The Environmental Performance-Reporting Relationship:
How does it work?
Bachelor Thesis
Vincent Christiaan Altorffer 10215751
University of Amsterdam BSc Accountancy & Control Supervisor: Mr. C. Clune MSc
Abstract
In order to meet the important global challenge of minimizing the environmental impact of our societies, firms will have to improve their environmental performance. Environmental reporting has been named as a way to contribute to this challenge. For environmental
reporting to have an impact on the environmental performance of firms, there needs to exist a relation between environmental performance and environmental reporting. Therefore, this literature review aims to answer the question: what is the nature of the relation between environmental performance and environmental reporting? The nature of the relation has been theorized from two different perspectives: economic voluntary disclosure theory, and socio-political legitimacy theory. This thesis ads to the existing literature by combining these theories with empirical evidence to come to a comprehensive conclusion on the workings of the relation between performance and reporting. As predicted by voluntary disclosure theory, good environmental performers will disclose more objective and comparable information to distinguish themselves from poor performers. Poor performers, on the other hand, follow the predictions of legitimacy theory by disclosing more information containing less objective and comparable information and more empty statements and general descriptions. However, these direct links between performance and reporting are distorted by three types of outside factors: societal factors, industry factors and individual factors. These findings lead to a model which describes the workings of the performance-reporting relationship.
Samenvatting
Om het wereldwijde doel te bereiken het effect van onze samenleving op het milieu te minimaliseren, zullen organisaties hun milieuprestaties moeten verbeteren. Milieurapportage is genoemd als manier om aan dit doel bij te dragen. Echter, milieurapportage kan hier pas aan bijdragen als er een relatie bestaat tussen milieuprestaties en milieurapportage. Met het oog op dit gegeven, tracht dit literatuuronderzoek de volgende onderzoeksvraag te
beantwoorden: wat is de aard van de relatie tussen milieuprestaties en milieurapportage? Twee theorieën voorspellen hoe deze relatie werkt: economische ‘voluntary disclosure’ theorie en socio-politieke legitimiteitstheorie. Deze scriptie draagt bij aan de bestaande literatuur door deze twee theorieën te combineren met empirisch onderzoek en hierdoor tot een veelomvattende conclusie te komen over hoe deze relatie tussen milieuprestaties en milieurapportage werkt. Zoals voorspeld door ‘voluntary disclosure’ theorie, zullen organisaties die goed presteren zich proberen te onderscheiden van slechte presterende organisaties door meer objectieve en vergelijkbare informatie op te nemen in de milieurapportage. Echter, slecht presterende organisaties zullen de voorspellingen van
legitimiteitstheorie volgen door een grotere hoeveelheid informatie te rapporteren met minder objectieve en vergelijkbare informatie, maar meer algemene informatie die moeilijk te
interpreteren of controleren is. Deze directe relatie tussen milieuprestaties en milieurapportage wordt verstoord door drie soorten externe factoren: maatschappelijke factoren, industrie-specifieke factoren en individuele factoren. Deze bevindingen leiden tot een model dat de werking van de relatie tussen milieuprestatie en milieurapportage beschrijft, rekening houdend met de invloed van deze externe factoren op de directe relatie.
Table of Contents Abstract 2 Samenvatting 2 Table of Contents 3 1 Introduction 4 2 Theoretical Contradictions 5
2.1 Voluntary Disclosure Theory 5
2.2 Legitimacy Theory 7
3 Relation Between Performance and Reporting 9
3.1 Does A Link Exist? 9
3.2 Other Factors of Influence 16
4 Discussion 20
5 Conclusion 22
1 Introduction
One of the biggest global challenges over the coming decades will be to limit the impact of our societies on the environment. To achieve this, organisations will have to improve their environmental performance. The best way to give organisations incentives to do so, has been the subject of a debate for some time (Sinclair, 1997). This debate often revolves around two alternatives that are each subject to a large amount of criticism: environmental legislation versus industry self-regulation (Sinclair, 1997). Alternatives to these two types of
environmental regulation, however, do certainly exist. One alternative that has been mentioned in this context, is environmental reporting (Sinclair, 1997). However, since environmental regulation tries to impact environmental performance, there needs to be a relation between environmental performance and environmental reporting. This raises the question whether or not this relation actually exists and if it does, how it then may function.
This interest in the subject of the link between environmental reporting and
environmental performance from an academic standpoint is not new: empirical research in this area has been undertaken as early as in the 1980s (Ingram and Frazier, 1980; Wiseman, 1982). Since that time, many theoretical and empirical papers have been written on this subject (Bewley & Li, 2000; Clarkson, Li, Richardson & Vasvari, 2008; Fekrat, Inclan & Petroni, 1996; Patten, 2002). However, what is missing is a clear link between these different papers, to see what we have learned about the association between environmental reporting and environmental performance until now.
The goal of this literature review is to find out what the nature of the association
between environmental performance and environmental reporting is. In this context,
environmental reporting and environmental disclosure are treated as one and the same (as they are in most of the literature), because they both mean making private environmental
information public. Thus, the research question of this thesis is: what is the nature of the relation between environmental performance and environmental reporting? By answering this research question, this thesis adds to the existing literature by combining theory and different types of empirical research by creating a formerly missing, comprehensive conclusion as to how the relation between reporting and performance functions.
Two competing theories try to explain this relation between environmental
performance and reporting: economic voluntary disclosure theory and socio-political legitimacy theory (Gray, Kouhy & Lavers, 1995; Li, Richardson & Thornton, 1997). These theories are combined with different types of empirical research (Clarkson, Li, Richardson &
Vasvari, 2008; Fekrat, Inclan & Petroni, 1996; Patten, 2002; Rockness, 1985; Wiseman, 1982). By studying the nature of the relation between environmental performance and environmental reporting or disclosure in this way, this review constructs a model that encompasses both the direct link between environmental performance and environmental reporting (which follows predictions from both voluntary disclosure theory and legitimacy theory), as well as three types of outside factors: societal factors, industry factors, and individual factors.
