Full length article
Do words reveal the latent truth? Identifying communication patterns of corporate losers

https://doi.org/10.1016/j.jbef.2020.100291Get rights and content

Abstract

Identifying communication patterns of narratives in disclosed financial statements is assuming importance in recent times. This paper explores the communication patterns through various channels of loser firms which have witnessed prolonged underperformance and compare the same w.r.t their peers from the same industry. We perform sentiment analysis in published communication of these firms, to scrutinize their tone and emotional-valence. Our findings suggest that loser firms have a negative mood while conveying earnings vis-à-vis their comparable firms.

Further, it is ratified that loser firms lack trust, joy, and surprise elements and they depict a surge of fear and sadness in their communication. Moreover, the themes of discussion using the topic modeling approach are indicative of concerns that may have affected the performance of losers. This study has the potential to unveil early warning signals for firms who are in bad shape, which otherwise remains undetected through numerical disclosures.

Introduction

Effective communication from firms is critical to managing impressions of key stakeholders, particularly during periods of prolonged underperformance (Salancik and Meindl, 1984). This study explores the communication patterns of 95 loser1 Firms and 95 matched peer firms using text and sentiment analysis of ‘Earnings (press) Releases’ (ER), ‘Management Discussion and Analysis’ (MDA), ‘Notes to Accounts’ (NTA) and ‘Independent auditors’ Reports’ (IAR). The financial state of any firm may not be genuinely reflected through various financial statements, because numbers are perceived to be the dry part of information (Kloptchenko et al., 2004) and are also susceptible to misrepresentation (Aharony et al., 2010, Ajit et al., 2013, Chen et al., 2011, Lennox et al., 2018, Sarkar et al., 2008). The chances of manipulating reported numbers are expected to be higher, particularly in markets with a large amount of information asymmetry, weak investor protection, and judiciary (eneish, 2001, Healy and Wahlen, 1999, Kedia and Philippon, 2005).

This study draws motivation from the information retrieval literature, which assumes that there are hidden signals in various means of corporate communications (Chen et al., 2017, Huang et al., 2014, Humpherys et al., 2011, Loughran and Mcdonald, 2016). The language of these public documents consists of certain subtle indications about the state of corporate, especially for firms in bad shape (Lu et al., 2013). The rationale for this study is primarily to explore whether the narrative discussions are in tandem with numeric disclosures for the perusal of stakeholders in the communication of prolonged underperforming firms.

Traditionally, this mammoth task of reading through the lines has been performed manually by a relatively handful of experts (He, 1999), thus making it a challenge to collate a range of diverse views. In the current age of digital technology, information retrieval is no more a concern, where text mining and related techniques offer an immense potential to derive meaningful insights (Huang et al., 2014, Kumar and Ravi, 2016, Loughran and Mcdonald, 2014). Firms broadly have two channels of communication: annual reports and earnings press release(s). Former, focuses on the macro environment and the overall ecosystem of the concerned business, with impetus on the future road map (Huang et al., 2014, Kloptchenko et al., 2004, Li, 2008). On the contrary, the latter are reflections of the current state of operational efficiency and profitability aspects of firms (Davis et al., 2012, Davis et al., 2011). Earnings press releases have a unified framework to announce quantitative and qualitative disclosures (Davis et al., 2012). There is a dearth of studies to analyze the latter part of the information which may have incremental insights over and above the numbers.

Scholars have acknowledged that the tone of communication has polarity embedded in such documents. Nevertheless, they have seldom been analyzed for firms in bad shape. Moreover, the mood or sentiment is skewed for annual reports where it is a common practice to showcase a bright outlook; the detailed analysis of such documents has also been lacking for corporate losers. In this study, we define corporate losers who have a net worth less than −100 mn (INR) for a consecutive span of 5 years (2013–2018)2 And thus are in bad shape. Net worth is the value of all the non-financial and financial assets owned by an institutional unit or sector minus the value of all its outstanding external liabilities and is one of the most popular and widely accepted measures to examine firm performance (Weisbrod and Hansen, 1968, Sekino and Watanabe, 2016).

In summary, this paper catechizes the following research questions: Do loser firms manifest a negative tone of communication? Does their implicit theme of discussion reflect any fundamental concern(s)? Do the themes of discussion or sentiments vary across various (sub)components of communication channels? Are there fundamental differences w.r.t these attributes in communication patterns of loser firms vis-à-vis comparable firms? While there are excellent reasons to believe that, as a firm’s prospects become less predictable, its communication with the world outside should reflect that (Brown and Utterback, 1985), few studies have examined this issue head-on. A firm’s prospect is an essential concern for all stakeholders who may be interested in monitoring the performance of firms confidently moving ahead or plagued by uncertainties, like changes in technology, market demands, regulatory requirements, material availability, and competition. This exploratory study presents some interesting insights into these questions.

Our study is different from previous studies and contributes to the extant literature in multiple ways. First, we focus on the firms which are reeling under spells of prolonged underperformance. This may echo the signals for future distress (if any). Second, the polarity of communications is compared for losers vis-à-vis the peer group to see whether the individual communication styles reflect the respective performance. Third, for documents that focus on the bigger picture, the study explores the themes of discussion, and whether it varies across the two groups of losers and the comparable firms (peer group). The remaining part of the paper is organized as follows: The next section discusses relevant literature on the issue, Section 3 discusses the data and methodology, Section 4 presents the results, while Section 5 discusses the results followed by summary and conclusions.

