jennings et al - 2006 - strong corporate governance and audit firm rotation- effects on judges' independence perceptions and litigation judgments

253–270 Strong Corporate Governance and Audit Firm Rotation: Effects on Judges’ Independence Perceptions and Litigation Judgments Marianne Moody Jennings, Kurt J.. Reckers SYNOPSIS: The

September 2006

pp 253–270

Strong Corporate Governance and

Audit Firm Rotation: Effects

on Judges’ Independence Perceptions

and Litigation Judgments Marianne Moody Jennings, Kurt J Pany,

and Philip M J Reckers

SYNOPSIS: The Sarbanes-Oxley (SOX) legislation mandated modest threshold levels

of corporate board independence and expertise, as well as audit partner (not firm) rotation One objective was to create an environment supportive of enhanced actual and perceived auditor independence This study examines whether perceptions of au-ditor independence and auau-ditor liability are incrementally influenced by further strength-ening corporate governance and by rotating audit firms Our experimental study ad-dresses these questions by analyzing responses of 49 judges attending a continuing education course at the National Judicial College The experiment manipulates

cor-porate governance at two levels (minimally compliant with current corcor-porate gover-nance requirements versus strong) and auditor rotation at two levels (partner rotation versus audit firm rotation) We find that strengthening corporate governance (beyond

minimal SOX levels) and rotating audit firms (compared to partner rotation) lead to enhanced auditor independence perceptions We also find that judges consider audi-tors less likely to be liable for fraudulently misstated financial statements when firm rotation is involved in a minimally compliant corporate governance environment.

Keywords: auditor rotation; corporate governance; judges.

Data Availability: Confidentiality agreements prevent the authors from distributing the

data.

INTRODUCTION

The Sarbanes-Oxley Act of 2002 (SOX, U.S House of Representatives 2002) provides

a variety of initiatives intended to enhance audit quality and restore investor confi-dence in capital markets With respect to auditors, provisions of the Act seek to enhance independence both in fact and in appearance First, SOX reforms the relationship between corporate boards of directors and external auditors By elevating the degree of independence and expertise of corporate board members, legislators reasoned that they

Marianne Moody Jennings, Kurt J Pany, and Philip M J Reckers are all Professors at Arizona State University.

Submitted: May 2005 Accepted: May 2006

Corresponding author: Philip M J Reckers

Email: philip.reckers@asu.edu

could reduce the pressures brought to bear by corporate management on the auditor to compromise independence Second, legislators considered mandatory audit partner and firm rotation By limiting the duration of auditor (or audit firm) client relationships, legislators reasoned that economic incentives associated with compromised independence might be lessened

SOX ultimately mandated new minimal levels of corporate board independence and expertise (discussed below) as well as engagement and review partner rotation, but it did not require that CPA firms be rotated.1 While corporations have shown some interest in adopting board governance standards beyond those required by SOX, they have resisted corporate policies mandating audit firm rotation Indeed, 89 percent of corporate boards still do not have an independent chairman (Business Roundtable 2006), and nearly 99

percent of Fortune 1000 public companies and their audit committees had no policy of

rotating audit firms as of November 2003 Only 4 percent were even considering rotating auditors a full year following passage of SOX (U.S General Accounting Office [GAO]

2003, 15) Nevertheless, both regulators and the business press continue to consider the proposition that long-term relationships between companies and their auditors create a closeness between the auditor and management that reduces the public’s perception of auditor independence and audit quality

We report the results of an experimental study of the effects of enhanced levels of corporate board independence and expertise and of audit firm rotation on U.S judges’ perceptions of auditor independence and auditor liability Our research instrument first pre-sents judges with background information about a public company, including audited fi-nancial statements, and it asks whether they perceive the external auditors as independent

under varied conditions of corporate governance (minimally compliant with regulatory cor-porate governance requirements versus strong) and audit rotation (partner versus firm).

