Master Thesis, master Accountancy The relationship between culture and the change in future risk reporting after an internal control failure

(1)

Master Thesis, master Accountancy

The relationship between culture and the change in future risk

reporting after an internal control failure

Name: Henri Kikkert Studentnumber: S2367017 Mail: h.kikkert@student.rug.nl Phone number: 0611506555

Address: Hereweg 86-2, Groningen Date: 14-06-2017

First supervisor: G.C. Helminck

Second supervisor: Prof. dr. J.A. Emanuels Assessor: dr. V.A. Porumb

(2)

2

Abstract

This research examined the influence of culture on change in risk reporting. In addition to this key relationship, we also investigated the moderation effect of industry on the relationship between uncertainty avoidance and change in risk reporting. In this study, 235 firm-year observations of failure firms and 65 firm-year observations of non-failure firms are collected. These firms are located in 25 different countries and the relevant SEC filings are from 2004 – 2012. The cultural influence is measured by using two indexes, Hofstede and Project GLOBE. No significant difference between the changes of the means of the risk reporting for the non-failure group and the non-failure group are shown in the study. This study indicated no relation between the cultural dimensions (individualism, power distance and uncertainty avoidance) and change in risk reporting as well. Either, there is no moderating effect of industry. The study is the first research in the current literature which examines the change in risk reporting. Therefore it is a contribution for the existing literature. Besides that, the results of this study can be used by several stakeholders in their decision-making process.

(3)

3

1.

Introduction

In recent decades, internal control failures have caused serious problems for organizations. These failures have led to different negative consequences for organizations like financial losses, reputational damage and sometimes even to bankruptcy or take overs. Enron (2001) and WorldCom (2002) were the first major accounting scandals (Gordon, 2002; Sidak, 2003). Among other things, these major accounting scandals have led to the introduction of the Sarbanes-Oxley Act (SOx) by the American government in 2002. This law is introduced to prevent future accounting scandals (Coates, 2007). American publicly traded companies are required to report about their internal controls and about the efficiency of their controls in their annual report. This is described in the SOx. Countries in Europe have introduced certain codes as well (Collier, 2005). But still firms such as Rabobank, Imtech, Volkswagen and Wells Fargo are examples of firms with internal control failures. Despite the several introduced codes, still prominent examples of organizations which internal control systems have failed, are existing in the last recent years.

Based on the stakeholder theory (Freeman, 1999), an organization has to operate according to the interests of the stakeholders. These interests can be contradictory. A balance between these conflicted interests of the various stakeholders of the firm should be found by the organization. Therefore it is important for organizations to report clearly about internal controls and risks, because stakeholders base their decision on information about internal controls and risks. When a firm has to deal with an internal control failure it is necessary for stakeholders to receive more information to be sure the firm has no other risks after this failure. Moreover, an internal control failure has negative consequences for a stakeholder (Hammersley et al., 2012; Gordon et al., 2012). Hammersley et al. (2012) show that organizations with consecutive internal control failures, have to deal with higher audit fees, a bigger chance of auditor resignation, an increase in credit ratings and an increase in the cost of debt. Moreover, Gordon et al. (2012) state that a consecutive internal control failure has a negative effect on the cost of equity of a firm. These consequences are negative for stakeholders as well. Therefore the expectation is that stakeholders would like to receive more information about internal controls and risks. With more information stakeholders can substantiated their decisions better. Therefore it is interesting to investigate if an organization reports more about internal controls and risks after an internal control failure. It is important for stakeholders to know if the internal controls are executed well and whether or not it is

(5)

5

likely another internal control failure will occur in the future. In this study will be investigated whether or not an internal control failure will result into changes in risk reporting.

In prior studies, there is written a lot about internal control problems, but there is still a gap in the existing knowledge about internal control problems. The existing research has focused more on the consequences and the causes of the internal control problems. Doyle et al. (2007), Hammersley et al. (2012) and Ashbaugh-Skaife et al. (2007) focused more on the causes and consequences of an internal control problem, based on data from reports of weaknesses in internal control over financial reporting. Some other scholars have focused more on another aspect of internal control problems. Dobler (2008) and Myllymäki (2014) focused in their researches more on the incentives for risk reporting and Hermanson (2000) focused more on the demand for risk reporting. However, one of the consequences of internal control problems is not discussed in the existing literature and this is the influence of internal control failures on future risk reporting (demand and) supply in (consecutive) annual reports. Therefore, in this study we will investigate if there is a relation between internal control failures and the change over time in risk reporting in consecutive annual reports to make a contribution on the existing literature. This study will identify some factors that influence the change in risk reporting after an internal control failure. Especially the relationship will be investigated between culture and the change in risk reporting.

In this research I will investigate the effect between culture and the change in future risk reporting after an internal control failure. Every country in the world has to deal with a different national culture (Hofstede, 2001). Because organizations are operating in different countries they have to deal with a different culture. Current literature shows difference in national cultures can influence differences in reporting about internal controls and internal control failures. Hooghiemstra et al. (2015) state in their research that a more individualistic firm reports generally more about internal controls. On the other hand, Hooghiemstra et al. (2015) have shown that a firm with a more uncertainty avoidance culture reports less about internal controls. Moreover, Kanagaretman et al. (2016) show that differences in national culture affects the reporting about Internal Control Material Weaknesses (ICMW). So, different cultures can influence the reporting behaviour of a firm. In this research we will also investigate the reporting behaviour of a firm. We will compare the reporting behaviour about risks before and after an internal control failure. The expectation is that national culture will

(6)

6

influence the change in future risk reporting after an internal control failure. The central question becomes: “What is the effect of national culture on the change in future risk reporting after an internal control failure?”

Additionally, in this study is investigated what the influence is of the industry wherein a firm is operating on the change in risk reporting. Organizations can operate in different industries and this could influence their reporting behaviour. For example, financial firms have to deal with a strong regulation. Therefore many guidelines are necessary to minimize risks. This could influence the change in risk reporting. Therefore this study investigates if the type of industry will influence the relationship between culture and change in risk reporting.

This study makes a theoretical contribution to the literature. First of all, this study is the first study, as far as I know, which investigates the relation between an internal control failure and the change in risk reporting. Moreover, it is certainly the first study which studies the relation between culture and change in risk reporting. Prior studies have focused more on which variables influences the amount and quality of risk reporting (e.g. Linsley et al., 2006; Dobler, 2008) and some studies show national culture influences voluntary disclosure (Hooghiemstra et al., 2015; Kanagaretman et al., 2016). But no study has investigated the relation between national culture and the change in risk reporting after an internal control failure. As a result, this study can be used as a starting point for future research in this research area.

