See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/228166840 Assurance on Sustainability Reports: An International Comparison Article in The Accounting Review · October 2007 DOI: 10.2308/accr.2009.84.3.937 CITATIONS 256 READS 448 3 authors, including: Roger Simnett UNSW Sydney 75 PUBLICATIONS 1,513 CITATIONS SEE PROFILE Ann Vanstraelen Maastricht University 46 PUBLICATIONS 1,319 CITATIONS SEE PROFILE All content following this page was uploaded by Roger Simnett on 06 May 2017. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. Assurance on Sustainability Reports: An International Comparison Roger Simnett Professor, School of Accounting, University of New South Wales, Australia [email protected] Ann Vanstraelen Professor, Department of Accounting and Information Management, MARC Maastricht University, The Netherlands and University of Antwerp, Belgium [email protected] Wai Fong Chua Professor, School of Accounting, University of New South Wales, Australia [email protected] Acknowledgements: The authors would like to thank the participants of the 2007 Midyear Auditing Meeting of the American Accounting Association, the 2007 International Symposium on Audit Research and seminars at Bentley College and Northeastern University, along with Martin Carree, Rogier Deumes and Thomas Thijssens for their helpful comments. The authors appreciate the financial support of the Australian Research Council and CPA Australia. The research assistance of Anna Kuo, Anna Huggins, Linda Pellegrino and Per Tronnes is also much appreciated. Assurance on Sustainability Reports: An International Comparison Abstract: Globally, companies increasingly publish separate general purpose, non-financial (sustainability) reports. Some of these are independently assured and assurers may or may not be from the auditing profession. We seek to understand this emerging voluntary assurance market. Using a sample of 2,113 companies (from 31 countries) that produced sustainability reports between 2002-2004, we use sequential logit analysis to identify the factors associated with the decision to voluntarily purchase assurance and the choice of assurance provider. We hypothesize that a company’s need to enhance credibility through assurance and choice of assurance provider will be a function of company, industry and country-related factors. Our results support the argument that companies seeking to enhance the credibility of their reports and build their corporate reputation are more likely to have their sustainability reports assured, although it does not matter whether the assurance provider comes from the auditing profession. We also find that companies operating in stakeholder orientated countries are more likely to choose the auditing profession as an assurer. Keywords: Voluntary assurance, sustainability reports, choice of assurer, assurance services Data availability: All data are publicly available 2 I. INTRODUCTION The last five years may come to be regarded by social historians as the watershed period in the development of a global consciousness about the reality and effects of climate change. The publication of the Stern Review (2007), the work of the UN Intergovernmental Panel on Climate Change (2007), and progress in the development of emissions trading schemes in Europe and elsewhere have all combined to focus community and government attention on the need to pursue a sustainable mode of capitalism. Since the Brundtland Report (1987), the catchcry of “sustainable” business has been associated with many and varied initiatives – from pressure exerted by lobby groups for environmentally-friendly disposal of unused assets, to corporations banning the use of child labor in supply chains and the current emphasis on the reduction of carbon emissions. Within the accounting arena, the sustainability agenda has been linked to early concepts such as social audits and human resource accounting in the 1970s, to intellectual capital, environmental and triple bottom line reporting in the 1990s, and to recent versions of the Global Reporting Initiative (2007). While corporate support for these different accounting/reporting initiatives has varied over time, there has been a consistent concern that traditional financial reports do not adequately represent the multiple dimensions of corporate value today. This has resulted in a search for new financial metrics (see Rappaport 1998; Stewart 1999), and/or additional non-financial measures of value/performance (see Kaplan and Norton 1992; Sveiby 1997). Associated with these developments has been a growing tendency for companies to issue general purpose stand-alone non-financial reports (entitled sustainability reports for this study). Some of these reports are independently assured, and the assurer may or may not be a member of the auditing profession. The primary purpose of this study is to better understand this emerging 3 assurance market and, in particular, the role of assurance in establishing corporate credibility. We also seek to inform the international assurance standard-setting process.1 The sustainability reports analyzed here are the non-financial equivalent to general purpose financial reports, and are intended to meet the information needs of the general public. Given the primary purpose of this study, there are two major categories of non-financial information that are excluded. The first are special purpose reports which are not intended for public consumption, and which are therefore of less public interest because they are used primarily for internal decision making or to meet a reporting requirement of a specific regulatory body. The second category of non-financial information excluded is non-financial information disclosed in annual reports. While companies have been reporting this information in both the annual financial reports to shareholders and in separate voluntary reports (Kolk 2003; Ballou et al. 2006), this study concentrates on the separate (stand-alone) voluntary reports because its main focus is to examine the assurance of this information. The decision to disclose non-financial information in an annual report clouds the assurance decision. For the period of analysis, no instance was identified of a separate assurance report for sustainability information contained in an annual report, which could cause confusion as a result of including two assurance reports in the same document. While there may be certain requirements to disclose non-financial information to certain groups (such as specific regulators), around the world, no regulation can be identified that requires this information to be disclosed in the form of a stand-alone report for the general public. As a result, there is no regulation requiring that the information in this report be assured. Thus, this study examines a true voluntary assurance decision. 1 One of the authors of this paper is the Chair of the Sustainability Expert Advisory Panel which has the brief of informing the International Auditing and Assurance Standards Board (IAASB) and helping develop assurance standards or guidance in the area of assurance for sustainability reporting. 4 A useful way of categorizing the non-financial information contained in these sustainability reports is by reference to the six categories of non-financial indicators contained in the Global Reporting Initiative (GRI) 20072 categorization scheme: economic, environment, labor, human rights, product responsibility and society. The reports examined for this study range from comprehensively reporting on a single dimension outlined above (such as environmental activities/performance) to reporting on all six dimensions. Variation in the content of the sustainability reports is an issue which may impact on the assurance function, and this study will consider its possible impact. Given that the purchase of assurance is a costly decision, it may be assumed that companies purchase such assurance because the benefits outweigh the costs. Benefits could include increased stakeholder or user confidence in the quality of the sustainability information provided and/or increased stakeholder trust in the level of organizational commitment to sustainability agendas. In order to better understand this emergent assurance market, we provide background information on the factors associated with the decision to produce sustainability reports. We do not develop and test hypotheses on these factors as the aims of this paper are to analyse and better understand the decision to assure and the choice of assurance provider. In pursuit of these aims, we formally test the factors associated with the organizational decision to assure, and, for those companies that do assure, the factors associated with the choice of assurance provider. We expect that there are company, industry and country characteristics that will create a greater need to improve the credibility of sustainability information, and therefore influence the decision to assure as well as the choice of assurance provider. The justification for concentrating our formal 2 This is widely regarded as the most dominant set of global reporting regulations for sustainability reporting: “… the number of companies around the world adopting GRI standards and issuing corporate sustainability reports, along with the fact that the GRI works closely with the United Nations, gives its reporting criteria the credibility necessary to be considered generally accepted” (Ballou et al. 2006, 66). 