The impact of legislation on audit market services in the Netherlands

(1)

The impact of legislation on audit market services in the

Netherlands

Name: R.H. Wijermars

Student number: 1793756

MSc: Accountancy

Address: Almastraat 36, 9716 CX Groningen Telephone: 06-57634525

Email: r.h.wijermars@student.rug.nl

Supervisors: drs. W. Kevelam RA, prof. dr. R.L. ter Hoeven RA Word count: 11.763

(2)

Abstract

Langendijk (1997) has previously studied the audit services market in the Netherlands. The current study expands on his earlier work and focuses on audit fee and disclosure quality research in the Netherlands. Insights into determinants are revealed, including the effect of mandatory audit firm rotation in response to the Auditors Profession Act of 2013. 319-year observations from annual reports 2011 to 2016 are examined where (mandatory) audit firm rotation has taken place. Increased attention for financial reporting followed various corporate governance- and financial scandals early 2000s. Since 2008, disclosure of audit fees has been compulsory in the Netherlands. The quality of audit fee disclosure is measured by a disclosure index. The quality of audit fee disclosure has increased gradually over the years, showing improvements in financial reporting. Total assets and risk have shown to be significant predictors of audit fees, supporting previous literature. This study contributes to the understanding of auditing and accounting in an increasingly interconnected global economy

(3)

3

1. Introduction

Business operations and financial performance is documented in annual reports. There is information asymmetry as managers and shareholders have contradictory interests. Auditors derive their existence by reducing this information asymmetry between organizations and people reviewing company’s financial statements. This shows that auditors are important in ensuring quality of financial reporting. As auditor’s act in the interest of the shareholder, the auditor’s compensation should also be disclosed in the notes of the annual report.

Previous lax laws and regulations have caused a wave of corporate governance and globally financial scandals in the early 2000s. These scandals have called more attention to financial reporting and transparency. Established in 2003 as a response to these scandals, the code ‘Tabaksblat’ aimed to improve transparency in the financial statements, increase accountability to the supervisory board and strengthen and protect the position of shareholders. Developed in 2006, Council Directive 2006/43/EG art. 49 1a, b recommends providing an overview of total fees charged by audit firms for each fiscal year. It also proposes making a distinction between fees for auditing financial statements, other assurance services, tax advisory services and other non-audit services. The directive aims to improve the comparability of financial statements in the EU, repair public confidence in financial statements and improve the functionality of financial markets by providing more consistent and specific information. This directive was incorporated into the Dutch Civil Code when provision 2:382a was introduced in June 2008. After this change, it became mandatory for large Dutch legal entities to disclose auditor fees in the explanatory notes of their financial statements.

According to Dieleman (2008, as cited in Langendijk, 2011) legislators had a clear focus when introducing provision 2:382a. He explains that audit fee notes in financial statements must address the independence of external auditors. Additionally, he assumes that the independence of external auditors is under pressure when other services are highly compensated in comparison with audits of financial statements; assuming that this leads to a lower quality of the audit of the financial statements. Langendijk (2011) illustrates that the independence of auditors can also be threatened by services provided through network organizations when

(6)

network organizations provide various other services; as self-interest is a threat to independence. This finding is also supported by Parmalat. Mandatory disclosure of audit fees in the firms’ financial statements provides an opportunity to explain the size and composition of the audit fees, which allows to assess auditor independence.

To promote financial stability and reform market supervision, the European Commission launched the Green Paper in 2010 to begin a debate about the roles of external auditors, governance, the independence of audit firms, supervision of external auditors and the structure of the audit market. The paper was also intended to create an internal European market for audit services (European Commission, 2010, as cited in Klijnsmit & Majoor 201, p. 160). Subsequently, this discussion led to proposals for new European regulations (European Commission, 2011) and a revision of the Council Directive from 2006 (European Commission, 2011b). The propositions were presented by European commissioner Barnier and aimed at restoring trust in audits. Barnier (European Commission, 2011c) sought to strengthen independence and professional sceptism by introducing a strict separation between audit and advisory services, mandatory audit firm rotation and more transparency measures.

Historically, the quality of audit fee disclosures has been questionable. According to Breij (2009, p. 30), this questionable disclosure quality is partly due to the vague description included in provision 2:282a of the Civil Code. The article’s legal text led regulators, executives, auditors and reviewers of financial statements to raise questions about the scope of the legal text. Parties were mainly confused about which parts of audit organizations were required to adhere to the article. Langendijk (2011) found that numerous combinations of the four components of auditor fees have been found, leaving people reviewing company’s financial statements to guess whether missing components led to costs or not. Additionally, comparative figures were not always included, a requirement stated in provision 2:363 of the Civil Code, subsection 5. Provision 2:382a is regarded as important at quotation procedures; it is expected to lead to increased price competition among audit firms. Consequently, this increased competition could impair the quality of service, an unintended consequence of increased transparency (Langendijk, 2011).

(7)

7

improve disclosure beyond the limited scope of provision 2:382a, but these recommendations received little attention (Breij, 2009). They provided a best practice example that audit firms could follow (van Offeren, Verdoes & Witjes, 2010). These recommendations are now incorporated in the Directive of Financial Reporting (RJ 390.3).

Another portion of this study encompasses mandatory audit firm rotation (MAFR). As result of the Auditors Profession Act in 2013, audit firms are obliged to mandatory rotate after auditing a public-interest entity (PIE) for ten consecutive years, with a cooling down period of four years (Autoriteit Financiële Markten, 2013). Mandatory audit firm rotation began on January 1st, 2016. Only in Italy MAFR has been implemented prior to the change in the EU Directive, which allows this study to analyse the intended benefits of MAFR. In the relationship between audit firms and client, incentives can arise that threaten auditors’ independence (Quadackers, L.M., 2012); rotation can help to address this problem. On the other hand, rotation can cause firms to (temporarily) under price-audits, a practice commonly referred to as lowballing (Ettredge & Greenberg, 1990). Longer-term relationships with clients may be necessary to offset discounted fees.

Because the Netherlands recently implemented MAFR, little is known about the consequences of the requirement. Since audit fee determinants differ between countries (Taylor & Simon, 1999), more research is needed to assess independence and contracting issues concerning the audit process (e.g. lowballing). Previous audit fee research has mainly focused on the US, UK and Australian markets (Hay, Knechel & Wong, 2006; Hay, 2013). The obligation to disclose audit fees in the explanatory notes is new for the Netherlands compared to these countries (Langendijk, 2011). To date, only four studies have examined the level of audit fee disclosures in the Netherlands (Breij, 2009, 2010; van Offeren, Verdoes & Witjes, 2010; Langendijk, 2010). Only one study has examined determinants of audit fees in the Netherlands (Langendijk, 1997). Also, previous research has focused on satisfaction concerning cost control of audit fees in the Netherlands (Groenen, Kerkhof & Langendijk, 2002). Therefore, it is important to examine the determinants of audit fees and the current quality of fee disclosures in the Netherlands in depth. Insight into the Dutch audit services market will expand understandings of accounting and auditing in the increasingly interconnected global

(8)

economy. As there isn’t much research on how these laws have impacted audit fees and disclosures this study aims to answer the following question:

‘What variables affect audit fees and the quality of fee disclosures in the Netherlands?’

