Two Heads Are Better Than One: The Effect of Combined Experience of the Engagement and Review Partners on Audit Quality

Combined Experience of the Engagement and

Review Partners on Audit Quality

Two Heads Are Better Than One: The Effect of

Combined Experience of the Engagement and

Review Partners on Audit Quality

Name: Remko Bosman Student number: S2554291 Address: Tuinbouwdwarsstraat 23a

Postal code: 9717 HT Groningen E-mail: r.bosman.1@student.rug.nl

Supervisor: dr. C.A. Huijgen Second assessor: dr. N. Hussain

Date: 19-08-2019 Word count: 9359 University of Groningen Faculty of Economics and Business Accountancy & Controlling Department

Abstract

Due to several accounting scandals, new guidelines were introduced by the Sarbanes-Oxley Act in the US in 2005 to restore the public confidence. One of the measures was the implementation of a

mandatory rotation of audit partners. The EU later followed with similar guidelines. This research contributes to the existing literature about the effects of the rotation of engagement partner and review partner on audit quality, based on a sample of German listed firms. Three combinations of audit partner tenures are created to examine their effects on the client’s discretionary accruals (DACCs): a short tenure of both partners, a long tenure of both partners, and long-short tenure of both partners. Prior literature suggests that a short partner tenure leads to higher DACCs compared to a long partner tenure. These findings can be interpreted as evidence of lower audit quality. We find that when both partners have a short tenure, compared to partners who both have a long tenure, the audit quality is impaired. Moreover, in a situation where one of both partners has a long tenure, the client’s DACCs are lower compared to the partners having short tenures. Thus, the experience of at least one partner with a certain client is important for preserving an appropriate level of audit quality.

Table of Contents

1. Introduction 6

1.1. Background 6

1.2. Academic contribution 7

1.3. Research question 7

2. Prior literature and hypothesis development 8

2.1. Theoretical background 8

2.1.1. Audit 8

2.1.2. Agency theory 9

2.1.3. Audit quality 10

2.2. Measures of audit quality 10

2.2.1. Discretionary accruals 10

2.2.2. Audit fees 11

2.2.3. Going concern report 11

2.3. Auditor rotation and audit quality 12

2.4. Benefits and costs of mandatory auditor rotation 12

2.5. Engagement partner tenure 13

2.6. Review partner tenure 13

2.7. Hypothesis development 14

3. Research methodology 15

3.1. Sample 15

3.2. Variables 15

3.2.1. Dependent variable: Audit quality 15

3.3. Independent variables 17 3.3.1. Partner tenure 17 3.3.2. Control variables 17 3.4. Data analyses 18 4. Results 19 4.1. Descriptive statistics 19 4.2. Correlations 22 4.3. Test of hypothesis 23

4.4. Additional tests 27

5. Discussion and Conclusion 35

5.1. Findings 35

5.2. Theoretical implications 36

5.3. Practical implications 36

5.4. Limitations and future research 37

1.

Introduction

1.1. Background

Nowadays, audit regulations are strict worldwide. These regulations have surfaced after a number of enormous corporate scandals. It originated with the failure of Arthur Andersen in the Enron scandal of 2002. Due to these scandals, investors lost billions of dollars and, therefore, U.S. senators demanded better laws for reliable governance. Companies became bankrupt through fraudulent accounting practices and when CEOs executed self-dealing transactions (Romano, 2005). Stock prices decreased during this period by 40%. This downturn in the stock market affected thousands of shareholders and employees. With those circumstances in mind, the public confidence in auditors decreased. This situation can be largely explained by both personal and professional interests that are in conflict with each other. This resulted in a public debate about the importance of an external auditor. There is doubt about the professionalism and independence of auditors towards their clients. All Big Four accounting firms were investigated legally by stakeholders who were affected to gain greater clarity. The main subjects of this debate were the auditor’s independence and the quality of an audit. Audit failures have a negative impact on the audit quality; the higher the number of failures, the lower the audit quality (Francis, 2004). Monsouri, Pirayesh and Salehi (2009) believed the impairment of auditor

independence is a crucial factor of these failures. Furthermore, auditor independence is affected by the auditor tenure; this is the duration of the auditor-client relationship (Chen, Lin & Lin, 2008).

After the financial scandals of Enron and WorldCom, two senators in the U.S. introduced the Sarbanes-Oxley (SOX) Act (Romano, 2005). The SOX Act applies to all U.S. listed companies and other companies that issue shares in the American Stock Market (Romano, 2005). The Senate

emphasized the need for an audit firm rotation to preserve the objectivity of the auditors. Previously, the audit engagement partner rotated once every seven years (Chi & Huang, 2005). The

implementation of SOX is a required rotation of the engagement partner and audit review partner every five consecutive years; this is mandatory for public firms. The maximum tenure has been shortened with the intention to increase audit quality and audit independence. In addition to the rotation rules of an audit partner for one specific client, another rule is implemented to preclude the partner from returning to the audit client after rotation of that engagement (IESBA, 2006). Different scholars argued that audit firms are paid by the company, which has a negative effect on audit quality. After the implementation of SOX, different opinions appeared regarding mandatory audit firm

evidence uphold that shorter audit firm tenures lead to a higher degree of earnings management along with lower audit quality, which contradicts the arguments stated by SOX.

1.2. Academic contribution

Several researchers have conducted studies about the factors that might influence audit quality. This research contributes to previous research on the effects of audit firm tenure and audit partner tenure on audit quality. It specifically focuses on the different audit partners: the engagement partner and the review partner. There is less research available about a combination of both the review and

engagement partners. This study examines three combinations of different tenures of the engagement partner and review partner. Specifically, there are three possible combinations to investigate the combination with the strongest impact on audit quality. This paper contributes to existing research about audit firm tenure and audit partner tenure by exploring both. In prior research, the subjects were often the audit partner tenure or audit firm tenure. For audit firms, this research provides insight into making decisions about the composition of engagement and review partner teams, which would also be useful for partners who endeavor the best audit quality. The results of this research could add value for audit firms to restrict the earnings management practices of their clients.

1.3. Research question

My expectation is that both review partner tenure and engagement partner tenure influence audit quality. My main research question is:

What is the influence of the engagement partner and review partner tenure on the quality of an audit?

The research question is answered based on the empirical data of German listed firms. The structure of this research is as follows. First, the theoretical background is discussed. In this section, the theory is described with the aim of formulating potential expectations about the outcomes. This is followed by a discussion of the prior research findings as well as a possible suggestion for the outcomes of this study. Furthermore, the hypotheses are formulated, including the directions. In the research method section, the method of data collection that is used for this study and the regression models are specified. Next, the results of the regressions are presented. Finally, the conclusion explores the explanations of the results.

2.

Prior literature and hypothesis development

2.1. Theoretical background

Various theories explain the demand for audit services. Some of these theories are more well-known or popular than others, and some of them are based on perceptions. This research uses the agency theory as it relates to audit quality.