This literature review is split into four sections. The next section will give an outline
of voluntary disclosure theory and legitimacy theory, and how these theories predict the environmental performance-reporting relationship works. Then, section three will evaluate a number of different empirical papers, which will be used to test the two theories outlined in section two. Section three will also evaluate factors outside of the direct
performance-reporting relationship that influence environmental performance-reporting and disclosure, which will then be combined with the empirical and theoretical links discussed earlier to construct a model of the relationship between performance and disclosure. Section four contains a discussion on the findings of this thesis, as well as suggestions for future research. Section five contains the conclusion.
2 Theoretical Contradictions
Voluntary disclosure of any sort, being financial, social, or indeed environmental, is
characterized by a choice that firms have: the choice to disclose certain information, while at the same time holding other information back. Both in economics and in socio-politics, there is a large body of literature surrounding this choice. In economics, the body of literature focuses on voluntary disclosure theory, while in socio-politics the literature focuses on legitimacy theory.
2.1 Voluntary Disclosure Theory
Verrechia (1983) sees the choice regarding the disclosure of information as an optimization process: managers will weigh the costs of disclosing information (proprietary costs) against the costs of withholding information. If the cost of disclosing information outweighs the cost of withholding information, the information will be held back, but if the cost of disclosing is
lower than the cost of withholding information, the information will be disclosed. This results in a partial equilibrium, because stakeholders that are affected by information asymmetry cannot tell whether information remains undisclosed because it is bad news, or good news that is not good enough to disclose, and thus incur the proprietary costs of disclosure. Dye (1985) also cites the possibility that firms choose to withhold some information from stakeholders when this information is regarded as “bad news”. Dye (1985) names three reasons why this could be the case: first in case of incomplete knowledge of stakeholders about management’s information, second in case of information already being subject to non-voluntary disclosure, and third in case of a principal-agent problem between shareholders and management, in which case the best resolution would be for management to not disclose certain information. Although this last option does not necessarily apply to all stakeholders (since not all stakeholders have a principal-agent relationship with a firm’s management), similar situations could occur between other stakeholders and management, for instance in the case where certain stakeholders can influence legislation. In essence, voluntary disclosure theory dictates that managers of firms will always have a tendency, or drive, to disclose as much good information as possible, and withhold as much bad information as possible (Bewley & Li, 2000).
Regarding environmental disclosure, the point of influence over legislation might well
be a very interesting one, especially with an eye on environmental lobbying. Since the studies mentioned earlier are primarily concerned with financial information, however, the question remains whether or not these theories also apply to environmental information. Li, Richardson and Thornton (1997) also make the point of stakeholders using (environmental) reports to take actions against polluting firms. More importantly, they tie the concepts of voluntary disclosure theory to environmental reporting. Li et al. (1997) propose a game-theory based model with three players: firms in an industry that have a given need for polluting, a group of risk-neutral investors, and an opponent that has the same information as the risk-neutral investors. This opponent is a party that has a set of concerns about the pollution emitted by the industry. The firm is rational, and will therefore choose a net-equity-maximizing
disclosure strategy. In this model, the environmental pollution caused by a firm constitutes a liability for that firm. Also, when a firm discloses information about its pollution, and the amount of pollution surpasses the tolerance level of the opponent, the opponent will try to impose a cost C on the firm. Thus, firms have two reasons to withhold negative information: to avoid a more negative assessment of net equity by the market (because of an increased environmental liability), and to avoid incurring cost C. Bewley and Li (2000) elaborate on this
model, by empirically testing the relation between environmental disclosure and five factors: knowledge of outsiders about environmental exposure, pollution propensity, political
exposure, quality of the auditor and the financial performance of the firm. They find that political exposure, pollution propensity and outsider’s knowledge correlate positively with the amount of voluntary environmental disclosure. Thus, the economic theory of voluntary
environmental disclosure can be interpreted as follows: more public exposure and more positive information (and thus, better environmental performance) lead to more disclosure and better environmental reports.
2.2 Legitimacy Theory
As outlined earlier, Bewley and Li (2000) find that not only political exposure and outsider’s knowledge show positive correlation with voluntary environmental disclosure, but also that pollution propensity of an industry or firm has a positive effect on the amount of disclosure and reporting. Bewley and Li (2000) are not alone in finding pollution propensity as an explaining factor in the amount of environmental disclosure released by firms, even when looked at from the perspective of voluntary disclosure theory. Li, Richardson and Thornton (1997) also find pollution propensity of industries and individual firms to be of influence on the amount of environmental information that is disclosed by firms, even though the model they propose is based on voluntary disclosure theory (which in turn derives from other economic theories, most importantly game theory). This raises questions on how this relation between environmental disclosure and pollution propensity might function, as it is not a logical relation from the perspective of voluntary disclosure theory: firms in polluting
industries are indeed more likely to create bad environmental news, and will therefore be less likely to disclose information of this sort.
Socio-political theory in the field of environmental disclosure focuses on precisely this
relationship between pollution propensity, environmental disclosure and environmental information (Clarkson et al., 2008). Gray, Kouhy and Lavers (1995) criticize the “traditional” approach to the theory of environmental disclosure as outlined earlier. Their study provides an overview of the development of environmental and social disclosure of UK firms between 1979 and 1991. They explain the rise of environmental disclosure through the use of
Lindblom’s legitimation strategy (Gray et al., 1995). In this study, only a very minor portion of the observed firms actually changed their environmental performance to influence their environmental reports. Instead, most firms used environmental disclosure to change the
perception of their environmental performance, rather than the actual performance itself (Gray et al., 1995). This points towards an explanation of the association between environmental reports and environmental performance that is the exact opposite of the explanation given by economic voluntary disclosure theory. Socio-political legitimacy theory does exactly this, and explains the association between performance and reporting as an inverse one: the more negative the actual environmental performance information is, the more firms will try to
change their perceived performance via environmental reports and disclosure1.