The current study fills this gap and contributes to existing research in several ways. The current study examines the quality of disclosure of audit fee notes in the annual reports of Dutch listed firms for the years 2012 to 2016. In addition, the study expands on the determinants of audit fees in the Dutch setting, including the effect of mandatory audit firm rotation. This study uses a disclosure index based on the Directive of Financial Reporting 390.3 (see Appendix C), which allows the current study to measure the extent to which companies follow best practices. Consequently, this measurement can examine the quality of audit fee disclosure. This examination will provide insight in whether 2:382a has led to an increase in disclosure quality over time; and to which extent the audit fees are distributed among the four components. Additionally, as this study provides a five-year cycle listed firms will be obliged to rotate. This obligation provides insight in the entire period of audit firm rotation, making it possible to generalize or compare the effects of mandatory audit firm rotation in an EU setting.

This paper is divided into seven chapters, further divided into additional subsections. Following the introduction in the first chapter, the theoretical framework will be discussed in the second chapter of in which ‘hypothesis development’ accounts for a separate third chapter. In the fourth chapter, the research design will be discussed. The fifth chapter will show the results followed by the discussion of the results and conclusion in the sixth chapter. The seventh chapter will conclude the study by discussing limitations and provide avenues for future research.

(9)

9

2. Theoretical framework

This chapter first discusses some main topics: how auditors facilitate disclosure, voluntary- and mandatory disclosure. The underlying theories will be discussed in these topics. This is followed by a paragraph about EU reform as this impacts the audit services market in the Netherlands. Further, some general topics regarding the audit market services in the Netherland will be discussed followed by the theory & hypothesis development.

2.1 How auditors facilitate disclosure

As briefly mentioned in the introduction, auditors are required to audit the financial statements as result of information asymmetry. Originated by Jensen and Meckling (1976), agency theory concerns resolving problems that arise between principals (shareholders) and agents (managers). These problems exist as agents have more information than principals and do not always act in the principals’ interest. This behaviour is evident as the ‘demand for financial reporting and disclosure arises from information asymmetry and agency conflicts between managers and outside investors’ (Healy & Palepu, 2001; p. 406). Agents (i.e. auditors) are hired to act on behalf of principals (i.e. shareholders). Auditors monitor agents’ behaviour, reducing information asymmetry and agency conflicts.

Financial reporting interest groups, including auditors, seek to reconcile societal demand and corporate actions (Chalmers & Godfrey, 2004). Auditors exist to enforce regulatory compliance and hold companies accountable to society (Chalmers & Godfrey, 2004). They are an independent third party and agent of society. Figure 1 from Healy and Palepu (2001) provides insight into how disclosure interrelates with information and financial intermediaries (such as auditors) in the capital market.

(10)

However, Watts and Zimmerman (1981) argue that auditors are imperfect intermediaries because they have conflicting incentives. The main conflicting incentive is that auditors act in the interest of managers who hire them rather than the interest of firms’ investors. In addition, corporate governance and financial scandals have led to doubts about external auditors’ independence. Mandatory audit firm rotation should contribute to restoring perceptions of auditors and the usefulness of financial reports.

According to O’Sullivan and Diacon (2002), the extent to which an audit firm provides non-audit services depends on the likelihood of agency problems. Companies with agency problems seek to avoid perceptions of impaired audit independence, leading to a reduction in non-audit services provided by auditors. On the other hand, Van Schaik (2003b) argues that the distribution of audit fees should not affect independence because auditors should be interested in the total value of fees and not the ratios between different categories.

2.2 Voluntary disclosure

In this study, disclosure is defined as ‘any deliberate release of financial (and non-financial) information, whether numerical or qualitative, required or voluntary, or via formal or informal channels’ (Gibbins, Richardson & Waterhous, 1990, p. 122, as cited in Hooghiemstra, 2012). Voluntary disclosure theory focuses on the intention of company management to provide information.

Spence’s (1973) signalling theory explains behaviour in the labour market and can also help to explain voluntary disclosure. The signalling theory is based on the agency theory, as signalling is a reaction to information asymmetry. Hummel and Schlick (2016) state that firms with high-quality disclosures are incentivized to signal their high performance and ‘distinguish themselves from lower-quality firms through voluntary disclosure’ (Watson, Shrives & Marston, 2002; p. 291). According to legitimacy theory, poorly performing firms prefer low-quality disclosures as a way of concealing their true performance and protecting their legitimacy (Hummel & Schlick, 2016).

Einhorn and Ziv (2008) studied how past disclosure behaviour affected subsequent disclosures. They found that firms that had previously disclosed information voluntarily tended

(11)

11

to provide comparable disclosures in subsequent financial statements. Disclosing information informs the market of a firm’s presence and implicitly commits the firm to future discloses. Furthermore, Chalmers and Godfrey (2004) argues that little disclosure signals the company as being uninformed and that therefore poor performing firms elect not to disclose in the first place. Chalmers and Godfrey (2004) state that reviewers of financial statements consider reporting reputation when assessing financial statements (Trueman, 1986). The researchers provide supporting evidence, stating that a firm’s disclosure history affects decisions about whether and what to disclose. From the previous paragraphs it is evident that not all firm’s voluntary disclose all information. However, there are external pressures to disclose.

2.3 Mandatory disclosure

Institutional theory concerns how institutional rules, norms and laws affect organizations. Complying with these informational demands is important to firms’ credibility and reputations. Organizations achieve legitimacy by complying with these rules, norms and laws (Deegan, 2006; Meyer & Rowan, 1977). Financial reporting standards have evolved as result of changes in societal norms and institutional pressure. Mandatory disclosure plays a significant role because the content of financial reporting is largely determined by laws and regulations. Regulating corporate disclosures can solve agency problems in capital markets (Healy & Palepu, 2001). Beyer (2010) argues that disclosure regulation is needed to address the misalignment of insider (management) and outsider incentives, which complicates the sharing of credible information. NIVRA established voluntary disclosure requirements, and NOvAA’s 2003 independence provision was amended with statutory provision 2:382a in 2008. Dimaggio and Powell (1983) have inferred that firms act in response to both formal and informal pressure. Formal pressure takes the form of the laws and regulations, and informal pressure is defined as pressure from public opinion. In the institutional environment, provision 2:382a requires companies to be more transparent. Public explosure in the media has also placed pressure on companies. The level of non-audit services threatening (perceived) independence is an item that received public exposure (van Offeren et al., 2010). Disclosure requirements such as provision 2:382a, accounting standards and auditors’ enforcement actions require firms to adhere to a

(12)

certain level of disclosure and enhance credibility. Chalmers and Godfrey (2004) mention requirements from professional bodies, such as European Parliament and the European Council in 2006/43/EG art. 49 1a, b, as well as community values and peer practices that are associated with reporting practices.

