed, the practical impact of Stolt-Nielsen was sharply limited in Oxford Health Plans LLC v. Sutter, 133 S. Ct. 2064, 2070–71 (2013), which held that arbitrators have broad authority to interpret agreements to allow for class arbitration in situations in which the parties have not “stipulated” that the contract is silent. From the general principle that parties are free to set arbitration terms as they “see fit,” the Court concluded that there must be a contrac­tual basis for allowing class arbitration. 44 Stolt-Nielsen, 559 U.S. at 683–84 (internal quotation marks omitted) (quoting Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U.S. 52, 57 (1995)). And that contractual basis could not be inferred from mere silence, the Court found, because “class-action arbitration changes the nature of arbitration to such a degree that it cannot be presumed the parties consented to it by simply agreeing to submit their disputes to an arbitrator.” 45 Id. at 685. This conclusion derived from the Court’s view that, because of the complexity and procedural formal­ity associated with class processes, “the relative benefits of class-action arbitration are much less assured [than those of bilateral arbitration], giving reason to doubt the parties’ mutual consent to resolve disputes through class-wide arbitration.” 46 Id. at 685–86. Thus, Stolt-Nielsen helped lay the groundwork for the idea that class process and arbitration are fundamen­tally in tension and that rules designed to preserve the availability of class processes may come into conflict with the FAA’s goal to encourage arbitration.

That idea bore fruit in AT&T Mobility LLC v. Concepcion, 47 563 U.S. 333 (2011). decided the term after Stolt-Nielsen. Concepcion involved the application of a California common law rule voiding class action waivers, in certain cir­cumstances, under state unconscionability doctrine. 48 See id. at 340 (describing California’s Discover Bank rule). Class plaintiffs, who were parties to an arbitration agreement banning class arbitration, argued that the rule voiding class action waivers was simply an extension of state law governing all contracts and that, indeed, it applied equally to class action waivers found in arbitration and nonarbitration agreements alike. 49 Id. at 341. Nevertheless, the Court found California’s rule preempted by the FAA. 50 Id. at 352. In doing so, the Court returned to the theme that “[r]equiring the availability of classwide arbitration interferes with fundamental at­tributes of arbitration and thus creates a scheme inconsistent with the FAA.” 51 Id. at 344. Thus, under the Court’s reasoning (if not its holding), any rule that operates to require class arbitration when the parties have not agreed to it arguably conflicts with the FAA’s policy to promote arbitra­tion as the Court envisions it. 52 See id. at 348–51.

The next major arbitration case to come across the Court’s docket involved, unlike Concepcion, claims arising under federal law. American Express Co. v. Italian Colors Restaurant concerned the enforceability of an arbitration contract between American Express and merchants who ac­cepted American Express cards and who sued American Express alleging violations of the federal antitrust laws. 53 133 S. Ct. 2304, 2308 (2013). Although the agreement explic­itly barred class arbitration, the merchant plaintiffs put in evidence demonstrating that individual actions would be cost prohibitive because of the cost of obtaining expert testimony necessary to prove a violation. 54 Id. Specifically, plaintiffs’ unrebutted evidence showed that the cost of an expert report needed to win in arbitration would cost “between several hundred thousand and one million dollars”—many times larger than any individual plaintiff could hope to recover. 55 Id. at 2316 (Kagan, J., dissenting). And the arbitration agree­ment in place prohibited not only class procedures but also other methods of cost sharing. 56 Id. Thus, the merchants argued that the class arbi­tration ban was invalid under longstanding Supreme Court pre­cedent in­dicating that arbitration agreements may be set aside if they do not allow for the “effective vindication” of federal claims. 57 Id. at 2310 (majority opinion). In essence, the plain­tiffs argued that if they could not bring their claims on a class-wide basis, the claims would not be brought at all.

The Court in Italian Colors did not take issue with plaintiffs’ argu­ment that they could not practically proceed on an individual basis, but the Court still found the agreement enforceable as written. Justice Scalia, writing for the majority, applied a two-step framework in reaching that result. First, he wrote that no “contrary congressional command” re­quired the Court to invalidate the arbitration agreement or its ban on class processes, pointing out that the antitrust laws do not mention class actions and, in fact, predate Federal Rule of Civil Procedure 23, which allows for class claims in federal court. 58 See id. at 2309. The Court also rejected the claim that Rule 23 itself provides an “entitlement to class proceedings for the vindication of statutory rights.” See id. at 2309–10.

Second, the majority opinion rejected the idea that the “judge-made” effective-vindication doctrine required the Court to invalidate the arbitration agreement in the present circumstances. 59 See id. at 2310–12. In doing so, the Court narrowed the effective-vindication doctrine to cover only situations in which the agreement eliminates the “right” to pursue a certain claim and not ones involving only the practical impossibility of “proving” a vio­lation, 60 Id. at 2311 (emphasis omitted). a distinction the dissent criticized for violating the principle that “[a]n arbitration clause may not thwart federal law, irrespective of how exactly it does so.” 61 Id. at 2313 (Kagan, J., dissenting); see also Lemley & Leslie, supra note 5, at 12 (“The majority opinion in Italian Colors read the word ‘effective’ out of the Effective Vindication Doctrine, turning it into what might be called the Nominal Vindication Doctrine.”).

Finally, the Italian Colors Court put a slightly different twist on the theme first sounded in Stolt-Nielsen and Concepcion. In the final part of its opinion, the majority worried that litigation over whether the effective-vindication doctrine applied would undermine the goal associated with arbitration—namely, procedural informality. 62 See Italian Colors, 133 S. Ct. at 2312 (“Such a preliminary litigating hurdle would undoubtedly destroy the prospect of speedy resolution that arbitration in general and bilateral arbitration in particular was meant to secure.”). In other words, the Court found that the legal wrangling occasioned by a more robust version of the effective-vindication doctrine was itself a reason to read that doctrine narrowly.

B. Concerns

So where are we now? Arbitration clauses are now ubiquitous in con­sumer contracts, many of which are also contracts of adhesion. 63 See infra notes 133–135 and accompanying text. Many employment and labor contracts include arbitration clauses as well. 64 See Nicole Wredberg, Note, Subverting Workers’ Rights: Class Action Waivers and the Arbitral Threat to the NLRA, 67 Hastings L.J. 881, 893 (2016) (noting that in 2008 “fifteen percent to twenty-five percent of employers nationally had adopted mandatory arbitration procedures”). Increasingly, these contracts ban class arbitration of disputes, meaning that arbitration must proceed on a bilateral basis, if at all. 65 See infra note 135 and accompanying text. As detailed above, the Supreme Court has made both arbitration clauses and provi­sions banning class arbitration extremely difficult to challenge in court.

Scholars have found much concerning about this picture. Three concerns stand out—namely, the risks that arbitration poses to victim compensation, deterrence, and the system of law itself. First, the rise of arbitration raises issues concerning the ability of the public, particularly consumers with small-value claims, to gain compensation for legal wrongs com­mitted against them. As Justice Breyer put it in his dissent in Concepcion, no rational consumer—and, even more so, no rational law­yer—will pursue a claim involving a $30 sum (the individual stakes in that case). 66 AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 365 (2011) (Breyer, J., dissenting). Thus, mandatory arbitration combined with bans on class pro­cesses will “sometimes have the effect of depriving claimants of their claims.” 67 Id. Indeed, a recent empirical study of American Arbitration Association cases found that “very few individuals bother to arbitrate mi­nor grievances.” 68 Horton & Chandrasekher, supra note 1, at 117. Following Concepcion itself, the percentage of plaintiffs seeking values of $5,000 or less fell to 13%. 69 Id. Prior to the opinion, “21% of plaintiffs sought $5,000 or less,” a change the authors describe as “statistically significant.” Id. And consumers won just 35% of cases filed, a drop from prior studies that had found a roughly 50% win rate. 70 Id. at 100; see also infra note 142 and accompanying text (discussing additional data collected by the CFPB).

Second, and coupled with the concern over compensation, is a con­cern over deterrence. Simply put, if companies no longer have to bear the costs of their legal violations, they will have less incentive to comply with the law. This concern is particularly acute because of the extent to which the American legal system relies on private enforcement to gener­ate optimal deterrence. Professor Einer Elhauge has expanded on this point in the context of the federal antitrust laws, the setting for Italian Colors. 71 See Elhauge, How Italian Colors Guts, supra note 2, at 771 (arguing Italian Colors “threatens to gut private antitrust enforcement in the United States by replacing it with ineffective forms of arbitration”). In Europe, the ability of private plaintiffs to enforce the antitrust laws has long been limited, not formally through restrictions on rights of action but effectively through limitations on discovery and class pro­cesses. 72 Id. The rise of “ineffective forms of arbitration,” Professor Elhauge predicts, will create a similar dynamic in the United States, with the effect of “immuniz[ing] businesses against US federal antitrust enforcement by any­one who contracts with them.” 73 Id. at 776. Professors Mark Lemley and Christopher Leslie have put the point similarly, arguing that “eliminating class action litigation and class arbitration may—as a practical matter—eliminate liability.” 74 Lemley & Leslie, supra note 5, at 37. Because of the United States’ historic reliance on private recovery, this would effectively “gut[]” antitrust enforcement. 75 Id. at 12, 37.

In its extreme form, the concern regarding deterrence bleeds into a third, deeper concern, which is that arbitration threatens the system of law itself, effectively rewriting certain laws so that they do not reach con­duct that is functionally immunized against suit. This is, of course, the fear motivating the effective-vindication doctrine the Court effectively did away with in Italian Colors. Indeed, it was a note Justice Kagan sounded frequently in her dissent, in which she accused the majority of allowing arbitration to serve as “a mechanism easily made to block the vindication of meritorious federal claims and insulate wrongdoers from liability,” thus undermining “the Sherman Act and other federal statutes providing rights of action.” 76 Am. Express Co. v. Italian Colors Rest., 133 S. Ct. 2304, 2320 (2013) (Kagan, J., dissenting).

Scholars have echoed and extended Justice Kagan’s point regarding the potential ill effects of mandatory bilateral arbitration for various sys­tems of federal law. 77 See Glover, Disappearing Claims, supra note 2, at 3076 (arguing Italian Colors allows corporate defendants to limit obligations “under various consumer protection laws . . . without lobbying Congress or relevant state legislatures for changes to those laws”). By allowing companies to erect procedural barriers to recovery, Professor Maria Glover writes, the Supreme Court has al­lowed those companies to effect a “reworking of substantive law” and, most dramatically, engage in “self-deregulation.” 78 Id. at 3076, 3078. And by inserting pro­cedural hurdles into consumer and employment contracts of adhe­sion, companies may achieve such deregulation outside the normal channels of lawmaking: “[I]f exculpation is buried in the fine print of a maze of difficult-to-understand procedural provisions, then the result is private legal reform largely removed from public scrutiny as well as judi­cial scrutiny.” 79 Id. at 3082. Striking a similar chord, Professor Judith Resnik points out that because of the confidentiality norms surrounding arbitration, to the extent that the arbitration system allows for legal development, it pushes it underground. 80 See Resnik, Diffusing Disputes, supra note 14, at 2895–904 (describing the confidentiality provisions in many arbitration clauses and the difficulty they cause for ana­lyzing arbitration data). Professor David Horton has made a similar point, noting the abil­ity to keep claims out of court enabled companies to make their own “alternative procedural universe,” including requiring confidentiality. Horton, supra note 2, at 460; see also Guyden v. Aetna Inc., 544 F.3d 376, 384–85 (2d Cir. 2008) (finding a confidential­ity provision in a consumer arbitration agreement was legal); Ting v. AT&T, 319 F.3d 1126, 1152 (9th Cir. 2003) (making a contrary finding because the confidentiality provi­sion impeded plaintiffs in obtaining information necessary to discover wrongdoing); Horton & Chandrasekher, supra note 1, at 123 (referring to arbitration as a “confidential, nonprecedential legal system”); Kathryn A. Sabbeth & David C. Vladeck, Contracting (Out) Rights, 36 Fordham Urb. L.J. 803, 806 (2009) (“Arbitration . . . is conducted out of public view and is less constrained by the letter of the law.”). But see W. Mark C. Weidemaier, Judging-Lite: How Arbitrators Use and Create Precedent, 90 N.C. L. Rev. 1091, 1093 (2012) (studying arbitration awards and finding arbitrators often produce writ­ten opinions and cite to precedent). This submergence of the law raises concerns regarding the public’s ability to comprehend the content of the law. 81 See, e.g., Glover, Disappearing Claims, supra note 2, at 3056 (“[P]rivatization threatens to impede public awareness of the substantive law, inasmuch as private proceed­ings frustrate the public’s ability to understand the state of the law, how particular laws are interpreted, and how claims are pursued.”); Sabbeth & Vladeck, supra note 80, at 807 (arguing arbitration allows “rights that are not enforced publicly [to] vanish from the public’s eye, making the public less educated about the laws governing society and proba­bly less likely to recognize and correct the laws’ violations”). Moreover, obscuring legal processes may make issues of individual arbi­trator bias, systemic imbalance, or simple, garden-variety incompetence harder to detect, discuss, and remedy. 82 See Glover, Disappearing Claims, supra note 2, at 3055–56 (describing the issues that may come with the lack of transparency in arbitration); Resnik, Diffusing Disputes, supra note 14, at 2816 (“Without public access, one cannot know whether fair treatment is accorded regardless of status. Without publicity, judges have no means of demonstrating their independence. Without oversight, one cannot ensure that judges, tasked with vindi­cating public rights, are loyal to those norms.”). The cumulative result, according to Professor Resnik, is stark: “a system that ought to be seen as unconsti­tutional, in which state-enforced dispute resolution is outsourced to hun­dreds of unregulated providers whose rules are hard to find, pro­cesses generally closed, and outcomes difficult to know.” 83 Resnik, Diffusing Disputes, supra note 14, at 2809.

Such are the potential problems posed by the rise of mandatory bi­lateral arbitration, particularly in consumer and employment contracts. The next Part turns to how agencies have responded.

II. Agency Responses

Although themselves primarily focusing on the federal courts, schol­ars of arbitration have noted the unlikelihood that the courts, and the Supreme Court in particular, will change course and pare back their ar­bitration jurisprudence, at least in the near future. 84 See, e.g., Glover, Disappearing Claims, supra note 2, at 3091 (arguing “any changes to the Court’s recent arbitration jurisprudence, and its far-reaching implications for the substantive law, are unlikely to emanate from anywhere except perhaps a future—and different—Court” and “[u]ntil such time, it is likely that this erosion of substantive law will continue unabated”). This Part shifts the focus and surveys actions taken by different federal actors: administrative agencies. Although some of the concerns expressed in Part I may well never completely go away absent doctrinal reform, federal administrative agencies have begun to enter the fray, taking actions that potentially mit­igate some effects of the turn to arbitration. This Part surveys how. Specifically, this Part examines agencies’ roles as enforcer and as regula­tor. Section II.A explores how, as enforcers, agencies have begun to initiate enforcement actions that serve as substitutes for private actions by pursuing cases that look similar to traditional class claims or that estab­lish important precedent. Section II.B then examines how, as regu­lators, agencies have begun to directly regulate the use of arbitration clauses.

A. Substitution

This section notes various kinds of agency action that, in effect, step in for actions by private parties that may be hindered by the rise of man­datory arbitration. Although scholars have noted the general point that agency enforcement action might substitute for actions by private parties, the literature has been somewhat negative on the use of agency enforce­ment mechanisms as a possible substitute, pointing particularly to the realities of agency budget constraints. 85 See, e.g., J. Maria Glover, The Structural Role of Private Enforcement Mechanisms in Public Law, 53 Wm. & Mary L. Rev. 1137, 1205–07 (2012) (describing in­stances of failed regulatory enforcement); see also Zachary D. Clopton, Redundant Public–Private Enforcement, 69 Vand. L. Rev. 285, 309 (2016) (pointing to a potential lack of agency resources as justification for overlapping public–private enforcement schemes); Floyd D. Weatherspoon, Incorporating Mandatory Arbitration Employment Clauses into Collective Bargaining Agreements: Challenges and Benefits to the Employer and the Union, 38 Del. J. Corp. L. 1025, 1054 (2014) (noting the Equal Employment Opportunity Commission’s (EEOC’s) lack of resources in conclud­ing the agency cannot adequately compensate for the decline in enforcement brought about by mandatory bilateral arbitration). See generally Richard B. Stewart & Cass R. Sunstein, Public Programs and Private Rights, 95 Harv. L. Rev. 1193, 1214 (1982) (“Public enforcement is . . . frequently inadequate because of budget constraints . . . .”). This argument seems to as­sume that agency enforcement operates on the retail level, stepping in for individual private actions on a one-to-one basis. 86 See, e.g., Weatherspoon, supra note 85, at 1054 (pointing to the number of agency enforcement actions as indication of a lack of agency enforcement ability). As the next two sub­sections show, however, agencies can and do take enforcement actions with wholesale effects—both by collecting money damages and by seek­ing injunctive relief—and these types of actions may curtail some of the consequences of mandatory bilateral arbitration while still conserving agency resources

1. Mimicking Class Actions. — The first way that agencies might miti­gate the effects of mandatory bilateral arbitration is by using their en­forcement powers in ways that mimic class relief. By doing so, agencies can remedy harms that would otherwise go unremedied, provide com­pensation to consumers or others, and strengthen deterrence. 87 Of course, this is not to say that all agencies should at all times pursue such claims when they could regardless of their statutory mission and other enforcement priorities. Rather, this Essay intends only to illuminate one set of enforcement strategies that agen­cies may consider in light of changes in the availability of private enforcement.