2.1.1. Audit

In the International Standards on Auditing (ISA), there is no specific definition of an audit that is used consistently. However, the main definition explained in ISA 200 states: “The objective of an audit of financial statements is to enable the auditor to express an opinion whether the financial statements are prepared, in all material respects, in accordance with an identified financial reporting framework.” An auditor strives to give additional assurance to the management of an organization about the financial statements and the business activities. Auditors perform essential activities to reduce the information risk, although the accounting information used for decision making may lack of credibility (Watts & Zimmerman, 1986). The adequacy of the risk control systems is assessed and elaborated in an

accompanying report. According to Matonti, Tucker and Tommasetti (2016), the auditor function can be described as follows: “The financial auditors have to certify the correctness of the bookkeeping entries and the financial reporting of management operations to verify that the accounts are maintained appropriately and that ultimately the annual reports give a true and fair view of the financial position, financial performance and cash flows of the company.” In recent years, more divergent tasks are involved in auditors’ proceedings. Matonti, Tucker and Tommasetti (2016) mentioned some specific tasks that are executed by auditors, namely the supervision of the financial disclosures process, the efficacy of the internal control systems, the internal auditing and risk

management and the auditing of the annual and consolidated accounts. A competent and independent external audit should perform an audit to verify if the financial auditor of the firm performs his tasks without mistakes or fraudulent actions (Arens, Elder & Mark, 2012). They also suggest that an audit is an amassment and evaluation of assurance whether or not the information meets the benchmarks that have been established in advance. For external auditors, it is essential to be independent because stakeholders presume that the auditors detect and correct the material mistakes in the financial statements.

2.1.2. Agency theory

The agency theory of Jensen and Meckling (1976) is used as a foundation for this research. According to the agency theory, an organization is a nexus of contracts between two or more parties, the

principal and the agent. The information asymmetries between the principal and agent must be reduced. Therefore, the agency theory dictates that principals would try to bridge the informational asymmetries by installing information systems to monitor the agent (Shapiro, 2005). The principal and agents both have their own interests. Because agents (management) have more inside information about their firm than principals (shareholders/stakeholders), information asymmetry arises. The focus of the agency theory is on problems that can arise when there is an agency relationship. The principal could use contracts for the agent to restrict his behavior (Jensen & Meckling, 1976). These incentives could be profit sharing or option bonuses. According to Jensen & Meckling (1976), another way to reduce the information asymmetry is to hire an external auditor to audit the financial statements of the company. The perception of an auditor is to sell his service to the client in a way that safeguards the credibility of the accounts (Hayes, Dassen, Schilder & Wallage, 2005). Profit-sharing or option bonuses could be lower because an audit is executed, thus leading to more monitoring of the

managers. Lee (1972) stated: “The most important requirement of the external auditor is to increase the credibility of financial statements generated from accounting information.” These theoretical aspects are relevant to investigating the effects of an audit partner tenure in an external audit firm that audits listed companies.

For this research, the agency theory is important because the main focus is audit quality. Audit firms and shareholders of the company prefer to have the highest audit quality to ensure reliable financial statements of the organization. Increased credibility guarantees more benefits based on more reliable reports for improving the quality of their investments (Hayes et al., 2005). Investors often make decisions to buy stocks based on fair and fraud-free information. Therefore, the information

asymmetry should be lower when the audit quality is higher. Francis (2004) argued that companies with greater information asymmetry and inherent uncertainty have an incentive to hire a more credible, high-quality auditor to disclose the firm’s intrinsic value.

2.1.3. Audit quality

DeAngelo (1981) defines audit quality as: “the market assessed the joint probability that a given auditor will both discover a breach in a client’s accounting system and report the breach.” Deis and Giroux (1992) added to this definition that an auditor should be able to resist the client pressure in different situations. The probability that an auditor detects errors or fraud in the financial statements depends on the capabilities and abilities of the auditor, and the probability of reporting these

misstatements depends on the independence of the auditor. According to Knechel, Vanstrealen and Zerni (2015), DeAngelo’s definition consists of two components. First, the likelihood of discovering misstatements and secondly, appropriate actions or measures upon this discovery of misstatements. Furthermore, audit quality has been split into perceived audit quality and actual audit quality (Taylor, 2005; Jackson, Moldrich, & Roebuck, 2008). To increase the actual audit quality, the auditor needs to reduce the level of risk of material errors in the financial statements (Siregar, 2012). Perceived audit quality shows the degree of confidence and the measured effectiveness of controls to mitigate the misstatements in the annual reports that are provided by the management. These definitions mostly refer to the auditor’s responsibility for the audit process and goals (Knechel et al., 2013). For example, NBA (2018) stated that there is a need for professional reasoning and fair presentation to achieve audit quality. Likewise, Palmrose (1988) defined the quality of an audit as: “the level of assurances - the probability financial statements contain no material omissions or misstatements” and argues that a higher level of assurances corresponds to a higher quality of audit services.

2.2. Measures of audit quality

In this research, a measure of audit quality is required to investigate the influence of audit partner tenure to audit quality. There is no clear instrument currently available to measure audit quality. Audit quality is measured through indirect indicators, also known as proxies. The following section

extrapolates three of these proxies: discretionary accruals (DACCs), audit fees and going concern reports.

2.2.1. Discretionary accruals

The first method to measure audit quality is detecting DACCs. These DACCs can be calculated through the modified Jones model (Dechow, Sloan & Sweeney, 1995). Krishnan (2003) demonstrated that DACCs are plausible to measure audit quality. The discretionary component of the total accruals is the part under a manager’s influence. It is, therefore, more complex to perform the audit at firms

with higher DACCs compared to firms with lower DACCs (Krishnan, 2003). Bartov, Gul and Tsui (2000) indicates a higher probability for a qualified audit report when the earnings management is measured by a model for DACCs where the absolute value of these accruals is relatively higher. There is evidence that the absolute and positive value of DACCs is reduced significantly when the audit partner tenure is longer, which leads to higher audit quality (Chen et al., 2008).

2.2.2. Audit fees

Audit quality can also be measured by the amount of audit fees charged by the auditor to the client. The auditor provides a necessary service to the client; therefore, a determined fee is chargeable to the client as compensation for the auditing activities. Datar, Feltham and Hughes (1991) indicated that higher risk clients often hire higher-quality auditors. Audit firms charge relatively higher fees to these higher-risk clients as the quality of the audit must be higher (Datar et al., 1991). To determine the amount of audit fees, the auditor is dependent on the costs incurred for audit services (Seetharaman, Gul & Lynn, 2002). Regarding these costs, big audit firms have a cost advantage when providing corresponding auditing services (Choi, Kim, Kim & Zang, 2010). Due to the market demand for large audit firms and high audit quality, these firms charge higher fees (Choi et al., 2010). According to Blokdijk, Drieenhuizen, Simunic and Stein (2006), a price premium is frequently associated with audits executed by large audit firms. The more audit effort is shown or possessed by an auditor, the higher the audit fees along with higher audit quality (Reichelt & Wang, 2010).

2.2.3. Going concern report

Audit quality can be measured by examining the probability that an auditor is issuing a going concern report (Knechel & Vanstraelen, 2007). A going concern opinion is the auditor’s estimate of the continuity of the company (Geiger & Rama, 2006). When the auditor has doubt about the continuity of the firm, a going concern opinion is provided (Chin & Chi, 2008). When the auditor provides a modified audit report, there is an indication that the auditor is not completely sure about some components of the financial statements. The auditor can add a going concern opinion to the audit report (Gray & Manson, 2011). According to Ruiz-Barbadillo, Gomez-Aguilar and Carrerra (2009), three factors affect the probability of a going-concern opinion, namely: company’s level of financial distress, the degree of audit competence and the level of auditor independence.