Disclosure is also linked to firms’ reputations for financial reporting, as corporate governance and financial scandals have increased societal pressure for transparency (Chalmers & Godfrey, 2004). Consequently, the researchers found a positive association between reputational damage resulting from non-compliance and the extent of a company’s response to demands for information from legal, professional and community groups. Furthermore, Chalmers and Godfrey (2004) state that audit firms also have reputational concerns and should comply with professional requirements (e.g. the accountant’s oath) and best practices. To maintain or build their reputation audit firms are expected to influence their clients’ disclosures.

2.4 Reform European Union legislation

European Union reforms of the statutory audit market were adopted in June 2014. All laws were applied in June 2016, apart from MAFR, which is subject to transition arrangements (European Commission, 2016). Among the requirements for statutory audit firms relating to PIE audit clients is the declaration ‘that prohibited non-audit services were not provided and that the audit firm(s) remained independent of the audited entity in conducting the audit’ (Pwc.com, publication EU audit reform). When lack of disclosure is evident in financial statements, the statutory auditor must report to authorities supervising PIEs, provide the audit firm’s total turnover and classify turnover into the categories of statutory audit and non-audit services and the entities to which they relate. When exceptional circumstances affect MAFR, an audited entity may file a request to the national regulator or supervisory authorities for a two-year extension of a ten-year term (Pwc.com, publication EU audit reform). The next paragraph will further discuss the independency issue.

(13)

13

2.5 Audit market services in the Netherlands

2.5.1 independence

Despite being mandatory, not all companies fully disclose audit fee information. Audit fee notes in the financial statements allow readers to examine the services provided and asses the independence of external auditors (van Offeren et al., 2010) (See Appendix A1,A2 and A3 for audit fees and audit fee disclosure in the Netherlands). Auditor independence is assumed when auditor’s focus on audit quality instead on their financial interests, thereby possibly impairing audit quality. Therefore, the independence of auditors is considered to be a proxy for audit quality (Francis, 2004). Offeren et al. (2010) state that the ratio of audit fee auditing the financial statement relative to the total audit fee is seen as an indication of (in)dependence of the auditor. Thereby assuming that a relatively low audit fee for auditing the financial statements could indicate a threat to independence. Additionally, van Schaik (2003b) discusses the importance of disclosing audit fees. One of the independence prescriptions1 implemented on January 1st, 2003 recommends disclosing audit fees. The prescription aims to better inform reviewers of the financial statements regarding the independence of auditors and assumes a threat to independence when other services are highly compensated relative to statutory audits. On the other hand, van Schaik (2003b) states that the distribution of audit fees should not affect independence because auditors should be interested in the total value of fees and not the ratios between different categories. Schaik (2003b) advocates caution when drawing conclusions about auditors’ independence based on audit fee disclosures. He states that companies and audit firms lack guidance from NIVRA and NOvAA for classifying audit costs, resulting in various activities that are applicable to more than one category and causing inconsistent disclosures of audit fees.

(14)

2.5.2 Comparability audit fee

Langendijk conducted several studies of audit services on the Dutch market (e.g. in 1994 and 1997). As a result of the introduction of provision 2:382a, information about audit fees is freely available, making it possible to conduct comparable studies in an EU setting, provided that financial statements offer information with a similar format. This similarity is still not the case, as can be inferred from Langendijk’s research on financial statements (as described in van Offeren et al., 2010) as well as Breij (2009, p. 30) and van Offeren et al.’s (2010) research stating that the form of statements of segmented accounting costs differ. Most companies present a clear table, but some provide only an inaccessible descriptive summary.

3. Hypothesis development

3.1 Determinants of audit fees

Many studies have investigated audit cost functions, product differentiation (e.g. Van Offeren et al., 2010; O’Sullivan & Diacon, 2002), pricing and competition in the audit industry (e.g. Anderson & Zéghal, 1994). Most of these studies have focused on the UK, US and Australia (Langendijk, 1997). Hay’s (2013) meta-analysis along with that of Hay, Knechel and Wong (2006), provides an extensive overview of audit fee research from 1977 to 2007. From this overview can be inferred that widely used determinants of audit fees encompass number of assets (size), complexity (mainly covering subsidiaries), risk and leverage ratio. Other less-examined determinants include auditor specialization, auditor change and regulation. The current study examines the relation of size, industry specialization (by the auditor), risk and auditor firm rotation for the audit market services in the Netherlands. The following paragraphs will separately discuss each chosen determinant.

(15)

15

3.1.1 Industry specialization

Industry specialization is an important topic of study for various reasons. Industry specialization (in the form of product differentiation) provides for client needs (Dunn & Mayhew, 2004; Mayhew & Wilkins, 2003) price differentiation for passing on economies of scale to the client (Ettredge & Greenberg, 1990) or differing quality (Watts & Zimmerman, 1986). It is in auditors’ interest to specialize in specific industries, as this could allow them to set a premium on audit fees (Craswell et al., 1995; Mayhew & Wilkins, 2003; Numan & Willekens, 2012). Finally, it is interesting to research industry specialization because product differentiation is not similar between firms or countries; research on the Netherlands is limited.

Langendijk (1997) analysed whether the six biggest audit firms in the Netherlands earned an audit fee premium; assuming that market share was related to industry specialization. He investigated audit fee premiums hoping to find ‘evidence of product differentiation in the market for audit services’ (Langendijk, 1997, p. 254). He expected an audit fee premium for the six biggest audit firms as they are very large international firms; which have invested in reputation capital indicating that they would ‘be able to obtain a return on this investment by imposing higher prices of their services’ (Simon, 1995, p. 357 as cited by Langendijk, 1997). However, he finds no fee premium for the six biggest audit firms as a group; only KPMG earning an audit fee premium.

Mayhew and Wilkins (2003) state that audit firms with large market shares can spread industry-specific training costs over more clients, producing economies of scale that are not easily duplicated by small market share firms. Palmrose (1986) as well as Ettredge and Greenberg (1990) provide evidence that auditors can perform audits less expensively when they develop industry specialization through an enlarged customer base. Hence, it is argued that these market share driven cost efficiencies can be (partially) passed along to clients.

