Forecasting Corporate Financial Performance Using Sentiment in Annual Reports for Stakeholders’ Decision-Making

Abstract

This paper is aimed at examining the role of annual reports’ sentiment in forecasting financial performance. The sentiment (tone, opinion) is assessed using several categorization schemes in order to explore various aspects of language used in the annual reports of U.S. companies. Further, we employ machine learning methods and neural networks to predict financial performance expressed in terms of the Z-score bankruptcy model. Eleven categories of sentiment (ranging from negative and positive to active and common) are used as the inputs of the prediction models. Support vector machines provide the highest forecasting accuracy. This evidence suggests that there exist non-linear relationships between the sentiment and financial performance. The results indicate that the sentiment information is an important forecasting determinant of financial performance and, thus, can be used to support decision-making process of corporate stakeholders.

Comprehensive assessment of firm financial performance using financial ratios and linguistic analysis of annual reports

“…However, the scope and structure of such information is different for well/poorly performing firms. Generally, the obtained results conform to the expectations of positive/negative sentiment categories (Davis et al, 2012;Hajek et al, 2014;Myskova & Hajek, 2016). The results of the present study suggest that firms that have good financial results verbally describe not only the achieved values of financial indicators, but they also elaborate the causes of the result in more detail.…”

“…Significant differences of word categories has been observed for firms with low/high earnings and stock returns (Li, 2008), stock market volatility (Loughran & McDonald, 2011), market-to-book ratio (Myskova & Hajek, 2016), return on assets (Davis et al, 2012), credit ratings (Hajek & Olej, 2013), Altman Z-score (Hajek et al, 2014). However, the dictionaries used previously are mostly limited with the focus on general linguistic categories, particularly on positive and negative sentiment.…”

“…A set of extensive dictionaries were developed by Loughran and McDonald (2011). These dictionaries have received much interest in recent related literature (Hajek et al, 2014;Myskova & Hajek, 2016;Henry & Leone, 2016). The following categories are included in these finance-specific dictionaries: positive, negative, uncertainty, litigious, and modal.…”

The Influence of CEO Letters’ Tone on Financial Indicators of the Company

“…пример на рис. 1) неоднократно успешно использовалась для осуществле-ния текстового анализа, демонстрируя высокую достоверность полученных ре-зультатов, что также стало одним из осно-ваний для выбора данного типа модели [Savas et al, 2015;Hájek, Olej, Myskova, 2014].…”

Text mining to detect indications of fraud in annual reports worldwide

“…Their results show that adding sentiment information leads to a more accurate prediction of financial distress than when the prediction is based on financial information alone. In a subsequent research with more sentiment word categories Hajek et al (2014) showed that a change in the development of a company influences the textual sentiment information in the annual report.…”