Introduction

Firms are increasingly held accountable for the impact of their activities on society (Hahn 2012). However, it is often difficult for external stakeholders to assess the actual sustainability performance of a company. To reduce information asymmetries between companies and their stakeholders, firms are expected to communicate their behavior and to comply with the norms of corporate sustainability transparency (similar, Philippe and Durand 2011). Accordingly, sustainability and corporate social responsibility (CSR) reportingFootnote 1 has become a standard topic in management and accounting (Hahn and Kühnen 2013). In 2011, 95 % of the 250 largest global companies published such a report (KPMG 2011) answering increasing stakeholder pressure to explain their business conduct (Haniffa and Cooke 2005; Fortanier et al. 2011; Gallo and Jones Christensen 2011). In general, transparent disclosure can enhance trust with different stakeholders whereas “a lack of governance transparency represents a source of information risk for external stakeholders.” (Archambeault et al. 2008, p. 377). However, prior studies identify an abundance of positive information in corporate sustainability reports and a lack of negative voluntary disclosures (e.g., Lougee and Wallace 2008; Holder-Webb et al. 2009; Deegan and Rankin 1996). Companies often seem to try to use overly positive, whitewashed sustainability reports merely for PR purposes as a tool for gaining and improving a company’s reputation and legitimacy (Higgins and Walker 2012; Castelló and Lozano 2011; Deegan 2002; Hooghiemstra 2000; Cho et al. 2010). However, the usefulness of such reports for reducing information asymmetries and improving transparency can be questioned since they do not help foster accountability and paint a true and fair viewFootnote 2 of a company’s non-financial performance.

Regulation seems to be of little help in this aspect since there is only very limited regulatory guidance on sustainability reporting in most countries (Manetti and Becatti 2009; Deegan 2004). Projects such as the Global Reporting Initiative (GRI), a multi-stakeholder forum dedicated to providing guidance on sustainability reporting, try to overcome this gap (GRI 2011). Today, the GRI guidelines are regarded as “the de facto global standard” (KPMG 2011, p. 20; emphasis in original) for voluntary sustainability disclosure. They challenge companies to provide transparent, complete, and balanced reports. This explicitly includes positive and negative corporate contributions to sustainability (GRI 2011). However, the effects of disclosing negative sustainability-related incidents in particular are largely neglected in scholarly research (Hahn and Kühnen 2013; for notable exceptions see, e.g., Chan and Milne 1999; Coram et al. 2009; Reimsbach and Hahn 2013). Whereas various sustainability-related studies, for example, in environmental reporting, have examined biases in the quantity of disclosure (e.g., Neu et al. 1998; Patten 1992; Blacconiere and Patten 1994; Clarkson et al. 2008) or in the choice of thematic content of disclosure (e.g., Cho and Patten 2007; Hughes et al. 2001; Patten 2002), the dimension of legitimizing strategies regarding language and rhetoric in sustainability reports has not yet been thoroughly examined.

This omission is serious since illuminating the impact of negative disclosure is highly relevant from a managerial perspective: Disclosing “negative aspects” can endanger corporate legitimacy if negative ecological and social consequences of corporate activity are not perceived as being in line with societal norms, values, and beliefs by the company’s stakeholders (Chan and Milne 1999; Deegan and Rankin 1996). Given that sustainability aspects represent value-relevant information to investors (Orlitzky et al. 2003; Dhaliwal et al. 2012), negative sustainability performance can thus translate into negative financial performance and increased stock market risk (Bansal and Clelland 2004). When not reporting negative aspects, on the other hand, companies might encounter increased skepticism regarding the reliability of their sustainability disclosure with potentially negative consequences for their accountability. Actively disclosing negative aspects of sustainability performance might even be regarded as a positive signal in terms of actively managing risk, thus helping to avoid future issues.

The focus of this study therefore was to analyze the communicative legitimation strategies companies use to report “negative aspects,” i.e., negative ecological and social impact caused by corporate activity. In general, we aim to shed light on reporting behavior related to negative sustainability-related aspects in voluntary GRI-oriented reports. We specifically strive to answer the following questions:

  1. (1)

    How do companies legitimize negative incidents in their sustainability reports, and how can the respective reporting strategies be judged in light of a “true and fair view” in sustainability reporting?

  2. (2)

    Which implications stem from the findings for further advancing the GRI guidelines?

Our study thus contributes to the growing body of literature on sustainability and CSR reporting (Hahn and Kühnen 2013). Shedding light on different types of communicative disclosure strategies for reporting negative incidents and discussing their potential for providing a true and fair view of a company’s non-financial performance enables us to develop implications for managers who want to provide a complete sustainability report in accordance with the GRI guidelines. Furthermore, we aim at providing meaningful guidance for regulators and policy makers trying to understand the nature of corporate reporting behavior and the opportunities and constraints of voluntary corporate disclosure. To achieve this, we conducted a qualitative content analysis of current GRI-oriented sustainability reports from companies listed on the US Dow Jones Industrial Average Index and the German DAX Index. By evaluating reports from companies listed on these two indices we aimed to achieve a broad representation including companies from various industries and corporate governance systems enabling a comprehensive overview of legitimation strategies connected to negative aspects.

The remainder of this paper is structured as follows. In “Theoretical Background” section, we give an overview of theoretical peculiarities of disclosing “negative aspects” as postulated by the GRI guidelines. Here, we refer to economics-based disclosure theories (especially voluntary disclosure theory and signaling theory) and to socio-political theories of disclosure (especially legitimacy theory) before outlining prior research on the relevance of corporate legitimacy and legitimizing strategies. In “Methodology” section, we describe the method and data of our study. In “Findings on Strategies to Legitimate Negative Aspects in Sustainability Reporting” section, we illustrate our findings regarding different legitimation strategies used in the sustainability reports before we discuss these findings focusing on different potential effects of these strategies in “Discussion” section. Here we also outline implications for the further development of the GRI guidelines. Finally, we present conclusions, limitations, and avenues for future research in “Conclusion” section.

Theoretical Background

Economics-Based Disclosure Theories, Sustainability Reporting and Negative Incidents in GRI-Related Disclosure

According to economics-based voluntary disclosure theory (Dye 1985; Lang and Lundholm 1993; Verrecchia 1983), companies voluntarily disclose information to reduce information asymmetries between managers and outside stakeholders to convey their (good) performance. Similarly, signaling theory suggests that in cases of information asymmetry, the better informed party tries to credibly convey information about itself to the less informed party (Spence 1973; Connelly et al. 2010) to reduce this asymmetry. The sustainability performance of a company can be regarded as asymmetric information because it is difficult for parties outside the company to gain credible information on sustainability aspects. Companies can reduce this information asymmetry by proactively reporting on their sustainability-related activities. In doing so, they might wish, for example, to ensure success in the capital markets because an increasing number of investors and rating agencies (which, for example, also influence the composition of indices such as the Dow Jones Sustainability Index or the FTSE4Good) integrate sustainability issues into their decisions (see, e.g., Global Sustainable Investment Alliance 2012).