Dunn and Mayhew (2004) found that being able to differentiate on other characteristics than price allows audit firms to gain market share as well as retain existing clients. Firms are incentivized to successfully develop specialization and meet client needs in a manner that competitors cannot easily reproduce. Mayhew and Wilkins (2003) state that firms with large market shares can develop more industry-specific knowledge and expertise, enabling them to

(16)

provide higher-quality services than firms with small market shares. Differentiation in audit quality can also be explained in terms of agency costs (Watts & Zimmerman, 1986; Craswell et al., 1995). Craswell et al., (1995) explain product differentiation by showing that companies being audited voluntarily bond with higher-priced industry specialists and prefer their services to other licensed auditors. This finding suggests that auditor quality differentiated by industry expertise is in demand.

Willenborg (2002) concluded that opposing effects passed on through economies of scale and premiums requested in exchange for audit quality make it difficult to predict how audit fees correspond with industry specialization. Numan and Willekens (2012) examined the effects of competition by studying how audit firms that specialized in certain industries set prices. They found that audit fees increased when auditors’ industry specializations are aligned with client preferences. The authors concluded that price premiums are realized due to auditor-client alignment grounded in industry knowledge resulting from specialization.

Similarly, Mayhew and Wilkins (2003) state that possessing a larger industry market share and offering higher service quality allows audit firms to charge fee premiums. The authors used Porter’s five forces model to examine how industry specialists could differentiate themselves from competitors through market share. They hypothesized that having a substantially higher market share than competition would lead to increased bargaining power. Subsequently, this bargaining power would result in clients being unable to obtain comparable high-quality service from competitors. Under these conditions, audit firms can realize an audit fee premium for differentiated services. When an audit firm is not in the position to differentiate itself, bargaining power is reduced, and increased pressure from competition causes clients to participate in cost saving measures driven by market share. According to Casterella, Francis, Lewis and Walker (2004), larger fee premiums tend to be offset for larger companies, while smaller ones are likely to have little bargaining power. Therefore, the size of the company could indicate the extent to which clients are willing to pay for specialized audit services.

Furthermore, Francis, Stokes and Anderson (1999) argue that market leadership is an unambiguously signal that enables auditors to set higher prices. Craswell et al., (1995) argue that industry specialization by the eight largest firms is costly and increases audit fees. They

(17)

17

found that among public listed Australian firms, the eight largest industry specialists earned a significant premium (34%) compared to non-specialist of the eight biggest auditors. They argue that the required return on investment in any industry dominates potential cost savings; this audit fee premium is passed on to clients and leads to higher audit fees.

No standard method exists to determine whether an audit firm is specialized. According to Krishnan (2003), industry specialization is a continuous measure of market share. Mayhew and Wilkins (2003) argue that any auditor with a market share above 20% or more is classified as an industry specialist. Mayhew and Wilkins (2003) also define industry specialization as auditor with the largest industry market share provided that this market share is at least 10% larger than that of the second largest auditor. Reichelt and Wang (2010, p. 657) use the following definition: ‘the largest annual market share and at least 5% greater industry market share than the closest competitor’. Aligning industry specialization with client specific characteristics and demands enables audit firms to service larger groups of clients with similar characteristics. Thus, it seems that audit firms meet clients’ needs by providing industry-specific value (Mayhew & Wilkins, 2003). In addition, research suggests that market share is a reasonable tool for measuring audit firms’ industry specialization. In the current study, auditors qualify as industry specialists when they possess a market share greater than 25%, to distinguish expertise between Big Four audit firms In line with the literature suggesting that industry specialization relates to audit fee premiums, the current study hypothesizes:

Hypothesis 1: Industry specialization is positively related to the size of audit fees.

3.1.2 Risk

According to Cobbin (2002), the literature states that several proxies are used extensively to assess risks associated with companies. Liquidity, profitability and gearing ratios are used as proxies for risk because these characteristics have been found to significantly influence fees. Gonthier-Besacier and Schatt (2007) identified exogenous dimensions of risk, such as sector of operation, and endogenous ones, such as nature of assets and a firm’s financial situation. In the current study, a lack of solvability (high debt and low equity) is qualified as risk because it is

(18)

expected that firms in worse financial positions will incur greater risks for auditors. This risk can be due to a lack of financial means in negotiating (new) audit assignments or reputational risk for auditors because the companies being audited may go bankrupt.

O’Sullivan and Diacon (2002) refer to several researchers (e.g. Simunic, 1980; Simunic & Stein, 1996) who argue that a risk premium is an important component of price setting for auditors. The risk premium corresponds with the expected value of future losses when errors are discovered in the audit. Therefore, it is expected that auditors will demand compensation for risks. The current study hypothesizes:

Hypothesis 2: Audit risk is positively related to the size of audit fees.

3.1.3 Size

The literature suggests that total assets are considered to be a significant determinant of audit fees (Ahmed & Courtis, 1999; Hay, 2013; Hay et al., 2006; Simunic, 1980). Therefore this study will measure size by the amount of total assets. Hassan and Naser (2013) state that large firms are more active than smaller firms. Consequently, these companies require more and additional audit services. Additionally, they draw more attention and will be inclined to provide more public information raising audit fees as a result. The current study hypothesizes:

Hypothesis 3: The size of Dutch listed companies is positively related to the size of audit fees.

3.1.4 Lowballing

Existing literature claims that lowballing following mandatory rotation is usual (e.g. Ettredge & Greenberg, 1990; Gregory & Collier, 1996). Corbella, Florio, Gotti and Mastrolia’s (2015) results indicate that the fees paid to the Big Four auditors decreased following audit firm rotation. Many companies began to rotate in 2014, even though MAFR was not in effect until 2016;2 this practice lowered audit fees. Cameran, Francis, Marra and Pettinicchio’s (2013)

(19)

19

research found that Italy, which has implemented MAFR since 1975, does not favour the process. Their research suggests that opportunistic pricing may be an issue, since final-year fees were substantially higher than normal. Moreover, the researchers found evidence of lowballing, since new auditors discounted fees while logging more audit hours. Regarding non-audit services in the US, Bell, Causholli and Knechel (2015) found that first-year audits involved audit fees and were of substandard quality. However, the audit effort was higher than in subsequent years. Additionally, Simon and Francis (1988) suggest that new auditor fees drop significantly in the first year, with a smaller drop in fees in the subsequent two years before price drops disappear. It has also been suggested that auditors lowball costs to establish large firms as a client’s (Castro, Peleias & Silvia, 2015).

