Mandatory disclosure of the engagement partner’s identity and audit quality

(1)

Amsterdam Business School

Mandatory Disclosure of the Engagement Partner’s

Identity and Audit Quality

Master Thesis - Draft Version 3

Name: Ester Kok

Student number: 6050050 Date of submission: June 19, 2014

1st_Supervisor: _{Prof. dr. D.M. Swagerman}

(2)

(3)

Abstract

The PCAOB has proposed to require audit firms to disclose the identity of the engagement partner in the audit report. They state that this will increase audit quality, however, they do not yet have any evidence for this. In the European Union the requirement is already implemented, however the European Commission never explained why they implemented it. Furthermore, there is too little research done about the impact of the requirement on audit quality and the research that is conducted is divergent. Therefore, this paper will investigate if the disclosure of the Engagement Partner’s Identity will increase audit quality. The expectation is that after the implementation of the requirement for Engagement Partners to disclose their identity in the audit report audit quality increases. This hypothesis is tested based on data research in the Netherlands where this requirement is implemented in 2005. Audit quality is measured by earnings management, which is computed using the performance matched discretionary accruals model. The findings are that there is no relationship between the disclosure of the identity of the Engagement Partner and audit quality. While this could mean there is actually no relationship, this could also be explained by the size of the sample or the fact that non-listed companies are used instead of listed companies.

(4)

Introduction

As a result of the numerous accounting scandals of past decades the function of the auditor and audit quality remain an ongoing debate. The society keeps asking questions about the function of the auditor and wonder about the added value of the audit report. In addition, the audit quality is questioned. It seems the expectation gap is still present. As a result of these discussions, the legislation around the audit is increasingly tightened.

One example is the implementation of the requirement for Engagement Partners (EPs) to disclose their name in the audit report. This requirement is already implemented in the European Union (EU) in 2006 (Directive 2006/43, article 28). However they never explained why this requirement is implemented. Now, the the Public Company Accounting Oversights Board (PCAOB) also proposes to implement this requirement (PCAOB 2011). Until now, in the United States of America (USA) only the name of the CPA-office was required to be presented in the audit report. The PCAOB has however arguments for their proposal. The PCAOB argues that the implementation of the proposed requirement will have a positive impact on audit quality. In their proposal they mention numerous arguments to support their statement. The first argument is that the transparency of the audit process increases. The investors will be able to see which person was responsible for the audit of a company. The second argument is that the EP’s accountability will increase. The EP’s name will be visible in the audit report so everyone can request justification from the EP. Furthermore, the EP’s responsibility for the quality of the audit will increase. As the name of the EP is explicitly disclosed in the audit report, the EP will feel more responsible for the issued report. And the last argument is that it gives an opportunity for the general public to evaluate the performance of an EP. The general public is able to find out easily who was responsible for the audit of a company when for example fraud is detected at that company (PCAOB 2011; King et al. 2012).

(7)

Although these arguments all seem very plausible, the PCAOB gives no proof of the increase in audit quality.

So the EU has given no arguments for their requirement and the PCOAB has arguments but no proof for these arguments. In addition, there has also conducted little research about the effect of EP disclosure on audit quality. Carcello and li (2013) however did investigate the relationship between signing the audit report by the EP and audit quality in the UK (PCAOB, 2009). They also took into account the audit cost. They did find a positive relationship for both audit quality and audit cost. The question they cannot answer is if the increase in audit quality outweighs the increase in audit cost. Although they investigated the impact of signing the audit report (which is not publicly visible) instead of disclosing the name of the EP it is expected that this last relationship will be even stronger because this is publicly visible. In addition King et al. (2012) have evaluated the proposal of the PCAOB 2011 by describing the pros and cons. They find not only arguments for the implementation of the requirement to disclose the identity of the EP in the audit report, but also arguments against the implementation of the requirement. They argue that it is possible that audit quality in appearance increases, but there is no effect on audit quality in fact. This is an unwanted effect. Both papers give some indications for the effect of the implementation of the requirement for EPs to disclose their identity in the audit report on audit quality, but this specific relationship is never investigated before with empirical research.

It seems therefore interesting to find out what the impact of disclosure of the EP’s identity in the audit report will be. Based on the recent debates around audit quality, the requirement of the EU, the possible new requirement in the USA and the fact that this relationship is never investigated before with empirical research, this paper investigates the relationship between the disclosure of the EP’s identity and audit quality. The research question of this paper is therefore as follows:

(8)

Does the disclosure of the Engagement Partner’s identity in the audit report increase audit quality?

If this question will be answered with yes the question still remains if the cost of the audit also increase. If this is the case, one should ask if the benefits outweigh the costs. But this paper concentrates on the relationship with audit quality, because there is limited data available about audit costs with regard to the used sample.

The research question will be answered by investigating the implementation of the requirement in the Netherlands in 2005 (Dutch Civil Code, 2014). Audit quality before 2005 will be compared to audit quality after 2005 using data research. Audit quality is measured by discretionary accruals (earnings management), which is computed using the performance matched discretionary accruals model (Kothari et al, 2005). The data that is used is manually gathered data from financial reports from non-listed companies in the Netherlands. Non-listed companies are used to exclude the effect of the implementation of International Financial Reporting Standards (IFRS) in Europe in 2005. Based on the arguments of the PCAOB and the research of Carcello and Li (2013), it is expected that the audit quality after 2005 (after the implementation of the requirement) will be higher than audit quality before 2005. The results however, show no relationship at all. This can be explained by the fact that the sample is relatively small and consists of non-listed companies.

In the next chapter the theory around audit quality, the proposal of the PCAOB including their arguments and previous research are outlined. The hypothesis is also presented. The following chapter describes the research design. Chapter 3 will show the results and based on these results the discussion in chapter 4 gives an answer to the research question. In chapter 4 are also some additional notes included. Finally, the conclusion and limitations of this paper are mentioned and recommendations are given.

(9)

1 Theory

In this chapter, first, the definition of audit quality is introduced and the relevance of audit quality is explained. This forms the basis for understanding the possible effects of the implementation of the PCAOB-proposal that requires to disclose the EP’s identity in the audit report. Next, the two PCAOB-proposals (2009 and 2011) are described. Then, the papers of King et al (2012) and Carcello and Li (2013) are discussed. Finally, based on this theory the research question is presented.

