The effect of intergrated reporting quality on audit fees : evidence from a South African setting?

The effect of Integrated reporting quality on audit fees, evidence from a South African

setting?

Submitted in partial fulfilment of the requirements

for the degree of Master of Science in Accountancy in the Graduate College of the University of Amsterdam 2018

AMSTERDAM, The Netherlands

STATEMENT OF ORIGINALITY

This document is written by Martijn Kanter who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is

original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the

ABSTRACT

The purpose of this thesis is to examine the relationship between integrated reporting quality and the level of audit fees within South African top 100 listed-companies. The relationship

predicted is based on integrated thinking, which increases the reliability of organisational information, provides an enhanced view on the organisational risks and a better internal

control system. The integrated reporting quality is measured via the EY Excellence in Integrated Reporting Awards. This thesis does not find evidence that integrated reporting

ACKNOWLEDGMENTS

I would like to acknowledge the internship provided to me by Ernst & Young LLP, and express my gratitude for a chance of academic and professional development. I will strive to use gained skills and knowledge in the best possible way. Finally, I must express my profound

gratitude to my family for providing me with everlasting support and continuous encouragement throughout my years of study. This accomplishment would not have been

possible without them. You have my deepest respect, thank you.

Content

1. Introduction………... 6

1.1 Background………... 6

1.2 Research Question and Contribution………..7

2. Literature review………. 9

2.1 The integrated reporting Framework………. 9

2.1.1. The fundamental Concepts………. 9

2.1.2. Guiding Principles ………..11

2.1.3. Content Elements ………..12

2.1.4. Assurance on Integrated Reporting……….. 13

2.1.5. Integrated Reporting Quality……… 13

2.2 Audit Fees………..……… 14

2.2.1. Influencing factors………..………..… 14

2.2.2. Audit Risk model………..………..… 17

3. Hypothesis Development………..………..… 18 4. Research Design………..……….………..… 20 4.1 The model……….………. 20 4.2 The Variables……….……….. 21 4.3 Data Construction……….……….. 23 5. Empirical Results……….……… 24 5.1 Descriptive Statistics……….………. 25 5.2 Correlation Matrix……….……….……. 26 5.3 Regression Analysis……….……….……….. 27

5.4 Robustness Test and Additional Test……….……….……… 30

6. Conclusion and limitations……….……….……….……….. 32

6.1 Summary ……….……….………. 32 6.2 Conclusion……….……….……….. 32 6.3 Limitations……….……….……….. 34 6.4 Future research……….……….……….. 35 References……….……….………. 36 Appendixes………….……….……….…….. 39

1. Introduction

1.1 Background

The past two decades have seen substantial development in the academic literature on accounting and accountability systems for the combined management, reporting of financial and non-financial performance (De Villiers, Rinaldi and Unerman, 2014). There is growing evidence supporting the claim that traditional corporate reporting is struggling to meet the information need of stakeholders. (Cohen and Holder-Webb, 2012). This struggle is underpinning information asymmetry which ultimately obstructs the efficient allocation of funds on the investment market. Consequently, an efficient allocation of funds on the investment market requires an adequate information disclosure towards investors (Healy & Palepu, 2001). Organisations therefore sequentially seek new opportunities to provide more information towards stakeholders by providing additional reports containing non-financial information next to the existing financial reports (Cohen et al., 2012). Finding that are confirmed by Adams, Fries and Simnett (2011), who indicated that current providence of information by organisations do not meet requirements of different stakeholders. Recent years have shown a rapid increase in the publication of sustainability reports, the percentage of organizations publishing sustainability reports grew dramatically from 20% to 80% over the last five years (Ioannou & Serafeim 2017).

Within this process a movement arose linking financial and non-financial information, leading to the introduction of Integrated Reporting by the International Integrated Reporting Council (IIRC). The IIRC was established in 2010 as a response to the global financial crisis, with a directive to develop a framework for integrated reporting and endorse the use in practice. Eventually published in 2013, The IIRC Framework stated that the prime objective of an integrated report is to describe value created over time for the investors of an organization. The IIRC’s long-term vision is a world in which integrated thinking is embedded within the mainstream business practice in the public and private sectors, facilitated by Integrated Reporting as the corporate reporting norm (IIRC, 2013 p.2). Integrated thinking leads to a better understanding of the firm’s risks, which could lead to a better and more complete risk management system (Moolman, Oberholzer, & Steyn. (2016).

The IIRC Framework is rapidly gaining popularity. Currently, more than 1500 organizations prepare an integrated report, including world-leading organizations as BP, Coca-Cola, HSBC, Tata Steel, Mitsubishi, Phillips, Unilever. The framework is supported by China's Ministry of Finance,

and the Securities and the Exchange Board of India (SEBI) urged listed firms to implement integrated reporting. In addition, Richard Howitt, Chief Executive Officer of the IIRC states that the EU’s non-financial reporting directive is seen as rapprochement towards integrated reporting European Union. Moreover, pioneering legislators in South Africa made the publication of integrated reports mandatory (King III), public corporations are obligated to publish their annual reports via the IIRC Framework. However, the Johannesburg Stock Exchange does not require organizations to provide an assurance statement for the integrated report. Due to the absence of a standard format and a principle-based framework, significant differences exist in the quality of integrated reports (IRQ) between organizations (Pistoni, Songini & Bavagnoli 2018). Although the audit quality has received enormous attention by the mass media, practitioners, firms, regulators and scholars, especially after the financial crisis of 2008 (Demartini & Trucco, 2017), little attention is given to the IRQ despite the rapidly gaining use practice and popularity.

1.2 Research question and contribution

To further understand the added value of integrated reporting for businesses, the fundamental question of what the costs and benefits of integrated reporting are, need to be answered. However, as stated by Cheng, Green, Conradie, Konishi, and Romi in 2014, little was known about the cost and benefits perceived by stakeholders and organizations preparing them because of the emerging character of integrated reporting. In the last three years, a broad interest of academia and a growing number of publications in the field of integrated reporting lead to a more comprehensive understanding of the benefits associated with integrated reporting. Current literature provides evidence for a positive relationship between the use of integrated reporting by an organization and the firm value (Lee & Yeo, 2016), improved internal ecision making and liquidity, capital market effects (Barth, Cahan,, Chen, and Vente,. 2017), a more correct prediction of earnings by analysts and reduction in the cost of equity capital (Zhou et al. 2017). Factors, which may provide an incentive for organizations to adopt the use of integrated reporting, in addition to the existing challenge of seeking improved channels to provide information towards stakeholders. Surprisingly, still little is known about the costs related to adopting integrated reporting, specifically between the quality of integrated reporting and the level of audit fees. A startling insight in light of the paper of Pistoni et al. (2018), who essentially came to the conclusion that a great difference exist between the quality of integrated reports. Therefore this thesis aims to clarify the costs associated with integrated reporting quality.