From the establishment of the NIVRA and NovAA (2002), it can be concluded that audit fees for new clients need to be sufficient for billable hours by qualified staff to safeguard the quality of audits. NIVRA and NovAA forbid the offering of audits below cost price. Schaik (2003a) explains that lowballing is a threat to independence in the form of reduced objectivity. Van Schaik (2003a) reasons that an insufficient first-year fee to cover expenses is acceptable because in subsequent years, an audit firm profits from experience with clients, resulting in a more efficient audit. Another reason Van Schaik (2003a) mentions for lowballing is that auditors may want to gain or retain clients by accepting a low audit fee for auditing the financial statements, but receive additional compensation for other audit- and advisory services. However, the NIVRA and NovAA forbid auditors from placing quotations depending on performing other audit services in subsequent periods. Dutch restrictions on lowballing are stricter than those in the US, UK and Canada (van Schaik, 2003a). The Institute of Chartered Accountants of England and Wales states that members’ objectivity and the quality of their work may be impaired as a result of self-interest in acquiring clients (ICAEW, 2002). Low audit fees can cause suspicion of low quality among outside parties (SEC, 1977). Therefore, van Schaik emphasizes the societal importance of the real and perceived independence of auditors. As stated in Elitzur and Falk (1996), lowballing does not occur often with disclosure of audit fees; audit quality is perceived to be impaired with low(er) audit fees. Hence, the recommendation to disclose is in the interest of audit firms because disclosure reduces the need

(20)

for auditors to lowball (van Schaik, 2003b).

Some literature claims that audit fees increase in the year of audit firm rotation. Raiborn, Schorg and Massoud (2006) state that first-year audit fees increase due to the costs of familiarizing audit personnel with a client and its practices. They argue that the higher audit fee is a result from the shortened rotation period in which these costs can be spread. In addition, Francis (1984) compared prior, initial and subsequent audit fees of the eight biggest audit firms (i.e. Big Eight) accounting firms in Australia. If lowballing would occur, initial fees observed should be lower than either prior-year or subsequent -years. Francis’s evidence was inconsistent with price cutting of initial audit fees by the Big Eight accounting firms in Australia. He found some evidence that initial audit fees were higher than continuing engagement fees.

Dutch legislation is comparably strict in an international overview regarding lowballing and increased disclosure. It is expected that this strict regulation will reduce lowballing (van Schaik, 2003b). However, lowballing will still occur in the Netherlands. Therefore, the current study hypothesizes:

Hypothesis 4: Changes in the audit firms used by companies are negatively related to the size of audit fees.

3.1.5 Mandatory- or voluntary firm rotation

Brown and Knechel (2016) examined the compatibility between auditors and clients. They find that the degree of compatibility can be influenced by auditors’ industry knowledge. Johnson and Lys (1990) state that the competitive advantage of the current auditors of existing clients can decay over time as result of adjustments in a client’s operations and activities that affect auditor-client alignment. Consequently, voluntary auditor realignment occurs when incumbent auditors’ competitive advantage decreases. Clients switch to auditors that are less expensive or of higher quality.

Following mandatory changes in audit firms in a financial year, the current study assumes that the current audit firm will possess more industry knowledge than a new one as the

(21)

21

result of better auditor-client alignment. Audit firms that specialize in certain industries will be able to realize economies of scale, operate efficiently and request audit fee premiums for their higher-quality audit services. However, this study expects lowballing to occur following (mandatory) audit firm rotation. Consequently, rotation will lead to lower audit fees. Therefore, the current study hypothesizes:

Hypothesis 5: Changes in audit firms (moderator) due to (mandatory) audit firm rotation are negatively related to the affect of industry specialization on the size of audit fees.

3.2 Determinants of audit fee disclosure

Numerous studies have examined the extent to which firm characteristics relate to the level of disclosure in annual reports. Ahmed and Courtis (1999) conducted a meta-analysis of 23 separate studies and found a significant relationship between corporate disclosure levels and size, listing status and leverage. Several other studies have concluded that firm size is positively associated with quality of reporting (Brammer & Pavelin, 2006; Chow & Wong-Boren, 1987; Eng & Mak, 2003; Hossain et al., 1995). In the following paragraphs, each determinant of this study regarding disclosure quality will be discussed.

3.2.1 Size

Several studies (e.g. Brammer & Pavelin, 2006; Lee, 2015; Marston & Shrives, 1991; Palmer, 2008) suggest that company size is positively related to the quality of audit fee disclosures. Additionally, Lee (2015) and Hassan and Naser (2013) conclude that company size is positively correlated with quantity and quality of disclosures.

The previously discussed increased information demand for large firms may be due to a more ‘diverse portfolio of activities’ (Ahmed & Courtis, 1999, p. 53), reacting to the informational expectations of a more widespread ownership or ‘the ability to draw upon an advanced internal data-gathering and reporting system’ (Ahmed & Courtis, 1999, p. 53). Larger firms face increased scrutiny from stakeholders, such as financial analysts and shareholders, enforcing better disclosure (Palmer, 2008).

(22)

Palmers’ (2008) arguments for increased and higher-quality disclosures from large rather than small firms are that large firms are better equipped to adopt regulations, have greater knowledge of the impact of regulations and have more advanced implementation processes. Moreover, Hooghiemstra (2012) argues that to solve agency problems such as entrenchment and suboptimal investments, it is expected that larger firms will disclose higher-quality information because they have greater agency problems (Dey, 2008, as cited in Hooghiemstra, 2012). Firms want to demonstrate that their actions are legitimate and that they try to be responsible. To maintain their reputation, large firms will need to disclose more information than smaller ones. Additionally, it is more likely that the management of a large firms will benefit from ‘better disclosure, such as easier marketability of securities and greater ease in financing’ (Singhvi & Desai, 1971, p. 131).

Since financial reporting standards for audit fees have been substandard (Langendijk, 1997; Breij, 2009; Schaik, 2003b), it is relevant to research whether the quality of larger firms’ disclosures is relatively higher. Following the existing literature, the current study hypothesizes:

Hypothesis 6: The size of Dutch listed companies is positively related to the quality of audit fee disclosures.

3.2.2 Complexity

Healy and Palepu (2001) argue that when firms are members of network organizations, it is ‘difficult for financial statements to fully reflect the complex relations and implicit commitments that underlie network relations’ (Healy & Palepu, 2001, p. 433), Hay et al. (2006) state that complexity measures can encompass number of subsidiaries and activities in foreign countries. In the current study, number of subsidiaries is used to define the level of complexity. Chalmers and Godfrey (2004) state that there may be a higher demand for disclosure of complex or risky industries. The researchers identified the mining and oil industries as complex and risky and found that companies in these industries voluntarily disclose more information than those in other industries. Liu and Lai (2012) mention that complex firms with limited transparency are subject to a significant amount of information asymmetry, which leads to a demand for

(23)

23

high-quality auditors to improve the quality of financial reporting. Hiring a high-quality auditor can reduce the costs of information asymmetry through improved credibility of financial statements. By improving disclosure quality, managers reduce agency costs, which can ‘reduce the cost of capital by reducing investor uncertainty’ (Watson, Shrives & Marston, 2002, p. 290). Based on the well-cited research of Healy and Palepu (2001) the current study hypothesizes:

Hypothesis 7: Complexity is negatively related to the quality of audit fee disclosures.

