Management deception, big-bath accounting, and information asymmetry: Evidence from linguistic analysis

https://doi.org/10.1016/j.aos.2018.02.004Get rights and content

Abstract

Accounting big baths are pervasive in practice. While big baths can improve the information environment and reduce information asymmetry, they can also degrade the information environment and obscure operating performance. In this study, we examine the role of management ethics. Specifically, we investigate whether managers’ truthfulness (or conversely, deceptiveness) affects how investors perceive big baths. Using linguistic analysis on earnings-conference calls to measure managerial deception and employing a difference-in-differences research design with propensity-score matching, we find that information asymmetry is significantly higher following big baths taken by deceptive CEOs, compared with big baths taken by less deceptive CEOs.

Introduction

How does a firm's information environment change after accounting big baths? Prior literature provides evidence on both positive (Elliott & Shaw, 1988; Francis, Hanna, & Vincent, 1996; Haggard, Howe, & Lynch, 2015) and negative (Bens & Johnston, 2009; Kirschenheiter & Melumad, 2002; Kothari, Shu, and Wysocki 2009; Moore, 1973) consequences of big baths on the information environment. As managers have discretion regarding whether to incur a large write-off, and can decide the timing and amount of the write-off, management's incentives are important in studying the effects of big baths. However, such incentives are unobservable. Investors may use managerial characteristics to infer management incentives. Among the most salient managerial characteristics in this setting is truthfulness; thus this study examines how truthfulness (or conversely, deceptiveness) affects investors' perceptions of big baths.

According to upper-echelons theory (Hambrick & Mason, 1984), the ethical attentiveness throughout the organization is instilled by its leaders (Patelli & Pedrini, 2015). A series of accounting fraud scandals over the last decades put leadership ethics at the forefront of the heated debate on financial-reporting truthfulness (Mihajlov & Miller, 2012; Tourish & Vatcha, 2005). Ethics is an intrinsic part of managers' behavior (Solomon, 1992). As firms' high-level decision makers, top managers are likely to follow a cognitive and rational approach that revolves around moral judgments about the issues when making ethical decisions, just as an individual making a choice when facing an ethical dilemma (Albert, Scott, and Turan 2015; Kohlberg, 1981; Reynolds, 2006; Vitell, Lumpkin, and Rawwas 1991; Weber, 1990). Big baths are managerial decisions that can be the result of managers' ethical considerations of the firms' welfare, or can be the result of managers' incentives to maximize their personal utility. Being truthful or deceptive to investors and other stakeholders also indicates management's ethical choice of how they view their responsibility to the firm's stakeholders.

On one hand, big baths can manifest themselves as exceptionally large negative discretionary accruals. On the other hand, big baths can consist of one-time, large write-offs, and may include restructuring charges, asset impairments, and litigation losses. These write-offs are generally reflected as “special items” in the financial statements. There are two ways to look at a big bath. If a company reports a loss that is larger than expectations, it could be the case that there are certain issues within the firm that warrant such actions and that managers are truthfully conveying such information to the capital market and other stakeholders. In line with that view, some analysts interpret big baths as managers’ positive response to existing problems (Elliott & Shaw, 1988). Big baths can also “clear the air” (Haggard et al., 2015). That is, by writing off assets when their carrying values are less than the market values, the reported values of the assets are realigned with their economic values. As a result, firm-level information asymmetry following a big bath should decrease.

However, big baths are sometimes used as an earning-management technique to shift current earnings to future periods. As Levitt (1998) points out, if big-bath charges are overly conservatively estimated with “extra cushioning,” they can miraculously be reborn as income when future earnings fall short. Big baths can also be used to secure bonus payments. Often, managers' rewards are tied to meeting certain performance targets. In an economic downturn, managers may follow the big-bath approach by bundling as much bad news into the current period as possible, aiming to make their targets easier to achieve in the next period.1 In cases of new management, big-bath accounting can be used to mitigate top executives’ job-security concerns.2 The new manager can benefit from taking a big bath, blaming the low earnings on the previous manager so as to display an improved financial performance in future (Moore, 1973).

