The Litigation Clock: Competing-Risks Survival Analysis of U.S. Federal Securities Class Actions

Abstract

Federal securities class actions almost always end in settlement or dismissal, yet existing empirical work studies these outcomes in isolation. This thesis asks how the Private Securities Litigation Reform Act of 1995 (PSLRA), judicial geography, and observable case characteristics jointly shape the timing and type of resolution in federal securities class actions. We apply, to our knowledge, the first full competing-risks survival framework to federal securities litigation, analyzing 12,968 cases from the Federal Judicial Center's Integrated Database (1990--2024) using cumulative incidence functions, cause-specific Cox and Fine-Gray subdistribution models, inverse probability of treatment weighting (IPTW), and shared frailty models. Three findings emerge. First, the PSLRA is consistently associated with elevated early dismissal: a piecewise model reveals a 79% higher first-year dismissal hazard (HR~=1.789) that decays to null beyond two years, consistent with an early pleading-stage filter. This association survives composition adjustment (IPTW HR~=1.519), though temporal confounding cannot be ruled out. The settlement-side association is directionally consistent but materially fragile. Second, circuit geography is the dominant structural predictor of outcomes, with settlement hazards ranging from 1.1 to 4.3 times the Second Circuit baseline. Third, MDL consolidation produces a sign reversal visible only through competing-risks analysis---a lower instantaneous settlement hazard coexists with higher cumulative settlement probability---demonstrating that single-outcome frameworks mischaracterize resolution dynamics.