Has the risk disclosure in the Oil & Gas, and Chemical industry changed during the financial crisis and is the change related to the national culture?

(1)

Has the risk disclosure in the Oil & Gas, and Chemical

industry changed during the financial crisis and is the change

related to the national culture?

Master thesis

Groenlo, Friday 15 August 2014

Keywords: Risk, Risk disclosure, National Culture, Readability

Author: Sven Rouwmaat Student number: S2418096

E-mail: s.rouwmaat@gmail.com Phone: 06-34769647

University: Rijksuniversiteit Groningen / University of Groningen Mentor: Prof. mr. drs. J.T, Degenkamp

2n reader: Prof. dr. J.A. van Manen Supervisor PwC: Paul de Klerk

(2)

Abstract:

The recent years we have had multiple cases in which the management of risk was deemed not sufficient. Examples of this are the financial crisis, Deepwater Horizon Oil Spill and Odfjell. There has been research that also concluded that the risk disclosure was below par. It could be said that the risk disclosure did not reflect the real risks. We have looked into the risk disclosure in two different ways. The first one is a very limited qualitative study existing out of a single interview with a director at a bank and the second is a more extensive quantitative study on the readability and length of the risk paragraph. The main question we wanted to answer was if there is a change in the size of the risk disclosure and how the influence of the cultural aspects has changed between 2007 and 2013. We have found that the bank does not provide addition capital or a costs reduction if an organization discloses more information regarding their risks. Our quantitative findings have shown us that the risk disclosure has significantly increased between 2007 and 2013, and we also found that the influence of the cultural aspects have changed between these years. We concluded that for smaller (not listed) organization it has no use to provide additional risk disclosure, even though previous research has stated that especially the smaller organization would benefit from this. We also conclude that the financial crisis has had a positive influence on the risk disclosure and that national cultural aspects are one of the reasons for this change.

(3)

Foreword

Before you lies my thesis for the Master Accountancy at the University of Groningen. Writing my thesis over the past 6 months at PwC Rotterdam has provided me with a challenge. This was also the challenge I was looking for when I started with this study. During my thesis I had to switch from a qualitative to a primarily quantitative research approach. This did cost me some time and I also lost some motivation for a couple of weeks. Overall I have worked with allot of pleasure on this thesis and learned allot during the process of it.

I especially would want to thank Prof. mr. drs. J.T. Degenkamp for his guidance. There were times that I got stuck and were his thoughts and suggestions helped me allot. I appreciate his fast responses and the feedback he provided me with.

I would also like to thank my supervisor at PwC, Paul de Klerk, for his suggestions and thoughts when I needed them. Finally, I would like to thank everyone else that has showed an interest in my thesis and helped me with the statistical aspects.

Sven Rouwmaat Groenlo, August 2014

(4)

1. Introduction

There is a common understanding within the academic community that the current risk disclosure is too general and of little use to investors (Abraham & Cox 2007, ASB 2009, Lajili & Zéghal 2005, Linsley & Shrives 2006, Solomon et al. 2000). Abraham & Shrives (2014) found that disclosures changes little and they don’t have a relation, or a limited one, with the actual risks that are faced by an organization. It can therefore be seen as “symbolic window dressing” (Carruthers 1995) that exist “without the organization’s following through” (Irvine 2008) and result in the decoupling of the real risks an organization faces (Cormier et al. 2005, Oliver 1991). Despite the above, it appears that there is a demand for improved risk reporting (ICAEW 2011). The need for sophisticated information to manage risk portfolios for companies suggests that they are holding back useful risk information. Why companies hold back information and the extent of it are not completely clear (Marshall & Weetman 2002, 2007).

The current financial crisis that started in 2008 makes it harder for organizations to finance their operations. The wrongful actions of big financial organizations (i.e. Lehman Brothers, Fannie Mae and Freddy Mac) started the current financial crisis (Pirson & Turnbull 2011). Magnan and Markarian (2011) have linked risk disclosure with the current financial crisis. The financial crisis made external financing for organizations more expensive (Ivashina & Scharfstein 2010). One possible way to counter this, is with (voluntary) risk disclosure. This will result in lower costs of finance for the organization (Abd-Elsalam & Weetman 2003, Francis et al. 2005). In our research we looked if the risk disclosure has changed because of the crisis and if the change can be related to the national culture.

International, the disaster at the drilling rig Deepwater Horizon in 2010 has had a lot of media attention due to its impact on the environment and (local) economy1_{. The commission that was tasked with}

investigating the Deepwater Horizon Oil Spill came to the conclusion that the disaster could have been prevented and that BP, Halliburton and Transocean have made a series of mistakes. According to the commission these mistakes have placed doubts on the safety culture of the entire industry. For BP the oil spill resulted in a financial disaster. Forbes has calculated the total costs in the first three years at $42.2 billion2_{. As a result of these financial costs that BP has suffered, they became more vulnerable for a}

(hostile) takeover by their competitors3_{. In the Netherlands there has been a lot of attention on the}

extraction of natural gas in the province Groningen and the effects it has on the region. Recent reports have linked the winning of gas to the increase of earthquakes.

The chemical industry is also one were the impact of risks is high. An inspection at organizations in this industry showed that a lot of them are still making mistakes4_{. But unlike the Oil industry, they did not}

have an incident with international attention in the last couple of years. We therefore also looked if the

1http://www.oilspillcommission.gov/sites/default/files/documents/DEEPWATER_ReporttothePresident_FINAL.pdf (Retrieved February 12th 2014) 2http://www.forbes.com/sites/afontevecchia/2013/02/05/bp-fighting-a-two-front-war-as-macondo-continues-to-bite-and-production-drops/ (Retrieved February 12th 2014)

3http://www.nu.nl/beurs/2983495/shell-overwoog-overnamebod-bp.html (retrieved February 19th 2014) 4http://www.nu.nl/binnenland/2945467/chemiesector-vaak-in-fout.html (retrieved April 29th 2014)

(7)

Deepwater Horizon incident has had an impact on the risk disclosure of the industry, by comparing the oil & gas industry with the chemical industry.

We have looked at the readability scores of the risk disclosures as a part of the disclosure quality and have used the length as an indicator for the increase of risk disclosure. Besides the cultural dimension of Hofstede, we have also looked if the financial position and size have had an effect. We have found that during the crisis the risk disclosure has significantly increased and that we see a change in the cultural dimensions. Overall we can say this change is positive, because the variables that showed a significant negative influence in 2007 have lost their significance in 2013. However the cultural dimension masculinity became significantly in 2013 and has a negative effect on the length.

In chapter 2 we will work out our theoretical framework and form our hypothesis. The methodology is developed in chapter 3 and chapter 4 shows our results. Our findings and conclusion can be found in the final chapter. Besides our findings, we will also describe our limitations and proposals for future research.

2. Literature review

2.1. Risk

For performing a risk disclosure study, a definition of risk is required (Linsley and Shrives 2006). There are multiple words for risk that are broadly used (e.g. hazard, threat or harm) (Lupton 1999). The synonyms mentioned above are associated with the negative side of risk. The Committee of Sponsoring Organizations of the Treadway Commission (COSO) states that risk cannot only have a negative outcome, but also a positive outcome (COSO 2004).