Section snippets

Identifying distress signals: conventional approaches used in accounting and finance literature

The practice of predicting the financial health of a company, particularly firms under distress, dates back several decades. Beaver, 1966, Beaver, 1968 uses univariate analysis for selected ratios and finds that some have outstanding predictive power. Altman (1968) develops and applies a multiple discriminant analysis model (MDA) called the Z-Score model. The next two decades see additional contributions to financial distress research. For example, Ohlson (1980) proposes a logit model, Taffler

Sample construction: identifying ‘losers’ and ‘peers’

We initially start with all 5150 firms listed in NSE and BSE and record their financial health in terms of net worth. To avoid the selection of firms by chance, we look at the firms’ well-being in the last five years. Firms with higher negative net worth rarely find opportunities of being bailed out and their restructuring process is cumbersome for emerging markets such as India. Of all 5150 listed firms, we select only those firms which had a negative net worth when averaged over the last five

Results & discussion

Table 2 shows a comparative analysis of the statistical summary of sentiment scores for the loser and its peer group for the four datasets considered for this research work. For ER, it is observed that the mean sentiment score for the peer group is positive, whereas the mean sentiment score for the losers’ group is negative. Furthermore, the maximum positive sentiment for the peer group is considerably more than that of the loser group.

For MDA, we observe that peers almost always have positive

Practical implications

Traditional accounting research has mostly explored the relationship between financial distress and historical accounting information. However, quantitative financial information comprises only approximately 20% of all the information contained in annual reports (Beattie et al., 2004). Accordingly, the financial disclosures in published statements may not be revealing the true state of corporate health. Therefore, to obtain a complete picture, it is necessary that one uses the qualitative

Conclusion

Textual analysis has been used across many disciplines, including computational linguistics, natural (or statistical) language processing, information retrieval, content analysis, or stylometrics (Loughran and Mcdonald, 2016). Earlier studies also find evidence that firms convey sentiments and subtly hint at their future performance through such communications (Davis et al., 2012). Mining text for sentiment (also known as affect or emotion) and understanding how it can affect readers’ thinking

CRediT authorship contribution statement

Rahul Kumar: Conceptualization, Methodology, Software, Data curation, Writing - original draft, Visualization, Investigation, Validation, Writing - review & editing, Formal analysis. Soumya Guha Deb: Data curation, Writing - original draft, Visualization, Investigation, Writing - review & editing, Formal analysis. Shubhadeep Mukherjee: Data curation, Writing - original draft, Software, Validation, Writing - review & editing.

References (64)

  • LuY.C. et al.

    Revisiting early warning signals of corporate credit default using linguistic analysis

    Pac. Basin Financ. J.

    (2013)
  • NguyenT.H. et al.

    Sentiment analysis on social media for stock movement prediction

    Expert Syst. Appl.

    (2015)
  • SunA. et al.

    Trade the tweet: Social media text mining and sparse matrix factorization for stock market prediction

    Int. Rev. Financ. Anal.

    (2016)
  • TafflerR.J.

    Empirical models for the monitoring of UK corporations

    J. Bank. Financ.

    (1984)
  • AjitD. et al.

    2013 Sebi Drg study earnings management in India

    (2013)
  • AltmanE.I.

    Financial ratios, discriminant analysis and the prediction of corporate bankruptcy

    J. Finance

    (1968)
  • AntweilerW. et al.

    Is all that talk just noise ? The information content of internet stock message boards

    Am. Financ. Assoc.

    (2004)
  • Bar-HaimR. et al.

    Identifying and following expert investors in stock microblogs

  • BeaverW.H.

    Financial ratios as predictors of failure

    J. Account. Res.

    (1966)
  • BeaverW.H.

    The information content of annual earnings announcements

    J. Account. Res.

    (1968)
  • BholatD.M. et al.

    Text mining for Central Banks

  • BrownJ.W. et al.

    Uncertainty and technical communication patterns

    Manage. Sci.

    (1985)
  • ChenH. et al.

    Effects of audit quality on earnings management and cost of equity capital: Evidence from China

    Contemp. Account. Res.

    (2011)
  • ChenT. et al.

    Improving sentiment analysis via sentence type classification using BiLSTM-CRF and CNN

    Expert Syst. Appl.

    (2017)
  • DavisA.K. et al.

    Beyond the numbers: An analysis of optimistic and pessimistic language in earnings press releases

    Financ. Account. Rep. Sect. Pap.

    (2011)
  • DavisA.K. et al.

    Beyond the numbers: Measuring the information content of earnings press release language

    Contemp. Account. Res.

    (2012)
  • DebS.G. et al.

    Low Leverage Policy: a boon or bane for Indian Shareholders

    J. Asia Bus. Stud.

    (2018)
  • DehejiaR.H. et al.

    Propensity score-matching methods for nonexperimental causal studies

    Rev. Econ. Stat.

    (2002)
  • eneishM.D.

    Earnings management: A perspective

    Manag. Financ.

    (2001)
  • Engelberg, J., 2008. Costly information processing: Evidence from earnings announcements. In: AFA 2009 San Francisco...
  • HeQ.

    Knowledge discovery through co-word analysis

    Libr. Trends

    (1999)
  • HealyP.M. et al.

    A review of the earnings management literature and its implications for standard setting

    Account. Horiz.

    (1999)
  • Cited by (0)

    View full text