Administration
Permanent URI for this collection
Browse
Browsing Administration by Subject "10-Ks"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Open Access The Incremental Information Content of Analysts' Research Reports and Firms' Annual Reports: Evidence from Textual Analysis(2019-11-22) Park, June Woo; Tsang, AlbertThis dissertation consists of three essays, investigating the properties of analysts research reports and firms annual reports, and their impact on capital markets using textual analysis methods. The first essay studies the validity of analyst report length, measured by page count, as a proxy for analysts research effort. Specifically, I find that longer reports are positively associated with recommendation upgrades more than downgrades, and with forecast accuracy. I further document an asymmetric market reaction to longer upgrades as compared to the same length downgrades. The findings support my hypothesis that by providing more and accurate information, analysts exert credibility-enhancing effort on their upgrades, as these are perceived by investors to be more optimistic and less credible than downgrades. The study suggests differing interpretations of analyst vs. annual report length as a proxy. In a second textual analysis essay, I examine the determinants of environmental disclosures (ED) in U.S. 10-Ks (i.e. annual reports) and its impact on a future stock price crash risk. I provide crucial evidence that ED is related to bad news (i.e. news that tends to be obfuscated by managers) by showing the autocorrelation of its change over time and its negative association with short-term market reaction. In the long run, however, an increase in ED shows a lower likelihood of significant stock price drops. The results are consistent with the notion that firms benefit from non-financial information disclosure. A third textual analysis essay compares the value of private versus public information sources in U.S. analysts earnings forecasts. Using a pattern search algorithm (i.e., regular expression) on the headlines of earnings forecasts, I find that additional private sources of information are associated with less forecast error, triggering greater market reaction. Moreover, I document that the combination of management and non-management private information sources minimizes forecast error and maximizes market reaction. Finally, I show that more accurate and informative forecasts are made by analysts who make greater efforts to access private information sources, even when they do not have other information advantages (e.g. brokerage firm reputation). Thus, I provide new insight into the determinants of forecast properties.