1.1 Audit Quality

1.1.1 The definitions of and proxies for Audit Quality

There is a lot of research done about audit quality, but there is still no general used definition (Knechel et al, 2013, p. 385). To understand audit quality, the different definitions used in existing literature are described. One widely used definition of audit quality is the definition of DeAngelo (1981a); (Knechel et al, 2013, p. 387). She states that audit quality is the chance that an auditor detects a breach in an organizations accounting system and also reports this detected breach. This definition has two parts. The first part of this definition describes the skills of an auditor and the procedures the auditor performs. Is the auditor able to find the breach? The second part of the definition is the independence of the auditor. If the auditor finds the breach will he or she report the breach, or are there incentives not to report the breach? (DeAngelo, 1981a, p. 186). A lot of other researchers use a similar definition: an auditor delivers high audit quality when he or she detects errors in the financial statements (Chan and Wong 2002; Gul et al. 2002; Behn et al. 2008; Chang et al. 2009). This definition is in line with the first part of the definition of DeAngelo (1981a). The Government Accountability Office (GAO) gives another

(10)

definition of audit quality: an audit has high quality when it is performed in accordance with Generally Accepted Auditing Standards (GAAS) and gives reasonable assurance that the audited financial statements are presented in accordance with the Generally Accepted Accounting Principles (GAAP) and are not materially misstated (GAO, 2013, p. 13). In other words, audit quality is high when what the auditor declares in the audit report is true.

This paper will follow previous research and defines audit quality with the definitions of DeAngelo (1981a), because it includes two important parts of audit quality: the skills and procedures of the auditor and the independence of the auditor. Previous research focuses only on the second part of the definition. They assume that the first part of the definition, the skills and procedures of the auditor, is a fixed value (Deis and Giroux, 1992, p. 464; DeAngelo, 1981b, p. 116). They assume therefore that changes in audit quality are fully due to changes in the independence of the auditor. The skills of the auditor and procedures of the auditor are no problem; the auditor should be able to find all the breaches. The question is whether they will report them or not (Deis and Giroux, 1992, p. 464). However, as Deis and Giroux (1992, p. 464) indicate, it is important to acknowledge that there are some variables which determine the skills of an auditor and procedures of the CPA-office, for example experience, education and office size. The skills of the auditor and procedures of the CPA-office are therefore in practice not entirely fixed. But this variable is difficult to measure. This can only be done with information about the background of the auditor and the CPA-office. For example, information about the education and experience of the auditor and information about the internal courses and procedures of the CPA-office. This information is generally not publicly available.

The above-mentioned definitions of audit quality are all positive definitions. But the measurements or proxies for audit quality are generally negative or, in other words, are proxies for poor audit quality (Peecher and Piercey, 2008). It is argued that audit quality failure is better to measure than good audit quality (Casterella et

(11)

al., 2009, 716). In contrast, Knechel et al. (2013, p. 388) state that it is difficult to detect audit quality failure. But previous research shows that it is indeed better to measure audit quality failure than audit quality itself. This is evident from the proxies that are mostly used to measure audit quality. For example, earnings management or discretionary accruals is a widely used proxy for audit quality (e.g. Chung and Kallapur (2003); Ferguson et al. (2004); Frankel et al. (2002) and Jackson et al. (2008)). It is stated that when the amount of discretionary accruals is high, the auditor has allowed management to use its power in order to affect the financial statements in favour of management. In other words, management has used earnings management, and this means that the auditor has failed. Furthermore, the amount of financial restatements is an often used measurement for audit quality (e.g. Stanley and DeZoort (2007); Bloomfield and Shackman (2008) and Kinney et al. (2004)). This is also a measurement for audit quality failure, because financial restatements are reporting errors. Therefore, if there are financial restatements it means the auditor did not find, or did find but did not report, the reporting error.

However, it is also possible to measure audit quality itself. In addition to the above mentioned proxies for audit quality, the propensity to issue going concern opinions is also often used as a proxy for audit quality (e.g. Basioudis et al. (2008); Calaghan et al. (2009) and Knechel and Vanstraelen (2007)). It is stated that when an auditor issues a going concern audit opinion before a company goes bankrupt the auditor was able to detect the going concern issue and reported this, and auditors who are better able to detect a going concern issue have higher audit quality.

In this paper the amount of earnings management is used as a proxy for audit quality. This measurement indirectly measures the skills of the auditor and procedures of the CPA-office. To prevent management from managing earnings, an auditor first has to detect the earnings management. It is expected that an auditor with more experience, more/better education and better auditing procedures is more likely to detect earnings management. So although earnings management primarily

(12)

is a measurement for auditor independence (will the auditor report the earnings management and prevent management from applying earnings management?), it indirectly says something about the skills of the auditor.

1.1.2 Why is audit quality relevant?

Why is audit quality (still) relevant? The answer is that the audit is something on which the users of the financial statements (investors, financial institutions etc.) want to rely. Hence, it is important that audit quality is sufficient. An unqualified audit opinion should give the user of the financial statements assurance that the financial statements are not materially misstated. History teaches that this is not always the case. The accounting scandals of inter alia, Enron, Worldcom, Parmalat and Ahold raised questions: Where was the auditor? Should the auditor have warned us? If they did not, what is the added value of the auditor? Due to these accounting scandals, the role of the auditor and audit quality are questioned.

Because of this a lot of research is conducted about audit quality. This great amount of research shows that audit quality is a complex concept, because it shows that audit quality is influenced by many factors. Auditor size is one factor that is often investigated (i.a. DeAngelo (1981a) Lennox (1999) and Francis and Yu (2009)). Most research found that if auditor size increases, audit quality will also increase. Another factor that is investigated in relationship to audit quality is the amount of non-audit services (i.a. Basioudis et al. (2008); Bloomfield and Shackman (2008) and Callaghan et al. (2009)). The expected relationship is that if the amount of non-audit services increases, non-audit quality decreases because non-auditor independence will decrease. But the findings of the researches are divided and therefore it is not clear if this relationship exists. Finally another example of a factor that is expected to influence audit quality is audit firm or engagement partner tenure (i.a. Chen et al. (2008); Stanley and DeZoort (2007) and Jackson et al. (2008)). In this case, the main findings are that in the first years of the audit, the audit quality increases

(13)

(because of more client experience) but at a moment in time audit quality decreases (because of a decrease in independence). These are only examples of factors that influence or may influence audit quality.

The above paragraph shows that there are many factors that influence audit quality. It also shows that there is conducted a lot of research about audit quality. The remaining question is: why is it still relevant to investigate audit quality? This is because we still do not know all the factors that affect audit quality. The fact that the PCAOB introduces another factor that could influence audit quality, the disclosure of the EP’s identity in de audit report, supports this. But if we know these factors and the effect they have on audit quality we can eliminate the ones that decrease audit quality and implement or require the ones that increase audit quality. This will finally increase audit quality and therefore decrease the chance that big accounting scandals as mentioned in the beginning of this section will happen again. However it is not completely that simple. There is another fact: audit costs. Audit quality increase means often also audit cost increase. And this factor should also be taken into account.

1.2 Proposals of the PCAOB

Before the PCAOB proposed to require to disclose the name of the EP in the audit report, they proposed to require the EP to sign the audit report (PCAOB 2009). They refer to the EU’s Eighth Company Law Directive (Eighth Directive) (PCAOB 2009, p. 2-3). In the Eight Directive, EP’s are required since 2006 to sign the audit report (Directive 2006/43, article 28). Most European countries even required this before 2006, for example Germany and France. Why do these countries require this and, and why does the PCAOB also wants to require this?