5 analysis on those companies that produce sustainability reports is that it is only these organizations that have to decide whether to assure, and only if they decide to assure do they then have to make a decision as to the type of assurance provider. We argue that assurance is related to a desire to improve the credibility of the disclosed information. In line with the stated aims of the research and extending recent international comparison research (for example, Choi and Wong 2007), we identify characteristics of industries and countries that are expected to result in a greater need for companies to demonstrate that the information they produce is credible.3 They are therefore more likely to have this information assured and, if so, choose a higher quality assurance provider. For evaluating the role of assurance and choice of assurance provider in the international market, this voluntary setting is therefore more suitable than a regulated financial statement setting. In brief, our results show the following. For the period 2002-2004, based on 40,993 companies for which we had the required financial data, we found that 2,113 produced sustainability reports for which we were able to gather other required information. Japan (527 companies), the U.K. (385 companies) and the U.S. (339 companies) are the three most heavily represented countries. In total, 655 of these 2,113 (31%) public reports are assured, and members of the auditing profession assure 275 (42%) of these. Our multivariate results demonstrate a strong link between companies with a higher need to enhance credibility and those having their sustainability reports assured. With regards to the choice of assurance provider, we find that companies domiciled in more stakeholder orientated countries are more likely to choose a member of the auditing profession. The factors associated with a greater need to add credibility 3 The company variables included represent dimensions that have been commonly found to be related to voluntary disclosure and assurance decisions; size, leverage and profitability (for example, Chow 1982, Hibbit 2003, Cormier et al. 2005). These are therefore included as control variables in the later analyses. 6 appear to be more closely related to the decision to have the information assured than they are with the choice of assurance provider. The remainder of this paper is organised as follows. In the next section, we provide a literature review on the demand for voluntary assurance. Subsequently, we develop our hypotheses, followed by a description of our data and empirical models. Finally, we present the results and conclude. II. LITERATURE REVIEW ON VOLUNTARY ASSURANCE The companies in this study have voluntarily produced a sustainability report, and have complete discretion as to whether this information is assured. Companies voluntarily produce such reports for a variety of reasons, including informing stakeholders, and thereby reducing the information asymmetry between the company and the market/public. Reporting also attests to organizational commitment, risk management and a desire to build corporate reputation. In all these cases, the effectiveness of achieving these desired outcomes hinges on the perceived and actual credibility of the information provided. Assurance plays an important role in this. This section examines the findings in the literature on the market demand for assurance, the value obtained by employing assurance, and the incremental value associated with ‘high quality’ assurance. Demand for, and value of, voluntary assurance Only a small number of empirical studies have examined the demand drivers for the voluntary adoption of assurance. Largely this is because the assurance of financial reports has been mandated by law for the better part of the twentieth century in most developed market 7 economies, and research has focussed on issues associated with this context. Chow (1982) was one of the first major studies to examine voluntary assurance. He investigated this issue from the agency theory perspective, arguing that agency costs are associated with the voluntary adoption of financial statement audits. His study focussed on the year 1926, that is, prior to the introduction in the U.S. of a legal mandate for assurance on historical financial information. He documents that proxies for agency costs associated with both shareholders and creditors, such as leverage, the extent to which accounting numbers are used in debt covenants, and company size, are positively associated with the voluntary adoption of financial statement audits. Abdel-khalik (1993) adopts a different perspective and views the demand for assurance as an effective within-company control mechanism to compensate for the loss of control induced by organizational design and the resultant loss of observability of subordinate behaviour. He argues that this loss of internal control may potentially give rise to moral hazard problems and an increasing likelihood of distorted communication. Consistent with the ‘organizational design’ hypothesis, his results indicate that larger companies are more likely to voluntarily demand assurance. The association between assurance and control is approached from a related but different angle by Blackwell et al. (1998), who argue that the demand for assurance stems from the need to mitigate information asymmetry with institutional creditors. Blackwell et al. find assurance to be perceived by institutional creditors as an effective means of control. Similarly, Carey et al. (2000) examine family businesses in Australia, and find that the voluntary demand for assurance is associated with information asymmetry and loss of control, such that the proportion of non family managers and non-family directors is positively associated with the demand for external 8 assurance. Consistent with Chow (1982), Carey et al. (2000) find that the demand for assurance is also associated with higher levels of debt. The above discussion indicates that assurance serves as a useful control mechanism to enhance the credibility of disclosed information and facilitate greater user confidence. Hence, it should result in more appropriate resource allocation decisions by information users. Demand for “high quality” assurance In most studies to date, differences in the quality of assurance services provided have been studied by examining differences between the use of Big N firms and non-Big N firms. It is argued that the Big N audit firms possess scale economies and greater capacity to invest in new technologies. They also have a greater investment in maintaining their reputational capital. Taken together these factors suggest that the Big N firms are less likely to behave opportunistically or myopically. As a result, they are better able to serve as an effective monitoring mechanism than smaller auditors (DeAngelo 1981; Watts and Zimmerman 1986). Moreover, due to their size, the Big N audit firms are less prone to fall victim to fee dependence, as the costs of compromising independence (litigation and reputation costs) outweigh the benefits (Craswell et al. 2002). Examining the quality issue by comparing differences between the Big N and non-Big N audit firms is sensible within a context in which the law provides monopolistic rights to these firms to conduct the audit of financial statements. Such a state-guaranteed monopoly does not exist in the area of assurance on sustainability reports, and assurance in this unregulated market may also be purchased from other providers such as environmental management firms. It is into markets such as this that visionaries such as Robert Elliott saw the accounting profession expanding its skills as information and assurance (attest) professionals (Elliott and Jacobson 9 2002). We therefore distinguish between assurance providers who are members of the auditing profession and other assurance providers. Similar to the arguments for the Big N versus non-Big N outlined above, we classify members of the auditing profession as the higher quality assurance providers. The auditing profession has well-developed ‘global’ standards, a body of ethics and independence requirements, as well as quality control mechanisms at both the firm and engagement levels that help ensure that the assurance provided is of a consistently high quality. In areas requiring specific subject matter expertise, which is commonly the case with assurance engagements on sustainability reports, there are specific standards in place to ensure that the engagement is not accepted if the engagement team does not have the required expertise, and procedures to be followed when using the work of experts. It is also a profession with an established history and reputational capital that is well known to communities. These factors help increase public confidence in the competency and legitimacy of the auditing profession as high quality assurance providers. However, there is usually an increased cost associated with having assurance provided by members of the profession as opposed to other providers4, and as such the client chooses which assurance provider to use on a cost versus benefit basis. The provision and assurance of sustainability reports Very little is currently known about assurance on sustainability reports, with the exception of some descriptive research that has been undertaken. The most comprehensive of these is KPMG (1999, 2002, 2005), which examines whether the Top 100 companies in a number of different countries produce publicly available sustainability reports, and whether these are assured. A comparison of these three KPMG surveys suggests that the frequency of these types of 4 While this is difficult to show systematically as the fees for the engagements are not disclosed, discussions with assurance providers in this area in at least three countries suggest the fees charged by audit firms can be commonly up to five times the fees quoted by environmental consultants for the same engagements. 10 reports is increasing, at least among the larger companies, and the reports are more commonly assured. Interestingly, there is noticeable between-country variation in both the provision and assurance of sustainability reports. They also indicate that the U.S. was one of a small handful of countries where the number of Top 100 companies issuing sustainability reports decreased between 2002-2005 (32% in 2005, down from 36% in 2002). In Canada, by contrast, there was a significant increase in the number of Top 100 companies producing such reports (41% in 2005 compared with 19% in 2002). The KPMG report (2005) notes that the U.S. and Canada both have an exceptionally low proportion of their sustainability reports assured: U.S.