Complexity of financial reporting standards and accounting expertise

Abstract

This study tests whether firms seek to mitigate the adverse effects of Financial Reporting Complexity (FRC) by investing in accounting expertise. We develop a measure of FRC based on the complexity of accounting standards that govern annual disclosures. We find that FRC is positively related to the accounting expertise on a firm’s board of directors and audit committee. We also find that accounting expertise mitigates the relation between FRC and negative reporting outcomes. Collectively, this study increases our understanding of the actions firms take to mitigate the negative consequences of FRC, and the role of accounting expertise in this setting.

Introduction

The U.S. Securities and Exchange Commission’s (SEC) Advisory Committee on Improvements to Financial Reporting defines financial reporting complexity (FRC) as: “the difficulty for... preparers to properly apply generally accepted accounting principles in the U.S. (U.S. GAAP) and communicate the economic substance of a transaction or event and the overall financial position and results of a company... ” (Securities and Exchange Commission, 2008, p. 18).¹

Prior literature documents a growing trend in FRC (Dyer et al., 2017) and examines its consequences. Generally, firms with high FRC have a less favorable information environment (e.g., Lehavy, Li, Merkley, 2011, Li, 2008, Peterson, 2012, You, Zhang, 2009) and increased financial misstatement risk (e.g., Filzen, Peterson, 2015, Hoitash, Hoitash, 2018) relative to firms with low FRC. Extant literature generally attributes these findings to high FRC levels reflecting the intentional choice of firm managers to obfuscate financial reports (e.g., Li, 2008, Lo, Ramos, Rogo, 2017).

Recently, some studies have argued that FRC primarily captures the complexity of firms’ business operations and accounting standards rather than the intention to obfuscate (Bushee, Gow, Taylor, 2018, Dyer, Lang, Stice-Lawrence, 2017, Guay, Samuels, Taylor, 2016). If high FRC is not an intentional choice, one could expect firms to seek to diminish the adverse effects of FRC on their information environment and financial reporting risk. However, there is little evidence on whether and how companies mitigate the negative effects of FRC. One exception is a study by Guay et al. (2016) documenting that firms increase voluntary disclosure as financial reporting complexity increases. Their findings are consistent with firms balancing complex mandatory financial reporting and voluntary disclosure to achieve an optimal information environment.

Our study extends this line of research by investigating whether firms invest in accounting expertise to mitigate the adverse effects of FRC. If high FRC is driven by the applicable accounting standards and regulations (Dyer et al., 2017), we expect companies to invest in accounting expertise to reduce the risk of negative reporting outcomes (e.g., restatements). If, on the other hand, high FRC arises from managers’ efforts to obfuscate financial reporting (Li, 2008, Lo, Ramos, Rogo, 2017), we would expect a negative or nonexistent relation between a firm’s level of accounting expertise and FRC. In other words, if FRC is used as a smokescreen, it is unlikely that companies would invest in expertise to counteract the obscuring effects of complexity. In contrast, we expect a positive relation between FRC and accounting expertise if high FRC is mainly driven by a combination of a firm’s business operations and accounting standards complexity. While Guay et al. (2016) examine voluntary disclosure as a channel to mitigate FRC, we focus on firms’ accounting expertise. An important difference is that increasing voluntary disclosure can be used to address transitory changes in annual report text (e.g., changes in sales, business acquisitions, or litigation), while increasing accounting expertise can be a response to permanent changes in firm complexity (e.g., changes in business operations and/or applicable accounting standards).