Why the EU requires this is unclear. They explain a lot of new requirements in the Eighth Directive, but article 28 is not one of them. However the PCAOB has some

(14)

arguments in their proposal. They believe that audit quality will increase when the EP signs the audit report (PCAOB 2009, p. 5). In the 2009 proposal, they underpin this statement with 2 arguments:

‘First it might increase the engagement partner's sense of accountability to financial statement users, which could lead him or her to exercise greater care in performing the audit. Second, it would increase transparency about who is responsible for performing the audit, which could provide useful information to investors and, in turn, provide an additional incentive to firms to improve the quality of all of their engagement partners.’ (PCAOB 2009, p.5, emphasis added)

If the EP knows he or she has to sign the audit report with his or her own name, he or she takes his or her responsibility and makes sure the audit quality is high (PCAOB 2009 p. 5). King et al. (2012) summarize the arguments of the PCAOB in figure 1.

In addition, the proposal refers to section 302 of the Sarbanes-Oxley Act (SOX), which requires principal executive officers (PEO) and principal financial officers (PFO) to declare that the report does not contain any material misstatements. This declaration is only an emphasis of the responsibilities the PEO and PFO already had. This would also be the case for the EP (PCAOB, p. 7). The EP is already responsible for the audit report, the signing of the report will only emphasize this.

After the comments they received at the 2009 proposal, the PCAOB changed their proposal and this is the proposal of 2011, which is investigated in this paper (PCAOB 2011). The PCAOB proposes in this proposal that only the identity of the EP is disclosed in the audit report, they don’t have to sign the report. The comments they received on the proposal of 2009 were divergent. Investors argued that requiring the EP to sign the audit report would increase transparency and audit quality. When

(15)

investors see who is responsible for the audit they are also better able to demand skilled auditors. Therefore they expect that CPA-offices will strengthen the skills of the EPs (PCAOB, 2011, p. 6-7). In addition, some accounting professors say that, although not scientifically investigated, it is likely that the implementation of the signing requirement will increase audit quality, but there are also accounting professors who argue that the implementation of the signing requirement will increase EP accountability at the expense of office accountability. If this is the case, the impact on audit quality is not clear (PCAOB, 2011, p. 7-8).

Figure 1: PCAOB Arguments

Accounting firms and audit associations where also less positive about the proposal. They commented that the EPs are already taking full responsibility and accountability. They already want to keep a good reputation, both for themselves and for the firm. Furthermore, they state that it has no benefit for the investors to know the name of the auditor. The audit committee knows the EP and they are better able

Mandatory Signature or Identity Disclosure of the EP Transparency of EP’s Identity Investors Perceive Audit Report as More Informative Improved Overall Audit Quality EP’s perceptions of Increased Accountability Motivation of CPA -firms to Increase Quality of all EP’s

Increased Due Professional

(16)

to evaluate the EP and act in favor of the investors. The accounting firms and audit associations are also afraid that the implementation of the requirement will cause investors to keep track of EP’s performances. This could result in investors drawing the wrong conclusions about financial statements based on the identity of the EP (PCAOB, 2011, p. 8).

The comments show some arguments for and against the proposal. But it is notable that they do not have specific arguments for or against signing the audit report by the EP. They more generally have arguments about the disclosing the identity of the EP, by either disclosing their name or letting them sign the audit report. Why did the PCAOB change their proposal? This is because of the argument of the accounting professors about the decrease of the accountability of the CPA-office. If only the identity of the EP is disclosed in the audit report, but the audit report is still signed on behalf of the office, it is expected that the accountability of the EP will increase but the accountability of the CPA-office will not decrease. Therefore, it is expected that audit quality will increase, despite the arguments of the accounting firms and audit associations (PCAOB, 2011, p. 9-10).

1.3 Previous research and hypothesis

1.3.1 King et al. (2012)

King et al. (2012) investigated the possible effects of the implementation of the two proposals of the PCAOB (PCAOB 2011 and PCAOB 2009) by a literature study. Their research has three goals:

1. Evaluating the pros and cons of the implementation of the requirements using existing literature

(17)

2. Discussing how the effects of implementing the proposal of 2009 differ from implementing proposal 2011

3. Suggesting further research opportunities

This section focuses on the first goal. King et al. (2012) introduce three theories to achieve the first goal: source credibility theory, accountability research and the affordance theory. This last theory only applies to the signature requirement; therefore it is ignored in this paper.

The source credibility theory originates in communication studies. This theory explains the transparency argument of the PCAOB. The definition of credibility is:

‘the attitude toward a source of communication held at a given time by a receiver’ (McCroskey 1997, 87)

Credibility is therefore a subjective term. It depends on the receiver, in this case the investors, how credible information is. The audit report is the information for investors and the source of information is the EP. The theory states that information users want more transparency about the source of the information. The proposed requirement generates this. However, users of information also rely more blindly on the information if there is more transparency. This could cause the effect which the academic accounting community warned for: investors will look more at the characteristics of the EP instead of the information in the audited financial statements. So based on the source credibility theory, the implementation of the proposal of the PCAOB would not have a positive effect. In addition, because investors believe (refer also to section 1.2) the implementation of the requirement to disclose the identity of the EP in the audit report will increase audit quality, there is a risk that audit quality in appearance increases, but audit quality in fact will not. This is an unwanted effect.

(18)

The accountability theory originates in the social psychology (King et al, 2012). The theory states that there are two minimum conditions that should be achieved to make sure that accountability has a positive effect:

1. The subject (in this case, the EP) should be accountable to an audience (in this case, the investors) with unknown views

2. The subject (EP) should be aware of the requirement to account for their decision in the future before they make that decision

However, these conditions are not enough. Previous research has found that accountability only increases performance if the performance was worse because of lack of effort and attention and not because of lack of resources or task complexity (Kennedy 1993, Libby et al. 2004). In case of the EP, audit quality will only increase if audit quality is at the moment not at an optimum (which is most likely) due to the fact that the EP lacks effort and attention. If this is true, the effort and attention will increase as a result of the implementation of the requirement, which in turn leads to higher audit quality, but the problem is that this also increases audit costs. If the costs outweigh the benefits, this is not a desirable effect. King et al. (2012) conclude based on the previous research that there is no support for the argument of the PCAOB that increased accountability will increase audit quality. The possible effect stays unclear. In addition, they emphasize that the research they used is about accountability in general. There is no research about accountability and EPs.

Based on the research of King et al. (2012) the effect of the implementation of the requirement to disclose the identity of the EP in the audit report is unclear. They have arguments against and arguments for the implementation. But especially the arguments against are noteworthy. There is a risk that only the audit quality in appearance will increase and audit quality in fact will not. In addition, there is a risk that audit quality will increase, but the audit cost will increase even more. So this gives enough reasons to investigate the effect of the implementation of the requirement in a country where it is already implemented.

(19)

1.3.2 Carcello and Li (2013)

Carcello and Li (2013) investigated the implementation of the signing requirement in the UK. Although this is different from the relationship which is investigated in this paper, the paper of Carcello and Li (2013) should give some concrete findings to determine the expectations for this paper.