In addition, the results of this study result into new and updated disclosure standards per country. The occasion of differences in change in risk reporting after an internal control failure between countries and cultures can be the difference in strictness of disclosure standards of reporting risks. By developing and updating disclosure standards and codes of corporate governance, the risk reporting can be improved. Therefore the interests of the stakeholders and investors can be protected.

Furthermore, failure firms (firms with an internal control failure) can use the results of this study. Failure firms with less change in risk reporting, can implement internal rules for themselves to report more about risks after an internal control failure. As a result, the firm can meet the stakeholder's demand for more information.

Last, stakeholders such as banks, auditors and investors, shall be interested in the results of this study. This study shows that certain cultures and certain countries change their risk reporting more than other cultures and countries. Stakeholders can use the results of this study in their decision-making process. Auditors can use this information to perform sharper

(7)

7

audits in certain countries which show a small change in risk reporting after an internal control failure. Thereby, capital providers will know in which country there are probably more risks than reported and they can use this information in their decision to provide capital to a firm or not. Besides that, investors can use the results of this study, because they know in which countries there may be more risks than reported and thereby the investor can decide better whether it is profitable to invest in a firm or not.

(8)

8

2. Literature review and background

In this chapter, the variables are described which are used in this research. First, the previous literature will be discussed of these variables and the expected relationships between these variables. Secondly, the stakeholder theory will be explained, which is used in this study. After that, the definition of culture is given and the models are explained which are used to measure culture. Then the independent variables individualism, power distance and uncertainty avoidance are explained. Last, the moderating variable industry will be described. 2.1 Literature review

In 2002, the American government introduced the SOx to prevent future major accounting scandals. Before the implementation of this law, Hermanson (2000) did research about the demand of risk reporting. Hermanson (2000) shows an increasement in risk reporting among the financial statement users. Despite this research, one scandal follows in a short time after another. As a reaction on the several scandals, the US government implemented the SOx (Coates 2007). Still after the implementation of the SOx, a lot of internal control failures have occurred. According to this, many more research is done to internal controls and internal control failures. The existing research has focused more on the consequences and the causes of the internal control problems. Doyle et al. (2007), Hammersley et al. (2012) and Ashbaugh-Skaife et al. (2007) focused more on the causes and consequences of an internal control problem, based on data from reports of weaknesses in internal control over financial reporting. Some other scholars have focused more on another aspect of internal control problems. Dobler (2008) and Myllymäki (2014) focused in their researches more on the incentives for risk reporting. However, one of the consequences of internal control problems is not discussed in the existing literature. This is the influence of internal control failures on future risk reporting (demand and) supply in (consecutive) annual reports.

In this study, is focused on organizations which have to deal with internal control failures. An organization with a material weakness is considered as an organization with an internal control failure. The Public Company Accounting Oversight Board (PCAOB) defines a material weakness as “A material weakness is a deficiency, or a combination of deficiencies, in internal control over financial reporting, such that there is a reasonable possibility that a material misstatement of the company's annual or interim financial statements will not be

(9)

9

prevented or detected on a timely basis” (PCAOB, Auditing Standard No. 51

). A good working internal control system prevents material weaknesses in a firm. Therefore the explanation is when a firm has a material weakness the internal controls of the firm have failed. In this study, organizations with a material weakness are investigated. Ashbaugh-Skaife et al. (2006) argue in their research that firms with internal control deficiencies (ICDs) face higher risks related to firms with strong internal controls. Ashbaugh-Skaife et al. (2006) define a firm with an ICD as a firm with an internal control problem. So, firms with an internal control problem face higher risks. This implicates these firms can report more about risks as well.

Furthermore, research is done about firms who reported material weaknesses in consecutive years. An organization with an internal control failure has a big problem, but an organization with internal control failures in consecutive years faces even bigger problems. Myllymäki (2014) states that a firm, who has faced a material weakness, has a greater likelihood on misstatements in the post material weakness period. Hammersley et al. (2012) and Gordon et al. (2012) state that organizations which report internal control failures in consecutive years, will face a lot of negative consequences such as higher audit fees, a bigger chance of auditor resignation, an increase in credit ratings, an increase in the cost of debt and a higher cost of equity. Based on this information, the expectation is that stakeholders would like to receive more information from a firm about risks and the effectiveness of their internal controls after an internal control failure. This to investigate whether or not a chance of another internal control failure is existing.

In the current literature, there is known that managers made a conscious trade-off between the costs and the benefits to disclose voluntary information (e.g. Bronson et al, 2006; Deumes et al, 2008; Hooghiemstra et al, 2015). The expectation is that after an internal control failure the manager made also such a trade-off for disclosing information. The costs for disclosing voluntary information for managers can be: reputation risks for the managers (Raghunandan et al, 1994), disclosing information to competitors and the costs for document and edit the information (Solomon et al., 1990). The benefits for disclosing voluntary information are the lower cost of capital. Based on the reported information, shareholders can make a better estimation which will reduce the risk of making a wrong decision (Deumes et al., 2008).

(10)

10

Advantage for a manger is, voluntary disclosure can result into a better reputation (Abraham et al., 2007; Kothari et al., 2009). In this study the hypnosis is, a manager will make a trade-off between the costs and the benefits of reporting about the risks of an organization after an internal control failure as well.

Dobler (2008) focused more on the incentives for managers for risk reporting in his research. Three reasons why a manager could not report risks of the organization where stated by Dobler. First, a manager has no risk information to report about. Second, a manager could not report, because the risk information is non-verifiable. Therefore the manager will choose to misreport or he cannot use the information. And last, a manager may not report about the risk information due to threats of commercial drawback, for example costs. These are the possible reasons why a manager could choose to report not about risk information. This can be an explanation why a manager will not change their reporting after an internal control failure. A manager could choose to hide risks about possible material weaknesses, but it is also possible that a manager has no information to report about.

In this research, the relation between culture and the change in risk reporting after an internal control failure will be investigated. There are a few studies which investigates the relationship between culture and voluntary disclosure. Hooghiemstra et al. (2015) have studied these relationship and they used in their study the framework from Hofstede (2001) to measure national culture. The results of the study of Hooghiemstra et al. (2015) suggest that a more individualistic firm report more about internal controls than firms who has a more risk-averse culture. Also Kanagaretnam et al. (2016) did a comparable research. They studied the effect of culture on the reporting behaviour of a firm about ICMW’s. They used the framework from Hofstede (2001) to measure national culture as well. Kanagaretnam et al. (2016) showed material weaknesses were earlier remediate by individualistic firms and firms with an uncertainty avoidance culture which had faced both material weaknesses. On the other hand, firms who are dealing with the power distance dimension of culture, have not the propensity to remediate the material weaknesses. So, some studies have shown that national culture influences voluntary disclosure (Hooghiemstra et al., 2015; Kanagaretman et al., 2016), but no one has investigated the relation between national culture and change in risk reporting after an internal control failure.

