Board characteristics and Financial
reporting fraud
Master thesis
Rijksuniversiteit Groningen
Faculty: Economie & Bedrijfskunde
Study: Accountancy & Controlling
G.T.J.M. Jansen.
Student number: 1876120.
E-mail:
g.t.j.m.jansen@student.rug.nl
First reader: drs. M. M. Bergervoet
Board characteristics and Financial
reporting fraud
Abstract
This study examines the association between the involvement of the CEO in
financial statement fraud and the characteristics of the board of directors during a
period of financial crisis and after several enormous financial scandals. The
involvement of the CEO in financial reporting fraud will be related to
characteristics of the board that represent the monitoring role of the board (e.g.,
board size, independence, gender, age, board meetings, and tenure). The results of
this study will provide evidence regarding multiple statements made in existing
literature and will confirm or disprove them. Which will provide useful information
for future regulations regarding the board’s composition. The idea that a large
percentage of women will contribute to the monitoring performance will be proven
non-significant for example.
Acknowledgments
After almost six months of hard work, I can proudly present to you my master
thesis. The process leading on to this moments hasn’t been easy, since I’ve
experience multiple setbacks. The data collections was especially tedious and
troublesome. But overall, the time flew by and I can honestly say that I’m proud to
present to you my finished work.
I would like to thank my first and second readers, Marcel Bergervoet and Kristina
Linke, since both of them gave me needed feedback and guidance regarding the
thesis. Furthermore I would like to thank the people of EY Utrecht, who made it
possible for me to write my thesis there. Last but certainly not least, I want to thank
my girlfriend, parents and friends for their support and advice during the last couple
of months.
Utrecht
July 2013
Table of contents
1
Introduction
p.4
2
Theoretical framework
p.6
2.1
Role of the board
p.8
2.2 Fraud Triangle
p.9
3
Hypotheses Development
p.12
4
Research Design
p.16
4.1 Data Sample
p.17
4.2 Control variables
p.19
5
Empirical Results
p.20
6
Discussion and future research
p.31
1
Introduction
There have been multiple, global, accounting- and fraud scandals in the past twenty years. Investors lost their trust in the financial reports and organizations which appeared to be healthy were declared bankrupt within months. Organizations like Enron, Parmalat, Worldcom, Refco and Ahold created an enormous demand for more transparency and better information from society. Those major financial frauds caused a massive loss of trust under stakeholders, in the integrity of US organizations (Carson, 2003). It contributed to stockholder loss, image- and legal problems for organizations and even bankruptcies (Beasley and Jenkins, 2003).
The fraud scandals showed multiple failures in the existing risk management, since organizations apparently could not prevent fraud from happening. And as a reaction, the attention for corporate governance has grown among academics and policymakers ever since (Bebchuck and Spamann, 2010; Blundell-Wignall et al., 2008; Cohen et al., 2002; Cornett et al., 2010; Hoitash et al., 2009; Pirson and Turnbull, 2011).
Authorities have also reacted to the apparent deficiencies. Several initiatives have been taken based on the assumption that financial reporting is related to corporate governance. Initiatives like the Sarbanes-Oxley Act and the already implemented ERM model provided by COSO have been developed to help organizations with the reliability of their internal control and to constrain the potential risk of fraudulent financial reporting. As mentioned by the Securities and Exchange Commission about the Sarbanes-Oxley Act, “The Act mandated a number of reforms to enhance corporate responsibility, enhance financial disclosures and combat corporate and accounting fraud...” (www.sec.gov). So they try to ensure that the investors can rely on the reporting of the organization on which they base their decisions. After the introduction of the Sarbanes-Oxley Act, most publicly held organizations tried to comply with the new regulation and hired more independent and/or female directors for their board (Dalton and Dalton, 2010). The Sarbanes-Oxley Act of 2002 contained some regulations concerning memberships and functions of corporate boards. For instance, the board of directors of public companies have to consist of mostly independent directors. And the audit-, corporate governance- and compensation committees may only consist of independent directors. Linck et al. (2009) found that public companies generally hired extra outside directors, just to meet those requirements. And a study of Dalton and Dalton (2010) concluded that women are taking significantly more presence in the boards after the enactment of the Sabanes-Oxley Act in 2002, although it does not directly addresses the presence of both genders in corporate boards. The Act does require the CEO to approve the financial statements, and so, makes the CEO responsible for any misrepresentations. Various scholars and organizations have studied the issue of fraud. Recent examples are studies by the Committee of Sponsoring Organizations of the Treadway Commission (COSO) and the Association of Certified Fraud Examiners (ACFE). The second has released the ‘Report to the Nations on Occupational Fraud and Abuse’ in which they have stated that financial reporting fraud is extremely costly. Although financial reporting frauds weren’t the most common form of fraud, less than five percent of all fraud cases, they were by far the most expensive, with a median loss of approximately $4 million per case. In the study performed by COSO, they have found that of the techniques used for fraud in the financial reports, the most common techniques
where the overstatement of assets (51%) and improper recognition of revenues (61%). Uzun et al. (2004) examined how board characteristics and other governance features could affect the presence of corporate fraud during 1978 until 2001 in the U.S. And found that the board’s composition is significantly correlated to the presence of corporate fraud. However, unlike Uzun et al.(2004), this study is only focussed on the involvement of the CEO in financial fraud and not the occurrence of corporate fraud as a whole.
Research regarding corporate governance assumes the relation between governance structures and the quality of earnings and their disclosure. Various authors have stated corporate governance as a cause of financial statement fraud (Karamanou and Vafeas, 2005; Farber, 2005; Larcker et al., 2007). Castellano and Lightle (2005) as well as Hunton et al. (2011) have considered the tone at the top as one of the possible elements within the organization which can be of influence on the organizational climate. The role of the CEO has also been seen as an important factor with influence on the ethical values within the organisation (Efendi et al., 2007; Zhang et al., 2008). And since the board plays an important role as monitor of the organizations’ management, is it a part of the corporate governance mechanisms (Cohen et al., 2002). Therefore is this study attempting to find relationships between the board’s characteristics and that monitoring role. The CEO will be the most likely to be of influence on the tone at the top since he or she has the most powerful position within the organization. And the CEO will also have an incentive to commit fraud, because investors will reward organizations that meet expected earnings, even if they are most likely resulting from earnings management (Bartov et al., 2002). It’s even proven that a CEO’s remuneration goes up, when the performance, on paper, goes up (Ozkan, 2011). So a CEO has the incentive and power to influence employees to commit fraudulent actions.
