MASTER THESIS ACCOUNTANCY AND CONTROLLING UNIVERSITY OF GRNONINGEN

(1)

DETERMINANTS ON THE AUDIT FEE AND AUDIT FEE DISCLOSURES

MASTER THESIS

ACCOUNTANCY AND CONTROLLING

UNIVERSITY OF GRNONINGEN

Student | V.S.A. Lachmansingh Student number | s3258114

Pages | 58

Word count (excluding appendices) | 12.568 Date | August 18, 2017

Supervisor |Drs. W. Kevelam RA

Co-reader | Prof. dr. R.L. ter Hoeven RA

(2)

Abstract

This study investigates the determinants influencing the audit gees and the quality of the audit fee disclosure of Dutch listed companies. The requirements that the European Commission issued in 2006 and 2014 constitute the basis for the investigation of these mandates. In particular, the study seeks to determine whether the mandatory requirements enhance audit quality. To this end, a measurement technique employing both qualitative and quantitative features is employed. According to the findings obtained, within the period considered here, the quality of audit fee disclosure is improved, auditors enjoy greater auditors’ independence and audit fee is reduced. The empirical data help the formulation of hypotheses, which are found to possess statistical significance, based on test results. Factors significantly influencing the audit fee include size, complexity, audit risk and auditors’ independence. Additionally, it is noted that audit risk and auditors’ independence impact the audit fee quality. A potential moderating effect is measured based on audit firm tenure, which is concluded to be insignificant for both associations. The annual reports of Dutch listed companies spanning the 2012-2016 period are used to measure the effects, and thus provide input to the current literature on the topic. The integration of qualitative and quantitative characteristics is another important literature input, indicating that, during the financial year 2016, the mandatory rotation of audit firms has a moderating effect.

(5)

1. Introduction

In the Netherlands, a requirement for mandatory audit firm rotation is enforced since 2015 (2014/56/EU). Several articles in the Dutch Financial Times distinguished the issue as a key debate about audit fee and quality, one of which postulates that there is a reduction in audit fee in the first year of auditor change. This turns out to be true, with an approximately 28% reduction in the audit fee in the first year, which attracted greater media attention to audit quality.

It is companies’ obligation to contract an audit firm and undergo auditing. In this regard, their goal is to ensure the maximum audit quality but without paying exorbitant fees. Since 1980, ample studies conducted on the topic of audit fees. The mandatory audit fee disclosure since 2008 (Article 382a of the Dutch Civil Law) emphasizes the great significance of audit fees and audit fee disclosure. This mandatory disclosure enables Dutch companies to see how their audit fees compare with those of other companies, thus helping to enhance competition among audit firms. Parallel to this, several researchers proposed models for explaining the discrepancies between the audit fees of different companies (Rezaei et al. 2016; Gerakos & Syverson 2015).

The European Union first enforced the above-mentioned mandatory disclosure in 2006, and subjected this to revision in 2014. The revision specifies that, besides mandatory disclosure, there must be mandatory rotation of auditors and audit firms as well. This revision effectuated only since 2016, thus warranting the study of mandatory audit fee disclosure and mandatory rotation of auditors and audit firms.

Due to the way in which the mandatory disclosure requirements genders competition, stakeholders strive to attain a balance in the audit fee range. This takes place in view of the way in which overly high payment impacts organizational value, while low payment impacts the audit quality.

The determinants influencing the audit fees and the quality of audit fee disclosure must be identified to be able to examine these aspects effectively and formulate relevant conclusions. This is what the present study aims to achieve; more specifically, the study explores the determinants of audit fee and audit fee disclosure among Dutch listed companies. The knowledge generated from such research will be useful to companies, auditors and stakeholders when attempting to assess the acceptability of the audit fee and the quality of the audit fee disclosure. Nonetheless, there may be a considerable level of subjectivity in the matter and uncertainty may not be altogether dispelled as to the determinants influencing the audit fee and the quality of audit fee disclosure.

(6)

Previous studies established several factors influencing the audit fee which are systematic correlating between each other. For example, the size and complexity of the client company and the size and independence of the audit firm are found to affect audit fee and disclosure quality in a significant way. However, few studies addressed audit fee and disclosure quality jointly, as the present study aims to do.

The fact that this study investigates the ways in which recent changes and developments affect Dutch listed companies is one of its main contributions. In terms of what these changes and developments comprises, they relate to European and Dutch law and regulations.

Data from Dutch listed companies are employed for the purposes of this research. Generalization of both the data and the results at the level of the European Union (EU) is possible because of the way the free transport of goods and the homogenous tax rate for goods imported from outside the EU underpin the economic union. In addition, this is possible because the Netherlands is a EU member state and, furthermore, because the financial reporting standards of Dutch listed companies, namely, the International Financial Reporting Standards (IFRS) are the same throughout the EU.

Among the earliest studies to address audit fees is Simunic (1980), a study which proposed an economic model characterized based on the following influential factors: size, complexity, audit risk, industry, and differentiation between listed and private companies. Given that it sparked extensive research (Francis 1984, 1989; Palmrose 1986; Chan et al. 1993; Langendijk 1997; Groenen et al. 2004; Furguson et al. 2006; Hay et al. 2006; Francis & Yu 2009; Rezaei et al. 2016), Simunic’s economic model is employed in this study as the basis for identification of audit fee determinants.

An important prerequisite for market efficiency is disclosure; indeed, minimization of information asymmetry depends significantly on accounting disclosure (Healy & Palepu 2001). Although many studies are examining the quality of financial report disclosure, the quality of audit fee disclosure is not thoroughly addressed, and therefore this study aims to fill the existing gap in the literature. Recent changes in law and regulations may explain why previous studies not paid too much attention to this topic. On the other hand, there is ample research on the correlation between disclosure quality and organizational governance, with emphasis on the discrepancies among listed and private companies (Lang & Lungholm 2000; Khlif & Souissi 2010; Reeb et al. 2013). The

(7)

aspects that earlier studies focused on, have little relevance for this study, which employs data from Dutch listed companies.

Operationalization of disclosure quality is the main issue distinguished in earlier studies (Dechow & Dichev 2002; Botosan 2004; Daske & Gebhardt, 2006). These studies embody various associations, since quality is a subjective attribute that comprises different preferences. Different variables related to financial reports are employed in different studies to measure disclosure quality, including size, timeliness, risk and restatements (Barth et al. 2001; Cohen et al. 2004; Nichols & Wahlen 2004; Barth et al. 2008; Jiang et al. 2011). Such variables constitute a suitable starting point to identify the determinants that impact on audit fee disclosure quality. Furthermore, to evaluate disclosure quality thoroughly and to minimize preference bias, the present study addresses qualitative features as well. Established in the Dutch Financial Reporting Framework (RJ 290), these features are rated on the basis of a disclosure index. Both disclosure quality in general and audit fee disclosure quality in particular can be measured more efficiently through this approach in conjunction with variables from earlier studies. This study formulated an index comprising twelve items and which is based on the foundation of the basic qualitative attributes delineated in RJ 290. Hence, the attributes can be examined both collectively and individually.

