Introduction

In the current study, we examine the corporate disclosure practices of a specific and under-researched area of corporate social responsibility (CSR)—that of countering corruption. Corruption is often viewed as negative for business as it affects a company’s ability to compete and increases operational costs such as those associated with a heightened legal risk (Hills et al. 2009). Disclosure on corruption issues serves multiple purposes—it helps hold companies accountable for their performance and achievements on its anti-corruption efforts; raises public awareness; exhorts pressure toward the adoption of similar principles and procedures by other companies; and is a means to understand what is effective in the combat against corruption and which developments are most needed (Hess 2009).

Despite the importance of countering corruption, CSR research has traditionally been focusing on issues such as environmental protection, health and safety at work, local community and stakeholder relations and management. It was only in 2002 that this issue was considered by the Global Reporting Initiative in its Sustainability Reporting Guidelines (GRI 2002) and only in June 2004 that the fight against corruption was added as the 10th principle of the United Nations Global Compact (UNGC 2009).Footnote 1 This suggests that fighting corruption in all its forms has only recently become an integral part of CSR initiatives and policies.

Although several studies investigating the effects of specific events on corporate social disclosure practices have been conducted, there is a lack of literature related to the social aspects of such disclosures (Coetzee and Van Staden 2011; Frost et al. 2005). Indeed, much of prior social disclosure research examines whether and how environmental disclosure reacted to environment-related events (e.g., Cho 2009; Deegan et al. 2000; Jantadej and Kent 1999; Patten 1992b; Savage et al. 2000), while few studied other events (e.g., Coetzee and Van Staden 2011; Islam and Mathews 2009; Vourvachis et al. 2016). Although there is a wealth of non-academic studies on anti-corruption disclosure, they rely mostly on corporate practices and place little emphasis on the factors explaining why and how this type of disclosure is produced (Gordon and Wynhoven 2003; KPMG 2008; Novethic 2006; Transparency International 2009, 2012). To the best of our knowledge, only a few academic studies have specifically examined the disclosure of corruption-related matters and most of them analyzed the drivers and the consequences of such disclosure (Barkemeyer et al. 2015; Healy and Serafeim 2016; Islam et al. 2015; Joseph et al. 2016), and no academic study appears to have examined the impact of specific events on these types of disclosures. As such, this paper attempts to address this gap in the literature.

Despite (1) the scarcity of academic publications concerning these themes and (2) the extant research on decoupling and greenwashing (see Graafland and Smid, forthcoming),Footnote 2 several institutions and organizations believe that the degree or extent of reporting on corruption can be a strong indicator of the quality and comprehensiveness of a company’s efforts in addressing bribery and corruption (Transparency International 2009; UNGC 2009). More specifically, Hess (2009) refers to the importance of CSR reporting in the fight against corruption in similar terms. Accordingly, the inclusion of anti-corruption indicators in CSR reporting should serve both internally and externally directed purposes. Regarding the former, the inclusion of such indicators would help ensure that the company is committed to combating corruption—not only the reporting process helps companies in the implementation of the indispensable changes and ensure their continued effectiveness, but it also assists different stakeholders in holding each other accountable. Concerning the latter, such disclosures enable increased accountability of leadership to the public and strengthen the knowledge about what works in fighting corruption and developing more adequate risk assessments.

As such, we examine the changes in disclosure practices on compliance and the fight against corruption at Siemens AG (hereafter “Siemens”), a large German multinational corporation, over a period of 11 years during which a major corruption scandal was revealed. This particular scandal led to several other incidents, including a number of investigations. We conduct a content analysis of the company’s annual reports (AR) and sustainability reports (SR) from 2000 to 2011 to specifically examine the changes of Siemens’ corruption and compliance disclosure using both quantitative and qualitative methods. Through the lens of legitimacy theory, stakeholder analysis, and organizational façades, we find evidence that Siemens changed its compliance and corruption disclosure practices to repair its legitimacy in the wake of the 2006 corruption scandal. Our primary findings, analyzed through a legitimacy theory lens, suggest that the AR shows peaks of disclosure amounts on corruption and compliance disclosures earlier than in the SR. We analyze these practices more closely by using stakeholder theory and the rational, progressive, and reputation façades framework (Abrahamson and Baumard 2008). More specifically, we apply these façades to the main stakeholder groups who were implicitly referred to in the disclosures pertaining to compliance and corruption issues and involved in the crisis to understand the general trends noted. We find that differences in AR and SR disclosures can be partly explained by analyzing the disclosures made about—and to—the different stakeholder groups. We find that the AR focuses more on internal stakeholders such as employees, while the SR focuses more on external stakeholders such as suppliers. We also find that the company uses the façades differently depending on which report is being analyzed.

In the following section, we present the theoretical lens of our analysis. “Case Background” section provides some background information about the Siemens case. Our methodology and analysis are presented in “Methods” and “Findings” sections, respectively. We discuss our findings, limitations and implications, and conclude in the final section.

Theoretical Framework

Legitimacy Crisis and Disclosure Strategies

Proponents of legitimacy theory argue that firms exist as a part of a broader system which determines whether the firms are legitimate or not and thus grants the license to operate within that system (Deegan 2002). The central issue is that society may revoke its social contract with the company if the company is perceived as falling short of its expectations (Deegan and Rankin 1996; Dowling and Pfeffer 1975). When an actual or potential disparity exists between the two value systems—that of society and the company—a threat emerges and questions the entity’s legitimacy (Dowling and Pfeffer 1975). External perceptions of legitimacy may change in the advent of threats due to specific events (Deegan and Rankin 1996; Deegan et al. 2000; Patten 1992a) such as changes in the community’s expectations (Lindblom 1994), changes in the composition and/or values of the public (O’Donovan 2002) or the occurrence of incidents (Cho 2009; Deegan et al. 2000). Evidence of such a rupture can be illustrated with consumers reducing or eliminating the supply of labor and financial capital to firms, or constituents lobbying the government for increased taxes, fines or laws to encourage the reduction of those actions which do not conform to the community’s expectations (Deegan and Rankin 1996).

