1
The Effect of Auditor Tenure on Audit
Quality: Evidence from Germany
Master thesis 2019 – 2020
Student: Renske Wijma Student number: S3823873 Study program: MSc Accountancy
Supervisor: Dr. C. A. Huijgen
Date: June 21, 2020
2
The Effect of Auditor Tenure on Audit
Quality: Evidence from Germany
Master thesis 2019 – 2020
ABSTRACT
This paper distinguishes the effect of audit firm tenure and audit partner tenure on audit quality in a German setting. The German audit profession is characterized by high reputation risk. Moreover, listed companies are required to report the names of individual engagement and review partners. The regulation of mandatory audit firm and partner rotation should contribute to a higher audit quality. However, the effect of audit firm and partner tenure is yet unclear. Because of lack of client-specific knowledge in the early years of the audit engagement, I expect short audit firm and audit partner tenure to be negatively associated with audit quality. I expect, based on the high reputation risk and the social identity theory, that long audit firm and partner tenure affect audit quality differently. Using discretionary accruals as a proxy for audit quality, results show that both short audit firm and partner tenure are positively (and significantly) associated with absolute discretionary accruals and thus impair audit quality. For long audit firm or partner tenure, I do not find significant evidence. This suggests that the tenure effect of both the audit firm and audit partner may be more pronounced in the earlier years of the audit engagement than in later years. Therefore, mandatory auditor rotation would not be necessary.
3
I INTRODUCTION
The purpose of the independent audit is to provide assurance for the quality of financial reports. The quality of financial reports is, in turn, essential to confident and informed markets and investors. Regulators are concerned that, as the length of the auditor-client relationship increases, auditors become less objective and too familiar with their client. This results in less independence. The trust of theusers of financial statements has begun to decline because of the well-known accounting scandals and decline in auditor independence (Junaidi et al. 2012). A mandatory audit firm rotation of every ten years and audit partner rotation of every seven years (Regulation (EU) No 537/2014) should avoid the threat of familiarity. This, in turn, should contribute to auditor independence and eventually improve audit quality. However, opponents of this regulation argue that a newly appointed auditor would make more mistakes and might fail because of a lack of understanding the client (Vanstraelen, 2000).
Many studies have analysed the relation between auditor tenure and audit quality. In general, results regarding short-term tenure show that the audit quality is lower in the initial years of the auditor-client relationship. However, with regard to long-term tenure, results are conflicting (Chen et al., 2008; Carey & Simnett, 2006; Jackson et al., 2008; Vanstraelen, 2000; Knechel & Vanstraelen, 2007).
Related studies in a German context are very rare. The German audit profession is characterized by low liability risk, which does not create strong incentives for auditors to resist client-induced bias in financial reporting (Quick et al., 2018). Reputation risk would then be a more important driver of audit quality. This is in line with the study of Weber et al. (2008). Bamber & Iyer (2007) argue based on social identity theory, that an employee of a service organization whose direct interaction with clients is a major part of his work, will start to identify himself with his clients. They point out that the incentives of the individual audit partner may differ from those of the audit firm. Therefore, they imply that unlike an audit partner, an audit firm may have stronger incentives to remain independent.
A study performed in the German setting is from Hohenfels (2016), who investigates how auditor tenure affects the investors’ perception of audit quality in Germany. Results of this study show that investors perceive lower audit quality during the early and later years of an auditor-client relationship. Another study performed in the German setting concerns the effect of engagement and review partner tenure/rotation on audit quality in Germany (Gold et al., 2012). Gold et al. (2012) find evidence of less income-decreasing accruals with an increase of review partner tenure, but not with engagement partner tenure. They further find that rotation of the review partner is associated with more income-decreasing accruals. However, they do
4
not find the same effect for engagement partner rotation. They underline the importance of distinguishing the roles of the two partners.
Yet, it is unclear whether audit firm tenure (hereafter firm tenure) has a different effect on audit quality than audit partner tenure (hereafter partner tenure). Therefore, I am interested in investigating the effect of both firm tenure and partner1 tenure on audit quality in the German setting and whether mandatory rotation would be recommended. Therefore, the research question that will guide this research is:
How does the length of the auditor-client relationship affect the audit quality of German listed companies?
The sample consists of German listed companies from 1999-2011. I use absolute discretionary accruals as a proxy for audit quality, following the modified Jones model (Dechow et al., 1995). I classify the tenures into short, medium and long, and include a set of control variables to perform the OLS regressions. Results show that both short firm and partner tenure are positively (and significantly) associated with absolute discretionary accruals and thus impair audit quality. For long firm tenure, I do not find a significant effect, which is consistent with my expectation. However, for long partner tenure, I do not find a significant effect as well. This is not consistent with my expectation. I argue that the social identity theory might not be generalizable and applicable to all individuals. Rather, it depends on personal characteristics whether an audit partner is prone to client identification.
I perform an additional test using the natural log of tenure, which shows that an additional year of partner tenure has a stronger positive effect on audit quality than an additional year of firm tenure. This suggests that audit quality is more dependent on partner tenure than on firm tenure. Other additional tests do not alter my main results.
My study contributes to the discussion whether and how auditor tenure affects audit quality. Overall, my findings do not support the view that required firm and partner rotation would be necessary to improve audit quality. If deterioration of audit quality is the motivation for the regulation of mandatory rotation, then my results do not support this argument.
The remainder of this thesis is organized as follows. Section II explains the key concepts concerning this study. Moreover, prior studies are reviewed to develop the hypotheses. Section III presents the research design, including the statistical model and variables measurement.
5
Section IV discusses the sample selection and results. Finally, Section V provides the conclusion of the research question and discusses limitations and area’s of future research.
II THEORETICAL BACKGROUND
AuditingAn audit of the annual accounts is compulsory for every entity of public interest. According to Knechel et al. (2013), an audit is a professional service delivered by experts in response to economic and regulatory demand. Management may have incentives to misrepresent the financial statements, mostly because financial reports are used to evaluate their performance. Therefore, the essence of auditing is verification (Antle, 1981). Agency theory is a commonly used theory to explain the difference in goals or risk aversion (i.e. moral hazard), or differences in available information (i.e. adverse selection) between two parties. The cornerstone of agency theory is the assumption that the interests of principals and agents do not align, and that there is information asymmetry between those two parties (Hill & Jones, 1992). The value of the audit services depends on the assumption that certified public accountants are independent from their clients (Shockley, 1981). The auditor is provided the role to ensure that the reported financial information represents a true and fair view to the users. Moreover, the task of the auditor is to ensure that this information complies with the applied reporting standards. Financial reports are meant to communicate financial information to the entities’ stakeholders. Relevance and reliability are the fundamental qualitative characteristics of useful financial information (IFRS, 2018). According to the Conceptual Framework of IASB, information is relevant if ‘’it is capable of making a difference to the decisions made by users’’. To do so, it should have a predictive or confirmatory value. Information is reliable, when ‘’it faithfully
represents the substance of what it is purposed to represent’’. This is the case when it represents
information that is complete, neutral and free from error information. An audit of financial reports can enhance the quality of the financial information to reduce the existing information asymmetries and potential conflicts of interest between parties (Johnson et al., 2002). However, the rise in accounting irregularities has reopened questions about audit quality, independence and auditor tenure (Gosh & Moon, 2005).
