The Level and Scope of Voluntary Assurance on Sustainability Reports and
its Effect on the Environmental Reputation of a Firm
Bart Verbon, s2335166
University of Groningen MSc Accountancy & Controlling Supervisor: M. Looijenga MSc EMA RA
Co-assessor: prof. dr. D.A. de Waard & dr. R.B.H. Hooghiemstra Word count: 11.352
June 2018
Keywords: Voluntary Assurance, Environmental Reputation, Sustainability Reports, Type of Assurer, shareholder/stakeholder-minded countries, Agency Theory, Stakeholder Theory, Legitimacy Theory
ABSTRACT
With an growing attention to sustainability, where regulations are forming and reporting on non-financial information is getting more mandatory globally, the demand for assuring sustainable information is rising. This research aims to provide new insights on why firms opt to assure a sustainability report voluntary and how firms can benefit from the additional assurance expenses. With explanatory theories as agency theory, stakeholder theory, cost-benefit analysis, and prior literature done in the U.S., the effect of voluntary assurance on the environmental reputation of a European firm is researched. In addition, moderating effects as the type of assurer and the difference in countries (stakeholder-minded versus shareholder-minded countries) is assessed. All the European sustainability reports from the global 500 largest firms ranked in the Newsweek Green Ranking 2017 will be used for a measure of environmental reputation. The results show the categories within the level and the scope of assurance are highly significant and positively associated with the environmental reputation of a firm. Besides this relation, no moderating effects are found on the difference between Big 4 and non-Big 4 auditors and on differences between Continental Europe and the U.K. and Ireland. This research contributes by empirically expanding on literature concerning voluntary assurance and shows that the level and scope of assurance on sustainability does bring a positive effect on the environmental reputation of a firm.
TABLE OF CONTENT ABSTRACT 2 TABLE OF CONTENT 3 I. INTRODUCTION 4 II. THEORY 7 III. METHODOLOGY 11 IV. RESULTS 18
V. DISCUSSION AND CONCLUSION 26
VI. REFERENCES 29
VII. APPENDIX 34
I. INTRODUCTION
This thesis paper examines the effects of voluntary assurance on sustainability reports on the environmental reputation of a firm. Currently, firms can opt for a variety of types of assurance for their non-financial information, and can even choose not to assure the information at all. This paper tries to find an answer on which level and scope of assurance would add value for a firm its sustainability report and will improve the environmental reputation most. Additionally, possible side-effects of the assurance supplier and the country differences will be analyzed. Moreover, the goal of this research is to clear the different outcomes on effects of voluntary assurance found by papers as Cho et al. (2014), who did not found a relation on voluntary assurance and market value, and Coram et al. (2009) who did, and to build on the prior research of Birkey et al. (2016), who found a relation between voluntary assurance and environmental reputation of a firm.
The possible improvement on environmental reputation due to the implementation or expanding voluntary assurance could benefit the firm financially. Hussainey and Salama (2010) found evidence that the environmental reputation of a firm is positively associated with investor’s ability to anticipate on future earnings. In its further research, they proposed looking for determinants on environmental reputation. Gurau (2013) aimed to find these factors for improving a firms’ environmental reputation. It was found that customers needed to be updated on green information and that the information should have objective proof of reliability and credibility.
Next to the environmental reputation benefits, the demand on non-financial information is rising. This increasing attention to non-financial information and sustainability information, in particular, is found extensively in the literature. Gamerslagh et al. (2011) suggest the reason for a company to disclose voluntary information is to reduce political and societal costs, both economic factors, which are based on the political cost theory of Watts and Zimmerman (1978). Plumlee et al. (2015) found evidence on the relation between voluntary environmental disclosure quality and firm value, formulated as expected future cash flow and cost of equity capital. The study of Louren ço et al. (2012) found high corporate sustainability performance is valued by investors, where conversely, profitable firms with a low corporate sustainable performance get penalized by the market. This corporate sustainability performance is measured with data retrieved from the firms’ sustainability reports according to the methodology of the paper.
Disclosures on non-financial sustainability information are published in corporate social responsibility (CSR) reports, corporate responsibility (CR) reports or integrated reports. In business, the growing amount of these kinds of reports are noticeable. KPMG holds an extensive survey on these CR reports from firms. From 2005 to the latest survey by KPMG (2017) on CR reporting, the total amount of CR reports from the G250 increased from 64% to 93%. Moreover, the amount of CR reports of the 250 largest companies (G250) from the Fortune 500 which are voluntary assured, increased from 37% to 67%. The Global Reporting Initiative (GRI), the standard on reporting on CR, recommends firms assure their CR reports,
to help ensure increasing the meaningfulness and accuracy of the disclosed data. Besides this recommendation, the increase of assurance could be explained by the experimental evaluation study done by Coram et al. (2009), where the value of voluntary assurance on positive non-financial performance is significant on stock prices. The credibility of assurance is needed to balance out the natural credibility of a non-assured negative non-financial performance.
Effects of this growing attention seen in literature and businesses also reach governmental entities. The European Union (E.U.) has approved a directive mandating large firms (i.e. public interest entities more than 500 employees) to disclose non-financial information, including environmental matters (2014/95/EU). Firms in the E.U. need to comply with this new requirement by 2018 (CSR Europe, 2017). Directives as these support initiatives such as the United Nations Sustainable Development Goals (i.e. encouraging companies to integrate sustainability information in their reporting). Prior to this directive, countries like France and Sweden implemented regulations to oblige firms to disclose non-financial information (Habek and Wolniak, 2013). Examples are also found outside Europe. South Africa for instance, mandated non-financial disclosure for listed firms on the Johannesburg Stock Exchange (Ackers and Eccles, 2015).
Besides governments are mandating disclosure on non-financial information for businesses, a growing amount of countries implement plans to reduce carbon emissions, in line with the recent Paris climate agreement of 2015 – limit the rise of the global average temperature to a maximum of 2 degrees Celsius by reducing greenhouse gas emissions to reduce the risks of climate change. One of these governmental plans were examined by Allan et al. (2014). Scotland implemented a plan to reduce greenhouse gas emissions with 37% of the 1990 baseline by 2020 by introducing a carbon tax. The momentum on taxing carbon emissions is also examined elsewhere. The plan to tax carbon emissions was tried by France but was eventually vetoed by the Constitutional Council (Zhang et al. 2016).
Due to the increasing attention on disclosing information on sustainability, firms more often opt for assurance on their CR, CSR or integrated reports (KPMG, 2017). The level of assurance offered by an assurance provider can be limited and reasonable as defined in the International Standard on Assurance Engagements (ISAE) 3000, where reasonable assurance is the highest level of assurance given the circumstances of the engagement. Limited assurance provides an acceptable engagement risk, but provides a lower level of certainty. The most common level of assurance is limited, where the publication of the Global Reporting Initiative (GRI, 2013) showed 400 reports are limited assured and 75 reports are reasonably assured by accountants of accountancy firms. Besides the difference in the amount of certainty, Hodge et al. (2009) found in its experimental questionnaire a reasonably assured sustainability report is perceived as more credible and confident than a limited assurance or no assurance at all.