Companies are increasingly known to use strategies to influence societal perception in reaction to legitimacy gaps. A common strategy involves making changes to a company’s disclosures. Patten (1992b) found increased environmental disclosures in the annual report following an environmental disaster in firms unrelated to the crisis, while Deegan and Rankin (1996) found increased positive environmental disclosures in firms’ annual reports related to their own environmental prosecution. Deegan et al. (2000) found that in the wake of an event that resulted directly from its own operations, a company is more likely to disclose incident-related information in its annual report, in comparison with other companies operating in the same industry. Cho (2009) noted an increase of disclosures in both a company’s annual report and corporate releases following environmental disasters along with more specific reporting strategies. Eweje and Wu (2010) also proposed different disclosure strategies over a longer period compared to Cho (2009) following an incident.

As such, the current study builds on prior literature that looks at a firm’s disclosure practices following an internal incident by analyzing in more depth how the stakeholders involved in a scandal were addressed in the disclosures.

Reporting to and on Stakeholders in Light of Legitimacy Crisis

According to Freeman (1984), a stakeholder can be defined as “any group or individual who can affect or is affected by the achievement of the firm’s objectives” (as cited in Welch and Jackson 2007, p. 183), and different stakeholder groups can be characterized between internal and external (O’Dwyer 2005; Savage et al. 1991). Darnall et al. (2009) also make this distinction between these two stakeholder groups highlighting that the former are management and non-management employees (Waddock and Graves 1997), while the latter can be regulatory stakeholders (Freeman 1984), interest groups and professional organizations (Etzion 2007), and entities involved in the supply chain such and suppliers and customers (Cox 1999). Competitors can also be considered an external stakeholder (Harrison and John 1996).

Organizations need the support from their stakeholders as it does from the broader society in order to survive—therefore, stakeholders require organizations to act in a certain manner to maintain such relationships. Social disclosures are an effective way for organizations to communicate with its different stakeholder groups (Gray et al. 1995). In general, the annual report is viewed as catering to the needs of shareholders. Studies such as Teoh and Shiu (1990) and Epstein et al. (1994) focus on investors’ needs when studying the needs of users of financial statements. In contrast, Thorne et al. (2014) found that stakeholder theory is adequate to understand why companies publish SRs. As suggested by Belal and Owen (2007) when interviewing managers from Bangladesh, social reporting in SRs is a way to manage stakeholder groups. Spence (2007) found a similar motivation when interviewing companies and their views of CSR. A quote from a CSR Manager in Financial Services was noteworthy when discussing the pressures from different stakeholder groups to report on CSR activities:

Stakeholders expect it, employees expect it, the communities expect it, our peers expect it, the government expects it. There is an awful lot of people out there who are watching. Because we are such a big company out there. Because we supply a lot of people with their life assurance, they want to know that we are a good company. And that is why we do a lot of what we do out there (as cited in Spence 2007, p. 867).

The interest of the current study, though, is in how companies report about the main stakeholder groups implicitly referred to in the disclosures pertaining to compliance and corruption issues when facing a legitimacy crisis associated with a corruption scandal that involves those very groups. A few studies have actually analyzed the reporting of different stakeholder groups but the analysis was limited to employees (Mäkelä and Näsi 2010; Williams and Adams 2013). In addition, these studies do not investigate how a company responds in light of a legitimacy crisis when the stakeholders such as employees themselves are involved in the crisis. Williams and Adams (2013) examine how a company reports about—and to—employees in light of British trends in employment following the government policies of Margaret Thatcher, and Mäkelä and Näsi (2010) analyze the disclosures surrounding a downsizing of a company. In slight contrast, our study investigates a legitimacy crisis that was partly due to the wrongful activity of certain stakeholders. As such, besides analyzing whether the company used reporting to influence stakeholders’ perceptions in the wake of a corruption scandal, we also question whether the main internal and external stakeholders—who were implicitly referred to in the disclosures pertaining to compliance and corruption issues and involved in the crisis—were treated differently from each other when reported in the AR and SR. In terms of a more specific strategy, we argue that one way to manage the different—and sometimes conflicting—stakeholder demands is through organizational façades.

Organizational Façades as a Reaction to Legitimacy Gaps

Organizational façades can constitute a useful way to study and understand how companies react to conflicting stakeholder demands (Cho et al. 2015). An organizational façade according to Abrahamson and Baumard (2008) is “a symbolic front erected by organizational participants designed to reassure their organizational stakeholders of the legitimacy of the organization and its management” (as cited in Cho et al. 2015, p. 82). Work processes, organizational structures, and pronouncements could all contribute to the creation of a façade (Starbuck and Nystrom 2006). Abrahamson and Baumard (2008) suggested three different façades that an organization may establish—rational, progressive, and reputation. As described in Cho et al. (2015), a firm projects the rational façade when using a market logic such as a cost-benefit analysis or another form of organized logic. The business case for sustainability is another way to think about this façade. The progressive façade is used when a company displays a new approach to solve a problem that highlights its innovation. A company may implement new standards to advance its sustainability program with, for example, the Global Reporting Initiative or implement a triple bottom line. The reputation façade is one that is found in company mission statements, which potentially enhances a company’s image and gives the impression that the company can achieve more than it realistically can. Themes related to social and environmental stewardship is an example of such a façade.

A company undergoing a legitimacy crisis involving different stakeholders could employ several façades at one time to cater to the different stakeholder needs. Cho et al. (2015) analyzed these façades in the AR and SR of two large multinational oil and gas companies during the years of an environmental debate. The authors found that the rational façade was featured more in the AR, while the progressive and reputation façades were more prevalent in the SR. These results, however, are preliminary since the analysis was more exploratory in nature.Footnote 3 The authors also argue that the companies were able to organize their façades through limited differences within them. However, inconsistencies were noted across each façade. The authors suggest that the progressive façade was key to ensure the companies painted a picture of themselves that was adequately cohesive for their stakeholders, hence mitigating the disparities between the rational and reputation façade.

