Should an audit firm rotation be mandatory? : a research of the influenceof audit firm tenure on audit quality and the difference in the effect of audit firm and audit partner tenure on audit quality

(1)

‘S

HOULD AN AUDIT FIRM ROTATION

BE MANDATORY

?’

A RESEARCH OF THE INFLUENCE OF AUDIT FIRM TENURE ON AUDIT QUALITY AND

THE DIFFERENCE IN THE EFFECT OF AUDIT FIRM AND AUDIT PARTNER TENURE ON AUDIT QUALITY

Master in Accountancy & Control - Specialization Accountancy

Name: Aschna Makhan

Student number: 10681108 Supervisor: Dr. Réka Felleg Word count: 15877

(2)

1 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8

Statement of Originality

This document is written by student Aschna Makhan who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

This thesis investigates if an audit firm rotation should be mandatory. Therefore, this thesis addresses the question to what extent audit firm tenure increases audit quality. Furthermore, this thesis addresses the question if there is a difference in the effect of audit firm and audit partner tenure on audit quality. Based on the questions, I formulated three hypotheses: (I) audit firm tenure is associated with higher audit quality, (II) audit partner tenure is associated with audit quality and (III) there is a difference in the effect of audit firm and audit partner tenure on audit quality. Based on a sample of clients who voluntary switched to a new audit firm in the period 2000-2004, gathered from Audit analytics and Compustat for the period 2000-2014, I found a significant and positive relationship between audit firm tenure and audit quality. This result implies that the audit quality increases when audit firm tenure lengthens. Therefore, there is no support for the calls to implement a mandatory audit firm rotation. Hence, hypothesis one is supported. With regard to the association between audit partner tenure and audit quality no significant relationship was found. Therefore, there is no evidence that an audit partner rotation an effect has on the audit quality, and is hypothesis two rejected. Finally, with regard to the difference in the effect of audit firm and audit partner tenure on audit quality no significant relationship was found between audit partner tenure and audit quality, while a significant and positive relationship was found between audit firm tenure and audit quality. Therefore, there is no evidence that a mandatory audit partner rotation an effect has on the audit quality, while there is evidence shortening the audit firm tenure through the implementation of a mandatory audit firm rotation a negative impact will have on the audit quality. Therefore, there is no support for the calls to implement a mandatory audit firm rotation, and is hypothesis three supported.

(4)

1 Introduction

In this research I examine the relationship between audit firm tenure and audit quality. Specifically, I analyse if there is a difference in the effect of audit firm and audit partner tenure on audit quality.

This research is motivated by the critics on the quality of the performed work of the audit firms, which is caused by the recent global economic crisis. The critics on the audit quality have led to a proposal by European Commissioner Michel Barnier to enhance the audit quality of the audit firms. In his ‘Green Paper’ (Groenboek, 2010), Barnier suggests that a mandatory audit firm rotation should by implemented. Barnier argues that an increase of non-Big4-audit firms to the market of Public Interest Entities (hereafter ‘PIE’) can improve the audit quality. Currently, the PIE market is mainly audited by Big4-audit firms.

Furthermore, Barnier argues that shortening the audit firm tenure through the implementation of a mandatory audit firm rotation can improve the independence of the audit firms. Barnier mentions that situations where a company uses the same audit firm for decades incompatible seem with standards of independence. This due to the risk of familiarity.

However, the proposal by Barnier is not approved by everyone. The European Parliament mentioned that it has his doubts about the proposal to implement a mandatory audit firm rotation. Therefore, they asked the ‘Impact Assessment Unit of the Directorate for Impact Assessment and European Added Value, within the Directorate General for Internal Policies’ of the General Secretariat of the European Parliament to analyse Barnier’s

arguments that shortening the audit firm tenure will enhance the audit quality. From their findings (Statutory audits of public accounts and of public-interest entities, 2012) they came to the conclusion that there is insufficient proof that a decrease of audit firm tenure through an implementation of a mandatory audit firm rotation will enhance the audit quality.

Although the empirical evidence on the relationship between audit firm tenure and audit quality is scarce, Chen, Lin and Lin (2008) examined besides the relationship between audit firm tenure and audit quality also the relationship between audit partner tenure and audit quality. The reason for their research is the lack of prior literature that investigates those relationships. Chen et al. (2008) argue that the several studies that investigated the

relationship between audit firm tenure and audit quality no evidence provide that the audit quality decreases when audit firm tenure lenghtens. Furthermore, Chen et al. (2008) mentions that a mandatory audit firm rotation can not be justified if, after considering the effects of audit partner tenure, long audit firm tenure does not have adverse effects on audit quality. The

(6)

reason for also including the research of audit partner tenure on audit quality is that the audit partner rotation already is mandatory in several countries like the United States, the United Kingdom, the Netherlands and Germany. Furthermore, a mandatory audit firm rotation is more costly than a partner rotation, due to the loss of client specific knowledge. From their findings, Chen et al. (2008) came to the conclusion that when the audit partner or audit firm tenure (after controlling for audit partner tenure) lenghtens the audit quality increases. Therefore, they argue that both rotations will have negative impact on the audit quality.

Based on aforementioned, in this thesis will be investigated if an audit firm rotation should be mandatory. Therefore, in this thesis will be studied to what extent audit firm tenure the audit quality increases. Due to the fact that the audit partner rotation is several countries already is implemented specifically will be investigated if there is a difference in the effect of audit firm and audit partner tenure on audit quality.Should it be the case that audit firm tenure is

associated with a higher audit quality in comparison with audit partner tenure, the effects of the implementation of an audit firm rotation will be worse on the audit quality in comparison with an audit partner rotation. Furthermore, if audit partner tenure is associated with the same level of audit quality as audit firm tenure or with a higher audit quality in comparison with audit firm tenure, then an audit firm rotation should not be mandatory. This question has not been addressed yet by academic researchers. Therefore, this research can be considered unique.

To be able to give an answer to these research questions, I formulated three hypotheses: (I) audit firm tenure is associated with higher audit quality, (II) audit partner tenure is associated with audit quality and (III) there is a difference in the effect of audit firm and audit partner tenure on audit quality. To test the hypotheses, I have used an empirical research method. I have gathered from Audit analytics and Compustat a sample of clients who voluntary

switched to a new audit firm in the period 2000-2004, for the period 2000-2014. To measure the audit quality, I have used the proxy ‘Discretionary accruals’ estimated by the use of the modified Jones model.

After performing the regressions, an association was found between audit firm tenure and the discretionary accruals. Therefore, I can conclude that there is a positive association between audit firm tenure and audit quality, and that hypothesis 1 is supported. With regard to hypothesis 2, no association was found between audit partner tenure and the discretionary accruals. Therefore, there is no evidence of an association between audit partner tenure and audit quality. Hence, hypothesis 2 is rejected. Finally, with regard to hypothesis 3, the results

(7)

reveal that there is a difference in the effect of audit firm and audit partner tenure on audit quality. The results show us a positive association between audit firm tenure and audit quality, while no association was found between audit partner tenure and audit quality. Therefore, hypothesis 3 is supported. In this research I have also performed an additional analysis. I have run a regression, based on the regression equation of hypothesis 3, with the use of the Jones (1991) model. The results from the additional analysis reveal that the results similar are in comparison with results based on the modified Jones model. However, the modified Jones model did not found an association between the size of the client and audit quality, while it was found in regression based on the Jones (1991) model. Hence, I recommend that future research on the differences in the results between the Jones (1991) model and the modified Jones model should be performed.

