Master thesis
T H E E F F E C T O F O W N E R SH IP C O N C E N T R A T I O N O N T H E Q UA LI T Y
O F S US T A I N A B I L IT Y A S SU R A N C E
Name: V.K.P. Rosiek Student number: 11347643 Thesis supervisor: E. Gjata Date: June 25th, 2018 Word count: 16241
MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam
2
Statement of Originality
This document is written by student Varsha Rosiek 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 supervision of completion of the work, not for the contents.
3
Abstract
Due to investor and public concerns about harm to the environment, there is a growing trend among companies to publish their sustainability reports and voluntarily obtain sustainability assurance. Even though there are many benefits that arise from voluntarily obtaining sustainability assurance, the quality of this assurance seems to vary considerably in practice. Literature regarding sustainability assurance is currently scarce, and the quality thereof has primarily been examined using different types of assurance providers as a proxy. This research aims to expand the existing literature on the quality of sustainability assurance, and it examines whether ownership concentration is a determinant that explains the variation in the quality of sustainability assurance.
In this study, it is hypothesized that reporting companies with concentrated ownership are more likely to be associated with higher quality assurance on sustainability reports than those with dispersed ownership. This is because shareholders with concentrated ownership can participate in organizational decision making regarding the audit, which strengthens the independence of the assurance provider and leads to higher quality assurance. Based on a sample of 100 companies from countries around the world, the regression analysis reveals that ownership concentration has a marginally significant relation with the quality of sustainability assurance. However, after running robustness checks, the results demonstrate no significant effect for ownership concentration on the quality of sustainability assurance. The results suggest that reporting companies are successful in signaling the credibility of their sustainability reports through the purchase of voluntary assurance; therefore, concentrated shareholders do not find it necessary to influence the quality of sustainability assurance. To conclude, this study provides suggestions for future research in the field of sustainability assurance.
Key words: Sustainability reporting, assurance, quality of assurance, ownership concentration,
4
T A B L E O F C O N T E N T S
1. Introduction___________________________________________________________ 06
1.1 Background_____________________________________________________ 06 1.2 Discussion and research question____________________________________ 07 1.3 Motivation and contributions_______________________________________ 09 1.4 Structure_______________________________________________________ 10
2. Literature review______________________________________________________ 11
2.1 Assurance over sustainability reports_________________________________ 11 2.1.1 Sustainability assurance in practice___________________________ 11 2.1.2 Decision to voluntarily purchase sustainability assurance__________13
2.2 Assurance standards______________________________________________ 14 2.2.1 The AA1000AS__________________________________________ 14 2.2.2 The ISAE3000___________________________________________ 15 2.2.3 Comparing AA1000AS with ISAE3000_______________________ 16 2.3 Criticism of sustainability assurance_________________________________ 17 2.3.1 Lack of a standard model for sustainability assurance____________ 17 2.3.2 Expectation gap of stakeholders_____________________________ 18 2.3.3 Independence of the assurance provider_______________________ 18 2.4 The quality of sustainability assurance________________________________ 19
2.4.1 The quality of sustainability assurance and the type of assurance provider________________________________________________ 19 2.4.2 The quality of sustainability assurance and country-level investor protection_______________________________________________ 20 2.5 Theory and hypothesis development__________________________________21
2.5.1 Agency theory___________________________________________ 21 2.5.2 Signaling theory__________________________________________ 22 2.5.3 Prior research regarding the influence of ownership on organizational
5
3. Research method_______________________________________________________ 26
3.1 Sample description_______________________________________________ 26 3.2 Regression model and variables_____________________________________ 29 3.2.1 Regression model_________________________________________ 29 3.2.2 The quality of sustainability assurance________________________ 30 3.2.3 Ownership concentration___________________________________ 31 3.2.4 Control variables_________________________________________ 31
4. Results_______________________________________________________________ 35
4.1 Descriptive statistics______________________________________________ 35 4.2 Correlation between the variables____________________________________ 36 4.3 Testing the assumptions___________________________________________ 38 4.4 Regression analysis_______________________________________________ 38 4.5 Robustness checks_______________________________________________ 41
5. Discussion and conclusion_______________________________________________ 44
5.1 Limitations_____________________________________________________ 45 5.2 Suggestions for future research______________________________________45
6. References____________________________________________________________ 48 7. Appendices____________________________________________________________ 56
7.1 Appendix 1: Coding rules for determining the quality of sustainability
assurance______________________________________________________ 56 7.2 Appendix 2: Assurance quality per company___________________________ 59 7.3 Appendix 3: Assumptions dependent variable Assurance quality___________ 60 7.4 Appendix 4: Assumptions dependent variable Assurance quality 2014_______ 61 7.5 Appendix 5: Assumptions dependent variable Assurance quality 2015_______ 62
6
1 . I N T R O D U C T I O N
This chapter provides background information on the subject of assurance with regard to sustainability reports. Thereafter, it presents a discussion that forms the basis for the research question. Lastly, the motivation and contributions of this study that are to be made by answering the research question are described.
1.1 Background
The increased awareness of investors regarding social, environmental, ethical, and corporate governance issues in addition to economic performance, has resulted in a demand for sustainability reporting (Herda, Taylor & Winterbotham, 2014). According to the Global Reporting Initiative (GRI),1 a sustainability report consists of the most critical impacts of the organization—positive as well as negative—on the environment, society, and economy (GRI, 2013). It further helps an organization to set targets, measure performance, and manage change to make their operations more sustainable. Other reasons to publish a sustainability report include informing stakeholders, thereby reducing the information asymmetry between the company and the public; attesting to organizational commitment; and seeking to build corporate reputation and risk management (Simnett, Vanstraelen & Chua, 2009). In addition, sustainability reports provide the necessary non-financial information to external and internal stakeholders, and they are considered to be evidence of responsible business practice (Krivačić, 2017). However, the public’s inexperience with the origin and meaning of sustainability reports and a lack of common sustainability reporting standards allow managers a platform for opportunistic behavior (Peters & Romi, 2015). Since disclosures of non-financial information are less heavily regulated than disclosures of financial information, firms have significant latitude in choosing whether to disclose information and what, where, when, and how to do so when choosing to disclose (Cohen et al., 2012). This has resulted in a shift in reporting towards the assured sustainability report (Mock, Rao & Srivastava, 2013).
