The effect of Board Gender Diversity and Board Independence on CSR Decoupling

The effect of Board Gender Diversity

and Board Independence on CSR

Decoupling

Abstract

With the growing importance of corporate social responsibility (CSR) for firms a question arises whether these firms follow through on their promises regarding CSR. Due to the gap in the current literature it cannot be explained why firms often have strong differences in their CSR performance and CSR disclosure. This study examines the gap that exists for many firms between their CSR performance and CSR disclosure, called CSR decoupling, and tries to find possible influences on the size of the gap. The firms are given a score for their CSR performance and CSR disclosure by social rating agencies, and the gap between the scores is taken as the amount that a firm decouples. In particular this study focusses on different board characteristics and their effects on CSR decoupling, board gender diversity and board independence, and the combined effect of the two are looked at more thoroughly. With a sample of 10972 firm years of firms from all over the world it is found that all three variables show some significance towards CSR decoupling, but that the results are quite inconsistent, leading to no conclusive answer to what their effects on CSR decoupling really are. This research does however find promising results that could spark interest for future research on what affects the amount of CSR decoupling in a firm.

Student name: Timon Scholte Lubberink Student number: S2756048

Supervisor: Dr. N. Hussain Date: 24-6-2019

1. Introduction

In today’s day and age firms care a lot about their image, corporate image after all, has been recognized as an important factor in the success or failure of many major organizations (Worcester, 2009). One way for a firm to change their image is by embracing corporate social responsibility. Corporate social responsibility in its simplest form is corporations’ broader responsibility towards society (Rao & Tilt, 2016). It is expected of firms to not only contribute to the world economy but also manage the interests of multiple stakeholders (Jamali, Safieddine & Rabbath, 2008). In recent years there has been an increase in popularity for firms to publish CSR reports. In these reports firms can report their progress on CSR programs focussing on issues like sustainability, equality and labor conditions within the firm.

While CSR reports mean to show a company’s intent on doing their part for society, many people believe that these companies are profiting from insincere claims of corporate social responsibility (Mattis, 2008). Firms may adopt CSR policies, but the implementation of these policies may not happen, this is called CSR decoupling. CSR decoupling is the divergence among CSR policies, implementation of CSR programs, and CSR impacts for various environmental and social issues (Graafland & Smid, 2016). Graafland and Smid (2016) further state that complete decoupling is a condition of full divergence among policies, programs, and impacts amounting to purely ceremonial CSR. Big CSR scandals are well reported in the news and this seems to fuel people’s opinions on company’s CSR programs. One example of a recent CSR scandal can be found by looking at Nike. Nike, one of the biggest clothing brands in the world has an extensive CSR program where they focus on sustainability and equality. However at the start of 2018, Nike was involved in a scandal where high-profile employees were demeaning to women and people from other countries (Bloomberg, 2018). It is said that executives knew about this and did not act accordingly while this all happened. So while the CSR reports indicate the company’s ambitions for equality, the actual mentality within the company seems vastly different. The disconnect between what is being said in the CSR reports and what the actual company culture is like brings the question whether CSR reports are actually a representation of what the company stands for, or whether it is just sweet talk. Aguinis & Glavas (2012) find that many firms partake in ‘symbolic rather than genuine actions’ when it comes to CSR and they decouple to appease stakeholders. This means that firms don’t fail at effectively implementing their CSR programs, but they had no intention to go further than talking about the programs, since implementation can be very expensive (Behnam & MacLean, 2011). This sentiment is shared by some stakeholders and investors, who are sceptical of the credibility and reliability of the disclosed information (Braam & Peeters, 2017). In this case firms want to please stakeholders by sharing their future CSR programs, while not actually wanting to go through the trouble of implementing them. The advantage of doing this is that the corporate image improves, since stakeholders want to see these programs implemented in the firms, but the firms actually don’t intend to implement these improvements. The short term advantage can however damage corporate image in the long run since, while decoupling might not be intentional, doing so can have a big negative reputational impact (Pope & Wæraas, 2016). The initial advantage of pleasing the stakeholders will be negated the negative consequences of stakeholders who monitor the firm finding out about the decoupling (Sauerwald & Su, 2019).

The decoupling of CSR can however be unintentional, many companies simply are not able to implement all their CSR programs effectively. One of the reasons for this is that due to rapidly changing stakeholder pressures, firms combine rapid adoption with low-quality implementation (Luo, Wang & Zhang, 2017). The changing stakeholder pressures also lead companies to adopt inconsistent standards that cannot be implemented simultaneously (Meyer & Rowan, 1977). This inconsistency between policies and implementation leads to decoupling, which is often not the intention, but rather the outcome of uncoordinated decision making within the firm (Crilly, Zollo & Hansen, 2012). In some cases firms decouple when different stakeholder groups press for conflicting policies, firms will then try to mitigate conflict with both stakeholder groups (George, Chattopadhyay, Sitkin & Barden, 2006).

Decoupling can also go the other way around, Lyon and Montgomery (2013) find that firms that are considered to be environmentally conscious disclose less of their smaller achievements, since there is little to be gained from this disclosure. Kim and Lyon (2015) describe this as undue modesty where instead of exaggerating their CSR performance (greenwashing) they brownwash. Bansal and Clelland (2004) find that firms with a high environmental legitimacy can expose unnecessary levels of unsystematic risk by disclosing more of their CSR activities. This leads many of these firms to not disclose some of their achievements at all (Morsing & Schultz, 2006). Bromley, Hwang and Powell (2011) explain that CSR practices can be implemented effectively, but if the achievements are not part of the firm’s primary goals, reporting them may be ineffective. Managers also worry that by promoting all their CSR activities, and not just the most important contributions, stakeholders might view their acts as self-serving (Peloza, 2005). Hawn and Ioannou (2016) also find that firms perform more CSR actions than they disclose, showing that firms refrain from disclosing all of the CSR activities they perform.

With the limited amount of research on CSR decoupling not all is known on what affects the amount of CSR decoupling in a firm. Graafland and Smid (2016) however, show that when the board is made responsible for the CSR goals, that CSR decoupling is reduced, and CSR reporting is stronger. What is seen here is that the board seems to significantly influence how much CSR decoupling takes place. Having established the role of the board, we want to know whether different board characteristics can influence the CSR decoupling further. In this research we will be looking at a possible relationship between the gender diversity of the board and the independence of the board with CSR decoupling. Gender diversity of the board is an increasingly popular topic in the literature, since many countries are introducing legislation for companies to have a minimum amount of women on the board. With the introduction of these kind of legislations it is therefore interesting to see the effects that altering the gender diversity of the board could have on a company. Independence of the board is already a popular topic in the literature, with some existing legislature on firms needing outside directors. With independence of the board being pushed by legislature it is interesting to see whether having more independent directors alters the amount of decoupling that takes place.

