Epic Systems v. Lewis: The Return of Freedom of Contract in Work Law?
Co-Associate Dean for Research & Faculty Development and Associate Professor, Seattle University School of Law
Of the U.S. Supreme Court’s most high-profile cases in the October Term 2017, Epic Systems Corp. v. Lewis was also among the most predictable: much like with the Court’s other major labor case, Janus v. AFSCME, court watchers could reliably anticipate the outcome as soon as it was apparent that a nine-member Court that included Justice Neil Gorsuch was going to hear the case. The Court did not surprise, splitting 5-4 to decide that pre-dispute individual arbitration clauses are enforceable, notwithstanding that the National Labor Relations Act (NLRA) protects employees’ “concerted activity,” and the Norris-LaGuardia Act (NLGA) renders unenforceable so-called “yellow dog contracts,” in which employees agree not to join a labor union as a condition of employment..
Epic Systems will yield a range of consequences for workers—all of them negative. Justice Gorsuch’s majority opinion reads narrowly the NLRA’s protection for “other concerted activities,” planting the seeds for further retrenchment in the National Labor Relations Board (NLRB), which is now controlled by Trump appointees, and in the courts. In addition, Justice Gorsuch’s reliance on the rhetoric of “freedom of contract” signals that his approach to labor and employment law is likely to ignore the reality of workplace power dynamics, just as progressive court-watchers feared (and corporate lobbyists hoped) during his confirmation process. And most concretely, Justice Gorsuch’s opinion in Epic Systems will make it harder for employees to enforce their rights. This also means the decision will give a leg up to unscrupulous employers who either deliberately violate employment law or are indifferent to compliance, as compared to law-abiding employers.
This Essay has three parts. First, it describes how the NLRB’s rule on individual arbitration clauses came about, and analyzes how the Epic Systems majority opinion relied on an ahistorical, decontextualized understanding of both the Federal Arbitration Act (FAA) and labor law to reverse the Board’s rule. Second, it discusses Epic System’s likely consequences for workers who are compelled to bring their claims in individual arbitration or not at all. Finally, it discusses the threats to workers posed by the Epic Systems majority’s reading of the NLRA and reliance on “freedom of contract” rhetoric.
I. Individual Arbitration & Labor Law
To understand the ahistorical nature of the Epic Systems decision, it helps to trace key legal developments beginning in the early twentieth century. The purpose of this section is twofold: first, it sketches the legal background that led up to Epic Systems; and second, in the course of doing so, it offers an object lesson in the Court’s repeated interference with workers’ collective action.
A. Regulating Work in the New Deal
In the late 1800s and early 1900s, the American labor movement was gaining members and power, and strikes and boycotts aimed at improving working conditions were on the rise. In addition to “primary” strikes involving only the employer and employees locked in a labor dispute, unions increasingly relied on “secondary” or “sympathy” strikes, which greatly increased strikers’ leverage by expanding the scope of labor disputes to include the employees of “neutral” employers. But employers found a way to derail these efforts: some federal courts were willing to enjoin these strikes under a range of legal theories, including that they violated the Sherman Act. Congress attempted to end this practice in 1914, declaring in the Clayton Act that “[t]he labor of a human being is not a commodity or article of commerce,” and sharply limiting the authority of federal courts to issue injunctions in response to labor disputes.
The Supreme Court limited the Clayton Act’s effect by interpreting it narrowly in a pair of cases decided in 1921. In Duplex Printing Press Co. v. Deering, the Court declared that “Congress had in mind particular industrial controversies, not a general class war,” and held that federal courts could still enjoin secondary strikes and boycotts. And in American Steel Foundries v. Tri-City Central Trades Council, the Court approved an injunction against striking workers who were picketing in groups of four to twelve, limiting strikers to one picketer at each entrance to the employer’s plant. The effect of these cases was to increase the frequency with which federal courts issued (often ex parte) injunctions against labor unions and members engaged in secondary strikes, boycotts, or picketing—a practice now referred to “government by injunction.”
Congress’s second effort to end government by injunction was more successful. In 1932, Congress passed the NLGA, which stripped federal courts of jurisdiction “to issue any restraining order or temporary or permanent injunction in a case involving or growing out of a labor dispute.” But that wasn’t all: the statute also declared “yellow dog contracts”—employment contracts that required employees to promise not to join a union on pain of job loss—to be contrary to public policy and unenforceable.
This time, government by injunction receded. Perhaps even more surprisingly, the Court did not review the constitutionality of the anti-yellow dog provision of the NLGA. This was significant because the Court had struck down legislative attempts to make it a crime for employers to impose yellow dog contracts on their employees in two previous Lochner-era cases.