3.2.3 Industry specialization

Dunn and Mayhew (2004) investigated companies hiring specialized audit firms ‘as part of their overall disclosure strategy’ (Dunn & Mayhew, 2004; p. 35). They state that ‘clients typically seek auditors who understand their industries’ (Goff, 2002, as cited in Dunn & Mayhew, 2004) and choose industry specialists. Their study demonstrates that industry specialists can assist clients in enhancing their disclosures. However, the researchers predicted that specialized firms are less important in regulated industries where enhanced disclosure adds little value. Behn, Carcello, Hermanson, D.R. & Hermanson, R.H. (1997), as cited in Dunn & Mayhew (2004) state that industry specialist their role in ‘enhancing disclosure quality is consistent with prior research that documents a strong link between client satisfaction and auditor industry specialization and that clients value auditor assistance that goes beyond basic [General Accepted Accounting Principles]’. Following this literature, the current study hypothesizes:

Hypothesis 8: Industry specialization is positively related to the quality of audit fee disclosures.

3.2.4 Risk

Ahmed and Courtis (1999) conducted a meta-analysis of 23 separate studies and found a significant relationship between the extent of corporate disclosures status and leverage. Ahmed and Courtis argue that levered firms have greater obligations to satisfy creditors’ demands for information (Wallace, Naser & Mora, 1994, as cited in Palmer, 2008). In accordance with agency theory, various other studies have found a positive relationship between leverage (low

(24)

solvability) and amount of disclosure (e.g. Hossain, Perera & Rahman, 1995; Lan, Wang, Zhang, 2013). Hummel and Schlick (2016) found that levered firms were more inclined to provide high-quality disclosures, possibly because creditors demand increased transparency from more levered firms. Agency theory suggests that firms with low solvability disclose more information to satisfy stakeholders (Watson et al., 2002). Therefore, the current study hypothesizes:

Hypothesis 9: Risk is positively related to the quality of audit fee disclosures.

4. Research design

4.1 Data and sample selection

The current study involves quantitative research based on empirical data retrieved from annual reports collected by four students from Rijksuniversiteit Groningen (RUG). The study set out to examine the quality of the audit fee disclosures of 75 companies in the Netherlands listed on the AEX, AMX and AScX stock exchanges (see Appendix B) over a five-year period (2012 to 2016). General company data was collected to gather information about the determinants related to the size of audit fees and the quality of fee disclosures. The listing in 2016 is the focus point, where after the four preceding years are examined (five for certain figures). Also, audit firm rotation will be examined and the distribution of fees for the audit services provided. Small and medium enterprises (SMEs) are excluded from the empirical research. This archival study possesses strong external validity because it references empirical data.

The sample consists of 319-year observations. Due to insufficient data 38-year observations have been omitted. Further, 10-year observations with a joint audit and 2-year observations with firms audited by a non Big Four audit firm have been omitted as this caused difficulties assessing the industry specialization variable. As there appeared to be outliers based on used regression models the Mahalanobis distance was examined, which shows outliers of the regression models; leading to removing 3-year observation with outliers with a Mahalanobis score greater than 100. Moreover, 3-year observations have been removed as they were deemed outliers with a 0% disclosure score. To reduce influence of outliers winsorizing is applied to

(25)

25

the disclosure score and the solvency ratio. The values which are more / less than three times the standard deviation of the mean are set to the average of +/- three times this standard deviation. Instead of winsorizing the square root is taken from the total assets and the number of subsidiaries; to reduce outliers. Moreover, a Ln transformation is conducted for the solvency ratio and the total audit fee to reduce the impact of skewed distribution (heteroskedasticity issues). These measures were taken as it improved the normal distribution of fees and disclosure quality. In Appendix F1 and F2 the histogram, normal probability plot and scatterplot are shown of the two models.

4.2 Research Method

To analyse the quality of financial reporting, the current study uses a disclosure index. Disclosure indices are based on content analysis (information in financial statements); they are an established method in the social sciences (Beattie et al., 2004; Marston & Shrives, 1991). Because it is difficult to directly assess disclosure quality, disclosure indices are used. These indices assume ‘that the amount of disclosure on specified topics proxies for the quality of disclosure’ (Beattie et al., 2004, p. 15). Using an unweighted score as the opposite of weighted items has not produced significantly different results (Chow & Wong-Boren, 1987), and ‘using an unweighted score permits an analysis independent of the perceptions of a particular user group’ (Chow & Wong-Boren, 1987, p. 536). It is common to use level of disclosure as proxy for quality of financial reporting by assuming that more information leads to more transparent reporting (Ahmed & Courtis, 1999; Eng & Mak, 2003; Cooke, 1992).

The disclosure index is the most appropriate method to measure the quality of audit fee disclosures because it can measure adherence to the Directive of Financial Reporting. The selection of disclosure indices is based on the Directive of Financial Reporting 390.3, a guideline for listing audit fees in financial statements. The disclosure index expanded on the mandatory items established in provision 2:382a (composition of fees). It also followed best practices of the NIVRA directive by considering audit firms’ networks and methods of listing the fees as well as whether firms provide qualitative explanations of size and development. The

(26)

three possible combinations include the following: (i) the information-element is mentioned (ii), the information-element is not mentioned and (iii) the information-element is not applicable. A nominal scale was used (Yes and No), and no weight was attached to the individual indices. The disclosure index consisted of 12 items (see Appendix C), and disclosure scores were measured as follows:

4.2.1 Conceptual models

This study consists of two models. The first model will examine the relation of four determinants to the size of the audit fee and the second model will present four determinants that are related to the disclosure quality of the audit fees. The chosen determinants are shown in the models on the next pages. The following paragraphs will operationalize all variables.

(27)

27

Conceptual model

Change of audit firm in the

Financial year (Moderator)

Model 1:

H5

-H4 -

Industry specialization

H1 +

Companies associated with risk

H2 +

Size audit fees

(28)

Model 2

H6 +

H7 -

H8 +

H9 +

Industry specialization

Complexity

Disclosure quality audit fee

Company size

(29)

29

4.2.2 Dependent variables

The dependent variables of this study are ‘size of audit fees’, operationalized as the natural logarithm of audit fees; following Simunic (1980); and ‘quality of audit fee disclosure’, as calculated by the disclosure index score.

