Identification Using Russell 1000/2000 Index Assignments: A Discussion of Methodologies

Abstract

This paper discusses empirical methods that rely on Russell 1000/2000 index assignments for identification. Using simulated data, the paper illustrates why the varying approaches reach conflicting conclusions about the effect of index assignment on a firm’s ownership structure and corporate policies. Some estimators likely suffer from bias (e.g., those that employ a sharp regression discontinuity estimation); others do not (e.g., those that either use a fuzzy regression discontinuity or an instrumental variable estimation). The paper also discusses changes in Russell’s index assignment methodology that began in 2007 and why these changes require modifications to the existing methodologies.

Keywords: instrumental estimation, regression discontinuity, Russell indexes

JEL Classification: D22, G23, G30, G34, G35