And finally, there was the NLRA itself, which was enacted in 1935 and upheld by the Supreme Court in 1937. The NLRA as originally enacted declared a preference for workers’ collective action:
[t]he inequality of bargaining power between employees who do not possess full freedom of association or actual liberty of contract, and employers who are organized in the corporate or other forms of ownership association substantially burdens and affects the flow of commerce . . . and the purchasing power of wage earners in industry.
In other words, the NLRA was expressly premised on an understanding of workplace power dynamics: because workers were dependent on employers for their livelihoods, they often lacked the leverage to negotiate meaningfully with their employers, and the employment “agreements” they reached reflected the employer’s will rather than any sort of two-sided give-and-take. This insight would be obvious to anyone who has ever held a low-wage job—but it was also at odds with the Supreme Court’s Lochner-era caselaw, which was committed to preserving employers’ authority to set work rules without interference from the state under the guise of freedom of contract between workers and employers.
Section 7 of the NLRA, which contains the core substantive protection for workers’ rights to engage in collective action was drafted broadly: “Employees shall have the right to self-organization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection.” That single sentence, with its list of protected activities, forms the backbone of workers’ labor rights. Moreover, while the Supreme Court has notoriously restricted workers’ NLRA rights where, for example, it perceived a conflict with employers’ property rights, it has also emphasized the capacious nature of the Act. For example, it has long been beyond doubt that the NLRA protects employees’ collective action whether or not they are unionized (or are looking to unionize), even when they will not personally benefit from their collective action, and when they are enforcing other statutory rights, such as the right to be free from workplace sexual harassment.
B. The Federal Arbitration Act and Employment
In 1925, during the Lochner era and amidst turmoil regarding workers’ collective action and the regulation of work, Congress enacted the Federal Arbitration Act (FAA). The statute was aimed at providing for “enforceability of arbitration agreements between merchants—parties presumed to be of approximately equal bargaining strength—who needed a way to resolve their disputes expeditiously and inexpensively.” Accordingly, it contains an exemption: “nothing [in the Act] shall apply to contracts of employment of seamen, railroad employees, or any other class of workers engaged in foreign or interstate commerce.” That language excluded from the FAA’s purview the employment relationships that Congress was understood to be able to regulate in 1925, in light of the Court’s narrow view of the Commerce Clause.
For decades, arbitration was relatively rare in the employment context, in part because of doubts about whether arbitration clauses in employment contracts were enforceable under the FAA. Those doubts were ultimately resolved in favor of arbitration in a series of cases decided beginning in 1991. First, in Gilmer v. Interstate/Johnson Lane Corp., the Court enforced an arbitration clause between an employee and employer that was imposed as a condition of registering as a securities representative with the New York Stock Exchange. The Gilmer Court wrote that “[m]ere inequality in bargaining power . . . is not a sufficient reason to hold that arbitration agreements are never enforceable in the employment context,” but left open the possibility of defenses such as fraud or coercion, and signaled that “claim[s] of unequal bargaining power” could be resolved “in specific cases.” However, because the arbitration clause in Gilmer was imposed by the NYSE rather than the employer, the Court chose not to address the scope of the FAA’s exemption of “workers engaged in foreign or interstate commerce.”
Then, in the 2001 case Circuit City Stores, Inc. v. Adams, the Court rejected the argument that Congress’s decision to exclude from the FAA all of the employment contracts over which it had jurisdiction in 1925 manifested an intent to exclude employment contracts from enforcement under the FAA more generally. The Court did not disagree about Congress’s likely intent, but it instead prioritized a plain-text approach to interpreting the FAA. Thus, since 2001, the Court’s view has been that the FAA’s exemption for employment contracts covers only workers engaged in interstate transportation. Finally, following Gilmer and Circuit City, the Court significantly limited states’ abilities to place limits on the use of arbitration clauses, and held that individual arbitration clauses were enforceable even when the costs of prosecuting a case on an individual basis exceed the possible recovery.
The result of these and other decisions has been steadily increasing numbers of employers that demand that employees agree to arbitrate their workplace disputes. Thus, one recent study found that whereas only about two percent of employees were covered by individual arbitration clauses in 1992 (the year after Gilmer), now more than half of employees are. Moreover, these clauses often require workers to resolve their disputes on an individual basis, placing many low-dollar-value claims out of reach as a practical matter.
