Last week, in a set of three consolidated cases captioned Epic Systems Corp. v. Lewis the Supreme Court held (5-4) that employment contracts requiring employees to arbitrate employment disputes individually, rather than through class actions, are enforceable under the Federal Arbitration Act (FAA). The Court divided along the usual ideological line; and the decision and opinion, written by Justice Neil Gorsuch, looks like a Trumpian victory over labor unions. In a read-from the-Bench dissent, Justice Ruth Bader Ginsburg inveighed against the Court’s return to pre-New Deal, Lochner-ean tropes and yellow-dog contracts. Forthcoming decisions this Term, including another important labor case, will reinforce the notion that the conservative majority simply has it in for the unions and for that matter trial lawyers. The Donald has the Court he campaigned for; lights out for the Resistance.
A week later and beyond the politics, I write to urge a jurisprudential point that has gotten lost in the punditocracy’s ample commentary. At one level, the majority opinion is a straight-up application of the Court’s numerous FAA precedents and, in particular, the late Justice Scalia’s 5-4 opinion in AT&T Mobility v. Concepcion (2011). As shown below, though those precedents cannot make a jurisprude of the originalist-textualist persuasion—Justice Scalia, say, or Justice Gorsuch—very comfortable. As a practical matter that’s water under the bridge. But the point is still worth rehearsing, both because the decision has momentous consequences and because we ought to keep ourselves honest.
From the FAA to Epic
In the olden days English and American courts commonly refused to enforce private arbitration agreements, which they viewed as an “ouster” of their jurisdiction. In 1925, Congress decided to change that. It directed federal courts to treat arbitration agreements as “valid, irrevocable, and enforceable.” The Act contains—or rather contained—two exceptions. Section 1 exempts certain groups of workers; and the so-called “savings clause” of Section 2 permits the invalidation of arbitration agreements “upon such grounds as exist at law or in equity for the revocation of any contract.”
Almost uniformly, legal scholars agree that the FAA meant to protect commercial arbitration agreements among merchants, as distinct from consumer or employment contracts. And so it did until the mid-1980s, when the Court decided to do a number on the Act. The FAA, the Court said in a recurrent formulation, established “a liberal federal policy favoring arbitration agreements.” To effectuate that policy, the Court held that the FAA preempts state laws that stand in the way of arbitration. It held that the Act applied in state as well as federal courts. It held that the question of whether disputes are arbitrable under an agreement belongs to the arbitrator, not courts. None of these and other, similarly ambitious pro-arbitration rulings flow from the text of the FAA, so long as you read it with a straight face.
Justice Antonin Scalia, in his early years on the Court, did not appear to care much about the FAA or think much of the Court’s extra-statutory constructions. He wrote very few opinions on the subject and, on one rare such occasion, declared that the idea of a preemptive FAA was obviously wrong. Somewhere along the way, though, he changed his mind. He embraced the preemption position on stare decisis grounds. And in his last years on the Court, he wrote the majority opinions in six exceedingly important FAA cases. Jointly and severally, those decisions give the statute a truly breathtaking scope and force.
AT&T Mobility is the most telling and consequential of those decisions. The arbitration provision at issue was very consumer-friendly, but it prohibited class-action arbitration. The Court had already held that states cannot enforce laws that directly impede arbitration agreements and their enforcement. California, however, had a doctrine that barred all class action waivers under a general contract doctrine of “unconscionability,” regardless of whether agreements were to be enforced in arbitration or in court. That ought to be protected by the above-quoted savings clause of Section 2—no?
No. Class action mechanisms, Justice Scalia wrote, interfere with the “fundamental attributes” of arbitration, such as speed and relative informality. Thus, unless the parties agree on class arbitration, the state’s general anti-waiver rule is preempted by the “purpose” and the “policy” of the FAA.
To my mind the holding is defensible, as a rule that blocks state attempts to circumvent the prohibition against state laws that directly interfere with arbitration. But that’s not the theory articulated in Concepcion (Justice Scalia always loathed the form of that argument). Instead, the opinion rummages around in consequences and judicially divined legislative intents. It’s hard to believe that Justice Scalia wrote it. For example, one would think that federal statutes preempt contravening state law, while judicially manufactured “policies” and “purposes” generally don’t.
Concepcion rendered the Epic outcome very nearly a foregone conclusion. Sure, that was about consumer contracts and this is about workers, who are arguably protected under Section 1. But the Supreme Court held many years ago that that Section covers only sailors and truckers. And sure, the alleged “grounds for revocation” in Concepcion was state law. Here it’s a federal statute—specifically, Section 7 of the 1935 National Labor Relations Act, which creates a right for workers to engage in “concerted activities for the purpose of collective bargaining or other mutual aid or protection.” But that, per Epic, does not make a difference, either. Class actions cannot be the “concerted activities” Congress had in mind in 1935, because they did not exist. And if Congress means to exclude contractual arrangements from the full force of the FAA, it has to say so clearly.
To repeat: the holding follows almost ineluctably from the Court’s precedents. Why, though, did nominally conservative justices fabricate an FAA of their own imagination, from next to nothing? Maybe it’s all politics; but I don’t think so. I think it’s a work-around—around one of the most destructive decisions in our history that is nonetheless a conservative-originalist article of faith: Erie Railroad.
