KPMG LLPNovember 25, 2002 Mr. Jonathan G. Katz
File No. S7-39-02
Dear Mr. Katz: We appreciate the opportunity to comment on the Commission's proposed rule, Improper Influence on Conduct of Audits, Release Nos. 34-46685/IC-25773 (the "Proposed Rule"). KPMG LLP commends the Commission on its efforts to enhance the means by which it may prevent individuals from fraudulently interfering with the independent audit of a public company. KPMG supports the overall intent of the Proposed Rule as an important step in restoring public confidence in independent audits of public companies, and in helping auditors to perform their important functions. However, we believe that certain aspects of the Proposed Rule require further clarification or amendment, as described below. As currently drafted, the Proposed Rule could lead to certain unintended consequences that could ultimately injure investors and impair the quality of independent audits. We respectfully request that the final rule be modified to avoid this effect. To mitigate these consequences we recommend, in summary, that the Commission:
A detailed discussion of our comments follows. I. The Proposed Rule Could Have The Unintended Consequence Of Sharply Reducing Auditor-Client Communication That Would Injure Audit Quality The Proposed Rule broadly defines several key terms in a manner that could unduly inhibit the dynamic interaction between auditors and their clients that is essential to the completion of an audit in accordance with generally accepted auditing standards. An audit of a public company is a complex process that requires the difficult exercise of professional judgment. These judgments are necessary and appropriately made after discussion with the client, who - among other things - provides needed information and explains the underlying assertions contained in the client's financial statements. The interaction between auditors and their clients during this process needs to be open and frank in order to appropriately test the financial representations of an issuer. However, the Proposed Rule's broad scope, and use of ambiguous terms as to what might be considered prohibited conduct and who might be subject to charges or discipline in connection with such activity, may impede this critical dialogue and could lead to less effective audits. A. The Proposed Rule Seeks To Read Out The Scienter Requirement Congress Intended In passing the Sarbanes-Oxley Act of 2002 (the "Act"), Congress emphasized that it must be unlawful for a person to interfere intentionally with the audit of a public company. Section 303(a) of the Act prohibits "any officer or director of an issuer, or any person acting under the direction thereof, to take any action to fraudulently influence, coerce, manipulate, or mislead" an auditor in the course of their audit of the issuer's financial statements. As with criminal prohibitions generally, Congress created a scienter requirement such that any interference in an audit must be done with a "fraudulent" intent. The Commission has stated that it will interpret the word "fraudulently" in Section 303(a) to modify only the word "influence" and not "coerce, manipulate, or mislead." We believe that this interpretation could lead to an unintended prohibition of a great deal of beneficial and innocent conduct - in essence by making it unlawful for an individual without any fraudulent intent to exert any influence over an audit, or at least to create the possibility that robust give-and-take could later be deemed to have been impermissible "coercion, manipulating or misleading." Though it can be argued that a separate intent requirement is provided by the language that requires conduct under Section 303(a) to be done "for the purpose" of rendering an audit report materially misleading, we believe that the Commission's proposal to remove the element of fraudulent intent could be interpreted as removing, or at least severely diluting, any scienter requirement. This is particularly true because any position taken by an auditor that is later subject to scrutiny will likely involve some item material to the financial statements, such that any alternative position urged by the issuer could be seen as having such a purpose to influence. The result of this expansive interpretation could be that expressions of honest, professional disagreement between auditors and their clients could be interpreted to be contrary to Section 303 and thereby subject the participants to sanction under the Proposed Rule. For example, as currently drafted, the Proposed Rule could be invoked to prohibit a chief financial officer from calling his company's auditor to debate a preliminary assessment with respect to the application of GAAP to some aspect of the company's financial statements. It is a fairly routine occurrence where, during the course of an audit and based on less than all existing evidence, a member of the audit team may make an initial assessment of an accounting issue. In some of those cases, the company may have additional information, which, when brought to the auditor's attention, might lead the auditor to reach a different conclusion on the issue. In those cases the CFO might try to persuade the audit team of the correctness of the company's position. An auditor may be convinced by the CFO's argument regarding the proper application of GAAP to the particular accounting issue and change his or her mind. Absent application of the scienter limitation, it is unclear whether this commonplace and useful interaction later could be characterized as unlawful "coercion" or an attempt to "manipulate" the outcome of the audit under the Proposed Rule. Faced with application through hindsight, the Proposed Rule could make