(11)

11

Currently, some research already investigates the effect of industry type on reporting. Petrovits et al. (2011) did a research about non-profit firms in relation with reporting about internal controls. They found that a weak internal control system will result into less donations and grants from the government. Based on these results, the expectation is non-profit firms will remediate the material weakness and increase their risk reporting. Moreover, Ge et al. (2005) used in their research different firms from different industries and distinguishes the firms per industry. A lot more material weaknesses were reported in computer and service industries in comparison to the food or chemical industry. This is one of the results the study of Ge et al. (2005) showed. Does the difference in industry results into a difference in risk reporting after an internal control failure as well? In this research we will investigate if the type of industry wherein a firm operates, influences the relationship between culture and risk reporting of the firm in this research. We will focus on financial and non-financial firms. Linsley and Shrives (2005) show non-financial firms have another way of reporting than non-financial firms. Financial firms have a more prescriptive way of reporting of risks. Financial firms, like banks, have to comply with the Basel regulation. Moreover, financial firms have already disclosed more about risks than non-financial firms. In this research, the variable industry (financial or non-financial) is used as moderating variable. 2.2 Stakeholder theory

To test the hypotheses in this study, we use the stakeholder theory. The stakeholder theory is described by Freeman (1984) as any individual or group who affects or is affected by the achievement of the firm’s objectives. According to the stakeholder theory, an organization has to operate according to the interests of the stakeholders. Important is that the organization finds a balance between the conflicting interests of the various stakeholders. In this study we will examine if there is a relationship between internal control failures and future risk reporting in consecutive annual reports. According to the stakeholder theory, the expectation is an organization will provide information to the stakeholders about their internal controls. Especially after previous internal control failures it is necessary for stakeholders to be sure, the internal controls are working well at this moment. They also like to know whether or not the organization has a chance to another internal control failure. This will have negative consequences for the organization and for the stakeholders as well (Hammersley et al., 2012; Gordon et al., 2012). The organization has to add more information in the annual reports about the internal controls than before the internal control failure. For a stakeholder it is necessary to have as much as possible information to make the best decision. When an

(12)

12

organization adds more information about the risks, the stakeholder or investor is better informed about the organization and can make a well-made decision.

2.3 Culture

In this research we will investigate the relationship between culture and the change in risk reporting after an internal control failure. Culture can be an influencing factor on the change in risk reporting. In this research we will investigate listed firms that are operating in the whole world. There are differences in different continents and different countries several cultural differences all over the world. Culture is earlier investigated, but to formulate a clear definition of culture is difficult (Jahoda 2012). Hofstede and Bond (1988) formulate the definition of culture as follows: ‘‘the collective programming of the mind that distinguishes the members 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’’. Licht et al. (2005) state that the essence of culture is value emphases. These are the shared ideas about what good is and right in the society. Licht et al. (2005) state that the prevailing practices and norms are important in the definition of culture as well. So the shared norms, believes and ideas, distinguishes a group with a certain culture from another culture.

It is hard to make culture measurable, but there are several researchers who made culture classifiable. Taras et al. (2009) compared several studies which developed methods to measure culture in their research. Examples of these studies are: Hofstede (1980), Inglehart (1997), Maznevski et al. (1995), and the GLOBE Project (House et al., 2004). Taras et al. (2009) showed many types of cultural dimensions and most of the dimensions are derived from the cultural dimensions from Hofstede (1980). Hofstede (1980) defines national culture as follows in his research: “The collective programming of the human mind that distinguishes the members of one human group from those of another. Culture, in this sense, is a system of collectively held values.” Hofstede (1980) developed a framework to measure national culture. He identifies four cultural dimensions: Individualism versus collectivism, power distance, uncertainty avoidance and masculinity versus femininity. In later research, he improved the framework and he added a cultural dimension to the framework: long-term orientation versus short-term orientation (Hofstede, 2001). Besides that, in 2010 the sixth cultural dimension, indulgence versus restraint, was added (Hofstede et al., 2010).

(13)

13

There are some critics on the framework of Hofstede (1980). Some suggest that there should be more cultural dimensions. In addition, some think it is an outdated framework. Triandis (2004) suggest in his research that there should be 18 dimensions. On the other hand, Hofstede (2011) refuted this opinion and he states that additional dimensions are only eloquent when they are both statistically and conceptually independent from the dimensions whose are already available. Some scholars have their doubts about the use of Hofstede’s framework (1980) in research as well. McSweeney (2002) state Hofstede ignores the fact that there may prevail different cultures within a country. Orij (2010) states the culture dimensions of Hofstede are still the same for 30 years.

Despite the critics, the framework of Hofstede is used in a lot of recent research (e.g. Han et al., 2010; Hooghiemstra et al., 2015; Kanagaretnam et al., 2016). In several studies with different research questions, the four cultural dimensions of Hofstede (1980) have been used meaningfully (Sondergaard, 1994). According to Soares et al. (2007), “Hofstede’s framework is the most widely used national cultural framework in psychology, sociology, marketing, or management studies”. Therefore, the framework of Hofstede is used too in this research.

Furthermore, the cultural framework of Project GLOBE developed by House et al. (2004) will be used in this research. House et al. (2004) have tried to improve the framework of Hofstede by using more cultural dimensions in their framework. Project GLOBE will be used as a second cultural dimension to measure culture. By using a second cultural framework the results of the tests and analysis will be more reliable.

In this research the cultural dimensions individualism, uncertainty avoidance and power distance are used. In corporate governance research, individualism and uncertainty avoidance are the most often studied cultural dimensions. Moreover, these cultural dimensions have the clearest consequences for the relation between the managers and the shareholders. These cultural dimensions influence the trade-off between the costs and the benefits to disclose information (Hooghiemstra et al., 2015; Hope, 2003). Besides that, we use the cultural dimension power distance, because power distance has a certain relationship with collectivism. Collectivism (opposed to individualism) and a large power distance are related in most societies (Hofstede, 2001). This research investigates whether this relationship applies to this research as well.

(14)

14

2.4 Individualism

First of all, Ashbaugh-Skaife et al. (2006) have argued in their research that firms with internal control deficiencies (ICDs) face higher risks relative to firms with strong internal controls. A firm with higher risks can report more about risks. According to the stakeholder theory, the expectation is that especially after an internal control failure, an organization will provide more information to the stakeholders about their risks. It is important for organizations to report clearly about internal controls and risks, because stakeholders base their decision on information about internal controls and risks. When a firm has to deal with an internal control failure it is necessary for stakeholders to receive more information to be sure the firm has no other risks after this failure.

