The Sarbanes-Oxley Act of 2002 (SOX) was enacted with widespread bipartisan support rarely seen today, receiving favorable votes of 423 to 3 in the House of Representatives and 99 to 0 in the Senate on the consolidated bill that expanded greatly upon the Senate bill, which was more focused and regulatory-oriented than the House bill. The overwhelming congressional approval reflected the unceasing chain of high-profile financial reporting scandals; two of the most significant scandals—AOL-Time Warner and WorldCom, which inflated its assets by an estimated eleven billion dollars—were disclosed on the cusp of Congress adjourning for the August recess during which senators and representatives anticipated they would be pummeled by their constituents’ outrage over the untrustworthiness of the financial reporting systems. These reporting failures came on the heels of similar reports involving numerous well-established publicly traded companies, e.g., Enron and Global Crossing. These reporting gaffs and the ensuing collapse of prices across securities markets propelled Congress to make the most sweeping reforms to the American securities laws since their inception in the Great Depression.
I. Waning Restatements, Rising Revisions, and Questions of Independence
The ensuing financial crisis introduced to the public the meaning and significance of an accounting restatement. Before and especially in SOX’s wake, accounting restatements became a failure event among reporting companies. Indeed, restatements did not recede from the public attention with President George W. Bush signing SOX on July 30, 2002. They dominated the financial news, with their number steadily increasing, peaking in 2006 when 13.6 percent of reporting companies reported a restatement; since then, the percentage has declined, reaching 4.8 percent in 2020. If the story stopped there, we could easily point to the decline in restatements as evidence that SOX had demonstrated it could move an important needle in measuring the quality of financial reporting by public companies. The correlation would be hard to overstate: After the implementation of SOX’s multiple measures, there followed a significant decline in accounting restatements. Walla!
Victory is not so easily recorded. Declines in the number of accounting restatements must nonetheless be assessed against the disturbing rise in a relatively new phenomenon, accounting “revisions.” The difference between the two is legally significant. Accounting restatements, so-called “Big Rs,” reflect information that earlier financial reports were materially incorrect, can no longer reasonably be relied upon, and must be restated; the SEC requires all restatements to occur promptly via filing Form 8-K, thereby giving the change a good deal of prominence. Items that do not have a material impact on an earlier financial report are beyond Form 8-K and are usually treated through disclosure in the footnotes of the fiscal period in which the gaff is discovered. If the error is not material to previously issued financial statements, it may nonetheless be of a nature that not correcting its current effect renders the current period’s reports materially misleading so that a registrant must still correct the error. This type of correction is sometimes referred to colloquially as a revision restatement or a “little r” restatement and results in no correction to earlier released financial reports. Note that the differing treatment turns on whether the item causing the earlier misstatement is material across financial periods.
Management is not neutral between treating the matter as a restatement or a revision. As seen, restatements result in the reported item having more prominence by compelling their prompt disclosure on Form 8-K. More significantly, restatements often arise not because of an innocent mistaken judgment, estimate, or assumption, but because of “misconduct.” When this occurs, Section 304 of SOX requires the firm’s CEO and CFO to reimburse the firm for any bonus or incentive compensation received during the twelve-month period following publication of the financial report embodying the restatement. Moreover, such a clawback does not require any wrongdoing or complicity by the disgorging officer in the events giving rise to the restatement. Liability is absolute. Further concern for restatements appears in the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, which amended Section 10D of the Exchange Act and mandated that the SEC prohibit public listing on any exchange unless the issuer implemented procedures to recoup from any executive incentive compensation received in the three-year period before the restatement’s publication. The SEC recently imposed that prohibition, further heightening the consequences of a correction being deemed a restatement. The SEC’s recent action evaluated whether the correction is material in more than the current period, but wisely focused on whether senior executives have been rewarded by a measurement standard later determined to have been not just erroneous, but materially erroneous. Thus, the intent of Congress behind the Dodd-Frank provision is fulfilled, finally, and without openly confronting the more controversial concern of whether revisions exist because of a lack of sufficient independence on the part of the auditor.
Such independence was questioned in a recent study of annual reports filed with the SEC from 2004 through 2015 that examined 1,842 reports that contained accounting misstatements of which 54 percent were treated as revisions and the remainder were reported as restatements. Employing a variety of tests commonly used by auditors to determine the materiality of a matter, such as whether the change is 5 percent or more of income, the study’s author concludes that one-third of the revisions were material, observing that it is “not uncommon for a firm to deem a misstatement immaterial … despite it meeting an observable materiality indicator.” She further finds that classifying a misstatement as a revision, rather than a restatement, was highly correlated with whether the firm had in place procedures for clawing back incentive compensation executives garnered during any period affected by a restatement.