Audit quality
Audit quality is one of the most important issues the audit profession is facing. A higher audit quality reduces the uncertainty associated with financial reports prepared by managers (Vanstraelen, 2000). The definition of audit quality has been addressed and reviewed by several scholars over the past decades. The most well-known and common used definition is from
6
DeAngelo (1981), who defines audit quality as ‘’the market-assessed joint probability that a
given auditor will both discover a breach in the client’s accounting system and report the breach’’. Hence, audit quality depends on the auditors’ skills and knowledge, and its
independence.
Francis (2011) argues that audit quality is influenced by six factors, including (i) audit inputs, (ii) audit process, (iii) audit firms, (iv) audit industry and markets, (v) institutions, and (vi) economic consequences of audit outcomes. Knechel et al. (2013) note that the perception of audit quality can differ between stakeholders. Users, auditors, regulators, and society may have a different perception of what audit quality should constitute. The user of financial reports may perceive high quality as that it contains no material misstatements. The audit partner that conducts the audit may define high audit quality as completing all required tasks according to the firm’s audit process. The audit firm may consider high audit quality as one for which the effort can be defended against challenge in an inspection or court of law. Regulators may perceive high audit quality when it is in compliance with the prescribed standards. And, society may view high audit quality to the extent of mitigating economic problems for a company or the market (Knechel et al., 2013)2.
Auditor independence
Maintaining auditor independence is vital to the audit profession, since auditors are considered as the ‘’gatekeepers’’ of the public securities markets. Auditor independence is important because it affects audit quality. If auditors do not remain independent, they may be less likely to report irregularities, which impairs audit quality. Over the years, accounting irregularities raised to question auditor independence. ‘Fundamental to all of the questions about auditor independence, is an inherent scepticism about how close the relationship between the auditor and the management of the audit client can be without creating, in fact or in perception, a mutuality of interest that could impair the auditor's independence’ (Sutton, 1997).
In general, any situation which increases the probability that an auditor will not truthfully report the results of the audit investigation can be seen as a threat to independence (Simunic, 1980). Tepalalgul & Lin (2015) argue that four threats to auditor independence exist, namely: (i) client importance, (ii) non-audit services, (iii) auditor tenure, and (iv) client affiliation with audit firms. In this study, I focus on the threat of auditor tenure.
2 In this study, I perceive high audit quality as the financial report containing no material misstatements. Hence,
7 Auditor tenure and mandatory rotation
With regard to auditor tenure, the introduction of a mandatory rotation of audit firm and individual audit partner is expected to improve audit quality. However, there are opposing views that question if this requirement would actually help. Proponents of mandatory rotation argue that, as the length of the auditor-client relationship increases, it may foster closeness between management and the auditor. Consequently, the auditor is more likely to agree with managers in important reporting decisions (Ryan et al., 2001). Shockley (1981) discusses other factors why auditor tenure might affect audit quality, namely complacency, lack of innovation, less rigorous audit procedures and a learned trust in the client. Proponents of required rotation claim that a mandatory auditor change would increase audit quality because a new auditor has a ‘fresh look’. It would also decrease the possibility of the auditor being influenced by the client. Furthermore, proponents are concerned that the auditor relies more on its previous efforts and experience with the client over time of an engagement, reducing the extent of audit procedures (Brody & Moscove, 1998).
However, opponents of the mandatory rotation argue that a newly appointed auditor would make more mistakes and might fail because of a lack of understanding the client (Vanstraelen, 2000). Johnson et al. (2002) argue that lack of adequate client-specific knowledge during the early years of an engagement, may result in higher audit fees for initial engagements. Moreover, it may lead to a lower likelihood of detecting misstatements and material errors. Client-specific knowledge such as its activities, accounting issues, and internal control structure, is crucial for auditors to detect material errors and misstatements. This indicates that mandatory rotation could harm auditor competence (Kwon et al., 2014). Furthermore, because of a lack of client-specific knowledge in the beginning years, the auditor depends more on client’s management. The auditor has to rely on information provided by management, which decreases independence. Longer auditor tenure may increase auditor competence as the auditor can base decisions on client-specific knowledge that he develops over time (Knechel & Vanstraelen, 2007). Opponents are also concerned about higher audit costs that entails auditor rotation. A long auditor-client relationship enables the auditor to perform the audit more efficiently, resulting in lower costs for the client.
8 The audit profession in Germany
In 1998, Germany implemented a legislation on mandatory audit partner rotation, which became effective in 20023 (United Nations, 2007). Audit partners of public listed companies are required to rotate after a period of seven years (Weißenberger, 2002). In May 2006, the EU Parliament introduced Directive 2006/43/EC to implement the obligation of audit partner rotation in all Member States. In this directive, there were no requirements regarding audit firm rotation.
This changed in April 2014. The EU Parliament introduced the regulation that requires a limitation on the duration of an audit firm-client relationship (Regulation (EU) No 537/2014). Article 17:1 imposes a maximum length of the relationship of ten years for firms of public interest. In some exceptional cases, the maximum duration can be extended to 20 or even 24 years4. According to Directive 2014/56/EU, each member state had to decide how to implement the new EU regulation in national law by June 2016. The German legislator implemented the EU requirements in the German Commercial Code (Handelsgesetzbuch).
The German legislator has taken a two tier approach to mandatory audit firm rotation, meaning that the standard rotation period for all public interest enterprises is ten years. Banks and insurance companies cannot extend firm tenures beyond these initial ten years. All other enterprises have the option to extend for another ten years via a tender or another fourteen years for a joint audit. For audit partners, Germany still applies the required rotation of seven years (HGB, paragraph 318).
In Germany, there is a limited auditor liability. The statutory auditor, its employees and the legal representatives of an audit firm involved in the audit are obliged to carry out a thorough and impartial audit, and to maintain confidentiality. Anyone misconducting these obligations can be enforced to compensate the client with a maximum of one million euros. In case of a public limited company the obligation to compensate can rise up to four million euros (HGB, paragraph 323).
3 Engagement and review partners starting in 1995 or earlier where forced to rotate ultimately by 2002 (Gold et
al., 2012).
4 The duration of the audit engagement can be calculated as the amount of years that an auditor has covered the
client’s audit engagement. This starts from the point that the auditor has been appointed for the first time (article 17:8).
9
However, compared to the US and the UK5 for instance, the liability risk for auditors in Germany is low. Moreover, third-party liability is not specified in the German Company Act. This means that it is hard to establish to shareholders and debtholders in cases of negligent violation of auditors’ duties. It must be proved that the auditor violated a law with the intention to harm the third-party. Overall, lawsuits against auditors in Germany are rare (Piot, 2005).
The low liability risk in Germany does not create strong incentives for auditors to resist client-induced bias in financial reporting. Therefore, reputation threats are a more important driver of audit quality (Quick et al., 2018). This argument is in line with the study of Weber et al. (2008), who investigate a financial reporting failure involving a public company and KPMG-Germany. They find that, after detecting the audit failure, KPMG-Germany lost a significant part of their clients in the next year (15.7% versus a three-year average of 7.7%). The share prices of their remaining clients declined about three percent. Based on these results, the authors suggest that German investors value auditor reputation and thus provide strong incentives to deliver high audit quality. This implies that the audit profession in Germany is characterized by a high reputation risk.
Literature review and hypothesisdevelopment
Several studies have investigated the relation between auditor tenure and audit quality. Regarding the effects of short term auditor tenure, many studies find consistent results. For instance, Johnson et al. (2002) examine whether short firm tenure is associated with lower-quality financial reports. Their results suggest that short audit firm-client relationships are associated with lower financial reporting quality, compared to medium firm tenure.