In the cultural dimension individualism, individuals are only taking care of themselves and their nearest family-members. On the other hand, there is the opposite dimension, collectivism. Collectivism means individuals focus more on groups than on themselves (Hofstede, 2001). Individualism is more focusing on “I” and collectivism is more focusing on “we” (Hofstede, 2001). A collectivistic culture is strongly integrated. According to this culture, individuals belong to certain in-groups. These in-groups protect the interests of the members, but these groups expect also a certain loyalty (Hofstede, 1984). In an individualistic society, people make more decisions in their own interest and they are not interested very much in groups. They are focused more on personal goals and personal success (Hofstede, 2001). Heine et al. (2007), show individuals think they are more talented and competent in individualistic societies. Hence, in individualistic societies they are focused more on the personal achievement than on the interest of the groups. Therefore, we can expect that managers in an individualistic society act more in the way which is good for their own career or for their reputation (Hofstede, 2001). The expectation is that managers will make a trade-off between the costs and the benefits of disclosing information (Hooghiemstra, 2015). Based on that trade-off they will disclose voluntary information or not. Costs are for example the reputational damage for the manager when he discloses wrong information. After an internal control failure the reputation of the manager will not be at the highest point. Stakeholders such as investors will have little confidence in the manager, because the firm has faced an internal control failure. A manager needs to disclose correct information about risks which is complete and useful to improve his current reputation. Wrong information will result in a decline in reputation. According to the point of view from the manager, non-favourable risk information has to be disclosed as well. The stakeholders need to be informed about the risks after an internal control failure. Based on the above mentioned words, the expectation is that a

(15)

15

manager in an individualistic society reports more about risks to save among other things his reputation.

In collectivistic societies, there is much cooperation and consideration of others (Hofstede et al., 2001). Managers will work together and work close with other stakeholders as well. Therefore, managers do not need to disclose voluntary information about risks, because stakeholders know most of the risks inside the cooperation. Therefore no change in risk reporting for firms in collectivistic societies will be expected.

Based on the section above, the expectation is that a firm will change their risk reporting positively after an internal control failure in an individualistic society.

H1: Individualism is positively associated with the change in risk reporting in their annual reports after an internal control failure.

2.5 Power distance

The dimension power distance, explains in which extent the less powerful members of the society accept the power of the more powerful members (Hofstede et al., 2010). Powerful managers could take decisions in their own best interest, because they have more power. In more powerful societies decisions will take place more central (Hofstede et al., 2010). Moreover, powerful managers can influence financial reporting choices (Kanagaretnam et al., 2016). In a high power distance society, there is a certain hierarchical order between the manager and the other employees. Everyone has and knows their place in the hierarchical order (Hofstede et al., 2010). Subordinates do not have discussions with their superior and they accept the orders from their superiors (Hofstede et al., 2010). People try to equalise the distribution of power in a low power distance society, (Hofstede et al., 2010).

Gray (1988) shows secrecy has a strong connection with power distance. In line with this, a high power distance is negatively related with the amount of voluntary disclosure of an organization. Moreover, Orij (2010) shows secrecy is positively related with power distance. Orij states a manager wants to hold his power relative to the stakeholders. Therefore a manager will not disclose voluntary information. According to Orij (2010), a manager wants to retain an information advantage relative to the stakeholders. This implies a manager will not change their reporting after an internal control failure to keep this information advantage. Doupnik et al. (2006) argue a society is insecure about disclosure when they score high on secrecy. As a result, such a firm will disclose less information relative to firms which scores low on secrecy.

(16)

16

Moreover, in a high power distance society, managers could take decisions more in their own interests, because they have more power than the others (Hofstede et al., 2010). They could easily influence financial reporting choices (Kanagaretnam et al., 2016). Managers can even influence this more when the internal controls are weak. Managers want to retain an information advantage relative to stakeholders (Orij, 2010). The expectation is that this relation does not change after an internal control failure. According to the stakeholder theory, an organization has to operate according to the interests of the stakeholders. Therefore, it is important for organizations to report clearly about internal controls and risks, because stakeholders base their decision on information about internal controls and risks. When a firm has to deal with an internal control failure it is necessary for stakeholders to receive more information to be sure the firm has no other risks after this failure. In a high power distance society, managers want to hold the information advantage relative to stakeholders. Therefore, managers will even look better which information they will report. A manager will be even more careful with selecting the risk information he reports about. As a result, a manager shall report less about risks relative to the previous year.

Based on the section above, the expectation is that a firm will change their risk reporting negatively after an internal control failure in a high power distance society.

H2: Power distance is negatively associated with the change in risk reporting in their annual reports after an internal control failure.

2.6 Uncertainty avoidance

In uncertainty avoidance societies, individuals feel not comfortable with uncertainty (Hofstede et al., 2010). People avoid uncertain situations as well (Hofstede et al., 2010). These uncertain situations are unpredictable, unstructured or unclear (Hofstede et al., 2010). An organization will avoid these uncertain situations when an organization scores high on the uncertainty avoidance index of Hofstede (Hofstede et al., 2010). Organizations will be seen as safe seeking, aggressive and emotional as well (Hofstede et al., 2010). By implementing codes, a firm can prevent these uncertain situations. Organizations have a more relaxed attitude when an organization is scoring low on the uncertainty avoidance index (Hofstede et al., 2010). Moreover, these organizations are not emotional and accept anxious situations. (Hofstede et al., 2010).

Managers will avoid uncertain situations in a high uncertainty avoidance organization. Dobler states that a manager does not report about risks when the risk information is

(17)

non-17

verifiable or the manager could not use the information. This could be reasons in a high uncertainty avoidance organization as well. The manager will make a trade-off between the costs and the benefits. The manager values the costs relatively higher. Tversky et al. (1998) state people prefer to avoid losses rather than acquire benefits. In an uncertainty avoidance society the loss aversion trend will be stronger as well (Bryan et al., 2012). Moreover, managers in high uncertainty avoidance societies are less comfortable with conflicts and they find it hard to accept competition (Bryan et al., 2012). New competitors could enter the market when the organization reveals precious information. Therefore a manager could hide some information about risks. There is a possibility that the reported information turned out to be wrong (Graham et al., 2005, Dobler, 2008). For this reason, a manager will disclose possible less information about risks in a high uncertainty avoidance society. Certainly after an internal control failure a manager will not take a risk to disclose wrong information.