2.1.1. Defining Risk

Knight (1921) established, in his groundbreaking work Risk, Uncertainty, and Profit, the distinction between risk and uncertainty. He states that risk and uncertainty haven’t been separated the way they should be and that we need to see uncertainty as drastically different from the known concept of risk. Functionally risk, within the economic discussions, covers two aspects that are explicitly different and is used too often. Knight continues to argue that these two are distinctly different phenomenon. Risk is often used for quantifiable events, but also for events for which it is not possible to measure them. Risk is uncertainty with a measurable aspect. When a part of the uncertainty is not quantifiable, we speak only of uncertainty (Knight 1921).

Although the work of Knight is almost a century old, it still has relevance. The definition that is used in textbooks for risk is ‘a set of outcomes arising from a decision that can be assigned probabilities’ and when probabilities can’t be assigned to a set of outcomes, we speak of uncertainties (Watson and Head 1998). The definition used by Watson and Head is basically the same as the one established by Knight in 1921. The distinction between these definitions is the measurable (quantifiable) aspect of it. According to Reddy (1996) these definitions reflect events that have occurred during the modern era, previous ideas of risk were connected to natural events (Lupton 1999).

(8)

2.1.2. Risk management

To be able to disclose information regarding your risks, an organization needs to have some kind of risk management in place. For an organization to meet their targets and comply with the set expectations, they need to manage their risks (COSO 2004). One of the frameworks that can be used to manage risk, is the Enterprise Risk Management framework developed by COSO (2004). Enterprise risk management can be seen as an instrument that provides support when the internal process is being disturbed (Emanuels en De Munnik 2006). They define enterprise risk management as “a process that is realized by the board of the

company, management and other personnel, and is used in formulating the strategies and across the complete organization, designed to identify potential events that may affect the company's impact and risk control so that they fall within the risk appetite, to provide reasonable assurance regarding the achievements of entity objectives.”

We previously stated that risk has some sort of uncertainty, but also a measurable aspect (and uncertainty has no measurable aspect). In addition we also stated that these uncertainties provide threats as well as opportunities. By managing the risks, an organization is in a position to effectively coop with these uncertainties and thereby strengthen its capacity to create value (Emanuels and De Munnik 2006). For those organizations that provide a strong ERM capability and discipline, a ERM will provide a competitive advantage (Stoh 2005).

2.2. Stakeholder theory

Freeman (1984) published the stakeholder theory in his groundbreaking book Strategic Management: A

Stakeholder Approach, the concept of “stakeholders”. The idea of the stakeholder idea originated at the

Stanford Research Institute (Freeman 2004). Windsor (1992) stated that the stakeholder theorists have different definitions for a “stakeholder”. A narrow definition was publicised by the Stanford Research Institute (1963), they defined stakeholders as groups “which the organization is depending on for its

continued survival” (Freeman and Reed 1983). A broader definition if given by Freeman (1984) where he

defined a stakeholder as “any group or individual that can affect or is affected by the achievement of a

corporation’s purpose”. With the development of this definition, Freeman took the viewpoint of senior

management. He came to the view that when a group or individual could have an effect on an organization, or be affected by it, that managers then have to look at the needs and wishes of these groups or individuals. This means that an organization needs an explicit strategy, so they can deal with these stakeholders.

There are some misinterpretations and myths that have been addressed by Phillips, Freeman and Wicks (2003). The first one is that “stakeholders are critics and other non-business entities”. This is the total opposite of how the theory was formulated and also does not coincide with the original development (Freeman 2004). Secondly there is the opinion that shareholders have a conflict with the other stakeholders. The point of the stakeholder theory is that the interests of all stakeholders are in line, and the shareholders are one of the stakeholders (Freeman 2004). And finally it has been stated that the

(9)

concept of stakeholder should be used for the formulation of a new non-shareholder theory of the organizations. Freeman (2004) believes that it “is more useful to consider stakeholder theory as a genre”.

2.3. Disclosure

Disclosure is the communication of information regarding a company’s strategies, characteristics, operations and other external factors that have the potential to affect the expected results (Beretta and Bozzolan, 2004). Corporations provide the required information through mandatory disclosure, but can also provide addition information through voluntary disclosure (Healy and Palepu 2001). For mandatory disclosure, the minimum requirements are recorded in the law and regulations. Healy and Palepu (2001) further state that corporate disclosure is critical for the functioning of an efficient capital market; this can also be directed to the stakeholders.

The disclosure of organization has not always been sufficient (Lindsley and Shrives 2006). Berretta and Bozzolan (2004) concluded that organizations mainly focus on giving information that is related to past events or risks that are currently relevant. The most valuable information, regarding future risks, is not disclosed. They also state that managers have the tendency to attribute the risk, that has a negative impact on the organization, to external conditions and don’t look at their own organization to find the cause. This indicates that the attribution theory (e.g. Aerts 1994) could be a factor in risk disclosure (Lindsley and Shrives 2006).

2.3.1. Quantity and Quality

Assessing the disclosure quality is a complex task. It is unclear what the characteristics are that can be used to assess the quality of disclosure and there is also not a clear definition of what disclosure quality is (Botosan 2004). Identifying the measurement of disclosure quality is still one of the more relevant questions that have to be answered (Healy and Palepu 2001). There are different definitions used for disclosure quality. For example, Starreveld (2004) defined information quality as something that is understandable, purposeful and reliable.

Beretta and Bozzolan (2008) have stated that there is not a clear distinction between quantity and quality. Generally it is assumed that the quality of disclosure can be determined by the quantity of disclosure. They conclude that quantity is not a good proxy for the quality of the disclosure. In our study we will therefore not look at the quality of the risk paragraph, but at the readability of it. An increased readability adds to understandability of the report. We therefor see the readability as a part of the quality.

Hypothesis 1: The financial crisis has had a positive effect on the length of the risk paragraph. Hypothesis 2: The length of the risk paragraph has a negative effect on the readability.

2.3.2. Effects

To show why it can be of great importance for an organization to disclose its information, we need to look at the effects that disclosure has. Baiman & Verrecchia (1996) and Verrecchia (1999) have shown that

(10)

firms who are committed to a higher level of disclosure reduce the information asymmetry between managers and the market. The reduction of the information asymmetry leads to an increased market liquidity.

Cheng, Liao and Zhang (2013) looked at the commitment provided by mandatory disclosure and information effect of voluntary disclosure. They found that smaller (a public float5_{lower than $75 Million)}

listed companies experience an increase in market illiquidity when they maintain their disclosure level. Based on the findings they state that mandatory disclosure can’t be fully replaced by voluntary disclosure, because the results suggest that mandatory disclosure provides a credible disclosure commitment mechanism that reduces market illiquidity. Voluntary disclosure however is useful to some extent, as Cheng et al. (2013) find that smaller organizations, who do commit themselves to voluntary disclosure, experience less increase in market illiquidity than firms who do not commit to disclosure. Following these results, Cheng et al. (2013) conclude that reducing the disclosure requirements could reduce the overall financial reporting costs, but have the unintended consequence that market illiquidity is increased as a result of the loss of commitment that is provided by mandatory disclosure. This confirms the earlier findings of Rock (2002) and Stulz (2009). One of the explanations that voluntary disclosure can’t fully replace mandatory disclosure is given by Hail (2011). Managers have the option to observe the content before disclosing the information, giving them the option to only disclose positive information.

