The shift of corporate social responsibility
A study of instrumental and political CSR within annual CSR
reporting and the effect of standardization
AUTHOR: Alina Simons
STUDENT ID: 11822805
SUPERVISOR: dr. Piet Verhoeven
DATE: 28th of June 2019
Abstract
While corporate social responsibility (CSR) has gained increasing attention over the past
years by corporations and researchers alike, there remains little consensus on its execution
and its conceptualization. Two common, yet contrasting, approaches to CSR by corporations
are instrumental and political CSR, which form the main concepts of the present research.
First, the present research provides an overview of instrumental and political CSR, which
simultaneously brings forth distinct content indicators with which the two concepts can be
identified within annual CSR reports. Based on this, the findings show that within the CSR
reports by Nestlé, Prada, and ExxonMobil, the deliberator content indicator of political CSR
is most frequently present. Additionally, an analysis of the relationship between the Global
Reporting Initiative (GRI) G4 reporting guidelines and the two concepts reveal that the G4
guidelines decrease the presence of instrumental CSR within the annual reports. This yields a
promising result for the standardization of the practice, which appears successful at improving
the quality of the reports.
Keywords: instrumental CSR; political CSR, reporting guidelines; standardization; corporate
social responsibility
Introduction
The concept of corporate social responsibility (CSR) has gained increasing attention over the
past years, developing into a trending practice for many companies around the world. This is
reflected in the steadily increasing volume of annual CSR reports, which form the centerpiece
of corporate CSR communication (Kilian & Hennigs, 2014). While 20% of the S&P 500
later (Governance & Accountability Institute, 2019). Despite this growing interest in CSR,
there remains little global consensus on its execution, which is due to the voluntary nature of
the practice and the fact that it remains unaudited (Kilian & Hennigs, 2014).
This lack of consensus is also reflected in the academic field of CSR research,
which has thus far been dominated by the absence of common frameworks and definitions
(Crane & Glozer, 2016). Generally, however, definitions of CSR refer to the responsibility of
enterprises to serve “people, communities, society, and environment” (Cai, Jo & Pan, 2012,
p.467) and to “prevent, manage and mitigate any negative impact that [their business conduct]
may cause” (European Commission, 2017). The growing attention on CSR is rooted in a
change in public consciousness (Carroll, 2015), which entailed a growing sense of awareness
of social and environmental issues and an increased distrust in corporations (Kilian &
Hennigs, 2014). This resulted in heightened stakeholder expectations (Crane & Glozer, 2016;
Ehrnström-Fuentes, 2016), accompanied by demands for companies to “do more than make money and obey the law” (Carroll, 2015, p.87).
In order to increase the quality of the CSR reports and
standardize the practice, various non-profit and non-governmental organizations started to
create a range of voluntary guidelines and standards of reporting. The need for global
guidelines developed as a response to numerous corporate “scandals” involving CSR
reporting, such as corporations “only disclosing positive information or omitting risks and negative impacts to polish an image” (Wagner & Seele, 2017, p.333).One notable example of
such global guidelines forms the independent, international organization Global Reporting
Initiative (GRI), which launched its first version of CSR reporting guidelines in 2000 (GRI,
2019a). At present, the fourth version of the guidelines, commonly referred to as the G4
guidelines, are the most widely adopted on a global scale and are aimed at increasing
accountability, transparency, and credibility of annual CSR reports (GRI, 2019b; Barkemeyer
communicate their impact on the environment and society in order to form a “force for
positive change,” while simultaneously increasing the firm’s value (GRI, 2019b). The view
of CSR as a force for positive change, as well as a tool for value creation, reflects two distinct
approaches to CSR, namely political and instrumental CSR. Instrumental CSR is defined as a
strategy “that first and foremost aims at improving the financial position of the firm” (Seele &
Lock, 2014, p.402). Here, CSR forms a tool trough which value creation is achieved by, for
example, increasing brand awareness, building a positive reputation, and enhancing the firm’s
image (Schultz, Castello & Morsing, 2013). Any positive environmental or societal effects are
regarded as a “byproduct” (Seele & Lock, 2014, p.402). Political CSR, on
the other hand, is defined as a “movement of the corporation into the political sphere in order
to respond to environmental and social challenges” (Scherer & Palazzo, 2010, p.910). Two
aspects are central to political CSR. First, the aspect of self-regulation, as transnational
corporations increasingly operate in ways that extent beyond the reach of local jurisdiction, or
expand their business to nations that do not have any regulatory policies in place (Scherer &
Palazzo, 2011). The second central aspect is dialogue and rational communication, which is, as Lock and Seele (2017) note, “at the heart of political CSR” (p.3). Here, especially
stakeholder participation, transparency, and accountability are emphasized (Seele & Lock,
2014).
Based on this, the following research questions can be identified:
RQ1: Which aspect of either political or instrumental CSR is most frequently present within
annual CSR reports?
RQ2: Is there a relationship between the presence of the GRI G4 guidelines in the CSR
Theoretical Framework
In order to answer the above stated research questions, several content indicators for
instrumental and political CSR were identified based on relevant literature. The purpose for
this was, first, to provide a contrasting overview of the two concepts that considers and
incorporates various frameworks and definitions. A second purpose was to provide a clear set
of content indicators with which political and instrumental CSR can be identified within CSR
communication content; an aspect that has thus far been missing. The third purpose was the
operationalization of political and instrumental CSR.
Instrumental CSR
Based on various literature dealing with the concept, instrumental CSR consists of three
content indicators. These three content indicators are here labeled as economic actor,
positivity selector, and one-way communicator. They aim at capturing the essence of
instrumental CSR, which can be summarized as the view of CSR as a tool for value creation,
the operationalization of CSR reports as vehicles for image enhancement, as well as a general
focus on one-way communication with stakeholders. The three content indicators will now be
discussed in more detail.
Economic Actor
The first indicator of the instrumental approach to CSR is here labeled as the economic actor,
which entails the previously discussed link between instrumental CSR and a potential increase in the firm’s value (Cai, Jo & Pan, 2012; Carroll, 2015; Du, Bhattacharya & Sen, 2011;
Ekatah, Bampton, & Halabi, 2011). This aspect originated in the 1970’s and is rooted in the
still frequently cited Friedman view, which states that corporations in a free society only bear
one social responsibility, namely to “use its resources and engage in activities designed to
Halabi,, 2011, p.250). While this view of CSR has since evolved and corporations now often
recognize their social and environmental responsibilities, investments in social demands in
this approach are still deemed appropriate only if they do not impose costs on the firm
(Garriga & Melé, 2004).