The goal of this study is to produce empirical evidence about the determinants of audit fee level and audit fee disclosure quality of Dutch listed companies. The study stands apart from other studies because, in addition to the mixture of qualitative and quantitative features, it also considers the potential moderating effect of mandatory audit firm rotation.

The research question that this study seeks to answer is: “To what extent do size, complexity, audit risk and independence influence the audit fee and the audit fee disclosure quality of Dutch listed companies?”

The study possesses relevance because it addresses a gap in the related literature through the developed measurement instrument rooted in both quantitative and qualitative features. The results obtained can form the basis for the formulation of hypotheses and predictions in future studies. The study is structured in the following way: the theoretical framework, including law and regulations, theoretical foundations and hypothesis formulation are addressed in Chapter 2; the research model and data collection are presented in Chapter 3; the results obtained from the data analysis are summarized and discussed in Chapter 4; concluding remarks, as well as study limitations and suggestions for further research, are provided in Chapter 5.

(8)

2. Theoretical Framework and hypotheses formulation

2.1 LAW AND REGULATIONS

Since this study concerns Dutch listed companies, an overview is provided in the following part of the Dutch law and regulations pertaining to the topic as well as current developments.

2.1.1 Mandatory disclosure of Audit Fees

Due to the many accounting scandals that occurred in previous times, auditors faced criticism from the European Commission for failing to produce accurate and valid audits. The European Commission preceded its criticism with the publication of a directive for regulation of auditors’ activities, stipulating prerequisites for auditors’ independence and supplying a monitoring tool for national authorities.

Issued on May 17th, 2006, Directive 2006/43/EC outlines the mandatory auditing of annual reports and consolidated annual reports with regard to the quality framework of the EU. It addresses the auditors’ independence in relation to the company and seeks to avoid auditors becoming financially dependent. Financial dependence is understood as direct or indirect financial interest arising from the balance between the auditor’s assurance and non-assurance services. The Directive further mentions that companies are not obliged to report these services in their annual report. To make the auditor-company relationship more transparent, amendments are brought to Directive 78/660/EEC and Directive 83/349/EEC that stipulated that both assurance and non-assurance audit fees of the audit firm needs to be included in the annual report and the consolidated annual report of companies.

In terms of enabling stakeholders to assess auditors’ independence, the Dutch legislature and the Royal Netherlands Institute of Chartered Accountants (NBA) both advocate the relevance of audit fee reporting.

Hence, a new section was added to the Dutch Civil Code on June 27, 2008 (BW2T9 382a), specifying that it is mandatory for companies to indicate the fees of professional auditing in their annual report. This new section derives from the enforcement of Directive 2006/43/EC of the European Community (EC).

(9)

The issued requirements delineated in Article 382a (BW2 title 9) of the Dutch Civil Code are: 1. The total amount audit fee must be presented in the financial statements separately for each

financial year in the following categories:

a. Audit fee charged for auditing the annual report; b. Audit fee charged for other assurance services; c. Fees charged for tax advice services; and d. Fees charged for non-assurance services.

2. If the company financial statements contain consolidated financial statements the fees as noted under 1 should also presented for the subsidiaries;

3. The company can make use of an exemption to present the audit fees if the fees are included in consolidated financial statements at a higher level.

The second subsection of Article 382a of the Dutch Civil Code specifies the exemption that if a company is part of a consolidation, then it does not necessarily need to publish audit fees in its annual report. Therefore, disclosure of the audit fees can be done in the consolidated annual report. The third and final subsection of Article 382a addresses the issue of audit fee disclosure, stipulating that implementation of the IFRS eliminates the necessity of publishing audit fees.

Audit firm is narrowly defined in the first subsection of Article 382a, and the costs of the operations of such a firm appear restricted to country level. Under a narrow definition of the law, disclosure of the fees of associated tax advisors, associated advisory agencies and foreign audit firms integrated in the organization is not mandatory. However, none of these entities constitute audit firms according to the narrow definition of audit firm, the application of which affords three options for disclosing the audit fee:

1. Audit firm level: Approach to disclose the audit fees for the (Dutch) audit services.

2. Organizational audit firm level: Approach to disclose the audit fees for all the audit services within the firms’ network.

3. Organizational firm level: Approach to disclose the fees for all services within the firms’ network included related companies and services.

(10)

Furthermore, the NBA uses the narrow definition of Article 382a to draw the conclusion that disclosure of audit fees and other professional fees can be done in a number of ways. Thus, the NBA proposed an approach to the implementation of Article 382a and developed a model for Dutch auditors, as illustrated in Figure 1.

Figure 1: NBA’s disclosure model

According to the NBA, the model above can aid stakeholders to evaluate/assess the independence of an auditor. Hence, the NBA denounces the narrow definition of audit firm and believes that its model is more informative for stakeholders and ensures greater transparency.

As previously indicated, earlier studies have limited their focus to the factors influencing audit fees, they do not address the issue of mandatory audit fee disclosure. Recent literature such as Cairney and Young (2006), Hay et al. (2006), Carson and Fargher (2007), Cahan et al. (2008), Francis and Yu (2009), Arens et al. (2013), Bills et al. (2014) and MohammadRezaei et al. (2016) are all derived from the earlier research of Simunic (1980) who developed a pricing model based on several hypotheses, within this research the expected relations were proven.

2.1.2 Mandatory rotation of audit firm

The European Commission initiated an investigation into the matter of audit quality in response to, on the one hand, stakeholder criticism regarding this matter in the aftermath of the financial crisis, and on the other hand, previous scandals. On October 13th 2010, the EU Commissioner, Michel Barnier, published the “Green Paper” entitled “Audit Policy: Lessons from the Crisis”. The purpose of this paper is to promote discussion about the role audit firms played in the financial reporting of Public Interest Entities (PIEs). With respect to the interaction between audit firms and companies,

2008

X Auditors Other X Network Total X network

Audit of financial statements 0 0 0

Other assurance services 0 0 0

Tax advisory services 0 0 0

Other non-assurance services 0 0 0

Total 0 0 0

2007

X Auditors Other X Network Total X network

Audit of financial statements 0 0 0

Other assurance services 0 0 0

Tax advisory services 0 0 0

Other non-assurance services 0 0 0

(11)

Barnier distinguishes the major issues to be auditors’ independence and audit firm tenure. Linking these issues to the 2008 financial crisis, Barnier proposes potential approaches to solving the issues, such as forbidding audit firms from combining assurance and non-assurance services and mandatory audit firm rotation to eradicate the oligopoly of four major audit firms dominating the audit market.

In agreement with Barnier, the European Commission acknowledged that audit quality was significantly susceptible to the threat of familiarity and in 2011 put forth the requirement of mandatory audit firm rotation every six years (or eight years in exceptional circumstances). Moreover, the European Commission also put forward a “cooling-off” interval preceding re-engagement of the same audit firm.

