Disclosure Committees: Implications for Disclosure Quality and Timeliness

ABSTRACT

To help companies comply with the certification requirements under Section 302 of SOX, the SEC recommends issuers form a disclosure committee. In this study, we examine the effects of disclosure committees on disclosure quality and timeliness. We find that the presence of disclosure committees is associated with higher quality and more timely corporate disclosure. In addition, we provide evidence that the benefits of disclosure committees on disclosure quality are greater if membership detail is publicly revealed and that benefits of the committee may be greatest for firms that experience a negative disclosure event. Lastly, we provide evidence that disclosure committees are associated with higher quality earnings announcements and lower likelihood of receiving a severe SEC comment letter. Collectively these results suggest disclosure committees are not merely ‘window dressing’, a conclusion with implications for practitioners, regulators, and academics interested in improving corporate disclosure practices.

KEYWORDS:

• Disclosure committee
• corporate governance
• disclosure quality
• disclosure timeliness

JEL classifications:

• M41
• M48

Acknowledgement

We appreciate helpful comments and suggestions from Beatriz Garcia Osma (editor), two anonymous reviewers, and the workshop participants at the University of Massachusetts Dartmouth, and the University of New Hampshire, the 2020 AAA Annual Meeting, 2020 Financial Accounting and Reporting Section Mid-Year Meeting, and 2020 Annual Meeting of the Northeast Business & Economics Association. Professors Nash and Xu acknowledge summer research support provided by the Peter T. Paul College of Business and Economics at the University of New Hampshire.

Disclosure statement

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

Notes

1 These assertions were also confirmed in our discussions with the Deputy General Counsel at a large accelerated filer.

2 While we expect disclosure committees to improve outcomes, institutional theory argues that they may merely fulfil symbolic roles (Cohen et al., 2008). Anecdotal evidence suggests if ‘a disclosure committee (does not) include the right people and (go) through the right process … its mere existence is a potential liability’ (EY, 2014, p. 3). Consequently, it is an empirical question whether disclosure committees provide substantive benefits.

3 In the extant literature, our measures of disclosure quality and timeliness are often referred to as measures of financial reporting quality and timeliness. For the purposes of clarity within this manuscript, we reference disclosure quality and timeliness consistently, but we acknowledge alternative terminology exists.

4 The lack of findings in Schmardebeck (2015) does not necessarily suggest that disclosure committees do not improve disclosure quality and timeliness. While some research has found that better disclosure reduces information asymmetry, others find that better disclosure can increase information asymmetry (Botosan & Plumlee, 2002; Brown & Hillegeist, 2007; Bushee & Noe, 2000). This is because better disclosure quality may attract transient investors and investors’ incorporation of disclosed information varies depending on disclosure types (e.g. annual or quality reports). Botosan and Plumlee (2002) concludes that ‘aggregating across different disclosure types results in a loss of information.’ Further, disclosure committees have an impact on disclosure beyond just earnings (EY, 2014). Therefore, we explore direct effects of disclosure committees on earnings and non-earnings related disclosures.

5 The evaluation of the effectiveness of internal controls are further specified in Section 404(a) of SOX, which requires management to document, test, and report on internal controls (SEC, 2004).

6 Disclosure committees typically include members of management and are considered a management committee. However, while rare, some companies prefer that the disclosure committee functions as a subcommittee of the board and audit committee (Deloitte, 2013). We leave further examination of structural differences to future research.

7 We use a professional data collection service to minimize human error and bias related to hand-coding disclosure information. MyLogIQ provides data from SEC filings and their data service is used by businesses and entities such as Ernst & Young, the Financial Executives Research Foundation and The Conference Board (EY, 2014; EY, 2015; FERF, 2017; TCB, 2016).

8 Among these sixteen observations, seven disclose its employee or director serving on another company's disclosure committee, three disclose that an acquired business had a disclosure committee, two disclose they do not have disclosure committees, and two disclose that accounting staff served as a ‘de facto’ disclosure committee. Lastly, two more observations had no mention of disclosure committees, hence are considered data errors by MyLogIQ.

9 It is possible that a firm with a disclosure committee will choose not to disclose having such a committee in any SEC filing. If this were the case, the firm would be falsely classified as not having disclosure committee in our sample. We recognize this as a limitation of our sample collection process.

10 Our use of restatements as a proxy is further supported by anecdotal evidence. A bank CAO responded ‘[a]s proof, we haven't had any restatements in the last few years,’ when asked to comment on the effectiveness of the disclosure committee (EY, 2014, p. 25).

11 We do not differentiate between restatements based on their cause or characteristics in the main analyses because Sellers et al. (2020) note including all occurrences ‘provides the most inclusive measure’ (p. 3).

12 A disclosure committee's responsibility to improve the textual contents of corporate filings was confirmed by a SEC Reporting and Technical Accounting Manager at a large accelerated filer

13 This index is a comprehensive measure of readability based on plain English attributes similar to those highlighted by the SEC and is designed to capture writing features that ‘bog’ readers down. A higher value of the Bog Index indicates a lower readability level. Schmardebeck (2015) measures readability by file size and word count.

