Rev Manage Sci (2008) 2:111127 DOI 10.

1007/s11846-008-0017-y ORIGINAL PAPER

The effect of a transparency report on auditor independence: practitioners self-assessment

Received: 14 May 2007 / Accepted: 25 March 2008 / Published online: 16 April 2008 Springer-Verlag 2008

Abstract The objective of this study is to investigate how practitioners assess several aspects of a transparency report on auditor independence published by auditing rms. This research was conducted using a survey approach with an experimental component, where the research instrument was distributed to 92 experienced practitioners. Results indicate no signicant perceived difference as to whether the transparency report is mandatory or voluntary or whether the report is audited or not. Also the transparency reports effectiveness is not assessed differently by practicing auditors or accountants. Further descriptive analysis over the full sample indicates the individually most important safeguards related to auditor independence are perceived to be quality assurance, internal quality controls and independence practice and compliance. However, the importance of fee information and information on audited entities is assessed to be more important by accountants compared to auditors due to different interests. Keywords Transparency Auditor independence European Union Report Audit quality Regulation JEL Classication M42

C. Pott (&) C. Watrin Institut fur Unternehmensrechnung und besteuerung, Westfalische Wilhelms-Universitat, Universitatsstr. 14-16, 48143 Munster, Germany e-mail: [email protected] T. J. Mock University of California, Riverside, 900 University Avenue, Riverside, CA 92521, USA T. J. Mock Maastricht University, Maastricht, The Netherlands

123

112

C. Pott et al.

1 Introduction The ex ante value of an audit to consumers of audit services which include current and potential owners, managers and customers of the client depends on the auditors perceived ability to discover errors or breaches in the accounting system (competence), and to withstand client pressure in the event an error or breach is discovered (independence). In this context, auditor independence can be dened as the conditional probability that, given a breach has been discovered, the auditor will report the breach (DeAngelo 1991, pp. 115116). Thus, any auditors impaired independence has the potential of lowering the quality of the audit. Recent scandals in the United States and the European Union have highlighted the fact that the audit is an important element in ensuring the credibility and reliability of nancial statements. However, signicant economic damage to the capital markets and the economy may have resulted due to impaired auditor independence. The independent auditor is expected to serve as an impartial analyzer of the nancial statements of companies, which investors have an interest in (European Federation of Accountants 2001, p. 3). Based on the reliability and creditability of the nancial information, investors make choices, which affect the efciency of capital markets. However, unreliable information will act as a brake on economic progress through harming interests of individual investors. Hence, governments regard audit quality as a public interest and thus try to maintain audit quality through regulatory requirements. Related to this, many professionals consider greater transparency by auditors as a key ingredient necessary to maintain condence in the audit process (Wyman 2004). One means thought to enhance audit quality is by providing transparent information about the audit rms. The Eighth European Directive states that an annual transparency report will be required by June 2008 throughout the European Union (European Parliament and Council 2006). This regulation requires audit rms to publish detailed information on operating policies and procedures, such as independence practice and fee information. However, most Member States have not yet incorporated the transparency report requirements into national law.1 An important group of readers of a transparency report will be major audit clients who will benet from a better understanding of an audit rms independence safeguards. The audit profession can also benet if all audit rms are informed as to independence activities invested into by other rms. We believe that our pool of participants composed of practising auditors and accountants provides the needed mixture of interest in transparency and the expertise necessary to understand the content of such a transparency report. Thus, this study captures impressions of the assessment of such a transparency report by an important subset of the persons that could be affected by additional transparency including those operating in a European jurisdiction. With this experimental study, we survey how auditors assess the effectiveness of more transparency when audit rms are in the position mandated to provide certain
1

The Professional Oversight Board of the UK launches consultation on statutory transparency reporting by auditors of public interest entities (http://asb.co.uk/pob/press/pub1162.html).

123

The effect of a transparency report on auditor independence

113

information. In comparison, we want to see how they assess the value of such information in situations where publication is voluntarily. This should provide insight into which information details audit rms might focus on in their transparency report in order to provide audit quality signals to the market. Furthermore, since our sample also includes practicing accountants, we provide additional evidence on how audit client representatives assess the effectiveness of a transparency report. To address the widely discussed topic of who audits the auditors (e.g., European Federation of Accountants 2003; Institute of Chartered Accountants of England and Wales 2003), we are also interested in how a transparency report is assessed when the report is audited. Thus, we experimentally manipulate whether the transparency report is audited by an independent assurance provider before it is published. Again, the assessment of auditors and accountants might be different due to their different roles in the nancial markets. The paper proceeds as follows: Sect. 2 discusses the theory and develops the research questions. Section 3 presents the research method; and in Sect. 4 the results are discussed. We conclude with a brief discussion of implications, limitations and future research in Sect. 5.