it dangerous for an auditor to change his or her mind after discussing a judgment with an officer or director of a client. Thus, absent clarification that the auditor will not be subject to challenge or discipline for the honest exercise of professional judgment, the unintended consequence of the Proposed Rule could be that auditors become locked into initial positions for fear of after-the-fact criticism that, by changing their honestly held opinion, they were unduly influenced by management. Another example, drawn from the Commission's release, involves an attorney who prepares a memorandum for an issuer's auditors that misstates the law. The Proposed Rule and the examples of "improper influence" in the proposing release would suggest that the lawyer might be guilty of unlawfully interfering with an audit, even if the "misleading analysis" was an innocent mistake. The question for the auditor will be how can he or she rely on any legal advice if he/she needs to be concerned that such advice could later be deemed an improper attempt to influence the audit. Similarly, under what circumstances can the auditor accept and rely on a legal opinion from the company's outside counsel as opposed to seeking a second legal opinion, with the corresponding cost to the client, in order to protect themselves? Additionally, we believe that the materiality requirement of Section 303(a) does not provide a meaningful limit to the number of situations in which the auditor's conduct will be subjected to second-guessing under the Proposed Rule. For the most part, auditors and their clients will only have significant discussions or disagreements over material issues, or at least items that regulators will subsequently consider to be subjectively material to the financial statements. As a result, the Proposed Rule would likely have an adverse effect on the communication between auditors and their clients and those who interact with the auditors during the course of an audit engagement, such as issuers' legal counsel. If "coerce, manipulate, and mislead" are given the unrestricted reading suggested by the Proposed Rule, the natural and vital give-and-take that exists, and must continue to exist, between auditors and issuers may well be artificially and unnecessarily constrained, because the participants may fear that they might run afoul of the Proposed Rule's prohibitions. The audit process could suffer as a result. Accounting is itself dependent on judgment and auditors need to know how that judgment was exercised by their clients in order to be effective. The only means of obtaining this crucial insight is through full and frank discussions with the client. Congress could not have intended such innocent and productive behavior to be unlawful under the Act. Moreover, Congress could not have intended the degree of open and honest dialogue between issuers and auditors to be limited by subjective interpretation of the nature of these discussions discerned from non-objective criteria, like the decibel level at which some of these conversations may take place. Since the passage of the securities laws, auditors have dealt with all kinds of clients, and it should not be presumed that they cannot continue to exercise professional judgment in the face of disagreement from management. The Proposed Rule must clearly reject such an interpretation and explain that normal, professional interactions with clients and their advisors - even ardent, but bona fide disagreements - will not be deemed to be violations of Section 303(a). To that end, the Commission should construe the statute as requiring a fraudulent intent consistent with the natural meaning of the statutory text, such that "fraudulently" modifies "coerce," "manipulate," and "mislead," in addition to "influence." The Commission should also make it clear that Section 303 does not create any liability (civil, criminal or regulatory) for an auditor who may be the victim of the prohibited conduct. We are very concerned that the Commission's important interpretation has been relegated to a footnote which could be overlooked by potential commentators. B. The Proposed Rule's Examples Of "Fraudulent Influences" Demonstrate An Excessively Broad Scope Of Activities Open To Subjective Interpretation Related to the concerns addressed above, the examples of prohibited conduct under the "fraudulent influence" standard set forth in the proposal - the one area for which the scienter requirement is retained in the Proposed Rule - demonstrate that the Commission has effectively eliminated the scienter requirement there as well. We will address some specific examples:
This example lists the "offering . . . [of] contracts for non-audit services" as an act that might constitute "fraudulent influence" of an auditor under the Proposed Rule. The Commission needs to clarify that what is being prohibited is an express quid pro quo and not any request for permitted non-audit services that an issuer may make to an auditor during an audit engagement. Issuers often hire their audit firms to provide non-audit services during the "audit engagement," especially as that term is broadly defined under the Proposed Rule to include periods before and after the actual audit. Congress expressly permitted auditors to provide certain non-audit services to their audit clients under the Act if they are pre-approved by the client's audit committee. This example, without clarification, would inappropriately invade the Act's carefully devised scope of services regime, by effectively exposing any company (or auditor) to criticism and possible discipline for retaining the audit firm to provide permitted services. This topic will be addressed further below. See infra Part II.
