The influence of third-party sustainability assurance on content completeness of carbon emission disclosure in sustainability reports

(1)

The influence of third-party sustainability assurance

on content completeness of carbon emission disclosure

in sustainability reports

Master thesis, MSc Accountancy

University of Groningen, Faculty of Economics & Business T.A. Krukkert

S2468042

Adriaan van Ostadestraat 73 9718RS Groningen Supervisor: Dr. D.A. de Waard Co-supervisor: Dhr. M. Looijenga

Word count: 8667 Keywords:

Sustainability assurance, content completeness, carbon emission, sustainability reports, EU NFI Directive, ASAP, NASAP, carbon emission disclosure

(2)

2

“When the Last Tree Is Cut Down, the Last Fish Eaten, and the Last Stream

Poisoned, You Will Realize That You Cannot Eat Money”

(3)

3

Abstract

This paper investigates the influence of third-party sustainability assurance on the content completeness level of carbon emission disclosure. Furthermore, it investigates the influence of the EU NFI Directive and NASAPs on the relationship described above. Based on previous literature, it is hypothesized that third-party sustainability assurance has a positive effect on the content completeness level of carbon emission disclosure. This study focuses on companies in the European Union and on the energy and water utility industry. The results of this study confirm the hypothesis. It is also hypothesized that the EU NFI Directive should have a strengthening effect on this relationship and NASAPs should have a moderative effect on this relationship. The results did not find a significant influence of the EU NFI Directive and the sample of NASAPs was too limited to accept or reject the hypothesis. The results of this study can be used directly by companies because it shows that third-party sustainability assurance increases the content completeness level of carbon emission disclosure and content completeness is one of the most important measures for the quality of sustainability reports (Helfaya et al., 2019). Further research could focus on other countries and other industry types. Furthermore, the influence of NASAPs on the content completeness of carbon emission disclosure should be further examined with a larger sample. The definition of content completeness level, as well as content completeness scorecards regarding other topics, could be further developed to measure the content completeness level of sustainability reports as a whole.

(4)

4

1. Introduction

1.1 Introduction to the research topic

There is an increase in demand in third-party sustainability assurance since the worldwide growth in the publishing of sustainability reports (Deegan, Rankin & Tobin, 2002; Lee & Hutchinson, 2005; Alciatore & Callaway, 2006). Furthermore, Schoolderman, Looijenga & Nandram (2017) state in their article that the public demand for transparency of long term value creation is getting bigger. The increase in demand for transparency of long term value creation by companies has led to the introduction of the EU Directive 2014/95/EU, regarding the disclosure of non-financial information by certain large companies and groups (hereafter: EU NFI Directive). Since reporting year 2017, the new EU NFI Directive is applicable. The EU NFI Directive states the need for improvement in the disclosure of social and environmental information by companies. Companies with 500 or more employees are mandated to disclose certain information regarding social and environmental performance (European Parliament and of the council, 2014). The EU NFI Directive further states that third-party assurance on the sustainability report or the sustainability section within the integrated report could be required by the Member States.

Following a news article (Van Bergen, 2011) there is a discussion about the value of sustainability reports and which role accountants could play in the judgment of these reports. The reason for this discussion was the issues of Shell in Nigeria in 2008 and BP in the Gulf of Mexico in 2010. Wim Bartels from KPMG argued that sustainability assurance is needed because without assurance, a company that publishes a sustainability report could only give statements about positive things and even publish false statements. Sustainability assurance should prevent this and implicates that the company is serious about the issue of sustainability reporting and the sincerity of the report. During the assurance service, an accountant will advise a company in which areas of corporate social responsibility the company can improve as well. Furthermore, assurance serves as a threat that there is a possibility that an accountant will not give an unqualified opinion. Earlier research strengthens these statements and described that third-party sustainability assurance increases the quality of sustainability reports (O’Dwyer & Owen, 2007; Mock, Strohm, & Swartz, 2007; Corporateregister.com, 2008). Furthermore, they argue that sustainability reports will include more negative practices of companies when third-party sustainability assurance has been performed.,

Braam & Peeters (2018) found that the corporate sustainability performance of the company influences the choice of assurance provider and the scope of the assurance. Companies that perform better, chose more often for third-party assurance, as well as a broader scope in assurance.

Despite the increase in demand in third-party sustainability assurance since the worldwide growth in the publishing of sustainability reports, sustainability reporting and assurance is still a new phenomenon (O’Dwyer & Owen, 2007; Hummel, Schlick & Fifka, 2017; Farooq & De Villiers, 2019). The level of maturity is rather low in comparison to the financial reporting, which has developed over the last 100 years to the maturity level of where it is now. Therefore, and because sustainability reporting is largely voluntary, the criteria for sustainability reporting are vaguer. One of the hardest challenges for third-party assurance

(7)

7

providers was the operationalizing of the greenhouse gas (GHG) emissions assurance (Martinov-Bennie, Frost & Soh, 2012).

Because sustainability assurance is largely voluntary and lacks standards and regulations, there is competition between different kinds of sustainability assurors. Farooq & De Villiers (2019) distinguish two types of assurors. Namely, accounting sustainability assurance providers (ASAPs) and non-accounting sustainability assurance providers (NASAPs). ASAPs are professional accountants and the majority of the NASAPs are consultants. ASAPs refer to qualified accountants and NASAPs refer to certification providers, special sustainability consultants and other types of consultants.

Sustainability reports contain more non-financial qualitative information, which is for now too complex to give a high level of assurance on. The lack of standardization and common terminology on this subject is increasing the complexity (Martinov-Bennie et al., 2012). A logical consequence is greenwashing, which means that companies present themselves more sustainable and socially responsible than they really are. Companies rather like to report about what is going well than to report about what is going wrong. For example, they will provide statements about how diverse their personnel portfolio is, but will not provide statements about how many jobs disappeared because of increased automation (Van Bergen, 2011).

Despite these downsides, Bartels argues that it is better to remain to voluntary reporting and disclosure and using guidelines like the Global Reporting Initiative (GRI). Otherwise, it will become a ticking-the-boxes kind of report, while the main goal should be to inform readers about how the company is working towards a sustainable future and what impact its sustainability practices have on the future (Van Bergen, 2011; Schoolderman, Looijenga & Nandram, 2017).

The lack of standardization in assurance statements makes it hard to conclude the actual level of assurance that is obtained. The developments of standards and guidelines reduce the comparability and usefulness of sustainability assurance as well. The exploratory phase of sustainability assurance and the competitive market, that involves NASAPs as well, strengthen this effect (Martinov-Bennie et al., 2012).

What stands out is that assurance provided by accountants is more cautious and offers a higher level of assurance (O’Dwyer and Owen, 2005 and O’Dwyer, 2011). There is a tendency that accountants focus more on quantifying the statements made by the company, while consultants (NASAPs) qualify the statements and try more to change the sustainability practices of a company (Dillard, 2011).

