What exactly do we mean by audit quality?

1. Introduction

Nothing is arguably more important in auditing practice, regulation, and research than the concept of audit quality. Yet, despite its centrality to the understanding of auditing, defining it has proved elusive, and there is little consensus on the concept or the measurement of audit quality. I review several perspectives on audit quality, and discuss at length why audit outcomes (audit reports and the quality of the client’s audited financial statements) are the most relevant basis for assessing audit quality. I conclude with some ideas on how auditors might move beyond the limitations of the current binary (pass/fail) audit report to convey more information about the quality of a client’s audited financial statements.

Three groups have an active interest in audit quality: regulators, audit firms, and audit scholars.¹ Regulators and audit firms appear to view audit quality largely in black and white terms. The audit is ‘OK’ if you follow generally acceptable auditing standards and the specific standards of the International Auditing and Assurance Standards Board (IAASB), and not OK if you do not. At least this seems to be the logic of audit inspection report comments wherein audit quality is deficient (not OK), in the opinion of the inspectors, if the evidence gathering process is deficient. It is a procedural or process way of thinking about quality. Audit quality is good if the process is good, i.e. if you follow the standards.² However, regulators also take a broader view as they are concerned with an audit firm’s internal control structure for ensuring that an audit engagement complies with the firm’s policies (and auditing standards) for gathering sufficient audit evidence. Academic researchers are also interested in audit quality. Some scholars focus on the audit process and study it through experiments using the ‘judgment and decision making’ (JDM) paradigm. Other scholars do empirical research and study observable audit outcomes, namely, the audit report and the quality of client’s audited financial statements.

2. A framework for understanding audit quality

Francis (2011, Table 1) provides a starting point for thinking about the elements that create high-quality audits. It uses different units of analysis from micro (inputs) to macro (outcomes) as a way to think about the drivers of audit quality, and is similar to later frameworks by the IAASB (2014) and the Public Company Accounting Oversight Board (PCAOB) (2015).

Table 1. Units of Analysis in Assessing Audit Quality (Adapted from Francis, 2011).

Audit Inputs Audit tests Engagement team personnel

Audit Process

Implementation of audit tests by engagement team personnel, including reliance on internal auditors and other external auditors

Audit Firms

Engagement teams work in audit firms Audit firms hire, train, incentivize and compensate auditors, and develop audit guidance (testing procedures) Audit reports are issued in name of audit firms

Institutions^21

Institutions affect auditing and incentives for quality such as legal regimes and litigation, licensing bodies (State Boards of Accountancy), organizations like the AICPA, FASB, SEC, PCAOB, and governance structures (audit committees)

Audit Outcomes and their Economic Consequences^22

Audit outcomes and effects on clients and users of audited accounting information

Francis (2011, p. 125) summarizes the framework as follows:

Audit quality is affected at each level of analysis in Table 1. Audits are of higher quality at the input level when the people implementing audit tests are competent and independent, and when the testing procedures used are capable of producing reliable and relevant evidence. The quality of audit inputs flows through to the audit process, where audits are of higher quality when the engagement team personnel make good decisions regarding the specific tests to be implemented and appropriately evaluate the evidence from these tests in leading to the audit report. Audit quality is affected by the accounting firm in which the auditors work. Firms develop the testing procedures used on audit engagements, and create incentives that affect the behavior of engagement team personnel. Last, the incentives of accounting firms and individual auditors to produce high-quality audits are affected by the institutions that regulate auditing and punish auditors and accounting firms for misconduct and low-quality audits.

Francis (2011) concludes that audit quality is not a singular concept or measurement. Instead, it is better understood as a multi-dimensional concept with multiple attributes that contribute to audit quality.³ A similar view is reflected in recent suggestions by regulators for measuring and publicly disclosing indicators of audit quality (PCAOB, 2015; Financial Reporting Council, 2020).⁴

3. Is audit quality equivalent to ‘audit process’ quality?

An important element of the framework in Francis (2011) is the quality of evidence-gathering in the audit process. As noted above, firms and regulators largely focus on the audit process when they currently think about audit quality in the inspections of engagements. Did the auditors follow audit standards? Did they collect sufficient evidence? Did they interpret the evidence correctly? Adhering to standards in the audit process is important, and the presumption is that ‘audit process’ deficiencies will result in low-quality audits.

