Individual vs. group decision-making: Evidence from a natural experiment in arbitration proceedings

The importance of understanding the systematic differences between group and individual decisions has been well recognized in the literature. However, the vast majority of empirical evidence on this issue is derived from laboratory experiments, and hence does not reflect professional incentives and career concerns, both of which may play a crucial role. To fill this gap, I exploit a unique regulatory change that exogenously decreased the number of presiding arbitrators from three to one for a specific class of cases in the Financial Industry Regulatory Authority arbitration and an original data set of arbitration awards. The findings indicate that panels of three arbitrators render more extreme “all or nothing” awards compared to sole arbitrators. An arbitrator fixed effects model confirms that this tendency is also present within arbitrators, thus ruling out that the effect is driven by differential selection of arbitrators into panels. Rather, evidence suggests that groups provide individual arbitrators with a “shield of anonymity” which mitigates their concerns about adverse reputational effects of extreme decisions.