Little research is done on the reasons why organizations choose to assure their CSR or integrated reports. In addition, research on the scope of assurance on CR or CSR and its effect is limited. With the rising attention on non-financial information and its assurance, it is important to research the effects of voluntary assurance on these sustainability reports. As
stated before, Coram et al. (2009) studied this effect of voluntary assurance on sustainability reports. However, Cho et al. (2014) could not accept their hypothesis on the relation between assurance on CSR reports and market value of large firms, because firms with a CSR report without assurance had similar firm value. This inconsistency in results of assurance on sustainable reporting requires further research.
This paper will continue on the findings of the recent paper of Birkey et al. (2016) and will fill the gap by expanding the view of their research by including a country comparison between continental Europe and the U.K together with Ireland. In the paper of Birkey et al. (2016), the environmental reputation of U.S. companies was examined and the type of assurer was included in the research. The contribution of this paper is to extend the view elsewhere in the world and include continental Europe and the U.K., with a split for the respective stakeholder and shareholder mindset of countries, where Continental Europe is stakeholder-oriented and the U.K. together with Ireland is shareholder-oriented. With stakeholder-oriented countries compared with other countries, the stakeholders have more influence on the firm its operations and financial performance (Chen, 2009). Subsequently, Dhaliwal et al. (2012), found a moderating effect of differences in stakeholder-oriented and shareholder-oriented countries on the relation between nonfinancial information disclosure and analyst earnings forecast accuracy. Where stakeholder-oriented countries increase this positive relation.
In the relatively young field on the effect of the level and scope of assurance on sustainability information and the relation between CSR, with its increasing importance of the included information and environmental reputation, the research question of this paper will be:
To what extent does the level and scope of assurance on sustainability information of an organization affect the environmental reputation of a firm?
Birkey et al. (2016) have found evidence that assurance on CSR increases the amount of environmental reputation of a firm, although for a dated time period and with a U.S. perspective. To support this finding, Val Smith, head of Corporate Sustainability at Citigroup, says: ‘The pressure to act from investors and other stakeholders is ever increasing and inaction can have significant business and reputational impacts on a company’ (KPMG, 2017). In the research of Birkey et al. (2016), the used time period was 2009 and 2010. This paper will include more recent data in the sample, which will be elaborated later in the research method section.
To further differentiate the extent of a voluntary assurance on a reasonable or limited level, Ballou and Heitger (2005) highlighted the differences in scope of an audit engagement. They stated the majority of the assurance in corporate sustainability reports are on quantitative performance measures such as greenhouse gas pollution and environmental fines instead of a full scope, where the entire report is assured. Moreover, the in-depth study of CorporateRegister (2008), highlights the broadness of the scope of assurance on CSR reports and divided the scope in full scope and selected information.
The outcome of this research could benefit firms who are thinking about expanding their level and scope of their assurance on their CSR, CR or integrated reports. With new insights
on a relation between increasing the level and scope of assurance and environmental reputation, this decision could be altered to a higher amount of assurance when it helps to improve the reputation of the firm.
After this introduction the paper continues as follows: in the next section the prior research and theories will be discussed as a foundation for the theoretical framework and hypotheses; the third section contains the methodology and the research approach; section IV shows the findings and results of the conducted analysis in this paper; and section V consequently, provides a discussion of the findings, further research, limitations and gives a conclusion of this paper.
II. THEORY
To research the effects on the firms’ environmental reputation by voluntary assurance on sustainability reports, a theoretical structure will be used based on the paper of Birkey et al. (2016). The additional relational variables will be included in this structure, using findings of other related research. Hypotheses will be developed for this research using the combination of explanatory theories. These theories are to the agency theory, stakeholder theory, and the cost-benefit analysis.
Voluntary assurance
Research on effects of voluntary assured sustainability reports is limited (Bagnoli and Watts, 2016). However, some theories will help us explain why firms voluntary assure their sustainability or integrated reports. The descriptive stakeholder theory for instance. Parties, such as governments, consumers or political groups are legitimate stakeholders and can influence firms (Donaldson and Preston, 1995; Freeman, 1984). In this way, the suggested relation of this paper that voluntary assurance influences the environmental reputation of a firm could be explained through the demand of stakeholders as governments, green investors, or consumers who ask for more certainty and transparency. When firms listen to their stakeholders that ask for assurance, the reputation, and in the case of this research, the environmental reputation of a firm could rise.
Another theory which could help understand the relation between voluntary assurance and reputation is the agency theory and its information asymmetry. Chow (1982) states that the link to lower agency costs and the information asymmetry is assurance. Regarding the
reputation of a firm and the agency theory, the more agency problems exist between investors and the firm, the more monitoring needs to be done (Meuleman, 2009). This low
transparency could lead to a lower reputation of the firm since components of reputation rankings consist of transparency and availability principles, it will affect the score of the ranking (Newsweek methodology, 2017).
The research done on voluntary assurance on sustainability reports which divided assurance in level and scope is solely explorative. Jones et al. (2014) assessed the sustainability reports of the ten best food retailers of the U.K. and found differences in scope of the assurance. Where some reports do not provide assurance at all, other reports have assurance on only key elements, areas or data. These findings are aligned with the scope of assurance on the selected information stated by CorporateRegister.com (2008). They state that these differences and individual approaches of the firms make one common methodology effectively not possible for now. Besides, the vast differences in scope of assurance could even counter the positive effect of having a sustainability report assured at all (Gürtürk and Hahn, 2016). To highlight the importance of the level of assurance and if it will lead to similar importance compared to the scope of assurance, Hodge et al. (2009) implied that the level of assurance might have an impact on the perceived confidence of the users of the sustainability reports, but no
significant relation is found. Therefore, it is important to divide the voluntary assurance in the amount of scope and level.
Environmental reputation
The effect of the relation between voluntary assurance on sustainability and environmental reputation could be explained by the following theories.
Legitimacy theory researched by Suchman (1995), explains the relation of assurance and reputation of a firm. The paper noted that if an audience perceives a legitimate firm they find it more worthy. So, as the paper also implies, the theory says it is a perceptive reaction of the observers from what the firm created. In this case, the theory can describe a relation between the choice of a firm to voluntary add assurance on a CSR or integrated report could lead to an assumption that the firms get more valuable in the form of a higher reputation. Additionally, Guthrie and Parker (1989), found evidence that legitimacy theory had an influence on the social disclosures in annual reports.