–3% and Canada– 10%, relative to U.K.–53%, Australia–43%, continental Europe–41% and Japan–31%. This low rate in the U.S. may be partly a result of the U.S. attestation standards restricting the auditing profession from providing assurance on these reports (mainly because of concerns about lack of suitable criteria), although this explanation doesn’t account for the inability of other types of assurance providers to fill the void in the U.S. Given that other assurance providers exist, it remains unclear why assurance rates are lower in North America. The presence of cross-country variation is further supported by a study sponsored by CPA Australia (2004), which identified and categorised 161 assurance reports on general purpose sustainability reports. That study shows a marked variation over geographic regions with respect to who provides assurance on these reports. In the four major regions classified in the CPA Australia study, audit firms provided assurance on 87% of reports in Japan, 60% in continental Europe, 23% in the U.K., and 15% in Australia. This brief review has highlighted that there is relatively little research on the voluntary purchase of assurance for sustainability reports. Nevertheless, this is an interesting subject matter on which to explore the reasons for the observed patterns in this global assurance market, both 11 with regards to the reasons associated with having this subject matter assured and the choice of assurance provider. III. DEVELOPMENT OF HYPOTHESES This study explores why companies producing sustainability reports have this information voluntarily assured and their choice of assurance provider. As independent assurance is a costly process, we expect that the companies that have these reports assured will be those for which the net benefits are greater.5 As the literature review suggests, assurance confers several benefits. It helps to reduce agency costs6 and confers greater user confidence in the accuracy and validity of the information provided. We focus particularly on country-specific and industry factors that influence the benefits companies gain from purchasing assurance. In the international accounting and finance literature, there is growing evidence of the importance of the national legal environment in explaining financial market development, corporate ownership structures, corporate policies and the properties of accounting information in different countries/jurisdictions (Choi and Wong 2007). In this respect, Francis et al. (2006) provide evidence that governance structures are to some extent endogenous to broader country-level institutions and that company-specific incentives do not entirely explain variation in governance structures. More specifically, we examine whether the organizational benefits resulting from assurance are a function of the legal environment of the country of domicile and the industry to which the company belongs. Why these two variables? Recent literature has shown that 5 As the costs of these assurance services are not disclosed, the analysis concentrates on the characteristics of companies that are expected to gain greater benefit from such services. 6 Studies (see Chow 1982; Abdel-khalik 1993; Blackwell et al. 1998; Carey et al. 2000) find that higher agency costs reflected by size and leverage are positively associated with the voluntary demand for audit. Hence, we include these factors as control variables in the models that we estimate. 12 companies “adapt to poor legal environments to establish efficient governance practices” (Durnev and Kim 2005) and that “Big Five auditors fulfil a strong governance function in weak legal environments” (Choi and Wong 2007). Extending this argument from the financial audit research, we expect that companies operating in a weaker legal environment will be more likely to purchase assurance to increase user confidence in the credibility of the information contained in their sustainability reports. The demand for assurance is expected to be lower in countries with stronger legal environments because there are more country-level protection mechanisms in place in these environments (Choi and Wong 2007). This is supported by other studies which have found that firm level governance characteristics are more pronounced in weak legal environments (for example, Klapper and Love 2004; Lang et al. 2004).7 Further, prior research has shown that there is an industry association between the level of environmental and social risks experienced by companies and the level of environmental and social disclosure (e.g. Adams et al. 1998; Patten 2002). Hence, we expect that companies belonging to industries having a greater environmental or social impact are more exposed to environmental or social risks and will have a greater need to manage these risks by purchasing assurance to increase user confidence in the credibility of the information contained in the sustainability reports they produce. In examining choice of assurance provider we draw a distinction between companies that choose assurance from outside the auditing profession (e.g. environmental consultants), and companies that choose assurance from the auditing profession. We classify members of the auditing profession as the high quality assurance providers (see DeAngelo 1981; Watts and 7 It is recognized that there is the alternative view, that in weak legal environments where the processes in place do not support and discipline auditors, they may be seen to compromise their audit quality, which may decrease the perceived benefits and therefore the demand for this service (Choi and Wong 2007). Consistent with this argument, Francis and Wang (2008) find that Big 4 auditors provide a lower level of audit quality (measured using proxies for earnings quality) in countries with a weaker institutional environment and a lower risk of litigation. 13 Zimmerman 1986). This classification is supported by the fact that the auditing profession has in place well-developed ‘global’ standards, independence and ethical requirements and quality control mechanisms to help ensure the quality of any assurance reports that are issued by their members. The argument is further supported by the fact that firms (especially the major firms) within the profession also bring a high level of reputation capital to their engagements. A counter argument is that specialist providers that are not members of the auditing profession (such as environmental consultancies) may possess a higher level of subject-matter expertise. However, such specialised expertise can always be (and nowadays commonly are) bought or employed by audit firms. Further, price could be an indication of quality in well-informed markets and, as outlined earlier, members of the auditing profession are believed to charge higher prices for the assurance of sustainability reports than assurance providers who are not members of the auditing profession. Given the above arguments and the absence of contradictory data, we hypothesize that members of the auditing profession are the higher quality provider. Formally, our hypotheses are as follows: H1a: Companies with a greater need to increase user confidence in the credibility of sustainability reports will be more likely to have their sustainability reports assured. H1b: Companies with a greater need to increase user confidence in the credibility of their sustainability reports will be more likely to choose assurance from the auditing profession. Besides the need to increase user confidence we argue that the business culture of a country, and in particular whether a country is more stakeholder or shareholder orientated, can influence the demand for assurance on sustainability reports and the choice of assurance provider. A stakeholder orientated or communitarian culture is one in which a broad spectrum of stakeholders are seen by society as possessing a legitimate interest in corporate activities. 14 Stakeholder groups in these countries will therefore have considerable influence upon the activities of companies. By contrast, a shareholder orientated or contractarian business culture is one in which companies are primarily seen as instruments for the creation of shareholder value; other stakeholder groups have less legitimacy and therefore less influence on corporate activities (see Bradley et al. 1999). Given this distinction, we agree with the view that “management in communitarian orientated societies would be more likely to perform and disclose social responsibility activities as part of strategically managing stakeholder relationships” (Smith et al. 2005). In line with this argument, Holder-Webb et al. (2008) suggest that U.S. companies lag behind international companies with respect to social responsibility reporting as they operate in a cultural environment that has a greater shareholder orientation. By extension, we expect that the demand for assurance and the choice of the auditing profession as assurance provider is higher in stakeholder orientated countries compared to shareholder orientated countries. This is reflected in the following two hypotheses: H2a: Companies domiciled in countries that are more stakeholder orientated are more likely to demand assurance on sustainability reports compared with companies domiciled in countries that are more shareholder orientated. H2b: Companies domiciled in countries that are more stakeholder orientated are more likely to choose assurance from the auditing profession compared with companies domiciled in countries that are more shareholder orientated. IV. DATA Our research attempted to identify as many sustainability reports as possible that were published covering the period 2002-2004. The major part of the data collection was over 2005 and 2006, 15 and 2004 was the latest complete year of observations at that time. The major source of these reports was Corporate Register (www.corporateregister.com) which is a comprehensive directory of published corporate environmental and social reports. This source was supplemented by the Global Reporting Initiative database (www.globalreporting.org)8 and the companies on the Dow Jones sustainability index (www.sustainability-index.com), other databases and general searches. In total, we identified 2,662 general sustainability reports issued over the period 2002-2004. In order to obtain the required financial information we used the Global Compustat database which contained 64,256 observations (of which 40,287 had the complete financial information) for the 31 countries identified as having companies producing sustainability reports over the three years. We then attempted to match each of the 2,662 general sustainability reports with a complete observation from Global Compustat, and were able to gain the necessary additional information for 1,407 of the 2,662 observations. For those observations with missing data we then searched the Datastream database and company websites, and were able to obtain the required information for a further 706 company observations, resulting in 40,993 complete company observations for the purposes of providing background information on the reporting decision, and a final sample of 2,113 observations for 867 listed companies over the three years for which we had all the necessary data for testing the hypotheses. Of the 2,113 sustainability reports, 655 (31.00%) contained independent assurance reports. We identified that the assurance provider was a member of the auditing profession in 275 (41.98%) cases. V. RESEARCH MODEL 8 At the time of initial search the GRI maintained their own database. The maintenance of this was later taken over by Corporate Register. 16 Our hypotheses are tested using a sequential logit model. This model serves the sequential notion of the decisions that are taken as this model is defined as a sequence of independent binary logit models. The sequential logit model estimates the effect of explanatory variables on the probabilities of passing a set of transitions.9 First, for those companies producing a sustainability report, deciding whether or not to have this information assured. Second, for those companies deciding to have this information assured, whether or not to choose an assurance provider from the profession. In addition to these two decisions, there is the prior decision to produce a sustainability report. For completeness sake, these three decisions and the number of observations under each branch are illustrated in the decision tree presented in Figure 1. INSET FIGURE 1 HERE The model tested is as follows: ASSURANCE/PROVIDER = f(LEGAL, INDUSTRY, STAKEHOLDER, CONTROL VARIABLES) (1) In the first step of the sequential logit analyses, ASSURANCE takes the value of 0 in the case of the sustainability report not being assured and 1 where the report is assured. In the second step which only includes that subset of observations with assured sustainability reports, PROVIDER takes the value of 0 where the assurance provider does not belong to the auditing profession, and 1 in the case of assurance provided by a member of the auditing profession. LEGAL represents the quality of the legal environment and is measured using the “rule of law” measure developed by the World Bank (Kaufmann et al., 2007).10 The ‘Rule of Law’ score measures the extent to which agents have confidence in and abide by the rules of society, and in 9 For further technical details, refer to Ben-Akiva and Lerman 1985, Liao 1994 and Ophem and Schram 1997. We estimated the sequential logit model with the “seqlogit” module in STATA. We also re-ran the analyses with separate logistic regression analyses which gave the same results. 10 We use this measure rather than the La Porta et al. (1998) measures used by Durnev and Kim 2005 and Doidge et al. 2007 as this provides a relevant unique measure for each of the years 2002-2004. 17 particular the quality of contract enforcement, the police and the courts, as well as the likelihood of crime and violence. The industry categories used are mining, production, utilities, finance and other, on the basis of a firm’s Global Industry Classification Scheme (GICS) code and their main reported operating activities. We consider the mining, production, utilities and finance industries to be more exposed to environmental and social risks and therefore firms in these industries possess a greater need to increase user confidence in the credibility of their reported activities. MINING companies extract non-renewable resources with major environmental consequences while UTILITIES produce the largest amounts of greenhouse gas emissions and are exposed to community concerns about climate change. Further, the PRODUCTION industry is a major user of energy and can produce significant amounts of industrial waste products. Finally, FINANCE industries materially influence the financial well-being of societies and have a large ‘social footprint.’ As a result, stakeholder groups have a keen interest in the activities of these companies. Consistent with Ball et al. (2000), we consider companies domiciled in common law countries to have a more shareholder orientated corporate governance model and those in code law countries to have a more stakeholder orientated model. Firms in common law countries deal with shareholders at arms length and as a result an increased demand for information can be expected. In code law countries there is a greater degree of insider owners, such as banks, who get their information directly from management (Ball et al. 2000). Hence, we use a dummy variable for code law/common law as a proxy for STAKEHOLDER orientation. Based on insights from the literature on voluntary demand for assurance, we include the following company-related control variables in our model: size of the company (Ln(SALES)), 18 profitability of the company as measured by the return on assets (ROA), and leverage of the company as measured by long-term debt on total assets (LEVERAGE). The results are run for each separate year and on a pooled basis.11 VI. RESULTS Descriptive results As outlined earlier, we identified 2,113 sustainability reports from 31 countries in order to test hypotheses 1a and 2a. Table 1, Panel A, shows that the three main countries represented are Japan (527 observations), the U.K. (385 observations) and the U.S. (339 observations). In analysing the contents of the reports, we considered the six dimensions of the GRI framework: economic, environment, labor, human rights, product responsibility and society. A clear dichotomy was found between reports that covered multiple issues and reports that covered a single issue.12 As outlined in Table 1, Panel B, 1,612 of the 2,113 reports (76.3%) were broad, covering multiple dimensions. The other 501 reports covered a single dimension of the reports and in every one of these instances this was environment. These proportions are consistent with those reported in the KPMG (2005) survey, where 81% of the stand-alone sustainability reports covered multiple issues and 19% single issues. Because of the concerns of the potential heterogeneity of the sample, particularly on the question of assurance and the choice of assurance provider, the results will be reported both including and excluding the single issue (hereafter environmental) reports. 11 The pooled analysis is run both with and without year dummies (YEAR03 and YEAR04). 12 This coding was initially undertaken by a person experienced in reading these reports. An independent second coding of 200 of the year 2004 reports’ contents was undertaken by one of the authors, and not one discrepancy from the initial coding was identified. 19 Table 1, Panel C, shows that the industries represented are production (1,174 observations), utilities (365 observations), finance (268 observations), mining (119 observations) and other (187 observations). Table 1, Panel D, shows that 707 of these reports relate to year 2004 reports, 719 to 2003 reports, and 687 to reports from 2002. As expected, once a company produces a sustainability report they usually continue to produce this annually. Because of repeat observations over the years, and because in an emerging service the drivers behind this service may be evolving, we report the analysis for each year as well as the pooled analysis for 2002 2004. Of the 867 separate companies producing reports over any of the three years, 552 (63.67%) produced reports for all three years of our sample. Furthermore, Table 1, Panel D, shows that 655 (31.00%) of these sustainability reports (for 304 different companies) contained independent assurance reports13, and this subset of firms forms our sample for testing hypotheses 1b and 2b. For 275 (41.98%) of these 655 cases, the assurance provider was a member of the auditing profession.14 Table 1, Panel E shows that all assurance by the auditing profession was provided by the Big 4, with the major assurance provider being PricewaterhouseCoopers, accounting for 34.18% of assurance provided by the auditing profession. The assurers from outside the profession were comprised of a number of global networks (six were identified that provided assurance in more than one country, providing in total 31.49% of the assurance from outside the profession), and a large number of local assurers with varying qualifications. Of the 146 companies that had their sustainability reports assured for the three years, 54 were assured by the auditing profession, of which 50 (92.59%) kept the same assurance provider over the three 13 The proportion of sustainability reports containing an independent assurance report is consistent with the 30% identified by KPMG (2005). 14 This is a lower proportion than that reported by KPMG (2005), which identified that 58% of assurance observations were provided by major accountancy firms. These differences are most likely due to differences in sample composition, with KPMG’s sample comprising the top 100 companies from 16 countries. 