This thesis attempts to further our understanding of Integrated Reporting Quality (IRQ) explicitly, through examining the relationship between the IRQ and the effect on audit fees. More precisely, this thesis is based on research conducted by analysing data, partly hand collected from South Africa, over a period between 2013 and 2017. Attempting to answer the following research question:

RQ: How does the Integrated Reporting Quality affect the level of audit fees.

Gaining insight into this topic by providing an answer to this research question is relevant since organizations are currently in search of an improved way to provide information towards stakeholders. In line with this movement, the International Integrated Reporting Council (IIRC) introduced Integrated Reporting and is aiming to set it as standard within the field, currently achieving this goal by actively gaining support within the practice.

To begin with, this thesis answers directly to the papers of de Villiers et al. (2014), Cheng et al. (2014) and Velte & Stawinoga (2017) providing a pivotal understanding to proposed future research questions in the field of integrated reporting. Therefore empirically contributes to this strand of literature by providing insight into the new area of integrated reporting and complements existing literature, providing evidence of the costs related to IRQ. Research encouraged by Hay, Knechel and Wong (2006), which acknowledged further is required on issues related to audit fees, such as the quality of integrated reporting. This is in line with the outcomes of Cheng et al. (2014), de Villiers et al. (2014), Simnett and Huggins (2015) and Dumay et al. (2016), who stressed that assurance needs to be provided on the content of IR. To fundamentally evaluate the assurance quality on integrated reporting, the integrated reporting quality itself needs to be to be further investigated. Secondly, the results can be of interest to managers, investors, standard-setters, regulators to assess the costs aspect related to the integrated reporting quality, since currently little is known between the relationship of integrated reporting quality and the level of audit fees.

2. Literature review:

2.1 The integrated reporting Framework

The IIRC describes itself as ‘a global not-for-profit coalition of regulators, investors, organizations, standard setters, the accounting profession and NGOs', with the mission to set integrated reporting within mainstream business practice as the standard in the business. To achieve this mission, the IIRC introduced the integrated reporting framework, a document that elaborates on a organizations financial and non-financial performance, with a special focus on environmental, social, and governance performance (The IRRC, 2017). Organisations can use the integrated reporting principles-based framework to measure value creation in the so-called ‘capitals.’

The IIRC defines integrated reporting as follows: ‘’An integrated report is a concise communication about how an organization’s strategy, governance, performance, and prospects, in the context of its external environment, lead to the creation of value over the short, medium and long-term.’’ (IIRC, 2013, p.7)

Integrated thinking can lead to improved decisions at the management and board levels as well as more focused reporting and better communication with the firm's stakeholders (Barth, Cahan, Chen & Venter 2017).

Although the momentum of the IIRC and its launched integrated reporting made an enormous leap forwards the last couple of years, the final adoption, and standardization of integrated reporting is still more than uncertain and therefore an intriguing process (Humphrey, O‟Dwyer and Unerman, 2017). A so-called ‘breakthrough moment' is happing according to the IIRC itself, either the integrated reporting is going to be the standard reporting mechanism or is losing it momentum to alternatives. Moreover, the IIRC is excessively populated and financed by accounting professions itself, a salient detail that awakes attention within the context.

2.1.1. The fundamental Concepts

As described in the section above, capitals and value creation are fundamentals within the basic concept of integrated reporting. An integrated report aims to provide insight about the resources and relations used and affected by the organisation, these are collectively referred

to as ‘’the capitals’’ (IIRC, 2013, p.4) Organisations can use the principles-based framework to measure value creation of; Financial (e.g., equity or debt), Manufactured (e.g. produced services or goods), Intellectual (e.g. knowledge-based intangibles), Human (e.g. human capital), Social and relationship (e.g. relationships with communities) and Natural (e.g. environmental resources). An integrated report differs from other reports since it focusses on the short, medium and long-term value creation, and underlines on the significance of integrated thinking within the organization. When fully implemented integrated thinking improves the decision-making process, the communication, and the information within the organization.

Figure 1: The Value Creation Process (IIRC, 2013, p.13)

Figure 1. shows that each capital is required to create value for the organization. The input and output of capitals are financial, manufactured, intellectual, human, social & relationship and natural, which are influenced by the mission and vision of the organization. The IIRC emphasizes on disclosing the relationship between the different capitals to gain insights into the opportunities and threats for the organization. Consequently, the objective of integrated reports is to disclose all resources and relationships that materially impact the value-creation activities of the firm in the short, medium, and long terms (Cohen & Simnett, 2015).

To determine the materiality in an integrated report the concept of reporting boundary is vital. Figure 2. shows the entities and stakeholder that need to be considered in determining the reporting boundary. Within this process there are two aspects: The financial reporting entity

and the Risks, opportunities, and outcomes attributable to or associated with other entities/stakeholders beyond the financial reporting entity that has a significant effect on the ability of the financial reporting entity to create value (IIRC, 2013, p.19).

Figure 2: Entities considered in determining the reporting boundary: (IIRC, 2013, p.13)

2.1.2. Guiding Principles

The following guiding principles are given by the Integrated Reporting Framework in order to effectively produce an Integrated Report (IIRC, 2013 p.6):

Strategic focus and future orientation: an integrated report should provide insight into the organization’s strategy and how that strategy relates to the organization’s ability to create value in the short, medium, and long term and its use of and effects on its forms of capital. Connectivity of information: an integrated report should show, as a comprehensive value creation story, the combination, interrelatedness and dependencies among the components that are material to the organization’s ability to create value over time.

Stakeholder responsiveness: an integrated report should provide insight into the quality of the organization’s relationships with its key stakeholders and how and to what extent the

organization understands, takes into account, and responds to their legitimate needs, interests, and expectations.

Materiality and conciseness: an integrated report should provide concise information that is material to assessing the organization’s ability to create value in the short, medium, and long term.

Reliability and completeness: an integrated report should include all material matters, both positive and negative, in a balanced way and without material error.

Consistency and comparability: the information in an integrated report should be presented on a basis that is consistent over time and in a way that enables comparison with other organizations, to the extent it is material to the organization’s own value creation story.

2.1.3. Content Elements

To effectuate the preservation and enlargement of multiple forms of capital and the long-term value creation by the firm, the IIRC underlines the relevance in disclosing the relationship between different classes of capital in the value creation process (White, 2010). The following content elements are given by the Integrated Reporting Framework in order to effectively capture the value creation process in the Integrated Report (IIRC, 2013 p.6):

Organizational overview and external environment: What does the organization do and what are the circumstances under which it operates?

Governance: How does the organization’s governance structure support its ability to create value in the short, medium and long term?

Business model: What is the organization’s business model?

Risks and opportunities: What are the specific risks and opportunities that affect the organization’s ability to create value over the short, medium and long term, and how is the organization dealing with them?

