Principles-based accounting standards and audit outcomes: empirical evidence

Enron engaged in financial engineering, such as special purpose entities (SPEs) and limited partnerships, to avoid reporting debts on its balance sheet. Enron also recognized profits from sales of assets to SPEs (Thomas 2002). Benston et al. (2006) state that while Enron’s accounting met the technical requirements of the (rules-based) Generally Accepted Accounting Principles (GAAP), it violated the intent of GAAP. The implication is that financial engineering would be less successful under a principles-based regime. The SEC called for a study on the feasibility and economic analysis of adopting standards that are principles-based or objectives-oriented (SEC 2003). Maines et al. (2003) and SEC (2003) discuss benefits and costs of principles- vs. rules-based standards.

The FASB’s several current projects are consistent with the growing interest in the principles-based approach, such as revisions to the Conceptual Framework undertaken jointly with the International Accounting Standards Board (IASB), as well as revisions of accounting for goodwill (Accounting Standards Codification, ASC hereafter, 350), stock-based compensation (ASC 718), revenue recognition (ASC 606), and leases (ASC 842).

Folsom et al. (2017) use earnings informativeness, predictive value of earnings for next period cash flows, and earnings persistence to proxy for financial reporting quality.

Ling (2018) reports that principles-based standards are negatively related to audit fees using S&P 500 firms. Our study differs from Ling (2018) in significant ways. We discuss the differences in detail in the next section.

Dichev et al. (2013) find that high-quality accounting standards are an important element of high financial reporting quality. They also report that CFOs rate accounting standards as the second most influential factor in earnings quality, after a firm’s business model.

While the use of professional judgment by managers is expected to have a favorable effect on financial reporting quality, it may demand higher audit effort as a result of increased inherent risk arising from intentional or unintentional misapplication of the underlying principles under principles-based standards. We follow prior research and use audit fees to proxy for audit effort (Srinidhi and Gul 2007; Rice and Weber 2012; Lobo and Zhao 2013).

Echoing this view, Leslie Seidman, former chair of the FASB states, “Precise guidance is necessary in the U.S. which has a more litigious culture. The U.S. financial reporting system can’t function over the long run with accounting standards that provide only principles” (Tysiac 2012). While the AICPA, Deloitte & Touche, Grant Thornton, and KPMG were in favor of the FASB’s (2002) proposal for principles-based standards, BDO Seidman opposed stripping away implementation guidance and placing primary reliance on broad principles. Microsoft supported the proposal and said that “the amount of interpretive and implementation guidance in accounting standards to ensure comparability is the biggest culprit in driving much of the detail and complexity in accounting standards.” (Comment letters are available at https://​www.​fasb.​org/​jsp/​FASB/​CommentLetter_​C/​CommentLetterPag​e?​project_​id=​1125-001).

We describe PSCORE in detail in Section 3.

As a robustness test, we re-estimate the audit fee model on a sample matched on the firms’ propensity to rely on principles-based standards and find that client firms that rely more on principles-based standards (top two quintiles of PSCORE) pay about 25.0% less in audit fees (about $160,050) than client firms that rely less on principles-based standards (bottom two quintiles of PSCORE). Our results are also robust to controlling for firm fixed-effects, to a “change-on-change” model that controls for time-invariant confounds, and to another matched sample based on firm size. We also conduct a path analysis and find that principles-based standards impact audit fees directly as well as indirectly through financial reporting quality while the direct effect is larger.

Trompeter (1994) finds that audit partners are more likely to require income-decreasing adjustments when accounting standards are more specific (rules-based), suggesting less aggressive reporting by managers with rules-based accounting than principles-based accounting. However, Agoglia et al. (2011) find that CFOs are more likely to capitalize leases, i.e., less likely to engage in aggressive financial reporting, when applying a less precise (more principles-based) lease classification criterion than when applying a more precise (more rules-based) criterion.

Jamal and Tan (2010) find that under a rules-based standard, auditor-type (principles-oriented, rules-oriented, or client-oriented) does not influence the propensity to report the lease as an off-balance sheet (operating) lease. However, under a principles-based standard, the propensity to report the lease as operating is lowest when the auditor is principles-oriented.

