Measuring the informativeness of earnings announcements: The role of event windows

This paper examines the informativeness of earnings announcements by focusing on the role of event windows within which the information content is measured. This focus on event windows is primarily motivated by recent trends towards the announcement of earnings during after-market hours. It is also motivated by a lack of any systematic and rigorous investigation into the length of the event window in which the informativeness of earnings is maximal. Using the Ball and Shivakumar's (2008) abnormal R-Sq, Beaver et al. (2020) USTAT and other metrics to measure the informativeness of earnings announcements, we document that the narrower one-day event windows exhibit greater information content. However, unlike Beaver et al. (2020), we show that asymmetric windows not centered around the announcement period are more informative. Specifically, we find that for the aggregate sample and for after-hours announcements, the window of the after-hours announcement period into the following trading day has the maximum information content compared to that of the traditional three-day window. However, for daytime announcements, the window of maximum information content includes only the announcement trading day. This novel insight, that the event window which maximizes the informativeness of earnings is a function of the timing of earnings' announcements (during non-trading versus trading hours), even after using calibrated announcements, is a key takeaway of this paper. (c) 2021 Board of Trustees of the University of Illinois. Published by Elsevier Inc. All rights reserved.