Subsequently, the instrument discloses that the earnings were discovered to be fraudulently misstated, and that a lawsuit was initiated The discovery stage of the lawsuit reveals ad-ditional information regarding the fraudulent actions of management and the conduct of the audit Judges are then asked to assess auditor liability

We find that both strong corporate governance and audit firm rotation result in increased judicial perceptions of auditor independence We also find that firm rotation (compared to partner rotation) cause judges to consider auditors less liable for fraudulently misstated financial statements in an environment of minimally compliant corporate governance, a condition typical of many firms The second section presents background information and develops hypotheses The third section presents our research method The fourth and fifth sections present and discuss our results, respectively

BACKGROUND AND DEVELOPMENT OF HYPOTHESES

Background

The importance of auditor independence, both actual and as perceived by others, has been widely accepted both in theory and by regulators De Angelo (1981a, 1981b) defines audit quality as the market-assessed joint probability that auditors will discover a breach

1 Section 207 of SOX required the Comptroller General of the United States to conduct a study to review the potential effects of requiring mandatory rotation of registered public accounting firms The subsequent study issued in November 2003 (U.S General Accounting Office 2003), concludes that the benefits of mandatory firm rotation are not certain and that more experience with the effects of SOX’s other requirements is needed before any requirements relating to audit firm rotation can be considered further.

in the accounting system and report that breach She reasons that auditors who have an economic interest in their clients (i.e., lack financial independence) may be less apt to report

a discovered breach or may apply less effort to discover one She defines economic interest

as a future ‘‘quasi-rent’’ stream in which quasi-rents represent the present value of future revenues (less costs) over the expected duration of an auditor-client relationship The pos-sibility of potentially earning long-term quasi-rents can result in a situation in which the auditor’s professional independence is impaired and / or is perceived to be impaired Most legal challenges faced by auditors are based on plaintiff arguments that an in-appropriate audit report was issued due to deficient standards of work performance and / or reporting linked to failed independence A lack of actual independence is a lack of objec-tivity in weighing audit evidence or in reporting.2 The lack of actual independence is ordinarily unobservable because audit working papers seldom, if ever, acknowledge a lack

of objectivity Accordingly, legal independence assessments must be based on judges’ (or

juries’) perceptions of auditor independence, given circumstantial evidence and

environ-mental conditions The Securities and Exchange Commission (SEC) standard for auditor independence also uses a subjective determination based on perception, not necessarily through specific circumstances that compromise independence or through auditor acknowl-edgment of independence issues.3In 2000, the SEC expressed concern that the forces that hamper independence are insidious and difficult to document, but are nonetheless forces that impair judgment The SEC reaffirmed the ‘‘appearance standard’’ that has been applied

in both regulatory matters and litigation.4 SOX’s list of prohibitions on auditor activities and conflicts of interests codified the appearance standard.5 In our experiment we thus manipulate two factors related to auditor independence—corporate governance and auditor rotation

Judicial Decisions

Judges, as a group, are highly experienced in the art of evaluating evidence Further, given their role in society, judges normatively should decide cases strictly on their merits: the facts and the law However, research suggests that judges’ objectivity can be compro-mised In particular, Guthrie et al (2001, 779) point out that legal scholars representing a variety of schools of thought have long argued that judges do not find facts or apply legal principles in a completely accurate and unbiased fashion Judges’ views have been found

to be subject to:

2 See Elliott and Jacobson (1998) for a discussion of independence concepts.

3 Rule 2.01(b) of Regulation S-X states that ‘‘[t]he Commission will not recognize any certified public accountant

or public accountant as independent who is not in fact independent,’’ 17 C.F.R § 210.2.01(b) (1985) In support

of this rule, the SEC has a wide variety of releases (http: / / www.sec.gov / about / offices / oca / ocaprof.htm presents

a listing of various releases and reference materials).

4 For SEC statements, see Security and Exchange Commission 2000 Exchange Act Release No 33-7919: Auditor Independence Requirements Federal Register 65 (Dec 5) at 76,008, 76,017 U.S Government: Washington, D.C (also available at www.sec.gov ⬍http://www.sec.gov⬎) For SEC judicial statements on independence, see United States Supreme Court 1984 United States versus Arthur Young and Co United States Supreme Court

465 at 805, 817–18: Washington, D.C (also see supremecourtus.gov) and Second District of New York 2001 Complete Management Inc Securities Litigation Federal Supplement 153 (2d) at 314, 334–35 U.S Govern-ment: Washington, D.C Shareholder actions include Second District of Texas 2002 Enron Corporation Secu-rities, Derivative and ERISA Litigation Federal Supplement 235 (2d) at 549 U.S Government: Washington, D.C and Security And Exchange Commission 2001 Exchange Act Release No 43862: Accounting and Au-diting Enforcement release No AE-1360: Matter of KPMG Peat Marwick LLP Administrative Proceeding File

No 3-9500 (Jan 19) U.S Government: Washington, D.C (also see S.E.C Docket 74 (2001) 384 at www.sec.gov

⬍http://www.sec.gov⬎).

