Keshk, Walied; Lu, HungYuan Richard; Mande, Vivek (2020). How have US banks adopted the Financial Accounting Standards
Keshk, Walied; Lu, Hung‐Yuan Richard; Mande, Vivek (2020). How have US banks adopted the Financial Accounting Standards Board's Level 3 fair value disclosure rules? Accounting & Finance 60, April Supplement S1, 693-727.
1. Read from the beginning until page 701 (hypothesis 3), and the conclusion on page 720. Summarize in your own words (about 100 words).
2. Explain the difference between Levels 1, 2, and 3. Keshk et al (2020) categorize levels 1,2 together, so you are free to do so but are not required. You may cite and reference other sources to answer this question.
3. If you were an auditor, which type of estimate requires more work to audit Levels 1,2 or Level 3? Explain.
This study examines how US banks have complied with the Financial Accounting Standards Board’s (FASB’s) requirements on Level 3 fair value disclosures. There has been a rise in the use of fair values in financial reporting during the past decade. The FASB’s Accounting Standards Codification (ASC) 820, Fair Value Measurements and Disclosures, however, provides only principle-based guidance on how fair values should be measured using internal and external information. Under ASC 820, fair values of assets/liabilities are classified into Levels 1, 2 and 3 of the fair value hierarchy.
While Levels 1 and 2 fair values rely on market-based inputs, Level 3 measurements require model-based inputs. As such, Level 3 valuations are regarded as highly subjective and less reliable than other fair values. Firms are, therefore, required to provide more detailed information about the inputs used to calculate Level 3 fair values (ASC 820-10-50-8) and describe the processes in place for obtaining and validating these measurements (ASC 820-10-55-105). To increase their usefulness to investors, quantitative information about the valuation inputs must be presented in a tabular format (ASC 820-10-50-2bbb). In response to users’ concerns that the variability arising from these measurements can have a big impact on financial statements, firms must also disclose qualitatively the sensitivities of fair value measurements to changes in inputs (ASC 820-10-50-2 g). Additionally, to assist users, firms must prepare a table that reconciles the beginning balance of fair values to their ending balances by each level in the hierarchy (ASC 820-10-50-2c). When firms use third-party specialists for developing estimates of fair values, ASC 820 requires a description of the processes in place for evaluating and validating the valuation methodologies and the inputs provided by outside specialists (ASC 820-10-55-105). Compliance with the above rules, however, has proved to be a challenge for many firms (e.g. Ernst & Young 2012). The Securities and Exchange Commission (SEC) routinely issues comment letters pointing out instances of noncompliance and deficiencies in both fair value measurements and related disclosures.1 Significant deficiencies have also been found by the PCAOB in the audits of fair value measurements and disclosures (e.g. Acuitas, Inc., 2016). In many instances, these deficiencies occurred because effective internal controls were not in place in these firms (Ettredge et al., 2011; Acuitas, Inc., 2016). This study examines two research questions.2 First, we investigate the extent of noncompliance with Level 3 fair value disclosures provided by banks about their recurring fair value measurements. We focus on the banking industry because banks hold more financial instruments measured at fair values than do firms in most other industries (Lu and Mande, 2014a). Second, we test whether there are systematic differences between compliant and noncompliant banks. Because fair value measurements and disclosures are complex, we expect that compliance will be a challenge for smaller banks. The work of auditors is also paramount in ensuring that client firms are compliant with fair value disclosure rules. Therefore, we expect compliance in banks to increase as audit quality increases. As audit committees are responsible for the oversight of financial statements, we hypothesise that banks having effective audit committees will be more compliant with financial reporting rules. Finally, we predict that institutional investors, who often are the primary users of fair value information, will be more likely to monitor banks’ disclosure behaviour. Our results show that there is widespread noncompliance with the basic disclosure requirements of ASC 820. Over 40 percent of banks do not disclose the valuation inputs table. Additionally, a majority of banks do not fully discuss the processes used for validating fair value estimates, whether developed internally or obtained from third-party specialists. A majority of banks also do not provide disclosures related to the sensitivities of fair values to changes in valuation inputs. Our results also show that, in support of our hypotheses, noncompliant banks are smaller, have less effective audit committees and internal