INTRODUCTION

The issue of auditor independence is not new. Mayhew and Pike¹ state that the problem stems from a lack of clarity about one issue; namely for whom does the audit firm really work? They pose the question, 'does the auditor work for the investing public as intended by the securities acts of 1933 and 1934, or alternatively, does the auditor work for the client firm?' According to the authors, this lack of clarity has resulted in auditors' independence being impaired. It is difficult, however, to point to tangible evidence supporting this view. For example, there are no Securities and Exchange Commission (SEC) enforcement cases addressing independence as an issue in the recent financial reporting failures including WorldCom. There remains, however, a general perception that impaired auditor independence is partially responsible for recent financial reporting failures and investor losses. Concerns have long been expressed that auditor provision of consulting services to clients results in actual, as well as perceived, loss of independence (eg, see Kleinman, Palmon and Anandarajan²). Arthur Andersen, the eponymous founder of the Arthur Andersen accounting firm, pithily summarised the concerns raised about auditor provision of consulting services in 1934. He said, Some regard [this] movement as wholly unsound, the opening of a territory into which the accountant could not venture except at great risk of loss of professional standing. Others saw in the movement an opportunity for greater and more constructive service to business. (cited by Sweeney³).

Whether from the auditor's or society's vantage point, the fundamental problem remains the same, that is, what is the trade-off between service to the society as reflected by high-quality audits and the real or perceived threat to that independence arising from, or being exacerbated by, auditor provision of nonauditing services. Such threats are in addition to those stemming from long auditor–client association, incentive structures within CPA firms that reward client retention, and the like (eg, Kleinman, Palmon and Anandarajan²). It has also been argued that the fundamental problem of auditor independence arises from the fact that, in the United States at least, the client hires, pays and can displace the auditor. While this freedom is somewhat restricted by potentially negative publicity owing to SEC-required auditor change reporting requirements and suspicions as to why the auditor may have been displaced, the fact that the auditor serves at the pleasure of the client remains, despite recent reforms regarding the much greater empowerment and restructuring of audit committees.

As Norris⁴ documents, alleged breaches of independence due to profitable business relationships with clients can be costly to the CPA firm (Ernst and Young, in this case). As the Enron and WorldCom episodes further document, independence loss may be catastrophic for client employees, shareholders, creditors, and the markets and society generally. The collapse of these two huge firms further sensitised the audit world (see Sweeney³), the corporate world,⁵ and the SEC to the risks to auditor independence that are posed by consulting practices. In 2000, then SEC chairman Arthur Levitt instigated a renewed evaluation of auditor independence.⁶ This led to the SEC's issuance of new independence rules and disclosures in November 2000.⁷ After the collapse of Enron, and the fury over the level and nature of the involvement between Arthur Andersen and Enron, as well as revelations about the internal processes at Andersen with regard to retaining Enron as a client, this issue was revisited in both houses of the US Congress. The subsequent accounting scandals involving WorldCom became the trigger for adoption of the allegedly draconian, but important, Sarbanes–Oxley Act (SOX).

Much has been written about Section 404 of the SOX, especially the great expense that it inflicts upon SEC registrants preparing their annual reports. We note that it is not the preparation of the reports, however, that increases the expenses, rather it is the requirement to document the controls over financial reporting and the concomitant certification that has impacted the corporate bottom line. Had the management been carefully monitoring its control systems previous to the enactment of Sarbanes–Oxley, the expense incurred to improve the impacted systems would not have resulted. Less remarked on, though, is whether the restrictions that SOX imposes on the auditing profession in its relationship to its clients have significantly improved auditor independence. It is difficult for researchers to address this question because researchers require empirical evidence rather than intelligent surmise to support a position.

The real question then is whether the firms acted to shore up independence, and if so, what actions were taken. Useful studies here could involve researchers examining (by means of case studies) the steps audit firms took to enhance independence. A major criticism by practitioners is that the independence discussed in the academic literature relates to 'perceived independence'. Clearly, a perception of impaired independence may not mean actual independence impairment. Also, clearly, studies documenting the behaviours of CPA firms with regard to independence maintenance should shed useful light on this issue. Currently, no research to date has attempted this, perhaps due to the politically charged nature of the issue.

We seek to shed some light on auditor independence-related issues by asking: What has the academy learned about SOX's impact on auditor independence in the five year's since its enactment? Have researchers even examined key ramifications of SOX's requirements? If not, what areas need scrutiny? These are the issues considered here. In essence, we examine whether there is a gap between what regulators want to see addressed and what researchers have actually looked at with regard to auditor independence in the post-Sarbanes–Oxley period. We also suggest that insights into the structure of US auditing may be gleaned by examining apparently successful auditing innovations overseas. We reserve this for the end because the focus of our paper is on the domestic audit market, its opportunities and constraints. In terms of future research however, developments overseas offer a rich source of potential alternative structures that may be worth investigating for application here.