Patten (1992) uses this legitimacy theory to examine the effects of the Exxon oil spill
off the coast of Alaska in 1989 on the environmental reports of other companies in the same industry (petro-chemicals). He tries to determine whether or not the damage to the legitimacy of one firm (in this case Exxon) drives other, similar, firms to come up with proof of their legitimacy and relevance relative to the damaged firm. He determines this in the light of the legitimacy theory of Shocker and Sethi. This theory dictates that social institutions, such as businesses, operate through social contracts, which will only last as long as they deliver socially desirable ends to the society in which it operates, and distribute benefits to the group from which they derive power (such as stockholders in the case of publicly traded companies) (Patten, 1992). This does not only extent to economic benefits, but according to Patten (1992) also social benefits, of which environmental benefits are a part. Patten (1992) finds that petroleum companies did indeed increase environmental disclosure in response to the Alaska oil spill, which implicates that there is at least partial evidence to support an approach from the perspective of legitimacy theory.
The fact that both voluntary disclosure theory and legitimacy theory seem to have at
least partial evidence supporting them, asks for a further exploration of empirical literature on the relation between environmental reporting and environmental performance. Not only to test what theory seems to be more dominant, but also to see if a relation seems to exist between reporting and performance on a larger scale, and in different industries and geographical areas. In paragraph 3 a larger body of empirical literature will be used to answer these questions.
1 This section focuses on explaining the link legitimacy theory has to environmental performance and environmental reporting. Therefore, only papers have been cited that use legitimacy theory in this particular context. For a broader background on legitimacy theory see Suchman (1995).
3 Relation Between Performance and Reporting
Before questions surrounding the nature of the link between environmental reporting and environmental performance become relevant, the existence of a relation between reporting and performance first needs to be empirically tested. Since reporting is essentially a way for stakeholders to evaluate the performance of an entity (in both financial and non-financial terms), it is crucial that there is a link between the report and the subject of the report in order for reporting to be useful and convey meaning. A body of empirical literature has been formed since the 1980s surrounding the link between environmental reporting and
performance, starting with Ingram and Frazier (1980) and Wiseman (1982). Since then, a distinct body of literature has developed around the relation between reporting and performance.
3.1 Does A Link Exist?
It is crucial, in order to find out whether or not an association between environmental performance and environmental reporting exists, to find a reliable measure for both performance and disclosure. As touched upon earlier, the first real attempt at finding an empirical link between environmental performance and corporate disclosure was done by Ingram and Frazier (1980) (Patten, 2002). Their measure of environmental performance came from the environmental performance measures from the 1977 Pollution Audit report compiled by the Council on Economic Priorities (CEP), a non-profit public interest organisation
dedicated to evaluating the social performance of corporations (Rockness, 1985). For a way of gaining testable data from the corporate disclosures of the companies from the CEP report, Ingram and Frazier adapted four testable dimensions previously used in related literature: evidence, time, specificity, and theme, with a total of twenty content categories. The reports were then judged by two judges, who checked for the existence of each of the twenty content categories in the reports. Their findings were then used to compute standardized scores, which could be tested relative to actual performance. Ingram and Frazier (1980) tested the relevant firms in the CEP report in this way, and found little basis for the existence of an association between the standardized reporting scores and the CEP performance measures. They concluded that “the results of this study indicate, at best, only a weak association between quantitative measures of disclosure content and independent measures of social performance” (1980, p.620).
In a direct response to Ingram and Frazier (1980), Freedman and Jaggi (1982) tried to re-examine the association between performance and reporting based on three central points of criticism. First, the reporting information was limited to annual reports, and ignored 10K filings (special, mandatory reports filed with the SEC) and special social reports. Second, the use of two independent judges was questioned due to possible bias in the classification process. Third, the selection of the time periods was questions due to inconsistencies in the type of disclosure associated with the two periods. On the basis of these, Freedman and Jaggi (1982) used a different method to establish a pollution disclosure index, while still using the CEP report’s performance index. Instead of using four dimensions with twenty content categories to construct a reporting index, Freedman and Jaggi looked at the occurrence of emission standards, and environmentally related capital expenditures (current, past, and present), as well as descriptive disclosure. Weights were attached to the different occurrences, resulting in a reporting index that differs from that of Ingram and Frazier (1980). However, despite the use of a different reporting index, Freedman and Jaggi’s results supported Ingram and Frazier’s findings that there is no significant association between environmental
performance and disclosure based on both Spearman Rank test and Pearson product-moment correlation test. Because Freedman and Jaggi’s reporting index puts considerable weight on emission standards, their explanation for the absents of an association is that pollution
disclosures have not been strictly enforced, so that environmental performance was not a main issue for the management of firms. Although it is possible that environmental performance was not an issue for management and therefore the environmental performance lagged behind legislation, this is only a possible explanation for the absents of a positive association between reporting and performance. Also, the heavy weight put on environmental regulation might not be the right way to compare reporting to performance, because not all measures of
environmental performance are regulated.
Wiseman (1982) puts less emphasize on environmental regulation to evaluate
environmental reports. In her 1982 paper, Wiseman developed a new index that is known as the Wiseman Index (Wiseman, 1982; Clarkson, Li, Richardson & Vasvari, 2008). The index uses 18 different measures of reporting divided into 4 categories (economic factors, litigation, pollution abatement, and other environmentally related information) and gives them each a score between 0 and 3: from no disclosure (worst) to quantitative disclosure (best). On the basis of this, an index is created ranging from a minimum of 0 (18*0) to a maximum of 54 (18*3). Wiseman (1982) tested the association between the resulting index and the CEP rankings using Spearman’s Rank Correlation Coefficient. No statistically significant results
were found, apart from a statistically significant association between the Economic Disclosure Index and CEP rankings in the pulp and paper industry. Wiseman herself commented on the significance of this association with relative scepticism, saying that “this was not the case in other industries, and the negative direction of this coefficient in the other industries may imply the opposite conclusion” (1982, p.61). Because Wiseman’s aim was to test the quality of disclosure by firms, she concluded that there was a good chance that environmental disclosure might actually misrepresent a firm’s performance, but that further research was necessary to determine the actual quality of environmental disclosure by firms. The observation that environmental disclosure might misrepresent performance is a very interesting one, since statistical insignificance implies no association, meaning for instance incompleteness or vagueness of reports, and not an inverse relation.