The current study empirically examines in more depth the use of façades by analyzing the different stakeholder groups in corporate reporting. The current study also examines reporting practices over a longer period by analyzing a company’s reporting before and after a legitimacy crisis rather than around a socio-environmental/political debate (see Cho et al. 2015). Since work processes can contribute to a façade as can pronouncements, the current study does not make a distinction between the actions of a company and its disclosures.

How does a company discuss its internal and external stakeholders involved and implicitly referred to in the disclosures associated with a legitimacy crisis in its corporate reporting while reassuring its stakeholders reading the AR and SR? Following legitimacy theory, we expect that the disclosures will increase following a legitimacy crisis. As suggested by Vourvachis et al. (2016), legitimacy-threatening events heighten the social and political exposures of companies and/or industries and the response, as consistently documented, is an increase in disclosure. However, while the affected companies increase event-related disclosure, they may also increase disclosure of issues not directly related to the specific event in order to divert attention from it (Deegan et al. 2000).

Based on the findings reported by Cho et al. (2015), the rational façade follows a market logic, whereas the progressive and reputation façades focus more on corporate image. Therefore, we expect that the former to be found mainly in the AR whereas the reputation façade is more likely to be exhibited in sustainability reports. We also expect that the AR will focus more on internal stakeholders, while the SR will focus more on external stakeholders based on stakeholder theory. The AR contains the financial statements and caters predominately to the shareholders who are more interested in the financial situation of the business. In contrast, the SR allows the company to focus on its social and environmental activities, thus responding to a wider stakeholder base. While these stakeholders may demand different types of information, the SR report has been found to be a form of greenwashing (Cho et al. 2012; Guidry and Patten 2010; Lyon and Maxwell 2011)—thus, a more reputational report is expected. The progressive façade is expected to have mixed results since it may associate with both the rational and reputational façades. We also expect that the stakeholders will be treated differently between the AR and SR. For example, employees in the AR are expected to be disclosed “rationally” while in the SR they would be discussed in a more “reputational” way.

According to Cho et al. (2015, p. 83), legitimacy theory suggests that companies deliberately obfuscate potentially controversial issues and actions by way of selection of information to be disclosed, omission of pieces of information, and/or biased accounts. These authors further explain that a legitimacy theory perspective would lead to conclude that outcomes are likely to be detrimental for broader society. In contrast, however, organizational façades—which are seen as courses of action through which companies manage conflicting stakeholder demands—combined with organized hypocrisy “may indeed make room for potentially positive outcomes for broader society” (Cho et al. 2015, p. 84). They also suggest the possibility of a reputation façade expressing an ideal that can be subject to aspiration and pursued. As they put it, “the façades can free the organization to experiment and innovate beyond the rational boundaries of the market’s judgment” (Cho et al. 2015, p. 84). Notwithstanding, one has to acknowledge that just because beneficial societal outcomes can ensue from them on top of their fundamental legitimating functions, this does not mean that façades will necessarily procure such outcomes. These authors finally argue that when considering a company’s options within the broader societal context, the situation may leave it “with little choice but to engage in organized hypocrisy and establish discrepant organizational façades for this purpose” (p. 91), and that such possibility is not adequately acknowledged by legitimacy theory.

We selected Siemens as our empirical case company as it went through a legitimacy crisis. Not only was Siemens’ corruption scandal grand in terms of both scale and geographical outreach, Siemens’s “efforts in setting up anti-corruption processes and universally applying strict anti-corruption processes worldwide were unprecedented” (Schembera et al. 2015, p. 14). Further, Eberl et al. (2015, p. 1208) identify Siemens’ corruption case “as an extreme, unique case of bribery worth studying”, namely because it is today perceived as a model for how to deal with a corruption scandal and carry through compliance processes. Although this scandal and related incidents have already been the object of some recent studies (Eberl et al. 2015; Schembera et al. 2015; Schembera and Scherer 2014), none have focused on how the company has reacted in terms of its reporting practices. Hence, we view the study of the impact of Siemens’s corruption scandal as the most appropriate to contribute to the literature examining the impact of specific corruption-related events on the disclosures of a company.

Case Background

In November 2006, the Munich public prosecutors conducted searches at the offices of Siemens and its employees’ private homes in search of evidence concerning suspicions of public corruption including embezzlement, bribery, money laundering, and tax evasion. As a result, the company incurred a fine of €201 million in October 2007. According to the court decision, a former manager of the communications group colluded with other colleagues and bribed foreign public officials for the purpose of obtaining contracts on behalf of the company in Russia, Nigeria, and Libya which totaled 77 cases during the period from 2001 to 2004 (Siemens 2008a, p. 179). Investigations from the Munich public prosecutor continued throughout 2006 and involved several companies from the Siemens group in different geographical areas such as Germany, Greece, Switzerland, Liechtenstein, and Italy. Some of the cases that led to these investigations dated back to 2004 and 2005. In 2007 new corruption allegations appeared, involving Siemens companies in China, Hungary, Indonesia, Nigeria, Norway, and the USA. In December 2008, legal proceedings against the company’s Supervisory and Managing Board from the Munich legal prosecutor in Germany and the USA were terminated with the imposition of an additional fine of €395 million. Also in December 2008, Siemens pleaded guilty for violating the US Foreign Corrupt Practices Act (FCPA).Footnote 4 Three of its subsidiaries (Siemens Argentina, Bangladesh, and Venezuela) agreed to pay to the US Department of Justice a combined total criminal fine of US$450 million (Siemens AG with US$448.5 million fine and each of the subsidiaries agreed US$0.5 million).

Several other anti-corruption investigations continued or started after 2009 in Afghanistan, Argentina, Austria, Brazil, Greece, Russia among other countries. Most of these legal proceedings were related to events that occurred before 2006 but were only uncovered after the 2006 scandal or because of its investigation. Hence, the post-event period considered in the current study is not exempt from corruption-related situations. Given the difference in terms of how the 2006–2008 period (when the last settlement occurred) unfolded when compared to the period between 2009 and 2011, we divided the post-event period into these latter sub-periods in order to obtain a comprehensive understanding of the problem and provide a more refined analysis.