We first define and discuss the importance of auditor independence. The SEC defined independence in a November 2000 release that provided general guidelines with respect to tasks auditors could undertake for their clients. These guidelines were designed to eliminate practices and service lines of business that constrain auditor independence. SOX then focused on and addressed what auditors could do with respect to nonaudit services. The time between passage of SOX and this writing is brief. It remains important, however, to examine the extent to which accounting researchers have provided insight into the SEC and other regulators on the efficacy of these guidelines. We also highlight areas that, although considered important by the SEC, have not been addressed by researchers.

WHY IS THE ISSUE OF AUDITOR INDEPENDENCE IMPORTANT?

Prior to the Sarbanes Oxley Act (2002), management frequently sought to highlight their short-term successes because of incentive plans in place. (Scholarly literature has explained this behaviour using agency theory.) Auditors are supposed to act as safeguards to investors by preventing such self-seeking behaviour. Prentice⁸ notes that there are several problems unique to the United States. First, auditors are paid by the firm's managers rather than by the stockholders. The second problem, according to Prentice, is that managers may attempt to buy off their auditors by giving them short-term incentives (eg, increasing their total fees by providing them with consulting work unrelated to their audit).^{9, 10} There is another theory that that audit partners would not want consulting work because of their concern over the perception of independence. Further, the Public Oversight Board (POB) required this matter to be addressed by the firm and audit committees on an annual basis. No research has attempted to reconcile these two views. This is another potential field for research that would be illuminating to both academics and practitioners alike.

WHAT IS AUDITOR INDEPENDENCE?

What is auditor independence? The SEC in November 2001 defined independence from a conceptual viewpoint. (See site: http://sec.gov/rules/final/33-7919.htm that has the November 2001 release.) The definition is primarily based on that provided by Mautz and Sharaf.¹¹ The latter define independence as a characteristic of virtue of a good practitioner, who must be independent of any interest that might affect his/her judgment. Kleinman and Palmon¹² review different conceptions of auditor independence and conclude that the definition, and certainly achievement, of independence is profoundly problematic since the objectivity and integrity of judgment, for example, that the The American Institute of Certified Public Accountants (AICPA) describes as being at the heart of independence is impossible of realisation. In Kleinman and Palmon's view, while 'true' independence is extremely difficult given the wide range of contradictory, social, professional, financial and legal pressures that face the auditor and the audit firm, there may be a range of deviations from true independence that the auditor may be affected by that is not large enough to lead the auditor to behave in a nonindependent manner. Thus, while thoughts and emotions may be engaged in a way that leads the auditor to cognise in a nonobjective manner, his/her behaviour may not be affected.

Conceptual confusion aside, independence is one of the central values upon which the auditing profession's legitimacy is predicated.¹³ In other words, independence cannot be compromised. Francis¹⁴ argues that an auditor's ability to be independent has been eroded by changes in the environment that commenced in the 1970s. For example, the Federal Trade Commission's 1978 actions encouraging competition for clients among audit firms led to an erosion of standards of conduct regarding client solicitation. Increased competition, while normatively a good thing, may have led to increased pressure on incumbent audit firms to compromise their standards in the hope of retaining clients. Accordingly, we believe that the Federal Trade Commission's action in 1978 exacerbated any previous problems regarding independence that may have afflicted the profession. For example, subsequent to the ruling, it was not unusual for accounting firms to 'low ball' audit fees to get a client. We have heard anecdotal evidence that some firms would offer to perform an (initial) audit for no fee, betting on future growth.¹⁵ One firm was known for not charging an audit fee for start up tech companies. This continued in the mid-1980s, when CPA firms began to view internal audit activities as an expanding service area for both new and existing clients. Most firms were not that successful in providing outsourcing services to audit clients because of the resistance of the internal audit departments and the audit committee's concern over independence. Enron was an anomaly.

Geiger et al.¹⁶ noted that it was perceived that the provision of consulting services may improve audit quality by providing external auditors with considerable knowledge about the client, its operations and its industry. The greater the external auditor's insight into the client, the better their ability to understand business transactions and identify key audit risks. This was also the argument that the AICPA¹⁷ made in its 1997 White Paper. The audit firms that did provide consulting services did so using the rationale that more detailed understanding of the firm and its operations would also enhance the subsequent quality of the audit.

In November 2000, concerned in part by the increasing percentages of audit firm revenue that were earned by the audit firms from performance of nonaudit services for their clients, the SEC adopted a new rule. The latter prohibited accounting firms from providing certain nonaudit 'consulting' services to their audit clients. The rule also requires public companies to disclose in their proxy statements the fees paid to their independent auditors for audit and nonaudit services. The purpose of adopting the rule was to mitigate the threat to independence posed by auditors providing both auditing and consulting services to a client (eg, Schneider et al.¹⁰).

This was not the first time, however, that the SEC addressed this issue. After the Foreign Corrupt Practices Act (FCPA) of 1977, the SEC had registrants report these fee relationships. The AICPA POB required member firms to report the percentage of audit fees to nonaudit fees. This information was also required to be presented to registrants' audit committees. The SEC's November 2000 independence-related action provided more clarity than had existed previously, and updated the requirements.

