Assurance on sustainability reports: the
variability and differences in quality
Master thesis, MSc Accountancy
University of Groningen, Faculty of Economics and Business
June, 24th 2019 Nena Tans S3456277 Rabenhauptstraat 2-2 9725 CD Groningen Tel: +31 6 81 14 99 55 e-mail: n.l.tans@student.rug.nl Word count: 9337
Assurance on sustainability reports: the
variability and differences in quality
ABSTRACT: This study investigates the variability and differences of the quality of sustainability assurance over the years, and to what extent this quality is influenced by the type of provider of assurance, the industry where the company is operating in, and the legal system of the country where the company is domiciled in. The sample for this research consists of all Dutch and UK listed companies available in the GRI database who received assurance on their sustainability report, covering the reporting years from 2013 until 2017. In total 253 firms are selected. The research question is empirically tested through a one-way ANOVA test and regression analyses. The results do not provide evidence for differences in quality over the years. Additionally, this study found relationships between the provider of assurance and the quality, and between the industry where the company is operating in and the quality. These results suggest that companies receiving assurance from a provider of the audit profession and companies operating in more polluting industries, receive assurance of higher quality. Moreover, there is no evidence found for the expectation that companies domiciled in stakeholder-oriented countries receive higher quality assurance compared to companies domiciled in shareholder-oriented countries.
Keywords: sustainability reports, sustainability assurance, quality of sustainability assurance, assurance provider, industry, shareholder/stakeholder-oriented countries, agency theory, stakeholder theory
TABLE OF CONTENTS
1. INTRODUCTION ……… 4
2. THEORY AND HYPOTHESIS DEVELOPMENT ………. 9
2.1. Quality of external sustainability assurance ………… 10
2.2. Assurance provider ……….. 12 2.3. Industry ……….. 13 2.4. Legal system ……….. 14 3. METHODOLOGY ……….. 16 3.1. Sample selection ……… 16
3.2. Measurement of the variables ……… 18
4. RESULTS ……… 22 4.1. Descriptive statistics ……… 22 4.1.1. Normality ………. 24 4.1.2. Correlation ……… 24 4.2. Hypotheses testing ………. 26 4.3. Additional analyses ……….. 29 4.3.1. Outliers ………. 29
4.3.2. Standardization assurance reports …………. 31
5. CONCLUSION AND DISCUSIION ………. 32
5.1. Findings ………. 32
5.2. Implications ……….. 34
5.3. Limitations and recommendations ……… 34
5.4. Overall conclusion……… 35
REFERENCES ………. 36
Appendix 1: Content analysis quality ……… 40
Appendix 2 : Frequencies quality scores Netherlands………….. . 42
1. INTRODUCTION
n recent years, sustainability has become a frequently discussed subject in the news. Companies are reporting more and more about it, and sustainability reporting has become an often used subject for multiple studies (Hahn & Kühnen, 2013; Dhaliwal et al., 2011). Looking at the top 100 companies from 49 countries worldwide (N100), 72% of the companies are reporting on Corporate Responsibility (CR) (KPMG, 2017). This percentage is even higher when investigating the 250 world’s largest companies (G250), based on revenue, 93% of these organizations are reporting on CR. CR information consists of information about economic, social and environmental aspects (Gallego-Alvarez & Quina-Custodio, 2016). The CR information is discussed in sustainability reports and the company reports on its contributions, positive or negative, towards the goal of sustainable development (GRI, 2006).
The demand for transparency and disclosure of environmental and social issues from various stakeholders has increased rapidly. Simnett, Vanstraelen & Chua (2009) argue that traditional financial reports do not represent the multiple dimensions of corporate value anymore. As a result, companies either add financial information to the financial report, or report on non-financial information in a separate manner. Since 2018, large Dutch organizations are required to report on non-financial information, which may include CR information (Ministerie van Economische Zaken, 2017). In the UK, companies are required to report on a number of mandatory issues (The companies Act 2006 Regulations 2013). Their strategic report should include information on gas emissions, diversity and human rights and they have to show that the supply chains are free from slavery and report on it. However, most companies report far beyond these requirements (PWC, 2015). The UK belongs, with 92%, to the top 3 of countries with the highest rates of CR information in annual financial reports (KPMG, 2017).
Despite the increased sustainability reporting, stakeholders have remained skeptical due to their view that these reports could only contain positive news about environmental issues while in practice this might not be the case (Bilal Farooq & de Villiers, 2019). In addition, this skeptical view is also due to the lack of consistency and completeness of the sustainability reports (Martínez-Ferrero et al., 2018).
In response, the increase in reports on sustainability is followed by an increase in demand for external assurance on the sustainability report (Cohen et al., 2011; Zorio et al., 2013).
The amount of companies investing in external assurance on their sustainability reports has grown steadily every year (KPMG, 2017). External assurance on sustainability reports can increase the recognition, trust and credibility for a company’s stakeholders (Simnett et al., 2009; GRI, 2013). It also reduces agency costs due to an increased confidence about the accuracy and validity of the report (Kolk & Perego, 2010). Furthermore, investors give value to assured sustainability reports, as it say something about the seriousness and confidence of the published reports (Cuadrado-Ballesteros et al., 2017; Hodge et al., 2009; Zorio et al., 2013; Fuhrmann et al., 2017). External and independent reviews are a voluntary service provided by parties such as accountancy firms, engineering firms, and sustainability firms, with the purpose of adding assurance (GRI, 2013).
Due to the voluntary choice whether to assure the sustainability reports or not, there are no laws and regulations or required reporting principles for the content of both the sustainability reports as well as the assurance reports. Despite the lack of laws and regulations, there are some guidelines and/or standards companies can use. Globally accepted examples are the Global Reporting Initiative (GRI), AccountAbility, the Sustainability Accounting Standards Board (SASB), and the International Integrated Reported Council (IIRC). For the audit of sustainability reports, auditors may use the ISAE standard, although this standard is not specific for sustainability purposes. Auditors in the Netherlands do have a specific standard for sustainability assurance, namely the NV COS 3810N.
The absence of legally binding regulations, standards, and reporting principles about reporting requirements, which guidelines or standards to use, how to give assurance and who is qualified to provide assurance, leads to differences in the quality of assurance services provided (KPMG, 2013). Hummel et al. (2019) argue that the voluntary choice raises questions about the quality of the assurance provided. The presence of different assurance providers and different standards influences the quality of the assurance services provided (Perego & Kolk, 2012).
In addition, it becomes difficult for investors to make comparisons of performances among companies and reports over time, when the quality and content of the sustainability and assurance reports differ (Eccles et al., 2012). The variation in the content of the sustainability reports could also result in differences in the quality of sustainability assurance (Simnett et al., 2009).