Strategy and resource allocation: Where does the organization want to go and how does it intend to get there?

Performance: To what extent has the organization achieved its strategic objectives for the period and what are its outcomes in terms of effects on the capitals?

Outlook: What challenges and uncertainties is the organization likely to encounter in pursuing its strategy, and what are the potential implications for its business model and future performance?

Basis of presentation: How does the organization determine what matters to include in the integrated report and how are such matters quantified or evaluated?

2.1.4. Assurance on Integrated Reporting:

The absence of a standard format and a principle-based framework lead not only to significant differences in the quality of integrated reports but also to variations on the assurance provided (Pistoni, Songini & Bavagnoli 2018). Current literature suggests that auditing and assurance have not been of interest to the academic world since the research is more focused on the external reporting issues (Dumay et al. 2016). However, as described before, assurance of integrated reporting is currently evolving (Velte and Stawinoga 2017). KPMG (2013) and Cohen and Simnett (2014) found evidence that the assurance corporate social responsibility reports are developing increasingly. This is in line with the outcomes of Cheng et al. (2014), de Villiers et al. (2014), Simnett and Huggins (2015) and Dumay et al. (2016), who stressed that assurance needs to be provided on the content of IR. To fundamentally evaluate the assurance quality on integrated reporting, the integrated reporting quality itself needs to be to be further investigated.

2.1.5. Integrated Reporting Quality (IRQ)

IRQ can be defined as the level of insight the integrated report provides to a user on the organization strategy and value creation process. As described in the introduction, differences exist in the quality of integrated reports (IRQ) between organizations (Pistoni, Songini & Bavagnoli 2018). Those differences originated from the absence of a standard format and a principle-based framework which enables the organizations to deviate in their integrated report. Moreover, assurance of the non-financial information of integrated reports is not compulsory.

Although the concept of audit quality has been discussed extensively in the current academic literature, little is known about the IRQ and perceived effects. Available research shows that a better integrated reporting quality is associated with higher investment efficiency and greater analysis target price forecast accuracy (Barth et al. 2017), results that support the objective the IIRC to improve external information and better internal decisions. Lee and Yeo (2016) and Zhou et al. (2017) both claim positive economic benefits as an effect of higher IRQ, which indicates companies can attribute value to the perceived quality of their integrated reports.

2.2. Audit Fees

Audit fees are charged by the external auditor, who is appointed with the objective to examine the organization's statements independently. The statements are examined which either they show a true and fair view in all material respects of the company's financial position, results, and cash flows, in conformity with international generally accepted accounting principles (GAAP). Jensen and Mackling (1976) research on the agency problem indicates that the external audit is a control mechanism, which mitigates issues that occur as an effect of the separation of ownership and management. In current academic literature, the topic of audit fees is extensively covered. The majority of research is based on the influential work by Simunic (1980), who conducted a study on the auditor attributes and number of clients in relation to a higher or lower audit fee.

2.1.1. Influencing factors

In the section below I will briefly discuss those factor associated to influence audit fees, as proven by results of the prior scientific examination. Starting with the three most predominant factors.

Organisational size, risk and the complexity

The three most primarily factors associated to have an influence on audit fees are the organizational size, risk and the complexity (Hay et al. 2006). The organizational size is commonly measured in total assets or converted into the natural logarithm of this number to improve linear the relationship with audit fees. Another way to measure the organisations size is via the operating cash flow. Audit fees have a proven positive relationship regarding organizational size, where organizational size can predict the variation on an average of 70%. (Bell, Knechel, and Willingham 1994). As stated in the overview paper by Hay et al. (2006), 87 studies that included assets as a control variable for size, find with all but 2 having a significant

positive coefficient. Prior research has showcased that a more complex client is associated with a harder audit and therefore with a more time-consuming audit (Simunic, 1980). However the complexity of organizations have been measured differently over time and studies, common indicators complexity are the variables foreign-controlled assets, the number of subsidiaries, proportion of foreign assets and industry classification. The relationship between organizational complexity and audit fees are significant and positive, as shown in the results of Hay et al. in 2006. Although the measurement differs across the various studies, empirical evidence overwhelmingly supports a positive relationship between complexity and audit fees. The last prime factor positively associated with audit fees is inherent risk. The audit is perceived as riskier by the auditor who therefore requires specialized audit procedures, to lower the level of audit risk. Three variables are commonly used to measure an organization's inherent risk, namely; inventory divided by total assets, receivables divided by total assets, and the combination of inventory and receivables divided by total assets (Hay et al., 2006).

In addition to thefactors size, risk and the complexity, prior research indicates other factors which influence audit fees, which will be discussed in the paragraph below.

Profitability

Profitability has a proven to have a positive relationship regarding audit fees, and it reflexes the possibility that the auditor might be exposed to a loss in case of bankruptcy by the client (Simunic, 1980). The lower the profitability of an organization, the higher the associated risk exposure to the auditor is. Low performance of the organization, therefore, leads to a higher audit fee charged by the auditor. Variables that are commonly used to measure a organizations profitability are the ratio of return on assets, return on equity (measured as the net income divided by total assets) and a dummy variable indicating a loss or profit disclosed by the organization in the financial statements (Hay et al. 2006). The expected influence of the return on equity (ROE) on audit fees is negative since a higher ROE indicates a lower risk for the auditor, where in contrast the loss dummy is associated with higher audit fees.

Leverage

Leverage is an indicator of the level of debt financing and therefore is linked to the risk of an organizations bankruptcy, which would imply a loss for the auditor (Simunic 1980). Researchers find an effect between the level of leverage and the audit fees an organization pays to the auditor. Variables that are commonly used to measure a organizations leverage are the ratio

of debt to the total assets (leverage) and the quick ratio (Hay et al. 2006). The ratio of debt to the total assets is associated with a higher audit fee, where the relation between the quick ratio and audit fees is found to be negative. In addition, leverage is also measurable as the ratio of debt to equity (equity-debt ratio).

Governance

Corporate governance deals with the effective implication of a control-environment and therefore had an adverse effect on the audit fees. The better the corporate governance of an organization is in place, the less work has to be performed by the external auditor. Measurements of corporate governance make use out of the following commonly used variables, board independence (Tsui, Jaggi, and Gul 2001), audit committee expertise (Krishnan and Visvanathan 2009), and audit committee board tenure (Chan, Liu, and Sun 2013).

Internal Control

Another factor associated with audit fees is the internal control. Internal control verifies the reliability of the financial accounting disclosure and safeguards internal assets from intentional losses and non-intentional losses. A robust internal control system enhances the reliability of information, which is affected by the balance and freedom of materiality (IIRC).