4.2.3 Independent variables

‘Company size’ is measured as the square root of total assets, based on Mayhew & Wilkins (2003), which has shown that transformed assets are highly correlated with audit fees. They further note that transformed assets are considered to be a better measure of industry concentration than an untransformed measure. The study of Mayhew & Wilkins (2003) also shows that it is assumed that industry market share is associated with industry specialization. Therefore, a Big Four firm qualifies as ‘Industry specialist’ when he audits more than 25 percent of the market share; measured by the square root of total assets. An ordinal scale (1-4) is used to distinguish Big Four firms and year dummies (2012-2016) are used to examine differences between years. ‘Risk’ is measured based on the solvency of a company, the natural logarithm of the equity divided by debt. ‘Audit firm rotation’ indicates whether an audit firm rotated in the financial year, which also acts as a moderator by interacting with industry specialization. ‘Complexity’ is defined as the square root of subsidiaries as Langendijk (1997) has shown that this specification provides a good fit in ordinary least squares regression.

4.2.4 Control variables

‘Audit commission existence’ is used as control variable as this has shown to be a determinant of audit fees in previous research (Hay et al., 2006; Hay, 2013). The existence of an audit commission is also used as a control variable for the second model as this study expects that this leads to higher financial reporting quality. In addition, year dummies (2012 through 2016) are included as control variable, where 2012 will act as the reference dummy; allowing a comparison with later years. Table 1 shows the definitions used in the correlation- and regression analysis.

(30)

Table 1. Definitions variables correlation- and regression analysis

SPEC Industry specialization audit firm

ROTATE Audit firm rotation

RISK Audit risk of a firm, measured as the natural logarithm of the solvency ratio (equity/debt)

SIZE The square root of total assets

COMPLEX The square root of subsidiaries of the auditee

ACE Audit Commission Existence

YEAR Year in which an audit has taken place

DIX Disclosure index score

4.2.5 Statistical method

To test the hypotheses a correlation- and regression has been conducted. First, the Pearson correlation coefficients are calculated as this is a first indication of a significant (linear) association between two variables; in addition, multicollinearity can be checked. The value of the correlation coefficient (which lies between -1 and 1) will indicate how the variables relate to each other. If the independent variables correlate strongly, multicollinearity can be an issue. Multicollinearity refers to high correlation among the independent variables (Huizinga, p.309) which will negatively affect the reliability of the analyses. When a correlation coefficient is greater than 0.7 or less than -0.7, there is a possible multicollinearity issue. In addition, the tolerance and ‘Variance Inflation Factor’ (VIF) will be examined to assess multicollinearity. A

(31)

31

rule of thumb is that VIF values greater than 10 signal multicollinearity (Huizing, 2007, p. 309); which equals 1 / tolerance.

Further, the data will be processed by means of two hierarchical regression analyses in SPSS. The regression analysis is used to estimate a (possible) linear relationship between the dependent variables and one or more independent variables. The first model is a hierarchical regression with 3 steps. Step 1 consists of the control variables; step 2 consists of the control variables + the independent variables + moderator; step 3 consist of the control variables + independent variables + moderator + interaction term. For the first model, all independent variables (including control variables) are standardized. For the second model, a hierarchical regression is used with step 1 consisting of the control variables and step 2 consisting of the control variables + the independent variables.

This leads to the following regression for model 1:

Ln audit fee = β0 + β1 x SIZE + β2 x SPEC + β3 x RISK + β4 x ROTATE + β5 x (ROTATE x SPEC) +



And the following regression for model 2:

(32)

5. Results

5.1 Descriptive statistics

Table 2,3,4 and 5 elaborate on audit fee- distributions and disclosure quality. For the sake of brevity tables concerning a full overview of disclosure scores and (audit) firm characteristics are shown in Appendix D and E. Appendix E indicates that there appeared to be more industry specialist in the last few years and that relatively more audit firm rotations have taken place. The level of risk and size of the firms remained relatively stable.

Table 2 shows the development of the size of the audit fees over the years 2011-2016 and how the audit fees have been distributed. The mean total audit fee has dropped from €5.8 million in 2011 to €4.5 million in 2016. This is due to a severe drop in non-audit services provided since 2012.

Table 2. Audit fee distribution 2011-2016 (in €k) Year n Mean total

audit fee % change SD total audit fee % Audit financial statements Mean fee NAS % NAS SD NAS 2011 56 5.817 - 10.690 79,0 1.272 22,1 1.184 2012 57 5.714 -1.8 10.471 74,6 1.451 25,4 1.041 2013 60 5.156 -9,8 9.204 77,8 1.143 22,2 595 2014 64 4.928 -4,4 8.991 81,6 907 18,4 400 2015 67 4.670 -5,2 8.609 84,0 747 16,0 409 2016 71 4.530 -3,0 8.008 86,9 504 13,1 178

(33)

33 Table 3 shows the audit fee distribution of 2016.

Table 3. Audit fee distribution 2016 (in €k)

Mean % of total audit fee SD

Audit of the financial statement 4.019 86,9 7.437

Total fees for other audit assignments 430 9,3 886

Total fees for advisory services in the tax field 61 1,3 160

Total fees for other non-audit services 117 2,5 406

Table 4 gives an overview of the disclosure score during the years 2012 through 2016. Table 4. Disclosure score (%) overview 2012-2016, mean and standard deviation

Year n Mean SD 2012 57 68,27 19,75 2013 60 69,86 20,30 2014 64 70,39 21,12 2015 67 70,17 20,93 2016 74 73,36 19,60

(34)

Table 5 gives an overview of the disclosure scores on audit fee distribution in 2016 Table 5. Disclosure score (%) distribution audit fee 2016

n %

Audit of the financial statement 71 100,0

Total fees for other audit assignments 66 93,0

Total fees for advisory services in the tax field 54 76,0 Total fees for other non-audit services 55 77,5

(35)

35

5.2 Correlation analysis

The correlation matrix in table 6 shows the correlation coefficients with a two-tailed significance.

Table 6. Correlation matrix

FEE DIX SPEC RISK SIZE ROTATE COMPLEX ACE YEAR

FEE - DIX ,297** - SPEC ,108 -,034 - RISK -,348** ,002 -,173** - SIZE ,649** ,078 ,210** -,340** - ROTATE -,005 ,003 ,190** ,018 -0,025 - COMPLEX ,661** ,134* ,126* -,223** ,470** -,003 - ACE ,218** ,038 ,078 -,003 ,103 -,001 ,111* - YEAR ,008 ,074 ,196** ,067 ,020 ,269** ,000 ,012 - * p < .05. **p <.01 N = 319

From the correlation analysis can be concluded that there are no multicollinearity issues. Also, the requirements of the additional test (tolerance + VIF) are fulfilled. The significant relations will be discussed. The disclosure index score is very strong positively correlated to the size of audit fees and strong positive correlated to complexity; implying that high disclosing firms generally pay higher audit fees and that firms with more subsidiaries generally disclose more. Industry specialization is very strong negative correlated to risk, very strong positive correlated to size, audit firm rotation and year; and is strong positive correlated to complexity; indicating that firms’ high in equity are less likely to be audited by an industry specialist; that