From investors' perspectives, the action of taking big baths by a firm is observable, but the motivations behind the action are not entirely clear. Hou (2015) finds that in a well-diversified market, idiosyncratic information risk is priced when information is subject to managers’ discretion and thus ambiguous. Can investors infer the motivations of managers by observing their types – truthful or deceptive – and associating their actions with the types? By taking advantage of newly-developed technologies, investors are now analyzing the linguistic patterns displayed in management speech. Investors have been using algorithmic textual analysis, CIA lie-detection techniques, and more recently, audio analysis of management speech, to seek an edge with stock calls, sector sentiment, and overall market direction.3

Earnings-conference calls, in which managers discuss their firms' financial performance with analysts and investors, are important information sources to search for signs of management deception. If a manager is discovered to be deceptive when discussing her firm's financial results, and the firm takes a big bath at the time, will investors perceive this big bath to have low credibility? Will investors associate this big bath with such motivations as meeting earnings targets or securing bonus payments? If this is the case, we would expect to observe an increase in information asymmetry following big baths taken by deceptive managers. Conversely, if a firm's manager takes a big bath and she is considered truthful, investors may perceive this big bath as having high credibility. In this case, the information asymmetry may decrease.

A primary reason both practicing accountants and researchers care about information asymmetry is that it reflects the information environment.4 Clearly an important issue related to big baths is whether these accounting events improve or deteriorate the firm's information environment. However, to broaden the scope of our study, in additional analyses we also test for trading volume (another commonly employed outcome variable in this line of literature).

This study builds on Larcker and Zakolyukina (2012) who find that certain words are significantly associated with management deception. For instance, deceptive CEOs use more “reference to general knowledge” and “extreme positive emotion” words, and use fewer “anxiety” and “shareholder value” words. The linguistic approach proposed by Larcker and Zakolyukina (2012) is based on psychological theories linking deception to linguistic behavior (Vrij, 2008), and is built up by applying a well-developed and frequently used psychosocial dictionary – LIWC. There are an increasing number of applications of LIWC analyses in deception detection, and also in personality, forensic, clinical, relationship, and cultural assessments (Chung & Pennebaker, 2012, chap. 12). Providing further validity to Larcker and Zakolyukina (2012), Loughran and McDonald (2013) demonstrate that the credibility of managers is diminished by having an overly positive S-1 in the IPO process, consistent with Larcker and Zakolyukina (2012) who find that deceptive CEOs use significantly more positive emotion words in conference calls. Another piece of evidence substantiating the usefulness of CEOs’ linguistic patterns in signifying deception is provided by Hobson, Mayew, Peecher, and Venkatachalam (2017), who demonstrate that once given instructions on the “cognitive dissonance” in the CEOs remarks, auditors are able to more precisely detect fraudulent companies as well as the unidentified “red flag” sentences in earnings-conference calls.

We use Larcker and Zakolyukina (2012) approach to identify truthful and deceptive managers in this paper, and examine whether the change in information asymmetry around a big-bath event is a function of managerial deception. As CEOs play the largest role in corporate decision making, we focus on the linguistic pattern of CEOs.

In our primary analyses we employ a difference-in-differences research design coupled with propensity-score matching of treatment and control firms. This approach provides strong control for potential confounding events as well as omitted-variable biases. We find evidence that investors are able to discern managers’ deception levels from conference calls and that information asymmetry is affected accordingly. Specifically, we find that information asymmetry (proxied using Amihud, 2002 illiquidity measure and bid-ask spreads) increases significantly after big baths taken by deceptive CEOs as compared to those taken by less deceptive CEOs. In additional analyses, we find that this effect is more pronounced when a CEO who has been truthful in the past becomes deceptive in the bath year. Our inferences are robust to a variety of regression specifications and other robustness tests.

The study adds to the big-bath line of research by examining a potentially important factor that could affect the impact of big-bath taking on information asymmetry. Second, our study contributes to research on how investors use ex-ante credibility of CEOs to interpret financial reporting quality (e.g., Loughran & McDonald, 2013). We do this by applying textual-analysis techniques to accounting issues to infer management's intentions using subtle linguistic cues. Finally, the study adds to management-ethics research by examining management deception and its associated capital-market consequences in the setting of big-bath taking, an economically important event as documented in prior research.5 This paper thus also contributes to the literature by studying the financial outcomes of CEO idiosyncratic characteristics and psychological patterns.