Following the proposal in 2011, on September 2nd_{, 2014, directive 2006/43/EC is amended by} directive 2014/56/EU. This amendment consists the elaboration on the mandatory rotation of auditors and audit firms. The directive is effective as per June 17, 2016, and applies from that date. The reason why the initial directive is revised was to protect audit quality and to improve audit quality, with this achieved as a result of the way in which it ensured auditors’ independence and objectivity with respect to the companies. Regarding the issues of independence and objectivity, the amendment of the 2014 Directive is clearer and more straightforward than the 2006 Directive. The correlation between audit quality and the issues of auditors’ independence and objectivity directly reflected the previous criticisms of audit quality.

Directive 2014/56/EU did not have a significant impact on Dutch law and regulations. Around four months following Barnier’s 2010 Green Paper, Ronald Plasterk, a member of the House of Representatives, submitted a proposal entitled “De accountancy na de crisis” to the government on February 15th 2011. Drawing on the Green Paper, this proposal emphasised similar protective measures of mandatory rotation and interdiction of non-assurance services. In October 2011, the proposal was integrated into a new government bill and is effectuated before Directive 2014/56/EU, on January 1st_2014.

Most existing studies focused on the correlation between auditor switching and changes, while the matter of auditor changes, received much less attention. Nonetheless, the auditor switching literature contains several reasons for companies to change from auditors. Continually prior research regarding audit tenure seems relevant and therefore where it is contributing to this research it is taken into consideration.

(12)

Chow and Rice (1982) find that firms tend to switch auditors after receiving a qualified opinion, but they also find that firms that switched auditors (upon receipt of a qualified opinion) compared to other firms which received a qualified opinion who did not switch are not more likely to receive an unqualified opinion the following year. Changing audit firm is not a solution because it does not necessarily result in an unqualified opinion in the future.

Francis (1984) explored a hypothesis about price reduction in Australia based on sample of 26 companies that switched auditors during the period 1974-1978. Receiving a qualified opinion and auditor switching is found to be significantly correlated, but the price reduction hypothesis is invalidated.

Based on first-year audits and associated fees, Butterworth and Houghton (1995) examined the extent to which auditor switching determined a downward trend in audit services (low balling/price cutting). The researchers observed that in the year when auditor switching occurred, there was a marked decrease in audit fees, although trends such as low balling or price cutting were not apparent.

Arrunada and Ares (1997) examined the effect of mandatory audit firm rotation on audit fees, observing that in comparison to the preceding auditor, a first-year auditor led to 15-25% increase in incremental costs.

Carcello and Nagy (2004) investigated in their research association between audit firm tenure and audit quality. Audit quality was operationalized within this research as fraudulent financial reporting. The results of this research are conflicting with the point of view of the European Commission and therefore also with directive 2014. The results are showing more errors in reporting within the first three years of the audit firm tenure, where there are no results found of an increase in errors with a prolonged audit firm tenure. Based on this research and the contradictory evidence the mandatory rotation of audit firm can be argued.

Stanley et al. (2007) scrutinized Carcello and Nagy’s (2004) work, with the exception that they addressed audit quality in terms of restatements within financial reporting (10-K filings). Their findings were consistent with that of the previous study, concluding that the longer the audit firm tenure was, the lower the likelihood of restatements.

(13)

Chen et al. (2008) examined the correlation between audit firm tenure and earnings quality at both the audit firm and audit partner level. Neither audit firm nor audit partner tenure was found to adversely affect earnings quality of the audited company.

Lim and Tan (2010) find a negative association between audit firm tenure and audit quality. Their results indicate that an increase in familiarity results in a decrease of auditors’ professional skepticism and that a newly appointed audit firm more likely finds deficiencies within the business processes.

Finally, Ewelt Knauer et al. (2013) find also a negative association between audit firm tenure and audit quality. Their results indicate that an increase in audit firm tenure is leading to a routine audit approach by the auditor resulting in not revealing and noting potential frauds. Furthermore, their results indicate a declining independence from the auditor to the client as the audit firm tenure increases.

It is clear from the above that the regulatory authorities and earlier studies put forward different viewpoints and therefore the findings of the present study are bound to be noteworthy.

2.2 THEORETICAL FOUNDATIONS

The theoretical foundations on which this study is based are presented in the following part. As previously mentioned, this study aims to determine the impact of recent law and regulation development about mandatory audit fee disclosure and mandatory audit firm rotation, to identify the determinants influencing audit fees and audit fee quality disclosure.

2.2.1 Audit quality, agency theory, and moral hazard

The changes in law and regulations are outlined in section 2.1. Critical views about audit quality (understood as auditors’ independence and objectivity in relation to the company) prompted the EC to introduce Directive 2006/43/EC and Directive 2014/56/EU with the purpose of improving audit quality within the EU.

DeAngelo (1981) is a seminal research on audit quality. In this study, audit quality was understood to be reflected in how professionally competent auditors were and the level of auditors’ independence to the company. A narrower definition of audit quality is difficult to achieve due to its inherent complexity. Nonetheless, in wider terms, audit quality can be generically described as auditors’ ability to detect and disclose errors in the client company’s accounting system (DeAngelo 1981; Watts & Zimmerman 1983).

(14)

The commonly used theory of the principal-agent relationship underpins the agency theory. This theory impacts auditors in two ways, namely, the weight attributed to auditors’ report and opinion and the independence afforded auditors, which may be the source of a conflict of interest among auditors, companies and stakeholders. According to the agency theory, auditors’ opinion is a means for communication with stakeholders. However, the companies in their position of auditor contractors can stifle this opinion along with auditors’ independence. Earlier studies addressed this potential problem, which continues to be an important element of audit quality.

Moral hazard may arise as the agent and the principal, who in this case are managers and shareholders, respectively, do not have access to the same information and therefore the agent may prioritize personal interests over those of the principal. According to several authors, agency costs increase in direct proportion with the magnitude of agency conflicts, leading to greater demands for audits of high, or perceived high, quality (Palmrose 1984; Francis & Wilson 1988; DeFond 1992; Craswell et al. 1995; Johnson & Lys 1990).

Francis (2004) conducted a comprehensive review of earlier studies on audit quality. The audit quality results were placed on a scale from very low to very high. Audit quality scaled as very low was understood as lack of auditor conformance to regulatory requirements (GAAP) or failure to publish precise auditing reports.

Within this study an overview is constructed, based on the examined researches, compelling with the variables which are influencing the audit quality. Mentioned variables are size, risk, stakeholders’ response to going concern reports and the fees of the auditor regarding other services. These variables are also conducted within this study. Relevance with audit quality, the agency theory, moral hazard and adverse selection with this study is found and related to the examination of the audit fee and the quality of the audit fee disclosure.

2.2.2 Accounting disclosures and information asymmetry

Agency issues and information asymmetry restrict the optimal allocation of value in efficient capital markets (Beyer et al., 2010). Information asymmetry can be minimized through accounting disclosure, which ensures that information is safely and homogenously disseminated among shareholders and managers (Healy & Pelepu 2001).