However, Wiseman was not alone to conclude that environmental disclosure might
misrepresent environmental performance. Rockness (1985), too, investigated the relation between disclosure and performance of the steel, oil, and pulp and paper industries. As a measure of performance Rockness also used the CEP rankings, but rather than the Wiseman Index, Rockness used Nunnally’s Q-sort ranking method to evaluate the environmental disclosure per firm. This ranking procedure requires participants of the study, in this case financial analysts, MBA students, environmental activists, and environmental regulators, to read and sort all disclosures into three piles of equal size from best to worst and to rank the individual firms from best to worst in each of the three piles. This gives a selection of the firm with the best perceived performance to the worst perceived performance for each of the participants. Rockness (1985) compared these results to the CEP rankings, yielding
Spearman’s Rank Order Correlation Coefficients for each group of participants and industry. None of these coefficients showed any significant positive correlation between CEP rankings and participant rankings, leading to the conclusion that the information in environmental disclosure is at best an incomplete report of real performance. But, as cited earlier, Rockness, like Wiseman, points to the fact that although the coefficients were not statistically
significant, the majority of the coefficients were still negative. In this particular context, this implies that participants evaluated the disclosures of companies with the worst environmental performance as having the best, while at the same time evaluating companies with the best performance as having the worst. According to Rockness (1985), this does point to the possibility of an inverse relation between reporting and performance, again despite the statistical insignificance.
association between environmental reporting and environmental performance. Yet these papers are not enough to conclude that no association between reporting and performance exists. There are two reasons for this. First, social accounting, and environmental reporting as a part of that, was still relatively young in the 1980s. Therefore, it is highly plausible that no accepted practices, guidelines, or standards on environmental reporting had matured yet, particularly inside individual firms. In fact, the scope and quality of disclosures varied so considerably between companies, that over the period after 1980 the SEC sought an
information-exchange agreement with the Environmental Protection Agency (EPA) in order to identify inadequate and inaccurate disclosure (Fekrat, Inclan & Petroni, 1996). Second, all studies use the same CEP report to create an index for environmental performance. At the time, the CEP report was deemed the only reliable, and for this reason usable, source for environmental performance evaluations (Wiseman, 1982). However, the CEP report only contains 37 observations of 31 U.S. companies from the industries relevant to the studies previously examined, being steel, oil refining, paper and pulp, and chemicals (Freedman & Jaggi, 1982). Thus, the use of the CEP report severely limits the amount of data used in the studies for not only performance measures, but consequently also reporting measures (because both performance and reporting data have to come from the same company). All data comes from the same relatively small pool of observed firms, which is not enough to extrapolate the conclusion from the examined studies to a broader, global setting. Because of this, a look should be taken at more recent studies to come to a comprehensive conclusion on the existence of an association between environmental reporting and environmental
performance.
A 1996 study by Fekrat et al. uses a methodology similar to Wiseman’s (1982),
including the same number of 26 U.S. firms from a variety of different industries, including the automobile and oil industries. This set up gives insight into the effects of the information-exchange agreement and the possible development of best practices and guidelines in
individual firms. Despite over a decade of possible development, the study still shows a considerable amount of variance between individual firms and industries (Fekrat et al., 1996). In fact, the study confirms the findings of Wiseman (1982). According to Fekrat et al. “it might be fair to say that not much has changed since Wiseman’s study was published” (1996, p.186). Although the evidence presented in the study does point towards this conclusion, several flaws from earlier research persisted, most notably of which is the small sample size used in the study.
between reporting and performance is threefold. Two of the problems, being inadequate sample selection and inadequate performance measures, have been touched upon earlier. However, Patten (2002) ads to this a third concern: a failure to control for other factors, such as company size or other potential industry effects. In his research, Patten controls for these three factors by taking a larger sample size (131 firms of which 63 came from
environmentally sensitive industries), using toxic release per dollar revenue (to control for the inadequate performance measure), and adding various control variables in the regression analysis, such as a control variable for the interaction effect between environmental
sensitivity of an industry and the extensiveness of reporting. The resulting model was tested through regression analysis. The analysis resulted in an F-statistic of 16,88 (significance 0,00) and an adjusted R-squared of 0,37922. Particularly firm size, industry type and the amount of toxic release (adjusted for company size) had an influence on the amount of disclosure. The fact that toxic release has a significant positive influence on the amount of disclosure is evidence for an inverse link between environmental performance and environmental
disclosure (Patten, 2002). This is in contrast to previous studies, which consistently reported no link between reporting and performance.
In an extensive paper, Al-Tuwaijiri, Christensen and Hughes (2004) use a
simultaneous equations approach to determine the effects of environmental disclosure, environmental performance, and economic performance on each other. The important
difference to prior research, is their critical stance on the use of CEP data, for which they give three reasons. First, the CEP’s formulae to determine environmental performance of firms are not publicly available, thus impairing the ability to interpret these measures. Second, the weights assigned to environmental factors are not constantly assigned across different
industries, impairing the use of CEP data for use in cross-sectional sample spanning multiple industries, such as the study of Al-Tuwaijiri et al. Third, the CEP at least partly uses
environmental reports to determine environmental performance, which prohibits the use of the CEP index because the data of reporting and performance would no longer be independent of each other. Instead of the CEP index, Al Tuwaijiri et al. (2004) use a quantitative measure for environmental performance: the ratio of recycled toxic waste to total toxic waste generation. The higher the ratio of recycled waste versus total waste, the better the environmental performance of the firm. For a measure of environmental reporting a disclosure-scoring method was used, similar to that of Wiseman (1982). The measure uses four key indicators: amount of toxic waste generated and recycled, financial penalties for the violation of federal environmental laws, the clean-up responsibility of hazardous-waste sites, and reported oil or
chemical spills (Al-Tuwaijiri et al., 2004). The different indicators were then assigned
weights from 0 to 3: no disclosure to quantitative disclosure. Both environmental performance and environmental reporting measures were taken for a sample of 198 firms that produce at least one pound of toxic waste per $10.000,- of revenue generated, in order to ensure
environmental exposure. Again, the sample is considerably larger than in older research. The regression analysis performed on the dataset resulted in a statistically significant positive relationship between environmental performance and more quantifiable environmental
reporting, consistent with voluntary disclosure theory (Al-Tuwaijiri et al., 2004). This result is surprising, given that earlier research has either found no statistically significant relationship, or a significant negative relationship between reporting and performance.