Francis (2004) mentioned two different errors regarding going concern opinions. For Type 1 errors, a going concern or adverse audit opinion is expressed while the financial statements provide a good picture of the reality, and the firm is healthy. An unqualified audit opinion would have been more appropriate. For Type 2 errors, an unqualified audit opinion is provided while the financial statements

do not present an appropriate picture of reality. Big audit firms are more accurate in judging the continuity, with the consequence that these firms can better prevent the Type 1 and Type 2 errors (Francis & Krishnan, 1999; Francis, 2004; Francis & Yu, 2009).

The quality of reporting decisions increases if there are fewer errors regarding going concerns opinions (Geiger & Rama, 2006). Therefore, the audit meets the standards and requirements, which leads to higher audit quality (Francis, 2004). The conclusion is that fewer errors regarding the going concern opinion result in a higher quality of the audit report and accordingly a higher audit quality. 2.3. Auditor rotation and audit quality

Kim, Min and Yi (2004) find that unexpected accruals are higher and thus, the audit quality is lower, in the first years after the auditor rotation. Other studies concluded that longer audit tenure is not associated with a lower audit quality (Meyers, Meyers & Omer, 2003; Chi & Huang, 2005; Chen et al, 2008). Greiger and Raghunandan (2002) document that significantly more audit reporting failures occur in the first years of an audit engagement with a certain client, rather than when this auditor has a served this client for a longer period. Fraudulent financial reporting is presented more in the first three years of an audit engagement between auditor and client. Consequently, the likelihood of higher audit quality is lower (Carcello & Nagy, 2004). Johnson, Khurana and Reynolds (2002) concluded, “lower audit quality (larger abnormal accruals) in the first three years following auditor changes relative to ongoing engagements of four or more years which is consistent with lower initial audit quality on new engagements.” Overall, the existing literature presents different findings, depending on the situation. Nevertheless, most research seems to indicate a trend of lower audit quality in the first years after an audit firm rotation or audit partner rotation. However, the intention of the implementation of this mandatory auditor rotation is to improve the audit quality and the financial statements of the companies (Carey & Simnett, 2006).

2.4. Benefits and costs of mandatory auditor rotation

Diverse arguments for and against mandatory auditor rotation arose after several studies. Carey and Simnett (2006) mentioned the benefits including the support of the independence of the audit partner, which enables a more objective view regarding the audit, resulting in recognition and problem

resolution. Another argument is that the rotation can aid the decreasing auditor independence and the declining quality of financial reporting due to the long auditor-client relationships (Gavious, 2007). With a longer audit tenure, client preferences could be accepted, which results in substandard audits and poor earnings quality (Myers et al, 2003). Auditor independence could increase audit quality. Acca (2011) contends, “Although it is hard to argue that a firm can be an external auditor to a

company for 30 years without becoming part of the ‘organogram’ of the company.” On the other hand, when the audit tenure is longer, auditors have more knowledge of their client’s financial

accounting, which is beneficial. Following this logic, a rotation may be a detriment. Solomon, Shields and Whittington (1999) and Johnson et al. (2002) argued that an auditor rotation includes an

additional cost for the audit firm. Moreover, the new audit partner could have less knowledge about the client’s risks, operation and financial reporting, which leads to lower audit quality.

2.5. Engagement partner tenure

The engagement partner is a member of the audit engagement team and is responsible for making auditing, accounting and reporting decisions that influence the financial statements. This partner also keeps in contact with the management team and the internal audit committee (George, 2004). The audit partner tenure can be defined by the length of years for which an audit partner is responsible for the audit services to a specific client (Chi & Huang, 2005).

Chi and Huang (2005) and Myers et al. (2003) suggested that a longer audit tenure is not related to a decrease in audit quality. Fanny and Siregar (2007) find that audit partner tenure is associated with lower DACCs, therefore, there is evidence that longer audit partner tenure is related to higher audit quality. Sirigar, Amurallah, Wibowo and Anggraita (2012) concluded that audit partner tenure has a negative effect on DACCs for the period before mandatory auditor rotation was implemented.

However, in the period after the implementation, there was a positive relationship between tenure and DACCs. Geiger and Raghunandan (2002) concluded that: “there were significantly more audit

reporting failures in the earlier years of the auditor-client relationship than when auditors had served these clients for longer tenures.”

2.6. Review partner tenure

The engagement quality review is important for the audit review process and must serve as an

evaluation of the performance of the engagement partner and the engagement team (Epps & Messier, 2007). The general purpose is to serve as quality control for audit engagements, as well as the audit partner (Epps & Messier, 2007). The review partner evaluates the performance of the engagement partner and the team members. The review partner has less contact with a certain client and this reduces familiarity threats (Schneider, Church & Ramsay, 2003). Review partners have less

accountability to the client but have the knowledge about the client’s industry, so the judgments and decisions are objective and high-level (Matsumura & Tucker, 1995). For these reasons, the review

partner performs a crucial role in the audit review and is essential to providing higher audit quality. The tenure of the review partner can be defined as the length of years in which the review partner is responsible for the review of the engagement partner. (Chi, 2005). In Germany, a review partner must sign the audit report together with the engagement partner. Both partners are responsible for the audit opinion. This responsibility is not applicable in many other countries where the review partner’s signature is not mandatory. Gold, Lindscheid, Pott and Watrin (2012) find evidence that review partner tenure is positively associated with audit quality. A possible reason for this finding is that the involvement of the review partner with the client is limited and review partner rotation results in a loss in expertise (Gold et al, 2012).

2.7. Hypothesis development

Several researchers have conducted studies about the effects of audit partner tenure on the audit quality. Prior research suggests that the audit quality is lower in the first two to three years of an audit engagement. For that reason, there is the expectation that a short tenure results in lower audit quality.

This leads to the following hypotheses:

H1: A short tenure of the engagement and review partner has a negative effect on audit quality, compared to a long tenure of both partners.

H2: A short tenure of both partners has a negative effect on audit quality, compared to a long (short) tenure of the engagement partner combined with a short (long) tenure of the review partner.

H3: A long (short) tenure of the engagement partner combined with a short (long) tenure of the review partner has a negative effect on audit quality, compared to a long tenure of both partners. H4: A short tenure of the engagement partner combined with a long tenure of the review partner does not have an effect on audit quality, compared to a long tenure of the engagement partner combined with a short tenure of the review partner.

3.

Research methodology

3.1. Sample

The database used for this study contained information about audit firms, and engagement and review partners. Furthermore, several accounting variables of clients were provided in order to estimate accruals and control variables. In order to analyze the effect of engagement partner tenure and review partner tenure on audit quality, an archival research method is used. The data for my research consists of 856 German listed firms with 5,931 observations. Data was collected for the years 1998-2010. These firms are relevant to this research because the German government already implemented the regulation that mandatory audit partner rotation is required after seven consecutive years since 1998 (Deutscher Bundestag, 1998).

3.2. Variables

3.2.1. Dependent variable: Audit quality

In prior research, audit quality has been measured in several ways, as discussed in Chapter 2. To measure audit quality, I use a modified version of the Jones model of Dechow (1995) to calculate DACCs. In this model, non-DACCs are subtracted from the total accruals to arrive at the discretionary part of the total accruals.