As a reaction on ia. the mentioned scandals, regulation has become more strict, but has it the desired effect? The Sarbanes-Oxley Act, for example, requires a majority of independent directors within the board. But does the new composition of the board allow them to control top management in a better way? Or can a CEO still do how he or she likes without much hindrance of the board? This study will try to find an answer to those questions.
An important difference between previous research and the current study is the fact that the data that will be used in this research comes from a period of financial distress, which is often not the case in the existing literature. The sample used in similar studies by Adams and Ferreira (2008) and Abbott et al. (2009) for example, stopped in 2003. And that of Uzun et al. (2004) stopped as early as 2001. The data sample of this study has been taken from the first year of the current financial crisis, 2007, for a period of six years.
During the ongoing financial crisis, there are still companies which are on the edge of bankruptcy and who show signs of fraudulent behaviour. As a reaction, the Nederlandse Beroepsorganisatie van Accountants (NBA) launched an alert in November 2012, which is directed to make auditors more aware of the possibility of financial reporting fraud. Berry Wammes, CEO of the NBA has explained that they have done so, because of the effect that the financial crisis has on the
reasoning of organizations. When organizations are in distress, they will use more, sometimes even illegal methods to keep the organization alive. Methods like using different valuation methods, but also simply falsify numbers, which is qualified as financial reporting fraud.
The purpose of this study is to see if the stricter legislation does have the desired effect. Namely, a better corporate governance and so the prevention of fraudulent activities by top management. Is the CEO still involved in financial statement fraud or not? And what period better to collect the data from than a crisis, where management has the most incentives to circumvent the rules. This study will contribute to the existing corporate governance literature, because it provides a better understanding of the factors that contribute to financial statement fraud and it will provide information on the effectiveness of the implied rules regarding the board of directors to prevent fraud. Secondly, this research will contribute to existing literature about financial reporting quality, since financial statements which contain fraud do not properly reflect the actual status of the organization. Thirdly, this research also contributes to the agency theory, since it provides empirical evidence about board of director characteristics which will or can influence the actions of management, and so can influence the agency costs.
With the gained knowledge about the composition of the board, there may be relationships to be found between the board’s characteristics and the involvement of the CEO in financial statement fraud. Which can be a step closer to understanding the reasons behind, among other things, financial statement fraud, the board’s monitoring performance and its control over the CEO. This study could, as such, contribute to future regulations.
In order to achieve the results that will contribute to the existing literature as mentioned above, this study will investigate if the board of directors can control the actions of the CEO, and prevent him or her from committing fraudulent behaviour. Furthermore, this study will try to find the specific characteristics which can take the credit for it. And so, it could provide evidence that shows that the established regulations are aimed at the wrong characteristics. The main question that is asked during this study is therefore:
Which characteristics of the board of directors can influence the involvement of the CEO in fraud regarding the financial statements?
The remainder of this paper is organized as follows. Firstly, I will provide the scientific contribution that this study delivers. In section 3, the hypotheses will be formed. And after that there will be an explanation of the used methods and statistics. The outcomes of the study are discussed in the Results section. A summary and conclusion will be given in the final section.
2
Theoretical Framework
In the past there have been several studies on the influence of board characteristics. Numerous studies have investigated the relation between board characteristics and performance (Mahadeo et al. 2012; Chiang, 2005; Srinidhi et al., 2011; Boo and Sharma, 2008, Jensen, 1993). Chiang (2005) for example, found that the size of the board is of positive influence on the performance of the organization, Srinidhi et al. (2011) have examined the effects of females on the board and
Jensen (1993) argued that the CEO could control the board when he also is the chairman. And Boo and Sharma (2008) suggested that an independent director can monitor more effectively because of the lack of emotional connection with top management. All these studies have focused on the board’s influence on performance. But there has been, however, considerably less (to none) research on the influence of board characteristics on the CEO involvement in fraud.
Dechow et al. (1996) examined firms which were subject to Accounting and Audit Enforcement Releases by the Securities Exchange Commission during the period of 1982 until 1992, they focused on earnings management and found evidence that firms which are manipulating earnings, a financial reporting fraud method, often have less independent board members and more often a CEO who is also chairman of the board (Dechow et al., 1996). This is in line with the findings of Jensen (1993) and Farber (2005). The results from Farber (2005), Beasley (1996) and Dechow et al. (1996) also show that a small percentage of outside board members is associated with fraud organizations. Just like this study, the study of Beasley (1996) revolved about financial statement fraud and board characteristics. And in his study he also found evidence that the risk of financial statement fraud decreases as the number of directorships in other firms declines. These results are consistent with the idea that independent governance mechanisms are more effective in monitoring the top management and show some implications of differences in board composition. But Beasley (1996) only focussed on the percentage of outside members within the board during the period 1980 – 1991. In his study Beasley (1996) gives multiple characteristics that need to be investigated. He mentions, among others, that the number of meetings can be of influence on the relation between the board and financial statement fraud. Some of the variables mentioned by Beasley (1996) will be included in this study.
Literature regarding the role and effectiveness of the board is extensive. As shown, empirical evidence has provided the influence and importance of board characteristics (Beasley, 1996; Dechow et al., 1996; Chiang, 2005). However, most of the existing literature overlap regarding their variables. Beasley (1996), Dechow et al. (1996) and Uzun et al. (2004) for example all focused on the independence aspect of directors. Also, none of those studies were directed at the characteristics of the board in relation to the CEO’s involvement in financial reporting fraud, but at the combination of corporate governance and fraud, fraud and earnings management, or board characteristics in relation to performance or fraud as a whole (Dechow et al. 1996; Bowen et al. 2005; Uzun et al. 2004).
Furthermore, the results from all of those studies came from periods without the presence of a crisis, which is the case in the period examined by this study. The data that will be used to from my conclusions comes from the period 2007 until 2012. So with this study I will try to find trends in the board of directors during a period of financial crisis, when management will be most likely tempted to participate in fraudulent conduct. Since the last major financial crisis was almost a hundred years ago, little is known about organizations in such a situation.
The purpose of this study is to give an insight into the characteristics of the boards within organizations where financials statement fraud has been detected, so to see if there are possible striking points to be found. And more specifically, to see if there are board characteristics which ensure better control over the CEO’s actions. Because a CEO who is involved in fraudulent actions is the best and biggest example of failure by the board in its monitoring function.