Another paper by Clarkson, Li, Richardson and Vasvari (2008), however, confirms the
earlier findings of Al-Tuwaijiri et al. An important aim of the study was to develop a reliable proxy for environmental performance. One of the problems with measuring environmental performance is that the measure of performance can differ per industry. For example, toxic waste recycled as a percentage of total toxic waste generated is a good measure of
environmental performance for the petro-chemical industry, while total amount of waste water generated per level of sales is a better measure of environmental performance for the pulp and paper industry. Therefore, Clarkson et al. use a measure of environmental
performance that differs per industry studied. The study looks at five industries (pulp and paper, chemicals, oil and gas, metals and mining, and utilities) and alters between the method used by Al-Tuwaijiri et al. (ratio of recycled toxic waste to total toxic waste) and the ratio of TRI to sales. The TRI is the Toxic Release Inventory of the Environmental Protection Agency, an index that looks at toxic release and other waste management by public U.S. companies (Clarkson et al., 2008). Because the TRI is not only concerned with toxic release, but also with general waste management, it is a more appropriate measure of environmental performance for firms that have a more diverse impact on the environment, such as a combination of toxic waste and polluted water. The other important issue is developing a reliable measure for environmental reporting. According to Clarkson et al. (2008), voluntary disclosure theory predicts that good environmental performers aim to differentiate themselves from bad environmental performers. Therefore, good environmental performers will use measures of environmental performance that are hard to mimic. This means using objectively verifiable measures, meaning quantitative measures that are compared to industry
benchmarks, something that poor environmental performers will be reluctant to disclose (Clarkson et al., 2008). The study therefore uses a scoring model made up of 95 line items
developed by an expert in environmental reporting that reflect the guidelines for objective environmental reporting set by the Global Reporting Initiative (GRI). The GRI is an initiative started by the Coalition for Environmentally Responsible Economies and the United Nations Environmental Program, which tries to develop global standards for environmental reporting based on principles such as comparability, auditability and accuracy (Clarkson et al., 2008). The 95 line items consist of 79 hard measures of disclosure (meaning objectively measurable) and 16 soft measures (meaning more general), therefore favouring quantitative disclosure. The total sample size used in the study is 191 firms: 27 from pulp and paper, 63 from chemicals, 18 from oil and gas, 42 from metals and mining, and 41 from utilities. The data gathered for these firms were used in both inter-industry multiple regression and
intra-industry multiple regression analysis. The inter-intra-industry multiple regression analysis aimed to determine the existence of an overarching relationship between reporting and performance, while the intra-industry multiple regression analysis aimed to control for industry difference such as pollution propensity. According to Clarkson et al. (2008), the relationship between reporting and performance follows the predictions of economic voluntary disclosure theory. Better environmental performers will make more precise, quantitative information available than poor environmental performers, both on an inter-industry and intra-industry level.
What is important to note about the research by Clarkson et al. (2008) is that the
existence of a positive relationship between environmental reporting and performance is confirmed in line with voluntary disclosure theory, but that they do not shut the door on a role for legitimacy theory within this relationship. What they find, is that voluntary disclosure theory is correct in explaining how much quantitative, objective information is released if environmental performance is good, but legitimacy theory provides an explanation for what is being said if environmental performance is poor. This observation essentially harmonizes both earlier and contemporary research to provide an answer to the question whether or not a relation between environmental performance and environmental reporting exists, as well as offer an explanation for the inability of earlier researchers to find a statistically significant relationship between reporting and performance. This explanation is twofold: first of all, the amount of available data was not enough to do thorough research, and no strong enough distinction was made between “hard” (quantitative, objective, and comparable) and “soft” (vague and subjective) disclosure. This is a likely explanation for the feeling of both Wiseman (1982) and Rockness (1985) that despite statistical insignificance, an inverse relation between reporting and performance seemed likely.
relationship between reporting and performance. This could have three reasons: the
relationship simply did not exist, the relationship did not have enough time to mature given the young nature of environmental reporting at the time, or there were data related problems in the four papers. Clarkson et al. (2008) added to this a fourth possibility: flaws in the research design. The study by Fakret et al. (1996) showed that the relatively young nature of environmental reporting and the consecutive action by the EPA was unlikely to be the cause of the problem, given that even with updated data from the same set of firms used in earlier studies, the study failed to give statistically significant results. To compensate for the two research related problems (data and research design), later research both used different data sets and altered the design of the studies. Patten (2002) used a larger sample size and used a different study design in which he controlled for possible outside factors such as firm size. This resulted in a statistically significant, inverse relationship between reporting and performance. However, Patten only looked at the amount of disclosure, and made no distinction between “hard” and “soft” disclosure. Both papers by Al-Tuwaijiri et al. (2004) and to a larger degree Clarkson et al. (2008) do make a distinction between types of
disclosure. Both of these studies find a statistical significant positive relation between reporting and performance. All eight papers therefore help to come to a comprehensive answer to the question whether or not a relation exists. A relation does exist, but the relation is more subtle than predicted by theory: the total amount of disclosure is statistically
significant, yet not to an extreme extent, inversely related to environmental performance. However, the amount of objective environmental disclosure is used by good environmental performers to differentiate themselves from poor performers, and statistically significant positively related to environmental performance, as predicted by disclosure theory.
3.2 Other Factors of Influence
Although empirical evidence backs up the existence of a relation between reporting and performance as predicted by theory, to answer the question what the nature of this link is, it is important to determine what factors influence the relation between environmental reporting and environmental performance. Both theory and evidence suggest that factors like industry and political exposure are of influence on the amount of environmental information disclosed by firms. As said before, Li, Richardson and Thornton (1997) find that pollution propensity of an industry is of influence on the amount of environmental information that is disclosed by firms. Bewley and Li (2000) find the same, but also include political exposure as explaining
factor. This means, that the link between environmental reporting and environmental
performance might be strengthened or weakened depending on factors external to the relative environmental performance of the firm, at least on an inter-industry level.