1 The GRI is a global non-profit organization that developed guidelines for sustainability reporting and assurance.
Their guidelines are regarded as the most important guidelines, as most companies that report on sustainability use their standards.
7 Simnett et al. (2009) mention that assurance serves as a useful control mechanism that enhances the credibility of disclosed information and facilitates greater user confidence, which results in more appropriate resource allocation decisions by information users. Coram, Monroe, and Woodliff (2009) agree with this, stating that assurance enhances the perceived credibility of non-financial disclosures. In addition, according to Alon and Vidovic (2015), assurance on sustainability reports has the potential to positively affect stakeholders’ perspectives regarding companies’ activities, and it can enhance the credibility of the information disclosed therein. However, the quality of sustainability assurance seems to vary considerably in practice (Bagnoli & Watts, 2015). The primary reason for this variation is that there is no standard model for assurance on sustainability reports, which results in different assurance providers selecting and performing different scopes of assurance procedures (Herda et al., 2014). Furthermore, there are no clear standards in the area of sustainability assurance (Kolk & Perego, 2010).
1.2 Discussion and research question
The purpose of this study is to examine the effect of ownership concentration on the quality of sustainability assurance. Similarly to the paper of Herda et al. (2014), an assurance statement in this paper is considered to be a high quality assurance statement when it contains the key elements for high quality assurance statements, as suggested by international organizations such as the GRI, AccountAbility2, the Federation of European Accountants3 (FEE), and the
International Federation of Accountants4 (IFAC). Furthermore, ownership concentration refers to
the amount of stock that investors own. Shareholders with concentrated ownership, compared to shareholders with dispersed ownership, have a stronger monitoring power over a firm’s managerial decisions, since those owners have the incentive to proactively safeguard their investment (Francis & Wilson, 1988). Those concentrated shareholders can take certain actions,
2 AccountAbility is a global non-profit organization, and it has developed international assurance standards for
sustainability reporting, namely AA1000AS.
3 FEE is a European organization that represents the accounting profession and it has provided guidance on
sustainability assurance, such as minimum required content for an assurance statement.
4 The IFAC is an organization for the accounting profession and it has developed ISAE3000, which is an
8 such as voting, electing board members, and participating in decision making regarding the audit (Boyd, 1994; Glac, 2014).
In a mandatory assurance environment, on the one hand—such as the assurance of a financial report—the company cannot choose whether to assure or not. Instead, it is mandated by law to assure its financial report (Herda et al., 2014). If the quality of an audit is high, then shareholders gain access to information that is more useful and credible, which in turn reduces information asymmetry (Kane & Velury, 2004). Prior research finds that shareholders use their concentrated ownership to participate in organizational decision making to obtain a higher audit quality (Dao, Raghunandan & Rama, 2012; Chan, Lin & Zhang, 2007). The reason behind this is that concentrated shareholders can participate in decision making regarding the audit, thereby strengthening the independence of the auditor, which leads to a higher audit quality. The results of these studies suggest that shareholders with concentrated ownership are more likely to foster independent auditing, which increases audit quality.
In a voluntary assurance environment, on the other hand—such as assurance of a sustainability report—the company can choose whether or not to assure its sustainability report (Cheng et al., 2015); it is not mandated by law. According to the signaling theory, reporting companies that obtain sustainability assurance can positively distinguish themselves from those that do not obtain sustainability assurance. This is because obtaining sustainability assurance reduces information asymmetry by allowing shareholders to differentiate more credible information from less credible information (Simnett et al., 2009). On the one hand, voluntarily purchasing sustainability assurance can thus be used as a tool to signal the credibility of its sustainability information. In this case, it is not expected that concentrated shareholders participate in organizational decision making regarding the audit and as such do not strengthen the independence of the assurance provider, and thereby do not affect the quality of sustainability assurance at all. They will interpret the signal as the company already reporting credible sustainability information, On the other hand, when it comes to a voluntary assurance environment, the reporting company can choose the assurance provider itself (Simnett et al., 2009), which impairs the independence of the assurance provider when the chosen provider works for the reporting company itself. As the independence of the assurance provider is a
9 known aspect for ensuring the quality of assurance (Ball, Owen & Gray, 2000), it is expected that shareholders will use their concentrated ownership to participate in organizational decision making regarding the audit, thereby strengthening the independence of the assurance provider to obtain a higher quality assurance, in a similar manner to that evidenced in prior research (Dao et al., 2012; Chan et al., 2007).
Based on the findings of prior studies (Dao et al., 2012; Chan et al., 2007), the answer to whether concentrated ownership affects the quality of assurance is clear in a mandatory, financial research setting. However, when it comes to a voluntary, non-financial research setting, there are contradictory arguments, and thus no answer can be given directly based on the knowledge of prior research. Therefore, this needs to be empirically examined. Thus, since it is unclear whether concentrated ownership actually affects the quality of sustainability assurance, this needs to be examined, which leads to the following research question:
RQ: Does ownership concentration affect the quality of sustainability assurance or not?
1.3 Motivation and contributions
This study contributes to the sustainability disclosure literature in the following ways. First, there has been a call for future research to examine the quality of sustainability assurance rather than solely its adoption (Kolk & Perego, 2010). This study responds directly to this call by investigating whether ownership concentration affects the quality of sustainability assurance. Second, according to Zorio, Garcia-Benau and Sierra (2013), the literature regarding sustainability assurance is scarce. Prior research related to sustainability assurance investigates the determinants of the choice to buy sustainability assurance (for example Simnett et al., 2009; Kolk & Perego, 2010). Other studies examine the quality of sustainability assurance by means of a content analysis, where assurance statements are compared with key elements for high quality assurance, as suggested by international organizations such as the GRI, AccountAbility, FEE, and IFAC, and based on the sustainability disclosures the assurance statement has been regarded as being of either a low or high quality (for example, O’Dwyer & Owen, 2005; Perego & Kolk, 2012). Up to now, differences in the quality of sustainability assurance have been mostly explained by the type of assurance provider, where the type of assurance provider is used as a
10 proxy for the quality of assurance (for example, Zorio et al., 2013; Peters & Romi, 2015). These studies assume that assurance providers from the accounting profession provide a higher quality assurance than those from the non-accounting profession because of their independence and ethics requirements, their maturity of standards, and the quality control mechanisms in place. Furthermore, there is only one study that identifies another proxy for the quality of sustainability assurance, namely the study of Herda et al. (2014), where country-level investor protection is used as a proxy for assurance quality. This study, on the other hand, is the first to identify whether ownership concentration affects the quality of sustainability assurance.