As a result, we analyzed the AR and stand-alone SR by Siemens based on the following timeline:

  • 2000–2005: pre-event period (Period 1);

  • 2006–2011: post-event period divided as follows:

    • 2006–2008: post-event time period from when the first corruption scandal emerged to its last settlement (Period 2); and

    • 2009–2011: post-event time period during which the company was under ongoing corruption investigations (Period 3).

Methods

Legitimacy Analysis

We use content analysis—a method commonly applied in social disclosure/reporting research (Coetzee and Van Staden 2011; Gray et al. 1995)—to examine the ARs and stand-alone SRs available on Siemens’ website. We measured the extensiveness of disclosure using the number of sentencesFootnote 5 (Branco et al. 2008; Buhr 1998; Deegan et al. 2000, 2002; Hackston and Milne 1996) related to compliance and corruption. Consistent with Branco and Rodrigues (2008) and Cho and Patten (2007), we made the assumption that each item of disclosure was equally relevant and added the disclosure scores rather than weighting them. We analyzed the topics of both corruption and compliance since these themes appeared together.

In addition, we adopted the methodology of considering pre- and post-event periods of similar duration, which is well established in the literature examining the effects of events on CSR disclosure (Branco et al. 2008; Coetzee and van Staden 2011; Deegan et al. 2000; Islam and Mathews 2009; Jantadej and Kent 1999). Consistent with these studies, we consider that this is the optimal approach of analyzing the effects of an event on corporate disclosure in a meaningful way.

Stakeholder and Organizational Façades Analysis

In order to better understand the differences in disclosure volume peaks, we performed (1) a stakeholder analysis to determine whether the AR and SR emphasized the stakeholder groups differently and (2) a façades analysis to determine whether those stakeholders were treated differently.

In the stakeholder analysis, we focused on employeesFootnote 6 and suppliers because they stand out as the stakeholder groups for whom corporate policies regarding compliance and corruption issues are of most interest and concern. A company can conceive and implement compliance and corruption-related policies specifically centered on these two stakeholder groups, proposing certain types of conducts regarding how employees should behave and how relations with suppliers must be handled, as well as instruments to make such policies effective. For example, in a document offering guidance on how to report on the fight against corruption, the UNGC proposes a “Matrix of Reporting Elements” in which the main stakeholder groups specifically mentioned correspond to the ones we consider (UNGC, 2009, p. 14).Footnote 7

We searched for instances when these stakeholders were referenced by using the keywords “supplier” and “employee” in the AR and SR from 2000 to 2011. We coded these sentences to identify major themes. For example, a sentence such as “Compliance with the Code of Conduct for Siemens Suppliers will be monitored by means of an additional corporate responsibility monitoring module in connection with the periodic quality audits of suppliers” (Siemens 2007b, p. 47) was coded as “compliance with code of conduct for suppliers” and “supplier audits”. The themes were reviewed for references to compliance and corruption. We also broadened those themes to include references to responsibility and sustainability. All of these themes were analyzed together comparing each year with the previous year to identify the addition of new themes throughout time.

We then performed another level of coding around the different façades—rational, progressive and reputation. Based on the literature review, we used the following guidelines to distinguish the different categories. The rational façade was evident when the disclosures quantified one of themes, referred to a legal standard, and displayed a business case or organized logic to explain its actions. This façade also tends to focus more on the negative aspects of the corruption scandal by simply stating the events as they happen. For example, the following sentence from the 2006 AR was classified as a rational façade: “Another former employee was apprehended in Austria and extradited to Germany” (p. 131). The progressive façade highlights new tools and the latest sustainability trends, such as the GRI Index, that the company is using to solve problems. Other tools that were noted were training, surveys, and the code of conduct. For example, the following sentence in the 2000 AR was classified as a progressive façade due to implementing employee training: “Siemens employs extensive internal controls, enterprise-wide uniform reporting guidelines and additional measures including employee education and training, to ensure that its financial reporting is conducted in accordance with applicable regulations and accepted accounting principles” (p. 38). The reputation façade highlights the positive by discussing the benefits and goals of its actions. The tone allows the company to appear that it can do more than what it can realistically achieve. For example, highlighting that the company expects all employees to uphold its principles is an example of a reputation façade since it is unrealistic to assume that all employees would uphold its code of conduct at all times. It is possible that two façades could be evident at the same time. The previous example given related to employee training was also classified as a reputation façade since the tool was also referenced to its goal of proper financial reporting. The authors conducted at least two iterations of the coding to ensure consistency.

Findings

Legitimacy Theory

Period 1: 2000–2005

As shown in Fig. 1, the content analysis of the company’s ARs and SRs indicates that the lowest amount of disclosures was recorded in the 2000–2005 pre-event period, which is expected. Figure 2 shows similar results but also breaks the disclosures into the categories of corruption and compliance.

Fig. 1
figure 1

Disclosure volume on Siemens scandal, 2000–2011

Full size image
Fig. 2
figure 2

Disclosure volume on Siemens scandal divided between the themes of compliance and corruption, 2000–2011

Full size image

Period 2: 2006–2008

As expected, results indicate a significant change in Siemens’ disclosure patterns in 2006 as shown in Fig. 1. The volume of compliance and corruption disclosure included in both the AR and SR exhibits a highly significant increase with the advent of the 2006 corruption scandal as seen in Fig. 2. These results provide evidence on the company’s intent and strategy to repair its legitimacy in the wake of a threatening event such as a worldwide corruption scandal. They also suggest that the increase in disclosure volume was even more significant for the years following the scandal burst (i.e., 2007 and 2008). Similar to Cho (2009), the current study analyzes both ARs and SRs, and documents different trends for each type of report during this second period that are worth noting:

  • AR disclosures for the year of the corruption scandal (i.e., 2006) are far more extensive than those included in the SR;

  • peaks and general increases in disclosure extensiveness occurred at different moments in both ARs and SRs. More concretely, peaks and increases in disclosure volume in SRs exhibit a one-year lag when compared to AR;

  • peaks in disclosure extensiveness during the 3-year time period occurred for the AR in 2007, which is the year when the company had its first condemnation and for the SR in 2008 when the 2006 corruption case ended following the settlement.

These results thus suggest that from 2006 to 2008, Siemens disclosed information earlier in its AR compared to its SR. We also find that AR had more total disclosures than the SR for the years 2006 and 2007 in Period 2. Explanations for the different trends are provided in the following section covering the stakeholder analysis.

Period 3: 2009–2011

Figures 1 and 2 show an inconsistent evolution of disclosure in the 2009–2011 post-event sub-period. As expected, the extensiveness of SR disclosure on compliance and corruption overall decreased significantly in the year following the 2008 settlement (except for 2011), except for compliance items in the 2009 AR and corruption items in the 2010 and 2011 AR as well as in the 2011 SR. While we acknowledge that the 2006 event has likely raised other corruption-related events from 2009 to 2011, we conjecture that increases in disclosure are associated with possible new corruption-related situations such as ongoing investigations, especially in 2011 when specific corruption-related disclosure substantially increased.

Stakeholder Analysis

While the stakeholder analysis only covers a portion of the corruption and compliance sentences, it does reveal that the SR focuses more on the suppliers, while the AR focuses generally more on employees (see Table 1). Suppliers are mentioned in relation to the themes of compliance and corruption 320 times in the SR compared to 33 times in the AR. This finding is more pronounced following the crisis for employees. Employees are mentioned 13 times before 2006 in the AR and 19 times before 2006 in the SR; however, the numbers drastically increase with 240 instances after 2005 in the AR and 145 in the SR.

Table 1 Compliance and corruption disclosures in AR and SR according to stakeholder groups
Full size table

As found in the previous section, the AR peak is in 2007, while the SR peak is in 2008. The analysis shows that part of the AR peak can be attributed to the disclosures covering employees, which peaked in 2007. The supplier disclosures peaked in 2008, which has a direct impact on the SR. These figures show that, in general, the AR and SR emphasized different stakeholder groups such that the AR focuses on those stakeholders with more direct ties with the company while the SR emphasizes those stakeholder groups that are more external to the company.

The next section will answer the following question—how does Siemens try to keep legitimacy with the main stakeholder groups implicitly referred to in the disclosures pertaining to compliance and corruption issues and involved in the crisis using organizational façades in the AR and SR? The following sections will report in more detail how each of the stakeholder groups either supports or deviates from the previous assumptions. We also analyze how the disclosures change over time. Table 2 details the different façades over time for employees and suppliers, and only includes compliance and corruption disclosures for the stakeholder groups.

Table 2 Compliance and corruption disclosures in AR and SR according to stakeholder groups and organizational façades
Full size table

Stakeholder and Organizational Façades Analysis

Disclosure Volume in Total and Over Time

When comparing employees to suppliers on total disclosure volume measured by number of occurrences, Siemens focuses on different façades in the AR and SR. The rational façade is the focus in the AR for employees, while the reputation façade for employees is the focus in the SR with both the rational and progressive façade displaying a strong presence in the SR as well. The rational and reputation façades are key in the AR for suppliers, and the progressive façade is the focus in the SR for suppliers with also a strong focus of the rational façade and slightly less the reputation (Table 2).

When focusing on volume over time in Table 2, different trends begin to emerge for both employees and suppliers. The façades occur in waves in the AR for employees, while the façades occur more simultaneously in the SR for employees with peaks happening a year later than the AR. In the AR, the rational façade begins the year of the crisis in 2006, followed by a peak in 2007 while gradually decreasing in number until 2011. Both the progressive and reputation façade begin in 2000; however, the peak for the progressive façade is in 2009 (i.e., the year after the final settlement regarding the 2006 event) and the peak for the reputation façade is in 2010 and 2011 (i.e., the post-settlement period). The peak of the façades in the SR occur in 2008 which decreases in the subsequent years. Like the AR, the rational façade occurs late beginning in 2007, while the progressive and reputation façade are evident since 2001. The façades are not evident between 2004 and 2006.

The façades occur in waves in the AR for suppliers as well with less of a presence of the progressive façade. While the progressive façade is limited in number, it does peak in 2007 like the rational façade and the reputation façade peaks later in 2010 and 2011. The rational and reputation façades do not appear in the AR until 2006, the year of the crisis while the progressive façade first appears in 2007. The façades occur more simultaneously for the supplier SR with the progressive façade having its peak in 2009, a year later than the rational and reputation façade, but the progressive façade is the focus with the reputation façade until 2011.

In summary, the AR focuses on one façade at a time in light of a crisis, while the SR focuses on several façades at one time, albeit in a delayed fashion. The AR audience, which is primarily shareholders, is concerned about the financial statement impact, so the company may be attempting to mitigate the concerns of lawsuits. Shareholders need the information quickly to know how to invest their money, which is why the disclosures began in 2007. The AR audience may also need to get eased into the reputation façade explaining why the progressive façade is needed to act as an intermediary. For the SR, Siemens cannot disclose the problems to the SR audience in the rational façade before the company can disclose its actions in the progressive façade and its goals in the reputation façade. The next section will analyze how the content differs among the façades.

Differences Between the AR and SR: Timing of the Content in the Façades

Employees

Several themes are similar between the AR and SR, while the timing was different in certain areas. For the rational façade, details of the investigation began in the 2006 AR, while the investigation was mentioned in the 2007 SR along with the ombudsman, amnesty program, number of employees completing trainings, employees getting fired, and number of employee sanctions. The ombudsman was mentioned in the 2006 AR, a year earlier than the SR, while the quantification of its actions, such as its compliance training, began in the 2008 AR, a year later than the SR. However, similar items were quantified for the first time in both the AR and SR in 2008 such as the number of employees associated with the compliance program and the number of employees who are part of the amnesty program.