5 SOX ( ن 107-204).

● many of the same heuristic biases that afflict others,

● direct exposure to issues in courtrooms over which they presided,

● matters gleaned from other court cases and the popular press, and

● beliefs and attitudes built up over a lifetime of experiences.6

A significant body of research supports the conclusion that beliefs and attitudes are decision-influencing because individuals often ignore relevant information (or differentially weigh positive and negative information) to support their prior beliefs (e.g., Mahoney 1976, 1977; Lord et al 1979; Wilson et al 1993; Pham et al 2001) Similar findings are likely

to apply to judges

Historically, judges’ attitudes and their decisions are significantly correlated (e.g., Champagne et al 1981; Danelski 1966; Robbennolt 2005) In accounting, laboratory ex-periments find that judges’ attitudes and their judgments relating to auditor liability are related directly and interactively with other environmental factors (Jennings et al 1991a, 1991b, 1993; Anderson et al 1997) For example, an extensive literature is developing within accounting pertaining to the outcome-bias phenomenon among judges (see Lowe and Reckers [2006] for a summary)

Beliefs and attitudes arise through learning, whereby a person acquires a reaction to

an action over a period of time Once learned, the attitude is triggered automatically when one is exposed to the action or thinks about it (Bagozzi et al 2003) Since attitudes are

‘‘learned’’ reactions acquired over time, they can also change over time, but they change only modestly in normal times (for reviews see, Ajzen 1996; Eagly and Chaiken 1993) Reckers et al (2006) document a significant change in judges’ attitudes and beliefs regard-ing auditors over the last decade (similar to that found in other groups) and argue that Enron and related debacles precipitated a significant change in previously engrained atti-tudes and beliefs Thus, a timely issue is whether SOX reforms produce a counter-effect in judges’ views and influence their judgments on auditor independence and shared liability

Corporate Governance

Perceptions of an auditor’s independence and the magnitudes of liability judgments are likely to be affected by the strength of corporate governance SOX reforms, among other things, target the relationship between corporate boards and external auditors The extent

of independence and expertise of corporate board members potentially alters pressures that might be brought to bear by corporate management on the auditor to compromise indepen-dence Although SOX and the resulting changes in stock exchange listing requirements include increased corporate governance standards for registrants, significant flexibility still exists in the manner in which such reforms are implemented To illustrate areas of contin-uing flexibility, consider the following:

● The chairperson of the board may be independent of management or be part of management, including the corporate CEO

● The proportion of independent directors on the Board may vary

● The level of financial expertise of members of the Board may vary

● The level of financial expertise of members of the audit committee may vary

● The diligence of the audit committee may vary (e.g., number of meetings per year, activities undertaken, etc.)

6 See Guthrie et al (2001), Champagne et al (1981), and Redding and Repucci (1999).

Prior research has shown a direct correlation between financial reporting quality and each of the following:

● board and audit committee independence,

● financial expertise, and

● audit committee diligence.7

We investigate the impact of enhancements of corporate governance beyond minimal SOX requirements on judicial perceptions of auditor independence and auditor liability

Audit Firm Rotation

Numerous researchers have argued that audit firm rotation makes auditors appear to be more independent (e.g., Arel et al 2005; Brody and Moscove 1998; Kemp et al 1983; Ramsey 2001; Winters 1978; Wolf et al 1999) Threats to auditor independence are not new; more than 40 years ago, Mautz and Sharaf (1961, 208) warned auditors that:

the greatest threat to his independence is a slow, gradual, almost casual erosion of this ‘‘honest disinterestedness’’ the auditor in charge must constantly remind his assistants of the importance and operational meaning of independence.

Accordingly, a lengthy tenure is often perceived to limit an auditor’s ability to exercise

objectivity and may lead to poor audit quality and audit failures.