Suggestions that could determine managerial disclosure decisions in terms with the agency theory can be found within the voluntary disclosure literature (Hooghiemsta et al.). The literature states that decisions for voluntary disclosure are based on the trade-off between the expected benefits and the costs of disclosing the information. By disclosing information, the information risk will be reduced and resulting in a lower stimulant for investors to price protect themselves (Ogneva, Subramanyam & Raghunandan, 2007). An important aspect of the disclosure theory is that the decisions of a manager to voluntary disclose information is driven by the interpretation of investors. The choice to not disclose information by a manager can have the result that the investors will price protect themselves, because they have a lower amount of information available (Verrecchia, 2001; Hope, Kang, Thomas & Vasvari, 2009). Investors will price protect themselves, because they have an information disadvantage and require a compensation for this disadvantage. Internal control has a positive effect on the quality of the financial reporting and adds to the reduction of the information risk. Investors are thereby reassured that the financial information is reliable as a result of the internal control system. This results in a lower incentive for the investors to protect themselves by asking for a higher reward (e.g., Hooghiemstre et al., Ogneva, Subramanyam & Raghunandan 2007).

The capital needs theory goes deeper into the financing of an organization and the motivation for disclosure and their motivation for disclosure. The theory states that the need for capital is the primary reason for organizations to voluntary disclose information (Abd-Elsalam and Weetman 2003). More disclosure about the operations and organization reduces the uncertainty of external providers of capital

(11)

and will result in lower cost of capital. As a result, an organization will have more possibilities to finance their operations (Abd-Elsalam en Weetman 2003, Francis et al. 2005).

Another reason for managers to disclose additional information, is the potential consequences for the their reputation (Skinner 1994). A manager wants to preserve (build) his credibility and will avoid potential damage to it (Graham et al. 2005). Damage to a manager’s reputation can occur when he only discloses favorable information; this results in a decrease in the effectiveness of the disclosed information and increases the information risk (Hooghiemstra et al.)

On the other side could the disclosure of information lead to legal costs (e.g. when an organization discloses information that is incorrect or when they holds back important information). Besides the reputation damage and legal costs, there are also costs for the collection, processing, assessing and distribution of the information. Finally, an important consequence of the voluntary disclosure is that you also provide your competitors with proprietary information (Verrecchia, 2001). The release of internal control information may include the main risks of a company and how they are managed (Deumes & Knechel, 2008).

Hypothesis 3: A positive financial position of an organization has a negative influence on the disclosure of risks.

2.3.3. Size and disclosure

Hermanson (2000) has found that reporting information on the internal control is important and valuable. The users of this information have stated that the information with regard to internal control influences their decision, because it improves the financial reporting of an organizations and this gives a better view of the continuity of the organization.

There have been multiple studies that found a positive relationship between the size of an organization and the number of disclosures within the general and environmental disclosure (e.g. Firth 1979). Beattie et al. (2004) found that there is a positive relationship with size and disclosure in their sample of UK organizations and in non-UK organizations this relationship has been established by Hossain et al (1995). McMullen (1996) has researched whether organizations voluntary disclose their Management Reports on Internal Control (MRIC). One of the arguments was that the disclosure of the MRIC would improve the financial reporting. In his research he found that 11,5% of the small organizations voluntary disclosed their MRIC and that none of these organizations had problems with their financial reporting. In his research he classified companies with less than $250 million in total assets as small. From the 114 large companies that did have problems with their financial reporting, only 10,5% had voluntary disclosed their MRIC. In his conclusion, McMullen (1996) states that there is an indication that there is a relation between the presence of a MRIC and the absence of financial reporting problems for smaller organizations.

Previous empirical research has provided evidence of the effect of size (Adihikary & Tondkar 1992, Ahmed & Courtis 1999, Belkaoui & Kapik 1989, Cooke 1992, Firth 1979, Healy & Palepu 1994, Richardson

(12)

& Welker 2001, Robb et al. 2001). The research of Beretta and Bozzolan (2004) into the quality of risk disclosure confirms that there is an association between size and risk disclosure with their sample of Italian organizations. Their theoretical framework used the dimensions density (the ratio between the numbers of sentences in which risk information is provided over the total number of sentences included in the MD&A) , depth (the content of information disclosed regarding the expected economic impact of identified risks upon future performance), outlook and relative quantity. They found no significant evidence that size has an effect over density or outlook, but did find it with depth. Beretta and Bozzolan (2004) did not find any significant relation between the industry and the disclosure.

Following the research of Beretta and Bozzolan (2004), Linsey and Shrives (2006) found proof that the level of information disclosure in annual reports of UK based organizations increases with the size of the organization and that this reflects the findings of general disclosure studies (e.g. Cooke 1992).

There are multiple explanations for why size has an effect on the disclosure. Smaller organizations have less resources to collect and disclose information compared to large organizations, this results in higher cost of reporting if they want to meet the same level of disclosure (Buzby 1975). As a result, the benefits of lower agency costs do not outweigh the costs of reporting. This argument however, can have lost its value due to the developments of information systems. Another argument is that larger companies tend to have a higher proportion of outside capital (Leftwich, Watts and Zimmerman 1981). This is also mentioned in the agency theory by Jensen and Meckling (1976). This results in higher agency costs for large organizations and increases the likelihood that large organizations provide more disclosure regarding their operations.

Another explanation is that the disclosure of information requires a higher level of financial communications and thus an increased need for skilled employees (Cooke 1989). Cooke (1989) further argues that larger organizations have more capabilities to attract employees with the required skills and thus is disclosure enabled by the size of an organization. Deumes and Knechek (2008) further state that because larger organizations have more activities, they have more inherent risks and this makes it more obvious that large organizations report more on their internal control.

Hypothesis 4: The size of an organization is a positive influence on the length of the risk paragraph.

2.4. Culture and the link with disclosure

2.4.1. Culture

Mueller et al (1991) stated that it makes sense that the environment in which an organization operates affects the (financial) reporting and disclosure. The culture is part of the environment (Hope 2003) and is defined by Hofstede and Bond (1988) as “the collective programming of the mind that distinguishes the

member of one category of people from those of another. Culture is composed of certain values, which shape behavior as well as one’s perception of the world.“ The perception that the culture of countries from the

same continent is similar has been addressed by Hofstede (1984). He analyzed that even within Europe we see that cultures can be very different. Another example is found by O’Grady and Lane (1996), who

(13)

found in their study about businesses that expanded their operations from Canada to the United States that there is a psychic distance paradox. The paradox is about the assumptions that we have about physically close countries and that they prevent us from learning the critical differences between the cultures.