Mishkin (2006) defined information asymmetry as the situation in which a party makes an incorrect transaction decision due to not having precise information about the other party to the transaction.

(15)

Mishkin (2006) further noted that there was an unwillingness among stakeholders to pay for quality information and, in this way, minimize information asymmetry. The primary reason for this was that stakeholders exploited the information that other parties accounted for.

Revenue maximization is the ultimate goal of shareholders. Managers who possess the necessary motivation and managerial competence and experience can help achieve this goal and thus enhance company profitability. However, it frequently happens that managers’ interests are incompatible with those of shareholders and take advantage of information asymmetry to pursue those interests. Indeed numerous economic studies indicated that managers are motivated to enhance the company’s net income if they also stand to gain (Healy 1985).

Shareholders can derive information about their company from a number of sources, the main one of which is the financial statements. Indeed, the disclosure of financial statements is intended to provide accurate and clear financial information to the stakeholders so that the latter can assess the company’s capability of attaining its goals and how likely it is for them to recover their investments. Shareholders will make as much use as possible of the company’s valid information, potentially leading to a decrease in the required return.

Investigating the implications of mandatory disclosures, Christensen et al. (2013) put forth three justifications to legitimize such disclosures. The results revealed the existence of externalities, cost reductions at the level of the whole economy and tight sanctions. The purpose of all justifications is to ensure the success of mandatory disclosure when managed adequately (Larosiere et al. 2009). The present study seeks to achieve this as well.

Summarizing, capital market efficiency can be enhanced through minimization of information asymmetry via mandatory disclosure. However, this is only possible if mandatory disclosure is managed appropriately. This agrees with the revised 2014 EC Directive and is the focus of the present study. Achievement of the research aim should be possible as the mandatory disclosure requirement is clearly stated in the revised directive alongside with the duties of the companies, auditors and overseeing bodies. Measurement and interpretation of the potential effect of this requirement are undertaken in this study.

(16)

2.3 DEVELOPMENT OF HYPOTHESIS

The hypothesis associated with the independent variables and their conjectured correlation with audit fees and audit quality is formulated in the following part on the basis of a review of the relevant results of earlier studies.

2.3.1 Audit fee

The audit fee represents the amount of money an audit firm charges to produce an opinion as to the accuracy of a company’ financial statements regarding the financial situation of that company. Simunic (1980) specified that the number of subjects intended for auditing and the managers’ expectations about the audit process determine the audit fee. For the purposes of the present study, the audit fee is derived from the annual reports.

2.3.2 Size

Numerous studies indicated that company size has a considerable impact on audit fee (Simunic 1980; Francis 1984; Palmrose 1986; Chan et al. 1993; Langendijk 1997; Groenen et al. 2004; Ferguson et al. 2006; Hay et al. 2006; Rezaei et al. 2016). The audit risk usually depends on client size and risk minimization is time- and effort-consuming from the perspective of auditing. In earlier studies, company size operationalization was undertaken based on total assets. However, since the use of total assets presents a number of problems, Chan et al. (1993) suggests that company size can be more accurately characterized in terms of net revenue. Furthermore, owing to presentation and/or selection of financial reporting standards, net assets may indicate different sizes for companies that are similar (Chan et al. 1993). An additional implication distinguished in this study is off-balance sheet assets and financing. Previous studies addressed both determinants, and therefore the present study will take into account total assets and net revenue, since the net revenue is not employed so far in any study for size operationalization in the case of Dutch companies. Therefore, this study stands to make a valuable and relevant contribution to the current literature. H1: The amount of revenue is of positive associated to the amount of audit fee.

(17)

Several studies identified a positive correlation between company size and the goal to enhance audit quality (Myers et al. 2003; Chen et al. 2008; Khlif & Souissi 2010). Meanwhile, company size and financial reporting accuracy are also found to be positively correlated in another study (Carcello & Nagy 2004). Thus, it is clear that company size and financial reporting quality are connected. Hence, the quality of audit fee disclosure is assessed in this study based on company size.

H5: The amount of revenue is of positive associated to the audit fee disclosure quality.

2.3.3 Complexity

Company complexity and audit fee are reported to be positively correlated in a number of studies (Simunic 1980; Francis 1984; Wallace 1984; Simon 1985; Palmrose 1986; Chan et al. 1993; Langendijk 1997; Groenen et al. 2004; Ferguson et al. 2002, 2003, 2006; Hay et al. 2006). Indeed, Arens et al. (2013) even argued that the phenomenon whereby auditors allocate time and resources to understand their clients increases in a directly proportional way with respect to the complexity of the client.

The nature of the company is also observed to be linked to company complexity (Hay et al. 2006), and Langendijk (1997) and Groenen et al. (2004) corroborated this in their work conducted in the Netherlands. In these studies, complexity is operationalized to include the number of subsidiaries, with a number of other international studies highlighting the correlation between the two.

Including the number of subsidiaries among the determinants of audit fees has been legitimized in the related literature, with data from interviews with audit partners (Chan et al. 1993) providing evidence of the existence of a positive correlation between the two (Simunic 1980; Francis 1984; Francis et al. 1986; Langendijk 1997; Groenen et al. 2004; Hay et al. 2006). The gathered data point to the fact that audit partners automatically consider a greater amount of resources in the case of companies that have subsidiaries. The audit partners justify this on the grounds of the variability of financial reports, the publication of which is subjected to compliance with different statutory and professional requirements.

There are some additional instances of elevated audit fee due to subsidiaries, apart from the above-mentioned variability of statutory and professional requirements. If different auditors undertake the auditing of the subsidiaries, additional supervision and audit procedures are necessary, particularly with regard to transactions between companies and mandatory disclosure.

(18)

2.3.4 Audit risk

Audit risk is among the recognized determinants of audit fee (Simunic 1980; Francis 1984; Simon 1985; Chan et al. 1993; Ferguson et al. 2003; Groenen et al. 2004; Hay et al. 2006). Therefore, a heightened audit risk results in an increase in the audit fee, which takes the form of a premium intended to attenuate the audit risk.

According to the Dutch Audit Standards (COS 200 paragraph 23), material misstatement of financial statements could mean that auditors’ opinion is unqualified. Eimers (2006) stressed the significance of mentioned risk, which is challenged upon the uncovering of financial scandals (e.g. Enron, Ahold, etc.). In the aftermath of such scandals, answers are demanded from the auditors about the efficiency of the audit and its procedures and the stakeholders may even put forward a claim and hold the auditors to account to some extent.

Another dimension of audit risk aside from the general financial risk is the reputation risk. Kleinsmit et al. (2003) comprehensively examined this type of risk, and it is established to be of importance for determining audit risks.

Different studies undertakes the operationalization of audit risk factors with different financial ratios:

1. Current-ratio (Francis (1984) and Langendijk (1997);

2. Debt-ratio (Francis (1984) and Ferguson et al (2002, 2003 and 2006); and 3. Quick-ratio (Francis (1984) and Langendijk (1994).