This further entails that a firm position itself as a solely economic entity (Lock &
Seele, 2017), which affects the way responsibilities are assigned. In this view, responsibility is
assigned along the so-called liability logic (Djelic & Quack, 2011; Schultz, Castello &
Morsing, 2013). This results in a reactive state of firms towards environmental and social
responsibilities, which are commonly only assumed when legal reasoning will explicitly find guilt, fault or harm of the firm’s involvement (Schultz, Castello & Morsing, 2013).
The prominence of the liability logic already implies a further implication, namely the
assumption of intact governmental and regulatory systems. Thus, the instrumental approach to
CSR presumes that formal rules and regulations, or in other words, hard laws, are executed
effectively were applicable (Scherer & Palazzo, 2010). Consequently, in this view, firms do
not have to engage in proactive measures such as self-regulation, as the governmental and
regulatory systems already effectively exert their influence as ‘checks and balances’ on the
firm.
Finally, the positioning as a solely economic actor has implications for the way
legitimacy is perceived by practitioners. Legitimacy, in the broader sense, can be defined as the “alignment or congruence between organizational activities and societal expectations”
(Crane & Glozer, 2016, p.1235). The instrumental approach to CSR, however, is dominated
by what is commonly referred to as pragmatic legitimacy, which can be defined here as the “organization’s ability to instrumentally manipulate and deploy evocative symbols in order to
gain societal support” (Schultz, Castello & Morsing, 2013, p.683). Thus, normatively
conforming behavior is not ethically motivated, but rather practiced for the sake of increasing
2013).
Positivity Selector
The second indicator of the instrumental approach to CSR is here labeled as the positivity
selector, which entails the commonly researched tendency of firms to exclusively discuss
positive results and aspects of its business conduct within CSR reports (Scalet & Kelly, 2010;
Sethi, Martell & Demir, 2016). For example, in a study conducted by Scalet and Kelly (2010),
the response of 30 firms to negative events related to their business were analyzed. The
findings revealed that the majority of firms, namely 80%, did not address the negative events
in any of their CSR reporting, even when those events received media attention (Scalet &
Kelly, 2010). An explanation for the low response rate was theorized as being rooted in an
institutional context, as there are no mandatory guidelines for CSR and thus no negative
consequences for ignoring negative events or business performance (Scalet & Kelly, 2010).
This is a result that was further investigated by Sethi, Martell
and Demir (2016), who found that while firms generally provide information on the major
environmental, social and governance issues at hand in their CSR reports, they tend to ignore
controversial aspects of their business, as well as any connected negative impact it has on the
said issues. Sethi, Martell and Demir (2016) note that this selective focus on positive aspects of the firm’s performance can result in decreased public trust in the quality of the CSR
reports. Generally, however, it reflects what has been described earlier as pragmatic
legitimacy, as this selective reporting is likely aimed at creating a positive reputation and
enhancing the overall image among stakeholders by exclusively focusing on the attractive
features of CSR (Schultz, Castello & Morsing, 2013).
One-way communicator
one-way communicator and refers to the practice of firms to exclusively focus CSR
communications on standalone media, such as annual reports. While some scholars have
deemed the use of CSR reports as powerful and important (Kilian & Hennigs, 2014; Patten &
Zhao, 2014; Sethi, Martell & Demir, 2016), others criticized the reports for only being used as
an image enhancement tool (Cho, Michelon & Patten, 2012; Aras & Crowther, 2008). While
both of these views may be applicable, it has generally been argued that two-way
communication is crucial for successful CSR, as the practice itself is focused on society and
its members, making dialogue and understanding important elements (Seele & Lock, 2014).
However, due to the ambivalent evidence regarding if
CSR can create economic value for the firm, adopting a one-way, instrumental
communication approach appears to be the easiest and safest option for many practitioners
(Kim, 2017). This was supported in a study by Pollach (2015), who analyzed interviews with
16 CSR directors of large industrial companies. Here, all except for one of the directors
approached CSR in an instrumental manner (Pollach, 2015). Thus, current CSR communication efforts appear to remain a “one-way method of supporting and reinforcing
corporate actions and identity” (Morsing and Schultz, 2006, p. 328)
Political CSR
Based on various literature, political CSR can be said to consist of three content indicators.
These three content indicators are here labeled as quasi-political actor, deliberator, and
two-way communicator. Similarly to the content indicators identified for instrumental CSR, the
three political CSR content indicators aim at capturing the essence of the concept. This can be
summarized as the view of CSR moving the corporation into the political sphere through
self-regulation, the focus on deliberative communication and collaboration with stakeholders, and
the active usage of two-way communication channels. The three content indicators will now
Quasi-political actor
Contrary to instrumental CSR, a firm adopting the political approach to CSR does not only
position itself as an economic actor, but also recognizes its role as a quasi-political actor. Both
spheres are regarded as intertwined, deeming that corporations “bear not only economic,
social, and environmental responsibilities, but also political responsibilities” (Seele & Lock,
2014, p.401). This approach is rooted in globalization and the structural changes that come
with it (Lock & Seele, 2017; Scherer & Palazzo, 2017). Namely, as discussed earlier,
transnational corporations increasingly operate in ways that extent beyond the reach of local
jurisdiction (Scherer & Palazzo, 2011), which creates a “regulatory vacuum” in which
nation-states cannot exert any power over corporations (Lock & Seele, 2017, p.2). Thus, pressure is
increasing for companies to fill this gap by stepping into a new political role (Lock & Seele,
2017; Scherer & Palazzo, 2017). Consequently, political CSR shifts the focus away from the
strategic business case for CSR, towards a state in which firms proactively collaborate with
stakeholders and third-party organizations to “regulate the market through democratic
principles” (Ehrnström-Fuentes, 2016, p.434). This stands in direct contrast to instrumental
CSR and its’ focus on hard laws. Instead, firms that practice political CSR recognize the
regulatory vacuum in which they operate and voluntarily choose to engage in self-regulation
through soft law (Scherer & Palazzo, 2010).
Similar to instrumental CSR, the recognition of a firm as
a quasi-political actor has implications for the way legitimacy is perceived. However, in this
political approach to CSR the dominating concept is moral legitimacy. In this view,
legitimacy is achieved through creating location-dependent norms and rules through active
deliberation and open discourses between the firm and its stakeholders (Ehrnström-Fuentes,
2016; Scherer & Palazzo, 2010). Through this process of deliberation, the firm and its stakeholders ideally reach a “mutual agreement on what kind of corporate conduct should be
considered morally legitimate” (Ehrnström-Fuentes, 2016, p.436). Thus, while instrumental
CSR commonly follows an inherently economic logic of legitimacy, political CSR commonly
follows a pro-social logic (Ehrnström-Fuentes, 2016).