There are several limitations to the ‘audit process is audit quality’ viewpoint. First, audit standards provide general guidance for the conduct of an audit. While there are specific procedures that auditors are required to do, audit standards are best described as largely principles-based with relatively few ‘bright line’ rules to be followed, which gives auditors considerable flexibility in structuring audit testing and the evidence-gathering process.⁵ This limits the effectiveness of inspection reviews of ‘the audit process’ as a way to infer audit quality. Simply put, there is not a single correct way to gather evidence. This makes it especially problematic for inspectors to ex-post assess the ‘quality’ of the audit process and can easily lead to disagreements between auditors and regulators about the reasonableness (quality) of the work done by the auditor. In other words, given that auditing standards are largely principles-based, and given the absence of comprehensive and detailed standards for the entire audit process, the quality of the evidence-gathering process becomes a matter of considerable subjective judgment, for both the auditor and the regulator.⁶

Second, while audit standards require that auditors must collect sufficient evidence to support the audit opinion, there is no clear stopping-rule on how much evidence is needed to support the audit opinion. Since audit standards are largely principles-based, they give the auditor a great deal of flexibility in making this assessment. Again, this can lead to disagreements between regulators and auditors about the quality and quantity of the evidence collected.

Third, although the presumed linkage of audit process deficiencies with low-quality audit outcomes (audit failures) is logical, empirically it is more a matter of faith than established fact. In short, we really do not know exactly how the quality of the audit process maps to the quality of audit outcomes (i.e. the audit report and the quality of the audited financial statements).⁷

For all of the above reason, the ‘audit process is audit quality’ viewpoint is problematic and has serious limitations in terms of its potential ability to tell us something about the bigger issue, which is the quality of the audit report.

4. Binary model of audit quality

Audit quality is admittedly difficult to assess, and the framework in Francis (2011) is useful in thinking about the multiple drivers of audit quality. There is, however, a simple and straightforward way of thinking about audit quality. It starts with a basic question: what is it that auditors are required to do? What auditors are required to do by law is to opine on the client’s financial statements in an audit report. So, when we talk about audit quality, what we are really talking about is the quality of the audit report, and, by implication, the quality of the audited financial statements on which the audit report opines. This corresponds to ‘audit outcomes’ in the framework in Table 1. While complying with audit standards is necessary, it does not get to the heart of what audit quality is really all about. It is the quality of the audit report that matters, not the audit process, per se.

My thesis is that audit quality is ultimately about the quality of the client’s audited financial statements on which the auditor opines in the audit report. In most audits, the auditor issues a standard clean opinion, and this is probably the ‘correct’ audit report in a binary sense, i.e. clients do not fail or engage in fraudulent behavior. However, there are two ways in which there are arguably low-quality audits, or what can be termed audit report failures. The first potential audit report failure occurs when the auditor issues a standard clean opinion but there is an undetected or unreported material fraud or misstatement in the audited financial statements. These cases typically make headlines and raise public doubts about the quality of audits.

The second potential audit report failure occurs when the auditor issues a standard clean opinion, but the client subsequently fails or files for bankruptcy within 12 months of the audit report date. This sometimes called a Type 2 going concern (GC) reporting error, failing to qualify when warranted. Audit standards require auditors to assess if the client firm is a viable on-going entity and to issue a warning in the form of a ‘going concern’ audit report if there is substantial doubt the firm can continue 12 months beyond the audit report date.

