Word power: A new approach for content analysis

Abstract

We present a new approach for content analysis to quantify document tone. We find a significant relation between our measure of the tone of 10-Ks and market reaction for both negative and positive words. We also find that the appropriate choice of term weighting in content analysis is at least as important as, and perhaps more important than, a complete and accurate compilation of the word list. Furthermore, we show that our approach circumvents the need to subjectively partition words into positive and negative word lists. Our approach reliably quantifies the tone of IPO prospectuses as well, and we find that the document score is negatively related to IPO underpricing. Disciplines Finance | Finance and Financial ManagementThis journal article is available at ScholarlyCommons: https://repository.upenn.edu/fnce_papers/242Word Power: A New Approach for Content Analysis Narasimhan Jegadeesh Di Wu * July 2013ABSTRACT We present a new approach for content analysis to quantify document tone. We find a significant relation between our measure of the tone of 10-Ks and market reaction for both negative and positive words. We also find that the appropriate choice of term weighting in content analysis is at least as important as, and perhaps more important than, a complete and accurate compilation of the word list. Furthermore, we show that our approach circumvents the need to subjectively partition words into positive and negative word lists. Our approach reliably quantifies the tone of IPO prospectuses as well, and we find that the document score is negatively related to IPO underpricing. * Narasimhan Jegadeesh is the Dean's Distinguished Professor at the Goizueta Business School, Emory University, and NBER, email: Jegadeesh@emory.edu This paper presents a new approach to determine the strength of various words in conveying negative or positive tone that is particularly suitable for finance and accounting applications. We objectively determine term weights based on market's reactions to 10-K filings. We apply our approach to determine the relative weights for the words in the lexicon to quantify the tone of 10-Ks.We find several new results with our approach. For much of our analysis, we compute term weights using the positive and negative word list compiled by Loughran and McDonald.Interestingly, the quantitative score that our approach assigns to various 10-Ks have very low correlation with the scores assigned by Loughran and McDonald using the same list of words.1 For example, the Harvard list classifies tax, cost, and liability as words with negative connotation but these words typically have a neutral connotation in finance and accounting.2 Available at http://www3.nd.edu/∼mcdonald/Word Lists.html 2 More importantly, our approach provides a more reliable measure of document tone than the earlier approaches. In particular, we find a significant relation between document tone and market reaction for the positive words, while none of the other papers in the literature have been successful in doing so. Additi...

Readability of 10‐K Reports and Stock Price Crash Risk

“…The 10-K reports are one of the most comprehensive and credible channels through which managers convey their superior information to outside investors. Although many of the key numbers in the financial statements have already been disclosed well before the 10-K report filing dates, other sections of the reports, such as the Management Discussion and Analysis, provide investors with new and important supplementary information (Brown and Tucker 2011;Feldman et al 2010;Griffin 2003;Jegadeesh and Wu 2013;Kothari et al 2009a;Li 2010a;Loughran and McDonald 2011;You and Zhang 2009). Managers are required to explain in the 10-K reports the key driving forces responsible for changes in current performance, which helps investors to determine whether current performance is indicative of future performance.…”

CEO-Speeches and Stock Returns

“…To assess the quality of their dictionary, the authors show that 73.8% of the General Inquirer dictionary's negative words do not have a negative meaning in financial documents and, in later work, demonstrate that none of the most frequently occurring negative words in the 10-K disclosures are included in the Henry (2008) dictionary (Loughran & McDonald, 2015). Due to its comprehensiveness and its appropriateness for financial documents, the LM dictionary has become the most widely used dictionary in business research and has been used to assess the textual sentiment of 10-K filings (Loughran & McDonald, 2011), earnings conference calls (Davis et al, 2015), news articles (García, 2013), or IPO prospectuses (Ferris et al, 2013;Jegadeesh & Wu, 2013 Kearney and Liu (2014) and Loughran and McDonald (2016).…”

“…Several measures to gauge textual sentiment have been utilized in the literature. Jegadeesh and Wu (2013) and García (2013) employ direct measures of positivity and find statistically significant market reactions. McDonald (2011, 2016), however, point out that positive words are frequently used to frame negative words, whereas negative words are unambiguous in their usage.…”

“…Recent research has acknowledged the value not only of quantitative data disclosures but also of qualitative information, predominantly in the form of the textual sentiment of business communication. Sentiment is typically examined via content analyses which have been applied on several types of business communication such as annual reports (Feldman et al, 2008;Jegadeesh & Wu, 2013;Loughran & McDonald, 2011, 2015, earnings press releases Davis & Tama-Sweet, 2012;Henry, 2008;Henry & Leone, 2016;Huang et al, 2014), IPO prospectuses (Demers & Vega, 2008;Ferris et al, 2013;Jegadeesh & Wu, 2013), CEO letters (Boudt & Thewissen, 2016), and earnings conference calls (Davis et al, 2015;Doran et al, 2012;Larcker & Zakolyukina, 2012;Price et al, 2012). In general, these studies find that qualitative information is indeed processed by investors and helps to predict future accounting returns, stock returns, stock volatility, and stock trading volume.…”

Word Power: The Impact of Negative Media Coverage on Disciplining Corporate Pollution

“…Second, news media use specific words to indicate their opinion about a focal corporation (Jegadeesh and Di Wu 2013;Tetlock et al 2008). Not all news reports have the same impact on readers; hence, assuming every negative report is equal to every other negative report that is misguided.…”