Thus, at first sight, there seems to be few incentives to report negative aspects of a company’s sustainability performance, especially as sustainability reporting is still usually voluntary and largely unregulated in most countries (Manetti 2011; Deegan 2002). However, with regard to voluntary disclosure and signaling theory reporting, disclosing negative incidents might enhance trust in the company’s information policy and signal a proactive and honest disclosure, whereas not reporting negative aspects might conversely lead to speculation that the report is overly positive, whitewashed, and hence not a reliable source of information. In line with this thinking, Reimsbach and Hahn (2013) recently showed that proactive corporate disclosure of negative sustainability-related aspects is not necessarily regarded as a negative signal by non-professional investors and that such disclosure might even be regarded as a risk mitigation mitigation tool. Furthermore, an increasing number of sustainability reports are externally assured by independent auditors (KPMG 2011). This increases the pressure to provide a balanced and true picture of the company’s sustainability performance, including negative aspects.

Nevertheless, research shows that companies tend to prefer to emphasize positive information in voluntary sustainability reporting (Lougee and Wallace 2008; Holder-Webb et al. 2009). Thus, reports seem to be prone to ambiguity and arbitrariness due to the lack of regulatory safeguards. The GRI tries to overcome this issue by providing guidelines for reporting including negative aspects. It introduced the principle of “balance” to challenge companies to report on positive and negative contributions “to enable a reasoned assessment of overall performance” (GRI 2011, p. 13). However, the GRI does not provide an explicit outline or understanding of what constitutes a negative contribution or impact. We therefore propose the following preliminary characterization of negative aspects regarding sustainability reporting to identify these issues in company reports: Negative aspects in sustainability reporting include any corporate statement referring to factual and/or potential corporate conduct that had or has a (potentially) negative impact on the realization of sustainability. Since this characterization is rather broad, we refer to the GRI “performance indicators” to illustrate examples of such negative aspects. In terms of the ecological dimension of sustainability, negative aspects stemming from corporate operations are associated with the pollution of the ecological environment (air, soil, water, etc.) in terms of the production of toxic waste, the release of harmful emissions or hazardous substances, and so on. When referring to the social dimension, negative aspects could be, for example, work-related accidents, bribery or fraud, breaches of data privacy, and incidents of child labor or discrimination. When turning to negative economic (especially financial) aspects, regulatory guidelines (e.g., US GAAP and IFRS) provide certain rules for financial reporting that potentially influences legitimation strategies so that they might not be comparable to the ones we discuss with regard to environmental and social issues. Consequentially, negative aspects of financial performance were not analyzed. Instead, we assume them to be covered in traditional (mandatory) financial reporting which was not part of our study.

Socio-Political Theories of Disclosure and Sustainability Reporting

According to social-political theories of disclosure (e.g., Gray et al. 1995), voluntary disclosure is not merely used to inform capital market participants but also to manage impressions. The aim is to help companies face social and political pressure exerted by non-market stakeholders, such as NGOs, policymakers, or the media. The disclosure of certain sustainability information can be an instrument for generating favorable impressions of an organization’s sustainability performance, thus preserving organizational legitimacy (Bansal and Clelland 2004; Bebbington and Larrinage-González 2008; Brown and Deegan 1998; Deegan 2002; Deegan et al. 2002; Hooghiemstra 2000; O’Donovan 2002). According to legitimacy theory, legitimacy can be characterized as “a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions” (Suchman 1995, p. 574). Achieving and preserving legitimacy is relevant because companies act in extensive exchange relationships with different stakeholder groups and depend upon resources from their environment (e.g., from public authorities, residents, activists, customers, or employees), and these stakeholders have increasing expectations regarding how companies conduct business.

In a very general sense, legitimacy theory suggests that no organization has an inherent right to exist and that society confers legitimacy upon each organization (Deegan 2002). This idea is transferred to two main viewpoints in legitimacy theory: the institutional perspective and the strategic perspective (Suchman 1995). In the institutional tradition (e.g., DiMaggio and Powell 1983; Meyer and Rowan 1977; Zucker 1987), legitimacy is regarded as a set of constitutive beliefs. For example, institutional dynamics which stem from the industry environment in which a firm operates, generate external pressure on companies to behave in a way that is perceived legitimate by external institutions. In this sense, companies, at most, can have only a very limited influence on society’s perception. The strategic tradition (e.g., Ashforth and Gibbs 1990; Dowling and Pfeffer 1975; Pfeffer and Salancik 2003) adopts a more managerial stance, assuming that the organization can take actions to influence its legitimacy.

The perceived legitimacy by different stakeholders helps to ensure access to, for example, resources and markets and thus leads to the continuity of the company (Palazzo and Scherer 2006). However, the perceived legitimacy is potentially threatened by negative incidents (or those incidents that are perceived to have a negative connotation) associated with the respective legitimacy-seeking entity. On the one hand, actively disclosing negative aspects could put corporate legitimacy at risk if the negative aspects are not in line with societies’ expectations of corporate behavior and performance. For example, Bansal and Clelland (2004) showed that companies disclosing environmental sustainability performance experience higher unsystematic risks. On the other hand, not disclosing negative incidents could lead to a public backlash if the incidents are uncovered by independent third parties, whistleblowers, or entities and threaten the corporate legitimacy (Reimsbach and Hahn 2013; Våland and Heide 2005). Accordingly, Bansal and Clelland (2004) also showed that unsystematic risks seem to be especially pronounced when the negative aspects are uncovered by external third parties. In the 1990s, for example, Nike’s reputation was hit hard following discussions of the poor working conditions in the company’s value chain (Zadek 2004), and numerous oil companies were repeatedly publicly accused of environmentally and socially harmful business practices especially in developing countries (Boele et al. 2001). These examples illustrate that—metaphorically speaking—society can revoke a company’s social “license to operate” and thus put the company’s economic success at risk if it perceives that the organization is not operating in an acceptable way.