Deliberator
The second indicator for political CSR is here labeled as the deliberator and applies when
firms proactively take part in setting and redefining societal standards through assuming
co-responsibilities with governments and other third parties (Scherer & Palazzo, 2007). This
stands in direct contrast to the liability logic inherent to the instrumental approach to CSR, as
firms who practice political CSR proactively recognize their own responsibilities for all
aspects of their business, such as the entirety of their supply chains. Thus, rather than taking
advantage of the regulatory vacuum for the sake of profit, corporations which practice
political CSR are open to work together with stakeholders and third parties to regulate the
market regardless of any economic benefits (Ehrnström-Fuentes, 2016). Consequently, political CSR “changes the modus of responsibility from the reactive model […] to a
proactive concept of societal involvement” (Scherer & Palazzo, 2007, p.1110).
Two-way communicator
In contrast to instrumental CSR, dialogue and deliberation, as proposed by Habermas (1984),
form the centerpiece of the political approach to CSR (Lock & Seele, 2017). Thus, the third
and final indicator is titled two-way communicator and refers to the necessity to engage in
continuous open dialogue with stakeholders and independent third parties to achieve moral
legitimacy (Seele & Lock, 2014). This view is based on Habermas’ theory of communicative
action (1984), which states that legitimacy can only be achieved through ethical discourse
(Seele & Lock, 2014). Seele and Lock (2014) extended this theory of communicative action in light of political CSR, describing four deliberative demands of the concept, namely “open
further suggest a toolbox for CSR communication (see figure 1), aimed at practitioners who
seek to meet the four demands of political CSR.
Figure 1. A typology of CSR communication tool
As shown in figure 1, Seele and Lock (2014) make a distinction between deliberative and
instrumental, as well as published and unpublished tools. While the instrumental tools
correspond to the instrumental approach to CSR, the deliberative tools correspond to the
political approach to CSR. As the present study is focused on openly accessible CSR
material, especially the published communication tools are of interest in this context. The
CSR websites, brochures, and reports reflect the one-way communicator content indicator
discussed earlier. While these channels are useful for the publishing of information relevant
for stakeholders and shareholders, they do not allow for direct engagement with them. In
contrast, web blogs, wikis, and social media are listed as political CSR communication tools.
These are channels that can be used for both, the publishing of relevant information, as well
as the direct engagement with stake- and shareholders. Thus, if a firm extends its CSR
communication beyond purely instrumental tools towards achieving open discourse and
two-way communication with stakeholders, it is a clear indication for the presence of political
CSR. Concluding, the overview of the six content indicators
on the finding by Sethi, Martell and Demir (2016), as well as Scalet and Kelly (2010), that
positive aspects of CSR engagement and business conduct appear to dominate current CSR
reporting, the first hypothesis is as follows:
H1: The positivity selector content indicator is more frequently adopted within the CSR
reports than the economic actor, one-way communicator, quasi-political actor, deliberator, and
two-way communicator.
Content indicators and framing
These six content indicators of instrumental and political CSR are based on the general
literature and are, when considered from a communications perspective, comparable to
frames. While there is no single, agreed-upon definition of framing, it generally refers to
conceptual tools with which textual information can be conveyed, interpreted and evaluated
(Semetko & Valkenburg, 2000). Framing further entails that certain aspects of perceived
reality are selected in order to enhance their salience (Semetko & Valkenburg, 2000). This
enables the approach to framing that Semetko and Valkenburg (2000) adopt within their
research, namely the pre-defining of frames as content analytic variables, which were then
used to verify the extent to which the frame occurs in the analyzed material.
When this definition and approach to framing is
transferred to the content indicators brought forth in the present research, two similarities are
notable. First, the six content indicators were developed to form a tool with which political
and instrumental CSR can be identified within CSR communication content. This reflects the
definition of framing, as the content indicators aid in the interpretation and evaluation of
information featured in CSR reports. Second, the content indicators reflect the salience of certain aspects of instrumental and political CSR. Thus, similarly to Semetko’s and
concepts are present within the CSR reports.
Standardization
As discussed earlier, the independent organization GRI is working towards improving the
quality of CSR reports through helping corporations standardize their way of reporting. To
achieve this, GRI is launching sets of guidelines that fit all forms of organizations world-wide
and are periodically updated and reviewed with the aim to ensure the “most up-to-date guidance for effective sustainability reporting” (GRI, 2015, p.3). The most recent version of
the guidelines, the G4 guidelines, launched in 2014 and were developed in collaboration with
a selection of international CSR reporters and users, as well as various representatives of, for
example, the business, labor, and civil society (GRI, 2015). The main focus for the
development of the G4 guidelines was on materiality, which entails that only the information
that is most relevant for stake- and shareholders should be included in the final CSR reports
(GRI, 2015). This emphasis on materiality aims at making the reports “more relevant, more
credible, and more user-friendly” (GRI, 2015, p.3).
Corporations seeking to adopt the guidelines can choose between two options,
the ‘core’ and the ‘comprehensive’ option. Regardless of which option will be chosen, it first
entails a process of self-identifying the corporation’s material aspects, which are defined as aspects that either “reflect the significant economic, environmental, and social impacts, or
substantively influence the assessment and decisions of stakeholders” (GRI, 2015, p.11).
While the ‘core’ option requires the corporation to report at least one indicator for all its self-identified materiality aspects, the ‘comprehensive’ option requires the corporation to report all
indicators for the materiality aspects (GRI, 2015). Depending on if these requirements are met, the corporation’s CSR report will be awarded with an ‘in accordance’ (GRI, 2015). Thus,
it is important to note that an ‘in accordance’ does not indicate the quality of the report, but
While GRI successfully increased the volume of CSR reports that adhere to the
guidelines, critics question if there has actually been a shift away from impression
management, towards accountability and change (Barkemeyer et. al., 2014; Wagner & Seele,
2017). In other words, do the GRI guidelines further facilitate the instrumental approach to
CSR, or do they advance the political role of corporations in a globalized world? This
question also formed the main objective of a research conducted by Wagner and Seele (2017).