2 Theory and research questions From a capital market perspective, transparent reporting is important for both the allocation and institutional efciency of the capital market. Transparent information can align informed auditors assessments of independence policies and procedures to investors perceived assessments about the auditors independence. Investors should have enough information about the audit rms practice, policies and procedures in maintaining auditor independence to be able to evaluate these issues precisely. Providing investors with reliable and reasonable information about the auditor-agent of a public rm can have a positive effect on overall economic performance by increasing investment opportunities in the global capital markets. A growing body of evidence indicates that the development of nancial markets facilitates performance, but also depends on the level of corporate transparency. Love (2003) studied the hypothesis that nancial development affects growth by decreasing information related imperfections in the capital markets. This research indicates that a decrease in information asymmetry has a positive effect on economic growth. Information asymmetry also exists between the audit rm and other market participants. Information provided in a transparency report may thus also have similar economic effects. Furthermore, nancial information has helped investors overcome moral hazard and adverse selection problems as related to auditors (e.g., Rajan and Zingales 1998). In this respect transparent reporting would provide clients and potential investors with comparable and detailed information on all audit rms and might thus, revive competition in the audit market as well as reduce opportunistic behavior of auditors. For example, information on the education of auditors and fee information are important details potential investors and clients have signicant

123

114

C. Pott et al.

interest in. In this respect the information provided might decrease perceptions related to impaired auditor independence (e.g., Schneider et al. 2006). As argued by Black (2000) and Ball (2001), strong nancial accounting regimes focusing on the credibility and accountability of information are a prerequisite to the very existence of vibrant securities markets. To address the need for greater transparency as one major safeguard, the European Parliament and Council (2006) require audit rms to publish a transparency report annually. Member states must ensure that audit rms carrying out audits of public entities publish on their website an annual transparency report that includes at least the following: 1. 2. 3. 4. A description of the legal structure and ownership If the audit rm belongs to a network of rms, a description of the network and the legal and structural arrangements in the network A statement on the governance structure of the audit rm A description of the internal quality control system of the audit rm and a statement by the administrative or management body on the effectiveness of its function An indication of when the last quality assurance review took place A listing of public interest entities for which an audit has been carried out during the last year by the audit rm A statement about the audit rms independence practices, which also conrms that an internal review of independence compliance has been conducted A statement on the policy followed by the audit rm concerning continuous education of auditors Financial information indicating the economic importance of the audit rm, such as the total turnover divided into fees from the audit of annual and consolidated accounts, and fees charged for other assurance services, tax advisory services and other non-audit services (Article 40).

5. 6. 7. 8. 9.

The required transparency report is seen differently throughout member states. For instance, the German Chamber of Public Accountants (2004) is generally supportive towards more transparency, but questions the level of detail required. Furthermore, the German position points out that this report only addresses audit rms and not individual auditors. They also argue that the date which indicates when the last quality assurance review took place is not very important and suggest that it is more important to know if the quality assurance review was successful. More supportive of such transparency regulation is the government of the United Kingdom. The United Kingdom delegation suggests that the report is an appropriate compromise between all of the stakeholders interests and that there are also important cost implications. Clearly, the benets of greater transparency should outweigh costs (United Kingdom Department of Trade and Industry 2005, p. 10). However, prior research has shown that the benets of increased accounting regulation may exceed the costs (Gwilliam et al. 2005). The European Federation of Accountants (2003, p. 17) believes that a transparency report provides a safeguard only if appropriate oversight is arranged. For example, a review of KPMG Australias processes and policies in respect of independence, conict resolution and quality controls was conducted by two

123

The effect of a transparency report on auditor independence

115

academic experts (Houghton and Trotman 2003). This document is publicly available and offers sound and detailed information on KPMGs policies and procedures and thus, serves as an oversight arrangement. Similarly, one of the requirements of the revised Combined Code in the UK and its related guidance is that audit committees should take on a more active role in overseeing the external audit. An important part of this oversight relates to the independence of the auditor. The Combined Code and the Smith Guidance specify the matters to be taken into account in carrying this out. These include seeking from the audit rm, on an annual basis, information about policies and processes for maintaining independence and monitoring compliance with relevant requirements (paragraph 4.22) (Institute of Chartered Accountants of England and Wales 2003). This could be achieved by an audit of the audit rms policies and procedures in place as stated in the transparency report. Thus, an audit of the audit rm would take place. Lastly, since a transparency report is required to be published by audit rms, auditors are directly affected by this requirement. Accountants, who represent the clients interest, are provided with the additional information of the transparency report. Thus, whereas auditors have to bear the cost and time to provide a transparency report, accountants receive additional insight into their auditor or other audit rms. Moreover, since some safeguards of the transparency report are related to very sensitive issues, e.g. client fee information, auditors might assess this information in a more negative light. However, accountants might be quite interested in information, which helps them to understand their own audit rm or other audit rms business and independence activities. Prior research has shown that better-informed clients lead to improved efciencies in the audit process (Chancy et al. 1997). These different interests of auditors and accountants might be lead to differences as how individual safeguards are assessed. Related to the above discussion and to prior research that suggests that greater transparency helps to safeguard imperfections in markets and to address the controversial discussion throughout member states, we investigate the following overall research questions: RQ1: Is a transparency report perceived to be more effective in improving auditor independence (by practicing accountants versus practicing auditors) if regulated or if voluntarily published? RQ2: Is a transparency report perceived to be more effective in improving auditor independence (by accountants vs. auditors) if audited or not audited? RQ3: How do accountants vs. auditors assess the importance of individual safeguards?