This research contributes to the prior literature that investigates the relationship between audit firm tenure and audit quality. Because of the fact that the results of this research can be generalized to countries that has the similar accounting and auditing standards as of the United States, this research provides valuable insights to the suggested mandatory audit firm rotation in the Netherlands that will take place on the 1st of january of 2016 (NBA, Scheiding controlediensten & andere werkzaamheden en verplichte kantoorroulatie, December 2012) and countries like the US where the mandatory audit firm rotation not yet has been adopted. Therefore, this research is relevant for parties, such as the PCAOB, the AICPA, the SEC, the European Commission, European Parliament and audit firms.

Due to the fact that the information about a partner’s engagement most of the time not observarble is from public documents such as an audit report or other proxy filings, there is a lack of research done on the relationship between audit partner tenure and audit quality (Litt, Sharma, Simpson and Tanyi, 2014). Furthermore, Chen, Lin and Lin (2008) mentioned in their research that despite the adoption of audit partner rotation in several countries, there still has been a lack of attention in academic research to investigate the relationship between audit partner tenure and audit quality. This thesis therefore also provides a contribution to the prior research done on the relationship between audit partner tenure and audit quality.

The research of Chen et al. (2008) was the only research I found that investigated the influences of the both audit firm tenure and audit partner tenure on audit quality. However, the legal environment in Thailand is different compared to other countries. There are some differences in the audit partners’ reporting incentives and the audit firms’ mechanisms to control or monitor the partners’ behaviour between Thailand and other countries. The audit

(8)

firms in Thailand must be formed as unlimited liability partnerships or proprietorships, which is not required in the United Stated or the United Kingdom. The individual audit partners in Thailand face therefore higher levels of potential liability compared to audit partners of other countries. Furthermore, the strength of their law enforcement is weaker in comparison to other countries. The differences affect the audit partners’ and audit firm’s incentives, which have a negative impact on their audit quality. The results of their research may therefore not be generalizable to countries with a different institutional setting, like the United Stated or the United Kingdom (Chen et al., 2008, p. 417).

This research uses a unique setting. I have used the data of clients registered with the SEC, which has been gathered from the databases Compustat and Audit Analytics. The results of this research can therefore be generalized to countries that has the similar accounting and auditing standards as of the United States. Compared to Chen et al. (2008), this research specifically examines if there is a difference in the effect of audit firm and audit partner tenure on audit quality. This research can therefore be considered unique and provides new evidence and valuable insights.

Besides the contribution, this research has also limitations. Due to the fact that the

information about a partner’s engagement most of the time not observarble is from public documents, a lack of research is done on the relationship between audit partner tenure and audit quality. With the use of my method, followed by Litt, Sharma, Simpson and Tanyi (2014), I was able to determine the audit partner tenure per client. However, regarding hypothesis 2, no association between audit partner tenure and audit quality was found. This can be explained by the selected clients of this data, which were the clients that voluntary switched to new audit firm in the period 2000-2004. Furthermore, this can be explained by the effectiveness of the mandatory audit partner rotation. The audit partner tenure therefore has a maximum length of five years. For future research, I therefore recommend to use my method with a larger sample size or by using another country, like the UK, where you can collect the data of an audit partner tenure with a longer length than five years.

(9)

The remainder of this research is structured as follows. In the next section, I discuss the theoretical background of this research. I explain what audit quality is, what the influences on audit quality are and how it can be measured. Furthermore, in the section I review related studies with regard to audit firm and audit partner tenure and develop hypotheses. The research design and methodology of this research are described in the third section. In the fourth section, I discuss the empirical findings. Finally, in the fifth section follows the discussions and conclusions based on the empirical findings. Furthermore, the section includes a brief summary, the contributions, the limitations and recommendations for future research.

(10)

Context factors

Outputs Inputs

2 Literature review and hypotheses

2.1 Audit quality

Since the beginning of the recent global economic crisis the profession of the auditors started been criticized by the society. Banks reported between 2007 and 2009 enormous losses on their positions due to the high risks that were hiding in those positions (Groenboek, 2010, p. 3). The society wondered how it was possible that the auditors could give unqualified audit reports in those periods. Could the recent global economic crisis have been prevented by the auditors (Veerman, 2010)? The trust in the auditors and their audit quality has been decreased.

Due to the critics on the audit quality, which was the cause of the recent global

financial economic crisis, the IAASB (2011) made a model that consists the influences on the audit quality from the view of different stakeholders. Based on their model1 (IAASB, 2013) there are three kinds of influences, which are inputs, outputs and context factors.

The inputs to the audit quality are auditing standards and the auditor’s personal attributes such as auditor skill and experience, ethical values and mind sets. Furthermore, the audit process also is an important input. It concerns matters as the soundness of the audit methodology, the effectiveness of the audit tools used and the availability of adequate technical

support. The outputs to the audit quality are the auditor’s report and the auditor

communication to those charged with governance. The auditor’s communication regards matters such as the deficiencies in the client’s internal control or the quality of the client’s financial reporting practices. Finally, the context factors also are an influence to the audit quality. Examples of context factors are the client’s corporate governance, law and regulation and the client’s financial reporting framework (IAASB, 2011).

Audit quality is by De Angelo (1981) defined as “the market-assessed joint probability that a given auditor will both (a) discover a breach in the client’s accounting system, and (b) report the breach.”. Based on the definition of De Angelo (1981) there are two criteria to measure the audit quality, which are (I) the professional competence of the auditor and (II) the independence of the auditor. The probability that an auditor will discover a breach in the client’s accounting system is seen as a measure of the professional competence

1_{The influences on the audit quality aren’t one-way and can be interrelated with each other.}

Audit quality

(11)

Audit

quality

Professional competence Independence of the auditor, and the conditional probability of reporting the discovered breach as a way to measure the independence of the auditor (1981, pp.185-186). The definition of De Angelo (1981) consists the criteria which in accordance to the International Ethics Standards Board for Accountants Code of Ethics for Professional Accountants (hereafter ‘the IESBA Code’) an auditor must meet to ensure their audit quality (Verordening gedragscode, 2013, A-100; NIVRA and the NOvAA, 2012 and IFAC, 2013).

The criterion profession competence is under the IESBA Code the fundamental principle professional competence and due care. According to this principle, an auditor should perform professional services with due care, competence and diligence. Furthermore, the auditor should have a continuing duty to maintain professional competence and skills at a level required to ensure that a client or employer receives sufficient professional services based on the applicable technical and professional requirements (Verordening gedragscode, 2013, A-100; NIVRA and the

NOvAA, 2012 and IFAC, 2013).