We can characterize the above two cases as audit report failures. In both cases, audit report quality is binary in nature: the auditor either did or did not issue the ‘correct’ audit report. However, it turns out things are not so black and white, even in these instances. Not all undetected frauds/misstatements are deemed to be audit report failures. The reason is that audits are not a 100 percent guarantee of correctness, and an auditor can in good faith have followed appropriate auditing standards and still failed to detect client fraud and/or material misstatements.⁸

While auditors try to control and minimize their audit report risk, the fact is that audits are inherently risky and cannot achieve certainty or zero errors. A definitive audit report failure occurs only if the auditor has been negligent and grossly deficient in applying audit standards and normal testing procedures. Consistent with this, in SEC Enforcement Actions against companies in the United States for misleading reporting, the auditors of such companies are rarely named as co-defendants in SEC Enforcement Actions. More generally, the proven instances of audit report failures are quite small. Based on litigation and SEC Actions, audit report failures in which the auditor is clearly culpable are less than one percent of audits, at least in the United States.⁹

A similar situation exists with respect to going concern audit reports. The auditor does not have a crystal ball and cannot predict the future with certainty. In some audits, it will be clear that the client was in the danger zone and that a going concern warning should have been issued. Failure to do so is probably an audit report failure. But, in other cases, the warning signals are not so clear in which case the auditor is unlikely to be judged guilty of failing to issue a going concern audit report. Even in the case of business failures (which are not common), Carcello and Palmrose (1994) find that the auditor is sued by investors less than 50 percent of the time when they had issued a standard clean opinion instead of a going concern report.¹⁰

To conclude, most audit reports are standard clean opinions, and are probably the correct report in a binary sense. Potential audit report failures occur when there are unreported/undetected client frauds and misstatements, and when the client fails within 12 months of the audit report. However, we see that even in these cases, there is no clear-cut auditor negligence, hence there is not necessarily a reporting failure for which the auditor is culpable. While the binary model of audit report quality seems intuitive, it is actually quite difficult to determine if there is an audit report failure, given that audits are not guarantees of certainty. There is an inherent residual risk that the ‘wrong’ audit report might be being issued, even though the auditor has conducted the audit in an acceptable manner.¹¹

5. Audit report quality: 50 shades of grey?

Summarizing the above discussion, the black and white (binary) model of audit quality report is not particularly useful in delineating good and bad audits, even when there is client fraud or bankruptcy risk. Most audit reports are standard clean opinions, and technically these are the ‘correct’ report. However, we know that there is wide variation in the quality of audited financial statements, particularly the quality of audited earnings, yet all of these financial statements receive the same standard clean audit report.¹² There is a spectrum or continuum of quality in audited financial statements because of the managerial discretion, choices, judgments, and accrual estimations that exist in applying accounting standards. The binary pass/fail audit report treats all firms on the spectrum as if they are of the same quality, yet we know they are not the same.

A large body of research investigates the quality of audited earnings. Nissim (2022) provides a review of this literature. Briefly, research finds that earnings are of higher quality if they are more persistent (sustainable), more predictive of future earnings and future cash flows, if they have lower levels of unexpected (abnormal) accruals, are less likely to be restated in subsequent periods, are not aggressively managed to meet benchmark earnings targets (such as analysts’ forecasts) and have greater value-relevance for stock prices (for listed firms). More generally, firms with high levels of accruals are more likely to fail, more likely to have material misstatements and frauds, and are more likely to have restatements and regulator sanctions (Dechow et al., 1996; Dechow et al., 2010).

The presumption I make is that the auditor should play a crucial role in monitoring the quality of the accounting policy choices and accrual estimates made by the client in applying accounting standards – even if these policies and estimates do not technically violate accounting standards.

Arthur Levitt, who was Chief Accountant of the Securities and Exchange Commission in the United States from 1993 to 2001, was highly critical of firms using their discretion to aggressively manage reported earnings to meet earnings targets such as analysts’ forecasts. In a widely cited speech titled ‘The Numbers Game’ he argued that even though firms might be in technical compliance with accounting standards, at some point aggressively managed earnings become misleading to investors about the true performance of firms and may be tantamount to fraud (Levitt, 1998).