The decision to disclose environmental and other non-financial information seems to
depend on three different types of factors: societal factors, industry factors, and individual
factors (Lee & Hutchison, 2005; Halme & Huse, 19972). Societal factors are factors that differ
on a country-to-country or regional basis, such as public policy pressure (Lee & Hutchison, 2005). A heightening in public policy pressure can arise when concerns arise around the environmental impact that firms have, such as through environmental accidents (Walden & Schwartz, 1997). Generally, a higher amount of public policy pressure will result in more release of information by firms (Walden & Schwartz, 1997). Publicity and legitimacy are societal matters that influence disclosure in the same manner (Lee & Hutchison, 2005). Not only individual societal factors, but also country differences as a whole have been shown to influence the extent to which firms disclose environmental and other social information. Chen and Bouvain (2009) tested the differences in Corporate Social Responsibility (CSR) reporting (of which environmental reporting is a part) via a number of factors, including size,
multinationality, industry, and country. They found that the country variable was significant at the 0,01 level in explaining both the quantity and the subjects of CSR reports, using four different statistical methods (Chen & Bouvain, 2009). This means that societies that are more concerned with the environment, will generally demand more public disclosure of
environmental and other social information by firms (Halme & Huse, 1997). Next to this, societal factors influence the kinds of information released by firms (Chen & Bouvain, 2009).
Not only societal factors, but also industry factors influence the decisions to disclose certain information (Lee & Hutchison, 2005). The important driver of the industry factor is industry characteristics (Lee & Hutchison, 2005). Generally, industries are either high profile or low profile, depending on characteristics such as industry visibility, political risk and level of competition (Hackston & Milne, 1996). Thus, the higher the visibility, political risk and level of competition, the more likely a firm is to be regard as being present in a high profile industry. Although classification is to some degree subjective and may differ per study, high profile industries typically include extractive industries (mining), petro-chemical industries, and pulp and paper (Hackston & Milne, 1996). Low profile industries, on the other hand, include industries such as food, health and personal products, and hotels (Hackston & Milne,
2 In Halme & Huse, the three types of factors are named differently, but the concepts are the same.
1996). It is easy to see how the concept of pollution propensity ties into the distinction between high profile and low profile industries. High profile industries tend to be more pollution intensive. Pollution propensity has been sighted as an important determinant in the amount of environmental information disclosed (Li, Richardson & Thornton, 1997; Bewley & Li, 2000). Empirical studies have mostly centred around industries with a high pollution propensity, to assure at least some measurement of environmental performance such the ratio of toxic waste recycled to total toxic waste generated. Also, the concept of industry visibility might have an influence on the particular industries that are studied, since the environmental impact of heavy polluting industries like petro-chemicals is subject to more debate than for instance that of the service industry.
Individual, or firm-specific factors also play a role in determining the amount and
content of environmental and other social disclosure (Lee & Hutchison, 2005). Culture and ethical attitudes towards sustainability and the environment are considered firm specific factors. Culture in this context can mean both the culture within a company, as well as the culture that a firm operates in, and thus ties into the concept of societal differences. A study by Mathews and Reynolds (2001) on the management of firms in the timber industry in the Nordic countries (Finland, Norway and Sweden) and the U.S. used the answers to a large number of questions and statements to find the impact of culture on environmental reporting. They found that the differences between the U.S. and the Nordic countries were significant, while there were no significant differences in responses between Finland, Norway and Sweden. This proves the significant impact general culture can have (rather than individual country variables) on the disclosure of environmental information. Also, they found that the attitude of the management of individual firms was also of impact, showing that firm-specific culture also impacts environmental reporting and disclosure. Cormier & Magnan (2003) also find that firm specific factors like ownership concentration, financial condition, and firm size influence environmental disclosure. Ownership concentration does so negatively: the more concentrated the ownership, the less environmental information the firm will disclose. Financial condition and firm size, on the other hand, influences environmental disclosure positively: the better the financial condition of a firm, the more it will disclose. Firm size is often used as a control variable in empirical studies on the relationship between
environmental reporting and environmental performance for this reason (Neu, Warsame, Pedwell, 1998; Cho & Patten, 2007).
Thus, many factors other than performance influence the decision to disclose or not
course, is that causality will move from environmental performance to environmental reporting. Although none of the examined literature explicitly mentions the causal link, the causality is often hinted towards in the literature (Gray, Kouhy & Lavers, 1995; Bewley & Li, 2000). This link is logical, since reporting and disclosure should, at least in theory, give a reflection of the performance of a firm. However, in both settings where environmental reporting is mostly voluntary, as well as settings where environmental reporting is mandatory to a larger degree, different environmental reports have been shown to be difficult to compare (Lee & Hutchison, 2005; Frost, 2007). Because of this, based on both the existing literary concepts of the relation between environmental performance and environmental disclosure, and the many other factors of influence, the relation might work as shown in figure 1.
This model shows the causal relationship between environmental performance and
environmental reporting and the implications poor performance and good performance have for environmental reporting. Environmental performance can, for the sake of simplicity, be either good or poor, while environmental reporting has two main aspects. The ‘amount’ aspect of environmental reporting can be both the size of individual reports or disclosure, as well as the frequency with which reports are issued. The ‘content’ aspect represents the overall objectivity and comparability of environmental information in the reports or disclosure. Poor performance will affect amount positively (more information will be disclosed), but will affect the content of the reports negatively, meaning less objective and
Fig. 1. The relationship between environmental performance and environmental reporting,
and the influence of outside factors on this relationship.
Environmental Performance Poor performance Good performance Environmental Reporting Amount Content
+
+
-
Factors of Influence - Societal factors - Industry factors - Individual factorscomparable information will be reported or disclosed. Good performance, on the other hand, will have a positive influence on the content of environmental reports and disclosure. The three outside factors distort this relationship depending on where an organisation is located, in what industry it is active and on individual firm factors.