After examining the work of Langendijk (1994), Groenen et al. (2004) concluded that solvability and liquidity as independent variables were not correlated in the context of the Dutch market. Nonetheless, in a subsequent study addressing audit fees, Langendijk (1997) reported the same correlation.

(19)

Current ratio

The current ratio is widely used in academic literature (Francis (1984) e.g.) to measure audit risk. According to Craswell et al. (2002), a low current ratio enhances the risk that the auditor will not be paid for the services provided. Besides the remuneration, the risk regarding the independence threat increases together with a simultaneous increase of the audit fee to safeguard certain threats. Levitan et al. (1985) argues that a high current ratio may determine auditors to provide an unqualified opinion.

H3a: Audit risk (proxy current ratio) is positive associated to the amount of audit fee.

Dechow (1994) and Khimich and Sloan (2011) reported a positive correlation between the quality of reported earnings and financial report disclosure quality. As previously mentioned, the current ratio is among the control variables employed for the assessment of reported earnings quality. According to the findings of earlier studies the current ratio seems to influence the audit fee disclosure quality; the positive association expected will be measured in this research.

H6a: Audit risk (proxy current ratio) is positive associated to the audit fee disclosure quality.

Debt ratio

The financial leverage ratio indicative of the extent to which debt is used to fund a company’s assets (i.e. the level of a company’s conformance to long-term debt covenants) is known as the debt ratio. Several researchers (e.g. Francis 1984) uses this ratio in the calculation of the audit fee. The correlation between number of debts and loss risk within the company provides an explanation for the theoretical perspective. This risk augments the audit risk. Furthermore, Francis (1984) argues, in the presence of a high debt ratio, the financial risk impacts the audit due to a negative interpretation. Furthermore, in contrast to a lower debt ratio, mentioned interpretation enhances susceptibility to financial risk. In a study conducted in the US, Schultz et al. (1987) debt ratio auditor perception in relation to the audit risk.

Kamp (2006) suggests that solvency is also an influential factor in the specific context of the Netherlands. Indeed, audit fee and debt ratio can likely be positively correlated. Employing the debt/equity ratio, Ferguson et al. (2003, 2006) and Carson et al. (2007) adopted an identical approach.

(20)

Several studies identified a positive correlation between the quality of reported earnings and the disclosure quality of the financial report (Krishnan 2003; Kothari et al. 2005). The debt ratio is among the control variables applies for the measurement of the quality of reported earnings. According to the findings of earlier studies, it is concluded that the debt ratio affects the quality of audit fee disclosure. This study will evaluate the conjectured positive correlation between the two. H6b: Audit risk (proxy debt ratio) is positive associated to the audit fee disclosure quality.

Quick ratio

The quick ratio reflects a company’s ability to meet its short-term debts. To assess financial risk, Francis (1984) employed this ratio alongside the debt ratio. Meanwhile, Schults et al. (1978) reported that financial risk and audit risk were correlated based on their investigation of auditor perception regarding the financial risk associated with the audit risk. Although the current ratio is relevant for the purposes of the present study, assessment of current assets (primarily inventory) is not the same for all companies and the existing variability causes discrepancies in the gathered data. Therefore, this study uses the quick ratio in keeping with earlier international studies (Ferguson et al. 2006; Carson et al. 2007).

H3c: Audit risk (proxy quick-ratio) is negative associated to the amount of audit fee.

H6c: Audit risk (proxy quick-ratio) is negative associated to the audit fee disclosure quality.

2.3.5 Independence

The correlation between auditors’ independence and audit fee is the focus of several studies (Simunic 1984; Groenen et al. 2004; Hay et al. 2006; Francis 2006; Patel et al. 2009). Since it includes the assurance fees that auditor’s bill, the current study broadens the scope, thereby determining how the ratio of non-assurance fees to audit fees in earlier research impacts auditors’ independence.

(21)

From the 1970s onwards, extensive criticism and debate are directed at non-assurance services that supplement the auditors’ traditional services. The US government detected the problems and took action during the late 1970s (Francis et al. 1979). On the basis of the audit fee and the fees for additional services, the government concluded that complete auditors’ independence was not possible. The results of KPMG’s internal and external audit of Rentokil led to the same conclusion in the Netherlands. The scandals associated with Enron/WorldCom prompted the US to outlaw (prohibit) the integration of non-assurance services and traditional auditing services in 2002. In the Netherlands, this occurred seven years later.

Francis (2006) argues that two reasons legitimize the way in which non-assurance services undermine auditors’ independence and hike audit fees. The first reason is that stakeholders can behave more critically in response to auditor provision of non-assurance services due to the potential occurrence of conflict of interests in the audit firm. The second reason is that the auditor can become financially dependent on the company since the auditor is the one billing the total fee to the company. Meanwhile, in a different study, Moore et al. (2006) discovered that moral enticement and auditor provision of non-assurance services are positively correlated.

Simunic (1984) proposes the low-balling effect as an additional determinant of auditors’ independence. Simunic (1984) concludes that other non-assurance services primarily covered the auditor costs. The competitive market affords an explanation for this – to make up for losses from assurance services, the auditor provides assurance services at reduced cost and non-assurance services for higher cost. Hammond et al. (2009) obtained similar results.

In contrast, Groenen et al. (2004) in the Netherlands identified a positive correlation between audit fee and the proportion of non-assurance services. Such a positive correlation may be due to the extra protective measures required when a mixture of assurance and non-assurance services are provided. However, the study of Groenen et al. (2004) presents a shortcoming in that a control variable for advisory services is employed. In the present study, the non-assurance services scaled on an ordinal base will be assessed to determine whether the audit fee and non-assurance services are correlated.

H4: Independence of the auditor (proxy amount of non-assurance services) is positive associated to the amount of audit fee.

(22)

According to several studies, non-assurance services can have a negative effect on auditors’ independence (Chung & Kallapur 2003; Frankel et al. 2002; Beattie & Fearnley 2002; Markelevich & Rosner 2013). When they highlighted that non-assurance services are basically commercial in nature, several regulators confirmed this idea. Owing to the potential implications of this aspect, the present study will adopt it as an independent variable in the investigation of the correlation between auditors’ independence and audit fee disclosure.

H7: Independence of the auditor (proxy amount of non-assurance services) is positive associated to the audit fee disclosure quality.

2.3.6 Controlling variables

In the examination of the above-mentioned factors, it is likely that additional determinants of audit fee and audit quality may be identified. Therefore, this study will address such combined effects and/or correlations. Among the potential further determinants is whether a company engages the audit services of a big-four audit firm or not. Taking into account such additional determinants will ensure that the results obtained will be reliable and valid. The controlling variables that will be assessed in this study are company auditing conducted with the aid of the big-four audit firms and company reports addressing loss or profitability.