O’Dwyer & Owen (2007) argue that assurors recognize the difficulties in providing a high level of assurance on the completeness of sustainability reports. It has been argued that it is impossible at this moment to guarantee the completeness of sustainability reports, because of a lack of research on this topic. On top of that, Adams (2004) argues that the lack of completeness is one of the most critical issues on sustainability reporting. Last, Martinov-Bennie et al. (2012) state that the operationalizing of carbon emission disclosure in sustainability reports is one of the hardest assurance challenges for assurors.

But do sustainability reports reviewed by third-parties include more GHG content and therefore cover a more complete picture of GHG practices a company is involved in as well? And is there a difference between ASAPs and NASAPs in content completeness levels of carbon emission disclosure in sustainability reports? And at last, does the introduction of the

(8)

8

EU NFI Directive make a difference? In this paper, I will research the influence of performing third-party assurance on the content completeness of the carbon emission disclosure in sustainability reports. This leads to the following main research question:

RQ: Does third-party sustainability assurance improve content completeness of carbon emission disclosure in sustainability reports and is there a noticeable difference in influence between ASAPs and NASAPs and what changed since the introduction of the EU NFI Directive?

1.2 Academic contribution

Limited research has been done regarding the issue of content completeness of carbon emission disclosure in sustainability reports. Rossi & Tarquinio (2017) implicate that assurance improves the quality of sustainability reports. This is mostly because sustainability assurance is a new subject. Prior constructivist literature argues that third-party assurance can increase users’ perceptions and trust of reliability and credibility of voluntary reports (Zadek, Raynard, Forstater & Oelschlaegel 2004; Park & Brorson, 2005; Adams & Frost, 2008; Hodge, Subramaniam & Stewart, 2009; Simnett et al., 2009; Fonseca, 2010; Pflugrath, Roebuck & Simnett, 2011; Gillet, 2012; Casey & Grenier, 2014; Rossi & Tarquinio, 2017). O’Dwyer & Owen (2007) argue that it is difficult for assurors to provide in a high level of assurance on the completeness of sustainability reports. Furthermore, Adams (2004) argues that the lack of completeness is one of the most critical issues on sustainability reporting.

Problems with recent sustainability reports are that companies only provide statements about positive practices while many companies are also involved in non-sustainable practices (Martinov-Bennie et al., 2012). Ironically, the objectives of assurance are the improvement of users’ trust and credibility in these reports (Fonseca, 2010). This emphasizes the need for future research on third-party assurance of sustainability reporting. Variables that can influence the content and quality of sustainability reports and third-party assurance are often pointed out as interesting fields of research (Kolk & Perego, 2010; Cohen & Simnett, 2015; Rossi & Tarquinio, 2017).

As mentioned before, the EU NFI Directive is applicable since 2017. This is the first possibility to measure the influence of the EU NFI Directive on the content of sustainability reports. Therefore, findings regarding this topic are a relevant contribution in this field of research.

This paper might be interesting for readers to gain information about the content level of sustainability reports and third-party sustainability assurance as well. It is useful for practitioners to understand how some variables influence the level of content of sustainability reports and external sustainability assurance as well. The remainder of the paper consists of the literature review & theoretical composition in chapter two, the research methodology in chapter three, the data collection and results in chapter four and will end with the discussion and conclusion in chapter five.

(9)

9

2. Literature review and theoretical composite

This section discusses prior research regarding the topic of sustainability assurance and develops a theoretical framework for the research. The variables discussed in the introduction will be discussed more in depth. The theoretical background will be the basis of the hypotheses, which will be developed in this section.

Earlier literature about sustainability assurance refers to several theories as being relevant for the topic. The legitimacy theory, stakeholder theory, and agency theory are mentioned in many articles regarding sustainability assurance. These three theories will be discussed in the following paragraphs.

2.1 Applicable theories

2.1.1 Legitimacy theory

Earlier research in sustainability reporting often refers to the legitimacy theory, Deegan (2002) and Brown & Fraser (2006) state that organizations have implicit contracts with society. Organizations and their operations are legitimized by fulfilling these contracts. Differences in the extent of disclosed information about the sustainability performance of a company are the result of differences in public policy and institutional pressures (Milne and Patten, 2002; Cormier, Gordon & Magnan, 2004). Suchman (1995) and Palazzo & Scherer (2006) define legitimacy as the assumption that the actions of a company are conform to a socially constructed system of norms, values, beliefs, and expectations. When there is congruence between the social values associated with the activities of the company and the norms of acceptable behavior in society, there is organizational legitimacy. If there is a disparity between these two value systems, there is a threat to organizational legitimacy (Mathews, 1993). The legitimacy theory states that it helps companies to manage threats to the continuance of the social contract by issuing and obtaining sustainability reports and sustainability assurance too (Dowling & Pfeffer, 1975; Patten, 1991, 1992; Gray, Kouhy & Lavers, 1995; Darrell & Schwartz, 1997). They also provide companies with the resources and the support that is perceived to be desirable and appropriate. This is the result of legitimacy(Black & Quach, 2009; Ackers & Eccles, 2015). Furthermore, when the values of a firm are perceived as inconsistent with the values of society, society can revoke the contract. This, for example, means that suppliers will not meet the demand for materials, labor, capital, etc. and customers will not buy the products of the company anymore (Deegan, 2002). So, legitimacy is the behavior of the company that is accepted by society. Companies can only continue when they are legitimate (Suchman, 1995). The legitimacy theory explains self-interested behavior as well. An understanding of the actions of companies is provided by the legitimacy theory. Therefore, the choices of a company regarding sustainability disclosure could be explained using the legitimacy theory (Spence, Husillos & Correa-Ruiz, 2010). External assurance can be used by companies to enhance the confidence of society in the credibility of their sustainability reports and to maintain their legitimacy (Park & Brorson, 2005; Kuruppu & Milne, 2010; Smith, Haniffa & Fairbrass, 2011; Perego & Kolk, 2012; Cohen & Simnett, 2015; De Beelde & Tuybens, 2015; Bepari & Mollik, 2016). Thus, the perceptions of society regarding the legitimacy of the behavior of a company can be influenced. The next example makes that very clear. Sustainability disclosure is increasing. At first, so was greenwashing. Earlier described as the phenomenon where

(10)

10

companies present themselves more sustainable and socially responsible than they really are (Kolk & Perego, 2010). As a response, companies hired third-party assurance providers. This increased the credibility of their sustainability reports (Jones & Solomon, 2010; Farooq & De Villiers, 2019).