In this study, we consider the perceptions of judges relating to situations with and without audit firm rotation Regulators have openly suggested the need for audit firm ro-tation Lynn Turner (2002), former Chief Accountant of the Securities and Exchange Com-mission, speaking before the United States Senate Committee on Banking, Housing, and Urban Affairs, indicated that to truly protect the independence and integrity of the audit, Congress should require mandatory rotation of the audit firm Turner’s remarks followed those of Ellen Seidman (2001), Director of the Office of Thrift Supervision, who opined that audit firm rotation every 3–4 years was desirable, in that it would allow a ‘‘fresh look’’

at the organization Similarly, nonregulatory bodies such as the Conference Board (2003) have suggested the need for firm rotation A recent GAO study (GAO 2003) observed that further analysis is needed to determine the benefits of mandatory rotation, because the benefits are harder to predict and quantify than the additional costs However, the combi-nation of no regulatory requirement of audit firm rotation and few companies voluntarily establishing such a policy makes archival research directly addressing rotation virtually impossible

Still, independence in appearance can be examined In this paper, we address an issue that has not been addressed—the likely effects of audit firm rotation on judges’ perceptions

of auditor independence and auditor liability We do so using an experiment that manipu-lates (1) minimally compliant versus strong corporate governance and (2) firm versus part-ner rotation across subjects

Hypotheses

Perceptions of Independence

This study addresses two hypotheses related to perceptions of auditor independence, and one hypothesis and one research question related to auditor liability The first issue addressed is whether further strengthening of corporate board independence and expertise

7 See Arel et al (2006) for a summary of this research.

will sufficiently alter the environment in which an auditor works to lead judges to perceive enhanced levels of auditor independence The second issue is whether audit firm rotation (compared to partner rotation) will lead judges to perceive enhanced levels of auditor in-dependence Our first two hypotheses (stated in the alternative form) thus are:

H1: Judges will perceive the extent of auditor independence to be greater under a

situation with incrementally stronger corporate governance

H2: Judges will perceive the extent of auditor independence to be greater under a

situation with incrementally greater auditor rotation (i.e., firm rotation rather than partner rotation)

If a synergy exists between the two manipulated enhancements, then an interaction

hypothesis would also be in order We did not see a basis, ex ante, to predict such a synergy,

and no interaction was found

Auditor Liability

After eliciting participants’ perceptions of auditor independence and financial statement credibility,8 we informed participants that the financial statements are misstated We then asked participants to assess auditor liability An extensive literature exists within and outside

of accounting regarding outcome (hindsight) bias Knowledge of a misstatement can po-tentially restrict the ability of judges to objectively evaluate auditor performance retrospec-tively Extant research offers a cognitive explanation of the phenomenon Judges (and juries) process information in a temporally backward mode, from the given outcome to the ante-cedent conditions That is, individuals focus their attention on the given outcome and try

to explain its occurrence by creating causal links to predecessor events and actions Once this causal framework is developed, individuals experience difficulty considering how al-ternative outcomes could have occurred (Schkade and Kilbourne 1991; Baron and Hershey 1988; Fischhoff 1975) Events surrounding the recent demise of Arthur Andersen and sim-ilar allegations of impropriety of other audit firms increased public skepticism as to the independence of the auditors In the context of a failed audit, judges will more likely reconstruct a scenario of auditor guilt in an environment lacking controls, such as audit firm rotation By way of analogy, Lowe et al (2002) find that jurors considered auditors more responsible for a failed audit when they failed to use available quality control decision aids

With respect to auditor liability, culpability in the form of a lack of indepen-dence ordinarily can only be inferred from surrounding circumstances; a lack of inde-pendence cannot be proven Our case materials do not provide explicit proof that the firm lacks independence Yet, the audit firm rotation condition suggests greater independence safeguards Partner rotation only, on the other hand, is the historic status quo and as such

is a condition more consistent with auditor independence being compromised to achieve personal gain Accordingly, we hypothesize that audit firm rotation will lead to lower as-sessments of auditor legal liability.9More specifically:

8 In addition to asking participants about perceived auditor independence, participants were asked the extent to which they perceived the environment to protect the public, and the perceived likelihood of unintentional errors and intentional misstatements Findings were highly consistent across these four queries, and thus only the responses to the independence question are reported.

9 Bonner et al (1998) report that although alleged independence violations are relatively infrequent, when such

a violation exists, it has the highest likelihood of resulting litigation.