Geiger & Raghunandan (2002) examine the association between the length of the audit-firm-client relationship and audit reporting failures. They investigate whether the audit firm has issued a going concern6 report shortly beforecompanies went bankrupt. The results of this study suggest that auditors are less likely to modify their opinions for the financial statements immediately preceding bankruptcy during the initial years of the audit engagement.
Furthermore, Carcello & Nagy (2004) investigate the relation between firm tenure and audit quality by examining the relation between tenure and fraudulent financial reporting. They
5 In the US, penalties can rise up to fifteen million USD. In the UK, there are four ways of auditor liability
restrictions: fixed monetary cap; cap based on market capitalization of client; cap based on multiple of audit; or proportionate liability. If proportionate liability is chosen, limitation on liability must met three criteria: consent of shareholders must be obtained, limit on liability must be fair and reasonable (decided by English court), and Government will have the power to regulate what to be agreed (Chung et al., 2010).
6 An auditor issues a going concern report/opinion when the client’s financial state is in a condition that there is
10
find a higher frequency of fraudulent financial reporting in the early years of the audit-firm-client relationship. These studies are in line with the argument that a newly appointed auditor would make more mistakes and might fail because of lack of client-specific knowledge which reflects a higher level of information asymmetry. Therefore, my first hypothesis is:
H1: Short audit firm tenure is associated with lower audit quality, compared to medium audit firm tenure.
Knechel & Vanstraelen (2007) study the impact of firm tenure on the likelihood that an auditor issues a going concern opinion, using a sample with bankrupt and financially distressed companies that survived in the end. Their results show no evidence that firm tenure negatively affects audit quality. Furthermore, they find that long firm tenure reduces the likelihood that the auditor issues a false going concern opinion7. This may indicate that the auditor can base its decisions on client-specific knowledge learned over time. Jackson et al. (2008) use, amongst others, the auditors’ propensity to issue a going concern opinion as proxy to test the relation between tenure and audit quality. They find that the quality of the audit increases with firm tenure.
Studies using discretionary accruals as measurement for audit quality find that the amount of discretionary accruals declines with longer firm tenure (Myers et al., 2003; Chen et al. 2008; Manry et al., 2008). Myers et al. (2003) also find some evidence that longer firm tenure is associated with less extreme income-increasing accruals. The authors suggest that as the relationship lengthens, auditors limit the ability of management to create reserves to manage future earnings. Chen et al. (2008) investigate the relation between firm tenure and audit quality in Taiwan, where audit firm/partner rotation is not mandatory. Using performance-adjusted discretionary accruals as proxy for audit quality, the results show a significantly decrease in discretionary accruals with firm tenure. Gold et al. (2012) argue that the information asymmetry in the early years of the audit engagement may decrease as the firm tenure lengthens. The auditor gains more knowledge of the client and its industry over time, which provides an advantage in detecting material misstatements in financial reports. This would lead to an increase of audit quality.
These studies, together with the arguments of the opponents of required audit firm rotation, suggest that longer firm tenure improves or does not affect audit quality. I suggest that, based on the high reputation risk the German audit profession faces, medium firm tenure has
7 Perceived by Knechel & Vanstraelen (2007) as issuing a going concern opinion to a company that does not file
11
the most beneficial effect on audit quality. Consequently, long firm tenure will have no further effect on audit quality. Therefore, my second hypothesis is:
H2:Long audit firm tenure does not affect audit quality, compared to medium audit firm tenure.
The argument that a newly appointed auditor would make more mistakes is also applicable to partner tenure. Chi et al. (2009) examine the effect of mandatory partner rotation on audit quality in Taiwan. They find that the audit quality of companies whereof the audit partner just has changed, is significantly lower than the audit quality of the same companies a year before under a previous audit partner. Therefore, the third hypothesis is:
H3: Short audit partner tenure is associated with lower audit quality, compared to medium audit partner tenure.
Even though aforementioned studies find a positive relation between long firm tenure and audit quality, it can be argued that long partner tenure affects audit quality differently. Carey & Simnett (2006) use three measures of audit quality to investigate whether there is a diminution in quality associated with long partner tenure in Australia. The measures they use are the auditor’s propensity to issue an going concern audit opinion; the extent to which key earnings targets meet the benchmark; and the amount of abnormal working capital accruals. For the first and second measure, the authors find evidence that is consistent with deterioration in audit quality associated with long partner tenure. But, they found no evidence of an association using abnormal working capital accruals as a proxy for audit quality.
Bamber & Iyer (2007) use social identity theory to investigate how external auditors are influenced by their clients. Social identity theory, originally formulated by social psychologists Tajfel & Turner (1985), explains that part of a person’s concept of ’self-image’ comes from the groups to which that individual belongs. The social identity of the individual results from a self-categorization process, through which individuals group themselves with others. According to this theory, an individual can classify his/herself into various social groups such as their family, nationality, organization, and many others. There are three processes that create this group mentality: (i) categorization: people categorize others in order to understand and identify them; (ii) identification: the individual adapts the identity of the group to which he belongs and acts similar to that group, and ; (iii) comparison: the individual compares his group (ingroup) against other groups (outgroup).
12
Bamber & Iyer (2007) argue that social identity theory predicts that an employee of a service organization, whose direct interaction with clients is a major part of his work, will start to identify himself with his clients. Their findings are consistent with this statement and show that client identification increases with the length of auditor tenure. They suggest that this increases external auditors’ desire to preserve the client relationship, which yields to more lenient accounting decisions. Stefaniak et al. (2012) examine, based on social identity theory, whether internal and external auditors’ perceived relationship with an organization affects their objectivity. They find that identification has a different influence on internal and external auditors’ evaluations. Internal auditors’ employer identification is related to less lenient evaluations, as external auditors’ client identification is associated with more lenient evaluations, even without clients’ management expressing their preferences.
Bamber & Iyer (2007) also point out that the incentive of the individual audit partner may differ from that of the audit firm. They find that as the auditor-client relationship continues, the likelihood of the individual auditor leaning towards the client’s preferences increases. Conversely, a continuing firm tenure is associated with a decrease of this likelihood. Taking these results together, the authors imply that unlike an audit partner, an audit firm has stronger reputation incentives to remain independent. Furthermore, taken the high reputation risk audit firms have to care about in the German audit profession, and a low litigation risk for audit partners, it can be expected that a long partner tenure affects audit quality differently than a long firm tenure.
These studies, together with the arguments of the proponents of mandatory audit partner rotation, suggest that long partner tenure negatively affects audit quality. Therefore, my fourth and last hypothesis is:
H4: Long audit partner tenure is associated with lower audit quality, compared to medium audit partner tenure.
III METHODOLOGY
Sample selection and data
This research is conducted based on archival data. The dataset consists of German non-financial listed companies with financial reporting data available in Worldscope for the period 1999 - 2011. Per client-year observation, the names of the engagement and review partner, and the audit firm and office are included in the database. They are collected from the annual reports taken from Thomson Research. The database initially consists of 6,304 client-years. Observations whereof financial data to determine discretionary accruals are not available, are
13
deleted. To control for industry fixed effects I use the classification developed by Barth et al. (1998), to estimate discretionary accruals. Industries with less than 20 firms in one group are deleted. The final sample consists of 5,891 client-year observations.