3 Method 3.1 Design Two independent variables were manipulated in a 2 9 2, between-subjects, experimental design. The rst manipulation was whether the transparency report

123

116

C. Pott et al.

is mandatory or voluntary and the second was whether the transparency report is audited or not. Participants were assigned to one version randomly. 3.2 Participants Participants in the experiment were practicing accountants and auditors from the US participating in a conference on current regulatory developments. In this conference policy setters from the U.S. Securities and Exchange Commission (SEC) and the Financial Accounting Standards Board (FASB) discussed latest regulatory developments. Participating in this conference is one means of maintaining required continuing professional education (CPE) for practicing accountants and auditors. Participants had an average of 21 years of professional experience in auditing and/ or accounting with a minimum of 5 years experience and thus they are a highly experienced sample of preparers and auditors of nancial statements.2 We obtain the assessment of non-European professionals, because we are interested in how participants who were not expecting a transparency report to become a mandatory requirement assessed the reports effectiveness. Within the sample of professional participants we are taking the advantage of participants who are relatively nave observers in terms of the transparency report and simultaneously provide the professional expertise concerning what a transparency report means to auditors and preparers. The research was conducted as a survey with an underlying experimental design and was distributed to all conference participants. Sixty-two questionnaires were completed at the conference and 29 were mailed in afterwards.3 As noted, participants had an average of 21 years of professional experience in auditing and/or accounting. Participants indicated their knowledge about auditor independence regulation at a relatively high average of 5.45 on a scale of 0 to 6. Table 1 shows the sample sizes by experimental treatments and by the participants current accounting role (accountant vs. auditor). As is evident in the table, approximately the same sample sizes occurred for each cell. 3.3 Procedure Each participant received a copy of the research instrument as part of the conference materials. An announcement requesting subjects to voluntarily participate in the
2

Biases can arise from a sample, which does not represent the participant population. Representativeness expresses the degree to which sample data accurately and precisely represents a characteristic of a populations parameter variations at a sampling point. If the sample is representative in some important target variables, then we have increased condence in the validity of the sample to represent the target population. We conduct Chi-Square tests on the sample variables knowledge and experience, which tests whether the frequency distribution of these target variables observed in the sample is consistent with the particular theoretical distribution, since data on the whole population is missing. Results indicate an equal distribution for knowledge (v2 = 86.747) and experience (v2 = 40.615).

A Chi-square test was used to assess response bias. The probability of observing the observed difference if responding subjects are different from non-respondents is approximately v2 = 34.801. This probability is lower than conventional criteria for statistical signicance, so we can conclude that respondents and non-respondents are not signicantly different.

123

The effect of a transparency report on auditor independence Table 1 Experimental treatments and cell sample sizes Oversight arrangement Audit Accountant Regulation approach Mandatory transparency report Voluntary transparency report 11 8 12 14 10 12 12 12 Auditor No audit Accountant

117

Auditor

study was made by the conference director. Since there was no specic time scheduled during the conference to complete the research materials, participants were also provided with a self-addressed envelope in case they preferred to mail their response instead of completing the instrument during the conference. The instrument contained the following materials. First, there was an introduction to the current debate in Europe as to whether or not a transparency report about audit rms own policies and procedures concerning auditor independence should be required. They were informed that EU member states have a broad range of opinions on whether the report should be mandatory or voluntary. The discussion included concerns over the details required and cost versus benet implications. This information was followed by a case study of a hypothetical audit rm R&P, LLP which was said to be located in the United Kingdom. The rst experimental manipulation (Regulation Approach) was implemented as follows. One group of participants were told to assume that in the United Kingdom the regulations include the mandatory publication of an annual transparency report on its own homepage to document the policies and procedures that R&P carries out to maintain auditor independence (mandatory). The other group was informed that in the United Kingdom the professional guidance includes the voluntary publication of an annual transparency report on its own homepage to document the policies and procedures that R&P carries out to maintain auditor independence (voluntary). Related to this experimental treatment, the main objective is to nd the balance between doing justice to the audit rms demand for self-regulation (Marten 1999; Benau et al. 1999) and the need of enforcement for reliable transparent reporting. Mandatory transparency reporting requires high quality signals to decrease information asymmetries between audit rms and stakeholders (Helm et al. 2003). External reviews are discussed as being important to ensure high quality transparency reports and enhanced public condence in transparent quality controls (Marten et al. 2003). Therefore, appropriate oversight, such as a detailed review process might be necessary to ensure the effectiveness of auditor independence policies and procedures (Houghton and Trotman 2003). The second experimental treatment (Oversight Arrangement) was manipulated by informing the participants in the audit treatment that R&P has engaged two independent experts to audit their policies and procedures and the nal audit report of these experts was posted on R&Ps homepage. Furthermore, participants were