The usage of these controlling variables is also consistent with earlier studies (e.g. DeAngelo 1981; Francis & Yu 2007; Chan et al. 2008) with regard to the controlling variable employed for audit firms. In the earlier studies, auditor size, audit fee and audit quality are observed to be significantly correlated. Operationalization of this correlation for measurement based on an ordinal scale will be undertaken in the present study in keeping with the literature (Simunic 1980; Francis 1984; Palmrose 1986; Groenen et al. 2004; Francis & Yu 2007).

Simunic (1980) revealed that reported losses and audit fee were positively correlated; therefore, loss or profit measurement is undertaken in the present study based on the controlling variable. Meanwhile, Frankel et al. (2002) and Lim and Tan (2007) discovered a similar correlation with audit quality.

2.3.7 Moderating variable

As outlined in paragraph 2.1.2, the impact of audit firm rotation on the audit fee has been the focus of several studies (e.g. Chow & Rice 1982; Francis 1984; Butterworth & Houghton 1995; Arrunada & Ares 1997). However, due to the fact that these studies did not observe any effects of price cutting

(23)

or low balling, their findings lack consistency. For example, Arrunada and Ares (1997) pointed towards the fact that the audit fee increased in the first year of appointment of a new auditor. Furthermore, another limitation of earlier studies is that they did not take into account either the 2006 Directive (2006/43/EC) or the revised 2014 Directive (2014/56/EU). Since the present study conducts a measurement of the potential moderating effects that the audit fee is exposed to, in line with one of the research objectives of addressing mandatory audit firm rotation, it will depart from the existing literature.

Audit firm tenure and audit quality were reported, have often been identified as negatively correlated (Lim & Tan 2010; Knauer et al. 2013). However, the literature does not corroborate this result, as demonstrated in Arrunada and Ares (1997), Carcello and Nagy (2004), Stanley et al. (2007) and Chen et al. (2008). Hence, there is still a lack of clarity regarding the nature of the impact of audit firm rotation on audit quality. Aside from not being rooted in the revised 2014 Directive, the earlier studies did not address the situation in relation to Dutch listed companies. Since the present study conducts a measurement of the potential moderating effects that the audit fee is exposed to, in line with one of the research objectives of addressing mandatory audit firm rotation, it will depart from the existing literature.

The developed hypotheses and the operationalization of these hypotheses are systematically reflected in a conceptual model as illustrated in figure 2.

(24)

Figure 2: Illustration of the 2 conceptual models

Model 1: LnFee = α + β1 LnRevenue + β2 Subs + β3 LnCurr + β4 Debt + β5 LnQui + β6 LnNAS + β7 Loss + β8 Big 4 + β9 LnModerator + ε

Model 2: Disc.Qual = α + β1 LnRevenue + β2 LnCurr + β3 Debt + β4 LnQui + β5 LnNAS + β6 Big 4 + β7 LnModerator + ε

Audit fees

Size (H1)

Complexity (H2)

Auditrisk (H3a, H3b, H3c)

Independence (H4)

Dummy Variables - Reported Loss - Audit by Big4 Moderator effect Change of auditor Disclosure Quality

Size (H5)

Auditrisk (H6a, H6b, H6c)

Independence (H7)

Dummy Variables - Reported Loss - Audit by Big4 Moderator effect Change of auditor

(25)

Above conceptual models are translated in to the following OLS analysis: Where:

LnFee = Natural logarithm of audit fee; LnRevenue = Natural logarithm of revenue; Subs = Amount subsidiaries;

LnCurr = Natural logarithm of Current ratio (current assets scaled by total assets); Debt = Debt ratio (long term debt scaled by total assets);

LnQui = Natural logarithm of Quick ratio (current assets minus inventory scaled by current liabilities) ;

LnNAS = Natural logarithm of non-assurance services; Loss = Control variable, 1 for loss;

Big 4 = Control variable, 1 for audited by Big 4 audit firm;

LnModerator = Natural logarithm of change of auditor, 1 for change in the year; Disc.Qual = Disclosure quality score of audit fee disclosure in percentage

(26)

3. Research Design

The independent variables and the statistical measurements that will be carried out as a way of assessing the formulated hypotheses are outlined and discussed in the following part. In addition, the data and sample employed for the purposes of this study are described as well.

3.1. RESEARCH METHOD

Most of the existing studies have addressed aspects related to audit fee in relation to American and British companies, but the present study differs in this regard since it focuses on Dutch listed companies. This study also stands out because it is the first to investigate the correlation between the quality of audit fee disclosure and the audit fee. In addition, both of these two aspects are examined with regard to their response to the moderating effects of mandatory audit firm rotation. To assess the developed hypotheses, this study employs a quantitative research approach in both of the constructed models, in keeping with earlier studies. However, in addition to a quantitative approach, this study also adopts a qualitative approach in order to determine the quality of audit fee disclosure. The use of a qualitative approach may have a negative impact on external validity and potential preferences, but the likelihood of this happening in the present study is low due to the fact that the regulator indicated the use of qualitative measurement. A comprehensive description is provided regarding the model devised to permit qualitative assessment.

3.2 DATA COLLECTION AND SAMPLE

An empirical quantitative analysis is carried out with the purpose of evaluating the validity of the hypotheses formulated in Chapter 3 and shown in Figure 2. Operationalization of this analysis is achieved with the help of Ordinary Least Squares (OLS) regression analysis. In order to attain the established research objectives, the necessary data are derived from Dutch listed companies headquartered in the Netherlands. Furthermore, aside from the AEX, which is the most extensive stock exchange market, companies within the AMX and AScX markets are assessed as well. The period of time over which data from company annual reports are derived is December 31st, 2012 – December 31st, 2016. The methods of data collection employed are search of the Orbis database and manual extraction from the financial statements.

(27)

3.3 MEASUREMENT

3.3.1 Measurement of independent (descriptive) variables

A number of independent variables are derived from earlier studies with the purpose of assessing the audit fee and the quality of audit fee disclosure. These independent variables are size, complexity, audit risk, and auditors’ independence.

Earlier studies pointed towards company size as one of the variables with the greatest significance. Its measurement is based on a comparison between the amount of company assets and the audit fee.

Several studies highlighted revenue as being the variable of the greatest importance with regard to audit fee and audit quality (Simunic 1980; DeAngelo 1981; Palmrose 1986; Langendijk 1997; Groenen et al. 2004; Francis & Yu 2007). These studies argued that the increase in audit fee and disclosure quality was directly proportional to the company size. The present study seeks to expand existing knowledge as a result of investigating the impact of revenue, which has not been examined before. To this end, an approach similar to that used in earlier studies in relation to company size is employed. The investigation of revenue is further warranted in view of the way in which it has greater reliability as it can be more easily compared among different companies. Moreover, in keeping with Chan et al. (1993) a natural logarithm is employed to achieve the statistical transformation of revenue to make it more linear with the audit fee and audit quality.

