The expanded audit report:

(1)

The expanded audit report:

improving audit quality?

Marjon Alberts S3273601

Master’s thesis Accountancy Faculty of Business Economics

University of Groningen

Supervisor: Prof. Dr. D.A. de Waard RA Second assessor: Dr. R.C. Trapp

(2)

ABSTRACT

The issuance of the expanded audit report is another attempt of regulators to restore public confidence in the auditing profession. This revision increases the level of transparency for auditors, and might therefore affect auditor efforts. In turn, I expect an improvement in audit quality. Additionally, the relation between several characteristics of the expanded audit report and audit quality was examined. A quantitative study was performed using a total of 875 unique firm observations. The findings suggest that the issuance of the expanded audit report is associated with a significant improvement in audit quality. The analyses failed to provide significant results on the relationship between the length of the audit report and audit quality. However, the number of key audit matters is significantly positively associated with audit quality. In sum, besides enhancing public confidence in the auditing profession, the expanded audit report seems to also have improved audit quality.

(3)

1. INTRODUCTION

The value and existence of an audit derives from public confidence in an auditor’s capabilities. The current debate about the auditing profession arose in 2001 following multiple high-corporate failures occurring shortly after the issuance of unqualified audit reports (Herbohn, Regunathan & Garsden, 2007), and the unsuspecting role of auditors in the financial crisis (Sikka, 2009). The purpose of an audit and the role of auditors was, and still is, subject to widespread criticism. On the one hand, audit firms are subject to reputational blows due to negative press coverage. For example, in 2017 extensive media attention was drawn to the association of PwC with an accounting scandal at their client BT (Thompson, 2017), the investigation of KPMG for the audit of Rolls Royce (Kollewe, 2017), an £1,8 million fine for EY after admitting misconduct in auditing Tech Data (Kunert, 2017) and a hack at Deloitte that compromised confidential information which went unnoticed for months (Hopkins, 2017). On the other hand, in recent years, oversight boards have publicly put the level of audit quality up for debate, stating it is insufficient (FRC, 2016b; AFM, 2017). In sum, the auditing profession lost the confidence of the public and in response, the auditing industry tends to revise standards in order to regain public confidence (Sikka, 2009). One of the measurements initiated by regulators is to expand the audit report, which requires auditors to provide users of the financial statements with more insights in the audit process. In this study, I aim to examine whether the expanded audit report is associated with an improvement in audit quality and if certain characteristics of the expanded audit report are determinants for higher audit quality.

The discussion about the value of auditors, and the litigation against auditors is, according to Porter (1993), an implication of a gap between society’s expectations of auditors and the actual performance of auditors. In fact, it is demanded by society that auditors provide an early warning signal in case of possible corporate failures. However, the legislated role of auditors is to provide an affirmation of the financial statements and auditors are limited in acting beyond that role (Herbohn et al., 2007). The established expectation gap that exists illustrates the differences between what the users of the financial statements expect from auditors and what auditors actually provide (Gold, Gronewold & Pott, 2012). The audit report lacks informative value to narrow that gap (e.g., Church, Davis & McCracken, 2008; Gray, Turner, Coram & Mock, 2011; Humphrey, Loft & Woods, 2009). For the audit report to be effective and of value, it is necessary that it is read and understood. A study by Coram, Mock, Turner and Gray (2011) shows that readers of the financial statements are solely interested in the existence of the audit report and whether it is unqualified or not, rather than in the content of it, which shows that the symbolic value, rather than the informative value of the audit report is emphasized. The perceptions of investors that an audit report lacks informative value is due to its standardized nature. The expanded audit report is expected to increase the informative value of the audit report, and thus the relevance of the audit.

As first issued by the Financial Reporting Council (FRC) in the United Kingdom, the expanded audit report provides more insights in the scope and outcomes of an audit. The standard, ISA 700, is required for entities of public interest and became effective as of financial year 2013 (FRC, 2016c). A year later, a similar standard was introduced in The Netherlands by the Nederlandse Beroepsorganisatie voor Accountants (NBA), namely NV COS 702N. The new standards aspire to de-standardize audit reports. Despite the important role of auditors in society, they communicated solely through standardized reports. Additionally, the outcome of an audit was perceived to be unobservable by society (Knechel, Krishnan, Pevzner, Shefchik & Velury, 2013). The expanded audit report aims to increase the usefulness of the audit report, as it involves more firm-specific information and elaborates on the responsibilities of the auditor. The disclosures that are required possibly increase

(5)

the accountability of the auditor (Reid, Carcello, Li & Neal, 2015). The additional information also may enhance the professional skepticism of auditors and incentivize auditors to collect more proper audit evidence to support their opinion (PCAOB, 2016). Another possible development of the expanded audit report is that the decreasing degree of standardization pushes audit firms to compete to deliver the most added value to users of the financial statements (Clikeman, 2001), and therefore increase auditor efforts. An audit report that enhances audit quality is one that commands confidence, as confidence in the audit report is linked to confidence in the quality of the audit process (Holm & Zaman, 2011). Additionally, the NBA (2014) stated that an expanded audit report increases the relevance of the audit and, in turn, audit quality. Hence, the role of the audit report in association with audit quality is emphasized, not only by academic research but also by regulating bodies.

The importance of this research draws from several matters. Primarily, the research adds to literature regarding audit quality by examining the effect of a relatively new development in the auditing profession. The effects of the expanded audit report on audit quality is an infrequently researched subject, and studies have presented this as a suggestion for further research (Reid et al., 2015; Coram et al., 2015; Reid, 2015).

Additionally, the effect of the expanded audit report has been previously researched in the United Kingdom in the year following up the implementation year of ISA 700. This study, conducted by Reid et al. (2015), shows an improvement in audit quality. However, it is unknown if the effects of the expanded audit report on audit quality persist for multiple years. Taking into account the learning curve that exists in the auditing profession, the effects may differ over time (O’Keefe, Simunic & Stein, 1994).This research provides multiple year insights, and answers whether the effects on audit quality actually persist. Furthermore, the study contributes to the study of Reid et al. (2015) on a country-level. This thesis contains data and corresponding results from both the United Kingdom and The Netherlands. In the latter, the expanded audit report and its effects on audit quality is not previously researched and is still unknown. Therefore, the study provides new insights relevant for the audit profession, regulators and users of the financial statements in The Netherlands.

Thirdly, the audit report is an important subject of research as it is the primary evidence of an audit available to the public. At a social level, the recognition of audit quality is necessary to maintain the viability of the audit. As stated by Carpenter and Dirsmith (1993), the audit profession largely depends on the ability to assert and legitimize the value of the audit and its outcomes. The existence and justification of auditing is dependent by the image generated by the profession. Apparent auditor behavior, and maintaining a high level of professionalism is an essential element in the social image of auditing (Herrbach, 2001). The audit report is a tool for auditors to communicate their professional skepticism, create legitimacy and add significance to the profession. Examining developments concerning the audit report creates possibilities for practical implications valuable to the auditing profession and regulatory bodies that set auditing standards, but also to users of the financial statements to gain an understanding of audit quality and its determinants.