Carcello and Li (2013) investigated, based on the PCAOB-proposal of 2009, what the effect is of requiring the EP to sign the audit report in his or her own name on audit quality and audit costs. They investigate this effect in the UK, where this requirement is implemented in 2009. They use earnings quality and the propensity to issue a qualified audit opinion as a measurement for audit quality and compare audit quality in the year before the implementation of the requirement and the year after the implementation of the requirement. In addition, they examine the effect of the requirement on audit costs, because they argue that if audit quality increases but the audit costs increase even more, the implementation of the proposal of the PCAOB would not be desirable.

They find an increase in audit quality due to the implementation of the requirement. Earnings management has declined and the propensity to issue a qualified audit opinion has increased, but as expected (also based on the research of King et al (2012)) they find a significant increase in audit cost (audit fees). However, the research can not say if these cost have increased more than the benefits or not. Therefore, based on the research of Carcello and Li (2013) it is not clear if the implementation of the requirement for EPs to sign the audit report is a good idea.

As already mentioned, they investigated the implementation of the requirement to sign the audit report. As noted in section 1.2, this requirement increases the risk that the accountability of the CPA-office decreases which could lead to no increase of audit quality. This is the reason why the PCAOB changed the

(20)

proposal. Based on the research of Carcello and Li (2013), this seems not the case. In addition, there is another major difference between the implementation of the signing requirement in the UK and the implementation of the disclosure requirement. The signature is not publicly visible as the name of the auditor is. For example, the audit report with the EP’s signature in the UK does not even have to be filed with the Companies House (PwC Legal, 2010). The disclosure of the EP’s identity in the EU is publicly visible, even on the financial statements of companies’ websites. So actually everyone can see who is responsible for the audit of a particular company. This is important, because this increases the reputational damage of the EP in case of audit failure. Therefore, it is expected that the impact of the implementation of the requirement of disclosing the EP’s identity on audit quality is even bigger than the impact of the implementation of the requirement for EPs to sign the audit report. However, it is possible that the impact on audit costs is also bigger.

The paper of Carcello and Li (2013) finds that the implementation of the requirement for EPs to sign the audit report increases audit quality but also audit costs. The question remains if the benefits outweigh the costs. For now the paper of Carcello and Li (2013) gives an indication of what can be expected from the requirement to disclose the identity of the EP in the audit report. In the following section the hypotheses will be introduced.

1.3.3 Hypothesis Development

As mentioned in the introduction the research question that is answered in this paper is:

This research question is answered by data research and more specific hypothesis test. This is further explained in section 2.1.

(21)

Based on the previous sections in chapter one there is not an unambiguous expectation. The EU does not give specific arguments. The 2009 and 2011 proposal of the PCAOB, the researches of King et al. (2012) and Carcello and Li (2013) have divergent expectations and findings. In particular Carcello and Li (2013) give indications based on concrete evidence that the disclosure of the EP’s identity in the audit report will increase audit quality. The other arguments are all based on general theories about accountability (King et al., 2012) or intuition (PCAOB, 2011). Furthermore, there is no indication for a negative relationship between audit quality and the disclosure of the identity of the EP based on the previous research and the expectations of the PCAOB. Therefore the question seems to be: is there a positive relationship or is there no relationship at all. Therefore the hypothesis of this paper is:

H1: Audit quality will increase after the implementation of the requirement for EPs to disclose their identity in the audit report.

The paper of Carcello and Li (2013) has added another hypothesis to their research. They also test the relationship between the implementation of the requirement and audit costs. This would also be a relevant addition to this research, however the data for this is very limited available. Hence, the choice is made to not include this in this paper. Therefore, only one hypothesis is used to answer the research question.

This chapter described the relevant theory about audit quality and the proposal of the PCAOB. Based on this theory, the above hypothesis is developed. In the next chapter the method of research is described.

(22)

2 Research Design

The next chapter will present the research design. First, the research method is described. Then, based on the theory in section 1.1.1 and developed models from previous research, the proxy for audit quality is explained. Furthermore, the model that is used to test the hypothesis is developed, including the controls variables. In the last sections the sample selection method and the sample are described.

2.1 Research Method

As mentioned in the previous sections, the research question will be answered by hypothesis test. The reason for this is that with this method an independent answer can be found. The hypothesis test will be based on data from financial statements of companies and this is objective data. This is also an often used method to investigate audit quality (e.g. Chung and Kallapur (2003); Ferguson et al. (2004); Stanley and DeZoort (2007); Bloomfield and Shackman (2008) and Basioudis et al. (2008)).

Based on the paper of King et al. (2012) other test methods are considered. King et al. (2012) give some suggestions for further research. For example interviewing or surveying investors and EPs. By interviewing or surveying investors, their view about this topic could be investigated. However, this view is already known based on the comments on the proposal of the PCAOB (2011). The view of the EPs is indirectly known based on the comments of the accounting firms and audit associations. Furthermore, it is expected that both views are not objective. For example, it is unlikely that an EP will admit that his or her audit quality and accountability is at the moment not at an optimal level, whether this is fact true or not. Therefore, for this paper is chosen to investigate the research question based on objective data. However, the methods suggested by King et al. (2012) are good ways to explain the relationship between EP name disclosure and audit quality if a relationship is found.

(23)

The hypothesis test is conducted with linear regression. This is done, because this research attempts to explain audit quality based on the disclosure of the name of the EP. Linear regression is an often used and suitable way to do this. Furthermore, as will be explained in the next section, a regression model is used to determine audit quality. Previous research that used this regression model or a comparable regression model also used linear regression to answer the research question ((e.g. Chung and Kallapur (2003); Ferguson et al. (2004); Frankel et al. (2002) and Jackson et al. (2008)).

2.2 Regression Models

2.2.1 PMDA-model

As explained before, audit quality is measured by earnings management. Refer also to section 1.1. However, earnings management itself is also not directly measurable. Therefore a proxy for earnings management is needed. In previous research, the Jones Model (1991) is an often-used measurement for earnings management. Jones (1991) developed this model for her research at earnings management. She uses discretionary accruals as a measurement for earnings management. What she does different compared to previous research (DeAngelo, 1986 and Healy, 1985) is that she does not assume that non-discretionary accruals are constant. She argues that changes in economic circumstances do affect non-discretionary accruals (Jones, 1991). By measuring total accruals and non-discretionary accruals a residual is left and the model states that this residual are the discretionary accruals (Jones, 1991). The residual is therefore the proxy for earnings management.

In the first model (Jones, 1991) the discretionary accruals in revenue are not included. Because of this, Dechow et al. (1995) had some comments at the Jones

(24)

Model (1991). They argue that accounts receivables can be used to manage revenues and are therefore discretionary accruals. They should be corrected in revenues. Hence, they add accounts receivable to the model. This modified model is called the Modified Jones Model (Dechow et al., 1995, p. 199).

