The impact of the Poldermodel on stakeholder transparency in CSR reports of Dutch firms.

(1)

The impact of the Poldermodel on stakeholder transparency in

CSR reports of Dutch firms.

Do less internationalized firms score higher?

Name:

Marnix Wilms

Student number: s2572109

Address:

Eikenlaan 214

9741EV Groningen

marnixwilms@gmail.com

+31 650543787

Course: Master thesis

Study program: International Business & Management

Theme: Corporate Social Responsibility (CSR)

Supervisor: Kees van Veen

Referent: Marques Morgado

(2)

(3)

page 3 van 48

Abstract

There are many determinants that influence the amount and quality of CSR reports. Therefore CSR reports vary significantly across firms, countries and region. Until now research has mainly focused on explaining these differences by looking at company specific factors or accounting differences between nations. However these types of research cannot explain all the differences between firms on how well they report their CSR activities. Previous studies have argued that national context might also influence the quality of CSR reporting. Therefore is the aim of this study is to examine if Dutch firm with less internationalized corporate governance structures will have higher stakeholder transparency in their CSR report. The reason for the proposition lies within the believe that the Poldermodel left behind a management culture with a high degree of stakeholder orientation. Using a sample of 81 of the biggest Dutch companies we rate the degree of internationalization on three indicators related to corporate governance. These indicators are: the degree of Dutch ownership, Dutch board membership and Dutch stock market listing. Using a multiple linear regression model, we link these indicators to the firm scores on stakeholder transparency they achieved on the 2014 Transparency Benchmark. Our finding are that Dutch board members have a small but positive influence on stakeholder transparency. Dutch listings and ownership does however not seem to be related to higher stakeholder transparency.

Key words:

(4)

page 4 van 48

1. Introduction

At the present time the majority of firms report about their corporate social responsibilities (CSR) activities in some degree (Hess, 2007). This surge in CSR disclosure derives from the believe that CSR can enhance a firm’s value and operating efficiency by market gains, capital market benefits, risk management as well as regulatory benefits (Malik, 2014). However, Roberts (1992) already noticed that if companies want to fulfill their social responsibilities it will not be enough to just obey the existing laws in the area of doing business but companies also have to make choices to comply to community standards. Understanding the reasons and underlying motivations behind these choices are critical. (Manner, 2010). Considering cultural and corporate governance factors would benefit this stream of research (Haniffa & Cooke, 2005; Chan, Watson & Woodliff, 2014). Until now research has focused on answering these questions by looking at company factors or accounting differences for answers. Therefore this study aims to answer the until now unanswered call of Meek, Roberts & Gray (1995) and van der Laan-Smith, Adhikari & Tondkar (2005), which is to recognize and explain the large differences between firms, in specific countries or regions in the types of CSR information resident firms disclose.

(6)

page 6 van 48

degree of internationalization within Dutch firms lead toward lower stakeholder transparency in their CSR reporting.

The focus on stakeholder transparency is another contribution of this research. Research up till now almost exclusively focused on broad terms such as disclosure and voluntary disclose. However, with better informed stakeholders, skepticism towards CSR reporting is mounting. One way to take away this skepticism is for firms to be more open and transparent in their CSR disclosures (Arvidsson, 2010). Furthermore it is transparency and not just disclosure that cultivates firm understanding (Perez, 2014) and stakeholder transparency in particular that helps to build mutual trust and commitment in the firm-stakeholder relationship (Jahansoozi, 2006) and gives advantage in corporate reputation (Perez, 2014). All these factors are precisely what firms aim to achieve according to the stakeholder and legitimacy theories. Owen, Swift & Hunt (2001: 267) already recognized the growing importance of stakeholder transparency calling it “an essential prerequisite for successful pursuit of the process” were the process refers to gaining value from social reporting. However research into determinants of stakeholder transparency is still lacking. It is the fact that data from the annual Transparency Benchmark that the Dutch Ministry of Economic Affairs organizes since 2004 has now become available. That allows us to look at stakeholder transparency were other research could not. Using a third party benchmark gives us a better measurement of transparency then other self-made measurements that are used in previous research (Dubbing, Graafland & Liedekerke, 2008; van de Burgwal & Vieira 2014). All together we feel confident that by including the Poldermodel and stakeholder transparency we look at firm determinants for transparency in a novel way.

(7)

page 7 van 48

2. Literature review

2.1 Transparency

In order to explain transparency in the context of CSR reporting, we first have to give the definition of CSR and CSR reporting. CSR has had multiple definitions ranging from very broad to very narrow (Fisher, 2004). Regardless of the exact definition, CSR is of great importance for society, because as stated very well by Carroll (1999) “it addresses and captures the most important concerns of the public regarding business and society relationships”. The definition of CSR reporting follows the same logic. According to Hackston & Milne (1996) CSR reporting is “the provision to diverse stakeholders of non-financial information relating to a company’s interactions with its physical and social environment, as stated in corporate annual reports or separate social reports” There is however growing skepticism towards CSR reporting. Stakeholders often believe that CSR is just window-dressing or in other words a PR-activity that sounds admirable but is not acted upon (Cai, Jo & Pan, 2012). This line of thinking is not surprising since it is also possible to use CSR disclosures to obscure instead of enlighten the public by burying them with meaningless information (Rawlings, 2009). Furthermore, firms have been found to only adopt compliance programs such as CSR reporting as additional insurance against prosecution (Hess, 2007). When it is impossible to rate the effectiveness a such a compliance program, firms can have the illusion of having such a program in place and using it as cover to take even less responsibilities for their actions (Hess, 2007).

(8)

page 8 van 48

window dressing, CSR would just be another abbreviation (Dubbink, Graafland & van Liederkerke, 2008).

2.2 Why do companies choose to be transparent

Previous section clearly illustrates that transparency has benefits for the various stakeholders of a firm. But besides mandatory regulations what other factors make firms decide to be more transparent in their CSR reporting. For one, the pressure faced by firms from society to openly communicate information about their involvement in corporate social responsibility has risen substantially within the last years. Following many corporate scandals such as excessive management bonuses, amplified pollution, fast draining of natural resources and so forth, the public awareness towards social and ethical behavior of firms has increased (Arvidsson, 2010). Some recent failures of high profile firms such as Lehman Brothers, AIG and British Petroleum have shown that the reach of a firm is much larger than just financial investors. Stakeholders such as employees, suppliers, customers but also the communities and environment in which the firms operates are also impacted (Van der Laan Smith, Adhikari, & Tondkar, 2005). This has led to a growing discussion about the depth and breadth of the social and ethical issues firms should report about and be held accountable for. In light of these recent developments firms have increasingly started to report their CSR activities more extensively.