The remainder of this thesis is organized as follows: in chapter 2 the background, theory and development of the hypotheses underlying this thesis will be described. Subsequently in chapter 3, the methodology of this thesis is elaborated on. The results are presented in chapter 4. Lastly, the conclusion and limitations of this thesis will be discussed in chapter 5.

(6)

2. BACKGROUND, THEORY AND HYPOTHESIS DEVELOPMENT

The purpose of this study is to examine whether the expanded audit report has led to an improvement in audit quality, and whether several characteristics of the expanded audit report are associated with higher audit quality. ISA 700 and COS 702N ask for an expanded audit report, wherein auditors are required to provide information about the scope of the audit, the materiality threshold and key audit matters identified during the audit (FRC, 2016; NBA, 2014). The increasing degree of transparency necessary is expected to lead to an improvement in audit quality. In this paragraph theory surrounding the subject of this research are discussed, and subsequently the hypotheses for this research are developed.

2.1 The determinants of audit quality

Little is understood about audit quality, but it is widely discussed and researched (Knechel et al., 2013). Despite the popularity, a uniform definition of audit quality does not exist (Francis, 2011). The definition of audit quality used in this study is the one formulated by DeAngelo (1981a: 186): ‘’the

quality of audit services is defined to be the market-assessed joint probability that a given auditor will both a) discover a breach in the client’s accounting system, and b) report the breach.’’ This

definition of audit quality contains two components: the audit process (discovering a breach), and the audit output (reporting this breach). In this study, the assumption is made that an association exists between audit output, namely the expanded audit report, and audit quality.

In addition to the absence of a uniform definition, the measurement and characteristics of audit quality are also frequently discussed in literature. Especially, determinants of audit quality are an area of interest. Audit quality is affected by inputs, processes, legal institutions and economic consequences of audit outcomes (Francis, 2011). As an example, previous research has shown that the following characteristics affect audit quality:

 whether the auditor is a Big 4-firm or not (e.g., DeAngelo, 1981; Francis & Yu, 2009);  tenure of the auditor (e.g., Johnson, Khurana & Reynolds, 2002; Ghosh & Moon, 2004);  the level of audit fee (e.g., Hoitash, Markelevich & Barragato, 2007; Stanley, Brandon &

McMillan, 2015);

 audit firm rotation (e.g., Jackson, Moldrich & Roebuck, 2008; Cameran, Prencipe & Trombetta, 2016).

Besides academic research, institutions that are involved with the regulation and oversight of auditors are also invested in examining audit quality. For example in the United Kingdom, the FRC assesses audit quality on a yearly basis in so-called audit quality inspection reports (FRC, 2016). Additionally, in The Netherlands the Dutch Authority for Financial Markets (AFM) has also published several reports wherein the quality of audits performed by Big 4-firms was assessed (e.g., AFM 2010; AFM 2014; AFM 2017). The results of these reports show that multiple audits were insufficiently substantiated, stating that audit quality is not up to par with what is desired. These institutions, investors, society and the auditing profession itself demand an improvement in audit quality. In response, audit firms are changing their corporate cultures, and regulators are issuing new accounting standards. The expanded audit report is one of the measures, as it serves to increase the level of transparency about the audit.

(7)

2.2 The agency theory

Audits exist as a monitoring tool to reduce potential conflicts of interests between owners and managers of a firm (DeAngelo, 1981a; Wallace, 2004). The underlying theory that describes this phenomenon is the agency theory. Investors are generally seen as outsiders, as they do not play an active role in managing the firm. Investors instruct managers to operate on their behalf by delegating decision making authority (Jensen & Meckling, 1976). As a consequence of this separation of ownership and control, information asymmetry arises which illustrates that the management of a firm has more information than outsiders of a firm. To reduce information asymmetry, mechanisms such as legislation, contracting, monitoring and information intermediaries exist (Healy & Palepu, 2001). Auditors function as information intermediaries and have a monitoring role. The existence of audit reports contribute to reducing information asymmetry (Wallace, 2004). An audit report provides stakeholders with an attestation by an independent individual of the information presented by the management (Datar, Feltham & Hughes, 2000). In applying the agency theory, the value of an audit report is illustrated and emphasized.

In practice, the perceptions about the value of an audit report differ. In acquiring financing, capital providers often require firms to hire an independent auditor – implying that capital providers find that auditors enhance the credibility of financial statements (Healy & Palepu, 2001). Additionally, it implies that audit reports are appreciated in regulated and unregulated environments. On the contrary, investors of a firm often question the value of audit reports. Studies have shown that the market shows no significant response to the issuance of an audit report (Church et al., 2008; Dodd, Dopuch, Holthausen & Leftwich, 1984). These studies illustrate that investors are mainly interested in the underlying financial figures rather than in the communication of an auditor reporting about these figures. However, research shows that the existence of audit reports is appreciated by investors as it increases the reliability of financial statements, and is, in that viewpoint, valuable (Soltani, 2000; Gray et al., 2011; Mock, Bédard, Coram, Davis, Espahbodi & Warne, 2013). Additionally, an experiment conducted by Coram et al. (2011) shows that financial analysts and investors are mostly concerned with establishing whether an audit opinion is unqualified or not, rather than concerned with the actual content of an auditor’s report. Investors value the existence of the audit report, but rarely read it (Asare & Wright, 2012). In conclusion, the audit report serves as an affirmation of the expectations of the market, and in that way, serves a useful role. It has symbolic value, but conveys little useful information (Church et al., 2008).

The existence of information asymmetry is not only present in an organization, but also exists between auditors and users of the financial statements (Antle, 1982). In addition to the limited value of the standard audit report, the report does not provide users with suitable information to comprehend the responsibilities an auditor has (Porter, 1993). Therefore, an expectation gap exists between auditors and users of the audit report. The gap entails that users have a different viewpoint of what the auditor’s responsibilities are, what an audit entails and what message is conveyed in the audit report (Church et al., 2008). The existence of the expectation gap can erode confidence in the report, affect investment decisions negatively and cause unnecessary litigation (Asare & Wright, 2012). To narrow the expectations gap, the users of financial statements should be educated and the misperceptions should be corrected with regards to the responsibilities of the auditor (Gray et al., 2011). Even though several attempts were made to narrow the expectation gap, the gap persists (Church et al., 2008). Auditors and users of the financial statements should have an aligning understanding of the scope of the audit for the audit report to have value (Asare & Wright, 2012).

