The concept of audit independence is fuzzy; the rules governing it are complex and burdensome; and a reexamination is long overdue.

Fortunately, the reexamination has begun. An AICPA task force is reconsidering the flamework of independence; the SEC has a concepts release pending; and last year the six largest accounting firms issued a white paper on the subject. Fresh approaches to changes in the rules are possible. With that in mind, we address the concept of audit independence and its application.

The Concept

The profession's authoritative literature on audit independence does not contain an unmistakably clear concept for establishing independence rules. By an unmistakably clear concept, we mean one that consistently defines what is included and what is excluded from the category audit independence.

There are, of course, many rules designed to judge independence in specific situations, but these kinds of rules are too rooted in the specific situations to be applicable to all other situations. To take a simple example, the rule that an auditor may not acquire any direct or material indirect financial interest in the auditee cannot tell you whether close familial relationships with client executives are compatible with independence. The rule about financial interests therefore fails to define consistently what is included and excluded from the category of audit independence.

The AICPA's ethics code discusses objectivity along with independence, but in a way that raises more questions than it answers. Objectivity is treated as achievable in pan by avoidance of conflicts of interest. However, the code's independence requirements also detail all sorts of interests that are incompatible with independence--in other words, that create conflicts of interest. The code even states that "the maintenance of objectivity and independence requires a continuing assessment of client relationships," which suggests that both involve avoiding conflicts of interests. Thus, independence and objectivity are treated both as separate concepts and as sharing a central identifying feature.

The language on independence in the auditing standards holds that the auditor "must be without bias with respect to the client .... "The idea of being without bias seems like a good starting point for an independence concept, but we must ask with respect to what should the auditor be without bias.

The question arises initially because of the undefined term "the client." The client could be defined as management, or it could be the owners, or it could be both. The financial interests of stockholders can be different from those of management, as when managers want to hide poor performance from those who could demand their replacement. However, these sets of interests can also be similar, as when management reports higher earnings than are warranted and thereby obtains financing needed to continue operations. Being without bias toward management would therefore be insufficient if there were bias in favor of the stockholders' financial interests.

Nevertheless, widening the definition of the client to include the board and shareholders doesn't do the job. It leaves open the possibility that the auditor could have a bias with respect to other parties in the financial reporting process-for example, potential investors. A bias in favor of such parties might work this way: Suppose the auditor had stock in a company that was negotiating to purchase the auditee. The lower the reported earnings and asset values, the better the price for the purchasing company.

One could respond to the exception just cited with an additional level of generalization--namely, that the auditor should be without bias with respect to all parties in the financial reporting process. But that would call for an openended list of such parties, including, for example, the SEC. More importantly, some "parties" on the list would have to be divided. For example, as we have noted above, the financial interests of stockholders could be identical to those of a management group that inflates earnings in order to obtain financing needed to keep the company in business, but every shareholder might not be so mercenary. Some may stand on principle if given the option. The mercenary and the principled stockholders would have as much right to separate entries on the list as, say, management and the board of directors.

A Way to Revision

The discussion thus far shows that selecting characteristics that might satisfy the criterion for an unmistakably clear concept of independence eventually entails testing them against possible circumstances that could occur on an audit. An audit involves a preparer, an auditor, the auditor's report, and a written assertion (the financial statements). We have already discussed the possibility of basing the concept of audit independence on absence of bias with respect to the client, which would include the preparer. The only possibility remaining is the written assertion.

What does it mean to say that the auditor is independent of the assertion or without bias with respect to the assertion? It means, that there is no bias of any kind with respect to the assertion, and it also means, by inference, that there are no interests causing bias. However, it makes no sense to ban the auditor's interest in the reliability (or truth) of the assertion. Therefore, the auditor should have no interest in the financial statements except their reliability.

Note that the idea of absence of bias with respect to the assertion makes it easier to appreciate why auditors should not have material stock holdings in the auditee. By being a stockholder, the auditor's relationship to the financial statements changes. He or she can benefit or suffer from what those financial statements say, no matter what the moral stance of other stockholders or the predilections of management. Although the interests of specific parties in the financial reporting process need not be consistent, the auditor's relationship to the assertion is a constant. If it is unbiased, independence is present.