Since the absence of publicly available data makes it impossible to directly measure the level of accounting expertise in a given firm (e.g., expertise of the accounting department and outside consultants), we use the accounting expertise on its board of directors and audit committee as a proxy for a firm’s overall level of accounting expertise. This choice of accounting expertise measure introduces additional factors that may contribute to the difficulty in finding a relation between FRC and accounting expertise. Namely, we expect no relation between FRC and board accounting expertise, if FRC is mainly driven by business operations or transitory issues (e.g., industry-related trends or litigation). For firms with high operational complexity, the benefits of having boards with operational and industry expertise (Coles, Daniel, Naveen, 2008, Klein, 1998) might outweigh the benefits of adding more accounting expertise, considering that board size is limited (Cheng, 2008).² In addition, from a research-design standpoint, we expect no consistent relation between board accounting expertise and FRC if board accounting expertise does not reflect a firm’s overall investment in accounting expertise.

We conduct our analysis in three steps. First, we create a new measure of FRC based on the length of accounting standards and SEC disclosure regulations applicable to a firm’s annual report. Second, we use this measure to test whether FRC is related to the levels of accounting expertise on the board of directors and audit committee. Finally, we examine how FRC and accounting expertise jointly affect three negative financial reporting outcomes: internal control weaknesses, accounting restatements, and SEC comment letters. We expect positive relations between FRC and the incidences of these financial reporting problems; however, these relations should be comparatively weaker for firms with high levels of accounting expertise.

Prior studies have generally viewed FRC from the investors’ perspective, focusing on the textual complexity of annual 10-K reports. In contrast, we study FRC through the lens of a company that must apply complex accounting rules to prepare financial reports. We use the length of the accounting standards and regulations applicable to a firm’s annual report to measure FRC.³ Our FRC measure is based on the intuitive conjecture that accounting items with longer and more detailed disclosure standards are more complex (e.g., financial instruments, goodwill, pension liability). Consequently, financial reports with more complex accounting items have a comparatively higher FRC. Conceptually, our measure attempts to capture the amount of knowledge of accounting standards required to prepare a company’s annual report.

Unlike other measures of FRC in the extant literature (e.g., annual report length and readability), our measure captures the complexity of applicable accounting standards set by the FASB and the SEC; it is not derived from the annual report text generated by a firm’s management (and related parties). Compared to other measures of FRC based on 10-K text, our measure is less likely to be subject to management discretion, more likely to be exogenous with respect to any single corporate event, and more likely to capture persistent changes in complexity. Finally, it may be more difficult to mitigate complexity emerging from accounting standards than it is to mitigate complexity emerging from overly long and/or difficult-to-read textual disclosures (e.g., 10-K text can be simplified, shortened, and communicated more clearly; for more discussion on this topic, see Bonsall et al., 2017).

Our FRC measure is constructed at the firm-year level. First, we use the FASB’s eXtensible Business Reporting Language (XBRL) Financial Reporting Taxonomy to link each Generally Accepted Accounting Principles (GAAP) monetary accounting item (i.e., number) reported in a firm’s 10-K (e.g., inventory) to the relevant text in the FASB’s Accounting Standards Codification (ASC) and the SEC’s Regulation S-X that governs the disclosure of that item.⁴ Next, we count the number of words in the relevant ASC and Regulation S-X text to assign a complexity score to each item. Our firm-level FRC proxy is the sum of the complexity scores across all unique items reported in the firm’s 10-K report. Finally, in order to reflect the relative complexity between firms, we standardize the FRC measure by industry and year.

To calculate the extent of accounting expertise on the board and audit committee, we use the definitions of accounting experts from Cohen et al. (2014), Krishnan and Lee (2009), and DeFond et al. (2005). Specifically, we classify an individual as an accounting expert if he/she currently has or has had in the past at least one accounting qualification.⁵ We then use both the number and ratio of accounting experts on the board and audit committee as proxies for a firm’s accounting expertise.

We document that FRC is positively associated with accounting expertise on the board and audit committee. Turning to the negative consequences of FRC, we find that our FRC measure is positively associated with the incidence of internal control weaknesses disclosures and restatements, consistent with prior literature (e.g., Doyle, Ge, McVay, 2007, Hoitash, Hoitash, 2018, Peterson, 2012). We also find that the likelihood of receiving an SEC comment letter is comparatively high for firms with high FRC. Finally, we document that accounting expertise completely mitigates the negative effects of FRC on the likelihood of internal control problems and SEC comment letters, but not the likelihood of accounting restatements. Taken together, these results point to a strong association between adverse financial reporting outcomes and FRC. Nevertheless, high levels of accounting expertise on the board of directors and audit committee attenuate the adverse effects of FRC.