This also indicates that an organization’s legitimacy is not an objective fact but lies in the eye of the beholder (i.e., it is subject to the perception of those who convey legitimacy). There might therefore be a gap between actual corporate conduct (and disclosure about this conduct) and stakeholder perception. This opens up opportunities for corporate practices influencing the perceived legitimacy and helping companies to ensure the supply of this particular intangible resource (Deegan 2002). Accordingly, legitimation strategies can be characterized as strategies intended to secure legitimacy as a valuable resource by using specific communication techniques in corporate disclosure (e.g., Dowling and Pfeffer 1975; Ashforth and Gibbs 1990; Suchman 1995; Hooghiemstra 2000). This view hence follows a strategic perspective of legitimacy.

Corporate Legitimation Strategies and Negative Incidents

When specifically looking at the occurrence of negative incidents, legitimation strategies aim at restoring legitimacy (when the incident has already caused a loss of legitimacy) or at proactively preserving legitimacy (in cases where the incident has, for example, not been made public before the company itself discloses it). In the latter sense, the crisis communication literature suggests that early and transparent communication of negative incidents can help minimize losses of credibility (e.g., Allen and Caillouet 1994; Ulmer and Sellnow 2000). Table 1 offers an overview of the most prominent categorizations of legitimation strategies, with a specific reference to negative occurrences in corporate behavior.

Table 1 Conceptualizations of legitimation strategies specifically referring to negative incidents
Full size table

Benoit (1997), for example, proposes a well-cited variety of legitimation strategies in crisis communication aiming to repair a bruised corporate image. Similarly, Suchman’s (1995) framework of organizational legitimacy contains a subset of strategies aiming at repairing legitimacy when negative publicity has occurred. Furthermore, Cho (2009) identifies three generic legitimation approaches in response to environmental disasters. The respective strategies introduced by these scholars are, however, primarily connected with those negative aspects that already caused damage to organizational legitimacy so that they are usually not suitable for use for voluntary corporate disclosure of negative aspects as in most GRI-oriented sustainability reports.

Other frameworks offer a more proactive set of legitimation attempts. Under the label of impression management, Merkl-Davies and Brennan (2007) discuss concealment and attribution as the main approaches for disclosing negative incidents in corporate narratives. However, neither of these strategies is particularly suited for analyzing balanced reporting of negative aspects in sustainability disclosure because they all aim at drawing attention away from the respective negative fact. A less biased set of legitimation strategies is proposed in the conceptual paper by Lindblom (2010), who—based on hypothetical examples—discusses the communication of organizational changes, the attempt to change stakeholder perceptions, the association with symbols having high legitimacy, and adjustments in societal expectations as means for strategic disclosure. Our study ties into this thinking by providing empirically grounded insights into those strategies connected to corporate (sustainability) reporting and to the proactive disclosure of negative aspects as required in balanced sustainability reporting according to the GRI. As discussed in the introduction, non-financial disclosure has proliferated due to the increasing pressure and scrutiny of various stakeholders. Companies use non-financial reporting to explain their business in a positive light and thus respond to this scrutiny (Haniffa and Cooke 2005; Fortanier et al. 2011; Gallo and Jones Christensen 2011). Research shows that the disclosure of certain sustainability information can help to preserve organizational legitimacy by generating beneficial impressions of the company’s sustainability performance (Bansal and Clelland 2004; Bebbington and Larrinage-González 2008; Brown and Deegan 1998; Deegan 2002; Deegan et al. 2002; Hooghiemstra 2000; O’Donovan 2002). However, very few papers specifically investigate negative aspects in voluntary reporting which might be because positive information prevails in sustainability reports.

Some interesting findings in this regard instead stem from traditional accounting research. Here, several studies indicate that companies use rhetorical devices and prefer to emphasize the positive aspects of their performance while, for example, blaming the external environment for bad news in order to positively influence stakeholder perceptions (e.g., Clatworthy and Jones 2003; Smith and Taffler 2000; Yuthas et al. 2002; Sydserff and Weetman 2002). Merkl-Davies and Brennan’s (2007) framework was based on extensive review of research on accounting narrative disclosures. According to this framework, impression management in reporting practices focuses on either the quantity or on the quality (e.g., language and verbal tone, attribution) of the information. The focus of our study is on the latter. Only recently, Higgins and Walker (2012) studied persuasive strategies in sustainability reports. Similarly, Castelló and Lozano (2011) analyzed sustainability reports and identified different types of rhetoric to gain legitimacy. Both studies, however, follow the general trend of investigating legitimacy related to the mainly positive information dominant in corporate sustainability disclosure. In a rare attempt to go beyond positive information disclosure in sustainability-related reporting, Cho et al. (2010) found that disclosure by low-environmental performers is coined by more “optimism” and less “certainty” than the disclosure by those companies with a better performance. Regarding language and rhetoric, Neu et al. (1998) showed that in annual reports, narrative disclosures of environmental aspects are the preferred form of reporting since they can be specifically targeted to manage public impression. Our study digs deeper into these aspects by illuminating corporate disclosure strategies related to negative aspects. We specifically focus on language and tone of sustainability reporting since these aspects have been largely neglected thus far.

Methodology

Research specifically focused on systematically analyzing the strategies used by companies to disclose negative sustainability-related aspects is very scarce. We therefore decided to apply an exploratory research design to illuminate modes of disclosing negative incidents in corporate sustainability reports. Thus far, there is no holistic categorization of legitimation strategies referring to voluntary disclosure of negative incidents. Thus, we decided to explore different legitimation strategies by means of an inductive research approach using qualitative content analysis to uncover the ways in which negative incidents are disclosed and rhetorically backed in corporate non-financial reports. The proposed classification could then later be used as a framework for further (quantitative) research.

Sample Selection and Material

Our sample consisted of sustainability (or CSR or Corporate Citizenship) reports of companies listed on the US Dow Jones Industrial Average Index and the German DAX Index (as of March 1st, 2012). We included the latest available reports (i.e., stand-alone sustainability reports or integrated reports) at the time of analysis in mid-2012 that followed the GRI guidelines. This encompassed the PDF-versions of reports covering the years 2011 or 2010. We opted to include companies from two broad indices to achieve a balanced representation of various industries. Furthermore, including reports from companies based in a more shareholder-oriented Anglo American context and from companies based in a more stakeholder-oriented Continental European context with potentially different approaches to sustainability management and related disclosure (Matten and Moon 2008) should also add depth in the data at hand.