Doogar et al. (2010) find, on average, audit fees are lower under AS5 among accelerated filers of Big 4 auditors. Sin et al. (2015) examine both Big 4 and non-Big 4 clients and report two findings. First, under AS2, non-Big 4 clients are more likely to disclose material weaknesses in their internal controls than Big 4 clients. One reason for this result is that riskier clients have moved from the Big 4 to the non-Big 4 after the SOX of 2002. Second, after the transition to AS5, the gap in the likelihood of internal control weakness disclosure narrows between the Big 4 and non-Big 4 auditors.

McEnroe and Sullivan (2012) survey the largest accounting firms and CFOs of Fortune 1000 firms. They report that respondents’ perception does not significantly differ between principles-based and rules-based standards with regard to reliability, timeliness, and faithful representation. Auditors and CFOs rank principles-based standards higher than rules-based standards on relevance, professional judgment, understandability and cost-effectiveness. However, rules-based standards are ranked higher on verifiability, comparability, neutrality, and consistency.

Our study differs from Ling (2018) in multiple ways. While she focuses on larger firms, we do not impose restrictions on firm size. More importantly, she does not examine the relation between principles-based standards and going concern opinion, audit lag, or audit risk. Also, we conduct path analysis to examine the relations between principles-based standards and audit fees mediated by financial reporting quality.

We are aware of prior research that finds an increase in audit fees after IFRS adoption by European and Australian firms (Kim et al. 2012; De George et al. 2013). These findings are consistent with greater exposure to audit complexity as a result of transition to IFRS. While IFRS is perceived as more principles-based than the US GAAP, their settings differ from our context in several ways. First, Folsom et al. (2017) state that there is considerable cross-sectional variation of rules-based vs. principles-based standards in the US GAAP. However, it is not clear whether the switch to IFRS from local GAAP provides sufficient variation in the rules-based nature of the standards to empirically discern the relation between reliance on principles-based standards and audit fees. Second, Folsom et al. (2017) also note that IFRS adoption could fundamentally change how firms account for transactions. The fundamental change in accounting practice could impact audit pricing in ways that are not necessarily related to the more principles-based nature of IFRS. Finally, the significant differences in information and legal environments between the United States and the other countries limit the relevance of findings from multi-country research to the US setting.

For example, clients could take the posture, “Show me where it says I can’t” (Agoglia et al. 2011).

While the US GAAP contains several bright-line rules, such rules are seldom used in auditing standards. The outcomes of the auditing process are an opinion guided by the auditor’s professional judgment of audit evidence. Thus, the maintained assumption in our analyses is that the nature of the auditing standards is fairly constant, i.e., principles-based, consistent with the IAASB’s view (ICAEW 2006, 20).

Charles et al. (2010) report a positive relation between audit risk and audit fees. In addition, prior studies find that greater risk of earnings management (lower financial reporting quality), especially via income-increasing accruals, is associated with higher audit fees (Gul et al. 2003; Abbott et al. 2006).

Process accountability is the expectation that one must justify one’s beliefs, feelings, and actions to others, regardless the outcome of the actions (Lerner and Tetlock 1999). Process accountability is an important issue, especially under principles-based standards, because auditors need to have a deeper understanding of not only the accounting treatment proposed by the client but also the relevant accounting principles in the standards. Peecher et al. (2013) note that auditors are accountable, especially to the PCAOB, the SEC, and others, for audit outcomes (outcome accountability) and the quality of their professional judgment process (process accountability).

On the other hand, low epistemic motivation can result in “closing of the mind” to new information (Kruglanski 1989).

This is consistent with prior research that finds that process accountability mitigates auditor judgment biases and increases audit effort (Ashton et al. 1992; Kennedy 1993, 1995).

To address the concern that transaction complexity could be related to PSCORE, Folsom et al. (2017) orthogonalize RBC1 to two measures of complexity (the number of times the FASB describes the underlying transaction as complex and the number of words the firm takes to define the underlying transaction) and rescale this orthogonalized RBC1 to be between 0 and 1 before calculating PSCORE.