20 years. Of the remaining 92 companies that were assured by the non-auditing profession, 55 (59.78%) kept the same assurance provider over the three years. Of the 655 companies that had their sustainability report assured, 497 had their financial statements audited by a Big 4 audit firm. Of these 497 companies, 198 (39.84%) also selected a Big 4 audit firm as the assurer of the sustainability report. Of the 158 companies whose financial statements were audited by other than a Big 4 audit firm, 78 (49.37%) selected a Big 4 audit firm as the assurer of their sustainability report. This suggests that there is little successful cross selling of this assurance service. Of the 198 organizations that continued to use a Big 4 firm as the assurer of their sustainability report, 129 (65.15%) chose the same Big 4 firm that audited their financial statements. - Insert Table 1 about here - The descriptive statistics are shown in Table 2. Our sample of companies producing sustainability reports is significantly larger, with a mean sales value of USD$15,042 million for those that do compared with USD$1,107 million for those that do not (t=77.77, p<0.001). They are also more profitable, with a return on assets of 3.2% compared with -2.3% (t=11.01, p<0.001), and their level of long-term debt to total assets is higher at 19.2% compared with 13.1% (t=17.14, p<0.001). With regards to the variables of interest for our hypotheses, their average LEGAL score is higher at 1.47 compared with 1.31 (t=13.92, p<0.001), suggesting that companies in stronger legal environments are more likely to produce these reports. With respect to STAKEHOLDER, 56.1% of the sample without sustainability reports and 51.6% of those producing sustainability reports come from code law countries and are therefore considered to be stakeholder orientated (Ball et al. 2000), showing that a higher proportion of shareholder 21 orientated companies are likely to produce sustainability reports. The correlation matrix is shown in Table 3 for both the entire sample and the sample with sustainability reports. Except for the expected negative correlations between the major industry classifications, there is no absolute correlation above 0.4. - Insert Tables 2 and 3 about here - Multivariate results As we outlined earlier, because the emphasis in this paper is on the assurance decision, we include an analysis of the factors associated with the decision to produce sustainability reports as background information only. A significant prior literature (see for example Berthelot et al., 2003; Cormier and Magnan, 1999) already focuses on understanding the incentives to produce sustainability or environmental reports and we do not seek to replicate that research here. This background information on the decision to produce sustainability reports is obtained by using sequential logit with the decision to report as the first stage of that analysis. The statistical model (Equation 1 above) is run with those companies who do not produce a sustainability report taking a value of 0 and those reporting taking a value of 1. Further, as outlined earlier, because of concerns about the potential heterogeneity of the reporting entities, we distinguished between those that produced an environmental (single issue) report, and those that produced a sustainability report covering multiple dimensions. For this reason, the results for the decision to assure and the choice of assurance provider are reported separately for all companies producing sustainability reports, and only companies producing multiple issue sustainability reports. The discussion of the multivariate results using sequential 22 logit analysis15 will cover all observations in the first place, and then an analysis of the results which excludes the environmental reports. It will also include an initial discussion of the results pooled for 2002-2004, followed by a discussion of the differences in the results for the single years. Background information on the decision to produce sustainability reports We find that large companies (t=54.72, p<0.001, 2-tailed) and more highly leveraged companies (t=2.28, p<0.05, 2-tailed) are more likely to produce stand-alone sustainability reports (see Table 4 Part 1). With regards to the industry characteristics we find that companies in the mining (t=16.59, p<0.001, 2-tailed), production (t=17.19, p<0.001, 2-tailed), utilities (t=13.51, p<0.001, 2-tailed), and finance (t=24.45, p<0.001, 2-tailed) industries are more likely to produce sustainability reports than companies in the other industries. In relation to country specific variables, companies residing in stakeholder countries (t=3.57, p<0.001, 2-tailed) and countries with stronger legal environments (t=6.64, p<0.001, 2-tailed) are more likely to produce sustainability reports. Decision to voluntarily purchase assurance For those companies that produced sustainability reports, the results of the sequential logit analysis of the factors associated with the decision as to whether or not to have these reports assured is outlined in Table 4 Part 2. In examining the control variables, we find that large companies (Ln(SALES)) are significantly more likely to have their sustainability reports assured compared to small companies (t=3.81, p<0.001, 2-tailed), while financial risk (LEV) was not associated with this decision (t=1.22, p>0.1, 2-tailed). Profitability (ROA) was significant in 15 We also run the analyses with separate logistic regression analyses, with the same results. 23 2004, causing this variable to be marginally significant for the pooled analysis examining all sustainability reports (t=1.71, p<0.1, 2-tailed), but was not significant for any of the periods when environmental reports were excluded. With respect to our hypotheses we find strong support for H1a, which states that companies with a higher need to enhance credibility will be more likely to have their sustainability reports assured. In particular, we find that companies in three of our four categories of industries that are expected to require enhanced credibility of published reports: MINING (t=4.47, p<0.001, 2-tailed), UTILITIES (t=2.32, p<0.05, 2-tailed) and FINANCE (t=2.64, p<0.01, 2-tailed), but not PRODUCTION (t=0.00, p>0.10, 2-tailed), are more likely to have their sustainability reports assured. With regards to H2a, we find that STAKEHOLDER is significant (t=6.03, p<0.001, 2 tailed), indicating that those companies in stakeholder countries are more likely to have their sustainability reports assured. However, this finding must be considered in light of the additional analysis reported later in this paper, which shows that this result is primarily attributable to a U.S. effect. LEGAL is positively significant (t= 1.84, p<0.1, 2-tailed), suggesting that those in the stronger legal system are more likely to be assured. Opposite of what was expected on the basis of findings for the quality of financial report assurance (Choi and Wong 2007), it appears that in weaker legal environments assurance is not used to increase user confidence in the credibility of sustainability reports. As outlined earlier, a potential alternative explanation is that the public’s perceived credibility of this type of assurance service is low in countries with a weak legal environment, and that therefore the benefits of assurance for companies do not outweigh the costs. Of relevance here is the decrease in this variable’s significance over the period 2002-2004, suggesting that while this may have been an appropriate alternative explanation in 2002, it does 24 not appear to be the situation in 2004. A similar result is found when environmental reports are excluded from the analysis.16 - Insert Table 4 about here - Choice of assurance provider The results of the sequential logit analysis for choice of assurance provider for those companies that have their sustainability reports assured is contained in Part 3 of Table 4. In analysing the control variables we find a significant positive association between the size of the company and the choice of a member of the auditing profession as assurance provider (t=2.90, p<0.01, 2-tailed), while there is no significant association between profitability of a company in the form of return on assets and assurance provider (t=-1.39, p>0.1, 2-tailed). Members of the auditing profession are also found to be more likely to be the assurance provider for companies with lower leverage (t=-2.90, p<0.01, 2-tailed). It is possible that this is a result of the auditing profession being less likely to associate with companies with higher levels of financial risk. The results show little support for H1b, i.e., that those companies with a higher need to enhance credibility are more likely to choose assurance from the auditing profession. In particular, legal environment (LEGAL) is marginally significant (t=-1.65, p<0.10, 2-tailed), while companies in the mining, production, utilities and finance industries are no more likely to choose a member of the auditing profession as their assurance provider than other companies. The lack of significant results holds across all years, both including and excluding environmental reports.17 16 For the decision to voluntarily purchase assurance, similar results are found for the pooled analysis when dummy variables are included for 2003 and 2004. The 2004 dummy variable is marginally significant, (t=1.79, p<0.10, twotailed), suggesting 2004 sustainability reports are more likely to be assured than 2002 reports, consistent with the descriptive statistics reported in Table 1. 17 MINING was marginally significant for the pooled analysis excluding environmental reports (t=1.88, p<0.10, twotailed). 