In sum, we analyzed 40 reports (19 from Dow Jones Industrial Average companies and 21 from DAX companies; see Table 3 in Appendix). This total is less than the 60 companies included in both indices. The companies not included in our analysis did not publish a sustainability-related (or integrated) report, did not publish a report according to the GRI guidelines, give website-only version of their reports or merely published some form of a short overview in the analyzed period. Overall, this led to a data sample of about 4,000 pages in total. The first research aim was to explore how companies legitimize negative incidents in their sustainability reports. The data we used enabled us to reach this aim. Other data, for example, interviews, which are typically used in qualitative case studies, would not have provided us with much added value because we specifically aimed to uncover the disclosure strategies used in corporate sustainability reports and not, for example, the rationale underlying the use of the respective strategies. Other data sources, such as external reports on the companies, might have the potential to yield interesting insights beyond illuminating corporate legitimation strategies and the rhetoric adopted. However, such external resources cannot be used for triangulation purposes in our specific case because they would not mirror corporate rhetoric but outsider rhetoric, which was beyond the scope of this research.

Data Analysis

We used qualitative content analysis (see Duriau et al. 2007; Mayring 2000, 2010) on the data extracted from the sustainability reports. Due to the exploratory nature, we used an inductive, interpretative approach for analyzing the data. Following Mayring’s (2000; 2010) step model of inductive category building we first defined our research question (see “Introduction” and “Theoretical Background” sections) and object (i.e., the data; see “Sample Selection and Material” section). Second, we chose “negative aspects” as relevant selection criterion to determine which parts of the material were relevant for our analysis. To identify these aspects we independently searched for any passages in the reports that might be related to negative incidents or aspects using the software tool MAXQDA. This included the extensive search for generic keywords such as “negative,” “incident,” “accident,” “adverse,” “harm,” “risk,” and “conflict”. The keywords were previously identified and discussed by the two main researchers with further input from two PhD students to cover a wide range of synonyms providing a broad variety of possible associations with negative incidents in general. Furthermore, we searched for concrete topics such as “corruption,” “spill,” “child labor,” “discrimination,” and potential synonyms by referring to the respective GRI indicators, which include such incidents that usually bear a negative connotation (GRI 2011). It became clear that negative aspects are reported in brief form so that we used single sentences (and sometimes short paragraphs) as analytical units. This allowed us to identify those passages in the reports that might be associated with negative aspects of corporate sustainability performance.

In the third step, we went through the material and marked the identified spots to evaluate whether they covered negative aspects of sustainability performance as characterized above. Differing assessments by the two coders were discussed on a case-by-case basis. We then independently scrutinized the identified statements and passages for recurring patterns in disclosure. These patterns were used to inductively deduce different categories (i.e., the legitimation strategies) from the material. Thus, the formulation of the categories was guided by the material at hand instead of following a set of predefined theoretical concepts. We started by giving every single identified statement a name that expressed its main content (such as “comparison with other companies,” “imprecise description of incident,” and “reference to industry”). In the beginning, this led to an abundance of different category names. In some cases, we found that the categories referred to identical issues. In such cases, the respective categories (e.g., “normal behavior” and “natural behavior”) were merged. In other instances, the categories were not identical, but they largely resembled each other (e.g., the initial codes “explanation of factual reasons” and “explanation of inevitability”). These were summarized in a main category (in this case “Rationalization”) while retaining the different subcategories. Finally, some of the identified strategies (e.g., “comparison with other companies” and “reference to industry”) also seemed to embody the same content (namely a reference to the respective industry). However, a closer analysis of these strategies revealed that the competitors or the industry was used as authority or they were used to abstract from the own company what we regarded as two differing approaches. Hence, we redefined the categories by referring to these strategies as “Authorization” and “Abstraction.” In the end, each category had specific characteristics and the identified spots were coded with these initial categories. New spots were either subsumed under existing categories or a new category was built.

Forth, we began coding the material and after about a quarter of the material had been searched we met to compare our findings of possible legitimation strategies. This revealed some items that were coded differently by the coders. These differences tended to be spots that contained combined legitimation strategies. We therefore added a separate “combination” category. The identified categories were then revised, and in the fifth step, the whole material was searched and coded according to the identified categories before turning to the interpretation of results in the sixth and final step (see “Findings on Strategies to Legitimate Negative Aspects in Sustainability Reporting” and “Discussion” sections). In sum, the underlying approach thus was a hermeneutic and iterative process including multiple interplays of critically reflecting the data, searching for reporting patterns, and questioning and refining the identified categories of legitimizing negative aspects (for a similar approach see, e.g., Castelló and Lozano 2011; for general notes see, e.g., Corbin and Strauss 2008).

In terms of the reliability of the overall research process, we ensured replicability and stability with comprehensive and detailed documentation of the whole research process as well as a multi-coder analysis of the given data. Since identifying rhetoric and legitimacy strategies in reporting may be characterized as “soft” elements referring to the deeper meaning that lies buried in the text (Duriau et al. 2007), the challenge is to make “the judgments of coders intersubjective, that is, those judgments, while subjectively derived, are shared across coders” (Potter and Levine-Donnerstein 1999, p. 266). Indeed, subsequent discussions in the process revealed items that had not been identified by both researchers as well as some differences in the coding and interpretation of the reporting patterns. These different judgments between the coders were individually assessed and resolved when the data were reviewed and the findings discussed after step four of our coding (see above), thus gradually aligning differences regarding the mental schemes of the coders. In the final coding a high inter-coder agreement was achieved which points to a strong reliability of our coding framework.Footnote 3 Furthermore, internal validity was enhanced by repeatedly checking each case against the source data (the single reports) and by intensive discussions within the research team (which included two other persons apart from the authors). The iteratively identified legitimation strategies were subject to plausibility testing by mirroring them with existing patterns of legitimation referring to negative incidents which enhanced our comfort with the construct validity. A certain degree of external validity can be claimed through de-contextualization and abstraction (Avenier 2010) of the identified legitimation strategies. However, such external validity can only be claimed in the context of companies reporting according to the GRI (and not for non-reporting companies nor for those companies which conceal their negative aspects).

Findings on Strategies to Legitimate Negative Aspects in Sustainability Reporting

We identified six strategies companies use to legitimize negative aspects in sustainability disclosure on our data: marginalization, abstraction, indicating facts, rationalization, authorization, and corrective action (see Table 2). An outline of these strategies will be discussed.

Table 2 Legitimation strategies for disclosing negative aspects in sustainability reports
Full size table

We did not encounter major differences in the reporting behavior of the Dow versus the DAX companies. This might be explained by the fact that all the companies in the sample are globally acting firms with a significant share of international business (and some of the DAX companies even are or were listed in the United States). Thus, there are similar expectations regarding the sustainability of their operations. As a result, instead of within-country differences in reporting approaches, differences between single companies can be expected. Such differences might stem from variations in their sustainability and CSR management (Roome 1992; Carlisle and Faulkner 2004). Consequentially, picking companies from two corporate governance systems did not produce any particular insights regarding differences in reporting behavior but helped us broaden the data, thus increasing confidence in the external validity of the findings.