Many researches that examined the relationship between culture and other variables are based on the work of Hofstede (1980). Based on his research, which focused on more than 100.000 employees in 39 countries, Hofstede came up with a couple of cultural dimensions. The extensive use of the cultural dimensions of Hofstede has been criticized by Harrison and Mckinnon (1999). They recommend conducting further developing these dimensions. Furthermore they also criticize the extensive use of the cultural dimensions of Hofstede. The use of the cultural dimensions of Hofstede has four possible weaknesses that can occur (Harrison and Mckinnon 1999). Firstly, there is a problem to consider the totality of the cultural domain. There are studies that do not include all the dimensions and also lack the motivation for the absence of one or more dimensions. Secondly, the dimensions of Hofstede fail to measure the possible differential intensity of cultural norms & values. According to Lachman, Nedd and Hinings (1994) “not all values are equally important (in all nations), or have the same impact in regulating

behavior. Cultural values ought to be differentiated in terms of the impact they have in legitimizing and directing choices of modes of organizing and patterns of managerial behavior”. A value that is more central

within a nation’s culture has relatively more impact on organizational and managerial practices compared to a nation where this same value is less important. We can therefore not include or exclude a dimensions solely by the score. Thirdly, because of the very simplistic treatment of culture, the dimensions fail to capture the greater depth and complexity of the concept culture. Results from previous research that show an association with a particular score on a dimension can therefore not necessarily be applied to other nations that have the same score on the same dimension. Finally, Harisson and Mckinnon (1999) state that the dimensions fail to capture the complexity of the human actions and that they do not explain the characteristics and dynamics of specific groups. Even though the dimensions of Hofstede have these weaknesses, there is no other research into culture that has sufficiently improved his approach. Therefor Hofstede’s dimensions are still used in the recent cultural studies (Jansen, Merchant and van der Stede 2009). Because of this, we will also be using the dimensions of Hofstede.

The dimensions that Hofstede distinguished during his initial research are Power Distance (PDI), Individualism versus Collectivism (IDV), Masculinity versus Femininity (MAS) and Uncertainty Avoidance Index (UAI). Hofstede has used these four dimensions for the positioning of the cultures of different countries. Previous research has mostly used the dimensions uncertainty avoidance en individualism (Hope 2003). More recently Hofstede has added the dimensions Pragmatic versus Normative (PRA) and Indulgence versus Restraint (IND). We can therefor say that Hofstede has (partly) addressed the criticism of Harrison and Mckinnon (1999) who stated that further development was needed. Below we will introduce the definition of the six dimensions6_:

(14)

1. Power Distance is a characteristic of culture that defines the extent to which the less powerful

person in society accepts inequality in the hierarchical distribution of power (Hofstede 1984). Power distance has an effect on the less powerful as well as the more powerful members of society (Gray 1988). It has an effect on both, because the more powerful members have the tendency to keep their power and the less powerful members have to strive for equalization of power. Hofstede (1984) states that the extent of power distance, that is tolerated, differs between cultures and some societies tolerate more inequality than others.

2. Individualism versus collectivism is the 2nd_{dimension we will describe. They, individualism and}

collectivism, are cultural characteristics that oppose each other (Hofstede 1984). A individualistic society is defined by the preference of the individual to be expected to take care of only themselves and their families. With collectivism, the opposite, are societies where individuals can expect that either their relatives or other members of a group will look after them in exchange for his (or her) loyalty. This dimension is defined by the terms “I” or “we”. It has been argued by Gray (1988) that this dimension shows the degree of interdependence of the society that is maintained among the individuals.

3. Masculinity versus Femininity are also characteristic of culture that oppose each other (Hofstede

1984). The score for this characteristic is related to the different social rules of women and men in the society. In this dimension the masculinity is related to the preference in the society for achievements, heroism, assertiveness and the material reward for success. A higher masculinity score means that the larger portion of the society has a competitive nature and therefor the culture is more competitive. The opposite preference for cooperation, modesty, caring for the weak and quality of life represents the femininity side. A lower score means that the larger portion of the society is more consensus-oriented.

4. Uncertainty Avoidance is a characteristic of culture that shows the degree to which members of a

society feel uncomfortable with uncertainty and ambiguity (Gray 1988). A situation that they consider to be unstructured, unclear or unpredictable influences the behavior of these members (Hofstede 1984). A nation that maintains rigid codes of beliefs and behavior and are intolerant of unorthodox behavior and ideas, exhibits a strong UAI. Societies with a weak UAI tend to be less aggressive, unemotional and have relatively tolerant behavior. Weak UAI societies also tend to be more relaxed in which practice counts more than principles. This dimensions represents the degree of how societies want to control what happens in the future (Gray 1988).

The 2010 research of Minkov has resulted in two additional dimensions. Originally the paper was published online in December 2010. The research is later published as a paper by Minkov and Hofstede (2012). By using the World Values Survey, Minkov and Hofstede (2012) have found evidence for two

(15)

additional dimensions. Below we will get into these two dimensions7_{, we explain them based on the}

description on the website of The Hofstede Centre8_.

5. Pragmatic versus Normative is the 5th_{dimension. This is not a completely new dimension, but a}

replication of the dimension previously known as Long-Term-Orientation (LTO). Because of Minkov’s research, the scores could be extended from 23 to 93 countries. The scores on the website9 _{are now based on Minkov and Hofstede (2012). This was previously based on the}

research of Bond and Minkov, but the construct were not fully identical.

There happens a lot around us that cannot be explained. This dimension describes how people will relate to that. Within a normative culture, the people will want to explain as much as possible. The people in this kind of society want to establish the absolute truth and require personal stability. These societies show a relative small tendency to save for the future, are focused on gaining quick results and show a great respect for traditions and social conventions.

When a society does not feel the need to explain everything, we speak of a pragmatic orientation. People in this kind of society believe that is impossible to understand the complexity of life to the full extant. These societies have the challenge to live a virtue life, instead of the challenge to know the absolute truth. Pragmatic orientated people see the truth as something that depends on the situation, context and time. They also are able to accept contradictions, adapt to circumstances, have a strong tendency to save and invest, determination and thriftiness to achieve results.

6. Indulgence versus Restraint is a new dimension that also came out of the research from Minkov

and Hofstede (2012). This dimension was added in the 2010 edition of Cultures and

Organizations.

Indulgent societies are characterized by a relatively free gratification on the basics and natural drives that are related to enjoying life and having fun. Restraint societies, as an opposite, show a suppressed gratification of the needs and are regulated by strict social norms.

2.4.2. The link with disclosure

Hofstede’s (1980) research was initially not intended to explain the relationship of the national cultures with the (financial) disclosure, but was mainly focused on examining the behavior of cultures within organizations. Because of this, Gray (1988) was one of the first researchers who have tried to link Hofstede’s cultural dimensions with disclosure and the development of accounting systems. For his framework, Gray has used the (first) four cultural dimensions of Hofstede and used them to further extend Hofstede’s research by establishing the connection between the cultural dimensions and the accounting values. Gray (1988) has found that the dimensions Professionalism versus statutory control; uniformity

7http://geert-hofstede.com/national-culture.html 8http://geert-hofstede.com/dimensions.html 9_{http://geert-hofstede.com/countries.html}

(16)

versus conformity; conservatism versus optimism and; secrecy versus transparency can be used to define the accounting culture of a country. The more recent study by Finch (2009) has concluded that not all four dimensions of Gray have a relation with the accounting system of a country. He states that only the last two accounting value dimensions can be used for the measurement and the disclosure of accounting information.