In their research, they take a closer look at the GRI G4 guidelines to find out to what extent they “contribute to a political understanding of CSR,” as opposed to the predeciding G3.1
guidelines (Wagner & Seele, 2017, p.334). In light of the extensions made to the G4
guidelines, which now include aspects such as stakeholder engagement, dialogue, and the
general level of disclosure of governance, the link to political CSR appears to be valid.
Indeed, the classical text analysis conducted by Wagner and Seele (2017) reveals that the G4
guidelines feature more aspects of political CSR than the G3.1 guidelines. Thus, in theory, the
G4 guidelines are more successful in supporting the political approach to CSR than the
previously published G3.1 guidelines (Wagner & Seele, 2017). The present research attempts
to extent the one conducted by Wagner and Seele (2017) to further determine if the G4
guidelines do, in fact, facilitate political CSR within annual reports, as opposed to those
reports not adhering to the G4 guidelines. Based on the study conducted by
Wagner and Seele (2017), the second hypothesis is as follows:
H2: The CSR reports not adhering to the G4 Reporting Guidelines feature more instrumental
CSR aspects, as compared to the CSR reports that do adhere to the G4 Reporting
Guidelines.
In order to answer the research questions, a deductive quantitative content analysis was
deemed to be the most suitable method, as the present study is exclusively concerned with
readily accessible and textual CSR communication and does thus not require any survey or
interview research. The content analysis was further chosen, because it allows for the
interpretation of large amounts of content, as the data will be extracted from multiple, lengthy
CSR reports.
Case
The research encompasses one representative company each for three industries, the fashion
(Prada), the energy (ExxonMobil), and the food processing (Nestlé) industry. These three
companies and industries were chosen based on three-fold reasoning. First, all three
companies are officially registered in the GRI sustainability disclosure database, which gives
direct access to each CSR report, as well as a clear indication regarding if any, as well as
which GRI guidelines were followed. The three companies specifically were among the few
who each had four CSR reports published of which two adhere to the G4 guidelines, and two
do that do not. Thus, these three companies correspond to the criteria of the sampling unit.
Second, all three companies classify as transnational corporations and
are thus operating in a globalized environment and the earlier described regulatory vacuum.
Based on their impact on the various communities in which they operate, as well as the global
environment, transnational corporations have been known to experience pressure from
stakeholders to conduct business in a responsible manner (Grougiou, Dedoulis & Leventis,
2013; Sethi, Martell & Demir, 2016).
Third, ExxonMobil, Nestlé, and Prada classify as companies
operating in often socially stigmatized industries that have been in the spotlight for their
negative environmental and societal impact. Key stakeholders such as activist groups,
contribution to, for example, the violation of human rights, corruption, emission of
greenhouse gases, or the pollution of soil, ground and surface waters (Claudio, 2007; Kim,
2015; Pagan, 1986). Thus, annual CSR reports form a valuable tool for such stigmatized
companies to respond to the heightened stakeholder pressures and expectations (Patten &
Zhao, 2014; Sethi, Martell & Demir, 2016).
Sample
The data sampling was based on a purposive sampling strategy and limited to Nestlé, Prada,
and ExxonMobil, as well as registration in the GRI sustainability disclosure database. A
nine-year period was chosen, namely nine-year 2009 to 2017, in order to be able to track differences
between reports that do not adhere to any GRI guidelines and those that adhere to the G4
guidelines. Due to differences in when the corporations began adopting the G4 guidelines, the
specific publishing years differ per company. For each of the three companies, four CSR
reports were retrieved directly from the GRI sustainability disclosure database. Of each of
these four CSR reports per company, two adhere to the G4 guidelines and two do not. Thus,
this creates a total sample size of N=12 CSR reports.
Each individual page of the CSR report was chosen as the unit of analysis.
ExxonMobil’s CSR reports together amount to 185 pages, Prada’s CSR reports amount to 214
pages, and Nestlé’s CSR reports amount to 360 pages. Thus, this creates a total sub-sample of
n=759 pages. However, as the title pages, content pages, appendixes, as well as pages solely
consisting of visual material were excluded, the total amount of pages that were analyzed is
n=591. Individual pages where chosen as the unit of analysis based on two-fold reasoning.
First, the aim of the present research is to analyze the extent to which individual aspects of
allow making inferences regarding how often an aspect was present in one report, not merely
if it was present or not. Second, only a limited amount of transnational corporations had four
CSR reports published on the GRI sustainability database, of which two adhere to the
guidelines, and two that do not. This creates a relatively small total sample size of 12 reports.
Thus, analyzing each page created a larger sub-sample of n=591 CSR report pages.
Procedure
As mentioned briefly in the theoretical framework, the six content indicators identified for
instrumental and political CSR were used for the operationalization of the two concepts
within a codebook (Appendix A). Because the operationalization of political and instrumental
CSR within quantitative research is still in its infant stage, the present research adopted Semetko’s and Valkenburg’s (2000) approach of developing several content indicators which
are then translated into binary, yes/no questions within a codebook. This way, it is possible to
analyze the extent to which instrumental and political CSR are present within the CSR repots,
as well as track any differences between the reports adhering to the G4 guidelines and those
that do not. Additionally, this approach is easily replicable and is able to cope with larger
sample sizes (Semetko & Valkenburg, 2000).
The codebook encompasses a total of 21 questions, to which the coder
has to answer all, except the first two, with either yes (1) or no (0). The codebook is set up as
follows; first, two general questions record to which case and CSR report the page belongs, as
well as if it adheres to the GRI G4 guidelines, or not. Then, the first section of the codebook
was dedicated to measuring the presence of instrumental CSR, which was further divided into
The first sub-section, titled economic actor, encompassed five questions, two aimed at
measuring the degree of financial aspects in the CSR reports and three aimed at measuring the
attribution of responsibility. All questions except for one were transferred and adapted from Valkenburg’s and Semetko’s (2000) research studying the framing of European politics. The
questions included, among other, “Is there a mention of the costs/degree of expense
involved?”, “Does the page suggest that some level of the government is responsible for the
issue/problem?” and “Does the page suggest solution(s) to the problem/issue?”. As noted
earlier, the operationalization of instrumental and political CSR is at its early stage and there
are thus no existing items that could be adopted. Instead, the decision was made to adopt the
items by Semetko and Valkenburg (2000), as they indicate the required aspects.