(8)

In conclusion, the audit report is not read, but only glanced at. The audit report is too general and standardized and contains little information about the specific work undertaken and findings obtained by auditors (Humphrey et al., 2009). An audit report could add value when wordings are tailored to the firm that is audited, and when it does not contain general statements. The expanded audit report as issued by the FRC (2016) and NBA (2014) enhances the value of the audit report, and further contributes to mitigating information asymmetry and aligning the understandings about an auditor’s responsibilities. For example, the requirements of these new standards entail that auditors disclose about the scope of the audit, the responsibilities of the auditor, a description of the materiality applied, and which key audit matters were identified and how these were addressed in the audit. In sum, the expanded audit report provides more firm-specific information as well as information about the audit process and thus, serves to reduce information asymmetry and narrow the expectations gap.

2.3 The accountability theory

Users of the financial statements base economic decisions on the financial statements, and as stated previously, auditors serve as an affirmation of this information. As established by Asare and Wright (2012), investors use their interpretation of the audit report to shift several risks towards the responsibilities of auditors. By expressing information in the audit report certain expectations are created amongst financial statement users, according to the theory of inspired confidence (Limperg, 1932). The expanded audit report demands that auditors provide more information. Consequently, these new insights are subject to the interpretation of users and therefore, create new expectations amongst users. Auditors can subsequently be held accountable for these interpretations and expectations. As stated by Reid et al. (2015), the expanded audit report increases the degree of auditor accountability.

Accountability can be defined as the giving and demanding reasons for certain behaviour (Roberts & Scapens, 1985). A change in accountability changes behaviour. A study by Lerner and Tetlock (1999) argues that a person that is accountable for his or her actions is more accurate, consistent and demonstrates greater self-insight than a non-accountable person. Previous studies have shown that the degree of accountability increases a person’s efforts and improves his or hers performance (Johnson & Kaplan, 1991; Ashton, 1992; Kennedy 1993; Asare, Trompeter & Wright, 2000). Tan and Kao (1999) imply that accountability may increase an auditor’s efforts. In turn, an increase in accountability causes a more adequate judgement of a firm’s financial statements. Additionally, Asare et al. (2000) show in an experiment that auditors who function under circumstances wherein they are accountable perform better than auditors that did not function in an environment where accountability plays a role.

In general, users of the financial statements have little insight about how credibility to the financial statements is added, as audit evidence is not available to the public (Sikka, 2009). The audit process is seen by many as a so-called black box (Herrbach, 2001). The new standards regarding the expanded audit report provides more insights in this black-box. Auditors might perceive the increasing degree of transparency as pressure to justify their actions in order to support their judgements and decisions that should be disclosed in the audit report (Peecher, Solomon & Trotman, 2013). The expanded audit report requires transparency about audit activities. For example, an auditor is required to elaborate on how key audit matters are addressed in the audit (FRC, 2016). As a consequence, users of the financial statements gain a better understanding of the audit process, and might be able to – to a certain degree - evaluate the performance of auditors, which adds to the expected increase of accountability.

(9)

The additional information provided in the expanded audit report has to be reliable, not only to maintain public confidence but also to restore it. Auditors often use the audit report as a defensive tool in order to reduce their level of accountability (Asare & Wright, 2012). The expanded audit report provides little room for defensive mechanisms, as standardized statements no longer suffice. As a consequence, the required firm-specific information about the audit process enhances professional scepticism, and may incentivize an auditor to collect additional proper audit evidence (PCAOB, 2016). Prior research has argued that an increase in accountability leads to an improvement in audit quality (e.g., Carcello & Li, 2013; Peecher et al., 2013; Deis & Giroux, 1992). Based on the findings of the studies mentioned so far, it can be expected that as a consequence of the expanded audit report, audit quality will improve. Therefore, the following hypothesis is stated:

H1. The expanded audit report is associated with an improvement in audit quality

2.4 The legitimacy theory

The auditing profession is largely dependent on the recognition of audit quality by the social environment (Herrbach, 2001). The corporate failures that occurred over the last several years and the insignificant role of auditors in the financial crisis has negatively affected the value society attaches to the role of auditors. As stated by Holm and Zaman (2011), the auditing profession is subject to a crisis in confidence. The profession has lost its legitimacy, and is trying to regain its position in society. The legitimacy theory describes that individuals as well as firms aim to act in accordance with social values in order to legitimise their actions (Guthrie & Parker, 1989). With regards to the auditing profession, Holm and Zaman (2011) state that auditors are invested in ensuring the public they can be trusted and thus, are invested in creating legitimacy.

Research has shown that investors desire an expanded audit report with more information about the scope of the audit, the audit process and the firm (Innes, Brown & Hatherly, 1997; Humphrey et al., 2009; Gold et al., 2012). The expanded audit report provides these types of information. The perceptions of users of the financial statements can be positively affected by including more information in the auditor’s report (Coram et al., 2011). Auditors can use the audit report to legitimize their audit activities. Additionally, proponents of the expanded audit report state that it contributes to an enhancement of public confidence in financial reporting, the audit itself and increases the relevance of the audit report (EY, 2014). Auditors can use the expanded audit report as a tool to restore legitimacy to the audit profession.

Over the past few years, oversight boards have publicly put the level of audit quality up for debate (FRC, 2016; AFM, 2017). Additionally, negative press coverage after accounting scandals have damaged the reputation severely and has put the auditing profession under increasing federal oversight (Bazerman, Loewenstein & Moore, 2002). In essence, each scandal covered by the press is to society another piece of evidence that the level of audit quality is insufficient. In response and in pursuit of restoring the legitimacy of the profession, auditors are engaged in modes of impression management (Holm & Zaman, 2011). Past research shows that the length of several reports has been used as a determinant of legitimacy in multiple studies (e.g. Guthrie & Parker, 1989; Deegan, Rankin & Voght, 2000; Campbell, 2004). The wish to be seen as socially legitimate is used as an explanation for various disclosure levels. Typically, when an organization is subject to criticism, it undertakes corrective strategies to reposition itself in society. A study by Deegan et al. (2000) shows that when a firm has been subject to criticism, the firm is likely to provide a greater level of disclosures. This implies that the disclosure of information is used to restore legitimacy.

(10)

The expanded audit report provides an opportunity to auditors to enhance the public opinion, by signaling transparency, professional skepticism and independence. New mandates provide strategic opportunities to act on demands of the environment, and provide opportunities for professional reinvention and for enhanced legitimacy. Additionally, as the new standards are labelled as principles-based, the level of detail provided in the audit report is subject to professional judgement. Kipp (2017) shows that nonprofessional investors perceive audit quality to be higher when disclosures are presented with a higher level of detail. In adherence to the legitimacy theory, it can be implied that auditors use the report to manage the perceptions of users of the financial statements. By issuing a lengthy audit report, auditors respond to the demand of society to be more transparent. However, as stated previously, certain statements in the audit report create expectations amongst users. Auditors feel pressured to live up to these expectations and to justify what they are accountable for. Thus, additional disclosures demand an increase in auditor efforts in line with the accountability theory. In turn, audit quality is expected to be higher. Therefore, the following hypothesis can be stated:

H2. The length of the expanded audit report is positively associated with higher audit quality

A prominent feature of the expanded audit report is a paragraph about key audit matters. Key audit matters are, as defined by the FRC (2016b) ‘’those matters that, in the auditor’s professional

judgement, were of most significance in the audit of the financial statements of the current period and include the most significant assessed risks of material misstatement identified by the auditor, including those which had the greatest effect on: the overall audit strategy, the allocation of resources in the audit; and directing the efforts of the engagement teams.’’ Auditors are required to

state per key audit matter why it was considered and how it was addressed in the audit (FRC, 2016a; NBA, 2014).