The purpose of this study is to explore the variability and differences of the quality of sustainability assurance over the years. There are different proxies to measure the assurance quality, since the quality itself is not directly observable (Martinez-Ferrero et al., 2018). In addition, different users of the sustainability reports have different perspectives on how to define high quality audits (Knechel et al., 2013). Following prior research (O’Dwyer & Owen, 2005; Perego & Kolk, 2012; Zorio et al., 2013; Martinez-Ferrero et al., 2018), the quality of the assurance report is used as a proxy to measure the quality of sustainability assurance, as the audit quality can be viewed through the accuracy of the audit reports (Knechel et al., 2013). The quality of the assurance report is evaluated quantitatively using a content analysis. Knechel et al. (2013) argue that audit quality is a perceived trait rather than a directly observed one. They state that a ‘balanced scorecard’ with key attributes of auditing is a way to facilitate perceptions about the quality. In this research, the content analysis with key attributes is used to observe the assurance quality.
This research assesses several independent variables which are expected to influence the quality of sustainability assurance. The first independent variable is the provider of assurance. The GRI database distinguished three types of assurance providers: audit firms, engineering firms and consultancy/boutique firms (GRI, 2018). For this research, the providers are divided into two groups: audit firms (from the audit profession) and non-audit firms, such as engineering and consultancy/boutique firms (Kolk & Perego, 2010; Martinez-Ferrero et al., 2018). This distinction is due to the expectation that different types of assurance providers cause different levels of assurance quality, because the assurance providers undertake a variety of approaches and all have different backgrounds (O’Dwyer & Owen, 2005).
The second variable is the industry in which the sustainability reporting company operates in. Companies in certain industries will disclose more information about sustainability than those in other industries, because of the divergence in environmental activities (Hahn & Kühnen, 2013). Stakeholders of companies in more polluting industries have a higher demand for having the sustainability reports assured (Birkey et al., 2016). It is expected that companies in more polluting industries have a greater demand for, and a greater likelihood of high quality assurance, because the stakeholder pressure for these particular industries is relatively higher (Hahn & Kuhnen, 2013). Therefore, the industries distinguished by the GRI database are classified into two categories: more polluting industries and less polluting industries. The third assessed variable is related to the legal system of the country where the company is located in, with the focus on whether the country is shareholder-oriented or stakeholder-oriented. The legal system of the country where the company is located in, may influence the demand for high quality assurance (Simnett et al., 2009; Cho et al., 2014; Zhou et al., 2016). The distinction is twofold, countries are separated in stakeholder-oriented countries (code law legal system) and shareholder-oriented countries (common law legal system). Therefore, data from the Netherlands is used as a stakeholder-oriented country and data from the UK is used as shareholder-oriented country (Ball et al., 2000). Taken together, the research question is as follows:
’’Does the quality of sustainability assurance differ over time; and to what extent is the quality influenced by the provider of assurance, the industry, and the legal system of the country where the company is located in?’’
There are several studies available on sustainability reporting, however, the studies and literature on the quality of sustainability assurance are limited. Prior literature about sustainability assurance quality has mainly focused on other settings. Perego & Kolk (2012), for example, used a panel of 250 Fortune Global 250 firms over a period of 10 years and investigated how the diversity in assurance standards and different providers influenced the quality. Zorio, García‐Benau & Sierra (2013) developed an index to measure quality and looked at the assurance reports of Spanish companies. Another study from Martinez-Ferrero et al. (2018) also investigated the quality of sustainability assurance reports, but focused on the expertise and experience of the assurance providers as determinants. This research extends prior literature about sustainability assurance and contributes to its predecessors by providing additional research on sustainability assurance quality.
As mentioned before, certain advantages of assurance on sustainability reporting are increased trust and credibility for stakeholders as well as the fact that it may reduce agency problems and costs (Simnett et al. 2009; Kolk & Perego, 2010; GRI, 2013). In addition, investors value assured sustainability reports because the assurance says something about the reliability of the sustainability reports (Cuadrado-Ballesteros et al., 2017; Hodge et al., 2009; Zorio et al., 2013; Fuhrmann et al., 2017). When the majority of sustainability assurance reports are of poor quality, these advantages can fade away. Therefore, next to contributing to academic literature, this research is also relevant for practitioners and regulators, since it can contribute to the knowledge of practitioners and regulators of the expected differences in quality and content. This study could be useful to stimulate the quality and content of reports to be unambiguous and it will provide input for reporting principles, which contributes to higher quality assurance reports, as there is a need for mandatory reporting principles/standards (Eccles et al., 2012; Kolk & Perego, 2010).
The remainder of this paper is organized as follows. The next chapter consists of a theoretical framework and a literature review. In chapter 3 the research methodology is presented. Thereafter, the results are given in chapter 4. Finally, the conclusion and discussion is presented in chapter 5.
2. THEORY AND HYPOTHESIS DEVELOPMENT
Often used theories which are widely used in sustainability assurance research are the agency theory, the stakeholder theory, the legitimacy theory, the stewardship theory and some studies use the isomorphism theory, which focuses primarily on the perspective of the company. This research is mainly based on two theories that explain the need for high quality assurance on sustainability, respectively the agency theory, and the stakeholder theory.
The agency theory explains the conflict of interests between the principle and the agent, which may lead to information asymmetry due to different goals and risk preferences (Eisenhardt, 1989; Hussain et al. 2016). Both the principal and the agent want to maximize their individual utility (An et al. 2011). The principal-agent relationship can be applied in different situations, for example between employer and employee, company and debt holder, and between the shareholders and CEOs. In this study, the agency theory specifically applies to the stakeholders of the investigated companies (principles) and the firms reporting on sustainability (agent). The voluntary disclosure of sustainability reports with the purpose of informing stakeholders, partly mitigates the information asymmetry between the company and the stakeholders (Simnett et al., 2009). This line of reasoning is also argued by Kolk & Perego (2010), who state that the demand for sustainability reporting comes from the need to reduce agency costs, which is associated with information asymmetry. However, Cuadrado-Ballesteros et al. (2017) found that the agency problem may not be mitigated with reporting on sustainability alone, due to incorrect or incomplete reporting of sustainability. They found that the information asymmetry can only be mitigated with the issuing of high quality assurance on sustainability. In addition, Cuadrado-Ballesteros et al. (2017) found that having high quality assurance on sustainability reports results in a lower information asymmetry. The audit and assurance of the sustainability report can serve as a monitoring response to the agency conflict when the audit is voluntary and the benefits for the company are greater, due to the reduced agency costs and higher confidence of users (Carey et al., 2000). Hence, reporting on sustainability alone may not mitigate the information asymmetry, while receiving high quality assurance on the reported sustainability information does contribute to further mitigation of the information asymmetry.