4 Discussion
Both voluntary disclosure theory and legitimacy theory have their place in explaining the relation between environmental performance and environmental reporting ceteris paribus. Good performers will try to distinguish themselves from poor performers by releasing information that is harder to mimic, as predicted by voluntary disclosure theory, while poor performers will try to legitimize their activities by releasing larger quantities of information that is hard to verify or contains many promises and empty statements.
However, the extent to which this relationship actually works this way in practice is
highly circumstantial. Other factors influence the decision what parts of information about performance will actually be disclosed. These factors of influence are threefold. First, societal factors will depend on the society at large in which a firm operates: a European firm will disclose and report differently from Asian or American firms in the same way that
international firms will disclose and report differently from national or regional ones. Second, industry factors depend on the industry that a firm operates in. Firms will weigh reporting decisions depending on factors that are tied into political and social scrutiny depending on the industry they operate in. Third, individual firm factors will play a role in the decision process. This leaves room for individuals, mainly in higher firm management, to influence
environmental reporting.
A good example of how big a role these non-performance factors can play in the
environmental reporting process, is a study by Clarkson, Overell and Chapple (2011), which replicates the study of Clarkson et al. (2008). However, while the earlier study used a sample of large international firms, the replication uses a smaller sample of 51 firms exclusively from Australia. Although both studies use the same methods and even both have partly the same participating researchers, the results are vastly different. The earlier, bigger study found that better environmental performers disclosed more and better environmental information. The Australian study, on the other hand, found statistically significant results suggesting that poor environmental performers consistently reported more and better (objective and comparable)
environmental information. The fact that these studies, which are for all intents and purposes identical, save for the data set used, find opposite results, raises questions about how big a role non-performance factors play, and consequently to what extent current environmental reports can be trusted.
The papers analysed in this literature review result in the conclusion mentioned earlier,
ceteris paribus, but it is very important to note that this conclusion comes from a consensus
based on very conflicting evidence. If a metaphor were to be used to describe the way in which the empirical evidence results in the aforementioned conclusion, it could be said that the evidence leads to the conclusion in much the same way that the average of one and ten is five and a half. This does not discredit the conclusion any more than the use of the numbers one and ten would change the truth of the average being five and a half, but it does show that variability in the relation between reporting and performance is very high.
This variability is not only observed in this literature review. Neu et al. (1998) also
observed large variations in reporting and disclosure based on the relative power of different stakeholder groups. If power plays such a defining role in determining what is disclosed and what is not, the problem with current environmental reporting may be sought in its voluntary nature. Voluntary reports are not only very different from each other and for this reason hard to compare, but voluntary initiatives also do not tend to produces consistent and widespread systematic practice of reporting (Frost, 2007; Gray, 2001; Lee & Hutchison, 2005). This failure to produce systematic practice that creates accurate environmental reports is by no means surprising. As stated earlier, organisations will try to legitimize actions that are considered to be harmful to their social contract with society. Thus, if environmental reporting is of a voluntary nature, completely truthful reporting with regards to poor
environmental performance will not occur, because an organisation will not chose to harm its social contract with society when presented with a choice.
Oppositely, by its very nature, useful and high impact disclosure of environmental and
other social information should hurt: it should give rise to difficulties, because the current form of organisations is not environmentally friendly or sustainable, and the very goal of social accounting is to show what elements of the reporting entity must change in order to change this (Gray, 2001). Of course, voluntary disclosure is highly unlikely to maximize or even achieve this effect (Gray, 2001). Because voluntary disclosure is unlikely to achieve the effect of confronting an organisation and its stakeholders with the reality of its environmental performance, it could be very beneficial to move towards a system that is of a more
For these reasons, future research should focus on determining what parts of environmental information are most impacted by the distorting factors. Based on this information, standards should be developed to lessen or even eliminate the effect of these distorting factors. Also, the issues surrounding the overall trustworthiness of voluntary environmental reporting in its current form should be resolved: what should be changed to environmental accounting within the possibilities of the current social accounting framework to let it better achieve its goal, while at the same time preserving the benefits of voluntary disclosure (such as the signalling effect predicted by voluntary disclosure theory). However, searching for changes within the accounting framework might not be enough: as said before, the power of different stakeholder groups is a determining factor in the individual variation of reporting. The power of stakeholders is not surprising, since the current accounting
framework emphasizes the usability of information. However, since information flows from performance to reporting, making this information usable might actually distort the
performance-reporting relationship. For this reason, it is important to not only search for possible changes within the current accounting framework, but also to remain critical of the framework itself.
5 Conclusion
This literature review has tried to find out what the nature of the link is between
environmental performance and environmental reporting. The workings of this association have been theorized from two competing perspectives: the economic perspective on the one hand, and the socio-political perspective on the other. The economic perspective uses voluntary disclosure theory to explain the relation between performance and reporting. This theory predicts that good environmental performers will try to disclose as much
environmental information as possible as a signal to the market that it is unlikely to face high environmental liabilities in the future compared to poor environmental performers. For the same reason, poor performers will try to withhold as much information as possible. The socio-political perspective uses legitimacy theory to explain the relation between performance and reporting. This theory predicts the opposite voluntary disclosure theory: organisations work on the basis of a social contract with society, which requires organisations to create social value for society. If the creation of social value by an organisation is questioned, the organisation will try to legitimize its actions by disclosing more information to change the
public opinion.
The scope of this paper is to come to a comprehensive idea coming from the contexts
of these two theories combined with empirical papers as to how theory and practice come together to explain the nature of the link between environmental performance and
environmental reporting. Empirical evidence suggests that both voluntary disclosure theory and legitimacy theory have their place in explaining the relation between performance and reporting. Legitimacy theory explains the amount of reporting and disclosure by
organisations, by correctly predicting that poor environmental performers will disclose more information than good performers. Voluntary disclosure theory ads to this, by correctly predicting the content of the reports from good and poor performers: good performers will try to differentiate their reports from bad performers by disclosing information that is harder to mimic, such as exact and easily comparable figures, and performance relative to industry benchmarks.
To find the full scope of the nature of the relation between environmental performance
and environmental reporting, however, the relation has to be put into context. The underlying reasoning is that reporting is the result of performance, ceteris paribus, but that reporting is not homogeneous in comparable situations due to factors outside the performance-reporting relationship. These outside factors are of three types: societal factors, industry factors, and individual factors. Societal factors change the performance-reporting relationship based on geographical and cultural location of the firm: they can change the relationship based on whether or not a firm operates internationally, nationally or locally, as well as based on where a firm originates from. Industry factors change the performance-reporting relationship based on the industry a firm operates in: operations can take place in either a high profile or low profile industry, which changes factors such as pollution propensity and political pressure. Individual factors relate to an organisation’s culture and management. Even between organisations in the same countries and industries, reporting can be different based on the varied choices individuals within a firm make. All these factors in different categories can have a multitude of different effects on the performance-reporting relationship.
The findings in this review lead to the performance-reporting relationship as visualised
in figure 1. Because of the large amount of distorting factors that affect the relationship between performance and reporting, the reliability of environmental reporting and disclosure has been taken into question. Research will have to be done to determine the precise impact, in both area and scope, each of the disturbing factors has on the performance-reporting relationship. Furthermore, the reliability of current environmental reporting remains unclear,
as well as what needs to be changed to minimize the impact of factors that should not have an influence on the performance-reporting relationship, both within and outside the current (social) accounting framework. For these reasons, it is vital to remain critical of the way in which environmental information is reported and disclosed.
References
Al-Tuwaijiri, S. A. , Christensen, T. E. & Hughes, K. E. (2004). The relations among
environmental disclosure, environmental performance, and economic performance: a simultaneous equations approach. Accounting, Organizations and Society, 29(1), 447-471.
Bewley, K. , & Li, Y. (2000). Disclosure of environmental information by Canadian manufacturing companies: A Voluntary Disclosure Perspective. Advances in
Environmental Accounting and Management, 1(1), 201–226.
Chen, S. , Bouvain, P. (2009). Is corporate responsibility converging? A comparison of corporate responsibility reporting in the USA, UK, Australia and Germany. Journal of
Business Ethics, 87(1), 299-317.
Cho, C. H. , Patten, D. M. (2007). The role of environmental disclosures as tools of
legitimacy: a research note. Accounting, Organizations and Society, 32(7), 639-647. Clarkson, P. M. , Li, Y. , Richardson, G. D. , Vasvari, F. P. (2008). Revisiting the relation
between environmental performance and environmental disclosure: An empirical analysis. Accounting, Organisations and Society, 33(1), 303-327.
Clarkson, P. M., Overell, M. B. , Chapple, L. (2011). Environmental reporting and its relation to corporate environmental performance. ABACUS, 47(1), 27-60.
Cormier, D. , Magnan, M. (2003). Environmental reporting management: a continental European perspective. Journal of Accounting and Public Policy, 22(1), 43-62. Dye, R. A. (1985). Disclosure of nonproprietary information. Journal of Accounting
Research, 23(1), 123-145.
Fekrat, M. A. , Inclan, C. , Petroni, D. (1996). Corporate environmental disclosures: competitive disclosure hypothesis using 1991 annual report data. The International
Journal of Accounting, 31(2), 175-195.
Frost, G. R. (2007). The introduction of mandatory environmental reporting guidelines: Australian evidence. ABACUS, 43(2), 190-216.
Freedman, M., & Jaggi, B. (1982). Pollution disclosure, pollution performance and economic performance. Omega, 10(2),167–176.
Gray, R. , Kouhy, R. Lavers, S. (1995). Corporate social and environmental reporting: a review of the literature and a longitudinal study of UK disclosure. Accounting,
Auditing & Accountability Journal, 8(2), 47-77.
Gray, R. (2001). Thirty years of social accounting, reporting and auditing: what (if anything) have we learnt? Business Ethics: A European Review, 10(1), 9-15.
Hackston, D. , Milne, M. J. (1996). Some determinants of social and environmental
disclosures in New Zealand companies. Accouting, Auditing & Accountability Journal,
9(1), 77-108.
Halme, M. , Huse, M. (1997). The influence of corporate governance, industry, and country factors environmental reporting. Scandinavian Journal of Management, 13(2), 137-157.
Ingram, R. W. & Frazier, K. (1980). Environmental performance and corporate disclosure.
Journal of Accounting Research, 18(2), 614-622.
Lee, T. M. , Hutchison, P. D. (2005). The decision to disclose environmental information: a research review and agenda. Advances in Accounting, 21(1), 83-111.
Li, Y. , Richardson, G. D. , Thornton, D. B. (1997). Corporate disclosure of environmental liability information: theory and evidence. Contemporary Accounting Research, 14(3), 435-474.
Mathews, M. R., Reynolds, M. A. (2001). Cultural relativity and accounting for sustainability: a research note. Accounting Forum, 25(1), 79–88.
Neu, D. , Warsame, H. , Pedwell, K. (1998). Managing public impressions: environmental disclosures in annual reports. Accounting, Organizations and Society, 23(3), 265-282. Patten, D. M. (1992). Intra-industry environmental disclosures in response to the Alaskan oil
spill: a note on legitimacy theory. Accounting, Organizations and Society, 17(5), 471-475.
Patten, D. M. (2002). The relation between environmental performance and environmental disclosure: a research note. Accounting, Organisations and Society, 27(1), 763-773. Rockness, J. W. (1985). An assessment of the relationship between US corporate
environmental performance and disclosure. Journal of Business Finance &
Accounting, 12(3), 339-354.
Sinclair, D. (1997). Self-regulation versus command and control? Beyond false dichotomies.
Law & Policy, 19(1), 529-556.
Suchman, M. C. (1995). Managing legitimacy: strategic and institutional approaches. The
Academy of Management Review, 20(3), 571-610.
Verrechia, R. E. (1983). Discretionary Disclosure. Journal of Accounting and Economics,
5(1), 179-194.
Walden, W. D. , Schwartz, B. N. (1997). Environmental disclosure and public policy pressure. Journal of Accounting and Public Policy, 16(1), 125-154.
Wiseman, J. (1982). An evaluation of environmental disclosure made in corporate annual reports. Accounting, Organizations and Society, 7(1), 53–63.