C. The D.R. Horton/Murphy Oil Rule at the Board and in the Courts
Against this backdrop of increasing reliance on individual arbitration by employers, it is hard to understate the importance of the NLRB’s D.R. Horton decision. The decision held that employers could not require their employees to commit to resolve workplace disputes on an individual basis. Instead, the Board held that arbitration clauses in employment contracts were consistent with the NLRA only if they left employees with at least one forum—either judicial or arbitral—in which they could resolve disputes on an aggregated basis. In other words, employers could still require employees to commit to resolve disputes in arbitration—but if they did that, they also had to agree to allow employees to band together to arbitrate their claims, including in class or collective actions. As a practical matter, the D.R. Horton rule likely would have limited the use of employment arbitration, because arbitration involves very limited appellate review, and employers would have chosen to litigate class or collective actions in court rather than risk large, unappealable losses.
The Board’s reasoning in D.R. Horton was straightforward. First, it pointed to the long-accepted principle that Section 7 covers collective action that takes place outside of the workplace, including through litigation. As the Board put it, “employees who join together to bring employment-related claims on a classwide or collective basis in court or before an arbitrator are exercising rights protected by Section 7 of the NLRA.” Indeed, the Supreme Court had already agreed, in Eastex, Inc. v. NLRB, that Section 7 protects “employees’ efforts to ‘improve terms and conditions of employment or otherwise improve their lot as employees through channels outside the immediate employee-employer relationship.’” In Eastex, the Court upheld the Board’s view that Section 7 protects unions’ and employees’ rights to advocate for pro-worker public policy even when they might not benefit directly from the particular policy being advanced. As far back as the 1940s, the NLRB had also held that Section 7 protects employees when they resort to court or agency enforcement processes, and courts of appeals, though not the Supreme Court, have repeatedly affirmed these decisions.
Relatedly, the Board noted that other caselaw prohibits employers from demanding that employees promise to resolve their disputes individually rather than through collective action—that is, employers cannot demand that employees individually waive their statutory rights. That is true under the NLRA, but the Board wrote that the NLGA also has a role to play; if one accepts that class or collective litigation to improve working conditions is a form of collective action, then a pre-dispute individual arbitration clause is a form of yellow-dog contract.
Next, the Board considered whether the NLRA’s protection for collective litigation was in conflict with the FAA. Concluding that the two statutes could be reconciled, the Board relied on the Supreme Court’s statements in cases including Gilmer that the FAA does not require the waiver of substantive rights; moreover, the FAA contains a “savings clause” stating that arbitration agreements are “valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” The Board reasoned that the substantive right protected by the NLRA was the right to act collectively—meaning that an order invalidating an individual arbitration clause “does not conflict with the FAA, because the waiver interferes with substantive statutory rights under the NLRA, and the intent of the FAA was to leave substantive rights undisturbed.”
Finally, the Board emphasized that it was not mandating class arbitration—to the contrary, it was holding “only that employers may not compel employees to waive their NLRA right to collectively pursue litigation of employment claims in all forums, arbitral and judicial.” Instead, employers could choose between litigating their employees’ collective claims, or arbitrating them. Thus, an employer could presumably require employers to sign, as a condition of employment, either an agreement committing employees to arbitrate their disputes, but allowing them to do so on a concerted basis; or an agreement committing employees to arbitrate their individual disputes, but permitting them to file cases seeking relief on behalf of multiple employees in court.
The Board applied the D.R. Horton rule in dozens of cases, facially invalidating employer policies requiring employees to accept individual arbitration or lose their jobs. One of these cases, Murphy Oil, re-affirmed D.R. Horton while also mooting questions about whether one of the Board members who decided D.R. Horton had been invalidly appointed to the Board without Senate consent. In addition, numerous employees who sought to sue their employers under statutes such as the Fair Labor Standards Act (FLSA) invoked the D.R. Horton rule to oppose their employers’ motions to compel (individual) arbitration.
Through both of these routes, cases testing the Board’s D.R. Horton rule soon made their ways to the courts of appeals, where they met with initial rejection and ultimately a mixed reception. In the appeal from D.R. Horton itself, the Fifth Circuit seemed to credit the NLRB’s reasoning that class or collective litigation or arbitration is protected by Section 7, but nonetheless held that the FAA required enforcement of individual arbitration agreements. In reaching that conclusion, the court first wrote that under Supreme Court precedent, “arbitration has been deemed not to deny a party any statutory right.” From there, the Fifth Circuit wrote that the “use of class action procedures . . . is not a substantive right,” citing a number of cases rejecting arguments that different statutory schemes contained rights to class action procedures. Turning to the Supreme Court’s more recent arbitration decisions, the D.R. Horton court then concluded that “the Board’s rule does not fit within the FAA’s saving clause,” apparently because a prohibition on class action waivers in arbitration would make arbitration too unappealing to employers: “Requiring a class mechanism is an actual impediment to arbitration and violates the FAA.” Finally, having set up a direct conflict between the NLRA (and the NLGA) and the FAA, the Fifth Circuit held that the NLRA’s guarantee of collective action was not specific enough to override the FAA’s policy favoring arbitration.