Conservatism versus Optimism

In accordance with an couple of the dimensions from Hofstede, the values that are identified by Gray (1988) are also opposites of each other. Within this dimension, the conservatism stands for a cautious approach and optimism for a risk-taking approach for the (financial) reporting. There is a strong variation in conservatism between countries. Choi and Mueller (1984) have found that countries like France and Germany are strongly conservative compared to the United States and United Kingdom. It is argued by Gray (1988) that these differences are reinforced by an couple of factors. These factors are ‘the relative development of capital markets, the differing pressure of the user interests, and the influence of tax laws on accountants in the countries’ (Gray 1988).

Gray (1988) has linked conservatism closely with Hofstede’s uncertainty avoidance (4). One way of dealing with uncertainties, or avoiding them, is by taking an conservative approach with the (financial) reporting. Furthermore Gray (1988) also argues that conservatism has an (weaker) link with individualism(2) and masculinity(3). The link exists based on the emphasis that the individual and masculine behavior influences the level of conservatism in a way that it makes them more optimistic. The connection between power distance and conservatism was not established.

Secrecy versus Transparency

Gray (1988) has defined secrecy as a more confident and restricted preference to disclose information and transparency is a more transparent and open approach to the disclosure of information. The managers and/or accountants can influence this dimension, because they are the ones who determine what information is disclosed to outside parties. The secrecy can vary between countries from a low level to a high level of disclosure (Choi and Mueller 1984). The level of secrecy is influenced by the development of the capital markets and the nature of share ownership (Watts 1977). This can lead to additional voluntary disclosure of (financial) information. Because the US is more shareholders orientated than Europe, we expect an increased amount of voluntary disclosure from US based organizations. An explanation for this could be that the shareholders demand more information than they legally have to provide and thus voluntary disclosing this information. Because governments can oblige organizations to disclose more information, there is a somewhat causal relationship between secrecy and disclosure. Governments will have different rules and regulations regarding the disclosure, which will cause a difference in the level of disclosure per country.

Gray (1988) has linked secrecy mostly to uncertainty avoidance (4). This means that when a country scores high on the uncertainty avoidance, they have a higher preference of secrecy. This is because when people have a tendency to avoid risks, they will restrict more information. Gray has also indicated a strong

(17)

connection with Hofstede’s Power Distance dimension (1). By restricting information from outsiders, the people who already are in the positions with more power can make sure that they can preserve their power over the others. Zareksi (1996) has argued that societies with a high power distance have more businesses that discourage the extensive sharing of information. This could imply a high amount of secrecy. Gray (1988) further states that there can also be a relationship with masculinity (3), because societies with a lower masculinity are striving for a more open community. This relationship, between masculinity and secrecy, is however less important than the one of secrecy with the power distance and uncertainty avoidance. As a final possible relationship, Gray (1988) sees a possible relationship of secrecy with individualism (2). This is because people within a individualistic society are more preferable towards a transparent society than one with allot of secrecy. This means that when an organization puts its focus on individualistic behavior, the employees of this organization require more information from external parties in comparison with a focus on collectivism. The assumption is therefore that a high individualism will make it more likely to have a lower degree of secrecy.

For our hypothesis we will mainly focus on the dimension of secrecy versus transparency. This is because the relationship of Hofstede’s dimensions with the Conservatism is limited. For instance, there is no relationship between Conservatism and Power distance. This in combination with the weak and unclear relationship between masculinity and individualism with regards to conservatism made us mainly use the secrecy dimension of Gray’s framework. We can do this because there is a relation between conservatism and secrecy. They both generate, in some way, the same cautious behavior.

Research of Jaggi and Lowe (2000) has also suggested that countries that have adopted Common law a higher disclosure of information than what is the case in countries with Code Law. However they did find a significant relationship between the culture, accounting disclosure and countries that have adopted Common Law (Jaggi and Low 2000). This research is extended by Hope (2003) and they came to the same conclusion that there is no significant relationship between disclosure and Common Law.

There has been a wide array of research done to test the hypotheses of Gray (1988). Finch (2009) has summarized this research in one article. A part of this research has been aimed at the relationship that exists between the culture and the disclosure. The research of Zarzeski (1996) was conducted for seven countries and came to the conclusion that secrecy; individualism, masculinity, and uncertainty avoidance have an influence on the disclosure of organizations. Gray and Vint (1995) have further found that disclosure has a significant negative relationship with uncertainty avoidance, but they did not find a significant relationship for the cultural dimensions power distance and masculinity. Furthermore, Hooghiemstra et al. have looked at the effects of culture on the disclosure of internal control information. They have found that individualism has a positive effect on the disclosure and that there is a negative association between uncertainty avoidance and disclosure. We refer you to the research of Finch (2009) for a more complete overview of all the research based on Hofstede’s (1980) and Gray’s (1988) research.

Hypothesis 5: The national cultural dimension Individualism is positively associated with the length of risk disclosure.

(18)

Hypothesis 6: The national cultural dimension Individualism is negatively associated with the readability of risk disclosure.

Hypothesis 7: The national cultural dimension Uncertainty Avoidance is negatively associated with the length of risk disclosure.

Hypothesis 8: The national cultural dimension Uncertainty Avoidance is positively associated with the readability of the risk disclosure.

(19)

3. Methodology

At first we wanted to perform a qualitative study into the risk disclosure of smaller organization’s that operate within the Oil & Gas-, and Chemical industry. Due to a low interest in interviews from both financial institutions, other stakeholders and organizations in the industries we have decided to change our research to a quantitative study. The main reason to change our research was the time it would take to acquire the needed amount of interviews and the rather limited time for this thesis.

To give a small insight into our original research proposal, we have done an interview with a director at an organization within the financial service sector. The interview was done on May 7th _{2014. For this}

interview we have setup an initial questionnaire (Appendix C) that was used as a guideline for our interview. The main purpose of the interview is to get an insight into the role that risk disclosure, in any form, can perform for the financing of smaller organizations. The interview has been done in Dutch and the used quotes are translations of this interview. Because we only have one interview, the qualitative results have very poor validity.

Below we will get into the sample selection, data selection and the variables we have used for our qualitative research.

3.1. Sample

We have selected 100 organizations for our sample divided over two industries; this results in 50 organizations to represent the Oil & Gas industry and 50 for the Chemical industry.

For the selection of the organizations we used the Orbis database. This database gives us the ability to make a first selection based on multiple criteria. We selected all organizations that are ‘ultimate owners’, have the financial data for 2007 & 2013, and have the liquidity ratio for 2007 & 2013. For the industry selection we have used the primary US SIC – code (US SIC 13 & 291 for the Oil & Gas industry and 28 for the Chemical industry)10_{. This selection left us with 262 organizations for the Oil & Gas industry and 552}

organizations for the Chemicals and allied products.

