Readability of Notes to Consolidated Financial Statements and Corporate Bond Yield Spread

Abstract

We examine the association between the readability of ‘Notes to consolidated financial statements’ (Notes) in annual reports and corporate bond yield spread. We find that less readable narrative disclosure of Notes is significantly associated with greater bond yield spread. In addition, the association becomes stronger for firms in high-tech sectors or those with higher equity volatility, whereas it becomes weaker for firms with higher profitability or those with the reporting location of Notes being outside of the 10-K format file. We also provide evidence that Notes readability is helpful in explaining the puzzle of a positive yield spread always appearing when the bond approaches its maturity in practice. Overall, our results suggest that the Notes readability has a substantial association with bond yield spread and its term structure.

Keywords:

• Notes to consolidated financial statements (Notes)
• Narrative disclosure readability
• Notes readability
• Bond yield spread
• Credit risk
• Term structure

JEL codes:

• G32
• M41
• G33

Acknowledgements

We are greatly indebted to Hervé Stolowy (the editor), two anonymous reviewers, Yan-Shing Chen, and the conference participants at 2015 AAA Annual Meeting for their insightful suggestions and comments on earlier drafts of the paper. We also thank Po-Nien Chiang for the excellent research assistance. We gratefully acknowledge the financial support from the Ministry of Science and Technology in Taiwan (NSC 101-2410-H-030-029-MY2 & MOST 106-2628-H-009-004-MY3) and the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental Data and Research Materials

Supplemental data for this article can be accessed on the Taylor & Francis website, doi:10.1080/09638180.2020.1740099.

Appendix A1. Example of Incomplete 10-K Format File.

Appendix A2. Example of Reporting Location of Notes in the EX-13 (10-K Filing).

Appendix A3. The Endogeneity Discussions of the Relationship between Notes Readability and Bond Yield Spreads.

Appendix A4. Supplemental Tables and Figures.

Notes

1 Financial disclosure contents refer to the decision to disclose, and whether to disclose all or selected information. For readability regulations, the Securities and Exchange Commission (SEC) simply encourages firms to adopt the plain English principle in their annual reports.

2 Strategic reporting is defined as managers intentionally reporting in a way that highlights favorable information and obfuscates negative information about a firm to its external investors. Hence, narrative disclosures in annual reports (including Notes) could be viewed as a tool of strategic reporting by managers.

3 It may be argued that the MD&A section includes more relevant information regarding firms’ current and future performance and is more relevant for assessing future risks that drive the spreads. However, the disclosure in the MD&A section is less regulated, suggesting that managers have more discretion in the disclosure of its contents. Managers may thus have less incentive to engage in strategically textual reporting in this section.

4 According to structural-form credit models of Merton (1974) and Duffie and Lando (2001), a firm’s default risk is positively related to asset value variance, default threshold, and incomplete accounting information while has the opposite association with asset value level. The illustrated comparison of these two models can be referred to Chen et al. (2015).

5 Following Yu (2005) and Lu et al. (2010), the incomplete information spread is defined as the gap between the term structure of yield spread for bond issuers with high incomplete information and for bond issuers with low incomplete information at the respective maturity.

6 Many studies also demonstrate that there is a positive relationship between annual report quality and investment efficiency, such as Bushman and Smith (2001), Healy and Palepu (2001), and Biddle and Hilary (2006). Biddle et al. (2009) also show that higher annual report readability makes the firm diverge less from the predicted investment levels and be less sensitive to macroeconomic conditions.

7 The structural-form credit models (Merton, 1974; Duffie & Lando, 2001) are single-period models which derive the probability of default from a random variation in the unobservable value of the firm's assets. However, it must be noted that Merton’s (1974) credit risk model has the assumption of complete information while Duffie and Lando’s (2001) credit risk model extends the limitation of that assumption. Hence, Merton’s (1974) model theoretically underestimates bond yield spread when the firm’s bond is approaching its maturity date due to the complete information assumption.

8 The incomplete information spread introduced by Yu (2005) and Lu et al. (2010) are based on (analyst-based) AIMR disclosure ranking and trading-based information asymmetry (e.g. adjusted PIN score; Duarte & Young, 2009). Huang and Zhang (2012) propose that an industry-specific subcommittee of security analysts evaluates a firm’s disclosure quality (namely, its AIMR ranking) in three dimensions: annual published information, quarterly and other published information, and investor relations. The AIMR ranking thus primarily describes the completeness of a firm’s disclosed information in its published reports based on the subjective judgment of security analysts. Hence, the AIMR ranking may be affected by the biases of survey respondents (Chen & Liao, 2015).

9 Li (2008) demonstrates that the readability variation of Notes section is slightly greater than that of the whole annual report, suggesting that the linguistic features of Notes are not entirely the same and vary across firms over time. Hence, Notes readability plays an important role in external investors’ knowledge about the quality of a firm’s financial statements and its related explanations.

10 The incomplete information spread of Notes readability is defined as the gap between the term structure of yield spread for bond issuers with low Notes readability (high incomplete accounting information) and for bond issuers with high Notes readability (low incomplete accounting information) at the respective maturity. The illustration of incomplete information spread of Notes readability is shown as Figure A1(a) in online appendices.

11 Since this study focuses on the readability of the Notes section which is only a subsection of the annual report, it is not appropriate to employ the 10-K document file size (Loughran & Mcdonald, 2014) as the readability measure of the Notes section.