For the progressive façade, the AR begins in 2002 discussing its code of conduct for employees while the SR begins discussing its principles based on an employee survey in 2001. Following the crisis, the AR mentions its business guidelines first in 2006, while in 2007 both the AR and SR mention similar themes such as the anti-corruption guide, employee reporting channels, anti-corruption/compliance training, the amnesty program, asking questions to compliance officers, and helpdesks. The employee survey was mentioned a year later in the AR in 2008 compared to 2007 in the SR.

For the reputation façade, the theme of the company’s guidelines being binding for all employees began in the 2001 SR and in the 2002 AR, which continues after the crisis in 2006 for the AR and 2007 for the SR. In 2007, the AR and SR both mention the need for clear or precise rules, in the former stating it as key to its compliance program and the latter stating it as a goal:

Clear rules for all employees, extensive training, confidential reporting channels for suspected irregularities and a helpdesk for compliance-related questions are key elements of our Compliance Program (Siemens 2007a, p. 43).

Our goal is to provide managers and employees with a set of clearer and more precise rules (Siemens 2007b, p. 24).

Suppliers

The number of suppliers signing a declaration to uphold its standards began in 2003 in the SR before the crisis, while the AR reported a number of suppliers complying with its standards in 2006. After the crisis, 2007 was the first year reported in the SR with similar information. While the timing is somewhat different for the rational façade, the progressive façade announces similar tools in both the AR and SR in 2007 such as supplier self-assessments and audits and channels to communicate problems. The reputation theme of expecting compliance from suppliers following the crisis begins in 2006 for the AR and 2007 for the SR. Similar themes of meeting anti-corruption regulations and integrating suppliers in sustainability topics were noted in both the AR and SR.

Differences Between the AR and SR: Content of the Façades

Employee: Rational Façade

In the AR, disclosures began in 2006 when the corruption investigation commenced. There is more of a discussion concerning the legal affairs related to the employee violations in the AR compared to the SR. For example, Siemens mentions the law firm who was hired (Siemens 2006a), warrants against employees, employees being sentenced, plea bargaining deals, court fines (Siemens 2007a), updates on convictions (Siemens 2008a), employees in custody, and filing criminal charges (Siemens 2009a). Siemens also specifies that the sanctions could be against the company in addition to the employees (Siemens 2006a) while only the employees were mentioned in the SR, costs for advisors (Siemens 2007a), number of countries that have held management conferences, number of answers it gave to employee questions, response rate of employee survey (Siemens 2008a), and number of employees with compliance violations (Siemens 2009a). In the AR, it also mentions that there is a compliance component in the bonuses while bonuses are not mentioned in the SR. What is of particular interest is that in the 2007 AR Siemens admits that the company did not make the importance of its rules clear while the 2007 SR mentions that their goal is to have clearer rules:

But as the current investigations show, we didn’t succeed in making the importance of our compliance rules clear to all our employees (Siemens 2007a, p. 42).

Within the reporting period, we began to revise and expand the Business Conduct Guidelines. The findings of our internal investigations will be incorporated into the project. Our goal is to provide managers and employees with a set of clearer and more precise rules (Siemens 2007b, p. 24)

In the SR, the rational façade began in 2007. Siemens mentions different figures from the AR such as the percent of employees getting fired and percent loss of salary, the number of employee sanctions (Siemens 2007b), and the turnaround time to obtain answers (Siemens 2008b). It also mentions that employees are receiving reprimands and warnings (Siemens 2007b).

In summary, the AR discloses that Siemens is partly responsible for the employee violations, while the SR shifts more of the responsibility onto the employees and portrays Siemens as an enforcer implementing its newest actions.

Employee: Progressive Façade

The tools that were specific to the AR were incentivizing actions through bonuses (Siemens 2008a) and training employees for proper reporting (Siemens 2000a). In the SR, it provides more details surrounding the tools such as using real and new case studies (Siemens 2007b) instead of fictitious case studies as mentioned in the 2009 AR. It also mentions different tools such as compliance screening for employees being considered for key positions within the company and an employee newsletter reporting on compliance issues (Siemens 2008b). In summary, more responsibility is placed on employees in the SR compared to the AR as is the case with the rational façade. For example, it highlights in the 2008 SR that employees conduct incident-driven inspections and that employees were included in the compliance program launch instead of just improving it as noted in the 2009 AR.

Employee: Reputation Façade

In the 2009 AR, the “improve it” desk was noted specifically, while the 2009 SR noted that the company made improvements based on employee recommendations without specifically mentioning the “improve it” desk. In the SR, more themes were noted like managerial employees pledging to uphold the rules every two years (Siemens 2003b), human rights being connected with training employees in 2007 (and not just suppliers as mentioned in the 2009 AR), the importance of having the backing and support from employees, the need for employees to ask themselves questions (Siemens 2008b), and the company becoming a benchmark in compliance according to an employee survey (Siemens 2009b). The 2008 SR also mentions that employees have embraced its compliance program largely due to its compliance communications which highlights a general theme that employees and the company are not in a two-way relationship per se but more in a one-way relationship in which the company dictates to the employees what must be done.

Employee: Summary

Employees are a key stakeholder group for the AR, while they are more peripheral for the SR audience. For the rational and progressive façade in the AR, Siemens took more responsibility for the scandal. The relationship between Siemens and its employees was also viewed more as a partnership as a part of the reputation façade compared to the SR. Siemens wants to portray to the AR audience that its relationships with its stakeholders are healthy since they are needed for business operations, whereas it wants to portray to the SR audience that it is being tough and that the stakeholders are getting back on more solid footing.

Supplier: Rational Façade

In the AR, Siemens mentions the antitrust investigation, sanctions on suppliers (Siemens 2007a), and creating an equal playing field for bidding via self-regulation:

Greater transparency, overall social responsibility and monitoring will be essential for creating an even playing field for all the suppliers bidding for contracts. This will happen through self-regulating mechanisms that customers and suppliers have committed themselves to implementing (Siemens 2011a, p. 76).