In line with the research of Hay et al. (2006) who showed that client risk and audit fees are positively related, Hogan & Wilkings (2008) found a negative relationship between audit fees and internal control. Internal control risk can be defined as the chance that the internal control system that does not deal with all material misstatements (Hogan & Wilkings 2008). Moreover, Hoitash et al. (2008) found that control problems indicate lower risk management quality, which is related to higher audit fees. This relationship can be explained as the external auditor has to perform more tasks to provide assurance when the internal controls are not in place, leading to a higher audit fee (Jiang et al. 2015). To understand the influence of internal control risk on the audit fee, the audit risk model can provide further insight.

2.2.2. Audit risk model

The audit risk model provides a model to understand the relationship between the audit risk, inherent risk, control risk, and detection risk. In the Statement on Auditing Standards (SAS) No. 47 by American Institute of Certified Public Accountants, the audit risk model is described as follows:

Audit Risk (AR) = Inherent Risk (IR) x Control Risk (CR) x Detection Risk (DR)

Audit risk (AR) is set by the auditor itself, which can be defined as the willingness to accept an unqualified opinion is provided on a materially misstated financial statement. Inherent risk the probability of material misstatement before taking in regarding the usefulness of the internal control system. This material misstatement can be on an account balance or class of transaction. The control risk is the probability of material misstatement which is not detected or prevented by the internal control system, this risk is assessed by the auditor. The last variable in the audit risk model is detection risk, which is the level of risk that the auditing procedures will not detect all material misstatements.

In line with the audit risk model, prior research finds that audit fees are significantly higher for firms that have reported internal control deficiencies (Hogan & Wilkings, 2008). The same study shows that audit fees are higher when internal control problems have been discovered by the auditor itself and that the inherent risk and information risk is estimated higher than their industry counterparts.

3. Hypothesis development

For the hypothesis development in this thesis, we draw on is the work of Wagenhofer (1990) full-disclosure equilibrium. In this paper, it is stated that private information, which is valuable to both the financial market and opponents, will always be always disclosed. Consequently, organizations are willing to provide information towards the outside world. In line with this full-disclosure equilibrium, the legitimacy theory of Shocker & Sethi (1973) explains that organizations have a social contract with their stakeholders, and have to comply with a society’s norms, values and boundaries by implementing adequate structures, internal controls and processes (Dowling & Pfeffer 1975). Consequently, the going concern and existence of the organizations rest on the ability to meet society’s expectations via suitable systems. By implementing effective internal controls, the reliability of the organizational information will be enhanced (IIRC 2017), and the quality of the integrated report will be higher. Literature shows that integrated reports with clear connectivity between the corporate governance and the financial reporting, lead to a higher legitimacy of an enterprise in the view of stakeholders (Velte & Stawinoga 2017). In line with those findings, earlier studies find a negative relationship between audit fees and internal control, which indicates that poor internal control will result in a higher audit fee (Hogan & Wilkings, 2008).

In this thesis the effect of integrated reporting quality on the level of audit fees is researched. Whereby the dependent variable will be translates as audit fees, and the independent variable will be Integrated Reporting Quality.

Firstly, De martini & Trucco (2016) found that integrated reporting quality is linked to the audit risk, since better information quality decreases the information asymmetry in the market and consequently, the information risk of firms. Moreover, other studies find an effect between the level of audit fees and audit risk (Houston et al. 1999; Hay et al. 2006; Hogan and Wilkins 2008, where audit risk is linked to the substantial internal control problems firms face by Hogan and Wilkins (2008). In line with the outcomes of previous studies mentioned above, I formulate the following hypothesis:

H1: Organisations that publish an Integrated Reporting with higher quality have a lower audit fee, in comparison to the organisations who publish an Integrated Reporting with lower quality

In addition to hypothesis 1, two other hypothesis are formulated. The three most primarily factors associated to have an influence on audit fees are the organizational size, risk and the complexity (Hay et al. 2006). In this thesis the influence of the risk will be further researched, and will be used as a moderating factor. Leverage is an indicator of the level of debt financing and therefore is linked to the risk of an organizations bankruptcy, which would imply a loss for the auditor (Simunic 1980). The moderating variable is going to test the influence of the leverage on the relationship between audit fees and integrated reporting quality. Leverage is the total debt scaled by the organisations capital, a higher leverage therefore implies more debt financed organisation. Leading to the second hypothesis:

H2: The negative relation between integrated reporting quality and audit fee will be less distinct for organisations with higher level of leverage.

As mentioned above, the three most primarily factors associated to have an influence on audit fees are the organizational size, risk and the complexity (Hay et al. 2006). Return of equity (ROE) is an indicator that can specify the organizational risk. ROE equals the net income minus bottom line and preferred dividend requirements, divided by the common equity. The expected influence of the ROE on the relation integrated reporting quality on audit fees is positive. Since a higher ROE indicates a lower risk for the auditor (Simunic 1980 and Hay et al. 2016). In line with Hogan et al. (2008), I expect audit fees are a proxy for audit effort, and therefore a lower audit risk will lead to a lower level of audit fee. Leading to the third and last hypothesis in this thesis:

H3: The negative relation between integrated reporting quality and audit fee will be less distinct for organisations with lower level of return on equity.

In the following section the research method will be described via a conceptual framework, later in the paragraph the variables and the data construction will be discussed.

4. Research Method

4.1 The model

The following conceptual model is used to understand the relationship of the independent variable integrated reporting quality (IRQ) on the dependent variable audit fees.

Figure 3: Conceptual Model

De martini & Trucco (2016) found that integrated reporting quality is linked to the audit risk, since better information quality decreases the information asymmetry in the market and consequently, the information risk of firms. Other studies have confirmed the correlation between the level of audit fees and audit risk (Houston et al. 1999; Hay et al. 2006; Hogan and Wilkins 2008, where audit risk is linked to the substantial internal control problems firms face by Hogan and Wilkins (2008). In addition, research shows that organizations with internal control deficiencies have a higher level of inherent risk and information risk, and therefore are positively related to audit fees. The alteration of the audit fees depends on the severity of the internal control problems (Jiang and Son 2014) since the external auditor has to perform more work to provide the same level of assurance when there is a bad internal control (Hoitash et al. 2008).

As stated by Hay et al. (2006), ‘’A common methodology has developed for examining the determinants of audit fees that has been used in well over 100 published journal articles’’. This thesis draws back up on a that standard audit fee model used by Beck and Mauldin (2014), which is based on the prior research of Simunic (1980), Francis and Simon (1987), Raghunandan and Rama (2006), Hogan and Wilkins (2008), Doogar et al. (2010).

The model includes control variables known as determinants of audit fees and uses a modified independent variable namely, Integrated Reporting Quality. To examine the components that influence the audit fee we use the following equation in the regression analysis:

Integrated

ln_FEEitS = β0 + β1IRQit + β2ln_LEVERAGEit + β3ln_ASSETSit + β4ROEit + β5RECINVit + β6CASHFLOWit

+ β7BSIZEit + β8BPRESit + β9VOLATILIYit + β10INDUSTRYit + ε

Appendix b details all variable’s definition and data source.