H3: Subsequent to an audit that has failed to identify existing fraud, judges will

per-ceive lower auditor liability under conditions of audit firm rotation

Predicting the effects of the strength of corporate governance on judges’ perceptions

of auditor legal liability is more complex because no directly germane prior research exists The strength of corporate governance speaks most directly to the culpability / liability of management and corporate board members and only indirectly to auditor liability Two lines of argument can be advanced as to the potential effects of corporate governance on judges’ assessments of auditor liability First, as discussed above, extant hindsight research (using judges as well as juries) supports the notion that individuals process information in

a temporally backward mode, reconstructing events to make sense of the negative outcome (Lowe and Reckers 2006) In auditing, arguably the most sinister scenario is one of a conspiracy between a non-independent auditor and a corporate client lacking a culture of integrity An attribution of a lack of independence and legal culpability can be most easily reached in an environment of otherwise relatively weak corporate governance (i.e., an en-vironment in which management pressure on the auditor is likely to be high) and rotation

of partners but not firms (a condition in which the audit firm has the highest economic incentives to retain the client) This reasoning suggests that auditor liability would be greatest under the joint condition of partner rotation and minimal board strength resulting

in a significant interaction between the experiment’s two manipulated factors

Alternatively, a less cynical possibility is that auditor liability could be greatest under the joint condition of partner rotation and strong corporate governance That is, if the financial misstatements cannot be attributed at least in part to minimally compliant corporate governance, by default, then greater blame is laid at the feet of the auditor when corpo-rate governance is strong This result is arguably consistent with attribution theory A basic attribution theory tenet is that causal attributions are made either to the environment or to the individual—a zero sum game Thus, if corporate governance seems strong, then the misstated financial statements must be the fault of the auditor Added support for this second perspective may accrue from prevailing ‘‘comparative contribution’’ judicial philosophy Under proportionate liability rules, judges’ allocations of damages for the plaintiff’s loss are made by using a percentage of total fault for each party (Raoke and Davidson 1996) Judges may determine that other parties (e.g., corporate board, lax corporate internal con-trols) are responsible in part for the damages and can assign blame through fault apportionment

Accordingly, this second line of thought suggests that the effects of auditor rotation (partner versus firm), once again, may be conditional on the strength of corporate board governance But it is unclear whether the significance of the auditor rotation manipulation would be greatest in the weak or strong corporate governance condition, or equally signif-icant under each governance condition That is, we are unsure if auditor rotation will be a main effect or interaction effect Accordingly, we investigate the issue as a nondirectional research question:

RQ1: Subsequent to an audit that has failed to identify existing financial reporting

fraud, will judges’ perceptions of auditor liability be a joint product of the strength of corporate governance and the form of auditor rotation? And, if so, what will be the nature of the interaction?

TABLE 1 Profile of Subjects

(n ⫽ 49)

Mean

Age Gender (% Female) Years as a judge Owner of stock

48.3 25%

2.61 94%

METHOD Participants

The experiment was conducted during a continuing professional education training program at the National Judicial College Table 1 provides a general demographic data profile Participants under the various experimental conditions did not differ on any of the profiling items Nine participants who failed manipulation checks were removed from our analyses The findings are based on the responses of 49 judges

Research Task and Dependent Variables

Judges were provided with experimental materials describing the scenario of a typical audit The materials included background information on the corporate audit client that manufactured a variety of industrial products Materials also addressed the relationship (fees and tenure) of the audit firm and audit client.10 The CPA firm had issued unqualified opinions throughout its audit tenure Manipulated variables included auditor rotation and the corporate governance structure Additionally, the scenario included summarized income statement and balance sheet information for the current year Finally, to add stress to the situation, we included the need for the company to restate earnings three years earlier due

to inappropriate revenue recognition practices The background information is presented in the Appendix

After receiving the common background information and the auditor rotation and cor-porate governance manipulations, judges were then asked to reply to the following question

to permit an assessment of their perceptions of auditor independence:

To what extent do you believe the external auditors, K&L, are independent?