Referring back to the legitimacy theory, the paragraph containing key audit matters gives auditors the ability to transfer transparency, independence and professional skepticism. A study by Kachelmeier, Schmidt and Valentine (2015) indicates that the inclusion of key audit matters in the audit report could decrease auditor liability. A study by Brasel, Doxey, Grenier and Reffett (2014) finds similar outcomes with regards to perceived auditor liability. Christensen, Glover and Wolfe (2014) state that the inclusion of key audit matters increases the relevance of the audit report, and thus, the audit. Additionally, the audit report is an important part of an investor’s perception of auditor credibility (Power, 2003). The results of a study by Carver and Trinkle (2016) indicate that a more detailed audit report improves the perceptions of investors on auditor credibility. The story model as developed by Pennington and Hastie (1992) implies that evidence accompanied by an explanation is perceived to be more credible. Subsequently, Brown, Majors and Peecher (2014) state that the inclusion of key audit matters decreases assessments of auditor negligence as the information is interpreted as a story of the auditor’s competence. Additionally, Brown et al. (2014) find that in the presence of key audit matters in the audit report, audit quality is higher. The reasoning behind this is that when an auditor performs negligence in the presence of key audit matters, that damages are more severe than when an auditor does not disclose key audit matters. Therefore, by disclosing a certain amount of key audit matters, auditors enhance their credibility and decrease their liability as perceived by investors. In sum, the more key audit matters presented in the audit report, the more legitimacy is created as this signals expertise and professional skepticism. Also, the theory presented implies that the inclusion of key audit matters affects the credibility of the audit report and auditor liability. However, disclosing key audit matters, which contains firm-specific information and provides insights in the audit process, creates certain expectations amongst users in line with the theory of inspired

(11)

confidence (Limperg, 1932). The communication of key audit matters might incentivize auditors to demonstrate professional skepticism necessary to signal high quality audits (PCAOB, 2016). In alignment with the accountability theory, auditors might feel pressured to justify the key audit matters and audit activities performed. In turn, the disclosure of key audit matters and the corresponding audit activities may increase efforts. Pressure occurs to justify these activities and efforts increase in performing them. In turn, audit quality is expected to improve. Therefore, the following hypothesis can be formulated:

H3. The number of key audit matters in the expanded audit report is positively associated with higher audit quality

The theoretical framework has explained why an increase in audit quality is expected. However, an increase in audit quality is not the ultimate aim of the expanded audit report. In essence, the expanded audit report only desires to increase the level of transparency. Regulations regarding the audit report do not include changes in the actual scope of the audit. However, based on the theories discussed above, the assumption is made that the expanded audit report will improve audit quality. The analyses to be performed serve to confirm or reject these assumptions.

(12)

3. METHODOLOGY

In this section, the methodology is elaborated on. This chapter presents what sample is used and how the corresponding data is collected. Thereafter, the dependent variables are discussed. Control variables are presented in the third paragraph. In conclusion, the fourth paragraph provides insights in which analyses are used and illustrates the model specifications.

3.1 Sample and data collection

A dataset of firms that are required to comply with the new regulations regarding the expanded audit report was provided by my thesis supervisor. The dataset initially contains 323 firms, which are listed on the AEX-index, AMX-index and ASCX-index in The Netherlands or the Financial Times Stock Exchange (FTSE) in the United Kingdom. The data provided comprises the years 2013 up to and including 2016. In addition, 2012 is added to the dataset to facilitate a comparison in audit quality, which is required to test the first hypothesis.

In addition, several companies were eliminated from the sample. First, companies where fundamental data is missing are excluded from the sample. Second, financial firms are eliminated from the sample due to their unique and complex nature (Reid et al., 2015). In sum, the final sample size is depicted in Table 1.

Table 1: Sample distribution

Initial number of firms 323

Financial institutions and other related firms 75

Firms with missing data 73

Number of firms after elimination 175

Number of years used in this study 5

Total number of observations 875

The data collected derives from two sources. As earlier stated, some data originates from my thesis supervisor. Secondly, data is obtained using public databases (e.g., Compustat and Orbis).

3.2 Dependent variables

This paragraph elaborates on the dependent variables of this research. Three hypotheses were formulated, each containing one dependent variable. In the first section, I will elaborate on audit quality. Secondly, the length of the audit report is discussed and finally, the number of key audit matters is amplified.

3.2.1 Audit quality

As stated previously in the theoretical framework, there is no uniform definition of audit quality. Similarly, there is no uniform way to measure it (Francis, 2011). Audit quality is multidimensional and unobservable. Many models have been developed to capture it. In this study, the modified Jones model (1991) is applied. In this approach, audit quality is proxied by absolute discretionary accruals. Originally, discretionary accruals are a proxy for earnings quality, and in academic research frequently used as a proxy for audit quality (Knechel et al., 2013). Audit quality cannot be easily separated from earnings quality, since discretionary accruals are highly subject to the probe of auditors, the proxy is suitable to reflect audit quality (Knechel et al., 2013).

(13)

This modified Jones model has been frequently used in prior literature in order to capture audit quality (e.g., Dechow, Sloan & Sweeney, 1995; Carcello & Li, 2013, Reid et al., 2015). At first, the model requires the computation of total accruals of a firm. Total accruals are represented by the following formula (Reichelt & Wang, 2010):

TACCi, t = ∆ CAi, t + ∆ CLi, t + ∆ Cashi, t + ∆ STDebti, t – D&Ai, t

wherein: TOT_ACCi, t = total accruals

∆ CAi, t-1 = current assets

∆ CLi, t = current liabilities

∆ Cashi, t = cash and cash equivalents

∆ STDebti, t = short term debt

D&Ai, t = depreciation and amortization

Subsequently, total accruals are scaled by total assets as prescribed by the model (Jones, 1991). The modified Jones model states that accruals consist of nondiscretionary accruals and discretionary accruals. The equation that can be derived from this assumption is illustrated below (Dechow et al., 1995):

TOT_ACCi, t = β0 (1/Ai, t-1) + β1 ∆ REVi, t + β2 PPEi, t + ei,t

wherein: TOT_ACCi, t = total accruals

Ai, t-1 = the natural log of total assets

∆ REVi, t = the change in revenues from prior year divided by total assets

PPEi, t = gross property, plant and equipment divided by total assets

ei, t = error term

Nondiscretionary accruals are represented by β0 up to and including β2 in the abovementioned model.