4.2 The Variables

In the following paragraph, the dependent, independent and controlling variables will be described. This thesis will make use out of partly hand-collected data from South Africa over a period of the last five years, where the Johannesburg Stock Exchange (JSE) requires organizations to publish an Integrated Report annually. The benefits of the South African case setting is that concerns about self-selection are mitigated since there is no voluntary disclosure.

Dependent Variable:

Based on research of Beck and Mauldin (2014), the dependent variable will be measured in natural log form to mitigate the effect of non-linear relations. The audit fees will be measured in at the firm i and in the year t. (ln_FEES)

Independent Variable:

In line with research conducted by Barth et al. (2017) and Zhou et al. (2017), I use the data from EY which annually values the quality of integrated reports in South Africa. The EY integrated Reporting Awards evaluates the 100 JSE organisations each year in term of market value equity. The quality category of an integrated report is publicly announced and can be found in the EY Integrated Reporting Awards.

The rating of those awards are based on the focus on the quality of the disclosure, especially on the level the integrated reports give the user insight in the organisations strategy and value creation processes. As stated by Barth et al. (2017) the IRQ measurement is not simply a disclosure index that captures the presence or absence of particular items. According to the chair of the EY panel that rates the reports, the ratings focusses on the quality of the disclosure, specifically whether the integrated report gives readers a sense of the firm's strategy and value creation process. Moreover, in addition to the independent nature of the rating and its focus on quality the EY adjudicators are experts.

Moreover, research of Zhou et al. (2017) shows that this measurement of IRQ based on the 2012 draft IIRC Framework correlates with the EU quality categories, those finding add validity

to use of the IRQ in this thesis. The Integrated Reporting Quality will be measured at the firm i in the year t (β1IRQ), with an indicator variable between 1 and 5. Corresponding with the

rankings of the EY integrated Reporting Awards the number 1 stands for top 10 firms, 2 for Excellent, 3 for Good, 4 for Average, 5 for Progress to be made.

Control Variables:

The control variable ln_ASSETS, is measured as natural log of the total assets. LOSS is measured as an indicator variable with the value 1 (loss) or 0 (profit). CASHFLOW is the operating cash flow divided by total assets. RECINV is measured as the inventory plus accounts receivable divided by total assets, which reflex operating risk (Beck and Mauldin 2014).As stated by Hay et al. (2006): ‘’Results suggest that inherent risk is an important driver of audit fees but the combination of inventory and receivables may be a better proxy than considering these accounts separately.’’ BDSIZE is measured as the number of director is the board. BDMEET is the number of board meetings per fiscal year. BDIND equals the percentage of independent board member as stated by the organisation. INDUSTRY equals the company's general industry classification. VOLATILITY represents a stock's average annual price movement to a high and low from a mean price by firm i in year t.

Moderating variable

The first moderating variable used in this thesis is LEVERAGE. Leverage is an indicator of the level of debt financing and therefore is linked to the risk of an organisations bankruptcy, which would implies a loss for the auditor (Simunic 1980). This variable changes the effect on form the dependent variable on the independent variable, see the figure shown at hypothesis 2. The moderating variable is going to test the influence of the leverage on the relationship between audit fees and integrated reporting quality. Leverage is the total debt scaled by the organisations capital, a higher leverage therefore implies more debt financed organisation.

The second moderating variable used in this thesis is return on equity (ROE). ROE equals the net income minus bottom line and preferred dividend requirements, divided by the common equity. The expected influence of the ROE on the relation integrated reporting quality on audit fees is positive. Since a higher ROE indicates a lower risk for the auditor.

4.3 Data Construction

This paper combines data from multiple sources, explicitly, DataStream, EY Excellence in Integrated Reporting Awards and integrated reports published by organizations. Data between 2013 and 2017 is used.

As stated above, this thesis makes use out of public data collected from EY South Africa for the EY Integrated Reporting Excellence Awards. The organizations covered by the sample can be found in Appendix C. These organizations represent approximately 92.5% of the total market capitalization in South Africa (EY 2017), and consequently covers most of the JSE capitalization despite the fact the sample is relatively small.

Since the non-standard databases of the Integrated Reporting Excellence Awards was used, other data was collected in a later stadium. Firstly, ISIN codes of the companies published in Integrated Reporting Excellence Awards are identified, whereafter data for the variables were added. A significant part of all data (except the IRQIT) is collected via DataStream. In addition, this thesis makes extensively use out of hand-collected public data to supplement missing data in the dependent and independent variable. This manually gathered data is collected from the organizations own websites, which were search with the terms; Integrated Report, Financial statement, Annual Report. Hand-collected data is used to supplement missing data in the dependent and independent variable. All annual reports and integrated reporting reports are downloaded from the organization's website or the archives of Integrated Reporting Examples Database.

5. Empirical Results:

In the following part, the empirical results of this thesis will be outlined. First, a summary will be provided about the descriptive statistics, and second the results of the Pearson correlation matrix and Spearman rank correlation. The outcomes of the OLS regression will be discussed in the final paragraph.

5.1 Descriptive statistics

Table 1 provides descriptive statistics for the sample used in this thesis, with a total score of 486 observations, including 131 different organizations. Outliers are winsorized at the maximum or minimum level of three times the standard distribution from the mean. In table 1 dummy variables that only have the value 1 or 0 are excluded.

The mean of the natural log of audit fees is 9,84; this is not in line with the research of Goncharov et al. (2013) who find a mean of 4.92. This alteration can be straightforwardly explained since the audit fee in this thesis is measured in South African Rand and consists of the 92.5% of the total market capitalization of the market, in contradiction to the Goncharov et al. (2013). The mean for IRQ is 3,18, which correspondence with a ‘good’ level of integrated reporting in South Africa, according to the EY Excellence in Integrated Reporting Rewards. Nevertheless, the standard deviation of 1,2 shows a great difference between the measured quality of the integrated reports within the sample, in line with the research of Barth et al. (2017) and Pistoni et al. (2018). Furthermore, table 1 shows a mean of 30,8% for leverage with a standard deviation of 18,43%, the high standard deviation illustrates the various ways South African companies are financed. The return on assets show a mean of 0,07 with a standard deviation of 0,08, inconsistent with the finding of Hogan et al. (2008). Nonetheless, this dissimilarity can be clarified merely since the sample group is different, as well are 2013 until 2017 years of economic prosperity, in contradiction with 2008.

Table 1 also shows 17,36 for the natural log of total assets, 0,16 for return on equity, 0,19 for inventory plus receivables divided by total assets, 0,19 for the operating cash flow divided by total assets, 12,2 for the number of board members and 96% for board presence during meetings.