Several variables from the database are used to calculate the total accruals (TACCR). Healy (1985) and Jones (1991) use the balance sheet approach to calculate the total accruals. The computation of the total accruals is as follows:

TACCRt = ΔCAt – ΔCASHt – ΔCLt +ΔSTDEBTt – DEPt (1)

Where:

TACCRt = Total accruals in year t

ΔCAt = Current assets in year t less current assets in year t-1

ΔCASHt = Cash and cash equivalents in year t less cash and cash equivalents in year t-1 ΔCLt = Current liabilities in year t less current liabilities in year t-1

ΔSTDEBTt = Short term debt in year t less short term debt in year t-1 DEPt = Depreciation and amortization expense during year t

After the computation of the total accruals, a regression analysis is used to estimate the parametersα0, α1 andα2 for every single industry based on the SIC code of the clients. Barth, Beaver and Landsman (1998) developed an industry classification, which distinguishes fifteen different industries.

The modified Jones model regresses the total accruals on the change in revenues (corrected for the change in receivables) and the amount of property plant and equipment. In order to estimate the non-DACCs (Jones, 1991), the equation is shown below:

TACCRt/TAt-1 = α1 * 1/TAt-1 + α2 * (ΔREVt-ΔARt)/TAt-1 + α3 * PPEt/TAt-1 (2)

Where:

After the total accruals and the specific parameters are calculated, the calculation of the company specific non-DACCs is as follows:

NDACCRt/TAt-1 = α1 * 1/TAt-1 + α2 * (ΔREVt-ΔARt)/TAt-1 + α3 * PPEt/TAt-1 (3)

Where:

NDACCRt /TAt-1 = Non-DACCs in year t divided by total assets in year t-1 TACCRt = Total accruals in year t

∆REVt = Revenues in year t less revenues in year t-1

∆RECt = Receivables in year t less receivables in year t-1

PPEt = Gross property plant and equipment in year t

TAt-1 = Total assets in year t-1 α 1,α2 andα3 = Parameters to be estimated

To calculate the DACCs, I subtract the company-specific total accruals from the non-DACCs. DACCRt/TAt-1 = TACCRt/TAt-1 – NDACCRt/TAt-1 (4)

Where:

DACCRt = Discretionary accruals in year t

3.3. Independent variables 3.3.1. Partner tenure

The independent variables are engagement partner tenure and review partner tenure. The partner tenure can be defined as the length of years for which the partner is responsible for the audit services to a specific client (Chi & Huang, 2005). The partner tenures are divided into two categories,

therefore, dummy variables are created for long and short tenures. Tenures of two years or less are coded with 1 and tenures of more than two years are coded with 0.

3.3.2. Control variables

Earlier research shows different other factors that could influence audit quality. Control variables are included to account for these factors. These variables are firm size, leverage, Big Three, ROA and accounting standard.

The first control variable is client firm size, measured as the natural logarithm of total assets of a firm in a certain year. Reynolds and Francis (2001) argued that firm size is negatively associated with total accruals; therefore, this control variable is relevant in this research. Lee and Choi (2002) find that smaller firms use earnings management more often to avoid losses. The reason is that bigger firms are more vulnerable to reputation loss. Therefore, a reputation-damaging event leads to lower

involvement quality of the stakeholders (Rhee & Valdez, 2009).

The second control variable is the financial leverage of a client, measured as the total debt divided by the total assets (Johnson et al., 2002). According to Defond and Jiambalvo (1994), companies with more debt use their flexibility to engage in earnings management to avoid penalties of breaching debt covenants. To avoid these violations, managers of highly leveraged firms are inclined to create income-increasing accruals (Defond & Jiambalvo, 1994).

The next control variable is audit firm size, which can be a proxy for audit quality (DeAngelo, 1981). Big Four audit firms’ clients have lower accruals, which means higher audit quality because they are

able to restrict aggressive earnings management (Becker, DeFond, Jiambalvo & Subramanyam, 1998). High-reputation auditors are stimulated to restrict a high level of earnings management to defend their reputation and legal exposure (Francis & Wang, 2008). For smaller audit firms, the dependency of a large client is relatively higher, which could lead to preferential treatment through the auditors of these firms (Reynolds & Francis 2001). In Germany, The Big Three audit firms are EY, KPMG and PwC. A dummy variable is created which is equal to 1 if the audit firm is a Big Three audit firm and is equal to 0 otherwise.

Return on assets (ROA), which is measured as net income divided by total assets (Manry et al., 2008), is the next control variable. Kothari, Leone and Wasley (2005) investigated the influence of ROA on DACCs. They provided evidence of a positive relationship between performance measured by ROA and the size of accruals.

The accounting standard that a client uses is the last control variable. Those who adopt IFRS are associated with more DACCs and earnings management compared to German GAAP (Van Tendeloo & Vanstrealen, 2005). The sample consists of data between 1999 and 2009. After 2005, German firms were required to implement IFRS as a financial reporting standard. Until this point, the firms could choose the standard they preferred to use for financial reporting: German GAAP or IFRS. A dummy variable is set equal to 1 if the firm uses IFRS or US GAAP as a standard and equal to 0 if firms use German GAAP.

3.4. Data analyses

To account for non-normal distribution of the variables, extreme outliers of continuous variables are winsorized. Further, the multicollinearity among all variables is tested through the pair-wise

correlation matrix. Regression analysis is conducted to find significant or non-significant effects between the dependent variable and the independent variables and control variables. In the regression analysis, the absolute values of the DACCs are used, since it is not obvious which motivations firms may have for earnings management, whether upwards or downwards.

The following model is tested:

|DACC|= β0 + β1* LongShort / ShortLongt + β3 * ShortShortt + β5 * Firmsizet + β6 * Leveraget +β7 * ROAt + β8 * Standardt + β9* Big Threet + ε (4)

4.

Results

4.1. Descriptive statistics

The initial sample consists of 3,185 firm-year observations. I have selected the years 1999 to 2009 to calculate the engagement partner tenure and review partner tenure. To compute tenure, a switch of partners should first occur because partners’ names are not known before 1999. I eliminated firms with less than three year observations to prevent noise. As mentioned before, observations from the years 1998 and 2010 are excluded because information about partners is not available for those years. To ensure the partner rotation does not result from an audit firm rotation, observations with an audit firm rotation have been removed.

The industry classification of Barth et al. (1998) is adopted in this research to estimate the alphas in equation (2). Table 1 provides an overview of the different industries and the number of observations per industry. All firms’ observations from the database are used to estimate the alphas. Financial institutions, insurance, and real estate are not included because the modified Jones model is not appropriate to measure audit quality for financial companies. With 1,806 observations, industry 7 (durable manufacturers) has the most observations and with 159 observations, industry 5

(pharmaceuticals) has the least. The 159 observations are still enough to conduct a regression to calculate the alphas for this industry.

|DACC| = Absolute value discretionary accruals

LongShort / ShortLong = A long (short) tenure of the engagement partner combined with a short (long) tenure of the review partner

ShortShort = A short tenure of both partners Firm size = Log total of the assets of a firm

Leverage = Total debt / total assets

ROA = Net income / total assets

Big Three = Firm is audited by a Big Three audit firm

Table 1

Industry classification Barth et al. (1998).