Measurement of audit quality
In order to measure audit quality, I use discretionary accruals as a proxy. Discretionary accruals are accruals that do not relate to normal operating activities. A higher level of discretionary accruals may indicate that management has been able to engage in earnings management and to use its power over the auditor to report the financial statements for their own personal benefit (Jackson et al., 2008). Earnings can be managed either upward or downward, depending on year-specific situations. When audit quality is poor, the financial statements are more likely to contain items that do not reflect the company’s operating results. This means that the quality of reported earnings reflects the quality of audit work (Chen et al., 2008). Furthermore
,
it is argued that a large amount of discretionary accruals is indirect evidence of lower earnings quality and thus also lower audit quality (Francis & Yu, 2009). Therefore, I perceive a higher amount of discretionary accruals as an indicator of lower audit quality. To estimate discretionary accruals, I use the modified Jones model of Dechow et al. (1995). They show that this model has the highest statistical power in detecting earnings management. First, the expected total accruals have to be estimated. They are computed as follows:(𝐸)𝑇𝐴𝑡= 𝛽1(1 / 𝐴𝑠𝑠𝑒𝑡𝑡−1) + 𝛽2(∆ 𝑆𝑎𝑙𝑒𝑠𝑡− ∆ 𝐴𝑅𝑡) + 𝛽3𝑃𝑃𝐸𝑡+ 𝜀𝑡 (1)
Where (𝐸)𝑇𝐴𝑡 is expected total accruals, 𝐴𝑠𝑠𝑒𝑡𝑡−1 are total assets in the preceding year, ∆ 𝑆𝑎𝑙𝑒𝑠𝑡 is change in net sales compared to the preceding year, ∆ 𝐴𝑅𝑡 is change in net accounts receivable compared to the preceding year, and 𝑃𝑃𝐸𝑡 is net property, plant and equipment of the present year. Following Defond & Jiambalvo (1994) and Subramanyam (1996), the coefficients 𝛽1, 𝛽2 and 𝛽3 are parameters from the estimation of expected accruals using all companies in the industry (Barth et al. 1998). Discretionary accruals are then estimated as:
𝐷𝐴𝑡 = 𝑇𝐴𝑡− [𝛽̂1 (1 / 𝐴𝑠𝑠𝑒𝑡𝑠𝑡−1) + 𝛽̂2 (∆ 𝑆𝑎𝑙𝑒𝑠𝑡− ∆ 𝐴𝑅𝑡) + 𝛽̂3𝑃𝑃𝐸𝑡] (2) where all variables are defined as earlier. The residual from the regression of equation 1, 𝜀𝑡, shows the level of discretionary accruals. The variables are winsorized at 1st and 99th percentiles of their distributions, to filter the data from the most extreme percentiles (Francis, 2005; Chi et al., 2009).
14
Measurement firm tenure and partner tenure
Firm tenure equals the number of years of the audit-firm-client relationship. Because it is unknown how many years this relationship exists in the initial year the client occurs in the dataset, current firm tenures until the first rotation are not included. First, I determine the year of rotation of the firm. Then, I measure the total length per audit firm-client relationship from the initial year of the engagement. Then the tenure is categorized into short, medium or long using a dummy variable. If the firm tenure falls into the category the dummy equals 1; otherwise 0. Following Carey & Simnett (2006), I perceive firm tenure as short when the firm has audited the same client for one or two years. A firm tenure of six years or more is perceived as long. Tenures of six years or more, where no initial rotation has taken place, are also categorized as long, since it is certain that these tenures can be categorized as long. As shown in table 1, there are 745 audit firm rotations in the sample of 5,891 client-year observations. 95 additional cases could be appointed as long firm tenure with certainty. In total, there are 840 categorized firm tenures. 31,7% of the tenures is short, 24,6% is medium, and 43,7% is long.
Like firm tenure, current partner tenures in the first client-year in the dataset until the first rotation are not included. German companies of public interest are required to disclose the identity of both the engagement and review partner in the audit report (HGB, article 322). This enables to determine how many consecutive years an individual audit partner has performed the audit for the same client as either engagement or review partner. Partner tenure is measured as the amount of years the key audit partner (engagement partner) has performed the audit for the same client. When a partner initially started as review partner, the partner already has some client-specific knowledge when switching to the role of engagement partner for the same client. Therefore, if the engagement partner has been involved into the audit at the same client as review partner in prior years, these years are included when measuring engagement partner tenure.
First, I determine the moment of rotation of the engagement partner. Then, I measure the total length per audit partner-client relationship from the initial year of the engagement. Similar to firm tenure, partner tenure is perceived short when the partner has audited the same client for two years. Tenure is perceived as long when the audit engagement is six years or more. Tenure is categorized into short, medium or long using a dummy variable, just as for firm
15
tenure. As shown in table 1, there are 1,527 audit partner rotations in the total sample. 56,4% of the partner tenures is short, 31,4% is medium, and 12,2% is long8.
TABLE 1
Summary firm and partner tenures
Audit firm Audit partner
Rotations* Non-rotations 745 5,146 1,527 4,364 Total 5,891 5,891 Short-tenure*** Medium-tenure Long-tenure *** 266 208 366 861 479 187 Total 840 1,527
* Dummy variable when rotated (=1) or non-rotated (=0).
** Tenure of 1 or 2 years is categorized as short, tenure of 3-5 years is categorized as medium and tenure of 6 years or more is categorized as long.
*** Cases where no rotation has taken place, but the firm-client relationship exists for at least six years are also categorized as long tenure.
Control variables
Although the primary variables of interest in this study are firm tenure and partner tenure, other variables may affect audit quality as well. Therefore, I control for several variables. The first one is BIG3. Prior studies document an association between measures of higher quality auditors and higher quality of financial reporting (Gul et al., 2009). According to Tepalalgul & Lin (2015), Big-4 audit firms provide a higher quality of audit services than non-Big-4 firms. Therefore, it could be expected that clients audited by a Big-4 firm will have a lower level of discretionary accruals. However, in my sample period, there are only three big audit firms. Therefore, I include a dummy variable if the client is audited by a Big-3 or not. The second control variable is SIZE. Larger clients will have more assets to sell in the event of financial distress (Jackson et al., 2008). These clients would be less likely to manage earnings, resulting in a lower level of discretionary accruals. The third one is LEVERAGE. High levels of leverage
8 The amount of long partner tenures is relatively small, compared to the amount of short and medium partner tenures. The reason for this is the mandatory audit partner rotation every seven years since 2002.
16
indicate a higher level of risk, which could lead to a higher level of discretionary accruals (Carey & Simnett, 2006; Jackson et al. 2008). The fourth one is GROWTH. Prior studies, for instance Chen et al. (2008), have found a positive association between growth and discretionary accruals. The fifth one is CFO. Poor operating cash flows are commonly associated with a higher likelihood of financial distress and therefore, a higher level of discretionary accruals (Carey & Simnett., 2006). The sixth one is ROA. According to Kothari et al. (2005), companies’ prior performance may influence the levels of accruals. The last one is STANDARD. The applied reporting standards may affect how accruals are processed. Table 2 summarizes the description and measurement of the control variables.