This example could well deem numerous innocent acts to be violations of the Proposed Rule. The Commission needs to emphasize that the provision of "inaccurate or misleading legal analysis" would constitute a violation only if it is done with a fraudulent intent, and that the auditor bears no exposure for reliance on expert legal advice that later turns out to be inaccurate.
While no one would suggest that corporate officials should be able to use the threat of discharge to cause an auditor to change his or her honestly held professional opinions, it is not clear that the current Form 8-K Item 4 disclosure regime does not adequately address this concern. It is also of concern that this example would seemingly prohibit an issuer from discontinuing the engagement of an auditor in order to obtain more specialized industry knowledge from another accounting firm. Restricting the ability of issuers to change auditors for any number of legitimate business reasons is counter to the ultimate goal of improving audit quality. Again, the Commission needs to emphasize that the cancellation or threatened cancellation must be part of a quid pro quo for a specified reporting result in order to constitute "fraudulent influence."
This example of a presumptive violation could discourage audit firms from deciding to replace any partner on an engagement for any number of legitimate reasons. A firm may make a replacement in order to have a more experienced partner assigned to an audit that has become unexpectedly complex, or simply to address other personnel or staffing issues within the firm. Audit firms may also desire to change a partner where there is a personality conflict between the partner and one or more of the issuer's executives that is impeding communication during the audit. However, the presumption created by the Proposed Rule - that removal of a partner from an audit engagement reflects an issuer's improper attempt to influence the audit - will likely make audit firms unwilling to consider even beneficial changes to avoid any later assertion that the audit firm violated the final rule by acceding to the client's desire for a change in audit partners. An audit firm could legitimately be concerned about defending these types of personnel decisions to the Commission or other regulators, not because the change was not well-grounded and improved audit quality, but because such a change would presumptively raise the specter of suspicion that the audit firm had succumbed to client pressure. As a result, partners that are not the optimal choice for a particular engagement might be kept on the engagement to the detriment of audit quality, or the auditor would be left with no option but to resign, which may not be in the best interests of the issuer's shareholders. The examples of "fraudulent influence" that are provided by the Commission in its release could make positive and benign conduct presumptively suspicious. As a result, we believe that the Commission should abstain from listing any specific examples of conduct that would be presumptively prohibited under the final rule. In any event, the Commission should make it clear that the conduct described requires fraudulent intent. Additionally, the final rule should specify that there is no presumption that any listed conduct is violative of the Act. Auditors, issuers, and the Commission itself are all better served when the rules contain the necessary clarity to define what conduct is prohibited and what is not. Although there will inevitably be "gray areas," the requirement of a fraudulent purpose will give the Commission ample authority to handle real instances of misconduct. C. "At The Direction Of" Language Should Not Apply To "Partners Or Employees" Of The Same Audit Firm The Commission has stated that it interprets Congress's use of the term "direction" in Section 303 to "encompass a broader category of behavior than `supervision.'" Persons whose conduct might be considered at the direction of officers and directors for purposes of the Proposed Rule would include "customers, vendors or creditors" as well as "partners or employees of the accounting firm." We agree with the Proposed Rule's broad interpretation as a general matter that liability under the Proposed Rule should reach conduct by persons acting as agents of management or at the direction of management with the intent to cause the auditor to issue a false audit report. However, we believe that the Commission should limit this interpretation so as not to cover auditors and employees within the same audit firm. The inclusion of "partners or employees of the accounting firm" would sweep up an unlimited number of internal interactions between auditors that could adversely affect the conduct of an audit. For example, a client may speak with a tax partner in the firm that also audits the issuer to discuss a different view about the accounting for taxes associated with a transaction as determined by the audit partner. When the tax partner conveys the client's reasoning and the audit partner later decides, in the exercise of professional judgment, that the client's position is valid, that appropriate interaction could be deemed to be an impermissible coercion "at the direction of an officer or director" of the issuer, because