Table 1 Descriptive Statistics

Variable Observations Mean Std. Deviation Minimum Maximum

LnAuditFee 486 9,847245 1,280935 6,77422 12,79492 IRQ 486 3,18 1,210 1 5 Leverage 486 30,84843 18,43142 0,09 82,56 ROA 486 0,071898 0,084165 -0,2526 0,3921 LnAssets 486 17,36086 1,384929 13,66916 21,42954 ROE 486 0,157017 0,188137 -0,5518 0,9547 RecInv 486 0,19027 0,187417 0 0,74396 Cashflow 486 0,08866 0,079019 -0,085 0,40983 BSize 486 12,2 2,927 6 20 BPres 486 0,9601 0,04508 0,75 1 Volatility 486 0,2432 0,06399 0,119 0,4986

All variables further defined in Appendix B.

Table 2 Correlation Matrix

LnAudit Fee IRQ Lever age ROA LN Assets ROE Rec Inv Cash Flow

BSize BPres Volati lity LnAuditFee 1 -,032** (,004) ,200** (,000) -,272** (,000) -,538** (,000) -,134** (,003) -,047 (,296) -,018 (,695) ,363** (,000) -,036 (,435) ,004 (,931) IRQ -,153** (,001) 1 ,076 (,094) ,110* (,015) ,199** (,000) ,007 (,874) ,030 (,505) -,092* (,043) -,142** (,002) ,037 (,417) -,094* (,037) Leverage ,202** (,000) -,066 (,148) 1 -,021 (,642) -,180** (,000) ,057 (,209) -,014 (,755) -,081 (074) -,019 (,670) -,035 (,446) -,181* (,000) ROA -,269** (,000) 0,45 (,323) -,058 (,205) 1 -,346** (,000) ,756** (,000) ,275** (,000) -,491** (,000) -,158** (,000) -,084 (,066) -,241** (,000) LnAssets 0,641** (,000) -,202** (,000) ,175* (,000) -,294** (,000) 1 -,107 (,019) -,359** (,000) -163** (,000) ,292** (,000) -,021 (,641) ,076 (,093) ROE -,153** (,001) -,024 (,592) 0,050 (,273) 0,733* (,000) -,062 (,172) 1 ,144* (,012) ,355** (,000) -,010 (,820) -,082 (,071) -273** (,000) RecInv ,051 (,263) ,009 (,838) -,067 (,139) ,222** (,000) -,342** (,000) ,114** (,001) 1 ,492** (,000) -,173** (,000) -,083 (,069) ,157** (,001) Cashflow -,069 (,130) -,136** (,003) -,113* (0,12) ,600** (,000) -,200** (,000) ,410** (,000) ,037** (,000) 1 -,027 (,559) -,138** (,000) ,119* (,008) BSize ,394** (,000) -,155** (,001) -,012 (,749) -144** (,001) ,400** (,000) -,063 (,165) -,163** (,000) -,083 (,069) 1 -,011 (,811) -,040 (,377) BPres -,026 (,563) 0,43 (,343) -,028 (,532) -,071 (,119) -,008 (,876) -,023 (,609) -,062 (,170) -,155** (,012) -,052 (,249) 1 ,027 (,546) Volatility -,026 (,564) -,145** (,001) -,181** (,000) -,244** (,000) ,007 (,885) -340** (,000) ,054 (,237) ,066 (,147) -,070 (,125) -,009 (,836) 1

Pearson correlations appear below the diagonal, non-parametric Spearman correlations appear above the diagonal * Correlation is significant at the 0.1 level (two-tailed) ** Correlation is significant at the 0.05 level (two-tailed) *** Correlation is significant at the 0.01 level (two-tailed)

5.2 Correlation Matrix

In table 2 the results of the Pearson correlation matrix and Spearman rank correlation are shown. Below the diagonal, the Pearson correlation can be found which tests for correlation between variables, above the diagonal the non-parametric Spearman correlation can be found which tests for possible multicollinearity complications. There are three different level of correlation significance. Namely, * correlation is significant at the 0,1 level (two-tailed), ** Correlation is significant at the 0.05 level (two-tailed), *** Correlation is significant at the 0.01 level (two-tailed). A lower number resembles a higher significant correlation and vice versa.

If the non-parametric Spearman correlation equals 0, no correlation exists between the variables. Values between 0 and 0,10 are characterized as small, between 0,10 and 0,30 as medium, between 0,30 and 0,50 as large and between 0,50 and 0,70 as large correlation. When the value is above the 0,70 the presence of multicollinearity is expected (Field, 2014). As can be found in table 2 there is multicollinearity between the return on assets and the return on equity (,756) and a medium correlation with Cashflow (-,491). Therefore the return on assets will be removed from the variables set during the regression analysis. Multicollinearity between the two variables (ROA and ROE) can be explained by the fact that they measure the same influencing factor, namely, profitability. The negative medium correlation between CashfFlow and RecInv (,492) won’t be a large threat for the regression model. The Spearman correlation matrix can also be used to identify extreme values, however, since the sample group is already winsorized this is not the case.

Table 2 shows a negative significant 0,05 level correlation between the LnAuditFee and the integrated reporting quality (-,153), in line with the expectations of the hypothesis. LnAuditFee also shows a significant correlation with Leverage (,202), a higher audit fee, therefore, indicates a higher level of leverage. In the same time, LnAuditFee shows a significant correlation with LnAssets of, 641, which means a that a company with more assets is likely to pay a higher level of audit fees. LnAuditFee also shows a significant correlation with ROE of -,153, which indicates that a lower return on equity is associated with a higher level of audit fees. The last variable that correlates significantly with LnAuditFee is BSize (,394), a bigger board, therefore, is associated with a higher audit fee. Overall, the Pearson correlations are consistent with expectations and previous studies (Barth et al. (2017), Hay et al. (2006), Hogan et al. (2008) and Pistoni et al. (2018)).

5.3 Regression Analysis

Table 3 provides insight into the empirical results based on the OLS regression. The model used in this thesis explains 51% of the variance in audit fees (Adjusted R Square: ,507071), which shows a high descriptive power. This thesis hypothesizes that the relationship between the integrated reporting quality and the level of audit fees will be negative. As predicted the expected sign is negative, which implies that a higher integrated reporting quality leads to lower audit fees, considering all other variables are similar (-0.002, t stat = -,071). However, the results of the regression analysis show a small coefficient that is not significant. This indicates that the hypothesis 1 of this thesis does not hold and that the null hypothesis cannot be rejected. The low significance might be a result of the small sample group of 485 observations since a substantially larger sample group better clarifies a significant relationship.

