David Horton of California, Davis has written Do Arbitrators Follow the Law? Evidence from Clause Construction, 126 Colum. L. Rev. Forum — (forthcoming 2025). Here’s the abstract:
Courts and scholars have long disagreed about whether arbitrators follow the law. In the past three decades, the stakes in this debate have soared as the U.S. Supreme Court has expanded the impact of the Federal Arbitration Act and arbitration has practically become the default dispute resolution regime for consumer and employment cases. However, it is difficult to assess whether arbitration is lawless. For one, the process is private, usually confidential, and often generates unreasoned unwritten awards. In addition, determining whether an arbitrator decided a case “correctly” is highly subjective. Thus, the literature on point relies on crude proxies such as surveys of arbitrators, the frequency with which judges vacate awards, and arbitral citation practices. This Essay offers a different perspective. During the 2010s, it was unclear whether an arbitration provision that did not mention class actions allowed such procedures. But in 2019, the Court tried to resolve this issue by holding in Lamps Plus, Inc. v. Varela that the parties must affirmatively authorize class arbitration and that neither “silence” nor “ambiguity” suffices. Lamps Plus created a proving ground for comparing judges and arbitrators because both types of decision-makers engage in “clause construction” (determining whether an arbitration clause allows class procedures) and the American Arbitration Association requires its awards on the topic to be reasoned and published. The Essay capitalizes on this window by assembling a dataset of recent court opinions and arbitral awards. It discovers that judges generally read Lamps Plus as establishing a bright line rule that generic arbitration clauses function as class waivers, but 27% of arbitrators found a tacit agreement to allow class procedures. The Essay then explores the implications of its findings for the lawlessness hypothesis.