Industry Observations Percentage

1. Mining and construction

241 4.06%

2. Food 291 4.91%

3. Textiles. Printing and Publishing 434 7.32% 4. Chemicals 200 3.37% 5. Pharmaceuticals 159 2.68% 7. Durable Manufacturers 1806 30.46% 8. Computers 977 16.48% 9. Transportation 287 4.84% 10. Utilities 208 3.51% 11. Retail 621 10.47% 14. Services 706 11.91%

An overview of the descriptive statistics of the variables is displayed in Table 2. This table shows the number of observations, mean, standard deviation, minimum and maximum for both the dependent and independent variables, as well as the control variables. To prevent outliers from affecting the results, the DACC is winsorized by 1% at both tails. The average absolute DACCs is 0.0513 with a standard deviation of 0.077. For the partner combinations, 20.5% is a LongLong combination, 25.9% a LongShort/ShortLong combination and 53.6% a ShortShort combination. The control variables include firm size, leverage, and ROA and are all winsorized at the same level. The first control variable is firm size, which has a mean of 5.241 with a standard deviation of 2.196. Leverage appears to have a mean of 0.214 and the standard deviation is 0.204. The mean of ROA is -0.006 and the standard deviation is 0.206. Furthermore, 52% of the firms are audited by a Big Three audit firm. 70.1% of the firms use IFRS or US GAAP.

Table 3 contains a two-sample t-test to compare the different DACCs means for the partner tenure combinations. The table shows that 1,706 observations have a short engagement partner tenure and review partner tenure, and 653 observations have a long engagement partner tenure combined with a long review partner tenure. The ShortShort tenure combinations have the highest DACC, with a mean of 0.059. The mean of LongShort is lowest, which is 0.049. The test provides a positive and

significant difference in DACC between ShortShort and LongLong (0.008. p<0.05) and ShortShort related and LongShort/ShortLong (0.014. p<0.01). The comparison of means indicates higher DACCs when the partner tenures are both short compared to partner tenures that are both long. This t-test does not include the influence of control variables. Therefore, more corroborate conclusions can be drawn after the multiple regression analysis.

Table 2

Descriptive statistics

Observations Mean Standard deviation Minimum Maximum

Dependent variable DACC 3.185 .0513 .077 .001 .708 Independent variables LongLong 3.185 .205 .404 0 1 LongShort/ShortLong 3.185 .259 .438 0 1 ShortShort 3.185 .536 .499 0 1 Control variables Firm size 3.185 5.241 2.196 -.543 12.477 Leverage 3.185 .214 .204 0 3.748 ROA 3.185 -.006 .206 -5.160 1.092 BIG Three 3.185 .520 .499 0 1 Standard 3.185 .701 .458 0 1

** = Coefficient is significant at level 0,05 (two-tailed). *** = Coefficient is significant at level 0,01 (two-tailed).

4.2. Correlations

Before conducting the regression analysis, the correlations between the variables have to be established. The Pearson correlation matrix in Table 4 presents the pairwise correlations among all variables. The dependent variable correlates significantly with LongLong and LongShort/ShortLong with negative values of -0.0354. and -0.0369. There is a significant negative correlation between LongLong and ShortLong/LongShort, with a value of -0.3005. The other significant correlations with the control variables are firm size and leverage in relation to the LongLong combination, leverage and ROA in relation to LongShort/ShortLong, leverage and Big Three in relation to ShortShort, Big Three and standard in relation to Leverage, and standard in relation to Big Three.

Table 3

Two-sample t-test

Tenure combinations LongLong LongShort/ShortLong ShortShort LongShort ShortLong

Observations 653 826 1,706 498 328 Mean DACC .051 .050 .059 .049 .050 Tenure combinations LongLong - LongShort/ShortLong .001 (.756) ShortShort .008** (.0109) .014*** (.001) LongShort -.001 (.726)

** = Correlation is significant at 0.05 level (two-tailed). ***= Correlation is significant at 0.01 level (two-tailed).

To ensure no serious multicollinearity between the independent variables of the model, the correlation coefficients must be between -0.7 and 0.7. When the correlations are higher or lower, there is

multicollinearity. In this model, all values are between -0.7 and 0.7, thus serious multicollinearity does not exist.

Multicollinearity can also be detected by investigating the VIF-values. When the VIF value is higher than 10, there is multicollinearity. The mean VIF-value of this model is 1.27 with a maximum of 1.7. Hence, the conclusion is that there is no serious multicollinearity in this research.

4.3. Test of hypothesis

To test the hypothesis, a multilinear regression analysis is conducted. The independent variables are all dummy variables, and to test hypothesis 1, LongLong (LL) is excluded in the first model since this is the default position. To test hypothesis 2, hypothesis 3 and hypothesis 4 respectively,

LongShort/ShortLong (LS/SL), LongLong (LL), and ShortLong (SL) are the default positions. The F-values are significant for every regression analysis. Adjusted R-square variates between 0.014 and 0.063. This means for hypothesis 1, for instance, that 5.3% of the changes of the dependent variable are explained through the independent variables.

Pearson correlation matrix

1. 2. 3. 4. 5. 6. 7. 8. 1. DACC 2. LongLong -.0354** 3 .LongShort/ShortLong -.0369** -.3005** 4 .ShortShort .0611 -.5454 -.6355 5. Firm size -.1923 .0376** .071 -.0928 6. Leverage -.0849 .01663** .0296** -.0128** .1311 7. ROA -.1031 .0534 .0436** -.0814 .2173 -.1339 8. Big Three -.0813 -.0182 .0553 -.0339** .3543 -.0359** .0868 9. Standard -.0003*** .0937 .1282 -.1885 .1294 -.0468** -.0256** .0904

The hypotheses below are tested by means of the regression analysis:

H1: A short tenure of the engagement and review partner has a negative effect on audit quality, compared to a long tenure of both partners.

H2: A short tenure of both partners has a negative effect on audit quality, compared to a long (short) tenure of the engagement partner combined with a short (long) tenure of the review partner.

H3: A long (short) tenure of the engagement partner combined with a short (long) tenure of the review partner has a negative effect on audit quality, compared to a long tenure of both partners. H4: A short tenure of the engagement partner combined with a long tenure of the review partner does not have an effect on audit quality, compared to a long tenure of the engagement partner combined with a short tenure of the review partner.

In Table 5, the regression results are shown. The first hypothesis is accepted because ShortShort partner combinations show significantly higher DACCs than LongLong partner combinations. This is in line with the predicted direction of hypothesis 1. The relationship between ShortShort and DACCs is positive and significant where LongLong partner combinations are the reference group. The coefficient of the ShortShort dummy variable is positive and significant 0,009 (p<0,05). The control variables, firm size, leverage and ROA, appear to be negatively related to DACCs and are all significant (-0.007, p<0.01), (-0.033, p<0.01) and (-0.076, p<0.01). The accounting standard is positive related to DACCs and significant (0.009, p<0.05).

The second hypothesis is also supported. ShortShort partner combinations show significantly higher DACCs than LongShort/ShortLong partner combinations. The dummy of ShortShort partner

combinations is positively related to DACCs. The relationship is significant (0.007, p<0.1) and it has the direction as expected. In this case, LongShort/ShortLong partner combinations are the reference group. The evidence is somewhat weaker for this hypothesis, compared to the first. The control variables, firm size (-0.006, p<0.01) and ROA (-0.022, p<0.01) again have a negative significant relationship to DACCs. In contrast to the first hypothesis, the effect of leverage is positive and significant (0.030, p<0.01) and the effect of accounting standard is positive but insignificant. The third hypothesis is rejected because LongShort/ShortLong partner combinations appear to have no higher DACCs compared to LongLong partner combinations, which is the reference group. There is a positive relationship between LongShort/ShortLong combinations and DACCs as expected before, but it is insignificant. The control variables have the same significant relations as in the first

hypothesis. The only difference is that the accounting standard is not significant but still has the expected direction.

