«Corporate Audits and How to Fix Them Joshua Ronen A uditors are supposed to be watchdogs, but in the last decade or so, they sometimes looked like ...»
0.03 percent of market value, and average fees as a percentage of sales decrease as firms become larger. In absolute terms, based on 2008 AuditAnalytics available data, the average audit fee was $634,000 for companies with market capitalization below $1 billion, and $7.5 million for the companies with market capitalization above $1 billion—although fees for some firms can greatly exceed these averages. From the auditor’s perspective, however, combined audit and consulting fees can be quite substantial. The combined fees of the biggest four audit firms were $16.89 billion (94 percent of the total fees earned by all audit firms): Deloitte & Touche, $3.84 billion; Ernst & Young, $3.88 billion; KPMG, $3.29 billion; and Pricewaterhouse Coopers, $4.81 billion.
In a number of ways, it seems reasonable to think about the auditing profession in two parts: the largest firms and their large auditors, and the rest of the One of the indices of audit quality that accounting researchers (improperly, it will be argued) often examine is the amount of “abnormal accruals.” Accounting accruals generally are defined to be the excess of reported income over the difference between the cash receipts and cash disbursements that are related to operations. The portion of such excess that seems justified by the magnitude of sales or fi xed durable assets (for example, credit sales are recognized as revenue even though not received in cash) is referred to as “nondiscretionary” or “normal” accruals. The excess of the actual amount of accruals over normal accruals is referred to as “discretionary” or “abnormal” accruals.
194 Journal of Economic Perspectives Figure 1 Public Companies and Their Auditing Firms, 2002 and 2006
80 16% 45% 69% 98% 98% 95% 10% 92% 90% 71% 10% 44% 22%
Source: Figure 2 of GAO (2008), Audits of Public Companies: Continued Concentration in Audit Market for Large Public Companies Does Not Call for Immediate Action: Report to Congressional Addressees.
Note: The figure shows the percentage of companies (distinguished into groups by size) using auditing firms in the smaller, midsize, and largest categories.
market. In a survey conducted by the GAO (2008), 82 percent of the large companies saw their choice of auditor as limited to the Big Four because those firms have the technical expertise, capacity (like international offices), and reputation to undertake complex audits. The need for an independent auditor further limits firms’ choices. Almost all (96 percent) of the large companies reported that they had used one of the Big Four for some nonaudit services, which under current rules means that it becomes more difficult to hire that firm as an auditor. In addition, many large firms would prefer not to use the auditor of a main competitor, which further constrained their choices. Further, many small and mid-sized audit firms shun large public companies in view of the higher risk of such an audit going wrong—and the higher costs that would be incurred if one did go wrong. As a result, the GAO study found that many industries and regions have audit markets even more concentrated than the totals in Figure 1 might suggest: for example, Ernst & Young alone accounts for 77 percent of fees collected in the agricultural sector. Little wonder that in the GAO survey, 60 percent of the large companies Joshua Ronen 195 viewed competition in the audit sector as inadequate. Most small public companies, on the other hand, seemed to be satisfied with their audit choices.
The existing empirical evidence does not show that the seemingly high level of concentration in this market is leading to an exercise of market power and higher fees (for example, GAO, 2008). Perhaps a more interesting question for present purposes is whether high concentration can compromise incentives for high-quality audits—an issue to which we will return.
Empirical Evidence on Audit Quality
Have auditors been doing their jobs sufficiently well? The quality of their performance is a contentious issue because while experts can judge individual audits as being of high or low quality on a case by case basis after the fact, no systematic data exists on whether audits as a group are of high or low quality. 5 The empirical literature thus has attempted to use a variety of measures to proxy for audit quality, all of which have their difficulties.
The usual proxies rely on counting events that might signal a poorly performed audit: the quantity of certain kinds of litigation; SEC investigations of and enforcement actions against auditors (including those in which an auditing firm does not admit wrongdoing but pays a fine and agrees to halt a certain practice); and restatements of corporate earnings (reissuance of past financial statements to correct errors or misapplication of GAAP). Francis (2004) reviews this kind of evidence and concludes that these proxies for failed audits occur at a low frequency—less than 1 percent annually, based on a population of around 10,000 publicly listed companies in the United States. A major problem with these proxies, even if otherwise valid (and below I argue they mostly are not), is that they are binary in nature, suggesting that audits are either “good” or ”bad.” However, quality varies along a continuum, and it would be desirable to know the average quality of audit for the economy as a whole. At some times, trends in the proxies may reveal trends in the average quality. In many cases, however, like a down stock market or a change in the enforcement regime, these proxies may rise or fall for reasons other than an average audit quality.6 Yet another proxy sometimes used for the quality of audit has been the size of the auditor, with the presumption that no single client is so important to a large auditor that it will risk losing its reputation by misreporting (DeAngelo, 1981). But this argument sounds somewhat dated by the late 2000s! For example, if the audit firm’s incentive structure rewards each of its offices separately for its profitability, the local audit team’s independence may be compromised even if the whole firm is The PCAOB, as discussed further below, reviews audit engagements and the quality control systems of accounting firms, but only limited information contained in its reports is made public. In particular, the public portion of the reports does not include any discussion of potential defects in the firm’s quality control systems unless the defects are not addressed satisfactorily within 12 months.
I thank Timothy Taylor, the managing editor of this journal, for making this point.