On the other hand, an organization has to operate according to the interests of the stakeholders according to the stakeholder theory. Therefore, it is important for organizations to report clearly about internal controls and risks, because stakeholders base their decision on information about internal controls and risks. When a firm has to deal with an internal control failure it is necessary for stakeholders to receive more information to be sure the firm has no other risks after this failure. Based on this reasoning, the expectation is that managers will report more about risks in a high uncertainty avoidance organization. But a manager finds it difficult to accept competition in uncertainty avoidance organizations. Moreover, a manager will prevent disclosing wrong information as well (Graham et al., 2005, Dobler, 2008). Thus, a manager does not want to report wrong information and information to his competitors. Therefore a manager does a favour to his stakeholders and reports less about risks,

Based on the section above, the expectation is that a firm will change their risk reporting negatively after an internal control failure in a high uncertainty avoidance society.

H3: Uncertainty avoidance is negatively associated with the change in risk reporting in their annual reports after an internal control failure.

2.7 Industry type

In this research, the moderating variable is the type of industry wherein a firm is operating. In this research a distinction is made between financial and non-financial firms. Linsley and Shrives (2005) state financial firms have another way of reporting than non-financial firms. Financial firms have a more prescriptive way of reporting about risks. Financial firms have to

(18)

18

comply with the Basel regulation (Linsley et al., 2005). This regulation expect from banks to employ a template with risk exposures. Banks achieve transparency in reporting as well. Moreover, financial firms have to disclose already more about risks than non-financial firms. Therefore the expectation is that a financial firm cannot change their risk reporting as much as the non-financial firm can do.

In this research, a financial firm will strengthen the relationship between uncertainty avoidance and change in risk reporting. In a high uncertainty avoidance organization, managers will avoid uncertainty. Therefore managers will avoid uncertain situations as well. The expectation is that a financial firm will show less change in risk reporting in an uncertainty avoidance society. Financial firms do relatively much on risk reporting and they have a lot of regulations to reduce risks. These regulations ensure that a financial firm have already reported a lot about risks. As earlier mentioned, stakeholders want to have more information about risks after an internal control failure. The risk reporting of financial firms must meet strict conditions. Therefore a company finds it difficult to report about risks. Moreover, there is a chance the information reported is not true or is not useful. Therefore, a manager can choose to do not report about the risks. As a result, the expectation is that a financial firm will strengthen the relationship between uncertainty avoidance and change in risk reporting.

H4: A firm that operates in the financial sector strengthens the negative relation between uncertainty avoidance and the change in risk reporting in annual reports after an internal control failure.

(19)

19

3. Research design

In this methodology section, the data which is used in this research are described. In addition, the dependent variable, independent variable, moderating variable and control variables will be explained. Next, the measurability of the variables is explained. Last, the regression models are displayed.

3.1 Data

This study is a quantitative study. In this study, 235 cases of companies with a material weakness are collected, to reach a valid data set. These firms are retrieved from the database Audit Analytics. The year of the material weakness is the event. The organization can be more than once in the sample when an organization has multiple material weaknesses in different years. It does not matter that a company in the sample more than once occurs, because we look in this study to the change in risk reporting after an internal control failure. An organization can change the risk reporting after each material weakness in several years. The firms are selected based on their business country region. The business country region is the country wherein the firm is operating and wherein the company in general is doing their business. The selected firm-year observations are after 2002, because SOx have been implemented in 2002. The firm-year observations in the sample are from 2004 till 2012. We have also selected a control sample. This sample contains 65 firm-year observations with no material weakness in a certain year. This sample is used to investigate if there is a difference in change in risk reporting between non-failure firms (firms without an internal control failure) and failure firms. An independent sample test is used to test if there is a significant difference between the means of the two groups. In this research, further analyses can be performed when there is a significant difference between the means of the two groups.

3.2 Variables

3.2.1 Change in risk reporting

In this research, the dependent variable must be made measurable. Therefore, a model has to be developed to compare the risk reporting before the internal control failure and after the failure. There are different approaches to analyse annual reports in order to measure how much a firm reports about risk. Linsley and Shrives (2006) used in their study a content analysis. They used number of sentences to measure risk reporting. Others have developed an index to measure the risk reporting of a firm (Hooghiemstra et al., 2015). In this research a content analysis is used as well. A content analysis can be performed in several ways. Number

(20)

20

of words, certain word combinations, and page proportions can be used. In this study the number of words will be count about risk reporting. These words are described in the relevant risk parts of the 10-K filing or 20-F filing of the organization. Every firm from Audit Analytics is obliged to prepare a report in accordance with certain guidelines from the SEC. The 10-K form is a comprehensive way of the financial condition and the performance of the organization for a certain annual period2. An organization reports in certain parts of this 10-K form about the risks of the organization. Risk information in 10-K filings include item “1A: risk factors” and item “7A: quantitative and qualitative disclosures about market risk”. This part includes information about market risks3 . In 20-F reports, comparable items can be found under item “3D: risk factors” and under item “11: quantitative and qualitative disclosures about market risk”. The numbers of words are count in these relevant parts. In the year of the material weakness (t=1) the words will be count about risk reporting and in the year before the material weakness (t=0) the words will be count. These two years will be compared with each other. The change in risk reporting is measured as follows:

((amount of words about risks t=1 – amount of words about risks t=0) / amount of words t=0).

The change in risk reporting is thus a relative change in risk reporting. When a firm has reported more (or less) words about risks after the internal control failure than before, the firm has changed their risk reporting.

3.2.2 Culture

Culture is measured by using the cultural dimensions of Hofstede and Project GLOBE. As earlier mentioned, the dimensions power distance, individualism and uncertainty avoidance are used as dependent variables. Hofstede has scored the different countries on these cultural dimensions. These scores are ranged from 0 to 112. In table 1 the scores of the different culture dimensions per country are described for both the dimensions of Hofstede and Project GLOBE. The power distance is very low in a country when the score on power distance is 0. On the other hand, the power distance is very high in a country when the score on power distance is 112. In this research, firms from 25 different countries are used. Panama and Chile have no score on the GLOBE index and are removed from the dataset.