Due to immense progress in Information Technology (IT), end consumers but also suppliers, governments and other parties are often very well informed about the activities of a firm. It is argued that this leads to higher pressure for the whole supply chain to be more transparent. A good example of this pressure is the rise in socially responsible investing (SRI). In 2008 already 10 percent of all US investment was screened using SRI guidelines or a similar process (Galema, Plantinga & Scholtens, 2008). However as shown before are investors are the only stakeholder of a firm and increasing transparency is one way of dealing with the big diversity of stakeholders, such as shareholder, customers, suppliers, lenders and employees. If the management succeeds to react and report appropriately it might lead to better and more beneficial relationships with these stakeholders (Wilmhurst & Frost, 1999).

(9)

page 9 van 48

different firm approaches and determinants of CSR transparency. These theories will now be explained in more detail.

2.3 Legitimacy theory

In the early eighties it was first suggested that CSR disclosures comes from the need for firms to legitimize their corporate activities (Cho & Patten, 2007). At this time the theory focused on how corporate management reacts on changing community expectations (Wilmhurst & Frost, 1999). The assumption underlying this theory is that stakeholders within the community set boundaries on acceptable firm behavior. Firms as incumbents of this community have to stay within these boundaries in order to keep or achieve legitimacy (Cong & Freedman, 2011). Given to rise in community awareness, legitimacy theory states that firms will take control to make sure their corporate activities are compliant to the community needs and wishes. An annual report focused on CSR disclosure or a special CSR report can be used to strengthen the communities perception that the firm management team is taking responsibility towards the issues of importance to the community (Wilmhurst & Frost, 1999). This in return can benefit firms by improved reputation, visibility in the consumer market, support of local communities or positive treatment from regulatory bodies (Hess, 2007)

To realize these benefits Arvidsson (2010) stated that it is not enough for firms to just comply and communicate to be complying to existing law and regulations in order to be recognized as a social responsible firm. It also must make their CSR activities visible towards internal and external stakeholders in order to achieve legitimacy. She names three main points that firms have to take into account in order to achieve legitimacy. First she declares that legitimacy is achieved or maintained with institutional stakeholders by stakeholder disclosure. Secondly the transparency policy is a reflection of the values and priorities of the firm. Third, only by disclosing open and timely about a firms CSR practices it can become a competitive advantage (Ardvidsson, 2010).

2.4 Stakeholder theory

(10)

page 10 van 48

identification of moral or philosophical guidelines for the operation and management of corporations”. According to stakeholder theory the dialogue with stakeholders is essential to identify these moral guidelines (Piechocki, 2004).

The main argument for stakeholder theory is grounded in the fact that without the support and/or resources provided by the various stakeholders of an organization, the survival and success of an organization are in jeopardy (Clarkson, 1995). The fact that firms have dependence on stakeholders gives them influence in the organization and on the behavior as such. It is therefore argued that a successful firms will address the needs of all important stakeholders and that this include developing a strong social responsible reputation through performing and disclose openly and transparent about CSR activities (Chan, Watson & Woodliff, 2014). The example of the increase in social responsible investment (SRI) is a clear illustration that complying to social pressures gives a firm ‘license to operate’. The difficulty is that the relationship between the firms and these licensors is not static and need constant negotiation, debate and dialogue between the firm and stakeholder (Hess, 2007). Dawkins & Fraas (2013) looked at the basic reasons for firms to engage in CSR transparency from a stakeholder perspective. They found that disclosure and transparency are important for the different stakeholders since it allows them to monitor the activities of the firm and therefore close (part of) the information asymmetry. Financial or other annual performance indicators are a good way to look at short-term results. This might be enough for certain stakeholder groups but other groups might need guarantees that the firm is also anticipating on, and actively looking to prevent, long term issues (Hess, 2007). Successful firms have to comply to these information needs in order to gain support from these stakeholders.

2.5 Transparency and corporate governance

(11)

page 11 van 48

corporate governance (Cong & Freedman, 2011). The same surge in corporate scandals that led to an increase in research on CSR also led to a call for stronger corporate governance (Cong & Freedman, 2011). Previous research has indeed concluded that factors influencing corporate governance are closely related to voluntary disclosure (Eng & Mak, 2003; Haniffa & Cooke, 2005; Barako, Hancock & Izan, 2006). These same authors found two main factors regarding to corporate governance and its influence on CSR reporting; ownership structure and board composition.

Since society is constantly changing the rule of law cannot always keep up. In order to address this issue governments often use tools that come closer to self-regulation. Social reporting is one of this subject where law cannot always reflect the current expectations of society and an increasing number of scholars therefore advocate a social reporting regime stronger based on corporate governance then on regulations (Hess, 2007). In comparison with financial information that is disclosed by companies and is often standardized and obligated by numerous (accounting) laws. Information featured in CSR reporting including non-financial and non-regulated features are not standardized nor regulated by law (Golob et al., 2013) Even with upcoming guidelines such as the GRI (Global reporting initiative), ISO, Social Accountability 800 it remains a voluntary decision to be transparent (Golob et al., 2013). This voluntary aspect gives management more discretion to report about CSR in the way they find appropriate. This be the reason why ownership structure and board composition are found to have an influence on the CSR reporting practices of firms (Barako, Hancock & Izan, 2006; Oxelheim & Randoy, 2003)

The great variety of scores reported by the Transparency Benchmark on stakeholder transparency reinforces this point. Firms scores range from the minimum of zero to the maximum of twenty points1. Research into environmental disclosures of Dutch listed firms revealed that firms size, industry and profitability can only explain part of the variance between firms and their disclosure practices (van de Burgwal & Vieira, 2014). Other research has shown that internationalization of corporate governance structures might provide additional explanation. (Barako, Hancock & Izan, 2006; Oxelheim & Randoy, 2003; Halme & Huse, 1996; Chan, Watson & Woodliff, 2014; Johnson & Greening, 1999). These authors all tried to connect a higher degree of international governance to higher degree of voluntary disclosure. We however follow the research of Haniffa & Cooke, (2005) who looked at the

1

(12)

page 12 van 48

specific context of Malaysia and found a strong relationship between Malaysian board membership and share ownership and the amount of CSR disclosure by Malaysian firms. In the next section we will explain why we feel that the Dutch context is also appropriate for this type of research.

2.6 Why look at transparency in the Dutch setting

Institutional domains and national culture shapes how corporate governance is executed (Aguilera & Jackson, 2003). Based on this they explain the difference between corporate governance practices in continental Europe and Anglo-America. On a lower level, Egan (1997) showed that corporate culture is influencing managerial decisions and therefore also corporate governance. Using French, Swedish, British and German companies she argues that different business cultures and managements styles impact managerial decisions. We think that the Dutch Poldermodel has shaped an unique Dutch business culture. To explain this in more detail we will first provide a short explanation of the Poldermodel and how it continuous to be off impact.