The second sub-section, titled positivity selector, encompassed three questions aimed at measuring the presence of positive and negative aspects of the companies’ business
conduct or CSR activities. The questions included “Does the page mention, highlight or discuss any positive events, outcomes, or achievements?”, “Does the page mention, highlight
or discuss any negative events, business aspects, or outcomes?” and “Does the page discuss
any activities toward reducing/mitigating the negative environmental and/or societal impacts of the company’s business?”. The latter was transferred from a research conducted by Sethi,
Martell and Demir (2016), which studied the tendency of corporate embellishment in CSR
reports, creating a good fit for the positivity selector indicator. The third sub-section, namely
one-way communicator, was combined with the third sub-section of the political CSR section
and shall thus be discussed later.
The second section was dedicated to measuring the presence of political CSR.
Like the first section, it was divided into three sub-sections based on the content indicators
brought forth in the theoretical framework. The first sub-section, titles quasi-political actor,
encompassed three questions aimed at measuring the presence of, and/or involvement in political initiatives, as well as the attribution of responsibility. The questions included, “Does
the page discuss any political initiatives, events, or programs?”, “Does the page discuss any involvement in political initiatives, events, or programs?”, and “Does the page discuss any
shared responsibilities with the government?”.
The second sub-section, titled deliberator, encompassed
four questions aimed at measuring the presence of deliberative stakeholder engagement, self-regulation, as well as disclosure of information. The questions included, among other, “Does
the page include any stakeholder input or feedback?”, “Does the page discuss any form of
self-regulation?”, and “Does the page disclose any information on internal governance?”. The
questions measuring stakeholder engagement were transferred and adapted from a research
conducted by Manetti (2011), who measured the quality of stakeholder engagement in
sustainability reporting. Additionally, all four questions were extracted and created based on
the results of a research conducted by Wagner and Seele (2017), who found that the GRI
standards are indeed, theoretically, able to encourage political deliberation through asking
corporations to disclose governance information, as well as increasing the level of stakeholder
engagement. Finally, the last sub-section of the codebook was titled direction of
communication, which combined the two final content indicators of instrumental and political
CSR, namely one-way and two-way communicator. For this combined measure, four
questions were created aimed at measuring if, and how, a report either promotes one-way
communication processes, or two-way communication processes. The questions included, among other, “Does the page mention any channels or methods used to engage in stakeholder
dialogue?”, “Does the page actively stimulate and encourage stakeholder dialogue?”, and
“Does the page refer stakeholders to a CSR website, report, or brochure for additional
information or questions?”. The latter was created based on a research conducted by Seele and
Lock (2014), who identified CSR websites, reports, and brochures as one-way communication
tools that are correspondent with the instrumental approach to CSR.
the overall data was coded by an independent coder, resulting in 60 coded pages of the total of
591 pages. Afterwards, the percent agreement, Kappa, as well as KAlpha were calculated for
each variable. All of the 21 items had a Kalpha higher than .60 and 20 out of the 22 items had
a Kalpha higher than .80. The average percent agreement was determined as 91%. Therefore,
the intercoder reliability was sufficient. An overview of the percent agreement, Kappa, and
KAlpha values per variable can be found in Appendix B.
Analyses and Variables
First, it has to be noted that two variables were excluded from analysis entirely, namely the
negativity (B4) and the negativity solution (B5) variables. While they are part of the positivity
selector content indicator, they do not contribute to the overall purpose of indicating the
amount of times a positive aspect of the business conduct or CSR engagement is discussed
within the reports. Thus, to ensure a high content validity, the variables were neither used
within the analyses, nor within the development of scales.
In order to answer the first research question, namely which aspect of either
the political or instrumental CSR is most frequently adopted within the CSR reports, seven
variables were selected for analysis. The seven variables chosen for analysis include the
categorical case indicator variable, as well as the six dichotomous, interval-measured
variables quasi political actor, deliberator, two-way communicator, economic actor1,
positivity, and one-way communicator. Then, a one-way analysis of variance was chosen, as
the aim is to compare the means of these six variables to find out which one is more
frequently present in the 12 CSR reports than the others. Before the analysis was run, the
corresponding assumptions were checked, which revealed that the groups are not of equal size
and equal variances can thus not be assumed. A Welch and a Brown-Forsythe robust test of
equality of means was employed for the further interpretation of the one-way analysis of
variances.
To prepare the data for the second analysis, two exploratory principle component
factor analyses (PCA) were run with an orthogonal rotation (Varimax). The purpose of these
analyses was the development of scales based on the underlying dimensions of political and
instrumental CSR. This method was adopted from the study conducted by Semetko and
Valkenburg (2000), who employed a PCA analysis to develop scales for the measurement of
different news frames. Based on their approach to the PCA analysis, a factor loading
threshold of .50 was also adopted in the present study, as well as an orthogonal rotation. The
reasoning for adopting Semetko and Valkenburg’s (2000) method for the development of
scales is two-fold. First, all the items used to measure political and instrumental CSR are
binary, yes-no categories. As Semetko and Valkenburg (2000) note, binary data are often
measured with more measurement error, resulting in overall lower correlations between
variables and thus an increased risk in masked underlying factor structures. Their choice of a
PCA analysis with orthogonal rotation nonetheless resulted in a clearly identifiable factor
structure. Thus, to decrease the risk of deflated correlations based on binary data, the choice
for the present research was also made for a PCA analysis. Second, political and instrumental
CSR are both operationalized with three content indicators each. While all content indicators
are operationalized to measure their corresponding concept, they each capture varying aspects
of them. For example, the three instrumental CSR content indicators measure, among other,
direction of communication, attribution of responsibility, finances, and positive aspects of
CSR. Thus, while the variables are expected to relate highly with each other, as they measure
the same concept, they are not expected to cluster on the same factor, as they each capture a
different underlying dimension of either political or instrumental CSR.
The first PCA analysis was run with the seven items used to measure the economic
The analysis yielded a factor solution in which the nine items clustered into three components.
The factor solution, which explained 53.9% of the variance of the items, is presented in table
1. Only the items with factor loadings higher than .50 were included in the development of the
instrumental CSR scales. As Table 1 shows, the item B2 (government responsibility) did not
meet this threshold. The item is excluded as Semetko and Valkenburg (2000) note, factor
loadings lower than .502 indicate that an item is “empirically and conceptually more distant
to the remaining items that loaded on the same factors” (p.99).
Based on the results, three dimensions of instrumental CSR were instituted, which are
here labeled as economics, one-way communications, and positive aspects. These three
dimensions were then translated into scales. The first scale, measuring the dimension
economics, is composed of items B1 and B1.2 and ranges from 0 to 2 (M=.41, SD=.71). The
second scale, measuring the dimension one-way communications, is composed of items B2.2
and D1.1 and ranges from 0 to 2 (M=.05, SD=.23). The third scale, measuring the dimension
positive aspects, is composed of items B2.3 and B3 and ranges from 0 to 2 (M=.48, SD=.53).