controls, are associated with lower audit quality and have lower levels of institutional monitoring. By documenting the extent of noncompliance, this study sheds light on the complexity and cost burden associated with fair value reporting mandates. Our results should help audit committees, auditors, regulators and other capital market participants identify more easily banks that are likely to be in violation of fair value reporting rules. We believe that our findings will also be useful to the FASB which is currently engaged in an ‘improvements project’ for simplifying fair value disclosure rules.3 Finally, our theory and findings should be relevant to regulators in the United States, Australia and other countries that have fair value disclosure requirements.4 The rest of this study is organised as follows. Section 2 reviews the literature on disclosure quality and develops our hypotheses. Section 3 presents our sample and empirical model, followed by a discussion of the results in Section 4. Section 5 then concludes the study. 2. Literature review and hypotheses development 2.1. Factors influencing disclosure quality Prior work has extensively examined factors influencing the quantity and type of voluntary disclosures provided by firms to investors. Studies show that firms voluntarily provide additional information when they desire access to capital markets. Firms providing superior disclosures reap benefits in the form of higher share prices, lower bid-ask spreads and lower costs of equity and debt capital (Welker, 1995; Botosan, 1997; Sengupta, 1998). Related literature finds that managers voluntarily disclose good or bad news in a strategic fashion. For example, Einhorn (2007) argues that managers strategically hide good news from their investors in order to lower stock prices prior to option grants. There have been, however, fewer studies examining factors affecting disclosures that are mandatory. Schwartz and Soo (1996) document widespread noncompliance with disclosures about auditor changes in Forms 8-K, especially when the departing auditor is a Non-Big N firm. Ettredge et al. (2011) find that noncompliance in disclosures about auditor changes is positively related to low quality corporate governance. Chen et al. (2015) document noncompliance with disclosure requirements relating to litigation losses mandated under ASC Topic 450, but when the firms are audited by an industry specialist, the authors find that compliance with GAAP rules increases. Research using Australian firms shows similar evidence of noncompliance with mandatory reporting requirements. Mayorga and Sidhu (2012) examine how the largest 20 Australian listed firms comply with disclosure requirements relating to major assumptions and estimation uncertainties associated with the valuation of a firm’s assets and liabilities. Their results show that firms do not fully comply with these requirements, thereby limiting the informativeness of the disclosures. Using a sample of large Australian firms, Gallery et al. (2008) document sizable variation in the quality of mandatory predisclosures about the effects of adopting International Financial Reporting Standards (AIFRS). Among 50 large Australian firms with goodwill on their books, Carlin and Finch (2010) find that more than 10 percent of them failed to provide the discount rate used for estimating the fair value of goodwill. Failure to provide mandatory disclosures can have significant negative consequences for firms in the form of penalties, lawsuits and stock price declines (Ettredge et al., 2011). Although we should not expect that firms will behave strategically in their noncompliance, there is some evidence that they do. Alexander et al. (2011), for example, find that firms do not disclose the effects of a new accounting standard under Staff Accounting Bulletin (SAB) 74 if the disclosures reveal bad news. Along the same lines, Ettredge et al. (2011) find that firms behave opportunistically with regard to disclosing bad news about auditor changes noting that it is ‘surprising that managers are willing to risk the disapproval of the SEC by partial compliance with 8-K regulation’. 2.2. Noncompliance with fair value disclosure requirements There is ample anecdotal evidence that firms face challenges in implementing fair value reporting requirements. Ernst & Young (EY) (2012), for example, document pervasive noncompliance with fair value disclosures concerning: presentation of quantitative Level 3 inputs, disclosures of those inputs and valuation techniques, levels of disaggregation at which fair value information is presented and inputs developed by third parties. In its study, EY surveys Level 3 fair value disclosures of 60 companies in 2012. However, the small size of the sample (only 14 banks are included) limits their study’s usefulness. EY also does not examine what firm characteristics are associated with fair value noncompliance. SEC comment letters also provide evidence of noncompliance with fair value disclosure rules. The SEC’s focus has mainly been on disclosures about Level 