Kothari et al. (2005) investigated different discretionary accrual models. They state that the above-mentioned models do not take into account the performance of a company. A company with a high performance is more likely to apply earnings management. Kothari et al. (2005) argue that it is necessary to distinguish normal earnings management and abnormal earnings management. The relatively high amount of earnings management of a company with high performance is normal. So the model should correct for this. This is done by adding the variable return on assets (ROA) (Kothari et al., 2005, p. 165). They also add a constant to the model. The previous Jones and Modified Jones model did not contain a constant. Kothari et al (2005) point that adding a constant controls for heteroskedasticity. Their model is called the performance matched discretionary accruals model (PMDA-model) and is the model that is used in this paper to determine discretionary accruals, refer to Equation 1.

Another often-used model to measure earnings management, the model of Dechow and Dechev (2002), is also considered for this paper. The model establishes a link between the changes in working capital and the changes in the cash flows from operating activities. Although this is an often used model, for this paper it can not be used because of the limited available data when it comes to the cash flows from operating activities.

(25)

Equation 1: PMDA-model ܶܣ௜,௧= ߚ଴+ ߚଵ _ܣݏݏ݁ݐݏ1

௜,௧ିଵ+ߚଶ(∆ܴܧܸ௜,௧− ∆ܴܧܥ௜,௧) + ߚଷܲܲܧ௜,௧+ ߚସܴܱܣ௜,௧+ ε௜,௧

Where

TAi,t = Total accruals for company i in year t scaled by the total assets

of company i in year t-1

Assetsi,t-1 = Total assets for company i in year t-1

ΔRevi,t = Change in net sales for company i in year t scaled by the total

assets of company i in year t-1

ΔReci,t = Change in accounts receivable for company i in year t scaled by

the total assets of company i in year t-1

PPEi,t = Total tangible fixed assets for company i in year t scaled by the

total assets of company i in year t-1

ROA = Return on assets (=Net income/Total assets) for company i in year t

ε1,t = Error term for company i in year t = proxy for earnings

management

So the error term from the PMDA-model is used as a proxy for earnings management. This term can be either positive or negative, depending on the type of earnings management. One could argue that a negative error term means income lowering earnings management and is therefore only a conservative way of reporting earnings. However, it is actually earnings management so therefore both positive and negative error terms are earnings management. To prevent the positive and negative error terms to cancel each other out, the absolute value of the error term is used as a proxy for earnings management. More about the consequences of this in the next chapter.

(26)

2.2.2 Regression model & controls variables

In the linear regression model, earnings management is the dependent variable and the disclosure of the EP is the independent variable. The disclosure of the EP is measured with a dummy variable, measuring one if the identity of the EP is disclosed and measuring zero if the identity of the EP is not disclosed. Earnings management is measured as described in the previous section.

To the above-described model some control variables are added. Previous research has shown that there are several variables that influence earnings management. The first variable that influences earnings management and is therefore included in the model as a controls variable is the size of a company. This variable is used frequently in previous research as a controls variable. It is stated that a larger company has less earnings management, because the risk of reputational damage is larger for both the auditor and the company and larger companies have usually more resources for a better internal control environment. This controls variable is also used by for example Geiger and Raghunandan (2002), Carcello and Nagy and Chen et al. (2008). Although there are other proxies for company size, this paper used the same proxy as the before mentioned papers: the natural log of total assets. This is the most used proxy for size.

The next controls variable that is added to the model is leverage. Previous research has found that companies with a high debt ratio apply more earnings management than companies with a lower debt ratio. Reason for this is that companies with more debt in relation to total assets have generally more incentives to meet certain conditions in the agreements with their bank or other investor. Management can use earnings management to meet these conditions (Johnson et al., 2002; Ghosh and Moon, 2004).

The last controls variable that is added is the size of the audit firm. The distinction is made between Big 4 audit firms (KPMG, EY, PwC and Deloitte) and other audit firms. This distinction is a frequently used distinction. In previous

(27)

research it is found that audit quality of Big 4 audit firms is higher than the audit quality of smaller audit firms. This is because Big 4 audit firms have more recourses (Davidson and Neu (1993); DeAngelo (1981a) and Lennox (1999)). This variable is a dummy variable with a value of one if the financial report is audited by a Big 4 audit firm and a value of zero if the financial report is not audited by a Big 4 audit firm.

Summarizing, the following model is used to test the hypothesis:

Equation 2: EP-disclosure model

ܧܯ = ߙ + ߚ଴ܧܲܦ݅ݏܿ+ ߚଵܵܫܼܧ + ߚଶܮܧܸ + ߚଷܤ݅݃4

Where:

EM = The amount of earnings management, measured by the absolute value of the residual of equation 1

EPDisc = Dummy variable which is one if the identity of the EP is disclosed and zero if not

SIZE = The size of a company measured by the natural log of total assets

LEV = The amount of leverage measured by the debt ratio of a company (liabilities/total assets)

Big4 = Dummy variable which is one if the financial statements are audited by a big 4 audit firm and zero if not

2.3 Sample

2.3.1 Sample country, period and companies

As explained in the introduction, the relationship between the disclosure of the identity of the EP in the audit report and audit quality will be investigated in the

(28)

Netherlands. This is because the Netherlands is a country in which the requirement is implemented in 2005. The Netherlands is therefore one of the last European countries who implemented this requirement just before the requirement was implemented in the whole EU. The requirement was for example already applicable for some years in France and Germany (King et al. 2012, 537). Therefore, the Netherlands is suitable because a relatively recent comparison can be made. However, it has to be noted that the Netherlands is very different from the USA. There are for example differences in financial reporting (US GAAP versus IFRS) and in corporate scandals (Coffee Jr., 2005). Furthermore, there are institutional differences which lead to differences in risk management (Bodnar et al. (2003). These are only a few examples of all the differences between the USA and the Netherlands. This paper is written in response to the proposal of the PCAOB to implement the requirement in the USA. But as evidenced by the before mentioned differences it is not possible to conclude about the proposal of the PCAOB based only on evidence from the Netherland. However, the results of this paper can give an indication and in combination with previous research (King et al. (2012) and Carcello and Li (2013)) and possible future research it can contribute to the question whether the PCAOB should implement the requirement or not. Moreover, it can contribute to evidence about the relationship between the disclosure of the identity of the EP in the audit report and audit quality in general.

So the research is conducted in the Netherlands. Audit quality before and after 2005 is compared because the requirement to disclose the identity of the EP’s identity in the Netherlands is implemented in 2005 (Dutch Civil Code, 2014). Therefore, the year 2004 is compared to the year 2005. It has been considered to investigate a longer period than two years. But this paper follows Carcello and Li (2013) and investigates only the year before and after the implementation of the rule. The reason for this is that there is less available data before 2004. Furthermore, companies that change from auditor are not used in this research because this could

(29)

influence audit quality. Therefore, if the period will be extended, there are more companies that changed the auditor and are not usable.