But was this update to audit to nonaudit fee reporting meaningful? Palmrose and Saul¹⁸ noted that the lack of further SEC action on fees when the reporting requirement was originally instituted (circa 1977–1980s) reflected the failure of published research to demonstrate that nonaudit service provision compromises auditor independence (See also Kleinman, Palmon and Anandarajan²). Palmrose and Saul¹⁸ further note that even in the wake of the Enron affair, there was no clear evidence that the provision of nonaudit services led to an impaired audit. As Schneider et al.¹⁰ state, only the auditor him/herself knows whether he/she retained his/her independence! In the extreme, that is certainly true. Also, Kleinman and Palmon¹² argue, based on a comprehensive analysis of the social psychological, sociological and environmental webs that entangle the auditor, that the auditor may lose his/her independence without being aware that it has been lost (see also Kleinman and Palmon¹⁹). The implication of Palmrose and Saul's¹⁸ argument was that researchers had not provided the SEC with conclusive evidence providing impetus for further action. The SEC, however, conducted its own research. In point of fact, one of the reasons that the SEC rescinded its audit and nonaudit fee disclosure requirement in 1980 was because its sponsored research revealed no conclusive evidence that independence had been impaired. Don Kirk, a member of the POB and former Chair of the FASB, also did an independence project in the late 1990s and found that independence had not been impaired by rendering nonaudit services (see Craig²⁰). While it may seem that the issue is the perception of impaired independence, rather than the fact of independence impairment, understanding what actually happens between the auditor and client is difficult absent catastrophic client failure. Nevertheless, even if we grant that distinction, perception of lack of independence is a serious issue in itself given that the role of the auditor is to serve as a guarantor, of sorts, of the informational integrity of the firm's financial statements.

Given the CPA firms' vigorous efforts to derail the SEC's limited efforts at reining in audit firm consulting services in 2000 (eg, Norris²¹), a great deal of supporting evidence would have been required to sustain the SEC effort. Does this situation pertain today? Have accounting researchers tackled the key issues that would provide fodder for the regulatory bodies' deliberations as to whether to take further action to enhance auditor independence? And, if not, what remains to be done?

THE SOX AND ITS IMPACT ON AUDITOR INDEPENDENCE

In the summer of 2002, as we now well know, there were several significant accounting scandals in the United States. The business failures of Enron, WorldCom, Global Crossing and other large firms and the subsequent admission of key officials that they had deliberately misled investors, led the public to question the role and integrity of these companies' auditors. The auditor of Enron, Arthur Andersen, suffered a criminal indictment (later overturned) that forced the firm's dissolution. These events led the public to strongly question the independence of the auditor. The most far-reaching legislation resulting from the scandals is formally entitled An Act to Protect Investors by Improving the Accuracy and Reliability of Corporate Disclosures Made Pursuant to the Securities Laws, and for Other Purposes. Its short title is the SOX. The SOX also created the Public Company Accounting Oversight Board (PCAOB) to set standards for, and to police the behaviour of, auditors. Further consequences included strengthened auditor independence rules and placing the audit committee, not management, in charge of the auditor's engagement.

The effect of SOX has been far reaching. Not only has it transformed corporate governance in the United States, but also it has had a profound influence abroad. SOX requires publicly traded companies to disclose the amount and nature of fees paid to their external auditors. These disclosures can be incorporated into a company's annual proxy statement or included in its annual 10K filing. The purpose of the fee disclosures is to allow investors to evaluate for themselves whether the proportion of fees for audit and nonaudit services causes them to question the auditor's independence.^{10, 22} (We note that prior to the SOX and subsequent to 1980 when the SEC rescinded ASR 250 and 264, public companies were disclosing the audit fees and nonaudit fees in proxy statements. SOX only formalised this requirement.) In January 2003, the SEC adopted rules to fulfil the requirements of the SOX. In particular, the SEC took numerous actions. We present these under topic headings.

Nonaudit services

The SEC revised the rules relating to the nonaudit services that, if provided to an audit client, would impair the accounting firm's independence. SEC rules also require disclosures to investors of information related to audit and nonaudit services provided by and fees paid to, the auditor. (We, however, note that some public companies were voluntarily disclosing these data prior to SOX) Sections 201 and 202 of SOX require that an audit committee must approve all nonaudit services with the exception of nonaudit services where the total fees amount to less than five per cent of total revenues paid by the audit client to its auditor. (We note that SOX only formalised this requirement. POB required that nonaudit services be disclosed to audit committees. Most audit committees required that nonaudit services be approved prior to auditors performing the services. The reasoning was to address whether independence of the auditors would be impaired.) In particular, Section 201 of the SOX lists nine nonaudit services that would potentially impair the firm's independence. (Most of this information was also addressed in the November 2000 SEC independence release.) The prohibited services (unless approved by the audit committee) include:

• Bookkeeping or other services related to the accounting records or financial statements of the audit client.
• Financial systems design and implementation. The new rules prohibit the accounting firm from providing any service related to the audit client's information system unless it can be proved that the results of these tests will not be subject to audit tests.
• Appraisal or valuation services including any process of valuing assets that could be subject to audit tests.
• Actuarial services involving the determination of amounts recorded in the financial statements and related accounts that could be audited.
• Internal auditing that relates to the audit client's internal accounting controls, financial systems or financial statements.
• Auditor acting, temporarily or permanently, as a director, officer or employee of an audit client, or performing any decision-making, supervisory or ongoing monitoring function for the audit client.
• Expert opinions or other expert services to an audit client for the purpose of advocating that audit client's interests in litigation or in regulatory proceedings.