196 Journal of Economic Perspectives not. Large audit firms may face a situation where gains from not asking too many hard questions are reaped by the auditor at hand, in terms of higher fees, while losses from a poor audit will be carried by the firm as a whole. Nor can the fact that Big Four audits have been found to have higher fees (DeFond, Francis, and Wong, 2000; Ferguson, Francis, and Stokes 2003) be seen as an indication of higher quality. Higher audit fees may imply greater effort (hours) or greater expertise (higher billing rates), but they could also reflect more monopolistic price power or even incentives for the auditor to lean toward management’s view of how the firm’s transactions should be reflected in the financial statements.7 While the number of clearly failed audits is relatively low, major difficulties arise when using such proxies to evaluate auditors. These outcomes emerge from an interaction among different parties: corporate management, the auditing firm, government regulators, and other watchdogs like investors and the financial press.
For example, if management becomes for some reason more likely to engage in aggressive management of earnings in a way that tiptoes up to the line of abuse— and sometimes crosses it—auditors may not be well-prepared to react to such a change. If and when auditing failures occur and are observed depends to some extent on the actions of all these parties.
Ultimately, the important question is whether investors feel that they can trust the results of audits. From that perspective, even a small number of clearly failed audits can cause major disruptions. After all, a lawsuit affects not just one relationship between an auditing firm and a client, but the extent to which any of the audits from that firm—or even whether certain decisions being made by auditors as a group— can be trusted. Such lawsuits have been common in recent years. In 2002, Arthur Andersen, at that time one of the five largest firms in the United States, dissolved after it was indicted on obstruction of justice charges. KPMG recently faced potential criminal indictment in connection with the provision of tax-related services, but it entered into a deferred prosecution agreement with the Department of Justice (GAO, 2008). BDO Seidman, a midsize auditing firm, is appealing a $521 million state judgment related to a private audit client, and according to its chief executive the judgment would threaten the firm’s viability (Associated Press, 2007).
The current economic downturn seems sure to increase skepticism about audit quality because, when many firms are losing money, lawsuits blossom and often focus attention on whether auditors should have issued reports with more qualifications. Data on federal securities class action filings show 110 filings during the first half of 2008, which considerably exceeded the semiannual average of 63 filings between July 2005 and June 2007, as well as the annual average between January 1997 and December 2007. Of course, not every lawsuit is justified or will eventually be upheld. But when stock prices are falling, a resulting wave of lawsuits will put Another argument sometimes made is that firms audited by the Big Four accounting firms tend to have lower abnormal accruals, which can be interpreted to imply higher earnings quality on the part of the audited firms (Becker, DeFond, Jiambalvo, and Subramanyam, 1998; Francis and Krishnan, 1999).
However, these lower levels of abnormal accruals could also lead to worse predictions of future cash flows, depending on how they are defined (Brochet, Nam, and Ronen, 2008).
Corporate Audits and How to Fix Them 197 auditors and their reports under a microscope, and some of the audits will surely come off badly.
Current Incentives for Quality Audits There are currently a number of potential incentives to encourage quality audits, including regulation, rules that attempt to ensure a degree of auditor independence, along with incentives that audit firms have for competing and building reputation. But each of these methods has its weaknesses, and even taken together, they have not overcome widespread doubts about the quality of audits.
Regulatory Structure Auditing occurs under a regulatory structure that was overhauled by the Sarbanes–Oxley Act of 2002. The Act created the Public Company Accounting Oversight Board (PCAOB), a private-sector, nonprofit corporation that oversees the auditors of public companies. This board has sweeping powers: the power to register public accounting firms; to set auditing, quality-control, ethics, independence, and other standards; to conduct inspections of registered public accounting firms; to conduct investigations and disciplinary proceedings and to impose sanctions on registered public accounting firms if justified (including fines of up to $100,000 against individual auditors and $2 million against audit firms). The PCAOB can also set rules to regulate the nonaudit services that audit firms may offer their clients, such as consulting or tax services. In addition, the board is empowered to require audit firms to provide testimony or documents in its possession and may suspend or debar audit firms or auditors that fail to comply; it may also seek the SEC’s assistance in issuing subpoenas for testimony or documents from individuals or entities not registered with the PCAOB.8 Between 2004 and 2006, the Public Company Accounting Oversight Board performed 497 inspections of registered U.S. “triennial” firms—firms auditing no more than 100 entities and which are subject under law to being examined at least once every three years—the number of which ranged from 893 at the end of 2004 to 986 at the end of 2006 (PCAOB, 2007). Although numerous deficiencies were identified, not much light has been shed on the extent to which these deficiencies affected actual quality of the audited financial statements. Moreover, the PCOAB A prominent rule from Sarbanes–Oxley is that the chief executive and fi nancial officers must certify personally that financial statements “fairly present…, in all material respects, the fi nancial condition and results of operations of the issuer” (18 U.S.C. § 1315(b)); violations can result in a maximum fi ne of $1 million and up to 10 years in prison, or $5 million and 20 years if the wrongful certification is “willful” (18 U.S.C. § 1350(c)). This latter requirement has been linked to a record number of restatements correcting errors in fi nancial statements, suggesting some degree of effectiveness, although statutes already provided significant penalties in terms of prison time but did not deter the major violations of 2001 and 2002 (Shapiro, 2005). This experience also illustrates that failed audits are not just the outcome of efforts by audit firms, but result from an interaction among regulators, firm executives, audit firms, and other gatekeepers.
198 Journal of Economic Perspectives conducted disciplinary proceedings involving violations of audit standards against no more than 17 audit firms in this time, only one of which was a Big Four firm (Deloitte & Touche) (PCOAB, 2008).
For all the powers of the Public Company Accounting Oversight Board, it does not change the basic contractual arrangement under which auditors operate: corporations still engage their own auditors and compensate them for their services. Thus, although the Act addressed some of the deficiencies of the pre-existing system, the perverse incentives built into the relationships among the auditor, the client, and the public still linger.