The period in which the research is conducted is clear. Now, the companies from which the data is collected will be described. The data for the PMDA-model is gathered from financial reports from Dutch non-listed companies. The advantage of using listed companies instead of listed companies and the reason that non-listed companies are used is that this eliminates the effect of the implementation of IFRS for listed companies in the European Union. Previous research has found that the implementation of IFRS in the EU has a positive effect on audit quality (Barth et al., 2008; Chen et al., 2010). If listed companies where used and a positive effect on audit quality was found, it would not have been possible to conclude that this effect was fully due to the implementation of the requirement to disclose the EP’s identity in the audit report. Therefore non-listed companies are used instead of listed companies. Unfortunately this has also a disadvantage. The disadvantage is that the risk of reputational damage for the EP will be smaller regarding non-listed companies. Listed companies are generally followed by more people and have more stakeholders than non-listed companies. Therefore, when the EP fails, this will be noted by far more people than in the case of non-listed companies. The reputation will therefore be more damaged than in the case of non-listed companies. Based on this, it is expected that the effect of the implementation of the requirement to disclose the EP’s identity in the audit report on audit quality will be weaker for EPs who audit non-listed companies than for EPs who audit listed companies. Hence, this will increase the chance that this research will not find a relationship between the implementation of the requirement and audit quality, when this relationship actually does exist for listed companies. However, if a relationship is found for non-listed companies the effect will certainly apply for non-listed companies.

(30)

2.3.2 Sample method

So the period that is sampled is 2004-2005 and the data is gathered from non-listed companies in the Netherlands. The data is collected from financial reports of non-listed companies available on http://www.company.info. A list of ‘Besloten Vennootschappen’ (B.V.) is generated, which are legal entities comparable to Private Limited Companies (Ltd.) in the United Kingdom and Limited Liability Companies (LLC) in the United States. This category of companies is chosen, because B.V.’s are most similar to listed companies (Naamloze Vennootschappen (N.V.’s)), but are not listed. A list is made of the three thousand largest B.V.’s, excluding financial service firms. This last category of firms is excluded because these companies have generally different ratio’s that could significantly influence the results. From this list one hundred companies are selected. One hundred companies are chosen taking into account the available data, the fact that the data is manually collected and the size of the sample that is needed to have a representative sample of the population. The one hundred companies that are selected had to meet some conditions. First of all, in the audit report of 2004 the EP’s identity may not be disclosed and in the audit report of 2005 the EP’s identity has to be disclosed. Second, the financial statements of 2004 and 2005 have to be audited by the same audit firm. This is because previous research has found that the change of an audit firm influences the audit quality (Geiger and Raghunandan, 2002, p. 70). Furthermore, the net sales should be disclosed. The net sales are required for the PMDA-model. Finally, the financial statements had to be prepared in accordance with the Dutch GAAP and not in accordance with IFRS. The reason for this is explained in section 3.2.1. Based on these conditions one hundred companies are selected. From these companies the data is gathered from the financial statements of 2004 and 2005. This is a total of two hundred company years.

First the data for the PMDA-model is scanned for outliers. One extreme outlier is detected. Because of the extreme value of this company, caused by almost

(31)

empty financial statements in 2003, this company is removed from the dataset. A total of 198 years is left. Because of the already small sample the other less extreme outliers are winsorized. These extreme values are substituted by the first less extreme value. Only the extreme outliers are winsorized, because of the small sample and therefore the relatively great effect of winsorizing on the data. The data that is corrected for outliers is filled in the PMDA-model and a linear regression is conducted. The unstandardized residuals are saved as the proxy for earnings management. Second, the data for the EP-disclosure model, including de residual, is scanned for outliers and again only extreme outliers are corrected using the winsorizing method. The final sample is therefore 198 company years.

This chapter described the research design, including the method, proxies, models and sample. In the next chapter the results of the regressions are presented.

(32)

3 Results

The following chapter presents the empirical results from the PMDA-model (equation 1) and the EP-disclosure model (equation 2). First the results of the PMDA-model are shortly described. Second the descriptive statistics of EP-disclosure model are described. Finally the regression results of the EP-disclosure model are presented.

3.1 PMDA-model

In this section the descriptive statistics and results of the PMDA-model are shortly described. Table 1 and 2 present the descriptive statistics. The mean, standard deviation, minimum and maximum are presented from the 5 variables of the PMDA-model in table 1. As shown in table 1, the mean of the total accruals is -0,045 which indicates that the majority of the accruals is income lowering. The means of the other variables are all positive, which indicates that the used earnings management of the sampled companies is mainly income lowering. But, as shown by the minimum and maximum values the variables vary from minus to plus.

Table 1: Descriptive Statistics PMDA-model

Mean Std. Deviation N Min Max Total Accruals -,045 ,179 198 -,967 ,571 1/Assets t-1 ,000 ,000 198 ,000 ,000 (ΔREV- ΔREC) ,191 ,541 198 -1,520 2,470

PPE ,280 ,296 198 ,000 2,931

(33)

Table 2 presents the R, R Square, adjusted R Square and standard error of the estimate of the PMDA-model. The adjusted R2 _{is really close to zero which indicates}

that the model has a weak explanatory power. This makes the model less strong.

Table 2: Model Summary PMDA-model

R R Square Adjusted R Square Std. Error of the Estimate

,144a ,021 ,000 ,179

In table 3 the results of the linear regression are shown. The coefficients, standard error, t-value and significance are presented. It shows that none of the variables is significant. This should be kept in mind. The results of the PMDA-model are weak. This is probably due to the fact that the sample is small, a larger sample could have given better results.

Table 3: Coefficients PMDA-model

Coefficient Std. Error t Sig. (Constant) -0,032 0,025 -1,266 0,207 1/Assets t-1 -27989,161 509053,793 -0,055 0,956 (ΔREV - ΔREC) 0,008 0,024 0,328 0,743 PPE -0,074 0,044 -1,69 0,093 ROA 0,09 0,12 0,753 0,453

With this PMDA-model the unstandardized residuals of every company year are determined. The histogram of these residuals is presented in graph 1. The residual is normally distributed. However, as explained in section 3.1.1, the proxy for earnings management that is used in the EP-disclosure model is the absolute value of the residual. Therefore the unstandardized residual from the PMDA-model is converted

(34)

to absolute values. The distribution of this value is presented in graph 2. The change of this distribution is considered by conducting the regression on the EP-disclosure model in the next chapter. Due to the fact that the distribution is only positive and has actually only one side of a normal distribution a one-sided test is conducted.

Graph 1: Earnings management normally distributed

3.2 EP-disclosure model

3.2.1 Descriptive statistics

The descriptive statistics of the EP-disclosure model are presented in table 4 and 5. Again the mean, standard deviation, minimum and maximum of the variables are

(35)

shown. One should keep in mind that the EM variable is absolute, so it can not be concluded based on the numbers in table 4 that the majority of the companies uses earnings management to increase income. As is presented in the previous section, the majority of the companies actually uses it to lower income.