Section 202 of SOX requires public disclosure of nonaudit services approved by the audit committee and requires that all fees paid to the accountant be disclosed. (Again we note that SOX only formalised this process.)

Partner rotation

It is also required that certain partners on an audit engagement team rotate after no more than five to seven consecutive years, depending on the partner's involvement in the audit (certain small accounting firms are exempt from this requirement). This had been a POB requirement since the 1980s until the POB's demise. SOX just adopted the POB rules. An audit partner is defined as a member of the audit engagement team who has responsibility for decision making on significant auditing, accounting and reporting matters that affect the financial statements.

Audit engagement team

An accounting firm would not be considered independent from an audit client if certain members of the management of that issuer had been members of the accounting firm's audit engagement team within the one-year period preceding the commencement of audit procedures. (See SEC November 2000 release, which also had this rule.)

Additional auditor compensation

An accountant would not be considered independent from an audit client if any audit partner received compensation based on the partner procuring engagements with that client for services other than audit, review and attest services. (This would be akin to a contingency fee, which is prohibited.)

Audit committee approval

Require that the company's audit committee pre-approve all audit and nonaudit services provided to the issuer by the auditor. The audit committee now also has a major role to play in the hiring and firing of auditors.

Additional provisions of SOX were put in place to establish code of ethics for management to ensure greater independence (Sections 305 and 406). In addition management is required to test the controls over financial reporting (Section 404). The confluence of these provisions with the auditor–client relational aspects of SOX together provide further insulation between the auditor and its findings and the power of management to displace or threaten the auditor. The confluence of these provisions, as well as SOX's insistence on greater board independence in its own right, act to mitigate, if not remove, what Arthur Levitt called a 'fraternal culture' on the boards of directors. Instead, Levitt went on to note that a 'culture of skepticism' was developing on corporate boards.²³ The success of these efforts is noted by the number of CEOs who are being ousted by their now more assertive boards (see Barrionuevo²³). Our personal belief is that this fraternity still exists.

WHAT ARE THE FINDINGS OF CURRENT RESEARCH IN THE POST-SOX PERIOD?

Initial research sought to examine investor perceptions of auditor independence in the post-SOX period. See Table 1 for a summary of research.

Table 1 - Selected post-SOX studies related to auditor–client relationships.

Full table

In a survey involving loan officers, Geiger et al.¹⁶ found that respondents viewed the influence of SOX on auditor independence negatively. Respondents' perceptions of auditor independence and financial statement integrity were negative, thus reducing their willingness to grant a loan. Hodge²⁴ observed that the nonprofessional, individual investors in his sample perceived auditor independence to have declined over time, even in the post-SOX period. Given the recency and vividness of the Enron collapse, it is doubtful that auditors would be seen in a positive light, especially with Andersen's disgrace still gracing the newspapers' front pages. Drawing conclusions about the adequacy of SOX so quickly after Enron, therefore, is very problematic. In a pre-SOX survey, Hussey and Lan²⁵ surveyed finance directors and concluded that they, overall, agreed that prohibiting nonaudit work by auditors would significantly enhance perceptions of auditors' independence, with the respondents still pessimistic. The additional constraints placed on audit firms through SOX, however, and the heightened sense of auditor legal liability that came with the fall of Andersen, however, may result in tempered pessimism in some future replication of the Hussey and Lan study.

Legislators and lawmakers appear to believe that before SOX, auditor independence was widely compromised.²⁶ There are, however, no prior archival studies that provide robust evidence of auditors compromising their independence (eg, see Kleinman, Palmon and Anandarajan²). The next area of research examined different components of SEC legislation to improve auditor independence. In the post-SOX period, Chung and Kallupar²⁷ found a negative relation between absolute abnormal accruals and an office level measure of client importance to the audit firm. This was a study that measured auditor independence indirectly by using abnormal accruals. An abnormal accrual is defined as the incremental accrual over and above the expected accrual under the circumstances. Abnormal accrual was computed as the difference between the actual, total, accrual and the expected accrual. The theory was that higher abnormal accruals reflected lower earnings quality. Lower earnings quality in turn can be attributed to impaired auditor independence. The researchers found a negative association. This implied that lower discretionary accruals were associated with bigger clients. The study's conclusions imply that auditors are more stringent with larger clients, not less stringent as theory espouses. Chung and Kallapur concluded, post-SOX, that existing incentives may be sufficient to motivate auditors to be independent. This may reflect heightened fear of auditor liability given Andersen's fall, as well as unclear guidance from the PCAOB and the SEC regarding AS-2. An alternate explanation for this finding, however, would be that the clients reined in their own impulses to book 'abnormal accruals' given the high level of public attention that Enron brought to large firm accounting practices. Section 404 reviews of controls over financial accounting and Section 302 certification probably fostered this more cautious environment.