The stakeholder theory emerged as a common used theory in corporate social responsibility (Mcwilliams & Siegel. 2001). This theory puts more focus on aligning the interests of stakeholders, and thereby expands the agency theory, which mainly focusses on shareholders (An et al., 2011). According to the stakeholder theory, the company should address multiple goals with respect to the stakeholders, rather than only those goals with regard to the shareholders. The demand for sustainability reporting and assurance by stakeholders groups, such as investors, employees, suppliers, debt-holders, governmental bodies, and society in general, has increased. According to Comyns (2016) the sustainability reports are produced as a response to the information needs of stakeholders and stakeholder pressure. This theory emphasizes the importance of meeting the interests of the relevant stakeholders, in order to achieve success (Zhou et al., 2016)
The choice of voluntarily reporting on sustainability and the demand for high quality sustainability assurance can be derived from both of these theories. The agency theory helps to explain the demand for mitigating the information asymmetry, to which the stakeholder theory is added to explain the demand from the stakeholders, and not only the shareholders (An et al. 2010). To conclude, high quality assurance is needed to further mitigate the information asymmetry between the company and their stakeholders (Cuadrado-Ballesteros et al., 2017).
2.1. Quality of external sustainability assurance
In recent years, companies are reporting more and more on sustainability. Despite this increase, stakeholders have remained skeptical and the possible benefits of the sustainability reports stayed away (Kolk & Perego, 2010; Martinez-Ferrero et al., 2018; Bilal Farooq & de Villiers, 2019). In response, the demand for external assurance on the sustainability reports increased (Cohen et al., 2011; Zorio et al., 2013). However, due to the lack of legally binding regulation and standards or principles, the quality of these assurance reports can vary.
The purpose of this study is to explore the variability and differences of the quality of sustainability assurance over the years. Despite several studies on assurance quality, it remains a vague understanding as researchers have not yet determined a common definition or a common way to measure the quality of assurance (Hummel et al., 2019). The content analysis is a widely used method in studies on the quality of sustainability assurance to measure quality. Previous research has found that the quality of sustainability assurance is weak or that there are significant differences in quality. However, these studies address different causes, or examined different determinants of sustainability assurance quality. One of the first researches on this subject was done by Kamp-Roelands (2002), who examined assurance statements on sustainability reports and developed a framework for the audit of sustainability reports. At the time this research was done, the variety in the audit approach and audit report already existed. According to Kamp-Roelands (2002), this variety leads to a loss of confidence and to a loss of the demand for the services. In later years, O’dwyer & Owen (2005) examined if the assurance statements give attention to transparency and accountability to stakeholders. They found that a minority of the statements referenced these subjects. Following O’dwyer & Owen (2005), the absence of these subjects can be related to the voluntary nature of the assurance. This line of reasoning is also used by Hummel et al. (2019), who argue that the voluntary choice raises questions on the quality of the assurance. Perego & Kolk (2012) found that, for their data sample, the quality of assurance is highly dependent upon the standards that are used, and found a significant variability in the used standards. The variability in executed processes and the use of standards is also found by Gurturk & Hahn (2016). This variability results in fewer benefits. Besides, due to the diversity in sustainability information and the subjective nature of it, the sustainability reports may be complex to evaluate where specific skills are needed for. This variation in the content of the sustainability reports could also affect the quality (Simnett et al., 2009).
In this regard, even though the quality appears to be improving (Perego & Kolk, 2012), it is still expected that there are differences in the quality of assurance on sustainability reports. These differences are likely to depend on different determinants.
2.2. Assurance provider
External assurance on sustainability reports can be performed by a variety of parties, such as audit firms, engineering firms, sustainability services firms, or environmental consultancy firms/boutique firms (GRI, 2013). Several researches into quality of sustainability assurance showed that the differences in quality are related to the type of assurance provider. However, different viewpoints exist. It is expected that different types of assurance providers causes differences in the quality over years and between companies, because the assurance providers undertake different approaches and have different backgrounds (O’Dwyer & Owen, 2005). Martinez-Ferrero et al. (2018) found, for example, that the assurance quality is higher when the assuror is of the auditing profession. They relate this to the greater skills and training within the audit profession. The audit profession has stringent education and experience benchmarks before practitioners are allowed to enter the profession and they constantly need to update their education (Huggings et al., 2011). Assurance providers from the audit profession are assumed to provide higher quality assurance due to the fact that these have well-developed standards, independence, ethical requirements, and quality control mechanisms that are mandatory (Simnett et al., 2009; Huggins et al., 2011). Furthermore, audit firms have a high level of reputation capital to live up to, as they need to protect the brand name of the organization (Zhou et al., 2016).
However, results of Perego & Kolk (2012) show that there is almost no difference in quality whether it is assured by an audit firm or a non-audit firm. Arguments for this are that practitioners of the audit profession do not have enough knowledge on the subject matter to assure sustainability information (Huggins et al., 2011). Engineering firms, sustainability service firms or environmental consultancies have a competitive advantage because of their specific skills and knowledge. They are specialized in assurance on sustainability information and may have more experience and expertise with opinions and concerns of stakeholders, while audit firms provide assurance on more and broader topics and might have less experience with including all stakeholders in the audit (O’Dwyer, 2001).
Consequently, non-audit firms may provide more comprehensive assurance statements (Hodge et al., 2009; Hummel et al., 2019). Deegan et al. (2006a) found that these assurance statements from non-audit firms mostly include recommendations or commentary on the processes and systems, where audit firms do not have those comprehensive assurance statements.
However, the audit teams these days are composed of members from diverse backgrounds. In addition, the teams from audit firms could hire specialists to include this special knowledge and subject matter competencies in the team (Simnett et al., 2009). The work of an external specialist is also recorded in the standards for the practitioners of the audit profession, namely ISA 620. Therefore, it is expected that assurance providers from the audit profession are more likely to give high quality assurance than assurance providers who are not from the audit profession.
H2: Assurance providers of the audit profession are more likely to give high quality assurance than assurance providers who are not of the audit profession
2.3. Industry
Reporting on sustainability only is not enough to increase the confidence of the stakeholders, since they remained skeptical due to the view that these reports could contain only positive news about environmental issues while in practice this might not be the case (Bilal Farooq & de Villiers, 2019). Due to incorrect or incomplete reporting, Cuadrado-Ballesteros et al. (2017) found that the confidence and credibility can only be increased with the issuing of high quality assurance on the sustainability reports. The assurance may be used to improve the environmental image of companies (Birkey et al., 2016).
It is assumed that different types of industries have different environmental or social impacts, and disclose more about sustainability than other industries (Hahn & Kühnen, 2013). KPMG (2017) shows that, for example, the Mining industry, which is a more polluting industry, has high CR reporting rates. The results of Simnett et al. (2009) and Perego & Kolk (2012) show that industries with activities yielding greater environmental or social impacts will have more risks, and thus a greater need assurance to manage these risks, and to increase the confidence of the stakeholders.
Stakeholders of companies in polluting industries have a higher demand for the assurance of their sustainability reports (Birkey et al., 2016). Perego & Kolk (2012) also found that the quality of the assurance is higher for companies in more polluting industries. Chemicals, Energy (Utilities), Financial Services, Mining and, Water Utilities are examples of companies operating in more polluting industries (Simnett et al., 2009).
In this research, it is expected that companies in more polluting industries have a greater demand for and likelihood of high quality assurance. The reason for this is that the stakeholder pressure for these industries is likely to be higher (Hahn & Kuhnen, 2013). Besides, the demand for high quality assurance may be higher, due to the desire to enhance the credibility and confidence of the sustainability report (Patten, 2002, Simnett et al., 2009; Kolk & Perego, 2010).