Agencies’ ability to compensate individuals through class-like reme­dies has recently received some notice, though no scholar has linked the phenomenon to mandatory bilateral arbitration. The Securities and Exchange Commission’s (SEC’s) use of “Fair Funds” to compensate invest­ors has received the most attention. 88 See, e.g., Black, supra note 5, at 320–22, 337–42 (reviewing the SEC’s disgorge­ment and penalty power and the business community’s response to the SEC’s use of “Fair Funds”); Velikonja, supra note 5, at 331 (noting “[m]ore than half the time, the SEC com­pensates investors for losses where a private suit is either unavailable or impractical”); Zimmerman, supra note 5, at 548 (analyzing the procedural concerns with the SEC’s use of “Fair Funds”). Professor Elhauge has also written about disgorgement in the anti­trust context. See Einer Elhauge, Disgorgement as an Antitrust Remedy, 76 Antitrust L.J. 79, 79–83 (2009) [hereinafter Elhauge, Disgorgement]. The SEC has several bases of author­ity that it uses to seek restitution and disgorgement for federal securities law violations and to distribute recovered funds to investors. 89 See Velikonja, supra note 5, at 339–44 (discussing the SEC’s authority to seek disgorgement using the federal courts’ equitable powers and under specific congressional grants of authority); Zimmerman, supra note 5, at 527–29 (same). The SEC’s practice of seeking restitution for investors has increased dramatically since the Sarbanes-Oxley Act, which among other things specifically “authorize[d] the SEC to add civil fines paid in enforcement actions to disgorgement funds—called ‘fair funds’—and distribute them to the victims of securities violations.” 90 Velikonja, supra note 5, at 341; see also Sarbanes-Oxley Act, Pub. L. No. 107-204, § 308(a), 116 Stat. 745, 784 (2002) (codified as amended at 15 U.S.C. § 7246 (2012)). Whether to seek restitution and allow investors to recover is generally within the discretion of the agency. 91 Velikonja, supra note 5, at 342 (“The decision to distribute funds to investors is at the discretion of the SEC . . . .”). From 2002, when the Sarbanes-Oxley Act was passed, to 2013, the SEC has created 243 Fair Funds holding $14.46 billion in recovered fines and disgorged profits. 92 Id. at 350. As opposed to private actions, which gener­ally concern disclosure violations by publicly traded companies, SEC ac­tions resulting in Fair Fund distributions typically deal with violations in­volving Ponzi schemes, insider trading, market manipulation, and even foreign corruption, among others. 93 Id. at 353.

A comparatively understudied set of recent actions by the Federal Communications Commission (FCC), Federal Trade Commission (FTC), and Consumer Financial Protection Bureau (CFPB) working in tandem to recover consumer losses associated with mobile phone “cramming” also shows the potential for agency action to fill some of the void. The example is interesting because it shows how agencies can recover large restitution-style awards in situations in which private recovery would likely be cost prohibitive due to bans on class arbitration.

Cramming “occurs when telephone companies allow third parties to place [unauthorized] charges on their consumers’ telephone bills.” 94 Report and Order and Further Notice of Proposed Rulemaking, Empowering Consumers to Prevent and Detect Billing for Unauthorized Charges (“Cramming”), 27 FCC Rcd. 4436, 4439 (2012). “Crammers” take advantage of the otherwise lawful practice of “third-party billing,” whereby consumers can expressly agree to certain charges—for example, in the mobile context, charges associated with third-party applications—being placed on their bills. See Caroline E. Sweet, Note, The Hidden Scam: Why Consumers Should No Longer Be Forced to Shoulder the Burden of Liability for Mobile Cramming, 11 J. Bus. & Tech. L. 69, 70 (2016). Third-party crammers sometimes obtain consumers’ “consent” to the charges through fraud; other times, they may simply begin placing unau­thorized charges on the consumers’ bills, hoping the consumers will not notice. 95 See Sweet, supra note 94, at 70–71. Although evidence of overall consumer harm is hard to come by, cramming appears to be a widespread practice. A 2014 Senate com­mittee investigation estimated that “over time, wireless cramming has likely cost American consumers hundreds of millions of dollars.” 96 Cramming on Mobile Phone Bills: A Review of Consumer Protection Practices and Gaps: Hearing Before the S. Comm. on Commerce, Sci. & Transp., 113th Cong. 18 (2014) [hereinafter Senate Cramming Report]. In another government investigation, this one at the state level, mobile carriers reported to the California Public Utility Commission that the refund rate for third-party charges billed to California customers was around 12% in 2011. FTC, Mobile Cramming: An FTC Staff Report 13 (2014) [hereinafter FTC Cramming Report],
federal-trade-commission-staff-report-july-2014/140728mobilecramming.pdf [http://perma.
cc/6Y9P-WWKR]. That is compared to a rate of about 0.2% for charges placed on credit-card bills. Id. Evidence suggests that reported refund rates actually may significantly understate the extent of the problem, due to underdetection and difficulties obtaining compensation. Id. at 14.

Many consumers likely do not even know when unauthorized charges are placed on their bills. Third-party charges are often for relatively small monthly sums (typically under $10) and can be hidden in long bills. 97 See FTC Cramming Report, supra note 96, at 17; Sweet, supra note 94, at 74–75. Many consumers do not realize that third parties can place charges on their cellphone bills at all, and carriers usually do not expressly inform the consumers when selling service plans. 98 See FTC Cramming Report, supra note 96, at 17–18 (noting “consumers may not know that their bills can be charged for third-party services”); Sweet, supra note 94, at 74 (“Consumers’ virtual cluelessness on this issue is due—at least in part—to a lack of ade­quate disclosure by mobile phone carriers.”). Even if consumers do even­tually notice unauthorized charges, it may be only after those charges have appeared on their bills for months. 99 See Senate Cramming Report, supra note 96, at 19. In addition, consumers may have difficulty receiving voluntary refunds from carriers or third-party ven­dors, even when they discover unauthorized charges. 100 See FTC Cramming Report, supra note 96, at 14.

Because of the low dollar value of the claims and the ubiquity of class-arbitration bans in consumer cellphone contracts, persons denied a refund (or who do not recognize they are being overcharged at all) may have little hope for redress. 101 See generally Marguerite Reardon, Why You Can’t Sue Your Wireless Carrier in a Class Action, CNET (Mar. 27, 2012), [] (noting the increasing prevalence of arbitration agreements containing class action waivers in the cellphone industry). Indeed, despite the relatively large-scale nature of the consumer harm, there is little recent evidence that private class actions have been widely attempted. Although it is difficult to specu­late why certain claims are or are not brought, one possibility is that law­yers know such actions are likely to end up in individual arbitration. Indeed, in the rare—and perhaps singular—cramming case that did set aside a class action waiver as unconscionable, the court’s reasoning likely would not survive the Supreme Court’s decision in Concepcion. 102 See Stern v. AT&T Mobility Corp., No. CV 05-8842 CAS (CTx), 2008 WL 4382796, at *11–12 (C.D. Cal. Aug. 22, 2008). That case involved a Washington state un­conscionability doctrine that, similar to the California rule at issue in Concepcion, invali­dated class waivers in arbitration agreements in situations in which class processes were necessary to protect consumers’ interests under state law. See id. at *11.

In light of these issues, the FCC, FTC, and CFPB have recently teamed up to reach several very large settlements with mobile carriers, with the funds available for consumer redress. 103 That carriers should be targeted may seem surprising, since the real bad actors in cramming might be thought of as the third-party vendors who initiate the charges. But for years, carriers, who collect 30–40% of the total value of unauthorized charges, Senate Cramming Report, supra note 96, at 19, have been accused of adopting lax screening, prevention, and enforcement policies against vendors engaged in cramming. See, e.g., Sweet, supra note 94, at 80–82 (noting carrier practice “do[es] little to prevent mobile cramming”). The government’s suits also accuse carriers of failing to explain to customers that third-party charges may appear on their bills, disclosing those charges in an incon­spicuous and misleading way on the bills themselves, and refusing to provide refunds to affected consumers. See, e.g., Complaint for Permanent Injunction & Other Equitable Relief at 4–11, FTC v. AT&T Mobility, LLC, No. 1:14-cv-03227-HLM (N.D. Ga. Oct. 8, 2014); see also Complaint for Permanent Injunction & Other Equitable Relief at 6–10, FTC v. T-Mobile USA, Inc., No. 2:14-cv-00967 (W.D. Wash. July 1, 2014). The FTC has gone after some third-party vendors as well in a series of less high-profile cases. See Sweet, supra note 94, at 77–78. The settlements with carriers include all four of the major national carriers. In 2014, AT&T and T-Mobile agreed to pay $105 million and $90 million, respectively. 104 Press Release, FCC, Verizon & Sprint to Pay $158 Million to Settle Mobile Cramming Investigations (May 12, 2015), []. The next year, Verizon agreed to pay $90 million and Sprint $68 mil­lion. 105 Id. The terms of the settlements provide for consumer redress. AT&T, for example, provided $80 million as part of its settlement “to fund a consumer redress program to give refunds to victims of its unlaw­ful cramming activities.” 106 AT&T Mobility LLC Unauthorized Third-Party Billing Charges, 29 FCC Rcd. 11,803, 11,803 ¶ 3 (2014). Unclaimed money will go to the government “for additional consumer redress, consumer education, or other uses.” 107 Press Release, FTC, T-Mobile to Pay at Least $90 Million, Including Full Consumer Refunds to Settle FTC Mobile Cramming Case (Dec. 19, 2014), http://www. [].
Nearly five million consumers have claimed against the AT&T fund alone. 108 See Margaret Harding McGill, Almost 5 Million Ask FTC for Slice of $80M AT&T Settlement, Law360 (May 1, 2015, 3:07 pm), []. The consent decrees also contain a number of prospective com­mitments by carriers to supply consumers with greater information about the possibility of third-party charges and to more conspicuously disclose such charges when they are incurred. 109 See Press Release, FCC, supra note 104.

A recent draft article by Professors Michael Sant’Ambrogio and Adam Zimmerman shows another way that agencies may step in and sub­stitute for private class actions in federal court: by administering class actions themselves (what the authors call “the agency class action”). 110 See Michael Sant’Ambrogio & Adam Zimmerman, Inside the Agency Class Action, Yale L.J. (forthcoming 2017) [hereinafter Sant’Ambrogio & Zimmerman, Inside the Agency Class Action] (manuscript at 23–25) (on file with the Columbia Law Review). That article builds on prior joint work by the authors. See Michael D. Sant’Ambrogio & Adam S. Zimmerman, The Agency Class Action, 112 Colum. L. Rev. 1992 (2012). Agency class actions involve agencies acting not as litigants (as in the above examples) but as adjudicators of class claims, using some of the same procedural tools that federal judges do in resolving class or aggre­gate claims. Professors Sant’Ambrogio and Zimmerman demonstrate that greater use of the agency class action may be an efficient way to resolve large numbers of claims that might otherwise languish on agency dockets. 111 See Sant’Ambrogio & Zimmerman, Inside the Agency Class Action, supra note 110 (manuscript at 1–2) (describing significant backlogs at agencies in areas such as veter­ans claims and immigration, among others).

As a tool for addressing arbitration, the promise of the agency class action is less clear, though further exploration is warranted. First, most (though not all) of the authorities Professors Sant’Ambrogio and Zimmerman sur­vey involve agency power to adjudicate claims between private parties and the government itself, when arbitration agreements are less likely to come into play. 112 For some exceptions, see id. (manuscript at 50–51) (discussing the FCC’s unused power to adjudicate class claims between private parties alleging violations of the Communications Act). The ability of agency class actions to resolve the kind of consumer and employment claims at the heart of the arbitration debate is less clear. Second, because the litigants in an agency class action are both private parties who (by hypothesis) have agreed to resolve their disputes through individual arbitration, that agreement may divest the agency of authority to resolve the claim, just as it does a federal court. 113 On related questions, see generally Weston, supra note 5, at 105 (analyzing whether “predispute arbitration agreements displace a party’s right to access state and federal administrative agency procedures”). Nevertheless, in exploring their role as adjudicators of class disputes, agencies should be sensitive to the effect arbitration may have on the availability of private modes of class resolution.

2. Precedent Setting. — Another way agencies can take wholesale ac­tions to mitigate the effects of mandatory bilateral arbitration is through individual suits brought in order to establish precedent. Pursuing such a strategy allows federal law to continue to be developed in the open and also provides individuals with potentially favorable case law when they do enter arbitration.

The primary example of an agency acting in such a way involves the Equal Employment Opportunity Commission (EEOC). In EEOC v. Waffle House, the Supreme Court held that the Commission could pursue “vic­tim-specific” relief in the federal courts even when the victim herself was subject to a binding arbitration contract. 114 534 U.S. 279, 295 (2002). The Court has also held that the EEOC may obtain class-like relief even when the hypothetical class would not meet the requirements of Federal Rule of Civil Procedure 23. Gen. Tel. Co. of the Nw. v. EEOC, 446 U.S. 318, 320 (1980). Thus, under Waffle House, the EEOC may essentially step into the shoes of an injured employee in or­der to “enjoin an employer from engaging in unlawful employment prac­tices, and to pursue reinstatement, backpay, and compensatory or puni­tive damages.” 115 Waffle House, 534 U.S. at 287.

The EEOC has used its Waffle House authority purposefully to con­tinue to push employment discrimination law forward. Take the issue of transgender discrimination. For the last several years, the EEOC has been using various means to expand the number of ways that transgender in­dividuals can bring viable Title VII claims, including ruling in Macy v. Holder that discriminating against someone because that person is transgender is actionable regardless of the precise reason for the discrim­ination. 116 EEOC Decision No. 0120120821, 2012 WL 1435995 (Apr. 20, 2012). Courts have held, somewhat more narrowly, that a transgender individual can bring a claim for sex discrimi­nation only if the specific reason for the adverse employment action was to enforce sex stereotypes in the workplace. See, e.g., EEOC v. R.G. & G.R. Harris Funeral Homes, Inc., 100 F. Supp. 3d 594, 595 (E.D. Mich. 2015) (requiring that a “person’s failure to conform to sex stereotypes was the driving force behind the termination”). As part of its litigation strategy surrounding the issue, the EEOC has purposefully involved itself in cases in which the victim of the discrimination would likely have to go to arbitration. For example, Broussard v. First Tower Loan, LLC involved a claim by a transgender man who alleged he was terminated because of his transgender identity. 117 Civil Action No. 15-1161, 2016 WL 879995 (E.D. La. Mar. 8, 2016). As the EEOC explained to the court in that case, it chose to intervene be­cause it recognized that the court was likely to send Broussard to arbitra­tion, but whether Defendant violated Broussard’s Title VII rights needed to be decided in the federal courts, making available the poten­tial of establishing precedential district court or even appellate court de­cisions regarding the correct interpretation of Title VII. A private arbitra­tion proceeding, in which the EEOC would not participate at all, would severely undermine this important public interest. Arbitrators are not bound by the rules of evidence, may not be experts in employment law, and are not subject to meaningful judicial review on appeal. 118 Equal Employment Opportunity Commission’s Memorandum in Opposition to Defendant First Tower Loan’s Motion to Stay EEOC Claims Pending Arbitration at 9, Broussard, No. 2:15-cv-01161-CJB-KWR (E.D. La. Nov. 24, 2015), 2015 WL 12532236.

Broussard and related litigation thus show an agency quite self-consciously taking actions in part designed to push forward the development of the law in light of the presence of mandatory bilateral arbitration. In addition, it shows an agency pursuing both “victim-specific” relief and injunctive relief that may not practically be available in arbitration. 119 Id. at 17 & n.6.

B. Regulation

The second category of strategy agencies have adopted in response to the rise of mandatory bilateral arbitration involves regulation. This category may include general data gathering by the agency as well as (1) mandating that certain information be shared with the agency or the public and (2) directly regulating how arbitration agreements may be written, up to and including banning such agreements altogether. Such regulation may be grounded in either an express delegation from Congress to regulate arbitration specifically or an implied power that the agency holds.