Reports that companies appear to be characterizing material reporting errors as revisions shines a light on whether SOX has facilitated real progress in an area that weaves through key provisions of the Act focused on the linchpin of the trustworthiness of financial reporting: the independence of the reporting company’s outside auditor. If the prevalence of accounting restatements measures the auditor’s professional skepticism when evaluating management’s financial reports, in a world in which restatements and revisions are interchangeable in the real-time setting of the auditor’s work, we can seriously question just what SOX has achieved.
The bedrock of SOX’s mechanisms to promote the auditor’s independence is found in its first three chapters setting forth the legal foundation for accounting and audit standard setters, centralizing responsibility for the auditor in the audit committee, and introducing numerous restrictions with the intended effect to rid auditors of the powerful conflicts of interest that pervaded their historical relationships with their audit clients. The rules and conventions for reporting significant accounting information, known as generally accepted accounting principles (GAAP), have been overseen by the Financial Accounting Standards Board (FASB) for over three decades. Until SOX, the FASB was funded by contributions from the accounting firms, a practice that hardly shielded standard setting from the influence of the auditing profession’s clients. At the same time, the rules for the professional referees, i.e., auditors, to attest that management was adhering to GAAP were overseen by the auditors themselves through the professional standards body within their trade organization, the American Institute of Certified Public Accountants. It is easy to see how a wonderfully incestuous symbiotic relationship could produce the reporting meltdowns that spawned SOX. SOX addressed these organic weaknesses of the earlier standard setters.
SOX amended Section 19 of the Securities Act to require that the private sector’s accounting standard setter, then and now the FASB, must be funded exclusively by the public fisc, that a majority of that body’s members not be either current or recent members of auditing firms, and that the body could act by majority vote. Furthermore, SOX removed oversight of auditing standards and auditors from the profession and placed them under the newly created Public Company Accounting Oversight Board (PCAOB), whose five members were to be appointed by the SEC, after consultation with the Chairman of the Federal Reserve and Secretary of Treasury, and whose budget is now publicly sourced. The broad objective therefore was to insulate both standard setters from the influence of the regulated.
Recent history, however, suggests we should not be too complacent as to whether these structural initiatives are themselves sufficient. For example, the SEC’s Investor Advisory Committee recently leveled pointed criticism at the FASB for misallocating its resources on narrow technical issues that arise in financial reporting and not on thornier more sweeping questions. The significance of SEC oversight was apparent after the election of a new president and the appointment of a new SEC chair, who won headlines and plaudits for unceremoniously removing the sitting chair of the PCAOB and for swiftly underscoring the message that the previous board neglected key areas of the PCAOB’s mission. Nonetheless, concerns for true independence within this regulatory quilt persists from such facts as that, over the PCAOB’s twenty-year life, the SEC has frequently been less a governor on many of the PCAOB’s key initiatives than a cheerleader. Underscoring this is a disturbing statistic. In the last twenty years, eleven individuals have led the SEC’s Office of Chief Accountant, and each, immediately prior, had been a partner with a Big Four accounting firm, most of whom promptly returned to their former firms at the conclusion of their SEC tenure. Recent events likely reflect that the true linchpin for auditor professionalism is the commitment to that cause by the SEC’s chair, and not SOX’s provisions.
II. SOX and the Forces Girding the Lack of Auditor Independence
Firms making up the S&P 500 had a collective market capitalization, on January 31, 2023, of $35.9 trillion, or 80 percent of all publicly held firms traded in the United States. Among this elite group, only three companies were audited by a non–Big Four accounting firm in 2017. In 2022, the Big Four replicated that massive footprint, auditing almost 45 percent of all reporting companies, and were even more dominant among the largest firms, auditing more than 80 percent of so-called large accelerated filers. Concern regarding the industry’s concentration has long existed, arising from fears that the dominating Big Four auditors would misbehave as oligopolists, competing neither on the price nor the quality of their services. The lack of competition is supported by the absence of turnover of auditors by their clients. The audit firm’s relationship with its client may reflect unquestioned obedience more than robust competition among auditors on price and quality. Furthermore, there are also the sobering lessons learned from the collapse of then–Big-Five member Arthur Andersen, the demise of which disrupted the timely reporting process for—and posed regulatory and operational challenges to—hundreds of public companies. Are the continuing Big Four too big to fail? A quite different perspective is that concentration in the place of competition can steel the auditor’s independence from its clients who, because it is insulated from competition, can demand clients comply with high GAAP and generally sound reporting practices. Studies of this question, including two by the Government Accountability Office, provide mixed results on which of these two outcomes prevails within the industry.
