Is Bad News Difficult to Read? A Readability Analysis of Differently Connoted Passages in the Annual Reports of the 30 DAX Companies

Abstract

This study examines the strategic use of readability to obfuscate negative news in a German financial communication context. Combining a manual and an automated content analysis, the authors assess the tone and readability of three parts (chairman’s address, share-price development, and development in the fiscal year) of the 2014 annual reports of the 30 companies listed in the German stock index DAX. The results indicate that positively connoted passages in annual reports are not necessarily easier to read than negatively connoted passages. Furthermore, the readability of the annual report varies depending on the part and its function within the report.

The Language of Optimism in Corporate Sustainability Reports: A Computerized Content Analysis

“…Greenwashing can also be recognized by the language used, as it tends to employ overly optimistic language focused on positives only. Thus, greenwashing literature has parallels with research on selective disclosure where positive information is exaggerated and negative information is downplayed (Jahdi & Acikdilli, 2009; Laskin, 2018b; Porter & Kramer, 2006; Thoms et al, 2020). However, in previous research, such selective disclosure has usually been studied based on the content rather than based on the language used.…”

Readability of Annual Reports on the Vienna Stock Exchange: A Test of Management Obfuscation Hypothesis

“…Nevertheless, the formulas convince through their automation capability, simplicity, and independence from the readers (Bayerlein & Davidson, 2011;Bruce et al, 1981;Crossley et al, 2019;Redish, 2000). We believe that for readability studies, the absolute result of one single text is of less interest than the influence of time or other factors on the score of a bigger sample of texts (Thoms et al, 2020). Readability formulas also form the primary tool of readability research of written company disclosures (Adelberg, 1979;Courtis, 2004;Jones, 1988;Li, 2008;Rutherford, 2003;Smith & Taffler, 1992;Subramanian et al, 1993;Thoms et al, 2020).…”

“…We believe that for readability studies, the absolute result of one single text is of less interest than the influence of time or other factors on the score of a bigger sample of texts (Thoms et al, 2020). Readability formulas also form the primary tool of readability research of written company disclosures (Adelberg, 1979;Courtis, 2004;Jones, 1988;Li, 2008;Rutherford, 2003;Smith & Taffler, 1992;Subramanian et al, 1993;Thoms et al, 2020). Alternative procedures for assessing readability include usability tests with representative readers (Redish, 2000), asking readers to assess text where every nth word is deleted 1 (Smith & Taffler, 1992;Taylor, 1953) and other, more modern and multi-dimensional computational linguistics analysis tools such as the Coh-Metrix (Chang & Stone, 2019;Crossley et al, 2019).…”

“…Reading, analysing and interpreting data becomes a costly affair for investors and, therefore, weakens the negative effect of such data on a firm's reputation and share (Bloomfield, 2002;Courtis, 1998;Li, 2008). Adjusting the readability, a trigger for understandability, of the letter to the shareholders and/or interview with the board is a potential way to spread good and obfuscate bad news, results or outlooks (Baker & Kare, 1992;Courtis, 1995;Li, 2008;Rutherford, 2003;Thoms et al, 2020). If the management obfuscation hypothesis held true, lower performance (outlook) of a company would result in lower readability measures.…”

“…Previous results on this hypothesis are mixed and unclear. There is a broad consensus that annual reports are generally difficult to read, yet the relation between performance measures and readability was observed to be significant only in some studies (Adelberg, 1979;Baker & Kare, 1992;Hrasky et al, 2009;Jones, 1988;Li, 2008;Subramanian et al, 1993;Thoms et al, 2020), while others find no evidence towards obfuscation (Courtis, 1986(Courtis, , 1995Rutherford, 2003;Smith et al, 2006).…”