If so, then it is unclear whether the results stemmed from (a) audit firms becoming stricter; (b) clients being less willing to push the envelope with regard to corporate reporting; or (c) some combination of the above. This is clearly an issue that future research should address. It may be very difficult, however, to gain access to the intra-firm dialogues of auditors and of clients as well as their discussions with each other (see,²⁸ however, for an excellent study of this type).

While the Chung and Kallapur study adopted a macro view, other studies adopted a micro view by examining the influence of different aspects of the recent SEC requirements. One SEC requirement was partner rotation. In an experimental study, Marchesi and Emby²⁹ examined whether duration of tenure with a client would influence a goodwill impairment decision. They placed actual audit partners in two groups. One group was told that they had a long-standing relationship with a client, while another group was told that they were new partners. They found that partners with longer association believed clients assertions and therefore concluded that purchased goodwill was not impaired. Newer partners, in contrast, were more likely to conclude that there was impairment. Marchesi and Emby concluded that within-firm rotation of partners should help improve auditor independence. Gul et al.³⁰ found that, in the presence of fees for nonaudit services, the short-term independence of the auditor seemed to be impaired, but not the long-term independence of the auditor. These results are consistent with the Kleinman and Palmon^{12, 19} theoretical framework for understanding auditor–client relationships.

The bulk of the research to date, however, has examined the impact of the provision of nonaudit services in the post-SOX period. Ashbaugh et al.³¹ found no significant association between firms meeting analyst forecasts and auditor fees. Hence, they concluded that there is no evidence to support the claim that auditors violate their independence as a result of their clients purchasing relatively more nonaudit services. This is corroborated by Ruddock et al.³² Using Australian firms, Ruddock et al. examined whether auditor independence was impaired by provision of nonaudit services. They postulated that if independence was impaired, then audit reports would show less conservatism. They did not find this and concluded that the provision of nonaudit services did not impair the auditor's independence. These results were similar to those of Kinney et al.³³ and DeFond et al.³⁴ Finally, Whisenant et al.³⁵ concluded that economies of scale (and more effective audit due to a better understanding of the client's affairs) were not fostered by auditor consulting service provision.

SOX also attempted to increase auditors' independence from management by assigning authority to hire and fire auditors to the audit committee. Prior to SOX, management exerted a significant influence over auditor hiring and firing. Boards of directors were and still are dominated by management (but see Barrionuevo²³), even though audit committees are now made up of independent directors, not insiders. Hence, this situation enables management to have significant influence over auditor hiring and firing decisions. Even though the Blue Ribbon Committee on Improving the effectiveness of Corporate Audit Committees (1999) argued that this impaired auditor independence and suggested that authority to hire and fire auditors be transferred to the audit committee, this was not done till SOX's enactment. Mayhew and Pike¹ created a laboratory experimental situation in order to shed light on whether transferring the power to hire and fire auditors from managers to investors significantly changes auditor independence. They used auditors and, hence, their study has external validity. They concluded that when managers have reduced power to hire and fire auditors, based on reactions in an experimental situation, the independence of auditors increased. Hence, they provide evidence to corroborate the stance of SOX that reducing the power of managers to hire auditors improves independence. Clearly, other arrangements may exist, for example, auditors being hired and fired based upon relationships with board members and powerful shareholders.

The recent spate of corporate reporting scandals tied to backdating of options grants, however, suggests that management still has many opportunities to manipulate corporate accounts in their favour. With these scandals — many post-SOX — the cry again arose 'Where were the auditors?' The surfacing of these scandals is of too recent vintage to have generated academic research. We note, however, that academic research uncovered the current scandals and triggered the SEC's scrutiny of stock option dating practices. This is an example where management withheld information from the auditor and provided an audit trail that supported management's position. The SEC's enforcement has been against corporate legal counsel and CEO's not, as far as we are aware, audit firms. Management certainly has many other opportunities to collude in hiding transactions from the auditor. Given that three key parties (the boards of directors, which at least nominally are required to approve management compensation and increasingly should be populated by 'independent' directors; top management, the putative recipients of this corporate largess; and the auditors whose responsibility is to ensure that the options are booked appropriately and were granted pursuant to board action (but see Nocera³⁶)) are all involved, this is an issue ripe for study. An interesting question for study is just how far back the practice of backdating options goes.

Lu³⁷ investigated how companies' (a) threats to dismiss auditors and (b) their engagement in opinion shopping influenced auditor independence. The results do not indicate loss of independence by either the predecessor nor successor auditor. This is consistent with older research cited by Kleinman, Palmon and Anandarajan.² While Lu found no impact of opinion shopping and threatened auditor dismissal on auditor independence, it is interesting to explore in detail the reasons for auditor changes since SOX. This is a subject for future researchers.