Because several examples below of action pursuant to an express delegation deal with the CFPB, a brief introduction to that agency’s au­thority will be useful. Congress created the CFPB as part of the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010. 120 David A. Hyman & William E. Kovacic, Why Who Does What Matters: Governmental Design and Agency Performance, 82 Geo. Wash. L. Rev. 1446, 1486 (2014). The CFPB has broad responsibility for issues involving consumer financial products such as credit cards, mortgages, and student loans. 121 Id. at 1452. Among the powers invested in the CFPB is the authority to study the use of cer­tain predispute arbitration agreements and, if necessary, to regulate such agreements. The operative provisions provide as follows:

(a) Study and report

The Bureau shall conduct a study of, and shall provide a re­port to Congress concerning, the use of agreements providing for arbitration of any future dispute between covered persons and consumers in connection with the offering or providing of consumer financial products or services.

(b) Further authority

The Bureau, by regulation, may prohibit or impose condi­tions or limitations on the use of an agreement between a cov­ered person and a consumer for a consumer financial product or service providing for arbitration of any future dispute be­tween the parties, if the Bureau finds that such a prohibition or imposition of conditions or limitations is in the public interest and for the protection of consumers. The findings in such rule shall be consistent with the study conducted under subsection (a). 122 12 U.S.C. §§ 5518(a)–(b) (2012).

The Dodd-Frank Act also gives the SEC parallel power to regulate certain securities-related arbitration agreements, though to this date the SEC has not taken any actions on that authority. 123 See Jason Scott Johnston & Todd J. Zywicki, The Consumer Financial Protection Bureau’s Arbitration Study: A Summary and Critique 4 (Univ. of Va. Sch. of Law Pub. Law & Legal Theory Research Paper Series 2015-51, 2015), (on file with the Columbia Law Review) (describing the SEC’s power to develop reports but acknowledging that “nothing indicates that the SEC will be reporting anytime soon on arbitration clauses in broker-consumer contracts”). The CFPB, by con­trast, has issued a preliminary and final report to Congress on arbitra­tion. 124 See CFPB, Arbitration Study Preliminary Results: Section 1028(a) Study Results to Date (2013),
results.pdf []; CFPB, Arbitration Study: Report to Congress, Pursuant to Dodd-Frank Wall Street Reform and Consumer Protection Act § 1028(a) (2015) [hereinafter CFPB Final Report], [].
It has also outlined a number of proposals to regulate arbitration agreements to a Small Business Review Panel, the first step toward a rulemaking on the issue. 125 See CFPB, Small Business Advisory Review Panel for Potential Rulemaking on Arbitration Agreements: Outline of Proposals Under Consideration and Alternatives Considered (2015) [hereinafter CFPB Proposal],
cfpb_small-business-review-panel-packet-explaining-the-proposal-under-consideration.pdf [].
As expected, the Bureau subsequently re­leased a set of proposed rules in the Federal Register. 126 See Arbitration Agreements, 81 Fed. Reg. 32,830 (proposed May 24, 2016) (to be codified at 12 C.F.R. pt. 1040). The future of the proposed rules following the 2016 elec­tion is unclear. See generally Alan S. Kaplinsky & Michael Guerrero, The CFPB Is Under Siege by All Three Branches of the Government, Hill (Feb. 17, 2017, 4:00 pm),
branches-of-government [].

1. Information Gathering. — As part of its study and report to Congress, the CFPB engaged in an extensive information-gathering effort, with a number of interesting results. Among other things, the Bureau examined hundreds of consumer-financial-products agreements; 127 CFPB Final Report, supra note 124, § 1.3 (noting the agency analyzed “approx­imately 850 consumer financial agreements”). analyzed case files from almost 2,000 consumer disputes filed with the American Arbitration Association from 2010 to 2012; 128 Id. surveyed over 1,000 credit card holders regarding their knowledge of the arbitration contracts to which they were parties; 129 Id. reviewed nonarbitration individual, class, and small-claims cases and settlements from the same time period; 130 Id. § 1. studied public enforcement actions for potential overlaps with private disputes; 131 Id. and performed econometric analysis to study the effects of arbitration on the price and availability of consumer credit. 132 Id. (describing analysis based on “computer-assisted searching and extensive manual review”). Some of the study’s results are summarized below.

a. Incidence of Arbitration Agreements. — The study found that many consumer financial contracts contained arbitration provisions, though in some cases the percentage was less than one might expect. For example, 15.8% of credit-card issuers—representing 53% of all outstanding credit-card loans—used arbitration clauses in their consumer contracts. 133 Id. § 2. Those numbers were artificially depressed by an antitrust settlement in which several large issuers that had been using arbitration clauses agreed not to for a cer­tain period; had those issuers continued to use arbitration, the percentage of outstanding loans subject to an arbitration requirement would have been almost 95%. Id. The study similarly estimated that wireless carriers covering over 99% of the market and payday-loans stores covering 98.5% also used arbitration. 134 Id. § 2.3. However, only 7.7% of banks, covering 44.4% of deposits, used arb­itration in their checking-account contracts. 135 Id.

b. Contract Features. — The study contained many findings on the typical features of arbitration contracts. With regard to class arbitration bans, those findings were overwhelming:

93.9% of the credit card arbitration clauses, 88.5% of the checking account arbitration clauses, 97.9% of the prepaid card arbitration clauses, 88.7% of the storefront payday loan arbitration clauses, 100.0% of the private student loan arbitration clauses, and 85.7% of the mobile wireless arbitration clauses in our sample contained terms that expressly did not allow arbitration to proceed on a class basis. 136 Id. § 2.5.5. The study also found that a majority of sampled contracts contained language delegating authority to the arbitrator to rule on the validity of the arbitration clause, id. § 2.5.4, and that a minority contained “contingent minimum recovery” provi­sions allowing consumers to obtain a certain minimum recovery (usually over $2,500) if the consumer won an arbitration award greater than the last settlement offer received, id. § 2.5.11.

c. Consumer Knowledge and Beliefs. — The study’s consumer survey of credit-card holders found that over three-quarters of consumers whose agreements contained arbitration provisions stated that they did not know whether their issuer requires arbitration of disputes. 137 Id. § 3.4.3. Only 6.8% of such consumers correctly identified that they could not pursue dis­putes with their issuer in court. 138 Id. And more than half were also under the “largely mistaken” belief that they could participate in class actions against their issuer, which is consistent with prior studies showing wide­spread consumer ignorance regarding the presence and effect of manda­tory bilateral arbitration clauses. 139 Id. See generally Jeff Sovern et al., “Whimsy Little Contracts” with Unexpected Consequences: An Empirical Analysis of Consumer Understanding of Arbitration Agreements, 75 Md. L. Rev. 1 (2015) (finding only a fraction of consumers understand the legal significance of arbitration clauses when they sign contracts that contain such clauses). Under 2% of respondents said that they would in fact pursue legal claims against their credit-card issuer in any forum when given a hypothetical scenario in which the consumer had been charged erroneous fees and had been denied compensation from the company. 140 CFPB Final Report, supra note 124, § 3.1. However, over 50% said they would cancel their card. 141 Id.

d. Arbitration Filings and Outcomes. — The study contained many find­ings on the kinds of claims filed in arbitration and how those claims were resolved. For example, the study found that the median affirmative con­sumer claim filed was $11,500 and that about twenty-five claims a year involved $1,000 or less. 142 Id. § 5.2.1. Somewhat surprisingly, consumers had counsel in over 60% of total disputes, with 50% of disputes involving an attorney who had handled an arbitration in a similar area during the time period studied. Id. In decided cases, arbitrators granted relief to consumers in approximately one-fifth of cases involving a consumer af­firmative claim or request for debt forbearance. 143 Id. § 5.2.2. Arbitrators rendered awards in just less than a third of cases, with most of the rest appearing to settle. Id. Consumers bringing af­firmative claims recovered about twelve cents for every dollar claimed. 144 Id. For cases involving $1,000 or less, arbitrators granted relief to four con­sumers out of nineteen. 145 Id. The low rates of consumer awards in adjudi­cated cases potentially reflect other findings suggesting that many debt-related arbitration claims are the result of boilerplate filings by a single firm, which then does not effectively pursue them. 146 See Horton & Chandrasekher, supra note 1, at 119–20.

2. Information Forcing. — The first aspect of the CFPB’s proposed regulation 147 In October 2015, the CFPB released a Small Business Advisory Review Panel “Outline of Proposals” on arbitration, which it called the first step toward regulation of arbitration agreements. See CFPB Proposal, supra note 125, at 4. The Bureau has now released the proposed rules themselves for public comment. See Arbitration Agreements, 81 Fed. Reg. 32,830, 32,830 (proposed May 24, 2016) (to be codified at 12 C.F.R. pt. 1040). of arbitration involves rules forcing companies to disclose certain information about arbitration. Specifically, the proposed rules would require covered companies using arbitration agreements to submit to the Bureau both arbitration claims and any resulting awards from arbi­trations filed against them. 148 81 Fed. Reg. at 32,868. The CFPB would post those documents (after making privacy redactions) on the Bureau’s website. 149 Id. at 32,869. The pro­posed rules would allow the Bureau to “monitor how arbitrations and arbitration agreements evolve, and allow it to see whether they evolve in ways that harm consumers.” 150 Id. The CFPB has also stated that such data collection would help the Bureau decide whether “further interventions, up to and including prohibiting the use of arbitration agreements in all cases,” 151 CFPB Proposal, supra note 125, at 22. which (as discussed below 152 See infra note 158 and accompanying text. ) the proposed rules do not themselves propose, “may become appropriate.” 153 CFPB Proposal, supra note 125, at 22. In addition, the proposed rules would allow the Bureau to monitor potential violations of consumer-financial-protection laws, aiding its own functions as an agency. 154 81 Fed. Reg. at 32,869. They would also allow the Bureau to monitor and investigate potential abuses of the arbitration system itself and ensure that the arbitration of claims is consistent with basic fairness norms. 155 Id. For similar reasons, the CFPB has expressed hope that publication may help detect and deter any abuses occurring within the system of arbitration, including arbitrator bias. 156 See id.

3. Regulation Pursuant to an Express Delegation. — The CFPB’s most sweeping proposal would use its authority to “impose conditions or lim­itations on the use of” 157 12 U.S.C. § 5518 (2012). arbitration agreements in order to allow consumers to file class action litigation in court without threat of being sent to arbitration. Specifically, the CFPB would require any arbitration agreement “included in a contract for a consumer financial product or service. . . subject to the proposals to provide explicitly that the arbitration agree­ment is inapplicable to cases filed in court on behalf of a class unless and until class certification is denied or the class claims are dismissed.” 158 CFPB Proposal, supra note 125, at 17. The proposal thus effectively bars companies from including bans on class pro­cesses in their arbitration agreements and allows a judicial forum for class claims. It does not, however, outright ban arbitration, and it would allow consumers the choice of filing a class claim in either court or arbitration. 159 Id. at 18.

The CFPB’s press release accompanying its initial proposal provided two reasons for its proposal to allow class claims to proceed in federal court. First, the proposal would provide a “day in court” that consumers otherwise might not get if required to enter bilateral arbitration. 160 Press Release, CFPB, CFPB Considers Proposal to Ban Arbitration Clauses that Allow Companies to Avoid Accountability to Their Customers (Oct. 7, 2015), http://www. [].
Se­cond, the release specifically noted the broader concern regarding de­terrence, noting that “[w]hen companies can be called to account for their misconduct, public attention on the cases can affect or influence their individual business practices and the business practices of other companies more broadly.” 161 Id.; see also 81 Fed. Reg. at 32,855 (“[T]he Bureau preliminarily concludes, based upon the results of the Study, that individual dispute resolution mechanisms are an insufficient means of ensuring that consumer financial protection laws and consumer financial contracts are enforced.”). In supporting its determination, the CFPB’s proposed rules specifically discussed the importance that class action procedures have in enforcing the consumer protection laws, drawing on data it had collected in its study. 162 81 Fed. Reg. at 32,858–59. As the Bureau explained, “[C]lass ac­tion settlements are a more effective means through which large num­bers of consumers are able to obtain monetary and injunctive relief in a single case” and “as a result of a class settlement, companies frequently change their practices in ways that benefit consumers who are not mem­bers of the class.” 163 Id. at 32,858. The CFPB’s use of its express authority thus shows an agency cognizant of the changes that arbitration has wrought on the civil justice system and on class actions in particular.

4. Regulation Pursuant to an Implied Power. — Unlike the CFPB, which has expressly delegated authority to regulate arbitration, other agencies have acted pursuant to less clear authorityinterpreting their organic statutes to regulate arbitration agreements when the statute itself is ar­guably ambiguous.

The most high-profile effort to do so has come out of the National Labor Relations Board (NLRB). In 2012, the NLRB held in D.R. Horton that an employment agreement mandating arbitration and banning class arbitration illegally restricted an employee’s right “‘to engage in . . . con­certed activities for the purpose of collective bargaining or other mutual aid or protection.’” 164 D.R. Horton Inc., 357 N.L.R.B. 2277, 2278 (2012) (alteration in original) (quoting 29 U.S.C. § 157 (2012)). The Board reasoned that bringing class claims was a type of “concerted activity” protected under the National Labor Relations Act (NLRA) and that an employment agreement restricting that right was therefore in violation of the labor laws. 165 Id. at 2278–82. The NLRB also held that its decision did not conflict with the FAA, invoking several exceptions to the Act that it believed covered the Board’s action. 166 Id. at 2283–89. The Fifth Circuit va­cated the Board’s decision in D.R. Horton, citing the FAA as its basis for refusing to defer to the Board’s determination. 167 See D.R. Horton, Inc. v. NLRB, 737 F.3d 344, 348 (5th Cir. 2013) (concluding “the Board’s decision did not give proper weight to the Federal Arbitration Act”). The legal issues raised by this and similar cases will be discussed below. However, the Board has stuck to its losing position in the Fifth Circuit, continuing to enforce its order in D.R. Horton in disputes outside circuits where it has lost. 168 See Murphy Oil USA, Inc., 361 N.L.R.B. No. 72, 1–2, 2014–2015 NLRB Dec. (CCH) ¶ 15878 (Oct. 28, 2014). See generally Samuel Estreicher & Richard L. Revesz, Nonacquiescence by Federal Administrative Agencies, 98 Yale L.J. 679, 735–43 (1989) (using the NLRB as an example of how agency nonacquiescence to individual circuit court rulings promotes proper development of the law). The Seventh and Ninth Circuits have recently sided with the NLRB, cre­ating a circuit split on the issue that will be taken up by the Supreme Court. 169 See infra note 249 (describing the split and citing cases).

Another agency that has regulated arbitration using somewhat un­clear authority is the FTC. Under the Magnusson-Moss Warranty Act, passed in 1975, the FTC has substantial authority over the terms included in consumer warranty agreements. 170 Pub. L. No. 93-637, 88 Stat. 2183, 2185–86 (1975) (codified at 15 U.S.C. §§ 2301–2312 (2012)); see also Jonathan D. Grossberg, Note, The Magnuson-Moss Warranty Act, the Federal Arbitration Act, and the Future of Consumer Protection, 93 Cornell L. Rev. 659, 662–64 (2008) (noting the Act delegated to the FTC the power to establish rules gov­erning such warranties). Among its various powers is the ab­ility to require companies offering warranties to provide “[i]nformation respecting the availability of any informal dispute settle­ment procedure offered by the warrantor and a recital, where the war­ranty so provides, that the purchaser may be required to resort to such procedure before pursuing any legal remedies in the courts.” 171 15 U.S.C. § 2302(a)(8). The FTC also has the au­thority to “prescribe rules setting forth minimum requirements for any informal dispute settlement procedure which is incorporated into the terms of a written warranty.” 172 Id. § 2310(a)(2).

Shortly after Congress passed the Magnusson-Moss Act, the FTC in­voked its powers to prohibit binding arbitration provisions in warranty contracts. 173 See 40 Fed. Reg. 60,190, 60,210 (Dec. 31, 1975). The Commission did so based on its belief that “Congressional intent” was that all informal dispute procedures would “not be legally binding” and that “even if binding mechanisms were con­templated,” the Commission was “not now convinced that any guidelines which it set out could ensure sufficient protection for consumers.” 174 Id. The FTC has since reaffirmed that view in several rulemakings, despite mixed results in the courts regarding the Commission’s legal authority to regu­late arbitration. 175 See 80 Fed. Reg. 42,710, 42,718–20 (July 10, 2015) (to be codified at 16 C.F.R. pts. 700–701, 703) (reaffirming the ban on binding arbitration and noting decisions ques­tion­ing the FTC’s authority). Generally, these courts have held that Congress has spoken to the enforceability of arbitration agreements in the FAA and that it takes a clear statutory command to override the FAA. See, e.g., Walton v. Rose Mobile Homes LLC, 298 F.3d 470, 475 (5th Cir. 2002) (“There is no doubt that Congress has expressed a clear intention in favor of arbitration for contractual claims.”).