123

118

C. Pott et al.

informed that after having audited the policies and procedures as described in the transparency report, the two experts provided the following assurance: Based on our audit, no signicant issues have come to our attention that would cause us to believe that R&Ps policies and procedures to maintain auditor independence is not operating properly. The version (no audit) without an audit did not include any of the preceding information. After having read the introduction and the case study, participants were asked to assess whether or not the policies and procedures described in the transparency report provided an overall framework to improve the rms independence. The question stated the following: The policies and procedures listed in the transparency report are usually thought of as being safeguards to auditor independence risk. What is your viewpoint on how these policies and procedures provide an overall framework to improve the rms independence? Participants were asked to assess the effectiveness of the proposed transparency report on a 7 point Likert-scale, where 0 was labeled as not effective, 3 as moderately effective and 6 as very effective. At the end of the instrument, participants were asked to indicate on a seven point scale which of the safeguards reported in the transparency report might be more important than the others in order to maintain auditor independence. The question was the following: How important is each of the individual safeguards as related to each other? Participants were allowed to go back to the transparency report to reread the information listed in the transparency report, because the listed safeguards were abbreviated as Governance Structure, Internal Quality Control, Quality Assurance, Audited Entities, Independence Practice and Compliance, Continuing Professional Education and Fee Information. Finally, we collected participants demographic information. We asked whether they were currently practicing as an auditor or accountant, their experience in years and for their knowledge of auditor independence regulation self-assessed on a 7 point Likert-scale where the scale ranged from little knowledge and moderate knowledge to high knowledge. The scale was labeled with examples such as casual investor as an example for little knowledge and CPA and audit committee member for high knowledge.

4 Results 4.1 Research questions 1 and 2 Research question 1 asks whether a transparency report is perceived to be more effective in improving auditor independence if regulated or if voluntarily published (Regulation Approach). Whether a transparency report is perceived to be more effective in improving auditor independence if audited or not audited by an expert was asked in research question 2 (Oversight Arrangement). As part of research question 1 and 2 we also distinguish between whether the participant is working as an auditor or accountant (Type of Participant).

123

The effect of a transparency report on auditor independence

119

The descriptive statistics based on the question, where participants were asked to indicate their viewpoint on how the information given in the transparency report provided an overall framework to improve the rms independence are shown in Table 2. A mandatory regulation approach of the transparency report is assessed as being almost the same level of effectiveness as a voluntarily published transparency report (mean = 3.8 vs. 3.89, respectively). However, participants felt an oversight arrangement in the form of an audit of the transparency report was more effective (mean = 4.02) than if the transparency report was not audited (mean = 3.67). The data were analyzed statistically in several stages. A Mann-Whitney U test (also called the MannWhitneyWilcoxon test or the Wilcoxon rank-sum test) was performed. This test is a non-parametric test for assessing whether two samples of observations come from the same distribution. The null hypothesis is that the two samples are drawn from a single population, and therefore that their probability distributions are not signicantly different. The test assumes the two samples are independent. For the full sample analysis, we performed a KruskalWallis one-way analysis of variance by ranks testing equality of population medians among groups. To test whether any of the differences in Table 2 are signicant, the results of a MannWhitney U test and a KruskalWallis one-way analysis of variance of the assessments of the effectiveness of a transparency report are presented in Table 3. As indicated, there is no signicant difference in the subjects assessments as to whether or not the transparency report is more effective to maintain auditor independence whether the transparency report is mandatory or voluntary (z = -1.622, P = 0.538). Results also indicate that the oversight arrangement is also not statistically signicant using conventional signicance thresholds (z = -0.616, P = 0.105). Lastly, whether the participant is practicing as an accountant or an auditor did not results in a signicant difference in how effective a transparency report was assessed to be (z = -0.025, P = 0.98). Interaction effects of two independent variables, e.g. regulation approach and oversight arrangement (v2 = 3.678, P = 0.298), regulation approach and type of participant (v2 = 0.385, P = 0.943) or oversight arrangement and type of participant (v2 = 3.229, P = 0.358), do not indicate signicant differences in the perceived effectiveness of a transparency report. Also the joint effect of all three

Table 2 Assessments of the perceived effectiveness of the transparency report Oversight arrangement Audit Accountant Regulation approach Mandatory Voluntary Marginals 3.82 (0.87) 4.25 (0.71) 4.02 (1.16) 4.00 (1.28) 4.07 (1.49) 3.9 (0.88) 3.75 (0.97) 3.67 (1.26) 3.5 (1.57) 3.58 (1.56) 3.8 (1.18) 3.89 (1.27) Auditor No audit Accountant Auditor Marginals

Scale of 0 (not effective) to 6 (very effective) Values represent means (standard deviation)