The research question is therefore: What is the effect of gender board diversity and board independence on CSR decoupling.

All the data regarding the CSR and financial information was collected from the Thomson Reuters’ ASSET4 database, while the board characteristics were taken from the BoardEx database, leading to a sample of 10972 firm years from firms all over the world. The findings

show that board gender diversity, board independence and the combination of both all show some significant relationships towards decoupling in different settings, however the hypothesized relationships can largely not be accepted, since the findings are inconsistent across the board. The study does show promising signs of potential influences on the amount that firms decouple.

The contribution of this study is twofold, gender diversity of the board and independence of the board are well discussed topics, however not much previous research has been done on the potential effect that they might have on CSR decoupling. While their effects on other aspects of CSR are widely discussed in the literature, this study provides some insight into what board and firm characteristics influence CSR decoupling and its components.

On the other hand this study contributes to stakeholders and the companies themselves. CSR decoupling is a new phenomenon that many companies will deal with, but will want to avoid doing. The results of this research could help firms and stakeholders find potential influences that can help minimize decoupling, and increasing the firm’s transparency.

2. Literature Review

Prior literature on CSR in firms is quite wide-ranging, with many different CSR aspects that have been studied. The focus here however was to find previous articles on CSR decoupling, greenwashing, brownwashing and to a lesser extent CSR disclosure and CSR performance. In table 1 a range of articles that have been reviewed for the theoretical background of this research are shown, the theories used in these studies can range from agency theory, institutional theory, organizational theory and stakeholder theory. The studies also differ in the measurement they use, for CSR decoupling specifically, there are two different measurements. The first measurement is the difference between internal and external actions (Hawn & Ioannou, 2016; Schons & Steinmeier, 2016) where the difference is equal to the decoupling. The second measurement has recently gained popularity, here researchers use data from CSR rating agencies to find out the extent to which firms decouple their CSR (Pope & Wæraas, 2016). By using the data from these agencies very large samples can be used to find relationships between a variable and the effect it has on a firm's CSR decoupling. The data is gathered by experts who objectively give firms a score on how close their ‘talk’ is to their ‘walk’.

Companies experience external pressure when it comes to their strategy and CSR choices. One way to explain this external pressure is through stakeholder theory (Freeman, 1984). Stakeholders are all of the agents for whom the firm’s development and good health are of prime concern (Pesqueux & Damak-Ayadi, 2005). This includes customers, suppliers, shareholders, employees, special interest groups, government authorities and the media (Clement, 2005). Scholars tend to regard organizational CSR-related decision making as an externally determined process (Vashchenko, 2017). This means that organizations’ decisions are heavily influenced by the external pressures from their stakeholders. Firms want to have a good relationship with their stakeholders to reap the benefits that come with this like attracting employees and retaining them more easily or being granted preferential investment opportunities (Cecil, LaGore, Mahoney & Thorne, 2013). Being able to satisfy all groups that have a stake in the firm is essential in maximizing firm value (Freeman, Wicks & Parmar, 2004). Cornell and Shapiro (1987) further suggest that the relationship between stakeholders and a firm can be seen as a contract and that a firm’s value depends on its ability to fulfil the contract. Weaver, Treviño & Cochran (1999) suggest that external pressures can easily lead to decoupled practices, but that top management commitments can further encourage, or discourage decoupling. Their research shows that managements’ commitments can significantly influence decoupling in a firm.

With the increasing awareness that people have for the environment, equality, labour conditions and various other social issues, firms want to inform their stakeholders on their CSR actions to keep a good relationship (Greenwood & Lim, 2017). Firms are very dependent on the resources that stakeholders provide them, like customer loyalty, employee expertise and capital from investors, if these resources were to be withdrawn from the firm it could suffer from financial hardships (Harjoto et al. 2015). To keep stakeholders satisfied firms will commit to CSR and various other actions that are in the interest of the stakeholders so that they conform to their norms of what is appropriate behaviour (Clarkson, 1995). With the boards of firms being the decision makers, they can directly influence the relationship with their stakeholders by choosing to what extent they are willing to fulfill the expected social responsibilities (Jain & Jamali, 2016).

Corporate governance can be labelled as the processes and operating relationship that best achieve organisational goals (Chakrawal, 2006). Furthermore it is the acceptance of management that they are the trustees for the true owners and beneficiaries of the firm (Chakrawal, 2006). Corporate governance therefore keeps the decision makers of a firm in check, so that they do what is best for the firm’s continuity. With stakeholder interest for CSR rising, corporate governance mechanisms need to ensure that the board will follow through with the expected CSR programs. Many of the reviewed studies try to find causes and motivations for CSR decoupling, but the role of the board is often overlooked. Characteristics of the board have shown to be of big influence on the way a firm is being run. Gender diversity on the board significantly improves board effectiveness (Adams, de Haan, Terjesen & van Ees, 2015). Furthermore these boards allocate more time on board monitoring, supporting the idea that gender diverse boards are more independent from top management (Adams & Ferreira, 2009). Managers are often presented with choices that could either bring personal benefits or benefits for the firm as a whole, and firms that have boards that are efficient at monitoring the managers see their managers choose the option that is best for the firm more often (Loureiro, 2012). Another way of influencing the way the board operates is by appointing more outside

directors. Outside directors are meant to ensure that firms act in the best interest of the owners and stakeholders (Haniffa & Cooke, 2005). Outside directors are also found to improve the decision making of the board (Dahya & McConnell, 2005), making board independence a way for stakeholders to ensure that the board does what they need it to do. An improved monitoring process is therefore beneficial for a firm since it leads to more managers doing what is best for the firm’s continuity.

Schons & Steinmeier (2016) distinguish two different stakeholder groups in their study, high-proximity stakeholders and low-high-proximity stakeholders. High-high-proximity stakeholders are groups like employees, managers and oversight boards, while low-proximity stakeholders are groups like customers, non-governmental organizations and the general public. They argue that because low-proximity stakeholders are not easily able to distinguish between ‘walk’ and ‘talk’, firms who merely perform ceremonial CSR will still see the rewards, while not actually performing the tasks. This shows that decoupling could be a legitimate strategy for firms when they are approaching their low-proximity stakeholders. But while decoupling seems to have a positive effect for a firm whenever it is directed at low-proximity stakeholders, high-proximity stakeholders have the ability to distinguish between ‘walk’ and ‘talk’. Due to this ability, decoupling directed towards the high-proximity stakeholders show no positive effects (Schons & Steinmeier, 2016). Instead large gaps between symbolic and substantive CSR actions significantly diminish the gains received from decoupling.