Not Independent 0 1 2 3 4 5 6 7 8 9 10 Completely

Subsequently, participants were informed that current-year net income was misstated

We then solicited the judges’ views on likely auditor liability A disgruntled employee ‘‘blew the whistle’’ on a fraud involving overstated income effected by prematurely and inappro-priately recognizing revenues on sales that were subsequently cancelled or substantially reduced A lawsuit was filed when the corporation’s stock price declined precipitously after the fraud disclosure The experimental materials informed the judges that the discovery process revealed that weak internal controls had been overridden to effect the fraud The

10 In addition to audit fees of $758,000, fees for nonaudit services performed by the CPA firm were $2,325,875,

of which $2,072,374 was related to tax services.

FIGURE 1 Details of Corporate Governance Manipulation

Minimally

Board of Directors

Size

Number Independent of Management

Chairman

15

8 Not Independent (Company Founder)

15

12 Independent

Audit Committee

Size

Members all independent?

Expertise (and literacy)

Meetings in 2002

Summary Description

3 Yes Minimally Compliant Relatively Weak

5

Yes Strong Strong

auditor had previously informed the audit committee of weak controls Judges’ responses

to the following question were then elicited

To what extent do you believe the external auditors are legally liable and plaintiff investors should be allowed to recover substantial damages?

Not at All 0 1 2 3 4 5 6 7 8 9 10 To a Great Extent

Independent Variables

Corporate Governance Structure

While any number of variables within corporate governance might be manipulated, we selected two levels that comply with current corporate governance requirements (SEC 2003), are realistic,11 and are documented by prior research as affecting the financial re-porting process The low level is minimally compliant with regulatory corporate governance requirements and the high level goes beyond regulatory requirements (‘‘strong’’) Specifi-cally, Figure 1 shows that, when compared to the ‘‘minimally compliant’’ condition, the

‘‘strong’’ condition has a chairman who is independent of management, a higher proportion

of independent directors on the Board, a higher level of financial expertise of members of the audit committee, and more audit committee diligence

Auditor Rotation

This variable has two levels: partner rotation and audit firm rotation Because SOX requires periodic partner rotation (within the same CPA firm), we select this as the low-independence condition In contrast, a policy of rotating audit firms is our high-independence condition This manipulation allows us to directly test whether the level of rotation affects judges’ beliefs In either type of rotation, partner versus firm, a study must

11 See Taub (2004) for a discussion of continuing differences in strength of corporate governance and audit committees.

TABLE 2 Effects of Auditor Rotation (Hypotheses 1) and Corporate Governance (Hypothesis 2) on Perceptions of Auditor Independence

Analysis of Variance for Auditor Independence

a One-tailed tests for main effects, two-tailed for interaction.

b Minimally compliant with SOX versus strong corporate governance as operationally defined in Figure 1.

c Audit partner rotation versus audit firm rotation.

d Joint or synergistic influence of governance and rotation experimental treatments on perceptions of

independence.

also address the current year within the rotation cycle For example, if a company rotates every four years, then the CPA firm involved could be in any one of the first through the fourth years of the relationship To provide a strong and realistic test, we tested the final year prior to rotation—regardless of whether it was partner rotation (same firm retained)

or firm rotation In the firm rotation condition, the firm will lose the client within the next year, regardless of how current accounting matters are handled Thus, the CPA firm has the least to lose by resisting client pressure Also, the CPA firm personnel are aware that the manner in which the accounting issue is resolved will be obvious to successor auditors (with another firm), who will be expected to review this year’s audit documentation This scenario is in contrast to the partner rotation situation in which any such review will be performed by different partners within the same firm or, less probably, in conjunction with the peer review process

Manipulation Checks

Manipulation checks for both manipulated variables were included at the end of the task The percentage of subjects who failed either manipulation check was 15 percent (9

of 58) The results of our analyses are substantively the same if these subjects are included

or excluded Accordingly, the subjects who responded incorrectly to either manipulation check are dropped, resulting in a total usable sample of 49

Experimental Design

Subjects were randomly assigned to one of four forms of the questionnaire in a 2⫻ 2 between-subjects design that manipulates mandatory auditor rotation (partner versus firm rotation) and corporate governance (minimally compliant versus strong) We use a between-subjects design to make it difficult for between-subjects to identify the exact nature of the variables being manipulated (see Pany and Reckers 1987) One-tailed tests are applied to directional hypotheses; two-tailed tests are applied to the nondirectional research

RESULTS Hypotheses 1 and 2

Table 2 presents the analysis of variance results for the first two hypotheses Hypothesis

1 is supported, given that the mean assessment of independence in the strong corporate