A recent study, using a very clever research design, sheds a good deal of light on the favorable effects of competition when a firm considers changing accountants. The methodology enabled the investigators to identify SEC reporting firms contemplating a change of accountants so that the “bidding process” engaged in by the incumbent auditor and its competitors could be matched with outcomes, such as the ensuing audit fees paid by the registrant, auditor disclosure of material internal control deficiencies, and accounting restatements. The authors concluded:
We find that Big 4 competitive bidding is positively associated with audit quality as measured by a lower likelihood of misstatement. In addition, we find that Big 4 competitive bidding is associated with modest fee reductions in the two years following an auditor change, regardless of whether the incumbent auditor wins reappointment or not. Thus, we provide evidence indicating that incumbent auditors respond to competitive pressure by reducing fees without sacrificing audit quality.
The above conclusion supports regulatory policy that encourages audit clients to periodically consider changing audit firms, as even in the highly concentrated market of auditing services for large public firms, the profile of the study, the lash of competitive bidding produced statistically observable positive effects. These effects persisted even in the face of another potentially harmful development, the rise of auditors providing non-audit services to their clients.
Industry concentration is not the only force that can weaken auditor independence. SOX amended Exchange Act Section 10A to prohibit a reporting firm’s auditor from providing eight types of non-audit services, to empower the PCAOB to identify any other prohibited non-audit services, and to require the auditor to obtain preapproval from the reporting firm’s audit committee for any permissible non-audit services. These steps were clearly taken for their prophylactic effects, guided by evidence that auditors garnered a substantial part of their revenues by providing non-audit services to their audit clients and frequently the services rendered were an important focus of the audit itself. David Duncan, the Arthur Andersen partner overseeing Enron’s audits, was the poster child for such concerns, with the bulk of his compensation determined by bonuses received for cross-selling non-audit services to Enron. We can imagine that Duncan’s verve to question Enron’s many suspicious accounting practices would have been strengthened by an SEC requirement that his termination as the auditor by Enron, for whatever cause, triggered prompt public disclosure and most certainly a visit by the SEC. On the other hand, if Enron instead retaliated against Duncan (for challenging its practices) by terminating Arthur Andersen’s lucrative consulting arrangements, this would not have triggered a reporting obligation and, at the same time, would have substantially reduced Duncan’s and his employer’s financial benefits. Hence the prophylaxis now found in Section 10A(g) and (i) of the Exchange Act. The SOX-induced prohibition produced noticeable and rapid effects; among SEC reporting companies, the auditors’ non-audit fees peaked at $4.24 billion in 2002, fell by 26.4 percent to $3.12 billion in 2003 and fell another 18.3 percent to $2.55 billion in 2004, and, at the end of a nineteen-year study, stood at $1.75 billion in 2020.
The above statistics should not be read to suggest that the audit industry’s salad days are behind it. Reports of yearly revenue sources by the Big Four suggest otherwise. In 2021, the Big Four’s revenues reportedly totaled $115 billion world-wide from consulting and tax services, more than double the $53 billion received from its audit practice. Indeed, in the ten years ending in 2021, the Big Four’s global revenues from consulting and tax work rose 96 percent, eclipsing by many multiples the 17 percent growth in audit revenues. The contrasting statistics underscore the salutary impact of SOX on the provision of non-audit services to SEC-reporting companies amidst the fast-paced expansion of this feature of the Big Four’s work to non-reporting companies. Nonetheless, the amount of non-audit services provided to reporting companies of $1.75 billion dollars is not trivial. Further concern is whether, in a firm culture in which growth is enjoyed, and likely more aggressively pursued, the rendition of non-audit services will have harmful effects on the independence of those engaged in the slower growth lane of auditing. Hence, we see concern recently professed by the SEC over the growth of non-audit services.
