Testing for Change in Procedural Standards, with Application to Bell Atlantic v. Twombly

Quantifying change in legal standards-in the sense of change in the propensity of judges to decide cases a certain way-presents a vexing problem. In response to a change in the behavior of courts, plaintiffs and defendants will adjust their patterns of filing and settling cases. Models of the selection of disputes for litigation predict that when legal standards change, the rate at which plaintiffs prevail in litigation will not predictably change; if so, changes in legal standards cannot be measured with data on court outcomes. I consider both the Priest and Klein divergent-expectations model and the Bebchuk asymmetric-information model to develop a methodology for measuring changes in procedural standards in the presence of selection effects. I apply this methodology to Bell Atlantic Corp. v. Twombly, whose effects have been subject to speculation and debate. I find that Twombly precipitated no significant change in dismissal rates, even after accounting for selection effects.