As expected, this thesis found a significant relationship between the control variables and the depended variable. Between leverage and the audit fee, I found a positive significant relationship (-0,019***, t stat = 3,553), which implies a higher level of leverage is associated with a higher level of audit fees. This is in line with previous research of Hay et al. (2006) and Hogan et al. (2008). Besides, the relationship between the total assets and the level of audit fee are in line with the same research total assets, in this thesis we find that the level of total assets relates positively significantly with the level of audit fee (0,618***, t stat = 16,210). In this thesis I also found that the return on equity is negatively significant related to the level of audit fees (-0,267***, t stat = -6,141), RecInv is positively significant associated with the level of audit fees (0,185***, t stat = 5,091), CashFlow positively significant related to the level of audit fees (0,103***, t stat = 2,627), and a positive significant relation between the number of board members and the level of audit fees (0,158***, t stat = 4,449). The outcomes of the regression analysis on the control variables of leverage, total assets, return on equity, RecInv, Cashflow and board size are in line with the predictions made based on prior research.

The control variable that did not have a significant relationship with the level of audit fees is the percentage of board member present at meetings (0,012, t stat = 0,360).

In contradiction with previous research this thesis shows a significant negative relationship between the volatility and the level of audit fees (-0,062*, t stat = 4,449), and a significant negative relationship between the loss indicator dummy and the level of audit fees (-0,144**,

t stat = 2,182). Differences that might find their origin in the small sample group used in this thesis, moreover are those results not as significant as the other control variables.

Table 3 Regression H1

Variable Expected Sign Coefficient P-Value

IRQ - -,002 ,943 Leverage + ,119 ,000 LnAssets + ,618 ,000 ROE - -,267 ,000 RecInv + ,185 ,000 Cashflow + ,103 ,009 BSize + ,158 ,000 BPres - ,012 ,719 Volatility + -,062 ,093 LossProfit Adjusted R Sqr. F-Value N + ,50707 46,356 485 -,114 ,030

Table 4 Regression Including moderating factor H2

Variable Expected Sign Coefficient P-Value

IRQ - -,002 ,959 Leverage + ,123 ,000 IRQLeverage - -,035 ,291 LnAssets + ,612 ,000 ROE - -,261 ,000 RecInv + ,185 ,000 Cashflow + ,098 ,013 BSize + ,167 ,000 BPres - ,013 ,679 Volatility + -,054 ,153 LossProfit Adjusted R Sqr. F-Value N + 0,5081 46,540 485 -,119 ,003

To test the hypothesis 2 the variable IRQLeverage is constructed, this variable consist of the centered IRQ multiplied with the centered Leverage. The new variables are constructed with the help of mean centering, in line Irwin and McClelland (2001). Mean centering is commonly

used as a practice for estimating, testing, and interpreting regression models that include moderated relationships (Iacobucci, Schneider, Popovich, & Bakamitsos (2017). The centered IRQ variable is constructed as follows: mean of IRQ minus the original carried value of the variable. The centered Leverage variable is similarly constructed, namely, mean of Leverage minus the original carried value of the variable.

The model used in this thesis to explain hypothesis 2 predicts 51% of the variance in audit fees (Adjusted R Square: ,5081), which shows a high descriptive power. Table 4 shows the results from the regression analysis including the moderating factor of hypothesis 2. The results show no significant relationship between the variable IRQLeverage and the dependent variable audit fees (-0,035, t stat = -1,058). As an effect, the null hypothesis cannot be rejected. Therefore the hypothesis 2 does not hold. This thesis cannot prove an effect of leverage on the relationship between the level of audit fees and integrated reporting quality since the outcomes of the regression analysis are not significant for the variable IRQLeverage.

Table 5 Regression Including moderating factor H3

Variable Expected Sign Coefficient P-Value

IRQ - -,002 ,956 Leverage + ,120 ,000 IRQROE - -,009 ,796 LnAssets + ,618 ,000 ROE - -,266 ,000 RecInv + ,184 ,000 Cashflow + ,100 ,013 BSize + ,159 ,000 BPres - ,011 ,726 Volatility + -,060 ,112 LossProfit Adjusted R Sqr. F-Value N + 0,5071 46,360 485 -,121 ,002

To test the hypothesis 3 the variable IRQROE is constructed, this variable consists of the centered IRQ multiplied by the centered ROE. The centered ROE variable is similarly constructed as the other centered variables, namely, mean of ROE minus the original carried value of the variable. The model used in this thesis to explain hypothesis 3 predicts 51% of the variance in audit fees (Adjusted R Square: ,5071), which shows a high descriptive power. Table

5 shows the results from the regression analysis including the moderating factor of hypothesis 3. The results show no significant relationship between the variable IRQROE and the dependent variable audit fees (-0,009, t stat = -,258). As an effect, the null hypothesis cannot be rejected. Therefore the hypothesis 3 does not hold. This thesis cannot prove an effect of ROE on the relationship between the level of audit fees and integrated reporting quality since the outcomes of the regression analysis are not significant for the variable IRQROE.

5.4 Robustness Check and Additional Tests

In this thesis control variables are used in addition to the cause-effect variables. Control variables are used to predict if the hypothesis holds when the model takes into account effects of other variables. As commonly practiced control variables no part of a separate analysis, but embedded in the analysis performed through this thesis. The use of control variables facilitates as a validation for the sample group used in this thesis since the control variables have a significant relationship with the dependent variable. In case there would be no correction between the control variables and the predicted dependent, the sample group is likely to contain error and is most likely to be inconvenient for the predicted model. Though as mentioned above, the control variables used in this thesis do have a significant relationship with the predicted dependent variable. In addition to using control variables, this thesis regression analysis is split in sub-analysis. Split sample analysis is used in case of predicted an interaction or moderator effect to facilitate an enhanced interpretation of the interaction.

Although this thesis regression did not observe a significant relationship between the level of integrated reporting quality and the level of audit fees, an additional test is carried out to prevent the possibility that the results of this analysis are affected by the absence of additional control variables. By carrying out this additional analysis, the regression coefficients on robustness and plausibility. Since the three primary factors associated to influence audit fees are the organizational size, risk and the complexity (Hay et al. 2006). The new control variable industry is added to the regression analysis.

A common assertion made by auditors and researchers is that some industries are more difficult to audit than others (Simunic 1980; Turpen 1990; Pearson and Trompeter 1994). This variable captures the company's general industry classification, measured as 1 industrial, 2 utility, 3 transportation, 4 banking, 5 insurance and 6 other financial. The variable controls the effect of the differences between industry and potential cross-sectional dependence of

residuals. A dummy variable is created to measure the effect of industry classification. As stated by Hay et al. (2006), ‘financial institutions and utilities have relatively large assets but are generally easier to audit than companies with an extensive inventory, receivables, or knowledge-based asset’’. Therefore financial organizations usually are more straightforward to audit then utility organizations. In line with previous research, a dummy variable used that measures if the organizations are classified as a financial or utility business. The results of the additional performed regression analysis can be found in Appendix E. The overall value of the R-square increased with 5%, although there is still no significance in the relationship between the level of integrated reporting quality and the level of audit fees. Besides, the moderating factors did not show a significant relation either.