We therefore see at play the twin forces of an anticompetitive industry structure for audit services provided to public companies and the ever-present lure of fast growth in non-audit revenues from audit clients. And the recent history with the PCAOB reflects at least that regulatory zeal is far from steady. At the same time, evidence indicates that the process of competitive bidding among auditors has enhanced the quality of reporting, even when that bidding process does not cause a change in the firm’s historical auditor. Moreover, the level of non-audit fees garnered from reporting clients appears to have stabilized, though remaining at levels that nonetheless cause pause whether they impact the engagement auditor’s independence. Thus, industry structure alone, even when joined by independently funded regulators, is unlikely to be the antidote to enhance auditor independence. This conclusion is supported by SOX containing multiple directives to improve financial reporting, each stealing the auditor’s independence from the audit client.
An important SOX intervention mandates that the engagement partner’s tenure not exceed five years. Consequently, not less frequently than every fifth year, the audit will be under the direction of “fresh eyes”; the new engagement partner by definition is not someone who has, over prior years, developed social and psychological ties to the client’s management or is associated with the numerous accounting judgments, assumptions, and estimates that were employed in preparing the earlier financial statements. A collateral benefit of the programmed “fresh eyes” is it supports independence on the part of the current, soon the outgoing, engagement partner; the advent of a change in supervisory personnel can be expected to temper the current auditor bending inappropriately to the client’s reporting decisions. In this way, rotation introduces to auditing a form of peer review with the consequential effect of stimulating the outgoing partner to conduct a higher quality audit in the year or years preceding succession. On the other hand, fear abounds that changing auditor partners will weaken audit quality because such rotation sacrifices the client-specific knowledge the departing auditor gained through her tenure. Lacking such deep knowledge regarding the client’s business and systems, the new auditor may well be less likely to identify financial reporting problems. Probing which of these two competing visions of the consequences of partner rotation dominate has generated a good deal of investigation.
Because any change in the engagement partner, while being disclosed to the PCAOB, is not otherwise public information, most research on the effects of audit partner rotation has focused on practices and experience outside the United States, where engagement partners are publicly disclosed. Within the plethora of such studies, one, which focused on Hong Kong–listed firms and which appeared in a leading American accounting journal, lends strong support for the “fresh eyes” position. Because China not only requires public companies to disclose the individual engaged in the audit, but also requires their clients to set forth their pre-audit profits, the investigators were able to assess whether changes in audit partners were associated with material differences between pre- and post-audit financial reports. The study found that “audit adjustments occur more often when the engagement partner is scheduled for mandatory rotation at the end of the year” and “that audit adjustments occur more often during the incoming partners’ first year of tenure than in other years.” Each is consistent with the earlier described twin benefits of the “fresh look” hypothesis.
It is never clear how findings about financial reporting outside the United States translate to the United States—the U.S. regulatory quilt is broader and snugger than that found in other countries, the risk of lax auditors being disciplined by regulators or in private suits is greater in the United States, and there is a deeper capital market monitoring in the United States. To this end, the recent study by Professors Brandon Gipper, Luzi Hail, and Christian Leuz is significant not just because of the comprehensiveness of its inputs (3,333 audit clients from 2008 to 2014 involving 2,385 engagement partner rotations), but because it mines PCAOB proprietary information that bears on a range of matters that include the engagement partner, the hours each logged in engagement, and the audit fees. The study finds no statistically significant association between audit partner rotation and the proxies used to measure quality of financial reporting. Restating this observation, the authors found no statistically significant evidence of adjustments, restatements, or qualifications leading up to or following a change in audit partners. Interestingly, audit fees tend to drop in the year of partner rotation but rise over the remaining tenure of the partner; the trend is different with respect to hours logged, with the hours logged by the engagement partner declining over her tenure. Apparently, the change in audit partners is an opportunity seized by the client to renegotiate the audit fees, while at the same time, the new auditor may be “low balling” to avoid the client shopping for a new firm, but understanding recovery of that discount over the remaining partner’s tenure. A further significant finding is that, certainly with respect to large and hence complex clients, the audit firm far in advance of the rotation of partners assigns the upcoming partner to “shadow” the outgoing partner, thereby seeming to moderate the earlier described concern that audits may deteriorate after a rotation because of the loss of firm-specific knowledge.
There are several threads of analysis that can explain these findings. The absence of a connection between partner rotations and the proxies for reporting failure—as found by Gipper, Hail, and Leuz—is consistent with five years being sufficiently short to prevent co-opting the auditor’s independence. Moreover, their findings are consistent with the belief that an engagement partner’s awareness that there will be a “fresh look” has a beneficial effect on the partner maintaining her independence.