On Computing Probabilities of Dismissal of 10b-5 Securities Class-Action Cases

Abstract

The main goal of this paper is to propose a probability model for computing probabilities of dismissal of 10b-5 securities class-action cases filed in United States Federal district courts. By dismissal, we mean dismissal with prejudice in response to the motion to dismiss filed by the defendants, and not eventual dismissal after the discovery process. The proposed probability model is a hybrid of two widely-used methods: logistic regression, and naïve Bayes. Using a dataset of 925 10b-5 securities class-action cases filed between 2002 and 2010, we show that the proposed hybrid model has the potential of computing better probabilities than either LR or NB models. By better, we mean lower root mean square errors of probabilities of dismissal. The proposed hybrid model uses the following features: allegations of generally accepted accounting principles violations, allegations of lack of internal control, bankruptcy filing during the class period, allegations of Section 11 violations of Securities Act of 1933, and short-term drop in stock price. Our model is useful for those insurance companies which underwrite Directors and Officers liability policy.