Worth mentioning is the alteration in the control variables RecInv and CashFlow, both variables became insignificant as an effect of the added control variable industry. RecInv measures an organization's inherent risk, a possible another way to measure an

organization's inherent risk is via the receivables divided by total assets, as described by Hay et al. (2006). CashFlow measures an organizations size, a possible supplementary way to quantify an organizations size is via the total sales, as described by Hay et al. (2006).

6. Conclusion and limitations

In the section below I will discuss the conclusion following from result of the empirical analysis. First, a summary will be given, and then the conclusions will be outlined. The paragraph will end with the limitations of this thesis and recommendations for future research.

6.1 Summary

Jensen and Mackling (1976) research on the agency problem indicates that the external audit is a control mechanism, which mitigates issues that occur as an effect of the separation of ownership and management. Subsequently, an efficient allocation of funds on the investment market requires an adequate information disclosure towards investors (Healy & Palepu, 2001). Within this field, recent research indicates that organizations seek new opportunities to provide more information towards stakeholders by providing additional reports containing non-financial information next to the existing financial statements (Cohen et al., 2012). Finding that is confirmed by Adams et al. (2011), who indicated that current providence of information by organizations does not meet requirements of different stakeholders. As an effect the movement arose linking financial and non-financial information, leading to the introduction of Integrated Reporting by the International Integrated Reporting Council (IIRC). Literature shows that integrated reports with clear connectivity between the corporate governance and the financial reporting, lead to a higher legitimacy of an enterprise in the view of stakeholders (Velte & Stawinoga 2017). The IIRC Framework is promptly gaining acceptance. Presently, more than 1500 organizations prepare an integrated report annually. Although the momentum of the IIRC and its launched integrated reporting made a gain the last couple of years, the final adoption, and standardization of integrated reporting is still more than ambiguous and consequently an exciting development (Humphrey, O’Dwyer and Unerman, 2017).

As mentioned in the introduction, the goals of the IIRC is to set integrated thinking as a standard within businesses, where integrated thinking can provide a more holistic picture for organizations. This broader picture leads to a better understanding of the firm’s risks, which could lead to a better and more complete risk management system (Moolman et al. (2016). In addition, Demartini & Trucco (2016) found that integrated reporting quality is linked to the audit risk, since better information quality decreases the information asymmetry in the market and consequently, the information risk of firms.

Since the three primary factors associated to have an influence on audit fees are the organizational size, risk and the complexity (Hay et al. 2006), I expected a negative relationship between integrated reporting quality and audit fees. Leading to the first hypothesis: organizations that publish an Integrated Reporting with higher quality have a lower audit fee, in comparison to the organizations who publish an Integrated Reporting with lower quality.

Leverage is an indicator of the level of debt financing and therefore is linked to the risk of an organizations bankruptcy, which would imply a loss for the auditor (Simunic 1980). Leading to the second hypothesis; the negative relation between integrated reporting quality and audit fee will be less distinct for organizations with a higher level of leverage. At the same time, profitability has a proven to have a positive relationship regarding audit fees and also reflexes the possibility that the auditor might be exposed to a loss in case of bankruptcy by the client (Simunic, 1980). Leading to the third hypothesis: the negative relation between integrated reporting quality and audit fee will be less distinct for organizations with a lower level of return on equity.

The final sample used in this thesis contains 486 observations, including 131 different organizations, over the years 2013 to 2017. A Pearson correlation matrix is used to test the correlation between the different variables. For the first hypothesis a standard linear regression analysis is used, for the second hypothesis the variable IRQLeverage was added, for the third hypothesis, the variable IRQREO was added. Lastly, a robustness check was used, and additional analysis was carried out with an additional control variable Industry.

6.1 Conclusion

In this thesis three hypotheses are tested, the research result indicates that the quality of integrated reporting does not decrease the level of audit fees. No significant relationship was found between the quality of integrated reporting and the level of audit fees, providing no evidence for the first hypothesis. This unexpected outcome might find its origin in multiple explanations. First of all, the field of integrated reporting is relatively new. Auditors might not have discovered yet a way to exploit a better-integrated reporting quality, in assessing the risk and eventually determining the audit fee. It is possible that future research will find different outcomes with the same organizations in a later stadium. Secondly, Hay et al. (2006) describes that the demand for auditing and how this demand affects control mechanisms in an organization creates the potential problem of endogeneity. For example, the existence of

governance mechanisms may generate more need for external auditing, which could mitigate the effect of integrated reporting quality on the level of audit fees. As stated by Hay et al. (2006) Unfortunately; ‘’the current theoretical underpinnings of basic audit fee models do not deal well with issues of endogenous demand’’. Lastly, audit fees for companies with an overconfident manager can be lower if managers demand fewer audit services (Duellman, Hurwitz & Sun (2015). Research shows that organizations are willing to spend less on external assurance when managers have overconfidence, signaling to external parties with a lower audit fee. In future research, the involvement of management overconfidence on the relationship between the quality of integrated reporting and the level of audit fees need to be investigated.

The second hypothesis tests if leverage moderates the relationship between integrated reporting quality and audit fee. The coefficient was found to be more negative then single variable IRQ. This effect can be explained since higher leverage indicates an organization is financed with more external fund, increasing the risk for the auditor. However, the results show no significant relationship between the dependent variable audit fees and the constructed variable of IRQleverage. Therefore the second hypothesis does not hold.

The third hypothesis test if the return on equity moderates the relationship between integrated reporting quality and audit fee. The results show no significant relationship between the dependent variable audit fees and the constructed variable of IRQROE. Therefore the second hypothesis does not hold.

6.2 Limitations

This study is subject to some limitations. Firstly, this study has a focus on the South African case since integrated reporting is mandatory since 2010, this contracted focus lead to a sample group of 485 different observations. Moreover, some control variables were left out of the regression analysis due to a limited time provided to write this thesis. Since the data provided by the databases were not complete, gaps had to be filled in with handpicked data or be left out. While control variables are expected to have a significant effect on the dependent control variable, three control variables did not have a significant relationship. Possible inclusion of other control variables might increase the validation of the sample used in this thesis.

6.3 Future Research

Since integrated reporting is a relatively new field within academic research, still little is known about specific relations and conditions surrounding the implementation of integrated reporting. As described in the theory section, significant differences exist in the quality of integrated reports (IRQ) between organizations (Pistoni et al. (2018). Future research could, therefore, focus on the factors that are at the source of the differences in quality. Other future research could focus on the potential problem of endogeneity between the integrated reporting and audit fees. Moreover, since this thesis makes use out of data from South Africa, other studies might conduct the similar research other markets. Since more organizations are adopting integrated reporting voluntarily, more data will be available in the foreseeable future.