A ‘good’ auditor will constrain overly aggressive accounting choices by a client, while a ‘lax’ auditor might not.¹³ These judgments by auditors are of fundamental importance in shaping the quality of the client’s audited financial statements. While following the standards in the audit process is required, even more important is the role of the auditor in monitoring the reasonableness of the client’s accounting policy choices and accrual estimates, and, where needed, putting on the brakes and constraining aggressive choices made by the client.¹⁴

This discussion gives rise to the idea of using financial statement quality, especially earnings quality, as a way ‘to infer’ the quality of the auditor and the audit report. The PCAOB (2015, p. 13) suggests as much in their discussion of audit quality indicators. In other words, financial statement quality is a spectrum, fifty shades of grey, rather than a black and white binary model of correctness.¹⁵ The same is true of quality on audit engagements: it is not black and white but rather a spectrum of quality.

This idea also underlies archival research on audit quality, in which certain auditor attributes are expected a priori to be associated with better quality audits (e.g. auditor industry expertise), and are tested for their association with the spectrum of financial statement quality. Better auditors are expected to be positively associated with better earnings quality. The implication is that a continuum measure of audit report quality can be inferred from the continuum of the quality of audited financial statements. Empirically, a large body of research in auditing supports this linkage between auditor attributes that predict auditor quality, such as audit firm size or industry expertise, and the continuum of earnings quality (Francis, 2023).

6. More on earnings and accruals

It is important to elaborate on the importance of accrual estimates in the financial statements, and why they should play a central role in the auditor’s assessment of the client. The client’s financial statement earnings numbers are a combination of relatively straightforward factual data such as the cash flows from the purchase and sale of inventory, and more complex estimations of items which are called ‘accruals.’ A simple example of an accrual is a sale made on credit, with the payment to be received at some point in the future. The sale is recognized before cash is collected. Then, at the end of the fiscal year, the accounting system makes an accrual estimate of the balance of credit sales that is unlikely to be collected, called a bad debt expense.

Most accruals are more complex, such as the estimation of pension expense for a defined benefit pension plan linked to future salaries; required fair market valuation of assets (or liabilities) when there is no external market price; asset impairments; the amount of executive compensation arising from the issuance of stock options; estimated future warranty costs related to current-period sales; and the valuation of complex financial instruments such as derivatives.

Research shows that accrual-based earnings are a better measure of an organization’s operating performance than operating cash flows, and are more informative to investors (Dechow, 1994). However, accruals also introduce uncertainty that can potentially reduce the quality of earnings.¹⁶ There are no black and white accounting standards for accrual estimations. Instead, they are the subjective judgments and estimates of managers who can have personal incentives to use these accrual estimates to ‘distort’ or ‘manage’ earnings numbers. While careful readers of financial statements can identify accruals, they cannot easily determine their reasonableness. For example, managers might manage earnings to further their own self-serving objectives, such as achieving earnings targets for performance bonuses, or to meet forecasted earnings targets. This is why the careful audit and review of accruals (and constraining accruals aggressiveness) is where the audit can potentially have an important effect on the quality of audited accounting information, by providing a check on the reasonableness of managerial discretion with respect to accounting policies and accrual estimations.

Auditors are now required to report on what are called ‘key’ or ‘critical’ audit matters in their audit report, which are those areas of the audit that were important or especially difficult to assess. Not surprisingly, most of these disclosures relate to accrual estimates such as revenue recognition, asset impairments (including goodwill), and fair market values (Audit Analytics, 2021).

The magnitude of accruals in earnings will vary from firm to firm, depending on the industry and the nature of a firm’s operations. Accruals can be quite small for some organizations, and very large for others. I calculated the distribution of accruals for around 6,000 US listed firms on the Compustat database for the period 2000–2021. Total accruals are defined as: Total Accruals = Net Income from Operations – Operating Cash Flows. For most firms, total accruals have a negative sign, meaning they are expenses and reduce earnings. Examples are Depreciation Expense and Bad Debt Expense. The median value of accruals is 17 percent of revenues. Another way of measuring accruals is their size relative to operating cash flows: the median ratio is 67%, with an inter-quartile range of 35% to 147%. Accruals are clearly a large and significant component of earnings.¹⁷ Given the magnitude of accruals, it is not surprising research on financial statement quality has focused primarily on earnings quality.¹⁸