The data identified some differences with relation to the industries studied. A comparably large number of negative aspects could be found in reports from the chemical industry, whereas the financial service providers merely disclose any such aspects. This may be due to the chemical industry being more prone to negative incidents or being more proactive in its reporting behavior because it has been under constant public surveillance for quite some time (especially regarding ecological impacts). It could also be explained by negative incidents connected with the chemical industry (e.g., spills, accidents, and environmental damages) being more easily tracked to a specific company than negative incidents, for example, in the financial services industry, or a combination of the above reasons. Given the qualitative and exploratory nature of the research and the limited number of cases the current study cannot draw finite conclusions with regard to the significance of this observation.

Marginalization

By using the marginalization strategy, companies potentially seek to legitimize a negative incident by rendering it non-relevant, unimportant, or negligible. In our material companies typically use judgmental phrases and adjectives such as “no serious …” and “no significant …” or “minor” and “small” (see Table 2) to minimize the incident’s importance as illustrated by the example of Daimler (2012, p. 49): “No serious violations of data protection were detected at Daimler locations in Germany in 2011, nor were any cases worth mentioning detected at any Group companies abroad.” As can be seen, the marginalization strategy is coined by a concurrence of the description of the negative aspect and its evaluation. Instead of simply describing the incidents that have occurred, the company already evaluates the negative aspect (by rendering it unimportant), hence forestalling possible deviant evaluations by third parties. In the respective passages, the evaluation is carried out implicitly and by the company itself, and no evaluation criteria are provided. In sum, although the relevance of the negative aspect is downplayed, responsibility for its occurrence is somewhat implicitly acknowledged since the incident itself is not negated.

Abstraction

The abstraction strategy is characterized by generalizing the existence of a negative aspect as being prevalent throughout (typically) a whole industry and hence distancing the reporting corporation from the negative aspect. The following statement as well as the examples provided in Table 2 all include a general vagueness about which specific company caused which negative incident: “Some non-governmental organization stakeholders continued to raise concerns about select ingredients in some consumer products” (Johnson & Johnson 2011, p. 7). Thus, the association between a single company and the existence of the negative aspect is diluted. Instead, the negative incident is described as depicting a collective problem or challenge. The abstraction strategy therefore redirects attention from a single company to the entire industry and legitimizes the negative aspects by its widespread occurrence. This strategy thus shifts the blame by attributing guilt for their existence to peer companies.

Indicating Facts

A very common strategy is to simply mention the existence of a negative aspect as a fact without providing explanations or justifications for it: “In 2011, the number of product spillages amounted to 0.30 per 10,000 shipments.” (BASF 2012, p. 93). The company quantifies the occurrence of the negative incident but does not evaluate it and leaves its judgment to the report’s readers. Although this can be regarded as a comparably pure and unbiased form of disclosure it might be challenging for the readers to estimate the reported fact especially if no benchmarks (such as industry standards and references) are given to evaluate the figures provided by the company.

Rationalization

Rationalization usually aims at explaining and justifying aspects by referring to the utility or function of specific actions or practices (Vaara 2006). In our context, the occurrence of negative aspects is justified by using an economic rationale, i.e., it is warranted by the change in economic figures such as demand, sales, or production enlargements and hence by providing some form of logic and appeal to reason (Higgins and Walker 2012).

van Leeuwen (2007) differentiates two different types of legitimation by rationalization: instrumental rationalization and theoretical rationalization. Instrumental rationalization legitimizes the aspect or practice in question by highlighting the benefits, functions, or purposes (mostly the effectiveness or efficiency) as can be seen from the following example by Bayer (2011, p. 58): “The absolute rise is largely due to an increase in the volume of once-through cooling water owing to increased production at the (…) sites.” In such a case of theoretical rationalization, legitimation is grounded in “the way things are” (i.e., the occurrence of the respective aspect is an inevitable fact). The first set of rationalizing quotes in Table 2 highlights instrumental rationalization as companies try to explain negative corporate impact such as rising CO2 emissions by reaching corporate aims such as production enlargements. The second set in contrast depicts a focus on explanations saying “how things are.” The quotes emphasize some form of “normal” or “natural” behavior or development: “As a global company with a diverse business portfolio, the Bayer Group is exposed to numerous legal risks, especially in the area of product liability.” (Bayer 2011, p. 48). This “fact-of-life” rationalization—like instrumental rationalization—does not have to be rational in itself but at least has to appear rational to stakeholders to have a legitimizing effect.

Authorization

The authorization strategy aims at legitimization by referring to authority. These authorities can be natural persons (such as the CEO of a company) or impersonal references (such as regulations) (van Leeuwen 2007). Often regulatory bodies, regulations themselves, and academic research represent legitimating authorities. They deliver external explanations, validations, and judgments of the negative aspect as illustrated by RWE (2012, p. 114): “The heating of the cooling water creates some negative impact in the case of rivers. The limits have been specified by the regulatory authorities so that there are no significant impacts on rivers.” In contrast to the marginalization strategy, the company itself does not judge the existence of the negative incident or its impact, but rather a third party (or several third parties) validates the incident’s occurrence, hence providing an apparently more objective justification for it. This directly points to the fact that the effect this strategy has on the legitimacy of the reporting company depends on the legitimacy of the authority the company is referring to. Our analysis revealed several references to different authorities including those with a supposedly high authority in terms of societal approval (e.g., regulatory bodies). However, another authority frequently used is the company’s peer group, the company’s industry. From a critical perspective, the industry and its standards are not objective third parties. Notwithstanding (like all other authorities), they can act as a kind of legitimacy clue that serves to prevent moral judgments (Tost 2011).

As indicated in the “Methodology” section, an interesting observation is that some of the identified legitimation strategies sometimes emerged in combination (i.e., different strategies were used simultaneously for disclosing specific negative aspects). Some reports, for example, the RWE (2012, p. 114), contained combinations of the marginalization and authorization strategies as follows: “The limits have been specified by the regulatory authorities so that there are no significant impacts on rivers.” The respective authority (such as a specific regulation) is used to justify the company’s evaluation of the negative incident as being “marginal.”

Corrective Action

As in financial reporting (Erickson et al. 2011), the corrective action strategy was used extensively in our material. The company provides ideas, intent, or measures for how to tackle or avoid the negative aspect in the future (see the examples in Table 2). The examples emphasize that the company judges the negative effect or impact as being so important to take corrective action. Hence, the company—at least implicitly—takes responsibility for the occurrence of the negative aspect. Whereas some examples provide a general and quite vague corrective action strategy, others express a more concrete approach. We therefore divided the corrective action strategy into types 1 and 2. Type 1 indicates examples where corrective action is only evasively described such as “To more effectively monitor undisclosed subcontracting, we have taken steps to enhance our Standards for Suppliers, audit reporting and training processes” (Walmart 2011, p. 41). Type 2 quotes emphasize examples where the specific corrective action is clarified and precisely named as displayed by Cisco (2011, pp. C5): “Wastewater discharged from compressors exceeded local limits. (…) New filters have been installed in the sewage system, and a specialist has been hired to improve the wastewater control system.”

