An Analysis of the Wall Street Journal's Coverage of Corporate News and the Research Design of Event Studies

The typical event study of corporate news restricts its sample of announcements to events reported in the Wall Street Journal (WSJ) and listed in the WSJ Index. In this paper we examine whether such samples are representative of the events omitted from the WSJ. Focusing on convertible calls in the 1980s, we analyze possible differences between a sample based only on the WSJ Index and a larger but similarly constructed sample of calls reported in the Dow Jones News Service but not in the WSJ Index. Numerous parametric and nonparametric tests reveal no significant cross‐sample divergences in key attributes of the calling firms and called bonds, or in the distributions of risk‐adjusted returns around announcements. This evidence suggests that WSJ‐based samples for other corporate news items can be representative of cases not covered by the WSJ. Further, the evidence indicates that the results of event studies using WSJ‐based samples may support inferences about events not reported in the newspaper.