However, the D.R. Horton rule fared better in the Seventh and Ninth Circuits. First, in Lewis v. Epic Systems Corp., the Seventh Circuit considered the Board’s rule in the context of a putative collective action filed under the FLSA. The employer, Epic, responded to Lewis’s complaint by seeking to enforce an individual arbitration clause that it had imposed by means of an email stating that any employee who “continue[d] to work at Epic” was deemed to have agreed to waive “the right to participate in or receive money or any other relief from any class, collective, or representative proceeding,” and instead to bring wage-and-hour claims in individual arbitration. Unlike the Fifth Circuit, the Seventh Circuit agreed with the NLRB’s view of the interaction between the NLRA and the FAA, including that an individual arbitration clause was an illegal agreement under the NLRA—and therefore unenforceable according to the FAA’s savings clause. But the Seventh Circuit’s most important point of disagreement with the Fifth Circuit was its understanding of the substantive rights that the NLRA guaranteed: “While the FLSA and [the Age Discrimintation in Employment Act] allow class or collective actions, they do not guarantee collective process. The NLRA does. . . . Just because the Section 7 right is associational does not mean that it is not substantive.”
Following on the Seventh Circuit’s heels, the Ninth Circuit also affirmed the D.R. Horton rule in Morris v. Ernst & Young, LLP. Ernst & Young also involved employees who were required to sign individual arbitration clauses as a condition of employment, but who then sought to litigate an FLSA claim in court. As the Ninth Circuit saw it, the “NLRA obstacle is a ban on initiating, in any forum, concerted legal claims—not a ban on arbitration.” Thus, the court reasoned, the D.R. Horton rule was not at odds with the FAA’s pro-arbitration policy: “The problem with the contract at issue is not that it requires arbitration; it is that the contract term defeats a substantive federal right to pursue concerted work-related legal claims.” Accordingly, the Ninth Circuit agreed with the NLRB and the Seventh Circuit that Ernst &Young’s individual arbitration clause fell under the FAA savings clause because it illegally waived a substantive federal right.
In September 2016, the United States solicitor general sought certiorari after the Fifth Circuit refused to enforce the NLRB’s decision in Murphy Oil, in a decision that largely tracked the analysis from its D.R. Horton decision. (Needless to say, the solicitor general urged in the petition for certiorari that the Board’s rule was correct and should be upheld.) Likewise, the employers in Epic Systems and Ernst & Young also sought Supreme Court review. The Court consolidated the three cases and granted review in January 2017.
When it came time to file merits briefs, the (post-2016 election) solicitor general’s office did an about-face; it filed an amicus brief in support of the employers in the three cases, arguing that the NLRB should have given greater weight to the FAA’s pro-arbitration policy. This was one of a handful of cases in which the Trump administration reversed an Obama administration position—a list that notably included the other major labor case on the OT 2017 docket, Janus v. AFSCME. At the same time, the solicitor general authorized the NLRB’s general counsel, Richard Griffin Jr., to litigate on the Board’s behalf in the Supreme Court. (Griffin was appointed by President Obama to a term that ended in October 2017, shortly after argument in Epic Systems.) In other words, the federal executive branch took two opposing positions in Epic Systems.
Paul Clement argued for the employers before the Supreme Court. His main argument was that Section 7 protects “collective action by the employees in the workplace,” but not in courts or arbitrations. In response to questioning by Justice Breyer, Clement spelled out what he meant by this:
[F]rom the very beginning [of the NLRA], the most that has been protected is resort to the forum, and then, when you get there, you are subject to the rules of the forum. So, for example, if an atypical worker decides that he wants to bring a class action on behalf of a handful of fellow employees . . . the employer doesn’t commit an unfair labor practice by [arguing that the worker doesn’t satisfy the numerosity or typicality requirements of the class action rule, Fed. R. Civ. P. 23].
In other words, Clement conflated two things: the right not to be subject to a pre-dispute waiver of the right to engage in collective action and the (mostly non-existent) right to preclude an employer from countering its employees’ collective action, or even to compel some third party to accommodate workers’ collective action. But the fact that workers’ concerted activity can yield employer counter-moves is both well known to anyone with even minimal familiarity with labor law and distinct from the question of whether employees can execute advance waivers of their Section 7 rights. That is, an employer cannot ask employees to execute advance waivers of their Section 7 rights, but it is free to respond to employees’ collective action once it begins, such as by permanently replacing economic strikers, or locking employees out of the workplace in order to secure a favorable collective bargaining agreement. And while the NLRA can limit the responses available to employers (for example, employers may temporarily, but not permanently, replace workers who go on strike in response to unfair labor practices), the NLRB has never suggested that the NLRA would preclude an employer from opposing a motion for class certification. Moreover, the NLRA does not constrain entities other than employers, unions, and employees, so the suggestion that the NLRB might deem it inconsistent with the NLRA for a district court to apply Rule 23’s requirements had no basis in reality.