25 The results provide strong support for H2b, that companies domiciled in countries that are more stakeholder orientated are more likely (t=6.03, p<0.01, 2-tailed) to choose assurance from the auditing profession compared with companies domiciled in countries that are more shareholder orientated. As distinct from the analysis of the assurance decision, this finding holds when the U.S. observations are excluded, as reported in the following additional analysis. VII. SENSITIVITY ANALYSIS While it is necessary to include all countries in order to properly analyse and gain an understanding of this international assurance service, it is possible that certain countries may be influencing the results. In analysing assurance services by the countries that constitute a significant proportion of our observations, we identify that assurance of sustainability reports is far lower in the U.S. than in other major countries. This may be due to the country’s attestation standards existing at the time, concerns over whether suitable criteria exist for such engagements and/or fear of litigation (Ballou et al. 2006).18 With the U.S. also rating as strongly shareholder orientated, we indeed find the results for the variable “stakeholder” in the assurance decision to be different depending on whether or not observations from the U.S. are included. The exclusion does not affect the interpretation of any other results, as outlined in Table 5. - Insert Table 5 about here - The results when including the U.S. observations show that companies in stakeholder countries are more likely to have their sustainability reports assured. This finding is entirely attributable to the U.S., as after excluding U.S. observations we find that there are no longer 18 The American Institute of Certified Practicing Accountants (AICPA) standards on attest engagements (AT Section 101) allow such assurance engagements in certain circumstances including that the practitioner must have adequate technical training and proficiency, knowledge of the subject matter and reason to believe that the subject matter is capable of evaluation against criteria that are suitable and available to the user. This is consistent with the requirements of the International Standard on Assurance Engagements (ISAE) 3000. 26 significant results for STAKEHOLDER with regards the decision to assure, across all years and both including and excluding environmental reports. However (and irrespective of whether the U.S. observations are included or excluded), once a decision to assure has been made, companies in STAKEHOLDER countries are more likely to choose a member of the auditing profession as their assurance provider. We also observe that a large number of observations come from two other countries, Japan (527 (24.94%)) and the U.K. (385 (18.22%)). Including dummies for U.K. and Japan does not change the results, with the exception that, for the decision to assure, the legal variable is no longer significant, while at the same time the U.K. dummy is significantly positive. The correlation between LEGAL and U.K. is 0.3091, suggesting that what we see as a legal effect on the decision to assure is primarily due to observations from the U.K. VIII. CONCLUSION This paper aims to develop an understanding of the international market for assurance services provided on general purpose stand-alone sustainability reports and the factors associated with the demand for such assurance and the choice of assurance provider. We use sequential logit analysis to: First, provide background information on the factors associated with the decision to produce these comprehensive reports; Second, for those that do produce, the factors associated with having such information assured; and Third, for those that do have this information assured, the factors associated with the choice of assurance provider. In particular, by providing insights into the market for assurance of sustainability reports and the market share captured by members of the auditing profession, we aim to inform the international assurance standard setting process in this new and growing field. 27 The results of this study generally support our empirical predictions that the incidence of assurance of sustainability reports is higher for companies with a greater need to enhance credibility. Our results demonstrate the demand for assurance is higher amongst companies engaging in more highly visible industrial activity and companies with a larger ‘social footprint’, with companies in MINING, UTILITIES and FINANCE all being more likely to have their sustainability reports assured. We also find that companies in stakeholder countries are more likely to have their sustainability reports assured. However, this latter finding must be considered in light of the additional analysis reported in this paper which shows that this result is attributable to a U.S effect. The results further suggest that those in the stronger legal system are more likely to be assured, although the decrease in this variable’s significance over the period of the study suggests that this factor is less significant in 2004 than it was in 2002, possibly as a result of the evolution and greater acceptance of this assurance service in the global market over this period of time. With respect to the choice of assurance provider, we do not find that companies domiciled in countries with weak legal environments or belonging to industries with higher environment and social risks are more likely to choose a member of the auditing profession as their assurance provider. Combining the findings of the decision to purchase assurance and the choice of provider, the important decision for industries needing to enhance credibility appears to be the decision to assure the information in the sustainability report, and the determination of whether the assurance provider is a member of the auditing profession is less important. We do find strong evidence of an association between the choice of assurance provider and the stakeholder orientation of a company’s country of domicile, with companies from stakeholder orientated countries being more likely to choose a member of the auditing profession 28 as their assurance provider. As distinct from the analysis of the assurance decision, this finding holds irrespective of whether the U.S. observations are included or excluded. This finding shows the importance of considering country-specific characteristics in gaining an understanding of the international assurance market. This is further reinforced by the fact that the initial result of a positive association between stakeholder-oriented countries and the decision to assure (but not the choice of assurance provider) is attributable entirely to the U.S. Considering country-specific characteristics is particularly important for understanding global assurance services, where the U.S. has historically had regulations in place which are unique to that country. Hence, our results contribute to the growing body of literature highlighting the importance of county-specific factors when considering accounting and assurance issues at an international level. These conclusions must be moderated by the following considerations. First, while the search techniques for identifying publicly available sustainability reports were comprehensive, they were not exhaustive. In particular, there is the possibility of a bias against reports that were not translated into English. Nonetheless, there were still a significant number of observations from non-English speaking countries, including the fact that the highest number of observations came from Japan. We also note that the representation of countries in this study is consistent with surveys examining the extent to which companies prepare sustainability reports (KPMG 2005). Additionally, the requirement to supplement the reports by financial and other information for the purpose of the analysis means that smaller companies are less likely to be included in our analysis as they are less likely to appear on the types of databases used to provide the additional financial and other information. 29 References Abdel-khalik, A. R. 1993. Why do private companies demand an audit? A case for organizational loss of control. Journal of Accounting, Auditing and Finance 8 (Winter): 31-52. Adams, C., W. Hill and C. Roberts. 1998. Corporate social reporting practices in Western Europe: Legitimating corporate behaviour? The British Accounting Review 30 (March): 121. Ball R., S. P. Kothari, and A. Robin. 2000. The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and Economics 29 (February): 151. Ballou, B., D. L. Heitger, and C. E. Landes. (2006). The future of corporate sustainability reporting. Journal of Accountancy 202 (December): 65-74. Ben-Akiva, M. and S. Lerman. 1985. Discrete Choice Analysis: Theory and Application to Travel Demand, Cambridge: MIT Press. Blackwell, D. W., T. R. Noland and D. B. Winters. 1998. The value of auditor assurance: Evidence from loan pricing. Journal of Accounting Research 36 (Spring): 57-70. Bradley, M., C. A. Schipani, A. K. Sundaram, and J. P. Walsh. 1999. The purposes and accountability of the corporation in contemporary society: Corporate governance at a crossroads. Law and Contemporary Problems 62 (Summer): 9–86. Brundtland Report, The. 1987. World Commission on Environment and Development Our common future. Oxford, U.K.: Oxford University Press. Carey, P., R. Simnett and G. Tanewski. 2000. Voluntary demand for internal and external auditing by family businesses. Auditing: A Journal of Theory of Practice 19 (Supplement): 37-51. Choi, J., and T. J. Wong. 2007. Auditors’ governance functions and legal environments: an international investigation. Contemporary Accounting Research 24 (Spring): 13-46. Chow, C. W. 1982. The demand for external auditing: Size, debt and ownership influences. The Accounting Review 57 (April): 272-91. Cormier, D., M. Magnan, and B. Velthoven. 2005. Environmental disclosure quality in large German companies: economic incentives, public pressures or institutional conditions? European Accounting Review 14(1); 3-39. CPA Australia. 2004. Triple bottom line – A study of assurance statements worldwide. Melbourne, Australia 30 Craswell, A., D. Stokes, and J. Laughton. 2002. Auditor independence and fee dependence. Journal of Accounting and Economics 33 (June): 253-275. DeAngelo, L. 1981. Auditor size and audit quality. Journal of Accounting and Economics 3 (December): 183-199. Doidge, C., G. A. Karolyi, and R. Stulz. 2007. Why do countries matter so much for corporate governance? Journal of Financial Economics 86 (October): 1-39. Dow Jones Sustainability Indexes. 