After a very difficult period, the Dutch economy made a very swift upturn in the 1990’s which became known as the Dutch miracle (Karsten, van Veen & van Wulfften Palthe, 2008). Part of the credit for this remarkable recovery is given to the Poldermodel. This relates to the fundamental characteristics of the Dutch institutional system where there is a strong tendency towards negotiation and compromises between seemingly opposing partners (Karsten, van Veen & van Wulfften Palthe, 2008). This process is strongly anchored within the Dutch society since the formerly Dutch republic was already based on consensus between the seven provinces that eventually became the Dutch republic2. Therefore is consensus and negotiation strongly interweaved in the Dutch society.

That the Poldermodel still has an influence is reflected in the fact that in the open and vulnerable Dutch economy. Socio-economic decisions are still determined primarily by agreements between a permanent consultative body of employers and employees organizations and the Sociaal economische raad (SER) which is a permanent advisory body of the government. The basis for this tight cooperation is the strong belief that over a long time span employers and employees need each other to achieve the most profitable result for both groups (Glasbergen, 2002). A practical residual from this model is that it created incentives and opportunities for multiple parties to engage in constrictive discussions and as Glasbergen

2

(13)

page 13 van 48

(2002) puts it “a pluralistic view of public space” meaning that the public space is not just ruled by the government, but other actors such as corporations also play a significant role in this space. Besides mutual commitment, compromising and prioritizing Karsten, van Veen & van Wulfften Palthe, (2008) give “joint fact finding” as one of the key characteristics of the Poldermodel. This characteristic implies that firms already have to work together with stakeholders in the early stages of negotiation. This first stage is described as the exchange of information and joint fact finding. This sharing of information with stakeholders is one of the main arguments why we believe Dutch managers will have greater awareness towards stakeholder transparency within their CSR reporting.

We believe the connection between the mindset of Dutch managers and the amount stakeholder transparency in CSR reporting of a firm, can be made by including the Upper Echelon Theory (UET). Managerial values and beliefs, which are important determinants of personal decision making according to the original UET (Hambrick & Mason 1984) are shaped for a great deal in the context of the national culture (Ronen and Shenkar, 1985). In particular educational systems and labor market characteristics influence this managerial identity (Glunk, Heijltjes & Olie, 2001). Because in the Netherlands (Dutch)executives have spent most of their working life in the context of the Polder model, we suggest that this has positively influenced their perspective towards stakeholder transparency. Previous research into UET has yielded strong connections between organization outcome and team composition (Carpenter, Geletkanycz &Sanders, 2004). Showing that top management teams (TMTs) can influence firm outcome. Research of Garcia-Sanchez, Cuadrado-Ballesteros & Sepulveda (2014) found that the CSR efforts of a company is a strategic choice and therefore decisions regarding voluntarily disclosure and transparency is a main task for the board of directors. Combing this previous research, we expect the positive perspective on stakeholder transparency of Dutch managers to be reflected the mindset of the TMT which in turn will influence the decision to include a high degree of stakeholder transparency in CSR reports.

2.7 Higher degree of internationalization

(14)

page 14 van 48

diversity shows itself in greater variety of cultures, customers, regulations and competitors. Such an increase in diversity put strains on the domestic uniform mindset of the top management team (TMT) (Sanders & Carpenter, 1998). One of the most important factors impeding firms to deal with the increasing complexity of internationalization is a failing governance structure (Sanders & Carpenter, 1998). To deal with this growing complexity many firms appoint board members of different nationalities (Masulis, Wang & Xie, 2010).

It might therefore not be surprising that in 2005 already 46,4% of the top management (only executive boards) of the biggest 30 firms in the Netherlands were of foreign nationality (van Veen & Marsman, 2008). In our research we also take the supervisory board into account and then the internationalization of boards within the Dutch setting tells a different story. Until late 2004 the supervisory board was appointed using a cooptation mechanism, which means that the supervisory board could choose its own members. This led to a strong “old boys network” which was a very dense network of directorates where it was very difficult to get in as an outsider (van Veen & Elbertsen, 2008). However 2004 a new law was implemented whereby the appointment of new members in the supervisory as well as the executive board need shareholder approval. By breaking this “old boys network” we expect that the number of foreign supervisory directors will have increased since 2004.

Ample research has focused on the influence this internationalization process has on CSR disclosures (Barako, Hancock & Izan, 2006; Oxelheim & Randoy, 2003; Halme & Huse, 1996; Chan, Watson & Woodliff, 2014; Johnson & Greening, 1999). They tried to prove that greater internationalization of the corporate governance structures would lead to higher CSR disclosures/transparency. While we argue that because the specific Dutch management style that is a residual from the Poldermodel, firms with less international governance structures have higher stakeholder transparency.

2.8 Corporate governance determinants of stakeholder transparency.

(15)

page 15 van 48

This is also in line with research into internationalization measurements, which states that there are three main indicators relating internationalization to governance structures (Dorrenbacher, 2000). These measurement are previously used to proxy the impact of corporate governance on corporate social reporting (Halme & Huse, 1996; Haniffa & Cooke, 2005; Barako, Hancock & Izan, 2006) and include:

1. Number of stock markets on which a company is listed 2. Amount or proportion of shares owned by foreigners

3. Number or proportion of non-nationals in the board of directors.

We will use all three of these indicators because this improves the validity of the construct. Using a single item indicator often misrepresents the construct because it only represents a partial portion of the whole domain (Sullivan, 1994). Combining legitimacy and stakeholder theory to the Dutch setting as described above leads us to the development of some testable hypothesis. These hypothesis and their theoretical background will be further explained in the next section.

3. Hypothesis development

(16)

page 16 van 48

stakeholder transparency are driven by board composition. Therefore we expect that firm with a more international board composition will perceive stakeholder transparency in CSR reporting off less importance than a firm with a board composed of Dutch natives. Thus;

H1: A more international board will lead to a lower degree of stakeholder

transparency in CSR reporting and vice versa.

Country of origin is a significant factor in predicting the amount of voluntary CSR disclosure (Newson & Deegan, 2002). The same holds true for the country in which a firm is listed. Other research has also argued that multiple listing increase CSR disclosure (Hackston & Milne, 1996; Meek, Roberts & Gray, 1995). The most important reason for firms to have a multiple listing is the availability to new capital. In order to be able to attract capital in foreign markets firms have to step up their transparency (Meek, Roberts & Gray, 1995; Oxelheim & Randoy, 2003). A second reason why firms with multiple listing are proposed to have greater CSR disclosure is that these firms have to react to pressures and media coverage from a broader range of interest groups (Haniffa & Cooke, 2005). The third motivation for having dual or multiple listings is to credibly commit to another(better) legal regime. Stating that better public, political and employee relations are an credible incentive for firms the list on more than one stock exchange (Hossain & Reaz, 2007). However, multiple listing would only increase the degree of stakeholder transparency if the second or third listing would be at a stock market that ascribes higher value to stakeholder transparency. Using the same line of reasoning as for ownership. We believe that actors on the Dutch capital market will demand the highest degree of stakeholder transparency from firms. Thus:

H2: A firm listed on more than one stock market will not have higher stakeholder

transparency in CSR reporting then a firm only listed on the Dutch market.

(17)

page 17 van 48

leads firms to provide information they believe stakeholders require (Cong & Freedman, 2011). In other words, according to the stakeholder theory, stakeholders are the actors that determine the scale and scope of stakeholder transparency. We believe that the Poldermodel works both ways. By this we mean that not only Dutch firms are more used to stakeholder collaboration, Dutch stakeholders have also gotten used to be informed and consulted about firms policies and actions. Therefore we argue that a company with a higher degree of Dutch shareholders will be more pressed to be transparent towards its stakeholders. Because of this we believe that firms with a greater extent of Dutch ownership will have higher stakeholder transparency within their CSR reports. Thus:

H3: A firm with a higher degree of shares owned by Dutch nationals will have higher

stakeholder transparency in CSR reporting and vice versa.

Resulting from these hypothesis in the next part we will describe the research design and methodology in more detail.

4. Methodology

In exploring the link between corporate transparency towards stakeholders and the internationalization of a firm there are two concepts that need to be made quantifiable. The first one being the degree of internationalization in ownership, stock market listings and board composition (independent). And the second one is the transparency towards stakeholder in the firms CSR reporting (dependent). Furthermore, there are also some imperative control variables to take into account when performing this type of research. These control variables include: industry, firm size and profitability and will be elaborated further on in this section.

(18)

page 18 van 48

Figure 1 - Conceptual model – Authors interpretation of Dorrenbacher (2000)

How these different concepts and definitions are measured is explained in the following pages.

4.1 Stakeholder transparency

Third party intermediaries, government supported or not, are needed to make CSR reports standardized and comparable (Dubbink, Graafland & Liedekerke, 2008). These intermediaries provide verification, reliability, comprehensibility, comparability and objectivity (Dubbink, Graafland & Liedekerke, 2008). By making CSR reports more comprehensible, stakeholders that might not have the necessary expertise to comprehend the raw information on their own can then also interpreted and compare CSR practices of firms. Furthermore a comparable structure is the only way to allow stakeholders to reward or punish firms that perform above or below average (Hess, 2007). Because of these advantages stakeholder prefer CSR information to be conveyed by third parties (Hess & Dunfee, 2007). In the Netherlands there is the annual Transparency Benchmark that the Dutch Ministry of Economic Affairs has organized since 2004. This benchmark is conceived “to assess the extent to which businesses account for their activities in the area of Corporate Social Responsibility (CSR) in their annual reports3”. The idea behind this Benchmark is to encourage firms to showcase their CSR activities in an open and transparent way and to increase the awareness on CSR policies and improve the performance of firms in this area. It benchmarks around 400 firms on transparency and uses one of the big four auditing firms as third party intermediary. For this research we will therefore use data from this annual Transparency Benchmark.

3

(19)

page 19 van 48

The Transparency Benchmark is devised of two parts of equal importance. (Criteria Transparency Benchmark, 2014). The first part is content oriented and focuses on describing the company, business model, policies, results and the management approach. This part of the benchmark centers on what is in the report. The second part is quality oriented and looks at how well firms reported about the criteria in the first part. It looks at five sub-criteria for quality, namely: relevance, clearness, reliability, responsiveness and coherence. We will look at the chapter that is called “responsiveness” which measures stakeholder transparency. This part measures if a company take stakeholders information needs into account and how firms involve these stakeholders and contains the following sub criterion:

 Insight into stakeholders

 Guaranteeing involvement of stakeholder

 Calculation of stakeholder information requirement  Involvement of stakeholders in CSR aspects

As mentioned earlier the definition given in Rawlings (2009) regarding transparency is “information that is truthful, substantial, and useful; participation of stakeholders in identifying the information they need; and objective, balanced reporting of an organization’s activities and policies that holds the organization accountable” and Stakeholder engagement which as discussed earlier is ‘when firms and their stakeholders are engaged in a dialogue on suitable firm behavior’ is seen as the central and most important factor of CSR reporting (Hess, 2007). We feel confident that this definition is adequately addressed by these four criteria in the transparency index. We use these sub-criteria because publishing a CSR report is not the enabling device for stakeholder engagement but stakeholder engagement is an important ingredient to publish a decent CSR report (Hess, 2007).

The complete list of criteria underlying how firms are scored on stakeholder transparency is displayed in Appendix I.

4.2 Internationalization of corporate governance

(20)

page 20 van 48

Even with this great body of research into the degree of internationalization (DOI) it still remain arbitrary which factors or indicators to use when estimation the DOI of a firm (Dorrenbacher, 2000). For some types of research the use of indexes such at the transnationality index or the network spread index (Ietto-Gillies, 1998) are sufficient. There is however no index related to the internationalization of governance structures. Therefore we will use the three structural indicators given by Dorrenbacher (2000) that measure the extent of internationalization in corporate governance structures. These indicators are previously used to proxy the impact of governance and culture on corporate social reporting (Haniffa & Cooke, 2005; Barako, Hancock & Izan, 2006; Meek, Roberts & Gray, 1995). These indicators will now be introduced one at a time and an explanation why these are suitable is given for each of them.

1. Number of stock markets on which a company is listed

Firms with dual or multiple listing are argued to score higher on CSR related disclosures (Hackston & Milne, 1996). Because when internationally listed firms enter a new capital market they face more diverse transparency pressures in comparison to firms that are only listed at one market (Meek, Roberts & Gray, 1995). We argue that under normal circumstances this should then also hold true for stakeholder transparency. In this paper we have argued that having a Dutch listing is the most demanding with regard to the quality of CSR reporting. However, it might not be just about the number of listings. Oxelheim & Randoy (2003) have argued that having a Anglo-American listing would demand the highest CSR disclosure of a firm. We will therefore also test for differences in stakeholder transparency between firms with an Anglo-American listing and those who have not.