H3: Companies in polluting industries are more likely to have high quality assurance than companies in less polluting industries
2.4. Legal system
Two types of countries exist with regard to their legal systems: countries with a common law legal system and countries with a code law legal system (Kolk & Perego, 2010). Common law countries are shareholder-oriented countries, where companies are primarily acting in the interests of the shareholders. Code law countries are stakeholder oriented and focus on the relationship with all stakeholders. More emphasis is placed here on social and environmental responsibilities. In this research, companies domiciled in the United Kingdom are used as companies from a shareholder-oriented country, where the companies domiciled in the Netherlands are used as companies from a stakeholder-oriented country (Ball et al., 2000; Zhou et al., 2016).
The legal system of the country where the company is located in, can influence the demand for high quality assurance (Simnett et al., 2009; Kolk & Perego, 2010; Cho et al., 2014; Zhou et al., 2016). Smith et al. (2005) found that companies domiciled in stakeholder-oriented countries have higher quality sustainability reports than shareholder-oriented companies with less emphasis on social and environmental issues.
Sustainability reports with higher quality information may also lead to a higher quality assurance, since the content of the sustainability reports could affect the quality of the assurance (Simnett et al., 2009). In a stakeholder-oriented country, other groups than only shareholders have influence on the companies’ activities. Also, these countries are, in general, more interested in CR subjects (Simnett et al., 2009). In addition, the stakeholder pressure and the demand to assure this information and get high quality assurance is likely to be higher (Simnett et al., 2009; Zhou et al., 2016). It is therefore expected that companies domiciled in a stakeholder-oriented country are associated with a higher assurance quality compared to companies domiciled in a shareholder-oriented country.
H4: Companies domiciled in a stakeholder-oriented country are more likely to have high quality assurance compared with companies domiciled in a shareholder-oriented country
3. METHODOLODY
In this section, the used method to test the hypotheses is discussed. In the remainder of this section the sample (3.1.), and the measurement of the dependent variable, independent variables and the control variables (3.2.) are explained.
3.1. Sample selection
This research contains archival research and is focusing on all listed Dutch and United Kingdom (UK) companies available in the GRI database. The choice for these countries is made due to the different legal systems they have (Kolk & Perego, 2010). The Netherlands has a code law legal system, with a stakeholder-oriented view. The United Kingdom has a common law legal system, which is shareholder-oriented (Ball et al., 2000; Zhou et al., 2016). To test the variability and differences of sustainability assurance quality over the years, the sample of this research covers the reporting years from 2013 until 2017. This time period is the most recent one available in the GRI database; not all reports publicized in 2018 are already available. GRI (2013) uses the definition of the International Federation of Accountants (IFAC) of assurance engagement: ‘’an engagement in which a practitioner expresses 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’’
(IFAC, 2005, p. 6). GRI database tracks critical reporting and associated organizational data, including information on the external assurance and the assurance reports which can be derived from the database. The sustainability assurance reports from the database includes reports where different criteria and assurance standards are used, for example IASE 3000 or AA1000. The data can be voluntary added to the GRI database. During the research, it appeared that not in all cases the assurance reports for all years were available on the database. Therefore, for each company available in a certain year, the website of the company was consulted for the reports of other years when available. Data was obtained from a total of 471 sustainability/integrated reports from the Netherlands and 1.291 sustainability/integrated reports from the UK, as shown in Table 1.
Table 1
Sampling distribution by country
Year Netherlands United Kingdom
2013 141 254 2014 134 269 2015 81 305 2016 97 274 2017 18 189 Total 471 1291
For this study, the companies without external assurance on their sustainability reports are eliminated from the initial dataset. Besides, this research is focusing on standalone sustainability reports only. This means that the integrated reports are also eliminated from the initial dataset. The decrease in the amount of reports, as shown in table 2, is relatable to the fact that more companies are reporting their sustainability information in integrated reports. The reason for eliminating integrated reports is the fact that assurance reports on integrated reports also contain assurance on financial information, which is subject to several laws and regulations. With sustainability reports alone, the effects of the voluntary choice and the missing laws and regulations on the quality can be measured resulting in a more uniform way of reporting. Table 2 shows an overview of the sample selection.
Table 2
Sample selection
Year Total amount of companies Total amount of companies with
with standalone sustainability standalone sustainability reports
reports and external assurance and external assurance in the
in the Netherlands United Kingdom
2013 32 45 2014 35 41 2015 33 40 2016 24 36 2017 14 15 Total 138 177
In some cases the assurance reports from the database were not available online. These missing sustainability assurance reports were eliminated from the sample, because it is not possible to conduct a content analysis on these reports. In table 3, the total sample is broken down to its constituents.
Table 3
Breakdown total sample Netherlands United Kingdom Total amount of sustainability reports 571 1.291 Less: Companies with integrated reports and/or (433) (1.114)
companies without external assurance
Less: Assurance reports that were not available (24) (38)
Final sample 114 139
In conclusion, the final sample of the research consists of 114 sustainability assurance reports of the Netherlands and 139 sustainability assurance reports of the United Kingdom.
3.2. Measurement of the variables
Quality
There are different proxies to measure the assurance quality, since the quality is not directly observable (Martinez-Ferrero, 2018). Besides, different users have different perspectives about how to define high quality assurance (Knechel et al., 2013). According to Knechel et al. (2013) the audit quality can be viewed through the accuracy of the audit reports. They argue that audit quality is a perceived trait rather than a directly observed one. They state that a ‘balanced scorecard’ with key attributes of auditing is a way to facilitate perceptions of the quality. Therefore, and following prior research (O’Dwyer & Owen, 2005; Perego & Kolk, 2012; Zorio et al., 2013; Martinez-Ferrero et al., 2018), the quality of the assurance report is used as a proxy to measure the quality of sustainability assurance. The quality of the assurance reports is measured using a content analysis, which is relatable to the balance scorecard argued for by Knechel et al. (2013).
The content analysis of Perego & Kolk (2012) is used with the expansion of 4 criteria points from Gurturk & Hahn (2016). Gurturk & Hahn (2016) included the given recommendations, progress in reporting, reporting on reservations, and reflection on examined work. Gurturk & Hahn (2016) argue that the added criteria give an indication of the quality of the assurance, whereas the initial criteria say more about the content of the assurance reports. The initial content analysis of Perego & Kolk (2012) is based on the framework developed by O’Dwyer & Owen (2005). This framework identifies, in line with the guidelines of AccountAbility, GRI standards and the Federation of European Accountants (FEE), the minimum requirements of a high quality assurance statement. Perego & Kolk (2012) made a few improvements to this framework to avoid bias. In the content analysis, assurance reports can receive scores from zero to 33, with 33 representing the highest quality. Most of the points can be derived from the criteria points scope, objective, criteria used, materiality, responsiveness to stakeholders, suggestions for improvements, and the extent of stakeholder participation in the assurance process. If these criteria points are clearly described, the assurance report can receive two points for each criteria point. An assurance report can even receive 3 points when they provide reasonable assurance, but this is only the case for a few assurance reports. The used content analysis is shown in Appendix 1.