Relationships to Objectivity,

Integrity, and Economic Interest

The definition of audit independence offered above is a version of objectivity in that being without bias is an accepted definition of objectivity, and we have defined audit independence in terms of a condition that creates absence of bias. However, the definition of independence above differs in three ways from the definition of objectivity. First, it focuses on the absence of interests that could create bias, rather than on the absence of bias itself. Second, it explicitly allows and implicitly requires a prejudice toward the reliability of the assertion. Third, it is bounded by the auditor's relationship to the assertion. Being without bias on any other matter is not at issue, and being without bias in nonaudit circumstances is not at issue.

Integrity is separate from independence as we have defined it, but it also bears on the quality of the audit by affecting potential bias. Ks the professional guidance says, integrity is an element of character. The dictionary gives honesty as a synonym, and the word also means incorruptible. Therefore, if an auditor had perfect integrity, there would be no need to be concerned about independence; bias could not breach that perfect integrity. For audit purposes, perfect integrity would be fungible; it would serve on any engagement. Independence as we have defined it depends on the auditor's relationship to the specific assertions under audit.

While perfect integrity cannot be assumed, neither does it follow that auditors are shy the normal measure. This means that the public is protected from bias during an audit both by an auditor's independence and by his or her integrity. Neither must be perfect for the public to be fully protected because they complement one another to ward off bias.

The suggested concept of independence would preclude interests that enable the auditor to benefit economically from the assertions under audit and thereby be a source of bias. However, every auditor also has economic interests that weigh against the introduction of bias. There are penalties for performing substandard (biased) audits, including the potential loss of livelihood. Like integrity, such economic interests complement independence, providing the public with additional protection against biased auditing.

The economic interests weighing against the introduction of bias go beyond the auditor's personal economic interests. Audits are typically conducted by firms, and firms have enormous interests in the independence of their members. Their reputations are their most valuable assets, and they are also subject to penalties, such as awards to plaintiffs from litigation and SEC prohibitions from accepting new clients for some specified period. Every auditor is aware of joint and several liability, the importance of the firm's name, and the potential effect of his or her performance on the firm as a whole. No rational partner responsible in any way for the audit independence of others in the firm would treat the economic interests of the individuals assigned to the audit as the sole interests at risk.

Appearance of Independence

Thus far, in describing the concept of independence we have ignored the concept of appearance of independence. There is a place for appearance of independence in a conceptual structure on audit independence, but not as the separate coequal of the fact of independence, i.e., not as the notion that the auditor "should be independent in fact and in appearance." The role of appearance of independence should be limited to determining the fact of independence. This is necessary because the relationships that could indicate an interest other than the reliability of the financial statements may become complex or distant. How far removed from the auditor, for example, must a family tie to a member of management be not to have an influence on the performance of the audit?

A traditional approach to resolving such issues is to apply the reasonableperson concept. Under this concept, the auditor's relationships are judged by whether a reasonable person "having knowledge of all the facts and taking into consideration normal strength of character and normal behavior under the circumstances" would conclude that the relationship presents an unacceptable threat. In other words, the criterion is whether reasonable persons would conclude that circumstances indicate (i.e., appear to them to indicate) an unacceptable risk to independence. This is very different from treating the that base prohibitions or permissions on the appearance of independence--for example, the receipt by an employee or partner of a gift or other unusual consideration from a client.

Applying the Concept

The concept of independence is applied through two basic activities, setting rules to govern the behavior of individual auditors and their firms and administering the rules, which includes applying the concept to situations not explicitly covered by the rules. Unfortunately, the rulemaking and administrative procedures now in place are marked by overlapping jurisdictions.