12 In addition to the readability measures used in this study, Bonsall IV et al. (2017) propose a new measure of readability, the Bog Index, that captures the plain English attributes of disclosure.

13 Since the data on AIMR rankings is not available after 1996, it is difficult to provide robustness evidence with the control variable of AIMR rankings during our sample period. Unlike AIMR rankings, annual report readability describes the quality of the textual and lexical presentation rather than the completeness of the disclosed information (e.g. quantitative information). Further, the measurement of annual report readability is more objective (e.g. readability formula), and the availability of data on annual report readability data is less limited. Based on the above discussion, annual report readability not only represents another manifestation of a firm’s incomplete accounting information but also provides a more objective and less limited method of measuring disclosure quality.

14 A Central Index Key or CIK number is a number given to an individual or company by the United States Securities and Exchange Commission. The number is used to identify the filings of a company, person, or entity in several online databases, including EDGAR. The numbers are ten digits in length.

15 Among the firms with 10-K URLs (uniform resource locator) during the period 1991 and 2012 (observation number: 114830), 13.12% (observation number: 15066) had an EX-13 URL, suggesting that at least 13.12% of all firms have ‘incomplete’ presentations of their 10-K format files. Examples of an incomplete 10-K format file and an EX-13 URL can be seen in Appendix A1 and A2, respectively (see online appendices).

16 That is, this study hand-collects the complete textual contents of Notes and MD&A in other links (e.g. ANNUAL REPORT/ EX-13) to replace the simplified descriptions of Notes and MD&A in the original 10-K format file. In doing so, this study also discusses whether the reporting location of Notes changes the association between Notes readability and corporate bond yield spread.

17 Table 2 also reports summary statistics of Notes_FOG, MDA_FOG, Notes_SMOG and MDA_SMOG which are used for robustness checks.

18 Bps refers to basis points. A basis point equals 0.01%. Moreover, corporate bond yield spreads may be extremely high during a period of fragile macroeconomic condition. (e.g. subprime crisis period) and may become extremely low (or negative) when the U.S. government experiences a serious budget deficit (e.g. in 2011). Therefore, this study winsorizes the bond yield spreads at the top and bottom 1% and then implements the related empirical analyses, similar to Chen and Liao (2015), Chen et al. (2019), Güntay and Hackbarth (2010), and Oikonomou et al. (2014).

19 The average of the Notes_FOG variable in this study (15.96) is lower than the figure reported in Li (2008) (18.96), suggesting that the Notes readability during our sample period (2003-2012) is higher than during Li’s (2008) sample period (1993-2003). This is mainly because the sample period of this study is after 1998, which is when the SEC encouraged listed firms to adopt plain English in their filings. The sample period of Li (2008) from 1993 to 2003 covers the period before the encouragement of plain English. Since our sample is during the period of plain English encouragement, the results imply that the encouragement of plain English increases Notes readability.

20 The R-squared seems to be low given the model specification with two-way fixed effects. One reasonable explanation is that a firm can issue a variety of bonds and thus our sample is bond-year observations rather than firm-year ones. Hence, the regression model with non-bond-fixed effect using the annual bond sample observations may not significantly improve the R-squared. In addition, it is common that the regression model includes the non-bond-fixed effect using annual bond observations in the bond yield spread literature, such as Chen and Liao (2015) and Chen et al. (2019). This is mainly because the bond sample has been screened by the criteria of straight corporate bond issues (e.g. Yu, 2005; Lu et al., 2010; Chen & Liao, 2015; Chen et al., 2019). Meanwhile, the main bond-level characteristic differences for a firm’s bond issues (e.g. coupon rate, time to maturity, bond age, issued amount, and bond rating) have been controlled in our model setting.

21 For example, Notes readability and MD&A readability are both correlated with firm size, which may contribute to the observed high correlations between Notes readability and MD&A readability. Hence, after controlling for firm size variable, the correlations between Notes readability and MD&A readability may become weaker and thus Notes readability is likely to have its own specific impact on firm credit risk.

22 The number ‘1.2461’ is the standard deviation of Notes_FK, shown in the first line of Panel B in Table 2.

23 Since we follow Yu (2005) and Lu et al. (2010) and set the largest maturity knot of the term structure at 30 years, the sample bond observations with maturity longer than thirty years are excluded in Equations (5) and (6).

24 The subprime crisis in 2007 and the European debt crisis in 2011 shifted the distribution of firm value to the left. Firms generally have larger credit spreads during the period of fragile macroeconomic condition, and firms with extremely high credit spreads may appear. This study winsorizes the YS variable to deal with the extreme value problem and then discusses whether fragile macroeconomic condition moderates the incomplete information spread of Notes readability for short-maturity bonds.

25 We use the number of uncertain words (absolute measure) rather than the percentage of uncertain words relative to all words (relative measure). This is mainly because bondholders may particularly emphasize the text uncertainty in some important paragraphs of the Notes section (e.g. paragraphs relating to a firm’s debt-paying ability). The relative measure may underestimate the impacts of uncertain words in these relevant paragraphs. As a result, this study uses the absolute measure to capture the ambiguous tones of the Notes section.

Additional information

Funding

We gratefully acknowledge the financial support from the Ministry of Science and Technology in Taiwan (NSC 101-2410-H-030-029-MY2 & MOST 106-2628-H-009-004-MY3) and the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan.