Assurance decisions in integrated reporting in South Africa

Table of Contents

1. Introduction

2. Introduction to Integrated Reporting and Assuranc
2.1 Definition of Integrated Reporting
2.2 Development of Non-Financial Reporting
2.2.1 Development of Sustainability Reporting
2.2.2 Development of Integrated Reporting
2.2.3 Development of Integrated Reporting in South Africa
2.3 Advantages and Disadvantages of Integrated Reporting
2.4 Assurance of Integrated Reports

3. Literature Review and Hypotheses Development
3.1 Previous Research on the Determinants of Assurance
3.1.1 Methodological Approach
3.1.2 Country-level Variables
3.1.3 Industry-level Variables
3.1.4 Corporate Governance Variables
3.1.5 Financial Variables
3.2 Hypotheses Development
3.2.1 Research Questions and Independent Variables
3.2.2 Board Structure
3.2.3 Diversity of Board Members
3.2.4 Professional Background of Board Members
3.2.5 Shareholder Structure
3.2.6 Additional Hypotheses

4. Methodology and Data
4.1 Data Collection and Sampling
4.2 Model Specification and Variable Measurement
4.3 Multiple Imputation of Missing Data

5. Results
5.1 Descriptive Statistics
5.1.1 Means, Standard Deviations, Minimums, and Maximums
5.1.2 Development of Assurance Characteristics across Industries
5.1.3 Correlation Analysis on Textual Quality
5.1.4 Multicollinearity Tests
5.2 Multivariate Results
5.2.1 Decision to Adopt External Assurance
5.2.2 Decision to Mandate an Assurer from the Accounting Profession
5.2.3 Decision to Mandate Higher Levels of Assurance

6. Conclusion and Discussion
6.1 Main findings
6.2 Implications for the Main Constituents in Integrated Reporting
6.3 Limitations and Avenues for Future Research

Bibliography

Appendix

Preface

The copyright of the Master thesis rests with the author. The author is responsible for its contents. Bocconi University is only responsible for the educational coaching and cannot be held liable for the content.

This thesis is part of my study in the Double Degree Programme in the Master of Science in International Management at Bocconi University and the Post Graduate Programme in Management at the Indian Institute of Management Ahmedabad.

I would like to express gratitude to my thesis advisor Prof. Ariela Caglio (Bocconi University), and my thesis coaches Prof. Paolo Perego (Free University of Bozen-Bolzano) and Gaia Melloni (University of East Anglia) for their invaluable suggestions and feedback for my thesis topic and research methods.

Lastly, I would like to thank my family and friends for their trust and constant support.

Table of Figures

FIGURE 1. OVERVIEW OF CHAPTERS AND COMPONENTS

FIGURE 2. OVERVIEW OF VARIABLES IN PUBLICATIONS

FIGURE 3. OVERVIEW OF HYPOTHESES VARIABLES

List of Tables

TABLE 1. TYPES OF CORPORATE REPORTS

TABLE 2. OVERVIEW OF PUBLISHED STUDIES

TABLE 3. SAMPLE EXPLANATION

TABLE 4. DEFINITION OF VARIABLES

TABLE 5. SHAPIRO-WILK NORMALITY TEST

TABLE 6. SUMMARY STATISTICS

TABLE 7. DESCRIPTIVE STATISTICS OVER TIME

TABLE 8. DESCRIPTIVE STATISTICS BY INDUSTRY

TABLE 9. DESCRIPTIVE STATISTICS BY TYPE OF ASSURER

TABLE 10. CORRELATION ANALYSIS ON TEXTUAL QUALITY

TABLE 11. CORRELATION MATRIX FOR MODEL 1

TABLE 12. CORRELATION MATRIX FOR MODELS 2 AND 3

TABLE 13. COLLINEARITY DIAGNOSTICS

TABLE 14. RESULTS FROM RESEARCH MODELS

TABLE 15. OVERVIEW OF HYPOTHESES AND RESULTS

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

1. Introduction

Impacting generations of managers, Milton Friedman famously said: “The social responsibility of business is to increase its profits” (Friedman, 1970). Since the financial crisis of 2007-2008, it is apparent that managers cannot continue to follow this paradigm. Companies are trapped to optimise short- term financial performance, while “overlooking the greatest unmet needs in the market as well as broader influences on their long-term success” (Porter & Kramer, 2011). As a result, the narrow approach of value creation creates not only adverse economic, but also negative environmental and social externalities.

The focus on financial reporting (FR) is seen as one of the roots of this vicious cycle, and the recent trend towards sustainability reporting (SR) does not represent more than an enhanced collection of sustainability information. Thus, integrated reporting (IR), which creates a link between both types of information and aims to promote companies to develop strategies that recognise sustainability aspects as significant drivers for financial performance, can be regarded as a promising innovation (International Integrated Reporting Council (IIRC), 2014).

However, IR is far from becoming the norm. In 2015, only 15% of the world’s largest 250 companies by revenue published integrated reports (IRs) and the absence of IR standards and limited understanding of IR assurance present key barriers for widespread adoption (KPMG, 2017). IR research has thus far mainly focused on the normative arguments for IR. Henceforth, this thesis represents a significant effort to close the large gap in research on IR assurance (Dumay, Bernardi, Guthrie, & La Torre, 2017; Cheng, Green, Condradie, Konishi, & Romi , 2014). It is the first study investigating the determinants of assurance decisions in IR. Additionally, it represents the only study focusing on South Africa, the first and only country to enforce a mandatory IR disclosure regime globally (De Villiers, Venter, & Hsio, 2017).

Building on the findings of previous studies on the determinants of voluntary assurance, there are three research questions investigated in the thesis. First, what are the determinants of a firm’s decision to/not to adopt IR assurance? Second, given that firms have adopted IR assurance, what are the determinants for choosing a particular type of assurer? Third, given that firms have mandated a particular type of assurer, what are the determinants for choosing higher levels of assurance?

The sample consists of the IRs of 139 of the largest 160 companies listed at the JSE over the six-year period after the adoption of IR in South Africa in 2011. The research examines the significance of corporate governance (CG) variables, including board characteristics, shareholder characteristics, primary listing and IR assurance experience as potential determinants for assurance. The factors are complemented by a range of industry-level and financial control variables as well as further variables frequently used in previous research on voluntary assurance.

The thesis provides practical insights for the principal constituents of IR. Companies can broaden their understanding of how assurance decisions within firms are taken. IR assurance providers can use the insights to more effectively market IR assurance. Regulators in South Africa and outside South Africa, who are interested in adopting mandatory corporate disclosure initiatives similar to IR, can expand their understanding of the dynamics of IR assurance.

As portrayed in Figure 1, the thesis is divided into six chapters. Chapter 2 provides an introduction to integrated reporting. The concept of IR is examined in light of other types of corporate reporting and an outline of the development of non-financial reporting globally and South Africa is presented. Additionally, advantages as well as disadvantages of IR and current assurance practice is described. As a basis for the development of hypotheses, Chapter 3 starts with providing a literature review of previous publications on voluntary assurance in FR and SR. Subsequently, hypotheses for the three research questions investigated in the thesis are developed. Chapter 4 elaborates on data collection and the research methodology. Besides, it provides a detailed description of how missing data were imputed. In Chapter 5, the descriptive and multivariate results of the three research models are presented and analysed. Additionally, a correlation analysis between assurance characteristics and textual attributes of IRs is conducted to investigate the correlation between assurance and disclosure quality. Chapter 6 provides a conclusion and describes the implications of the findings for the main constituents of IR. Furthermore, the chapter points to limitations and suggests avenues for future research. Lastly, based on the findings, an interview was conducted with a leading expert on IR, which is attached in the Appendix.