For all scales, a score of 0 equals that none of the items are present and a score of 2 equals
that all items are present.
Table 1: Factor loadings instrumental CSR scale items
1 2 3
Economics One-way comms. Positive aspects Finance .858 .037 .071 Finance Specific .876 -.054 -.036 Govern. Resp. .093 -.158 .438 Stake. Resp. -.062 .645 .023 Problem Solution .274 .087 -.624
Positive .140 .265 .671
One –way .062 .752 -.091
The second PCA analysis was run with the ten items used to measure the
quasi-political actor, deliberator, and two-way communicator content indicators of quasi-political CSR.
The analysis yielded a factor solution in which the ten items clustered into four components.
The factor solution, which explained 62,5% of the variance of the items, is presented in table
2. Only the items with factor loadings higher than .50 were included in the development of
the political CSR scales. As Table 2 shows, the item C3 (shared/full responsibility) did not
meet this threshold and is thus excluded from the development of the scale.
Based on the results, it was possible to institute four dimensions of
political CSR, which are here labeled as two-way communications, political connectedness,
disclosure, and collaboration. These four dimensions were then translated into scales. The first
scale, measuring the dimension two-way communications, is composed of items D1.2 through
D3 and ranges from 0 to 3 (M=.17, SD=.53). The second scale, measuring the dimension
political connectedness, is composed of items C1 and C2 and ranges from 0 to 2 (M=.15,
SD=.48). The third scale, measuring the dimension disclosure, is composed of items C6 and
C7 and ranges from 0 to 2 (M=.56, SD=.75). The fourth scale, measuring the dimension
collaboration, is composed of items C4 and C5 and ranges from 0 to 2 (M=.33, SD=.56). For
all scales, a score of 0 equals that none of the items are present. For the political
collectedness, disclosure, and collaboration scales, a score of 2 equals that all items are
present, whereas for the two-way communications scale, a score of 3 equals that all items are
Table 2: Factor loadings political CSR scale items
1 2 3 4
Two-way comms. Political connect. Disclosure Collaboration Mention Politics 0.60 .922 -.007 .041 Invol. Politics -.047 .906 .059 .084 Shared/Full Resp. -.170 .063 .212 .340 Feedback .170 -.091 -.107 .769 Collaboration .076 .157 .052 .706 Self-regulation .038 -.017 .844 .083 Disclosure .065 .055 .839 -.019 Two-way comms. .850 .015 .039 .073 Stimulation .644 -.029 .047 -.068 Evidence .770 .040 -.006 .145
Then, a reliability analysis was run with the items of the seven, newly computed
scales. The political connectedness scale was found to be highly reliable (α=.805), while the two-way communications scale (α =.652), the disclosure scale (α =.623), and the economics
scale (α =.699) were found to be moderately reliable. However, the one-way communications scale (α=.103), the positive aspects scale (α =-.085), as well as the collaboration scale (α
=.338) were found to be relatively unreliable. An explanatory reason for the low reliability of
the three scales is that the operationalization of instrumental and political CSR is still in its
early stages and is thus likely to not yet represent all underlying dimensions of the two
concepts sufficiently. To increase the reliability of the scales, future research should add more
items measuring the positive aspects, collaborations, and one-way communications
dimensions in order to capture them more exhaustively.
All seven scales were then used within the analysis
aimed at answering the second research question, namely if there is a relationship between the
adoption of the G4 guidelines and the presence of either instrumental or political CSR in the
reports. To answer the second research question, a t-test was deemed most suitable, which
was run twice with the G4 guideline variable as the grouping variable. The first t-test analysis
featured the three instrumental CSR scales as dependent variables, which represent the three
dimensions of instrumental CSR, namely economics, one-way communications, and positive
aspects. The second t-test analysis featured the four political CSR scales as dependent
variables, which represent the four dimensions of political CSR, namely two-way
communications, political connectedness, disclosure, and collaboration. The independent
samples t-test was deemed most suitable since the grouping variable is dichotomous and the
aim is to detect a difference between two groups, namely the reports adhering to G4
guidelines and those that do not. Additionally, even though the eight scales used within the
test are inherently ordinal and thus violate the assumption of normality, they are still suitable
for the independent samples t-test, based the overall large sample size of N=591.
A one-way analysis of variance was employed to answer the first research question, namely
which aspect of either political or instrumental CSR is most frequently present within annual
CSR reports. As table 3 shows, the deliberator content indicator was found to be most
frequently present within the CSR reports, F(2,588)=5.32, p=.005, n²=.02. The positivity
selector was found to be the second most frequently present, F(2,588)=8.58, p=.001, n²=.02.
Then, in descending order, came the economic actor, F(2,588)=22.24, p=.001, n²=.07,
quasi-political actor, F(2,588)=14.11, p=.001, n²=.05, and the one-way communicator,
F(2,588)=11.44, p=.001, n²=.04 content indicators. The two-way communicator content
indicator, however, did not show a statistically significant difference in means.
Table 3: Most frequently present content indicators
Mean S.D N Deliberator .56 .51 591 Positivity selector .43 .51 591
Economic actor .33 .47 591
Quasi political actor .15 .36 591
Two-way Communicator .11 .31 591
One-way communicator .04 .20 591
Because the aim of the analysis was to analyze the difference in means based on all of
the reports, rather than analyzing the difference in means between them, conducting a
post-hoc analysis was not necessary. Thus, based on these findings, the first hypothesis stating that
the positivity selector will be most frequently present within the CSR reports has to be
rejected, as the deliberator content indicator was found to be the most frequently present
instead. To answer the second research question, namely if there is a
CSR, two independent-sample t-tests were conducted. An overview of the descriptive
statistics for the seven dimensions of instrumental and political CSR can be accessed through
table 4. The results of the first t-test analysis, which can be accessed through tables 5A and
5B, show that two of the three dimensions of instrumental CSR showed a statistically
significant difference in means, namely the economics and the positive aspects dimensions.
First, the reports not adhering to the GRI G4 guidelines score higher on the scale measuring
the economics dimension (M=.52, S.D.=.70) than those reports that do adhere to the
guidelines (M=.35, S.D.=.68). Second, the reports not adhering to the GRI G4 guidelines also
score higher on the scale measuring the positive aspects dimension (M=.57, S.D.=.50) than
those reports that do adhere to the guidelines (M=.43, S.D.=.54). For both, the Cohen’s d
value (economics scale, d = .25; positive aspects scale, d = .27) suggested a small effect size.