Assurance provider
The stated hypothesis of the variable is as follow: ‘’Assurance providers of the audit profession
are more likely to give high quality assurance than assurance providers who are not of the audit profession’’. To be able to test whether assurance providers from the audit profession are
more likely to give high quality assurance, a distinction is made between audit firms (from the audit profession) and non-audit firms. The type of assurance provider can be found in the GRI database. The GRI database distinguished the type of assurance providers into audit firms, engineering firms and consultancy/boutique firms (GRI, 2018). The engineering and consultancy/boutique firms belong to the non-audit firms. The audit firms, belonging to audit firms (from the audit profession) have a dummy variable with the value of 1. The dummy value of non-audit firms equals 0.
Industry
The third hypothesis in this study is as follow: “Companies in polluting industries are more
likely to have high quality assurance than companies in less polluting industries”. Therefore,
the industries in the GRI database are classified in two categories: more polluting industries and less polluting industries. Chemicals, Energy (Utilities), Financial Services, Mining and, Water Utilities are industries belonging to the more polluting industries (Simnett et al., 2009). According to Simnett et al. (2009) these industries, including the financial services industry, have the highest environmental and social risks. The other industries distinguished by the GRI database belong to the less polluting industries. A dummy variable is used for this classification, where polluting industries have the value of 1 and less polluting industries have the value of 0.
Legal system
The third assessed variable is about the legal system of the country where the company is domiciled in, from the perspective whether the country is shareholder-oriented or stakeholder-oriented. The following hypothesis is stated: ‘’Companies domiciled in a stakeholder-oriented
country are more likely to have high quality assurance compared with companies domiciled in a shareholder-oriented country’’. A distinction is made between a stakeholder-oriented (code
law legal system) country and a shareholder-oriented country (common law legal system). Therefore, data from the Netherlands is used as a stakeholder-oriented country and data from the United Kingdom is used as shareholder-oriented country (Ball et al., 2000). This is converted into a dummy variable, where the stakeholder oriented country is equal to 0 and the shareholder oriented country is equal to 1.
Control variables
Besides the independent variables, some control variables are added in the regression model. These variables are included to control for the possible effects on the sustainability assurance quality. The added control variables are based on Simnett et al. (2009), Zorio et al. (2013), and Martinez-Ferrero et al. (2018). The following control variables are used: firm size (FSIZE), firm profitability (ROA), and the firms leverage (LEV).
Firm size (FSIZE) is measured by the natural logarithm of the total assets (Martinez-Ferrero et al., 2018). Firm profitability (ROA) is measured by the Return on Assets. The leverage of the firm (LEV) is measured by the ratio of total debt to total equity. Studies found that size and leverage are associated with the voluntary choice and demand to get external assurance (Chow, 1982; Simnett et al., 2009). The variables are collected from the Compustat database, a globally used financial data platform, which includes all information needed for the control variables. In Table 3, the variables are summarized.
Table 3
Variable description
Variable Measurement Database
Quality (QUALITY) Measured through a content analysis GRI of the sustainability assurance reports,
where the reports can receive scores from zero to 33, where 33 is the highest quality level.
Assurance provider (PROV) Dummy variable which equals 1 if GRI the assurance provider is an
audit firm (from the audit- profession) and 0 if the provider is not an audit firm
Industry (IND) Dummy variable which equals 1 if GRI the industry belongs to a more
polluting industry and 0 if it belongs to a less polluting industry
Legal system (LEGAL) Dummy variable which equals 1 if GRI the company is domiciled in a stake-
holder oriented country and 0 if the company is domiciled in a shareholder oriented country
Firm size (FSIZE) Natural logarithm of the total assets Compustat Firm profitability (ROA) Return on Assets Compustat Firm leverage (LEV) Ratio of total debt to total equity Compustat
To test whether the hypotheses are true, regression analyses are conducted. Taken the variables together, the following regression model is developed:
4. RESULTS
The results of the study will be discussed in this section. First some descriptive statistics of the variables, normality tests and a correlation matrix are provided and explained, then in the second paragraph, the results of the hypotheses testing will be provided.
4.1. Descriptive statistics and correlations
The descriptive statistics are shown in Table 4. The frequency statistics show that 139 firms are from the United Kingdom and 114 firms from the Netherlands, which results in a total of 253 firms. This means that 54,9% of the firms are domiciled in a shareholder-oriented country. For 181 (71,5%) of the 253 firms, the assurance provider was a member of the auditing profession. Besides, 33,2% of the firms are operating in more polluting industries. Regarding the years, the declining percentage of assurance statements accompanying stand-alone sustainability reports is notable. However, the decline is in line with the earlier mentioned reason, that more and more firms report their sustainability information in integrated reports, which are not included in this research.
Table 4 Frequencies variables Frequency Percentage UK 139 54.9 NL 114 45.1 Total 253 100.0 Audit firms 181 71.5 Non-audit firms 72 28.5 More polluting 84 33.2 Less polluting 169 66.8 2013 50 19.8 2014 60 23.7 2015 57 22.5 2016 49 19.4 2017 37 14.6
Furthermore, as presented in table 5, the mean quality of the assurance reports is 14.439, indicating that the average score given to the assurance reports of the firms in the sample is 14.439. One of the control variables used in this study is the size of the company, measured as the natural logarithm of total assets. The total asset numbers are obtained from the Compustat database and were transformed in excel to a natural logarithm. The mean natural logarithm of total assets is 9.379. The range from 4.8 (total assets of 117 million euro) until 14.8 (total assets of 2.634.139 million euro) shows the large variance in the size of firms which are included in the sample.
Table 5
Descriptive statistics
N Min Max Mean Std. Dev.
Quality 253 6.0 24.0 14.439 2.8911
Ass. Provider 253 0 1 Dummy Dummy
Industry 253 0 1 Dummy Dummy
Country 253 0 1 Dummy Dummy
SIZE 253 4.8 14.8 9.379 2.0019 LEV 253 0.001 66.26 4.304 8.7941 ROA 253 -1.37 0.415 0.033 0.1395
The majority of assurance reports fall between 11 and 17 points, as shown in table 6. In paragraph 4.3. an additional analysis is conducted where these statistics are split for NL for each year to look for standardization in the assurance reports.
Table 6
Frequencies scores quality
Score Frequency Percentage Score Frequency Percentage
6 1 0.4 16 19 7.5 8 2 0.8 17 16 6.3 9 10 4.0 18 6 2.4 10 4 1.6 19 11 4.3 11 19 7.5 20 4 1.6 12 14 5.5 21 3 1.2 13 33 13.0 22 4 1.6 14 59 23.3 23 2 0.8 15 45 17.8 24 1 0.4 Total 253 100.0
4.1.1. Normality
To perform statistical analyses, it is assumed that the dependent variable is normally distributed. In order to statistically test if the quality of sustainability assurance is normally distributed, a Kolmogorov-Smirnov test is conducted. This test is used because the sample is greater than 50, for which this test is the most appropriate.
