Litigation and selection with correlated two-sided incomplete information

This article explores the selection of disputes for litigation in a setting with twosided incomplete information and correlated signals. The models analyzed here suggest that Priest and Klein's conclusion that close cases are more likely to go to trial than extreme cases remains largely valid when their model is interpreted as involving correlated, two-sided incomplete information and is updated (i) to incorporate take-it-or-leave-it offers or the Chatterjee-Samuelson mechanism, (ii) to take into account the credibility of the plaintiff's threat to go to trial, and (iii) to allow parties to make sophisticated, Bayesian inferences based on knowledge of the distribution of disputes. On the other hand, Priest and Klein's prediction that the plaintiff will win 50% of litigated cases is sensitive to bargaining and parameter assumptions.