The discretionary accruals are identified by the error term (= ei, t). The error term is calculated by

running a regression analysis, and deriving the residuals from this model. Ultimately, the residuals represent the discretionary accruals. Following Reid et al. (2015), the absolute value of these residuals are calculated. Thus, the absolute discretionary accruals is the proxy for audit quality which is used in this study. Essential is that higher audit quality is evidenced by lower absolute discretionary accruals (Reid et al., 2015).

3.2.2 Length of the audit report

The dependent variable that is used in the second hypothesis, is the length of the expanded audit report. The expanded audit report requires additional disclosures. Therefore, the length of the audit report significantly increased. Since the audit report was effective as of financial year 2013, the hypothesis will test data from the years 2013 up to and including 2016. Thus, the total sample for this hypothesis is 700 unique firm observations. Length can be measured using several standards. For example, the number of pages or the number of sentences. In this study, the measurement that will be applied in testing the second hypothesis is the number of words.

3.2.3 Key audit matters

The third dependent variable needed to test the third hypothesis is the number of key audit matters. The regulations concerning the expanded audit report requires auditors to disclose audit risks that affected the scope of the audit the most. The precise definition of key audit matters was provided earlier, in chapter 2.

(14)

Likewise the second hypothesis, key audit matters are required since the implementation of the expanded audit report. Therefore, the data incorporated in testing the hypothesis comprises the years 2013 up to and including 2016.

3.3 Control variables

In this paragraph, the control variables that are incorporated in the hypotheses are presented. Control variables aid in eliminating external factors that could impact the dependent variables. The variables elaborated on in this paragraph address firm-specific factors, auditor characteristics, industry factors and country factors.

Firm characteristics can either affect the sophistication of the firm’s information system, or the reporting incentives of the management (Johnson, Khurana & Reynolds, 2002). The firm-level characteristics that are accounted for are: firm size, profitability, and leverage. Firm size affects the financial reporting process, since larger firms are subject to the demands of more stakeholders, often exposed to more risks and have access to higher quality information systems. Additionally, the clientele of larger firms generates more revenues from accounting firms, and might therefore also impact audit quality (DeAngelo, 1981b). Secondly, profitability can either enhance or deteriorate reporting. The resources a company earns can be invested in higher quality information systems, enhancing the quality of financial reporting. However, a firm that is performing well might feel less pressure to elaborate in more detail. Finally, leverage is used as a control variable. Leverage represents a factor that can affect the reporting incentives of the management, and thus financial reporting quality which is related to audit quality (Johnson et al., 2002).

The auditor characteristic that is accounted for is whether or not a company was audited by a Big 4-firm. Numerous studies have shown a positive relationship between audit quality and the use of a Big 4-firm (e.g., DeAngelo, 1981a, Francis, 2011). Therefore, this variable is included in the analysis as the possibility exists that it affects the dependent variables. This control variable was incorporated in the analysis by creating an indicator variable, in which the value 1 confirms the use of a Big 4-firm. Industry factors are also controlled for using two-digit SIC-codes. Studies have shown that companies that face similar risks tend to adopt comparable reporting strategies – resulting in a similar financial reporting quality. Additionally, audit quality has been proven to be subject to industry-specific knowledge of the auditor (Minutti-Meza, 2013). According to Minutti-Meza (2013), when auditors have industry-specific knowledge, a more firm-specific audit is performed. In turn, higher audit quality is established. Therefore, industry factors are incorporated as a control variable.

Finally, the dataset used in this study comprises of two countries: the United Kingdom and The Netherlands. The countries are similar to some extent, but differ from each other regarding regulatory oversight, culture and economic characteristics. Therefore, I control for country factors by incorporating a third indicator variable. The value 0 represents firms listed in the United Kingdom, and the value 1 represents firms listed in The Netherlands.

(15)

3.4 Testing and model specifications

In this paragraph, the manner in which the hypotheses are tested are elaborated on. The tests performed, and model specifications are presented. Each hypothesis is discussed separately below. Univariate analyses are applied to test the first hypothesis. The variable of interest in this test is, logically, absolute discretionary accruals. The first hypothesis is tested using a paired samples T-test and in addition, a repeated measures ANOVA. Both tests tend to examine whether the means in dependent samples differ. Thus, these analyses are able to establish whether an improvement or deterioration in audit quality has occurred. A distinction is made between data of firms in 2012, and data of firms for the years 2013 up to and including 2016. Both groups contain the same sample of firms but for different years, facilitating a comparison and can therefore be considered as related (or dependent) groups. In sum, the paired samples T-test and repeated measures ANOVA are suitable statistical approaches to test the first hypothesis.

In the second hypothesis, I aim to find a positive association between the length of the expanded audit report and audit quality. A linear regression analysis is performed to establish a (positive) relationship. Including the control variables, the following regression model is used:

|LENGTH_ARi, t| = β0 + β1 ABS_ACCi, t + β2 SIZEi, t + β3 PROFi, t + β4 LEVi, t + β5 BIG4i, t

+ β6 INDUSTRYi, t + β7 COUNTRYi, t

wherein: LENGTH_ARi, t = number of words of the expanded audit report

ABS_ACCi, t = total absolute discretionary accruals

SIZEi, t = the natural log of total assets

PROFi, t = profitability, measured as the return on assets

LEVi, t = leverage, measured as liabilities divided by equity

BIG4i, t = indicator variable, 1: non-Big 4 auditor

INDUSTRYi, t = industry fixed effects, based on two-digit SIC-codes

COUNTRYi, t = indicator variable, 0: United Kingdom and 1: The Netherlands

In the third hypothesis, which is similar to the second one, I aim to evidence that the number of key audit matters disclosed in the expanded audit report is positively associated with audit quality. Linear regression analysis is also applied in testing this hypothesis. The corresponding model is as follows:

|KAM_ARi, t| = β0 + β1 ABS_ACCi, t + β2 SIZEi, t + β3 PROFi, t + β4 LEVi, t + β5 BIG4i, t

+ β6 INDUSTRYi, t + β7 COUNTRYi, t

wherein: KAM_ARi, t = number of key audit matters in the expanded audit report

ABS_ACCi, t = total absolute discretionary accruals

SIZEi, t = the natural log of total assets

PROFi, t = profitability, measured as the return on assets

LEVi, t = leverage, measured as liabilities divided by equity

BIG4i, t = indicator variable, 1: non-Big 4 auditor

INDUSTRYi, t = industry fixed effects, based on two-digit SIC-codes

COUNTRYi, t = indicator variable, 0: United Kingdom and 1: The Netherlands

To give a more in-depth overview of the variables used in the testing of the hypotheses, a description of all variables can be found in appendix 1.

(16)

4. RESULTS

In this chapter, an overview is provided of the analyses and the corresponding results that were derived. The first paragraph covers general processes and adjustments that are required to establish a suitable dataset for the analyses to be performed. In the second paragraph, descriptive statistics of the dataset are provided. The third paragraph provides the results regarding the first hypothesis. The fourth paragraph comprises the results of the analyses that were performed in testing the second and third hypothesis.