5. Conditions. — In a recent article, Professors Lemley and Leslie propose that the FTC and the Department of Justice’s (DOJ’s) Antitrust Division consider conditioning approval of proposed mergers on the merging parties’ agreement to renounce mandatory arbitration of anti­trust claims against themor, at least, their agreement not to include contract terms that make pursuing such claims unlikely. 176 Lemley & Leslie, supra note 5, at 53–55. Since mer­gersespecially those that trigger government scrutinyincrease market concentration, they both inhibit the kind of competition on contract terms that might mitigate the threat posed by ineffective forms of arbitra­tion and heighten the need for effective antitrust enforcement. Con­ditional approval of mergers to ban arbitration may therefore pro­vide a more surgical option targeted to contexts in which we may have the greatest concern over the effects of arbitration. 177 See id. at 56–57 (“[C]hallenging predispute arbitration agreements during mer­ger review will enable the government to target arbitration agreements imposed in those highly concentrated market conditions most likely to incubate anticompetitive behavior.”). Such action is also likely legal, as the antitrust agencies have broad authority to condition merger approval, and that authority should be unaffected by the Supreme Court’s arbitration jurisprudence. 178 See id. at 57–60 (“[A] merger condition imposed by the antitrust agencies is likely to survive judicial scrutiny.”). That is becausewith some exceptions not relevant herethe federal government can ask merging parties to waive legal rights (including, presumably, the right to enter arbitration agreements) in exchange for the government’s promise not to attempt to block a merger. 179 See id. at 60 (noting there is no “constitutional right to impose arbitration agree­ments on others” and therefore the government conditioning mergers on the nonimposi­tion of arbitration agreements should not be problematic).

Although Professors Lemley and Leslie’s focus is primarily on FTC and DOJ review and on antitrust claims, their proposal could potentially extend to other agencies and areas as well. As they note, the SEC has used its authority to approve initial-public-offering filings in order to re­quire companies to “disavow any mandatory individual arbitration agree­ments in securities cases.” 180 Id. at 57. Another agency with the potential to condition transactions in such a way is the FCC. The FCC has authority to approve mergers that involve the transfer of FCC licenses, 181 See 47 U.S.C. § 214(a), (c) (2012) (granting the FCC authority to approve, among other things, acquisitions of FCC licenses). and it has used that authority quite aggressively in other areas. 182 See generally Rachel E. Barkow & Peter W. Huber, A Tale of Two Agencies: A Comparative Analysis of FCC and DOJ Review of Telecommunications Mergers, 2000 U. Chi. Legal F. 29 (describing the FCC as a “shadow DOJ” that analyzes telecommunication mergers for their effects on competition in the industry).
                 For a recent skeptical view of the FCC’s use of its license-transfer-approval power, see generally Brent Skorup & Christopher Koopman, The FCC’s Transaction Reviews and First Amendment Risks, 39 Harv. J.L. & Pub. Pol’y 675 (2016).
Especially given the prevalence of mandatory bilateral arbitration clauses in consumer communications contracts—including those with mobile and Internet providers 183 See supra note 134 and accompanying text. —it would not be surprising if the Commission took the opportunity to regulate arbitration in future merger proceedings.

Agencies may also have authority to condition participation in gov­ernment programs on complying with regulations regarding arbitration. One example comes from the Department of Education, which proposed conditioning participation in the federal Direct Loan program—which provides subsidized student loans to students attending participating col­leges—on colleges not requiring arbitration of certain disputes. 184 See William D. Ford Federal Direct Loan Program, 81 Fed. Reg. 39,330, 39,421 (proposed June 16, 2016) (to be codified at 34 C.F.R. pt. 685). The Department’s rules have since been finalized. See William D. Ford Federal Direct Loan Program, 81 Fed. Reg. 75,926, 76,088 (Nov. 1, 2016) (to be codified in scattered parts of 34 C.F.R. pt. 685). In re­sponse to claims that the FAA bars such a mandate, the Department points to its nearly unfettered authority to regulate behavior by plan par­ticipants through conditions. 185 William D. Ford Federal Direct Loan Program, 81 Fed. Reg. at 76,022.

This Part has surveyed the legal tools agencies have available to them if they wish to address some of the potential effects of arbitration out­lined above. 186 See supra section I.B. The next Part turns to an evaluation of how those tools might be brought to bear.

III. Evaluation

A. The Case for Agency Involvement

This Part investigates the case for agency action addressing the issues brought on by the rise of mandatory bilateral arbitration. It argues that agencies have an underappreciated institutional role to play in the cur­rent system. This in turn suggests that other institutional actors, includ­ing Congress and the courts, should adjust their postures vis-à-vis agen­cies and allow the latter more room to operate within their potential sphere.

Two related caveats are in order. First, the institutional analysis pre­sented here is necessarily comparative. It asks not whether agencies are perfectly situated to perform various roles but what advantages and dis­advantages they have compared to other institutional actors, including courts and private litigants. 187 In this way, this Essay’s analysis draws on a prominent strand of public law scholarship that has investigated institutional-choice issues in terms of the competencies and pathologies associated with different institutional actors. See generally Thomas W. Merrill, Preemption and Institutional Choice, 102 Nw. U. L. Rev. 727 (2008) (providing a comparative institutional analysis of preemption law); Cass R. Sunstein & Adrian Vermeule, Interpretation and Institutions, 101 Mich. L. Rev. 885 (2003) (arguing institu­tional competencies should inform theories of legal interpretation). For an important work on comparative institutional analysis, see generally Neil K. Komesar, Imperfect Alternatives: Choosing Institutions in Law, Economics, and Public Policy (1994) (explor­ing institutional capacities in the context of legal interpretation). Second, because this Essay discusses “agen­cies” and “courts” in general, many of the arguments necessarily involve comparisons about competencies and behavior on average and over time. Of course, different agencies may perform better or worse in general, and the behavior of the same agency may fluctuate over time. 188 See, e.g., Noll, Regulating Arbitration, supra note 6, at 28–29 (describing the experience of the EEOC in the 1980s to illustrate how an agency’s enforcement of its stat­ute may not always be faithful to congressional intent). In future work I plan to discuss these issues more specifically with regard to certain agencies, particularly those agencies operating in the area of communications law and Internet regulation. That said, agencies (as well as other actors) do share certain institutional ca­pabilities—regardless of how those capabilities may be exercised at par­ticular moments of time. This section focuses on how those capabilities might be brought to bear in the area of arbitration. The next section then turns to potential agency pathologies and how they might be guarded against.

1. The Reassertion of the Public in Arbitration. — First, agency involve­ment, even simply in an oversight role, can help address one of the deepest underlying critiques of the move to mandatory bilateral arbitra­tion: that arbitration has privatized dispute resolutionand, with it, the development of the lawwithout sufficiently safeguarding the public interests involved. This is a concern that is especially acute for statutory causes of action. 189 See Horton, supra note 2, at 456–65 (claiming the Court has all but abolished judicial review of privately made procedural rules). See generally Resnik, Diffusing Disputes, supra note 14 (arguing arbitration has eviscerated statutory rights). Indeed, due to both legally enforceable confidential­ity provisions contained in arbitration agreements and informal norms regarding public access to arbitration proceedings, much of what hap­pens in arbitration remains hidden from public view and thus from accountability. 190 See Resnik, Diffusing Disputes, supra note 14, at 2894–96 (describing formal and informal norms governing secrecy in arbitration).

The federal courts, for their part, could have developed mechanisms for safeguarding “the public” in arbitration. 191 See id. at 2811. But led by the Supreme Court, they have opted for exactly the opposite by sealing off arbitration into a purely private space that both is extremely difficult to escape and has boundaries that are not actively policed by the state. 192 See supra section I.B. The culmina­tion of this trend was the Court’s decision in Italian Colors, in which the Court effectively limited the application of the effective-vindication doc­trine—meant to safeguard the public’s interest in the vindication of fed­eral statutory claims—to extreme situations in which a party formally gives up its right to certain claims. 193 See supra notes 53–62 and accompanying text.

Agency involvement in arbitration provides another angle through which to assert the public’s interest in dispute resolution. Although agencies are of course not perfect representatives of the public inter­est, 194 See, e.g., Daniel Carpenter & David A. Moss, Introduction to Preventing Regulatory Capture: Special Interest Influence and How to Limit It 1, 19 nn.39–40 (Daniel Carpenter & David A. Moss eds., 2014) (discussing when agencies depart from the public interest, including for reasons of regulatory “capture”). they provide another lever through which to exercise power that is at least publicly accountable. 195 See Stephenson, Public Regulation, supra note 12, at 135–38. They thus provide a means to police the bounds of a dispute-resolution system that many fear has become subser­vient to private interests, particularly the interests of big business. This is not entirely different from the role that agencies played historically in the early to midtwentieth century, when, of course, agencies were central to the New Deal reassertion of public, government power in the face of a business community and Supreme Court that sought to enforce laissez-faire economic policies and preserve common law arrangements from the nineteenth century. 196 See, e.g., Cass R. Sunstein, Beyond Marbury: The Executive’s Power to Say What the Law Is, 115 Yale L.J. 2580, 2593 (2006) [hereinafter Sunstein, Beyond Marbury] (dis­cussing the “post-New Deal transfer of effective lawmaking power from common law courts to federal bureaucracies”).

2. Agency Action May Help Partially Address Concerns Regarding Deterrence, Compensation, and the Rule of Law. — Part II demonstrated the kind of roles agencies can play in the postarbitration world. This subsec­tion explores those roles more explicitly against the backdrop of various potential problems that the rise of mandatory bilateral arbitration poses to the regulatory system. Broadly, those roles include: gathering infor­mation, undertaking litigation activity in light of arbitration’s effects, and, potentially, regulating arbitration agreements themselves.

First, agencies can perform a valuable information-gathering func­tion. As scholars have noted, one of the obstacles for those interested in arbitration has been a lack of good data on such items as win rates in ar­bitration, the kind of cases that are brought, and how outcomes compare to those of litigation, among many other things. 197 See, e.g., Horton & Chandrasekher, supra note 1, at 61–62 (noting both detrac­tors and proponents of arbitration have “only the dimmest sense of what actually happens in the extrajudicial forum”). This information is critical because arbitration’s potential costs in terms of the erosion of deterrence in various areas 198 See supra section I.B. cannot be accurately measured without data involving both arbitration and its alternatives. While private data gathering should continue, agencies may often be in a better position to gain information relevant to arbitration. One of the primary benefits agencies provide is their ability—through large staffs with dedicated re­sources—to acquire information relevant to policy determinations. 199 See generally Matthew C. Stephenson, Information Acquisition and Institutional Design, 124 Harv. L. Rev. 1422 (2011). And even apart from resources, agencies may have greater practical and legal ability (through, for example, subpoena power) to acquire nonpub­lic information. 200 See John B. Benton, Administrative Subpoena Enforcement, 41 Tex. L. Rev. 874, 874–83 (1963).

Relatedly, and as the CFPB has proposed, agencies can impose in­formation-forcing requirements that facilitate ongoing research into ar­bitration and can provide information regarding trends in arbitration and arbitration agreements over time. 201 See supra notes 147–153 and accompanying text. This kind of information can in turn facilitate the kind of dynamic regulatory approach that agencies—as opposed to Congress or courts—are often best suited to take. 202 See infra notes 237–240 and accompanying text.

Agencies can also provide compensation to injured consumers or others when private recovery is difficult because of mandatory arbitration that is often coupled with bars on class processes. As noted above, agen­cies have, for many years, and increasingly as of late, served as adminis­trators for funds designed to compensate individuals. 203 See supra section II.A.1. And many agencies have the power, either ancillary to their other enforcement powers or expressly given, to bring disgorgement actions or other class-like claims in order to provide such compensation. 204 See generally Zimmerman, supra note 5, at 521–27. By targeting agency resources toward harms that are unlikely to be compensated through mandatory bilateral arbitration, as the FCC and FTC have done in the area of mobile cramming, 205 See supra notes 103–109 and accompanying text. agencies can thus provide a valuable corrective. 206 See Elhauge, Disgorgement, supra note 88, at 84 (arguing in the context of anti­trust enforcement that disgorgement can help “take up the slack” caused by the increased difficulty of certifying class actions under Rule 23).

Concerns over compensation often go hand-in-hand with a concern regarding deterrence. If companies can keep the fruits of their ill-gotten gains (and if they know it), they will lack the proper incentives to comply with the law. Although some have noted that agency enforcement actions can partially correct for this lack of deterrence when there are overlap­ping private and public enforcement mechanisms, it has generally been assumed that agency enforcement is too episodic to fully compensate for a downtick in private recovery efforts. 207 See supra note 85 and accompanying text. However, through disgorgement actions or other wholesale actions, individual agency actions can pack more punch. Agencies may therefore be in a better position to make up for a lack of deterrence caused by the rise of mandatory bilateral arbitra­tion than has previously been appreciated.

Besides providing individuals with compensation and shoring up in­centives to comply with the law, agency action can also facilitate the con­tinued development of the law itself. Some have rightly worried that moving many types of disputes into arbitration will cause the stunting of the law itself—by pushing the resolution of disputes into private, some­times secret, and nonprecedential settings, the law may ossify as judges no longer issue binding interpretations of statutes. 208 See supra notes 76–83. This is perhaps especially troubling in areas—like antitrust or antidiscrimination law—that have largely relied on judge-made “common law” for their refine­ment. 209 See, e.g., Margaret H. Lemos, The Consequences of Congress’s Choice of Delegate: Judicial and Agency Interpretations of Title VII, 63 Vand. L. Rev. 363, 385 (2010) [hereinafter Lemos, Consequences] (describing how, in Title VII of the Civil Rights Act, “[s]ubstantive issues—such as the critical details of what constitutes prohibited ‘discrimination’—were left to the courts”); Margaret H. Lemos, The Other Delegate: Judicially Administered Statutes and the Nondelegation Doctrine, 81 S. Cal. L. Rev. 405, 462–63 (2008) [hereinafter Lemos, The Other Delegate] (remarking that antitrust law involves a system of “‘regulation by lawsuit’” in which the courts set the basic rules (quot­ing 9 Alexander Bickel & Benno C. Schmidt, Jr., History of the Supreme Court of the United States: The Judiciary and Responsible Government 1910–21, at 130 (1984))). Agency litigation, whether through disgorgement actions or (as with the EEOC actions discussed above 210 See supra notes 116–118 and accompanying text. ) through stepping into the shoes of private individuals, can help offset this effect. Indeed, agencies can choose, as the EEOC has, 211 See supra notes 116–119 and accompanying text. to focus their limited resources in areas where public officials believe legal development would be most useful.

Finally, when the agency actions discussed above may be insufficient, agencies can regulate primary behavior directly by promulgating rules that condition the use of arbitration agreements or even ban such agree­ments altogether. Such regulation may take many different forms, including:

  • Banning the use of predispute arbitration agreements in all con­tracts of a specified kind;
  • Allowing predispute arbitration agreements but specifying that they are inapplicable to class claims brought in court (the CFPB pro­posed rule 212 See Arbitration Agreements, 81 Fed. Reg. 32,830 (proposed May 24, 2016) (to be codified at 12 C.F.R. pt. 1040). );
  • Allowing predispute arbitration agreements but requiring them to allow for class claims to be brought in arbitration;
  • Allowing predispute arbitration agreements but requiring certain minimum payouts (or attorneys’ fees) if the claimant prevails, subject to certain conditions;
  • Allowing predispute arbitration agreements but providing they are inapplicable to certain kinds of disputes (for example, disputes un­der a certain dollar amount);
  • Allowing predispute arbitration agreements but banning certain restrictions other than class action restrictions, such as limiting discovery or shortening the statute of limitations;
  • Allowing predispute arbitration agreements but banning prodefendant fee-shifting arrangement or bars to recovery of exemplary or punitive damages, or both.

These are only some of the many ways that agencies might regulate arbitration agreements. Of course, the specter of regulation raises certain questions. Do the traditional justifications for delegation to agencies ap­ply when they regulate the dispute-resolution process—not something we traditionally associate with agency control? And, if agency authority over arbitration is not expressly delegated, how much slack should the courts allow agencies in interpreting their statutes in ways that implicate arbitra­tion? The next two subsections turn to those questions.

3. Congress Should Consider Further Targeted Delegations. — This sub­section applies the traditional justifications for delegation of lawmaking authority to agencies to the special case of delegations to regulate arbi­tration agreements. It focuses specifically on three considerations that are normally thought to support Congress’s decision to delegate deci­sionmaking authority to an agency: agencies’ greater expertise, flexibility, and accountability compared to other institutional actors, including Congress itself as well as the courts. 213 See, e.g., Lemos, The Other Delegate, supra note 209, at 445–55. And it asks whether these con­siderations also justify a greater agency role in regulating arbitration. These justifications are necessarily comparative. In other words, the ques­tion is not whether an agency is an ideal decisionmaker but simply whether we can expect an agency to perform better than other actors. 214 See Stephenson, Public Regulation, supra note 12, at 126 (“Evaluating the bene­fits of delegating to agencies . . . requires an assessment of how well agencies would per­form th[e] task—not in comparison to a hypothetical ideal decisionmaker, but in comparison to the primary institutional alternatives, Congress and the courts.”); see also Lemos, The Other Delegate, supra note 209, at 444 (“Pro-delegation commentators offer functional arguments in defense of agency lawmaking, focusing on institutional character­istics of agencies that make them valuable partners in the lawmaking enterprise—and in some respects better lawmakers than Congress.”).