(36)

large companies, companies who changed auditor and firms with more subsidiaries are more likely to be audited by an industry specialist. Also, in later years there are relatively more industry specialists. Risk is very strong negatively correlated to the size of audit fees, industry specialization, size and complexity; meaning that firms high in equity generally pay lower audit fees, are less likely to be audited by an industry specialist, are likely to be smaller; and likely have fewer subsidiaries. Size has a very strong positive correlation with the size of audit fees and complexity; implying that larger firms and firms with more subsidiaries generally pay higher audit fees. Audit firm rotation is very strong correlated with year; implying that in the last few years relatively more audit firm rotations have taken place. Complexity is very strong correlated to the size of audit fees and strong correlated to the existence of an audit commission; indicating that firms with more subsidiaries pay a higher audit fee in general; and are more likely to have an audit commission. Audit commission existence has a very strong correlation with the size of audit fees; meaning that firms who have an audit commission generally pay higher audit fees.

5.3 Regression analysis

The values of the full model will be used to accept / reject hypotheses for both models. Table 7 shows the regression coefficients of the first model. From table 7 can be concluded that 3,3% of audit fees are explained by the control variables (Adjusted R2 _{= 0,033); 42% more (Δ R}2₌ ,420) can be explained by adding the moderator and independent variables; no significant improvement has been made by adding the interaction term (p > 0.05 F Change). The full model explains 45.7% of audit fees (Adjusted R2_{= ,457). The value of the constant in the full model} shows that the average natural logarithm of the audit fee is 7,387.

(37)

37

Table 7. Hierarchical regression analysis of predictors of audit fees

Predictor variables Step 1 Step 2 Step 3

Constant 7,447*** 7,394*** 7,387*** zACE ,316*** ,233*** ,241*** Year 2013 -0,60 ,024 ,028 2014 -,038 ,036 ,064 2015 -,033 ,008 -,004 2016 ,010 ,065 ,081 zRotate ,032 ,030 zSpec -0,92 -,086 zSize ,873*** ,879*** zRisk -2,32*** -,235*** Interaction term -,115 Adjusted R2 ,033 ,453 ,457 R2 change ,048 ,420 ,006 F - value 3,153*** 30,235*** 27,791*** *p < 0.05; **p < 0.01; ***p < 0.001.

Hypothesis 1 assumes that industry specialization is positively related to the size of the audit fees; reasoning that audit firms request an audit fee premium as they specialize. From table 7 can be concluded that there exists a negative but non-significant relation (β = -,086; p >

(38)

0.05); rejecting our first hypothesis. Hypothesis 2 assumes that risk is positively related to the size of the audit fees; reasoning that audit firms demand compensation for possible reputational damage or their ability to provide future (audit) services. Note that the coefficient is inversely related to the expected relationship. From table 7 can be concluded that there exists a negative and significant relation (β = -,235; p < 0.001); supporting our second hypothesis. Hypothesis 3 assumes that the size of Dutch listed firms is positively related to the size of the audit fees. As larger firms have a wide(r) range of activities and are being scrutinized more by stakeholders it is expected that they incur a higher audit fee. From table 7 can be concluded that there exists a positive and significant relation (β = ,870; p < 0.001); supporting our third hypothesis. Hypothesis 4 assumes that the rotation of an audit firm is negatively related to the size of the audit fees, as lowballing is expected to apply. From table 7 can be concluded that there exists a positive but non-significant relation (β = ,030; p > 0.05); rejecting our fourth hypothesis. Hypothesis 5 assumes that the rotation of an audit firm negatively moderates the positive industry specialization – audit fee relation, as lowballing is expected to apply. From Table 7 can be concluded that there exists a negative but non-significant relation (β = -,115; p > 0.05) for the interaction term; not supporting our fifth hypothesis.

Table 8 shows the regression coefficients of the second model. From Table 8 can be concluded that the model doesn’t fit on the data (Adjusted R2 _{= -,011 and ,000), meaning that} neither the control variables nor the independent variables explain the disclosure quality of audit fees. Both F-values are non-significant (p > 0.05). The value of the constant in the full model shows that the average disclosure score is 63%.

(39)

39

Table 8. Hierarchical regression analysis of predictors of disclosure quality audit fees

Predictor variables Step 1 Step 2

Constant 65,581*** 63,008*** ACE 2,970 1,958 Year 2013 1,473 2,132 2014 2,227 2,741 2015 1,988 2,480 2016 5,270 6,004 Complex ,340 Risk 1,171 Size ,737* Spec -3,011 Adjusted R2 -,011 ,000 R2 change ,008 ,023 F-value ,445 ,995 *p < 0.05; **p < 0.01; ***p < 0.001.

Hypothesis 6 assumed that industry specialization is positively related to the quality of audit fee disclosure; as the presence of an industry specialist can enhance disclosure quality. From table 8 can be concluded that there exists a negative but nonsignificant relation (β =

(40)

-3,011; p > 0.05); rejecting our sixth hypothesis. Hypothesis 7 assumed that risk is positively related to the quality of audit fee disclosure; as levered firms are expected to have more information asymmetry, which they would like to reduce by improving disclosure. From table 8 can be concluded that there exists a positive but non-significant relation (β = 1,171; p > 0.05); not supporting our seventh hypothesis. Hypothesis 8 assumed that size of Dutch listed firms is positively related to the quality of audit fee disclosure as larger firms are more incentivized and have more resources to disclose. From table 8 can be concluded that there exists a positive and significant relation (β = ,737; p < 0.05); supporting our eighth hypothesis. Hypothesis 9 assumed that complexity is negatively related to the quality of audit fee disclosure, as the number of subsidiaries impairs disclosure quality. From table 8 can be concluded that there exists a positive but non-significant relation (β = ,340; p > 0.05); rejecting our ninth hypothesis.

6. Discussion & Conclusion

The themes of my study are the height- and disclosure quality of audit fees and the underlying determinants. The goal of this study is to analyse whether there is a relation between (audit) firm characteristics and the size of the audit fees; also considering audit firm rotation. Moreover, whether (audit) firm characteristics are related to the quality of audit fee disclosure. The research question of this study is:

‘What variables affect audit fees and the quality of fee disclosures in the Netherlands?’

In order to answer this research question, an empirical financial statement analysis has been conducted consisting of 75 annual reports in the reporting period 2011 to 2016.

2:382a was introduced to inform the reader of the financial statements of the independence of the external auditor; assuming that high NAS are a threat to independence (Langendijk, 2011). Moreover, Langendijk (2011) expected that the introduction of 2:382a would lead to price competition and possibly impair audit quality of annual reports as a result.

The impact of legislation on audit market services in the Netherlands