In the SR, several figures are given such as the percentage of suppliers committed to upholding the code of conduct, number of supplier audits (Siemens 2007b), and the number of self-assessments (Siemens 2008b). The topics of the code of conduct, audits, and self-assessment were mentioned in passing in the AR. The SR also mentions the challenges associated with supplier compliance, terminating supplier relationships (Siemens 2007b), and the breaches followed by the audits and remedial measures (Siemens 2008b). Details are also given about how it constructed the code of conduct, e.g., explaining how the company takes into account the business size and business activity when constructing a code of conduct. The SR also mentions that supplier standards are part of the discussion with associations (Siemens 2008b). In summary, more figures are in the SR, and the SR provides more of a context surrounding the implementation of different tools. The AR focuses more on the legalistic nature of the supplier activities as is the case with the employees.

Supplier: Progressive Façade

No new information was noted in the AR. In the SR, more details are given about the code of conduct, such as its revisions (Siemens 2006b), training (Siemens 2007b), and accepting outside codes of conduct (Siemens 2008b). The 2007 SR also mentions compliance structures in relation to Chinese suppliers. Tools that were highlighted in the SR are as follows: inspections, online qualification module, integrity pacts (Siemens 2008b), online sustainability training, and a supplier sustainability award (Siemens 2009b). Integrity pacts were specifically mentioned in the SR, which suggests an agreement with the supplier and Siemens. However, the AR suggests that the relationship between Siemens and the suppliers is a one-way relationship by saying that the suppliers have to abide by the principles of integrity.

Supplier: Reputation

The AR discloses again Siemens’ role following the legitimacy crisis by stating the principles are a way to show its trustworthiness to suppliers (Siemens 2009a). It also discloses a positive image of its suppliers by stating the suppliers are committed to transparency and are abstaining from bribes through collective action (Siemens 2011a). The SR discloses a more disciplined image of Siemens by stating it only does business with suppliers that uphold the code of conduct (Siemens 2007b). In summary, the AR portrays a less harsh image of suppliers while the SR portrays a stronger image of Siemens.

Supplier: Summary

In contrast to the employee disclosures, the relationship between Siemens and its suppliers as a part of the progressive façade is less of a partnership in the AR compared to the SR. Suppliers are a key stakeholder group for the SR audience, so highlighting a positive relationship is beneficial for Siemens. However, a more positive image is displayed of suppliers in the AR as a part of the reputation façade, while the SR portrays a more disciplined image of Siemens. The AR audience values suppliers that uphold the law as this will result in less lawsuits, while the SR audience wants to perceive the company as tough on suppliers in addition to having a good relationship since more constructive change may result.

Discussion and Conclusion

The interest of the current study was to examine how Siemens reported about the main stakeholder groups that were implicitly referred to in the disclosures pertaining to compliance and corruption issues and involved with a legitimacy crisis associated with these aspects. Following legitimacy theory, we expected that the disclosures would increase in the wake of the legitimacy crisis. Consistent with prior research (Cho 2009; Deegan and Rankin 1996; Eweje and Wu 2010; Islam and Mathews 2009; Patten 1992b), our findings suggest that Siemens did engage in legitimacy strategies by increasing disclosure when faced with an event threatening its legitimacy—the occurrence of the 2006 corruption scandal.

The findings pertaining to the period 2006–2008 are consistent with past studies that show a significant increase in disclosures immediately after the occurrence of an incident (Branco et al. 2008; Cho 2009; Coetzee and Van Staden 2011; Eweje and Wu 2010; Islam and Mathews 2009; Patriotta et al. 2011). The results obtained for the period 2009–2011 are also consistent with those obtained in Cho (2009), who showed evidence of significant decreases in disclosure in the post-event period. We also find evidence supporting differences in disclosure strategies between the two main sources of content analysis, the AR and stand-alone SR. Concerning the stand-alone SR, findings are consistent with previous studies since the peaks of disclosure extensiveness occur in the year following the event as in Deegan et al. (2000) and Deegan and Rankin (1996).Footnote 8 However, the current study notes that the disclosure peak of the SR was one year later than the AR. The findings contrast with our general expectations and the findings of Frost et al. (2005), who performed a (non-academic) study on corporate social disclosure trends in Australian firms and concluded that issues related both to non-compliance and specifically to corruption and bribery had a higher presence in SR than they had in AR.

Frost et al. (2005) and our own findings may be in part explained by one of the reasons mentioned by Branco and Delgado (2012) as rationales for the lack of disclosure on the fight against corruption—companies are likely to discuss as little as possible on this topic because they fear that raising such sensitive issues could generate suspicions of corruption-related problems. As Frost et al. (2005) did not analyze disclosure reactions to negative events, the companies examined by these authors had no substantial reason to report on the fight against corruption in a detailed manner and probably reported in a manner consistent with what we call a reputation façade. On the contrary, Siemens had to react to a negative event pertaining to corruption in a context in which suspicions were already there and explain the problem at hand as well as the specific measures taken to prevent the occurrence of similar future events.

The paper then sheds more light on the one-year lag of the SR compared to the AR by using stakeholder theory and organizational façades. The company may assume that annual report readers constitute a different audience than the sustainability readers and that the former have different informational needs. We could assume that sustainability reports are more targeted to external stakeholders such as the community, consumers, suppliers, and certain types of investors as opposed to annual reports, which are more financial and technical in nature and target a different public such as shareholders, banks, tax authorities, and financial analysts as well as internal stakeholders such as employees. Whereas for stakeholders other than investors explaining the occurrence of the event and presenting the measures taken to prevent the occurrence of future similar ones could be done by way of a rhetoric consistent with what we call reputation façade, for investors the discourse had to be more specific and direct, consistent with a rational façade. And this had to be done by way of the AR, which is the document that investors are more likely to read and analyze.