2_{https://www.sec.gov/fast-answers/answers-form10khtm.html} 3

(21)

21

Table 1 Culture dimensions scores per country

Hofstede index GLOBE index

PDI IDV UAI PDI IDV UAI

1 AUSTRALIA 38 90 51 4,74 4,23 4,39 2 BRAZIL 69 38 76 5,33 4,50 3,60 3 CANADA 39 80 48 4,82 4,32 4,58 4 CHINA 80 20 30 5,04 5,29 4,94 5 COLOMBIA 67 13 80 5,56 4,77 3,57 6 FRANCE 68 71 86 5,28 4,15 4,43 7 GREECE 60 35 112 5,40 4,26 3,39 8 HONG KONG 68 25 29 4,96 4,72 4,32 9 IRELAND 28 70 35 5,15 4,89 4,30 10 ISRAEL 13 54 81 4,73 4,58 4,01 11 KOREA (SOUTH) 60 18 85 5,61 5,37 3,55 12 MALAYSIA 104 26 36 5,17 5,06 4,78 13 MEXICO 81 30 82 5,22 4,88 4,18 14 NEW ZEALAND 22 79 49 5,44 5,51 3,89 15 PHILIPPINES 94 32 44 5,52 5,06 2,88 16 RUSSIA 93 39 95 4,99 5,27 5,31 17 SINGAPORE 74 20 8 5,33 4,78 3,78 18 SOUTH AFRICA 49 65 49 5,16 4,56 4,09 19 SPAIN 57 51 86 5,52 4,65 3,97 20 SWEDEN 31 71 29 4,85 4,44 5,32 21 SWITZERLAND 34 68 58 4,90 4,02 5,37 22 TAIWAN (CHINA) 58 17 69 5,18 5,09 4,34 23 TURKEY 66 37 85 5,57 4,96 3,63 24 UNITED KINGDOM 35 89 35 5,15 4,18 4,65 25 USA 40 91 46 4,88 4,22 4,15 3.2.3 Industry

To measure industry there is made a distinction between financial firms and non-financial firms. The financial sector includes banks, pension funds, investment companies, insurance companies and intermediaries. Subsequently a score of 1 is assigned if the organization is categorized in the financial sector. A score of 0 is assigned if the organization is categorized in a non-financial sector.

3.2.4 Control variables

In this research some control variables are used. First, we control for firm size, because firm size matters when examining voluntary disclosure about risks. Firm size is measured as the natural logarithm of the total assets of the firm. Second, we control for BIG4 auditor. A BIG4 auditor will influence the change in risk reporting. A firm audited by a BIG4 auditor gets a score of 1 and a firm audited by a non-BIG4 auditor gets a 0. This makes BIG 4 auditor as a dummy variable. Last, we control for block holders. The expectation is that block holders will influence the reporting after a material weakness.

(22)

22

3.3 Regression model

The research is conducted by using a multiple regression analysis. In the research model, βi is the coefficient and εi is the random error term. In table 2 the definitions of the variables and the abbreviation of the variables are outlined.

CHANGE_RR = β0 + β1*IDV + β2* PDI + β3* ZUAI + β4 * ZUAI * ZFIN_IND + β5 * AUD + β6 * BLOCK + β7 * TOT_ASS + εi

β4 * ZUAI * ZFIN_IND reflects the moderating effect of industry on the relationship between uncertainty avoidance and change in risk reporting. The regression model will be used twice, both for the cultural scores of Hofstede and for the cultural scores of Project GLOBE.

Table 2 Variable definitions

Variables Description

Dependent variable

CHANGE_RR Difference in risk reporting between t=0 and t=1. t=1 is the year of the material weakness. The change in risk reporting is measured as follows: Change in risk

reporting = ((amount of words about risks t=1 – amount of words about risks t=0) / amount of words t=0). The change in risk reporting is a percentage.

Independent variables

IDV Individualism score of a country based on Hofstede (2001). A higher (lower) score reflects a culture that is relatively more individualistic (collectivistic). Individualism was measured in this study for Project GLOBE on the basis of a collectivist score. The score was measured as follows: (collectivism I Societal

Practices (Institutional Collectivism) + Collectivism II Societal Practices (In-group Collectivism)) / 2

ZUAI Uncertainty avoidance score of a country based on Hofstede (2001) and Project GLOBE. A higher score reflects a culture that is relatively more uncertainty avoidant. This is the standardized score.

PDI Power distance score of a country based on Hofstede (2001) and Project GLOBE. A higher score reflects a culture that has relatively more power distance.

Moderating variable

ZFIN_IND Financial or non-financial industry. The financial industry includes banks, pension funds, investment companies, insurance companies and intermediaries. Subsequently a score of 1 is assigned if the organization is categorized in the financial sector. 0 is assigned if the organization is categorized as a non-financial sector. This is the standardized score.

Control variables

AUD Auditor of the firm. BIG4 auditor in the annual report after the failure gets a score of 1. Firm with no BIG4 auditor gets a score of 0.

BLOCK Identify whether outside block holders are present, being an outsider owning 5% or more.

TOT_ASS The amount of total assets as reported on the balance sheet at the end of the year in which the material weakness was identified in US dollars ($). In the regression analysis and correlation analysis the natural logarithm of the total assets of the firm is used.

(23)

23

4. Results

First, in this chapter the outcome of the comparison between the failure firms and the non-failure firms are described. Then the descriptive statistics are discussed. Thereafter the results of the correlation test and the regression analysis are outlined. Last, the effects of culture and industry on the change in risk reporting are discussed.

4.1 Control sample

First of all, in this study the difference in change in risk reporting between non-failure firms and failure firms is investigated. The sample contains 64 non-failure firms and 225 failure firms. Four firms had in one of the years (almost) no risk reporting. Therefore these four failure firms are removed from the dataset. These would have influenced the results. Furthermore, seven firm-year observations are removed. These firm-year observations had no score on the GLOBE index or on the Hofstede index. Those are firms from Puerto Rico, Bermuda, Panama and Chile. In addition, the winsorizing-technique is used to replace the value of the outliers in order to ensure that the outliers do not have an extreme impact on the results. The values which are more or less than average +/- 3*standard deviation for this value are replaced. At the end, there are in the sample 225 failure firms and 64 non-failure firms left.

In table 3 the results of the independent sample test are outlined. There is a difference between the means of the two groups: failure firms have on average 16.55% change in risk reporting and non-failure firms have on average 10.96% change in risk reporting. As a result, there is a difference between the means of the two groups, but this is not a significant (p=0.204) difference (equal variances not assumed). This implies that firms change not their risk reporting after an internal control failure.

Table 3

Independent sample test (CHANGE_RR)

Number of firms Mean St. deviation

Failure firms 225 ,1655 ,4853

Non-failure firms 64 ,1096 ,2370

t df Sig. (2-tailed)

Equal variances assumed ,890 287 ,374

Equal variances not assumed 1,273 216,356 ,204

4.2 Additional theory and analysis

As seen in section 4.1 and in table 5, there is no significant difference between the means of the failure group and the non-failure group. Therefore in this research an additional specific analysis have to be made. After all, there is no significant difference between the means.