In asking this question, this Essay highlights two underappreciated aspects of the current system for regulating arbitration. First, the current system, at least at the federal level, is not one of nonregulation or con­gressional control but rather a system characterized by delegation to the courts. 215 See generally Lemos, The Other Delegate, supra note 209 (discussing delegations to courts as a general phenomenon). The FAA is short, and its text answers few specific questions. Moreover, as discussed above, since 1925 the courts have pushed the scope of the Act far beyond that envisioned by the text and its framers. 216 See supra notes 28–33 and accompanying text. The result, as Justice Thomas has pointed out in several concurrences, is a rather free-floating arbitration jurisprudence heavily driven by policy-based considerations and judgments about how best to advance vague purposes. 217 See AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 352–53 (2011) (Thomas, J., concurring) (joining the majority opinion “reluctantly” and expressing the view that the Court’s opinion was not adequately grounded in the text of the FAA). Or, as Justice O’Connor put it even earlier in the Court’s de­velopment of the FAA, “over the past decade, the Court has aban­doned all pretense of ascertaining congressional intent with respect to the Federal Arbitration Act, building instead, case by case, an edifice of its own creation.” 218 Allied-Bruce Terminix Cos. v. Dobson, 513 U.S. 265, 283 (1995) (O’Connor, J., concurring); see also Resnik, Fairness in Numbers, supra note 1, at 113 (explaining examination of case law, more so than statutory text, is required to understand the FAA). Consider, for example, some of the questions impli­cated by the Supreme Court’s recent decisions:

  • Whether class arbitration is consistent with the goals associated with the FAA’s federal policy in favor of arbitration; 219 See Stolt-Nielsen S.A. v. AnimalFeeds Int’l Corp., 559 U.S. 662, 685–87 (2010).
  • Whether there is an “‘inherent conflict’ between arbitration” and another federal law’s “underlying purposes”; 220 Gilmer v. Interstate/Johnson Lane Corp., 500 U.S. 20, 26 (1991) (quoting Shearson/Am. Express Inc. v. McMahon, 482 U.S. 220, 227 (1987)).
  • Whether consumers are helped or harmed by certain procedural restrictions in arbitration agreements, including class action restrictions; 221 Concepcion, 563 U.S. at 351–52.
  • Whether certain procedural restrictions in arbitration agree­ments undermine the “effective vindication” of federal laws. 222 Am. Express Co. v. Italian Colors Rest., 133 S. Ct. 2304, 2311–12 (2013).

These questions are not legal questions as much as they are policy and empirical ones. They involve assessing the facts on the ground and exercising judgment regarding how those facts bear on policy goals. And modern administrative law tends to treat such questions as usually not best addressed by courts. 223 See generally Sunstein, Beyond Marbury, supra note 196 (discussing the Court’s deference to executive branch legal interpretation).

Second, the current system for regulating arbitration is largely trans-substantive. That is, what the Supreme Court says about the FAA in the context of one area of law will also apply to another. If the Court eviscer­ates the effective-vindication doctrine as it applies to the antitrust laws, as critics charge that it did in Italian Colors, 224 See, e.g., Elhauge, How Italian Colors Guts, supra note 2. it also largely forecloses that doctrine in the context of the employment laws, for example. 225 Italian Colors itself, for example, drew on cases in the employment and consumer contexts. See Italian Colors, 133 S. Ct. at 2310–12.

Thus, delegating to agencies—in contrast to the current regime of de facto delegation to the courts—would allow for more domain-specific judgments than are possible in courts given the trans-substantive nature of the FAA. Agencies, of course, operate in specific areas as defined by Congress, and they can tailor regulation as needed to their area. And it is quite plausible that the answers to various questions bound up with the regulation of arbitration—such as whether class processes are necessary to the advancement of statutory objectives—vary depending on whether one is talking, say, about consumer protection or employment discrimi­nation. The answers might also depend on the category of the harm in­volved and whether, for example, the resulting claims are small or relatively large.

To say that agencies can make domain-specific judgments about how to regulate arbitration also necessarily implies that agencies might some­times validly decide not to regulate arbitration at all. Because the current system is one of almost total nonregulation of arbitration, this Essay has focused on the costs of unfettered arbitration, particularly in the context of adhesion contracts. But arbitration also has benefits, including the potential for speedier and less expensive resolution of disputes, that ul­timately may be reflected in the price of consumer goods, employment wages, and the like. And in particular areas, agencies might reasonably decide that those benefits outweigh any costs, including the broader so­cietal costs such as lessening of deterrence. 226 See supra notes 71–76 and accompanying text (cataloging several concerns regarding reduced deterrence). This is likely to be the case especially when claims are currently being efficiently adjudicated in arbi­tration or, perhaps, when the agency itself thinks that private enforce­ment is not as necessary to the functioning of the statutory scheme.

Importantly, however, the judgment about whether and how to reg­ulate arbitration for specific areas takes expertise, and expertise, of course, is a classic advantage of administrative agencies. 227 See, e.g., Lemos, The Other Delegate, supra note 209, at 445–48; Pardo & Watts, supra note 10, at 424–31 (explaining how, in the bankruptcy context, agencies have an expertise advantage); David B. Spence & Frank Cross, A Public Choice Case for the Administrative State, 89 Geo. L.J. 97, 135 (2000) (arguing a “crucial reason” for delegating to agencies is their expertise). Expertise is largely about information, and agencies often possess—or can more cheaply acquire—more information than the courts, which are largely passive receivers. 228 See, e.g., Lemos, The Other Delegate, supra note 209, at 445 (“Agencies can be staffed by experts in the field and, over time, they accumulate substantial experience with the issues devoted to their care. The same is not true of courts, at least not in any across-the-board way. Judges are generalists, not experts.”). Nor is there reason to think that the general point regarding agency expertise is not applicable in the context of regulating arbitration, which at least at the federal level is largely about calibrating enforcement of statutory norms. The argument is similar to the one that has been developed in the context of agency authority to create or limit private rights of action. 229 See, e.g., David Freeman Engstrom, Agencies as Litigation Gatekeepers, 123 Yale L.J. 616 (2013) (reviewing major arguments for and against allowing agencies to wield their expertise in order to determine whether a private right of action should exist in a given context); Richard J. Pierce, Jr., Agency Authority to Define the Scope of Private Rights of Action, 48 Admin. L. Rev. 1 (1996) (arguing courts should uphold agency de­terminations regarding the scope of private rights of actions so long as the determinations are reasonable); Stephenson, Public Regulation, supra note 12, at 148–49 (arguing courts interpreting a statute that is ambiguous as to whether a private right of action exists should infer delegation of authority over the question to agencies). Deciding whether to allow private arbitration of statutory claims, arbitration of such claims but with a carve-out for class litigation, or something else “involves complex policy judgments, an expert understanding of the nature and likely effect of different enforce­ment strategies, and a sensitivity to the need for a consistent and efficient ap­proach to enforcing statutory norms.” 230 Stephenson, Public Regulation, supra note 12, at 127. As Professor Matthew Stephenson argues, “The government decisionmakers with the best information about and most sophisticated understanding of these issues are likely to be the executive administrators charged with oversee­ing the public enforcement of the statutory scheme in question”—in other words, agencies. 231 Id. Part of this relative advantage is for the simple reason that agencies know about their own enforcement capabilities, in­cluding their abilities to substitute for private enforcement in some of the ways described above. 232 See supra section II.A.

For similar reasons, agencies’ informational advantages may make them better regulators than not only the courts but also Congress. 233 See Matthew C. Stephenson, Statutory Interpretation by Agencies, in Research Handbook on Public Choice and Public Law 285, 286–90 (Daniel A. Farber & Anne Joseph O’Connell eds., 2010) [hereinafter Stephenson, Statutory Interpretation] (suggest­ing one reason for delegation to agencies is that “agencies may have better access to in­formation about the connections between policy choices and actual outcomes” than Congress). In recent years, members of Congress have introduced several bills that would ban predispute arbitration clauses entirely in certain areas. 234 See Myriam Gilles, The Day Doctrine Died: Private Arbitration and the End of Law, 2016 U. Ill. L. Rev. 371, 401 & n.186. These bills generally have not proceeded very far in the legislative pro­cess, and some people have expressed concern that banning arbitration completely would be premature and overbroad. 235 See, e.g., Peter B. Rutledge, Point: The Case Against the Arbitration Fairness Act, Disp. Resol. Mag., Fall 2009, at 4, 4 (making the case against the Arbitration Fairness Act, in part because it “does not simply address the bad cases while preserving the good [but] [i]nstead . . . proposes a systemic overhaul that categorically bans predispute agreements entirely”). Given the current state of knowledge regarding the welfare effects of arbitration overall, a better path may be that taken by Dodd-Frank, which, as discussed, re­quires a two-step process involving information gathering by the CFPB and then possible regulation. 236 See supra note 122 and accompanying. That approach harnesses agencies’ gen­erally better capabilities to acquire information relevant to tailoring regulatory regimes and to then put those regimes into place.

Greater flexibility is another traditional advantage agencies have over Congress and the courts. 237 See, e.g., Pardo & Watts, supra note 10, at 443 (arguing that “agencies are better able to adapt rules to respond to new information, different facts, or changed circum­stances than Congress (which is constrained by political roadblocks and institutional bar­riers) or the courts (which are constrained by stare decisis)”). Because agency rulemaking, though not without difficulties, is generally more adaptable than either the legislative process or a system of judicial precedent, agencies can more easily “im­plement policy changes in response to new information, changing cir­cumstances, or shifting political preferences.” 238 Stephenson, Public Regulation, supra note 12, at 140. Such adaptability may be particularly important when dealing with issues involving the optimal calibration of enforcement, where information about changes to the overall regulatory system brought about by, for example, the greater availability of class actions may be difficult to come by ex ante. 239 Cf. id. at 139–40 (making a similar point about agency regulation of private rights of action). This point is even more pronounced when one focuses on delegation to courts versus agencies specifically, since courts are both strongly con­strained by notions of stare decisis and unlikely to be able to efficiently gather information regarding the real-world effects of their decisions once made. 240 See Pardo & Watts, supra note 10, at 444 (discussing how stare decisis limits courts’ flexibility in interpreting the Bankruptcy Code); Stephenson, Public Regulation, supra note 12, at 141 (“Courts are likely to know considerably less than either the legisla­ture or the executive about how well private enforcement is working in practice.”).

A final advantage to greater agency authority as compared to courts’ is agencies’ ability to partner with self-regulatory or private organizations in order to gather information, develop consensus standards, and the like. The primary model here is the SEC’s relationship with organizations like the New York Stock Exchange that regulate their members subject to the SEC’s oversight and control. 241 See Richard H. Pildes, Putting Power Back into Separation of Powers Analysis: Why the SEC-PCAOB Structure Is Constitutional, 62 Vand. L. Rev. En Banc 85, 88–89 (2009), []. The parallel model in the arbitration context would involve agency partnership with organizations like the American Arbitration Association, which, apart from providing individual arbitrators to decide disputes, also perform “regulatory” functions in­cluding promulgating ethical standards and rules of procedure for arbi­trations falling under their umbrella. 242 See, e.g., Am. Arbitration Ass’n, The Code of Ethics for Arbitrators in Commercial Disputes (2004),
003867&revision=latestreleased []; Am. Arbitration Ass’n, Commercial Arbitration Rules and Mediation Procedures (2013), [http://perma.
Such organizations are natural wellsprings of knowledge, but their ability to impart such knowledge to courts is limited and episodic. Agencies, by contrast, not only could be directed to work with such organizations by legislation delegating agen­cies authority but also could naturally draw on such organizations’ knowledge when writing the rules.

Finally, as opposed to across-the-board legislation like the Arbitration Fairness Act, 243 The Arbitration Fairness Act would ban predispute arbitration agreements alto­gether for consumer and employment claims. See Hiro N. Aragaki, Does Rigorously Enforcing Arbitration Agreements Promote “Autonomy”?, 91 Ind. L.J. 1143, 1166 (2016) (citing Arbitration Fairness Act of 2015, S. 1133, 114th Cong. § 3 (2015)). there is good reason to think that Congress is more likely to delegate domain-specific authority to an agency. First, for some of the reasons given above, a congressional ban on arbitration would not actually be optimal policy for areas where the benefits of arbi­tration outweigh the costs. And whatever the policy merits, familiar rea­sons suggest that targeted delegations to agencies may simply be more likely than legislation that spells out the rules itself. Indeed, evidence suggests that Congress often delegates authority to an agency not only to harness the superior informational resources of agencies but because it is easier to forge consensus on delegation than on policy details. As Professor Margaret Lemos has explained,

Even on issues that Congress is competent to resolve, a variety of factors may make decisions difficult or politically unfeasible. For example, a majority of legislators may agree on a general policy direction but not on the details. By leaving implementa­tion to an agency, the enacting coalition is able to achieve its larger policy goals without fighting—and perhaps splintering—over the issues that divide it. 244 Lemos, Consequences, supra note 209, at 369; see also, e.g., Joseph A. Grundfest & A.C. Pritchard, Statutes with Multiple Personality Disorders: The Value of Ambiguity in Statutory Design and Interpretation, 54 Stan. L. Rev. 627, 641 (2002) (“When faced with a conflict among competing legislative coalitions, carefully crafted ambiguous language can allow legislators with divergent interests to adopt competing, inconsistent interpretations of the same statutory text.”); Stephenson, Statutory Interpretation, supra note 233, at 286–87 (“The cost of drafting statutory language that is both clear and complex is perhaps the most straightforward explanation for statutory imprecision.”).

The domain-specific nature of agency delegation also arguably makes such legislation more likely to pass through Congress than a trans-substantive revision of the FAA. Even in an era of hyperpartisanship and political gridlock, 245 See Jody Freeman & David B. Spence, Old Statutes, New Problems, 163 U. Pa. L. Rev. 1, 15 (2014) (noting the decline of political moderates in both houses of Congress and stating “polarization in Congress is at its highest level since the end of Reconstruction”). domain-specific legislation does continue to get passed, albeit at a lessened pace. Some of this legislation is a response to real or perceived crises and presents the opportunity for logrolling on issues in what are sometimes sprawling bills. Dodd-Frank, which provided one of the main examples of Congress delegating authority to regulate arbitration to an agency, is one prominent example. 246 See id. at 16 (discussing the crisis motivation behind passing Dodd-Frank).

Thus, the traditional justifications for delegation to agencies at least arguably apply with as much force to delegations of regulatory authority over arbitration as they do to other types of delegation, especially when contrasted with the current system of de facto delegation to the courts. 247 In addition to the above, political accountability provides another reason for preferring agency regulation to that by the courts, if not Congress. See Lemos, The Other Delegate, supra note 209, at 448–50 (“Due to constitutional protections of life tenure and guaranteed salary, federal judges are insulated from direct political control.” (footnote omitted)); Pardo & Watts, supra note 10, at 432 (“Another reason frequently given for allowing administrative agencies to engage in policymaking when filling the gaps in stat­utes relates to political accountability.”); Stephenson, Public Regulation, supra note 12, at 136–39; see also Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 865–66 (1984) (“While agencies are not directly accountable to the people, the Chief Executive is, and it is entirely appropriate for this political branch of the Government to make such policy choices . . . in light of everyday realities.”). On accountability vis-à-vis Congress, see Stephenson, Public Regulation, supra note 12, at 135–36 (arguing “delegating authority from one elected branch (Congress) to another one (the executive) does not entail any obvious accountability loss, unless one believes that decisionmaking by the former inher­ently involves a greater degree of political accountability than decisionmaking by the latter”). This suggests that Congress should continue to examine express delega­tions to agencies over either reasserted congressional control or a con­tinued system of judicial regulation. The next subsection turns to how courts should respond when confronted with agencies invoking an im­plied, not express, authority to regulate arbitration.