Arguing for the U.S. solicitor general, Jeffrey Wall made a similar argument to Clement, and also emphasized the FAA’s “clear congressional command” to enforce arbitration agreements, which Wall argued meant that the NLRB could not “interpret the NLRA’s ambiguity [as reflecting congressional intent to invalidate collective action waivers in employment arbitration agreements] . . . in the face of the FAA and federal rules like Rule 23.” Finally, both Clement and Wall chided the NLRB for the newness of the D.R. Horton rule, with Clement suggesting that if the NLRA really precluded individual arbitration clauses, then the AFL-CIO should have argued as much in amicus briefs in previous employment arbitration cases.
During Griffin’s argument for the NLRB, several justices asked questions that suggested they accepted Clement’s premise about the scope and meaning of the D.R. Horton rule. For example, Justice Alito asked whether Rule 23 abrogated Section 7; later, Chief Justice Roberts asked whether an arbitral forum rule that imposed a 50-person numerosity requirement on putative class arbitrations meant that “you have a right to act collectively, but only if there are 51 or more of you.” As a result, Griffin had to spend much of his time at the lectern explaining Labor Law 101, with occasional assists from Justices Kagan and Breyer. Likewise, Chief Justice Roberts asked Daniel Ortiz, counsel for the employees in Epic Systems and Ernst & Young, a variation on his 50-employee hypothetical. As with Griffin, the exchange seemed to yield more confusion than clarity, although Justice Sotomayor usefully observed that an employer’s intent in choosing an arbitral forum with particularly restrictive rules governing class or joint litigation was relevant to whether an arbitration agreement violated Section 7. Thus, an employer could violate the NLRA by intentionally restricting forum access to defeat collective litigation—but freestanding forum-imposed limits on collective litigation were entirely consistent with the Act.
While he asked no questions during oral argument, Justice Gorsuch wrote the majority opinion in Epic Systems. He first emphasized that the NLRA had existed for “77 years” before the NLRB adopted the Epic Systems rule, implying that the rule was therefore illegitimate—though without mentioning that individual arbitration of employment disputes was not prevalent until much more recently. Next, Justice Gorsuch reasoned that the FAA savings clause did not apply because it “recognizes only defenses that apply to ‘any’ contract,’” and “courts may not allow a contract defense to reshape traditional individualized arbitration by mandating classwide arbitration procedures without the parties’ consent.” Thus, the majority’s view was that even if individual arbitration clauses are illegal because they violate Section 7, the savings clause would not apply. Here, the majority analogized Epic Systems to Concepcion, which involved a state law that declared class action waivers to be unconscionable in either litigation or arbitration. But that analogy seems to fall short: whereas the law in Concepcion was aimed exclusively at preserving class actions, Section 7 applies equally to any employment contract that asks employees to waive their rights to engage in any activity that qualifies as protected concerted activity under the NLRA. Or, to put it another way, the Concepcion Court was dealing with a statute whose main function was to respond to class waivers in arbitration contracts. But—as the Epic Systems majority went on to argue at length—the NLRA was enacted to preserve a range of activities, of which the right to concerted litigation was at most one aspect.
Next, the majority turned to whether Section 7 encompassed the right to concerted litigation or arbitration at all, concluding that “[t]he notion that Section 7 confers a right to class or collective actions seems pretty unlikely when you recall that procedures like that were hardly known when the NLRA was adopted in 1935.” The Court majority used that argument about likely legislative intent (which did not discuss other procedures for concerted litigation, such as joinder) to frame its views on the core text of Section 7: “the right to self-organization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection.” According to the majority, the “catchall” term “other concerted activities for . . . mutual aid or protection” should be “understood to ‘embrace only objects similar in nature to those objects enumerated by the preceding specific words’” under the ejusdem generis canon. And those specific words, the Court continued, “serve to protect things employees ‘just do’ for themselves in the course of exercising their right to free association in the workplace, rather than ‘the highly regulated, courtroom-bound “activities” of class and joint litigation.’” That formulation—“things employees just do for themselves in the workplace”—came from Judge Sutton’s partial dissent in a Sixth Circuit case that affirmed the D.R. Horton/Murphy Oil rule. There are many things that might be said about this reading of the scope of protected concerted activity, but one is that it is premised on a narrow reading of the remainder of Section 7 that assumes, for example, that the rights to “self-organization” and to “assist labor organizations” can be exercised only at work—and further, that it did so without delving into NLRB or case law about the actual scope of those rights.