Section snippets

Big-bath literature

Prior research has examined the timing, motivation for, and consequences of taking big baths. Moore (1973) finds that discretionary accounting decisions that reduce income are more likely to be made in a period of management changes. The write-offs, many of which have substantial economic consequences, reflect decisions by corporate management. Kirschenheiter and Melumad (2002) construct a theoretical model to demonstrate that managers under-report earnings the most when the news is

Big baths

Identifying big-bath events is not straightforward. Generally speaking, the trade-off is between having a large enough sample that utilizes an objective (or “less subjective”) approach versus the researcher using her intuition to attempt to capture the spirit of big baths by coming up with a self-constructed measure.11

Sample selection and effectiveness of propensity-score matching

Table 1 presents the sample-selection process for the final treatment firms. Table 2 exhibits the logistic regression of the big-bath indicator on important predictors of such big baths.19 Of all the matching variables used in the propensity-score matching under the Accruals Approach in defining big baths, BTM_LAG (lag of book to market ratio), TURNOVER_A_LAG (lag of annual share turnover), LEVERAGE, and

Additional analyses and robustness tests

In this section, we first provide robustness tests related our primary analyses. In particular, we examine the effects of our matching procedures as well as other research-design choices. Next, we attempt to investigate the “mechanisms” through which investors learn about the big baths. Finally, we explore the effects of investors' prior experience with CEOs’ deceptive (or truthful) speech.

Conclusion

While some studies find that big baths can improve the information environment, others find that they degrade the information environment. To expand upon prior literature, this paper looks at big baths in conjunction with management deception. In particular, we examine whether management deception can decrease the credibility of big baths and alter investors’ perceptions. We find that information asymmetry (proxied for using the Amihud illiquidity measure and bid-ask spreads) increases

Acknowledgements:

We thank the Michael Lee-Chin research grant for financial support. We thank Stefan Anchev, Muhammad Azim, Stephanie Cheng, Mahfuz Chy, Bingxu Fang, Shushu Jiang, Ross Lu, Gordon Richardson, Barbara Su, Aida Wahid, Eyub Yegen, Wuyang Zhao, two anonymous reviewers, and seminar participants at the Rotman School of Management and the Canadian Academic Accounting Association for their comments and inputs. Hope gratefully acknowledges funding from the Deloitte Professorship.

References (87)

  • R.Y. Goyenko et al.

    Do liquidity measures measure liquidity?

    Journal of Financial Economics

    (2009)
  • K.S. Haggard et al.

    Do baths muddy the waters or clear the air?

    Journal of Accounting and Economics

    (2015)
  • F. Hodge et al.

    Management reporting incentives and classification credibility: The effects of reporting discretion and reputation

    Accounting, Organizations and Society

    (2006)
  • T. Loughran et al.

    IPO first-day returns, offer price revisions, volatility, and form S-1 language

    Journal of Financial Economics

    (2013)
  • K.J. Murphy et al.

    Financial performance surrounding CEO turnover

    Journal of Accounting and Economics

    (1993)
  • R.E. Petty et al.

    The elaboration likelihood model of persuasion

    Advances in Experimental Social Psychology

    (1986)
  • C.W. Smith et al.

    The investment opportunity set and corporate financing, dividend, and compensation policies

    Journal of Financial Economics

    (1992)
  • S. Utama et al.

    Institutional ownership, differential predisclosure precision and trading volume at announcement dates

    Journal of Accounting and Economics

    (1997)
  • L.S. Albert et al.

    Turning inward or focusing out? Navigating theories of interpersonal and ethical cognitions to understand ethical decision-making

    Journal of Business Ethics

    (2015)
  • K. Allee et al.

    The structure of voluntary disclosure narratives: Evidence from tone dispersion

    Journal of Accounting Research

    (2015)
  • J. Amernic et al.

    Measuring and assessing tone at the top using annual report CEO letters

    (2010)
  • K. Balakrishnan et al.

    Shaping liquidity: On the causal effects of voluntary disclosure

    The Journal of Finance

    (2014)
  • R.S. Baumhart

    How ethical are businessmen?

    Harvard Business Review

    (1961)
  • D.A. Bens et al.

    Accounting discretion: Use or abuse? An analysis of restructuring charges surrounding regulator action

    Contemporary Accounting Research

    (2009)
  • N. Bhattacharya et al.