Regarding the criterion independence, this is under the IESBA Code the fundamental principle independence. According to the IESBA Code, an auditor needs to be both (a) independent in state

of mind and (b) independent in appearance. With regard to the independence of mind, it means that an auditor should only focus on the task that has to accomplished without being affected by influences that comprise the professional judgement, allowing an individual to act with integrity and exercise objectivity and professional scepticism. With regard to the

independence in appearance, an auditor should avoid all facts and circumstances that a reasonable and informed third party, having knowledge of all relevant information, could reasonably conclude that the auditor’s integrity, objectivity or professional scepticism has been compromised (Verordening gedragscode, 2013, A-100; NIVRA and the NOvAA, 2012

and IFAC, 2013).

The professional competence and independence of the auditor can for a third party only be measured from the output of the performed audit, which are financial statements. Examples of proxies to measure the professional competence and independence of the auditor are audit reporting failures like issuing a going-concern opinion for distressed companies (Geiger and Raghunandan, 2002 and Carey and Simnett, 2006), the fraudulent financial reporting (Carcello and Nagy, 2004), the client’s financial restatements (Stanley and Dezoort,

(12)

2007), audit report lags (Lee, Mandy and Son, 2009), the just beating/ missing of earnings benchmarks (Carey and Simnett, 2006), the planning realization rates (Bedard and Johnstone, 20100 and the discretionary accruals (Johnson, Khurana and Reynolds, 2002; Meyers, Meyers and Omer, 2003; Chen, Lin, and Lin, 2008; Chi, Huang, Liao and Xie, 2009; Manry, Mock and Turner, 2008 and Carey and Simnett, 2006)

Summarizing, there is a lot of critics on the audit quality, which is caused by the recent global economic crisis. Based on aforementioned, I can conclude that there is not one universal and correct proxy to measure audit quality. The proxies to measure audit quality do have

something in common. It should be able to measure the professional competence and independence of the auditor as mentioned in the definition of De Angelo (1981).

(13)

12 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8 2.2 Audit firm rotation and audit firm tenure

Due to the recent global economic crisis and the critics on the audit quality, European Commissioner Michel Barnier did a proposal to implement a mandatory audit firm rotation. Barnier argues that situations where a company uses the same audit firm for decades

incompatible seem with the standards of independence. This due to the risk of familiarity (Groenboek, 2010). However, there are a lot of contradictions regarding the implementation of the rotation.

In the US published the Public Company Accounting Oversight Board (hereafter ‘PCAOB’) in August 2011 a concept release where they mention to be a proponent to

implement a mandatory audit firm rotation for listed companies. The release has, however, let to a lot of divided reactions. Proponents of the rotation, like old SEC chairmen Arthur Levitt, see the rotation as a strengthening of the independence of the auditors towards their clients. Opponents of the implementation are enterprises and audit firms. Enterprises are against the rotation due to the increase of audit fees and audit firms argue that the rotation will have negative impact on the audit quality (Arthur Levitt: bezwaar kantoorroulatie niet

steekhoudend, 2012).

In a letter (Request for Public Comment: Concept Release on Auditor Independence and Audit Firm Rotation, 2011) to the PCAOB the American Institute of Certified Public Accountants (hereafter ‘AICPA’) indicated to also be against the implementation of a mandatory audit firm rotation. The AICPA argues that there is no proof that shortening the audit firm tenure will lead to an enhancement of audit quality and therefore recommend the PCAOB to perform further study on this relationship and do not recommend them to pursue a mandatory audit firm rotation requirement. The AICPA believes that a mandatory audit firm rotation will lead to an increase of significant costs and have unintended consequences to potential hinder the audit quality, rather than the intended goal to enhance the audit quality. Furthermore, The AICPA and the Securities and Exchange Commission (United States Securities and Exchange Commission, 2003) argue that a mandatory audit firm rotation could result in limiting companies in specialized industries to engage audit firms who have the proper knowledge and experience to perform the audit. They also mention that the negative impact on multi-national audits will be severe, due to loss of institutional knowledge and experience, and that it will result in significant higher costs.

Arruňada and Paz-Ares (1997) investigated the potential effects of a mandatory audit firm rotation. Their findings show us that the start-up costs for the auditor was at least 15% of all the costs incurred during the time needed to be familiar with the client for companies in

(14)

industries in which the auditor already had experience, and 25% for companies in industries of which the auditor had no experience.

Regarding the implementation of the mandatory audit firm rotation, there is a lot of

discussion. Parties like Barnier and the PCAOB argue that shortening the audit firm tenure will enhance the audit quality. This because of the strengthening of the audit firms’

independence towards their clients. However, according to the AICPA, there is no proof that shortening the audit firm tenure will enhance the audit quality. The AICPA and the audit firms believe that, instead of enhancing the audit quality, shortening the audit firm tenure will have negative impact.

Regarding the association between audit firm tenure and audit quality I found a couple of studies that investigated this relationship, so examined Geiger and Raghunandan (2002) the relationship between audit firm tenure and audit quality through the proxy ‘audit reporting failures’. Reason for their research and the use of the proxy ‘audit reporting failures’ is the call from the SEC to investigate this relationship and the interests regarding the mandatory audit firm rotation. To measure the proxy ‘audit reporting failures’ Geiger et al. (2002) examined the audit report data in the 10-K filings from the period 1996-1998. They examined if the audit reports of companies entering into bankruptcy had been modified for

going-concern uncertainties during the years. When the audit report hasn’t been modified, it gives an indication of an audit reporting failure. The results of their research show us that the audit reporting failures more likely are to occur in the initial years of an audit engagement. Furthermore, the results show us that the audit reporting failures decreases when audit firm tenure lengthens. Therefore, there is no support for the calls to implement a mandatory audit firm rotation to enhance the audit quality.

Carcello and Nagy (2004) examined the relationship between audit firm tenure and audit quality through the use of the proxy ‘fraudulent financial reporting’. To measure the fraudulent financial reporting Carcello et al. (2004, p.59) read all the Accounting and Auditing Enforcement Releases issued by the SEC between 1990 and 2001 to identify the firms and/or officers that have been charged with a violation of Rule 10(b)-52 by the SEC. The findings of their research show us that fraudulent financial reporting more likely is to occur in the first three years of audit firm tenure. Furthermore, they found no evidence that fraudulent financial reporting more likely is to occur given long auditor tenure. Carcello et al.

(15)

(2004) therefore argue that an implementation of a mandatory auditor rotation will have negative impact on the audit quality.