Stakeholder theory has been studied by the accounting literature, but the research has been limited to analyzing disclosures pertaining to employees (Mäkelä and Näsi 2010; Williams and Adams 2013). The current study also investigates how a company responds in light of a legitimacy crisis when the stakeholders such as employees themselves are involved in the crisis which has not been addressed by the accounting literature. We found that the AR focused more on internal employees, while the SR focused more on external stakeholders, which is consistent with stakeholder theory. The earlier peak in the AR can be partly attributed to the disclosures covering employees, which peaked in the earlier years as noted in Table 1. The supplier disclosures that peaked in the later years had a direct impact on the SR. Since the readers of the SR are assumed to be more peripheral to the company, it is natural to observe that the SR focused more on an external stakeholder group such as suppliers while the AR focused on employees who are more internal to the company.

The previous assumptions were further tested by analyzing the specific legitimacy strategies used by the company, particularly organizational façades, which can be used when facing conflicting stakeholder demands. We found that the AR focuses on one façade at a time in light of a crisis, while the SR focuses on several façades at one time, albeit in a delayed fashion. These findings contribute to Cho et al. (2015) by furthering the study’s preliminary results into the differences between how the AR and SR report on organizational façades, a subject which has not received adequate attention by legitimacy theory. We find that the progressive façade was used differently by the AR and SR. When analyzing the disclosures surrounding employees, the peak of the progressive façade occurred between the peak of the rational and reputation façade, while the progressive façade peaked simultaneously with the other two façades. Slight differences were noted around the supplier disclosures. The AR audience, which is primarily shareholders, is concerned about the financial statement impact, so the company may be attempting to mitigate the concerns of lawsuits. Shareholders need the information quickly to know how to invest their money, which is why the disclosures began in 2007. The AR audience may also need to get eased into the reputation façade explaining why the progressive façade is needed to act as an intermediary. For the SR, Siemens cannot disclose the problems to the SR audience in the rational façade before the company can disclose its actions in the progressive façade and its goals in the reputation façade.

We also analyzed the content of the different façades. The qualitative findings suggest that the AR is consistent with its supplier and employee disclosures as is the SR. The differences result from the type of report, having the rational façade as the focus of the AR and the progressive and reputation façade the focus of the SR. These findings are consistent with Cho et al. (2015). However, the difference here is that the reputation façade was important for suppliers in the AR, whereas the rational façade was important for both employees and suppliers in the SR—with the rational façade being more important than the reputation façade in the SR. Associations and NGOs have an interest in the SR, as noted in the 2008 SR, so having some rational disclosures related to the crisis would be of interest to them.

We note from stakeholder theory that social disclosures are an effective way for organizations to communicate with its different stakeholder groups (Gray et al. 1995). The previous literature does not address how a company responds in light of a legitimacy crisis when the stakeholders such as employees themselves are involved in the crisis. We found that Siemens chose to communicate to its stakeholders differently in the AR and SR by analyzing the content of the different organizational façades. The AR is more concerned about sharing the responsibility for the wrongdoing between the company and the stakeholder, while more responsibility is placed on the stakeholder in the SR. Siemens is presented as more of a partner with its stakeholders in the AR, while in the SR Siemens is more of an enforcer. Siemens wants to portray to the AR audience that its relationships with its stakeholders are healthy since they are needed for business operations, while Siemens wants to portray to the SR audience that it is being tough and that the stakeholders are getting back on more solid footing. These findings are consistent across the different façades with each façade emphasizing a different aspect. The rational façade focuses on the violations and on who was responsible, the progressive façade focuses on the tools and the level of involvement of the different stakeholders, and the reputation façade focuses on Siemens aspirations in relation to itself and its stakeholders.

Like all studies, ours is subject to several limitations. It does not allow for conclusive generalization given the unique case and context that we examine, as well as its exploratory nature. In particular, shareholders and society stakeholder groups had the most mixed findings, which may be due to the different and conflicting pressures on these groups. For example, society in the AR was expected to be treated in a more rational way, while society in general is expected to be treated in more reputational terms. These conflicting pressures may result in unexpected results at times. However, since the Siemens case is sometimes mentioned as a “leading example in the industry” (Eberl et al. 2015, p. 1222), we believe that our study may offer important insights, given that it is, so far as we are aware, the first to apply an organizational façades approach to the analysis of the reaction of a corporation to an event such as the one we have studied.

Second, the analysis is limited to the extent that public company information is made available only online in the form of corporate reports. Despite these limitations, however, the current study contributes to existing research in several ways. It first brings additional evidence to the scarce research body on the social dimension of corporate social disclosures. Further, and to the best of our knowledge, this is the first study focusing on the specific impacts of such a relevant social event as corruption on corporate social disclosures. The specific results of the current study document the changes in Siemens’ disclosure practices over time for its different stakeholder groups in light of a legitimacy crisis. It also documents its choices regarding its portrayal of different stakeholder groups, which to our knowledge is also novel.

In particular, we consider our study as an important contribution to the literature using the analysis of organizational façades in the examination of corporate social disclosure practices. This type of analysis provides an “innovative” perspective in the social and environmental accounting research (Michelon et al. 2016). As far as we are aware, only one study has conducted such an analysis (Cho et al. 2015). Our work provides insights on how a company uses corporate disclosure to manage stakeholders’ perceptions in the wake of an event such as the one analyzed. We show that organizational façades analysis is of utility when undertaking such an analysis, thus contributing to this stream of research.

Finally, the current study uncovered several issues for further investigation. Additional research could be conducted concerning the impact of corruption-related events in other companies in different contexts. The present work also revealed the need for more research concerning factors influencing disclosure on corruption. Although studies have considered the materiality and importance of social disclosure in the AR for its users (Campbell et al. 2003), a lack of research exists concerning differences in informational needs of the users of corporate social disclosures in both the AR and SR. For example, what do the different stakeholder groups expect to find when reading the AR and SR? This question could be asked directly to different stakeholder groups. Finally, additional research could ask the question regarding how disclosures related to corruption differ from environmental disclosures. In particular, is corruption a more socially taboo topic than environmental concerns, and do the trends between these corporate social disclosures differ in a significant way.