Three Essays on Earnings Anouncements
King, Alexander Z
MetadataShow full item record
This dissertation includes three essays on earnings announcements. Essay one investigates the appropriate window around which the effect of earnings announcement should be measured. It finds the earnings window typically used of the day before, day of, and day after the earnings announcement contains uninformative periods. These uninformative periods downwardly bias the measured information content of earnings announcements. Further, this essay explores alternative measurement windows and identifies the smaller window of the trading day before, overnight period of, and trading day after the earnings announcement as the window with the highest measured earnings information content. These results are independent of firm size, book-to-market ratio, and industry. Essay two explores the role changes in earnings announcement timing have on intra-industry information transfer. This essay finds that, conditional on being one of the first announcers in an industry, advancement or delays in a firm’s earnings announcement increase the amount of information transferred to same-industry firms. This effect is stronger for delays in earnings announcements than for advancements. Finally, investors in subsequently announcing firms overreact to the announcements of preceding announcements, with such overreaction being corrected at the time of the subsequent announcer’s earnings release. This effect is strongest when both the preceding and subsequent announcer report earnings surprises of the same direction. Essay three studies the role investor neglect has on the market efficiency with which a firm’s stock reacts around its own earnings announcement. Using both analyst following and institutional holding as proxies for investor attention, this essay finds additional attention reduces abnormal stock return, and abnormal trading volume around a firm’s earnings announcement and post-earnings announcement drift after that announcement. These results largely persist across different neglect proxies, including principal component analysis, and after controlling for firm size.
earnings announcement timing