TABLE 2
Description and measurement control variables
Variable Description Measured as
BIG-3 Audited by Big-3 firm Dummy variable when audited by Big-3
=1, otherwise =0
SIZE Size of the client Natural logarithm of total assets
LEVERAGE Degree of debt financing Ratio of total liabilities to total assets GROWTH Growth of total assets Ratio of growth in total assets to total assets
from previous year
CFO Cash flows from operations
scaled by total assets
Ratio of cash flows from operations to total assets
ROA Return on total assets Ratio of net income to average total assets
STANDARD Reporting standard Dummy variable if client uses German
GAAP =1, otherwise (IFRS/US GAAP) =0
Research model
I first examine the relation between auditor firm tenure and audit quality (hypothesis 1 and 2), controlling for the variables mentioned earlier. I use the absolute value of discretionary accruals. The extent of the absolute value measures a company’s success in managing earnings either upward or downward (Reynolds & Francis, 2000). Using OLS regression, I estimate the following equation:
17
𝐷𝐴 = 𝛽0+ 𝛽1𝑆𝐻𝑂𝑅𝑇𝐹𝑇+ 𝛽2𝐿𝑂𝑁𝐺𝐹𝑇+ 𝛽3𝐵𝐼𝐺3 + 𝛽4𝑆𝐼𝑍𝐸 + 𝛽5𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸 + 𝛽6𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽7𝐶𝐹𝑂 + 𝛽8𝑅𝑂𝐴 + 𝛽9 𝑆𝑇𝐴𝑁𝐷𝐴𝑅𝐷 + 𝛽10 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽11 𝑌𝑒𝑎𝑟 + 𝜀 (3) Where 𝑆𝐻𝑂𝑅𝑇𝐹𝑇 is a dummy variable for firm tenures of 1 or 2 years and 𝐿𝑂𝑁𝐺𝐹𝑇 is a dummy variable for firm tenures of six years or more. To control for industry fixed effects, I use two-digit industry codes based on the classification of Barth et al. (1998) and include them as dummy variables. I further include dummy variables for year fixed effects.
To test hypothesis 1, the coefficient of interest is 𝛽1. This coefficient captures the amount of absolute discretionary accruals for a short firm tenure relative to a medium firm tenure. Coefficient 𝛽2 is of interest to test hypothesis 2. This coefficient captures the amount of absolute discretionary accruals for a long firm tenure relative to a medium firm tenure.
Then I examine the relation between partner tenure and audit quality (hypothesis 3 and 4), controlling for earlier mentioned variables. Using OLS regression, I estimate the following equation:
𝐷𝐴 = 𝛽0+ 𝛽1𝑆𝐻𝑂𝑅𝑇𝑃𝑇+ 𝛽2𝐿𝑂𝑁𝐺𝑃𝑇+ 𝛽3𝑅𝑂𝑇𝐴𝑇𝐸𝐷 + 𝛽4𝐵𝐼𝐺3 + 𝛽5𝑆𝐼𝑍𝐸 + 𝛽6𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸 + 𝛽7𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽8𝐶𝐹𝑂 + 𝛽9𝑅𝑂𝐴 + 𝛽10𝑆𝑇𝐴𝑁𝐷𝐴𝑅𝐷 +
𝛽11 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽12 𝑌𝑒𝑎𝑟 + 𝜀 (4) where 𝑆𝐻𝑂𝑅𝑇𝑃𝑇 is a dummy variable for partner tenures of 1 or 2 years and 𝐿𝑂𝑁𝐺𝑃𝑇 is a dummy variable for partner tenures of six years or more. 𝑅𝑂𝑇𝐴𝑇𝐸𝐷 is a dummy variable if both firm and partner rotate in the same, which equals 1 and otherwise 0. This variable is included because audit firm rotations may have a larger effect than audit partner rotations.
To test hypothesis 3, the coefficient of interest is 𝛽1, similar to firm tenure. Coefficient 𝛽2 is of interest to test hypothesis 4.
IV RESULTS
Descriptive statistics
Table 3 presents the descriptive statistics of the variables. The mean and median of DA are 0.00 percent of the beginning total assets, and the mean and median DA equals 9.00 and 6.00 percent of the beginning total assets. The mean (median) firm tenure is 5.19 (4.00) years. The longest firm tenure is 13 years, which is equal to the sample period. The mean and median partner tenure are 2.80 and 2.00 years, respectively, and the longest partner tenure is 10 years. Almost half of the sample observations is audited by a Big 3 audit firm (49%). The range of the
18
variables LEVERAGE and GROWTH is rather large, compared to the range of the variables TA,
CFO, SIZE and ROA. About 28% of the client-year observations applies German GAAP as
reporting standard9. The reason for this low percentage is the mandatory application of IFRS since 2005.
TABLE 3
Descriptive statistics
Variable Mean SD Minimum Q1 Median Q3 Maximum
DA DA TA FIRM TENURE PARTNER TENURE BIG3 CFO LEVERAGE GROWTH SIZE ROA STANDARD ROTATED 0.00 0.09 -0.06 5.19 2.80 0.49 0.06 0.60 0.11 5.05 -0.01 0.28 0.09 0.13 0.09 0.13 3.60 1.88 0.50 0.24 1.08 1.88 2.16 0.19 0.45 0.29 -0.55 0.00 -0.53 1.00 1.00 0.00 -2.19 -0.87 -0.99 -1.71 -3.40 0.00 0.00 -0.05 0.02 -0.11 2.00 1.00 0.00 0.00 0.42 -0.07 3.60 -0.02 0.00 0.00 0.00 0.06 -0.05 4.00 2.00 0.00 0.08 0.59 0.02 4.82 0.03 0.00 0.00 0.06 0.11 0.00 8.00 4.00 1.00 0.14 0.73 0.12 6.21 0.06 1.00 0.00 0.59 0.59 0.40 13.00 10.00 1.00 6.82 79.94 121.32 12.48 1.27 1.00 1.00
Table 4 shows the correlation between the variables. The univariate correlation between the coefficient ofDAand the coefficients of FT SHORT, FT LONG, PT SHORT and PT LONG are 0.0607, -0.0360, 0.0416 and -0.0345, respectively. All four coefficients have a significance level of p<0.01. The highest correlation is between de coefficients of ROA and CFO (0.5255), which has a significance level of p<0.01. Regarding multicollinearity the strength of this correlation is acceptable.
19
TABLE 4
Pearson correlation matrix
1 2 3 4 5 6 7 8 9 10 11 12 DA (1) X FT SHORT (2) 0.0607*** X FT LONG (3) -0.0360*** -0.0560*** X PT SHORT (4) 0.0416*** 0.4307*** 0.1921*** X PT LONG (5) -0.0345*** -0.0207 0.2262*** -0.0749*** X BIG 3 (6) -0.0643*** -0.0381*** 0.0272** 0.0009 0.0226* X CFO (7) -0.0240* -0.0055 0.0009 0.0126 0.0120 0.0519*** X LEVERAGE (8) 0.0272** 0.0060 -0.0080 0.0017 -0.0015 0.0005 -0.0656*** X GROWTH (9) 0.1316*** 0.0315** -0.0119 0.0501*** -0.0040 -0.0027 0.3688*** -0.0114 X SIZE (10) -0.2222*** -0.0782*** 0.0518*** -0.0339*** 0.0636*** 0.3459*** 0.1268*** -0.0121 0.0220* X ROA (11) -0.1411*** -0.0411*** 0.0346*** -0.0170 0.0415*** 0.0556*** 0.5255*** -0.1770*** 0.0706*** 0.2296*** X STANDARD (12) -0.0040 -0.0360*** -0.1389*** -0.0510*** -0.0883*** -0.0399*** 0.0195 0.0325** -0.0224* -0.0694*** 0.0315** X ROTATED (13) 0.0618*** 0.3219*** -0.0818*** 0.1650*** -0.0542*** -0.0313** 0.0250* 0.0087 0.0481*** -0.0797*** -0.0357*** -0.0087
20
Main analysis
Relation between firm tenure and absolute discretionary accruals
Table 5 presents the regression results for the full sample10, testing the relation between short and long firm tenure and absolute discretionary accruals. I use equation 3 to test my first and second hypothesis. Regarding my first hypothesis, the variable of interest is SHORT. I expect this coefficient to be positive, which means that firm tenure in the early years of the audit engagement has an increasing effect on discretionary accruals and thus decreases audit quality. Results show that the coefficient of SHORT equals 0.0211 with a significance level of p<0.01. This indicates that short firm tenure is associated with a higher level of discretionary accruals, as predicted by Hypothesis 1. Regarding my second hypothesis, the variable of interest is
LONG. I expect this coefficient to be insignificant, meaning that long firm tenures of six years
or more will have no further effect on discretionary accruals and thus do not affect audit quality. The coefficient of LONG is insignificant, which is consistent with Hypothesis 2. The control variables GROWTH, SIZE, ROA and STANDARD have the expected association with discretionary accruals and are significant. BIG 3 and LEVERAGE are not significant.