Table 7
Kolmogorov-Smirnov test of Normality
Statistic df Sig.
Quality .162 253 0.000
When normality is tested using a Kolmogorov-Smirnov test, a non-significant p-value indicates that we may assume that de data is normally distributed. As presented in table 7 above, the p-value of the dependent variable is 0.000, indicating that de dependent variable quality is not normally distributed. However, following the Central Limit Theorem this is not a problem in this study, since the assumption of normality is especially important in small samples. For a sample greater than 30, the Central Limit Theorem assumes that the assumption of normality has been met.
4.1.2. Correlation
In order to detect multicollinearity issues, the correlations between the different variables are presented in table 8 with the use of a Pearson Correlation matrix. Multicollinearity could unintentionally influence the estimations of the regression coefficients, since there is a partial overlap. Multiple significant correlations are found, with some of them expected to be
correlated. The multicollinearity of the variables is too high if the Pearson correlation coefficient is greater than 0.7. The coefficients in the correlation matrix of this research are below 0.7.
Table 8
Pearson correlation matrix
Variable 1 2 3 4 5 6 7 Quality 1 Ass. Prov. .138* 1 Industry .178** .073 . 1 Country -.061 .289** .003 1 Size .002 .127* .303** -.083 1 Leverage .060 .137* .329** .179** .252** 1 ROA -.078 -.112 -.281** -.045 -.038 -.069 1
* correlation is significant at the 0.05 level (2-tailed) ** correlation is significant at the 0.01 level (2-tailed)
In addition, the variance inflation factor (VIF) of each variable to detect issues of
multicollinearity are calculated and shown in table 9. The low values of VIF indicate that there are no issues of multicollinearity, since the values are less than 5. A VIF higher than 5 indicates issues of multicollinearity.
Table 9
Variance inflation factor
Collinearity Statistics
Variable Tolerance VIF
Assurance provider .883 1.133 Industry .774 1.293 Country .870 1.150 SIZE .850 1.176 Leverage .826 1.210 ROA .908 1.102
analyses for hypothesis 2, 3 and 4 are conducted and discussed below.
Hypothesis 1: “The quality of sustainability assurance differ over time”
To analyze the differences between the years (different groups), a one-way ANOVA is conducted. An ANOVA-test provides the possibility to compare two or more groups. The results are presented in table 10. A condition for an ANOVA test is that the variances for each group are the same. Based on the ‘Test of Homogeneity of Variances’ it can be concluded that the different years have equal variances, since the Sig. value is ,684. If the Sig. value is greater than ,050, the variances are equal. The ANOVA showed no statistically significant difference in quality between the different years, F(4,248) = ,317; p = ,867. The p-value (0,867) is insignificant, and therefore the first hypothesis is rejected.
Table 10
One-way ANOVA test for hypothesis 1
Descriptives 95% Confidence
interval for mean
Year N Mean Std.dev. Std. Error Lower Upper Min Max
2013 50 14,18 3,0552 ,4321 13,31 15,05 6,0 21,0 2014 60 14,58 3,1098 ,4015 13,78 15,39 8,0 23,0 2014 57 14,52 2,9038 ,3846 13,76 15,30 9,0 24,0 2016 49 14,65 2,8472 ,4067 13,84 15,47 9,0 22,0 2017 37 14,14 2,3940 ,3936 13,34 14,93 9,0 21,0
Test of Homogeneity of Variances
Levene Statistic df1 df2 Sig.
,570 4 248 ,684
ANOVA Sum of squares df Mean Square F Sig.
Hypothesis 3: “Companies in polluting industries are more likely to have high quality assurance than companies in less polluting industries”
Hypothesis 4: “Companies domiciled in a stakeholder-oriented country are more likely to have high quality assurance compared with companies domiciled in a shareholder-oriented country”
First, a regression analysis without control variables is performed. When the control variables ‘Size’, ‘Leverage’, and ‘Return on Assets’ are excluded from the regression analysis,
statistically significant relations are found between the quality and type of assurance provider (p= 0,015), and between the quality and the industry that the company is operating in (p= 0,007). These results indicate that audit firms have a positive effect on the sustainability assurance quality, as well as companies that operate in more polluting industries. The relationship between the legal system of the country where the company is domiciled in and the quality is only statistically significant at the 10% significance level (p= 0,099). The regression model without control variables is shown in model 1 in table 11 and explains 4,7% of the variation in quality (Adjusted R square = 0,047).
The results for hypothesis 2 (model 2), hypothesis 3 (model 3), and hypothesis 4 (model 4) including the control variables are provided in table 11. The results are for each independent variable individual. However, in practice these variables appear together, and therefore, are analyzed together in regression model 5.
Based on the linear regression analysis of model 5, the assurance provider (p= ,027) and the industry (p= ,033) are statistically significant related to the quality. Therefore, based on this model, these hypotheses can be accepted. The other hypothesis, related to the legal system of the country, is not statistically significant (p= ,141). Based on this analysis, there is no reason to accept hypothesis 4. The variables in this model explain 3,5% of the quality (Adjusted R-square = 0,035).
Variable Model 1 Coeff. (Std. Err.) t (p-value) Model 2 Coeff. (Std. Err.) t (p-value) Model 3 Coeff. (Std. Err.) t (p-value) Model 4 Coeff. (Std. Err.) t (p-value) Model 5 Coeff. (Std. Err.) t (p-value) Intercept 13,041 (,452) 28,833 (,000) 12,556 (,900) 13,950 (,000) 13,153 (,913) 14,412 (,000) 12,796 (,897) 14,271 (,000) 12,973 (,913) 14,209 (,000) SIZE ,153 (,098) 1,561 (,120) ,113 (,103) 1,092 (,276) ,163 (,101) 1,606 (,109) 0,014 (,112) ,124 (,901) LEV ,001 (,022) ,037 (,971) -,005 (,022) -,244 (,808) ,007 (,023) ,316 (,752) ,000 (,023) ,014 (,989) ROA -1,454 (1,310) -1,110 (,268) -,896 (1,367) -,656 (,513) -1,720 (1,307) -1,316 (,189) -,454 (1,372) -,331 (,741) PROV 1,005 (,412) 2,438 (,015) ,690 (,412) 1,674 (0,095) ,976 (,437) 2,231 (,027) IND 1,021 (,378) 2,703 (,007) ,857 (,450) 1,904 (0,058) ,960 (,449) 2,138 (,033) LEGAL ,618 (,373) 1,657 (,099) ,256 (,388) ,659 (,511) ,605 (,409) 1,478 (,141) R-square 0,058 0,034 0,037 ,025 0,058 Adjusted R-square 0,047 0,018 0,022 ,009 0,035 F-value 5,097 2,180 2,390 1,574 2,538
removing or winsorizing the outliers from the dependent variable is tested. Second, to look for standardization in the assurance reports of the Netherlands, the quality scores for each year are presented in descriptive statistics.