Stanley and DeZoort (2007) researched the relationship between audit firm tenure and the client’s financial restatements. Reason for the use of the proxy ‘financial restatements’ is that they are de facto of reporting failures. When audited financial statements are restated is validity of the audit opinion and the underlying audit process subject to question, because the originally filed financial statements were not free of material misstatements. Therefore, the audit quality can be measured in a direct way. To measure the proxy ‘financial restatements’ Stanley et al. (2007) gathered the data of companies with annual restatements involving intentional or unintentional misapplications of GAAP. Stanley et al. (2007) gathered the data of annual SEC filings within 10-K Wizard from the period 2000-2004. From the findings of their research, they came to conclusion that the financial restatements decreases when audit firm tenure lengthens. Therefore, they find no support for the calls to implement a mandatory audit firm rotation

Lee, Mandy and Son (2009) researched the relationship between audit firm tenure and the audit report lags. Reason for the use of the proxy ‘audit report lags’ is that they directly are associated with the audit efficiency and the timeliness of companies’ announcements of earnings. To measure the proxy ‘audit report lags’ Lee et al. (2009) gathered the data available in Audit Analytics from the period 2000-2005. The data concerns audit report lags that has a length between 14 and 180 days. The results of their research show us that the audit report lags decline when audit firm tenure lengthens. This indicates that audit firms with long tenure are able to audit their clients more efficiently. Therefore, Lee et al. (2009) argue that audit firm switches will lead to audit delays and an increase of informational inefficiencies in the form of delayed information to the markets.

Johnson, Khurana and Reynolds (2002) investigated the relationship between audit firm tenure and the quality of financial reports. To measure the quality of financial reports Johnson et al. (2002) gathered the data of all corporations in the United States that did not change fiscal year-ends in the period 1986-1995. Furthermore, they used the proxy ‘the absolute value of unexpected accruals’ so that the extent of management interventions can be measured. Secondly, they used the proxy ‘the persistence of the accrual components of earnings’ so that the quality of the accruals can be measured. To measure the accruals, Johnson et al. (2002) used the modified Jones model. The results from the research of Johnson et al. (2002) show us that that medium audit firm tenure (four to eight years) and short audit firm tenure (two to three year) are associated with financial reports which has

(16)

lower quality. Furthermore, Johnson et al. (2002) found no evidence that long audit firm tenure (nine years or more) is associated with reduced financial reporting quality

Meyers, Meyers and Omer (2003) examined the relationship between audit firm tenure and the quality of earnings. Through the independent variable ‘earnings quality’ can the degree of earnings management be measured. Therefore, Meyers et al. (2003) assert that earnings quality can be used to draw inferences about audit quality. To measure the earnings quality, Meyer et al. (2003) gathered the data available in Compustat from the period 1988-200. Furthermore, they used the proxy ‘discretionary accruals’, estimated by the use of the Jones (1991) model. The results of their research show us that the earnings quality higher is when audit firm tenure lengthens. Therefore, Meyer et al. (2003) conclude that there is no support for the calls to implementation a mandatory audit firm rotation

Finally, Chen, Lin, and Lin (2008) also investigated the relationship between audit firm tenure and earnings quality, after controlling for audit partner tenure. The reason for their research is that there is a lack of prior research done on the effects of the audit firm rotation and the audit partner rotation. In their opinion a mandatory audit firm rotation can’t be justified if, after considering the effects of audit partner tenure, long audit firm tenure does not have adverse effects on audit quality. The reason for also including the proxy ‘audit partner tenure’ is that the audit partner rotation already is mandatory in several countries like the United States, the United Kingdom, the Netherlands and Germany. Furthermore,

according to Chen et al. (2008), a mandatory audit firm rotation is more costly than a partner rotation. This due to the loss of client specific knowledge. Like Myers et al. (2003), Chen et al. (2008) also assert that the earnings quality can be used to draw inferences about audit quality. To measure the earnings quality, Chen et al. (2008) gathered the data available in the Taiwan Economic Journal database from the period 1990-2001. Furthermore, they used the modified Jones model. From the findings of their research, Chen et al. (2008) came to the conclusion that the earnings quality increases when the audit partner or the audit firm tenure lenghtens. Therefore, Chen et al. (2008) believe that both rotations will have an negative impact on audit quality.

Proponents of the implementation of a mandatory audit firm rotation, like Barnier, the

PCAOB and SEC chairmen Arthur Levitt argue that long audit firm tenure will have negative impact on the audit quality. This due to the risk of familiarity with the client. Therefore, according to the proponents, shortening the audit firm tenure will enhance the audit quality. However, audit firms and the AICPA argue that the mandatory audit firm rotation will have

(17)

negative impact on the audit quality. According to the AICPA, there is no proof that shortening the audit firm tenure will enhance the audit quality. Furthermore, prior research (Geiger and Raghunandan, 2002; Carcello and Nagy, 2004; Stanley and DeZoort, 2007; Lee, Mandy and Son, 2009; Johnson, Khurana and Reynolds, 2002; Meyers, Meyers and Omer, 2003; and Chen, Lin, and Lin, 2008) show us that there is no evidence that long audit firm tenure is associated with lower audit quality, and that shortening the audit firm tenure will enhance the audit quality. Although normative statements suggest that long audit firm tenure is associated with a lower audit quality, empirical findings suggest a positive association. Therefore, I expect that audit firm tenure is positively associated with audit quality. Hence, I formulated the following hypothesis:

Hypothesis 1

(18)

17 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8 2.3 Audit partner rotation and audit partner tenure

In July 30, 2002 the Sarbanes-Oxley Act (SOX) implemented stringent independence rules to restore the public trust and investor confidence and enhance the auditor independence,

objectivity and profession scepticism. One of these rules is the mandatory audit partner rotation, which is required every five years for the lead and concurring partner and every seven years for the other partners. The AICPA believes that the requirements of the audit partner rotation are effective and provide the fresh look, which has a positive impact on the auditor’s independence. Furthermore, according to the AICPA, the audit partner rotation does not carry the same degree of disruption and loss of institutional knowledge as that of a

mandatory audit firm rotation. Moreover, the AICPA believes that the level of disruption and loss of knowledge will be even worse if the existing partner rotation is coupled with a

mandatory audit firm rotation (United States Securities and Exchange Commission, 2003 and Request for Public Comment: Concept Release on Auditor Independence and Audit Firm Rotation, 2011).

Due to the fact that the information about a partner’s engagement most of the time not

observarble is from public documents such as an audit report or other proxy filings, there is a lack of research done on the relationship between audit partner tenure and audit quality. Nonetheless, I found a couple of studies that investigated this relationship. Chen, Lin and Lin (2008) examined besides the relationship between audit firm tenure and earnings quality, also the relationship between audit partner tenure and earnings quality. From the findings of their research Chen et al. (2008) came to the conclusion that the earnings quality increases when audit partner tenure lengthens. Therefore, Chen et al. (2008) argue that the implementation of an audit partner rotation will have a negative impact on the audit quality. Chi, Huang, Liao and Xie (2009) examined the relationship between audit partner tenure and the audit quality through the use of the proxies ‘abnormal and signed abnormal accruals’. To measure the proxies ‘abnormal and signed abnormal accruals’, Chi et al. (2009) gathered the data available in Taiwan Economic Taiwan Journal database from the period 1999-2004. Furthermore, they used the Jones (1991) model and the modified Jones model. The results of their research show us that the audit quality of companies subject to a mandatory audit partner rotation does not differ from the companies which are not subject to a mandatory audit partner rotation. Furthermore, the results show us that the audit quality from the companies subject to a

mandatory audit partner rotation lower is than the year before the rotation. Therefore, there is no evidence that shortening the audit partner tenure through the implementation of an audit

(19)

partner rotation will enhance the audit quality.