123

120

C. Pott et al.

Table 3 MannWhitney U and KruskalWallis one-way test of the assessment of the perceived effectiveness of the transparency report Source of variance Regulation approach Oversight arrangement Type of participant Regulation approach 9 oversight arrangement Regulation approach 9 type of participants Type of participants 9 oversight arrangement z-statistic v2-statistic P value -1.622 -0.616 -0.025 NA NA NA NA NA NA 3.678 0.385 3.229 4.428 0.538 0.105 0.98 0.298 0.943 0.358 0.729

Regulation approach 9 oversight arrangement 9 type of participants NA NA not applicable

independent variables, regulation approach by oversight arrangement by type of participant, does not lead participants to assess the effectiveness of the transparency report differently (v2 = 4.428, P = 0.729). Overall, these results indicate that neither practicing auditors nor practicing accountants see a signicant difference in either whether a independence report is voluntary or mandated or whether there is assurance on it or not. However, the difference concerning the greater effectiveness of audited reports does approach statistical signicance. 4.2 Research question 3 Research question 3 asks how practitioners (practicing auditors and accountants) assess the importance of individual safeguards. As noted, one post experimental question asked participants to indicate the importance of each listed safeguard. The importance of each safeguard was measured on a scale from 0 to 6, indicating that the single safeguard is not important to very important. Descriptive statistics on the overall assessment (N = 91) are shown in Table 4. Over the full sample, the safeguard related to the audit rms independence practice, which also conrms that an internal review of independence compliance has been conducted, is rated to be the most important safeguard (mean = 4.84). Slightly less important is the safeguard related to the internal quality control system of the audit rm and a quality assurance safeguard (mean = 4.78). An indication of when the last audit rms quality assurance review took place (mean = 4.59) is also felt to be very important. A safeguard related to the policy followed by the audit rm concerning continuous education of auditors (mean = 3.86) is assessed to be important, as well as the safeguard related to the governance structure of the audit rm (mean = 3.79). Of moderate importance is a listing of public interest entities for which an audit has been carried out during the last year by the audit rm (mean = 3.2). The least important safeguard maintaining auditor independence is nancial information showing the importance of the audit rm, such as the total turnover divided into fees from the audit of annual and consolidated accounts and fees charged for

123

The effect of a transparency report on auditor independence Table 4 Descriptive statistics: assessments concerning the perceived importance of safeguards Safeguard Full sample (N = 91) Mean Standard deviation Independence practice and compliance (IPC) Internal quality control (IQC) Quality assurance (QA) Continuing professional education (CPE) Corporate governance structure (CG) Audited entities (AE) Fee information (FI) 4.84 4.78 4.59 3.86 3.79 3.2 2.78 1.28 1.32 1.26 1.49 1.65 1.75 1.88

121

Difference in means CPA vs. accountants Mean Standard Mean Standard Chi-square test deviation deviation 5.16 4.82 4.64 3.92 4.10 2.78 2.26 1.20 1.32 1.43 1.63 1.69 1.95 1.83 4.44 4.73 4.54 3.78 3.41 3.71 3.41 1.29 1.34 1.05 1.31 1.53 1.31 1.76 8.46*** 0.13 1.03 0.46 4.40** 5.53** 9.39***

Auditors (N = 50)

Accountants (N = 41)

Scale of 06 with 0 labeled as not important, 3 as moderately important, and 6 as very important * Signicant at 10%; ** signicant at 5%; *** signicant at 1%

other assurance services, tax advisory services and other non-audit services (mean = 2.78). Comparing the assessments for practicing auditors versus practicing accountants, we nd signicant differences in four of the seven safeguards reported on in the transparency report. Reporting on an audit rms governance structure is assessed as being more important to practicing auditors than to practicing accountants (mean = 4.10 vs. 3.41). Furthermore, information on an audit rms independence practice is perceived to be more important by practicing auditors compared to practicing accountants (mean = 5.16 vs. 4.44). However, practicing accountants assess transparent reporting on audited entities and fee information to be moderately more important (mean = 3.71 and 3.41 respectively) compared to auditors (mean = 2.78 and 2.26 respectively). 4.3 Additional results related to knowledge and experience level differences As part of the demographic questions, we also asked participants to indicate their level of knowledge concerning auditor independence regulation. We did not have any participants indicating a low knowledge level. Thus, we only report the assessments of practitioners with moderate and high knowledge level on auditor independence regulation. Concerning the effectiveness of a transparency report, we do not nd individual or interaction differences in whether a participant indicated a moderate or high knowledge level (untabulated). We provide descriptive statistical details of the safeguards by knowledge level in Table 5. We nd three signicant differences between these two groups of participants. First, practitioners with a higher knowledge level assess independence practice and compliance activities as being more important than participants with a

123

122

C. Pott et al.

Table 5 Descriptive statistics: assessments concerning the perceived importance of safeguards by knowledge Safeguard High knowledge level (N = 83) Mean Standard deviation 1.28 1.31 1.28 1.49 1.74 1.79 1.88 Moderate knowledge level (N = 18) Mean Standard deviation 1.36 1.37 1.19 1.32 1.21 1.34 1.51 KruskalWallis chi-square test v2 P value