Discussion

In the following, we discuss the identified strategies regarding to their potential to cater to balanced reporting in light of a “true and fair view” in sustainability disclosure and to their potential for legitimacy management before deriving implications for further advancing the GRI guidelines.

The Potential Impact of Symbolic Versus Substantial Legitimation on Balanced Sustainability Reporting

As discussed above, voluntary disclosure theory (similar to signaling theory) regards this kind of voluntary reporting as an instrument to reduce information asymmetries (Guidry and Patten 2012), whereas social-political theories of disclosure suggest that voluntary disclosure is used as an impression management tool to reduce a company’s exposure to external pressure and to secure legitimacy. At first sight and according to both theories, there should thus be little incentive for companies to voluntarily disclose negative aspects of their sustainability performance. However, we suggest that both theoretical anchors can be used simultaneously to explain this type of voluntary disclosure when looking at the legitimation strategies identified in this paper.

We argue that legitimation strategies, such as those introduced above, can also help companies to accommodate the pressure to incorporate sustainability transparency in cases of negative disclosure. The strategies potentially mitigate the risk of a public backlash following the disclosure and, in an ideal case, might even help to strengthen corporate legitimacy. According to voluntary disclosure theory, “firms whose performance exceeds a certain threshold will disclose, while those below the threshold will not” (Lang and Lundholm 1993, p. 249). Based solely on past performance, companies should thus have little incentive to disclose negative incidents. However, we argue that in the case of negative disclosure, the “threshold” is not solely defined by the past performance (i.e., by the negative aspect). Signaling theory suggests that negative effects can be mitigated to a certain extent if the addressees perceive the reporting of negative incidents as proactive (Blacconiere and Patten 1994). The disclosure of negative incidents might be regarded as a positive signal in terms of proactivity and awareness of risks, thus helping to avoid similar incidents in the future. Consequently, instead of focusing solely on the negative incident, decision makers might give credit to the company for dealing with the respective problem (Yang 2007). However, whether or not the addressee perceives the given information as plausible and trustworthy greatly influences the potential effect such signaling efforts have. Likewise, not disclosing anything or even concealing information might be regarded as a signal in itself (see also Campbell et al. 2001; Bloomfield 2002).

Voluntarily disclosing negative sustainability-related aspects of a company’s performance can thus not only be used to reduce information asymmetries (as suggested by voluntary disclosure theory) and to improve transparency, but it can also be regarded as a costly signal according to signaling theory. This is because sustainability information is deemed value relevant (see, e.g., Orlitzky et al. 2003; Dhaliwal et al. 2012), and negative disclosure can potentially harm corporate legitimacy and the financial bottom line (Bansal and Clelland 2004; Skinner 1994). Despite being published by the company, negative information can be expected to be perceived as more trustworthy than the disclosure of positive information because the latter might be regarded as self-laudatory by external stakeholders thus diminishing its credibility (see source credibility theory, Birnbaum and Stegner 1979; Walster et al. 1966). It can then also be helpful to face social-political pressure to increase sustainability transparency (Philippe and Durand 2011) and to build, maintain, or repair legitimacy.

In the following, we further argue that the different legitimation strategies might have different implications for corporate legitimacy. The literature on legitimacy differentiates between symbolic and substantial approaches (Ashforth and Gibbs 1990; Hrasky 2012; Kim et al. 2007; Milne and Patten 2002). Substantial legitimation includes a real change of corporate aims, structures, actions, or activities, whereas symbolic strategies merely aim at changing stakeholder perceptions of these processes and is thus prone to decoupling tendencies. Previous research indicates that symbolic management has a potentially weaker effect on corporate legitimacy than substantial management (Kim et al. 2007; Ashforth and Gibbs 1990). In the following, we discuss which of the identified strategies follow a symbolic practice and which strategies highlight a substantial approach. This leads to identifying different possible legitimation effects and helps to assess the respective strategies in terms of balance and impartiality in GRI-oriented disclosure.

First, the marginalization strategy aims at rendering a negative incident non-relevant, unimportant, or negligible without considering, for example, the seriousness of the respective incident. This strategy thus clearly can be termed a symbolic approach. Although marginalization aims at reducing the offensiveness of the act, such conduct “cannot always be expected to improve one’s image” (Benoit 1997, p. 184). Assuming that, for example, the seriousness of the problems could become public, simply marginalizing these problems could induce a public backlash and thus risk legitimacy (see, e.g., the case of the Exxon Valdez oil spill; Benoit 1997; Williams and Olaniran 1994). Nevertheless, this strategy expresses at least some sort of corporate responsibility for the negative incident since this strategy does not deny the incident’s existence or place the incident outside the company’s sphere of influence.

Second, the strategy of abstraction, in contrast, can be seen as an evasion of responsibility. Even if the aspect or incident is negatively evaluated by others (such as NGOs), whether the specific company caused the respective negative effect remains unclear. Instead, the negative incident is attributed to a greater entity, for example, to the whole industry. The abstraction strategy hence aims at influencing the reader’s perception of the association between the company and the occurrence of the negative aspect thus distancing the company from the negativity. This strategy promises to be a successful legitimation strategy as long as criticism is not concretely directed at the reporting company. If, however, relevant actors (such as NGOs, the media, or the public) aim at accusing specific business actors for their behavior, legitimacy would again be at risk (see, e.g., the case of the campaign against Nike’s so-called sweatshops in the 1990s; Zadek 2004).

Third, other than the strategies discussed above, indicating facts cannot as clearly be assigned to one of the two approaches (symbolic vs. behavioral). On the one hand, indicating facts can simply offer an unvalued description of negative aspects. Readers might perceive such disclosure as objective and thus as an expression of a substantial management approach. This form of disclosure can then be considered unbiased, which has a higher legitimation effect than the strategies discussed so far. On the other hand, indicating facts and providing data might also be actively used to influence readers’ perceptions by achieving an “appearance of rationality” (Higgins and Walker 2012, p. 198). Other than in traditional financial reporting, the descriptive, objectified, and factualized disclosure in sustainability reporting is sometimes accused of being “a rather simplistic pursuit of ‘objective’ measurement largely adapting to traditional accounting goals” (Joseph 2012, p. 93). Such an assumed objectivization, however, might be deceiving if it is, for example, not being put into a context so that the given facts cannot be judged adequately. The indication of facts is then insufficient to inform stakeholders adequately (Marshall and Brown 2003) and to promote informed decision-making (Smith et al. 2008). It hence can also mark a symbolic legitimation approach if it is used to pretend it is an objective disclosure while in fact hiding relevant information from the report’s readers. In this case, readers might miss an evaluation of the reported aspect (or at least a benchmark for this evaluation) so that they might also perceive this form of disclosure as insufficient. This could then endanger corporate legitimacy especially in those cases when a company refuses to disclose necessary supplementary information such as industry benchmarks because the company’s performance is below the industry standard.