Relation between partner tenure and discretionary accruals
Table 6 presents the results for the full sample, testing the relation between short and long partner tenure and absolute discretionary accruals. I use equation 4 to test my third and fourth hypothesis. I control for industry and year fixed effects, similar to equation 310. Regarding my third hypothesis, the variable of interest is SHORT. I expect, like short firm tenure, its coefficient to be positive. The results show that the coefficient of SHORT equals 0.0928, with a significance level of p<0.01. This indicates that short partner tenures are also associated with a higher level of discretionary accruals, which is in line with Hypothesis 3. Regarding my fourth hypothesis, the variable of interest is LONG. I expect this coefficient to be positive, meaning that long partner tenure will also have an increasing effect on discretionary accruals. However, the coefficient 𝛽2 is negative and insignificant. Therefore, Hypothesis 4 is not supported. The control variables GROWTH, SIZE, ROA and STANDARD are significant and have the expected sign. The control variable ROTATED has also the expected sign and is significant. This stresses the importance of including this control variable. Otherwise, the positive effect cannot be assigned to one of the variables of interest.
21
TABLE 5
Results from regression of absolute discretionary accruals on short and long firm tenure dummies 𝐷𝐴 = 𝛽0+ 𝛽1𝑆𝐻𝑂𝑅𝑇𝐹𝑇+ 𝛽2𝐿𝑂𝑁𝐺𝐹𝑇+ 𝛽3𝐵𝐼𝐺3 + 𝛽4𝑆𝐼𝑍𝐸 + 𝛽5𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸 + 𝛽6𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽7𝐶𝐹𝑂 + 𝛽8𝑅𝑂𝐴 + 𝛽9 𝑆𝑇𝐴𝑁𝐷𝐴𝑅𝐷 + 𝛽10 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽11 𝑌𝑒𝑎𝑟 + 𝜀 Variables Expected Sign Coefficient Constant 0.146*** (0.00651) SHORT + 0.0211*** (0.00585) LONG ? 0.00148 (0.00534) BIG 3 - 0.00214 (0.00250) CFO - 0.000884 (0.00615) LEVERAGE + 0.00121 (0.00109) GROWTH + 0.00674*** (0.000676) SIZE - -0.00821*** (0.000647) ROA - -0.0462*** (0.00784) STANDARD -/+ -0.0115*** (0.00367) Observations 5,891 R-squared 0.071 Industry FE Included Year FE Included
22
TABLE 6
Results from regression of absolute discretionary accruals on short and long partner tenure dummies 𝐷𝐴 = 𝛽0+ 𝛽1𝑆𝐻𝑂𝑅𝑇𝑃𝑇+ 𝛽2𝐿𝑂𝑁𝐺𝑃𝑇+ 𝛽3𝑅𝑂𝑇𝐴𝑇𝐸𝐷 + 𝛽4𝐵𝐼𝐺3 + 𝛽5𝑆𝐼𝑍𝐸 + 𝛽6𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸 + 𝛽7𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽8𝐶𝐹𝑂 + 𝛽9𝑅𝑂𝐴 + 𝛽10𝑆𝑇𝐴𝑁𝐷𝐴𝑅𝐷 + 𝛽11 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽12 𝑌𝑒𝑎𝑟 + 𝜀 Variables Expected sign Coefficient Constant 0.145*** (0.00652) SHORT + 0.00928*** (0.00346) LONG + -0.00211 (0.00674) ROTATED + 0.00971** (0.00410) BIG 3 - 0.00200 (0.00250) CFO - 0.000525 (0.00616) LEVERAGE + 0.00118 (0.00109) GROWTH + 0.00667*** (0.000677) SIZE - -0.00818*** (0.000647) ROA - -0.0461*** (0.00784) STANDARD +/- -0.0116*** (0.00367) Observations 5,891 R-squared 0.071 Industry FE Included Year FE Included
Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
Additional analyses
To assess the robustness of my main results, I conduct four sets of additional tests. First, I use the method of Dechow & Dichev (2002) as an alternative measure of audit quality. Second, I use another classification of short, medium and long tenure. Third, I perform the OLS regressions again, using the natural log of both firm and partner tenure. Finally, I test whether the application of IFRS affects the relation between audit firm/partner tenure and discretionary accruals. Other studies also distinguish the effect between big audit firms and non-big audit
23
firms. However, since my control variable BIG 3 is insignificant in both my main regressions, depicted in Table 5 and 6, this additional test would not mean anything.
Measuring audit quality by quality of accruals
Consistent with Carey & Simnett (2006) and Kim et al. (2015), I employ the model of Dechow & Dichev (2002) as an alternative measure of audit quality. This model analyses the relationship between accruals and cash flows from operations. Since accruals are used to adjust for temporary differences in cash flows, Dechow & Dichev (2002) adjust the Jones model by using past, present and future cash flows in the estimation of the total accruals. First, I estimate per firm-year observation the amount of expected accruals, which is measured as follows:
(𝐸) 𝑊𝐶𝐴𝐶𝐶𝑡 = 𝛽1𝐶𝐹𝑂𝑡−1+ 𝛽2 𝐶𝐹𝑂𝑡+ 𝛽3 𝐶𝐹𝑂𝑡+1+ 𝜀𝑡 (5)
Where CFO is cash flows from operations in year -1, 0 and +1. Similar to 𝐷𝐴, the coefficients 𝛽1, 𝛽2 and 𝛽3 are parameters from the estimation of expected accruals using all companies in the industry. The residual from the regression of equation 5, 𝜀𝑡, shows the level of abnormal accruals. Then I estimate the abnormal amount of accruals, using the following equation:
𝐴𝑏𝑛𝑊𝐶𝐴𝐶𝐶𝑡 = 𝑊𝐶𝐴𝐶𝐶𝑡− ( 𝛽̂1 𝐶𝐹𝑂𝑡−1+ 𝛽̂2 𝐶𝐹𝑂𝑡+ 𝛽̂3 𝐶𝐹𝑂𝑡+1 ) (6)
I perform the equations 3 and 4 again, using the absolute value of accruals as dependent variable11. The results are presented in Table 7 and show that the coefficient of SHORT for both firm tenure and partner tenure is positive (0.0180 and 0.00996, respectively) and significant ( p<0.05 and p<0.01, respectively). The coefficient of LONG for firm tenure is 0.00128 and for partner tenure is -0.0532. Both are insignificant. This means that the results from the alternative Dechow & Dichev (2002) proxy are consistent with my original results based on the Modified Jones model.