the tax partner may be deemed to have been improperly influenced. Such an interpretation would be harmful because it would inhibit the essential consultations that employees within an audit firm engage in on a daily basis. Indeed, this interpretation of the Proposed Rule would impede on auditors' duty to inform and consult engagement partners about an audit because audit partners may have to be skeptical of the information given to them by others within the accounting firm. However, we believe that it is proper to sanction an individual partner who seeks to "fraudulently influence, coerce, manipulate or mislead" his audit partner, but adequate remedies for this scenario already exist under the securities laws. See Section 10b-5 of the Exchange Act Rules. Because of the potential dangers posed by including partners and employees of an audit firm within the application of the "at the direction of" language of the Proposed Rule, we believe that they should be excluded. D. The Interplay Of The Proposed Interpretations Serves To Multiply The Adverse Impact on Communications During the Audit Process Each of the issues raised above, individually, will inhibit the forthright relationship between auditors and their clients that is necessary for the completion of an audit in accordance with generally accepted auditing standards. Together, they will have a cumulative effect that may be exponentially greater than any harm caused by each individual item. For instance, as noted above, the elimination of any scienter requirement in the Proposed Rule might be more harmful when it is combined with the inclusion of partners and employees within the same audit firm. Thus, the Commission should not view each of its proposed interpretations in isolation, but should consider the interaction of the various components of the Proposed Rule and the increased risk to audit quality they pose collectively. II. The Proposed Rule Inappropriately Changes Scope Of Services Restrictions That Were Carefully Constructed By Congress Title II of the Act permits auditors, with some restrictions, to provide non-audit services to their audit clients when pre-approved by the audit committee. In drafting this section of the Act, Congress carefully balanced what it perceived to be the implications that the provision of non-audit services may have on the independent audit function with the benefits those services provide to corporations and, indeed, provide in enhancing the quality of auditing services that serve investors. However, the Proposed Rule threatens to upset this balance through the broad effect of its provisions. The Commission should tighten the scope of the Proposed Rule so as not to upset Congress's considered judgment as to the scope of services auditors may provide, particularly when the Commission has commenced a separate rulemaking to address scope of services issues directly. The Proposed Rule effectively places a presumptive, and likely unintended, ban on almost all non-audit services provided by an auditor to its client. This result is the product of the expansive reading of two provisions of the Proposed Rule given by the Commission. The first component has been discussed briefly already: the Commission's characterization of offering a contract for non-audit services as an example of "fraudulent influence." Without elaboration, this provision may create the presumption that any time a client hires its auditor's firm to provide non-audit services approved under the Act, it could be found to be attempting to fraudulently influence its auditor. Because a contract for non-audit services appears to carry this presumption of guilt, clients will likely be wary of hiring their auditors to provide any services except for those that are strictly deemed "audit services." Conversely, the audit firm may well fear accepting appropriate non-audit contracts for fear of being perceived to be improperly influenced. As Congress determined in passing the Act, this would be unwise. Audit firms provide many "non-audit services" that benefit corporations and their shareholders, and, in fact, often supplement and enhance the audit process. If a separate firm has to be hired to provide all of these services, valuable and significant efficiencies might be lost, and the cost of the delivery of such services will unnecessarily be increased for corporations and shareholders. At a minimum, the Commission should emphasize that such an offer would only be a violation of the Proposed Rule if it was part of an express quid pro quo between the client and the auditor. The second component that adds to the potential specter of presumptive guilt in providing non-audit services is that the applicable time frame in the Proposed Rule is virtually all-encompassing. The proposing release explains that the definition of "engaged in the performance of an audit" would not only include the actual period when the audit is being conducted, but also "negotiations for retention" as well as a period "after the professional engagement has ended," when the audit firm could be asked to consent to the reissuance of its prior report. In other words, the definition would encompass a significant period before and after the actual audit engagement. Because this period is so broadly defined by the