The SEC sets requirements for independence on SEC engagements and responds to queries about acceptable practice. The AICPA also sets independence requirements and responds to queries. Moreover, its quality control standards obligate every firm to adopt policies and procedures that provide reasonable assurance of compliance with professional requirements on independence. The same standards obligate each firm to evaluate the application of its system of quality control, which includes its independence policies and procedures, and to perform timely evaluations of the appropriateness of the policies and procedures that constitute its quality control system. Reviewers under AICPA practice monitoring programs triennially evaluate the design of each firm's quality controls (including independence policies and procedures) and compliance with those controls. A Public Oversight Board monitors the effectiveness of the peer review program of the AICPA's SEC Practice Section, which includes all firms with AICPA members in SEC practice, and the SEC monitors it too, even though its monitoring relies significantly on the work of the Public Oversight Board.

Thus, for each firm in SEC practice, there are typically at least three sets of requirements--the SEC's, the AICPA's, and the individual firm's. However, there may be more. Ks pointed out in the AICPA Guide "Quality Control Policies and Procedures for CPA Firms," independence requirements may also originate from state CPA societies, state laws, state boards, and regulatory agencies other than the SEC. Indeed, John L. Carey's history of the profession recounts a period of several years during which the rules of the state societies of Illinois and New York, following the lead of the SEC, were clearly more stringent than those of the AICPA. The Model Accountancy Bill jointly advocated by the AICPA and the National Association of State Boards of Accountancy explicitly gives state boards authority to set rules of professional conduct with respect to "independence, integrity, and objectivity." To promote uniformity, NASBA formulates and publishes a Model Code of Professional Conduct that contains rules on independence and on integrity and objectivity.

The multiple sources of independence requirements suggest that redundancy is a problem. It would help if a greater number of independence provisions of codes that are virtual repetitions of material in the rules of other bodies either were reduced to a reference to the imitated rules or openly acknowledged wherein the provisions are designed to capture the same letter and spirit. The independence provisions of the ethics codes adopted by the state boards of South Dakota and Tennessee, for example, make clear their relationship to NASBA's model code. No one gains from the pretense that rules intended to be the same or virtually the same have distinguishing qualities.

That would be minimal rationalization. A more thorough approach would be to trim the sources of independence rules. This can be done de faao even when retaining overlapping jurisdictions by one party delegating the practical exercise of its authority to another.

There is a precedent in the case of accounting principles for the SEC to delegate the practical exercise of some of its authority to a private-sector body--in this case, the FASB. The Commission exercises active oversight and influence, and gives up none of its statutory authority over the accounting principles used by registrants.

To follow that precedent, the differences between the Commission's independence requirements and those of the AICPA would first have to be resolved, but the task does not seem insurmountable. The primary differences are over bookkeeping services, the degree a retired partner is "paid out" that eliminates him or her from independence requirements, those non-audit services judged by the SEC to involve a mutuality of interest, and joint ventures and contracttor/subcontractor relationships.

The SEC-FASB precedent is evidence that different arrangements on independence are possible. A wider range of options, more thorough in their rationalizing effects, would include possible changes in the relationship between the rulemaking and administrative functions. In order to illustrate these possibilities, we shall first try to put in perspective the difference between events covered by the rules and those not covered by the rules. The difference originates in the fact that no set of rules, no matter how detailed, can explicitly cover all the possible fact situations. The barest exercise of imagination demonstrates that possible circumstances can create fact situations in virtually endless variety. Is the brother of a deceased stepfatherin-law a close relative for purposes of determining the independence of an auditor with a family relationship to a member of top management of a client company? The cousin of that stepfatherin-law? Does it matter how recently the stepfather-in-law passed away? How far away from the auditor he resided? Whether the auditor had seen the brother or cousin socially or how frequently? It is no surprise that the reasonable-person approach was developed to deal with such situations or that the judgments that emerge must appear to some either mincing or expedient.

These are mutants of generic situations that can lead to bias with respect to the financial statements. The generic situations are fairly obvious. No Solon or Solomon is needed, for example, to ban immediate family relationships with the top management of a company under audit, or direct and material indirect financial interests in a company under audit, or an auditor taking on the role of management. So we face two major classes of fact situations, the general and obvious, on the one hand, and the more specific and subtle, on the other.