IPC IQC QA CPE CG AE FI

4.97 4.80 4.64 3.84 3.79 3.01 2.53

4.27 4.66 4.38 3.88 3.77 3.94 3.77

4.3266 0.1367 0.9987 0.0217 0.0568 4.0097 7.0341

0.0375 0.7116 0.3176 0.8828 0.8117 0.0452 0.0080

Scale of 06 with 0 labeled as not important, 3 as moderately important, and 6 as very important

lower regulatory knowledge (mean = 4.97 compared to 4.27). The difference is signicant with P = 0.0375. Second, a lower knowledge level leads to assess information about the audited entities and fees as being more important than higher knowledge (mean = 3.94 compared to 3.01 and mean = 3.77 compared to 2.53, respectively). Both pairs of means are signicantly different with P = 0.0452 and P = 0.008, respectively. Participants also indicated their professional experience in years. Additional analysis on the experience level groups (e.g., two decades experience vs. four decades experience; one decade experience vs. four decades experience) did not indicate different assessments of the effectiveness of the transparency report (untabulated). Table 6 provides descriptive statistical details of safeguards by subjects experience in years. We grouped the sample into experience in years up to nine years (one decade), ten to 19 years (two decades), 20 to 29 years (three decades) and 20 to 40 years (four decades). Panel A provides means and standard deviations for the seven safeguards. Panel B shows results of a performed Fishers Least Signicance Difference Test. Results indicate a signicant difference between the assessments of the safeguard Internal Quality Control (difference in mean -0.92, P \ 0.005). Subjects from 20 with up to 29 years work experience assessed this safeguard to be more important than subjects with experience between ten to 19 years (mean = 4.30 compared to 5.22). More signicant differences in subjects assessment can be reported for the more sensitive information such as details about audited entities and fee information. A signicant difference in assessment can be reported for one decade and two decades experience compared to subjects having the most experience on details about the audited entities (difference in means are 1.82 and 1.2, P \ 0.005, respectively). Where practitioners working in the profession for a maximum of 9 years felt this information was quite important (mean = 4.25) and subjects working from 10 to

123

Table 6 Descriptive statistics details of safeguards by subjects experience in years

Experience in years IQC SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean QA CPE CG AE FI

Professional

Safeguards

IPC

Mean

SD

Panel A: means and standard deviation (SD) 1.03 1.21 1.26 1.52 4.82 1.52 4.56 1.37 3.78 1.50 4.17 1.61 5.22 0.80 4.59 1.27 3.92 1.63 3.92 1.61 3 2.43 4.30 1.48 4.51 1.22 3.63 1.36 3.42 1.73 3.63 5.12 0.83 5 1.19 4.75 1.28 3.75 1.48 4.25 1.03 1.69 1.73 1.75 4.62 2.81 2.33 2.60 1.06 1.87 1.86 1.82

09

1 decade

4.75

1019

2 decades

33

4.69

2029

3 decades

27

4.85

3040

4 decades

23

5.04

Comparison of means QA CPE CG AE FI

Safeguards

IPC

IQC

Panel B: Fishers least signicance difference test 0.48 0.41 0.43 -0.08 -0.05 0.03 0.14 -0.25 -0.15 -0.75 1.20* 0.57 -0.29 -0.50 0.64 0.96 -0.42 1.82* 0.82 -0.17 1.25 1.11 0.32 0.61 1.81* 2.29* 2.02* 0.48 0.21 -0.28

The effect of a transparency report on auditor independence

1 decade

2 decades

0.05

0.82

1 decade

3 decades

-0.10

-0.09

1 decade

4 decades

-0.29

0.29

2 decades

3 decades

-0.15

-0.92*

2 decades

4 decades

-0.35

-0.52

3 decades

4 decades

-0.19

0.39

Scale of 06 with 0 labeled as not important, 3 as moderately important, and 6 as very important

* Signicant difference in means computed with Fishers least signicance difference test

123

123

124

C. Pott et al.

19 years assessed this information to be less but still moderately important (mean = 3.63), the most experienced practitioners did not assess this information to be very important (mean = 2.43). Lastly, subjects in the one decade experience (up to nine years work experience) felt fee information to be very important (mean = 4.62). This assessment is signicantly different from all other assessments (two decades mean = 2.81, three decades mean = 2.33 and four decades mean = 2.6, P \ 0.005).