Lyon & Montgomery (2013) explain that the number of sources of information available to stakeholders and the variety of social media increase the degree of scrutiny on CSR strategies. This makes it easier to detect decoupling and ‘greenwashing’. ‘Green firms’ are therefore less likely to promote their green accomplishments while ‘brown firms’ are more inclined to promote their full environmental impact. With the increasing demands for transparency, firms will be less able to reap the rewards of decoupling in the future, since all stakeholder groups will be better able to distinguish between symbolic and substantive CSR actions. Firms will have to focus on being able to follow through on their intentions for improving their CSR, or face the consequences of having angry stakeholders.

3. Hypothesis development

CSR decoupling and board characteristics

From reviewing previous articles on decoupling, greenwashing and brownwashing, it becomes clear that many studies focus on the managerial motivations for and the outcomes of decoupling. There seems to be a lack of studies trying to find potential influences that increase or decrease decoupling. However by looking at studies on both CSR performance and CSR disclosure, the main components of CSR decoupling, it is found that board characteristics have a big influence on both disclosure and performance. Therefore with the help of stakeholder theory I will try to draw a connection that is based on the influence that gender diversity of the board and board independence have on CSR performance and CSR decoupling and hypothesize a potential relationship that these variables might have on CSR decoupling.

Gender diversity of the board

Many recent studies find positive effects that female board members can have on a firm. It is theorized that female board members can provide new perspectives that improve the decision making of the board (Swartz & Firer, 2005). This is also found by Coffrey & Wang (1998) who state that having women on the board leads to better effectiveness of the board. Conyon & He (2017) and Green & Homroy (2018) find that gender diversity in the board improves firm performance. Boards with female directors also publish more non-financial information (Jizi, 2017). Not only does having female board members encourage the publishing of more non-financial information, it also leads to a better disclosure of this information (Cabeza-García, Fernández-Gago & Nieto, 2018). Aside from the disclosure, female board members also are more attuned to stakeholders’ interest in the firms’ commitment to corporate social responsibility (Cook & Glass, 2015).

Overall there seems to be a positive influence that female board members have on firms. While there is currently no research that goes into the relationship between gender diversity of the board and CSR decoupling, these previous studies give an inclination that a relationship is possible. CSR decoupling can be seen as a negative influence on a firm, because of the reputational risk that it is associated with. Therefore it is expected that the relationship between the gender diversity of the board and CSR decoupling is negative. This means that more women on the board of directors should lead to less CSR decoupling. This leads to the following hypothesis:

H1: There is a negative relationship between gender diversity of the board and CSR decoupling. Independence of the board

Independence of the board is an often discussed topic when looking at board characteristics that affect the way a firm is run, and often leads to positive results for a firm. García-Meca and Sánchez-Ballesta (2010) find that when the board consists of more independent directors, that voluntary disclosure increases. Fama and Jensen (1983) argue that with a larger proportion of independent directors, boards will be more effective in monitoring managerial opportunism, and it can be expected that voluntary disclosure increases. Not only disclosure seems to improve for firms with outside directors, Cabeza-García et al. (2018) find that outside directors

can affect CSR initiatives as well. Furthermore Dah and Jizi (2018) find that board independence not only facilitates firm’s CSR reporting, but also positively influences their CSR performance.

Outside directors are meant to ensure that firms act in the best interest of the owners and stakeholders (Haniffa & Cooke, 2005). With the specific characteristics that come with an outside director, like having more experience and their independence from the CEO and other executives, outside directors are especially significant in corporate social activities (Hafsi & Turgut, 2013). While the research on CSR decoupling is sparse, with all the known effects of independent boards on CSR activities and disclosure, it can be theorized that board independence lowers a firm’s CSR decoupling. This leads to the following hypothesis:

H2: There is a negative relationship between board independence and CSR decoupling. The combined effect of board gender diversity and board independence

Board gender diversity and board independence seem to both have a positive effect on a firm’s CSR activities and disclosure. There is however a possibility that these effects might be strengthened or weakened when applied at the same time. Some studies look for a combined effect of these two variables, like Post, Rahman and Rubow (2011) who find that firms with more outside directors, and more than three or more female directors have higher scores for environmental corporate social responsibility (ECSR) and Kinder Lydenberg Domini (an environmental rating). Jizi (2017) finds that both gender diversity and independence of the board help maximize value from social projects and favourable affecting CSR engagement. Liao, Luo and Tang (2015) find that whenever a board is both sufficiently diversified and independent that it may be able to balance the conflicting expectations of their stakeholders. Differing stakeholder pressures may lead firms to decouple (George et al., 2006; Bartley & Egels-Zandén, 2016), therefore by testing the combination of both independence and gender diversity of the board there might be a stronger relationship. This leads to the following hypothesis:

H3: The combination of board independence and gender diversity of the board will have a negative effect on CSR decoupling.

4. Methodology 4.1 Sample

The data used in this research will come from two sources. First there is archival data from the BoardEx database, which will encompass all the data used to see the effect of board gender diversity and board independence. Second there is archival data from Thomson Reuters’ ASSET4 database, which will encompass all the data used to find the CSR decoupling. The BoardEx database is a consolidation of publicly available information concerning directors on the boards of public and large private firms. The use of this database is widespread among academic research, and the data is verifiable as it is collected from the firms themselves, and thus it is very reliable. The ASSET4 database consists of information on a firm’s quality of CSR policies, use and quality of implementation of these policies, quality of reporting and the impacts of CSR policies. This information is gathered by professional social rating agencies, who independently assess a firm’s performance on these criteria, and has been used by other researchers before (Hawn & Ioannou, 2016; Schons & Steinmeier, 2016). All the data in this database comes from trained analysts who collect the raw data from public, unbiased sources and transform it into units that can be used for quantitative analysis. With these units firms can be compared on their performance on CSR in an objective way. To calculate the decoupling of a firm the difference between the quality of performance and the quality of disclosure will be used.

