Text Analysis of Earnings Press Releases

Download or read book Text Analysis of Earnings Press Releases written by Xuan Huang and published by . This book was released on 2011 with total page 108 pages. Available in PDF, EPUB and Kindle. Book excerpt: This dissertation includes 3 chapters, which are about text analysis of earnings press releases. Chapter 1 investigates whether managers exercise discretion over how they present in the headline of earnings press releases. I find that on average headline-salient firms have higher earnings or profitable earnings. Further analysis shows that announcement with salient headlines in the earnings press releases are associated with a stronger market reaction on the announcement day and less subsequent drift. These results suggest that headlines can be a tool for perception management. Chapter 2 and 3 investigate whether managers employ tone of earnings press releases for the purpose of perception management. Chapter 2 finds that managers of firms with high abnormal accruals and firms just meeting or beating earnings targets tend to use more positive words in earnings press releases to hype the discretionary accounting numbers. This evidence implies that managers strategically use tone as a complement to earnings management to manage investor perceptions. Chapter 3 explores whether managers use tone in earnings press releases to enhance quantitative financial reporting or to mislead investors. We estimate abnormal positive tone and finds abnormal positive tone is associated with lower future earnings and cash flows from operations, a higher incidence of meeting/beating three types of earnings thresholds, and a higher incidence of future earnings restatements. Tone is also abnormally positive before firms undertake major corporate transactions such as SEO or M & A. Earnings-announcement-date returns increase with abnormal positive tone, controlling for discretionary accruals. Furthermore, instead of PEAD, after earnings announcements with abnormal positive tone there is post-event return reversal.