Figure 1. Overview of Chapters and Components

Abbildung in dieser Leseprobe nicht enthalten

2.Introduction to Integrated Reporting and Assurance

First, the chapter starts by defining the concept of IR. Second, an overview of the development of non-financial reporting and IR is provided. Third, the advantages and disadvantages of IR are investigated and lastly, current IR assurance practice is described.

2.1 Definition of Integrated Reporting

In order to understand the concept of IR, it is useful to differentiate IR from other types of corporate reporting (Table 1). Financial statements and narrative reports primarily focus on FR and are mainly targeted at financial stakeholders. In contrast, SR and IR also include non-financial information and are more stakeholder-oriented.

Table 1. Types of Corporate Reports

Abbildung in dieser Leseprobe nicht enthalten

Based on the International Federation of Accountants (2015), *depending on the model of IR

The sustainability information included in SRs has been shown to be relevant for shareholders and stakeholders. However, the information provided in the reports is often seen as overwhelming in quantity (Clarkson, Li, Richardson, & Tsang, 2015). As a result, even though a breadth of information is provided, shareholders and stakeholders face difficulties in utilising the information. IR aims to solve this lack of conciseness by linking financial and non-financial information. The idea is to provide a clear insight on how an organisation creates value over the short, medium and long- term (Cheng, Green, Condradie, Konishi & Romi, 2014; IIRC, 2005).

Dumay et al. (2017) claim that there are three models of IR. The first is the IR model outlined in the King Report on CG in South Africa. The King Model emphasises financial and sustainability performance and takes a stakeholder perspective. The second model is the International Integrated Reporting Framework (IIRF) and was published by the IIRC. In contrast to the King Model, the IIRF focuses on financial stability and sustainability and takes a shareholder perspective. The third model is the IR model developed by Eccles & Krzus (2010). The model states that IR should utilise the internet by providing stakeholders with a substantial amount of financial and non-financial information for conducting their own analysis.

Whereas the approach of Dumay et al. (2017) is useful to understand the various sources of thought leadership in the development of IR, it fails to recognize the interdependencies between the three models. In particular, there is significant exchange between the IIRC and the authors of the King Report as Mervyn King, acts both as the Chairman of the Council of the IIRC as well as the Chairman of the King Committee (IIRC, 2018).

2.2 Development of Non-Financial Reporting

2.2.1 Development of Sustainability Reporting

The roots of non-financial reporting can be found in the Exxon Valdez Oil Spill in 1989. As a response, social investment professionals launched the Coalition of Environmentally Responsible Economies (CERES), which set a 10-point code of corporate environmental ideals to be publicly endorsed by companies. At the same time, the United Nations Conference on Environment and Development (UNCED) for the first time put sustainable development as a public policy goal in the spotlight (Rowbottom & Locke, 2016). In 1997, CERES launched the Global Reporting Initiative (GRI), an independent international standards organisation that develops standards for reporting on environmental, social and economic performance. The standards have been updated several times in the last decades, and in 2017 63% of the largest 100 companies and 75% of the Global Fortune 250 reported to apply the GRI framework (KPMG, 2017; Rupley, Brown, & Marshall, 2017).

2.2.2 Development of Integrated Reporting

The debate about IR can be traced back to the year 1991 when the American Institute of Certified Public Accountants (AICPA) formed a special committee to address questions and concerns about the usefulness of FR. The special committee (Jenkins Committee) led to the creation of the Enhanced Business Reporting Consortium (EBRC), whose work was instrumental in the development of IR principles. In 2004, the Prince's Accounting for Sustainability Project (A4S) was established and directed at elevating the adoption of a global framework of mandatory connected reporting requirements for listed companies. Three years later, the framework, which envisioned a connection between environmental, social and governance variables to strategy, enabling companies and stakeholders an improved assessment of risks, was published. Interestingly, even before that, in early 2000, some companies including Novo Nodirsk started issuing IRs.