4.1 Dataset adjustments

As stated in the previous chapter, the base for the dataset was provided by my thesis supervisor. In addition, data was collected using several databases. The reliability of the base for the dataset is secured by a student who is employed by the University of Groningen, and who checks all data before it is incorporated in the dataset. All data is hand collected by several students that were, prior to the start of the collection process, provided with instructions. The data collected consists of several variables subtracted from annual reports and the corresponding audit reports. In addition, the reliability of data that is collected through databases is safeguarded through random checks. A small sample of firms from the dataset was selected, and the variables were checked manually by collecting the corresponding annual reports and audit reports. No inconsistencies were detected. Therefore, the data obtained from the previously mentioned databases is assumed to be reliable.

Prior to the analyses of the data obtained, data is checked to assure that the general assumptions are met, which are required in order to conduct statistical analyses. First, variables were subject to the process of winsorizing. In this process, outliers in the dataset were detected. The lower and upper bounds of the variables were computed using the mean plus/minus three times the standard deviation. The outliers were given the corresponding lower and upper bound values. Winsorizing allows the results of analyses to be more reliable. When the outliers in a dataset are ignored, the results might provide a distorted view.

In addition, for several variables the natural logarithm was calculated, which is specifically important to continuous variables. The variable that was converted to a natural logarithm was firm size, as measured by total assets. The calculation of the natural logarithm for other continuous variables was deemed unnecessary, due to their relatively small distributions. Finally, for the continuous variables, it was tested whether the data was normally distributed. A non-normal distribution can cause distortions in the results of the analyses. After performing this test, it is concluded that the variables were relatively normally distributed.

4.2 Descriptive statistics and multicollinearity

The descriptive statistics regarding the variables incorporated in the dataset used in this study are shown in table 2. The descriptive statistics serve as a summary of the data used, and do not provide implications for testing whether the hypotheses can be accepted or rejected. The descriptive statistics provided are the sample size, the mean, the standard deviation, the minimum and the maximum.

(17)

Table 2: Descriptive statistics

N Mean St. dev. Min Max

Continuous variables

Absolute discretionary accruals 875 0,044 0,046 0,000 0,186

Length 700 2850,034 1310,425 567 8897 # KAMs 700 3,667 1,936 0 9 Size 700 21,702 1,464 13,336 23,990 Profitability 700 0,065 0,071 -0,170 0,297 Leverage 700 1,944 2,047 -4,202 7,480 Indicator variables Big 4-firm 700 0,989 0,106 0 1 Industry 700 42,62 21,823 10 99 Country 700 0,22 0,416 0 1

The total sample consists of 175 unique firms with observations divided over five years. Thus, the total number of observations is 875. However, the second and third hypothesis are tested with a sample of 700 observations. The year 2012 is only used as a comparison and solely applied in testing the first hypothesis, whether or not an improvement in audit quality has occurred.

Using the Pearson correlation analysis, all variables were checked for multicollinearity. This analysis provides insights in whether correlation exist between variables, and to which degree the variables are related. Controlling for multicollinearity is important, as the existence of multicollinearity can distort the outcomes of the analyses. The results of the Pearson correlation analysis can be found in appendix 2. Multicollinearity between variables exist when the coefficients exceed the value 0,7. The table shows a correlation between the length of the audit report and the number of key audit matters. However, this relationship can be logically explained, since more key audit matters requires a more lengthier audit report. The relationship will not endanger the outcomes of the analyses, since the variables are not applied in the same regression models. Therefore, the conclusions that will be drawn from the results will not be jeopardized.

4.3 Results from the univariate analyses

In this paragraph, the first hypothesis of this study is tested – whether the expanded audit report resulted in an improvement in audit quality as measured by absolute discretionary accruals. The hypothesis is tested using two different statistical measures, namely, the paired samples T-test and the repeated measures ANOVA.

4.3.1 Paired samples T-test

To give a general overview of the absolute discretionary accruals in the years to be tested, table 3 provides some general descriptive statistics. As stated earlier, lower absolute discretionary accruals evidence higher audit quality. Compared to the year which is controlled for, 2012, the mean value of absolute discretionary accruals has decreased.

Table 3: Mean differences in audit quality

Year Mean N 2012 0,056 175 2013 0,039 175 2014 0,037 175 2015 0,038 175 2016 0,048 175

(18)

The paired samples T-test is used to determine whether the mean in a related group is the same. This statistical procedure is suitable, since it enables one to compare a variable prior and post to a certain event. In this study, the event concerns a change in the regulatory environment. More specifically, compliance with the expanded audit report. The variable of interest is absolute discretionary accruals. First, I conducted a paired samples T-test per year compared to the control year, 2012. The results are shown in table 4.

Table 4: Yearly results T-test

Mean diff. T-statistic p

2012 - 2013 0,0165 3,008 0,003

2012 - 2014 0,0184 3,747 0,000

2012 - 2015 0,0178 3,289 0,001

2012 - 2016 0,0075 1,291 0,198

The first column provides the mean differences between absolute discretionary accruals in the corresponding year versus the control year. The second column presents the T-statistics, and the third column the p-values. An improvement in audit quality can be assumed if the p-value is below 0,05. The results imply an improvement in audit quality, since the significance levels are below 0,05 and a decrease in absolute discretionary accruals is shown earlier in table 3. An exception occurs in the comparison of 2012 and 2016. The significance level exceeds the established boundary, and therefore, an improvement in audit quality due to the expanded audit report cannot be assumed in 2016. In order to provide a more holistic understanding whether or not an improvement in audit quality has occurred, I divided the sample into two groups. The first group contained the absolute discretionary accruals of 2012 (i.e., audit quality prior to the expanded audit report), and the second group contained the absolute discretionary accruals of 2013 up to and including 2016. Again, a paired samples T-test was performed, providing the following results as depicted in table 5:

Table 5: Accumulated results T-test

Mean diff. T-statistic p

0,0151 5,565 0,000

The same conditions apply as discussed earlier for table 5. The first column shows that overall absolute discretionary accruals have decreased. Additionally, the p-value falls within the established boundary. Therefore, the paired samples T-test implies that audit quality has improved. Based on the paired samples T-test, hypothesis 1 can be accepted.

4.3.2 Repeated measures ANOVA

A similar analysis that is performed, is the one-way repeated measures ANOVA. This test allows to detect differences between related means. The test is especially suitable for changes in mean scores over multiple points in time, and for several groups in which the participants are equal. The results of the repeated measures ANOVA are shown below in table 6.

Table 6: Results repeated measures ANOVA

F-statistic p

Model 2,430 0,000

ID Company 2,410 0,000

ID Year 3,280 0,011

(19)

The firms in the sample were assigned to different groups based on the identification variable ISIN-number, and subsequently by year. The p-values are below 0,05 which means the results are considered to be significant. In essence, the analysis implies that the mean of absolute discretionary accruals differs. Since the descriptive statistics presented in table 4 illustrated a decrease in absolute discretionary accruals, and based on the results of the repeated measures ANOVA, an improvement in audit quality has occurred. Based on the repeated measures ANOVA, hypothesis 1 can be accepted.