All information needed for the control variables will either be collected from these same sources, or from the public financial statements of the firms themselves. All observations that did not have enough information for any of the variables that were used, were cut from the sample. The final sample consists of 10972 firm-year observations, from firms all over the world. Coverage years will be 2009 up until 2017.

4.2 Measures Dependent variable

The dependent variable is the actual CSR decoupling, this will be measured by taking the difference between performance and disclosure. As mentioned before these scores will be given by professional social rating agencies. To get the performance score the social rating agencies divide CSR performance into different criteria like corporate governance, social and environmental performances. To get the scores many different aspects are looked at, but for most aspects it comes down to whether a policy is in place, how well it is implemented, whether the policy is monitored and whether the firm has set objectives for this policy. All of these aspects are individually given a score, and at the end the scores are aggregated and the firm will be given one main score for their CSR performance overall, as well as the scores for corporate governance, social and environmental performances individually. The scores for CSR disclosure will largely follow these same aspects, but will be given scores on whether all the information on these performances are disclosed.

The scores for CSR performance and CSR disclosure are both on a scale from 0 to 100, and differences can have larger implications when looking at the actual scores. A difference of 10 between performance and disclosure will be more significant when the aforementioned are

scored at 10 and 20, than when they are scored at 80 and 90. To make these difference have more of an impact, both performance and disclosure will be standardized and given a Z-score. For both of these variables the Z-score will be the equivalent of the actual score, minus the average, and the result of this will be divided by the standard deviation. After both the performance and disclosure were given a Z-score, the absolute difference between these was taken as the decoupling. The reason for taking the absolute differences, is so that a big positive difference for one of the criteria would not compensate for negative differences on the others. If this were to happen, it would seem like there was no decoupling taking place, while in actuality decoupling was taking place a lot, just in different directions. Initially I want to find the effects of all the variables on decoupling in general, so that later it can be compared with the two directions of decoupling which will be done in the additional analysis.

Independent variables

The first independent variable is the gender diversity of the board, this will be measured by taking the percentage of women on the board of directors.

The second independent variable is the independence of the board, this will be measured by taking the percentage of outside directors on the board of directors. The data for both of these variables has been gathered from the BoardEX database, which gathers the data from public sources that firms publish, ensuring their credibility.

The third independent variable is a combination of the two previous variables. The ratios of both variables were multiplied by each other to create a new variable that can measure the combined effect of gender diversity of the board and independence of the board.

Control variables

Apart from testing the main hypotheses, to ensure robust findings various control variables will be tested in the conceptual framework. These control variables are added to control for various internal and external influences in a firm’s environment. The effect of these variables are not the main focus of the study, but will provide more insight in what causes companies to decouple their CSR.

Implementation of CSR programs can be very costly for a firm, with no immediate pay-off. Organizational slack exists whenever a firm does not fully utilize its resources. Firms with a higher level of organizational slack, can use these extra capabilities to invest into programs that do not have an immediate pay-off (Levinthal & March, 1981). Furthermore Waddock and Graves (1997) find that with better economic performance, there is more availability of slack, which gives firms the opportunity to invest in corporate sustainable development and CSR programs (Bansal, 2005). Economic performance can however also lead to less investment in CSR, since profitable companies are able to handle the reputational damage that comes when they decouple their CSR (Delmas & Burbano, 2011). Economic performance will be measured by taking the return on assets.

Firms may be ineffective when trying to implement their CSR programs, due to rapidly changing stakeholder pressures (Luo, et al. 2017). This indicates that there is a lack of knowledge on effective implementation. This seems especially common in smaller firms, which are often organized on a more informal basis and lack experience in implementing CSR

programs (Lepoutre & Heene, 2006). Furthermore, many small business managers don’t pay attention to CSR or believe that their social and environmental impact is negligible (Petts, Herd, Gerrard & Horne, 1999). Larger firms also tend to face more pressures over their social and environmental impact, and the reporting of these impacts (Christmann & Taylor, 2006). Therefore firm size seems to be an influence on a firm’s effective implementation of CSR programs. The size of the firm will be measured by taking the logarithm of the total revenues. The log of the total revenues will be taken to reduce the effect of outliers.

Board size has been shown to influence the amount of corporate philanthropy that a firm practices (Brown, Helland & Smith, 2006). Corporate philanthropy is seen as an important part of a firm’s CSR practices (Porter & Kramer, 2002), and should therefore be controlled for. Board size will be measured by the amount of directors on the board of directors.

Financial leverage was used as a control variable in the study by Walker and Wang (2012), as it is seen to have an influence on all kinds of performance for a firm. Financial leverage will be measured by dividing the total debt by the total assets.

Hussain, Rigoni and Orij (2016) show that both sales growth and R&D intensity have a significant relationship with CSR performance. With the limited amount of research on CSR decoupling, it is still uncertain what has an effect on decoupling, and with these variables showing a clear effect on CSR performance they are seen as relevant enough to include in the regressions. R&D expenditures however are more likely to be higher for firms that have a significantly higher revenue since they have more money to invest. Therefore for this study I will be using R&D intensity, since the disparity between firm’s incomes is exceptionally high in the data used. R&D intensity will be calculated by dividing the R&D expenditures by the net revenues of that same year. R&D intensity is then winsorized, to minimize the effect of the outliers. Both the top and bottom 1% of results are taken out to ensure that the outcome is reliable. Sales growth is also taken as a control variable, and is calculated as the ratio of sales growth compared to the year prior.

Lastly dummy variables were created for firm years, industries and country that the firm operates in. These dummy variables were added to represent subgroups in the data that could show significant differences from the data set as a whole. For all firm years a dummy variable was created to show if there are any significant differences that these years show in relation to the full data set.

The industry sectors were classified through the Standard Industrial Classification (SIC) codes. Hawn and Ioannou (2016) argue that in industries where CSR is more material, the rewards and punishments these firms get for their CSR performance is more significant. They argue that due to this fact the industries where a lot of pollution is expected should be looked at separately to look at their influence. Ramus and Montiel (2005) find evidence that service firms not only have a weaker implementation of environmental policies in comparison to manufacturing firms, they also find significant evidence that service firms greenwash more. Graafland & Smid (2016) furthermore find a significant difference in CSR policy implementation for firms in the finance sector. With these studies in mind the industry dummies are the manufacturing industry, service industry, finance industry and a combination of very polluting industries under one dummy.

Region specific differences are accounted for by taking the 5 biggest distinct groups and relating their results with that of the full data. The region dummies that are included are the United States, Great Britain, Australia, Japan and a combination of countries in Europe included under one dummy variable.

Table 2 shows all the variables that will be tested during the linear regressions, all the variables above are listed once more for an easier overview.

4.3 Model specification

To calculate the effect of the independent variables on the dependent variable there will be five regressions taking place. Firs there will be a regression with only the control variables to isolate their effects. For every hypothesis there will be a regression, to find the isolated effect of the independent variables on CSR decoupling, this will be three individual regressions. After this all independent variables will be tested together to see if all effects together show big differences. The full formula comes out to be:

Decoupling = β0 + β1 (Female) + β2 (Indep) + β3 (EconPer) + β4 (FirmSize) + β5 (BoardSize) + β6 (SalesGrowth) + β7 (RDINT) + β8 (Leverage) + β9 (YearDummies) + β10 (IndustryDummies) + β11 (CountryDummies) + ε

Here β stands for the coefficients of the variables, and ε stands for the standard error term, which is an unavoidable error from within the data.

5. Results and Analysis Descriptive statistics

In table 3 the descriptive statistics of the various variables are shown. An interesting observation is that there is a very wide gap for the amount that firms decouple their CSR, with the maximum being an exceptionally high deviation from the mean.

Correlation analysis

Before any regression analyses were performed, I wanted to analyse the relationships between the different variables themselves. The results of this are shown in the Pearson correlation matrix shown in table 4. Most of the variables show a very small correlation with the others, with a few exceptions. Independence of the board’s correlation with gender diversity of the board is one of the notable correlations, both having a high correlation and being very significant. Another notable correlation is between firm size and its effect on both gender diversity of the board and independence of the board, showing a strong significant correlation. The combined effect of board gender diversity and board independence as expected shows a very strong significant relationship with the two aforementioned variables. Surprisingly leverage does not show any significant relationship with gender diversity of the board or board independence, but does show a significant relationship when these two variables are combined. Sales growth, economic performance (return on assets) and leverage show a few correlations which are both weakly correlated and show no significance. With all three of these variables showing little to no correlation with CSR decoupling, gender diversity of the board and independence of the board.

Regression analyses

There were five regressions performed, to find the influence of certain variables on the CSR decoupling individually, or grouped together in a certain way. These regressions are all shown in table 5. During the first regression I only looked at the effects of the control variables on CSR decoupling.Out of the six control variables that were tested, four did not find a significant relationship with CSR decoupling. The variables that were found to have a significant effect on CSR decoupling, being firm size and board size, only show a very weak effect. The interesting result is that both firm size and board size were hypothesized to reduce decoupling, but the results show that these variables slightly increase decoupling in a firm.

In the second regression, the first hypothesis was tested, the effect of the percentage of women on the board on CSR decoupling. As seen in the table percentage of women on the board has a statistically significant positive relationship with decoupling (0,002581). With the effect again showing only a very weak relationship. This is however another unexpected result, seeing as gender diversity of the board was hypothesized to lead to both better CSR performance (Cabeza-García et al., 2018) and better CSR disclosure (Cook & Glass, 2015). With this result I cannot find support for hypothesis 1, and this will therefore have to be rejected. The control variables show very little change when adding the variable for gender diversity.

The third regression tests the second hypothesis, the effect of the percentage of independent board members on CSR decoupling. The regression shows that there is a slight negative relationship between independence of the board towards CSR decoupling (-0,0004106), however this relationship is shown to not be significant. A notable change is found in the relationship between firm size and CSR decoupling, with the relationship no longer being significant. This result does not support hypothesis 2, which states that independence of the board will reduce decoupling. While a negative effect was found, the significance is not existent, and therefore hypothesis 2 will be rejected.

In the fourth regression hypothesis 3 was tested, here the combined effect of gender diversity and board independence on CSR decoupling is shown. When combining both of these variables, the newly created variable shows a vastly stronger relationship towards CSR decoupling than either of the variables showed individually (0,3058687). The relationship shows a very strong significantly positive relationship, which again was not hypothesized. Both gender diversity of the board and board independence were hypothesized to show a negative result on CSR decoupling, and the combined effect was thought to show a stronger, negative relationship. The data does not support the idea that the combined effect of board gender diversity and board independence has a negative effect on CSR decoupling, hypothesis 3 will therefore have to be rejected.

In the last regression all variables were taken into account during the regression. The results show a few changes from the previous regressions. Gender independence of the board now shows the hypothesized negative relationship towards CSR decoupling (-0,000841), however the correlation is not found to be significant. Independence of the board now shows a much stronger negative relationship towards CSR decoupling (-0,0017204), whilst also becoming very significant. This regression actually does find support for the second hypothesis, meaning that independent board members lead to less decoupling, however this effect was not significant during the second regression which makes the significance uncertain. The biggest change however is that the combined effect of board gender diversity and board independence now shows an incredibly strong relationship while still being very significant (0,7787609). This regression result is interesting since both gender diversity of the board and board independence are shown to be negatively related to CSR decoupling, but the combined effect of both shows an extremely strong positive relationship. This result implies that combining board gender diversity and board independence would vastly increase the amount a firm decouples, but on their own they have little influence.

Throughout all the regressions there does not seem to be a clear development happening in the year dummies, 2017 is the year where the least amount of CSR decoupling took place, since all other years have a positive correlation to it. However the difference in decoupling changes sporadically from year to year, and does not seem to indicate any significant changes over time. Significant differences are found between industries, both the finance industry, and the combination of heavily polluting industries show a significantly negative relationship towards the full dataset, implying that these industries decouple less which is in line with past research (Graafland & Smid, 2016; Hawn & Ioannou, 2016). The manufacturing industry shows an insignificant negative relationship, and the service industry an insignificant positive relationship. All the regional dummies show less decoupling in comparison to the full dataset with both Australia and Japan having the strongest relationship as well as the only significant relationships. This result implies that in these dummy regions less decoupling takes place in comparison to the full data set.

The adjusted r-squared shows how close all the individual data points are to the regression line. The closer the adjusted r-squared is to 1, the better the model explains the variation of the regression line. In these regressions however, r-squared is also shown to be a very small number, and therefore does not explain this variation very well. This result could stem from the fact that all the variables show very limited effects on CSR decoupling themselves. The F-value indicates whether the used variables actually influence the dependent variable, and in every regression this shows a very strong significance of P>0.001.

For every regression a robustness test was performed to test for multicollinearity. This was done through a test that checks for the variance inflation factor (VIF). The VIF value never got close to 10 in any of the regressions, for any of the variables. This means that there is no evidence for multicollinearity in the data used for the regressions.

The overall results show that the influence of almost all of the variables on decoupling is very slim, with some not showing any significance at all. With the combined effect of board gender diversity and board independence being a notable exception, showing a very strong positive relationship. The only control variable that consistently shows a significant relationship towards CSR decoupling is board size, showing a weak, but very significant positive relationship.

Additional analysis: Positive and Negative decoupling

To ensure the robustness of the findings some additional analyses were performed, first the sample was split into two groups, which differ in the direction that the decoupling took place. For simplicity’s sake the group that has a higher score for CSR disclosure than for CSR performance will be called positive decoupling (PosDecoupling). The group where the CSR performance gets a higher score than the CSR disclosure will be called negative decoupling (NegDecoupling). After splitting the dataset into positive and negative decoupling, all the previous steps for the regressions were taken. The positive decoupling dataset had a sample of 5643 observations while the negative decoupling dataset had a sample of 5329. All results were again checked for the VIF value, and there is no sign of any multicollinearity.

In table 6 and table 7 the descriptive statistics for the positive decoupling and the negative decoupling are shown. What is interesting is that the maximum for positive decoupling is quite a big higher than for negative decoupling, but this does not seem to affect the mean and standard deviation much. The negative decoupling group seems to have a higher average independence of the board, and R&D intensity seems to fluctuate much more, having a much higher mean than the positive decoupling group, but also an incredibly high standard deviation.

After the descriptive statistics, the relationships between all the variables were checked through a correlation analysis for both the positive and negative decoupling groups, the results are shown in table 8 and table 9. The relationships seem to quite similar across the board, however there are a few distinct differences. Board gender diversity is shown to have almost double the correlation for negative decoupling (0,1453) than for positive decoupling (0,0750), showing that board gender diversity is much more influential for negative decoupling. Board size is shown to have a negative relationship towards positive decoupling, and a positive relationship towards negative decoupling, meaning that board size is more likely correlated to firms understating their CSR performance. In comparison with the initial correlation matrix a much bigger correlation is found for almost all variables in both decoupling directions towards decoupling. On top of this more variables have a significant relationship with decoupling when looking at the direction of decoupling rather than absolute decoupling, sales growth, R&D intensity and leverage are among the variables that find some significance now that was not shown before.

In table 10 and 11 the results of the regressions are shown for the positive decoupling and negative decoupling datasets, the results show a few interesting differences between the two groups as well as with the initial results. Gender diversity of the board shows a very similar relationship for decoupling as a whole, positive decoupling and negative decoupling. When looking at its effect by itself it shows a positive significant relationship, however when taking all of the independent variables in account it becomes a negative relationship for positive decoupling and decoupling as a whole, but it stays positively related to negative decoupling. Gender diversity stays significant for both positive and negative decoupling, but not decoupling as a whole when all variables are regressed together. Independence of the board was found to be insignificant in the initial regressions, however here it is shown to increase positive decoupling significantly, while still not having a significant relationship with negative decoupling. This is an interesting result since it implies that independent board members will increase the quality of CSR disclosure by a bigger margin than they improve the CSR performance, which results in positive decoupling. Another big difference is that the combination of board gender diversity and board independence, shows a strongly positive, significant relationship with CSR decoupling in both directions when taken by itself. However when taken with the other independent variables in the same regression it suddenly becomes insignificant for positive decoupling, while still being very significant for negative decoupling and decoupling overall.

Firm size shows a significant, strongly negative relationship towards both positive and negative decoupling which could come from the fact that bigger firms are better at dealing with the different stakeholder pressures they face (Lepoutre & Heene, 2006). This result was however not found in the initial dataset with absolute decoupling, so the relationship remains to be speculated. R&D intensity only shows a significant relationship for negative decoupling, with it being negatively correlated. This could mean that firms with a higher R&D intensity are more occupied with CSR performance than CSR disclosure, but the effects are very small. All other variables show very similar relationships across the board, however both the positive decoupling and negative decoupling regressions show a higher adjusted R-squared and a much higher F-value, giving more credibility to these results.

Additional analysis: Components of CSR decoupling

After looking at the differences between positive and negative decoupling, I was interested to see whether the different components of CSR decoupling show significant differences too. In this section all the previous analyses will be performed on the individual components of CSR decoupling: Social decoupling, environmental decoupling and governance decoupling. The results were once again tested for multicollinearity, and no sign of multicollinearity was found. Table 12 shows the descriptive statistics and the correlation matrix for the three components of CSR decoupling. The complete dataset was used for these analyses and therefore the descriptive statistics and correlations between all independent variables and control variables are the same, which is why only the descriptive statistics and correlations that involve the three different decoupling aspects are shown.

The descriptive statistics are very similar for all three components, with the means and standard deviations being extremely close, the only difference can be found in the maximum decoupling of each component, with environmental decoupling having a much higher maximum. The correlation matrix shows similar relationships for each component of CSR decoupling, with governance decoupling in general having the strongest, most significant correlations. Gender diversity of the board only has a significant correlation with governance decoupling (-0,0917), showing no significance towards social decoupling 0,0173) or environmental decoupling 0,0183). Independence of the board is very strongly correlated to governance decoupling (-0,1420) while only showing weak correlations with social decoupling (-0,0273) and environmental decoupling (-0,0244), all correlations are significant however. This could show that independence of the board is a much higher influence on governance decoupling than on the other two aspects. The combination of board independence and board gender diversity is only not significant to environmental decoupling (-0,0159), while showing significant negative correlations with social decoupling (-0,0258) and governance decoupling (-0,0859).

In table 13, 14 and 15 the regression results are shown for social decoupling, environmental decoupling and governance decoupling respectively. The results are very different for each component, and are therefore worth discussing further. Social decoupling shows the weakest results, with only board independence and the combination of board gender diversity (-0,1893172) and board independence (-0,0005742) having a significant relationship when their effects are isolated. All other variables show no significant relationship, and the R-squared value and the F-value show that these results are not very trustworthy.

None of the independent variables show any significant relationship towards environmental decoupling. Only firm size and board size show a significant relationship, with both having a positive effect, therefore increasing environmental decoupling.

Governance decoupling is the only component where the independent variables all show a significant relationship towards decoupling. Both gender diversity of the board (-0,0046842) and board independence (-0,0031367) show the initially hypothesized negative relationships and are very significant. These results show support for the first and second hypotheses, which means that board gender diversity and board independence lower the amount of decoupling that takes place. However the biggest surprise is that the combination of board gender diversity and board independence shows a very strong negative relationship (-0,5699416) when its effect is isolated in the fourth regression, but shows an even stronger positive relationship (0,824621) when regressed with all the variables. This result is very strange and the reason for why this happened remains unanswered.

6. Conclusion

There is a growing concern from stakeholders that want firms to contribute more than just monetary gains. Corporate social responsibility is therefore quickly becoming an integral part of the operations in a firm, and stakeholders are expecting that firms will manage their interests (Jamali, Safieddine & Rabbath, 2008). This leads to many firms implementing CSR policies to appease their stakeholders. However, while many firms may disclose their CSR policies, implementation can be lacking in effectiveness (Graafland & Smid, 2016). The intention might be there, but firms often do not follow through on their promises regarding CSR. This disparity is what is called CSR decoupling (Graafland & Smid, 2016), it is essentially the difference between a firm’s CSR performance and their CSR disclosure. CSR decoupling can be very damaging to a firm’s reputation because when stakeholders find out that firms have not been truthful about their CSR performance they will most likely lose trust in the firm. According to stakeholder theory firms will therefore want to limit the amount of CSR decoupling that takes place in their firm, to keep the relationship with their stakeholders healthy.

The causes for CSR decoupling are largely still an uncertainty in today’s research. Most of the research that does look into CSR decoupling however goes into the motivations for, and the outcomes of decoupling. By doing this study I tried to look at whether certain board characteristics have an influence on the extent of CSR decoupling. Gender diversity of the board, independence of the board and the combination of both were the specific board characteristics of which a relationship was tested with CSR decoupling.

Both of these characteristics were tested because of the influences they have on different aspects of CSR in a firm. Gender diversity of the board has shown to have a positive influence on a board’s decision making (Swartz & Firer, 2005), board effectiveness (Coffrey & Wang, 1998) and disclosure of non-financial information (Cabeza-García et al. 2018). With the effect that gender diversity has on the disclosure of non-financial information in mind, it was hypothesized that gender diversity of the board would have a negative effect on CSR decoupling. This was hypothesized because CSR decoupling can lead to reputational damage, and many prior studies show an overall positive effect on a firm when they have female board members. Independence of the board has shown to increase voluntary disclosure (García-Meca & Sánchez-Ballesta, 2010), increase CSR initiatives in a firm (Cabeza-García et al. 2018) and improves both CSR reporting and CSR disclosure (Dah & Jizi, 2018). Because independence of the board has both a positive influence on CSR disclosure and CSR performance it was hypothesized that the independence of the board would have a negative effect on CSR decoupling. The combination of both these variables was hypothesized to have a negative influence on CSR decoupling, and it was expected to show a stronger relationship. These assumptions were made because firms with both independent boards and gender divers board show better environmental scores (Post et al., 2011) and better stakeholder expectation management (Liao, Luo & Tang, 2015). Both of these could lower the amount of CSR decoupling that takes place.

In the results of the analyses, I find that while both gender diversity of the board and independence of the board influence CSR decoupling, the effect is quite small. Gender diversity of the board shows an unexpected positive relationship with absolute CSR decoupling when regressed alone, which was not hypothesized. The implication here is that more female board members will increase the amount of decoupling in a firm, the amount by which it gets

increased however might be too insignificant for firms to pay attention to. During the last regression of absolute decoupling gender diversity of the board shows a negative relationship towards decoupling, however this relationship is not found to be significant. Positive decoupling and negative decoupling show similar results, with board gender diversity increasing decoupling for most of the regressions. For the three components of CSR decoupling board gender diversity only shows significance towards governance decoupling, with it lowering decoupling both times it was tested.

The conclusion here is that during the two regressions where gender diversity of the board was looked at, very different results came out. While the negative relationship in the last regression is a result that is inconsistent, since it is not significant I still think there is some evidence that gender diversity of the board might increase CSR decoupling slightly.

Independence of the board was found to have the effect that was hypothesized for absolute decoupling, which is the reduction of CSR decoupling in a firm, however it is not found to be significant whenever it is regressed on its own. This is an unexpected result, and it implies that independence of the board has little to no effect on a firm’s CSR decoupling. However during the last regression with all variables taken into account, it suddenly shows a very significant, albeit weak, negative relationship towards CSR decoupling. This result is in line with what was hypothesized, but it only shows significance whenever regressed with all the available variables. When split into positive and negative decoupling, board independence shows differing results, sometimes being positive, and sometimes being negative, which results in not being able to conclusively know what effect independence of the board has on decoupling. For the three components of CSR decoupling I find weak evidence that independence of the board lowers social decoupling, but strong evidence that CSR decoupling lowers governance decoupling.

The combination of gender diversity of the board and independence of the board shows a significant, extremely strong positive relationship towards absolute decoupling. This implies that these variables combined, increase CSR decoupling in a firm very significantly. When testing for positive and negative decoupling, similar results are found, with the combination of gender diversity of the board and independence of the board strongly increasing decoupling. The components of CSR decoupling show differing results, sometimes showing positive relationships, negative relationships or the effect being insignificant. Overall it is advised for firms to limit either one of the two aforementioned board characteristics, since taken together might lead to a significant increase in decoupling as shown in the majority of regressions. In general I find weak or no evidence for my hypotheses, but for governance decoupling however I find strong evidence for hypothesis 1 and 2, and an inconclusive result for hypothesis 3. From a stakeholder theory perspective it makes sense that board gender diversity and board independence negatively correlate to governance decoupling, as both of these are governance measures to monitor the management (Adams & Ferreira, 2009). So while the board characteristics might have little influence on CSR decoupling overall, there is a case to be made on their negative effects on governance decoupling.

The research performed has a couple of implications for literature, first it is one of the few studies looking into the causes for CSR decoupling, with prior research either focussing on CSR disclosure, CSR performance or motivations for and outcomes of CSR decoupling

instead. Although the results were mostly quite weak across the board, this research shows that board composition can influence the amount that a firm will decouple their CSR. In fact it was mostly the variables which involved board characteristics that have shown consistently significant relationships with CSR decoupling. This finding could act as a diving board for more researchers to study what affects the amount by which a firm decouples, and specifically what board characteristics influence CSR decoupling, as well as its components.

Practical implications are few however, only the combination of board gender diversity and board independence showed a strong enough relationship to have practical implications for firms. With the other variables showing quite weak relationships with CSR decoupling, so firms might find that it is too much work for too little gains to avoid certain board compositions. However giving the firms the knowledge that board composition can definitely have an effect on CSR decoupling and its components might be of value in the future.