Bedard and Johnstone (2010) researched the relationship between audit partner tenure and audit planning and pricing. To measure the audit planning and pricing, Bedard et al. (2010) gathered risk assessments 2002-2003 and the audit planning/pricing decisions for an audit firm’s continuing public clients in 2002. Reason for the use of the proxy ‘audit planning and pricing’ is that prior research not yet has addressed whether engagement processes differ based on audit partner tenure. The findings of their research show us that following a partner rotation, new audit partners more effort invest to gain client knowledge in the first year of the engagement. Furthermore, the results show us that the realization of the audit planning lower is after an audit partner rotation. Bedard et al. (2010) therefore argue that the investments by new audit partners’ in client knowledge not are compensated after the rotation of audit partner has taken place. Besides aforementioned, the research of Bedard et al. (2010) also provide evidence that the realization of the audit planning rates higher is when the audit partner tenure is longer than five years. Therefore, there is no evidence for the calls to implement an audit

partner rotation.

Manry, Mock and Turner (2008) also examined the relationship between audit partner tenure and audit quality. To operationalize audit quality, Manry et al. (2008) gathered two years of data from audits of 202 clients conducted by three audit firms. For two of the audit firm the clients had fiscal years ended in 1999-2000. While the fiscal years for the third firm ended in 2000-2001. To measure the audit quality, Manry et al. (2008) used the proxy ‘discretionary accruals’, estimated by using the Jones (1991) model. The findings of their research show us that for small clients that have an audit partner tenure greater than seven years, the audit quality increases when audit partner tenure lengthens. With regard to large clients or shorter-tenure smaller clients, the research of Manry et al. (2008) did not find a relationship between audit partner tenure and audit quality.

Finally, Carey and Simnett (2006) also examined the relationship between audit partner tenure and audit quality. To measure the audit quality, Carey et al. (2006) gathered the data of the public companies listed on the Australian Stock Exchange in 1995. Furthermore, they used three proxies to operationalize the audit quality, which are: (I) the propensity to issue a going-concern audit opinion for distressed companies, (II) the discretionary accruals and (III) the just beating/missing of earnings benchmarks. Regarding the measures ‘I’ and ‘III’, the results show us that the audit quality decreases when audit partner tenure lengthens. With regard to the measure ‘II’, their research found no evidence of a relationship between audit partner tenure and audit quality.

(20)

To enhance the audit quality and the audit partner’s independence the Sarbanes-Oxley Act (SOX) implemented in July 30, 2002 a mandatory audit partner rotation. According to the AICPA, the requirements of the audit partner rotation are effective and enhances it the auditor’s independence. Prior research show us mixed results. Chen, Lin and Lin (2008) and Bedard and Johnstone (2010) show us that long audit partner tenure is associated with higher audit quality. While Carey and Simnett (2006) provide us with evidence that long audit partner tenure is associated with lower audit quality. Furthermore, Chi, Huang, Liao and Xie (2009) and Bedard and Johnstone (2010) show us that there is no evidence that shortening the audit partner tenure the audit quality enhances. The research of Manry, Mock and Turner (2008) furthermore found no evidence of a relationship between audit partner and audit quality with regard to large clients or shorter-tenure smaller clients, while they did found it with regard to small clients that has an audit partner tenure greater than seven years.

Therefore, it is clear that there is an association between audit partner tenure and audit quality. However, the opinion of the AICPA combined with the results of prior research show us mixed results regarding the direction of the association between audit partner and audit quality. Therefore, I formulated the following hypothesis:

Hypothesis 2

(21)

2.4 The difference in the effect of audit firm and audit partner tenure on audit

quality

Proponents of the implementation of a mandatory audit firm rotation see the rotation as a strengthening of the auditors’ independence towards their clients and therefore an

enhancement of the audit quality (Arthur Levitt: bezwaar kantoorroulatie niet steekhoudend, 2012; and Groenboek, 2010). Opponents of the rotation, however, argue that a mandatory audit partner rotation in several countries already is implemented and the requirements of the audit partner rotation are effective and that it enhances the audit partner’s independence. Furthermore, the opponents argue that the implementation of mandatory audit firm rotation more costly is comparison with a mandatory audit partner rotation (United States Securities and Exchange Commission, 2003 and Request for Public Comment: Concept Release on Auditor Independence and Audit Firm Rotation, 2011 and Chen, Lin and Lin, 2008).

Prior research suggests that audit firm tenure is associated with higher audit quality. However, regarding the relationship between audit partner tenure and audit quality, prior research show us mixed results regarding the direction of the association between audit partner and audit quality. Based on the results, I can only conclude that audit partner tenure associated is with audit quality. Because of the fact that an audit firm rotation not yet mandatory is, while mandatory audit partner rotation in several countries already is implemented, in this research I specifically examine if there is a difference in the effect of audit firm and audit partner tenure on audit quality. Based on aforementioned, I expect that there is a difference in the effect of audit firm and audit partner tenure on audit quality. Hence, I formulated the following hypothesis:

Hypothesis 3

(22)

3 Research Methodology

3.1 Data and sample size

In order to investigate to what extent audit firm tenure the audit quality increases and if there is difference in the effect of audit firm and audit partner tenure on audit quality, an empirical research method is used. I have analysed the data which was

available in the databases Audit Analytics and Compustat in the period of 2000-2014. The sample size of this research is 600 observations from 73 clients of audit firms registered with the SEC. I selected the data from clients of audit firms registered with the SEC to have enough of data to determine the audit partner tenure. This would be less had I chosen countries like the Netherlands or the United Kingdom. Information about a partner’s engagement with a client is not observable from public documents such as an audit report or other proxy filings. Furthermore, it is not visible in the

filings with the SEC (Litt et al., 2014), which the reason is of the lack of research done on the relationship between audit partner tenure and audit quality. To be able to gather the data of both audit firm and audit partner tenure, I followed the method by Litt et al. (2014) and selected the clients who voluntary switched to a new audit firm in the period 2000-20043. Due to the fact that the audit firm rotation in the United Stated not yet mandatory is consists the sample clients that voluntary switched to a new audit firm. Clients that switch to a new audit firm automatically have a new audit partner beginning at the year of change. Therefore, I am also able to determine the audit partner tenure. The period from 2000 has been selected because of the mandatory audit partner rotation, which under SOX in the U.S. as of May 6, 2003 effective became. The mandatory rotation of the lead and concurring partner

furthermore is applied retroactively and required at an audit partner tenure of five years (United States Securities and Exchange Commission, 2003). The year 2000 is therefore the earliest fiscal year-end subject to the rotation rules (Litt et al., 2014). From the selected clients I have gathered the financial data and the data regarding their voluntary audit firm switches in the period 2000-20144.