The reason why companies decide to assure their sustainability reports is to legitimize their sustainability activities and to enhance their sustainability reports credibility. This will also increase their social legitimacy. To enhance stakeholder legitimacy, companies could provide more details about the procedures that are used by assurance providers and insert key recommendations given by assurance providers to improve their sustainability assurance. In this case, greater clarity about the companies’ sustainability reports and sustainability activities is given (O’Dwyer, 2011). The legitimacy theory is interesting for the content completeness of carbon emission disclosure in sustainability reports because, regarding this theory, assurance statements about GHG information in sustainability reports must contain a certain quantity of information that is useful to satisfy the stakeholders and other users of the report. The amount of useful information improves the accuracy, credibility, and quality of sustainability reports. Consistent with the legitimacy theory, the sustainability assurance statement is important for enhancing the legitimacy, especially for the potential users of the assurance statement (Rossi & Tarquinio, 2017). Sustainability reporting and sustainability assurance are explained by the legitimacy theory as strategic and organizational tools. They influence the perception of society on the legitimacy of companies, reduce stakeholder pressures and respond to stakeholder concerns (Damen, 2016).

2.1.2 Stakeholder theory

Ullmann (1985), Roberts (1992) and Freeman (2010) argue that sustainability reporting is used to ensure the support of stakeholder groups that are essential to firm survival. There are various kinds of stakeholder theories.Jones (1995) identifies the instrumental stakeholder theory and Jones & Wicks (1999) define the convergent stakeholder theory. The instrumental stakeholder theory implies that managers should avoid opportunistic relationships with stakeholders and opportunistic policies and decisions. The convergent stakeholder theory combines the instrumental stakeholder theory with the normative stakeholder theory. The normative theory states that managers must only pay attention to key stakeholder relationships (Freeman, 1999). When the convergent stakeholder theory is effective, a manager should be able to create a moral business environment. Jones & Wicks (1999) argue that this is hard but possible. Adams (2004) states that the divergent goals that companies and their stakeholders have are the underlying problem in the lack of completeness in sustainability reports. Stakeholders must be heard about what they think is important regarding sustainability issues. If companies consult with their stakeholders about what should be in the report, it can be considered that all significant topics would be included. Third-party sustainability assurance could verify this (Gray et al., 1995; Adams, 2004; Guthrie, Cuganesan & Ward, 2008; Beattie, 2014).

Previous literature consists of controversial relationships between sustainability performance and sustainability reporting. Patten (2002) and Cho & Patten (2007) found a negative relationship between sustainability performance and sustainability reporting. What stands out is that weaker-performing firms are more likely to report their positive sustainability efforts. This is consistent with the legitimacy and stakeholder theories. On the other hand,

(11)

AL-11

Tuwaijri, Christensen & Hughes (2004) and Clarkson, Richardson & Vasvari (2008) found a positive relationship between sustainability performance and sustainability reporting with firms being more likely to report good news.

The stakeholder theory and the legitimacy theory have some intertwining perspectives, but the difference between these theories is that the legitimacy theory mainly focuses on society and the stakeholder theory focuses on multiple groups of stakeholders and how the relationships with these stakeholder groups can be managed (Damen, 2016).

2.1.3 Agency theory

The last theory that is important regarding sustainability assurance is the agency theory. The agency theory considers relationships between principals and agents. The principals have a contract with the agent to perform work on their behalf (Zorio et al., 2013). The example which is often referred to is the relationship between shareholders (the principals) and the manager of the company (the agent). The theory assumes that the agents do not always act in the best interest of the principles. This is the result of self-interested behavior. This implies that the interests of the agent differ from the interests of the principles. The agency problem occurs when it is hard or too expensive for principals to monitor the behavior of the agents with the condition that the principals and the agents have conflicting interests. The result is that the principals do not know if the agents acted in the best interest of the principals. (Jensen & Meckling, 1976; Eisenhardt, 1989). Two aspects of unobservable behavior are covered by the agency problem. These are moral hazard and adverse selection, which are respectively hidden action and hidden information (Eisenhardt, 1989). Sustainability assurance is important to mitigate the explained information asymmetry between agents and principals (Power, 1991; Zorio et al., 2013).

2.2 Variables

2.2.1 Performed sustainability assurance as an independent variable

Prior research argues that performing third-party assurance will improve the quality of the sustainability report of the company and will help the company develop its sustainability practices as well (Mock, Strohm, & Swartz, 2007; O’Dwyer & Owen, 2007; Corporateregister.com, 2008). Standards and guidelines in both assurance & reporting can contribute to improving the quality of sustainability reports as well (Adams, 2004).

A third-party assurance can tackle some of the described downsides by providing a statement about all relevant issues in the sustainability report of a company. Thus, relevant issues in which the company may not score well too. The GRI (Global Reporting Initiative) guidelines are the most dominant guidelines in writing and audit the sustainability report. By using the GRI Standards, a sustainability report will be better balanced and cover all relevant aspects of corporate social responsibility (Van Bergen, 2011; Martinov-Bennie et al., 2012; Melloni, Caglio & Perego, 2017).

Assurance providers use the GRI Standards, not as an assurance standard, but to evaluate and verify the compliance of the statements that are made in the sustainability report of the company (Martinov-Bennie et al., 2012). Besides the GRI Standards, dominant standards are being used in third-party sustainability assurance as well. These are the International Standard

(12)

12

on Assurance Engagements (ISAE) 3000 from the International Auditing and Assurance Standards Board (IAASB) and the Institute of Social and Ethical Accountability’s Account Ability 1000 Assurance Standard (AA1000AS) (Martinov-Bennie et al., 2012). Prior literature regarding the use of these standards is inconsistent. Martinov-Bennie et al. (2012) state that the ISAE 3000 and AA1000AS are often used both in a single third-party assurance. The ISAE 3000 focuses on the necessary procedures, while the AA1000AS focuses on the quality of the reporting process. Hummel et al. (2017) and Farooq & De Villiers (2019) argue that the ISAE 3000 standards are more likely to be used by ASAPs since professional accountants are obligated to use the ISAE 3000 in their work by the International Federation of Accountants (IFAC). NASAPs are more likely to use the AA1000AS and sell it as a specialist standard (Hummel et al., 2017; Farooq & De Villiers, 2019). The result of the complementary use of these standards in third-party assurance can increase the quality of the assurance (Martinov-Bennie et al., 2012). Farooq & De Villiers (2019) also argue that ASAPs must take a flexible approach against new assurance standards to develop new types of assurance like sustainability assurance. As of now, ASAPs focus on existing standards, which are considered restrictive within sustainability assurance. Although it is necessary for ASAPs to take a flexible approach regarding the development of new assurance standards, they can base new standards on the ISAE3000 for example.

2.2.2 Content completeness level of carbon emission disclosure in sustainability reports as dependent variable

Prior research regarding the completeness level of sustainability reporting and carbon emission disclosure in sustainability reports is scarce. Adams (2004) states that “completeness is concerned with both the extent of an organization’s operations covered in the report (it’s scope) and the extent to which significant impacts are presented in a report.” Completeness regarding sustainability reporting, where carbon emission reporting is one aspect of, can have different definitions. The definition given by the AA1000AS, for example, is as followed: “The AA1000 Completeness Principle requires that the Assurance Provider evaluate the extent to which the Reporting Organization can identify and understand material aspects (…) of its Sustainability Performance”. Whereas the GRI Standards describe completeness as “the information contained within the report must meet the test of completeness in terms of the reporting boundaries (entities included), the scope (aspects or issues reported) and the time frame (Emphasis added).”

Adams & Evans (2004) mentioned that the lack of completeness of current sustainability reporting is an interesting research topic. This incompleteness follows from the reporting-performance portrayal gap. This gap explains events that companies deliberately omit from their reports. These events are most of the time already covered by the media and have a negative impact on society or the environment. Adams & Evans (2004) note that this level of incompleteness does not occur within financial reporting as well. The essence of the principle of completeness regarding Adams & Evans (2004) is that a sustainability report of a company provides a balanced and comprehensive overview of its performance (Melloni et al., 2017). As described earlier, Martinov-Bennie et al. (2012) made it clear that the operationalization of carbon emission disclosure in sustainability reports is one of the hardest assurance challenges for assurors.

(13)

13

According to the GRI Standards, completeness contains primarily three dimensions: scope, boundary and time. Scope refers to the complete amount of sustainability topics that are covered in a report. Enough topics should be covered in the reports to reflect significant economic, environmental and social impacts of the company and to enable stakeholders to assess the reporting organization’s performance in the reporting period.

Boundary refers to the number of entities whose performance is represented by the report. These entities are, for example, subsidiaries, joint ventures, and subcontractors. So, impacts that are directly related to the activity of a company with a business relationship should be reported on.

At last, time refers to the period specified by the report in which the selected information must be complete. That means that impacts must be reported in the period they occur. It also includes reporting on impacts that do not have a large short-term impact but will have a significant and irreversible impact on the long-term.

Together, these three dimensions cover the material topics and should reflect all significant economic, environmental and social impacts caused by the company, so that stakeholders are able to assess the reported performance of the company.

Completeness is part of the AA1000AS standards too. It is common in sustainability reporting that a report lacks material information about the impact on communities and the environment which is valuable to stakeholders of the companies. Unlike financial reports, this lack of completeness is tolerated, because the AA1000AS are voluntary standards for sustainability reporting (Adams, 2004).

Figure 1: Quality dimension score by Helfaya et al., 2019

The figure above shows the result of the study of Helfaya, Whittington & Alawattage (2019). The authors found that a significant part of the perceived quality comes from the content

(14)

14

published in a sustainability report. As seen in the figure above, the content consists then of the various types of information, measures, themes and the volume of the report. The result is that content completeness is perceived to be the most important factor to contribute to the quality of a sustainability report (Helfaya et al., 2019). This is consistent with the earlier described legitimacy and stakeholder theory, as it becomes clear what the users of the sustainability report find important and what the actual perceived value of the report is.

Because sustainability assurance services will verify the statements that a company makes about their sustainability activities in their sustainability report and helps the company with the development of their sustainability report, the expectation is that performing sustainability assurance on sustainability reports will have a positive influence on the content completeness of sustainability reports. Therefore, assurors must be aware of the risk that sustainability assurance will not be a ticking-the-boxes kind of report. Carbon emission reporting is a part of sustainability reporting and this study will focus on the content completeness in that field in particular. This leads to the following hypothesis regarding the independent variable:

Hypothesis 1: Performing third-party sustainability assurance has a positive influence on content completeness of carbon emission disclosure in sustainability reports.

2.2.3 The influence of NASAPs and the EU NFI Directive as moderator variables

As described earlier, there is a difference between ASAPs and NASAPs. Within the field of sustainability assurance, there is a competition going on between ASAPs and NASAPs. The lack of standardization makes it possible that multiple types of assurors jump into the market (Martinov-Bennie et al., 2012). As described earlier, the difference between ASAPs and NASAPs is that ASAPs are more cautious and offer a higher level of assurance than NASAPs (O’Dwyer and Owen, 2005 and O’Dwyer, 2011). ASAPs try to quantify the statements of the company, whereas NASAPs qualify the statements and advise the company on how to change their sustainability practices (Dillard, 2011). Furthermore, Hummel et al. (2017) argue that the latitude of the assurance is related to the type of assurance provider. ASAPs are positively related to the latitude of the performed assurance. Therefore, the following hypothesis contains a mitigating effect regarding NASAPs:

Hypothesis 2: Non-accounting sustainability assurance providers have a mitigating effect on the influence on content completeness of carbon emission disclosure in sustainability reports.

As stated above, the increased demand in the transparency of the sustainability performance of companies has led to the introduction of the Directive. Because the EU-Directive requires companies with 500 or more employees to disclose information about certain sustainability topics and the EU Member States can oblige certain companies to let their sustainability reports assure by a third-party, the following hypothesis contains a strengthening effect regarding the EU-Directive:

Hypothesis 3: The implementation of the EU-Directive has a strengthening effect on the influence on content completeness of carbon emission disclosure in sustainability reports.

(15)

15

Figure 2 provides a conceptual overview of the hypotheses.

(16)

16

3. Research methodology

This research uses data from the GRI Database and focuses on the EU Member States. Sustainability reports and the sustainability section of integrated reports from the years 2015, 2016 and 2017 are used. Therefore, the difference between the pre-EU NFI Directive and the post-EU NFI Directive can be measured. Since this research focuses on carbon emission disclosure, the data that is used comes from the utility industry. The utility industries are considered the industries that produce the largest amounts of greenhouse gas emissions and is at the same time exposed to community concerns about climate change (Simnett et al., 2009). Therefore, this study will focus on the utility sector. The utility industry consists of companies that offer electricity, gas and water services. The sample consists of 61 companies, which together provide 137 sustainability and integrated reports. Only companies that provide sustainability reports of (one of) the years 2015, 2016 & 2017 will be included.

A quantitative method will be used for the data analysis. All the used parameters in the benchmark, that are both quantitative or qualitative nature, will, therefore, be assessed a score in the level of completeness. That means that if qualitative information is used, it has been quantified for data analysis. The GRI index provides the general necessary information to test the hypotheses. For example: if third-party assurance has been performed, the type of assuror and which assurance standards the assuror used. All the substantive data will be derived from sustainability reports. If a company uses an Integrated Report (IR), only the sustainability section will be used to derive data. All the parameters used can be found in Appendix 1. The carbon emission disclosure index from Chithambo and Tauringana (2017) is used in this research with some minor adjustments.

3.1 Variables

3.1.1 Dependent variable: Content completeness level of carbon emission disclosure in sustainability reports

As described in paragraph 2.2.2, the content completeness level consists of three dimensions regarding the definition of GRI. As described earlier, the carbon emission disclosure index from Chithambo and Tauringana (2017) will be used in a slightly adapted form. All the parameters used can be found in Appendix 1. There is a scoring system for every parameter. Every parameter in the scoring system is measured as included or not included which is quantified to

1 or 0. A maximum of 66 points can be achieved.

3.1.2 Independent variable: Performing third-party sustainability assurance

The influence of performing third-party sustainability assurance will be measured by yes or no, which will be quantified to 1 or 0. Either third-party sustainability assurance has been performed, or third-party sustainability assurance has not been performed. The GRI database provides this information. Because this is a quantified qualitative fact, which takes the value 1 or 0, it is a dummy variable. It indicates the presence or absence of a categorical effect.

3.1.3 Moderator: Sustainability assurance performed by NASAPs

When third-party assurance is given, two types of assurors can be distinguished. As described earlier, these are accounting assurance providers (ASAPs) or non-accounting assurance

(17)

17

providers (NASAPs). This information is also provided by the GRI database. Because it is either ASAP or NASAP, this will be quantified in 1 or 0. Therefore, it is a dummy variable as described in paragraph 3.1.2.

3.1.4 Moderator: Implementation of the EU NFI Directive

The influence of this moderator will be measured by comparing the content completeness scores of carbon emission disclosure of the data of 2015 and/or 2016 sustainability reports (prior EU-Directive) with the content completeness scores of carbon emission disclosure of the data of 2017 sustainability reports (post-EU-Directive). 2015 is included too because companies can use 2016 to already prepare for the EU-Directive implementation. Furthermore, there are companies in the sample that must apply to the EU NFI Directive and companies that do not have to. This variable will, therefore, be quantified to 1, if applicable, and 0, if not applicable. 3.1.5 Control variables

Several control variables will be used to strengthen the results. It will exclude that there are other possibilities that certain relationships might occur. The following control variables will be used: The earnings before (interest and) taxes (EBT) to control for financial performance and the logarithm of total assets to control for firm size. Prior literature gives mixed evidence on the relationship between financial performance and carbon emission disclosure. Liu and Anbumozhi (2009) and Berthelot and Robert (2011) found a positive relationship, while Prado-Lorenzo et al. (2009), Rankin et al. (2011) and Peters and Romi (2012) did not found a significant relationship between financial performance and carbon emission disclosure. Freedman and Jaggi (2005), Brammer and Pavelin (2008), Stanny and Ely (2008), Prado-Lorenzo et al. (2009), Berthelot and Robert (2011), Rankin et al. (2011) and Peters and Romi (2012) controlled for firm size in relation to carbon emission disclosure as well. Henriques and Sadorsky (1999) argue that greater firm size results in increased firm exposure to intense public scrutiny. Firm’s responsiveness regarding GHG issues increases as a result.

(18)

18

3.2 Analysis and regression model

The ordinary least square model (OLS) will be used to conclude the influence and relevance of the independent and control variables on the dependent variable. The OLS model is as follows:

𝐺𝐻𝐺 𝑆𝐶𝑖= 𝛽0 + 𝛽1 3𝑟𝑑𝑃𝑎𝑟𝑡𝑦𝐴𝑠𝑠𝑢𝑟𝑎𝑛𝑐𝑒𝑖+ 𝛽2 𝑁𝐴𝑆𝐴𝑃𝑖+ 𝛽3 𝐸𝑈𝐷𝑖𝑟𝑒𝑐𝑡𝑖𝑣𝑒𝑖+ 𝛽4 𝐸𝐵𝑇𝑟𝑎𝑡𝑖𝑜𝑖

+ 𝛽5 𝐿𝑂𝐺𝑇𝑜𝑡𝑎𝑙𝐴𝑠𝑠𝑒𝑡𝑠𝑖+ 𝜀𝑖

Where:

𝐺𝐻𝐺 𝑆𝐶𝑖 = the carbon emission score based on appendix 1;

𝛽1 3𝑟𝑑𝑃𝑎𝑟𝑡𝑦𝐴𝑠𝑠𝑢𝑟𝑎𝑛𝑐𝑒𝑖 = if third-party assurance has been performed;

𝛽2 𝑁𝐴𝑆𝐴𝑃𝑖 = third-party assurance performed by NASAPs

𝛽3 𝐸𝑈𝐷𝑖𝑟𝑒𝑐𝑡𝑖𝑣𝑒𝑖 = difference in GHG emission score before and after

implementation of the EU NFI Directive

𝛽4 𝐸𝐵𝑇𝑟𝑎𝑡𝑖𝑜𝑖 = firm’s financial performance measured in Earnings Before

(Interest &) Tax ratio

𝛽5 𝐿𝑂𝐺𝑇𝑜𝑡𝑎𝑙𝐴𝑠𝑠𝑒𝑡𝑠𝑖 = firm size measured in a logarithm of total assets

𝛽 1-5 = coefficients

𝜀𝑖 = residual error

(19)

19

4. Data collection and results

4.1 Descriptive statistics

The following tables show different insights into the chosen sample. The most important tables are shown in this chapter, whereas other tables are included in the appendix. Because all the data is hand collected, the sample consists of high-quality data. Because this research is exploratory, the data sample is representative for this study and large enough to give robust results. The sample consists of 137 reports from 61 different companies. These companies are spread over 20 different countries (Table 1, see appendix). What stands out in this table is that the companies are widely spread over the countries. What is interesting as well, is that there are relatively many energy and water utility companies from Italy and the UK that issued a sustainability or integrated report.

Table 2 gives insight in the number of companies that must apply to the EU NFI Directive, which are 41 of the 61 companies. Many energy and water utility companies are state-owned and therefore do not have to apply to the EU NFI Directive. The rules of obligation to the EU NFI Directive can be found in Member State Implementation of Directive 2014/95/EU (2017).

Table 2 EU NFI Directive applicable

Frequency Percentage

Yes 41 67.2

No 20 32.8

Total 61 100.0

51 of the 61 companies in the sample are from the Energy utility sector, whereas 10 of the companies are from the Water utility sector (Table 3, see appendix).

Table 4 (see appendix) gives us insight into the frequency and percentage of assurance given per year on the reports in the sample. A little remarkable is that in 2017 the percentage of third-party assurance performed has dropped to 52,6% against 57,4% in 2016. This might be because sustainability assurance is very complex and therefore cost time to perform the assurance and then to publish the report (Martinov-Bennie et al, 2012). An increasing scope of assurance will increase the complexity of activities to perform the assurance and therefore add time. For example, one report with third-party assurance regarding the financial year 2017 in the sample was published on the 8th of May 2019. Therefore, it could be that some sustainability reports of the financial year 2017 with third-party assurance have yet to be uploaded to the GRI database.