Table 4:Descriptive statistics of the seven dimensions of instrumental and political CSR
G4 guidelines present G4 guidelines absent
The dimensions Mean S.D. Mean S.D.
Instrumental CSR Economics .35 .68 .52 .70 One-way comms. .05 .22 .05 .26 Positive aspects .43 .54 .57 .50 Political CSR Two-way comms. .19 .58 .13 .42 Political connected. .13 .47 .17 .49 Disclosure .61 .78 .46 .71 Collaboration .35 .60 .30 .49
Table 5A: Independent samples statistics: Economics scale 95% confidence interval of difference M SD Lower Upper t df p G4 guidelines present .35 .68 0.050 0.281 2.796 458.607 .005 G4 guidelines absent .52 .70
Table 5B: Independent samples statistics: Positivie aspects scale
95% confidence interval of difference M SD Lower Upper t df p G4 guidelines present .43 .54 0.057 0.233 3.223 589 .001 G4 guidelines absent .57 .50
The results of the second analysis, which can be accessed through table 6, show that
only one of the four scales measuring the dimensions of political CSR showed a statistically
significant difference in means, namely the scale measuring the disclosure dimension. The
reports adhering to the GRI G4 guidelines score higher on the scale measuring the disclosure
dimension (M=.61, S.D.=.78) than those reports that do not adhere to the guidelines (M=.46,
S.D.=.71). Cohen’s d value (d = .2) suggested this to be a small effect size.
95% confidence interval of difference M SD Lower Upper t df p G4 guidelines present .61 .78 -0.274 -0.031 -2.462 509.989 .014 G4 guidelines absent .46 .7
Concluding, the results of the two t-test analyses offer support for the second, and
final, hypothesis of the present research, which states that the CSR reports not adhering to the
G4 Reporting Guidelines feature, on average, more instrumental CSR aspects, as compared to
the CSR reports that do adhere to the G4 Reporting Guidelines.
Discussion and Conclusion
The present research aimed at achieving three goals. First, to incorporate literature that has
thus far been fragmented from each other together into a clear and contrasting overview of
instrumental and political CSR. Second, to operationalize political CSR in a quantifiable way,
a concept that has thus far remained largely normative. Third, to find out which of the two
concepts, instrumental or political CSR, is more frequently present within the CSR reports.
Finally, the fourth goal was to take a closer look at the effects of the GRI G4 guidelines on
annual CSR reports, or, in other words, to find out if the G4 guidelines stimulate a political
approach to CSR. Consequently, the present research contributes to the field of CSR research
through not only offering a way to operationalize political and instrumental CSR in a
quantifiable and comparable way, but through also producing new empirical evidence of the
extent to which the two concepts are present within the CSR reports. Additionally, while the
present research is the first to test this hypothesis through analyzing actual CSR reports.
The present research features two research questions. The first question asked
which aspect of either instrumental or political CSR is more frequently present within the
CSR reports. Contrary to the expectations, the corresponding hypothesis that the positivity
selector is more frequently present than the economic actor, the quasi-political actor, the
one-way communicator, the two-one-way communicator, and the deliberator content indicators had to
be rejected. Instead, the results show that the deliberator content indicator was more
frequently adopted than the other content indicators. There are two explanatory reasons for
this result. First, the deliberator content indicator is composed of items that deal, among other,
with stakeholder engagement and the disclosure of internal governance information.
Stakeholder engagement and the disclosure of internal governance information correspond to
what Wagner and Seele (2017) note as extensions that were made to the fourth version of the
GRI guidelines in light of the new emphasis on materiality. The G4 guidelines thus require
corporations to report on these aspects in order to receive an ‘in accordance’. As half of the
examined 12 CSR reports adhere to the G4 guidelines and thus have to feature information on
these two aspects, the high score of the deliberator scale can be rationalized. Second, all
examined CSR reports were published by large, well-known, and transnationally operating
corporations, namely Nestlé, ExxonMobil, and Prada. As was argued before, their overall
high visibility, as well as high environmental impact, increases the experience of pressure
from stakeholders to pro-actively disclose information to maintain legitimacy (Kilian &
Hennings 2014; Sethi, Martell & Demir, 2016). As Ehrnström-Fuentes (2016) argues, moral
legitimacy is indeed achieved through engaging in deliberation. However, it is also possible to
view this finding from a more critical perspective, as this increased commitment to
deliberative CSR communication by high-impact corporations was criticized for not being a “true commitment,” but rather a way to “hide what they really do in their daily business”
there is a relationship between the GRI G4 guidelines in the CSR reports and the use of
either the instrumental or political approach to CSR. The findings offer support for the
corresponding hypothesis, namely that the CSR reports not adhering to the G4 Reporting
Guidelines score higher on the instrumental CSR scales, as compared to the CSR reports that
do adhere to the G4 Reporting Guidelines. More specifically, the findings show that the
reports not adhering to the guidelines score higher on the economics and positive aspects
scales, whereas the reports adhering to the guidelines score higher on the disclosure scale. By
taking a closer look at these three scales that showed a significant difference in means, several
inferences can be drawn. First, the finding that the G4-adhering reports score higher on the
disclosure scale fits into the line of reasoning already suggested for the finding of the first
research question. The disclosure scale is composed of two items indicating self-regulation
and the disclosure of internal governance information. As previously discussed, the disclosure
of internal governance information is an aspect required from corporations who adopt the G4 guidelines and seek to receive an ‘in accordance’ for their CSR report. Thus, based on this
finding, it can be argued that the GRI G4 guidelines are generally successful in increasing the
disclosure of information in CSR reports that have previously not followed any guidelines.
Second, the finding that
the non-G4 adhering reports score higher on the positive aspects scale and the economics
scale than the G4-adhering reports, reflects two common characteristics of instrumental CSR.