The fourth hypothesis is accepted because there is evidence of an insignificant relationship between the LongShort partner combinations, compared to ShortLong partner combinations and DACCs. The coefficient is positive but not significant. As expected before, the LongShort partner combinations do not have significantly higher DACCs, compared to ShortLong partner combinations. The three control variables that have a negatively significant relationships with DACCs are firm size (-0.004, p<0.05) and accounting standard (-0.021, p<0.05). In contrast to the other hypotheses, the ROA is positive and significant (0.073, p<0.05) related to DACCs.

*= Coefficient is significant at level 0.1 (two-tailed). ** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

Multiple regression analysis

Controls H1 H2 H3 H4 Default LL LS/SL LL SL ShortShort .009** (.020) .007* (.087) LongShort/ShortLong .003 (.244) LongShort .001 (.844) Firm size -.006*** (.000) -.007*** (.000) -.006*** (.000) -.004*** (.000) -.004** (.019) Leverage -.0.039*** (.000) -.033*** (.000) 0.030*** (.001) -.020*** .002) -.017 (.465) ROA -.029*** (.012) -.026*** (.002) -.022** (.004) -.022*** (.000) .073** (.013) Standard .005 (.246) .009** (.013) .005 (.227) .002 (.461) -.021** (.035) BIG Three 0.006 (.189) -.007 (.065) -.005 (.241) .003 (.234) .006 (.497) Constant .098*** (.000) .091*** (.000) .095*** (.000) .066*** (.000) .094*** (.000)

Year dummies included Yes Yes Yes Yes Yes

R-Square .026 .055 .038 .066 .021

Adjusted R-square .024 .053 .035 .063 .014

F-value 11.94 23.01*** 16.46*** 17.51*** 3.04***

4.4. Additional tests

To test the robustness of the results, additional tests are conducted. First, I used another model to measure earnings management. A similar model to compute earnings management is the Defond and Park model (Defond & Park, 2001). The model is appropriate since Defond and Jiambalvo (1994) suggest that working capital accruals are managed more easily than long-term accruals.

The following equation is used to compute working capital accruals: AWCAt/At-1 = WCt – ((WCt-1 / REVt-1) x REVt ))/At-1 (5)

Where:

AWCAt = Abnormal working capital accruals in year t

WCt = Net working capital in year t ((current assets – cash and cash equivalents) – (current liabilities – short term debt))

WCt-1 = Net working capital in year t-1 REVt = Revenues in year t

REVt-1 = Revenues in year t-1 At-1 = Total assets in year t-1

Table 6 shows the results of the robustness test. Hypotheses 1 and 2 can be accepted again because the coefficients of ShortShort tenures are positive and significant, as expected. ShortShort tenure

combinations show significantly higher working capital accruals than LongLong or

LongShort/ShortLong partner tenure combinations. For hypothesis 3, insignificant difference between LongShort/ShortLong and LongShort, respectively, and working capital accruals are found. These results do not differ from Table 5. Hypothesis 4 can be accepted because there is an insignificant difference in working capital accruals between LongShort and ShortLong.

Table 6

Robustness test: Defond & Park model

Default H1 LL H2 LS/SL H3 LL H4 SL ShortShort .184*** (.000) .179*** (.000) LongShort/ShortLong .004 (.937) LongShort .016 (.802)

Control variables included Yes Yes Yes Yes

Year dummies included Yes Yes Yes Yes

R-Square .070 .0661 .0445 .0522

Adjusted R-square .0672 .0639 .0406 .0452

F-value 29.25*** 29.76*** 11.41*** 7.50***

Observations 2355 2528 1477 824

*= Coefficient is significant at level 0.1 (two-tailed). ** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

Two further robustness tests were conducted. For the first, one year was taken a short tenure, and for the second three years or less was taken as a short tenure. Table 7 depicts a positive and significant difference in DACCs between ShortShort partner combinations and LongLong partner combinations. The difference between ShortShort and LongLong is significant at the level of 0.012 (p < 0.01), which is higher than in the original regression model. The DACCs are significantly higher for clients with ShortShort partner combinations. ShortShort partner combinations versus LongShort/ShortLong partner combinations also had a significant difference of 0.011 (p < 0.05). For this comparison, the DACCs were significantly higher for clients with a ShortShort partner combination than for

combinations with one long tenured partner. LongShort/ShortLong partner combinations versus LongLong partner combinations shows an insignificant difference. This is also the case for LongShort partner combinations versus Short/Partner combinations. These results are not different from the original results in Table 5. These robustness tests indicate that the effects are stronger between the different partner combinations for hypotheses 1 and 2. When the short partner tenure is measured as a tenure of only one year, it can be concluded that the DACCs were higher than partner tenures of more than one year.

For the second, tenures up to three years were considered as short tenures. Table 8 provides the results of this robustness test. For hypothesis 1, the difference in DACCs is insignificant between ShortShort partner combinations and LongLong partner tenure. This result is different compared to the results for hypothesis 1 in Table 7. There is only a positive and significant difference in DACCs between

ShortShort partner combinations and LongShort/ShortLong partner combinations at the level of 0.009 (p < 0.05). The DACCs are significantly higher for clients with ShortShort partner combinations compared to clients with LongShort/ShortLong. When the short partner tenure is measured as a tenure up to three years, it can be concluded that the DACCs were not significantly higher than partner tenures of more than three years.

** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

Table 7

Robustness test: Short tenure 1 year

Default H1 LL H2 SS H3 LL H4 SL ShortShort .012*** (.000) .011** (.039) LongShort/ShortLong -.000 (.929) LongShort .002 (.811)

Control variables included Yes Yes Yes Yes

Year dummies included Yes Yes Yes Yes

R-squared .059 .029 .077 .029

Adjusted R-squared .056 .026 .073 .020

F-value 25.67*** 8.85*** 28.85*** 3.42***

** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

H1 H2 H3 H4 Default LL LS/SL LL SL ShortShort .008 (.206) .009** (.017) LongShort/ShortLong -.002 (.569) LongShort -.004 (.392)

Control variables included Yes Yes Yes Yes

Year dummies included Yes Yes Yes Yes

R-squared .041 .051 .040 .041

Adjusted R-squared .039 .049 .033 .031

F-value 18.34*** 25.92*** 5.86*** 4.14***

Finally, two other robustness tests were conducted to examine the difference in results between positive DACCs and negative DACCs. The first test only used the absolute value of DACCs as a proxy for audit quality. Table 9 indicates that there is a positive and significant difference in DACCs between ShortShort partner combinations and LongLong partner combinations 0.11 (p < 0.05). There is also a significant difference in DACCs between ShortShort and LongShort/ShortLong partner combinations. This difference has a higher significant level than the original model 0.017 (p < 0.05). Hypotheses 1 and 2 can both be accepted for positive DACCs. Further, this test indicated insignificant results for hypotheses 3 and 4.

Second, only negative DACCs were used as proxies to measure audit quality. Table 10 depicts a positive and significant difference in DACCs between ShortShort and LongLong partner

combinations 0.005 (p < 0.1). These results for hypothesis 1 are significant at a lower level than the case for positive DACCs. The differences in DACCs between LongShort/ShortLong and LongLong partner combinations are now significant for model 3. This indicates that the DACCs are more negative when one partner had a short tenure as compared to both partners having a long tenure. Insignificant differences in DACCs were found regarding hypotheses 2 and 4.