Fourth, the rationalization strategy depicts negative aspects as being either economically necessary or a “normal fact of life.” Both can be used to influence readers’ perceptions. Presenting a negative effect (e.g., increasing CO2 emissions) as either “necessary” or “normal” aims at connecting it to already institutionalized and legitimate practices (e.g., pursuing economic growth) to signal conformity with established practices. The intended transfer of legitimacy to the negative incident could thus influence stakeholders’ perception of the incident’s legitimacy in society.

Fifth, the authorization strategy has a similar effect. Authorities (personal as well as impersonal ones) possess some form of legitimacy themselves and hence are used as an anchor for the negative incident’s legitimacy. This strategy again aims at associating the negative aspect with symbols (in this case authorized and legitimate practices and persons) having a high-legitimate status. Building on insights from social psychological research, we can assume that the presentation of such legitimation clues (i.e., normal or necessary behavioral patterns in the case of rationalization or authorities in the case of authorization) can prevent readers from undertaking active legitimation processes themselves since cognitive processes are usually executed automatically and effortlessly to save cognitive energy (Kahneman and Frederick 2002). Such a strategy then influences the reader’s perception and can be termed a symbolic legitimation approach. In sum, the rationalization strategy as well as the authorization strategy aim at connecting the negative aspect to already institutionalized and legitimate practices or authorities (such as regulation or the pursuit of economic growth) to transfer legitimacy from this practice or authority. These strategies could work well to generate legitimacy if the respective reference value is socially approved. This is typically the case for regulatory authorities and their regulation as well as for (academic) research. Industry standards, “normal” corporate behavior, or corporate persons such as the CEO, however, usually do not hold a comparably high social approval. Hence, the legitimacy effects of the rationalization strategy and the authorization strategy strongly depend on the respective reference.

Sixth, disseminating information about negative incidents including corporate actions taken to solve or reduce the issue or prevent its occurrence in the future can be a substantial legitimation approach. Thus, the corrective action strategy could have a stronger legitimizing effect than the other strategies (Reimsbach and Hahn 2013). Stakeholders could perceive the disclosing company as being prepared to deal with the issues at hand so that this strategy could signal proactivity and awareness of risk. In this case, negative disclosure is not classified as a “bad event,” or, as Yang (2007, p. 83) posits, “strategic communication means not to hide bad information but to disclose it in a way that is conducive to its solution.” Furthermore, corrective action might qualify as a “costly signal” according to signaling theory (Connelly et al. 2010; Bliege Bird and Smith 2005) since it likely requires some effort by the communicating company in dealing with the respective issue at hand. However, this is only the case if the announced action is really taken and caters to solving the problem, and furthermore if the public is willing to accept corrective actions (which was, for example, not the case for BP’s announced actions to deal with the oil spill in the Gulf of Mexico because this incident was perceived as too dramatic to allow corrections; Muralidharan et al. 2011). Any corrective action that is merely communicated and otherwise decoupled from true corporate conduct (Bromley and Powell 2012) is otherwise employed to thematically manipulate readers’ attention to focus on positive aspects to conceal the negative aspect. Merely pretending to take action could even have converse effects if the deceitfulness is uncovered. Similarly, unspecific corrective action might be considered insufficient by the report’s readers (corrective action type I).

Overall, corporate communications (including sustainability reports) are intended to generate positive impressions (Kim et al. 2007) and to gain or maintain corporate legitimacy (Deegan 2002) while reducing information asymmetries. However, if the communicative discourse is strategically used to decouple words from actions instead of being linked to actual behavior, it is ethically questionable (Grunig 2003), and the manipulation of stakeholders’ perceptions can endanger (corporate) legitimacy. Substantive strategies that link communicative discourses to actual behaviors do not display such ambiguities. As discussed above, some of the strategies identified in the current study (especially the indication of facts and the corrective action strategy and to some degree the abstraction strategy) seem to be better suited to gaining legitimacy than the other strategies. This links back to the above discussion about past and future sustainability performance. We propose that substantive legitimation strategies are better suited to display proactivity than symbolic legitimation strategies. Companies that are, for example, able to propose potential actions to deal with a disclosed incident (as in the “Corrective Action” strategy) can in the long term be expected to be more willing to actively report such incidents. We argue that these companies have a lower threshold level of disclosure (Verrecchia 1983; Lang and Lundholm 1993) because the perception of their sustainability performance is also likely to be linked to the respective actions being reported instead of solely resting on the negative incident itself (which might even be amplified if the reporting behavior is perceived as being evasive or excusatory). Based on this insight, we are now going to derive implications for the GRI.

Implications for the GRI Guidelines

An analysis of the GRI guidelines shows that there are already some links to reporting negative aspects of sustainability performance. In relation to the question of what to report, it seems that the GRI already provides a fairly inclusive definition and explanation of “materiality” and “completeness.” Furthermore, the extensive set of GRI performance indicators provides a broad overview of topics from different areas (economic, environment, social, labor practices, human rights, society, and product responsibility) and includes many references to potentially negative aspects of sustainability performance. It is likely not possible to increase the reporting of negative incidents by further characterizations and definitions. Mandatory regulations could perhaps to so, but this is beyond the remit of the GRI. Nevertheless, what the GRI perceives as a “negative aspect” remains somewhat unclear since a concrete characterization is missing. This could nurture heterogeneity in reporting patterns since companies might be unsure which aspects to disclose and it also leaves room for manipulations in reporting behavior. We thus propose to clarify negative aspects in sustainability reporting as any corporate statement referring to factual and/or potential corporate conduct that had or has a (potentially) negative impact on the realization of sustainability.

Apart from this we will now move beyond the question of what to report and instead concentrate on the question of how to report since this aspect is still vague in the GRI guidelines so far and leaves much room for interpretation. For example, the GRI guidelines provide no criteria for “impartiality,” which is, however, necessary to again avoid manipulation. In the remainder of this chapter we now derive several implications for designing GRI guidelines based on this finding. We propose to divide the reporting of negative aspects into four separate steps as illustrated in Fig. 1.