2007. Available at: http://www.sustainability-index.com/ accessed May 2008. Durnev, A., and E.H. Kim. 2005. To steal or not to steal: firm attributes, legal environment and valuation. Journal of Finance 60 (June): 1461-1493. Elliott, R.K., and P.D. Jacobson. 2002. The evolution and the knowledge professional. Accounting Horizons 16 (March): 69-80. Francis, J.R., I. Khurana, X. Martin, and R. Pereira. 2006. The role of firm-specific incentives and country-level factors in the voluntary demand for independent audits. Working Paper, University of Missouri-Columbia. Francis, J.R., and D. Wang. 2008. The Joint Effect of Investor Protection and Big 4 Audits on Earnings Quality around the World. Contemporary Accounting Research 25 (Spring) 157– 191. Global Reporting Initiative, The (GRI). 2007. GRI register. Available at: http://www.globalreporting.org/Home, accessed May 2008. Hibbit, C. 2003. External environmental disclosure and reporting by large European companies. Unpublished PhD, Limperg Instituut. Holder-Webb, L., J. Cohen, L. Nath, and. D. Wood 2008. The supply of corporate social responsibility disclosures among U.S. firms. Journal of Business Ethics, forthcoming. Intergovernmental Panel on Climate Change (IPCC). 2007. Climate change 2007: The physical science basis. Available at: http://www.pnud.cl/recientes/IPCC-Report.pdf, accessed May 2008. Kaplan, R. S., and D. P. Norton. 1992. The balanced scorecard - Measures that drive performance. Harvard Business Review 70 (January/February): 71-79. Kaufmann, D., A. Kraay, and M. Mastruzzi. 2007, Governance Matters VI: Aggregate and Individual Governance Indicators 1996-2006. The World Bank. 31 Klapper, L., and I. Love. 2004. Corporate governance, investor protection and performance in emerging markets. Journal of Corporate Finance 10 (November): 703-728. Kolk, A. 2003. Trends in sustainability reporting by the Fortune Global 250. Business Strategy and the Environment 12 (September/October): 279-291. KPMG. 1999. KPMG International survey of corporate sustainability reporting 1999. Available at: http://www.kpmg.com., accessed November 2006. KPMG. 2002. KPMG International survey of corporate sustainability reporting 2002. Available at: http://www.kpmg.com, accessed November 2006. KPMG. 2005. KPMG International survey of corporate sustainability reporting 2005. Available at: http://www.kpmg.com, accessed May 2008. Lang, M.H., K.V. Lins, and D. Miller. 2004. Concentrated control, analysts following, and valuation: Do analysts matter most when investors are protected least? Journal of Accounting Research 42 (June): 589-623. La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny. 1998. Law and finance. The Journal of Political Economy 106 (December): 1113-1155. Liao, T. F. 1994. Interpreting the probability models: logit, probit and other generalized linear models. In series: Quantitative applications in the social sciences, No.101, 96 pages. Ophem, H. van, and A. Schram, 1997. Sequential and multinomial logit: A nested model. Empirical Economics. 22: 131-152. Patten, D. M. 2002. The relation between environmental performance and environmental disclosure: A research note. Accounting, Organizations and Society 27 (November): 762773. Rappaport, A. 1998. Creating shareholder value - a guide for managers and investors. New York, NY: The Free Press. Smith, J., A. Adhikari, and R. Tondkar. 2005. Exploring differences in social disclosures internationally: A stakeholder perspective. Journal of Accounting and Public Policy 24 (March/April): 123-151. Stern, N. 2007. The economics of climate change: the Stern review. Cambridge. Cambridge University Press. Stewart, G. B. 1999. The quest for value: A guide for senior managers. New York, NY: Harper Collins. 32 Sveiby, K. E. 1997. The intangible assets monitor. Journal of Human Resource Costing and Accounting 2: 73-97. Watts, R. L., and J. L. Zimmerman. 1986. Positive Accounting Theory. Englewood Cliffs, NJ: Prentice Hall. 33 Figure 1: Decision tree and number of observations for sequential logit analysis Reporting N=40,993 Sustainability report N=2,113 No sustainability report N=38,880 No assurance N=1,458 Assurance N=655 Assurance from the auditing profession N=275 Assurance from outside the auditing profession N=397 Note: Seventeen assurance engagements involved assurers from both within and outside the auditing profession. Decision tree: Sample excluding the U.S. Reporting N=32,605 Sustainability report N=1,774 No sustainability report N=30,831 No assurance N=1,140 Assurance N=634 Assurance from the auditing profession N=272 Assurance from outside the auditing profession N=379 34 Table 1 Panel A: Frequency of sustainability reports and their assurance Country No sustainability report Frequency Sustainability report Frequency Assured Frequency Not assured Frequency Proportion assured Stakeholder or shareholder orientated Legal score 2002-2003-2004 (Average) Australia 2,260 90 42 48 46.66% SHAREHOLDER 1.77;1.83;1.82 (1.81) Austria 175 17 8 9 47.06% STAKEHOLDER 1.84;1.81;1.81 (1.82) Belgium 275 18 3 15 16.67% STAKEHOLDER 1.47;1.51;1.51 (1.50) Brazil 370 13 5 8 38.46% STAKEHOLDER -0.34;-0.33;-0.33 (-0.33) Canada 577 116 16 100 13.79% SHAREHOLDER 1.74;1.77;1.8 (1.77) Denmark 346 21 12 9 57.14% STAKEHOLDER 1.85;1.92;1.97 (1.91) Finland 285 41 12 29 29.27% STAKEHOLDER 1.87;1.90;1.93 (1.90) France 1,605 80 32 48 40.00% STAKEHOLDER 1.28;1.36;1.41 (1.35) Germany 1,603 73 13 60 17.81% STAKEHOLDER 1.71;1.71;1.73 (1.72) Greece 246 16 1 15 6.25% STAKEHOLDER 0.69;0.75;0.81 (0.75) Hong Kong 401 15 7 8 46.66% SHAREHOLDER 1.11;1.29;1.37 (1.26) India 655 8 5 3 62.50% SHAREHOLDER -0.02;-0.01;0.00 (-0.01) Italy 576 61 39 22 63.93% STAKEHOLDER 0.77;0.76;0.65 (0.73) Japan 8,743 527 182 345 34.53% STAKEHOLDER 1.32;1.32;1.34 (1.33) Malaysia 2,130 4 3 1 75.00% SHAREHOLDER 0.41;0.44;0.55 (0.47) Netherlands 402 47 17 30 36.17% STAKEHOLDER 1.75;1.73;1.77 (1.75) New Zealand 173 11 3 8 27.27% SHAREHOLDER 1.79;1.86;1.92 (1.86) Norway 356 17 4 13 23.52% STAKEHOLDER 1.86;1.91;1.97 (1.91) Pakistan 105 1 0 1 0.00% STAKEHOLDER -0.76;-0.81;-0.86 (-0.81) Philippines 301 3 0 3 0.00% STAKEHOLDER -0.59;-0.62;-0.65 (-0.62) Portugal 92 10 1 9 10.00% STAKEHOLDER 1.22;1.24;1.19 (1.22) Singapore 1,235 3 0 3 0.00% SHAREHOLDER 1.54;1.69;1.82 (1.68) South Africa 248 54 5 49 9.26% SHAREHOLDER 0.07;0.04;0.15 (0.09) South Korea 715 15 8 7 53.33% STAKEHOLDER 0.79;0.65;0.7 (0.71) Spain 299 37 16 21 43.24% STAKEHOLDER 1.23;1.29;1.20 (1.24) Sweden 713 23 3 20 13.04% STAKEHOLDER 1.82;1.88;1.87 (1.86) Switzerland 500 60 31 29 51.66% STAKEHOLDER 1.92;1.97;1.98 (1.96) Taiwan 2,026 2 0 2 0.00% STAKEHOLDER 0.83;0.84;0.81 (0.83) Thailand 862 6 0 6 0.00% SHAREHOLDER 0.22;0.07;0.05 (0.11) UK 2,557 385 166 219 43.11% SHAREHOLDER 1.75;1.75;1.73 (1.74) US 8,049 339 21 318 6.19% SHAREHOLDER 1.57;1.55;1.48 (1.53) Total 38,880 2,113 655 1,458 30.99% Panel B: Type of sustainability report Auditor Frequency Percentage Single issue report (Environmental report) 501 23.71% Multi issue report 1,612 76.29% Total 2,113 100% 36 Panel C: Industry No sustainability report Sustainability report Assured Not assured Industry Frequency Frequency Frequency Frequency Proportion Assured Production 20,175 1,174 324 850 27.59% Utilities 3,428 365 133 232 36.43% Finance 776 268 94 174 35.07% Mining 1,641 119 53 66 44.53% Other (services, etc) 12,860 187 51 136 27.27% Total 38,880 2,113 655 1,458 30.99% Panel D: Year No sustainability report Sustainability report Assured Not assured Year Frequency Frequency Frequency Frequency Proportion Assured 2004 13,056 707 241 466 34.08% 2003 12,599 719 216 503 30.04% 2002 13,225 687 198 489 28.82% Total 38,880 2,113 655 1,458 30.99% Panel E: Details of assurance providers* AUDITING PROFESSION: Frequency Percentage No. of countries where firm assures Deloitte & Touche 48 17.45% 9 Ernst & Young 63 22.90% 10 KPMG 76 27.63% 11 PWC 94 34.18% 18 Other - - Total involving Auditing Profession: 275 100% OUTSIDE AUDITING PROFESSION: Environmental Resources Management (ERM) 31 7.81% 10 URS 23 5.79% 4 SGS 22 5.54% 5 CSR Network 20 5.04% 4 Bureau Veritas 18 4.53% 7 Corporate Citizen Coy 11 2.77% 2 Other 272 68.51% 1 Total involving Outside Auditing Profession: 397 100% *Fifty three reports are assured by more than assurer. Six were assured by two Big 4 audit firms, 17 were assured by a Big 4 audit firm and another assurer outside the auditing profession (resulting in the two sub-totals adding to 672, not 655), and 30 were assured by a combination of assurers outside the auditing profession. Table 2: Descriptive statistics Entire sample Year 2002 (N=13,912) Year 2003 (N=13,318) Year 2004 (N=13,763) Year 2002-2004 (N=40,993) Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Sales 1598.42 167.31 7534.74 1881.57 200.54 8700.10 1999.64 203.89 9442.95 1825.12 189.55 8591.91 Leverage 0.137 0.082 0.161 0.136 0.084 0.159 0.129 0.077 0.155 0.134 0.081 0.158 ROA -0.043 0.018 0.261 -0.017 0.024 0.216 0.0004 0.032 0.195 -0.020 0.025 0.227 Legal 1.310 1.540 0.515 1.328 1.550 0.534 1.317 1.41 0.525 1.318 1.480 0.525 Stakeholder 0.561 1 0.496 0.551 1 0.497 0.563 1 0.495 0.559 1 0.496 Sample without sustainability reports Year 2002 (N=13,225) Year 2003 (N=12,599) Year 2004 (N=13,056) Year 2002-2004 (N=38,880) Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Sales 1002.93 147.19 4715.21 1125.73 176.87 5104.84 1193.85 178.74 5564.73 1106.83 166.77 5139.13 Leverage 0.134 0.076 0.161 0.133 0.078 0.159 0.127 0.072 0.155 0.131 0.075 0.158 ROA -0.047 0.017 0.266 -0.020 0.024 0.222 -0.001 0.031 0.199 -0.023 0.024 0.232 Legal 1.301 1.47 0.519 1.319 1.55 0.539 1.309 1.41 0.529 1.310 1.470 0.529 Stakeholder 0.564 1 0.495 0.554 1 0.497 0.565 1 0.495 0.561 1 0.496 Sample with sustainability reports Year 2002 (N=687) Year 2003 (N=719) Year 2004 (N=707) Year 2002-2004 (N=2,113) Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Mean Median Std.dev. Sales 13061.72 5098.87 24170.84 15126.22 5930.35 27585.41 16879.88 6985.28 30525.65 15041.76 5934.11 27610.8 Leverage 0.196 0.176 0.149 0.195 0.177 0.148 0.183 0.170 0.138 0.192 0.174 0.145 ROA 0.015 0.020 0.108 0.035 0.028 0.061 0.045 0.038 0.058 0.032 0.029 0.079 Legal 1.477 1.57 0.405 1.477 1.55 0.414 1.464 1.48 0.418 1.473 1.550 0.412 Stakeholder 0.509 1 0.500 0.513 1 0.500 0.526 1 0.499 0.516 1 0.499 Variable definitions: Sales: Sales in millions U.S.