Following this selection, we have deleted all the organizations that missed data for a variable (i.e. market capitalization). This resulted in the removal of 70 Oil & Gas organizations and 172 Chemical. We have removed 4 Oil & Gas organizations and 8 Chemical because we were not able to explicitly define the legal system of Kuwait (“mixed legal system consisting of English common law, French civil law, and Islamic religious law”) and Saudi Arabia (“Islamic (sharia) legal system with some elements of Egyptian, French, and customary law; note - several secular codes have been introduced; commercial disputes handled by special committees”).

10_{At the time of our selection, not all organizations have published the annual report for 2013. Reproducing this research could result in a larger} selection in the future.

(20)

The remaining 185 Oil & Gas, and 372 Chemical organizations were separately randomized11_{resulting in a}

random order for each industry. Because not all organizations could provide us with the required date (for example, no English annual reports were available or the annual reports were locked for copying information) for the reading scores, we have selected the first 50 viable organizations.

3.2. Dependent variables

Readability score

There are a couple of formulas that can be used to quantify the readability of a text. One of the methods is the Flesch Reading Ease Formula. This formula is designed for English texts and aimed on quantifying the texts in difficulty levels for adults. The formula scores the readability of texts from ‘very easy’ to ‘very difficult’ (Li 2008). A higher score indicates that the material is easier to read.

Another method to measure the readability of a text is the Gunning Fog Index. This measuring method has been used a lot to measure the readability of financial disclosure (Li 2008). This formula is used to determine the complexity of a text. When the results from this formula are higher, it means that it it’s difficult. There has been some criticism on the use of this method. Loughran and McDonald (2014) argue that words like ‘company, business and directors’ are in every annual report and will be classified as difficult words. The users of the annual reports will not be distracted by these words and they also have some criticism for the use of the amount of words per sentence. All together they conclude that the Gunning Fog Index is not a good measure for the readability of annual reports.

Figure 1: Typical Fog Index12

Even though the Gunning Fog Index has had a lot of criticism, it is still used a lot in recent research into the readability of the annual reports (i.e. Li (2008), Biddle, Hilary, and Verdi (2009), Miller (2010), and Lehavy, Li, and Merkley (2011), Dougal, Engelberg, Garcia and Parsons (2012), and Lawrence (2013).). Because the Gunning Fog Index is mostly used for the readability of annual reports, we will also use this method to measure the readability. Besides the Gunning Fog Index we will also use the Flesch Reading Ease in our analysis.

For measuring both scores, we used a program called R and use the koRpus package. R is a “language and

environment for statistical computing and graphics”13_{. The koRpus package is designed as an add-on for R}

to analyze texts14_.

11 http://textmechanic.com/Random-Line-Picker.html

12 http://scientopia.org/blogs/whizbang/files/2011/06/FogIndex.png

Fog Index Examples

6 TV Guides, The Bible, Mark Twain 8 Readers Digest

8-10 Most popular novels 10 Time, Newsweek 11 Wall Street Journal 14 The Times, The guardian 15-20 Academic papers

Over 20 Only the government sites can get away with this, because you can’t ignore them. Over 30 The government is covering something up

(21)

Figure 2: Flesch Reading Ease Score Index15

For the analyzing of the text we have replaced dots that don’t function as the end of a sentence with a comma. This is because the program recognizes each dot as an end of sentence and this influences the readability score. All the texts are saved as ‘plain text’ (.txt) files.

3.3. Independent variables

3.3.1. Size of the risk paragraph

The size of the risk paragraph will be determined based on the amount of words used and will also be determined by using R in combination with the koRpus package. The formula for this can be added into the same analysis as the readability.

3.3.2. Culture

We use the cultural dimension of Hofstede as variables for culture. The website of The Hofstede Centre has quantified each dimension for 93 countries. Previous research uses the (first) four dimensions, in our research we will also add the extended 5th_{and the new 6}th_{dimension. Our literature review showed that}

namely Individualism and Uncertainty Avoidance are related to the disclosure of information, but also that previous literature got criticism for not using all the variables. In our analysis we found multicollinearity with the dimension Pragmatic and have therefore removed this dimension. So for culture we will use the following variables:

 Power Distance Index (PDI)

 Individualism (IDV)

 Masculinity (MAS)

 Uncertainty Avoidance Index (UAI)

 Indulgence (IND)

3.3.3. Size

We use the size of an organization, as it was shown in prior research that it influences voluntary disclosure (e.g. Hooghiemstra et al., , Ahmed and Courtis, 1999). We measure the size of the company based on their total assets. All the assets are in US Dollars and gathered with by using ORBIS. We have used a natural logarithm of the total assets to get a normal distribution.

13_{http://www.r-project.org/}

14_{http://cran.r-project.org/web/packages/koRpus/index.html}

15 http://www.readabilityformulas.com/flesch-reading-ease-readability-formula.php

Flesch reading Ease Score Examples 90-100 Very easy 80-89 Easy 70-79 Fairly easy 60-69 Standard 50-59 Fairly difficult 30-49 Difficult 0-29 Very confusing

(22)

3.3.4. Liquidity ratio

Based on our literature review we expect that organizations with a lower liquidity ratio disclose more information. We found that prior research states that the need for capital is the primary reason for organizations to disclose information. We will use the liquidity ratio as an indication for the capital needs of an organization and as an indication if they need additional financing. We will acquire the liquidity ratio’s for both years from the Orbis database.

3.3.5. Debt ratio

Prior research has indicated that investors will price protect themselves when there is information asymmetry. We use the debt ratio, because based on this literature we expect that the external financers will price protect themselves (e.g. demand a higher reward for the risk) when there is no disclosure. Organizations will use the disclosure to reduce this information asymmetry. The debt ratio is not available in Orbis, we therefor will calculate it based on the information in Orbis. For this we will use the ‘Total Liabilities and Debt’ and ‘Total Assets’ for both years.

3.4. Control variables

3.4.1. United States

The research of Hooghiemstra et al. did not include organizations based in the United States. They excluded these organizations, because SOx legally mandates that US listed organizations report about their internal controls. We will use a control variable to see if there is a difference between US organizations and non-US organizations.

The use of Gray’s (1988) research is a limitation, because it does not include the developments that occurred after his research (e.g. the implementation of SOx legislation in the USA, adoption of IFRS in 2005). By using this control variable we will partly reduce this limitation, because we make a distinction between US firms and non-US firms. A publication from PwC (2013) showed that worldwide 113 countries have implemented IFRS in some extent. We therefore assume that non-US firms comply with IFRS and the US organizations are the only ones that have to comply with the SOx legislation. We defined all US firms with a 1 and organizations from other countries were given a 0.

3.4.2. Industry

The two industries in our data sample will be the second control variable. It will provide us with some insight if there is a difference between the Oil & Gas industry and the Chemical industry in terms of risk disclosure. Both industries were separated with a dummy variable; the chemical industry has been defined with a 0 and the Oil & Gas industry with 1.