3_{To have at least one audit partner rotation has the period 2000-2004 been selected (Litt et al., 2014)} 4_{Starting point of the gathered financial data per firm is the voluntary switch in the period 2000-2004}

(23)

22 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8 3.2 Variables

Audit firm tenure

As mentioned in chapter 2 is expected that the audit firm tenure is associated with audit quality (Hypothesis 1). To operationalize audit firm tenure, this research uses the proxy

‘Audit firm tenure’. Audit firm tenure is measured as the length of the relationship between an audit firm and their client. During this relationship the audit firm needs to be employed by the client. To be able to measure this relationship I have gathered the engaged date and dismissed date of each audit firm the client has employed during the period 2000-2014. At the engaged date the audit firm tenure is zero. To be able to measure the audit firm tenure I therefore selected the clients that switched during the period 2000-2004. I than gathered the voluntary

switches per client during the period 2000-2014.

Audit partner tenure

As mentioned in chapter 2 is expected that audit partner tenure is associated with audit quality (Hypothesis 2). To operationalize audit partner tenure, this research uses the proxy ‘Audit partner tenure’. Audit partner tenure is measured as the length of the relationship between an audit partner and their client. Besides the reason to be able to measure the audit firm tenure, the clients that voluntary switch to a new audit firm also has been selected to be able measure the audit partner tenure. The mandatory audit partner rotation in the U.S. became as of May 6, 2003 effective. Due to the fact that the mandatory rotation of the lead and concurring partner retroactively is applied and is required at an audit partner tenure of five years, the year 2000 is the earliest fiscal year-end subject to the rotation rules (Litt et al., 2014; United States

Securities and Exchange Commission, 2003). To be able to gather the data of at least one audit partner rotation I followed the method by Litt, Sharma, Simpson and Tanyi (2014) and selected the clients that voluntary switched to a new audit firm in the period 2000-2004.

Data regarding the audit partner tenure is not available in the databases Compustat and Audit Analytics. Therefore, I used the engaged date and dismissed date of each audit firm the selected client has employed during the period 2000-2014. Clients that switch to a new audit firm have a new audit partner beginning at the year of change. At the engaged date the audit partner tenure is therefore zero. Due to the effectiveness of the mandatory rotation of the lead and concurring partner in the sample of this research, the audit partner tenure has a maximum length of five years.

(24)

Audit quality

The definition of audit quality by De Angelo (1981) is based on two criteria, which are (a) the professional competence of the auditor and (b) the independence of the auditor. The

professional competence and independence of the auditor can for a third party only be

measured from the output of the performed audit, which are financial statements. Examples of proxies to measure the professional competence and independence of the auditor are audit reporting failures. Most of the researches done on the relationship between audit firm tenure and audit quality and the relationship between audit partner tenure and audit quality uses the proxy ‘discretionary accruals’ to measure the audit quality. Following Johnson, Khurana and Reynolds, (2002), Meyers, Meyers and Omer (2003), Chen, Lin, and Lin (2008), Chi, Huang, Liao and Xie (2009), Manry, Mock and Turner (2008) and Carey and Simnett (2006), this research therefore uses the proxy ‘Discretionary accruals’ to measure the audit quality.

The discretionary accruals are the portion of the total accruals unexplained by normal operational activities. Having high discretionary accruals is an indication of earnings

management. Therefore, a decrease (increase) of discretionary accruals is seen as an increase (decrease) of audit quality. There are a couple of popular methods used by academic research to determine the proxy ‘Discretionary accruals’.

The first method is the Healy (1985) model. Healy (1985) determines the discretionary accruals by estimating the deviation compared to the normal level of the accruals. Healy (1985) predicts that earnings management in every period systematic occurs. To measure the discretionary accruals Healy (1985) compares the mean total accruals scaled by lagged total assets. The difference between the total accruals in an event year scaled by lagged total assets and the nondiscretionary accruals in the event year are the discretionary accruals (Dechow, Sloan and Sweeney, 1995 and Bartov, Gul and Tsui, 2000)

The second method is the De Angelo (1986) model. This model is similar to the Healy (1985) model. The difference compared to the Healy (1985) model is that it uses the last period’s total accruals scaled by lagged total assets as the measure of nondiscretionary accruals, instead of the mean total accruals lagged by total assets (Dechow et al., 1995)

A common feature of the Healy (1985) and De Angelo (1986) model is that they both predict that the nondiscretionary accruals are constant over time (Dechow et al., 1995).

However, according to Kaplan (1985), the level of nondiscretionary accruals should change in response to changes in economic circumstances. Due to the limitation of the models of not

(25)

including the changes in economic circumstances in their calculation of the discretionary accruals, those models are in this research therefore not used.

The third method is the Jones (1991) model. The Jones (1991) model does not assume the nondiscretionary accruals to be constant over time and therefore tries to attempt to control for the effect of changes in a firm’s economic circumstances on nondiscretionary accruals. The Jones (1991) model for the nondiscretionary accruals in the event year is:

NDAit = α1(1 / Ait - 1) + α2(∆REVit / Ait - 1) + α3(PPEit / Ait - 1 ) (A)

Where:

NDAit = nondiscretionary accruals in year t scaled by lagged total assets for firm i;

Ait – 1 = total assets in year t -1 for firm i;

∆REVit = revenues in year t less revenues in year t -1 for firm i;

PPEit = gross property, plant and equipment in year t for firm i; and

α1, α2, α3 = firm-specific parameters.

To control for the differences in firm level, the variables are scaled by total assets. The firm-specific parameters α1, α2 and α3 are estimated in the event year by using the following equation:

TAit / Ait - 1 = a1(1/Ait - 1) + a2(∆REVit/ Ait - 1) + a3(PPEit / Ait - 1) + εit (B) Where:

a1, a2, a3 = OLS estimates of α1, α2, α3

TAit = total accruals in year t for firm i

ε

it = error term in year t for firm i, which are the discretionary portion of the

total accruals; and

other variables = as in equation (A).

The fourth method is the modified Jones model. The difference compared to the Jones (1991) model is that the conjectured tendency of the Jones (1991) model is eliminated to measure the discretionary accruals with error when discretion over revenues is exercised (Dechow et al., 1995). Therefore, the modified Jones model makes an adjustment in the original Jones (1991) model, which is adjusting the change in the annual revenues for the

(26)

change in net receivables in the event year. The modified Jones model for the nondiscretionary accruals in the event year is:

NDAit = α1 (1/Ait - 1) + α2 [(∆REVit - ∆RECit) / Ait - 1] + α3 (PPEit / Ait - 1) (C)

Where:

∆RECit = net receivables in year t less net receivables in year t -1 for firm i;

α1, α2, α3 = obtained from the original Jones (1991) model; and

other variables = as in equation (A).