Fig. 1
figure 1

Scheme for disclosing negative incidents

Full size image

This scheme builds upon the insights derived from analyzing the different reporting patterns. One of the most prominent GRI reporting principles for defining disclosure quality is “balance.” The GRI guidelines require that “the overall presentation of the report’s content should provide an unbiased picture of the reporting organization’s performance … avoid[ing] selections, omissions, or presentation formats that are reasonably likely to unduly or inappropriately influence a decision or judgment by the report reader” (GRI 2011, p. 13). Instead, “reports should clearly distinguish between factual presentation and the reporting organization’s interpretation of information” (GRI 2011, p. 13). As discussed above, most of the identified legitimation strategies follow a symbolic management approach potentially aiming at manipulating the readers’ perception. The GRI criterion of “balance” thus is not fulfilled in most of the analyzed reports. To overcome this drawback, the negative aspect first has to be objectively described (e.g., “we had X severe incidents in the company in 2012”) to provide the reader with the relevant facts. This, however, is only necessary but not yet sufficient to paint a complete picture for the above mentioned reasons (e.g., missing context or reference points hampering an adequate evaluation) since it might not yet allow for a holistic assessment. Second, this fact thus has to be explained (e.g., what were the circumstances of its occurrence and why did the incident happen) to provide this context. When aiming at balance, impartiality and, in sum, transparency, a reporting company could stop at this point so that these two steps could be regarded as mandatory minimum requirements in future guidelines.

The caveat of such a disclosure is, however, that companies might still risk their legitimacy and reputation if the relevant stakeholders are not satisfied by transparency alone but demand corrective action in addition. If we thus assume that the mere reporting of negative incidents is no end in itself, two further (voluntary) steps have to follow that could help secure corporate legitimacy. To allow for an assessment of the importance of the respective incident including identifying the urgency of finding a solution (or improvement) the incident (third) has to be evaluated. However, to maintain transparency, who is evaluating has to become clear. Furthermore, an evaluation does not necessarily induce action. Thus, fourth, consequences have to be derived and clarified by explicating which concrete action has been taken to leave the level of evaluation and proceed to action (as well as enhance traceability). In sum, the proposed scheme could help to achieve a more holistic reporting of negative incidents that is likely to trigger a true examination of incidents catering to a balanced reporting as well as improved transparency and accountability in sustainability disclosure.

Conclusion

The aim of our study was to shed light on the communicative legitimation strategies companies use to report negative aspects in sustainability disclosure. We identified six legitimation strategies in the reports at hand. Our analysis showed that some strategies are not well suited to providing a true and fair view of the companies’ non-financial performances as the strategies aim at changing stakeholders’ perceptions instead of truly altering corporate processes, practices, aims, or approaches. Such symbolic approaches do not guarantee accountability for corporate impacts. We thus proposed a schema for the reporting negative incidents fostering a more transparent and traceable reporting of negative incidents, which could be useful for policy makers and regulators to design specific guidelines. Furthermore, earlier research shows that symbolic management has lower legitimation effects than a substantial or behavioral approach (e.g., Kim et al. 2007). The thus far rather biased and influencing reporting about negative incidents could even endanger corporate legitimacy.

Against this background, our exploratory study contributes to academia as well as practice by providing in-depth insights into the peculiarities of voluntary disclosure of negative incidents, thus advancing the existing literature on sustainability and CSR reporting (Hahn and Kühnen 2013). Our findings are, however, limited to the companies on the Dow Jones and DAX indices we studied. These are large Western listed companies. Findings could vary, for example, for SMEs or companies from other cultural backgrounds or for those companies not using the GRI guidelines. This last aspect also points to potential self-selection bias. We specifically analyzed those companies that had GRI-type reports and that might be the best organizations in terms of sustainability reporting. However, even for those companies using the guidelines, there are no binding regulations. Furthermore, one of our research aims was to discuss potential improvements of the GRI guidelines based on our findings. This is only possible when specifically examining GRI-reporters (and not non-reporters, which are only a minority of the Dow and DAX population; see Table 3 in Appendix). Similarly, we did not use data from external sources to verify the companies’ reports. This may have resulted in a potential non-response bias (e.g., some problems not being reported). However, our aim was to explore how companies actively report negative incidents and not why some incidents are not reported or what types of incidents are not reported. Therefore, external resources could not be used for triangulation purposes in our specific case. Nevertheless, such data might be useful in a subsequent study analyzing potential omissions of negative incidents in corporate disclosure.

Table 3 Reports included in analysis
Full size table

Some of our analyzed companies provide additional information about different sustainability information on their homepages. We did not include this additional website information in our study due to practicality reasons, which limits our data. Given the importance of the “materiality”-principleFootnote 4 in the GRI guidelines, however, the data we used can be expected to cover all relevant aspects of a company’s sustainability performance. Any additional information which is not included in the reports themselves was considered non-material by the respective company and is thus not relevant for our study. Furthermore, although the sample covers an extensive number of sustainability-related statements, the sample might still be considered a rather narrow excerpt of the entire field of existing sustainability reports. However, given the qualitative and exploratory nature of our research, we consider the findings arising from our sample relevant. Indeed, a more extensive set of data would have hardly been manageable in a qualitative content analysis approach based on a multi-coder analysis by two researchers. Finally, the study is limited by the general drawbacks inherent in qualitative content analysis where subjectivity can never be ruled out completely (Mruck and Breuer 2003) despite our efforts, including a multi-coder analysis.

As we presented an exploratory study, the avenues for further research stemming from our findings are manifold. There is, for instance, only scant research on the perception of the reports’ contents by its addressees (e.g., Belal and Roberts 2010; Johansen 2010). As bias is in the eye of the beholder, it would be interesting to find out about readers’ perceptions of the legitimation strategies. Only if readers recognize potential manipulation in the respective reports is there a risk for corporate legitimacy. Moreover, corporate motivations and aims of writing a sustainability or CSR report warrant further investigation. A specific focus in this aspect could again lie on the so far largely neglected negative aspects in sustainability disclosure. Finally, it could also make sense to include a temporal dimension in analyses of legitimation strategies of negative aspects to illuminate whether legitimation strategies change during the course of time as well as to find out about possible effects on legitimacy. Even if the identified legitimation strategies might be effective in the short term, in the long run (and especially if environmental circumstances—such as public perception of the negative impact—change) the manipulation of readers’ perception could create a “legitimacy backlash.” Last, it could be interesting to shed light on possible determinants of disclosing negative incidents (such as industry, jurisdiction, company size, subject of the negative aspect, etc.).