A number of international studies reported that the audit fee correlates with on company complexity, among other factors. Furthermore, the number of subsidiaries is established as a suitable variable for the measurement of company complexity (Simunic 1980; Francis et al. 1986; Langendijk 1997; Groenen et al. 2004; Francis & Yu 2007). Indeed, it is found that the increase in the audit fee was directly proportional to the increase in the number of subsidiaries. Given these considerations, the number of subsidiaries is employed in the present study to measure company complexity, especially as no other studies on Dutch companies, employed this variable.

Another variable that previous studies identified as a determinant of audit fee and audit quality is the audit risk of a company. The auditor perceives as a premium on the audit fee the connection that is likely to exist between the audit risk and the potential (public) exposure of the auditor. No other study conducted in the context of the Netherlands considered the audit risk in relation to the audit fee and audit quality. Meanwhile, with regard to current ratio, Francis (1984) indicated it to be positively correlated with the audit fee, which is consistent with the forecast in the present study.

(28)

More specifically, Francis (1984) reported existence of indirect proportionality between current ratio and audit fee, the latter increasing as the former declined. Furthermore, in accordance with Ferguson et al. (2006) and Carson et al. (2007), the present study undertakes the measurement of debt ratio based on division of the long-term debt by the total assets. Previous studies reported the existence of direct proportionality between debt ratio and audit fee. As concerns the quick ratio, the study adopts the approach presented in Francis (1984), measuring the quick ratio based on division of current assets after the current assets took out the inventory. Previous research established that the audit fee increases with the decline in quick ratio. Analysis of the three ratios (current ratio, debt ratio and quick ratio) enables operationalization of the audit risk, in line with earlier studies.

The results obtained in earlier studies in relation to the measurement of the current ratio and the debt ratio revealed that the reported earnings and the disclosure quality are positively correlated (Dechow 1994; Khimich & Sloan 2011; Krishnan 2003; Kothari et al. 2005).

As discussed in Chapter 2, the EC, the Dutch legislative authorities as well as supervisory entities in both Europe and the Netherlands identified auditors’ independence as a variable of audit quality. Operationalization of auditors’ independence is undertaken in the form of imposition of restrictions on non-assurance services, in accordance with a number of previous studies (e.g. Simunic 1984; Groenen et al. 2004; Hay et al. 2006; Francis 2006; Patel et al. 2009; Chung & Kallapur 2003; Frankel et al. 2002; Beattie & Fearnley 2002; Markelevich & Rosner 2013). Furthermore, in line with Groenen et al. (2004), the natural algorithm of non-assurance service fees is employed for the measurement of auditors’ independence. Moreover, a natural logarithm is used to achieve the statistical transformation of the non-assurance service fee (Chan et al. 1993) to improve its linearity in relation to the audit fee and the audit quality.

3.3.2 Measurement of dependent variables

The audit fee and the quality of audit fee disclosure are the two dependent variables that are employed in the present study.

Audit fee

This study obtains the audit fee from the disclosed audited annual reports of the Dutch listed companies under investigation. Owing to the nature of their sources, the audit fee data can be considered to be valid and reliable. The directive pertaining to the mandatory audit fee disclosure

(29)

that applies to Dutch listed companies is implemented on 27th June, 2008. Groenen et al. (2004) employed this variable as well. Likewise, a natural logarithm is used to transform the variable in keeping with the approach adopted in earlier studies (e.g. Langendijk 1997).

Quality of audit fee disclosure

As outlined in paragraphs 2.1.2 and 2.2.1 audit quality and the quality of financial reporting is researched in previous academic literature. In contrast, the present study is concerned primarily with the quality of audit fee disclosure. Earlier studies focusing on financial reporting quality employed a range of different measurement techniques to evaluate the audit fee disclosure quality. Hirst et al. (2004), Nicholas and Wahlen (2004) and Cohen et al. (2004) employed the development of a disclosure index on the basis of qualitative and quantitative items as their technique. Value relevance models and accrual models are additional techniques that are employed (e.g. Healy & Wahlen 1999; Barth et al. 2001), but they are not useful for the purposes of the present study because they are concerned with financial information pertaining to stock market reaction and do not cover all the aspects that are discerned. Therefore, the disclosure score index designed for the measurement of the qualitative features which the Dutch regulatory outlined is the optimal technique for the assessment of the audit fee disclosure quality in the present study.

Of the twelve features that are discerned, six are compulsory features that the Dutch regulator specified, three are suggestions that the regulator extended, and the remaining three are qualitative features intended to make disclosure more useful for decision-making.

For each of the twelve features associated with audit fee disclosure, a score of one point can be given. The totality of the points yields a composed and developed scale, with 1 denoting that the audit fee disclosure quality is low and 12 denoting that disclosure is of high quality.

3.3.3 Measurement of moderating variable

The present study undertakes the measurement of the moderating effect to make up for the unreliable findings of earlier studies regarding the possible impact of mandatory audit firm rotation on the audit fee and audit quality, as well as the discrepancies between those findings and the revised Directive that the EC published in 2014. To this end, subtraction of the audit firm is undertaken on the basis of the audited annual reports, including auditors’ opinion. Furthermore, in the year of change, audit firm switch will be recorded as 1. The potential effects of the moderating variable is interpreted on the basis of the results.

(30)

4. Results

4.1 GENERAL

The information used within this research conducts 75 Dutch listed companies covering the period from 2012 until 2016. The data collected is mainly manual subtracted from the annual reports of the companies. Some variables such as the ratios and the amount of subsidiaries are subtracted from the Orbis database and are verified with the annual reports. The data can be assumed as representative since all Dutch listed companies are included.

Table 4.1a describes the partition of the research population for the period from 2012 until 2016.

Table 4.1a: Partition of research population N 2012 63 2013 63 2014 65 2015 64 2016 71 Total 326 % 100

The collected data has been imported in SPSS24 for further analysis. With Casewise Diagnostics there are 49 observations removed from the data since these samples were not compelling all the required data for the regression analysis. The removed observations were as mentioned not compelling all required data, mainly resulting from incorporation during the period of 2012 till 2014. The removed observations are limited compared to the total sample and therefore the remaining sample can be assumed as representative for Dutch listed firms.

4.2 ROBUSTNESS CHECKS

In order to use the results from the regression to ensure the validity of the research outcome, there is described within this research if the research population is complying with certain assumptions.

(31)

Normality:

In order to have derivable analysis and statistics, the research population should be normally distributed. This means from a visualization perspective the graph should have a bell-shaped curve. The statistics which are taken into consideration are kurtosis and skewness. In order to conclude on the normality distribution of the variables from a statistic perspective, the Kolmogorov-Smirnov test is conducted.

Based on the skewness of the variables Fee, Non-assurance services fee, Revenue, Quick ratio and Current ratio (described in table 4.2b) these variables are transformed using a natural logarithm. Using a natural logarithm for these variables the skewness is decreased to an acceptable level and is normally distributed, the frequency statistics (P-P plot histograms) of the transformed variables are described in appendix 2. The kurtosis of the research population is deemed as irrelevant since the data is subtracted from audited annual reports and therefore can be used as reliable data.