(24)

24

However, for both groups a change can be found in the risk reporting. Therefore, in the further research we take the two groups together and investigate if culture explains the change in risk reporting. In the coming sections, there will be no more distinction between the two groups.

4.3 Descriptive statistics

In this part of the thesis the descriptive statistics of the dataset are explained.

Table 6 Descriptive statistics Dependent variable (change_RR) per country

Country Number of firm year

observations

Mean SD Minimum Maximum

AUSTRALIA 5 .122 .194 .016 .466 BRAZIL 3 -.049 .051 -.105 -.005 CANADA 12 .036 .101 -.137 .250 CHINA 43 .181 .507 -.774 1.755 COLOMBIA 2 .088 .153 -.020 .196 FRANCE 5 .159 .328 -.157 .599 GREECE 2 .357 .449 .040 .675 HONG KONG 30 .215 .696 -.997 2.603 IRELAND 2 -.069 .182 -.198 .060 ISRAEL 17 .057 .116 -.139 .315 KOREA (SOUTH) 1 .196 . .196 .196 MALAYSIA 1 -.003 . -.003 -.003 MEXICO 10 .267 .279 -.025 .815 NEW ZEALAND 1 .152 . .152 .152 PHILIPPINES 1 -.078 . -.078 -.078 RUSSIA 4 -.112 .233 -.427 .096 SINGAPORE 4 .062 .287 -.132 .483 SOUTH AFRICA 1 -.129 . -.129 -.129 SPAIN 1 -.049 . -.049 -.049 SWEDEN 1 -.094 . -.094 -.094 SWITZERLAND 5 .161 .229 -.047 .524 TAIWAN (CHINA) 4 .014 .068 -.072 .083 TURKEY 2 .015 .016 .003 .026 UNITED KINGDOM 10 .112 .218 -.081 .669 USA 122 .182 .472 -.415 2.603 Total 289 .153 .443 -.997 2.603 Independent variables

Mean SD Minimum Maximum

CHANGE_RR .153 .443 -.997 2.603 PDI (Hofstede) 51.47 20.382 13 104 IDV (Hofstede) 62.27 31.057 13 91 UAI (Hofstede) 47.71 18.704 8 112 PDI (GLOBE) 4.970 .175 4.726 5.607 IDV (GLOBE) 4.545 .417 4.016 5.507 UAI (GLOBE) 4.331 .415 2.883 5.371 FIN_IND .120 .323 0 1 TOT_ASS 4,450,438,249 29,748,246,478 0 339,747,804,894 AUD .480 .501 0 1 BLOCK .248 .251 0 1

(25)

25

In table 6 the descriptive statistics of the dataset are outlined. This dataset contains 289 firm-year observations. These firms are operating in 25 different countries. The mean of the change in risk reporting from the total dataset is 15.3%. On average, firms report 15.3% more words about risks in the relevant SEC filing of the material weakness than the report before the weakness. The standard deviations of the cultural dimensions of Hofstede are relatively high, because there are a lot of firms in the dataset which score relatively high or low on this index. 48% of the firms have a BIG4 accountant and the mean of the total assets is $4,450,438,249. Besides that, firms have a block holders’ average of 24.8%, which means that a firm has on average 24.8% block holders. Last, 12% of the firms of the dataset are a financial firm.

4.4 Results hypotheses

In this part of the results section the results of the research are described. For testing the hypotheses the correlation test and the regression analysis have been used. These tests are described in the sections below as well.

4.4.1 Correlation test

In this section the results of the correlation test are described. The correlation matrix is presented in table 7 and table 8. Table 7 shows the correlations regarding the culture variables of Hofstede. Table 8 shows the correlations regarding the culture variable Project GLOBE. These tables show the correlation between the dependent and independent variable. A correlation matrix causes the problems that appear with multicollinearity. Furthermore, the correlation matrix reflects the (significant) influences of the variables. Multicollinearity within the correlation matrix is likely to be considered when variables have a higher value than 0.7 or less than -0.7. In this study, there is one possible case which seems to give rise to multicollinearity. This applies to the relationship between individualism and power distance (r=-0.773). An additional test is done to see if multicollinearity really exists in this study. A second way to test multicollinearity in a model is by calculating a VIF value. A VIF value greater than 10 seems to give rise to multicollinearity. In this study there is an average VIF value about 1.5 and every VIF value is below 3. This indicates that there is no multicollinearity in this model.

Also, these correlation matrices (table 7 and 8) show that there is no significant correlation between the independent variables power distance, individualism and uncertainty avoidance and the dependent variable change in risk reporting. Furthermore, the matrices show that there is also no significant correlation between the control variables and change in risk reporting.

(26)

26 Table 7

Correlation matrix (Hofstede)

1 2 3 4 5 6 7 CHANGE_RR 1 PDI .029 1 IDV .002 -.773** 1 UAI -.050 -.236** .098 1 TOT_ASS .009 -.129* .221** .181** 1 AUD -.060 -.220** .150* .207** .375** 1 BLOCK -.027 .012 -.071 .161** .201** .262** 1 ** p< .01 * p< .05 Table 8

Correlation matrix (GLOBE)

1 2 3 4 5 6 7 CHANGE_RR 1 PDI -.032 1 IDV -.016 .443** 1 UAI .027 -.140* .390** 1 TOT_ASS .009 -.012 -.197** -.297** 1 AUD -.060 -.021 -.163** -.295** .375** 1 BLOCK -.027 .121* .072 -.173** .201** .262** 1 ** p< .01 * p< .05 4.4.2 Regression analysis

In this section the results of the regression analysis are discussed. Table 9 and table 10 show the results of the regression analyses. In table 9 the results are treated with regard to the Hofstede variables and in table 10 the results are treated with regard to the GLOBE project variables. In both analyses, seven models are used to test the hypothesis. First of all, the controlling variables were tested in model 1. Then the independent variables and the moderating variable were tested separately in model 2, 3, 4 and 5. Furthermore, the moderating effect of industry on the relation between uncertainty avoidance and change in risk reporting was tested in model 6 and last the complete model was tested in model 7.

4.4.2.1 Control variables

Table 9 and table 10 show that in model 1 no control variable has a significant influence on the change in risk reporting. Model 1 shows that a BIG4 auditor (AUD) and block holders (BLOCK) have a negative relationship with the change in risk reporting (CHANGE_RR), but this relationship is not significant (p=.284 and p=.790). The last control variable in this study

(27)

27

is firm size (TOT_ASS). Model 1 shows that firm size have no significant impact on the change in risk reporting (p=.544) as well. Model 1, with all control variables, does not appear to be significant at all with a R² of 0,005 and an F value of 0,476. The low value of the R2 means that this model has a low declaration power.