TABLE 7Results of regression on absolute accruals and firm/partner tenure
Variables Firm tenure Partner tenure
SHORT 0.0180** 0.00996***
(0.00721) (0.00366)
LONG 0.00128 -0.00532
(0.00605) (0.00676)
Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
24 Alternative classification tenure
Following Johnson et al. (2002), I now classify tenures of three years or less as short. Because there are different perceptions of long tenure, I determine long tenures as seven years or more (Geiger & Ragunandan, 2002) and as eight years or more (Carey & Simnett, 2006). For firm tenure I also perceive long tenures as nine years or more (Carcello & Nagy, 2004) and ten years or more (Regulation (EU) No 537/2014). Because there are only a small number of partner tenures of nine years or more (6 of the 1,527 measured tenures), these tests would not be representative. Therefore, they are not included in the analyses.
The results of the different classifications are shown in Table 812. The coefficients of
FT SHORT and PT SHORT in panel A remain positive (0.0133 and 0.00791, respectively) and
significant (p<0.05). Regarding long tenure, both coefficients FT LONG and PT LONG remain insignificant in all panels. One difference is that the coefficient FT LONG changes into a negative value, but suffers significance. I also conduct the regression with long tenure as eleven years or more and twelve years or more. Even though the coefficient of FT LONG increases to more negative values in both tests, they remain insignificant. Overall, using an alternative classification of short and long tenures does not alter my main results.
Natural log auditor tenure
Carcello & Nagy (2004) and Chen et al. (2008) use the natural log of tenure to investigate the effect of an additional year of tenure on the level of absolute discretionary accruals. I also conduct this test. The results are presented in Table 913. The coefficient of FT equals -0.00787 (p<0.1), which suggests that an additional year of firm tenure is associated with a decrease in DAof 0.79 percent of beginning total assets. The coefficient PT equals -0.0131 (p<0.01), suggesting that an additional year of partner tenure is associated with a decrease in DAof 1.31 percent of beginning total assets. This means that an additional tenure year of the audit partner has a stronger effect on audit quality than an additional year of firm tenure. This is consistent with the results from my main tests in Table 5 and 6, where the coefficient of LONG firm tenure in Table 5 remains to have a positive effect onDA, and the coefficient of LONG partner tenure in Table 6 becomes negative, although not significant. This effect is similar to my second alternative test (Table 8). However, it should be noted that this additional test does not consider an asymmetric relation between tenure and discretionary accruals, whereas my main regressions do.
25
TABLE 8
Results from regression on absolute discretionary accruals and audit firm/partner tenure, using alternative classifications of tenure
Variables Coefficient
Panel A: SHORT = ≤ 3 LONG = 7 ≤
FT SHORT 0.0133** (0.00525) FT LONG 0.00172 (0.00595) PT SHORT 0.00791** (0.00320) PT LONG -0.00102 (0.00972) Panel B: LONG = 8 ≤ FT LONG 0.000997 (0.00687) PT LONG -0.0177 (0.0204) Panel C: LONG = 9 ≤ FT LONG 0.000533 (0.00777) Panel D: LONG = 10 ≤ FT LONG -0.00276 (0.00888)
Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
TABLE 9
Results from regression on absolute discretionary accruals and audit firm/partner tenure, using natural log of tenure
Variables Coefficient
FT -0.00787*
(0.00431)
PT -0.0131***
(0.00366)
Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
Applied reporting standard
An event that occurred halfway within the period of my sample is the obligation to adopt IFRS. Because I expected that the applied reporting standard would affect the level of discretionary accruals, I included this as a control variable in my regressions. As reported in Table 5 and 6, the applied reporting standard does have a significant effect onDA. Therefore, I am interested
26
in what the effect is of applying IFRS/US GAAP or German GAAP on my main results. To do so, I divide my sample into client-year observations where IFRS/US GAAP is applied and client-year observations where German GAAP is applied. Table 1013 presents the results and shows that for both short firm and partner tenures, applying IFRS/US GAAP has a positive and significant (p<0.01) effect on absolute discretionary accruals. Gray et al. (2015) found that managers in more individualism driven countries, such as Germany14, tend to be more aggressive in earnings management. In particular, they find that the positive association between individualism and discretionary accruals are persistent and observable in de post IFRS era. For clients that apply the German GAAP, I do not find any significant effect.
TABLE 10
Results on regression on absolute discretionary accruals and audit firm/partner tenure, testing applied reporting standards
Variables IFRS/US GAAP German GAAP
FT SHORT 0.02314*** 0.01356 (0.00665) (0.0120) FT LONG 0.00111 0.02539 (0.00550) (0.0244) PT SHORT 0.01028*** 0.006516 (0.00398) (0.00684) PT LONG -0.00280 0.01919 (0.00698) (0.02639) Observations 4,259 1,632
Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
V CONCLUSION
In this paper, I analyse the effect of firm tenure and partner tenure on audit quality in Germany. The German audit profession is characterized by high reputation risk and mandatory disclosure of audit partners identity. Mandatory audit firm rotation of every ten years and partner rotation of every seven years should contribute to auditor independence and improve audit quality. However, prior studies do not find consistent results whether and how auditor tenure affects audit quality.
13 Industry and year fixed effects, and the control variables are included, but because of brevity not reported. 14 Other individualism driven countries are, for instance, the United States, Ireland, the Netherlands, and
27
Because of the lack of client-specific knowledge in the early years of the audit engagement, I assume both short firm and partner tenure to have a decreasing effect on audit quality. Regarding long term tenures, I assume the effect of firm and partner tenure on audit quality to be different. Based on the high reputation risk Weber et al. (2008) found in the German audit profession, I expected long firm tenures will not have a detrimental effect on audit quality. In contrast, I assume long partner tenures to decrease audit quality, based on the social identity theory of Tjafel & Turner (1985).
Using the absolute value of discretionary accruals as a proxy for audit quality, I perform OLS regressions to test my hypotheses. I find supporting evidence that both short firm and partner tenure (hypothesis 1 and 3) do have an increasing effect on the level of discretionary accruals, resulting in a lower level of audit quality. This is consistent with the opponents’ view on mandatory rotation. Regarding my second hypothesis, I find no significant effect on the relation between long firm tenure and the amount of discretionary accruals, which is conform my expectation. Regarding my fourth hypothesis, I do not find a significant effect on the relation between long partner tenure and the amount of discretionary accruals. This is not consistent with my expectation. It is possible that the social identity theory is not generalizable and applicable to all individuals, but rather depends on personal characteristics whether an audit partner is prone to client identification. A suggestion for future research is to investigate whether and how audit partners’ personal characteristics affect audit quality.
As an additional test, I use the natural log of tenure. Results of this test show that both an additional year of firm and partner tenure have a significant improving effect on audit quality. Comparing both effects shows that an additional year of partner tenure has a stronger improving effect on audit quality than an additional year of firm tenure. This suggests that audit quality is more dependent on partner level than on audit firm level. Other robustness tests do not alter my main results. Overall, the results suggest that the tenure effect of both the audit firm and engagement partner may be more pronounced in the earlier years of the audit engagement than in later years.
My study contributes to the discussion whether auditor tenure affects audit quality. My findings do not support the view that mandatory firm and partner rotation would be necessary to improve audit quality. My findings are consistent with prior studies who also find that short firm tenure (Johnson et al., 2002; Geiger & Raghunandan, 2002; Myers et al., 2003; Carcello & Nagy, 2004) and partner tenure (Manry et al., 2008; Chen et al., 2008) are associated with lower audit quality. If deterioration of audit quality is the motivation for the regulation of mandatory rotation, then my results do not support this argument. However, mandatory change
28
of audit firm and partner may increase the perception of audit quality, which is an important component to the efficient operation of the financial markets.