This study contributes to existing research that implies the idea that boards are an important mechanisms for corporate governance that can contribute to an organization’s well-being (Zang, 2010; McNulty et al., 2013). It will complement the empirical evidence provided by i.a. Beasley (1996), Dechow et al. (1996), Chiang (2005), Farber (2005) and Uzun et al. (2004) and provide a better insight in the board characteristics that can be found at organizations where financial statement fraud has been detected. Unlike Uzun et al. (2004) who have used media coverage and focused on corporate fraud in relation to among others, the independence and size of the board until the year 2001, I will use Accounting and Audit Enforcement Releases made by the Securities Exchange Commission to directly examine the relation between the influence of board’s characteristics on the involvement of the CEO in financial reporting fraud. Because of the improved insight of board characteristics that will be obtained by this study, future regulations can be aimed better at creating an improved organizational environment where fraudulent behaviour will be prevented. And so, the outcomes of this research can be used in future regulations that are specifically aimed at the limitation of fraudulent behaviour, but also for improvement in overall corporate governance and the reduction of agency costs.
The dependent variable in this study is the involvement of het CEO in financial reporting fraud, which will function as a proxy for the monitoring performance of the board. And in this study, I will try to find the characteristics of the board that could have had an influence on the decision of the CEO to commit fraud in the financial reporting. From this point forward, the relation between the board and the CEO’s decision to conduct fraudulent behaviour, will be explained using the fraud triangle. Subsequently, the variables will be discussed and thereafter the hypotheses are formed and given direction when possible. These hypotheses show the expected influence of the board’s characteristics on performance of the board in their role to limit and control the actions of the CEO regarding financial statement fraud. But first I will explain the role of the board.
2.1
Role of the Board
The role of the board can be seen from an agency theory perspective. In their agency framework, Fama and Jensen (1983) gave an explanation of the board’s role. The board has the task to control the top management and to make sure that their actions are in line with the preferences of the shareholders, it has the task to monitor the top management at the service of the shareholders and to resolve conflicts between managers and stakeholders (Ahmed and Duellman, 2007; Garcia Lara, 2009; Monitoring Commissie Corporate Governance Code, 2009). Several studies, (Fama, 1980; Fama and Jensen, 1983) have theorized that the highest internal control mechanism that is responsible for monitoring top management behaviour, is the board. Some of the board’s characteristics can have consequences for financial reporting fraud as shown by several studies (Farber, 2005; Dechow et al., 1996). In recent studies, Lewellyn and Muller-Kahle (2011; 2012) imply that the characteristics of the board had a negative impact on their risk management decisions, and thus have a direct impact on the risk of financial fraud. Because of this direct relation between the board and financial fraud, this paper will examine whether there are remarkable board characteristics to be found at organizations which have been identified as
fraud-organizations. This will provide a better picture of the characteristics of the board that will inhibit fraudulent behaviour by top management. As Albrecht and Albrecht (2003) concluded that effective control is the best way to reduce fraudulent acts.
As mentioned above, there have been several events of financial reporting fraud in the last twenty years. Those events where the cause of a loss of trust in the financial statements of organizations. Stakeholders lost trust in financial statements (Carson, 2003) and therefore had trouble making further judgements. Financial reporting fraud usually has a great impact and is of great importance to investors, external auditors, and others stakeholders, since it shows that the management scams the stakeholders by improving the condition of the organization in an artificial way. In a recent study, Beasley et al. (2010) requested for more investigation regarding this subject: “.. to better understand many
of the underlying factors likely to affect the prevention, deterrence, and detection of fraudulent financial reporting.” This study seeks to fulfill this need partly.
2.2 Fraud Triangle
The fraud triangle (Cressey, 1953) describes the three factors that must be present for any form of fraud. A person has to have a motivation to commit the
action, he/she has to think that it will be fine (rationalisation) and there has to be an opportunity for him/her to do it (Bell and Carcello, 2000). In their study in 2003, Albrecht and Albrecht looked at possible factors that could increase the opportunity to commit fraud and they note that possibly the most important way to minimize fraudulent acts is to have an effective control structure (Albrecht and Albrecht, 2003). This idea is in line with the empirical evidence provided by Dechow et al., (1996) demonstrating that poor corporate governance is related to financial fraud. As said before, the board has the task to control the top management and to make sure that their actions are in line with the preferences of the shareholders (Monitoring Commisie Corporate Governance Code, 2009). So, when the board fails to fulfil its duty of controlling and constraining the top managers, there is a sign of poor governance, and a CEO can see an opportunity to commit fraudulent actions. This is in line with the findings of Beasley (1996) that several board characteristics enable the board to monitor the management more effective. This is also consistent with the assumption of Uzun et al. (2004) that a board that is more effective at its monitoring role, will be better in preventing fraudulent behaviour, and that therefore, there will be poor performing boards found in organizations where fraud is detected. And so it is important to investigate if there is evidence from practice for several theories of board characteristics that are expected to be of influence on the performance of the board in their monitoring role, and that is exactly what this study does. This study will look at the involvement of the CEO in financial statement fraud, and so it will look at the effect of several board characteristics in restraining the top management from committing fraudulent behaviour.
To investigate the impact of various board characteristics, this study will examine organizations were financial statement fraud has been found. Therefore, the hypotheses will have to be based on the board’s characteristics expected to be found within those organizations.
I will now briefly describe the board characteristics that will be examined in this study.
The board of directors has been a subject of multiple studies. There has been various research, for example, on the influence of independent directors. Dechow et al. (1996) provided evidence that earnings management was more prevalent in organizations with a low percentage of independent directors. Because of the increased performance of the board of directors and the lower emotional bonding with the CEO, I expect to find a negative relation between the percentage of independent directors and financial statement fraud carried out by the CEO. Since a better performing board will control the actions of the CEO in a better way and thus will prevent them from committing fraud.
When the CEO is also the chairman of the board of directors, he or she can steer their attention away from the fraudulent actions, since the chairman directs the agenda of the board. So a CEO as chairman weakens the monitoring performance of the board in theory and in practice (Tuggle et al., 2010). It will be logical therefore, to consider whether the CEO is also the chairman or not. This will be further elaborated in the hypotheses section.
Another characteristic that is the subject in a great amount of literature is the size of an object. Whether it is the size of an organization, a government or a group of people, makes no difference, they have all been used in studies. In that way, this study is no different. However, this study tries to find a direct relationship between the size of the board of directors and its influence on the behaviour of the CEO, scilicet the decision to commit financial statement fraud. In that way, this study is almost unique. Since multiple scholars found improved performance with a larger board size (Chiang, 2005; Raheja, 2005), I expect to find a negative relation between the size of the board and CEO’s involvement in financial statement fraud.