In earlier literature (see Kleinman, Palmon and Anandarajan² for a review), auditors were often said to lowball auditing bids because they expected to make up any losses on initial year audits in future years since subsequent years auditing costs would be lower, while the audit fee would be the same or higher. Since the low-balling took place when the audit firms had other services available to sell clients, this behaviour no longer has a clear economic rationale. Further, audit costs in subsequent years may not be lower because of a change in scope, but the auditor was locked in on a fixed fee for a certain period (eg, three years). With the spin-offs or outright sales of auditing firm consulting arms, the auditor has less of an economic rent to gain in future years. Instead, the client faces a severe penalty should they displace the auditor. While the successor auditor has the ability to review the prior auditor's working papers, these are very likely to be less informative than having had hands-on experience with the client previously.

SUMMARY AND FUTURE RESEARCH

In this paper, we only examine published studies. What has the research found that could be useful to regulators? To date, there have been few published studies. Initial studies examined investors' perceptions of the impact of auditor independence. Geiger et al. concluded that investors were still pessimistic. That study, however, was published directly after the implementation of SOX. It is now five years since that study has been published. Has investors' initial pessimism been alleviated? This is an area for future research. Another viewpoint is that, judging by the great increase in the stock market since the 2000 reporting scandals, there is considerable uncertainty as to whether the auditor plays an important role in the capital market or whether investors' are even concerned with auditor independence. Alternatively, investors may have been lulled into feeling comfortable about audit firm performance given the clear legal penalties that the audit firms face in the era of Sarbanes–Oxley. This, too, is an area for future research that has not been thoroughly examined. It would be interesting to see whether SOX itself, and the impact of SOX on auditor independence, is now viewed more positively by the investing public. An empirical study by Chung and Kallapur demonstrates that auditors are more stringent with larger companies in the post-SOX period. This is just one study. It is four years old and used data gathered almost immediately after the implementation of SOX. Researchers now have access to four more years of data. A replication of this study could reveal whether the tough stance of the auditors has softened in the intervening years. Research on the provision of nonaudit services has been relatively more extensive. The results, however, are conflicting. Frankel et al.³⁸ conclude that the provision of nonaudit services (and increased fees therefrom) impairs the independence of auditors. Ashbaugh et al., however, find results to the contrary. Both studies used data immediately after the SOX enactment. Hence, this is another area deserving of further scrutiny. Finally, other results tend to support the idea that audit partner rotation, required by SOX, increases independence (Marchesi and Emny). The studies also find that auditor rotation improves independence in the post-SOX period (Mayhew and Pike). Finally, dismissal threats do not appear to have a significant impact on auditors' decisions, implying an independent auditor stance in the post-SOX period.

The SEC requirement to enhance auditor independence has been far reaching. It involves providing guidelines on matters relating to the provision of nonaudit services, partner rotation, audit engagement teams, auditor compensation and the role of the audit committees. Overall, the paucity of current research on these issues has led the field to fall short of providing detailed guidance to the SEC and other regulators on the effectiveness of their guidelines in enhancing auditors' independence.

Now let us look at other research opportunities in this area. An interesting research question posed by Kinney³⁹ is still valid today, namely, do other countries have a better system? Italy has long proscribed management consulting services rendered by the audit firm to an audit client. Have the Italian rules been successful? Does Italy have more investor confidence in the auditors' independence, or in the efficiency and integrity of the system of audited financial reporting and disclosure? Kinney notes that comparative survey research across regulatory regimes could be instructive in determining whether the United State's might be improved, and how that might be accomplished. In England, the Cadbury Committee recommended the rotation of audit partners within firms with respect to audits. China is conducting an open market policy and Chinese firms are seeking independent audit services. China is characterised by government control and influence over CPA firms which, in turn, protects auditors from the threat of litigation. The government, in their desire to promote foreign investment, has, however, imposed penalties ranging from suspension of licenses for signing false reports to imprisonment for violation of criminal statutes and forced dissolution of firms.⁴⁰ Have the threats of these actions improved independence in China? What about the French system of having two audit firms act jointly as auditors for specific clients.⁴¹ Under this system, two audit firms serve as auditors for large clients, allowing one of them to be dismissed without gravely damaging the ability of the client to achieve a timely audit in the future since one of the two prior auditors is still on the audit. According to Herbinet,⁴¹ this system has led to markedly lower auditor concentration in France than elsewhere. Given US Treasury Secretary Paulson's concerns about auditor concentration, and the difficulty second- or third-tier auditors have in auditing giant corporations, having audit firms team up provides alternatives to the continued dominance of the Big Four. In the alternate, of course, such a system gives companies an ability to seek auditors beyond the Big Four, thus increasing the audit firm's perceived risk that they could be replaced by the client.

A further consequence of the French system, however, is that it will be more difficult for any particular audit firm to lose its independence to the client since such an outcome becomes more quickly disoverable by its co-audit firm. We note that Canada has had a co-audit firm for its banking system for a number of years.