$ Leverage: Long-term debt/total assets ROA: Return on assets: Net profit (loss)/total assets Legal: Legal environment: Rule of law score developed by the World Bank (Kaufmann, 2006). The ‘Rule of Law’ score measures the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, the police, and the courts, as well as the likelihood of crime and violence. Stakeholder: Dummy variable: company domiciled in code law country = 1. Table 3: Correlation matrix Entire sample (N=40,993) Ln(Sales) LEVERAGE ROA MINING PRODUCTION UTILITIES FINANCE OTHER STAKEHOLDER LEGAL Ln(Sales) 1 LEVERAGE 0.234 1 ROA 0.349 -0.033 1 MINING -0.237 -0.024 -0.082 1 PRODUCTION 0.055 -0.070 0.062 -0.220 1 UTILITIES 0.129 0.233 0.018 -0.067 -0.332 1 FINANCE 0.003 0.037 0.009 -0.034 -0.167 -0.051 1 OTHER -0.029 -0.072 -0.047 -0.141 -0.699 -0.214 -0.108 1 STAKEHOLDER 0.051 -0.115 0.108 -0.175 0.110 -0.034 -0.051 0.005 1 LEGAL 0.039 0.044 -0.125 0.104 -0.127 -0.019 0.014 0.104 -0.143 1 Sample with sustainability reports (N=2,113) Ln(Sales) LEVERAGE ROA MINING PRODUCTION UTILITIES FINANCE OTHER STAKEHOLDER LEGAL Ln(Sales) 1 LEVERAGE -0.012 1 ROA -0.026 -0.132 1 MINING -0.140 0.016 0.075 1 PRODUCTION 0.007 -0.184 0.100 -0.269 1 UTILITIES 0.004 0.364 -0.118 -0.110 -0.510 1 FINANCE 0.022 -0.191 -0.089 -0.091 -0.420 -0.174 1 OTHER 0.070 0.047 0.032 -0.075 -0.348 -0.142 -0.118 1 STAKEHOLDER 0.113 -0.114 -0.057 -0.161 0.220 -0.021 -0.149 -0.051 1 LEGAL 0.064 0.050 -0.012 0.042 -0.019 -0.010 -0.013 0.030 -0.226 1 Additional variable definitions: MINING: Dummy variable = 1 if company is in mining industry. PRODUCTION: Dummy variable = 1 if company is in production industry. UTILITIES: Dummy variable = 1 if company is in utilities industry. FINANCE: Dummy variable = 1 if company is in finance industry. OTHER: Company is in other than the mining, production, utilities or finance industries. Table 4: Results from each part of the sequential logit model SAMPLE INCLUDING ENVIRONMENTAL REPORTS SAMPLE EXCLUDING ENVIRONMENTAL REPORTS 2002 2003 2004 2002-2004 2002 2003 2004 2002-2004 Part 1: Reporting N=13,912 N=13,318 N=13,763 N=40,993 Constant -11.950 (-29.68)*** -12.393 (-31.30)*** -12.708 (-32.40)*** -12.284 (-53.99)*** Ln(SALES) 0.971 (30.75)*** 1.015 (31.85)*** 1.036 (32.09)*** 1.004 (54.72)*** LEVERAGE 0.143 (0.46) 0.502 (1.52) 0.712 (2.09)** 0.430 (2.28)** ROA 0.052 (0.13) 1.510 (2.11)** 1.675 (2.16)** 0.613 (1.61) MINING 2.865 (10.89)*** 2.631 (9.79)*** 2.214 (8.02)*** 2.573 (16.59)*** PRODUCTION 1.555 (9.77)*** 1.620 (10.28)*** 1.461 (9.68)*** 1.543 (17.19)*** UTILITIES 1.596 (8.44)*** 1.463 (7.83)*** 1.296 (7.08)*** 1.449 (13.51)*** FINANCE 3.658 (14.98)*** 3.442 (13.81)*** 3.338 (13.62)*** 3.476 (24.45)*** STAKEHOLDER 0.236 (2.27)** 0.186 (1.80)* 0.242 (2.37)** 0.211 (3.57)*** LEGAL 0.563 (3.57)*** 0.546 (3.76)*** 0.602 (4.19)*** 0.569 (6.64)*** Part 2: Assurance N=687 N=719 N=707 N=2,113 N=488 N=538 N=586 N=1,612 Constant -3.071 (-4.38)*** -3.591 (-4.70)*** -2.517 (-3.39)*** -2.995 (-7.16)*** -2.902 (-3.73)*** -3.095 (-3.82)*** -2.596 (-3.20)*** -2.825 (-6.25)*** Ln(SALES) 0.079 (1.38) 0.162 (2.66)*** 0.139 (2.33)** 0.129 (3.81)*** -0.004 (-0.07) 0.126 (1.78)* 0.134 (2.02)** 0.096 (2.51)** LEVERAGE 0.447 (0.66) 0.793 (1.23) 0.212 (0.30) 0.479 (1.22) 0.312 (0.40) 0.420 (0.58) 0.295 (0.39) 0.324 (0.73) ROA 1.187 (1.15) 0.060 (0.03) 3.541 (2.15)** 1.594 (1.71)* 1.563 (1.25) -0.901 (-0.61) 2.591 (1.61) 1.215 (1.36) MINING 1.275 (2.83)*** 1.198 (2.67)*** 0.948 (2.12)** 1.151 (4.47)*** 1.555 (3.06)*** 1.121 (2.42)** 1.101 (2.34)** 1.236 (4.56)*** PRODUCTION -0.081 0.143 -0.078 0.000 0.311 0.104 0.058 0.145 39 40 (-0.25) (0.45) (-0.27) (0.00) (0.77) (0.31) (0.19) (0.75) UTILITIES 0.348 (0.95) 0.559 (1.57) 0.470 (1.43) 0.465 (2.32)** 0.939 (2.11)** 0.685 (1.81)* 0.769 (2.17)** 0.775 (3.50*** FINANCE 0.633 (1.63) 0.725 (1.83)* 0.460 (1.20) 0.592 (2.64)*** 0.869 (1.91)* 0.536 (1.29) 0.504 (1.25) 0.589 (2.45)** STAKEHOLDER 0.607 (3.18)*** 0.657 (3.59)*** 0.700 (3.93)*** 0.639 (6.03)*** 0.742 (3.34)*** 0.903 (4.51)*** 1.042 (5.50)*** 0.902 (7.77)*** LEGAL 0.577 (2.36)** 0.320 (1.36) -0.072 (-0.34) 0.237 (1.84)* 0.691 (2.62)*** 0.267 (1.11) -0.079 (-0.35) 0.238 (1.78)* Part 3: Assurance provider N=198 N=216 N=241 N=655 N=143 N=167 N=217 N=527 Constant -2.755 (-1.91)* -0.775 (-0.59) -1.94 (-1.51) -1.774 (-2.35)** -4.273 (-2.33)** -1.192 (-0.85) -1.878 (-1.40) -2.092 (-2.55)** Ln(SALES) 0.150 (1.43) 0.086 (0.83) 0.254 (2.43)** 0.170 (2.90)*** 0.214 (1.58) 0.154 (1.31) 0.225 (2.02)** 0.200 (2.97)*** LEVERAGE -1.360 (-1.12) -3.537 (-2.72)*** -1.191 (-1.00) -1.953 (-2.90)*** -1.812 (-1.06) -3.579 (-2.33)** -1.316 (-1.05) -2.166 (-2.73)*** ROA 1.355 (0.47) -3.386 (-1.71)* -4.088 (-1.27) -1.754 (-1.39) 2.148 (0.50) -4.069 (-1.85)* -3.377 (-0.98) -2.211 (-1.51) MINING 1.225 (1.46) 0.454 (0.55) 0.492 (0.67) 0.642 (1.41) 2.108 (1.78)* 0.740 (0.86) 0.640 (0.86) 0.915 (1.88)* PRODUCTION 0.131 (0.22) 0.337 (0.58) -0.461 (-0.90) -0.040 (-0.12) 1.056 (1.08) 0.592 (0.94) -0.278 (-0.53) 0.258 (0.70) UTILITIES 0.617 (0.90) 1.108 (1.56) -0.168 (-0.28) 0.461 (1.22) 1.458 (1.25) 1.268 (1.60) -0.074 (-0.12) 0.646 (1.50) FINANCE 0.595 (0.86) 0.232 (0.34) -0.006 (-0.01) 0.233 (0.62) 1.424 (1.33) 0.409 (0.56) 0.187 (0.31) 0.459 (1.11) STAKEHOLDER 1.856 (4.05)*** 1.238 (3.36)*** 1.135 (3.16)*** 1.319 (6.03)*** 1.907 (3.41)*** 1.116 (2.79)*** 1.175 (3.11)*** 1.272 (5.38)*** LEGAL -0.185 (-0.36) -0.501 (-1.13) -0.477 (-1.21) -0.421 (-1.65)* -0.047 (-0.08) -0.689 (-1.48) -0.478 (-1.21) -0.479 (-1.79)* Log pseudolikelihood (Wald chi-squared) -2136.58 (1196.00)*** -2175.48 (1252.07)*** -2186.14 (1262.00)*** -6522.50 (3705.62)*** -1620.71 (989.40)*** -1718.87 (1049.75)*** -1868.61 (1114.93)*** -5233.79 (3142.50)*** *,**,*** Significantly different from zero at the α = 0.10; 0.05; 0.01 level for two-tailed tests. Table 5: Results from the sequential logit model for sample without the U.S. SAMPLE INCLUDING ENVIRONMENTAL REPORTS SAMPLE EXCLUDING ENVIRONMENTAL REPORTS 2002 2003 2004 2002-2004 2002 2003 2004 2002-2004 Reporting N=10,991 N=10,521 N=11,093 N=32,605 Constant -11.676 (-29.53)*** -12.182 (-30.64)*** -12.164 (-31.51)*** -11.918 (-53.11)*** Ln(SALES) 1.053 (28.74)*** 1.104 (29.75)*** 1.091 (30.19)*** 1.078 (51.22)*** LEVERAGE 0.719 (1.92)* 1.259 (3.12)*** 1.382 (3.35)*** 1.099 (4.82)*** ROA 0.217 (0.42) 2.297 (2.66)*** 1.479 (1.61) 0.799 (1.75)* MINING 2.728 (8.81)*** 2.470 (7.68)*** 2.104 (6.51)*** 2.433 (13.29)*** PRODUCTION 1.583 (8.95)*** 1.641 (9.31)*** 1.533 (8.91)*** 1.584 (15.72)*** UTILITIES 1.602 (7.48)*** 1.431 (6.85)*** 1.273 (6.07)*** 1.435 (11.83)*** FINANCE 4.044 (15.14)*** 3.740 (13.25)*** 3.548 (12.98)*** 3.773 (23.83)*** STAKEHOLDER -0.892 (-7.47)*** -0.961 (-7.96)*** -0.810 (-6.64)*** -0.898 (-12.94)*** LEGAL 0.623 (4.93)*** 0.584 (5.04)*** 0.539 (4.75)*** 0.570 (8.38)*** Assurance N=575 N=605 N=594 N=1,774 N=383 N=431 N=478 N=1,292 Constant -3.139 (-4.39)*** -3.769 (-4.78)*** -2.959 (-3.80)*** -3.236 (-7.49)*** -3.090 (-3.74)*** -3.464 (-4.05)*** -3.125 (-3.69)*** -3.223 (-6.75)*** Ln(SALES) 0.175 (2.77)*** 0.253 (3.74)*** 0.262 (3.79)*** 0.231 (6.07)*** 0.116 (1.56)* 0.243 (3.01)*** 0.280 (3.52)*** 0.223 (5.02)*** LEVERAGE 0.457 (0.65) 0.894 (1.32) 0.200 (0.27) 0.524 (1.27) 0.502 (0.61) 0.785 (1.03) 0.321 (0.40) 0.548 (1.18) ROA 0.990 (1.11) 0.131 (0.07) 3.185 (1.85)* 1.331 (1.56) 1.096 (1.16) -1.160 (-0.74) 1.870 (1.14) 0.761 (1.03) MINING 1.174 (2.45)** 1.032 (2.17)** 0.899 (1.91)* 1.066 (3.90)*** 1.417 (2.61)*** 0.940 (1.91)* 1.007 (2.08)** 1.126 (3.92)*** PRODUCTION -0.012 0.265 0.284 0.199 0.388 0.246 0.415 0.364 42 (-0.03) (0.75) (0.86) (1.00) (0.88) (0.68) (1.24) (1.70)* UTILITIES 0.461 (1.17) 0.673 (1.77)* 0.759 (2.10)** 0.651 (2.99)*** 1.028 (2.15)** 0.776 (1.90)* 1.038 (2.68)*** 0.942 (3.91)*** FINANCE 0.530 (1.29) 0.683 (1.64) 0.511 (1.23) 0.575 (2.41)** 0.725 (1.50) 0.459 (1.05) 0.484 (1.13) 0.534 (2.10)** STAKEHOLDER -0.020 (-0.09) 0.061 (0.30) -0.028 (-0.14) -0.008 (-0.07) 0.070 (0.29) 0.251 (1.12) 0.251 (1.18) 0.197 (1.53) LEGAL 0.437 (1.98)** 0.206 (0.95) -0.215 (-1.08) 0.112 (0.94) 0.512 (2.22)** 0.136 (0.62) -0.241 (-1.15) 0.090 (0.75) Assurance provider N=192 N=208 N=234 N=634 N=138 N=161 N=212 N=511 Constant -2.723 (-1.89)* -0.665 (-0.51) -1.977 (-1.54) -1.776 (-2.35)** -4.156 (-2.29)** -1.028 (-0.74) -1.697 (-1.27) -1.970 (-2.41)** Ln(SALES) 0.143 (1.36) 0.082 (0.79) 0.259 (2.49)** 0.170 (2.90)*** 0.201 (1.50) 0.147 (1.25) 0.232 (2.09)** 0.200 (2.97)*** LEVERAGE -1.240 (-1.00) -3.445 (-2.53)** -1.114 (-0.93) -1.828 (-2.58)** -1.731 (-1.00) -3.623 (-2.31)** -1.465 (-1.18) -2.185 (-2.72)*** ROA 0.870 (0.32) -3.99 (-1.86)* -3.759 (-1.18) -1.949 (-1.56) 1.408 (0.33) -4.856 (-2.01)** -3.218 (-0.94) -2.185 (-2.72)* MINING 1.235 (1.47) 0.550 (0.66) 0.596 (0.77) 0.716 (1.55) 2.113 (1.79)* 0.864 (0.99) 0.581 (0.75) 0.918 (1.85)* PRODUCTION 0.137 (0.23) 0.363 (0.62) -0.409 (-0.75) 0.009 (0.03) 1.058 (1.11) 0.631 (0.99) -0.386 (-0.70) 0.231 (0.62) UTILITIES 0.598 (0.88) 1.073 (1.50) -0.157 (-0.26) 0.463 (1.22) 1.434 (1.25) 1.255 (1.58) -0.196 (-0.31) 0.590 (1.35) FINANCE 0.605 (0.88) 0.205 (0.30) -0.005 (-0.01) 0.247 (0.65) 1.424 (1.34) 0.374 (0.52) 0.004 (0.01) 0.382 (0.91) STAKEHOLDER 1.870 (3.85)*** 1.174 (3.04)*** 1.065 (2.89)*** 1.279 (5.62)*** 1.914 (3.21)*** 1.048 (2.50)** 1.061 (2.80)*** 1.210 (4.96)*** LEGAL -0.184 (-0.35) -0.513 (-1.17) -0.481 (-1.24) -0.429 (-1.70)* -0.051 (-0.08) -0.705 (-1.52) -0.486 (-1.27) -0.493 (-1.87)* Log pseudolikelihood (Wald chi-squared) -1670.91 (1049.12)*** -1700.88 (1109.98)*** -1749.66 (1145.41)*** -5147.14 (3296.11)*** -1193.20 (821.45)*** -1277.95 (897.06)*** -1453.73 (982.96)*** -3950.37 (2686.14)*** *,**,*** Significantly different from zero at the α = 0.10; 0.05; 0.01 level for two-tailed tests. View publication stats View publication stats