The third characteristic to be included in this research is the number of meetings that the board of directors have within a fiscal year. I expect to find that the board of an organization with financial statement fraud, have not had a great amount of meetings since Epstein and Roy (2004) state that the time that a director spends on the organization is one of the key success-factors. And when the number of meetings is low, the directors don’t spend that much time on the organization. Because of the little time they spend on the organization, the risk of not detecting fraudulent behaviour by the CEO becomes bigger and thus will it create an opportunity for the CEO. Therefore, I expect to find a negative relationship between the board’s meetings in a year and the involvement of the CEO in financial statement fraud.
The following three independent variables are formed from the same theory. All three expectations are made on the idea that heterogeneous groups will make better decisions that homogeneous groups, since they are less limited to one way of thinking. When asked to form a subjective conclusion, a group which consist only of men, will come to a different conclusion than a group of women. And a group consisting of men and women will come to yet another
conclusion. In this paper I assume that in groups where group members have less in common, there will be more conflicting views, what will lead to a broader vision of reality. And that therefore, a group consisting of members with different tenures, members from different age groups, and groups existing out of more than one gender, will outperform homogeneous groups. So I expect to find a negative relation between the deviation within the board of directors and the CEO’s involvement in fraud within the financial statements. Because a better performing board will limit a CEO in their actions more than their underperforming peers.
And the final characteristic that will be examined within this paper is the number of similar secondary activities. The National Association of Corporate Directors’ Blue Ribbon Commission on Director Professionalism gave the recommendation that board members should limit their number of tasks, to make sure that they have enough time available for an organization. Which is in line with the statement of Epstein and Roy (2004) mentioned above. And therefore I expect that a board which consists of members who all have multiple other similar activities, will not perform as good as their less busy equivalents and therefore give the CEO a opportunity to perform fraudulent acts. Each of these variables will be explained more in detail in section 3.
- -
- -
- + + -
The hypotheses will be formed in the following section. Whereupon I will discuss the research design, evaluate the results, summarize those into a conclusion and give recommendations for future research.
CEO involved in fraud. Independent Directors. Board Size. Number of Meetings. Female Members. Age Diversification. Tenure Diversification. Number of different Boards. CEO as Chairman.
3
Hypotheses development
CEO duality
The board of directors have to monitor the actions of the CEO to ensure that he or she acts in favour of the shareholders’ interests. But what if the CEO is part of the board? There is CEO duality when the CEO is also the Chairman of the board (Baliga et al., 1996). After searching the power that a CEO can have to influence the board members, a study of Tuggle et al. (2010) has shown that the presence of the CEO in the board has a negative effect on their monitoring function. When a CEO takes place within the board that is monitoring him, he can influence the behaviour and decisions of the rest of the board members, especially when he is the chairman who can influence the agenda. By influencing the discussed topics as the chairman, the CEO can divert the attention away from the fraudulent activities. Prior research from Davidson et al. (2004) has shown that duality indeed affects earnings management. Moreover, multiple studies have suggested that CEO duality weakens the board and gives the CEO more power (Cannella and Lubatkin, 1993; Finkelstein and D’Aveni, 1994). An article by Masters (2009) listed the percentage of new CEOs who are also being appointed as Chairman, which was at that time just 18 percent. And since the studies of multiple scholars have showed the negative impact of CEO duality on the monitoring role of the board and earnings management (e.g. Davidson et al, 2004; Tuggle et al., 2010), a high percentage in duality in the data sample of fraud organisations should therefore provide a valid (practical) argument to investigate the connection between CEO duality and his or her fraudulent actions. Therefore, I expect the following:
H1. The presence of CEO duality has a positive effect on the involvement of the CEO in fraud, in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Independent directors
Boo and Sharma (2008) suggested, as several have done, that independent directors are more effective monitors, since they don’t have further interest in the company other than a possible way to enhance their image and price on the job market. In their study from 1996, Dechow et al. investigated firms which have been accused of earnings management and found evidence that firms which manipulated earnings often had less independent board members. Evidence from India has found a positive but insignificant relation between the percentage of independent board members and the performance of the firm (Kumar and Singh, 2012). The study from Uzun et al. (2004) provides evidence that organizations that have been identified as fraud-organizations have a significantly lower percentage of independent directors in their board than non-fraud organisations have. Despite the lack of literature that explains the presence of fraud directly by the percentage of independent members; Applying these results to this study, I expect to find a low percentage of CEO involvement in the fraud. Because of the better performance which is expected of an independent board, what should limit the CEO in its actions and prevent him or
her from committing fraudulent actions. Since it will be difficult to identify all relationships of the board’s members with the organization, the independence definition will be used as defined by Rule 5605(a)(2) of the Listing Rules of the NASDAQ Stock Market, Inc. H2. The percentage of independent directors is significantly small, in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Board Size
Jensen (1993) is one of the many scholars that have mentioned the impact of the board’s size. In the study of Uzun et al. (2004) they have found a positive relation between the size of the board and fraud organizations, but there was a lack of significance. And Beasley (1996) did not find support for his expectation to find a larger board in fraud organisations. And as mentioned before, empirical evidence has shown that good corporate governance and effective control can reduce fraudulent acts (Albrecht and Albrecht, 2003; Dechow et al., 1996). In his study in 2005, Raheja found using a model, that the size of the board has influence on its effectiveness. Although there are different findings on the influence of board size on organization performance, most of them find a positive relationship (i.e. Chiang, 2005). Anderson et al. (2004) found a positive relationship between the monitoring performance of the board and its size. Because of those results, I expect a bigger board, in number of directors, to be better in monitoring and controlling the actions of the top management en thus could have prevented them from committing fraudulent acts.
Therefore the hypothesis is as follows:
H3. The size of the board is significantly small in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Number of meetings
The number of times the board sits down to discuss the organization could affect the monitoring ability. When a board has to few meetings, it can be a signal that the directors do not have the incentives to execute their role properly, it can be a sign of laxity. Yatim et al. (2006) provide evidence that boards that meet frequently have a bigger chance of performing their role effectively. As mentioned by Epstein and Roy (2004), the time that a director spends on the organization and the board meetings, is one of the key success-factors. It can be hard to get the right information from the CEO and challenge him or her, so there has to be enough meeting time to be able to do just that (Epstein and Roy; 2004). The monitoring role of the board is related to financial statement fraud as Dechow et al. have showed (1996). Furthermore, empirical evidence have found a positive relation between the number of meetings and fraud organizations, but it has not been found to be significant (Uzun et al., 2004). Notwithstanding the fact that there is, to my knowledge, no significant relationship found between the number of meetings and the board’s
monitoring performance, I expect that an increase in meetings could have had a positive effect on detecting fraudulent behaviour from the CEO.
H4. The number of meetings per year have been significantly low, in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Gender diversity
Several governments have installed a minimum percentage for females in top management positions. The Dutch government has passed a law back in 2011 for example, stating that by 2016, 30 percent of the board of directors and the top management should consist of women. Empirical evidence has provided that in the US, a board with a larger number of females have a better oversight function and that it results in an increase in earnings quality (Srinidhi et al., 2011). Krishnan and Parsons (2008) further have found that a higher gender diversity in senior management will result in an increase in earnings quality. Empirical evidence also finds that women often serve in boards in organizations with better performances (Farrell and Hersch, 2005). Adams and Ferreira (2008) investigated the impact of females on the board in the US and found evidence that suggests that the presence of women in the board of directors is positively related to its monitoring performance. Because of the showed results the presence of women has on earnings quality, organizational performance and the expected impact on the monitoring role, I expect the improved functioning of the board will be noticeable in the monitoring role and therefore be more effective in preventing fraudulent behaviour.
H5. There is a high percentage of men in the board of directors in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Director age
The age of directors is a subject that has been discussed extensively. The current view of a director is that he is an ex-manager who now sits in multiple boards and enjoys his retirement (Kang et al. 2007). Adding diversity of age in the board can have its advantages. Older directors bring experience, financial resources and a large network. While younger directors have the motivation and energy to work towards the future (Kang et al. 2007). And the middle group occupies the positions which carry the biggest responsibilities (Kang et al. 2007). In his study, Murray (1989) develops the idea that a homogeneous board, low age diversity, is able to operate in a more efficient manner, in a stable environment. While a heterogeneous group is able to adapt better to a changing environment. Houle (1990) on the other hand argues that age diversity is needed, even in a limited age range. Huse and Rindova (2001) found that various stakeholders have different expectations of the board’s role. They argue that it is therefore important to understand the different stakeholders perspectives when they look at what is best for the organization (Huse and Rindova, 2001) and therefore it may be beneficial when directors of the
board are of different ages, since the stakeholders will be from different ages. Notwithstanding the non-confirmatory results of Murray (1989), I do have the expectation that a larger diversity in age could have resulted in a better fulfilment of the monitor role by the board and thus a limited possibility for the CEO to commit fraud.
H6. In an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud, there has been little difference, hence a small deviation, between the ages of board members.
Tenure
The difference in the number of years of experience between directors could be of importance. The degree in which directors identify themselves as part of the organization, what can be affected by tenure, can play a role in the cooperation between board members. A director with a long tenure may not share information with short tenure directors or he or she can retrieve more that he or she contributed when there is more affection with customers than with the organization (Hillman et al., 2008). In a recent study of Tuggle et al. (2010) there is found evidence that a board member with a longer tenure pays more attention to organizational issues, so a directors years with the organization and his attention to problems regarding the organization grow simultaneously. The results of Huse and Rindova (2001) show that it is of importance to understand the expectations of the stakeholders and a director with a longer tenure has more experience with the organizations’ stakeholders so he or she knows what they expect. To sum-marise, I expect that a director with more tenure can have more experience with organization specific problems and can make a better judgement regarding the actions that have to be taken by the top management. But I also expect that a board consisting of only long tenure members could fail to see the problem from different angles and because of that there also have to be short tenure directors within the board to view problems from a different angle. In summary, I expect that a more diverse mix of short- and long tenure directors in the same board could be of a positive influence on the monitoring role and that therefore, within the examined fraud-organizations there can be found a high percentage of boards with a low deviation of tenure where the CEO has been involved. Therefore the hypothesis is as follows:
H7. In an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud., there has been a small variation of tenure years between board members.
Multiple boards
The business of directors has been a considered subject in the past. The National Association of Corporate Directors’ Blue Ribbon Commission on Director Professionalism (NACD, 1996) gave recommendations regarding the board. They recommended that directors should limit their memberships in organizational boards. But it evidence has shown that “overboarded” directors can have positive outcomes for the performance of the board in the study of Harris and Shimizu (2004). Overboarded directors are those who are director in multiple other boards (Harris and
Shimizu, 2004). Harris and Shimizu also found that overboarded directors do not miss significantly more board meetings that not-overboarded directors, which often has been thought (NACD, 1996; Ferry International, 1998). A possible explanation for the increase in performance when overboarded members are present, can be that directors who act in better performing boards are offered more seats in other board (Rerris et al., 2003). Furthermore, Rerris et al. (2003) also found no evidence for a relation between securities fraud and overboarded directors. Nevertheless, I predict that the number of directorships per board member will have a negative effect on the performance and so can deliver a less effective monitoring role. Since a director with multiple similar tasks, could mistake one organization for another more easily than a director with only one or two boards,
H8. The directorships per board member has been high in an organization where financial statement fraud is detected during the period 2007-2012 and where the CEO is involved in the fraud.
Audit committee
Because the New York Stock Exchange introduced the requirement of an audit committee com-posed of independent directors in 1978, I won’t consider those committees, since multiple organisations within the sample will have an audit committee because it is required by law. That also means that I won’t consider the presence of financial experts required by the Sarbanes-Oxley Act.
4
Research Design
The cases of alleged financial statement fraud were collected to study the involvement of CEO’s in financial statement fraud after stricter regulations as the Sarbanes-Oxley act. Since there is little to nothing known about the involvement of the CEO in financial statement fraud, this study will be of an explorative nature. Explorative researches can be performed by qualitative as well as quantitative research approaches. However, since prior research concerning financial statement fraud mostly consist of quantitative research, a quantitative approach seems suitable for this study. This will have consequences for the research design. The variables which are included in this study are based on the results of prior studies. Jensen (1993) has, for example, mentioned the impact of a CEO as chairman, Boo and Sharma (2008), suggested an independent director to be more effective and Adams and Ferreira (2008) investigated the relation between gender diversity within the board and their monitoring performance.
The decisions regarding the archival data sources will be further elaborated in the following paragraph. The sample used to test the expected relations consists of 120 cases. All of those cases are collected from AAER’s.
4.1 Data sample
As I have mentioned above, is the existing literature on financial statement fraud mainly based on databases consisting of archival data (Beasley et al. 1996, 1999; Bonner et al. 1998, Linke, 2012), and therefore it seemed also very suitable for this research design, since it provides practical data. Because of the availability of data from the US, I have chosen to direct my research not on Dutch organizations, but organizations on which the SEC has filled an litigation, which are listed on a US stock exchange. All data used in this study is therefore hand collected from the SEC website (www.sec.gov). I am aware of the fact that there are multiple cultural, legal and governance differences between the U.S. and European countries (Hofstede, 1980; Hopt and Leyens, 2004;Burgstahler et al., 2006). But findings resulting from research conducted on US data, could also be useful in an European setting, since a comparison of fraud cases between Ahold and Enron suggests that the methods and preceding process are not so different (Linke, 2012).
This research only uses cases of financial statement fraud reported by the SEC. The SEC forms Accounting and Auditing Enforcement Releases (AAER’s) which, in this study, will be used as a proxy for fraudulent behaviour regarding the financial reporting, consistent with the researches of Beasley (1996), Bonner et al. (1998), Farber (2005) and Linke (2012). This data may consist a bias due to several reasons. One is that the accused often admit no guilt and secondly, that the SEC may not use consistent enforcement methods. Nevertheless, as Farber (2005) states, the AAERs are suitable as a proxy for financial statement fraud.
The use of AAERs in this study is appropriate, since the persons accused to be involved in the crime are always mentioned in the AAER. It can therefore be used to find out if the CEO is accused of being involved in the fraudulent activities or not.
The definition of financial statement fraud that has been used during this study is that of Beasley et al. (1999): “... the intentional material misstatement of financial statements or financial disclosures or the perpetration of an illegal act that has a material direct effect on the financial statements or financial disclosures”.
Used cases have been accused of violating the regulations regarding anti-fraud, periodic reporting, recordkeeping, and internal control measures included in sections 17(a) of the Securities Act of 1933 and sections 10(b), 13(a) and 13(b)(5) of the Securities Exchange Act of 1934, which is conform the included sections used in the similar study of Linke (2012). The used definition of Beasley et al. (1999) specifically describes cases that violate the mentioned regu-lations, but excludes cases which do not concern the manipulation or misrepresentation of financial statements. Insider trading for example, does not involve financial statement fraud, but does violate the mentioned laws. So cases of insider trading have not been included in the current sample.
The cases that have been collected contain allegations of financial statement fraud which occurred in the period January 1, 2007 until December 31, 2012. Fraud periods of less than 3 months in a year have not been taken into account.
Since the independence of a board member would be difficult to establish, I have used the independence definition as defined by Rule 5605(a)(2) of the Listing Rules of the NASDAQ Stock Market, Inc. The number of members who comply with this rule is named in every annual DEF-14A form, so there is no interpretation problem.
Regarding the number of extra boards in which a board member has taken place, I have looked at similar boards, because of the ease with which the relevant organizations can be confused. Therefore, I selected memberships in the following boards: board of trustees, board of directors and supervisory boards.
In the event of a switch of board members during the year, the board (member) with the most number of months that given year has been taken. This principle has also been used with the tenure of the board members, since a member has to have taken seat at least 6 months, to be counted as a tenure year.
The information regarding the organization and board characteristics that has been used, has been taken from the EDGAR database (www.sec.gov./edgar/searchedgar/webusers.htm). Information regarding the board of directors and the organizations have been collected by the use of either 8-K, 10-K or DEF-14A forms, or simply internet research by using the search-engine Google.com. Selected by using the mentioned criteria, a sample of 120 cases have been collected from the period to test the developed hypotheses. January 1, 2007 has been chosen as the starting date, since 2007 was the first year of the current global financial crisis.
The used population fully consists of fraud organizations. In a study of Palepu (1986) he has stated that the correction for a possible bias in a constant term is only needed when you try to develop a predictive theory. Since this study isn’t aimed at developing a predictive theory, there will be no consequences of a possible bias in the constant term and so the method of logit regression can be used to test the developed hypotheses. To test the influence of characteristics of the board on the involvement of the CEO in fraud in the financial reporting in the US during the period January 1, 2007 - December 31, 2012, I have used the logit cross-sectional regression model stated below.
CEO INVOLVEMENTi = α+ β1DUALITY+β2%INDEPENDENTі+β3SIZEі+ β4MEETINGSі+Β5%GENDERі+β6AGEі+β7TENUREі+β8MULTIPLEі+ β9ROAMUTATIONi+β10%OWNEi+β11BLOCKi+εі
Where:
i = Case 1 through 120.
CEO = A dummy variable, an organization is given a 1 when it’s CEO is INVOLVEMENT accused in an AAER or a 0 when he isn’t.
DUALITY = A dummy variable, an organization is given a 1 when the CEO is also the Chairman of the board, and a 0 when that is not the case.
% INDEPENDENT = The percentage of independent members within the board. SIZE = The size of the board measured by the number of directors. MEETINGS = The number of board meetings per calendar year.
% GENDER = The number of male members divided by the total number of board members.
AGE = Measured as the deviation of age within the board.
TENURE = Measured as the deviation of years of experience within that board. MULTIPLE = The mean of the total of comparable boards outside the organization where the director is a member of.
ROA MUTATION = The difference between ROA from the present and from the previous year.
OWNE = Total of common stock held by the directors divided by total common stock placed.
BLOCK = The total cumulative percentage of outstanding shares owned by blockholders.
ε = The residual.
4.2 Control variables
Since the dependent and independent variables already have been discussed extensively, I do not find it necessary to discuss them again. However, the control variables have not been discussed earlier in this paper. And therefore I will discuss them in the following section. The control variables have been taken into account, since previous literature has mentioned their impact on fraud and/or board performance.
Achievements (ACHIV)
A recent study of Tuggle et al. (2010) has proven that a board devotes less time in its monitoring and controlling role when the organizations has achieved good performance in ongoing years. Especially a positive deviation from the results in the year before will result in a decline of paid
attention from the board. Therefore, as a control, I will measure the difference between performance from the year before the first fraud year and the performance at the end of the year. In compliance with the research of Tuggle et al. (2010), I will measure performance on the basis of Return On Assets (ROA). The ROA is measured by dividing Earnings Before Interest and Taxes (EBIT) by the Total Assets of the organization.
Ownership (OWNE)
According to the agency theory of Fama and Jensen (1983), the ownership of common shares by the directors should have a positive effect on their performance, since they will benefit personally from a better performing organization. But, it could also provide incentives for manipulating earnings to boost the stock prices in an artificial way. Loebbecke et al. (1989) state that ownership of organization stocks can be a key incentive for fraudulent behaviour. Therefore, one of the control variables will be the differences in ownership of common stock by the board members, in accordance with the study of Beasley (1996).
Blockholders (BLOCK)
The last control variable used by Beasley (1996) is the extent to which large blockholders, who own at least 5 percent of the organization, are present amongst the owners. Jensen (1993) mentions that large investors are important, since they are independent enough to see the organizations management in an unbiased way, and also have the financial incentives. According to Jensen (1993) , an active investor will be “important to a well-functioning governance system”. Concentrated ownership has been argued to be critical for effective corporate governance by traditional theories, because they have the resources as well as the incentives to monitor and pressure the management (Edmans and Manso, 2011). Edmans and Manso (2011) showed that multiple small blockholders result in a better reflection of management efforts by stock price. So, managers could want to increase their effort to increase the stock price. Or they could make it seem like they have increased their effort, by adjusting the financial reports.
In the following section the results of the logit regression will be gone through, whereupon a conclusion will be formed and possible shortcomings will be discussed.
5
Empirical Results
Before we take a look at the actual results of the regression, we are first going to overlook the collected data. As can be seen in Figure 5.1, there have been 120 observations entered in SPSS, by which it formed the analyses to accept or reject the formed hypotheses.
To tests the formed hypotheses I had to collect all the necessary data over the organizations. But as I have mentioned in the Data Sample section, not every organization provided their data, which resulted in a sample of 120 cases of which I had almost all necessary data.
The average age of the board members from all collected fraud cases was 56,89 years, however I haven’t made a hypotheses about the average age, but the standard deviation. But an average of almost 57 years seems to confirm the prejudices from the ‘Old Boys Network’. Furthermore consist 62,5% of those respective boards fully of men. This does not seem to confirm the allegations concerning the ‘Old Boys Network’, since it is assumed that almost all boards are formed by men. But when I took a closer look, I found that 90,83% of the boards consisted for 75% or more out men and in total, more than 93% of all board members are male, which do confirm the allegations. But the standard deviation of 8 years does speak against the thought of the ‘Old Boys Network’, since it means that the board consists of persons which can differ as far as 8 years older or younger that the average. So there can be an age difference of 16 years within the board, which is not in line with the assumption that all board members are about the same age.
Of all 120 cases, the CEO was involved in the fraudulent activities in 43 cases, which is just under 36%. This result is very remarkable, because it is almost the exact opposite of the existing literature. Beasley et al. (2010) found for example, that the CEO was in some way involved in the fraud in 89% of their cases. But this was over a period from 1998 until 2007. So there is a possibility that the Sarbanes-Oxley Act of 2002 could have caused this change in CEO involvement, since this study only takes cases from five years or more after the introduction of the regulation. Beasley et al. (2010) used data before and after the implementation of the regulation which does not give a clear view of the working of the act. But a difference of more than 50% is very remarkable and raises questions about the comparability of the used data. In 57% of all cases, was the CEO also the chairman of the board. So it seems that this, maybe, is not a real problem, since the CEO is only involved in a small percentage of the fraud cases.
Table. 5.1 CEO Involvement * CEO Duality Crosstabulation
CEO Duality Total
0 1 CEO Involvement 0 Count 33 41 74 Expected Count 31,6 42,4 74,0
% within CEO Involvement 44,6% 55,4% 100,0%
% within CEO Duality 66,0% 61,2% 63,2%
% of Total 28,2% 35,0% 63,2%
1
Count 17 26 43
Expected Count 18,4 24,6 43,0
% within CEO Involvement 39,5% 60,5% 100,0%
% within CEO Duality 34,0% 38,8% 36,8%
% of Total 14,5% 22,2% 36,8%
Total
Count 50 67 117
The
Sarbanes-Oxley Act
requested that the board should
mostly consist of independent direc-tors. However, as mentioned before, the organizations have generally hired more indepen-dent directors to comply with the new standards (Linck et al., 2009). But it seems that there is still a considerable per-centage of the board members who are in some way depen-dent (almost 34%), so there is still some progress to be made, since the assumption has been made that independent members will lower the degree of earnings management (Dechow et al., 1996). And in this matter, it seems that the adjusted regulations doesn’t fully have the desired effect.
In the hypotheses section, I stated the expectation to find small boards in fraud organizations, based on the argument that a large board may contain more knowledge. And it seems that the average board consists of nearby 7 persons, so that is a small board according to Yermack (1996) and it seems that this hypotheses is therefore going to be accepted. Those boards held, on average, 9 board meetings per year, which is not even once a month. But if you compare it to other boards as that of Volkswagen (6 times in 2012) (volkswagenag.com/...), FedEx (7 during 2012) (sec.gov/...) or Coca-Cola (5 meetings in 2011) (sec.gov/...), you will see that 9 is not a small number of meetings. And the maximal number of board meetings per year within the sample was even 26!
In the Control Variables section, I have stated the expectation that a large percentage of block holders could restrict fraudulent behaviour, since block holders have the means and incentives to monitor the management better than owners of a small percentage of the firm’s stock. And as can be seen in the descriptive statistics, in the sample of fraud organizations, on average, only 26,86% of the firm’s stock is owned by block holders. So about 73% of the firm’s stock is owned by the directors and officers or by small stockholders, which don’t have the means or incentives to monitor the top management and thus do not restrict them in their actions. 23% of it is, on average, owned by the directors and officers. So that means that within the fraud organizations, in general, roughly 50% of the organizations outstanding stock is owned by small investors who do not have the means to monitor or control the top management.
From here on I will reveal the analysis outcomes. This will be done per hypotheses at a time. The control variables to be added in the final analysis of all variables which have shown significance, to see if their impact is influenced by other variables.
Because both the dependent as well as the independent variables are dichotomous, I had to perform a Pearson Chi-Square test, of which the results can be found in Table 5.2.