4.4 Results from the regression analyses

In this paragraph, the results from the regression analyses are presented. These were conducted in order to test the second and third hypothesis. Each table represents the results of one hypothesis. In model 1, the regression results are presented for the dependent variables and all control variables. Model 2 represents the entire regression model, including the dependent variable, independent variable and control variables.

4.4.1 Length of the audit report

In order to determine whether the length of the audit report is a determinant of audit quality, a multiple linear regression analysis is performed. The results of the first linear regression, testing whether the length of the audit report is positively associated with audit quality, are shown in table 7.

Table 7: Results linear regression H2

Model 1 p Model 2 p Constant -864,631 0,294 -787,440 0,354 Discretionary accruals -376,919 0,708 Size 194,289*** 0,000 191,390*** 0,000 Profitability -2854,564*** 0,000 -2848,589*** 0,000 Leverage 56,652*** 0,009 57,535*** 0,008 Big 4 -111,399 0,780 -111,822 0,779 Industry 0,110 0,955 0,086 0,965 Country -250,112** 0,025 -248,876** 0,026 Adjusted R2 _0,154 _0,152

* significant at the 10% level (= p < 0.1) ** significant at the 5% level (= p < 0.05) *** significant at the 1% level (= p < 0.01)

In model 1, it is shown that the control variables clarify 15,4% of the dependent variable. When the independent variable, absolute discretionary accruals is added the explanatory power of the model decreases with 0,2%. Subsequently, the model clarifies 15,2% of the dependent variable, as absolute discretionary accruals are incorporated. Implying that the length of the audit report does not improve the explanatory power of the model. The dependent variable in the above model is the length of the audit report, as measured by the number of words. The variable of interest is discretionary accruals and is added to the regression in model 2. The result provides an insignificant value. Therefore, the second hypothesis in this study cannot be accepted. The value implies that audit quality does not affect the length of the audit report. Thus, a more lengthy audit report does not imply higher audit quality. In conclusion, the second hypothesis is rejected.

Table 7 also provides results on the control variables that were incorporated in the model. Size, profitability and leverage are statistically significant. In turn, the positive beta-values regarding size and leverage imply a positive association with the length of the audit report. Subsequently, the larger a firm is, the more lengthier its audit report is. Additionally, the higher a company’s leverage is, the more lengthier its audit report is. Both can be explained by the fact that larger firms and more

(20)

leveraged firms are exposed to higher risks. Profitability is negatively associated with the length of the

audit report. The more profitable a company is, the less lengthier the audit report is. In turn, the less profitable a company is, the more lengthier its audit report. This can be explained by the fact that companies who might face the risk of distress are apparent of more auditor disclosures. Since the auditor is also required to impose explanations on the continuity of an organization. Also, the dummy variable is a significant result with negative values. Implying, that auditors/firms located in The Netherlands issue less lengthier report and that auditor reports in the United Kingdom are longer.

4.4.2 Number of key audit matters

The third hypothesis, whether the number of key audit matters is positively associated with higher audit quality, was also tested with a multiple linear regression model. The results of this regression model are shown in table 8.

Table 8: Results linear regression H3

Model 1 p Model 2 p Constant -3,845*** 0,000 -3,163*** 0,003 Discretionary accruals -3,328*** 0,009 Size 0,396*** 0,000 0,370*** 0,000 Profitability -1,588** 0,048 -1,535* 0,054 Leverage 0,086*** 0,002 0,094*** 0,001 Big 4 -0,836 0,100 -0,840* 0,097 Industry 0,005* 0,062 0,004* 0,074 Country 0,190 0,180 -0,179 0,204 Adjusted R2 _0,212 _0,219

* significant at the 10% level (= p < 0.1) ** significant at the 5% level (= p < 0.05) *** significant at the 1% level (= p < 0.01)

The explanatory power of model 1 is 21,2%, implying that to that extent the control variables clarify the dependent variable. When the variable of interest, absolute discretionary accruals is added, the value increases to 21,9%. The slight increase of 0,7% reflects the explanatory power of the independent variable absolute discretionary accruals on the number of key audit matters. The dependent variable in the above model is the number of key audit matters in the audit report. The variable of interest is discretionary accruals is significant. The value confirms a linear relationship between the number of key audit matters and audit quality. The β-value implies whether the relationship is positive or negative. The β-value for discretionary accruals is negative. However, less absolute discretionary accruals imply higher audit quality. A negative relationship between absolute discretionary accruals and the number of key audit matters is equal to a positive relationship between audit quality and the number of key audit matters. Thus, the third hypothesis is accepted.

Table 9 also presents the results on the control variables in relation to the number of key audit matters in the expanded audit report. Size, profitability, leverage, and industry are significant. Size and leverage are positively associated to the dependent variable. In turn, this means that the larger an organization is and the higher its leverage, the more key audit matters it discloses. Profitability is negatively associated with the number of key audit matters. Thus, the less profitable an organization is, the more key audit matters are disclosed and vice versa. Additionally, industry factors are also significant, which in turn means that the number of key audit matters vary per industry.

(21)

5. CONCLUSION & DISCUSSION

This chapter concludes the theories, hypotheses and findings that were elaborated on in this study. An overview is given of which hypothesis are accepted or rejected and a conclusion is formed in the first paragraph. Subsequently in the second paragraph, the limitations that were subject to this study are presented. In the third paragraph, some directions for future research are provided.

5.1 Conclusion

The main aim of the issuance of the expanded audit report was to increase transparency amongst the auditing profession. By incorporating several theories and findings from academic research, the possibility was postulated that an increase in transparency could lead to an improvement in audit quality. Prior research performed by Reid et al. (2015) supported this postulation with empirical data covering one year after the issuance of the expanded audit report. The aim of this study was twofold. First, to determine whether the implementation of the expanded audit report is associated with an improvement in audit quality. Second, to determine whether certain characteristics of the expanded audit report are determinants for audit quality (i.e., the length and the number of key audit matters). The study finds that the expanded audit report is associated with a significant improvement in audit quality in line with Reid et al. (2015). The proxy for audit quality, absolute discretionary accruals, shows a declining value in multiple years following up on the implementation year, reflecting higher audit quality. The assumption was made that the expanded audit report increases the perception of accountability, and therefore impacts audit quality (Deis & Giroux, 1992; Carcello & Li, 2013; Peecher et al., 2013). As stated by the PCAOB (2016), the expanded audit report incentivizes auditors to collect more proper audit evidence to substantiate the assumptions made in the audit report. In turn, a more adequate judgement of a firm’s financial statement is made. This study has evidenced these expectations, and shows a significant improvement in audit quality. Therefore, the first hypothesis was accepted.

