The Information in Management’s Expected Earnings Report Date: A Day Late, a Penny Short

Abstract

Since 1995, managers of thousands of firms have voluntarily disclosed the expected date of their firm’s next quarterly earnings announcement to Thomson Financial Services Inc. These disclosures are approximately 500% more accurate than the simple time–series expected report dates used in prior accounting research. These disclosures are also informative. On average, managers who miss their own expected date eventually report earnings that fall about one penny per share below consensus forecasts for each day of delay. Investors respond by sending the price of late–announcing stocks down at the missed expected report date and continue to send them down as the reporting delay lengthens, consistent with our “day late, penny short” result. Despite this, we find that the market response at the time earnings are announced still depends on whether the announcement is early, on time, or late relative to the firm’s own expected report date.

Analyst estimation revision clusters and corporate events, Part I

“…2 We match these clusters to corporate events collected by First Call (a division of Thomson Financial) in their Historical, Company Issued Guidance, and Events databases; CCBN in their Street Events database; and Multex (now a part of Reuters Financial Products and Services) in their Significant Developments file.…”

The Association between Earnings News and The Timing of Annual Earnings Announcements

“…Literature generally supports that good earnings news is released earlier than bad earnings news (Bagnoli et al, 2002;Begley & Fischer, 1998;Haw et al, 2003). Good (bad) news in our study refers to cases where announced earnings exceed (miss) most recent analysts' forecasts.…”

“…As mentioned, B&F compare market returns on the announcement dates between early and late announcements without considering magnitude and sign of unexpected earnings. Bagnoli et al (2002) compare ERCs between early and late announcers, but their definitions of early and late announcers are based on unexpected delays, not on actual delay.…”

“…If this is the case, we expect that worse news is reported later than bad news, and better news is announced earlier than good news. Using management's expected earnings report dates as benchmarks, Bagnoli et al (2002) find that reporting late means bad news, but reporting later means worse news. However, the relationship in good news is not empirically supported.…”

“…Researchers (e.g., B&F) derive the expected announcement date from a time series model, that is, the same period in the previous year. On the other hand, Bagnoli et al (2002) use expected earnings report dates voluntarily disclosed by managers as a benchmark. Another way to determine the expected announcement dates is to use predicted values from a cross-sectional regression of actual announcement delay on determinants.…”