The table below (Table 5) views the frequency and percentage of the type of assuror if applicable. In two reports it was clear that assurance was given, but the type of assuror was not given. Furthermore, in this sample, the frequency of NASAPs is rather small. Just 14,7% of the total sample. The percentage of NASAPs has also decreased throughout the years. In 2015 it was 16,7%, down to 16,1% in 2016 and 10% in 2017. This had a major influence on the regression analysis. That will be further explained in the regression section.

(20)

20

Table 5 Type of assuror

Frequency 2017 Percentage 2017 Frequency 2016 Percentage 2016 Frequency 2015 Percentage 2015 Total frequency Total percentage ASAP 18 90.0 26 83.9 19 79.2 63 84.0 NASAP 2 10.0 5 16.1 4 16.7 11 14,7 n.a. 0 0.0 0 0.0 1 4.1 1 1.3 Total 20 100.0 31 100.0 24 100.0 75 100.0

Table 6 shows a trend analysis (One-way ANOVA) of the total GHG score throughout the years. Interesting trends that occur in the table are found in the mean, the standard deviation, and the minimum and maximum. The mean increased from 18,11 in 2015 to 21,65 and 21,42 in 2016 and 2017 respectively. The standard deviation decreased from 10,287 in 2015 to 9,753 in 2017. In between those years, it rose to 10.806. That might be due to the great range of total GHG scores in 2016. It was the highest range in the trend analysis with a minimum total GHG score of 3 and a maximum total GHG score of 43. In 2015, the minimum total GHG score was the lowest with 1 point. The maximum in 2015 was 37. In 2017, the highest minimum was recorded with 4 points. The maximum total GHG score in 2017 was 39.

Interesting to see was that the highest scores were measured in sustainability reports that were disclosed the latest in the sample. This is in line with the theory described above, which states that the complexity of the subject causes a delay in the disclosure of sustainability reports.

Table 6 Total GHG Score trend analysis

N Mean Std. Deviation Minimum Maximum Total GHG score 2017 38 21,42 9,753 4 39 Total GHG score 2016 54 21,65 10,806 3 43 Total GHG score 2015 45 18,11 10,287 1 37

4.2 Regression analysis

4.2.1 Correlation tables

Regarding the correlation table of 2017 (Table 7), it is not likely that there is any multicollinearity between any of the two variables. The correlation levels between any of the two variables are under .60, which indicates that there is no chance of any multicollinearity between any of the two variables. A correlation between .60 and .80 would indicate a chance of multicollinearity and a correlation level above .80 indicates that it is very likely that multicollinearity will occur between any of the two variables.

(21)

21

Furthermore, the table indicates that it is likely that the influence of third-party sustainability assurance on the content completeness of carbon emission disclosure in sustainability reports is significant. This relation between (1) GHG Score 2017 and (2) Assurance statement is significant on an 0.01 level (two-tailed; p<0.01). There is also a mildly significant correlation between the EU NFI Directive and the GHG Score at a 0.10 level (two-tailed; p<0.10). Table 7 Correlation 2017 Mean Std Deviation (1) (2) (3) (4) (5) (6) (1) GHG Score 2017 21,42 9,753 - (2) Assurance statement ,53 ,506 ,595*** - (3) EU NFI Directive ,67 ,473 ,306* ,151 - (4) NASAP ,0526 ,22629 ,186 ,224 -,394** - (5) EBT ratio ,0365141696 ,0437836995 ,272* ,419*** ,164 ,139 - (6) LOG assets 9,879019640 ,6470761878 ,589*** ,322** ,158 -,052 ,286* - Note: *Coefficient is significant at the 0.10 level (two-tailed; p<0.10); **coefficient is significant at the 0.05 level (two-tailed; p<0.05); ***coefficient is significant at the 0.01 level (two-tailed; p<0.01).

The correlation table of 2016 (table 8, see appendix) shows a low possibility of multicollinearity between (2) Assurance statement and (1) GHG Score 2016 and between (1) GHG Score 2016 and (6) LOG assets. However, a low VIF score of 3,418 and 1,111 respectively show enough evidence that there is no multicollinearity between these variables. Furthermore, there is a highly significant correlation at a 0.01 level (two-tailed; p<0.01) between (2) Assurance statement and (1) GHG Score 2015. There is a significant relation between NASAPs and the assurance statement and between NASAPs and the EU NFI Directive as well. These are both significant at a 0.05 level (two-tailed; p<0.05). The regression analysis will give definite evidence of these relations.

Table 9 (see appendix) gives us enough evidence that there it is unlikely that any multicollinearity will occur between any of the two variables in 2015 since not any of the variables correlate with each other above 0.60. Furthermore, table 9 indicates a significant relationship between the (2) Assurance statement and (1) GHG Score 2015 at a 0.01 level (two-tailed; p<0.01). The same significant relation can be seen between (6) LOG assets and (1) GHG Score 2015. There is also a significant relation at a 0.10 level (two-tailed; p<0.10) between (2) Assurance statement and (4) NASAP and between (5) EBT ratio and (6) LOG assets.

(22)

22

4.2.2 Regression analysis

Table 10 shows the regression analysis results of the influence of third-party sustainability assurance on the content completeness level of carbon emission disclosure in sustainability reports. These results are highly significant in all the years at a 0.01 level (two-tailed; p<0.01). Therefore, hypothesis 1 can be accepted. The control variables EBT ratio and LOG assets control for firm performance and firm size respectively.

Table 10 Regression independent variable

Total GHG Score 2017 Total GHG Score 2016 Total GHG Score 2015 Assurance statement on disclosed information 9,100*** (2,559) 12,293*** (1,803) 9,454*** (2,306) EBT ratio -12,356 (29,226) -21,942 (20,345) -18,698 (16,919) LOG assets 6,828*** (1,896) 8,268*** (1,347) 7,617*** (1,775) Constant -50,367 (18,312) -66,425*** (13,187) -62,408*** (17,517) Observations 38 54 45 R² ,533 ,682 ,523 Adjusted R² ,491 ,662 ,489

Note: *Coefficient is significant at the 0.10 level (two-tailed; p<0.10); **coefficient is significant at the 0.05 level (two-tailed; p<0.05); ***coefficient is significant at the 0.01 level (two-tailed; p<0.01).

Table 11, 12 (see appendix) and 13 (see appendix) show a two-way regression that measures the influence of the implementation of the EU NFI Directive on the relation measured in table 10. In 2017 there is no significant relationship, but it shows a moderating effect of the EU NFI Directive on the relation between third-party sustainability assurance and the content completeness level of carbon emission disclosure in sustainability reports. Table 12 also shows a non-significant moderating effect. Interestingly, table 12 shows a mildly significant positive effect in the direct relation to the content completeness level. This relation is significant at a 0.10 level (two-tailed; p<0.10). Table 13 does show no significant relationship too. So, there are some minor changes over the years, but hypothesis 3 should be rejected because there is no evidence for a strengthening effect of the EU NFI Directive on the relation between third-party sustainability assurance and the content completeness of sustainability reports.