After reading Section 7 narrowly, the Court turned to the inaccurate version of the D.R. Horton/Murphy Oil rule on which Clement and Wall partially premised their arguments: that Section 7 guaranteed access to class procedures, rather than simply banning waivers that cover access to available concerted litigation vehicles. Having set up that straw man, Justice Gorsuch proceeded to blow it down: “[w]ithout some . . . specific guidance, it’s not at all obvious what procedures Section 7 might protect. Would opt-out class action procedures suffice? Or would opt-in procedures be necessary? . . . What standards would govern class certification?” Then, characterizing as a “slight reply” the response that the foregoing did not correctly reflect the D.R. Horton rule, Justice Gorsuch added that “if the parties really take existing class and collective action rules as they find them, they surely take them subject to the limitations inherent in those rules—including the principle that parties may (as here) contract to depart from them.” But this reasoning assumes its conclusion, and it is also inconsistent with the NLRA as it is understood in other contexts—for example, employees are subject to various restrictions on the right to strike, but that does not mean that employers may require individual employees to waive that right.
Given the majority’s reasoning thus far, it is unsurprising that it also rejected the argument that even if the NLRA’s text did not clearly encompass the right to engage in concerted litigation, the NLRB’s interpretation of Title VII was entitled to Chevron deference. Specifically, the Court wrote that the Board was not entitled to Chevron deference for three reasons: first, it sought to reconcile the NLRA with a statute it did not administer; second, the executive branch was divided on the NLRA’s meaning; and third, the Court could resolve the potential statutory ambiguity using “traditional tools of statutory construction.”
Justice Ginsburg dissented, joined by the three other liberal-leaning justices. She framed the ability of individual arbitration clauses—often imposed by adhesion contract as a condition of employment (or continued employment)—to make it practically impossible for employees to vindicate their rights as an example of the imbalance of power between workers and employers that led Congress to enact the NLRA. Thus, she identified individual arbitration clauses as simply another species of yellow-dog contract, illegal under the NLGA and the NLRA. Moreover, she wrote that the majority’s narrow reading of protected concerted activity was “conspicuously flawed” in light of the NLRA’s capacious language and NLRB and court decisions that interpret the statute’s protections broadly. After refuting this and other aspects of the majority’s opinion, Justice Ginsburg concluded that “[t]he inevitable result of today’s decision will be the underenforcement of federal and state statutes designed to advance the well-being of vulnerable workers.”
II. Consequences for Workplace Arbitration and Work Law
A recent Ninth Circuit decision illustrates Epic System’s consequences for workers, and shows why Justice Ginsburg’s warning was precient. O’Connor v. Uber involved a class of more than 240,000 individuals who had driven for Uber in California or Massachusetts. The drivers argued that they had been misclassified as independent contractors rather than employees, and that the company had therefore violated the law when it failed to pay benefits (such as mileage reimbursements and tips) required of employers under state law. These claims were unlikely to be worth more than a few thousand dollars per driver, but could have added up to massive liability for Uber.
In a series of orders, a district court had certified the large class over the objections of the company, which argued that nearly all of the drivers had accepted individual arbitration clauses when they signed up to drive for Uber. (A very small percentage of drivers had opted out of arbitration.) However, the Ninth Circuit rejected most of the district court’s reasoning in a 2016 decision, leaving the NLRB’s D.R. Horton/Murphy Oil rule as a basis to keep the class intact.
The Supreme Court’s Epic Systems decision deprived the O’Connor plaintiffs of their last serious argument in support of keeping the class together. Thus, it is unsurprising that the Ninth Circuit easily reversed the district court’s class certification orders. On remand, it is possible that the district court will recertify a class of drivers who opted out of arbitration—but such a class would include a relatively small number of drivers. For example, in one case involving California Uber drivers, the court found that only 270 drivers out of 160,000 had opted out.
What will happen to the drivers who did not opt out of Uber’s individual arbitration clause? It is possible that they will decide to pursue individual arbitration. In fact, counsel for the O’Connor class has pledged to represent every former-class-member driver who wants to proceed to arbitration, and some drivers have already done just that. However, this promise of competent legal representation in low-value individual arbitrations is unusual—here, it is likely an artifact of the investment of time and money that class counsel has already made in the case, and maybe also the prospect that Uber will see a global settlement as a better outcome than litigating thousands of individual arbitrations, which would also involve fronting substantial arbitral forum costs. (These costs include the arbitrator’s fee as well as the costs of conference room rentals and similar, and could easily exceed the value of many drivers’ individual claims.) The more typical result in cases affected by Epic Systems will be that employees simply will not pursue low-value claims that would otherwise be candidates for concerted litigation if not for individual arbitration clauses. The result will be to virtually immunize from liability many employers who operate close to the legal line, and even some of those who willfully violate the law.