7. Is audit quality unobservable?

In the essay, I have been critical of the narrow way in which auditors and regulators think about audit quality. I now turn my criticism to academics who might read this essay. It has become almost a ‘matter of faith’ to repeat the dictum that audit quality is unobservable. This idea is probably traceable to DeAngelo (1981) who argued that it is costly or even impossible for outsiders (investors) to assess audit quality, and for this reason outsiders will use publicly observable proxies for audit quality such as the size or reputation of the audit firm.¹⁹

It is true that outsiders cannot observe the audit process: the gathering and interpretation of evidence by the auditor. However, I have argued that the outcome of the audit is what matters most in assessing quality. Outsiders can assess the outcome of an audit – the auditor’s report and the quality of the audited financial statements – in exactly the ways I have described in this essay. It is even the case that sophisticated investors have developed stock trading strategies based on the quality of a firm’s earnings and accruals that builds on the scholarly work of Sloan (1996) and others. So, it is lazy thinking for academics and others to perpetuate the myth that audit quality is ‘unobservable.’ The quality of the audit outcomes is observable. Moreover, the expanded audit report has at least the potential to give outsiders even greater information for assessing quality (Gutierrez et al., 2018; Minutti-Meza, 2021).

8. Conclusion

To recap, the framework in Francis (2011) discusses the building blocks of high-quality audits and should help firms and regulators in assessing the many different dimensions that can potentially drive audit quality. The binary model of a correct or incorrect audit report is simple and intuitive, but it is not a particularly useful (or accurate) way of thinking about and measuring audit quality. We know there is wide variation in the quality of audited financial statements, yet most receive the same standard clean audit report. There is a spectrum of audited earnings quality, but there is no comment on this in the standard clean report.

A good auditor should press their clients to make better accounting policy choices and accrual estimates that improve earnings quality, while a ‘lesser quality’ auditor might allow their client greater discretion to make choices and estimates that can lower the quality of audited earnings. So, earning quality becomes a way to make a more nuanced assessment of audit report quality, instead of the binary (pass/fail) report. The 50 shades of grey in the spectrum of the quality of audited financial statements (especially earnings) also means there are 50 shades of grey in the quality of the audit reports on the client’s financial statements. Given the limits of binary audit report, earnings quality is de facto a better measure of audit report quality.

Finally, is it possible for the audit report to say more about the quality of the client’s audited earnings? In some respects, this is what the disclosures of ‘key’ or ‘critical’ accounting matters do. They do so in a qualitative rather than empirical way, but they alert the reader to potential problem areas in the financial statements.

A bolder, more informative audit report might tell the reader if the firm’s accounting policies and accrual estimates differ significantly from industry averages and historical norms. Are earnings reflective of typical of policy choices and accrual estimates in the industry? Or do earnings contain more aggressive choices than the industry average/norm (resulting in larger earnings). Alternatively, are the earnings and accrual estimates more conservative than the industry norm (resulting in smaller earnings)?

Such an audit report could include an additional paragraph after the expression of the auditor’s opinion. Here is a hypothetical example of what such a paragraph might look like:

While we believe the financial statements are prepared in accordance with the accounting standards of the IASB, they also reflect management’s accounting policy choices and estimates. In our opinion, these choices and estimates are (choose one):

1. Typical of the historical norms in the client’s industry;

2. More aggressive than the historical norms of the client’s industry, which results in earnings that are likely to be larger relative to those based on these historical norms; or

3. More conservative than the historical norms of the client’s industry, which results in earnings that are likely to be smaller relative to those based on these historical norms.

In many respects, such a reporting requirement would simply be an extension of what auditors are already required to do for public-interest-entities, which is to discuss with the client’s Audit Committee the client’s accounting policies. It would also represent a logical expansion of the required disclosures of ‘key audit matters’ under IAASB standards, and ‘critical audit matters’ under PCAOB rules.²⁰

This type of expanded audit report could be an experimental first step in moving beyond the limitations of the binary audit report. My prediction is that such an audit reporting requirement would lead to a race to the top. The reason is that most companies would want to avoid being called out as ‘aggressive’ relative to industry norms and their industry peers, and would prefer instead to be seen as ‘typical’ or even more ‘conservative’ relative to historical industry norms. It is also possible that the audit itself would become better if auditors were formally required to make this kind of assessment of the client’s policies and estimates relative to industry norms.