From Erie to Epic
From the Founding to 1938, federal courts decided “diversity” cases—that is, cases among parties from different states—in their own exercise of independent judgment, an approach that came to be known as “federal general common law.” In Erie Railroad v. Tompkins (1938), the U.S. Supreme Court overruled 100-plus years of precedent and declared that henceforth, federal courts must in diversity cases follow the law of the state where they sit. There can’t be any federal substantive rule of decision unless Congress ordains it. Hail federalism. Hail judicial deference and the separation of powers.
Erie’s central proposition is hornbook law—and obviously insane. We can’t have state courts screw around with stuff that matters and then have federal courts act as the ventriloquist branch of government. Accordingly, our Solons soon exempted from Erie’s reign the things they happened to care about—maritime affairs; interstate relations; federal institutions; labor law; eventually, civil rights and international relations. In the end the so-called Erie rule came to cover nothing but its originally intended target: the commerce of the United States. For everything else, there’s the MasterCard of federal common law. Some true-blue New Dealers—e.g., Henry Hart and Felix Frankfurter—kept complaining that all this was contrary to Erie. They never got the joke: Erie was just a polite way of saying that the unions win and the employers and producers lose. Around 1970 or so, the trial lawyers got into the game: crank up a class action under state law case in some hellhole jurisdiction and tell the federal courts—assuming they ever get to see the case—that they must give force and effect to some exotic state-law cause of action, manufactured for no other purpose than to loot out-of-state producers. Rote cite to Erie.
What’s a sensible judge or justice supposed to do? You can’t re-think or re-visit Erie. One, Erie is what you learned in law school. Two, it’s too deeply entrenched and entangled with too much case law to be revamped in the rough-and-tumble of litigation. Three, and most important, Erie stands for the “originalist” position that judges merely interpret the law; they don’t get to make things up. It’s the anti-Lochner: just as Lochner is indelibly wrong, Erie must be indelibly right, for substantially the same reason. Again, then: what’s the answer?
You borrow a page from the liberal playbook a find a work-around. Enter the FAA. From that inconspicuous acorn of a statute the Court has developed a massive body of law—a federal common law of arbitrability. It’s a way of contracting around state law and out of Erie’s formless wasteland.
And my, does that have bite. Unlike the pre-Erie federal general common law, the federal common law of the FAA preempts positive state law. Unlike ordinary contractual choice-of-law or forum selection clauses, which can’t be enforced if they violate a state’s “public policy,” arbitration clauses permit parties to contract around just about any state law—and, after Epic, federal statutes as well. (Does the FAA bar mandatory class action arbitration for “disparate impact” or sexual harassment complaints under Title VII of the Civil Rights Act? Surely not, Justice Ruth Bader Ginsburg urged in her Epic dissent. To borrow a phrase, we’ll see what happens.) Unlike the pre-Erie law, which offered parties a choice of impartial federal courts and their law, arbitration clauses effectively block access to any court, anywhere, subject only to extremely deferential and circumscribed review.
Small wonder, then, that arbitration provisions—usually barring class action arbitration—have become a staple of consumer contracts. Already, far larger numbers of workers are employed under arbitration contracts than under collective bargaining agreements. And so, over vast expanses of the U.S. economy, Erie Railroad has been obliterated. It took eight decades; but better late than never.
In her Epic dissent, the divine RBG urged Congress to overrule the decision by statute. That isn’t going to happen. Commercial arbitration lawyers used to be apprehensive about the Court’s extension of pro-arbitration rules to consumer contracts, for fear that that might bring Congress into the act and endanger even commercial arrangements. That fear has dissipated because Congress no longer gets into the act of anything. Much more likely, (blue) states will seek to evade the decision. The California courts have found clever ways around Concepcion, and the states will find ways around Epic as well. (Dan Hemel has suggested a few of them here.) For all the Supreme Court’s good work, we haven’t done what just about every other federal country on the planet has done: make the enforcement of arbitration agreements a federal question and cause of action, and be done with state courts. So state courts remain a haven. But so long as the Supreme Court majority remains serious about this business, it will prevail in the war of attrition.
There remains the jurisprudential point. There’s something to be said for a second-best legal regime that compensates for an earlier error that has become uncorrectable. But you do pay a price. The second-best rule tends to entrench the initial error. And if your second-best regime runs up against jurisprudential principles that you otherwise profess to hold sacred, and if you run that project on threadbare statutory and stare decisis foundations, then of course the enterprise looks like a partisan campaign.
The most plausible and honest answer is that the decision to which the federal common law of arbitrability is the answer, Erie Railroad, was also and in the first place a tendentious, partisan maneuver and, upon further reflection, not a cornerstone of an originalist Constitution but an egregious departure from it. And on that note, a final thought: the only parts of American public law that make any sense are arbitration law, and antitrust law. Both run up very hard against Erie’s supposed teaching; both were developed in a common-law fashion by mostly conservative judges and justices, without any help from Congress; both restore pieces of constitutional order. What does this tell us about our dogmatic insistence on Erie, and on legislative supremacy, and against anything that smacks of common law modes of argument?
It’s unlikely that the Supreme Court, or any of its justices, will articulate the predicament in those terms. But we can discuss amongst ourselves.