When a firm or investor is making costs or giving inputs, it wants greater future benefits or outputs of its investments (Dasgupta, 1974; Dr èze and Stern, 1987). The relation between assurance on CSR or integrated reporting and environmental reputation could also be analyzed via this principle. Fumbrun (1995) found benefits in having a high reputation in the new economy. A high reputation leads to more job applicants, lower risks and losses during crises, a higher influence on suppliers and it can justify higher prices of products. So, looking from an economic perspective, these benefits could be greater than the costs of an assurance on the CSR or integrated report. Bagnoli and Watts (2016) used this theoretical idea as well for the decision to choose for voluntary assurance on CSR reports. They state an equilibrium exists between the amount of pay for the voluntary level of assurance and the amount of engagement in socially responsible activities. A firm with a high level of these activities would not want to pay much for a high level of assurance on its CSR report when the firm is in a pooling equilibrium. This because it is not beneficial for the firm when another low engagement firm is providing an equal CSR report. It is difficult to subtract the actual level of the socially responsibility activities between the two types of firms and will, therefore, lower the payoff from having a high level of voluntary assurance.
Additionally, Nekhili et al. (2017) prove that between the relation of CSR disclosure and firm value, more variables can have an effect. The study shows that with family-owned firms, this relation becomes stronger. The paper also concludes that CSR is a strategic investment for the firm and also increases the involvement towards external stakeholders. This research can demonstrate voluntary assurance is another explanatory variable on market value, where reputation is a part of the market value and firm performance (Roberts and Dowling, 2002). With these insights and theories as a structure for the research, hypotheses can be developed. The first two are the main empirical explorative hypotheses and will answer the research question mentioned in the introduction. The last two are moderating hypotheses.
The main focus of this paper is to find a relation between voluntary assurance on sustainability reporting and the environmental reputation of a firm. The research done by Birkey et al. (2016), defined in its hypothesis development the variable voluntary assurance as assured or not assured. To expand the paper of Birkey et al. (2016), voluntary assurance will be split in the first hypothesis, scoping the level of assurance and the scope of voluntary
assurance on sustainability reports in the second hypothesis. Although they found evidence for a positive relation, both hypotheses are stated in a null hypothesis because of the explorative character of this research.
Hypothesis 1: ceteris paribus, the level of voluntary assurance on sustainability reports has no relation with the environmental reputation of a firm.
Hypothesis 2: ceteris paribus, the scope of voluntary assurance on sustainability reports has no relation with the environmental reputation of a firm.
Stakeholder/shareholder countries
Besides these main hypotheses, this research looks at moderating effects on the relation of the first two hypotheses. The first moderating hypothesis:
Hypothesis 3 stakeholder/shareholder-minded countries have no moderating effect on the relation between the level and, or scope of voluntary assurance on sustainability reports and the environmental reputation of a firm.
As Birkey et al. (2016) suggested for further research, seeing if cultural differences in countries have an effect on the relation between voluntary assurance and environmental reputation is interesting. It could make this relation more generalizable worldwide. Michelon (2011) used an international sample to expose differences in stakeholder- and shareholder-minded countries. The U.S. and the U.K. on the one side and continental Europe on the other side. He found in the extent of disclosure of sustainable information differ, where more sustainable disclosure, effects of media exposure and stakeholder commitment is found in continental Europe. Moreover, Simnett et al. (2009) found significant evidence it is more likely that firms in stakeholder-minded countries choose to assure their sustainability reports by an accounting firm than firms established in shareholder-minded countries. These differences in disclosures and assurance preferences could moderate the relation between voluntary assurance and environmental reputation.
Type of assurer
Hypothesis 4 Type of assurer has no moderating effect on the relation between the level and/or scope of voluntary assurance on sustainability reports and the environmental reputation of a firm.
Regarding the type of assurer, firms have many choices getting their sustainable reports assured. Firms can choose from accounting firms, where differences are found in the audit quality between the Big 4 firms and the smaller accounting firms. The voluntary assurance is also given by consulting firms with specialized expertise on assurance on sustainability. The analysis of O’Dwyer and Owen (2005), showed the willingness of the consultancy firms of supplying a high level of assurance on sustainability reports compared to large Big 4 accounting firms, due to the possible boldness of the consultancy firms and the more traditional cautious accounting firms with high professional standards. This suggests a difference in the quality of the audit between accounting firms and consultancy firms.
Research on these quality differences is done by Francis and Yu (2009). They found a higher level of experience and expertise from the Big 4 accounting firms compared with the other accounting firms. Perego (2009) found that, indeed, Big 4 accounting firms deliver a higher assurance quality compared with other assurance suppliers. These findings are interesting for further research and study if assurance providers have a moderating effect on the relation between voluntary assurance and environmental reputation of a firm.
Control variables
To complete the theoretical structure of this research, control variables will be added to test the robustness of the regressive model.
The first control variable is firm size. Bagnoli and Watts (2016), found a significant relation between the size of a firm and voluntary assurance of a CSR report. Also, firm size is positively related with the environmental reputation (Fombrum and Shanley, 1990; Garbett, 1988). Firm size is also used by Lourenço et al. (2014) as a significant control variable for its research on the relation between market value and reputation of a firm.
Melo and Garrido-Morgado (2012) have shown a relation between the type of industry and the environmental reputation. The study of Cho et al. (2012) also found a relation between industries types and environmental reputation, based on the industry types described by the Newsweek Green Ranking. This makes ‘industry’ the second control variable.
The third control variable is disclosure of sustainable information. Eccles et al. (2014) compared low sustainable, traditional firms without sustainable disclosure with high sustainable firms, with sustainability objectives and a high level of transparency in their disclosure. These high sustainable firms outperform the low sustainable firms regarding the stock market, company’s public perception, and brand value and reputation. Additionally, Michelon (2011) shows a positive relation between corporate sustainability disclosure and reputation in two drivers. These drivers are media exposure and commitment of stakeholders. Another study by Cho et al. (2012) found statistical positive relational evidence between environmental disclosure and corporate environmental reputation. This study also used the Newsweek environmental reputation scores to measure this variable. This relation is also supported in the U.K. (Toms, 2002).
A financial measure for the profitability of a firm, the return on assets (ROA) is the fourth control variable in this research. Evidence of the relation between ROA and reputation of a firm is found by Cho et al. (2012). They did a sensitivity test to measure the influence of financial performance indicators with environmental reputation, where the ROA was significantly positively associated with the environmental reputation of a firm. Brown et al. (2010) found with a cross-sectional analysis of the relation between firm performance and corporate reputation (Fortune’s most admired companies, which includes a sustainable category).
III. METHODOLOGY
To elaborate on the way the research is done, in this section the research methods will be discussed. This chapter starts with selecting a sample, then, the research method of the dependent variable will be explained, followed by the independent variables and ending with several control variables to support this research.
Sample selection
The sample will be selected from The Newsweek Green Ranking, which ranked the 500 largest publicly listed firms globally. Measurement on the size of these firms is done by market capitalization. This environmental ranking is one of the most recognized in the world in the media (Neumayr, 2016; Brainard, 2009) and due to the extensive publicly available information on the methodology useable for research, as used in prior papers like the paper of Birkey et al. (2016) and Cho et al. (2012). In this research, all the Continental-European firms for the stakeholder-minded countries and all the U.K. and Irish firms for the shareholder-minded countries will be selected out of the list. In the Newsweek edition, published in 2017 (which contains data from 2015), 120 Continental-European firms were ranked and 34 U.K. and Irish firms were found on the list.
