The ‘What’, ‘How’, and ‘Why’ of Earnings Announcement Disclosures and Formatting

Management enjoys tremendous discretion in what information they discuss in earnings announcements (EAs) and how they present this information. We develop an extensive dataset cataloging the nature of discussion of a wide range of financial metrics in EAs from 2004-2019. We begin by identifying frequently discussed financial metrics in EAs as: earnings, revenue, earnings before interest and taxes (EBIT), earnings before interest, taxes, depreciation, and amortization (EBITDA), operating cash flows, gross profit, gross margin, and non-GAAP earnings. We then use a Python script to identify the attributes of metric mentions which include the metric’s prominence, frequency of mentions, and textual highlighting (i.e., is the mention bulleted, bolded, centered), as well as the fog index, use of passive voice, and negation in sentences discussing the metric. We perform descriptive analyses of these attributes and explore how metric favorability, persistence, value-relevance, and past disclosure choices affect these attributes. We find that each of these factors predictably influences disclosure choices with metric favorability and past disclosure choices most strongly predicting disclosure behavior. Our evidence suggests that our measures capture substantive variation in metric mention attributes and that managers strategically use a variety of disclosure tactics across a range of financial metrics. We make our data publicly available to researchers and provide guidance into how this data may be used and adapted for future research