7. Limitations and Further Research

There are a few limitations that are bound to the research that was performed. Perhaps the biggest limitation of this study is that all data has been collected from databases. While earlier in the study I have outlined the professionalism of the people that collect data for those databases, it is still generally advised to hand collect at least one of your variables (Fogarty, 2006). Collecting the data for at least one variable by hand can reduce the dependency on the trustworthiness of the databases. For this study however, with its limited time frame it was impossible to hand collect even one variable, and therefore the choice was made to use the databases for all variables.

Future research could be done on the different components of CSR decoupling. My results show interesting differences between social, environmental and governance decoupling, and future research might find the causes for such strong differences. Furthermore it could show that different characteristics of the firm or board could affect different components of CSR decoupling. Lastly this study looks at firms from all over the world, and it could be interesting to see the differences between regions more closely. From my results western and more modernized countries seem to decouple the least, and it might be of value to find potential reasons for this. Different laws and regulations could show significant differences in their results. Analyses between shareholder-oriented and stakeholder-oriented countries, or common law countries and civil law countries could show differing results, which can all be researched in the future.

8. Sources

Adams, R.B., de Haan, J., Terjesen, S., & van Ees, H., 2015. Board diversity: Moving the field forward. Corporate Governance: An International Review, 23, pp. 77-82.

Adams, R.B., & Ferreira, D., 2009. Women in the boardroom and their impact on governance and performance. Journal of Financial Economics, 94, pp. 291-309.

Aguinis, H., & Glavas, A., 2012. What we know and don’t know about corporate social responsibility: A review and research agenda. Journal of Management, 38 (4), pp. 932–939. Bansal, P., 2005. Evolving sustainably: A longitudinal study of corporate sustainable development. Strategic Management Journal, 26 (3), pp. 197-218.

Bansal, P., & Clelland, I., 2004. Talking trash: Legitimacy, impression management, and unsystematic risk in the context of the natural environment. Academy of Management Journal, 47 (1), pp. 93-103.

Bartley, T., & Egels-Zandén, N., 2016. Beyond decoupling: Unions and the leveraging of corporate social responsibility in Indonesia. Socio-Economic Review, 14 (2), pp. 231-255. Behnam, M., & MacLean, T.L., 2011. Where is the accountability in international accountability standards: A decoupling perspective. Business Ethics Quarterly, 21 (1), pp. 45-72.

Braam, G., & Peeters, R., 2017. Corporate Sustainability Performance and Assurance on Sustainability Reports: Diffusion of Accounting Practices in the Realm of Sustainable Development. Corporate Social Responsibility and Environmental Management, 25 (2), pp. 164-181.

Brammer, S., & Pavelin, S., 2008. Factors influencing the quality of corporate environmental disclosure. Business Strategy and the Environment, 17, pp. 120-136.

Bromley, P., Hwang, H., & Powell, W.W., 2011. Decoupling revisited: Common pressures, divergent strategies, in the U.S. nonprofit sector. M@n@gement, 15 (5), pp. 469-501.

Brown, W.O., Helland, E., & Smith, J.K., 2006. Corporate philanthropic practices. Journal of Corporate Finance, 12 (5), pp. 855-877.

Cabeza-García, L., Fernández-Gago, R., & Nieto, M., 2018. Do Board Gender Diversity and Director Typology Impact CSR Reporting? European Management Review, 15 (4), pp. 559-575.

Cecil, L., LaGore, W., Mahoney, L.S., & Thorne, L., 2013. A research note on standalone corporate social responsibility reports: Signaling or greenwashing? Critical Perspectives on Accounting, 24 (4/5), pp. 350-359.

Chakrawal, A.K., 2006. Corporate Governance and Accountability. Asia Pacific Business Review, 2, pp. 88-97.

Christmann, P., & Taylor, G., 2006. Firm self-regulation through international certifiable standards: Determinants of symbolic versus substantive implementation. Journal of International Business Studies, 37 (6), pp. 863–878.

Clement, R., 2005. The lessons from stakeholder theory for US business leaders. Business Horizons, 48 (3), pp. 255-264.

Clarkson, M.B.E., 1995. A stakeholder framework for analyzing and evaluating corporate social performance. Academy of Management Review, 20, pp. 92-117.

Coffrey, B.S., & Wang, J., 1998. Board diversity and managerial control as predictors of corporate social performance. Journal of Business Ethics, 17 (14), pp. 1595-1603.

Conyon, M.J., & He, L., 2017. Firm performance and boardroom gender diversity: A quantile regression approach. Journal of Business Research, 79, pp. 198-211.

Cook, A., & Glass, C., 2015. Do minority leaders affect corporate practice? Analyzing the effect of leadership composition on governance and product development. Strategic organization, 13, pp. 117-140.

Cornell, B., & Shapiro, A.C., 1987. Corporate stakeholders and corporate finance. Financial Management, 16 (1), pp. 5-14.

Crilly, D., Zollo, M., & Hansen, M.T., 2012. Faking it or muddling through? Understanding decoupling in response to stakeholder pressure. Academy of Management Journal, 55 (6), pp. 1429-1448.

Dah, M.A., & Jizi, M.I., 2018. Board Independence and the Efficacy of Social Reporting. Journal of International Accounting Research, 17 (1), pp. 25-45.

Delmas, M.A., Burbano, V.C., 2011. The drivers of greenwashing. California Management Review, 54, pp. 64-87.

Fogarty, T.J., 2006. Publishing in Academic Accounting: Practical Advice and Healthy Iconoclasm, in Z. Hoque (ed.), Methodological Issues in Accounting Research: Theories and Methods. London: Spiramus Press, pp. 515-34

Freeman, R.E., 1984. Strategic Management a Stakeholder Approach. Pitman. Boston Freeman, R.E., Wicks, A.C., & Parmar, B., 2004. Stakeholder theory and “the corporate objective revisited”. Organization Science, 15, pp. 364-369.

García-Meca, E., & Sánchez-Ballesta, J.P., 2010. The association of board independence and ownership concentration with voluntary disclosure: A meta-analysis. European Accounting Review, 3, pp. 603-627.

George, E., Chattopadhyay, P., Sitkin, S., & Barden, J., 2006. Cognitive underpinnings of institutional persistence and change: A framing perspective. Academy of Management Review, 31, pp. 347-365.