Since there are no clear standards in the area of sustainability assurance, this type of research should be of interest to accounting policy makers and practitioners to ensure higher levels of reliability, comparability, and homogeneity of the current practice of sustainability assurance (Kolk & Perego, 2010). Examining whether ownership concentration affects the quality of sustainability assurance, and thus examining whether this is one of the determinants that explains the variation in assurance quality, may provide useful insights into the level of pressure that concentrated shareholders have in a voluntary assurance environment, compared to a mandatory assurance environment, on the reporting firm and hence the credibility of sustainability assurance. The practical relevance of this study is that it explains the variation in assurance quality, which possibly impairs the accountability and transparency to stakeholders and the credibility of the assurance practice. Explaining the variation in assurance quality should be of interest to reporting companies, since it can help them to provide high quality information; it could also be useful to investors who demand credible sustainability information (Herda et al., 2014). Furthermore, insights into the practice of sustainability assurance lead to a better understanding of the information provided in the sustainability report.
1.4 Structure
The remainder of this paper is organized as outlined next. First, a comprehensive background is provided for the study, and the hypothesis is presented. Next, the research method is discussed, including the research design and sample selection, followed by the empirical results. The study concludes with a summary of the results, limitations, and future research opportunities.
11
2 . L I T E R A T U R E R E V I E W
In this chapter, the literature review of the study is presented. The chapter begins with sustainability assurance in practice and previous research regarding determinants of the adoption of sustainability assurance. Thereafter, the assurance standards are introduced and critical remarks regarding sustainability assurance are discussed. Finally, previous research that describes the quality of sustainability assurance is presented and the hypothesis is introduced, which builds on prior research regarding the influence of ownership on organizational decision making and the agency theory.
2.1 Assurance over sustainability reports
This section describes what sustainability assurance is and why companies choose to assure their sustainability reports.
2.1.1 Sustainability assurance in practice
Sustainability reporting and external assurance of these sustainability reports are growing (Simnett et al., 2009). This growth can be seen as a response to both investors that rely on the information as an indicator of underlying corporate risks and future performance, and other stakeholders concerned with social and environmental performance (Kolk & Perego, 2010). Like sustainability reporting itself, external assurance of sustainability reports has become a standard business practice (KPMG, 2015). The International Audit Assurance Standards Board (IAASB) defines an assurance engagement as “an engagement in which a practitioner aims to obtain sufficient evidence in order to express a conclusion designed to enhance the degree of confidence of the intended users other than the responsible party about the outcome of the evaluation or measurement of a subject matter against criteria” (IAASB, 2011). In addition, the GRI describes assurance as “activities designed to result in published conclusions on the quality of the report and the information (whether it be qualitative or quantitative) contained within it” (GRI, 2013). KPMG’s (2015) survey has indicated that 90% of the world’s largest 250 companies are using corporate responsibility reporting, and 67% of these companies invest in external assurance
12 statements. Also, companies seeking independent assurance has increased with 9 percent-point from 2005 till 2015, and an even higher increase of 33 percent-point for the world’s biggest 250 companies has been reported. This indicates that assurance on sustainability reports is becoming a common practice, especially for the large companies with extensive stakeholder groups. In addition, these companies often choose assurance providers from the auditing profession. Furthermore, it is also possible for companies to have only a part of their sustainability report assured. KPMG (2015) has found that 50% of the companies choose to assure their whole report, 34% choose specific indicators, 5% specific chapters, and 11% a combination of chapters and indicators.
According to Simnett et al. (2009), variations exist in assurance on sustainability reports between countries. Kolk and Perego (2010) have found that companies in Europe and Japan are most likely to produce assured sustainability reports. Chen and Bouvain (2009) have found that assurance on sustainability reporting is most frequently done in the United Kingdom, while it is
done least in the United States. In line with this, Casey and Grenier (2015) and Peters and Romi
(2015) have also found that assurance in the United States is far below the international levels. Factors explaining the lacking levels of assurance in the United States are regulatory oversight functions as a substitute form to enhance credibility instead of assurance, strict bank monitoring of highly leveraged firms suppressing the demand for assurance, and attribution to the hesitation or ineffective marketing of the accounting firms (Casey & Grenier, 2015). Furthermore, Kolk and Perego (2010) explain this lacking amount of assurance in the United Sates by heightened litigation risk and questionable benefits. These factors reduce the probability that the benefits of assurance outweigh the costs.
Even though, in most countries, it is a voluntary decision to have a sustainability report assured, groups such as the GRI recommend subjecting sustainability reports to such assurance (Herda et al., 2014). This is because assured information is generally perceived to be far more credible than information that is not assured (Hodge, 2001). However, the voluntary nature of sustainability assurance causes problems, because companies may choose for assurance to increase confidence regardless the quality (Bagnoli & Watts, 2016).
13 2.1.2 Decision to voluntarily purchase sustainability assurance
The existing literature suggests a number of reasons behind a company’s choice to purchase assurance services. Simnett et al. (2009) have investigated factors associated with the decision to voluntarily purchase assurance and the choice of assurance provider. Using a sample of over 2,000 companies from 31 countries, they have found that companies that seek to enhance the credibility of their reports and build their corporate reputation are more likely to have their sustainability reports assured, although it does not matter whether the assurance provider is from the auditing profession. They have also found that the demand for sustainability assurance is the strongest in the mining, finance, and utilities industries due to greater exposure to social and environmental risks and thus a greater need to enhance credibility.
Another study that has examined determinants of the adoption of sustainability assurance is the study of Kolk and Perego (2010). They have explored the factors associated with voluntary decisions to have a sustainability report assured. Using a sample of 636 multinational companies included in the Fortune Global 250 list, they have found that companies operating in stakeholder-oriented countries and with a weaker governance enforcement regime are more likely to adopt a sustainability assurance statement, and the demand for assurance is higher in countries where sustainable corporate practices are better enabled by market and institutional mechanisms. The reason is that assurance can play a substitutive role in controlling the credibility and quality of sustainability reports in countries with weak legal systems, and that companies in stakeholder-oriented countries are more likely to have their sustainability reports assured as a way to manage stakeholder relationships. In line with this, Francis et al. (2011) have found that assurance services serve as a substitute for weak institutions that constrain the contracting process of firms with other parties. They have also found that in countries with strong legal systems there is less assurance, because in stronger institutions the benefits of assurance do not outweigh the costs. Other researchers, who do not specifically examine a number of factors associated with a company’s choice to purchase assurance services, have found other benefits of voluntarily purchasing sustainability assurance. For example, assurance is associated with a reduction in the cost of capital, along with lower analyst forecast errors and dispersion; it lowers a company’s interest rates by reducing information asymmetry between the company and its lenders; it
14 enhances the quality of environmental disclosures by reducing information asymmetry; and it improves environmental management systems in the long run (Dhaliwal et al., 2012; Casey & Grenier, 2015; Blackwell, Noland & Winters, 1998; Moroney et al., 2012). Assurance on sustainability reports can further address investors’ concerns about companies engaging in greenwashing5, and it allows firms to use the sustainability reports to distinguish themselves
from less socially responsible firms (Lyon & Maxwell, 2010; Barnea & Rubin, 2010; Bagnoli & Watts, 2016).
Furthermore, KPMG’s (2015) survey has indicated that an improvement of the quality of the reported information, an improved reporting processes, and reinforced credibility among stakeholders are the main drivers of having a sustainability report assured. In addition, according to the GRI, assurance increases recognition, trust and credibility, and provides more confidence to stakeholders because it shows the seriousness of the reporter towards the sustainability report. Also, sustainability reporting and assurance can serve as a basis for a dialogue with stakeholders which could lead to mutual understanding. Next to that, assurance improves stakeholder communication and it can used as a tool to signal greater management ability (GRI, 2013).
2.2 Assurance standards
The AA1000 Assurance standard (AA1000AS) and the International Standard on Assurance Engagements (ISAE3000), both of which are used by assurance practitioners, have taken a dominant role (Kolk & Perego, 2010). This section describes and compares those two assurance standards.
2.2.1 The AA1000AS
The AA1000AS was developed by AccountAbility, and it offers a specific framework for sustainability assurance. The aim of AA1000AS is to address the expectation gap6 related to the
5 Greenwashing is the manipulation of the circulation of information by firms to mislead the public (Lyon &
Maxwell, 2011).
6 This expectation gap means that there are different perspectives about the duties and responsibilities of the
15 responsibilities of the assurance provider, the commitment to the interest of the public, independence issues, and formulation of the assurance statement (Mock et al., 2013). The AA1000AS is thus known for its unique focus on stakeholder accountability. It addresses anyone who performs external verification services (Kolk and Perego, 2010), and provides a methodology for assurance practitioners to evaluate the nature and extent to which an organization adheres to the AccountAbility principles (Manetti and Becatti, 2009). It further provides findings and conclusions on the progress of sustainability performance and encourages continuous improvements. There are two levels of assurance: type 1 and type 2. Type 1 assurance evaluates the materiality, responsiveness, and participation of an organization with stakeholders, while type 2 assurance extends the type 1 assurance by also evaluating the reliability of sustainability information.
2.2.2 The ISAE3000
The ISAE3000 was developed by the IAASB and does not specifically focus on sustainability reporting, but is intended for audits other than historical financial information (IAASB, 2011). The IAASB is an independent, standard-setting body and an arm of IFAC, and it contributes to the development of professional accounting bodies (Mock et al., 2013). As part of the principles of the IAASB, all external verification activities should mention the level of assurance to reduce the gap between the reader’s perception of the reliability of the verification and the actual effectiveness of it (Manetti & Becatti, 2009). Assurance conclusions can be classified into two levels, namely reasonable assurance and limited assurance. These levels of assurance indicate the extent and depth of work that the assurance provider has undertaken (GRI, 2013) and are therefore linked to a degree of confidence that the assurance statement affords the users of the assurance report. Regarding limited assurance, the nature, timing and extent of procedures for collecting sufficient evidence are less relative to reasonable assurance, and for that reason a limited level of assurance is a lower level of assurance that requires less work and hence lower costs. The conclusion is expressed in a negative form, which can be recognized by phrases such as “report contained no inaccuracies or misleading statements,” and “nothing has come to our attention” (Mock et al., 2013). In contrast, reasonable assurance indicates a high level of assurance, where the conclusion is expressed in a positive form. Positive assurance can be
16 recognized by phrases such as “fair and balanced representation, provides a fair account, accurately portrays the performance” (Mock et al., 2013).
According to Hasan et al. (2005), cost-benefit considerations play a role in the decision to choose reasonable or limited assurance. Assurance providers experience difficulties when it comes to collecting sufficient and appropriate evidence to provide reasonable assurance, which results in high costs of the assurance engagement, compared to limited assurance. Therefore, a firm weighs the benefits of increased user confidence in sustainability information against the costs of choosing a higher level of assurance (Simnett et al., 2009). However, according to Hasan et al. (2015), assurance on sustainability reports is mostly provided at a limited level.
2.2.3 Comparing AA1000AS with ISAE3000
A comparison of the AA1000AS with the ISAE3000 reveals that the ISAE3000 provides rigorous guidance about how an assurance engagement should be undertaken. The standards of ISAE3000 are more focused on the performance information that the client provides and describes procedures to check for material misstatements in the scope of the document. The AA1000AS on the other hand, are more focused on the relevance of the reported information for stakeholders, and they also explore stakeholder involvement in determining the subject matter as well as suitable criteria for the assurance engagement and the report in more depth (Mock et al., 2013). Non-accounting assurance providers might prefer to use the AA1000AS, since the organization that supplies this standard is not involved with accounting standard setting and because it is free from professional standards, such as the code of ethics, which might apply to assurance guidelines supplied by the IAASB (Mock et al., 2013). In contrast to assurance providers from the accounting profession, non-accounting assurance providers are not familiar with those standards, since they do not conduct financial audits. Furthermore, according to Hodge et al. (2009) and Perego (2009), non-accountants appear to focus more on completeness, fairness, recommendations and opinions in a sustainability assurance statement, while accounting firms focus on reporting format and procedures. Nevertheless, the standards do not directly compete among themselves. Some assurance providers reference the AA1000AS, ISAE3000, and GRI guidelines in different combinations, since they overlap in the minimum content of assurance (Kolk & Perego, 2010). KPMG (2005) has argued that this combined use of the
17 AA1000AS and the ISAE3000 will result in a better approach, credibility, methodology, and conclusion. Furthermore, there are a number of national (draft) standards that have emerged, such as Standards Australia in Australia and the Royal NIVRA in the Netherlands (Kolk & Perego, 2010).
2.3 Criticism of sustainability assurance
Although the disclosure and assurance of non-financial information remain mostly voluntary under the current reporting regimen of many countries (Cheng et al., 2015), assurance on sustainability reports has raised some critical remarks. This section describes each of them. 2.3.1 Lack of a standard model for sustainability assurance
There are some aspects that threaten the credibility of sustainability information (Smith, Haniffa & Fiarbrass, 2011). Unlike assurance of financial reports, there are various sustainability assurance guidelines and thus no standard model for assurance on sustainability reports (Herda et al., 2014). In line with this, Manetti and Becatti (2009) have commented on assurance standards that they are not standardized enough yet, that the level of assurance is not clearly explained, and that the links with financial audits and materiality are not clearly defined. Reporting companies have thus considerable involvement in selecting the scope of the assurance procedures performed and in choosing the level of assurance, which leads to substantial variations in the quality of assurance on sustainability reports. As a result, report users may not rely on the assurance report without first understanding the details of the actual assurance work performed (Herda et al., 2014). Furthermore, the voluntary nature of assurance on sustainability reports means that management is not obligated to follow the assurer’s recommendations (Moroney & Windsor, 2012). Moreover, since assurance on sustainability reports is not mandated by law, companies can have their sustainability reports assured by different types of assurance providers, such as environmental management firms (Simnett et al., 2009). Since there is usually an increased cost associated with having assurance provided by members from the auditing profession,7 the client
7 Discussions with assurance providers in this area in at least three countries suggest that fees charged by audit firms
can be commonly up to five times the fees quoted by environmental consultants for the same engagements (Simnett et al., 2009).
18 chooses the assurance provider on a costs versus benefits basis (Simnett et al., 2009). Furthermore, these assurance providers vary in their ability to assure sustainability reports (Bagnoli & Watts, 2016).
2.3.2 Expectation gap of stakeholders
It is also important to take the expectation gap of stakeholders into account when providing assurance. For example, in the case of an audit of financial statements, most investors believe that the audit should provide absolute assurance with respect to detecting material misstatements as a result of error and fraud, while the auditors believe that the audit should only provide reasonable assurance (Epstein & Geiger, 1994). This expectation gap also plays an important role when it comes to the assurance of non-financial information (Green & Li, 2011). Green and Li (2011) have found that, compared to preparers and shareholders, assurance providers perceive the level of responsibility for the assurance report to be lower and the credibility of assurance to be higher. Furthermore, Roebuck, Simnett and Ho (2011) have found that the perception of the level of assurance is high for reports that relate to historical information, compared to reports with prospective information. However, no difference was found regarding the description of the work performed. In addition, Hasan et al. (2005) have found that the difference between limited and reasonable levels of assurance is not well understood by users of the assurance report. An assurer’s awareness of stakeholders’ perceptions is important, and sustainability reporting preparers, users, and assurers must thus come to a shared view regarding the level of assurance provided in a sustainability assurance engagement in order to bridge the expectation gap.
2.3.3 Independence of the assurance provider
The independence of the assurance provider can be questioned, and according to Ball et al. (2000), it is one of the main factors that determines the quality of the assurance provided. The reporting company itself, instead of the stakeholders of the company, determines the conditions under which assurance is provided (Ball et al., 2000; Smith et al., 2011). This reflects the large degree of influence of management on the assurance process, even though assurance is supposed to be a way of adding credibility to the sustainability information provided (Simnett et al., 2009). This influence on the assurance process can form a threat and lead to a lack of relevance and
19 completeness of sustainability reporting (Casey & Grenier, 2015). Instead, an increase in stakeholder participation would increase the relevance and independence of the assurance provided, as it reduces the bias of management influence.
2.4 The quality of sustainability assurance
A number of studies have examined the quality of sustainability assurance. However, in these studies, the difference in assurance quality has been mostly explained by the type of assurance provider, with a distinction being made between members of the auditing profession and others. This section describes those studies.
2.4.1 The quality of sustainability assurance and the type of assurance provider
Pflugrath, Roebuck and Simnett (2011) have examined the impact that assurance on sustainability reports by assurers with different professional affiliations has on financial analysts’ perceived credibility of the information reported. Using an experimental design with 106 participants, they have found that when sustainability information is assured by a professional accountant, financial analysts perceive the source of that information to be more credible in terms of its trustworthiness, expertise, and overall credibility. Therefore, financial analysts view professional auditors as providing a greater level of independent and expert assurance than sustainability experts. The researchers have further mentioned that this could be due to the accounting profession’s reputation, supported by a well-developed body of international standards and ethics, and quality control mechanisms at both the firm level and the engagement level. The standards help to ensure that assurance engagements on sustainability reports are appropriately staffed, with the engagement team having the requisite auditing expertise as well as the necessary, specific, subject matter expertise.
Another study that has examined the sustainability assurance quality is that of Zorio et al. (2013). In this study, they investigated whether explanatory variables could be identified for companies’ decisions to publish sustainability reports, to assure sustainability reports or not, and to hire assurance services from an auditor or a consultant and the subsequent quality of the assurance report. Using a sample of 130 Spanish companies in the period 2005-2010, they have found,
20 amongst other things, that auditors provide higher quality assurance on sustainability reports. However, this result only holds for a sample of 130 companies in a Spanish setting, and the only distinction that was made was between auditors and consultants.
Furthermore, Peters and Romi (2015) have examined whether sustainability-oriented corporate governance mechanisms impact the voluntary assurance of corporate sustainability reports. They have made a distinction between three common types of assurance providers within the sustainability report assurance arena, namely internal auditors, consultants, and professional accounting firms. Since professional accountants are subject to independence and professional conduct requirements, and their work is guided by professional standards, they promote the quality of assurance services. However, the authors have also mentioned that higher costs to the assurance purchaser are associated with this higher quality. Therefore, an alternative is to engage sustainability consultants to provide assurance (Simnett et al., 2009). Peters and Romi (2015) have further emphasized the role of internal auditors, since they have the ability to both add reliability to reported information and provide this at lower costs, compared to external assurance providers. Using a sample of 912 U.S. sustainability reports in the period 2002-2010, they have found, amongst other things, that internal auditors are used for sustainability assurance more often than consultants and professional accounting firms.
2.4.2 The quality of sustainability assurance and country-level investor protection
An exception is the study of Herda et al. (2014), which is the only study where the type of assurance provider has not been used as a proxy for assurance quality. In this study, they examined the impact of country-level investor protection on reporting companies’ voluntary sustainability assurance decisions. More specifically, they examined whether a firm’s decision to voluntarily assure its sustainability report is a substitution monitoring mechanism that replaces poor investor protection, and they assessed whether the quality of voluntary assurance is associated with the variation in investor protection environments. Using a sample of approximately 600 companies in the period 2005-2009, they have found that managers in low investor protection countries are more likely to obtain both assurance on sustainability reports and higher quality assurance. This is in line with the results of Kolk and Perego (2010), which
21 indicate that assurance can play a substitutive role in controlling credibility of sustainability reports in countries with weak legal systems.
Even though existing literature has examined the assurance statements in order to draw a conclusion on the quality of the assurance statement, it is still not clear about the actual determinants regarding the quality of sustainability assurance. As mentioned earlier, in most studies, the differences in the quality of the assurance provided have been based on different types of assurance providers. However, according to Herda et al. (2014), using assurance provider type as a proxy for assurance quality may speak only to the reputation effects of assurance providers as opposed to actual assurance quality. Different from the other studies, this study investigates whether ownership concentration is a determinant of the quality of sustainability assurance.
2.5 Theory and hypothesis development
This section begins with the agency theory and the signaling theory, which are frequently used theoretical perspectives in research on sustainability reporting and sustainability assurance. Then, prior research regarding the influence of ownership on organizational decision making is presented; that research, along with the agency theory, forms the basis for the hypothesis.
2.5.1 Agency theory
According to the agency theory, managers are the agents of a firm’s owners, and they do not bear all the consequences of their own decisions (Jensen & Meckling, 1976). This theory predicts that the separation of ownership and control results in both goal incongruence, where the goals of the principal and the agent conflict, and information asymmetry, where the principal does not know what the agent is doing (Eisenhardt, 1989). The risk is thus that the separation of ownership and control leads to suboptimal decisions that managers might make to serve their own interests (Fama & Jensen, 1983). Costs associated with this information problem between the agent and the principal are called agency costs (Eisenhardt, 1989). Agency costs can be defined as the sum of the monitoring expenditures by the principal, the bonding expenditures by the agent, and the residual loss (Jensen & Meckling, 1976). According to Leuz and Verrecchia (2000), a decrease
22 in the gap of available information and hence lower agency costs can be achieved by a higher level of voluntary disclosure and transparency in the firm’s overall reporting, together with an increase in the credibility and reliability of the information presented. Furthermore, according to Piot (2001), audit quality can be considered as one of the main monitoring mechanisms to regulate conflicts of interests and reduce agency costs.
The agency theory can also be applied in the context of sustainability reporting and assurance. According to Cohen et al. (2012), the disclosure of non-financial information, such as sustainability disclosures, is essential to reduce the information asymmetry that exists between management and important stakeholders; it also allows investors to better assess key areas of performance, and it supports a broader view of corporate performance that also encompasses society at large. However, these disclosures are only useful if they are perceived to be credible (Coram et al., 2009). If sustainability reporting is not viewed as credible, and thus discounted or ignored by investors and analysts, then sustainability assurance reduces information asymmetry and increases user confidence in previously considered information (Carey, Simnett & Tanewski, 2000). From the agency perspective, the demand for assurance results from the need to reduce agency costs associated with information asymmetry and the resultant loss of control due to a lack of observability of managers’ behavior (Kolk & Perego, 2010; Chow, 1982).
2.5.2 Signaling theory
The signaling theory is closely related to the agency theory. The signaling theory describes the behavior of two parties that have access to different information (Connelly et al., 2011), and it suggests that a firm tries to credibly provide information to other parties in situations of asymmetric information (Hahn & Kühnen, 2013). The information asymmetry between outside parties and the management of the firm relates to the fact that it is difficult for outside parties to assess the sustainability performance of firms. The firm must choose whether and how to report sustainability information, and the public must choose how to interpret the signal (Connelly et al., 2011). Since investors make their decisions based on the information available to them, firms signal positive information to the public, while negative information will not be signaled. Assuming that there is a distinction between high-performing firms and low-performing firms, it is beneficial for high-performing firms, on the one hand, to signal their true performance to
23 distinguish themselves from low-performing firms (Connelly et al., 2011). On the other hand, low-performing firms would not benefit from divulging their true low performance, and outside parties would thus know that the firms that signal their information are high-performing firms. In a similar way, purchasing third-party assurance on sustainability reports helps firms to reduce information asymmetry by allowing shareholders to discern more credible information from less credible information (Simnett et al., 2009) and to distinguish between socially responsible and less socially responsible firms (Bagnoli & Watts, 2016). Firms can thus obtain sustainability assurance to signal the credibility of their sustainability information (Simnett et al., 2009). Assurance on sustainability reports is particularly important for positive information, since reporting negative information is viewed as credible without assurance (Coram et al., 2009). 2.5.3 Prior research regarding the influence of ownership on organizational decision making and hypothesis building
Existing literature has found that ownership has an influence on organizational decision making, motivation, and power (Finkelstein, 1992). However, when the owners of a company are dispersed, and none of them own a significant percentage of the company’s stock, none of the owners will have significant incentives to spend their time monitoring and evaluating managerial decisions (Oh, Chang & Martynov, 2011). Managers consequently exert power over a company’s decision making, and they make decisions that serve their own interests. Even in the case of some minor shareholders who are willing to monitor managerial decisions, they would not have the voting power on the board to affect corporate decision making (Oh, Chang & Martynov, 2011). In contrast, shareholders who do own a large percentage of the stock will usually have the power to affect corporate decisions through different channels, such as voting (Glac, 2014), appointing directors to the board (Boyd, 1994), and shareholder activism (Smith, 1996). The presence of large shareholders is thus likely to enhance direct monitoring and significantly reduce managers’ power (Francis & Wilson, 1988).
Furthermore, prior research has found a positive association between shareholder involvement in organizational decision making and audit quality in a financial reporting setting. For example, Dao et al. (2012) have found that shareholder involvement in organizational decision making is associated with higher audit quality. This is because increased shareholder participation in
24 organizational decision making is likely to strengthen auditor independence and lead to higher audit quality. In these situations, shareholders participate in organizational decision making, and the management or audit committee does not make decisions regarding the audit without shareholder approval. Also, the investors become monitors of the auditor’s work, and the auditor becomes a direct agent of the shareholders rather than of the board of directors and/or management. This aligns the auditor’s incentives more with shareholders than in cases where shareholders do not participate in organizational decision making. Furthermore, Chan et al. (2007) have found an association between changes in ownership structure and audit quality. This is because when the institutional ownership of a firm increases as a result of a decrease in individual ownership, the firm is likely to switch from a lower-quality auditor to a higher-quality auditor. Compared to individual shareholders, institutional shareholders represent highly concentrated shareholders that can vote on important issues, such as the appointment of the auditor, and they can also bring significant pressure to bear on management to ensure that an effective corporate governance structure is in place. Individual ownership, on the other hand, is widely dispersed; individual investors do not have much influence in most listed companies, and they may not be able to closely monitor management. Overall, these results suggest that shareholders with a concentrated ownership are more likely to foster independent auditing, which increases audit quality.
Applied to this non-financial research setting, firms with concentrated ownership may be more likely to be associated with higher quality assurance on sustainability reports than firms with dispersed ownership, as concentrated shareholders can participate in organizational decision making to strengthen the independence of the assurance provider, which leads to a higher quality assurance, in a similar manner to that evidenced in prior research (Dao et al., 2012; Chan et al., 2007). Even though reporting companies can signal the credibility of their sustainability reports through the purchase of voluntary assurance, compared to reporting companies that do not assure their sustainability reports, the independence of the assurance provider might still be impaired in a voluntary assurance environment. This is because the reporting company can choose the assurance provider (Simnett et al., 2009) and hence an assurance provider that works for the company itself. In this case, the assurance provider is not independent of the reporting company, and shareholders will use their concentrated ownership to ensure the independence of the
25 assurance provider in order to obtain a higher quality assurance. This leads to the following hypothesis:
H1: Reporting companies with concentrated ownership are more likely to be associated with higher quality assurance on sustainability reports than reporting companies with dispersed ownership.
26
3 . R E S E A R C H M E T H O D
In this chapter, the research method is presented; it provides information about the nature of research to be conducted. The chapter begins with the sample size and availability of data. Thereafter, the regression model is presented. Finally, the variables used in the research are discussed.
3.1 Sample description
This paper identifies many sustainability reports and their corresponding sustainability assurance statements that are published. The archival method is appropriate to study whether ownership concentration is associated with the quality of assurance on sustainability reports (Cohen & Simnett, 2015). According to Mock et al. (2013), Corporate Register (http://corporateregister.com) is the world’s largest database for non-financial reports, and according to Simnett et al. (2009), it is a comprehensive directory of published corporate environmental and social reports. Therefore, Corporate Register was used to collect a sample of 180 companies that participate in the Corporate Responsibility Reporting Awards (CRRA)—the only global annual awards for corporate responsibility reporting. Even though companies that participate in CRRA are considered as companies that report high quality sustainability reports, a lot of these companies, such as Hennes & Mauritz, BMW, ExxonMobil, Royal Dutch Shell and Volkwagen, have suffered from some serious scandals due to harm to the environment (Oldenkamp, Van Zelm & Huijbregts, 2016). Shareholders’ awareness of these companies’ harm to the environment is likely to result in increased participation in organizational decision making in order to obtain higher quality assurance, which gives the information environment more transparency and can prevent such scandal from happening again. Furthermore, choosing companies that take part in the CRRA was likely to result in a sample of companies that report on sustainability and possibly buy external assurance on their sustainability reports.
The sample period covers the period 2011-2017 because PwC finds a boom in corporate sustainability reporting in this period (PwC, 2018). Also, since some companies do not publish their sustainability reports on an annual basis, this multi-year window provides the best
27 possibility of gaining a representative sample (Mock et al., 2013). However, the focus periods become 2014 and 2015, since there were no available data for the previous years and because of time constraints, as I scored the quality of assurance for each company by myself. Tables 1 and 2 summarize the sample companies per country and per industry respectively. Table 1 indicates that the sample comprises companies mostly from France (6.11%), the United States (11.67%), Japan (16.11%), South Korea (8.33%), the United Kingdom (10.0%), South Africa (5.56%), and Spain (7.22%). Table 2 demonstrates that the industries represented are retail (17.22%), healthcare (3.33%), electronics (8.33%), telecommunications (8.33%), financial services (23.33%), energy (5.00%), manufacturing (11.67%), business services (10.56%), and utilities (12.22%).
Additionally, consistent with related prior research (Simnett et al., 2009), Corporate Register was supplemented by the GRI database (http://www.globalreporting.org) to download the published sustainability report. According to Herda et al. (2014), the GRI is a network-based organization that produces a comprehensive sustainability reporting framework, which is widely used around the world. However, the sustainability reports included in the GRI database are not restricted to reports based on the GRI standards. Companies are thus able to voluntarily include their reports in the database; however, this also means that not all existing reports are included in the database. Since Corporate Register and the GRI database do not always provide the opportunity to obtain the sustainability reports, the relevant companies’ websites were visited to check whether their sustainability reports are available. When sustainability reports are not available in the databases as well as on a company’s website, the assumption is made that the company has not published a sustainability report.
Corporate Register and the GRI database further include information to examine which companies from the sample that publish sustainability reports have obtained external assurance on those reports. This was double-checked on each company’s website because of a possible omission of information in Corporate Register and the GRI database. For each company that publishes a sustainability report, the company’s website was consulted to retrieve the assurance statement. In total, from the initial sample of 180 companies, 180 general sustainability reports were identified, issued over the period between 2014 and 2015. This could be either a separate,
28 stand-alone sustainability report or, if not available, an annual report, if it contained this type of information (integrated report), consistent with related prior research (Kolk & Perego, 2010).