To cultivate the efforts for a widely accepted IR framework A4S and GRI founded the IIRC in 2010. The council consists of 40 members including the heads of the International Accounting Standard Committee (IASB), the Financial Accounting Standards Board (FASB), the International Federation of Automatic Control (IFAC), the CEOs of the Big4 companies (Deloitte, PwC, EY, KPMG) and the heads of major British accountancy bodies and Chief Financial Officers of multinationals such as Nestlé, Tata and HSBC. Whereas there are competing actors in the setting of IR frameworks, researchers consider the IR framework published by the IIRC in 2013 as the pre-eminent IR framework globally (Flower, 2015).

2.2.3 Development of Integrated Reporting in South Africa

In South Africa, the King Report on CG represents the primary driver for the adoption of IR. The King Report is a booklet produced by the former high court judge Mervyn King together with a group of experts since 1994 and has been updated four times. Whereas the King guidelines are not part of common law, they serve as an integral component of South African case law and create liability in case of misconduct. In 2009, the third update of the King Report (King III) appeared and originated IR principles, explaining the features of IRs and demand companies to assure these reports independently. One year later, IR has been mandated as part of listing requirements for the 400 companies listed at the JSE, making South Africa the first country in the world to adopt IR as a mainstream component of CG (IIRC, 2015; Rowbottom & Locke, 2016). However, it is vital to accentuate that companies were only required on an "apply or explain" basis to adopt the King III principles, meaning that companies are not required to apply the principles if they explain. The “apply or explain” principle explains why despite the demand for IR assurance, as of 2016 only a minority of South African listed companies assured its IRs.

In 2016, the most recent version of the King Report (King IV) was published, setting forth not just CG principles, but also practices, which provide guidance on how to incorporate the King Principles. The “apply or explain” policy has been changed to an” apply and explain” policy. Thus, all companies listed at the JSE need to comply with the IR principles outlined in King IV and need to explain how the individual principles are fulfilled (KPMG, 2016).

2.3 Advantages and Disadvantages of Integrated Reporting

On the one hand, research confirms the positive effect of IR on stakeholders, companies, and investors. A study commissioned by the IIRC in 2014 detected that 91% of companies enhanced their engagement with external stakeholders, 92% of companies enhanced their understanding of value creation and 81% of shareholders exhibited an increased understanding of the company’s strategy. These findings are consistent with Steyn (2014) who found that corporate reputation, investor needs, and stakeholder relations are seen as the significant benefits of IR by corporate executives. In line with that Zhou, Simnett, & Green (2012) detected that the analyst forecast error is significantly negatively related to IR disclosure, which leads to subsequent reductions of the cost of equity capital for certain reporting companies.

On the other hand, there is agreement on the shortcomings of IR. Dumay et al. (2017) and Flower (2015) argue that the IR principles developed by the IIRC are phrased in general terms. The discretion allows firms to focus on those variables which materially affect value-creation. These variables can vary significantly between industries, wherefore there is, in contrast to the other types of corporate reporting, unclarity about the precise design of IR.

2.4 Assurance of Integrated Reports

Assurance is seen as one of the primary tools to increase credibility and the quality of IR (IIRC, 2015; Institute of Directors Southern Africa, 2009). In South Africa, King III prescribes that IRs should be independently assured. As there is an absence of standards for the assurance of IRs, King III recommends the two globally leading standards in the assurance of sustainability information, i.e. AccountAbility’s AA 1000 Assurance Standard (AA1000) and the Auditing Standard Board’s International Standard on Assurance Engagement (ISAE3000).

AA1000 and ISAE3000 are regarded as suitable to build confidence in the non-financial information provided in IRs. However, the two standards are not providing guidance on how to deal with the narrative and forward- looking information inherent in IRs. Thus, the part of the information, which represents the significant innovation in IRs compared to traditional reports, is assured differently depending on cultural and organisational factors. Hence, the advantages of IR assurance are questionable (Simnett & Huggins, 2015; IIRC, 2014).