It can be concluded that when DACCs are positive, the effects for LongLong and ShortShort partner combinations are stronger than when DACCs are negative. When one partner has a long tenure in a situation with positive DACCs, the DACCs are lower than ShortShort partner combinations.

** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

Positive discretionary accruals

Default H1 LL H2 LS/SL H3 LL H4 SL ShortShort ,011** (,018) ,017** (,015) LongShort/ShortLong -,000 (,918) LongShort -,001 (,964)

Control variables included Yes Yes Yes Yes

Year dummies included Yes Yes Yes Yes

R-squared ,058 ,049 ,076 ,135

Adjusted R-squared ,054 ,044 ,069 ,124

F-value 12,77*** 11,73*** 10,91*** 12,11***

Negative discretionary accruals Default H1 H2 H3 H4 LL LS/SL LL SL ShortShort ,005* (,092) -,004 (,285) LongShort/ShortLong ,009** (,044) LongShort ,001 (,917)

Control variables included Yes Yes Yes Yes

Year dummies included Yes Yes Yes Yes

R-squared ,174 ,153 ,266 ,194

Adjusted R-squared ,170 ,148 ,259 ,181

F-value 38,89*** 34,18*** 40,63 13,99***

Observations 1114 1143 681 355

*= Coefficient is significant at level 0.1 (two-tailed). ** = Coefficient is significant at level 0.05 (two-tailed). *** = Coefficient is significant at level 0.01 (two-tailed). The p-values are shown in parentheses (--).

5.

Discussion and Conclusion

This chapter compares the results of this research to the expectations and hypotheses set out earlier. Furthermore, this section describes the conclusion regarding to the main research question. Finally, theoretical and practical implications are suggested and the limitations of this study and possibilities for future research are discussed.

5.1. Findings

Audit partner tenure is one of the subjects discussed in the financial and audit markets. Audit firms strive for the best audit quality so that they can gain assurance for the financial statements of the clients. Therefore, this study investigates the effect of both engagement partner tenure and review partner tenure on audit quality. These partners are primarily responsible for providing assurance. Likewise, the client’s stakeholders require a true and fair representation of the financial statements. The review partner contributes to the audit quality with quality control for audit engagements (Epps & Messier, 2007). Performance evaluation of the engagement partner and audit team is provided by the review partner (Gold et al., 2012).

Previous research has indicated that in the initial years, when an audit partner is involved with a new client, the quality of the audit is lower compared to an audit partner, which has a long relationship with a specific client. In the first years after an auditor rotation, the abnormal accruals are higher (Johnson et al., 2002). Francis (2004) argued that audit quality is lower when the audit firm serves as an audit to a new client. Based on a review of previous literature before conducting the analyses, the expectation was that shorter tenures affect audit quality negatively.

The findings of this research are consistent with the expectations of prior studies regarding audit partner tenure and audit firm tenure. Hypotheses 1, 2 and 4 are accepted because the expectations and directions can be confirmed after conducting regression analyses. Significant evidence is found for these hypotheses. The main finding of this research is that DACCs are higher when responsible auditor teams consist of a short tenure of the engagement partner and a short tenure of the review partner, compared to a long tenure of the engagement partner alongside a long tenure of the review partner. Discretionary accruals are used as a proxy for audit quality. Therefore, I can conclude that short partner tenure combinations result in lower audit quality, compared to long partner tenure combinations. The second important finding of this research is that a short tenure of both partners leads to higher DACCs compared to a combination of partner tenure when one of the partners has a long tenure. The reason is a higher level of experience when one of the partners has a long-term

relationship with the client. No evidence is found in support of higher DACCs if one of the partners has a short tenure compared to a long tenure of both partners. It can be concluded that a short tenure of one partner does not negatively affect the audit quality.

The regression analysis finds evidence for hypothesis 1 and the robustness tests confirm this finding. Therefore, there is enough evidence to accept hypotheses 1, 2 and 4. Hypothesis 3 cannot be accepted since the evidence for a significant effect is limited.

5.2. Theoretical implications

This study contributes to the existing literature about audit quality regarding auditor tenure and auditor rotation. Earlier research provides evidence for lower audit quality when the tenure of the engagement audit partner is relatively short (Manry et al., 2016;). Gold et al. (2012) studied how review partner tenure is positively related to audit quality and concluded that a short review partner tenure leads to lower audit quality. Prior research generally focused on the tenure of the engagement partner of the audit firm since in most countries, it is not mandatory to disclose the review partner’s signature in the audit report. In Germany, the signature of the review partner appears in the audit report as well. This research focused on both the engagement and review partners, which develops more insight into the effects of both engagement and review partner tenures on audit quality. To examine the effect of both tenures, different combinations of engagement partner and review partner tenures were developed.

5.3. Practical implications

In practice, the findings of this study can be relevant for different stakeholders, such as auditors and their clients. With the outcomes of this research, several practical recommendations can be

extrapolated. The experience of the audit partner with the client is not always sufficient after partner rotation (Gold et al, 2012). Therefore, the first recommendation is to solve this problem through a rotation of only the engagement partner or the review partner, instead of both. We find evidence that when both partners have a short tenure, the audit quality is lower. When one of the two partners has a long tenure, the experience with the client is higher, which results in higher audit quality.

The second practical recommendation is that the mandatory audit firm rotation must be reconsidered. In this situation, both partners perform audit services to a new client, which means that the

engagement partner and review partner lack experience. As mentioned before, this leads to a lower audit quality compared to a long tenure for one or both partners. Other researchers confirm this conclusion because a mandatory audit firm rotation is negatively related to audit quality through a

lower auditor-client relationship in the earlier year (Jackson, Moldrich & Roebuck, 2008). The audit quality can be improved through a mandatory rotation for only the engagement partner or review partner.

The last recommendation is related to the review partner in other countries. Nowadays, a signature of the review partner on the audit report is mandatory in Germany. Therefore, the stakeholders of the firms receive more information regarding to the audit and this leads to greater transparency. Other countries can also implement this regulation to create more information in the audit report. 5.4. Limitations and future research

This study contains several limitations. The first is that the focus in this research is on short-tenure audit partner combinations. It could be relevant to investigate the effects of long-term partner tenures on DACCs and audit quality. In this study, I only examined the effects of short tenured partners on DACCs, compared to long partner tenures. It may be relevant to investigate the effects of long tenured partners because the long tenure is the topic of debate nowadays. According to this future research, regulators can possibly change the regulations regarding mandatory auditor rotation to improve audit quality.

Secondly, the researcher faced two database limitations during this study. The database consists of data from 1998-2010. Consequently, the data might be less representative for current years. For further research, one might consider using a more recent dataset, ranging from 2010-2019 for example. Furthermore, the database consists of firms that are listed in Germany. Focusing on one country may be restricted because firms do not necessarily operate similarly in other countries. For future research, data from different countries can be used as an additional test.

Another limitation for this research is that only discretionary and working capital accruals are indirect proxies to measure audit quality. More measures are also available as relevant proxies, such as going concern opinion, audit fees and audit failures.