Additionally, the expanded audit report provides auditors with opportunities to legitimize the auditing profession, and enhance public confidence. Prior studies have shown that the length and level of detail of disclosures are determinants for legitimacy (Deegan et al., 2000; Campbell, 2004; Coram et al., 2011). Therefore, the length of the expanded audit report and number of key audit matters in the expanded audit report were examined. The study did not find a significant association between the length of the expanded audit report and audit quality. Therefore, the second hypothesis was rejected. The results did find a significant positive relationship between the number of key audit matters and audit quality. Key audit matters provide a relatively high level of detail about an auditor’s considerations during an audit, and therefore impacts audit quality. The paragraph about key audit matters in the expanded audit report might increase the expectations of users of the audit report, which pressures auditors to perform additional audit activities in order to meet these expectations. The length of the audit report also concerns general information, and in itself does not necessarily provide a high level of detail about the audit. The expanded audit report is principles-based, the information that auditors provide that specifically applies to the audit is dependent on an auditor’s judgement. More general and less audit-specific information result in less expectations amongst users, and in turn, this might clarify why no significant relationship is found between the length of the expanded audit report and audit quality.

(22)

In sum, the expanded audit report is deemed to have a positive effect on audit quality, and the key audit matter paragraph provides a determinant for the level of audit quality. However, emphasis must be placed on the main aim of the expanded audit report. Regulators attempted to solely increase the level of transparency. The scope of the audit, and other requirements affecting auditor efforts have not changed by issuing this new standard. The implication that this study provides is that when auditors have to justify and clarify their efforts, their performance improves.

5.2 Limitations

The main limitation of this study relates to the variable audit quality. As stated previously, audit quality is an abstract subject. It is unobservable, there is no uniform definition and no uniform approach to measure it (Francis, 2011). In recent years, regulators and practitioners aim to improve audit quality by issuing new regulation and development programs. The concept is thus affected by many events. In turn, the improvement in audit quality that is established in this study is perhaps not a sole result of the expanded audit report. To state that the improvement in audit quality is a direct consequence of the expanded audit report might be irrational. However, the expanded audit report is deemed to have contributed to an improvement in audit quality. Future research can incorporate a control group that is not obliged to comply with the expanded audit report, and can assess the improvement in audit quality based on these differences.

The second limitation of the study is the proxy that is used to measure audit quality. Absolute discretionary accruals are frequently used in academic research, but has also been criticized. An example of criticism is that absolute discretionary accruals are highly sensitive to measurement errors (Knechel et al., 2013). For future references, a second proxy for audit quality could be included to confirm the robustness of the results.

The third and final limitation of this study can be subjected to the fact that firms from two different countries are used in this study. In itself, this is not problematic (Reid et al., 2015). A control variable was incorporated in the analyses to account for these differences. The results were however not interpreted per country, and conclusions have been drawn on the whole sample. The conclusions should therefore still be interpreted in light of these differences. Future research might incorporate a comparison between both countries to provide a more holistic view of the country differences.

5.3 Future research

In the final paragraph of this study, a few directions for future research are provided. The first possibility for future research is based on an assumption that was made in this study. That assumption formed an important aspect in the development of the theoretical framework, namely the accountability theory. That the expanded audit report actually increases accountability is not substantiated with empirical evidence. Examining whether auditors experience an increase in accountability can therefore be an interesting research opportunity.

A second opportunity for future research concerns the direction of research. The study examined whether the expanded audit report affected audit quality, based on the assumption that auditor efforts increase due to a higher level of accountability. Auditor efforts in itself were not incorporated in this study. Future research could take auditor efforts into account, by for example, incorporating the abnormal audit fees model by Ferguson, Francis and Stokes (2006).

(23)

REFERENCES

AFM. (2010). Rapport algemene bevindingen kwaliteit accountantscontrole en kwaliteitsbewaking. Assessed on October 17, 2017 from:

https://www.afm.nl/~/profmedia/files/rapporten/2010/kwaliteit-accountantscontrole-kwaliteitsbewaking.ashx

AFM. (2014). Uitkomsten onderzoek kwaliteit wettelijke controles Big 4-accountantsorganisaties. Assessed on October 17, 2017 from:

https://www.afm.nl/~/profmedia/files/rapporten/2014/onderzoek-controles-big4.ashx

AFM. (2017). Kwaliteit OOB-accountantsorganisaties onderzocht. Assessed on October 17, 2017 from: https://www.afm.nl/~/profmedia/files/rapporten/2017/oob/deel-1.ashx

Antle, R. (1982). The auditor as an economic agent. Journal of Accounting Research, 503-527.

Asare, S. K., Trompeter, G. M., & Wright, A. M. (2000). The effect of accountability and time budgets on auditors' testing strategies. Contemporary Accounting Research, 17(4), 539-560.

Asare, S. K., & Wright, A. M. (2012). Investors', auditors', and lenders' understanding of the message conveyed by the standard audit report on the financial statements. Accounting Horizons, 26(2), 193-217.

Ashton, R. H. (1992). Effects of justification and a mechanical aid on judgment performance. Organizational Behavior and Human Decision Processes, 52(2), 292-306.

Bazerman, M. H., Loewenstein, G., & Moore, D. A. (2002). Why good accountants do bad audits.

Harvard business review, 80(11), 96-103.

Brasel, K., Doxey, M. M., Grenier, J. H., & Reffett, A. (2016). Risk disclosure preceding negative outcomes: The effects of reporting critical audit matters on judgments of auditor liability. The Accounting Review, 91(5), 1345-1362.

Brown, T., Majors, T., and Peecher, M. (2015). The Influence of Evaluator Expertise, a Judgment Rule, and Critical Audit Matters on Assessments of Auditor Legal Liability. Working paper, University of Illinois at Urbana–Champaign.

Cameran, M., Prencipe, A., & Trombetta, M. (2016). Mandatory audit firm rotation and audit quality.

European accounting review, 25(1), 35-58.

Campbell, D. (2004). A longitudinal and cross-sectional analysis of environmental disclosure in UK companies—a research note. The British Accounting Review, 36(1), 107-117.

Carcello, J. V., & Li, C. (2013). Costs and benefits of requiring an engagement partner signature: Recent experience in the United Kingdom. The Accounting Review, 88(5), 1511-1546.

Carpenter, B., & Dirsmith, M. (1993). Sampling and the abstraction of knowledge in the auditing profession: An extended institutional theory perspective.Accounting, Organizations and

Society,18(1), 41-63.

Carver, B. T., & Trinkle, B. S. (2016). Investors’ Reactions to the PCAOB’s Proposed Changes to the Standard Audit Report. Working paper, Clemson University, Mississippi State University.

Christensen, B. E., Glover, S. M., & Wolfe, C. J. (2014). Do Critical Audit Matter Paragraphs in the Audit Report Change Nonprofessional Investors’ Decision to Invest?

The expanded audit report: