Master Thesis IBM:

1

Master Thesis IBM:

The influence of Top Management Team

Composition on Illegal Behaviour

The consequences of TMT diversity on Corporate Illegal Behaviour.

Valentijn Gerben Burema S2915359

2

Acknowledgements

We would like to thank the university for allowing us to write our master thesis, which marks the end of my academic career in IBM as well as the close of an important chapter of my life. I would also like to acknowledge the contribution of my supervisor professor R.W. de Vries. His suggestions were of great significance to my regression decisions. Next, we would like to recognize Daan Sauren and James Houlahan for their help with time management and the use of the English language, respectively. Finally, I am extremely grateful to my parents and brother for their unwavering support throughout the process of completing my master thesis. They were invaluable sparring partners and provided me with all the support I required. My parents had a huge influence on my ability to submit a completed master thesis before the deadline.

Abstract

In this thesis, the relationship between top management team composition and illegal behaviour is studied. In the introduction, it is mentioned why it is so important to study the relationship between illegal behaviour and top management team composition diversity. In the chapter on prior literature and hypothesis development, the concepts of top management team composition and illegal behaviour are further introduced and elaborated. The last section of this chapter includes a conceptual model of the relationship between top management team composition diversity and illegal behaviour. The data chapter explains the origin of the dataset, our selection criteria, and the variables used in the hypothesis testing. In the results sections, we use logistic regression to find out whether a relationship between TMT composition diversity and illegal behaviour exists. In the discussion chapter, we mention our results, the findings of others and a possible justification for our findings. In the last chapter, the conclusion, we discuss the theoretical and managerial implications of our findings. A possible explanation for finding no significant evidence is also discussed with reference to the limitations of this thesis. In the last section of the conclusion, we provide multiple suggestions for future research.

3

Table of Contents

2.1.4 Upper Echelon Theory p.8

2.1.5 Shareholder p.8

2.1.6 Fraud Triangle p.9

2.2 Top Management Team p.9

2.2.1 Gender Diversity p.9 2.2.2 Age Diversity p.11 2.2.3 Nationality Diversity p.12 2.2.4 Tenure Diversity p.13 2.3 Conceptual Model p.14 3. Methodology p.15 3.1Variables p.15 3.1.1 Dependent Variables p.15 3.1.2 Independent Variables p.15 3.1.3 Control Variables p.17 3.2 Research Design p.18 3.3 Data Sources p.19

3.3.1 US SEC & DoJ p.19

3.3.2 BvD. Orbis p.19

3.3.3 Time Period & Distribution p.20

3.4 Data p.20

3.4.1 BvD Orbis Research Strategy p.20

3.4.2 Top Management Team Diversity p.22

4

1.

Introduction

Why is it important to study illegal behaviour? Because the impact on society is huge. In recent times, multinationals have committed illegal behaviour to improve their annual reports and increase the value of the enterprise. After this news clients, investors and other stakeholders lost their faith in the multinational enterprises (MNE). This shows that illegal behaviour is risky, as firms received fines and their reputation was damaged. The role of the top management teams, such as the CEO and directors is significant (CNBC, 2017), because in some firms top management knew of illegal acts or allowed the MNE to commit them. Therefore, members of the top management teams are personally seen as the offenders that enabled or committed the crime. The question remains what factors drove the top management team’s decisions to commit fraud, corruption or bribery. Moreover, what made the managers responsible for the scandal? Possible explanations for their behaviour exist; for example, the pressure to perform by shareholders. This begs the question which characteristics of a top management team changes their behaviour and allow them to collectively engage in illegal behaviour.

One of the most recent cases of bribery and corruption comes from Airbus (Euronews, 2020), in a case that took 8 years to investigate and prosecute. Airbus has since settled with the United Kingdom, the United States and France for a $3.9 billion fine for corruption. This is one of the largest fines handed out by governments. Airbus paid officials and other stakeholders to win contracts for a prolonged period of time. As a result, their bribery was classified as corruption under the USA Foreign Corrupt Practices Act (FCPA). Airbus also conducted fraud to hide their activities (Euronews, 2020). Top management played a role in the scandal, as they hired relatives of clients and government officials to get business deals. It is remarkable that the top management team of Airbus is homogeneous. Regarding gender and age, the top managers are all male and similar in age, but with different nationalities. This distribution of managers shows that there is very little diversity in this team. The only diversity is that 50% of the top managers have a different nationality. Other cases like the one mentioned above show similar distributions in diversity. So what is the influence of a homogeneous team on illegal behaviour? Is a top management team (TMT) more likely to engage in illegal behaviour if diversity is minimal?

5

and other fraudulent activities without their clients’ permission, for which the US Consumer Financial Protection Bureau fined the bank a total of $185 million (Corkery, 2016). However, after all the lawsuits and fines, the total cost for Wells Fargo is estimated to be around $2.7 billion. This shows the damage that illegal behaviour can do to a business and society. In his article, Matt Egan (2016) shows the attitude of Wells Fargo’s top managers. Wells Fargo had a diverse top management team with women comprising 31% in 2011. According to Nelson (2012), adding women to top management teams will resolve risky behaviour, such as illegal activities. How does this work in the case of Wells Fargo? Can we assume that by further increasing gender diversity the likelihood of illegal behaviour would be reduced? Or do other diversity factors play a role? The top management team of Wells Fargo was highly diverse in gender, but similar in age and nationality. So how do we prevent top managers from making the same choices as the managers of Wells Fargo and Airbus? Maybe other forms of diversity, such as age, tenure and nationality, could influence the appearance of illegal behaviour?

Can diversity play a role in changing the behaviour of top managers? According to Nelson (2012), gender diversity can influence their behaviour. Similar findings by Zhu (2013) and Milliken (2019) support the idea that diversity in top management team composition could reduce illegal behaviour (Jianakoplos & Bernasek, 1998). The influence of gender is shown by Swamy et al. (2001). They studied the relationship between gender and corruption and found that males have a higher chance of engaging in corruption when compared to women. Diversity in top management teams is not just gender diversity; it should also incorporate age, nationality and tenure diversity. Is a TMT truly diverse if only gender diversity is considered, or should we incorporate more factors?

6

The examples of illegal behaviour and the author’s findings on diversity mentioned above, show that each form of diversity could have an impact on the likelihood of illegal behaviour. Articles on top management team characteristics and illegal behaviour exist (Daboub, 1995; Williams, 2005). Unfortunately, neither author has studied a broader concept of top management team diversity: their articles focus on a single form of top management team diversity and its influence. The articles place importance on certain characteristics, such as military service and education, with respect to the likelihood of illegal activity. These characteristics are not forms of diversity. Simons et al. (1999) describe top management team diversity as diversity in the age, previous experience, education and tenure of its members. This shows that diversity is a broad concept. However, the question remains what the overall influence is of all forms of diversity on top management teams’ decisions that lead to illegal behaviour. Simons et al. (1999) also mention the importance of background on the choices of a top manager. Combined with the article by Nielsen & Nielsen (2012), their findings show that the background can play an important role in the influence top management team diversity has on illegal behaviour. The lack of knowledge with regard to diversity and its classifications leads to the following research question:

7

2. Literature Review

In this chapter, we explain what is seen as illegal behaviour. Additionally, we use the Friedman doctrine, alongside other theories, to identify possible causes of a multinational enterprise’s decision to engage in illegal behaviour. Each independent variable is explained separately and supported by the literature. In a separate explanation, the hypothesis is developed.

2.1 Illegal Behaviour

In 1980, Grasmick & Green wrote a criminology article on illegal behaviour which defined it as a series of unlawful acts that deprive others of their possessions or hurt them. Grasmick & Green’s (1980) article is a landmark for criminology. They argued that a person’s own perception of benefits and risks influenced criminal behaviour, rather than society’s perception. In this thesis, illegal behaviour is considered in terms of economic and financial behaviour. We identify and use three forms of illegal behaviour, namely fraud, corruption and bribery. Each of the forms of illegal behaviour and the definition of the top management team is explained in the following sections.

2.1.1 Fraud

In law, the definition of fraud is an intentional deception of others that will secure or allow for an unlawful or unfair gain by depriving others of their rights or possessions (Cloninger & Waller, 2000; Jizi, Nehme, & Elhout, 2016). Multiple types of fraud exist, two examples being accounting fraud and auditing fraud. In accounting and auditing fraud, a person registers incorrect or non-existent accounts, bills or services into the register of the firm. This could be done to show higher revenue, increase profits, or influence economic ratios (ROE).

2.1.2 Corruption

8 2.1.3 Bribery

Bribery is the act of offering bribes to influence or change decisions, or to quicken a decision-making process (Harstad & Svensson, 2011). In another definition by Zhou & Peng (2012), bribery is defined as cases of illegal behaviour that simplify business and the difficulty of receiving permits.

2.1.4 Upper Echelon Theory

One of the most common theories used to study the influence of a top management team is the Upper Echelon Theory of Hambrick and Mason (Hambrick & Mason, 1984). We use this theory because it explains the role and importance of top management teams in a firm. With this theory, we can study the influence of a TMT on illegal behaviour. Hambrick & Mason’s (1984) theory explains that the composition of a top management team has a large influence and helps determine the TMT performance and decisions. The influence of a TMT comes from the belief that decisions made by the boss or higher-level management have to be accepted and executed by lower-level employees (Raes et al., 2011). The underlying assumption in this theory is that a TMT decides or has the power to influence or change the path a firm takes and that the rest of the employees will follow their decisions (Iaquinto & Fredrickson, 1997). The reason the authors give for the influence of a TMT composition on firm choices is that managers make decisions based on their values and beliefs.

2.1.5 Shareholder Theory

9

Frooman (1997) discovered that illegal behaviour increases short-term wealth. Managers can gain wealth quickly by engaging in illegal behaviour. This wealth can relieve pressure faced internally due to the threat of bankruptcy, or externally by shareholders that demand maximization of shareholder wealth (Abdullahi & Mansor, 2015). We use these theories to provide a plausible explanation to why firms engage in illegal behaviour

2.1.6 Fraud Triangle

The fraud triangle consists of three parts: opportunity, pressure and rationalization (Abdullahi & Mansor, 2015). Opportunity is the chance to commit fraud. Pressure is the driver to commit fraud, e.g. imminent bankruptcy, meaning liquid assets are needed. Pressure on TMTs could also be explained by other theories. The last part of the triangle is rationalization. An example of rationalization of illegal behaviour would be: ‘I deserve this money as I gave my heart and soul to the firm’. The triangle needs to fulfil the criteria for all three sections before a firm commits fraud. Fraudulent behaviour poses a risk with respect to the continuity of the firm. Therefore, a firm is less likely to commit fraud if the pressure in the triangle is removed. By increasing TMT diversity, the risk-aversion will also increase; as a result, TMTs are less likely to commit illegal behaviour (Nelson, 2012).

2.2 Top Management Team

To describe top management teams we look at the definition of Nielsen & Nielsen (2011). They define the top management team (TMT) as consisting of the CEO, top-level executives, board members, and other upper-level management staff. This team’s primary purpose is to make all the important decisions regarding the future, orientation and operations of a firm (Nielsen & Nielsen, 2011). Ling & Kellermans (2010) describe a TMT as the people in charge of the firm. From these definitions of a TMT, we can see that a top management team is an important part of a firm’s hierarchy. As the Upper Echelon theory explains, this team stands at the top of the firm, decides future directions, and makes important decisions (Hambrick & Mason,1984). A factor that could influence the TMT’s decisions is diversity. Due to the differences between members, diversity can lead to negotiations and compromise to satisfy all of its members.

2.2.1 Gender Diversity

10

transgender (Kimmel & Aronson, 2010). TMT members are classified as either male or female. Neither Kimmel and Aronson nor others, such as Eckert and McConnell-Ginet (2003) or Butler (1990), mention that a cultural construct is the basis of identifying as male or female. The authors see gender as an identity to which people can refer. The beliefs with respect to identity differ per region. Besides this, the role identity plays in society is also different. In this thesis, the definition of identity is followed in which people identify as either male or female.

In the case of gender diversity in management, research has pushed for more diverse top management teams (Zhu, 2013; Miles & Erhardt 2014). According to these groups, gender diversity will eliminate or limit the appearance and impact of illegal behaviour in firms. We know that firm performance and CSR are positively influenced by TMT diversity (Perryman et al., 2016; Milliken, 2019). According to Perryman et al., (2016) firms with relatively more gender diversity display, higher levels of ‘morally acceptable’ behaviour when compared to firms with less diversity. Similar results to Perryman et al. (2016) are found in Zhu (2013) and Thorne & Koningsburg (2020). Though, hardly anyone asks what the influence of TMT diversity is on illegal behaviour.

Nelson (2012) also shows the influence of gender diversity on illegal behaviour. She discovered that women are more risk-averse than men. As a result, women are less likely to partake in illegal behaviour due to the risks involved (Jianakoplos & Bernasek, 1998). However, if we look at TMT diversity, it is not just gender diversity; it also incorporates age, nationality and tenure diversity. Each type of diversity has its own hypothesis. The following section explains why we propose that gender diversity influences illegal behaviour.

The topic of gender diversity and illegal behaviour is under-represented in research. Not many authors have studied the forms of illegal behaviour that we classified as fraud, corruption and bribery.

11

As mentioned earlier, Nelson (2012) proved that women are more risk-averse than men. Therefore, women are less likely to engage in illegal behaviour due to the risks involved (Jianakoplos & Bernasek, 1998). Moreover, Perryman et al. (2016) showed that an increase in gender diversity improved ethical behaviour and firm performance. If a firm is performing well, there is less incentive for the TMT to perform illegal behaviour (Cohen et al. 2010).

As a result, we expect a positive correlation between a more gender diverse TMT and a reduction in illegal behaviour. We can test this by relatively analysing the gender ratio and the appearance of illegal behaviour. This idea is further tested in hypothesis 1.

Hypothesis 1: The presence of more females in the TMT of a firm reduces the likelihood of illegal behaviour.

2.2.2 Age Diversity

The influence of age diversity in TMT has recently been studied by frequently cited authors Xu et al. (2018) and Tanikawa et al. (2017). Both author groups studied the influence of age diversity on TMTs. Xu et al. studied the influence of TMT age on the likelihood of CEO fraud, while Tanikawa et al. focused on the relationship between TMT age diversity and financial performance. Yi et al. (2018) and other authors measure TMT age diversity as the standard deviation of the average age of all TMT members in the sample (Kunze, et al., 2011; Xu et al., 2018).

Xu et al. (2018) show the importance of age. They found a relationship between TMT age diversity and engagement in fraudulent activities (Xu et al., 2018). Their data demonstrated a strong relationship. In their findings, they mention that the larger the age diversity between TMT members, the more likely the appearance of fraud. However, the authors discovered that this likelihood is reduced when the average board age increases: a higher average TMT age decreases the likelihood of CEO fraud. Tanikawa et al. (2017) found a negative correlation between TMT age diversity and firm performance: an increase in age diversity results in a lower return on equity(ROE) for the firm. However, just like the previous article, Tanikawa et al. found that this effect was mitigated when the average TMT age was increased. In both articles, the authors confirmed that TMT age diversity is correlated to firm performance and fraud.

12

possible correlation. Tanikawa et al. (2017) and Xu et al. (2018) both found that TMT age diversity has a negative impact if the diversity increases too much.

Xu et al. (2018) and Tanikawa et al. (2017) mention that the likelihood of committing fraud decreases with an increase in average board age. These findings indicate that a moderate age-diverse TMT is important, as a moderate diversity reduces the likelihood of committing illegal behaviour, while high age diversity increases illegal behaviour. This is supported by the findings of Tanikawa et al. (2017) and Xu et al. (2018): we can expect that age diversity in TMT will have a similar impact as it had in the previous studies. Based on this we can propose that TMT age diversity increases the likelihood of illegal behaviour, but that this correlation is mitigated when the average TMT age increases. This proposition is tested with hypothesis 2.

Hypothesis 2: An increase in age diversity in TMT increases the likelihood of illegal behaviour

2.2.3 Nationality Diversity

Nielsen & Nielsen (2012) define nationality as a person bearing a passport from a certain country. Nationality is a concept in which a person is identified by the documents of a country. Nationality includes rights, rules, protection, and belonging to a certain group of people. Nationality is also cultural. Hofstede (1984) wrote that each country has its own national culture, values and beliefs. This makes each nationality unique; each nationality has a different impact on TMT decisions. Because, nationality diversity incorporates the values, beliefs and norms of different cultures as they impact on TMT processes. Due to their unique perspectives, TMT members will make different choices in order to reach a compromise or resolve issues.

Nielsen & Nielsen (2012) argued for having multiple different nationalities in a TMT to improve the decision-making process. This idea is supported by Hofstede (1984). He argues that the differences in national culture influence a firm’s operations, as each person brings different norms and values. In his theory, each country is assigned scores regarding masculinity, power distance, uncertainty avoidance, etc. These differences in masculinity, power distance, etc. influence the behaviour and decisions of TMT members. Hofstede (1980) showed that these beliefs and values differ among age groups, nationalities and genders. Unique compositions with different nationalities will influence the decisions of an organization differently.

13

and firm performance exists. This relationship is enhanced when TMT members work together for a longer period. This discovery can be explained with reference to his national culture theory. In this theory, Hofstede (1984) argues for the consequences of national culture on a firm’s operations. In the theory, each country is assigned scores regarding, masculinity, power distance, uncertainty avoidance, etc. These scores give an overview of a nation’s culture. As each nation is different, we can expect different behaviour, beliefs, norms and moral codes from them.

If we accept the notion that nationality diversity in TMT is beneficial to the firm, such diversity should also help the firm in reducing illegal behaviour. We expect this because nationality diversity increases CSR, which is the opposite of illegal behaviour. This is supported by the findings of Milliken (2019). These findings show that nationality diversity is positively related to CSR. When more nationalities appear in a TMT, illegal behaviour decreases. This will be tested with hypothesis 3.

Hypothesis 3: Increasing nationality diversity in TMTs reduces the likelihood of illegal behaviour.

2.2.4 Tenure Diversity

Tenure is described as the length of time a person has worked for a firm (Tanikawa & Jung, 2016). This period is measured from the day a person started working at the firm until their retirement or leave. The diversity of tenure can be measured with the standard deviation tenure at the firm of a firms TMT members.

14

The findings of Williams et al. (2005) and the other authors stress the importance of including tenure diversity. In the articles mentioned above, the TMTs perform better with greater diversity in TMT tenure, and a more diverse team engages more in illegal activities.

Based on the reasoning above, we can expect that a more tenure-diverse TMT will not commit illegal behaviour, as they have already satisfied their requirements and do not perceive any pressure from shareholders (Williams et al., 2005; Yi et al., 2018). Therefore, by taking the findings of Tanikawa & Jung (2016) and other authors into consideration, we can propose the following:

Hypothesis 4: An increase in a firm’s TMT tenure diversity will increase the likelihood of illegal behaviour.

2.3 Conceptual Model

The hypothesis mentioned above results in the following conceptual model:

15

3. Methodology

3.1 Variables

3.1.1 Dependent variables

The dependent variable used is ‘illegal behaviour’, incorporating fraud, corruption and bribery under this single term (Braithwaite, J., Walker, J., Grabosky, 1987; Gilardi, 2002). The sources of information mentioned in paragraph 3.3 Data Sources were used to identify firms which have been found to have committed fraud, corruption or bribery. For this, the keywords ‘fraud’, ‘bribery’ and ‘corruption’ are used. This binary variable is called Illegal Behaviour in the dataset. The variable is assigned a score of 0 or 1. An illegal behaviour firm is a 0 and a firm without illegal behaviour is a 1. The variable is equally distributed in the dataset: 50% of the firms display illegal behaviour and 50% do not at the time of comparison with the fraudulent firm.

3.1.2 Independent variables

Each independent variable measures a part of the diversity in top management team composition. In this thesis, these variables explain the influence of the diversity in the composition of a TMT on illegal behaviour. The variables used to measure this diversity are gender, age, nationality and tenure. To measure the diversity of each variable, diversity indexes and scores are used. Age and tenure are measured with the diversity in standard deviation corrected by the board size. Gender is measured as a ratio of the percentage of women on a board. Finally, nationality is measured with the help of Simpson’s Diversity Index (Torchia et al., 2015; Simpson, 1949).

Gender diversity

16

which the whole board is female; and the lowest 0, when all members are male. In this thesis, the highest possible score for diversity is 0.5. In this case, 50% of the TMT members are male and 50% are female. In the dataset, gender is coded as GEN.Div. To measure the diversity in gender, we look at the gender distribution at board level.

Age diversity

Age is measured as the number of years a person has been alive, measured from the day they are born until the moment of illegal behaviour. The name of this variable in the dataset is age. This data can be found on LinkedIn, BoardEx, and other network websites. In this study, it is a scale variable, which indicates how old a TMT member is.

To calculate age diversity, we use the standard deviation of age in a board. The outcome generates the variable age diversity, noted as AGE.Div in the dataset. The higher this score is, the more diversity in age is present in the TMT. With a low value, hardly any diversity in age is present. The Age.Div value is used to test the hypothesis of age diversity.

Nationality diversity

Nationality is the country of origin or the passport that a TMT member currently holds. The variable nationality refers to the different nationalities in the dataset. The score of variable nationality diversity is determined by the number of different nationalities in a TMT. The diversity variable is calculated per board, as the influence of diversity in a TMT on illegal behaviour is tested. To calculate the nationality diversity, Simpson’s Diversity Index is used.

Simpson (1949) wrote two formulas to calculate the diversity in a population. The first formula is based on a large sample. However, the board size to test diversity in this study is small between 6-29 members. In this thesis, the second formula for a small sample size is used.

17

to 1, in which 1 indicates the highest possible diversity, each member possessing a different nationality.

Tenure

Tenure is essential for decision-making. An employee works for a length of time at the firm (tenure) and receives experience in relation to their job while working. Dokko et al. (2009) published an article about the utility of experience in a job. In their study, they found that experience influences job performance and people’s choices. This is also confirmed in the study of Quinones et al. and others (Goldman, 2008; Quinones, Ford, & Teachout, 1995). Quinones et al. used a meta-analysis to study the importance and role of experience and tenure. Information on tenure at a firm is found in BoardEx, company websites, and LinkedIn.

Tenure is measured as the time spent at the company that committed illegal behaviour. This variable is measured as TIC in the dataset. To measure diversity, this variable is transformed to create TEN.Div. The variable TEN.Div is calculated as the standard deviation of tenure in a TMT. The outcome of this equation is used as the value of tenure diversity for each firm. Similarly to age diversity, the higher the value of TEN.Div, the larger the diversity in tenure is. With a small value, little diversity exists. We refer to TEN.Div as tenure diversity.

3.1.3 Control variables

We used firm age, firm size and board size as control variables. Because these variables were identified by prior research, as variables that could influence the dependent variable. For each control variable, we explain how it is measured and we refer to authors that noticed their influence.

Firm age

18

the other variables. The variable Firm age is measured from the founding of the firm until the first year of illegal behaviour in the dataset.

Firm Size

Similar to firm age it is important to control for the likelihood that illegal behaviour appears in small, medium or large firms. Through a numerical variable, we measure firm size as a logarithm of the number of employees (Barrick et al., 2007; Galbreath, 2018; Wu, 2006). The number of employees is derived from current annual reports.

Board size

Each variable is controlled for TMT size. This is to check whether the size of the TMT influences the likelihood of illegal behaviour. We define board size as the number of people that are part of a TMT (Certo et al., 2006; Hermalin & Weisback, 2003). To study the relationship between TMT diversity and illegal behaviour, we selected the board size at the moment the firms started to engage in illegal behaviour, with a minimum board size of 6 members.

3.2 Research design

19

3.3 Data Sources

The most important data sources that we use are past litigations and enforcement of the US Securities and Exchange Commission (US SEC) and the US Department of Justice (DoJ). Their decisions are of great value to this thesis with respect to classifying firm involvement in fraud, corruption or bribery. The data on the identified firms is collected from BoardEx and Bureau van Dijk Orbis (BvD Orbis). The decisions of the US SEC and DoJ are the deciding factors in terms of whether firms engaged in illegal behaviour. Their decisions are supported by investigations of Europol and the US Federal Bureau of Investigation. The dependent variable in this thesis is based on the decisions of these authorities.

3.3.1 US SEC & DoJ

The US Securities and Exchange Committee monitors the firms noted on the US stock exchange, with respect to laws and regulations. For this thesis, the ‘Accounting and Auditing

Enforcement Releases’ database provides information regarding violations and sanctions. Other

authors, such as Beasley et al. (1999, 2000), Bonner et al. (1998), and Linke & Emanuels (2009), used this database and the database of the US Department of Justice (DoJ) to study forms of illegal behaviour. The Accounting and Auditing enforcement releases database can be found under the tab litigations & releases on the US Securities and Exchange Committee website. Similar to the ‘Foreign Corrupt Practices Act’ (FCPA) database, a lot of cases in the Accounting and Auditing Enforcement Releases are solved with settlements without admissions of guilt or denial. In this thesis, settlements are also classified as cases of illegal behaviour. The US SEC database is recognized as one of the most reliable sources of information on illegal behaviour (Beasley et al., 1999). The FCPA database provides reliable information on the practice of corruption and fraud in and outside of the United States of America.

3.3.2 BvD. Orbis

20

size is a similar situation: it is also found in the Bureau van Dijk Orbis dataset, under the heading

number of employees. For this variable, the most recent number of employees is used.

3.3.3 Time Period & Distribution

The period from which cases are selected encompasses the years 2015-2019. These cases are located in two regions: Europe and the United States. In each region, sample cases and other information are selected with the help of data provided by the US Securities Exchange Commission, US Department of Justice (DoJ), and the Stock Exchange in both regions. This method of data selection for fraudulent firms is also used by Kamarudin et al (2012; 2014; 2018).

3.4 Data

For the sample firms, 50% of the firms were selected due to their presence in the databases of the US SEC and the DoJ during the years 2015-2019. The amount of data was reduced by multiple requirements. First, we remove all non-business entries and non-multinational firms and because we are interested in firms in the USA and Europe, all firms from different regions are removed from the dataset. Further reduction of the number of firms comes from a minimum board size of 6 members, our classification of illegal behaviour and lack of information on board members and financial data. The final reduction requirement is an equal dataset between the USA and Europe. These requirements provided a total of 46 illegal behaviour cases suitable for research.

The other 50% of the sample was collected through the BvD Orbis dataset. This data, concerning firms without illegal behaviour, is collected as a peer to the firm that engages in illegal behaviour. The peer firms are matched in industry and their current employee size. The current size is chosen to allow for a better comparison of a firm’s behaviour, as we expect that firms go through a similar cycle to achieve their current status and size.

21

different systems, and different restriction on information sharing. Only firms which fulfilled all requirements for this study and have available data were used. The data collected from US SEC, US DoJ and BvD Orbis were made suitable to use in this study by assigning values of 0 to firms with and a value of 1 without illegal behaviour. The distribution of the firms with and without illegal behaviour per industry can be found in table 1 below.

Table 1: Distribution of sample firms Industry

Firms, illegal behaviour Firms, legal behaviour

Frequency Percentage Frequency Percentage

Electronic & Electrical Equipment

6 13,04% 6 13,04%

Financial services 10 21,74% 10 21,74%

Healthcare 4 8,7% 4 8,7%

manufacturing 6 13,04% 6 13,04%

OIL & GAS 3 6,52% 3 6,52%

Pharmaceuticals and Biotechnology 5 10,87% 5 10,87% Retail 4 4 8,7% 4 8 4 8,7% Software & Communications 8 17,39% 8 17,39% Total 46 100% 46 100% Total samples 92

3.4.1 Bureau van Dijk Orbis Research Strategy

The data on firms without illegal behaviour was collected with the help of the BvD Orbis database. In this database, each firm has an identifying industry code. The code used to identify a matching firm without illegal behaviour is NACE Rev 2. In the search strategy of Orbis, we took the following steps to identify firms of the same size, industry and region. First, we selected all active firms; then we selected Location>world region. In the following table, we chose the regions North America, Western Europe, and Eastern Europe. In this study, these are the regions of interest. Search strategy step 3 was to select the NACE Rev. 2 code of the firm with illegal behaviour. Finally, step 4 concerned the number of employees. Here, the most recent range of the number of employees was selected.

22

firm itself performed illegal behaviour in the same time period. If the peer firm passed these checks, it was added to the dataset.

3.4.2 Top Management Team Diversity

To continue the study, values are assigned to the top management team composition. The TMT members selected are those on the board at the start of the illegal behaviour. In a case of fraud in 2008, the data on the board of 2008 is selected and tested. The information on the board members was found in the BoardEx and Orbis database. For firms to be included in the sample, the board size needs to cross the minimum threshold of 6 board members. This information is confirmed with the help of annual reports and company website.

Table 2: Board Size

Variable Average Min Max

Board Size 12.435 6 29

This data is used to assign scores to each board of the selected firms. We can then identify which value plays a role in the influence of a TMT (Hambrick & Mason, 1984). These values relate to the characteristics of board members, such as gender, age, nationality and tenure. These provide insights into the diversity of the board of each firm.

The combination of the datasets allows for analysis of the relationship between TMT diversity and the charged firms. The results of this analysis provide further insights into the recommended diversity and the influence of diversity in gender, age, nationality and tenure.

To make our data appropriate for testing, the diversity scores are controlled for board size. This is done to remove the influence of larger boards; by adding this step, each TMT’s diversity score carries the same weight.

3.5 Methods

The collected quantitative data is used to test the hypotheses. With a logistic regression method, it is possible to answer our research question: what is the influence of TMT composition

23

The logistic regression analysis is performed to test the relationship between TMT diversity and illegal behaviour; the outcome will indicate the existence and strength of a relationship between the dependent and independent variables. To test with a logistic regression, the independent variables are transformed into ratios or indexes. In logistic regression, we differentiate between two values: no (0) and yes (1). We want to test if a positive relationship exists, in which an increase in the values of the independent variables will also increase the value of the dependent variable. In this thesis, we expect that diversity will reduce the likelihood of illegal behaviour. As we want to test whether an increase in diversity influences illegal behaviour, we have to code firms with illegal behaviour as 0 and those without illegal behaviour as 1. Due to the use of multiple predictors or indicators that influence the dependent variable, we used the following logistic regression:

Behaviour=β0+ β1GenDiv+ β2AgeDiv+ β3NationalityDiv+ β4TicDiv+ β5FirmAge+ β6BoardSize+ε

24

4. Results

4.1 Empirical Analysis

To test the hypotheses, the statistical program StataSE (version 16) is used. The goal of the study is to identify and explain the relationship between multiple independent variables and a binary dependent variable. Therefore, a logistic regression is applied in StataSE. This method allows for proper testing of the variables. We use a correlation matrix, to test for multicollinearity between the independent variables.

4.2 Descriptive Statistics

Table 3 shows the descriptive statistics. As can be seen, there are 92 observations for each variable. Each variable has multiple descriptive values. The number of observations shows the number of values used to calculate the other descriptive statistics. Behaviour, gender diversity and nationality diversity are nominal variables. Their mean values indicate the percentage of the sample that is legal behaviour, female, or has a different nationality. The mean for age diversity, tenure diversity and firm age are given in years; the mean for firm size is the mean of the logarithm of the number of employees; and the mean for board size is the number of people on the board.

Table 3: Descriptive Statistics

Variable Obs. Mean Std. Deviation Min. Max.

Behaviour 92 0.5 0.503 0 1 Gender Diversity 92 0.105 0.089 0 0.400 Age Diversity 92 7.399 2.054 2.575 14.997 Nationality Diversity 92 0.281 0.268 0 0.920 Tenure Diversity 92 6.159 4.828 0.477 35.087 Firm Age 92 58.217 54.883 5 226 Firm Size 92 10.215 1.755 6.346 14.604 Board Size 92 12.435 4.707 6 29

25

diversity values are 0 for both gender and nationality diversity. This means that no diversity is present in some firms. For gender, the highest possible diversity value is 0.5. The maximum sample value of 0.4 shows that 40% of the TMT members are female. However, if we look at the mean of 10.5% and the standard deviation of 8.9%, we see that most of the sample firms have a gender diversity within the range of 1.6% and 19.4% with respect to the percentage of females in the TMT. Nationality diversity shows a similar situation. These values show that most top management teams consist of a homogenous group of males and/or a single nationality.

4.3 Correlation Analysis

The correlation matrix is presented in table 4. In this matrix, the correlation between all the independent variables is measured. This table shows that all correlations are below the 0.800 threshold for multicollinearity. As a result, no variables are eliminated from the dataset. The table shows that there is no significant correlation between any of the independent variables. A correlation is significant if the p-value is below 0.1. Gender diversity and board size show a significant positive correlation with firm size, at *** with a p<0.01 level. This means that these variables influence each other: an increase in firm size also increases gender diversity and board size. A negative correlation indicates that if one variable increases, the correlated variable decreases.

Table 4: Correlation matrix

Variable Gender diversity Age diversity Nationality diversity TIC diversity

Firm age Firm size Board size Gender diversity -0.138 -0.108 0.053 0.092 0.289*** 0.080 Age diversity -0.138 -0.055 0.028 -0.138 -0.110 -0.048 Nationality diversity -0.108 -0.055 -0.149 -0.086 0.220 0.005 TIC diversity 0.053 0.028 -0.149 0.172 -0.050 -0.018 Firm age 0.092 -0.138 -0.086 0.172 -0.002 0.027 Firm size 0.289*** -0.111 0.220 -0.050 -0.002 0.549*** Board size 0.0800 -0.049 0.005 -0.018 0.027 0.549*** *** p: 0.0 ≤ x ≤ 0.01 ** p: 0.01 < x ≤ 0.05 * p: 0.05 < x ≤ 0.1 4.4 Regression Results

26

significance measurement indicates a very high level of probability, and no significance measurement indicates a low probability. In our logistic regression, all variables (described in chapter 3) are used. The descriptive statistic above (table 3) show the number of observations in the regressed sample. The regression results in table 5 show six models. The first model regresses only the control variables on behaviour, from the second model onwards the diversity variables are added. Model 6 shows the combined influence of all variables on illegal behaviour. The constant in table 5 shows the intercept with the y-axis.

Model 1 tests the influence of the control variables on illegal behaviour. Firm age (0.105) and board size (0.752) have a positive relationship with behaviour, while firm size (-0.129) has a negative relationship. However, none of the variables has a significant relationship.

To test the influence of gender diversity, this variable is added to the control variables in Model 2. The logistic regression results in table 5 show that this variable is not significant (0.727) with a negative relationship (-0.873). Similar results are shown in Model 6, with no significance (0.494). As the p-values are above 0.1 in both models, the variable of gender diversity is not significant at any level. Therefore, we reject hypothesis 1.

In model 3, the variable age diversity is added to test its significance. Both model 3 (0.369) and 6 (0.316) show a negative relationship with no significance. Model 6 shows a negative relationship (0.369) similar to model 3 (-0.097). As the regression results show no significant evidence of a relationship between TMT age diversity and illegal behaviour, the second hypothesis is also rejected.

Model 4 shows the influence of the variable nationality diversity. The coefficient is negative in model 4 (-0.945) and model 6 (-1.153). Both models show that there is no significant evidence to support the hypothesis. The p-values are above 0.1, for model 4 (0.255) and model 6 (0.180). Therefore, we reject hypothesis 3.

27

Table 5: Logistic Regression on the Relationship of the Variables with Behaviour

Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Coef P>│z│ Coef P>│z Coef P>│z Coef P>│z Coef P>│z Coef P>│z

Gender diversity -0.873 0.727 -1.803 0.494 Age diversity -0.097 0.369 -0.111 0.316 Nationality diversity -0.945 0.255 -1.153 0.180 Tenure diversity -0.016 0.721 -0.020 0.657 Control Variables Firm Age 0.105 0.930 0.147 0.902 -0.050 0.967 -0.011 0.993 0.180 0.882 -0.032 0.979 Firm Size -0.129 0.372 -0.114 0.452 -0.146 0.323 -0.086 0.565 -0.132 0.364 -0.068 0.675 Board size 0.752 0.319 0.726 0.338 0.776 0.306 0.639 0.403 0.753 0.317 0.599 0.442 Constant -1.324 0.885 -1.647 0.858 0.667 0.944 -0.345 0.970 -1.779 0.848 0.914 0.924

Hosmer and Lemeshow’s chi2 4.36 0.824 8.13 0.421 7.66 0.467 13.05 0.110 14.30 0.074 3.55 0.895

McFadden’s Pseudo R2 0.009 0.010 0.016 0.020 0.010 0.03

Observations 92 92 92 92 92 92

Correctly classified 53.26% 55.43% 53.26% 56.52% 57.61% 55.43%

28

4.5 Goodness of fit

To ensure that our models fit the data and can predict the dependent variable, it is necessary to check the goodness of fit. For this statistic, we look at the results of Hosmer and Lemeshow’s chi-square and the value for the McFadden’s Pseudo R-square. The pseudo-R-square is slightly different from a normal R-square value, as the variation in logistic regression is fixed, with a standard logistic distribution. McFadden’s Pseudo R-square shows the terms of proportion in the log-likelihood of behaviour. As a result, the Pseudo-R-square can be used similarly to a regular R-square. If the Pseudo-R-square-model can predict a large percentage of the observations, it will indicate a high value. With a low value, the model cannot explain the changes in the value of the dependent variable. To see how well the model fits, we look at the Pseudo-R square values in table 5. For example, table 5 shows a value of 0.009 for model 1 and 0.010 for model 2.

The Hosmer and Lemeshow’s chi-square is used to identify whether a model is statistically significant. Table 5 shows the significance of the Hosmer and Lemeshow’s chi-square test. For a good fit, the p-value must be below 0.05. In this thesis, the p-value of Hosmer and Lemeshow’s chi-square is above 0.05 for all models. The goodness of fit criteria is p<0.05.

29

5. Discussion

Many studies have been conducted on the relationship between TMT composition diversity and firm performance. However, the topic of TMT diversity and illegal behaviour is underrepresented. The general consensus is that diversity is good for a firm. It improves firm performance and improves CSR. There is contradictory evidence for the influence of gender, age, and tenure diversity. Authors like Kamarudin et al. (2014; 2016; 2018) found no relationship between gender and tenure diversity and fraud, while Williams et al. (2005) and Nielsen & Nielsen (2013) found that diversity increases firm performance and reduces the likelihood of illegal behaviour. The other diversity variables show no significant relationship and neither support nor refute prior literature.

Our first hypothesis is: the presence of more females in the TMT of a firm reduces the

likelihood of illegal behaviour, is rejected, as the regression results did not show any

significance. The results are in line with Kamarudin et al. (2018). In their paper, they discovered that gender and tenure diversity have no relationship with fraud. The results of our logistic regression analysis and the findings of Kamarudin et al. (2018) contradict the findings of Nelson (2012). Nelson found that women are more risk-averse and that they are also more receptive to social pressure and norms. A possible explanation for our findings is provided by Joecks et al. (2013). The authors state that a critical mass of 30% women in a board is required before changes or improvements in the behaviour of the TMT will be seen. This threshold provides a reasonable explanation as to why we did not find any significant relationship in our data; namely, we did not cross the diversity threshold, our average degree of diversity being 10.5%.

A similar situation appears for the second hypothesis: an increase in age diversity in

TMT increases the likelihood of illegal behaviour. The coefficient of age diversity is not

30

A possible justification for no significant evidence in our study is that we did not measure for a curvilinear relationship, but rather a logistic regression.

In the third hypothesis, we tested whether an increase in nationality diversity in TMT

reduces the likelihood of illegal behaviour. The regression results show that there is no

significant evidence to support this hypothesis. Our results are not in line with Nielsen & Nielsen (2012) and Hofstede (1984). Hofstede found that each person is influenced by their national culture and that this, in turn, influences their decisions. He states that a feminine culture is oriented towards consensus. Such a culture is based on cooperation, taking care of the weak, and practices which benefit society; while a masculine culture is competitive and based on success, achievements and material benefits. The idea of national influence is supported by Nielsen & Nielsen (2012). They found that nationality influences the decisions of an individual. Based on these findings, it was expected that nationality diversity reduces illegal behaviour. However, the results contradict our expectations: there is no evidence that unique nationalities influence illegal behaviour. This can be explained by the existence of single-nationality TMTs. Nielsen & Nielsen (2012) and Hofstede (1984) explained that every nationality has a unique influence on people’s behaviour. In order for nationality diversity to be present, more than a single nationality is required. The existence of many homogenous or low-nationality-diversity TMTs in our study could be an impediment to finding a significant relationship between nationality diversity and illegal behaviour.

The last hypothesis is: an increase in a firm’s TMT tenure diversity will increase the

likelihood of illegal behaviour. This hypothesis produced similar results to the three hypotheses

31

6. Conclusion

6.1 Theoretical implications

The main goal of this thesis is to identify the influence of diversity in TMT composition on illegal behaviour. Our contribution to these studies is a combined study of the influence on illegal behaviour of the diversity variables: gender, age, nationality and tenure. This research is under-represented, as most studies focus on the influence of TMT composition on firm performance or CSR. Our findings on gender and tenure diversity support those of Kamarudin et al. (2018), Xu et al. (2017), and Yi et al. (2018). In their study, Kamarudin et al. (2018) found no relationship between gender diversity, tenure diversity and fraud. We also believe that our results support the findings of Joecks et al. (2013). We believe that our results do not show a significant relationship between illegal behaviour and TMT composition diversity, as the minimum amount of diversity required to influence illegal behaviour was not achieved. This is in line with Joecks et al., who state that a minimum of 30% diversity is needed before a diversity variable will influence a dependent variable. Specifically, our findings also add to the contradictions in theory. Our results support the findings of Kamarudin et al. (2018) and Joecks et al. (2013) but contradict those of Nielsen & Nielsen (2013) and Nelson (2012). Based on the latter’s findings, we expected a significant positive relationship between gender diversity and nationality diversity, since the findings of Perryman et al. (2016) and Milliken (2019) show that TMT diversity increases CSR and firm performance. Moreover, firms engaging in CSR and exhibiting high firm performance are less likely to engage in illegal behaviour (Johnson, 2003). However, our results suggest that there is no relationship between illegal behaviour and TMT composition diversity.

32

or Tanikawa et al. (2017) identifies a threshold or ceiling which influences whether age or nationality diversity decreases or increases the dependent variable.

6.2 Practical implications

The coefficients in our findings only provide indications for a positive or negative relationship, since the results are not significant due to the low percentage of diversity in TMT’s. Therefore, TMT composition diversity shows no influence on illegal behaviour. As a result, the managerial or practical application of this study is reduced. However, firms should increase TMT composition diversity if they are serious about equality, CSR and firm performance (NY Times, 2019: Perryman et al., 2016).

Firms should recognize that a low level of diversity does not have any influence on illegal behaviour. As a result, firms should realise that diversity itself must be increased before its influence on illegal behaviour can be measured. This is in line with Joecks et al. (2014) and Milliken (2019).

6.3 Limitations

During our study, we observed a number of limitations. The most significant of these are detailed below. The first limitation was the difficulty of collecting data, as not all firms provided information on their TMT members: not all information on the firms was readily available in the databases, and some data points were missing or hard to find. Moreover, sometimes the information in a database was shown to be incorrect when compared to annual reports. This could have happened when the firm merged, changed its name, or was added to the stock exchange after its Initial Public Offering (IPO). With a large, more complete, combined and categorised dataset, further research would not be limited by missing and incorrect information.

33

A possible explanation for finding no significant results in the logistic regression is that little diversity exists in any of the observed variables shown in table 3. Therefore, it is doubtful that the existing diversity influences TMT decisions to engage in illegal behaviour, as gender diversity requires a minimum threshold of 0.3 to make an impact on the results (Joecks et al., 2013). Similar thresholds could exist for other forms of diversity, such as age and nationality. Most of the sample firms had a homogeneous TMT, either consisting of a large percentage of males or being composed entirely of men. The low percentage of women could explain why gender diversity has no significant influence on the likelihood of illegal behaviour. With only one or two women on a board, their influence is likely to be reduced by the large percentage of male TMT members. The same applies to for age and nationality diversity: they were also relatively homogeneous in the sample.

The last limitation we mention is that the company mission/vision is not considered. The company mission or vision identifies what the company stands for and what ideology the company has. A firm that focuses on the shareholder theory might make different choices which might increase the likelihood of illegal behaviour compared to firms with CSR-oriented theories such as the triple-p theory.

6.4 Future research

While testing our hypotheses, we observed a range of promising areas for future research. Due to the limitations mentioned in section 6.3, the variables show low values of diversity. We suggest that future research could be done on these variables when TMT diversity has been increased to above the required threshold, which could provide further insights into the relationship between TMT composition diversity and illegal behaviour.

34

We also noted contradictions in the influence of nationality and culture on illegal behaviour. Future research could be performed to determine whether nationality and culture have any influence on illegal behaviour, or if such behaviour can be better explained by other variables. Tenure diversity is similar: it could also be studied further to determine whether it has an effect on illegal behaviour, or in line with the findings of Kamarudin et al. (2018), to determine if it is more productive to focus on the influence of other factors.

35

References

Aguinis, H., & Glavas, A. (2012). What we know and don’t know about corporate social responsibility: A review and research agenda. Journal of Management, 38(4), 932-968. Agrawal, A., Chadha, S. (2005). Corporate Governance and Accounting Scandals. The

Journal of Law and Economics, 48 (2), 371-406

Anholt, R.H., Mackay, T. (2009). Principles of Behavioral Genetics. Academic Press. ISBN: 9780123725752

Ambrose, J. (2019). Most of world's biggest firms 'unlikely' to meet Paris climate targets. The Guardian

https://www.theguardian.com/environment/2019/sep/24/most-of-worlds-biggest-firms-unlikely-to-meet-paris-climate-targets

BBC. (2019). Sweden's Ericsson to pay over $1bn to settle US corruption probe. BBC

https://www.bbc.com/news/world-us-canada-50695438?intlink_from_url=https://www.bbc.com/news/topics/cm8m1lj59z3t/corruption &link_location=live-reporting-story

Bantel, K. A., & Jackson, S. E. (1989). Top Management and Innovations in Banking : Does the Composition of the Top Team Make a Difference ? Strategic Management Journal,

10(S1), 107–124.

Barrick, M.R., Bradley, B.H., Kristof-Brown, A.L., & Colbert, A.E. (2007), “The moderating role of top management team interdependence: implications for real teams and working groups”, Academy of Management Journal, Vol. 50, pp. 544-557.

Beasley, S.M., Carcello, J.V., & Hermanson, D.R. (1999), Fraudulent financial reporting 1987-1997. An Analysis of U.S. public companies. Committee of Sponsoring

Organizations of the Treadway Commission (COSO), New York.

Beasley, S.M., Carcello, J.V., & Hermanson, D.R. (2000). Fraudulent financial reporting: Consideration of industry traits and corporate governance mechanisms. Accounting

Horizons, 14 (4), pp. 441-454

Bhattacharyya, R. (2020). CSR spend of NSE-listed cos rises 18% in FY19. Economic Times https://economictimes.indiatimes.com/markets/stocks/news/csr-spend-of-nse-listed-cos-rises-18-in-fy19/articleshow/73555589.cms

Bonner, S.E., Palmrose, Z., & Young, S.M. (1998), Fraud type and auditor litigation: An analysis of SEC Accounting and Auditing Enforcement Releases, The Accounting

Review, 73 (4), 503-532

Braithwaite, J., Walker, J., & Grabosky, P. (1987). An Enforcement Taxonomy of Regulatory Agencies. Law & Policy, 9(3), 323–351.

Business Times. (2019). Ericsson agrees to pay over US$1b to resolve US corruption probe.

Business Times.

36

Business Times. (2020). US prosecutors accuses ex-Alstom executives of bribery. Business

Times.

https://www.businesstimes.com.sg/government-economy/us-prosecutors-accuses-ex-alstom-executives-of-bribery

Business Times. (2019). SEC permanently bars ex-Goldman banker over 1MDB bribery. https://www.businesstimes.com.sg/government-economy/sec-permanently-bars-ex-goldman-banker-over-1mdb-bribery

Butler, J. (1990). Gender trouble: Feminism and the subversion of identity. London: Routledge.

Calderón, R., Ferrero, I., & Redin, D. (2012). Ethical codes and corporate responsibility of the most admired companies of the world: Toward a third generation ethics? Business and

Politics, 14(4), 1-24.

Castelo B.M. (2013). Shareholder Theory. In: Idowu S.O., Capaldi N., Zu L., Gupta A.D. (eds) Encyclopedia of Corporate Social Responsibility. Springer, Berlin, Heidelberg Certo, S.T., Lester, R.H., Dalton, C.M., & Dalton, D.R. (2006), “Top management teams,

strategy and financial performance: a meta-analytic examination”, Journal of Management Studies, Vol. 43, pp. 813-839.

Chapman-Davies, A., Parwada, J.T., & Tan, E.K.M. (2014). The Impact of Scandals on Mutual Fund Performance, Money Flows and Fees. USNW Business School Research

Paper, NO.2014 BFIN 20

Citzen.org. (2018). Corporate Impunity.

https://www.citizen.org/wp-content/uploads/migration/corporate-enforcement-public-citizen-report-july-2018.pdf

Cloninger, D. O., & Waller, E. R. (2000). Corporate fraud, systematic risk, and shareholder enrichment. Journal of Socio-Economics, 29(2), 189–201.

https://doi.org/10.1016/S1053-5357(00)00061-5

Cohen, J., Ding, Y., Lesage, C., & Stolowy, H. (2010). Corporate Fraud and Managers’ Behavior: Evidence from the Press. Journal of Business Ethics, 95(SUPPL. 2), 271–315. https://doi.org/10.1007/s10551-011-0857-2

Coffee Jr, J.C. (2005). A Theory of Corporate Scandals: Why the USA and Europe Differ.

Oxford Review of Economic Policy, 21 (2), 198-211

Cooray, A., & Schneider, F. (2018). Does Corruption throw sand into or grease the wheels of financial sector development? Public Choice, 177, 111-133

Credit Suisse. (2007). Annual Report 2007.

https://www.credit-suisse.com/media/assets/corporate/docs/about-us/investor-relations/financial-disclosures/financial-reports/csg-ar-2007-en.pdf

Daboub, A.J., Rasheed, A.M.A., Priem, R.L., & Gray, D.A. (1995). Top Management Team Characteristics and Corporate Illegal Activity. Academy of Management

Review, 20 (1), 138-170.

37

MANAGEMENT IMPROVE FIRM PERFORMANCE? A PANEL DATA INVESTIGATION. Strategic Management Journal, 33(1), 1072–1089. https://doi.org/10.1002/smj

Dokko, G., Wilk, S. L., & Rothbard, N. P. (2009). Unpacking prior experience: How career history affects Job Performance. Organization Science, 20(1), 51–68.

https://doi.org/10.1287/orsc.1080.0357

Driel, H. V. (2019) Financial fraud, scandals, and regulation: A conceptual framework and literature review, Business History, 61:8, 1259-1299

Eckert, P., & McConnell-Ginet, S. (2003). Language and gender. Cambridge, UK: Cambridge University Press.

Elkington, J. (1997). Cannibals with forks. The Triple Bottom Line of 21st Century, Capstone Publishing Ltd., Oxford, UK

Erickson, M.L., & Gibbs, J.P (1978) Objective and Perceptual Properties of Legal Punishment and The Deterrence Doctrine. Social Problems, 25 (3), 1, 253–

264, https://doi.org/10.2307/800063

Friedman, M. (1970). The Social Responisibllity of Business is to increase its profits. The

New York Times Magazine.

Frooman, J. (1997). Socially Irresponsible and Illegal Behavior and Shareholder Wealth: A Meta-Analysis of Event Studies. Business & Society, 36(3), 221–249. Retrieved from https://journals.sagepub.com/doi/abs/10.1177/000765039703600302#article

Galbreath, J. 2018. Is Board Gender Diversity Linked to Financial Performance? The Mediating Mechanism of CSR. Business & Society, 57(5), 863–889.

Gelles, D., & Yaffe-Bellany, D. (2019). Shareholder Value Is No Longer Everything, Top C.E.O.s Say. The New York Times.

https://www.nytimes.com/2019/08/19/business/business-roundtable-ceos-corporations.html

Gilardi, F. (2002). Policy credibility and delegation to independent regulatory agencies: A comparative empirical analysis. Journal of European Public Policy, 9(6), 873–893. https://doi.org/10.1080/1350176022000046409

Goldman, E. F. (2008). The Power of Work Experiences: Characteristics Critical to Developing Expertise in Strategic Thinking. HUMAN RESOURCE DEVELOPMENT

QUARTERLY, 19(3), 217–239. https://doi.org/10.1002/hrdq

Grasmick, H. G., & Green, D. E. (1980). Legal Punishment, Social Disapproval and Internalization as Inhibitors of Illegal Behavior. In Journal of Criminal Law and

Criminology (Vol. 71). Retrieved from

https://scholarlycommons.law.northwestern.edu/jclc

Hausman, J. A. (1978). Specification Tests in Econometrics. Econometrica, 46, 1251-1271 Harstad, B., & Svensson, J. (2011). Bribes, lobbying, and development. American Political

38

Hepher, T., & Frost, L. (2020). Airbus bribery scandal triggers new probes worldwide. Reuters, https://www.reuters.com/article/us-airbus-probe/airbus-bribery-scandal-triggers-new-probes-worldwide-idUSKBN1ZX2MW

Hermalin, B. E., & Weisbach, M. S. (2003). Boards of directors as an endogenously determined institution: A survey of the economic literature. Economic Policy Review, 9, 7–26.

Hofstede, Geert (1984). Culture's Consequences: International Differences in Work-Related

Values (2nd ed.). Beverly Hills CA: SAGE Publications. ISBN 0-8039-1444-X

Jianakoplos, N.A., & Bernasek, A. (1998). Are Women more risk Averse?. Economic Inquiry,

36, 620-630. doi:10.1111/j.1465-7295.1998.tb01740.x

Jensen, G.F., Erickson, & M.L., Gibbs, J.P (1978). Perceived Risk of Punishment and Self-Reported Delinquency. Social Forces, 57 (1), 57–78, https://doi.org/10.1093/sf/57.1.57 Jensen, M.C., & Meckling, W.H., (1976). Theory of the firm: Managerial behavior, agency

costs and ownership structure. Journal of Financial Economics, 3 (4), 305-360 Jizi, M., Nehme, R., & Elhout, R. (2016). Fraud: auditors’ responsibility or organisational

culture. International Social Science Journal, 66(221–222), 241–255. https://doi.org/10.1111/issj.12128

Joecks, J., Pull, K., & Vetter, K. (2013). Gender Diversity in the Boardroom and Firm Performance: What Exactly Constitutes a ''Critical Mass?''. Journal of Business Ethics,

118¸61-72.

Johnson, N. D., Courtney, Lafountain, L., & Yamarik, S. (2011). Corruption is bad for growth (even in the United States). Public Choice, 147, 377–393.

https://doi.org/10.1007/s11127-010-9634-5

Johnson, H.H., 2003. Dose it pay to be good? Social responsibility and financial performance. Business Horizons 46 (6), 34–40.

Kamarudin, K.A., Wan Ismail, W.A., & Wan Mustapha, W.A.H. (2012). Aggressive Financial Reporting and Corporate Fraud. Procedia - Social and Behavioral Sciences.

65. 638–643. 10.1016/j.sbspro.2012.11.177.

Kamarudin, K.A., & Wan Ismail, W.A. (2014). The Effects of Audit Committee Attributes on Fraudulent Financial Reporting. Journal of Modern Accounting and Auditing. 10, 507-514.

Kamarudin K.A., Wan Ismail W. A., & Kamaruzzaman A.A. (2018) Board Members Diversity and Financial Statements Fraud: Malaysian Evidence. State-of-the-Art

Theories and Empirical Evidence. Springer, Singapore.

Karmann, T., Mauer, R. & Flatten, T. C., & Brettel, M. (2016). Entrepreneurial Orientation and Corruption. Journal of Business Ethics, 133(2), 223–234.

https://doi.org/10.1007/s10551-014-2305-6

Kimmel, M. & Aronson, A. (Eds.). (2010). The gendered society reader (4th ed.). New York: Oxford University Press.