5 Discussion Transparent information about an audit rms own policies and procedures to maintain auditor independence (Woolf 2006) is potentially an important means of reducing information asymmetry between audit rms, clients and investors. To ensure that the report disclosing each audit forms policies and procedures with respect to auditor independence is published by all public auditing rms, such a report will be mandatory in the European Union starting June 2008. The transparency report is meant to serve as a signaling mechanism, which may be more effective when oversight arrangements are put into place to ensure the reports quality. Findings from our survey (RQ 1 and 2) indicate that the effectiveness of such a report is perceived not to depend on whether or not the report is mandatory or voluntary. Furthermore, the effectiveness of a report that documents the audit rms independence practices and policies is not assessed to differ whether it is audited or not. Both of the results are unchanged when we distinguish between practicing auditors and practicing accountants. The results related to RQ 1 and 2 put the need for having established a mandatory requirement in the European Union into question. Based on our ndings, there is no assessed difference in the quality of the reports information whether the report is mandated or voluntary. One explanation might be that the participants in our study believe that a transparency report is not providing important details and thus, not an effective means of communicating auditor independence. The value of the publication of a transparency report has been anticipated by European auditors even before the EU Directive was issued. In fact, two audit rms have provided a transparency report annually since 2002 on their rms homepage without being required to do so. These reports provide information on KPMGs legal structure, ownership structure, decision making structure and leadership structure (KPMG 2005). Based on these global corporate governance details, KPMG then reports on audit quality mechanisms. Those include, again as now on required in the EU, detailed information on internal quality control policies and procedures, ethics and independence issues, clients and engagements including supervision, review and consultation structures. Similar information is offered by Grant Thorntons report of 2005 and 2006 (Thornton 2005, 2006). Since both audit rms are now reporting for a third year, it is certainly possible that other rms will also follow on a voluntary basis. Interestingly, both reports provide most of the required details. A rationale for voluntary reporting on an audit rms policies and procedures may be to increase investors and potential clients

123

The effect of a transparency report on auditor independence

125

condence in the audit rm, thus maintaining its reputation. The value put on this transparent reporting may be a function of an audit rms performance. If transparent information on an audit rm is withhold then potential investors of clients or potential client are likely to conclude that the information might be bad news (Ross 1979). Withholding transparent information increases market noise because the range of possible interpretations of the information. As a result, the expected cost from loosing reputation and market shares may be so high that audit rms are better served by providing a transparency report (Verrecchia 1983; Suijs 2005). Thus, the need for positive information of an audit rm may prevent punishment by the market. From the perspective of the regulator, the advantages of letting the market decide which information should be provided to interested parties include rst, more exible responses when a change in information is demanded because there is no need to activate a legislative apparatus and second, less cost and time intense implementation. This study also provided evidence as to the relative importance of the safeguards required in the EU mandated report (RQ 3). The information on when the last quality assurance of the audit rm took place is assessed as being very important by practitioners. This nding contradicts concerns of the German Chamber of Public Accountants (2004). The value of this information is high since it may imply a successfully performed quality assurance. This is a signal of quality that can be used as a basis for decision making. Also, investors of existing clients are interested in the reliability of the audit rm that is auditing the company they have invested into or they are planning to invest into. As expected, the most important safeguard is the audit rms independence practice, which also conrms that an internal review of independence compliance has been conducted. Related to the discussion of how audit rms can deal with independence issues, their own policies and procedures are felt as being very important. However, some safeguards importance is assessed differently by auditors and accountants. Information on an audit rms governance structure and independence practice is perceived to be more important by auditors compared to accountants. Accountants assess transparent reporting on audited entities and fee information to be more important compared to auditors. The fact that fee information was assessed by auditors as being rather not important may be due to the type of participants who, although representing stakeholders, are auditors, who are provided with fee information or do not need to be informed of fee information. Similar, KPMG and Grant Thornton also do not provide detailed information on fees and audited entities. This indicates that disclosures of fee information as well as the rms audited entities are considered to be very sensitive details. In addition, since there is a critical discussion occurring on as to whether particularly non-audit fees are not reasonable high, audit rms might be afraid of harming their business reputation. Also, given that independence issues are more effective as assessed by auditors due to their work activities in audit engagements; this group of participants feels an audit rms independence practice to be more important than accountants. These results indicate that individually perceived safeguards effects may differ depending on work-eld related interests. This should be considered when

123

126

C. Pott et al.

regulators propose requirements based only on their own assessments of what are perceived to be important safeguards. Decisions may be supported by information about theoretically sound measures. However, such measures are not well dened in a world with imperfect and incomplete markets (Ballwieser 2004, p. 72). Our study has obtained practitioner impressions about the newly issued transparency report requirement in the European Union. Clearly, audit rms are opposed to the idea of mandatory requirements to publish a transparency report. However, being audited by an expert might promote auditor independence. Interestingly, with less knowledge about auditor independence regulation, we nd fee information and information about the audited entities to be perceived as being more important safeguards to auditor independence. This is also supported by the nding that professionals with up to nine years experience assess fee information as very important compared to all other experience levels. Overall our ndings indicate that regulators should carefully weigh costs and benets of new requirements before implementation. Our study indicates that although a transparency report might offer valuable information for investors or audited companies, the profession itself is somewhat pessimistic regarding a positive effect from more transparent information. A related issue concerns the degree to which personal characteristics of the participants not captured in this study might have affected the results. For example, auditors may be more sensitive to fee information than accountants. Furthermore, our ndings are based on the assumption that our U.S. participants are relatively unaware of the introduction of a transparency report in the EU. However, given that audit rms often work internationally, auditor participants might have known about the proposed transparency report from in-house communication and this might have affected their assessments. Future research should provide evidence on how other stakeholders assess the transparency report. Transparent reporting might be seen differently by different types of stakeholders due to different interests. Comparing ndings related to audit rms providing transparent reporting and to types of audit report users might give valuable insight into different perceptions related to the audit report. Also, given that Member States have to adopt the EU transparency report requirements as a minimum, further research should compare individual Member States transparency report requirements and investigate the association between transparency report differences and their perceived effectiveness. A related issue concerns the degree to which personal characteristics of the participants not captured in this study might have affected the results. For example, auditors and accountants may be more sensitive to fee information than investors.