3 inputs mandated by ASU 2011-04 because of the high degree of judgment required from management in developing these inputs. The recent comment letters have requested that companies increase ‘granularity and transparency’ of disclosures about Level 3 inputs and provide additional information where the disclosures are incomplete or omitted.5 Additionally, the role of auditors in clients’ noncompliance with fair value rules has come under PCAOB scrutiny. In its inspections, the PCAOB has found that auditors routinely fail to adequately test clients’ controls around fair value disclosures and measurements, fail to identify weaknesses when qualified personnel are not used in the valuations, place too much reliance on third-party work and fail to adequately test the classification in the fair value hierarchy.6 There have only been a few academic studies examining noncompliance with fair value rules. Lu and Mande (2014a) examine banks’ compliance with one requirement of the FASB’s ASU 2010-06–Improving Disclosures about Fair Value Measurements. Prior to ASU 2010-06, firms were allowed to present fair value information in the notes by category of investments. ASU 2010-06, however, requires presentation by class of investments which involves a finer partitioning of fair value information.7 Using data from the first quarter of 2010, Lu and Mande report that 23 percent of the banks in their sample failed to provide fair value information disaggregated by class of investment. In a related study, Lu and Mande (2014b) find that fair value information is more value relevant when it is disaggregated by class of investment. In contrast to Lu and Mande, we mainly examine banks’ compliance with Level 3 fair value disclosure requirements included in the FASB’s ASU 2011- 04–Amendments to Achieve Common Fair Value Measurement and Disclosure Requirements in U.S. GAAP and IFRSs. Disclosures mandated by ASU 2011- 04 are significantly more complex than those mandated by ASU 2010-06 (see Appendix I for a comparison). Compliance with ASU 2010-06 in Lu and Mande’s study, only required firms to present information that was readily available to them, albeit in a ‘finer’ fashion. Not surprisingly, within a year of ASU 2010-06 going into effect, virtually all banks were fully compliant with the requirement to disaggregate fair values.8 Therefore, Lu and Mande (2014a) document that there was only initial noncompliance with one of the requirements of ASU 2010-06. The reporting requirements of ASU 2011-04, however, are significantly more onerous for firms. Not only does it require significant changes in how data are organised (e.g. presentation in tables), but firms are tasked with the collection of new data (e.g. information about valuation inputs from third parties) and are required to perform new analyses (e.g. sensitivity of changes in fair values to changes in inputs).
Our study examines compliance with Level 3 fair value disclosures provided by banks about their recurring measurements of financial instruments. Level 3 valuations are considered to be the most subjective and complex fair value measurements provided by firms on their financial statements. To assist users, the FASB, therefore, requires firms to provide detailed information about inputs and the validation of the fair value measurements. Furthermore, all quantitative information about Level 3 fair value measurements must be presented in a tabular format. The contribution of our study is twofold. First, we document the extent of noncompliance with fair value disclosure rules. Second, we study whether noncompliant banks are smaller and associated with lower levels of monitoring. Our results show that banks are compliant in providing fair values by classes of investments. However, there is widespread noncompliance with other basic fair value disclosure rules. Over 40 percent of banks do not provide an ‘inputs’ table while a majority of banks do not fully discuss the sensitivity of fair value estimates or how these estimates, whether developed internally or through third party specialists, are validated. These findings should concern the FASB because they show the presence of noncompliance with ‘first order’ requirements of ASC 820. We also noticed anecdotally other forms of noncompliance with regard to the information presented in the tables. The rates of noncompliance, therefore, are higher than those documented in this study. We also find that smaller banks face greater challenges in complying with fair value disclosure rules. Firms that have ineffective internal controls and lower levels of institutional monitoring are also less compliant. Finally, this study shows, albeit weakly, that effective audit committees and higher audit quality are associated with greater compliance. Our results should be of use to the SEC and PCAOB who have a great interest in ensuring compliance with the requirements of ASC 820. Our study should also benefit the FASB and other standard setters as they continue their work on improving the quality of fair value disclosures provided to investors.