III. What’s Next for Workplace Arbitration and Work Law?
Justice Gorsuch began his opinion with two rhetorical questions resting a premise that might best be described as a work of legal fiction: “Should employees and employers be allowed to agree that any disputes between them will be resolved through one-on-one arbitration? Or should employees always be permitted to bring their claims in class or collective actions, no matter what they agreed with their employers?”  The extent to which this question echoes the Lochner-era Court’s assumptions about individual workers’ supposed freedom of contract is breathtaking. It probably will not surprise many readers to hear that the current Court is skeptical of the very idea of the need for collective rights at work—and in light of the Court’s circumscribed description of the rights that Section 7 protects, there is a real risk that the Court will further cut back on the NLRA’s already-limited protections for workers’ concerted activity.
Despite these setbacks in labor rights, administrative agencies charged with enforcing work laws can still do their jobs even when employees are subject to individual arbitration clauses. This is an important route to both substantive enforcement of workers’ rights and the development of law despite the growing use of individual arbitration. But that rule was unsuccessfully challenged in an earlier case that resulted in a 6-3 decision in which the majority held—over the objections of Justices Thomas, Rehnquist, and Scalia—that agencies such as the EEOC could “obtain victim specific relief” in court for an employee who “waived his right to seek relief for himself in a judicial forum by signing an arbitration agreement.” While the Court may ultimately lack the appetite to revisit this decision, it is conceivable that Chief Justice Roberts and Justices Alito and Kavanaugh would side with the dissenters in Waffle House, and vote to restrict the authority of administrative agencies to obtain relief that would benefit an employee on whom an employer has imposed an arbitration agreement—a possibility that could dampen many employees’ willingness to file EEOC charges in the first place.
This possibility could also imperil one of the leading “blue state” responses to the rise of individual arbitration in employment: the authorization of representative actions, in which employees may step into the shoes of the state to enforce state work law on behalf of themselves and their coworkers. For example, California’s Private Attorney General Act (PAGA) authorizes statutory damages of $100 per affected employee per pay period, with double damages available against employers who are repeat offenders. Then, 75 percent of any amount collected goes to the state of California, to help fund the state’s own enforcement of work law. So far, the Supreme Court has denied certiorari in cases challenging aspects of PAGA, but of course it is impossible to predict what the Court will do in coming years.
As Uber v. O’Connor demonstrates, Epic Systems will do immediate, concrete harm to employees. Further, it likely signals further retrenchment of workplace rights at the Supreme Court. The one silver lining is that workers are rediscovering that they have collective power whether or not their concerted activity benefits from statutory protections. As this summer’s wave of teacher strikes showed, workers in a hostile legal and political environment can still move forward, together.
* Charlotte Garden is the Co-Associate Dean for Research & Faculty Development & Associate Professor at the Seattle University School of Law.
 Epic Sys. Corp. v. Lewis, 138 S. Ct. 1612 (2018).
 Janus v. Am. Fed’n of State, Cnty., & Mun. Emps., Council 31, 138 S. Ct. 2448 (2018).
 Epic Sys. Corp., 138 S. Ct. at 1612.
 15 U.S.C. § 17.
 Duplex Printing Co. v. Deering, 254 U.S. 443, 472 (1921).
 Am. Steel Foundries v. Tricity Cent. Trade Council, 257 U.S. 184, 206 (1921).
 See Melvyn Dobofsky, The State and Labor in Modern America (1994); William E. Forbath, Law and the Shaping of the American Labor Movement (1991).
 29 U.S.C. § 101 et seq.
 Coppage v. Kansas, 236 U.S. 1 (1915); Adair v. United States, 208 U.S. 161 (1908). These cases, and their parallels to Epic Systems are discussed in Part III.
 NLRB v. Jones & Laughlin Steel Corp., 301 U.S. 1 (1937).
 29 U.S.C. § 151.
 29 U.S.C. § 157.
 See, e.g., NLRB v. Fansteel Metallurgical Corp., 306 U.S. 240 (1939).
 9 U.S.C. § 1 et seq.
 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, 106 (2006).
 9 U.S.C. § 1.
 See, e.g., A.L.A. Schechter Poultry Corp. v. United States, 55 S. Ct. 837 (1935).