14 Non-accelerated filers are required to submit their 10-K filings within 90 days of the company's fiscal year-end. For fiscal years ending on or after December 15, 2003, companies deemed accelerated filers (AFs) were required to submit their 10-K filings within 75 days of fiscal year-end (SEC, 2002b). The rule further shortened the 10-K filing delay to 60 days for fiscal years ending on or after December 15, 2006 for large accelerated filers (SEC, 2005). We use the respective filing deadline for each filer type in a given year to determine whether a company has filed 10-K late.

15 Unreported analyses show our results are also robust to the use of additional controls, including financial, accounting and supervisory expertise, both on the board of directors and the audit committee (Bailey, 2019), as well as audit fees, tax-related fees, and other non-audit fees (Paterson & Valencia, 2011). In order to present a parsimonious model, we do not include the above-mentioned controls in the main model.

16 As an alternative research design, we follow Hines and Peters (2015) and Bills et al. (2016a), and perform within-firm analyses. We identify firms that adopt a disclosure committee during the sample period and create a subsample comprised solely of observations associated with these firms. The benefit of this approach is both treatment and control observations are from the same subset of firms that choose to adopt a disclosure committee. This helps to ensure that our results are not driven by firm characteristics associated with the decision to form a disclosure committee. We re-estimate Model (1) and find DCOMM is negative and significant for all four dependent variables (RSTMNT, BOG, FILELAG, and LATE).

17 Unreported results show that the number of firms with a disclosure committee as a percentage of annual observations is relatively stable throughout the sample time period, ranging from 4.51 to 6.01 percent. We also find disclosure committees are fairly evenly distributed across industries. Of the 67 two-digit industries included in the full sample, 56 have at least 1 firm-year where a disclosure committee was present. The mean (median) percentage of firm-years in an industry with a disclosure committee is 6.80 (4.65) percent, and in no industry do more than 25 percent of firm-years have a disclosure committee. To ensure our main results are not driven by differences between industries with the presence of disclosure committees and those without, we eliminate from the full sample industries where no firm has a disclosure committee and re-estimate Model (1). In unreported results we find our main inferences hold in this alternate sample.

18 Prior studies report that 13.2 to 17.8 percent of firm-years are subsequently restated (Bryant-Kutcher et al., 2013; Newton et al., 2013). The mean of RSTMNT is consistent with these studies. Because restatements can cover multiple periods, the number of firm-years where restated financials are issued is significantly lower (Bills et al., 2016b).

19 The untabulated univariate results are consistent with the multivariate regression results presented in Table 2 in that companies experiencing restatements, late filings and internal control weaknesses are more likely to adopt disclosure committees.

20 Because we downloaded the Restatement data in August 2021, we have allowed sufficient time for RSTMNT especially in the last year of our sample period (i.e. year 2019) to be complete. However, to ensure the robustness of our results, we exclude restatements in 2019 and the results remain robust.

21 In untabulated analyses we examine the relation between disclosure committees and Dechow and Dichev (2002)'s working capital accruals and find that the absolute value of working capital accruals is negatively associated with the presence of disclosure committees. This indicates that the presence of disclosure committee is associated with better accrual quality, providing support for H1.

22 We focus on comment letters that are indicative of a potential issue and ignore those that may be simply asking for a minor clarification. Following prior research, we do not consider comment letters that are resolved in a short period of time to be severe comment letters. Specifically, we exclude the first quintile of letters characterized by the shortest conversation time between the SEC and company (Gietzmann & Pettinicchio, 2014).

23 Because we downloaded the SEC comment letters in August 2021, we have allowed sufficient time for CL_REC especially in the last year of our sample period (i.e. year 2019) to be complete. However, to ensure the robustness of our results, we exclude the SEC comment letters issued in 2019 and the inferences remain the same. The DiD results remain robust at 4%. Although not conventionally significant, the p-values for OLS (entropy balancing) results hold at 12% (11%).

24 Of the 1,151 firm-years for which we obtained additional information, 203 provided the name and position of members, and 428 provided only the positions of members. Besides CEOs and CFOs, common committee members include vice presidents and directors of business segments. Refer to Appendix A for an example.

25 In our cross-sectional tests, we find weaker results with respect to disclosure timeliness. Potential explanations for this outcome include that the incentive to further improve timeliness may not be as great as to improve quality, the degree to which these outcomes can be improved is capped, and other parties (such as external auditors) can play a significant role in determining timing.

26 Only one of the eight interactions is significant. Specifically, the interaction term between DCOMM and AC_SIZE in the BOG model is significant, indicating the benefit in improving the 10-K readability is greater when audit committee is smaller.

27 In exploring this possibility, we assume that a company has dissolved the committee if the firm stops disclosing information about the committee in subsequent years.