(23)

23

Table 11 Regression moderator EU NFI Directive 2017

Model 1 Model 2 Model 3

Steps & variables B SE B SE B SE

Constant -62,403*** (20,775) -42,888** (18,338) -47,934** (19,256) Control EBT ratio 25,135 (31,466) -16,458 (28,793) -16,631 (28,887) LOG assets 8,392*** (2,129) 6,545*** (1,869) 7,075*** (1,968) Main effects Assurance on disclosed information 4,473*** (1,273) 4,538*** (1,279) EU NFI Directive 1,861 (1,217) 1,636 (1,247) Two-way interaction Assurance x EU NFI Directive -1,146 (1,292) R² ,359 ,563 ,574 Adjusted R² ,322*** ,511*** ,507

Note: *Coefficient is significant at the 0.10 level (two-tailed; p<0.10); **coefficient is significant at the 0.05 level (two-tailed; p<0.05); ***coefficient is significant at the 0.01 level (two-tailed; p<0.01).

Unfortunately, there were not enough by NASAPs assured sustainability reports to perform a regression analysis. The correlation table shows some significant results between NASAPs and overall third-party sustainability assurance. That implicates that there is some sort of relation between the two that must be further examined. Further research can give more information on this relation, but it needs a larger sample to measure this relationship. Therefore hypothesis 2 will not be rejected nor accepted. Further research is necessary to give exclusion to this relation. The tables provide evidence of the strength of the results as well. The correlation tables indicate that it is unlikely that there is any multicollinearity between any of the two variables. This is the first step in strengthening the results. The control variables EBT ratio and LOG

assets control for firm performance and firm size respectively. This strengthens the analysis

results even more. The next chapter will discuss and conclude the research findings and discuss the research limitations and opportunities for further research.

(24)

24

5. Discussion and conclusion

5.1 Discussion, conclusion and implications

5.1.1 Findings and conclusion

Summarizing the findings, the main research question can be answered. Hypothesis 1 can be accepted because the relation of third-party sustainability assurance on the content completeness level of carbon emission disclosure in sustainability reports is significant on a 0.01 level (two-tailed; p<0.01) in all three years. Therefore, the main relationship in this study is highly significant. This means that the content completeness level of carbon emission disclosure in sustainability reports will significantly increase when third-party assurance has been performed. Furthermore, the R² and adjusted R² increase significantly when the independent variable is added to the control variables (Table 10, 11, 12 and 13). These results implicate that third-party assurance is significantly responsible for a higher content completeness level of carbon emission disclosure in sustainability reports, also when this relation is controlled for firm size and firm performance.

Regarding hypothesis 2, where the moderating effect of NASAPs on the main relationship should be tested, it could not be completed. This is because the number of NASAPs in the data sample was insufficient to conduct a regression analysis. The correlation tables showed some significant correlation with third-party assurance, therefore hypothesis 2 should be further examined with a larger sample to conclude the effect on the relation between third-party sustainability assurance and content completeness level of carbon emission disclosure in sustainability reports.

Hypothesis 3 can be rejected, as the regression analysis did not show any significant results. This result is not expected, but the lack of any significant result can be explained by the risk of ticking-the-boxes (Van Bergen, 2011; Schoolderman, Looijenga & Nandram, 2017). The introduction of the EU NFI Directive could lead to assurance that meets only the requirements but does not exceed them.

Therefore, the research question - Does third-party sustainability assurance improve

content completeness of carbon emission disclosure in sustainability reports and is there a noticeable difference in influence between ASAPs and NASAPs and what changed since the introduction of the EU NFI Directive? - can partly be answered. This study provides enough

evidence that third-party sustainability assurance improves the content completeness level of carbon emission disclosure in sustainability reports. The introduction of the EU NFI Directive did not change much at this point and as described above, the difference in ASAPs and NASAPs could not be measured due to an insignificant sample of NASAPs.

The results are in line with the theories discussed in chapter 2. The increase in public demand for long term value creation discussed by Schoolderman, Looijenga & Nandram (2017) is also in line with the legitimacy theory and the analytical results of this study. Some aspects of the legitimacy theory are the same as the stakeholder theory. Third-party assurance verifies the statements that the company makes about its sustainability activities. Therefore, the results are in line with the stakeholder theory. This can also explain the results of the influence of the EU NFI Directive. Stakeholders could demand more from companies than strictly necessary following the EU NFI Directive. Therefore, there is no influence of the EU NFI Directive visible.

(25)

25

Furthermore, the agency theory is in line with the results of the study. The results implicate that assured reports give more information and therefore further mitigates the agency problem as opposed to non-assured reports. Concluding the findings above, third-party sustainability assurance increases the content completeness of carbon emission disclosure in sustainability reports. The new EU NFI Directive does not have a significant effect on this relation. The influence of NASAPs could not be measured due to a lack of NASAPs in the data sample. 5.1.2 Implications

The findings of this study have multiple theoretical and practical implications. The results of this study can directly apply to assurance firms and stakeholders of the companies. Third-party assurance will lead to more content about carbon emission of companies and since this content is assured, the information given will be more reliable too. This will give confidence to shareholders of the companies since they fulfill the increasing demand of shareholders regarding sustainability assurance. Furthermore, it gives confidence to end users of the products that are interested in the sustainability of the activities of companies as well. Since the relation between third-party sustainability assurance and the content completeness level of carbon emission disclosure is proven, companies can take direct advantage of this by hiring an assurance firm to assure their sustainability report. On the other hand, this proven relation gives assurance firms more credibility and therefore assurance firms should be able to increase their sustainability assurance branch.

Besides the practical implications of this study, this research has some academic implications as well. The findings of this research open a new field of research on this topic. Earlier research has been very limited. Rossi & Tarquinio (2017) stated that sustainability assurance improves the quality of the reports. The study of Helfaya et al. (2019) showed that the content of a report is the biggest driver of quality in sustainability reports. Together with the statement of Adams (2004) that one of the most critical issues of sustainability reporting the lack of content completeness is, shows that this research contributes to existing literature.

5.2 Research limitations and further research

5.2.1 Limitations

This study has some limitations. First, only the most polluting industries are included in this research. The results can differ in other industry types.

Furthermore, the limited sample led to some limitations in this study. The influence on NASAPs could not be measured due to unreliability. This is the effect of a very small sample size of NASAPs within the complete sample.

Besides the limited sample size and certain industry type, this study has focused only on EU member states to also measure the influence of new EU law and regulations as well. There were no significant results found regarding this relation. Furthermore, the EU NFI Directive is new, and available data is therefore limited.

There are limitations regarding the measurement of the content completeness of carbon emission disclosure level as well. There is no clear definition or scorecard regarding the content completeness level. With the used scorecard, this study tried to measure if nothing regarding carbon emission disclosure is excluded. However, it is almost impossible to measure what is

(26)

26

excluded. The scorecard used in this research tried to cover all the aspects of carbon emission disclosure in sustainability reports.