This has implications for the balance between the two basic activities by which the concept of independence is applied, i.e., setting rules and administering them. It suggests that a limited number of generic prohibitions can satisfy the need for rules and that to incorporate more specific circumstances into the rules will sooner or later create a code that is unwieldy. It also suggests, because many of the specific and subtle judgments will inevitably appear arbitrary, that the more detailed the code, the less respected it may become. This prospect is reason enough to consider the possibility of a code devoted to generic situations and an administrative system at the firm level to determine whether specific circumstances violate the code.

The SEC and the AICPA are prime candidates to set rules governing the generic situations, because both are national in their scope. However, the SEC has no authority over the independence of the auditors of non- registrants. This weighs in favor of having the AICPA set the generic rules, because the AICPA has authority over its members who audit non- registrants. The SEC, of course, would have to evaluate whether delegating the practical exercise of its authority to the AICPA would satisfy its statutory obligations under the securities acts without regard to the effects on auditors of non-registrants and their non- registrant clients. If the AICPA were the sole rulemaking body on independence, the situation would parallel what already exists with respect to accounting standards, where the FASB's standards apply to both registrants and non-registrants.

All this covers only the formal exercise of rulemaking authority. The SEC would exercise oversight and influence; there would be communications between the two bodies; and the AICPA would be aware that the SEC's dissatisfaction could lead it to assume full exercise of its authority over the independence of auditors of registrants.

Under this model, each accounting firm would be responsible to administer the generic independence rules set by the SEC or AICPA, making decisions on fact situations not explicitly covered by those rules. Each firm would be free to supplement the generic independence rules with additional requirements for their professionals, and peer review would provide what public assurance is needed.

This model, which we can call the devolved model, would not be as drastic a change as at first it might appear. Assurance that independence is in place already rests largely with the firms and peer review, and most of the administration required to ensure the fact of independence lies with individual firms. Firms, for example, must inform their people of new clients whose retention affects what securities are permissible investments. Firms must obtain information on family relationships so that they can judge the propriety of personnel assignments based on that information. When questions arise as to the application of the principles, the SEC or AICPA is consulted, but that is exceptional rather than regular.

Thus, the bulk of the devolved model is already in place. If the AICPA and SEC, or one of the two, were to restrict their role to issuing fundamental principles without treating interpretations of individual fact situations as if they too had the power of rules, the devolved model would be in place.

A fully devolved model would allow the firms to set generic as well as any supplementary independence rules and to administer both. The SEC and AICPA would again rely on peer review.

The devolved models would encourage firms to compete to establish more efficient and effective technologies to administer independence requirements. However, both devolved models presume that firms are equipped to set rules and administer them and would accept the expense that goes with it. Some small firms or individual practitioners may not be in that position. On the other hand, all firms would have access to the independence codes and interpretations that have been issued over the years; peer reviewers would suggest improvements when opportunities for them are noticed; and the AICPA could retain the capacity, to help those that wished to avoid the expense of going it alone. Such assistance could include not only generic rules, but also a procedures manual for administering them and technical personnel to respond to queries.

The devolved models are only two of what is likely to be a more numerous group that may be considered. An examination of practices in other countries might yield others that could improve the haphazard appearance of the structure now in place.

ALL WILL BENEFIT

The concept of independence could be much clearer, and the system by which rules are set and administered is inefficient. Although there is no evidence that for these reasons the public interest is at risk from lapses in independence, the inefficiencies are a burden on practice and the haziness of the concept is professionally embarrassing. These conditions are incompatible with the importance of independence to the profession. All interested parties would benefit from an effective reconsideration of the concept of independence and rethinking of how independence rules are set and administered.

Robert Elliott CPA, is a Partner in the executive office of KPMG Peat Marwick in New York Mr. Elliott is a member of the AICPA NYSSCPA and the AAA

Peter D. Jacobson, PhD, is Senior Editor in KPMG Peat Marwick's executive offce in New York