When we take a look at the outcome for the influence of CEO duality on the involvement of the CEO in the financial statement fraud, the first thing that you can see in Table 5.1 is that of the analysis is based on 117 of the 120 cases, since 3 of them did not consist of all the necessary data. As can be seen in the same table, is that of the 117 used cases in the analysis, in 41 the CEO was also the Chairman, but wasn’t involved in the fraud. The CEO was in just 26 cases the Chairman and also involved in the fraud, and 17 CEO’s were involved but not the Chairman of the Board. % within CEO Involvement 42,7% 57,3% 100,0% % within CEO Duality 100,0% 100,0% 100,0%
So it seems at first sight, that there is a increased chance on CEO involvement in the fraud when he or she is also the chairman/woman, since there were 9 extra CEO’s involved which also where chairman compared to the involved non-chairman CEO’s. The positive relation of ,285 also implies that this statement will be confirmed. And so this contributes to the existing idea that a CEO as chairman weakens the board (Cannella and Lubatkin, 1993; Finkelstein and D’Aveni, 1994). But there is no significant relation as can be seen in Table 5.2. So hypothesis 1 cannot be accepted and there is no significant proof found which confirms the expectation. Which implies that the negative impact on the monitoring role found by Tuggle et al. (2010) apparently does not impact the CEO’s decision to carry out fraudulent behaviour.
The second hypothesis consisted, just as the rest of the following hypotheses, of a dichotomous variable and a consistent variable. The analysis could therefore be performed using a logistic regression. As the results of the regression show in Table 5.3, is the negative relationship between the percentage of independent directors within the board and the involvement of the CEO in the financial statement fraud significant, since the B is a negative number and it is significant on a 5% level because it is not greater than ,05. This is no real surprise, since several studies have already demonstrated that fraud organizations more often have a low percentage of independent directors within the board (Uzun et al. 2004) and multiple studies found that during that period the CEO
was involved in a very large percentage of those fraud cases (Beasley et al., 2010; Linke, 2012). So I, rightfully, expected that a higher percentage of independent directors would do a better job
at preventing the CEO from committing fraudulent behaviour. But there is still a possibility that this relation is not fully explained by these variables, and so there are control variables to test the solidity of the relation.
Table 5.2 Chi-Square Tests (CEO duality)
Value df Asymp. Sig. (2-sided) Exact Sig. (2-sided) Exact Sig. (1-sided) Point Probability Pearson Chi-Square ,285a 1 ,594 ,699 ,368 Continuity Correctionb ,115 1 ,734 Likelihood Ratio ,285 1 ,593 ,699 ,368
Fisher's Exact Test ,699 ,368
Linear-by-Linear
Association ,282
c
1 ,595 ,699 ,368 ,134
N of Valid Cases 117
a. 0 cells (0,0%) have expected count less than 5. The minimum expected count is 18,38. b. Computed only for a 2x2 table
c. The standardized statistic is ,531.
Table 5.3 Percentage of Independent Directors
Table 5.10 contains the outcomes of the control analysis, which has been performed to ensure that the significant relation is not caused by other possible variables, the control variables for example. This hypotheses will not be accepted after all because of the significance level of ,296 of the ‘% Independent’ variable in Table 5.10, which means that the involvement of the CEO can’t be directly explained by the percentage of independent directors within the board. This does not mean that it is not in line with existing literature what states that a low percentage independent directors has been found in fraud organizations (Uzun et al., 2004), since this current research is specifically aimed at the involvement of the CEO and therefore cannot be compared to those results. But these results do raise questions about the existing literature, since the CEO was involved in most of those fraud cases and it seems that there is almost no significant role between the fraudulent actions of the CEO and the board’s characteristics. So are those results about the impact of the independent directors (Dechow et al., 1996) and the financial statement fraud not dependent on other variables? Since this study shows that there is no significant relation between the fraudulent activities of the CEO and the board’s independence.
Regarding the impact of the board size on the involvement of the CEO in the fraudulent actions, there have mostly been positive relations found between the board size and the effectiveness of the board. Anderson et al. (2004) even found that the monitoring role of the board is positively related to the size of the board (Chiang, 2005). I expected therefore that small boards could not limit their CEO in his or her behaviour and thereby giving him or her a better opportunity to commit fraud. The results of the logistic regression to test this hypothesis is stated above in Table 5.4. These results show that there is indeed a negative relation between the board size and the involvement of the CEO, which means that as a board grows in size, the CEO is less likely to be involved in fraudulent activities, just as is predicted in the hypothesis. And so, the third hypothesis can be accepted, because the relation is also significant. These results are therefore in line with those of Anderson et al. (2004) and Chiang (2005). Even after controlling for the possible impact of the control variables to ensure that the effect in CEO involvement is not caused by other variables, it seems that this relation is significant as can be seen in Table 5.10.
Lower Upper
Step 1a
%Independent -1,618 ,824 3,853 1 ,050 ,198 ,039 ,998
Constant ,492 ,576 ,730 1 ,393 1,635
a. Variable(s) entered on step 1: % Independent.
Table 5.4 Board size
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)
Lower Upper
Step 1a
Board size -,319 ,086 13,930 1 ,000 ,727 ,614 ,859
Constant 1,533 ,581 6,953 1 ,008 4,633
Beasley et al. (1996) found that the audit committees of fraud organizations had significantly less meetings than those of non-fraud organizations. Uzun et al. (2004) have tried to find this result regarding the board of directors, but they did not find any significance. It was therefore not unexpected when I found no significant relation between the number of meetings of the board and the involvement of the CEO in the fraud. What is unexpected however, is that the results show a small positive (un-significant) relation, which implies that more meetings are accompanied by a bigger chance on CEO involvement in fraudulent activities. A possible explanation could be that the board has held more meetings because of the fraudulent activities, since it is likely that a board will hold extra meetings when there are signs of fraud. But this is a very interesting subject for further research, since it brings confusion about the efficiency of board meetings. Could it be possible that a CEO needs those board meetings to influence the board? Which would be consistent with the idea of the power of a CEO over the board of directors, especially when he of she is also the chairman/women (Cannella and Lubatkin, 1993).
Females in a board will have a positive impact on the earnings quality (Srinidhi et al., 2011) and empirical results have further suggested that the presence of women in the board is also positively related to its monitoring performance (Adams and Ferreira, 2008). The fifth hypothesis in this study is among other things, based on those existing results. The hypothesis stated that a small percentage of men, and thus a large percentage of women, would have a positive effect on the
Table. 5.5 Number of meetings
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)
Lower Upper
Step 1a
Meetings ,004 ,042 ,008 1 ,928 1,004 ,924 1,090
Constant -,742 ,420 3,118 1 ,077 ,476
a. Variable(s) entered on step 1: Meetings.
Table 5.6 Percentage Male/Female
B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)
Lower Upper
Step 1a
% Male 2,213 2,178 1,033 1 ,309 9,147 ,128 653,213
Constant -2,651 2,068 1,643 1 ,200 ,071