My study has also limitations. First, the reason why I found no significant effect on long term partner tenure, might also be due to the limited amount of long partner tenure observations. Second, results from the German setting where auditor rotation is mandatory, may not be applicable to a setting where no such rotation is required. The incentives for auditors and clients might differ between the two settings. Finally, prior studies use a variety of proxies to measure audit quality and do not always find consistent results. It is arguable that the choice of proxy has an impact on the results.
29
REFERENCES
Antle, R. (1984). Auditor Independence. Journal of Accounting Research, 22(1), 1-20.
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(1), 1-34.
Bamber, E. M., & Iyer, V. M. (2007). Auditors’ Identification With Their Clients and its Effect on Auditors’ Objectivity. Auditing: A Journal of Practice & Theory, 26(2), 1-24. Brody, R. G., & Moscove, S. A. (1998). Mandatory Auditor Rotation. National Public
Accountant, 43, 32-36.
Carcello, J. V., & Nagy, A. L. (2004). Audit Firm Tenure and Fraudulent Financial Reporting.
Auditing: A Journal of Practice & Theory, 23(2), 55-69.
Carey, P., & Simnett, R. (2006). Audit Partner Tenure and Audit Quality. The Accounting
Review, 81(3), 653-676.
Chen, C., Lin, C., & Lin, Y. (2008). Audit Partner Tenure, Audit Firm Tenure, and Discretionary Accruals: Does Long Auditor Tenure Impair Earnings Quality?
Contemporary Accounting Research, 25(2), 415–45.
Chi, W., Huang, H., Liao, Y., & Xie, H. (2009). Mandatory Audit Partner Rotation, Audit Quality, and Market Perception: Evidence from Taiwan. Contemporary Accounting
Research, 26(2), 359-391.
Chung, J., Farrar, J., Puri, P., & Thorne, L. (2010). Auditor Liability to Third Parties After Sarbanes-Oxley: An International Comparison of Regulatory and Legal Reforms. Journal of International Accounting, Auditing and Taxation, 19(1), 66-78. DeAngelo, L. E. (1981). Auditor Size and Audit Quality. Journal of Accounting and
Economics, 3(3), 183-199.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting Earnings Management. Accounting Review, 193-225.
Dechow, P. M., & Dichev, I. D. (2002). The Quality of Accruals and Earnings: The Role of Accrual Estimation Errors. The Accounting Review, 77(1), 35-59.
DeFond, M. L., & Jiambalvo, J. (1994). Debt Covenant Violation and Manipulation of Accruals. Journal of Accounting and Economics, 17(1-2), 145-176.
Directive 2006/43/EC. Available at: https://eur-lex.europa.eu/legal-content/EN/ TXT/? uri=CELEX%3A32006L0043.
Directive 2014/56/EU. Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/? uri=celex%3A32014L0056.
Francis, J., LaFond, R., Olsson, P., & Schipper, K. (2005). The Market Pricing of Accruals Quality. Journal of Accounting and Economics, 39(2), 295-327.
Francis, J. R., & Yu, M. D. (2009). Big 4 Office Size and Audit Quality. The Accounting
Review, 84(5), 1521-1552.
Francis, J. R. (2011). A Framework for Understanding and Researching Audit Quality. Auditing: A Journal of Practice & Theory, 30(2), 125-152.
German Commercial Code. Handelsgesetzbuch. Available at: https://www.gesetze-im-internet.de/hgb /BJNR002190897.html.
Geiger, M. A., & Raghunandan, K. (2002). Auditor Tenure and Audit Report Failures.
30
Gold, A., Molls, F., Pott, C., & Watrin, C. (2012). The Effect of Audit Engagement and Review Partner Tenure and Rotation on Audit Quality: Evidence from Germany. Working Paper, 2012. Available at: http://ssrn. com/abstract= 1631947.
Gosh, A., & Moon, D. (2005). Auditor Tenure and Perceptions of Audit Quality. The
Accounting Review, 80(2), 585-612.
Gray, S. J., Kang, T., Lin, Z., & Tang, Q. (2015). Earnings Management in Europe Post IFRS: Do Cultural Influences Persist?. Management International Review, 55(6), 827-856. Gul, F. A., Fung, S. Y. K., & Jaggi, B. (2009). Earnings Quality: Some Evidence on the Role
of Auditor Tenure and Auditors’ Industry Expertise. Journal of Accounting and
Economics, 47(3), 265-287.
Hill, C. W. L., & Jones, T. M. (1992). Stakeholder – Agency Theory. Journal of Management
Studies, 29(2), 131-154.
Hohenfels, D. (2016). Auditor Tenure and Perceived Earnings Quality. International Journal
of Auditing, 20, 224 –238.
IFRS. (2018). Conceptual Framework for Financial Reporting. Available at:
https://www.ifrs.org/-/media/project/conceptual-framework/fact-sheet-project-summary-and-feedback-statement/conceptual-framework-project-summary.pdf
Jackson, A. B., Moldrich, M., & Roebuck, P. (2008). Mandatory Audit Firm Rotation and Audit Quality. Managerial Auditing Journal, 23(5), 420-437.
Junaidi, Setiyono Miharjo, & Bambang Hartadi. (2012). Does Auditor Tenure Reduce Audit Quality? International Journal of Business, 14(3), 303 – 315.
Johnson, V. E., Khurana, I. K., & Reynolds, J. K. (2002). Audit‐Firm Tenure and the Quality of Financial Reports. Contemporary Accounting Research, 19(4), 637-660.
Kim, H., Lee, H., & Lee, J. E. (2015). Mandatory Audit Firm Rotation and Audit Quality.
Journal of Applied Business Research, 31(3), 1089-1106.
Knechel, W. R., & Vanstraelen, A. (2007) The Relationship between Auditor Tenure and Audit Quality Implied by Going Concern Opinions. Auditing: A Journal of Practice & Theory, 26(1), 113–131.
Knechel, W. R., Krishnan, G. V., Pevzner, M., Shefchik, L. B., & Velury, U. K. (2013). Audit Quality: Insights from the academic literature. Auditing: A Journal of Practice &
Theory, 32(1), 385-421.
Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance Matched Discretionary Accrual Measures. Journal of Accounting and Economics, 39(1), 163-197.
Kwon Young, S., Lim, Y., & Simnett, R. (2014). The Effect of Mandatory Audit Firm Rotation on Audit Quality and Audit Fees: Empirical Evidence from the Korean Audit Market.
Auditing: A Journal of Practice & Theory, 33(4), 167–195.
Manry, D. L., Mock, T. J., & Turner, J. L. (2008). Does Increased Audit Partner Tenure Reduce Audit Quality?. Journal of Accounting, Auditing & Finance, 23(4), 553-572.
Myers, J. N., Myers, L. A, & Omer, T. C. (2003). Exploring the Term of the Auditor-Client Relationship and the Quality of Earnings: A Case for Mandatory Auditor Rotation? The
Accounting Review, 78(3), 779-799.
Piot, C. (2005). Auditor Reputation and Model of Governance: A Comparison of France, Germany and Canada. International Journal of Auditing, 9(1), 21-44.
Quick, R., Schenk, N., Schmidt, F., & Towara, T. (2018). The Impact of Corporate Governance on Auditor Choice: Evidence from Germany. Journal of Management and