To conclude, audit quality is more than ‘audit process’ quality, and audit quality is more nuanced than the ‘binary model’ of audit report quality allows for. The ultimate outcome of the audit is a report on the client’s audited financial statements. We know that financial statements (especially earnings) are best understood as a spectrum of quality that is technically allowable within the broad parameters of accounting standards. It follows logically that the quality of the audit report would be enhanced if it said something about where a client’s audited earnings are on this spectrum. A useful starting point in achieving this disclosure might be to benchmark a client’s accounting policies and estimates to the historical norms of the client’s industry.

Acknowledgement

I thank the editor, Gilad Livne, and Olof Bik and Willem Buijink for their many helpful comments and suggestions on earlier drafts of this essay. I also thank the Foundation for Auditing Research (FAR) for supporting my research program and the FAR Chair at Maastricht University. The views expressed are my own and do not reflect those of the Foundation or the audit firms supporting the Foundation.

Disclosure statement

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

Notes

1 Audit clients and financial statements users also have a stake in audit quality, but they are less actively engaged in debating its meaning.

2 There has been some discussion, even among audit firms, that they are becoming overly focused on a checklist mentality to satisfy audit inspectors, and this that may lead to loss of judgement and poor audit quality even though such mentality may improve inspection outcomes (Van Rinsum et al., 2018; Brivot et al., 2018).

3 Knechel et al. (2020) characterize audits as taking place within a multi-party collaborative network wherein auditors ‘collaborate’ with clients and institutions such as the client’s governance system (audit committees) and audit regulators in the production of audits. Their argument endogenizes institutions. In contrast, Francis (2011) views institutions as exogenous to auditing, i.e., an external source of monitoring and control over what auditors do.

4 The PCAOB (2015) in the United States issued a discussion document suggesting 28 possible indicators of audit quality, some targeted at the engagement level, and others at the aggregate firm level. No further action has been taken to date. The 28 indicators are grouped into three categories: Audit Professionals, Audit Process, and Audit Results. These groupings correspond to the categories of audit inputs, audit process, audit firms, and audit outcomes in the framework in Francis (2011). Examples of quality metrics for Audit Professionals include auditor workload, training hours, experience, industry expertise, auditor turnover, and engagement hours. The Audit Process metrics are broader and more firm-focused on incentives and quality control monitoring (internal reviews and PCAOB inspections). Metrics for Audit Results include the quality of the client’s financial statements and going concern reporting, as well as enforcement and litigation. In 2020, The Financial Reporting Council (FRC) of the United Kingdom also issued a discussion document on audit quality indicators (AQI). The document describes 52 AQI currently used by one or more of the six largest UK audit firms. Like the PCAOB, the indicators are grouped by audit inputs, audit process, and audit outcomes. Most of the metrics used are currently based aggregate firm-level data. The FRC encourages audit firms to continue developing their AQI, to make some of them more granular at the engagement level, and to publicly report AQI and assess their changes over time.

5 This is analogous to the ‘principles versus rules’ debate in accounting standards (e.g., see Livne & McNichols, 2009). Audit practices are far more principles-based than are financial reporting standards.

6 Regulators have incentives to find deficiencies, to justify their value. Unfortunately, it is difficult for auditors to ‘debate’ the different views with a regulator who has the power to publish inspection outcomes. Ultimately, auditors want to reduce inspection and reputation risks and so will adhere to what they believe the regulator expects. This could have an adverse effect on audit quality. I believe the auditor working in the field has far more information at hand on which to make their assessments than does the regulator who merely does a desk review of working papers.

7 For example, Aobdia (2019) uses propriety PCAOB data and finds that only 2 of 15 measures of financial reporting quality are significantly correlated with inspection deficiencies: meeting or beating earnings targets, and restatements. Audit hours (which are not publicly observable) are also correlated, but audit hours are an input factor rather than a measure of financial reporting quality. See Gaynor et al. (2016) for a discussion of the intertwined nature of the audit quality and financial reporting quality.