Namely, that a corporation positions itself as a solely economic entity (Baron, 1995; Lock &
Seele, 2017) and that investments in social demands are deemed appropriate only if they do
not impose costs on the firm (Garriga & Melé, 2004). Quite contrary, CSR in this perspective
is often instrumentalized as a tool for value creation, namely through image enhancement and
reputation management (Pollach, 2015; Sethi, Martell & Demir, 2016). This approach to CSR
was found to be the most prominent among practitioners in a study by Pollach (2015), who
from the competition, thus ultimately creating value for the firm. The finding that the non-G4
reports specifically score high on the positive aspects scale and the economics scale suggests
that, prior to adopting the G4 guidelines, also Nestlé, ExxonMobil, and Prada approached
CSR from a predominantly value-oriented perspective. More specifically, the high score on
the positive aspects scale suggests that, prior to adopting the G4 guidelines, there was a focus
on image enhancement and reputation management. Finally, these findings offer new
insights to the research conducted by Wagner and Seele (2017), who found that the G4
guidelines are, on a theoretical level, able to stimulate political CSR within annual CSR
reports. However, after analyzing actual CSR reports, as opposed to the guidelines
themselves, the present research found that the G4 guidelines only stimulate one aspect of
political CSR, namely the disclosure of information. It can thus be argued that instead of
stimulating political CSR, the GRI G4 guidelines decrease the usage of instrumental CSR.
More specifically, after adopting the guidelines, the reports by Nestlé, Prada, and ExxonMobil
discuss significantly less positive and financial aspects of their business conduct and CSR
engagement. Thus, to revisit the earlier stated critics’ question if the guidelines actually
facilitate a shift away from impression management (Barkemeyer et. al., 2014), the present
research suggests that such a shift did indeed take place.
Thus, although the findings did not confirm that the G4 guidelines stimulate a
predominantly political approach to CSR, it is notable that the process of standardization is
still promising for increasing the overall quality of the reports. Through decreasing the level
of instrumental CSR, but increasing the level of disclosure, the G4 guidelines help to prevent
the exploitation of CSR reports as mere “rhetorically persuasive instruments” (Schultz,
Castello & Morsing, 2013, p.683) and instead promote the overall accountability and
transparency of CSR reporting. Because as Wagner and Seele (2017) note, only full
and the possibility for stakeholders to “judge whether companies really ‘walk the talk’”
(p.336).
Limitations and implications for future research
Several limitations of the present research have to be considered, which, however,
simultaneously offer new insights and implications for future research. First, the scope of the
research was limited to a total of 12 CSR reports, which affects the overall generalizability of
the results. Connected to this is the fact that only CSR reports were analyzed, which does not
exhaust the full range of CSR communication activities companies employ. Thus, future
research should move beyond the exclusive analysis of CSR reports and also take, for
example, CSR websites and social media channels into account. This would also improve the
validity of the results, as the research included items measuring two-way communication,
while analyzing a medium that is inherently linear.
Secondly, it has to be noted that the operationalization of instrumental
and political CSR is still at its infant stage. Notable advancements were made in this field by,
for example, Wagner and Seele (2017), who offered valuable first insights into the
operationalization of political CSR within qualitative research, or Seele and Lock
(2014;2017), who developed a typology of instrumental and political CSR communication
tools. Nonetheless, the present research was the first to attempt the quantifying
operationalization of instrumental and political CSR through a codebook and the development
of various content indicators. This directly affects the validity of the items in the codebook, as
well as the later developed scales. Consequently, future research should further develop the
operationalization of the two concepts for quantitative research by, for example, extending the
herein created codebook and scales. Third, and closely connected to
the operationalization of the two concepts, instrumental and political CSR classify as both,
a codebook sufficiently, which, in the end, featured only simple, binary yes/no categories.
This had an impact on the results of the research, as the coding of the CSR reports only
revealed if an aspect of either instrumental or political CSR was present, not how exactly it
manifested itself textually. Consequently, even though political CSR and “communication are
inherently linked with each other, which makes a political CSR perspective an adequate tool
to analyze corporate communication practices” (Wagner & Seele, 2017, p.334), a qualitative
research design could potentially offer more possibilities for analysis and conclusive
inferences. Thus, a suggestion for future research would be the testing of the hypotheses
brought forth in the present research within a qualitative research design.
Practical implications
Finally, the findings of this research have several practical implications. First, while the GRI
G4 guidelines were found to decrease the presence of instrumental CSR within the reports,
they do not yet stimulate a political approach to CSR. Thus, GRI should continue to extent
and adapt the guidelines with a clearer focus on the stimulation of political CSR. This could
be achieved through, for example, adding more indicators to the guidelines that deal with the
creation of dialogue opportunities with stakeholders and the inclusion of stakeholder feedback
within the annual reports. Furthermore, the extended guidelines could also emphasize the
collaborative aspects of political CSR through, for example, adding indicators that encourage
the participating corporations to actively reach out to third parties for voluntary audits and
feedback on business conduct and CSR engagement. This way, the guidelines could further
stimulate corporations to continuously self-evaluate and self-regulate, rather than taking
advantage of the regulatory vacuum in which they might operate (Ehrnström-Fuentes, 2016,
p.434). Second, the findings also stress the importance of standardizing the
practice of CSR in a more general sense. While the GRI guidelines have increased the overall
unaudited (Kilian & Hennigs, 2014). This voluntary nature of CSR reporting contributes to
larger cross-country differences between the reports and stimulates the tendency of many corporations to “embellish their reports with ‘lightweight’ information,” while not discussing
issues that could be of greater importance for stakeholders (Sethi, Martell & Demir, 2016,
p.221). This is also reflected within the findings of this research, as the reports that do not
adhere to any guidelines feature more positive aspects of their business conduct and CSR
engagement than those that adhere to the G4 guidelines. In light of the current climate crisis
and the urgency for action connected to it, distributing lightweight information will not allow
stakeholders to hold corporations accountable for harmful behaviors. Instead, the mandatory
adoption of the GRI G4 guidelines on a global level could potentially contribute to reducing
harmful corporate behavior through requiring corporations to discuss them publicly. Because,
as Patten and Zhao (2014) argue, “as long as the reporting remains within a voluntary regime,
it will never be a tool of accountability” (p.142).
References
Aras, G. A., & Crowther, D. (2008). Corporate Sustainability Reporting: A Study in Disingenuity? Globalization and the Good Corporation, 279-288. doi:10.1007/978-94 007-0818-1_19
Barkemeyer, R., Comyns, B., Figge, F., & Napolitano, G. (2014). CEO statements in sustainability reports: Substantive information or background noise? Accounting
Forum, 38(4), 241-257. doi:10.1016/j.accfor.2014.07.002
Cai, Y., Jo, H., & Pan, C. (2011). Doing Well While Doing Bad? CSR in Controversial Industry Sectors. Journal of Business Ethics, 108(4), 467-480. doi:10.1007/s10551-011-1103-7
Carroll, A. B. (2015). Corporate social responsibility. Organizational Dynamics, 44(2), 87-96. doi:10.1016/j.orgdyn.2015.02.002