As a response, the IIRC, in 2014 issued a discussion paper dealing with the assurance of IRs and in the following held debates with around 400 people globally and received 63 written papers by academic researchers, companies, accounting associations and accounting firms. Despite the effort, however, the IIRC did not yet set any integrated assurance standards (Maroun, 2017). First, the development of IR standards is complex. Second, assurers express concerns about a regulator-led rather than market-led evolution of IR assurance as they are afraid of over-regulation, which might threaten the cost-benefit ratio of IR assurance (IIRC, 2018).

Assurance of IRs is conducted by three types of assurers: accounting firms, specialist consultancies and certification bodies. As IR is at a nascent stage, it remains unclear, which assurance provider will dominate the non-financial assurance of IRs. On the one hand, there is reason to assume that accounting providers will dominate the future of IR assurance. In the related field of SR, accounting providers were found to deliver a higher quality of assurance compared to non-accounting firms (O’Dwyer & Owen 2005, Simnett et al. 2009). On the other hand, Perego (2009) found that non- accounting firms are superior in assuring “recommendations and opinion”. These aspects include the aspect of materiality and completeness of the information included in the report. Thus, whereas accounting firms were found to be superior in validating information, non-accounting providers might have an enhanced ability to deal with the more interpretative nature of IRs.

3. Literature Review and Hypotheses Development

The chapter commences with a review of publications examining the determinants for the voluntary assurance of corporate reports. The major findings of the publications are portrayed and provide a fundament for the subsequent development of hypotheses.

3.1 Previous Research on the Determinants of Assurance

3.1.1 Methodological Approach

A thorough review of publications in the field of voluntary assurance led to the identification of relevant research. First, Google Scholar was run to become acquainted with previous research on the assurance of financial reports (FRs), sustainability reports (SRs) and IRs. Second, scientific databases, EBSCO, Emerald, JSTOR, SpringerLink, and ScienceDirect helped in identifying quality papers. The search terms were "financial reporting" AND "assurance", "sustainability reporting" AND "assurance" and "integrated reporting" AND "assurance". Third, the literature review of Velte & Stawinoga (2017) supported finding additional publications in the field of SR assurance.

In total 23 studies in the years, 1982 to 2018, dealing with determinants for the assurance of FRs and SRs have been identified (Table 2). 21 studies dealt with the assurance of SRs and two dealt with the assurance of FRs. All of the publications investigated the determinants for voluntary assurance, 10 inquired the determinants for the choice of a particular type of assurer (e.g., accounting firm vs non-accounting firm) and four examined the determinants for the mandating of higher levels of assurance.

Geographically, 11 publications used samples of companies across the globe, six investigated European companies, four studied companies in the United States and one publication dealt with Australian and Chinese companies respectively. Based on the review of published studies in the field of voluntary assurance this research contributes to closing two critical gaps. First, it is the first effort to investigate the determinants of assurance decisions in IR. Secondly, it is the only study focusing on the unique South African setting, representing the first and only country to enforce a mandatory IR disclosure regime (De Villiers, Venter, & Hsio, 2017).

The most frequently used independent variables in previous publications fall into four categories:

- country-level variables
- industry-level variables
- CG variables and
- financial variables.

In sections 3.1.2 to 3.1.5, the most frequently used variables within each category are provided. First, for each variable, the expected correlation with the decisions to voluntarily assure, to select a provider from the accounting profession and to mandate higher levels of assurance is shown. Second, the reasoning for the expected correlation is explained based on the publications which investigated the respective variable. Third, the findings of multivariate regression of each variable in previous publications are presented.

Even though country-level variables are not relevant for the research model due to the focus on one country (South Africa), they are included in the literature review as these variables were frequently found to be among the critical determinants for assurance decisions. Hence, they provide relevant background to the conclusion and discussion in Chapter 6.