3.4.3. Legal system

Prior research (e.g. Jaggi and Low 2000, Hope 2003) has shown that there is no connection between the culture and accounting disclosure in Common Law countries. Because these could influence the overall results, we will separate the common law and the code law countries.

(23)

We determined the legal system based on the information on the website of the Central Intelligence Agency16_{and the article of Mahoney (2001). There were two countries, Thailand and Mauritius, for which}

the defined legal system was different in these sources. We have made the choice to use the information from the Central Intelligence Agency, who defined their legal system as Civil Law with influences of Common Law. The law system was defined with a dummy variable; we used a 0 for code law and 1 for common law.

3.4.4. Years

We have used a dummy variable for the years. The year 2007 was given the dummy variable of 0 and for 2013 we used 1.

(24)

3.5. Framework & Statistical Models

Below you will the framework that is based on the literature review and the methodology. The effects of disclosure are a result of the risk disclosure interpretation of the different stakeholders. The triangle ‘Risk Disclosure, Stakeholders and Effects of Disclosure’ is not used in our statistical analysis, because these connections are based on the literature that was gathered for the interviews and would have gotten the main attention if our initial research could have been performed.

The above framework will be applied to our depended variables, the length and readability, to explain the change in the risk paragraph of organizations within the Oil & Gas, and chemical industry. Below are the three models we will use in our regression analysis.

Based on the model above, we have made the following three models that reflect our main regression analysis. The first model uses the Gunning Fog Index score as the dependent variable. A higher Gunning Fog Index Score indicates a lower readability. Thus a positive beta in the analysis will indicate a decrease of the readability. Model 1: ܴܾ݈݁ܽ݀ܽ݅݅ݐݕݏܿ݋ݎ݁(ܨܱܩ) = ߚ଴+ ߚଵܮ݁݊݃ݐℎ +ߚଶܻ݁ܽݎ+ߚଷܲܦܫ+ߚସܫܦܸ+ߚହܷܣܫ+ߚ଺ܯ ܣܵ+ ߚ଻ܫܰܦ+ ߚ଼ܮ݅ݍ. ܴܽݐ݅݋+ߚଽܣݏݏ݁ݐݏ+ ߚଵ଴ܦܾ݁ݐ. ܴܽݐ݅݋+ߚଵଵܥ݋ݑ݊ݐݎݕ + ߚଵଶܫ݊݀ݑݏݐݎݕ +ߚଵଷܮ݈݁݃ܽܵݕݏݐ݁݉ + +ߝ

Risk

Disclosure

Stake-holders

Effects of

Disclosure

Control variables

Industry

Law

SEC/US

Independent variables

National

Culture

Debt

Liquidity

Size

(25)

The second model is about the Flesch Reading Ease. For this model we expected mostly the same results as with our first model. Because a lower Flesch Reading Ease score indicates a lower readability, the opposite of the Gunning Fog Index model is expected for the beta.

Model 2:

ܴܾ݈݁ܽ݀ܽ݅݅ݐݕݏܿ݋ݎ݁(ܨ݈݁ݏܿℎ) = ߚ଴+ ߚଵܮ݁݊݃ݐℎ +ߚଶܻ݁ܽݎ+ߚଷܲܦܫ+ߚସܫܦܸ+ߚହܷܣܫ+ߚ଺ܯ ܣܵ+

ߚ଻ܫܰܦ+ ߚ଼ܮ݅ݍ. ܴܽݐ݅݋+ߚଽܣݏݏ݁ݐݏ+ ߚଵ଴ܦܾ݁ݐ. ܴܽݐ݅݋+ߚଵଵܥ݋ݑ݊ݐݎݕ + ߚଵଶܫ݊݀ݑݏݐݎݕ +ߚଵଷܮ݈݁݃ܽܵݕݏݐ݁݉ +

+ߝ

With our third model, we will analyze the length of the risk disclosures and what variables are of influence on this. This model has one less variable, because we don’t expect that the readability be of influence on the length.

Model 3:

ܮ݁݊݃ݐℎ = ߚ଴+ ߚଵܻ݁ܽݎ+ߚଶܲܦܫ+ ߚଷܫܦܸ +ߚସܷܣܫ+ߚହܯ ܣܵ+ߚ଺ܫܰܦ+ ߚ଻ܮ݅ݍ. ܴܽݐ݅݋+ߚ଼ܣݏݏ݁ݐݏ+

ߚଽܦܾ݁ݐ. ܴܽݐ݅݋ +ߚଵ଴ܥ݋ݑ݊ݐݎݕ + ߚଵଵܫ݊݀ݑݏݐݎݕ +ߚଵଶܮ݈݁݃ܽܵݕݏݐ݁݉ + ߝ

We have also done multiple regressions for each of the models for the separate years. These results have provided us with a possible indication if the culture is an explanation for the change in the dependent variables. In the next chapter we will first get into the qualitative results from our single interview and will later get into the statistical results.

(26)

4. Results

4.1. Qualitative

Our initial research was designed to look at the risk disclosure of smaller organization within the Oil & Gas, and Chemical industry. We noticed a very poor interest from organization in these industries and were also not able to get a decent amount of interviews with their stakeholders. This is why we have decided to switch to a quantitative research design. We were able to get one interview with one of the directors at a bank in the Netherlands. The risk disclosure has been and still is an important aspect of his work. In this section we show you the results of this interview.

“The bank does business with other parties on the basis of openness. This includes a careful and very often by accountants certified report on the risks and how they control them”

Banks need to have an open relationship with their clients, otherwise there is no basis for a healthy relationship. The risk disclosure and the control activities in the annual reports that are certified by accountants are an important aspect of the risk assessment of a client. If the bank finds that the disclosed risk information is not enough, they need to gather the information themselves to get a sufficient risk assessment. Only if all conditions of the bank are met, they will finance the operations of an organization.

“…This is especially the case in the current situation of regulation and supervision of banks and accountants. The financial sector (i.e. banks, accountants, assurance companies etc.) currently has the duty to care for a

sufficient assessment.”

The increase of regulation and supervision for the financial sector has made the risk assessment more important. It is now more than ever important that the financial sector performs its duty to care for a sufficient assessment, so that they reduce the chances of repeating past mistakes. This has changed how banks do their assessments. Were they previously could make due with an accountant certified report, they now have to verify the risks to prevent possible social and legal damages.

“The bank also puts requirements on their internal process. The sustainability of the company, the social structure of the company and the environmental impact of a project or loan proposal have become more

extensive.”

The bank has put certain requirements on their assessment of an organization or project. These requirements are not always present in the annual reports and the bank is not going to wait for this to become mandatory. Instead they are setting these requirements themselves and there are companies who are responding to these requirements by addressing these items in their annual reports.

The bank itself has a, voluntary, sustainability report and prior to that also publicized a social report. Companies with a social responsibility don’t always want to wait for new regulations regarding the disclosure of information. Instead they want to move with the social discussions. This shows that larger organizations with an increased social responsibility are more willingly to voluntary disclosure information regarding their operations.