Graph 2: Absolute value of Earnings management

What can be concluded is that the majority of the financial statements of the companies in the sample is audited by big 4 audit firms. This makes sense, as the selected companies are large B.V.’s who are often audited by big 4 audit firms. A larger company is often more complex, and large companies tend to choose big 4 audit firms (Knechel et al., 2008, 70). However, this is also due to the fact that most of the smaller audit firms already disclosed the name of the EP before 2005 and are therefore not selected. Furthermore the descriptive statistics show a mean debt ratio

(36)

(LEV) around 60%. The mean size is 17,931, which means an average total assets of 61.282.201.

Table 4: Descriptive Statistics EP-disclosure model

Mean Std. Deviation N Min Max EM ,108 ,119 198 ,000 ,594 EPDisc ,500 ,501 198 0 1 LEV ,599 ,229 198 ,071 1,670 Big4 1/0 ,960 ,197 198 0 1 SIZE 17,931 1,236 198 15,859 21,930

In table 5 the R, R square, adjusted R square and standard error of the estimate are presented. The adjusted R square is 1,7% which means that the model only explains 1,7% of earnings management. Again this is a low value, which makes the model less strong.

Table 5: Model Summary EP-disclosure model

R R Square Adjusted R Square Std. Error of the Estimate

,193 ,037 ,017 ,118

3.2.2 Empirical Results

In table 6 and 7 the empirical results of the EP-disclosure model are presented. In table 6 the correlations of the EP-disclosure model (1-tailed) are presented. In table 7 the correlations and one-tailed significances are presented. The significances with an

(37)

asterisk are significant at at least 0,05%. The other correlations are not significant. This means that only the variable LEV is significant. The correlation is ,148 which means that earnings management increases when there is a higher debt ratio. This is in accordance with the expectation based on previous research. The correlation Big4 1/0 and Size is almost significant. The correlation is positive, which underpins the argument in section 3.1 that bigger firms are more frequently audited by big 4 audit firms.

Table 6: Correlations EP-disclosure model

Correlations (Sig. 1-tailed)

EM EPDISC 1/0 DR t B4 1/0 Size EM 1,000 -,053 ,148 ,108 -,004 (0,229) (,019)* (,066) (,480) EPDisc 1/0 -,053 1,000 ,018 ,000 ,038 (,229) (,399) (,500) (,297) LEV ,148 ,018 1,000 -,025 -,039 (,019)* (,399) (,363) (,291) Big4 1/0 ,108 ,000 -,025 1,000 ,092 (,066) (,500) (,363) (,099) SIZE -,004 ,038* -,039 ,092 1,000 (,480) (,297) (,291) (,099)

In table 8 the coefficients of the model are presented. In the column Sig. (1-tailed) the p-value for the one-tailed test is presented. The values with an asterisk are significant at 0,05%. This shows again that only the LEV variable is significant. The Big4 1/0 is close to significant. It is notable that this variable is positive. This means that if the financial statements are audited by a big 4 office earnings management is higher. This is not in accordance with the expectations. However, it should be noted that there are only a few companies audited by a non-big 4 audit firm in the sample, so the sample is not representative regarding the comparison non-big 4 audit firms and big 4 audit firms. Although the coefficient EPDisc 1/0 is nog significant at all, the direction of the coefficient is as expected negative.

(38)

Table 8: Coefficients EP-disclosure model

Coefficient Std. Error t Sig. (2-tailed) Sig. (1-tailed) (Constant) 0,013 0,129 0,097 0,923 0,462 EPDisc 1/0 -0,013 0,017 -0,787 0,432 0,216 LEV 0,079 0,037 2,14 0,034 0,017* Big4 1/0 0,068 0,043 1,578 0,116 0,058 SIZE -0,001 0,007 -0,081 0,936 0,468

Concluding, the above means that the model does not reject the H0 hypothesis and

therefore the model does not support the expectation that the disclosure of the identity of the EP in the audit report increases audit quality.

In this chapter the descriptive statistics and results of the PMDA-model and EP-disclosure model are presented. The conclusion of this chapter is that the results are not significant and therefore we can not conclude that the disclosure of the identity of the EP increases audit quality. In the next chapter these results are evaluated in more detail.

(39)

4 Discussion

In the previous chapter the results of the two models are presented. In this chapter, first, these results are evaluated and the link is made to the research question. Furthermore, some additional notes are presented, containing some additional thoughts of the writer of this paper.

4.1 Evaluation of results

As was concluded in the previous chapter, the results of the EP-disclosure model are not significant. Therefore, it is not possible to conclude that the implementation of the requirement for EPs to disclose their identity in the audit report will increase audit quality.

This conclusion is against the expectations, as it was expected that a positive relationship would be found between EP disclosure and audit quality. One explanation could be the size of the sample. The final sample contains only 198 company years, due to the fact that the data had to be collected manually. But this could mean that the sample is not representative enough to give significant results. It is possible that with a larger sample the results would have been significant. What should be noted, is that if this is the case the implementation of the requirement for EP’s to disclose their identity in the audit report will probably increase audit quality. Refer also to table 8, which shows a negative relationship between EPDisc 1/0 and EM.

But as explained in section 2.3.1 the insignificant coefficient of EPDisc 1/0 can also be explained by something else, namely the fact that the companies that are used are not listed. The risk of reputational damage for the EP is for non-listed companies much lower than for listed companies. An explanation for the found results could be that the audit quality does not increase due to the disclosure of the identity of the EP

(40)

in case of non-listed companies, but this does not mean that this is also the case for listed companies.

Furthermore, based on the comments at the PCAOB-proposal of 2009 there is another explanation. The investors argued that audit quality would increase due to the requirement to disclose the identity of the EP in the audit report because the investors could ask for more skilled EPs (PCAOB, 2011). If this is the case, which is not certain because this is not investigated, audit quality will not directly increase. It will take some time for CPA-offices to increase the skills of the EPs of their office by for example more training or new recruitment requirements. This would mean the effect of the requirement is not directly visible. This explanation could be investigated by conducting the research again, but over more years after the implementation of the requirement.

Finally, it could also be the case that there is in fact not a relationship between the disclosure of the identity of the EP in the audit report and audit quality for both listed and non-listed firms. The reason for this is explained in chapter 2, where the arguments against the implementation of the requirement are described. For example it could be the case that the EPs are already as accountable and responsible as they can be so they can not improve this due to a new requirement. But to be able to conclude this the relationship between EP disclosure and audit quality should be further investigated.

Concluding, based on the results of this paper the research question:

Does the disclosure of the Engagement Partner’s identity in audit report increase audit quality?

can not be answered. But this is only based on the results of this paper. As explained above, there are numerous explanations for the results which would mean the answer to this question c. However, based on this paper there is not enough evidence to proof this.