    Does earnings quality affect information asymmetry? Evidence from trading cost

    Contemporary Accounting Research

    (2013)
  • A.W.A. Boot et al.

    Security design

    The Journal of Finance

    (1993)
  • R. Bowen et al.

    Do conference calls affect analysts' forecasts?

    The Accounting Review

    (2002)
  • S. Brown et al.

    How disclosure quality affects the level of information asymmetry

    Review of Accounting Studies

    (2007)
  • B.J. Bushee et al.

    The role of the business press as an information intermediary

    Journal of Accounting Research

    (2010)
  • T.L. Carson

    Self-interest and business ethics: Some lessons of the recent corporate scandals

    Journal of Business Ethics

    (2003)
  • J. Chae

    Trading volume, information asymmetry, and timing information

    The Journal of Finance

    (2005)
  • J.D. Chen et al.

    CEO accountability for corporate fraud: Evidence from the split share structure reform in China

    Journal of Business Ethics

    (2016)
  • T.E. Christensen et al.

    Creating a bigger bath using the deferred tax valuation allowance

    Journal of Business Finance & Accounting

    (2008)
  • C.K. Chung et al.

    Linguistic inquiry and word count (LIWC): Pronounced “Luke,”... And other useful facts. Applied natural language Processing: Identification, investigation, and resolution

    (2012)
  • A.M. Ciancia et al.

    The effect of CEO reputation and explanations for poor performance on investors' judgments about the company's future performance and management

    Accounting, Organizations and Society

    (2010)
  • R. Craig et al.

    Exploring top management language for signals of possible deception: The words of Satyam's chair Ramalinga Raju

    Journal of Business Ethics

    (2013)
  • A.K. Davis et al.

    The effect of manager-specific optimism on the tone of earnings conference calls

    Review of Accounting Studies

    (2014)
  • R. Dehejia et al.

    Causal effects in nonexperimental studies: Reevaluating the evaluation of training programs

    Journal of the American Statistical Association

    (1999)
  • M. Druz et al.

    Tips and tells from managers: How analysts and the market read between the lines of conference calls

    (2015)
  • M.G.S. Ekin et al.

    Business ethics in Turkey: An empirical investigation with special emphasis on gender

    Journal of Business Ethics

    (1999)
  • J.A. Elliott et al.

    Write-offs as accounting procedures to manage perceptions

    Journal of Accounting Research

    (1988)
  • J.H. Evans et al.

    Honesty in managerial reporting

    The Accounting Review

    (2001)
  • J. Francis et al.

    Causes and effects of discretionary asset write-offs

    Journal of Accounting Research

    (1996)
  • Cited by (52)

    • Task-oriented speech and information processing

      2024, Journal of Banking and Finance
    • Insider vs. outsider CEO and firm performance: Evidence from the Covid-19 pandemic

      2022, Finance Research Letters
      Citation Excerpt :

      Accordingly, we examine whether the presence of an outsider CEO is associated with such earnings management during the Covid-19 crisis. In doing so, we follow the prior literature (Haggard et al., 2015; Hope and Wang, 2018) and employ ‘special items’ as reported in the financial statements. The dependent variable is the inverted value of ‘special items’ (so that higher values reflect greater big baths) as a percentage of ‘lagged total assets’ (Big Bath1).

    • Are gender diverse boards more cautious? The impact of board gender diversity on sentiment in earnings press releases

      2021, Journal of Contemporary Accounting and Economics
      Citation Excerpt :

      Loughran and McDonald (2016) discuss how their dictionary outperforms other linguistic dictionaries and is more applicable in the accounting and finance domain. This “bag-of-words” approach, even though it does not consider the actual and contextual meaning of words, has the advantages of being easy to understand, implement, and replicate (Hope and Wang, 2018). Furthermore, the word list approach to sentiment analysis does not consider the frequency negation.

    • Accrual earnings management in response to an oil price shock

      2021, Journal of Commodity Markets
      Citation Excerpt :

      When big baths are used as an earnings management technique, they erode the information climate and obscure operating performance. However, if the asset market value is less than the book value, write-downs can improve the information environment and reduce information asymmetry (Hope and Wang, 2018). Leuz et al. (2003) investigated earnings management in different countries and found that the level of investor protection strongly affects earnings management behaviour and that Norway is among the countries with the lowest degree of earnings management.

    View all citing articles on Scopus
    View full text