Homoscedasticity

Homoscedasticity connects to the dependable relations between the variables. Homoscedasticity refers to the assumption dependable variables are equally spread over the predicted values of the independent variables. If the residuals are not equally spread there is a strong suggestion of heteroscedasticity.

The homoscedasticity is assessed by plotting the regression standardized residuals against the regression standardized predicted values and inspect whether the residuals are equally spread over the predicted values of the dependent variable. The scatterplot is shown below. This is visualized with a scatterplot (appendix 3). Based on the scatterplots it is visualized the residuals are equally spread.

Linearity

In order to perform an accurate regression analysis it is essential there exists a linear relationship between the dependable and the independable variables. The linearity is visualized with scatterplots (appendix 3). Based on the scatterplots the linearity is ordered, however, in the second regression model within this research, there seems less linearity is assumed compared to the first model. The evaluation of the regression results can be made with a certain precaution.

(32)

4.3 DESCRIPTIVE STATISTICS

The analysis of the research population has been described in table 4.3b and the analysis for each year has been described in appendix 4.

Table 4.3b: Analysis research population

Variable N Mean Minimum Maximum Std. Deviation Skewness Kurtosis

Fee 326 5.477,98 98,00 68.000,00 10.249,18 3,24 11,59 Disc.Qual 326 70,48 25,00 100,00 20,39 -0,23 -0,94 Loss 326 0,19 0,00 1,00 0,40 1,57 0,46 Big4 326 0,98 0,00 1,00 0,14 -7,094 48,64 NAS 326 668,65 0,00 10.000,00 1.133,82 3,98 24,39 Revenue 326 13.189.564 319,00 467.153.000 52.536.822 7,06 52,41 Subs 326 198,76 0,00 1.681,00 311,91 2,61 7,59 Curr 326 2,37 0,04 76,70 5,99 8,66 88,79 Quick 326 2,09 0,04 76,70 6,02 8,67 88,64 Debt 326 51,95 3,05 99,86 24,33 0,59 -0,23 Moderator 326 0,13 0,00 1,00 0,33 2,26 3,14

The variables are in accordance with the research of Groenen et al. (2004). Fee = Audit fee in EUR x 1.000;

Disc.Quality = Percentage of score on disclosure index; Loss = Control variable, 1 for loss;

Big4 = Control variable, 1 for audited by Big 4 auditor (KPMG, PwC, E&Y, DTT); NAS = Non-assurance services fee in EUR x 1.000;

Revenue = Revenue in EUR x 1.000; Subs = Amount of subsidiaries;

Curr = Current ratio (current assets scaled by total assets);

Qui = Quick ratio (current assets minus inventory scaled by current liabilities); Debt = Debt ratio (long term debt scaled by total assets);

Moderator = Change of auditor, 1 for change in year of observation.

Table 4.3b describes the mean of the audit fee is EUR 5.477.980, Groenen et al. (2004) reported in their research a mean of EUR 725.000. Furthermore, the mean of the subsidiaries is significantly lower compared to the mean reported in the research of Groenen et al. (2004). These differences can be explained since in the research of Groenen et al. (2004) not only Dutch listed companies were included in the research population. These results also differ compared to the research of Groenen et al. (2004) since the law and regulations have been significantly changed in the period 2012 till 2016 compared to 1998.

As described in table 4.3b from the 326 observations 98 percent are audited by Big four audit firms. 6 observations within the population are not audited by a Big four audit firm. The high percentage of this variable within the research population is showing the dominance of the big four audit firms

(33)

in the market of the Dutch listed companies. The big four audit firm variable has been used as a control variable within this research, considering there are no significant variances within the population regarding this control variable this variable has not been further analyzed within the research.

From table 4.3b the mean of the disclosure quality can be concluded that throughout all the observations, 70% is the average score on the disclosure quality index. This score seems quite low given the fact the mandatory disclosure is required since 2008 and it can be expected, the disclosure quality of the audit fees, is higher than 70% based on this research population (2012-2016). In order to investigate this further the development of this variable will be included in the analysis for each year (appendix 4).

In table 4.3b the mean of the moderator is showing a percentage of 13%. There can be concluded a change of audit firm rotation of all observations with an average of 13%, this percentage seems reasonable since the mandatory audit firm rotation is effective as per 2016.

Appendix 4 describes the descriptive statistics of the research population for each year (2012 till 2016). Within the table, there are some developments over the year’s notable and interesting prior to performing the regression analysis.

Variable Fee is reflecting a decreasing tendency from 2012 until 2016 resulting in a decrease in the mean of EUR 380.980 over a 6 year period. Variable NAS is also reflecting a significant decreasing tendency in the mean resulting in a decrease of EUR 462.000 (56%) over a 6 years period. Together with the increasing tendency of the disclosure quality, it seems the changes in law and regulation (elaborated in chapter two of this research) is opposing certain tendencies and results. Above results are reflected in a chart to illustrate (figure 3) the described tendencies derived from appendix 4.

(34)

Figure 3: Tendencies of audit fee and non-assurance fee within researched population (amount in EUR x 1.000)

In appendix 4 the variable disclosure quality is reflecting an increase over the 5 years period. The increase of the years can be expected and is reasonable, however the increase of 5% within the years can be concluded as low. As the mandatory requirement is effective since 2008 there could be expected in the first years of adoption a lower percentage. The research population ends with the year 2016 and is 8 years after effectuation of the mandatory disclosure, where the supervising bodies together with the stakeholders could expect a 100% score on the constructed index.

Lastly, in appendix 4 the moderating variable is showing an interesting development. In the years 2012 and 2013 there is reflected a 2% percentage of change in audit firms within the observations. This indicates that almost none of the Dutch listed companies switched from audit form in these years. Analyzing further, the years 2014 and 2015 are showing a significant increase in mention percentage of 14% to 16%. The variable is even showing a percentage of 26% in the year 2016. This indicates that more than half of the Dutch listed companies switched from audit firm in the period 2014 till 2016. The observed change seems reasonable since the mandatory requirement was published in 2014 and is effectuated as per 2016. The proportion can be expected since for the remaining part of the population the maximum audit firm tenure is not achieved. Above results are reflected in a chart to illustrate (figure 4) the described tendencies derived from appendix 4.

0 1000 2000 3000 4000 5000 6000 2012 2013 2014 2015 2016

Development Audit fee and Non-assurance fee

(35)

Figure 4: Tendencies of disclosure quality and moderator (audit firm rotation) within researched population (in percentages)

Multicollinearity of the variables is analyzed based on the Pearson & Spearman correlation matrix, which is described in table 4.3c In sequence, the variance inflation factors are the variables analyzed. Based on the descriptives in table 4.3c the correlation values of the LnCurr and LnQui variables may indicate multicollinearity. Both variables conduct a result above 0,7 (Field, 2013). However, when eliminating one of these variables or both from the regression analysis, the explanatory value (adjusted R square) is decreasing. Based on this finding we can conclude there is no indication for possible multicollinearity.

0 10 20 30 40 50 60 70 80 2012 2013 2014 2015 2016