4.4.2.2 The effect of national culture on the change in risk reporting

The effect of national culture on the change in risk reporting is outlined in model 2, 3 and 4. In model 1 hypothesis 1 is tested, in model 3 hypothesis 2 is tested and in model 4 hypothesis 3 is tested. For both power distance (PDI), individualism (IDV) and uncertainty avoidance (ZUAI), no significant relation is found with the change in risk reporting. This applies to both the Hofstede variables and the Project GLOBE variables. In all models the R2 has a value below .012 and this means that the declaration power of the model is low.

As shown in both models 3 there is no significant positive relationship between individualism and change in risk reporting, B = .000, p = .965 and B = -.022, p = .742. This means that an increase in individualism does not result into an increase change in risk reporting. Project GLOBE shows a negative relation between individualism and change in risk reporting. This is in line with the expectations in the corresponding hypotheses. Individualism was measured in this study for Project GLOBE on the basis of a collectivist score, the expectation was therefore opposite compared to Hofstede's model. Still the relation is not significant. As a result, hypothesis 1 can be rejected.

Also, in both models 2 is shown that there is no significant negative relationship between power distance and change in risk reporting, B = .000, p = .742 and B = -.079, p = .602. Project GLOBE shows, as expected, a negative relation between power distance and change in risk reporting, but this relation is still not significant. As a result, hypothesis 2 can be rejected. There is no relation between collectivism and power distance as well.

Furthermore, in both models 4 is shown that there is no significant negative relationship between uncertainty avoidance and change in risk reporting, B = -.019, p = .491 and B = .019, p = .785. The Hofstede variable shows, as expected, a negative relation between uncertainty avoidance and change in risk reporting. This relation is still not significant. Project GLOBE shows just a positive relation, but this relation is also not significant. As a result, hypothesis 3 can be rejected.

(28)

28

4.4.2.3 The moderating effect of industry

The moderating effect of industry on the relation between uncertainty avoidance and change in risk reporting is tested in model 5 and 6. The regression analysis of model 5 and model 6 has made use of standardized explanatory variables, the Z-score. The Z-score of industry (ZFIN_IND) and the Z-score of uncertainty avoidance (ZUAI) were calculated. Also the in the interaction-variable is calculated by multiplying the Z-score of industry with the Z-score of uncertainty avoidance.

In model 6 is included the standardized value of uncertainty avoidance and the interaction effect of uncertainty avoidance and industry. Hence, hypothesis 4 can be judged. When we look to model 6, there was no significant interaction between industry and uncertainty avoidance on change in risk reporting (B = .021, p = .411 and B = .004, p = .857). This means that industry has no significant stronger effect on the relation between uncertainty avoidance and change in risk reporting.

(29)

29

Table 9 (Independent variables Hofstede) Regression analysis

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

B Sig. Β Sig. Β Sig. B Sig. B Sig. B Sig. B Sig.

INTERCEPT .078 .663 .050 .789 .077 .179 .060 .739 .076 .669 .052 .774 -.006 .981 TOT_ASS .006 .544 .006 .532 .006 .558 .007 .501 .006 .550 .007 .486 .006 .571 AUD -.062 .284 -.058 .325 -.062 .285 -.057 .330 -.056 .339 -.058 .332 -.046 .440 BLOCK -.029 .790 .050 .798 .077 .179 -.021 .846 -.030 .786 -.003 .978 -.021 .853 PDI .000 .742 .001 .675 IDV .000 .965 .000 .731 ZUAI -.019 .491 -.020 .470 -.016 .570 ZFIN_IND .026 .324 .027 .312 .026 .324 ZUAI*ZFIN_IND .021 .411 R2 .005 .005 .005 .007 .008 .012 .011 F .476 .383 .356 .475 .601 .593 .435 # 289 289 289 289 289 289 289

Table 10 (Independent variables GLOBE) Regression analysis

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

B Sig. B Sig. B Sig. B Sig. B Sig. B Sig. B Sig.

INTERCEPT .078 .663 .470 .543 .185 .618 .065 .724 .076 .669 .066 .721 .403 .614 TOT_ASS .006 .544 .006 .552 .005 .587 .007 .517 .006 .550 .006 .527 .006 .553 AUD -.062 .284 -.063 .274 -.064 .270 -.059 .317 -.056 .339 -.053 .370 -.055 .352 BLOCK -.029 .790 -.021 .846 -.024 .829 -.027 .809 -.030 .786 -.032 .777 -.019 .863 PDI -.079 .602 -.056 .756 IDV -.022 .741 -.011 .892 ZUAI .019 .785 .006 .842 .007 .838 ZFIN_IND .026 .324 .025 .350 .025 .349 ZUAI*ZFIN_IND .004 .857 R2 .005 .006 .005 .005 .008 .009 .009 F .476 .424 .383 .374 .601 .412 .380 # 289 289 289 289 289 289 289

(30)

30

5. Discussion and conclusion

This section starts with a few paragraphs wherein the findings are discussed and are summarized. After that, a few theoretical and managerial implications are present based on the findings. Last, the most important limitations were addressed and in addition some recommendations were given for further research.

5.1 Findings

The first finding of this research is that there is no significant difference in the means for change in risk reporting between the non-failure group and the failure group. This means that a firm with a failure will not change their risk reporting more than a non-failure firm. This is an interesting result, because a positive change in risk reporting after an internal control failure was expected. After that, no explanation could be made in the difference in change reporting between the failure group and the non-failure group with the cultural variables. In both groups there was a change in risk reporting. Therefore, the change in risk reporting with the cultural variables was tried to explain in this research.

In this research, there was argued that a more individualistic firm had a positive relationship with change in risk reporting. But the results show that there was no relationship between individualism and change in risk reporting. Hooghiemstra et al. (2015) state in their research that more individualistic firms, report generally more about internal controls. The expectation is that individualistic firms disclose more voluntary information and that can mean that they already report a lot about risks. Therefore, it could be that an individualistic firm will not change their reporting. A firm will not report about risks which are not relevant.

Furthermore, there was argued that a firm with a more power distance culture will change their risk reporting negatively. The results show that there was no relationship between power distance and risk reporting. Zarzeski (1996) states already that power distance is negatively related with the disclosure of information. That means that firms in a more power distance culture reports less about information. Therefore it could be that firms already disclose less about risk reporting and that there is almost no negative change possible in risk reporting. This may also apply to hypothesis 3 which expects that uncertainty avoidance has a negative effect on change in risk reporting. It could be difficult to get a negative relation between uncertainty avoidance and change in risk reporting when a firm already disclose less voluntary information.

Master Thesis, master Accountancy The relationship between culture and the change in future risk reporting after an internal control failure