In total 149 firms are selected in this research, a similar quantity compared with the research of Birkey et al. (2016); 165 American firms in 2009 and 186 American firms in 2010. Further testing this sample on validity, the Newsweek Green Ranking list is compared with The Fortune 500 from 2015, another list that ranks firms on size. This list ranks the firms on total revenue per fiscal year. The Fortune 500 counts 131 European firms (Fortune, 2015). Moreover, the distribution is similar within both lists.
From the Continental European selected firms, 77.39% have an assured sustainability report. From the U.K. and Ireland selected firms 64.71% have any kind of assurance on the sustainability report. Table 1 shows more details on the sample.
TABLE 1 Sample
Countries Sample n Assured % Assured
Stakeholder-min ded countries Austria 1 1 100% Belgium 1 1 100% Denmark 1 1 100% Finland 1 1 100% France 32 27 84,38% Germany 30 19 63.33% Italy 7 5 71.43% Luxembourg 1 1 100% Netherlands 13 9 69.23% Norway 1 1 100% Poland 1 0 0% Spain 9 9 100% Sweden 3 3 100% Switzerland 13 11 84.62% Subtotal 115 89 77.39% Shareholder-min ded countries United Kingdom 29 20 68.97% Ireland 5 2 40% Subtotal 34 22 64.71% Total sample 149 111 74.50% Independent variable
The independent variable of the proposed research is the level of assurance. As stated before, there are two types of assurance, namely: limited assurance and reasonable assurance. This variable will be categorized in accordance with the Global Reporting Initiative data legend (GRI, 2018): No assurance, limited assurance, and reasonable assurance. A fourth category exists, where firms have combined the limited and reasonable assurance on different aspects in their sustainability report. Since this research splits scope and level of assurance, the combined category is equal to the highest category. So, these reports count as reasonably assured sustainability reports. These different categories will be retrieved from the GRI database. The database contains sustainability information from over 12,000 firms over multiple years. Although the size of the database, some firms could not be found here. These remaining sustainability reports will be hand collected. To assure the correctness of the
information retrieved from this database, the data will be compared with the data from Newsweek Green Ranking, where one part of the ranking is based on assurance on sustainability reports. Firms with inconsistencies between the two databases will be hand-collected as well.
The second independent variable measuring voluntary assurance is the scope of assurance. Firms choose what to assure in their sustainability or integrated reports. Firms can choose to only give assurance over their key performance indicators (KPIs) or over the entire report. Although papers mentioned the scope of voluntary assurance (Jones et al., 2014; Gürtürk and Hahn, 2016), no measurements were done. Because no scales are found in the literature, categories will be used for this variable. The categorization of this dependent variable will be as follows and as mentioned in the GRI database. No assurance, so no scope. The next step is assurance over the greenhouse gas (GHG). This category is named GHG only. Then, reports assured on their KPI’s and or sections of texts. These are named as specified sections. The last category contains sustainability or integrated reports with a scope of the full report, called entire scope. The sustainability reports will be gathered from the GRI Database. The remaining reports not available on the GRI database or inconsistent with the Newsweek Green Ranking database will be hand collected. The distinction in the scope of the GRI is similar with the definitions used in the paper of Jones et al. (2014), where the category entire report is comparable with the comprehensive scope and specified sections and GHG only with the narrow scope of the assurance engagement.
Dependent variables
The dependent variable is the environmental reputation based on the ranking of Newsweek Green Ranking. The ranking is rule-based and consists of eight parts and two possible penalties. These are: the combined energy productivity score, 15 percent of the total score, combined GHG productivity score, 15 percent of the total score, combined water productivity score, 15 percent of the total score, combined waste productivity score, 15 percent of the total score, green revenue percent range, 20 percent of the total score, sustainability pay link, 10 percent of the total score, sustainability board committee, 5 percent of the total score and audited environmental metric, 5 percent of the total score. The two penalties are monetary fines paid or payable, a reduction of 5 percent and products/services category, also 5 percent reduction of the total score. The data Newsweek uses are obtained from Bloomberg, FactSet, Thomson Reuters and the Carbon Disclosure Project and the overall ranking is made together with Corporate Knights (Newsweek, 2017).
The green ranking of 2017 uses mainly data from the year 2015. The combined productivity scores are using data from 2013-2015, to take into account the changes in emissions productivity over the years. Also, the monetary fines are taken from the years 2013-2015. The remaining parts are based on data from 2015. To compare this variable, all the independent, moderating and control variables contain data from 2015.
In this ranking, the audited environmental metric is taken into account. Due to the dependent variable of this research, the 5 percent of the total ranking score related to ‘audited environmental metric’ will be excluded. In this audit metric, a company receives a 100
percent or 0 percent score if the company is respectively audited on its sustainability report or not. In order to exclude this metric the newly calculated weight system is shown in Table X. It is calculated by dividing the Newsweek percentage by 95%. The penalty scores remain the same.
TABLE 2
Re-calculation of the Newsweek Green Ranking
Newsweek weight system
New calculated weight
system
Combined Energy Productivity Score 15% 15.78947%
Combined GHG Productivity Score 15% 15.78947%
Combined Water Productivity Score 15% 15.78947%
Combined Waste Productivity Score 15% 15.78947%
Green Revenue Percent Range 20% 21.05263%
Sustainability Pay Link 10% 10.5263%
Sustainability Board Committee 5% 5.263158%
Audited Environmental Metric 5% 0%
Total 100% 100%
Monetary fines paid or payable -5% -5%
Products/services category -5% -5%
Moderating variables
MOD_AS_TYPE will be measured with effect coding, due to the low sample size and because only the interaction effect of the moderators is the object of this research. If a sustainability report is assured by one of the Big 4 audit firms, i.e. Deloitte, EY, KPMG or PwC, the score will be a 1. If the sustainability report is assured by another assurer, the score will be -1. The data will be gathered from the GRI database. If the data could not be found in this database, it will be hand collected from sustainability reports.
The MOD_COUNTRY variable is gathered by using the Newsweek database and will be split into two groups. Continental Europe and the U.K and Ireland. Continental Europe has a sample of 115 firms of 14 different European countries, which represents the stakeholder-minded countries. These countries receive a -1. The U.K. and Ireland are represented by 34 firms, which are the shareholder-minded countries in this sample. These countries receive a 1. With effect coding, the two groups will be compared with each other.
Control variables
CTRL_SIZE is measured taking the natural logarithm on the sales of a firm at year end, aligned with the control variable in the paper of Roberts and Dowling (2002), Toms et al. (2005) and Birkey et al. (2016). The natural logarithm is used to normalize the distribution of the data. The data from the year 2015 will be collected from the database Orbis.
Another control variable where the data is originated from Orbis is the return on assets. CTRL_ROA is calculated by dividing the net income by average total assets over the year 2015.