An additional framework for understanding auditor research internationally comes from using the insights of La Porta et al.⁴² These researchers argued that legal systems in countries around the globe reflect the legal system of the former colonial power that had once governed that country, pre-independence. According to La Porta et al.,⁴² Anglo-Saxon countries offer the best investor protections. The Anglo-Saxon countries are characterised as having common law systems. Germanic code countries offer a middling level of investor protection, while the Civil Code, characteristic of the legal systems of former French colonies, provides relatively less protection. Research questions that stem from this include: are auditors more likely to toe the line on independence in common law countries than in Germanic code countries since investor protections are better in the former, thereby raising the prospect of audit firm litigation losses? Similarly, are auditors in Germanic code countries more likely to toe the line on independence than auditors in Civil Code countries? Given this background, it would be instructive to understand how the interplay between board members, managers and auditors plays out across countries that have adopted different legal codes. Are auditor–client relations clearly more constrained in common law countries than in Germanic countries, and so forth? We believe that it is important to understand these especially given strong trends toward globalisation of business and easing of cross-border stock exchange listing requirements. Understanding this, along with the interaction of national characteristics (see discussion of Italian, British and Chinese auditor regulatory regimes above) should be of great use in gaining a greater understanding of the determinants of auditor independence and behaviour across a variety of settings.

Finally, the AICPA suggests that independence-related research should analyse issues related to confidence in the independence of auditors, perceptions of financial statement accuracy and reliability, and discretionary decision making by financial statement users. In this paper, we identify deficiencies in existing research and provide guidelines on expanding the reach of auditor independence research by looking at developments in the international, as well as domestic, arena.