Furthermore, the focus of this research is limited. This research focuses on the content completeness level of carbon emission disclosure only. The content completeness level of the remaining part of sustainability reports is not covered within this study.

5.2.2 Further research

Following the exploratory nature of this study, there are multiple subjects regarding this field of research that can be further examined. Further research can test if these results apply to other industries as well. Future research can be focused on other countries outside the European Union as well to confirm if the results of this study are robust worldwide. Together with these research opportunities, further research could test the effect of NASAPs in an industry or country where assurance by NASAPs is relatively more common.

Furthermore, future research can further develop the definition of content completeness level and further develop content completeness scorecards and use them for other topics as well. With such a development, the content completeness level of complete sustainability reports could be measured.

(27)

27

6. References

Ackers, B., & Eccles, N. S. (2015). Mandatory corporate social responsibility assurance practices: The case of King III in South Africa. Accounting, Auditing & Accountability Journal,

28(4), 515-550.

Adams, C. A. (2004). The ethical, social and environmental reporting-performance portrayal gap. Accounting, Auditing & Accountability Journal, 17 (5), 731–757.

Adams, C. A., & Evans, R. (2004). Accountability, completeness, credibility and the audit expectations gap. Journal of corporate citizenship, (14), 97-115.

Adams, C.A. & Frost, G.R., 2008, December. Integrating sustainability reporting into management practices. Accounting Forum (Vol. 32, No. 4, pp. 288-302). Elsevier.

Alciatore, M. L., & Callaway Dee, C. (2006). Environmental disclosures in the oil and gas industry. In Environmental Accounting (pp. 49-75). Emerald Group Publishing Limited.

Al-Tuwaijri, S. A., Christensen, T. E., & Hughes Ii, K. E. (2004). The relations among environmental disclosure, environmental performance, and economic performance: a simultaneous equations approach. Accounting, organizations and society, 29(5-6), 447-471.

Beattie, V. (2014). Accounting narratives and the narrative turn in accounting research: Issues, theory, methodology, methods and a research framework. The British Accounting

Review, 46(2), 111-134.

Bepari, M., & Mollik, A. (2016). Stakeholders’ interest in sustainability assurance process: An examination of assurance statements reported by australian companies. Managerial

Auditing Journal, 31(6/7), 655-687

Berthelot, S., & Robert, A. M. (2011). Climate change disclosures: An examination of Canadian oil and gas firms. Issues in Social and Environmental Accounting, 5(1/2), 106-123.

Black, L., & Quach, A. (2009). Long-term business success through corporate responsibility. White paper by Australian Centre for Corporate Social Responsibility (ACCSR)

and SAI Global.

Black, L., & Quach, A. (2009). Long-term business success through corporate responsibility. White paper by Australian Centre for Corporate Social Responsibility (ACCSR)

and SAI Global.

Braam, G., & Peeters, R. (2018). Corporate sustainability performance and assurance on sustainability reports: Diffusion of accounting practices in the realm of sustainable development. Corporate Social Responsibility and Environmental Management, 25(2), 164-181.

Brammer, S., & Pavelin, S. (2008). Factors influencing the quality of corporate environmental disclosure. Business Strategy and the Environment, 17(2), 120-136.

Brown, J., & Fraser, M. (2006). Approaches and perspectives in social and environmental accounting: an overview of the conceptual landscape. Business Strategy and the

Environment, 15(2), 103-117.

Casey, R.J. & Grenier, J.H., 2014. Understanding and contributing to the enigma of corporate social responsibility (CSR) assurance in the United States. Auditing: A Journal of

Practice & Theory, 34(1), pp.97-130.

Chithambo, L., & Tauringana, V. (2017). Corporate governance and greenhouse gas disclosure: a mixed-methods approach. Corporate Governance: The International Journal of

(28)

28

Cho, C. H., & Patten, D. M. (2007). The role of environmental disclosures as tools of legitimacy: A research note. Accounting, organizations and society, 32(7-8), 639-647.

Clarkson, P. M., Li, Y., Richardson, G. D., & Vasvari, F. P. (2008). Revisiting the relation between environmental performance and environmental disclosure: An empirical analysis. Accounting, organizations and society, 33(4-5), 303-327.

Cohen J.R & Simnett R (2015) “A Forum on Csr and Assurance Services,” Auditing, 34(1).

Cormier, D., Gordon, I. M., & Magnan, M. (2004). Corporate environmental disclosure: contrasting management's perceptions with reality. Journal of Business Ethics, 49(2), 143-165. CorporateRegister.com. (2008). Assure view: The CSR assurance statement report. London. Retrieved from http://www.corporateregister.com. Accessed on 16th of March 2012.

Damen, I. (2016). Quality of assurance on sustainability reports. A study on

determinants of the quality of external sustainability assurance.

Darrell, W., & Schwartz, B. N. (1997). Environmental disclosures and public policy pressure. Journal of accounting and Public Policy, 16(2), 125-154.

De Beelde, I., & Tuybens, S. (2015). Enhancing the credibility of reporting on corporate social responsibility in Europe. Business Strategy and the Environment, 24(3), 190-216.

Deegan, C. (2002). Introduction: the legitimising effect of social and environmental disclosures–a theoretical foundation. Accounting, Auditing & Accountability Journal, 15(3), 282-311.

Deegan, C., Rankin, M., & Tobin, J. (2002). An examination of the corporate social and environmental disclosures of BHP from 1983-1997: A test of legitimacy theory. Accounting,

Auditing & Accountability Journal, 15(3), 312-343.

Dillard, J. (2011). Double loop learning; or, just another service to sell: A comment on “The case of sustainability assurance: Constructing a new assurance service”. Contemporary

Accounting Research, 28(4), 1267-1277.

Directive 2014/95/EU of the European parliament and of the council (2014). Official Journal of the European Union.

Dowling, J., & Pfeffer, J. (1975). Organizational legitimacy: Social values and organizational behavior. Pacific sociological review, 18(1), 122-136.

Eisenhardt, K. M. (1989). Agency theory: An assessment and review. Academy of

management review, 14(1), 57-74.

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. doi:10.1108/AAAJ-10-2016-2756

Fonseca, A., 2010. How credible are mining corporations' sustainability reports? A critical analysis of external assurance under the requirements of the international council on mining and metals. Corporate Social Responsibility and Environmental Management, 17(6), pp.355-370.

Freedman, M., & Jaggi, B. (2005). Global warming, commitment to the Kyoto protocol, and accounting disclosures by the largest global public firms from polluting industries. The

International Journal of Accounting, 40(3), 215-232.

Freeman, R. E. (1999). Divergent stakeholder theory. Academy of management review,

The influence of third-party sustainability assurance on content completeness of carbon emission disclosure in sustainability reports