2. Amount or proportion of shares owned by foreigners

(21)

page 21 van 48

3. Number or proportion of non-nationals in the board of directors.

Research into firm performance of Japanese firms has shown that a higher share of foreign directors leads to higher firm performance (Asaba, 2005). This leads us to conclude that it is very well possible that the nationality of board members can influence firm behavior. Research from Masulis, Wang & Xie (2010) support this conclusion. Finding positive and negative influences from having foreign independent directors (FIDs), they gives another dimension to the research conducted before. These same authors find that FID have poor meeting attendance and score lower on morally accepted CEO compensation. Another finding of this research is that firms with more FIDs are more prone to financial misreporting. These last factors are closely related to corporate governance and show support for our claim that internationalization of the board of directors does not automatically has a positive influence on corporate governance and thus transparency. Additionally it is proven that non-executives have a strong influence on CSR disclosure (Matolcsy & Chow, 2007) we will look at the supervisory board and the executive board.

In the paragraph about sample size and data collection the measurements of these variables will be further illuminated.

4.3 Controls

Previous research (Haniffa & Cooke, 2005; Barako, Hancock & Izan, 2006; Meek, Roberts & Gray, 1995; Eng & Mak, 2003; van de Burgwal &Vieira, 2014) also introduced other factors that influences the amount of voluntary disclosure, since stakeholder transparency is an extension on this research we will use these same factors. Because our focus is on the three factors influencing governance we will use the other determinants as control variables. We will use only the variables which had the most clear and predictable outcomes in previous research as controls. These variables are:

(22)

page 22 van 48

can under public pressure lead to new laws and legislation, large firms have stronger incentives to provide voluntary CSR disclosures in their annual reports. Larger firms are also more probable to draw attention from regulatory bodies, giving the same effect (Chan, Watson & Woodliff, 2014). In accordance with stakeholder theory larger firms will disclose more and better information because they are likely to be more complex and have a more dispersed ownership structure. Therefore bigger firms will have a more diverse stakeholders which information needs the firm has to satisfy (Meek, Roberts & Gray, 1995)

Industry. The assumption that the type of industry a firm is active in influences CSR disclosure is generally accepted. Arguments for this assumption are similar to those for firm size. For example Chan, Watson & Woodliff (2014) argue that some types of industry face much higher public pressure then firms in other industries. Furthermore, firms from industries with a greater environmental impact are more likely to disclose CSR related information (Brammer & Pavelin, 2008). Research also showed that in accordance with stakeholder and legitimacy theory, firms in these high profile industries will use greater CSR disclosures and more transparency to legitimize their operations towards society. (Cho and Patten, 2007). Extractive industries or heavy manufacturing are often under more pressure and scrutiny from Non-governmental organizations (NGOs) then other industries (Deegan & Gordon, 1996). This leads to more attentions in society and firms in these industries will therefor provide extensive CSR reports (Brown & Deegan, 1998).

(23)

page 23 van 48 4.4 Multiple linear regression model

To test for interrelationships between the stakeholder transparency score on the Transparency Benchmark and the factors of interest for this research a multiple linear regression is a frequently applied choice (Haniffa & Cooke, 2005; Barako, Hancock & Izan, 2006).

The regression model used is as follows:

TBSresponsivness4 = β0 + β1DutchExBm + β2DutchSuBm + β3DutchSO +

β4OverseasList + β5Listings + β6SizeEmp + β7SizeTA + β8ROE

+ β9ROA + β10Industry + ε1

TBSresponsivness = Transparency Benchmark score on stakeholder transparency DutchExBm = Percentage of Dutch Executive board members

DutchSuBm = Percentage of Dutch Supervisory board members DutchSO = Proportion of Dutch direct share owners

Listings = Number of stock market listings

OverseasList = Company listed on a Anglo-American exchange (0 for no, 1 for yes) SizeEmp = Company size measured by number employees (natural logarithm) SizeTA = Company size measured by total assets (natural logarithm)

ROE = profitability measured by return on equity ROA = profitability measured by return on assets Industry = Industry classification

In the next section the sample design, data collection and measurements will be explained in more detail.

5. Data collection and sample size

We started with a sample containing 133 large Dutch companies that had the necessary information available. After cross checking these companies with the firms that participated with the Transparency Benchmark 2014 we are left with 88 firms. After closer inspection a few more firms dropped out of the sample because of missing values. We therefore ended with a sample size of 81 firms. The respective firms are listed in Appendix II.

We use information that third parties such as ORBIS(Bureau van Dijk) and LexisNexis have subtracted from annual reports. Because TMT has full editorial control over annual reports the information contained within these reports can be presumed to be correct (van der Laan

4

(24)

page 24 van 48

Smith, Adhikari & Tondkar, 2005). The ORBIS database which hold company information for more than 150 million companies is a respected source in European business research (Ribeiro, Menghinello & De Backer, 2010). Almost all data could be gathered from ORBIS only for some missing values we consulted LexisNexis. Important to notice is that we also include a time lag in our research. Time lag has been recognized to have an impact but besides a few exceptions e.g. Barako, Hangcock & Izan, (2006) and Brammer & Pavelin, (2008) research into the determinants of CSR disclosure have not included this (Hackston & Milne 1996). Since the factors underlying the stakeholder transparency score in the Transparency Benchmark have changed over time we are only able to include a one year time lag. Meaning that all our independent variables are gathered for 2013 but the data for our dependent variable is from 2014. This should make sure our research does not suffer from any potential causality problems (Brammer & Pavelin, 2007).

5.1 Measuring the dependent variable - Stakeholder transparency

The data collected for our dependent variable is subtracted from the Transparancy Benchmark of 2014. The main reasons for using this database is already explained earlier in the methodology. Here we will focus on explaining how the scores on stakeholder transparency are measured.

The Transparency Benchmark consist of 40 criteria on which the annual report of participating firms are measured and the maximum score these firms can achieve is 200. There are four main criteria related to stakeholder transparency and a total of 20 points that can be earned regarding stakeholder transparency (Criteria Transparency Benchmark, 2014). These four main criteria and the amount of points that can be awarded are quoted below:

1. The company clearly explains how they involve stakeholders in the policy and activities of the company and how they take their legitimate interests and expectations into account. (10 points)

2. While arranging the report, the company was guided by the information needs of stakeholders (3 points)

3. The company carries a vision on relevant corporate social responsibility themes in its reporting (2 points)

(25)

page 25 van 48

4. The company dares to be vulnerable by sharing challenges, issues or dilemmas publicly (5 points)

The more extensive explanation of these criteria and the allocation of points can be found in

Appendix I. The scores for the Transparency Benchmark 2014 are granted by a team of researchers from EY (previously Ernst & Young). As input in the regression model we use the cumulative score that firms have achieved on these four criteria.