The keywords used to determine the impact of a standard on a firm have been reviewed by the Big 4 accounting firms’ national office staff, who have technical expertise in the standards they review. Donelson et al. (2012) perform three validation tests. First, they report that the four dimensions of RBC1 are all positively correlated, and a factor analysis reveals that the four attributes proxy for a common construct. Second, they find that RBC1 scores are lower for the standards classified as principles-based by the SEC than for the standards classified as rules-based by the SEC. Third, the RBC1 scores for the IFRS standards are lower than or equal to the RBC1 scores for the corresponding US GAAP standards, consistent with the perception that IFRS standards are less rules-based than the US GAAP. Folsom et al. (2017) perform two additional validation tests. First, they count the number of times each firm mentions the relevant keywords in its annual 10-K and find that, while keywords for industry-specific standards are rarely mentioned for firms that do not operate in that industry, they are frequently mentioned for firms that do operate in that industry, indicating the influence of each standard on a firm’s financial reports. Second, they calculate the correlation between the number of keyword hits and the dollar magnitude of the corresponding line item in the financial statements and find these correlations to be economically and statistically significant. Collectively, these validation tests provide some assurance that PSCORE is a reasonable measure of a firm’s reliance on principles-based standards.

We collect the securities class action litigation cases from the Stanford Securities Class Action Clearinghouse. These lawsuits were brought against the firm or the managers or both under Rule 10b-5 of the Securities Exchange Act of 1934 on the grounds that managers made misleading statements and/or failed to disclose adverse material information to the market in a timely manner (Billings and Cedergren 2015). There is also evidence that poor accounting quality is a determinant of litigation (Henninger 2001; Palmrose and Scholz 2004). Thus, the class action litigations suggest higher audit risk.

We drop firm-year observations with less than 200 trading days for RET and TURNOVER, following Krishnan and Zhang (2005).

The results remain qualitatively unchanged with the first two years of engagement for the indicator of initial year.

The larger the proportion of receivables and inventories (INVREC), the more likely it is that the clients have uncollectible accounts or inventory impairments, increasing audit risk. Similarly, lower profitability (ROA), lower liquidity (LIQUID), higher leverage (LEV), and loss (LOSS) are related to higher audit risk.

We obtain similar results when we replace earnings report lag with audit report lag.

PSCORE data are available from Rick Mergenthaler’s website: https://​www.​biz.​uiowa.​edu/​faculty/​rmergenthaler/​. Folsom et al. (2017) employ a set of keywords to determine the impact of each standard on a firm. These keywords were reviewed by the national office of a Big 4 accounting firms, and the review ended in 2006.

Audit fee data are available from 2000.

The untabulated results are similarly negative if we replace PSCORE with its quintile rank measure, PSCORERK. The coefficients of PSCORERK with AAER, LITIRISK, and KZRSK are −0.0003, −0.066, and − 0.001, respectively, all significant at the 0.01 level. The coefficient of PSCORERK with SUED is negative (−0.001) but insignificant.

We compute the decrease in audit fees as 20.5% (e^-0.2294-1), where −0.2294 is the coefficient on the indicator variable representing high reliance on principles-based standards observations. We then multiply the decrease of 20.5% by the mean audit fees of $1,380,389 based on the four quintiles (untabulated), which equals $282,965. When we replicate this regression of audit fees on HIPSRANK using the propensity score matched sample, the coefficient of HIPSRANK becomes −0.2881 and the decrease in audit fees becomes 25.03% (e^-0.2881-1). When we multiply the decrease of 25.03% by the mean audit fees of the matched sample (untabulated), $639,400, the dollar representation equals $160,050 for the sample.

We exclude observations with auditor changes from the previous year to the current year.

Following Picconi and Reynolds (2013), we estimate the audit fee model by year after partitioning the sample into quintiles based on firm size and our results are consistent with the results in Table 4.

We also estimate the changes in audit fee model using the propensity score matched sample and obtain consistent results.

We thank the anonymous reviewer for this suggestion.

We estimate the Dechow and Dichev (2002) measure following Francis et al. (2005).

When we use PSCORERK, the coefficient on PSCORERK is also negative and significant at the 0.10 level for all three measures of financial distress, as predicted.