References

1. Mayhew, B. W. and Pike, J. E. (2004) 'Does investor selection of auditors enhance auditor independence?' The Accounting Review, 79 (3), 797–822. | Article |
2. Kleinman, G., Palmon, D. and Anandrajan, A. (1998) 'Auditor independence: A synthesis of theory and empirical research', Research in Accounting Regulation, 12, 3–42.
3. Sweeney, J. (9/2004) 'A Firm for All Reasons: Having Rejected the Popular Wisdom of Its Profession, Deloitte Advances With Integrated Services', Found at http://www.consultingmagazine.com/CMCoverFeat-Deloittesept04.html, (link currently inactive)visited on 5/5/2007.
4. Norris, F. (4/17/2004) 'Big Auditing Firm Gets 6-Month Ban On New Business', Found at http://select.nytimes.com/search/restricted/article?res=F30D10FC385F0C
   748DDDAD0894DC404482, visited on 5/5/2007.
5. Abelson, J. (1/26/2002) 'Enron's many strands: boardroom reactions; Trying Not to Be the Next Enron, Companies Scrutinize Practices', Found at http://select.nytimes.com/search/restricted/article?res=FB0F16FC
   3F5E0C758EDDA80894DA404482, on 5/5/2007.
6. Levitt, A. (2000)Renewing the covenant with investors. Speech delivered at New York University Center for Law and Business. 28th September.
7. Securities and Exchange Commission (2000)Revision of the Commissioner's Auditor Independence Requirements. File no. S7-13-00, effective 5th February, 2001.
8. Prentice, D. (2006) 'A voice crying in the wilderness for auditor independence: Abe Briloff and section 201 of the Sarbanes–Oxley act of 2002', Journal of American Academy of Business, 8 (2), 190–195.
9. See Schneider et al.10 Kleinman, G. Palmon, D. and Anandrajan, A. (1998) 'Auditor independence: A synthesis of theory and empirical research', Research in Accounting Regulation, 12, 3–42.
10. Schneider, A., Church, B. and Ely, K. M. (2006) 'Non-audit services and auditor independence: A review of the literature', Journal of Accounting Literature, 25, 169–211.
11. Mautz, R. K. and Sharaf, H. A. (1961). The Philosophy of Auditing, American Accounting Association, Sarasota, FL.
12. Kleinman, G. and Palmon, D. (2001). Understanding Auditor–Client Relationships: A Multi-Faceted Analysis, Markus Weiner Publications, Inc., Princeton.
13. Sikka, P. and Willmott, H. (1995) 'The power of independence: Defending and extending the jurisdiction of accounting in the United Kingdom', Accounting, Organizations and Society, 20 (6), 547–581. | Article |
14. Francis, J. (1994) 'Auditing, hermeneutics and subjectivity', Accounting, Organizations and Society, 19 (3), 235–269. | Article |
15. We are grateful to an anonymous reviewer for this information.
16. Geiger, M. A., Lowe, D. J. and Pany, K. J. (2002) 'Outsourced internal audit services and the perception of auditor independence', The CPA Journal, 72(April), 4–20.
17. AICPA (1997)Report of the Special Committee on Assurance Services ( www.aicpa.org/assurance/scas/comstud/assind/index.htm ) on 30th August.
18. Palmrose, Z. V. and Saul, R. S. (2001) 'The push for auditor independence', Regulation, 24 (Winter), 4–18.
19. Kleinman, G. and Palmon, D. (2000) 'A negotiation-oriented model of auditor–client relationships', Group Decisions and Negotiations, 9 (Winter), 2000,17–45. | Article |
20. Craig, J. L. (1995) 'Strengthening the professionalism of the auditor: An interview with Donald Kirk', CPA Journal, February, 1995, found at http://www.nysscpa.org/cpajournal/old/16641850.htm, on 7th December, 2007.
21. Norris, F. (8/16/2000) 'THE MARKETS: Market Place; A top accounting executive asks the S.E.C. to amend a new proposal on auditor independence', Found at http://select.nytimes.com/search/restricted/article?res=F20614FF345
    B0C758DDDA10894D8404482, on 5/5/2007.
22. Schneider et al.¹⁰ surveyed the literature on NAS disclosures and concluded that such provision impairs the perception of auditor independence but not the reality. The authors note that contextual effects (eg, user cognitive style and the presence of audit committees) may affect those perceptions.
23. Barrionuevo, A. (3/15/2005) 'Changing the guard: The boards; the rise of the boards', Found at http://select.nytimes.com/search/restricted/article?res=F00917FB39580C7
    68DDDAA0894DD404482, on 5/5/2007.
24. Hodge, F. (2003) 'Investors' perceptions of earnings quality, auditor independence, and the usefulness of audited financial information', Accounting Horizons, 17 (supplement), 37–48. | Article |
25. Hussey, R. and Lan, G. (2001) 'An examination of auditor independence issues from the perspectives of U.K. finance directors', Journal of Business Ethics, 32 (July), 169–178. | Article |
26. Iyer, V. M. and Raghunandan, K. (2002) 'Auditors' employment with clients and interaction with their former CPA firm', Journal of Managerial Issues, 14 (Winter), 486–499.
27. Chung, H. and Kallupar, S. (2003) 'Client importance, non-audit services, and abnormal accruals', The Accounting Review, 78, 931–955. | Article |
28. Beattie, V., Fearnley, S. and Brandt, R. (2001) Behind Closed Doors: What Company Audit is Really About, Palgrave Macmillan, Basingstoke, UK.
29. Marchesi, M. F. and Emby, C. E. N. (2005) 'The impact of continuity on concurring partner reviews: An exploratory study', Accounting Horizons, 19 (1), 1–10. | Article |
30. Gul, F., Jaggi, B. L. and Krishnan, G. V. (2007) 'Auditor independence: Evidence on the joint effects of auditor tenure and non audit fees', Auditing: A Journal of Practice and Theory, 26 (2), 117–142. | Article |
31. Ashbaugh, H., LaFond, R. and Mayhew, B. W. (2003) 'Do nonaudit services compromise auditor independence? Further evidence', The Accounting Review, 78 (3), 611–639. | Article |
32. Ruddock, C., Taylor, S. and Taylor, S. (2006) 'Non audit services and earnings conservatism: Is auditor independence impaired?', Contemporary Accounting Research, 23 (13), 701–746. | Article |
33. Kinney, W. R. Jr., Palmrose, Z. V. and Scholz, S. (2004) 'Auditor independence, non audit services and restatements: Was the U.S. government right?', Journal of Accounting Research, 42 (3), 561–588. | Article |
34. DeFond, M. L., Raghunandan, K. and Subramanyam, K. R. (2002) 'Do non audit service impair auditor independence? Evidence from going concern opinions', Journal of Accounting Research, 40 (4), 1247–1274. | Article |
35. Whisenant, S., Sankaraguruswamy, S. and Raghunandan, K. (2003) 'Evidence on the joint determination of audit and non audit fees', Journal of Accounting Research, 41 (4), 721–744. | Article |
36. Nocera, J. (4/28/2007) 'Talking Business: Weighing Jobs' role in a scandal', Found at http://select.nytimes.com/search/restricted/article?res=F6061FFB395A0C
    7B8EDDAD0894DF404482, on 5/5/2007.
37. Lu, T. (2006) 'Does opinion shopping impair auditor independence and audit quality?', Journal of Accounting Research, 44 (3), 561–583. | Article |
38. Frankel, R. M., Johnson, M. F. and Nelson, K. F. (2002) 'The relation between auditors' fees for non audit services and earnings management', The Accounting Review, 77 (Supplement), 71–105. | Article |
39. Kinney, W. R. Jr. (1999) 'Auditor independence: A burdensome constraint or core value?', Accounting Horizons, 13 (1), 69–75. | Article |
40. Yang, L., Dunk, A., Kilgore, A. and Lin, Z. J. (2003) 'Auditor independence issues in China', Managerial Finance, 29 (12), 57–64. | Article |
41. Herbinet, D. (5/23/2007) 'Letter to the Editor: Competitive audit market would foster innovation', Financial Times, found at http://www.ft.com/comment/letters, on 5/23/2007.
42. La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R. W. (1997) 'Legal determinants of external finance', Journal of Finance, 52 (3), 1131–1150. | Article |

Acknowledgements

We acknowledge our gratitude to Don Warren for his comments and suggestions.