We acknowledge that a positive relation between FRC and accounting expertise may be attributed to firms acquiring accounting experts to advise management on strategic and operational matters, rather than to firms purposefully acquiring these experts to mitigate the risk of negative reporting outcomes. However, firms with high operational complexity may prioritize board members with specialized expertise (e.g., industry knowledge) over members with accounting expertise. Moreover, the board’s accounting expertise by itself (our proxy for the firm’s overall accounting expertise) may be insufficient to reduce financial reporting risk. If experts are added to the board for reasons other than FRC management and the firm does not increase internal accounting expertise, then it is unlikely that increases in accounting expertise on the board would be associated with the reduction of negative reporting outcomes that we document.

We confirm our inferences with several additional analyses. First, to mitigate a concern that our findings are driven by our choice of FRC measure, we examine the relation between complexity and accounting expertise using two alternative proxies for complexity, 10-K report length and readability index (e.g., Li, 2008, Miller, 2010). Using these proxies, we find similar, albeit weaker, results. Second, it is also possible that the association between FRC and a firm’s accounting expertise is driven by unobserved endogenous firm characteristics or short-term transitory changes in FRC not captured by our regression models. We attempt to isolate these effects by analyzing temporal changes in FRC. We find that persistent changes in FRC are related to changes in accounting expertise on the board of directors and audit committee. Third, our results suggest that FRC is driven by operational and accounting standards complexity (as opposed to managerial obfuscation) and that firms with high FRC enhance their governance by acquiring accounting expertise. If this premise is correct, we expect to see a stronger relation between FRC and boards’ accounting expertise for firms where external oversight demands better governance. We test this relation for low and high levels of institutional ownership, as institutional ownership are associated with enhanced monitoring and governance (e.g., Chen, Harford, Li, 2007, Chung, Zhang, 2011). We find that the relation between FRC and accounting expertise on the board of directors is magnified for firms with high levels of institutional ownership. Overall, our additional analyses provide increased confidence that our main results are not explained by a correlated omitted variable, since that variable would have to vary in time simultaneously with different measures of FRC, firm’s accounting expertise, and institutional ownership.

Our study contributes to the literature by demonstrating that firms reduce the adverse effects of FRC by investing in accounting expertise. Collectively, our findings indicate that firms use accounting expertise in addition to voluntary disclosure (Guay et al., 2016) to manage FRC. While increasing voluntary disclosure can be a potential response to transitory increases in the complexity of 10-K text, firm’s accounting expertise can be a response to permanent increases in the complexity of accounting rules. Moreover, our findings are consistent with FRC arising primarily from business and accounting standards complexity and not from intentional obfuscation. Further, our results highlight the role of accounting expertise in firms’ governance, and complement a growing literature on the determinants and consequences of board and audit committee expertise (e.g., Badolato, Donelson, Ege, 2014, Bryan, Liu, Tiras, 2004, Erkens, Bonner, 2013, Krishnan, Lee, 2009). Finally, our measure of financial reporting complexity directly captures the complexity of accounting standards and regulations that govern firms’ annual report disclosures. As such, it complements existing complexity measures (e.g., 10-K length and the Fog index, among others) in the prior literature.

The rest of the paper is organized as follows. In Section 2, we discuss the predicted associations between FRC, accounting expertise on the board and audit committee, and negative reporting outcomes. In Section 3, we describe the construction of our FRC and accounting expertise measures, as well as the design of our empirical analyses. We describe our data, sample selection, and descriptive statistics in Section 4. In Section 5, we present the results of our main analyses. In Section 6, we report additional analyses and robustness tests. We conclude in Section 7.