25 Pages Posted: 29 Feb 2004
Date Written: February 27, 2004
While selection effects have important implications for empirical studies of the litigation process, existing theories of case selection are incomplete. Existing theories focus on "ex post selection" - selection resulting from choices made by parties after a dispute arises, such as by settlement or jury trial demands. But parties also engage in "ex ante selection" - selection resulting from choices made before a dispute arises. A common form of ex ante selection occurs when parties include a pre-dispute arbitration clause in their contract, agreeing to have future disputes resolved in arbitration rather than in court. This paper develops a theory of ex ante selection of disputes for litigation, and examines implications of the theory for empirical studies of litigation. Studies comparing outcomes in arbitration and litigation provide evidence that ex ante selection occurs. This paper argues that the effects of ex ante selection are not limited to such studies, but also may affect studies that examine only the litigation process. The central intuition is that the disputes for which litigation is most likely to be problematic (and thus of interest to researchers) are the very disputes most likely to end up in arbitration. When parties expect litigation to be costly or damage awards excessive, they have an incentive to provide for arbitration to resolve future disputes. Such ex ante selection may mask characteristics of the litigation process that empirical studies are seeking to examine.
Keywords: Arbitration, selection bias, empirical studies
JEL Classification: C10, K12, K41
Suggested Citation: Suggested Citation