The last limitation is the focus on combinations of audit partner tenures. This research examines the effects of both mandatory and voluntary auditor rotation because information about mandatory or voluntary partner rotation was not available in the database. For future studies, it could be interesting to investigate the different effects between a mandatory rotation and a voluntary rotation of the engagement and review partner. The SOX requires a rotation of both partners after every seven years. When data is available about mandatory and voluntary auditor rotations, future studies might examine the effects of these types of rotations on audit quality. It would be interesting to investigate if

mandatory partner rotation has different effects on the audit quality compared to voluntary partner rotation. The regulations regarding audit partner rotation are intended to increase audit quality, therefore it could be interesting for regulators to conduct research about this issue.

6.

References

ACCA (the Association of Chartered Certified Accountants), 2011. Audit Under Fire: a Review of the Post-Financial Crisis Inquiries. Retrieved July 2017.

Arens, A.A., Elder, R.J. and Mark, B., 2012. Auditing and Assurance Services: an Integrated

Approach. Boston: Prentice Hall.

Barth, M.E., Beaver, W.H. & Landsman, W.R., 1998. Relative Valuation Roles of Equity Book Value and Net Income As a Function of Financial Health. Journal of Accounting and Economics, 25, pp.1-34

Bartov, E., Gul, F.A. and Tsui, J.S., 2000. Discretionary-Accruals Models and Audit Qualifications. Journal of Accounting and Economics, 30(3), pp.421-452.

Becker, C.L., DeFond, M.L., Jiambalvo, J. and Subramanyam, K.R., 1998. The Effect of Audit Quality on Earnings Management. Contemporary Accounting Research, 15(1), pp.1-24.

Blokdijk, H., Drieenhuizen, F., Simunic, D.A. and Stein, M.T., 2006. An Analysis of Cross-Sectional Differences in Big and Non-Big Public Accounting Firms' Audit Programs. Auditing: A Journal

of Practice & Theory, 25(1), pp.27-48.

Carcello, J. and Nagy A. 2004. Audit Firm Tenure and Fraudulent Financial Reporting. Auditing: A

Journal of Practice & Theory 23 (2), pp.55-69.

Carey, P. and Simnett, R., 2006. Audit Partner Tenure and Audit Quality. The Accounting

Review, 81(3), pp.653-676.

Chi, W. and Huang, H., 2005. Discretionary Accruals, Audit-Firm Tenure and Audit-Partner Tenure: Empirical Evidence from Taiwan. Journal of Contemporary Accounting & Economics, 1(1), pp.65-92.

Chin, C-L., and H-Y. Chi. 2008. Gender Differences in Audit Quality. Available at: http://papers.ssrn.com/sol3/ papers.cfm?abstract_id1149405.

Chen, C.Y., Lin, C.J. and Lin, Y.C., 2008. Audit Partner Tenure, Audit Firm Tenure, and

Discretionary Accruals: Does Long Auditor Tenure Impair Earnings Quality?. Contemporary

Accounting Research, 25(2), pp.415-445.

Choi, J.H., Kim, J.B. and Zang, Y., 2010. Do Abnormally High Audit Fees Impair Audit Quality?. Auditing: A Journal of Practice & Theory, 29(2), pp.115-140.

Datar, S.M., Feltham, G.A. and Hughes, J.S., 1991. The Role of Audits and Audit Quality in Valuing New Issues. Journal of Accounting and Economics, 14(1), pp.3-49.

DeAngelo, L.E., 1981. Auditor Size and Audit Quality. Journal of Accounting and Economics, 3(3), pp.183-199.

Dechow, P.M., Sloan, R.G. and Sweeney, A.P., 1995. Detecting Earnings Management. Accounting

Review, pp.193-225.

DeFond, M.L. and Jiambalvo, J., 1994. Debt Covenant Violation and Manipulation of Accruals. Journal of Accounting and Economics, 17(1-2), pp.145-176.

Deis Jr, D.R. and Giroux, G.A., 1992. Determinants of Audit Quality in the Public Sector. Accounting

Review, pp.462-479.

Deutscher Bundestag, 1998. Gesetz zur Kontrolle und Transparenz im Unternehmensbereich (KonTrag).

Epps, K.K. and Messier Jr, W.F., 2007. Engagement Quality Reviews: A Comparison of Audit Firm Practices. Auditing: A Journal of Practice & Theory, 26(2), pp.167-181.

Fanny, M., & Siregar, S. 2007. Pengaruh Pergantian dan Jangka Waktu Penugasan Auditor terhadap

Kualitas Laba: Studi pada Emiten Bursa Efek Jakarta. Presented at The 1st Accounting

Conference Universitas Indonesia, Depok, Indonesia.

Francis, J.R., 2004. What Do We Know About Audit Quality?. The British Accounting Review, 36(4), pp.345-368.

Francis, J.R. and Krishnan, J., 1999. Accounting Accruals and Auditor Reporting Conservatism. Contemporary Accounting Research, 16(1), pp.135-165.

Francis, J.R. and Yu, M.D., 2009. Big Four Office Size and Audit Quality. The Accounting

Review, 84(5), pp.1521-1552.

Francis, J.R. and Wang, D., 2008. The Joint Effect of Investor Protection and Big Four Audits on Earnings Quality Around The World. Contemporary Accounting Research, 25(1), pp.157-191. Gavious, I. (2007), Alternative Perspectives To Deal with Auditors’ Agency Problem, Critical

Perspectives on Accounting, 18, pp.451-467.

Geiger, M.A. and Raghunandan, K., 2002. Auditor Tenure and Audit Reporting Failures. Auditing: A

Geiger, M.A. and Rama, D.V., 2006. Audit Firm Size and Going-concern Reporting Accuracy. Accounting Horizons, 20(1), pp.1-17.

Gold, A., Lindscheid, F., Pott, C. and Watrin, C., 2012. The Effect of Engagement and Review Partner Tenure and Rotation on Audit Quality: Evidence from Germany. Working paper Gray, Iain & Mason, Stuart, 2011. The Audit Process: Principles, Practice and Case ,South Western,

Cengage Learning. 5th Ed.

Hayes, R., Dassen, R., Schilder, A., & Wallage, P., 2005. Principles of Auditing, an Introduction to International Standards on Auditing. Harlow: Prentice Hall-Financial Times.

Healy, P.M., 1985. The Effect of Bonus Schemes on Accounting Decisions. Journal of accounting

and economics, 7(1-3), pp.85-107.

IESBA. 2006. Agenda Paper 3-D February 2006 – New York, New York

ISA 200. Objective and General Principles Governing an Audit of Financial Statements. IFAC 2016 Ittonen, K., 2010. A Theoretical Examination of the Role of Auditing and the Relevance of Audit

Reports.

Jackson, A.B., Moldrich, M. and Roebuck, P., 2008. Mandatory Audit Firm Rotation and Audit Quality. Managerial Auditing Journal, 23(5), pp.420-437.

Jensen, M.C. and Meckling, W.H., 1976. Theory of the firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3(4), pp.305-360.

Johnson, V.E., Khurana, I.K. and Reynolds, J.K., 2002. Audit‐firm Tenure and the Quality of Financial Reports. Contemporary Accounting Research, 19(4), pp.637-660.

Jones, J.J., 1991. Earnings Management During Import Relief Investigations. Journal of Accounting

Research, 29(2), pp.193-228.

Kim, J.B., Min, C.K., & Yi, C.H., 2004. Auditor Designation, Auditor Independence, and Earnings Management: Evidence from Korea.

Knechel, W.R. and Vanstraelen, A., 2007. The Relationship Between Auditor Tenure and Audit Quality Implied by Going Concern Opinions. AUDITING: A journal of practice &