4.3.1. Outliers
Since the results can be highly sensitive to outliers, the outliers are removed from the dataset in this analysis. As shown in figure 1 below, there are 10 outliers identified. These outliers are all so called ‘true’ outliers, which means these data points actually deviates from the rest of the sample and have a high or low quality score, and are not errors.
Figure 1: Boxplot with outliers
The results of the regression analysis without the outliers are shown in table 12. These results show that, at a 5% significance level, the variables are all not significant. This means that the outliers affects the results.
Table 12
Regression analysis without outliers
Variable Coeff. Std. Err. T Sig..
Constant 13,086 ,797 16,410 ,000 SIZE ,044 ,098 ,448 ,655 LEV ,004 ,020 ,202 ,840 ROA -,955 1,206 -,792 ,429 PROV ,597 ,395 1,514 ,131 IND ,763 ,395 1,930 ,055 LEGAL ,241 ,362 ,665 ,507
However, since the outliers in this study are ‘true’ outliers, the results will lose power and will be biased if the outliers are removed from the dataset. Therefore, another additional analysis is conducted, where the outliers are winsorized. In this case, the outliers are
winsorized to the next largest or smallest quality score which is not expected to be an outlier. The results of this analysis are shown in table 13. Winsorizing the outliers does not influence the results, since the same variables are significant (p= ,026 for assurance provider and p= ,040 for the industry). Winzorization can increase the robustness of statistical inferences. Table 13
Regression analysis with winsorized outliers
Variable Coeff. Std. Err. T Sig..
Constant 13,014 ,884 14,715 ,000 SIZE ,014 ,109 ,125 ,900 LEV ,002 ,022 ,069 ,945 ROA -,434 1,329 -,326 ,744 PROV ,946 ,424 2,233 ,026 IND ,897 ,435 2,064 ,040 LEGAL ,608 ,396 1,534 ,126
4.3.2. Standardization of assurance reports
The results of the one-way ANOVA test does not support the expectation that there are differences in quality over the years. A possible explanation for the insignificant result is standardization of assurance reports in the Netherlands, since most of the firms in a stakeholder-oriented country choose an audit firm as assuror (Simnett et al., 2009). These audit firms often use templates of regulatory bodies such as NBA, although this is not obligated. When using the same templates, quality scores in the Netherlands should be the same in most cases, which will lead to insignificant results in this research. To analyze whether this is true or not, the quality scores of the Netherlands for each year are shown in appendix 2. The descriptive statistics in appendix 2 show that each year score 14 and score 15 together have the highest frequency (40% in 2013, 66,6% in 2014, 58,6% in 2015, 59,1% in 2016, and 53,9% in 2017). Thus, the majority of the quality scores are 14 or 15 each year in the Netherlands, which could indicate that the reports are standardized in the Netherlands.
5. CONCLUSION AND DISCUSSION
In this chapter the research question will be answered. The results of the tested hypotheses are briefly explained in the first paragraph. Secondly, the theoretical and practical
implications are discussed. Thereafter, the limitations of the research are discussed and recommendations for future research are given. Finally, an overall conclusion is given.
The purpose of this study is to explore the variability and differences of the quality of sustainability assurance over the years, and to explore if this quality is influenced by the type of provider of assurance, the industry where the company is operating in and the legal system of the country where the company is domiciled in. Substantiated from the agency theory and the stakeholder theory, reporting on sustainability and receiving high quality (voluntary) assurance on the reported information contributes to the mitigation of the information asymmetry between the company and its stakeholders. However, the absence of legally binding regulations, the voluntary choice to assure the sustainability information and the presence of different assurance providers and different standards raises questions about the quality of the assurance. In addition, different determinants could affect the sustainability assurance quality. In order to explore the differences in the quality of sustainability assurance and if it’s influenced by some determinants, the study used of the following research question: ’’Does the quality of
sustainability assurance differ over time; and to what extent is the quality influenced by the provider of assurance, the industry, and the legal system of the country where the company is located in?’’.
5.1. Findings
To answer the research question, four hypotheses were developed and tested. Two of the four hypotheses are statistically significant. First, it was hypothesized that the quality of
sustainability assurance differ over time. However, findings of the one-way ANOVA test do not support this expectation. The average quality score given to the assurance reports remains almost the same each year. This is not in line with the literature arguing that the absence of legally binding regulations, the voluntary choice of the assurance and the different providers and different standards leads to variation and differences in quality (Perego & Kolk, 2012; Simnett et al., 2009; Hummel et al., 2019).
A possible explanation for not supporting the first hypothesis is the standardization of assurance reports. Especially in the Netherlands, audit firms often use the same standard or templates provided by regulatory bodies. When using the same standard or template for assurance reports, assurance reports will receive the same quality score. In the additional analysis, the standardization is analyzed with descriptive statistics. The statistics show that the majority of the quality scores are 14 or 15 in the Netherlands. However, since there are still other quality scores realized, it is difficult to make conclusions out of these descriptive statistics.
The second hypothesis expected that assurance providers of the audit profession are more likely to give high quality assurance than assurance providers who are not of the audit profession. This hypothesis is statistically significant and can be supported. The finding is in accordance with prior literature arguing that assurors from the audit profession give higher assurance quality, since they have greater skills and training, stringent education, experience benchmarks, well-developed standards, and independence and ethical requirements
(Martinez-Ferrero et al., 2018; Hugging et al., Simnett et al., 2009). In addition, audit firms have a high level of reputation capital (Zhou et al., 2016).
The third hypothesis is statistically significant, meaning that companies operating in polluting industries receive higher sustainability assurance quality compared to companies in less polluting industries. This result correspond to the results of Perego & Kolk (2012), who found that the quality of the assurance is higher for companies operating in more risky industries. Based on the research of Simnett et al. (2009), Chemicals, Energy (Utilities), Financial Services, Mining and, Water Utilites are classified as more polluting industries, since these industries have greater environmental and social risks.
Lastly, this research proposes that companies domiciled in a country with a code law system (stakeholder-oriented) are more likely to have high quality assurance compared with
companies domiciled in a country with a common law system (shareholder-oriented). Based on prior literature, the United Kingdom is classified in this research as a shareholder-oriented country and the Netherlands as a stakeholder-oriented country. However, the results are not statistically significant and therefore cannot be supported.
A possible explanation for the insignificant result is the question whether the United Kingdom can still be seen as shareholder-oriented or not. Although literature states that the United Kingdom is shareholder-oriented, the UK belongs, with 92%, to the top 3 countries with the highest rate of CR information in annual financial reports (KPMG, 2017). This fact raises the question if the UK has not moved from shareholder-oriented to stakeholder-oriented, since the large development of and focus on sustainability/CSR.
5.2. Implications
There are several theoretical and practical implications in this study. First, this study extends prior literature on sustainability assurance and contributes to previous research in providing additional research on sustainability assurance quality. As mentioned earlier, studies and literature about the quality of sustainability assurance are limited and have mainly been focused on other settings. This study provides insights related to the variability and differences in sustainability assurance quality and some determinants that influence the quality.