(41)

4.2 Additional notes

So based on this paper there is no clear answer to the research question. However, considering the described theory and the previous section, one could argue that the name of the EP should be disclosed in the audit report. The arguments of the PCAOB seem reasonable, despite they do not have any evidence for this. In addition, the arguments against the requirement in the comments at the proposal are subjective. The against arguments are mainly from the accounting firms and audit associations, who are unlikely to admit that audit quality can in fact be improved. They also will not admit that they feel not fully accountable for their work. Furthermore, the against argument that investors will focus too much on the name in the audit report and less on the information in the audit report is only an incentive for EPs to make sure they keep a good reputation. As long as the EPs will not damage their reputation, the investors can not extract any information from the name of the EP. In that case, all the EPs are the same, investors are not able to conclude that one audit report has more value than another. If audit firms indeed think that investors will try to take information from the name in the audit report, they will try to keep a good reputation. The requirement also improves the transparency. One could argue that the users of the audit report have the right to know who is responsible for the audit report. Concluding, based on the empirical research there is no evidence for the requirement to disclose the name of the EP in the audit report, but reflected on practice it seems a good requirement.

(42)

Conclusion and Limitations

The PCAOB proposes to implement a requirement for EPs to disclose their identity in audit reports. This requirement is already effective in for example the EU and Australia. The reason the PCAOB is considering this requirement is that they argue that this would increase audit quality. But there is conducted limited research at this subject. Furthermore the research that is conducted is divergent. Also, the comments on the first PCAOB-proposal (2009) are not unambiguous. Therefore this paper investigated the relationship between the disclosure of the EP’s identity in the audit report and audit quality by answering the following question:

The research question is answered by data research in the Netherlands. A comparison is made between audit quality before the implementation of the requirement in the Netherlands (2004) and audit quality after the implementation (2005). A total of 198 company years of non-listed companies is used. Audit quality is measured with the performance matched discretionary accruals model. The disclosure of the EP is measured with a dummy variable. The results show that there is no relationship between the disclosure of the identity of the EP and audit quality. This could mean there is in fact no relationship between the disclosure of the EP’s identity and audit quality. But as this research is conducted with a sample of non-listed companies, it is not certain that this is also the case for non-listed companies as they have a higher reputational damage risk. Furthermore, the results could be due to the small sample size. Another explanation could be that the EP’s need some time to improve their skills after the implementation of the requirement and therefore the effect would not be visible immediately. Concluding, based on this research the research question cannot be answered yet.

As noted before, there are some limitations to this paper. These limitations also lead to some recommendations for further research. First of all, this research is

(43)

conducted in the Netherlands. As is explained in section 2.3.1 it is not possible to conclude only on evidence from the Netherlands about the implementation of the requirement in the USA by the PCAOB. However, this paper can give an indication. In combination with previous research and possible future research it can contribute to the question whether the PCAOB should implement the requirement or not. It also contributes to the general evidence about the relationship between the disclosure of the identity of the EP in the audit report and audit quality.

Furthermore, the sample is limited. The decision to only investigate around one hundred companies is consciously made, however this could be the reason for the insignificant results. It is therefore recommended to investigate the relationship between EP identity disclosure and audit quality in a larger sample. In addition, the sample that is used consists of non-listed companies. This was also a consciously made decision, but as explained in section 2.3.1 it is possible that the effect of the EP disclosure is bigger for listed companies. A recommendation for further research is therefore to investigate the relationship between EP identity disclosure and audit quality with a sample consisting of listed companies. Another recommendation, which is already mentioned in section 5.1 is investigating the relationship between the disclosure of the identity of the EP and audit quality over a longer period. In this way the increase in audit quality due to better training or stricter recruitment conditions is also included. Finally, as explained in section 2.1 King et al (2012) give some suggestions to investigate this relationship by interviewing or surveying EP’s and investors.

(44)

Reference List

Barth, M. E., W.R. Landsman and M.H. Lang (2008), ‘International accounting standards and accounting quality,’ Journal of Accounting Research, 46(3), 467-498.

Basioudis, I.G., E. Papakonstantinou and A.M. Geiger (2008), ‘Audit fees, non-audit fees and auditor going-concern reporting decisions in the United Kingdom,’ ABACUS, 44 (3), 284–309.

Behn, B., J. Choi and T. Kang (2008), ‘Audit quality and properties of analyst earnings forecasts,’ The Accounting Review, 83 (3), 327-760.

Bloomfield, D. and Shackman, J. (2008), ‘Non-audit service fees, auditor characteristics and earnings restatements,’ Managerial Auditing Journal, Vol. 23 Iss: 2 pp 15-141.

Bodnar, G.M., A. de Jong and V. Macrae (2003), ‘The impact of institutional

differences on derivatives usage: a comparative study of US and Dutch firms,’ European Financial Management, 9 (3), 217-297.

Callaghan, J., M. Parkash and R. Singhal (2009), ‘Going-concern audit opinions and the provision of nonaudit services: implications for auditor independence of bankrupt firms,’ Auditing: A Journal of Practice & Theory, 28(1), 153-169. Carcello, J. and C. Li (2013), ‘Costs and benefits of requiring an engagement

partner signature: recent experience in the United Kingdom,’ American Accounting Association, 88(5), 1511-1546.

Carcello, J.V. and A.L. Nagy (2004), ‘Audit firm tenure and fraudulent financial reporting,’ Auditing: A Journal of Practice and Theory, 23 (2), 55-69. Casterella, J.R., K.L. Jensen and W.R. Knechel (2009), ‘Is self-regulated peer

(45)

review effective at signaling audit quality?’ The Accounting Review, 84 (May), 713-735.

Chan, D. and K. Wong (2002), ‘Scope of auditors’ liability, audit quality, and capital investment,’ Review of Accounting Studies, 7 (2), 97-122.

Chang, X., S. Dasgupta and G. Hilary (2009), ‘The effect of auditor quality on financing decisions,’ The Accounting Review, 84 (4), 1085-1117.

Chen, C. Y., C. J. Lin en Y. C. Lin, (2008), ‘Audit partner, audit firm tenure, and discretionary accruals: Does long auditor tenure impair earnings quality?’ Contemporary Accounting Research, 25 (2), 415-445.

Chung, H and S. Kallapur (2003), ‘Client importance, nonaudit services, and abnormal accruals,’ The Accounting Review, 78 (4), 931-955.

Coffee JR., J.C. (2005), ‘A theory of corporate Scandals: why the USA And Europe differ,’ Oxford Review of Economic Policy, 21(2), 198-211.

Davidson, R.A. and D. Neu (1993), ‘A Note on the association between audit firm size and audit quality,’ Contemporary Accounting Research, 9 (2), 479 488.

DeAngelo, L.E. (1981a), ‘Auditor size and audit quality,’ Journal of Accounting and Economics, 3, 183-199.

DeAngelo, L.E. (1981b), ‘Auditor independence, ‘low balling’, and disclosure Regulation,’ Journal of Accounting and Economics, 3, 113-127.

DeAngelo, L.E. (1986), ‘Accounting independence as market valuation substitutes: A study of management buyouts of public stockholders,’ The Accounting Review, 61, 400-420.

Mandatory disclosure of the engagement partner’s identity and audit quality