8 The idea that audits can (or should) have zero errors is patently ridiculous. For example, society does not have a zero error tolerance for something for more serious than audits, and that is whether patients live or die in surgery. It is understood that there are risks in surgery, and sometimes, even though surgeons have been competent, patients die.

9 It is possible that other undetected audit report failures exist, but there were no publicized adverse client events like client fraud or business collapse to bring the problem to light. For example, Dyck et al. (2023) estimate the real rate of corporate fraud is around 10 percent for large, listed US firms, about three times greater than known/reported frauds. Even so, this does not necessarily mean there are more audit report failures since most instances of corporate fraud do not involve auditor negligence.

10 An auditor could avoid this problem by always issuing a going concern audit report. This would result in inaccurate audit reports in most instances (over-qualifying) and are called Type 1 reporting errors (qualifying when unwarranted). This strategy would render all reports as uninformative as they would not distinguish firms with no going concern problems from those that do have problems.

11 This residual risk is termed audit risk in audit standards, which is the risk an auditor incorrectly issues a standard clean audit report on the audited financial statements. The standards do not dictate zero audit risk in planning audits.

12 There are also balance sheet implications any time accrual estimates are problematic. For example, assets are overstated (net receivables) if bad debt expense is understated to increase earnings. More generally, balance sheet quality is a concern in all audits, and is of special concern when balance sheet numbers are used in debt contracting covenants (e.g., balance sheet classification shifting to avoid covenant violations). In other instances, undetected frauds have involved ‘missing’ assets, e.g., Parmalat in Italy, Patisserie Valerie in the UK, Wirecard in Germany, and WorldCom in the US. Even though balance sheets are clearly important, investors still focus predominantly on earnings as the key summary indicator of performance.

13 Of course, it is always possible a good auditor might allow aggressive accounting, but only after a very careful assessment of the reasonableness of such policies given the client’s circumstances (Amir et al., 2019).

14 Examples of aggressive accounting would include certain revenue recognition policies such as ‘bill and hold’ sales in the software industry, unusually low rates of bad debt expense and depreciation expense, and a reluctance to record asset impairments in a timely manner. Conservative accounting would be the opposite of these policies.

15 As noted, financial statement quality goes beyond earnings, and includes the balance sheet, cash flow statement, notes, and other information reviewed by the auditor. For example, Feng et al. (2023) show that voluntary disclosures of (unaudited) non-GAAP performance measures are of higher quality when the firm’s auditor is of higher quality.

16 For a general link of uncertainty with the conceptual framework project, see Barker et al. (2020).

17 To illustrate, consider the 2021 global earnings of British-based Shell PLC (in US dollars). Net income was $20.6 billion, consisting of $45.1 billion in operating cash flows, less net accruals (expenses) of $24.5 billion. Net accruals are 54 percent of operating cash flows (24.5/45.1).

18 As noted, financial reporting quality is more than just about earnings, but it is understandable why earnings and accruals research has crowded out other financial statement research topics.

19 Causholli and Knechel (2012) build on DeAngelo (1981) and suggest that audits are a ‘credence’ good whose credibility (quality) is conditional on the audit provider’s reputation.

20 Key audit matters (KAM) are those issues the auditor considers to be most significant in the audit, while critical audit matters (CAM) are specific areas of the audit that were especially challenging, subjective, or involved complex judgments. Both KAM and CAM are required to be discussed with audit committees.

21 Francis (2011) is written largely from a US perspective, but each country will have its own set of relevant institutions

institutions in EU countries (see https://finance.ec.europa.eu/system/files/2021-02/ceaob-composition_en.pdf).

22 Audit outcomes include the audit report and the audited financial statements, and real economic effects include capital market consequences such as access to and cost of debt and equity financing, and the efficient allocation of capital. Higher quality audits would be predicted to have greater capital market consequences. There is not much research on this topic beyond the effect of auditor choice on the market’s response to corporate earnings announcements, e.g., Teoh and Wong (1993).