The last method is the Industry model. This model also assumes that the

nondiscretionary accruals are not constant over time. Therefore, this model also attempts to control for the effect of changes in a firm’s economic circumstances on nondiscretionary accruals. Difference compared to the previous models is that it does not attempt to directly model the determinants of nondiscretionary accruals. Instead, this model assumes that the variation in the determinants of nondiscretionary accruals common is across firms in the same industry (Dechow et al., 1995; and Bartov et al., 2000). The Industry model for the

nondiscretionary accruals in the event year is:

NDAt = 1 + 2medianj(TAt/ At - 1) (D)

Where:

medianj(TAt/ At - 1) = median value of total accruals in year t scaled by lagged total assets for all non-sample firms in the same two-digit standard industrial

classification (SIC) code (industry j);

1, 2 = firm-specific parameters estimated using OLS;

and

other variables = as in equation (A) and (B).

Dechow Sloan and Sweeney (1995) and Guay, Kothari and Watts (1996) tested the five methods, as mentioned above, for their effectiveness to measure the discretionary accruals. From their findings they came to the conclusion that the Jones (1991) model and the modified Jones model the most effective and powerful methods are to determine the discretionary accruals. Peasnell, Pope and Young (2000) examined the effectiveness of the Jones (1991)

(27)

model and modified Jones model to measure the discretionary accruals. For their findings they came to the conclusion that both models powerful are at detecting subtle revenue and bad debt manipulations. Following Johnson, Khurana and Reynolds (2002), Chen, Lin, and Lin (2008) and Chi, Huang, Liao and Xie (2009), I therefore decided to use the modified Jones model to estimate the proxy ‘Discretionary accruals’.

Following the modified Jones model, the discretionary accruals are measured by the following equation:

DAit = TAit/ Ait - 1

-

[β0 (1 / Ait - 1) + β1 (∆REVit - ∆RECit / Ait - 1) + β2 (PPEit / Ait - 1 )]

Where:

DAit = discretionary accruals in year t scaled by lagged total assets for firm i;

other variables = as in equation (A), (B) and (C); and

β0, β1, β2 = firm-specific parameters estimated using OLS.

Following Healy (1985), Jones (1991), Dechow et al. (1995) and Guay at al. (1996), the total accruals over the event year are measured by:

TAit = ΔCAit – ΔCLit - ΔCashit + ΔLTDit - DEPit

Where:

TAit = as in equation (B);

ΔCAit = change in current assets in year t for firm i;

ΔCLit = change in current liabilities in year t for firm i;

ΔCashit = change in cash and cash equivalents in year t for firm i;

ΔLTDit = change in long-term debt in year t for firm i; and

(28)

Control variables

Besides audit firm tenure, prior literature also describes other factors that have an influence on audit quality. Those factors are in this research included as control variables.

Size of the audit firm

European Commissioner Michel Barnier argues that an increase of non-Big4-audit firms to the market of Public Interest Entities (PIE) can improve the audit quality. This through the implementation of a mandatory audit firm rotation (Groenboek, 2010). However, De Angelo (1981), Davidson and Neu (1993) and Choi, Kim and Zang (2010) provide evidence that BigN-firms a higher audit quality provide compared to non-BigN-firms. According to De Angelo (1981) and Davidson et al (1993, it is caused by the reputation risk, which is higher for BigN-firms. BigN-firms therefore have a higher incentive to be independent. Choi, Kim and Zang argue in their research that it is also caused by the fact that BigN-offices more clients have than non-BigN-offices, making them less dependent on a single client. According to Choi et al., BigN-offices therefore are less sensitive to the pressure of the customer and the risk for delivering a substandard or biased statement (2010, p. 94).

Based on Barnier’s argument and the results of prior research, I can conclude that the size of an audit firm is associated with audit quality. Barnier’s argument combined with the results of prior research, however, give mixed results regarding the direction of the association between BigN-firms and audit quality. Although Barnier argues that an increase of non-Big4-audit firms to the market of Public Interest Entities (PIE) can improve the audit quality, empirical findings suggest a positive association between BigN-firms and audit quality. In this research therefore is expected that the size of the audit firm positively is associated with audit quality. To measure the size of an audit firm, this research uses the proxy ‘BigN’. The indicator is set to 1 if the auditor is BigN-firm and set to 0 otherwise.

Client size

The AICPA and the Securities and Exchange Commission argue that the implementation of a mandatory audit firm rotation will have a significant negative impact on multi-national audits and that it will result in significant higher costs. This due to loss of institutional knowledge and experience (Request for Public Comment: Concept Release on Auditor Independence and Audit Firm Rotation, 2011 and United States Securities and Exchange Commission, 2003). Myers, Myers and Omer (2003) and Chen, Lin and Lin (2008) provide evidence that the size

(29)

of a client is associated with higher audit quality. With regard to larger and more mature clients, more information is public (Carcello and Nagy, 2004). According to Carcello and Nagy (2004), those clients therefore will act less fraudulent, which has a positive influence on the audit quality. From the findings of Geiger and Raghunandan (2002), however, there is a negative association between the likelihood of going-concern modified audit opinion and the client size. This has a negative impact on the audit quality.

Based on aforementioned, it is clear that there is an association between client size and audit quality. However, the results from prior research show us mixed results regarding the

direction of the association between client size and audit quality. Therefore, in this research is expected that the client size is associated with audit quality. To operationalize the client size, this research uses the proxy ‘Total Sales’. Following Geiger and Raghunandan (2002), the client size is measured by the total sales of the client in year t.

Industry specialization

The AICPA and the Securities and Exchange Commission argue that a mandatory audit firm rotation could result in limiting companies in specialized industries to engage audit firms who have the proper knowledge and experience to perform the audit (Request for Public

Comment: Concept Release on Auditor Independence and Audit Firm Rotation, 2011 and United States Securities and Exchange Commission, 2003). Krishnan (2003) and Taylor (2002) provide evidence that industry specialized audit firms are associated with higher audit quality. The discretionary accruals of their clients are lower, which is an indication of using less earnings management. Furthermore, the industry specialized audit firms are less

conservative in their risk assessments and the risk assessments regarding the industry specialized accounts are of higher quality compared to those of non-industry specialized auditors.

Based on aforementioned, in this research therefore is expected that industry specialized audit firms are associated with higher audit quality. To operationalize the industry specialization of the audit firm, this research uses the proxy ‘Industry specialization’.

The industry specialization of the audit firm is not observable. Therefore, the researchers must rely on proxies to estimate the industry specialization. Yardley, Kauffman, Chairney and

(30)

Albrecht (1992) developed a measure of the audit firm’s industry specialization that estimates the industry specialization by the proportion of audit firm’s audit fees earned from one

industry of all those served. Due to the fact that the information about the audit fees

unavailable was until recently, researchers used proxies as sales, assets or the square root of assets as the base to estimate the proportion of audit fees received from a particular industry (Krishnan, 2003). Krishnan (2003, p.3) used two methods to estimate the audit firm’s industry specialization, which are: (I) the audit fees an audit firm earns in an industry relative to the total audit fees earned by all auditors serving that particular industry and (II) an audit firm’s audit fees earned from an industry relative to fees earned from clients across all industries served.