Commission, auditors and issuers might not be able to determine with any confidence when an offer to provide non-audit services would be suspect under the Proposed Rule. At a minimum, the Commission should clarify that its broad reading of "engaged in the performance of an audit" is solely for the purpose of Section 303 of the Act, and does not extend to other Commission regulations (for example, those pertaining to auditor independence in Regulation S-X). The broad application given by the Commission to these provisions of the Proposed Rule may have the unintended consequence of prohibiting auditors from providing those non-audit services permitted under the Act. Congress considered and rejected such an outright ban on the provision of non-audit services in favor of a more balanced approach. The Proposed Rule, as currently drafted, could significantly disturb Congress's careful efforts reaching an appropriate balance of scope of services restrictions in the Act. III. The Proposed Rule Creates Uncertainty As To Auditors' Responsibility Under Section 10A Of The Securities Exchange Act Of 1934 The Proposed Rule does not address what auditors' obligations would be with regard to reporting "coercive, manipulative, or misleading" conduct by their clients. The combined effect of the Proposed Rule and existing duties under Section 10A could mean that auditors will be compelled to act, not as auditors, but as informants under the Proposed Rule. Given the inherent vagueness in the definition of the conduct prohibited by the Proposed Rule, auditors and audit firms will face uncertainty as to when client behavior rises to a level qualifying it as an attempt to improperly influence, coerce, manipulate or mislead. Given the ambiguity inherent in defining such terms and the subjective nature of those determinations, it is difficult to see how auditors will be able to decide when the client's conduct has "crossed the line" and has triggered a Section 10A issue.1 Is it when the chief accounting officer, or chief financial officer, or the controller disagrees with the auditor on a matter of judgment; or when he or she disagrees vehemently, or in a particular tone of voice? And if the auditor is offended by the officer's tone, what remedial action, short of the officer's dismissal, will adequately remedy the situation? And how will the auditors reach such decisions with uniformity? This is a duty that goes far beyond the requirements of the current law and is not what we believe Congress intended. This is especially true given the potentially broad scope of liability inherent in the Proposed Rule, which will make it difficult for even the most experienced auditor to determine what is and what is not prohibited conduct. The Proposed Rule provides an auditor no clear guidance as to what conduct constitutes "coercion" as currently interpreted by the Commission, such that a Section 10A reporting obligation arises. We believe the Commission needs to clarify when and if auditors should report conduct under the Proposed Rule in accordance with their obligations under Section 10A. IV. Technical Corrections We suggest technical corrections for the following items:
V. Conclusion Though the spirit of the Proposed Rule is laudable, the imprecise and expansive terms it utilizes could detrimentally impact the quality of independent audits. The Commission is surely, and should be, mindful that it is classifying behavior as unlawful, with serious consequences that will flow therefrom - including sanctions, increased risk of private litigation, impairing the audit process, and decreasing the willingness of qualified individuals to serve on audit committees. Issuers and auditors should be given fair notice as to what behavior will be sanctioned and what behavior will be permitted in order to establish an effective and equitable enforcement regime. Such notice can only be provided by giving precise definition to the operative terms of the Proposed Rule. As currently drafted, the overly broad application of the Proposed Rule threatens to create a "chilling effect" that could undermine the candid and professional client-auditor communication that is central to an effective audit. Additionally, the Proposed Rule, by potentially mischaracterizing the provision of non-audit services as a source of "coercion," could likely have the additional and unintended effect of curtailing auditors' ability to provide any non-audit services to audit clients, contrary to the Act and the intent of Congress. The Commission should seek to clarify the terms of the Proposed Rule to limit their application only to conduct that is actually intended to corrupt an audit, instead of casting a wide net that would make innocent and beneficial conduct unlawful. We believe that a final rule in which the operative terms are clarified in such a manner as to limit their application to conduct that is actually intended to corrupt an audit will be an important step in restoring public confidence in financial reporting and help auditors to perform their important functions. We look forward to the Commission's final rule that meets this objective. If you have any questions about our comments please contact Sam Ranzilla at (212) 909-5837 or Melanie Dolan at (202) 533-4934.
____________________________
|