Table 1: Sample companies across country
Country Frequency Percent
Canada 4 2.22 France 11 6.11 India 3 1.67 Belgium 2 1.11 Ireland 1 0.56 The Netherlands 9 5.00 Australia 6 3.33 Italy 5 2.78 Denmark 3 1.67 Israel 1 0.56 United States 21 11.67 Norway 2 1.11 Finland 3 1.67 Germany 9 5.00 Japan 29 16.11 South Korea 15 8.33 Turkey 1 0.56 Sweden 6 3.33 China 2 1.11 United Kingdom 18 10.00 South Africa 10 5.56 Spain 13 7.22 Portugal 2 1.11 Switzerland 4 2.22 Total 180 100.00
29
Table 2: Sample companies across industry
Industry Frequency Percent
Retail 31 17.22 Healthcare 6 3.33 Electronics 15 8.33 Telecommunication 15 8.33 Financial services 42 23.33 Energy 9 5.00 Manufacturing 21 11.67 Business services 19 10.56 Utilities 22 12.22 Total 180 100.00
A number of reports were eliminated from the initial sample for two reasons. First, corresponding financial information was missing regarding a number of reports. Second, the assurance report information in some reports was either missing or presented in a foreign, non-English language. After making the eliminations described above, there were 100 reports with their corresponding assurance statements remaining. Figure 1 provides an overview of the composition of the sample. The final sample consists of 97 public, and 3 private companies.
3.2 Regression model and variables
This section presents the regression model and describes the variables used in the analysis, along with their measurements.
3.2.1 Regression model
After collecting the data, the hypothesis was analyzed to determine whether there is evidence to not reject the hypothesis. To research whether the dependent variable is influenced by the independent variable, the following regression model was used:
30
Figure 1: Corporate Register companies’ sustainability reporting and assurance choices
where AQ relates to the sustainability assurance quality. The independent variable is ownership concentration (OWN), and the control variables include company size (SIZE), profitability (ROA), leverage (LEV), and assurance provider (AP). Table 3 provides an overview of the definitions of the variables used in this study. The regression analysis was conducted using the statistical software SPSS.
3.2.2 The quality of sustainability assurance
The dependent variable of this paper, namely the quality of assurance on sustainability reports, has been examined in previous literature, where assurance statements are compared with key elements for high quality assurance, as suggested by international organizations such as the GRI, AccountAbility, FEE, and IFAC (for example, Herda et al., 2014; Perego & Kolk, 2012). Since the quality of audit work is not directly observable, researchers have used different measures for audit quality (Dao et al., 2012). In this study, the measurement by Perego and Kolk (2012) was used, as it is the most recent instrument for measuring the quality of sustainability assurance that
31 includes minimal requirements for high quality, as prescribed by the GRI, IFAC, and AccountAbility. The scale, along with the coding rules for determining the quality of sustainability assurance, can be found in appendix 1. Based on the scoring methodology of Perego and Kolk (2012), each company can achieve a score ranging from 0 to 27 points. Reasonable assurance instead of limited assurance, the use of publicly available criteria instead of publicly unavailable criteria, and completeness regarding the material aspects are examples of key elements that indicate a high quality assurance.
3.2.3 Ownership concentration
Ownership concentration is the independent variable in this paper. It was measured by data on the largest shareholder, expressed as a percentage of a company’s share capital (Brammer & Pavelin, 2006). Data regarding ownership concentration were retrieved from the Orbis database of Bureau van Dyke, as of December 31st of the year to which the assurance statement relates. The Orbis database covers large firms worldwide, and it provides direct and indirect ownership information as well as extensive financial information (Deephouse & Jaskiewicz, 2013; Voget, 2011). Therefore, the Orbis database is appropriate to retrieve ownership information. For each company, I searched this database for information about the shareholders regarding the years 2014 or 2015; thereafter, I processed the ownership percentage of the largest shareholder in the dataset.
3.2.4 Control variables
Since company size, profitability, leverage, and assurance provider have an influence on the quality of assurance, these variables were included as control variables (Francis & Yu, 2009; Zorio et al., 2013; Kane & Velury, 2004; Adeyemi & Fagbemi, 2010; Piot, 2001; Skinner & Srinivasan, 2012). Data on the company size, profitability, and leverage control variables were also retrieved from the Orbis database, since it contains extensive financial information about firms worldwide (Deephouse & Jaskiewicz, 2013; Voget, 2011). Data were retrieved as of December 31st of the year to which the assurance statement relates. In the case of companies that had missing financial data in the Orbis database, their annual reports and websites were checked.
32 Furthermore, data on the assurance providers were found in the assurance statements. A brief description of the control variables in the dataset and their measurement is presented next. Company size was included as a control variable, as the existing literature finds a positive relationship between company size and audit quality. According to Zorio et al. (2013), the size of the company has been found to be significant in achieving higher quality assurance on sustainability reports. This may indicate that large companies negotiate for higher quality in the assurance process or that the assurance provider is influenced by higher visibility of its work if it is a large company, and as a result, the assurance quality increases. Consistent with these studies, a positive coefficient is expected for the variable SIZE, since these studies suggest that it is positively associated with the quality of assurance. Company size was measured by the natural logarithm of the total asset value (Moroney et al., 2012; Clarkson, Li, Richardson & Vasvari, 2008; Hay, Knechel & Wong, 2006; Blackwell et al., 1998). As this sample consists of companies from countries around the world, data on total assets are available in different currencies. In line with prior research, the exchange rates on December 31st were used to convert the data on total assets available in different currencies into millions of U.S. dollars (US$), which were then divided by 10,000 for the purpose of including them in the analysis (Mock et al., 2013). The exchange rates for different currencies were obtained from http://www.exchange-rates.org.
Profitability was also included as a control variable in the analysis. It was measured by return on assets (ROA), assessed by net income divided by total assets (Moroney et al., 2012; Simnett et al., 2009). According to Moroney et al. (2012), ROA is a forward-looking accounting measure of company performance, resulting from management’s productive use of company assets. Skinner and Srinivasan (2012) find that firms with better performance, as measured by ROA, have lower demand for a high audit quality. As this study suggests a negative association between the company’s ROA and audit quality, a negative coefficient is expected for the variable ROA. Kane and Velury (2004) find that leverage has a positive relationship with audit quality. In a similar way, Adeyemi and Fagbemi (2010) demonstrate a positive association between leverage and audit quality. Piot (2001) states that there is a higher demand for audit quality when leverage increases. These results suggest that audit quality may reinforce debtholders’ protection. There