Table 2. Overview of Published Studies

Abbildung in dieser Leseprobe nicht enthalten

A= financial report, SR=sustainability report, ASR=assurance (Yes/No), ASRT=type of assurer (accounting, non-accounting), ASRL=level of assurance, CNT=country variables, IND=industry variables, CG=corporate governance variables, FIN=financial variables

In Figure 2 an overview of the variables investigated in previous publications and described in 3.1.2 to 3.1.5 is provided.

Figure 2. Overview of Variables in Publications

Abbildung in dieser Leseprobe nicht enthalten

BMs=board members, CSR=corporate social responsibility

3.1.2 Country-level Variables

Variable 1: Level of Stakeholder-Orientation

Expected Correlation: +

Reasoning: Stakeholders in countries with higher levels of stakeholder- orientation possess a more legitimate interest in corporate activities, wherefore SRs represent an adequate tool to manage stakeholder relationships.

Findings: Simnett et al. (2009), Kolk & Perego (2010) and Zhou et al. (2013) detected that companies in stakeholder-oriented countries, defined as countries applying code law, are significantly more likely to adopt SR assurance compared to companies in shareholder-oriented countries and are more inclined to mandate accounting firms.

Variable 2: Level of Legal Enforcement

Expected Correlation: -

Reasoning: Assurance represents a substitution tool for lacking legal enforcement mechanisms.

Findings: Simnett et al. (2009), Kolk & Perego (2010) and Zhou et al. (2013) measured that voluntary assurance and the choice of accounting firms as assurers is more likely to happen in countries with a lower level of legal enforcement. Opposed, Fernandez-Feijoo, Romero, & Ruiz (2015) and Braam & Peters (2018) did not detect any relationship.

Variable 3: Culture

Expected Correlation: +

Reasoning: The cultural setting shapes the behaviour of companies and decision-makers.

Findings: Martinez-Ferrero & García-Sánchez (2017) constructed a variable measuring the level of cultural systems and found that companies in countries with a higher long-term orientation and indulgence and less individualism, masculinity, uncertainty avoidance and power distance exhibit higher rates of assurance.

Variable 4: National Corporate Social Responsibility (CSR) Score

Expected Correlation: +

Reasoning: The higher the national CSR Score, the more are firms concerned to engage with stakeholders, leading to a higher emphasis on IR and IR assurance.

Findings: Sethi, Martell, & Demir (2017) detected an adverse effect of National CSR Scores on the level of assurance.

3.1.3 Industry-level Variables

Variable 5: Environmentally-Sensitive Industries (Yes/No)

Expected Correlation: +

Reasoning: Environmentally-sensitive industries are more dependent on public support, and thus have a higher need to offer credibility and reliability to their sustainability information through assurance.

Findings: The majority of studies did not find any significant industry effects. Ruhnke & Gabriel (2013), Sierra, Zorio, & García-Benau (2013) and Cho, Michelon, Patten, & Roberts (2014) found that companies operating in environmentally-sensitive industries are more likely to assure their reports, whereas Liao, Lin, & Zhang (2018) and Gillet-Monjarret (2015) reached opposite conclusions. There also has not been seen a significant effect of industry factors on the choice of a particular assurer. Whereas Peters & Romi (2014) detected a weak relationship, the majority of researchers did not identify any significant effect (Carey, Simnett, & Tanewski, 2000). In contrast, the level of assurance was mostly found to be significantly related with the belongingness to particular industries (Kolk & Perego, 2010).

Variable 6: Financial Industry (Yes/No)

Expected Correlation: +

Reasoning: The financial industry is a particularly sensitive industry, which is more dependent on public support, and thus has a stronger need to offer credibility and reliability to sustainability information through assurance.

Findings: Whereas Cho et al. (2014) and Castelo-Branco, Delgado, Ferreira Gomes, & Pereira Eugénio (2014) discovered a significantly positive relationship to the financial industry, most studies did not identify any significant effect on voluntary assurance decisions (Casey & Grenier, 2014; De Beelde & Tuybens, 2015; Cho, Michelon, Patten, & Roberts, 2014).

[...]