5.2 Measuring the independent variables - international activity Number of stock markets on which a company is listed

The ORBIS database only gives the main stock exchange on which each firm is listed. To find additional listings we used company and trading websites, most noteworthy the Bloomberg Business website5. Following Bancel & Mittoo (2001) the Over The Counter (OTCs) listing are excluded, since this hardly requires companies to alter their disclosure standards. The number of listings will be included as a categorical variable. 0 if a firms has no listing at all, 1 when a firm is listed at one stock market. 2 for two listing and finally a 3 for firms with three listings. Secondly, previous research has uncovered differences between firms listed on mainland Europe and US/UK listed firms with regard to the quality of their CSR disclosures (Meek, Roberts & Gray 1995). It has been suggested this difference in CSR disclosure comes from more demanding stock markets in Anglo-American countries (Oxelheim & Randoy, 2003). To compare firms with an Anglo-American listing against firms with only a European listing we will include a dummy variable. 0 for only European listed firms and 1 for Anglo-American listed firms.

Amount or proportion of shares owned by foreigners

In the ORBIS database, the main shareholders are listed including their nationality. It was however required to use additional information from annual reports or internet sources6 since the ORBIS database has some missing values regarding the total percentage of shares owned.

We follow the methodology used Haniffa & Cooke (2005) and Oxelheim & Randoy (2003) and will include foreign ownership as the percentage of outstanding shares held by non-Dutch investors. We do not specify whether a foreign investor is a company, institution or an

5

http://www.bloomberg.com/europe accessed on multiple dates in May, 2015

6

(26)

page 26 van 48

individual. However since not all shareholders will be willing to carry the monitoring cost we will use sizable ownership as measurement. This means we will use a 1 percent threshold in ownership, before we count that shareholder. This has previously been used as proxy to measure shareholder monitoring and shareholder influence (Brandes, Goranova, & Hall, 2008) Since ORBIS does not use the weighted average of the total ownership stake, an investor (company, person or institution) can be counted double when looking at the total ownership figures7.We therefore only looked at direct ownership to keep the data as clean as possible.

Number or proportion of non-nationals in the board of directors.

For the nationality of the directors we use the “Dutch interlock database” which contains data for the 100 biggest organization in the Netherlands regarding their TMTs and is available at the RUG. The nationality of the member of the executive board in 2013 will be taken into account and as stated earlier we will also use the nationality of the supervisory board in 2013. We will specify the amount of Dutch nationals on the board as a percentage of all board members. This will be done separately for the executive board and the supervisory board. This way we can test if the two boards have different influences on stakeholder transparency.

5.3 Measuring control variables

Firm size. Firm size is also determined from data available in the ORBIS database complemented with information from LexisNexis. Measurements often used for in research to proxy firms size are total assets and the amount of total employees (van de Burgwal & Vieira, 2014) These same authors also stated that using multiple measures for firm size will have a positive effect on the robustness of the results.

(27)

page 27 van 48

Profitability. We follow previous research by Haniffa & Cooke (2005) and use the return on equity (ROE) and return on assets (ROA) as measurement for the profitability of a firm. This data will be gathered from the ORBIS database and is available until 2013.

6. Analysis

6.1 Descriptive statistics

Before we look further into the descriptive statistics of all the variables used in the regression. It might be interesting to first have a look at the individual scores for the stakeholder transparency criteria. As mentioned is the total score for stakeholder transparency compiled from four underlying criteria (see appendix II for a detailed view of these criteria). We will now first look how the firms in our sample scored at the individual indicators that are used to measure stakeholder transparency. Table 1 shows the aggregated score from the 81 firms that compose our sample and compares how much percent of the maximum score these firms scored per question. For example, criteria one (C1) has a maximum score of 10 points. Meaning that if all 81 firms in our sample scored the maximum amount of points, the total should be 810 points. We see however that the total aggregated score for our sample is only 398. This shows that the firms in our sample combined scored only 49 percent of the maximum amount of points.

Table 1, Descriptive statistics criteria stakeholder transparency

C1 C2 C3 C4 Total

Total max score 810 243 162 405 1620

Total score by sample firms 398 135 116 234 883 Percentage of max score 49,14 55,56 71,60 57,78 54,51

Table 1, Descriptive statistics stakeholder transparency

(28)

page 28 van 48

average the firms from our sample score just below 55 percent of all points that can be gathered regarding stakeholder transparency. This clearly shows that there is still a lot of room for improvement on stakeholder transparency. It might also already indicate that it was optimistic to hypothesize that Dutch firms are more prone to stakeholder transparency, or that very few of the sample firms are still truly Dutch.

Table 2 provides the descriptive statistics of the total score for stakeholder transparency and the independent and control variables. Most variables are percentages and this is why the zero is a reoccurring input for the minimum value. Most of these numbers speak for themselves but on few we will elaborate more. First, the NumberOfListings shows that on average the firms from our sample have 1.4 listings. Which might be a bit biased because of some firms are not listed at all. Two listing occurred the most often and almost always was the first listing in the Netherlands and the second listing in Germany. Second, ListedOverseas shows that only 10 percent of the firms have a listing overseas, in our sample this means the London or New York stock exchange. We have broken the industry variable down to four main categories, the table shows that with 40 percent services is the most common industry within our sample, closely followed by manufacturing with 35 percent. Mining, utility and construction stay far behind with only 12 percent of the firms being active in these sectors. The remaining firms are in wholesale and retail.

Table 2, Descriptive statistics

Variable N Minimum Maximum Mean Std. Deviation

TotalScoreTransparency 81 0 20 10,90 7,52 TotalPercentageDutchBM 81 0 100 75,34 25,66 PercentageDutchSB 81 0 100 73,64 29,20 PercentageDutchEB 81 0 100 76,94 32,66 NumberOfListings 81 0 3 1,40 0,88 ListedOverseas 81 0 1 0,10 0,30 OnePercentPlusDutchSharesDirect 81 0 100 66,03 36,80 FivePercentPlusDutchSharesDirect 81 0 100 66,97 43,24 LNEmployees 81 2,77 12,35 8,11 2,24 LNTotalAssets 81 9,71 21,12 15,31 2,40 ProfitabilityROE 78 -46 83,68 9,29 18,99 ProfitabilityROA 79 -27,44 29,93 3,21 7,58 MiningUtilityConstruction 81 0 1 0,12 0,33 Manufacturing 81 0 1 0,35 0,48 WholesaleRetailTrade 81 0 1 0,14 0,34 Services 81 0 1 0,40 0,49 Valid N (listwise) 73

(29)

page 29 van 48

When checking the data for normality8 a few observations had to be dropped from the regression which is why some variables have N < than 81. For all analyses we are however able to include at least N=73 which is more than enough to satisfy the notion of having between 5 and 20 observations per independent variable (Peduzzi, Concato, Kemper, Holford, & Feinstein, 1996). Finally we because strong positive skewness of the size variables we had to use the natural log of total assets and employees.

The data was also checked for mulitcollinearity by running a Pearson correlation test between all variables, which is shown in Table 3. Variables that are significantly correlated are not unexpectedly the variables that are included to measure the same factor. The two profitability variables ROA and ROE have a high and significant correlation (0.801). Dutch supervisory board members and Dutch executive board members are also significantly correlated with each other (0.648). Finally there is a strong and significant correlation between Dutch shareholders with an ownership of 1 percent or more and Dutch shareholders with an ownership of 5 percent or higher (0.871). There is a multicollinearity problem if the correlation between two variables exceeds the 0.8 threshold (Haniffa & Cooke, 2005; Gujarati, 1995). However, when running the regression model the Variance Inflation Factor (VIF) is well below the commonly accepted 10 points for all variables (Lin, 2008). Thus this indicates that mulitcollinearity will not be an issue when using the multiple linear regression model to examine the data for causal relationships among the variables. As extra precaution we did also ran the regression analysis while including only one of the high multicollinear variables at the time. The results of these additional tests did not differ much from our original results but the VIF was indeed found to be lower. This additional results and can be found in

Appendix III.

8

(30)

eTranspa rency eDutchS B eDutchE B OfListing s erseas entPlusD utchShar esDirect ntPlusDu tchShares Direct

oyees Assets ityROE ityROA

TotalScoreTransp arency Correlation 1 Sig. (2-tailed) PercentageDutchS B Correlation ,332** 1 Sig. (2-tailed) ,002 PercentageDutch EB Correlation ,252* ,648** 1 Sig. (2-tailed) ,023 ,000 NumberOfListing s Correlation -,178 -,272* -,286** 1 Sig. (2-tailed) ,112 ,014 ,010 ListedOverseas Correlation ,071 -,350** -,181 ,278* 1 Sig. (2-tailed) ,530 ,001 ,106 ,012 FivePercentPlusD utchSharesDirect Correlation ,074 ,315** ,152 -,357** -,090 1 Sig. (2-tailed) ,512 ,004 ,176 ,001 ,422 OnePercentPlusD utchSharesDirect Correlation ,097 ,527** ,297** -,423** -,205 ,871** 1 Sig. (2-tailed) ,390 ,000 ,007 ,000 ,067 ,000 LNEmployees Correlation ,291** -,098 -,187 ,051 ,213 ,080 -,036 1 Sig. (2-tailed) ,008 ,382 ,095 ,651 ,057 ,476 ,752 LNTotalAssets Correlation ,486** -,055 -,065 -,199 ,303** -,005 -,105 ,399** 1 Sig. (2-tailed) ,000 ,624 ,567 ,075 ,006 ,964 ,349 ,000 ProfitabilityROE Correlation -,074 ,016 -,028 -,049 ,159 ,105 ,177 ,006 ,078 1 Sig. (2-tailed) ,520 ,888 ,807 ,670 ,165 ,361 ,122 ,961 ,498 ProfitabilityROA Correlation -,128 -,033 -,075 ,082 -,017 ,112 ,165 ,054 -,083 ,801** 1 Sig. (2-tailed) ,260 ,775 ,510 ,470 ,882 ,327 ,145 ,637 ,465 ,000

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

(31)

page 31 van 48 5.1 Results

In order to test the three hypotheses a multiple linear regression model was applied. Table 4 shows the results of this regression model. We introduced all the variables belonging to one of the tested hypothesis at once while also including the control variables. While in Appendix III we have included the statistical outcomes that we found when entering only one independent variables one at the time instead of per hypothesis, while keeping the control variables constant.

Model one is included to test for hypotheses one and this hypothesis was included to see if a higher percentage of Dutch board members has a positive influence on stakeholder transparency. The R2 of this model is 0.492 and significant on the 0.01 level. This implies that 50 percent of variance in stakeholder transparency can be explained by our model. When looking at the individual variables we find that Dutch supervisory board membership is significantly to stakeholder transparency (significant at the 0.01 level). For the executive board the relationship between Dutch membership and stakeholder transparency is not significant. However when including the percentage of Dutch executive board members and excluding the supervisory board, this variable also becomes significant at a 0.01 significance level. Al together we can conclude that H1 is supported.

(32)

page 32 van 48

The third model test hypothesis 3 which stated that firms with a higher percentage of Dutch ownership would have higher a degree of stakeholder transparency. This third model includes the percentage of Dutch shareholders that have an direct ownership share of one or more percent in the firm. We also included a variable that is the same as the former but with a five percent ownership threshold. Again we will first look at the fit of the whole regression model. For model three at a significance level of 0.001 the R2 is 0.351. Telling us that 35 percent of stakeholder transparency is explained by this model, which is the lowest of all the models we have tested. From the two included variables only Dutch ownership above one percent has a significant and positive influence and only at a 0.1 significance level. When including only one of these variables in the model the coefficients and significance levels did not change considerably (see Appendix III). In conclusion is the only relation we find at a too low level of significance to support H3.

To test how much of the total change in our dependent variable is explained by our hypothesized independent variables, we have included all these variables in model four. This model appears to be a strong predictor with a R2 of 0.53. Meaning that with a significance level of 0.01 this model accounts for 53 percent of variance in stakeholder transparency. When including all independent variables only Dutch supervisory board membership continues to have a positive influence on stakeholder transparency. Stranger is the fact that with all variables included we find a negative relation between stakeholder transparency and non-listed firms, albeit with a significance level of only 0.1.

Looking at the control variables we see that total assets has a strong and significant influence on stakeholder transparency at a 0.01 level. This influence is as expected visible in all the models. None of our profitability variables proves to be significantly influencing stakeholder transparency. In almost all our models the industry dummy Mining, Utility and Construction has a strong and positive relation with stakeholder transparency with the significance fluctuating between the 0.05 and 0.1 level.

(33)

page 33 van 48

Scores between 1.5 and 2.5 are deemed acceptable, so again our model shows no problems regarding this.

(34)

Table 4, Multiple linear regression output. **significant at the 0.01 level *= significant at the 0.05 level b= significant at the 0.1 level; 1 Listings are included as categorical variable with One listing as baseline; 2 Industry is included as a set of dummy variables with services as baseline industry; 3European listed as baseline,

4_{All F scores are above the critical level for the F-distribution}

Variables B t Sig VIF B t Sig VIF B t Sig VIF B t Sig VIF

(35)

page 35 van 48 6.2 Conclusions and discussion

The aim of this study is to see if greater internationalization of corporate governance structures within Dutch firms, would lead firms to have lower degrees of stakeholder transparency in their CSR reporting. We argued that because of the Netherlands long history with the Poldermodel that Dutch nationality of ownership and directors in firms would has an positive influence on the degree of stakeholder transparency these firms exhibit. Instead of using company specific determinants we use three main indicators that relate internationalization to governance structures; listings, ownership, and board membership

Summarizing the results an placing them in context, our research shows a higher percentage of Dutch board members has a small but positive influence on stakeholder transparency. The strongest influence has the degree of Dutch supervisory board members, which is in line with research that found that non-executive directors have a positive influence on voluntary disclosure (Ho & Wong, 2001). When including the percentage of Dutch nationals on the executive board and excluding the supervisory board the former became also a predictor of stakeholder transparency. This led us to accept H1, which stated that a more international

board would lead to lower stakeholder transparency. This is a very interesting finding, since most of the previous research argued and in some cases supported this with results that a more international comprised board has a positive effect on CSR related measures (Oxelheim & Randoy, 2003; Halme & Huse, 1996; Chan, Watson & Woodliff, 2014; Johnson & Greening, 1999). Whereas our results shows support for the argument that the extensive stakeholder involvement within the Dutch society as result from the Poldermodel, leads to a higher degree of stakeholder transparency in Dutch dominated firms. Providing statistical evidence for the statement “that how the role of a firm and its stakeholders is defined within a society, influences the extent but also the quality of their CSR reporting” (van der Laan-Smith, Adhikari & Tondkar, 2005).

(36)

page 36 van 48

we also looked at the difference between European listed and overseas listed firms, since it is argued that US/UK based firms have different (stronger) views on CSR disclosure then European firms have (Meek, Roberts & Gray, 1995; Oxelheim & Randoy, 2003). Again when looking at stakeholder transparency this relation appear to be not existing. Finally, it is the fact that that non-listed firm did not show a significant negative relation with stakeholder transparency that is unexpected. These firms are reported to have the lowest scores on transparency (Cooke, 1989). When reviewing the data we see however that in our sample most of the unlisted firms are from the banking sector that is currently under heavy scrutiny for the government9 and the public. Probably as a result from these pressures we find all banking related firms show high scores on stakeholder transparency. Furthermore when including all the variables in model four the relationship does become negative as one would expect.

For ownership we find only a very weak link between a higher percentage of Dutch shareholders and higher stakeholder transparency. Interestingly enough it does not shows a statistically significant negative effect between domestic ownership, as some previous research has argued (Barako, Hancock & Izan, 2006; Chan, Watson & Woodliff, 2014). Dutch owners with a direct share of one percent or more in a firm seem to have a positive influence on stakeholder transparency. The effect is however not significant enough to support H3 but it does show that the impact of the relation between domestic ownership and

stakeholder transparency seems to be different in the Dutch setting. Research into block holders showed that less dispersed shareholdings would have a stronger influence on CSR reporting (Brandes, Goranova & Hall, 2008). In our research we also included block holders but did not find any significant relation. When gathering the ownership data we found that the ownership was very dispersed for the firms in our sample. The fact that very little firms had more than a few shareholders with five percent or more shares might be the reason why we did not find any significant affect for this variable.

From our control variables we see that particularly the firm size variable has a positive influence on stakeholder transparency. These findings are consistent with earlier research (Hackston & Milne, 1996; Roberts, 1992). Research into Dutch firms characteristics and their impact on environmental disclosure has also shown a similar impact (van de Burgwal & Vieira, 2014). None of our profitability variables proves to be significantly influencing

9_{The Committee of Parliamentary Inquiry into the Financial System}

(37)

page 37 van 48

stakeholder transparency. Which is also in line with previous research into Dutch firms (van de Burgwal & Vieira, 2014). In almost all our models the Mining, Utility and Construction dummy have also a strong relation with stakeholder transparency. This finding is consistent with comparable research and show that firms active in more sensitive industries report more in order to achieve legitimacy (Roberts, 1992: van de Burgwal & Vieira, 2014).

Concluding, we can state that it is important to remember that difference types of disclosure have different determinants impacting them (Hackston & Milne, 1996). Even now we have still limited knowledge behind the exact determinants of CSR disclosure. This research tried to fill part of of this gap by including the effect of the Poldermodel on how much firms focus on stakeholder transparency. The firms in our sample show the expected interaction between the control variables and stakeholder transparency as research in other settings has suggested. Only when looking at governance structures we found significant different result than previous research. Instead of a higher internationalized board composition, we found that having more Dutch board members has a positive influence on stakeholder transparency. While some result only have a small coefficient they do have an important contribution. These results clearly shows that when looking into topics such as CSR disclosure, transparency or such is in this case stakeholder transparency, national differences matter. This gives an extra dimension to discuss when examining the influence of corporate governance on CSR practices. Keeping in mind that more factors influence high quality CSR practices will also be of interest to firms trying to enhance market value through the mechanisms explained by legitimacy and stakeholder theory or when investors look at firm value.

6.3 Limitations and future research

(38)

page 38 van 48

When looking into foreign share ownership this research does not make a distinction between institutional or corporate investors, while previous research shows that they differ in influence they have on the domestic firm (Douma, George & Kabir, 2005; Barako, Hancock & Izan, 2006). The reason this is not taken into account is due to the time constrains of the research period for this same reason we only included direct shareholders. Because even our complete model accounts for only 53 percent of the measured stakeholder transparency, including institutional investors among other independent variables in future research might provide even more insight in the determinants of stakeholder transparency. The final note on research into transparency comes from Dawkins & Fraas (2013) whom imply that the growing interest for reporting mechanisms e.g. ISO, Social Accountability 800 or the GRI is clearly a reflection of the greater need and expectations that the stakeholders of a company have about disclosure and transparency. however this conformity towards normative standards, will make reporting of companies more and more alike and this might make it more difficult to pinpoint differences between the CSR reporting of individual firms.

One of the main issues remains the fact that we might be able to conclude that Dutch firms which have less internationalized corporate governance score higher on stakeholder transparency. This is interesting since it is not in compliance with the current literature about diversification of governance and its influence on transparency We have argued that the reason behind this is the special role that the Polder model has played in the Dutch history. Because we did however not explicitly explore a direct link between these variables, further trying to reveal this link might be an interesting follow-up research.

The impact of the Poldermodel on stakeholder transparency in CSR reports of Dutch firms.