An Analysis of Optimistic and Pessimistic Language in Earnings Press Releases

Seminar paper from the year 2011 in the subject Computer Science - Commercial Information Technology, University of Freiburg (Chair of Information Systems Research), language: English, abstract: Earnings press releases are the major news event of the season for companies and investors, analysts, financial media and the market. As framework of investor relations (IR) they communicate the financial performance in numerical and narrative forms. For example, earnings press releases are obligatory for firms listed on the New York Stock Exchange (NYSE). There are several rules and guidelines how to prepare them. An accurate earnings press release contains, apart from analyses of operating results, historical data, positive and negative factors affecting key financial indicators, a realistic and truthful forecast of future quarters. Whereas numerous studies focus on interpretation of numerical forms in earnings press releases, this paper examines the influence of optimistic and pessimistic language in earnings press releases on future firm performance with several studies. It also opposes different approaches to measure the tone. Based on the study "Beyond the Numbers: An Analysis of Optimistic and Pessimistic Language in Earnings Press Releases" published by Davis, Piger and Sedor the paper presents a textual analysis approach with DICTION 5.0. The authors have been the first scientists so far to examine the role language plays in the credible communication of information to investors. The dictionary-based content analysis program DICTION 5.0 is able to identify subtle aspects of language. The systematic textual analysis techniques are based on pre-existing search rules. It is able to analyze a larger sample size than possible by human coding or manual reading. Apart from this, statistical methods - like the naïve Bayesian learning algorithm, reducing a given sentence to a list of words - are introduced and compared with each other. Given the different approaches to analy