4. Courts Should Rethink How They Approach Implied Agency Authority. — This subsection attempts to cut through the legal thicket that has been created when agencies attempt to regulate arbitration—for example, by banning it entirely or regulating the terms of arbitration agreements—using authority implied under Chevron U.S.A. Inc. v. Natural Resource Defense Council, Inc. 248 467 U.S. at 842; see also infra notes 264–269 and accompanying text. Courts have, for the most part, been hostile to claims by agencies that they may regulate arbitration when the statute does not mention arbitration specifically. 249 See, e.g., D.R. Horton, Inc. v. NLRB, 737 F.3d 344, 356, 362 (5th Cir. 2013) (rejecting deference to the NLRB’s 2012 decision in D.R. Horton finding that the Fair Labor Standards Act barred a waiver of the right to class processes); Sutherland v. Ernst & Young LLP, 726 F.3d 290, 297 & n.8 (2d Cir. 2013) (same); Owen v. Bristol Care, Inc., 702 F.3d 1050, 1054 (8th Cir. 2013) (same); Davis v. S. Energy Homes, Inc., 305 F.3d 1268, 1279–80 (11th Cir. 2002) (refusing to defer to the FTC’s interpretation of the Magnuson-Moss Warranty Act (MMWA) barring arbitration of warranty claims); Walton v. Rose Mobile Homes LLC, 298 F.3d 470, 479 (5th Cir. 2002) (same). But see Morris v. Ernst & Young, LLP, 834 F.3d 975, 979 (9th Cir. 2016) (upholding the NLRB’s decision in D.R. Horton that contractual bans on class processes violate the labor laws, although on grounds different from those urged below); Lewis v. Epic Sys. Corp., 823 F.3d 1147, 1159–60 (7th Cir. 2016) (same); Kolev v. Euromotors W., 658 F.3d 1024, 1029 (9th Cir. 2011) (deferring to the FTC’s interpretation of the MMWA on grounds similar to those of the district court, although the opinion was later withdrawn), withdrawn, 676 F.3d 867 (9th Cir. 2012). The Supreme Court has recently granted three petitions for certiorari to resolve the circuit split over the NLRB’s authority. See Karmasek, supra note 13. These courts have generally found that the presence of a cross-cutting federal statute—the FAA—as well as the strong “federal policy favoring arbitration” undercut tradi­tional principles of deference to agency legal interpretations. 250 See, e.g., D.R. Horton, 737 F.3d at 358–60. In many cases they have not seriously engaged with the interaction of Chevron and the FAA, instead reading the Supreme Court’s recent arbitration opin­ions broadly as expressing extreme reluctance to set aside arbitration agreements on any grounds. 251 Commentators have offered various ways to vindicate agency authority under particular statutory schemes, though not on the exact grounds offered below. A number of articles have explored issues surrounding deference to the NLRB’s determination in the D.R. Horton case and subsequent invalidation by the Fifth Circuit. See Catherine L. Fisk, Collective Actions and Joinder of Parties in Arbitration: Implications of DR Horton and Concepcion, 35 Berkeley J. Emp. & Lab. L. 175, 179 (2014) (criticizing persuasively the Fifth Circuit’s D.R. Horton decision on various grounds but not generally taking up the interac­tion of the FAA with Chevron); Stephanie Greene & Christine Neylon O’Brien, The NLRB v. the Courts: Showdown over the Right to Collective Action in Workplace Disputes, 52 Am. Bus. L.J. 75, 128–29 (2015) (stating, without extensive discussion, that the NLRB should receive Chevron deference for its determination that bans on class actions violate the labor laws); Charles A. Sullivan & Timothy P. Glynn, Horton Hatches the Egg: Concerted Action Includes Concerted Dispute Resolution, 64 Ala. L. Rev. 1013, 1032 (2013) (arguing D.R. Horton provided the “correct” and therefore “reasonable” interpre­tation of the NLRA but not taking up whether Chevron should apply generally). One student note does take up, in more extensive fashion, whether Chevron should apply to the NLRB’s determination in D.R. Horton. See generally Note, Deference and the Federal Arbitration Act: The NLRB’s Determination of Substantive Statutory Rights, 128 Harv. L. Rev. 907 (2015). However, that note’s argument would grant deference to whether collective action is a “substantive right” under the NLRA and thus inherently unwaivable in arbitration. Id. at 921. The argument this Essay sketches below—which grants deference to an agency’s determination that its organic statute contains a command contrary to the FAA—sidesteps the inherently slippery sub­stance–process distinction, which is arguably one governed more by the FAA itself than the statute over which the agency has interpretive authority.
Outside the D.R. Horton context specifically, see Rhonda Wasserman, Legal Process in a Box, or What Class Action Waivers Teach Us About Law-Making, 44 Loy. U. Chi. L.J. 391, 428–31 (2012) (arguing that agencies should receive deference for portions of their rul­ings on arbitration but then identifying several “formidable arguments against deference” for portions of agency decisions attempting to reconcile the agencies’ interpretations with the FAA). As this Essay argues below, if an agency reasonably interprets its statute as being contrary to the FAA, there is in fact no further reconciliation issue for either the courts or agency to decide. See Weston, supra note 5, at 133–36 (suggesting Chevron should trump the FAA because the agency is the more expert and specialized body but not expanding on the doctrinal argument); Grossberg, supra note 170, at 680–83 (making the case for Chevron deference specifically in the context of the MMWA but not relying on a “contrary congressional command” argument).

This subsection argues that, as a general matter, agencies should have authority to reasonably interpret their statutes in order to regulate arbitration. In doing so, it lays out a more straightforward path than liti­gants and scholars have generally followed. In a nutshell, the argument is this: The FAA sets the background legal regime governing arbitration. It makes arbitration agreements enforceable like any other contract. But within their regulatory domains, agencies have the authority to depart from background law in order to regulate more stringently (within the reasonable bounds set by their statutes). In doing so, agencies are exer­cising a delegated authority that authorizes them to rearrange back­ground legal rights and responsibilities. The FAA need not be read to wall off agencies from doing so when it comes to arbitration. After all, the purpose of the FAA is to put arbitration agreements “on an equal footing with other contracts,” not to privilege them. 252 AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 339 (2011).

This approach differs from how the courts have addressed the issue. For the most part, it has been assumed that an agency regulating arbitra­tion must fit such regulation into one of the exceptions to the FAA or otherwise “harmonize” the statute being interpreted with the FAA; and, because the agency does not have interpretive authority over the FAA, courts have limited deference to the agency in such situations. 253 See supra note 249 and accompanying text (citing cases in which courts were hostile to agencies’ claims that they could regulate arbitration using an implied authority). Even the recent Seventh Circuit opinion upholding the NLRB’s decision in D.R. Horton sought to explain why the NLRB’s determination fits within the FAA’s “saving clause,” an argument made considerably more difficult by Concepcion. See Lewis, 823 F.3d at 1157–60. But, as explained above, 254 See supra note 58 and accompanying text (discussing Justice Scalia’s majority opinion in Italian Colors, which held there was no “contrary congressional command” permitting the Court to invalidate the arbitration agreement or its ban on class processes). the Supreme Court has already said that the FAA must give way to a “contrary congressional command” limiting arbitra­tion. 255 This also distinguishes cases such as Hoffman Plastic Compounds, Inc. v. NLRB, in which the Court has found that agencies may not order relief made illegal by federal stat­utes that are not administered by the agency. See 535 U.S. 137, 151–52 (2002). Those cases stand for the proposition that agencies may not override other federal policies when they trump those pursued by the agency. In the case of the FAA, however, the Court has already spoken as to which policy prevails: When a contrary congressional command ex­ists, the FAA—as a background default rule—gives way. Thus, the question is not so much one of harmonizing the FAA with later enactments but determining the existence and scope of the allegedly contrary command.

That question should be within the purview of the agency in the first instance. Whether there is a contrary congressional command bearing on the arbitration of certain claims is a question, essentially, of the mean­ing of the non-FAA statute being invoked and whether that statute “evince[s] an intention to preclude a waiver of judicial remedies for the statutory rights at issue.” 256 Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 628 (1985). That much is clear from the Supreme Court’s own case law. Take Italian Colors, which, as explained above, included a brief discussion of whether the antitrust laws (or other federal statutes) contained a command contrary to the FAA. 257 See supra note 58 and accompanying text. The Court answered that question by looking to the antitrust laws, asking whether Congress ex­pressed an intent or adopted a policy that limited the general right to enter arbitration agreements on terms set by the parties. 258 See Am. Express Co. v. Italian Colors Rest., 133 S. Ct. 2304, 2309–10 (2013) (“The antitrust laws do not ‘evinc[e] an intention to preclude a waiver’ of class-action procedure.” (quoting Mitsubishi, 473 U.S. at 628)). The same is true of CompuCredit Corp. v. Greenwood, which addressed whether the Credit Repair Organizations Act (CROA) established an unwaivable right to judicial proceedings (thus precluding arbitration). 259 132 S. Ct. 665, 668 (2012). There, the Court reviewed the text and structure of the CROA in order to determine whether Congress intended to limit arbitration. 260 See id. at 670–73. As in Italian Colors, the Court answered that it did not. 261 Id. at 673.

Italian Colors, CompuCredit, and the other Supreme Court cases on whether a contrary congressional command limits the arbitration right have not involved an interpretation of the underlying statute by an agency in charge of administering it, such as in the NLRB and FTC ex­amples above. 262 See supra section II.B.4. The presence of such an interpretation raises the ques­tion whether the agency’s preferred interpretation is entitled to Chevron deference. There is a strong argument, contrary to most decisions by the courts, 263 See supra note 249. that it should be.

Chevron, of course, requires the courts to accept reasonable agency interpretations of statutes over which the agency has authority, provided Congress has not “directly spoken to the precise question at issue.” 264 Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 842 (1984). Whether Congress has directly spoken to the issue is resolved by the court “employing traditional tools of statutory construction.” Id. at 843 n.9. In other words, if a court determines that Congress has not spoken directly to the question at issue and there is thus an ambiguity, the court must defer to the agency’s reasonable resolution of that ambiguity even if the court would have chosen a different interpretation on its own. 265 Needless to say, this brief introduction glosses over many subtleties of the doc­trine that are not relevant here. See, e.g., Matthew C. Stephenson & Adrian Vermeule, Chevron Has Only One Step, 95 Va. L. Rev. 597 (2009) (suggesting Chevron has only one “step,” contrary to traditional formulation); Cass R. Sunstein, Chevron Step Zero, 92 Va. L. Rev. 187 (2006) (exploring various issues regarding de­termining when Chevron applies). Chevron’s foundations are much debated. 266 See Kent Barnett, Codifying Chevmore, 90 N.Y.U. L. Rev. 1, 14–22 (2015) (cat­aloguing points of academic disagreement regarding the source of Chevron deference). However, a widely shared and powerful account holds that Chevron represents a rule of thumb re­garding Congress’s intent when it passes legislation investing an agency with lawmaking power—namely, that Congress intends the agency and not the courts to be the principal interpreters of that legislation. 267 See, e.g., Sunstein, Beyond Marbury, supra note 196, at 2589–90; see also David J. Barron & Elena Kagan, Chevron’s Nondelegation Doctrine, 2001 Sup. Ct. Rev. 201, 212 (“Because Congress so rarely makes its intentions about deference clear, Chevron doctrine at most can rely on a fictionalized statement of legislative desire, which in the end must rest on the Court’s view of how best to allocate interpretive authority.”); Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 517 (“[T]he quest for the ‘genuine’ legislative intent is probably a wild-goose chase anyway.”). That rule of thumb is itself rooted in the realist-inspired notion that the reso­lution of legal ambiguity involves policy judgment, coupled with the functional reasons (outlined above 268 See supra section III.A.3. ) that should lead us to prefer agen­cies over courts as policymakers. 269 See Stephenson, Public Regulation, supra note 12, at 149; Sunstein, Beyond Marbury, supra note 196, at 2591, 2596.

If it is right that the existence of a “contrary congressional com­mand” turns on the meaning of the underlying statute, then the same functional reasons that cause courts to defer to agency interpretations generally should counsel deference to an agency’s interpretation of that statute. 270 Cf. Stephenson, Public Regulation, supra note 12, at 149 (“The policy arguments favoring a judicial presumption that ambiguous statutes should be read to confer on agen­cies the discretion to determine the proper scope of private enforcement are precisely the same arguments that favor express congressional delegation of such authority: superior agency expertise, flexibility, and political accountability.”). Agencies are simply better-positioned institutional actors than courts are to determine whether the statutes they administer should be read to ban or otherwise regulate arbitration, as long as the statute in question can reasonably be read to do so.

So then, under normal deference principles, there is a strong case for allowing agencies to interpret their organic statutes to ban or other­wise regulate arbitration. The harder question is whether normal defer­ence principles apply. The difficulty is that there is arguably a clear statement principle that applies when ascertaining whether a contrary congressional command exists sufficient to override the FAA and its “federal policy” in favor of arbitration. If that clear statement rule is strong enough that it requires, in cases of ambiguity, statutes to be read not to encroach on the FAA, then it arguably overrides an agency’s usual authority under Chevron to resolve ambiguities as the agency sees fit. How to reconcile such clear statement principles with Chevron is the subject of much debate. 271 See generally Curtis A. Bradley, Chevron Deference and Foreign Affairs, 86 Va. L. Rev. 649 (2000) (discussing how to reconcile Chevron with various foreign affairs canons); Cass R. Sunstein, Nondelegation Canons, 67 U. Chi. L. Rev. 315 (2000) [hereinafter Sunstein, Nondelegation Canons] (discussing this problem generally).

In the case of regulating arbitration, there are good reasons to allow an agency’s Chevron authority to prevail. As an initial matter, it is unclear how much clarity the Court actually requires in the context of arbitra­tion. Although in practice the Supreme Court has read statutes parsimo­niously in order to avoid holding that later statutes create an unwaivable right to a judicial forum, the Court has stopped short—especially in re­cent cases—of announcing that it was placing a very heavy thumb on the scale in favor of the FAA. In Italian Colors, the Court did not mention any clear statement principle in its discussion of whether a contrary congres­sional command existed. 272 See Am. Express Co. v. Italian Colors Rest., 133 S. Ct. 2304, 2309–10 (2013). CompuCredit, the last case to focus squarely on the issue, comes closer to applying a clear statement principle. There, Justice Scalia’s opinion for the Court mentions that Congress has spoken with “clarity” when it has limited arbitration in some way, 273 See CompuCredit Corp. v. Greenwood, 132 S. Ct. 665, 672 & n.4 (2012). but the opin­ion does not state what level of “clarity” would be required to do so nor respond to the dissent’s assertion that the FAA contains no “magic words” requirement. 274 See id. at 680 (Ginsburg, J., dissenting). Older cases often recited that the “burden is on the party opposing arbitration . . . to show that Congress intended to preclude a waiver of judicial remedies,” but did not specify what the bur­den is. 275 E.g., Shearson/Am. Express, Inc. v. McMahon, 482 U.S. 220, 227 (1987). By contrast, in other areas the Court has been rather specific that it is requiring “Congress speak with a clear voice” 276 Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 17 (1981). or that its “in­tent . . . must be obvious from ‘a clear legislative statement’” when un­dertaking certain forms of legislation. 277 Seminole Tribe of Fla. v. Florida, 517 U.S. 44, 55 (1996) (quoting Blatchford v. Native Vill. of Noatak, 501 U.S. 775, 786 (1991)) (describing this standard in the context of abrogating state sovereign immunity).

That the Court would not require utmost clarity when considering whether a later statute bears on arbitration makes sense when one considers that any such clear statement principle is most plausibly rooted in the canon against implied repeals. 278 See Richard A. Nagareda, The Litigation–Arbitration Dichotomy Meets the Class Action, 86 Notre Dame L. Rev. 1069, 1093 (2011) (“Like implied repeals of statutes gener­ally, implied repeals of the FAA are disfavored, such that the inclusion of a private right of action in another statute—even an unwaivable right—will not operate to displace the FAA.”). That canon is itself related to the principle that “when two statutes are capable of co-existence, it is the duty of the courts, absent a clearly expressed congressional intention to the contrary, to regard each as effective.” Morton v. Mancari, 417 U.S. 535, 551 (1974). The aversion to reading later statutes as altering earlier ones is strongest when the earlier statute is the more specific. 279 See Morton, 417 U.S. at 550–51 (refusing to hold a generally applicable statute impliedly repealed a previously enacted, more specific statute). The FAA, 280 The FAA, recall, was first passed in 1925. See supra note 14 and accompanying text. however, is likely to be the earlier statute but also more general than subject-area-specific statutes that might be read to limit it. It thus runs smack dab into the other canon that the more specific provisions control and that “[w]here there is no clear in­tention otherwise, a specific statute will not be controlled or nullified by a general one, regardless of the priority of enactment.” 281 Morton, 417 U.S. at 550–51.

That the specific controls the general does not end the matter, how­ever. After all, one might object that the FAA is actually the more specific statute in that it mentions arbitration whereas the statute being invoked by the agency (by hypothesis) does not. Establishing that an agency’s Chevron authority allows it to regulate arbitration—or, at least, why there is a plausible legal basis that it should—requires a more nuanced theory of both the FAA and of Chevron. The purpose of the FAA, in the long­standing view of the Supreme Court, is to protect arbitration from judicial hostility and to place arbitration agreements “on an equal foot­ing with other contracts.” 282 AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 339 (2011). The Act thus establishes that, as a general matter, agreements to arbitrate are valid and enforceable, just as any other contract would be. It thus provides a background legal rule govern­ing the rights of those who enter such agreements.