Due to the lack of data available in the database Audit Analytics about the proxy ‘Audit fees’, this research uses the proxy ‘Sales’ to estimate the audit firm’s industry

specialization. Following the method used by Krishnan (2003), the industry specialization of the audit firm in this research is measured by the sales an audit firm earns in an industry (for firm i) relative to the total sales earned by all audit firms serving that particularly industry in year t. Based on the classification within Compustat contains this sample nine industries, which are:

Economic sector Code within Compustat

Energy 10 Materials 15 Industrial 20 Consumer discretionary 25 Consumer staples 30 Health care 35 Financials 40 Information technology 45 Utilities 55

(31)

30 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8 3.3 Conceptual and empirical model

3.3.1 Conceptual model

Based on chapter 2 and 3.2, the conceptual model of this research is as follows:

Source: Author, 2015. Compiled from chapter 2 and 3.2.

3.3.2 Empirical model

The extent to what audit firm tenure the audit quality increases and if there is difference in the effect of audit firm and audit partner tenure on audit quality in this research is investigated, after controlling for the other factors that also have an influence on the audit quality. Audit quality is in this research measured by the proxy ‘Discretionary accruals’. An increase of the discretionary accruals is an indication of earnings management. Therefore, an increase of the discretionary accruals indicates a decrease in audit quality. Based on chapter 2 and 3.2, the regression equation of hypothesis 1 is as follows:

LNDAmodifiedJonesmodelit = β0 + β1 Audit firm tenureit + β2 BigNit + β3 LNTotalSalesit + β4 LNIndustryspecializationit + εit (I) Independent variable (I):

Audit firm tenure

Control variables: Industry specialization

Size of the audit firm

Dependent variable: Audit quality

Control variable: Client size

Size of the audit firm Independent variable (II):

Audit partner tenure Independent variables (III):

Audit firm tenure Audit partner tenure

(32)

With regard to hypothesis 2, the regression equation is as follows:

LNDAmodifiedJonesmodelit = β0 + β1 Audit partner tenureit + β2 BigNit + β3 LNTotalSalesit + β4 LNIndustryspecializationit + εit (II) Finally, with regard to hypothesis 3 the regression equation is as follows:

LNDAmodifiedJonesmodelit = β0 + β1 Audit firm tenureit + β2 Audit partner tenureit + β3 BigNit + β4 LNTotalSalesit + β5 LNIndustryspecializationit + εit (III) Where:

LNDAmodifiedJones modelit

= discretionary accruals in year t scaled by lagged total assets for firm i;

Audit firm tenureit = number of consecutive years that the firm has been audited

by the same audit firm in year t for firm i;

Audit partner tenureit number of consecutive years that the firm has retained the audit

partner in year t for firm i (has a maximum length of five years

due to the mandatory audit partner rotation, that as of May 6, 2003 effective became);

BigNit = an indicator is set to 1 if the audit firm is an BigN-audit firm,

and set to 0 otherwise in year t for firm i;

LNTotalSalesit = natural logarithm of total sales in year t for firm i; and

LNIndustryspecializationit = natural logarithm of the industry specialization of the audit

firm in year t for firm i.

(33)

Based on chapter 2 and 3.2, the empirical model of this research is as follows:

Source: Author, 2015. Compiled from chapter 2 and 3.2.

Independent variable (I): Audit firm tenure

Control variables: LNIndustry specialization BigN Dependent variable: Discretionary Accruals (LNDAmodifiedJonesmodel) Control variable: LNTotal Sales Independent variable (II):

Audit partner tenure Independent variables (III):

Audit firm tenure Audit partner tenure

(34)

4 Results

Before performing the regression analysis the sample has been tested for their skewness and kurtosis and outliers. This sample contains more than 30 observations. Based on the Central Limit Theorem, it is therefore not required that the variables should be normally distributed (Mordkoff, 2011). To increase the quality of my analysis I decided to transform the variables that have a high skewness and/or kurtosis. This was applicable for the variables

‘Discretionary Accruals, ‘Total Sales’ and ‘Industryspecialization’.

Before transforming the variable ‘Discretionary accruals’ I have also tested the variable for its normally distribution. It was visible that the variable was not normally distributed. Therefore, I have tried three methods to get the variable more normally

distributed. The first method is using the untransformed variable and deleting the outliers5. With regard to the second and third method I deleted the extreme values identified by SPSS through the Casewise diagnostics. For the second method I used the untransformed variable and for the last method the transformed variable. Eventually, I chose the third method because it had the best results. With regard to the extreme values, I had eight extreme values in the regression of the association between audit firm tenure and audit quality (regression I), six extreme values in the regression of the association between audit partner tenure and audit quality (regression II) and seven extreme values in the regression of the effect of audit firm and audit partner tenure on audit quality (regression III). The extreme values have in this research been deleted before performing the regression analysis.

Table 4.1: Descriptive statistics – Regression I

Mean Median Std.

Deviation

Audit firm tenure 4,875 4,000 3,6786

BigN ,503 1,000 ,5004 LNTotalSales 12,3879 12,3783 2,68689 LNIndustryspecialization -15,3064 -14,4480 4,15113 LNDAmodifiedJonesmodel ,7936 1,2048 2,47738 Number of obs 82

(35)

34 | P a g e A s c h n a M a k h a n - 1 0 6 8 1 1 0 8 Table 4.2: Descriptive statistics – Regression II

Mean Median Std.

Deviation Audit partner tenure 2,165 2,000 1,6830

BigN ,503 1,000 ,5004

LNTotalSales 12,3879 12,3783 2,68689

LNIndustryspecialization -15,3064 -14,4480 4,15113 LNDAmodifiedJonesmodel ,8958 1,2167 2,39942

Number of obs 84

Table 4.3: Descriptive statistics – Regression III

Mean Median Std.

Deviation

Audit firm tenure 4,875 4,000 3,6786

Audit partner tenure 2,165 2,000 1,6830

BigN ,503 1,000 ,5004

LNTotalSales 12,3879 12,3783 2,68689

LNIndustryspecialization -15,3064 -14,4480 4,15113 LNDAmodifiedJonesmodel ,7732 1,1579 2,46945

Number of obs 83

In table 4.1, 4.2 and 4.3 I have included the descriptive statistics where the mean, median and standard deviation of the variables of the regression equations are presented. It is visible that the median of the discretionary accruals scaled by total assets in all of the regressions higher is than the mean of the discretionary accruals scaled by total assets. This indicates that in the majority of the sample used in this research the discretionary accruals higher are than the mean of the discretionary accruals scaled by total assets. This is inconsistent with the results of Chen, Lin, Lin (2008), Meyers, Meyers and Omer (2003), Chi, Huang, Liao and Xie (2009) and Manry, Mock and Turner (2008) and is caused by the skewness of the data. However, the skewness of the data is within an acceptable range and does not violate the normality

Should an audit firm rotation be mandatory? : a research of the influenceof audit firm tenure on audit quality and the difference in the effect of audit firm and audit partner tenure on audit quality

‘S

HOULD AN AUDIT FIRM ROTATION

BE MANDATORY

?’

Statement of Originality

Abstract

Table of Contents

1 Introduction

2 Literature review and hypotheses

Audit

quality

3 Research Methodology

ε

-

4 Results