References
Ball R (2001) Infrastructure requirements for an economically efcient system of public nancial reporting and disclosure, Brooking-Wharton Papers on Financial Services 127169 Ballwieser W (2004) The limitations of nancial reporting. In: Leuz C, Pfaff D, Hopwood A (eds) The economics and politics of accounting. Oxford University Press, Oxford

123

The effect of a transparency report on auditor independence

127

Benau G, Antonia M, Barbadillo ER, Humphrey C, Husaini WA (1999) Success failure? Reections on the changing Spanish audit environment. Eur Account Rev 8(4):701730 Black B (2000) The core institutions that support strong securities markets. Business Lawyer 55(4):1565 1607 Chancy PK, Jeter DC, Shaw PE (1997) Client-auditor realignment and restrictions on auditor solicitation. Account Rev 72(3):433453 DeAngelo LE (1991) Auditor independence, low balling, and disclosure regulation. J Account Econ 3:113127 European Parliament and Council (2006) Directive 2006/43/EC of the European Parliament and of the Council of 17 May 2006 on statutory audits of annual accounts and consolidated accounts, amending Council Directives 78/660/EEC and 83/349/EEC and repealing Council Directive 84/253/EEC, Brussels. Available at http://www.europa.eu.int European Federation of Accountants (2001) The conceptual approach to protecting auditor independence, Brussels. Available at http://www.fee.be/publications/main.htm European Federation of Accountants (2003) Discussion paper european co-ordination of public oversight, Brussels. Available at http://www.fee.be/publications/main.htm German Chamber of Public Accountants (2004) Position of the German Chamber of Public Accountants to European Commissions proposed 8th Directive, Berlin. Available at http://www.wpk.de/ magazin/2-s004/ Gwilliam D, Nacve R, Geoffrey M (2005) The costs and benets of increased accounting regulation: a case study of Lloyds of London. Account Business Res 35(2):129146 Helm R, Mark A, Fischer L-J (2003) Externe Qualitatskontrolle und Qualitatssignale in der WirtschaftsprufungEine empirische Evaluierung des Nutzens fur Mandanten. Die Wirtschaftsprufung 56(23):13011309 Houghton K, Trotman K (2003) Review of KPMG Australias Processed and Policies in Respect of Independence, Conict Resolution and Quality Controls. Available at http://www.kpmg.com.au/ Portals/0/houghton_trotman_2003-11.pdf Institute of Chartered Accountants of England and Wales (2003) Reviewing auditor independence, November 2003, Guidance for audit committees. Available at http://www.icaew.com/index.cfm? route=118446 KPMG (2005) KPMGs Transparency Report, March 2005. Available at http://www.kpmg.com Love I (2003) Financial development and nancing constraints: International evidence from a structural investment model. Rev Financ Stud 16(3):765791 Marten K-U (1999) Externe Qualitatskontrolle im Berufsstand der Wirtschaftsprufer. Der Betrieb 52(21):10731076 Marten K-U, Kohler AG, Meyer S (2003) Umbruch im Peer-Review-SystemDeutscher Status Quo und der Sarbanes-Oxley Act of 2002. Die Wirtschaftsprufung 56(1/2):1017 Rajan RG, Zingales L (1998) Financial dependence and growth. Am Econ Rev 88(3):559586 Ross SA (1979) Disclosure regulation in nancial markets: Implications of modern nance theory and signaling theory. In: Edwards FR (ed) Issues in nancial regulation: regulation of American business and industry pinpoint. McGraw-Hill, New York Schneider A, Church BK, Ely KM (2006) Non-audit services and auditor independence: a review of the literature. J Account Lit 25:169211 Suijs J (2005) Voluntary disclosure of bad news. J Business Financ Account 32(7/8):14231435 Thornton G (2005) Transparency Report 2005. Available at http://www.gti.org Thornton G (2006) Transparency Report 2006. Available at http://www.gti.org United Kingdom Department of Trade and Industry (2005) UK Government Response to the Consultation Process and Updated Regulatory Impact Assessment, London. Available at http://www.dti.gov.uk/ consultations/les/publication-1453.pdf Verrecchia R (1983) Discretionary disclosure. J Account Econ 5:179194 Woolf E (2006) Sarbox in Europe. Accountancy 138(1357):2425 Wyman P (2004) Is auditor independence really the solution? The CPA Journal, 1 April 2004, pp 67

123

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.