Now consider Chevron. Chevron tells courts that when a statute is am­biguous, the court must defer to the agency’s reasonable interpretation, because (under the implied delegation rationale) Congress has chosen the agency and not the court as the body to resolve the ambig­uity. 283 See supra notes 267–269 and accompanying text. And such a delegation is, by its nature, the power to disrupt back­ground legal rules within prescribed bounds.

This logic extends to defeat a clear-statement-type presumption that would deny Chevron deference to agency interpretations bearing on arbi­tration. A presumption against regulation of arbitration agreements pre­sumably serves values similar to the canon against implied repeals: pro­tecting legal continuity and preserving existing legal relationships and rights to the extent possible. 284 See David L. Shapiro, Continuity and Change in Statutory Interpretation, 67 N.Y.U. L. Rev. 921, 937 (1992) (defending the canon against implied repeal on these grounds). It is possible, of course, that the presumption simply reflects a strong policy in favor of arbitration specifically. However, it is difficult to see why that policy should be given quasi-constitutional status and elevated above policies reflected in later-enacted stat­utes that might bear on the subject. But when Congress writes a flexible stat­ute and delegates authority over that statute to an administrative agency, it is implicitly rejecting these very same values. That is because, as stated, a decision to delegate lawmaking authority to an agency is a decision to allow the agency to disrupt existing legal relationships (within the bounds set by the statute) in service of statutory purposes. 285 See, e.g., Bradley, supra note 271, at 672 (“Agencies are understood under Chevron to be engaging in delegated lawmaking, so there is no reason why they cannot change the law.”). And it is a decision to enable an agency—and not the courts—to decide in the ini­tial instance which legal relationships should be so disrupted. In addi­tion, unlike with “delegations” to courts, which are bound by strict no­tions of stare decisis, agency interpretations of their statutes remain contestable over time, 286 See id. (“A significant implication of both Chevron’s realist premise and its delega­tion presumption is that agencies may be entitled to deference even if they change their interpretation.”). with the effect that agency interpretations oper­ate less like repeals in the formal sense.

Thus, under the prevailing implied delegation rationale for Chevron, there is a good argument for allowing an agency’s interpretation of its statute to prevail. The question then becomes whether Congress should be barred from making the choice to impliedly delegate such authority to an agency or whether Congress should be forced to decide more particu­larly whether to regulate arbitration. This is the sense in which, as Professor Cass Sunstein has written, certain clear statement principles operate as “nondelegation canons [that] forbid administrative agencies from making certain decisions on their own” and thus thrust them back onto the legislative branch. 287 Sunstein, Nondelegation Canons, supra note 271, at 316. As Professor Sunstein and others have writ­ten, such nondelegation canons may make particular sense when consti­tutional issues are at play, when the interests of other sovereigns (such as foreign countries or Native American tribes) are at stake, or when there is otherwise a “highly sensitive decision[]” to be made. 288 Id. at 317; see also Bradley, supra note 271, at 677–78 (discussing issues for which the Supreme Court has determined policy-oriented statutory canons of construction trump Chevron deference).

Despite the Supreme Court’s insistence on the strong “federal policy favoring arbitration,” 289 Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983). there is a plausible case that the policy should not rise to the level of a nondelegation canon requiring evidence of ac­tual congressional deliberation in order to depart from the background rules set by the FAA. Whether to regulate arbitration in a particular area does not implicate sovereignty concerns. It does not, for example, trench upon state policy as an agency does when it interprets an unclear statute to preempt state law. As Professor Thomas Merrill has written, such cases involve decisions implicating dimensions on which “courts are likely to perform better than agencies[,] . . . such as preserving fidelity to the Constitution’s division of powers, . . . preserving stability in that di­vision, . . . preserving an overall balance between state and federal au­thority, and the need to discern accurately the . . . state law alleged to be in tension with federal law.” 290 Merrill, supra note 187, at 772. A decision about whether to regulate arbi­tration in a particular area, by contrast, is a plain-vanilla matter of federal policy. And on such questions, courts’ deference to agency legal inter­pretations under Chevron is particularly appropriate. 291 See id. Indeed, it is a commonplace feature of agency regulation. One would have little doubt, for example, that the FCC can regulate contracts that fall within its pur­view more stringently than the background antitrust law does, for example. 292 See generally, e.g., U.S. Telecom Ass’n v. FCC, 825 F.3d 674 (D.C. Cir. 2016) (up­holding an FCC ban on certain dealings by Internet service providers that would be per­missible and enforceable under the antitrust laws).

5. Summing Up and Stepping Back: Just How Much Authority Does Current Law Provide? — If there is in fact a good legal basis for agencies claiming greater authority to regulate arbitration than has been widely assumed, one question is just how far that authority extends. In other words, if agencies together pushed their authority to its maximum, in what areas could agencies regulate arbitration, and in what areas would the current system of court-led regulation still prevail absent further congressional delegation? 293 This is not, of course, to say that agencies should automatically extend regulation as far as they can push it. Still, it is useful to know the limits of that authority should it exist. Although this is not the place to authorita­tively canvass every potential site of agency authority over arbitrationwhich would necessarily involve consideration of the text, statutory pur­pose, and context of those agencies’ authorizing statutessome general points are worth attention.

First, the antitrust agencies—as well as specialized regulators such as the FCC—have relatively broad and malleable authority to impose re­strictions on arbitration as merger conditions, at least with regard to anti­trust claims. 294 See supra notes 176–183 and accompanying text. However, their ability to do so is necessarily episodic, for it requires a merger decision that triggers scrutiny by government authori­ties. 295 See Lemley & Leslie, supra note 5, at 56–57 (discussing the scope and limits of federal merger review). Thus, though the ability to impose arbitration-related merger re­strictions is theoretically broad, it is in practice likely to be somewhat limited. The types of conditions pursued by the SEC regarding securities-related arbitration, 296 See id. at 57–58. and by the Department of Education regarding certain federal loan programs, 297 See supra notes 184–185 and accompanying text (noting the Department of Education’s proposal to condition colleges’ ability to participate in the federal Direct Loan program on those colleges not requiring arbitration of certain disputes). are also likely to be somewhat limited, particularly with regard to the kinds of claims to which the restrictions might apply.

Second, outside of the conditions context, some agencies, such as the CFPB and SEC, have clear authority to regulate arbitration within their spheres. 298 See supra notes 122–126 and accompanying text (discussing the statutory author­ization for the CFPB and SEC to regulate arbitration). These present easy cases under the law, if not always as a policy matter. Outside of these easy cases, of course, is where the uncer­tainty lies. Because the Chevron framework requires attention to the par­ticular statute being invoked as the basis of the agency’s authority, 299 See supra notes 264–265 and accompanying text. a full analysis of federal agencies’ authority to regulate arbitration is out­side the scope of this Essay. Some generalizations, however, can be made.

For the reasons given above, 300 See supra notes 279–286 and accompanying text. whether agency authority exists will often depend on whether the activity in question (here, the formation of a contract) is subject to some form of specialized regulation or left to background law (and background regulators). In particular, when a do­main-specific regulator is drawing on authority to fashion separate rules in its area of expertise, there is a greater warrant for extending deference to its decision to regulate arbitration. 301 See supra notes 279–286 and accompanying text (discussing why domain-specific agency authority to regulate arbitration should prevail over the FAA). And whether such a regulator will exist depends, largely, on the industry or activity in which the activity arises. Take, as an example, the FCC and the regulation of Internet-based products and services. The FCC has broad authority to regulate “prac­tices . . . for and in connection with” telecommunications services in or­der to ensure that such practices are “just and reasonable.” 302 47 U.S.C. § 201(b) (2012). And the D.C. Circuit has repeatedly found that the authority granted to the FCC to define “just and reasonable” practices grants the FCC substantial dis­cretion. 303 See, e.g., Orloff v. FCC, 352 F.3d 415, 420 (D.C. Cir. 2003). Such a grant of authority—and similar grants to other special­ized regulators—could well extend to imposing regulations regarding arbitration.

The FCC’s authority extends only to providers of “telecommunica­tions service,” however—traditionally, telephony and other legacy com­munications services but now also providers of mobile and wired broadband internet like Verizon and Comcast. 304 See U.S. Telecom Ass’n v. FCC, 825 F.3d 674, 700 (D.C. Cir. 2016) (upholding the FCC’s determination that broadband internet is a telecommunications service subject to Title II of the Communications Act). To be sure, this au­thority encompasses an important part of the economy, and one in which com­panies have been especially active in requiring arbitration of consumer claims. 305 See supra note 134 and accompanying text. But even in the Internet area, it does not extend to mobile applications, websites like Amazon, or other “edge providers” that may find themselves in disputes with their consumers. 306 See Protecting & Promoting the Open Internet, 30 FCC Rcd. 5601, 5745–46, at ¶ 336 (2015) (extending FCC authority over broadband internet access providers but not other internet players). These providers, like other companies in the economy not subject to special­ized regulation, are still subject to consumer-protection regulation under the general authority granted to the FTC. 307 See Babette E.L. Boliek, FCC Regulation Versus Antitrust: How Net Neutrality Is Defining the Boundaries, 52 B.C. L. Rev. 1627, 1668–70 (2011) (describing generally the interaction between FCC and FTC regulation). But while that authority—which grants the FTC the ability to prohibit “[u]nfair methods of compe­tition in or affecting commerce, and unfair or deceptive acts or practices in or affecting commerce” 308 15 U.S.C. § 45(a)(1) (2012). —looks superficially similar to that held by the FCC, historically the FTC has been less willing to advance, and the courts less likely to entertain, broad agency claims to interpret that au­thority and implement those interpretations through rulemaking. 309 See generally Justin (Gus) Hurwitz, Chevron and the Limits of Administrative Antitrust, 76 U. Pitt. L. Rev. 209 (2014) (tracing the evolution of the “folk knowledge” that Chevron deference does not apply to the FTC’s interpretations of section 5 of the Federal Trade Commission Act). Thus, broad authority over arbitration-related consumer practices may be less likely to extend to firms that fall only within the residual jurisdiction of the FTC.

A similar dynamic can be observed in the labor and employment ar­eas. Although, as discussed above, the labor laws may provide the NLRB with potentially broad authority to regulate different kinds of restrictions in employment agreements subject to laws such as the National Labor Relations Act, 310 See supra notes 164–166 and accompanying text (discussing the NLRB’s position that an employment agreement mandating arbitration and banning class arbitration was an illegal restriction on rights protected by the labor laws). employment relationships falling outside such statutes may be more difficult to reach. For its part, the jurisdiction of the EEOC is more narrowly limited to addressing discrimination. 311 See Adam B. Kaplan, Father Doesn’t Always Know Best: Rejecting Paternalistic Expansion of the “Direct Threat” Defense to Claims Under the Americans with Disabilities Act, 106 Dick. L. Rev. 389, 398 & n.52 (2001). And the EEOC has broadly received less deference than other regulators and has a more limited ability to engage in prospective policymaking generally. 312 See generally Melissa Hart, Skepticism and Expertise: The Supreme Court and the EEOC, 74 Fordham L. Rev. 1937 (2006) (noting the Court often declines to give def­erence to EEOC positions).

For the moment, then, the ability of agencies to directly regulate ar­bitration is not likely to extend to all contexts in which scholars have raised concerns. But it is also likely to extend much further than many have assumed. And, of course, even when direct regulation is not possi­ble, agency enforcement efforts, including through seeking remedies like disgorgement that may mimic the effects of class actions, may help pick up some of the slack.

B. Some Worries (and How to Mitigate Them)

The last section introduced some of the benefits that may flow from greater agency involvement with arbitration. This section surveys some of the potential downsides and how we might deal with them.

1. The Rise of Class Arbitration Bans Undermines Part of the Case Against Agency Disgorgement Actions. — One concern that has been raised against agency disgorgement actionsor other agency actions that mimic class actions by providing collective relief for harmsis that they lack some of the procedural protections for victims, such as hearing rights and probing judicial review of settlements, that accompany class actions in court. 313 See Zimmerman, supra note 5, at 546–56 (“Unlike class action settlements, agen­cies lack rules to ensure that victims have an adequate voice in the restitution process.”). This concern is closely linked to a concern that using public en­forcement as a true “substitute” for private litigation entails risks to pri­vate plaintiffs’ autonomy. The autonomy costs may be especially heightened when the government is attempting to vindicate a truly “private” rightas in, for example, traditional parens patriae actions 314 See generally Seth Davis, Implied Public Rights of Action, 114 Colum. L. Rev. 1 (2014) (delineating different kinds of government suits, including those that substitute for private vindication of private rights, and discussing, inter alia, potential costs to litigant autonomy). and when private parties are formally precluded from further litigation based on the prior government action. 315 See Margaret H. Lemos, Aggregate Litigation Goes Public: Representative Suits by State Attorneys General, 126 Harv. L. Rev. 486, 500 (2012) (“Although the case law on the preclusive effect of public aggregate litigation is surprisingly sparse, the prevailing view is that the judgment in a state case is binding on every person whom the state represents as parens patriae.” (internal quotation marks omitted) (quoting Susan Beth Farmer, More Lessons from the Laboratories: Cy Pres Distributions in Parens Patriae Antitrust Actions Brought by State Attorneys General, 68 Fordham L. Rev. 361, 384 (1999))). Thus, Professor Zimmerman has argued that agencies should adopt more fulsome procedures governing class-like settlements and that judges should more closely scrutinize such settlements, and the resulting distribution funds, than they now typically do. 316 See Zimmerman, supra note 5, at 564–71 (“Agency settlements lack a consistent form of judicial review.”).

Although the concern for litigant autonomy is a real one, the rise of mandatory bilateral arbitration unsettles some of the underpinnings of proposals to import further procedural protections into agency dis­gorgement actions. For one, the rise of mandatory bilateral arbitration changes the relevant baseline from one in which a class action—with all its protections for victims—may well be available, to one in which no form of recovery whatsoever may practically be available. This is true even for the sort of low-value claims thought to be especially amenable to class resolution. 317 See id. at 559. And whatever one thinks of the value of these suits, the point remains that the relevant comparison is in some cases not be­tween agency collective resolution and resolution through private class actions but between agency collective action and no private recovery. Moreover, requiring agencies to engage in costlier forms of procedure and heightened judicial scrutiny, as some propose, 318 See id. at 568–71 (recommending “courts take a hard look at agency settlements to determine whether a distribution plan makes reasonable, fact-based distinctions be­tween differently situated parties”). will predictably re­sult in less recovery as settlements are set aside and as agencies devote fewer resources to their pursuit. Thus, for those who are concerned prin­cipally with the compensation of victims or protection of private rights, hamstringing agencies in contexts in which collective relief may not be practically available may be the least victim-friendly option.

More generally, it is unclear whether the procedural protections af­forded to settling class members provide the correct framework through which to judge actions by federal agencies, even in a world lacking man­datory bilateral arbitration. 319 Professor Zimmerman himself acknowledges this point. See id. at 548 (“Com­mentators and courts have observed that agencies, at least theoretically, may alter substantive rights more accountably than courts, particularly when they use rulemaking procedures subject to notice and comment.”). One aspect of class litigation that may re­quire greater scrutiny is that it involves private persons—most notably, lead plaintiffs and their attorneys—pursuing private and public ends. 320 See, e.g., Richard A. Nagareda, Turning from Tort to Administration, 94 Mich. L. Rev. 899, 969 (1996) (“The more troubling limitation of an administrative perspective, however, lies . . . in the central role that private attorneys play as settlement architects. Private attorneys may have expertise in a specific area . . . akin to that of a public adminis­trative agency . . . , but private attorneys simply do not have the same political legiti­macy.”). See generally Samuel Issacharoff, Class Actions and State Authority, 44 Loy. U. Chi. L.J. 369 (2012) (discussing the relationship between class actions and state authority). Since class actions result in rearranging legal rights and entitlements, we might be particularly wary of allowing them to operate largely in private hands and without judicial oversight. 321 See Nagareda, supra note 320, at 969 (discussing the greater political legitimacy of agencies as compared to private attorneys). But when agencies act, they al­ready have the imprimatur of the state; indeed, agencies’ primary role is resolving conflicts among competing interests while rearranging rights. 322 See id. (“The authority of an administrative agency to engage in rulemaking stems from a delegation of power from Congress which, in turn, traces its regulatory authority to Article I of the Constitution.”). The Supreme Court itself made a similar point in General Telephone Co. of the Northwest v. EEOC, 446 U.S. 318 (1980), which held that actions by the EEOC to enforce collective rights were not subject to Rule 23 of the Federal Rules of Civil Procedure. There, Justice White wrote that
“the EEOC is authorized to proceed in a unified action and to obtain the most satisfactory overall relief even though competing interests are in­volved and particular groups may appear to be disadvantaged. . . . The EEOC exists to advance the public interest in preventing and remedying employment discrimination, . . . in part by making the hard choices where conflicts of interest exist.
Id. at 331.
Administrative law has its own set of doctrines governing that public ex­ercise of power, 323 See, e.g., Gillian E. Metzger, Ordinary Administrative Law as Constitutional Common Law, 110 Colum. L. Rev. 479, 511–12 (2010) (developing an account of adminis­trative law requirements as a means to enforce constitutional values implicated by state action). and it is not clear that these doctrines need to be strengthened when the form of agency action shares some features of a class action.