(27)

“There are even those who say as mankind could better assess risks, to that extent is mankind progressed.”

For the banks it is about both the quantitative as the qualitative assessment, where by the quantitative part can be difficult. It is possible for industrial products, but when it comes to the hedging of interest risk it can become difficult. There is a mathematical reality with which you say, “this could happen when you hedge interest risks”. As the current financial crisis has proven, the banks have failed in doing this in the past and were also too optimistic. As a result of this the banks are attempting to quantify more risks and uncertainties. The qualitative assessment is made of the quality of the management, the current situation of the organizations and the quality of their operations.

“This all is very complex.”

It is mostly not possible to quantify the qualitative assessments of the organization and this remains to be assessed by the account manager who then has to make a judgment based on his own experience. So even though the financial sector is trying to quantify more risks and information, there is still a need for the interpretation of the qualitative information in combination with the quantitative information. For instance, an organization with a brilliant new concept that possible has an endless potential can be limited if the bank only looks at the quantitative information.

“So you could say that the function of gathering the information does not lies within the company, but within the financial institution.”

“I don’t think that it realistic to ask and to demand more of the smaller organization.”

Smaller organizations already provide additional information for the banks. This not always happens through an annual report, but is communicated to the account manager. Because these organizations do not have the capacity to always gather all the information regarding their risks, the responsibility of this task for the smaller organizations lies within the financial institution instead. A good account manager has multiple clients in the same sector and can because of that help an organization in his field of expertise.

“I have to understand this from this business plan and what the specific organization is doing and not because he has an additional risk report. I hope that the entrepreneur has something better to do, if he has

time to make a risk report I would be worried. I would not even find that desirable.”

As part of their sales strategy, banks (but also accountants) do a lot of market research in the various markets. This is also why the bank does not expect the smaller organization need to invest more so they can tell the bank what risks they are exposed to and to understand them. For the bank it not desirable that smaller organizations spend time reporting their risks. This would indicate that for smaller organizations it is more desirable for the bank that they invest their time into growth instead of reporting everything. The literature has shown that the smaller organization could benefit from additional voluntary disclosure with regards to their financing. But the additional disclosure has no effect, because the bank has certain strict requirements of the information that is needed to provide capital. It could only be possible when an

(28)

organization is highly innovative and especially has a high standard of operational safety. This would then be expressed in the type of client and the type of business they attract.

“Everyone wants to drive the safest car. But before you are able to buy a Mercedes-Benz, a couple of things have to happen. By the time you own the Mercedes-Benz, your other financial aspects will also be good. But if you come here with a 12 year old Ford Focus because you can’t afford more, but it works and is street worthy.

We then will gladly help you grow into to Mercedes-Benz level. By the time you have reached this, other investors will be interested. This gives you more financing possibilities and will also lower the price.”

Size and performance are the main influence on the finance capabilities for banks. A larger organization that has accomplished certain things has already proven itself and this increases their financing possibilities, but will also lower the price of the financing. The price will be reduced because multiple parties are willing to finance new operations and self-evident will positively influence the price of the newly attracted capital.

Furthermore the recent incidents at ChemiePack, Golf of Mexico and the problems at Odfjell have influenced the models used by the bank to determine the allocation of capital to a sector. Especially in the case of Odfjell, because it competes in the top of the sector and to be in that category you need to comply with all rules. In this category of the industry all the players are connected with each other and the compliance of these rules decides your “license to operate”.

“This is an industry crisis thing, that puts a huge question on the integrity level of their norms and value system in the technical sense that we have to monitor the risks.”

The organizations are checking this at each other. Odfjell is a storage company and a leading one at that, because of this Shell, BP etc. send their own auditors to them to verify if they are complying with all the norms and values of the industry. This results that the problems at Odfjell also strike back at the audit departments of the Shell’s, BP’s, Exxon’s, Akzo Nobel’s and so on, because they all had to caught these shortcomings at Odfjell.

“That's the problem with accountants. They have a huge story about all there was. Look if accountants have a crystal ball, we will be the first customers.”

There have been people who think that accountants need to look to possible future risks and provide their judgment over that information. The bank does not agree with this, even though he did state the accountants make long stories over past events. He used the current problems with the Hospitals in the Netherlands as an example. None of the hospitals will receive an approved annual report and this is due the forward-looking aspect of the health care. It is to complex and the changes are too radical that they could not stand in for it.

“It is unreasonable to ask the accountant to indemnify you from everything. You can expect the accountant to say, “no, that’s not correct what you’re doing there”, but that has not always happened in the past.”

(29)

The bank states that the accountant has to give a good audit opinion and report. Also should the organization, ceteris paribus, be able to continue for the next 12 months. He did not feel anything for an accountant that makes predictions for possible risks that could occur in 3 years and does not believe it would help the organization in any way.

4.2. Results statistical analysis

4.2.1. Descriptive statistics

4.2.1.1. Missing values

Prior to the randomizing of our population, we already had deleted organizations that were missing one or more values. This resulted in a dataset that has no missing values.

4.2.1.2. Outliers & Normal distribution

After we concluded that there are no missing values, we looked at our outliers. Outliers are values that deviate and will influence the results. Besides the outliers we also looked at the normal distribution of our variables. A normal distribution of the data is one of the assumptions that is needed for a regression. We were not able to remove our outliers or get a normal distribution of our data.

We have found a large variety of outliers in multiple variables. The amount of outliers in our dependent variables is limited to 7 unique sample units. Because we are taking the period between the years into consideration, we have decided to keep these outliers in our sample. To remove these lines would mean that we have to delete a total of 14 (7%) sample units.

Our independent variables are a different story. We were not able to remove the outliers and get a normal distribution for all the variables, because this would result in a significant decrease of sample size. The variable that showed a lot of outliers are IND (70 outliers), UAI (68), MAS (88), and liquidity ratio (40). We assume that the amounts of outliers for the cultural variables (mainly IND, UAI, and MAS) are so high as a result of the variety of countries in our sample. As a result of our sample selection, we had 56 organizations from the United States. This caused that more than half of the sample units show the exact same value for these variables. Removing these outliers would make all the cultural variables a constant and reduce our sample size with 88. This would also result in a data set existing out of only organizations from the United States and making the variables completely useless. Using a logarithm for these variables will not have a favorable effect. A logarithm for most of the cultural variables would result in a decrease of the normal distribution and, with the exception of PRA, have no effect on the outliers.

We have used a logarithm for the independent variable Total Assets; this has reduced the outliers and improved the normal distribution. The use of a logarithm for the other variables was of little affect. Ultimately we have decided to not remove any outliers, because removing the outliers would result in a data set of only US based firms and this would make all our cultural variables useless. Also our control variable Country and Legal System would not have any use.

4.2.1.3. Descriptive analysis

We will first get into the descriptive of our complete sample (table 1) and later get into more detail of our dependent variables. Our sample exists out of 200 sample units divided over two industries and two

Has the risk disclosure in the Oil & Gas, and Chemical industry changed during the financial crisis and is the change related to the national culture?