The data for CTRL_INDUSTRY is retrieved from the Newsweek Green Ranking database. The sample will be divided into two groups, in accordance with Cho et al. (2012). The first industry group is categorized as environmentally sensitive. The industries classified as such are mining, oil and gas, paper, chemical, metals and utilities. The classification is also used in the GRI database. In total, this group of environmentally sensitive firms contains 31 firms. This group will be compared with the remaining 118 firms. The environmental sensitive companies will receive a 1 and the non-environmental sensitive companies will receive a 0, making this control variable a dummy variable. Since the industry type can be found in both databases, the inconsistencies between the databases will be checked by the Standard Industrial Classification (SIC) Code List.
The last control variable CTRL_DISC. Using the same methodology as Luo and Tang (2013) and Ben-Amar and McIlkenny (2014) disclosure is measured by the Carbon Disclosure Project (CDP). Although the name suggests the disclosure contains solely information on carbon, the method includes disclosure on governance, strategy, targets, and risks on sustainability as well (CDP scoring methodology, 2017). When studying on effects of disclosure on sustainable information, CDP is useful according to Tang and Luo (2011) due to the more relevant, comparable and comprehensive information when compared with individual sustainability reports. CDP uses 175 environmental, social and governance (ESG) indicators to calculate a score standardized from 0 to 100. CDP is an annual questionnaire, where firms can participate in and provide CDP the requested data. The scores derived from the annual questionnaire can be interpreted in high (71-100), midrange (50-70) and low (less than 50). The low scores lack in disclosing climate-related risks, opportunities and/or carbon emissions. The midrange scores show an increase of understanding concerning climate change risks related to their firm. The high scores include next to the ESG disclosure implementation of the climate change risks by the executive management into the core business (CDP scoring methodology, 2017; ratesustainability.org, 2015). The CDP data is also provided by Databanks and ESG raters such as Bloomberg, RepRisk and Institutional Shareholder Services (Davis Polk, 2017). The data for this variable is gathered by the database of CDP. 138 of the total 149 firms of the sample are available. The 11 remaining firms chose not to participate in the CDP questionnaire of 2015 or did not respond in that year.
The overall measurement of the independent, dependent, moderating and control variables including their data source is displayed in Table 3.
TABLE 3
Variables, Measurement and Data Source
Variables Measurement Data source
ENV_REP 0-100 score. Recalculated as shown in Table X
Newsweek Green
Ranking 2017
LEVEL_AS 3 different categories 0-2. Not
assured, limited assured,
reasonably assured
Sustainability reports / GRI Database
SCOPE_AS 4 different categories. Not assured, GHG only, specified sections, entire report
Sustainability reports / GRI Database
MOD_AS_TYPE Effect coding. Big 4 as 1, non-Big 4 as -1.
Sustainability reports / GRI Database
MOD_COUNTRY Effect coding. 1 for
Continental Europe, -1 for the U.K. and Ireland.
Newsweek Green
Ranking 2017
CTRL_SIZE Natural logarithm of the sales Orbis Database
CTRL_ROA Net income divided by the average total assets
Orbis Database
CTRL_INDUSTRY Dummy variable Newsweek Green
Ranking 2017
CTRL_DISC 0-100 score CDP Database
Since this paper is expanding on the research of Birkey et al. (2016), the model used will be based on the ordinary least squares (OLS) multiple regression analysis to find the estimation of the independent variable environmental reputation. Since this paper extracts the assurance to the level of assurance and scope of assurance, the analysis will be held twice, showed below:
(1) ENV_REP(i) = A1 + B1 LEVEL_AS(i) + B2 MOD_AS_TYPE(i) +
B3 MOD_COUNTRY(i) + B4 CTRL_SIZE(i) + B5 CTRL_ROA(i) + B6 CTRL_INDUSTRY(i) + B7 CTRL_DISC(i)
(2) ENV_REP(i) = A1 + B1 SCOPE_AS(i) + B2 MOD_AS_TYPE(i) +
B3 MOD_COUNTRY(i) + B4 CTRL_SIZE(i) + B5 CTRL_ROA(i) + B6 CTRL_INDUSTRY(i) + B7 CTRL_DISC(i)
During the second analysis, the level of assurance will be replaced by the scope of the audit. Both analyses contain data collected from 2015 and i in the models represents the individual firm.
IV. RESULTS
The first part of the results section is the descriptive statistics of the used sample of this paper. Next, the variables will be tested for multi-collinearity through the Pearson’s correlation coefficient. After this, the ordinary least squares regression analysis on environmental reputation will be presented, on the total sample and with the division on stakeholder-shareholder-minded countries. Then, additional tests will be held, to check the robustness of the regression analysis.
Descriptive Statistics
Table 4.A presents a summary of the descriptive statistics for this paper. These figures are checked for outliers by adding and deducting three times the standard deviation on the mean of the variables. This winsorizing technique did not show outliers that needed to be removed from the sample.
When comparing the sample with the paper of Birkey et al. (2016), the mean of 39.61 is lower (-/- 9.01). This is due to a change in the methodology. The old methodology ranked the number one with a 100 score. The new method keeps room for improvement, even for the greenest companies. Despite the difference in means, the standard deviation of 16.53 is almost identical (+0.21 difference).
Further analyzing these statistics, the mean of the control variable CTRL_DISC corresponds with the mean according to Global Climate Change Report (CDP, 2015), where the annual average lies between 80 and 100. Note that the number of firms is lower at CTRL_DISC at 138 compared to the total sample of 149. All the 11 firms not enclosed in the sample are originated in Continental Europe. The control variables CTRL_SIZE and CTRL_ROA are with its means of 17.42 and 2.39 and their standard deviation respectively are similar to Birkey et al (2016) and Ben-Amar and McIlkenny(2014). CTRL_INDUSTRY shows 21% of the firms in the sample is environmentally sensitive, which is a lower percentage compared with Birkey et al. (2016). Table 4.B shows the group statistics of the MOD_AS_TYPE and the MOD_COUNTRY. The moderating variables are held against the dependent variable ENV_REP. Noted is the low quantity of non-Big 4 assurers (n=18).
TABLE 4.A Descriptive Statistics
Variables n Min Max Mean St. dev.
ENV_REP 149 4.21 73.01 39.61 16.53 LEVEL_AS 149 0 3 0.88 0.70 SCOPE_AS 149 0 3 1.66 1.11 MOD_AS_TYPE 149 0 1 0.63 0.48 MOD_COUNTRY 149 0 1 0.21 0.41 CTRL_SIZE 127 13.66 19.40 17.42 0.81 CTRL_ROA 149 -12.99 24.35 2.39 4.91 CTRL_INDUSTRY 149 0 1 0.21 0.41 CTRL_DISC 138 36 100 95.29 9.55 TABLE 4.B
Descriptive of the Moderating Variables
N Mean St. dev. MOD_AS_TYPE ENV_REP Non-BIG4 18 45.59 16.21 BIG4 94 42.93 13.73 MOD_COUNTRY ENV_REP Continental Europe 115 39.06 17.02
U.K. and Ireland 34 41.49 14.82
Correlations
To identify any correlations between the variables used in this paper, the correlations variables extracted from Pearson’s correlation coefficient are shown in Table 5. The results show one highly significant relation between the scope of assurance. This relation of the entire scope and the specified scope are expected high since the variables are subtracted from assurance. At the same time, the above-mentioned relation stays with .474 below the critical value of 0.8 (Field, 2013). The remaining variables show low correlations and/or no significant relations with each other. Hence, no evidence is found of multicollinearity. Further possible multicollinearities are checked by the regression analysis using the variance inflation factor (VIF) values.
TABLE 5
Pearson’s Correlation Coefficient
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) ENV_REP 1 (2) LEVEL_LIMITED .286*** 1 (3) LEVEL_REASONABL E .134 -.423*** 1 (4) SCOPE_GHG ONLY .036 .234*** -.099 1 (5) SCOPE_SPECIFIED .265*** .421*** .013 -.271*** 1 (6) SCOPE_ENTIRE .070 .180** .223*** -.196** -.474*** 1 (7) CTRL_SIZE .218** .190** -.022 .077 .081 .052 1 (8) CTRL_ROA .253*** .094 -.045 .017 .037 .019 -.182** 1 (9) CTRL_DISC .223*** .166* .101 .097 .114 .037 .209** .026 1 (10) CTRL_INDUSTRY .032 -.028 .091 .158 -.077 .004 .218** -.341*** .153 1
***. Correlation is significant at the 0.01 level (2-tailed) **. Correlation is significant at the 0.05 level (2-tailed) *. Correlation is significant at the 0.1 level (2-tailed)
Regressions Results
In table 6, the ordinary least squares regression results are presented and divided into 5 models. The first model consists of the control variables. Seen here are CTRL_SIZE, CTRL_ROA, and CTRL_DISC having a significant positive related to environmental reputation. The control variable CTRL_INDUSTRY is not significantly related.
The next model shows the regression analysis of the first hypothesis: does the level of voluntary assurance on sustainability reports have a relation with the environmental reputation of a firm. Presented here is the positive relation between limited and reasonable assurance and environmental reputation (resp. .275 and .274 significant at the 0.01 level). Model 3 shows the results of the second hypothesis: does the scope of voluntary assurance have a relation with the environmental reputation of a firm. When the scope of assurance is on specified sections, the relation is significant positive to environmental reputation (.336 significant at the 0.01 level). When the scope of the audit contains the entire sustainability report, the relation is also significant positive (.285 significant at the 0.01 level).
The results of the third hypothesis can be found in model 4 and 5. The independent variables remain significant, but no significant relations are found in the moderators, except when a sustainability report is entirely assured by a Big 4 auditor, a negative significant relation is found. When adding the moderating variables to the models, the multi-collinearity increases. While rising VIF levels, the values remain under the critical value of 10 (Field, 2010) and the variables are therefore acceptable, but further testing is needed. In model 4, the moderator MOD_COUNTRY got a multi-collinearity issue. Therefore, the variable is excluded in the model. Additional tests done on the moderating groups are presented in the additional analysis section.
Noted is the significant relation between the control variables and ENV_REP remain in each model, except for CTRL_INDUSTRY and CTRL_DISC (however, the CTRL_DISC are significant at the 0.11 level). Moreover, the state of these regressions is robust with an acceptable adjusted R², a significant F-value, VIF scores under 10. Additionally, the Kernel density estimation presented a normal distribution of the errors.
TABLE 6
OLS Regression Analysis on Environmental Reputation
Control variables Hypothesis 1 Hypothesis 2 Hypothesis 3 & 4 (1) (2) (3) (4) (5) LEVEL_AS Limited .275*** .315*** Reasonable .274*** .352*** SCOPE_AS GHG only .133 .187 Specified .336*** .339** Entire .285*** .432*** Moderations MOD_COUNTRY .051 BIG4 -.172 -.235 Limited*Country -.201 Limited*Big4 .122 GHG only*Country .147 GHG only*Big4 .053 Specified*Country -.020 Specified*Big4 Entire*Country -.221 Entire*Big4 -.311** Controls CTRL_SIZE .218** .161* .145 .162* .178** CTRL_ROA .327*** .297*** .294*** .289*** .281*** CTRL_DISC .210** .143 .149* .142 .168* CTRL_INDUSTRY .085 .063 .088 .066 .085 Observations 117 117 117 117 117 Highest VIF 1.266 1.474 1.900 6.172 6.550 F-value 6.849*** 6.798*** 5.611*** 4.621*** 3.524*** Adjusted R² .167 .229 .216 .218 .232
***. Correlation is significant at the 0.01 level (2-tailed) **. Correlation is significant at the 0.05 level (2-tailed) *. Correlation is significant at the 0.1 level (2-tailed)
Additional Analysis
Due to the high collinearity caused by the moderating variables, I will compare the means between the groups. Table 7 presents the comparison of the means of the dependent variable
ENV_REP on MOD_AS_TYPE and MOD_COUNTRY. Because the variable
CTRL_INDUSTRY also contains a group comparison, it will be added to this additional test. The tests were executed through a t-test for the mean differences and a One-way ANOVA test for mean differences, which allows comparison with more than two groups with each other. The t-test shows the CTRL_INDUSTRY has got no significant difference in the means between the groups. Likewise, the one-way ANOVA test does not show a significant difference in environmentally sensitive industries and other industries in CTRL_INDUSTRY. So no influence of types of industry on environmental reputation is found. Also, the other moderating variable MOD_COUNTRY do not show significant differences of means in the t-test and the one-way ANOVA test. The moderating variable MOD_AS_TYPE does show a significant difference in the means. Because this test can compare three groups (including no assurer) the test shows a significant difference. The t-test, however, does not show a significant difference between Big4 and non-Big4 assurers.
TABLE 7
Comparing the Means. T-test and One-way ANOVA
T-test Mean difference Standard error Significance
MOD_AS_TYPE 2.663 3.639 .415 MOD_COUNTRY 2.428 3.232 .346 CTRL_INDUSTRY 1.294 3.346 .700 One-way ANOVA F MOD_AS_TYPE 13.833*** MOD_COUNTRY .565 CTRL_INDUSTRY .150
***. Correlation is significant at the 0.01 level (2-tailed)
Next to the ordinary least squares regression and the one-way ANOVA tests on moderating variables, I conducted a one way ANOVA test on the independent variables LEVEL_AS and SCOPE_AS. The results of the test on LEVEL_AS are these: the mean of the ENV_REP on not assured sustainability reports is 28.27. 42.91 is the mean of sustainability reports with a limited assurance. 47.08 is the mean of sustainability reports with a reasonable assurance. This variance of this analysis is significant (F=14.036***). Therefore, together with the
results of the regression, the first hypothesis can be rejected. The test shows for SCOPE_AS the following means: an environmental reputation score of 28.27 for the non-assured
sustainability reports; 41.37 for the sustainability reports with an assured GHG disclosure; 45.00 for the sustainability reports with assurance on specified sections and; 41.59 for the sustainability reports with assurance on the entire report. Again, this analysis is significant (F=9.552***). Although the linear ordinary least squares regression did not show a
significant relation, with this ANOVA test there is evidence to reject the second hypothesis. The results are displayed in table 8. Additional information on the group means is found in the appendix at table 9, showing the combinations of level and scope into an environmental reputation score. Due to a small sample size, the mean score of a reasonably assured report with GHG only as the scope is not available, because it is not present in the sample.
TABLE 8
One-way ANOVA on the Independent Variables
LEVEL_AS F= 14.036*** N Mean Standard deviation Not assured 37 28.266 18.249 Limited assurance 100 42.912 14.343 Reasonable assurance 12 47.082 11.912 SCOPE_AS F=9.552*** N Mean Standard deviation Not assured 37 28.266 18.249 Assurance on GHG only 15 41.372 18.361 Assurance on Specified sections 59 45.003 13.684 Assurance on Entire report 38 41.590 12.942 27
***. Correlation is significant at the 0.01 level (2-tailed)
To test the robustness of the One-way ANOVA test, I conducted a Kruskal-Wallis Test, which calculates mean ranks of ENV_REP on the independent group variables LEVEL_AS and SCOPE_AS to show within the categories significantly different means. The two tests were significant respectively H=19.983*** and H=19.121***. When testing the moderating variables MOD_COUNTRY and MOD_TYPE with the Kruskal-Wallis Test, the results were not significant, in line with the results of the one-way ANOVA test. Also, the control variable CTRL_INDUSTRY did not have a significant difference in means, a similar result with one-way ANOVA.
Together with the results of the ordinary least squares regression, the t-test, and the one-way ANOVA test, the third hypothesis can be supported, that no moderating effects could be assumed between the countries. The evidence for the fourth hypothesis is weak, showing a little significant moderating effect on opting for a reasonable assurance and a Big 4 auditor. From all the comparison testing, the one-way ANOVA test showed a significant difference in means. Therefore, with limitations, the fourth hypothesis can be rejected.
The newly calculated green score to ENV_REP was made to eliminate the expected multi-collinearity between the assurance-based variables as LEVEL_AS, SCOPE_AS, and AS_TYPE. The results of replacing ENV_REP with the original Newsweek score show no significant extra multi-collinearity at the Pearson correlation test and no stronger significant relations with the independent variables during the ordinary least squares regression.
To test the additionally on the control variables, I will use alternative data for the control variable firm size. Literature uses next to the natural logarithm of sales, the natural logarithm of total assets or the number of employees to measure the size of a firm (Bilkey and Tesar, 1977; Calof, 1994; Lourenço, 2014). When replacing the CTRL_SIZE with the natural logarithm of total assets, the relation between LEVEL_AS and SCOPE_AS still remain positive significant. However, the new control variable firm size is not significantly related. Also, when replacing the firm size measured as the natural logarithm of sales with the number of employees, the relation between LEVEL_AS and SCOPE_AS are positive significant related to ENV_REP. Like the natural logarithm of total assets, the number of employees is not significantly related to ENV_REP, together with the control variable CTRL_DISC.
V. DISCUSSION AND CONCLUSION
The reason of this paper was to find an empirically expanded view of voluntary assurance affecting the environmental reputation of a firm. Where recent studies such as (Birkey et al., 2016) stopped at investigating if the type of assurer could be of importance, this paper extends this view by extracting voluntary assurance with two new variables; the level of assurance and scope of assurance on sustainability reports. Where the level of assurance is determined as limited or reasonable assurance and the scope of the audit is grouped in GHG only, specified sections or assurance over the entire sustainability report. Moreover, this paper examines if the type of assurer - i.e. Big 4 or non-Big 4 assurers - and
stakeholder-minded or shareholder-minded country characteristics would influence the relation between voluntary assurance and the environmental reputation of a firm. For
improvement of the statistic model, firm size, return on assets, the disclosure of sustainability information and industry were added, which all have shown a relation on the environmental reputation based on earlier research (Bagnoli and Watts, 2016; Cho et al., 2012; Eccels et al., 2014; Melo and Garrido-Morgado, 2012).
Overall, to what extent does the level and scope of assurance on sustainability information of an organization affect the environmental reputation of a firm? The result section of this paper with its regression model and the additional analyses provided evidence for a strong significant positive relation between the level of assurance on sustainability reports and environmental reputation on a firm, both when a firm chooses for a limited or a reasonable assurance. Therefore, as expected, the first null hypothesis can be rejected.
The second aspect of voluntary assurance on sustainability reports is the scope of the assurance. The results of this explorative paper show a partly significant positive relation between the scope of an assurance and the environmental reputation of a firm. When a firm opts for getting the entire report assured, or at specified sections assurance, like several green KPIs or specific paragraphs about sustainability, the environmental reputation of a firm increases. So, the second null hypothesis can partly be rejected.
As literature suggested, a difference should be present between stakeholder-minded and shareholder-minded countries. However, due to high collinearity statements if the before mentioned countries could have an effect on the relation between voluntary assurance and environmental reputation could not be made. Moreover, no significant difference is seen in the two groups, so, the third null hypothesis could not be rejected nor accepted.
Another expectation was the moderating effect of the type of assurer on the relation between assurance and environmental reputation. Although the low coefficients presented in the Pearson’s, relative high VIF values, no significant differences in the Big 4 and non-Big 4 groups, and no moderating effects shown by the regression analysis, the fourth null hypothesis could not be rejected nor accepted.
With the newly acquired knowledge on the beneficial effect of adding assurance on a sustainability report on a firm’s environmental reputation, firms can now justify the expense for this voluntary decision. Furthermore, this research provides new insights on the broader spectrum of assurance, where differences are made clear on the difference of limited and
reasonable assurance on a sustainability report at the one hand and GHG only until an entire scope of assurance at the other.
The outcomes of the research question implicate the level perspective, as expected, the higher the level of assurance, the better the environmental reputation. To put this insight in a practical perspective, the statistic mean difference between limited assurance and reasonable assurance is 4.2. Although this number looks small, this increase - ceteris paribus - in the
environmental reputation score based and re-calculated on the Newsweek Green Ranking results in an increase of a firms’ rank by 18 places between the European firms.
On the scope perspective, however, showed a different result than expected. The medium-sized scope, i.e. assurance on the specified sections, resulted in a higher environmental reputation score than when the assurance is done on the entire sustainability report, according to the additional one-way ANOVA test. This effect could be explained by Lyon and Maxwell (2011). They described the effect of greenwashing, which is the case when a firm chooses to only disclose the positive environmental information to obtain a green image. Since assurance of non-financial information is not mandatory, firms can choose to opt for assurance solely on the positive environmental information. So, firms with a small scope of assurance on their sustainability reports could, therefore, have a higher environmental reputation score. Taking this in practice, when choosing between a full scope or choosing for assuring GHG emissions alone, it does not improve the environmental reputation. However, opting for assurance on specified sections over the other two categories, the environmental reputation based and re-calculated on the Newsweek Green Ranking of a firm – ceteris paribus – increases with 14 places.
The lack of significant results on the moderating hypothesis could implicate the following. Beginning with the outcomes of the country moderator. The means of the shareholder-minded countries and the stakeholder-minded countries did not significantly differ from one another. This explains no moderating effect on the relation between voluntary assurance and environmental reputation of a firm. The reason that the two groups do not differ, could be explained by the fact that the U.K. and Ireland, the shareholder-minded countries are both member states of the E.U. Although within the E.U. regulatory requirements and legislation differ, Freedman and Jaggi (2010) found that firms in both the U.K. and Continental Europe disclose more GHG information than the firms in the U.S. In addition, their research results showed that the U.K. disclosed even more than Germany. This could explain why no significant difference is found between stakeholder-minded countries and shareholder-minded countries within the sample used in this paper. When comparing the U.K. with the U.S., a report published by PwC shows the progress of countries regarding disclosing sustainable information. In this report, the U.S has made guidelines on how to disclose on climate change issues, whereas the U.K. requires companies to report about non-financial information as GHG emissions since 2013 (PwC Sustainable Reporting Tips, 2015). Taking the before mentioned into account, including the results of this research, the U.K. does not seem to match the shareholder mindset on sustainability issues.
Then, the moderating effects of the type of assurer. Although the literature had found a relation between the type of an assurer and environmental reputation (Francis and Yu, 2009),
in this paper, no significant relation is found. Furthermore, the additional test showed no significant difference between a Big 4 accounting firm and a non-Big 4 firm. This can be explained by findings of Louis (2005). It claimed that ‘non-big 4 auditors have superior knowledge of local markets and better relation with their clients.’ Prior literature also supports findings without significant difference between Big 4 and non-Big 4 auditors on the provided quality of the audit on sustainability reports (Mock et al. 2013). They suggest it is because of the demand of quantity of information assured is more important than the quality of the audit.
The results and the implications of this paper are subject to some limitations. First, the list of the Newsweek Green Ranking contains the world’s largest firms. Therefore, it is not possible to generalize these outcomes on all European publicly listed firms who are disclosing non-financial information. Secondly, the sample size could be larger. Although all the present European firms of the ranking of 2017 were included in this research, the method of the Newsweek Green Ranking remained unmodified since 2014 (which contains data from 2012) and the rating was since then published annually. Increasing the sample could also eliminate the possible effect of changes in years as well. The last limitation contains the measurement of the scope of voluntary assurance on sustainability reports. The method used in this paper is in accordance with the GRI and shows a clear differentiation in the categories. However, within the category specified sections, variations occur. For example, a sustainability report with assurance over one KPI and one paragraph about changes in emissions will be categorized in specified sections as well as a sustainability report with assurance over the entire report except for one KPI and/or one paragraph.
Further research can address some of the limitations. Next to adding more years to increase the sample size and control the research on the impact of differences due to years to this research, a newly developed metric to measure a scope of an assurance could be made. Because no prior research on a metric on the scope of voluntary assurance is done, the metric should be developed completely new. One of the current problems no metric exists yet is the problem to unify the sustainability reports (CorporateRegister.com, 2008). For research purposes, most ideal would be a metric ranging from 0 percent, meaning no assurance, to 100 percent, meaning assurance on the entire sustainability report. A standardized score like this will make this variable good for comparing reports.
An important factor in assurance on sustainability information is the fact assurance is voluntary. In the current transition where the demand of assurance on non-financial information is increasing, a trend is seen that assurance on sustainability information will become mandated by governments (Accountancy Europe, 2017; Ackers and Eccles, 2015). It is interesting for future research to see the effects between the changes of the environmental reputation of a firm by assurance on a sustainability report ex-ante, where assurance on sustainability was voluntary, and ex-post, where the assurance on sustainability is mandated. This paper found a reason to assure a sustainability report. Finding more reasons for having your report assured could be interesting for further research. For example, Bill O’Mara, global head of audit at KPMG International, sees the increasing awareness of investors that non-financial sustainable information is relevant to the financial performance of a firm and
noticed that this causes a greater demand for assurance on these reports (KPMG, 2017). This claim is researched and confirmed for firms in Turkey (Kuzey and Uyar, 2017). Further research could combine the findings of this paper and search for a relation between voluntary assurance and firm performance. For example, the probability that voluntary assurance could influence capital market reactions is one example that could be investigated in the future (Ruhnke and Gabriel, 2013).
REFERENCES
Accountancy Europe, 2017. How to respond to assurance needs on non-financial information. Available at:
https://www.accountancyeurope.eu/wp-content/uploads/171005-Publication-How-to-respond-to-assurance-needs-on-non-financial-information.pdf
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.
Bagnoli, M., Watts, S.G., 2016. Voluntary Assurance of Voluntary CSR Disclosure. Journal
of Economics & Management Strategy. 26(1), 205-230.
Ballou, B., Heitger, D.L., 2005. The Rise of Corporate Sustainability Reporting: A Rapidly-Growing Assurance Opportunity. School of Business, Miami University.
Ben-Amar, W., McIlKenny, P., 2014. Board effectiveness and the voluntary disclosure of climate change information. Business Strategy and the Environment.
Birkey, R.N., Michelon, G., Patten, D.M., Sankara, J., 2016. Does assurance on CSR reporting enhance environmental reputation? An examination in the U.S. context.
Accounting Forum, 40(3), 143-152.
Brainard, C., 2009. Newsweek Ranks 500 Greenest Companies. Available at: https://archives.cjr.org/the_observatory/newsweek_ranks_500_greenest_co.php
Brown, D.L., Guidry, R.P., Patten, D.M., 2012. Sustainability Reporting and Perceptions of Corporate Reputation: an Analysis Using Fortune Most Admired Scores. Advances
in Environmental Accounting and Management, 4, 83-104.
Calof, J.L., 1994. The relationship between firm size and export behavior revisited. Journal
of International Business Studies, 25(2), 367-388.
Corporateregister, 2008. Assure View: The CSR Assurance Statement Report. https://www.corporateregister.com/pdf/AssureView.pdf
Cho, C.H., Guidry, R.P., Hageman, A.M., Patten, D.M., 2012. Do actions speak louder than words? An empirical investigation of corporate environmental reputation. Accounting,
Organizations and Society, 37(1), 14-25.
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.
Chen, C.X., 2009. Who really matters? Revenue implications of stakeholder satisfaction in a health insurance company. The Accounting Review, 84(6), 1781-1804.
Chow, C.W., 1982. The demand for external auditing: Size, debt and ownership influences.
The Accounting Review, 57, 272-291.