Finally, at least as a formal matter, agency action does not automati­cally entail preclusion of private rights to pursue litigation (or arbitra­tion) of claims. 324 Cf. Velikonja, supra note 5, at 390 (making a similar point in the context of SEC “fair funds” actions). This is important because the traditional trigger for due process protections in class cases is that the resolution of the class claim extinguishes absentee class members’ claims. 325 See Zachary D. Clopton, Transnational Class Actions in the Shadow of Preclusion, 90 Ind. L.J. 1387, 1424–25 (2015) (noting protections that must be afforded to absent class members before their claims can be extinguished). But as Professor Prentiss Cox has pointed out in a recent article, government action gen­erally does not preclude subsequent private suits, even if the government action somewhat resembles a class action in form. 326 See Prentiss Cox, Public Enforcement Compensation and Private Rights, 100 Minn. L. Rev. 2313, 2337–42 (2016). That said, as ex­plained further below, courts resolving a later private dispute may look to prior government action in shaping relief and making other determina­tions. 327 See infra notes 335–339 and accompanying text (noting courts, in attempting to limit double recovery, may reduce the amount of private awards in cases following gov­ernment enforcement actions). And when they do, they should rightfully take into account whether the autonomy interests of private parties have been adequately respected.

2. Duplication. — Another worry that has been attached to agency disgorgement actions is duplication of actions. The worry is simply enough statednamely, that if agencies pursue actions that at least in theory could also be pursued as private class actions, they will risk double recovery, waste agency resources, and cause overdeterrence. 328 This concern has been voiced most frequently with respect to SEC enforcement actions. See, e.g., Black, supra note 5; Winship, supra note 5. Professor Zimmerman has raised a more general concern that agencies lack sufficient rules to coordinate their ac­tions with actions by private parties. See Zimmerman, supra note 5, at 544–46. Several answers to this concern present themselves.

As an initial matter, existing evidence suggests that, even under ex­isting structures, duplications of recovery may not actually be a large concern. In a recent article, Professor Urska Velikonja found less waste­ful duplication caused by SEC “fair fund” actions in light of private litiga­tion than is usually presumed. 329 Velikonja, supra note 5, at 386. Specifically, of the 222 fair fund settlements identified, forty-seven involved SEC actions that could be viewed as duplicative; however, many of those cases involved SEC enforcement actions that settled long before the parallel private litigation. Id. at 388–89. Largely this was due to the SEC’s efforts to coordinate the distribution of recovered funds with those recovered through private actions. 330 See id. at 388. Outside the SEC context, there is also limited evidence for the duplication hypothesis. 331 For example, in its report on arbitration, discussed above, the CFPB found that in the over 1,000 cases it studied filed by state and federal regulators, 88% did not involve an overlapping class action complaint. CFPB Final Report, supra note 124, § 1.4.8.

Agencies can also take steps to further avoid wasteful duplication. As relevant here, agencies can develop criteria for judging whether manda­tory arbitration provisions stand in the way of effective recovery or the development of the law. 332 For example, as the CFPB has done, agencies can gather information about the incidence of mandatory bilateral arbitration provisions in the market, the size of the rele­vant harm, and other information bearing on whether private efforts are likely to be futile. See Arbitration Agreements, 81 Fed. Reg. 32,830, 32,840 (proposed May 24, 2016) (to be codified at 12 C.F.R. pt. 1040). Agencies can then use that information in de­ciding whether to pursue relief of their own. The FTC’s now-withdrawn “Policy Statement” on disgorgement provides a nice exam­ple. 333 See Notice, Policy Statement on Monetary Equitable Remedies in Competition Cases, 68 Fed. Reg. 45,820 (Aug. 4, 2003); see also Press Release, FTC, FTC Withdraws Agency’s Policy Statement on Monetary Remedies in Competition Cases; Will Rely on Existing Law (July 31, 2012), []. There, the agency laid out to the public certain criteria regarding the pursuit of disgorgement and restitution damages in antitrust cases. Included among them were the “value added” by the FTC’s pursuit, which involved con­sideration of whether “private actions likely will not remove the total unjust enrichment from a violation”; whether “practical or legal difficulties are likely to preclude compensation for those injured by a violation”; and whether injured persons are able to otherwise obtain “damages sufficient to erase an injury.” 334 68 Fed. Reg. at 45,823.

If agencies are careful about which cases to pursue, the risk of over­recovery can be effectively minimized. Nevertheless, there will inevitably be some cases in which duplicative relief is sought. In such situations, methods operating at the back end can be useful. The SEC’s efforts to coordinate its settlement funds with funds recovered through private liti­gation show one path. 335 See Velikonja, supra note 5, at 388. The FTC’s 2003 Policy Statement also reported that in two recent cases the FTC had combined funds it had recovered with privately recovered funds in order “[t]o simplify the process and avoid any appearance of duplicative payments.” 336 68 Fed. Reg. at 45,823. Courts, for their part, can limit double recovery by reducing the amount of private awards, ex­ercising appropriate discretion regarding whether such a setoff is appro­priate. 337 See SEC v. First Jersey Sec., Inc., 101 F.3d 1450, 1475 (2d Cir. 1996) (“A settle­ment payment may properly, however, be taken into account by the court in calculating the amount to be disgorged . . . .”); SEC v. Tex. Gulf Sulphur Co., 446 F.2d 1301, 1307 (2d Cir. 1971) (establishing a recovery fund to be held in escrow for five years, during which time any funds recovered by the SEC would be subtracted from it); see also Gen. Tel. Co. of the Nw. v. EEOC, 446 U.S. 318, 333 (1980) (noting in passing that “the courts can and should preclude double recovery by an individual”). Finally, in extreme cases, courts might deny class certification altogether in duplicative cases, on the grounds that, in light of a prior agency recovery, class litigation is not “superior to other available meth­ods for fairly and efficiently adjudicating the controversy.” 338 Fed. R. Civ. P. 23(c); see also Steven B. Malech & Robert E. Koosa, Government Action and the Superiority Requirement: A Potential Bar to Private Class Action Lawsuits, 18 Geo. J. Legal Ethics 1419, 1423 & n.25 (2005) (noting courts have denied class certifica­tion due to prior agency or state-attorney-general recovery in antitrust, employment law, tax, and unfair trade practice contexts). However, especially in light of the court’s equitable discretion to limit recovery, the court should not deny class certification without carefully considering the nature of the prior agency settlement and whether it truly made the victims whole. 339 Cf. D. Theodore Rave, Settlement, ADR, and Class Action Superiority, 5 J. Tort L. 91, 114 (2012).

3. Agency Tunnel Vision. — Another potential worry, especially with agency regulation strategies (as opposed to substitution), is that agencies will systematically tend to overregulate arbitrationthat is, they will re­strict arbitration or place conditions on its use when it is not socially beneficial to do so. One reason for such a fear is that agencies will overweight their own statutory missions“protect consumers,” for ex­ampleand tend to underweight the values associated with arbitration itself, such as speedier and less expensive dispute resolution. 340 This fear has been raised in the CFPB context. See Johnston & Zywicki, supra note 123, at 6; see also Stephenson, Public Regulation, supra note 12, at 130 (articulating a similar con­cern in the context of agency regulation of private causes of action that “agencies tend to be overzealous regulators that focus narrowly on their own mission without consideration of costs or competing goals”). Such a concern may underlie some of the judicial hostility to agency claims of authority to regulate arbitration, with the courts perhaps fearing that agencies will ignore or downplay the benefits the Supreme Court has housed within the FAA’s federal policy in favor of arbitration.

To the extent we are worried about agencies’ tendency toward over­regulation, 341 It is not clear whether agencies in fact exhibit tendencies toward overzealousness or expansion of their own powers without limit. See David B. Spence & Frank Cross, A Public Choice Case for the Administrative State, 89 Geo. L.J. 97, 119 (2000) (arguing the agency-overzealousness critique “overstates the magnitude and significance of the drift problem, ignores other agency incentives, and ignores the enhanced ability of politicians to address the problem”); see also Daryl J. Levinson, Empire-Building Government in Constitutional Law, 118 Harv. L. Rev. 915, 933–34 (2005) (examining “empire building” arguments with respect to the bureaucracy). Moreover, Congress retains the power to check agency bias ex ante—through limitations on the terms of the delegation—or through ex post oversight mechanisms. See Spence & Cross, supra, at 119–21 (noting that in addition to limitations on the terms of the delegation, various structures including “the Office of Management and Budget’s review of proposed agency regulations” exist to cur­tail regulatory creep). Thus, the problem may be less severe than is sometimes thought. existing administrative law mechanisms can help to miti­gate the issue. For one, agencies should be attentive, as always, to the costs as well as the benefits of regulatory intervention. Executive branch agencies regulating arbitration will be subject to cost-benefit review by the Office of Information and Regulatory Affairs (OIRA), at least for ma­jor regulations. 342 See generally Nicholas Bagley & Richard L. Revesz, Centralized Oversight of the Regulatory State, 106 Colum. L. Rev. 1260 (2006) (explaining OIRA’s function within the Office of Management and Budget and discussing OIRA review); Spence & Cross, supra note 341, at 121 (pointing to OIRA review as a check on agency bias toward regulation). And that review will, presumably, be appropriately attentive to the costs of regulating arbitration. These costs may include the efficiency benefits sometimes associated with arbitration as well as any costs associated with requiring the availability of collective actions, such as the risk of settlement of unmeritorious claims. 343 The efficiency benefits associated with arbitration are subject to dispute. One illuminating study showed that companies that use arbitration in their contracts with con­sumers rarely do so in their negotiated, business-to-business agreements with peers, casting some doubt on claims that arbitration is necessarily more efficient than litigation. See Theodore Eisenberg et al., Arbitration’s Summer Soldiers: An Empirical Study of Arbitration Clauses in Consumer and Nonconsumer Contracts, 41 U. Mich. J.L. Reform 871, 876 (2008). In addition, the gravity of the harm posed by agency overregulation may plausibly be less than that posed by erring in the other direction. When an agency over­regulates such that a claim proceeds in court and not arbitration, the harm is the sum of efficiency benefits associated with arbitration and not litigation of that claim. When an agency underregulates, the claim may as a practical matter not proceed in any forum at all, and the entire “value” of the claim is lost. Furthermore, most agency proposals would not prevent parties from capturing the efficiency benefits of arbitration by agreeing to arbitrate after a dispute arises—that is, they would regulate only predispute mandatory arbitration. I am grateful to Professor Christopher Leslie for raising these points in conversation. And even when OIRA review is not required, such as for independent agencies or non-“major” rules, 344 See Eloise Pasachoff, The President’s Budget as a Source of Agency Policy Control, 125 Yale L.J. 2182, 2203–04 (2016) (“[I]ndependent agencies need not submit their regulations to OIRA for review . . . .”); Samuel J. Rascoff, Presidential Intelligence, 129 Harv. L. Rev. 633, 710 n.414 (2016) (explaining that only “major” rules are subject to full OIRA review and how rules are deemed “major” under the relevant executive orders). agencies regulating arbitration should as a matter of best practices be attentive to the costs of regulation.

Agencies that ignore or completely discount the costs of regulating or banning arbitration may also be vulnerable to having their rules va­cated on judicial review, though the law is somewhat in flux on this mat­ter. 345 See Cass R. Sunstein, Cost-Benefit Analysis and Arbitrariness Review 5 (Harvard Pub. Law & Legal Theory Research Paper Series, Working Paper No. 16-12, 2016), (on file with the Columbia Law Review) (arguing that in certain circumstances courts may set aside agency action for not properly considering costs); see also Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (holding the EPA’s determi­nation that cost was irrelevant to a power-plant regulation was unreasonable). But see Jacob Gersen & Adrian Vermeule, Thin Rationality Review, 114 Mich. L. Rev. 1355, 1374 (2016) (arguing there is no obligation to conduct a rigorous cost-benefit analysis under the Administrative Procedure Act). And to say that agencies should receive Chevron deference is not to give them a blank check. Agencies that stray from their statutory man­date, rely on extrastatutory considerations, or otherwise behave unrea­sonably may also be vulnerable under Chevron even if, as argued above, agencies’ reasonable interpretations should be upheld. 346 See supra section III.A.4 (arguing that agencies should have authority to reasona­bly interpret their statutes in a manner that allows them to regulate arbitration). Thus, an agency that does not ground its decision to regulate arbitration in its authorizing statute can—and should—have its decision vacated on review.

Likewise, an agency that has the discretion to but chooses not to reg­ulate arbitration should prevail on review provided that decision is within the bounds laid out by Congress and is not otherwise arbitrary and capri­cious. In this same vein, agencies and the courts should not assume that agencies’ statutory mandates, even if they can be read to regulate arbitra­tion, are necessarily at cross-purposes with the goals associated with arbi­tration. For example, even a (hypothetical) statutory mandate to “reg­ulate consumer financial contracts for the benefit of consumers” requires consideration of whether arbitration in fact serves the interests of consumers—for example, by lowering costs and making credit more available. And actions that may appear at first glance to benefit con­sumers, employers, or any other group may in fact work to the detriment of that group in the long term. 347 See, e.g., Andrew A. Schwartz, Consumer Contract Exchanges and the Problem of Adhesion, 28 Yale J. on Reg. 313, 352 (2011) (describing the law and economics view that “‘judicial meddling’ in contracts of adhesion can have the perverse effect of harming the very people it is intended to assist by, for instance, driving up the cost of consumer credit”). Agencies when promulgating reg­ulations—as well as courts when reviewing them—would do well to keep that fact in mind.


This Essay has investigated the roles that agencies might play in re­sponse to the rise of mandatory bilateral arbitration. In doing so, it has made a general case for increased agency involvement in the area.

Although this Essay has taken mandatory bilateral arbitration as its jumping-off point, the weakening of private enforcement has taken other forms as well. For example, the Supreme Court has made it easier for defendants to win motions to dismiss at early stages in the proceeding; 348 See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 556 (2007) (adopting a more strin­gent “plausibility” standard for pleadings that makes it easier for defendants to win mo­tions to dismiss). made it more difficult for plaintiffs to certify certain kinds of class ac­tions; 349 See Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011) (holding that a group of over one million women did not meet the traditional certification requirements for a class action). created impediments to recognizing implied private rights of action; 350 See Armstrong v. Exceptional Child Ctr., Inc., 135 S. Ct. 1378, 1388–89 (2015) (holding private Medicaid providers did not have an implied cause of action to challenge a state’s reimbursement rates). restricted the territorial application of certain private causes of action; 351 See Morrison v. Nat’l Austl. Bank Ltd., 561 U.S. 247, 255 (2010) (holding con­gressional legislation, unless expressly stated otherwise, applies only within the territo­rial jurisdiction of the United States). and repeatedly threatened to restrict standing of those seeking to enforce categories of congressionally created rights. 352 See Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1549 (2016) (holding Article III standing requirements dictate that a plaintiff cannot bring a claim that alleges only a bare procedural violation). And no matter the composition of the Supreme Court going forward, many of these changes are sure to be sticky, at least to some extent.

Although, like mandatory bilateral arbitration, all the above devel­opments have been criticized on their own terms, less attention has been paid to what they mean for the overall system of regulation in the United States. Moreover, existing calls for reform have tended to accept existing institutional relationships that generate the content of legal rules, in­stead criticizing the content of those rules themselves.

This Essay has attempted to address some of these blind spots in the area of arbitration. As others have observed, the American regulatory system—unlike, for example, those prevailing in Europe—has historically relied on ex post private enforcement as an important tool in the en­forcement of public law. 353 See, e.g., Samuel Issacharoff, Regulating After the Fact, 56 DePaul L. Rev. 375, 377 (2007) (“What really sets the United States apart is the fact that its basic regulatory model is ex post rather than ex ante, a form of regulation that draws heavily on its common-law tradition.”); see also Elhauge, How Italian Colors Guts, supra note 2, at 771–72 (making this point in the context of antitrust specifically). Thus, changes to the availability of private enforcement may have important effects on the overall regulatory ecosys­tem, one which is inhabited by public regulatory bodies as well as private enforcers.

This insight suggests the need for greater sensitivity on the part of administrative law scholars as well as public regulatory bodies themselves when it comes to changes occurring in the realm of private enforcement. And though the natural inclination might be simply to call for greater public enforcement when private enforcement wanes, this Essay has hopefully shown that other strategies, including those that enforce or enable greater private enforcement without necessarily substituting for it, may also serve valuable roles.