Secondly, the results can be of practical relevance for companies and regulating bodies. The relatively low quality scores on the assurance reports (on average 14/15 points out of 33) could stimulate regulatory bodies to mandate reporting principles and/or standards, in order to ensure higher quality and to have unambiguous assurance statements. The need for mandatory reporting principles/standards is also appointed in prior literature (Eccles et al. 2012; Kolk & Perego, 2010). Furthermore, this study provides evidence that assurors from the auditing profession are more likely to give higher quality assurance, which could be of interest for reporting companies.
5.3. Limitations and future research
The study has several limitations, which will be discussed in this paragraph. The limitations can serve as recommendations for future research.
The first limitation is the measurement of the sustainability assurance quality through the assurance reports. As noted earlier, auditors often use the same standard or template (such as ISAE 3000).
As a result, recommendations or information on progress, for example, are not stated in the assurance report, because this is not prescribed by the standard/template. However, this does not necessarily mean that this not happened. It is possible that the audit is of high quality, while this is not reflected in the report. Furthermore, the second limitation is that the dependent variable quality is hand collected through a content analysis, which made it inherently subjective and may have impaired the measurement for quality (Van der Laan Smith et al., 2005). Future research could use another measurement of quality, for example the audit fee, since price could be an indication of quality in well-informed markets (Simnett, 2009).
Another limitation in the research may be the classification of the United Kingdom as shareholder-oriented, resulting in the question whether the UK is more stakeholder-oriented because of the large development of sustainability/CSR. It is possible that therefore, the UK and NL are too similar, which has implications for the stated hypothesis. Future research could use the United States as stakeholder oriented, since sustainability reporting is less developed and the demand for it is relatively low in the US (KPMG, 2017). Besides, using a larger sample by adding more countries could provide more insight in the effects of the legal system on sustainability assurance quality.
5.4. Overall conclusion
The research question of the study is: ’’Does the quality of sustainability assurance differ over
time; and to what extent is the quality influenced by the provider of assurance, the industry, and the legal system of the country where the company is located in?’’. This study shows that
the quality of sustainability assurance does not significantly differ over the years. However, with the average quality score of 14/15 out of 33, the quality scores are relatively low. Moreover, this study indicates that the type of assurance provider and the industry are determinants that influences the sustainability assurance quality.
REFERENCES
An, Y., Davey, H. & Eggleton I.R.C. (2011). Towards a comprehensive theoretical framework for voluntary IC disclosure. Journal of Intellectual Capital, 12 (4), 571-585.
Ball, R., Kothari, S.P. & Robin, A. (2000). The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and Economics, 29, 1-51.
Bilal Farooq, M. & de Villiers, C. (2019). The shaping of sustainability assurance through the competition between accounting and non-accounting providers. Accounting, Auditing &
Accountability Journal,32 (1), 307-336.
Birkey, R., Michelon, G., Patten, D.M. (2016). Does assurance on CSR reporting enhance environmental reputation? An examination in the U.S. context. Accounting Forum, 40 (3), 143-152.
Carey, P., Simnett, R., Tanewski, G. (2000). Voluntary demand for internal and external auditing by family business. Auditing: A Journal of Practice and Theory, 19, 37-51.
Cho, C. H., Michelon, G., Patten, D. M., & Roberts, R.W. (2014). CSR report assurance in the USA : an empirical investigation of determinants and effects. Sustainability Accounting
Management and Policy Journal, 5 (2), 130-148.
Chow, C.W. (1982). The demand for external auditing: Size, debt and ownership influences.
The accounting Review, 57 (April), 272-291.
Cohen, J., Holder-Webb, L., Wood, D., & Nath, L. (2011) Retail investors’ perceptions of the decision-usefulness of economic performance, governance and corporate social responsibility disclosures. Behavioral Research in Accounting, 23 (1), 109-129.
Comyns, B. (2016). Determinants of GHG Reporting: An analysis of global oil and gas companies. Journal Of Business Ethics, 136(2), 349-369.
Cuadrado-Ballesteros, B., Martínez-Ferrero, J., & García-Sánchez, I. M. (2017). Mitigating information asymmetry through sustainability assurance: The role of accountants and levels of assurance. International Business Review, 26(6), 1141-1156.
Deegan, C., Cooper, B.J., Shelly, M. (2006a). An investigation of TBL report assurance statements: Australian evidence. Australian Accounting Review, 16 (2), 2-18.
Dhaliwal, D.S., Li, O.Z., Tsank, A., & Yang, Y.G. (2011). Voluntary nonfinancial disclosure and the cost of equity capital: the initiation of corporate social responsibility reporting. The
Eccles, R.G., Krzus, M.P., Rogers, J., & Serafeim, G. (2012) The need for sector-specific materiality and sustainability reporting standards. Journal of Applied Corporate Finance, 24(2), 8-14.
Eisenhardt, K.M. (1989). Agency theory: an assesment and review. The Academy of
Management Review, 14 (1), 57-74.
Fuhrmann, S., Ott, C., Looks, E., & Guenther, T. W. (2017). The contents of assurance statements for sustainability reports and information asymmetry. Accounting and Business
Research, 47(4), 369–400.
Gallego-Alvarez, I. & Quina-Custodio, I. A. (2016). Disclosure of corporate social responsibility information and explanatory factors. Online Information Review, 40 (2), 218-238.
Global Reporting Initiative (2006). GRI 101. Retrieved from:
https://www.globalreporting.org/standards/gri-standards-download-center/?g=56c30739-a139-475e-bc40-03fa444e4ccf
Global Reporting Initiative (2013). The external assurance of sustainability reporting. Retrieved from: https://www.globalreporting.org/resourcelibrary/GRI-Assurance.pdf
Global Reporting Initiative (2018). Sustainability disclosure database: Data legend. Retrieved from:
https://www.globalreporting.org/SiteCollectionDocuments/GRI-Data-Legend-Sustainability-Disclosure-Database-Profiling.pdf
Gurturk, A. & Hahn, R. (2016). An empirical assessment of assurance statements in sustainability reports: smoke screens or enlightening information? Journal of Cleaner
Production, 136, 30-41.
Hahn, R., & Kühnen, M. (2013). Determinants of sustainability reporting: a review of results, trends, theory, and opportunities in an expanding field of research. Journal of Cleaner
Production, 59, 5-21.
Hodge, K., Subramaniam, N., & Stewart, J. (2009). Assurance of sustainability reports: Impact on report users' confidence and perceptions of information credibility. Australian accounting
review, 19(3), 178-194.
Huggings, A., Green, W. J., & Simnett, R. (2011). The competitive market for assurance engagements on greenhouse gas statements: Is there a role for assurers from the accounting profession? Current issues in auditing, 52 (2), A1-A12.
Hummel, K., Schlick, C., Fifka, M. (2019). The role of sustainability performance and accounting assurors in sustainability assurance engagements. Journal of Business Ethics, 153, 733-757.