 Gilmer v. Interstate/Johnson Lane Corp., 500 U.S. 20 (1991).
 Id. at 33.
 Id. at 40 (Stevens, J., dissenting) (citing 9 U.S.C. § 1).
 Circuit City Stores, Inc. v. Adams, 532 U.S. 105 (2001).
 See Am. Express Co. v. Italian Colors Rest., 570 U.S. 228 (2013); AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011).
 Alexander J.S. Colvin, The Growing Use of Mandatory Arbitration, Econ. Pol’y Inst. (Sept. 27, 2017), available at https://www.epi.org/publication/the-growing-use-of-mandatory-arbitration/.
 In re D. R. Horton, Inc., 357 NLRB 2277 (2012).
 Id. at 2279.
 Eastex, Inc. V. NLRB, 437 U.S. 556 (1978).
 For example, workers who make $20/hour are protected by the NLRA when they advocate raising the minimum wage to $15/hour.
 In re D. R. Horton, Inc., 357 NLRB at 2280.
 Id. at 2281.
 9 U.S.C. § 2 (emphasis added).
 In re D. R. Horton, Inc., 357 NLRB at 2286.
 Id. at 2288.
 Murphy Oil USA, Inc., 361 NLRB 774 (2014).
 Id. at 775 n.16 (describing and rejecting arguments about validity of Member Becker’s appointment and participation in D.R. Horton, but observing that “[i]n any case, the Respondent’s arguments . . . are now moot, given our independent reexamination of D.R. Horton today”).
 D.R. Horton, Inc. v. NLRB, 737 F.3d 344, 357 (5th Cir. 2013) (citing Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 627 (1985)). In fact, this reading was itself debatable; the case to which the D.R. Horton court cited disclaimed the proposition “that all controversies implicating statutory rights are suitable for arbitration.” Mitsubishi, 473 U.S. at 627.
 D.R. Horton, Inc., 737 F.3d at 359.
 Id. at 360.
 Id. at 362.
 Lewis v. Epic Sys. Corp., 823 F.3d 1147, 1151, 1154 (7th Cir. 2016).
 Id. at 1161.
 Morris v. Ernst & Young, LLP, 834 F.3d 975 (9th Cir. 2016).
 Id. at 984.
 Id. at 985.
 Petition for Writ of Certiorari at 9-19, Epic Sys. Corp. v. Lewis, 138 S. Ct. 1612 (2018) (No. 16-307), (citing Murphy Oil USA, Inc. v. NLRB, 808 F.3d 1013 (5th Cir. 2013)).
 Janus v. Am. Fed’n of State, Cnty., & Mun. Emps., Council 31, 138 S. Ct. 2448 (2018).
 Oral Argument at 10:06, Epic Sys. Corp., 138 S.Ct. 1612 (No. 16-285), available at https://www.supremecourt.gov/oral_arguments/argument_transcripts/2017/16-285_1qm2.pdf.
 Epic Sys. Corp., 138 S.Ct. at 1619.
 Id. at 1622.
 Id. at 1623.
 Id. at 1624.
 Id. at 1635 (quoting 29 U.S.C. § 157).
 Id. at 1625 (quoting Circuit City Stores, Inc. v. Adams, 532 U.S. 105, 115 (2001)).
 Id. (quoting NLRB v. Alternative Entertainment, Inc., 858 F. 3d 393, 415 (6th Cir. 2017) (Sutton, J., concurring in part and dissenting in part)).
 Id. at 1625.
 Id. at 1626.
 Id. at 1630 (internal quotation marks omitted).
 Id. at 1639 (Ginsburg, J., dissenting).
 Id. at 1646.
 I have previously written about the O’Connor v. Uber case and individual arbitration agreements in the gig economy more generally. See Charlotte Garden, Disrupting Work Law: Arbitration in the Gig Economy, 2017 U. Chi. Legal F. 205 (2018). This and the next sections of this article draw on that piece, which was published before the Ninth Circuit’s recent decision decertifying the O’Connor class based on Epic Systems.
 O’Connor v. Uber, No. 14-16078, 2018 WL 4568553, (9th Cir. Sept. 25, 2018).
 Mohamed v. Uber Techs, Inc., 836 F.3d 1102 (9th Cir. 2016).
 Gillette v. Uber Techs., No. C-14-5241 EMC, 2015 WL 4481706, at *4 (N.D. Cal. July 22, 2015).
 Epic Sys. Corp., 138 S.Ct. at 1619.
 EEOC v. Waffle House, Inc., 534 U.S. 279, 305 (2002).
 Cal. Lab. Code § 2698 et seq.