Amsterdam Business School
The influence of the board diversity on risk-taking behavior
Name: Bram Koole
Student number: 10884270
Thesis supervisor: ir. drs. A.C.M. de Bakker Date: 26 January 2017
Word count: 11765
MSc Accountancy & Control, specialization Control
Page 2 of 41
Statement of Originality
This document is written by student Cornelis (Bram) Koole who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Page 3 of 41
Abstract
This paper examines the effect of diversity in the board of directors on the risk-taking behavior during the period of 2006-2015. Diversity is measured by gender, nationality, age and company experience. Risk taking is measured by variables financial risk, liquidity risk, solvency risk and corporate risk policy. The sample consists of the (non-financial services) companies of the Amsterdam Exchange Index and the Amsterdam Midcap Index, the two biggest stock indices within The Netherlands. We find that age, gender, nationality and company experience could influence aspects of risk-taking behavior.
Keywords: Board diversity, corporate risk taking, gender diversity, age diversity, nationality diversity,
Page 4 of 41 Contents
1 Introduction ... 6
2 Literature and hypotheses ... 8
2.1 Resource-based theory... 8
2.2 Upper echelons theory ... 8
2.3 Agency theory ... 9
2.4 Glass cliff theory ... 9
2.5 Contingency theory ... 10
2.6 Risk-taking ... 10
2.7 Characteristics of the board of directors ... 11
2.7.1 Age ... 11 2.7.2 Gender ... 11 2.7.3 Nationality ... 12 2.7.4 Company experience ... 12 2.8 Hypothesis development ... 13 3 Research method... 15 3.1 Conceptual model ... 15
3.2 Regression equation for testing hypothesis ... 15
3.3 Explanation of the variables ... 16
3.3.1 Dependent variables ... 16 3.3.2 Independent variables ... 17 3.3.3 Control variables ... 18 4 Data ... 20 4.1 General ... 20 4.2 Selection ... 20 4.3 Descriptive statistics ... 21
Page 5 of 41
4.3.1 Core statistical variables ... 21
4.3.2 Correlation matrix ... 23
4.3.3 Multicollinearity ... 25
5 Results... 26
5.1 Main results ... 26
5.2 Discussion of test results... 27
5.2.1 Model 1: association with financial risk ... 27
5.2.2 Model 2: association with liquidity risk ... 29
5.2.3 Model 3: association with solvency risk ... 31
5.2.4 Model 4: association with corporate risk policy. ... 32
5.2.5 Summary of results ... 33
6 Conclusions and limitations ... 35
6.1 Conclusions ... 35
6.2 Limitations and further research suggestions ... 36
6.3 Contribution and further research ... 36
References ... 38
Appendices ... 41
Page 6 of 41
1 Introduction
Recent years there has been a lot of lobbying in the Netherlands for more women within the board of directors as there is a shortage of women within the board of directors. Because of this shortage the Dutch Ministry motivated companies to hire more women for the board of directors. The target was to have at least a percentage women of 30% in boards of directors in 2016. This has been founded in article 276 of Book 2 of the Dutch Civil Code. If the 30% is not met, the company has to report in its board report a) why the 30% has not been met, b) how it was attempted to get a well-balanced board of directors and c) what the company will do to get the board of directors more balanced in the Future. Though this article 276 is dilapidated per 1st of January 2016 the wish for more women within board of directors is still present. Dekker (2015) wrote that this percentage was 7,8% in October 2015 and that 30% was likely not be reached in 2016. This subject is interesting because it is not completely clear if gender is that important in influencing the behavior of the board and therefore the company. A lot has been researched on board diversity and risk-taking behavior, but most researches do not concern Dutch companies and research mostly cover only one characteristic being gender.
Also, the financial crisis during 2007-2011 triggered for more research on corporate governance and risk taking behavior, because too much risk-taking was one of the main causes of the financial crisis.
In particular this research will aim if and how board diversity influences the risk-taking behavior of the company. Jensen et al (1993) claim that the board of directors is an important internal control system within a company, since they have the final responsibility for the firm. Both homogeneous and heterogeneous boards can be diversified. A more homogeneous board has directors which have often the same characteristics like age, gender, nationality and experience. A more heterogeneous board has directors which have more different characteristics. The benefits of a heterogeneous board, when comparing it to a more homogeneous board, is that members’ skills and experience can complement each other (Ferreira, 2010).
This research is important because there has not been a profound research to the risk-taking behavior of the board of directors based on their characteristics within the Netherlands. The outcome of this research can help companies when selecting members for the board of directors that suit the situation the company is in.
Page 7 of 41 The contribution to current literature is that it is more comprehensive than most studies that research board gender diversity. Also, board diversity has not yet been researched on companies in the Netherlands.
The diversity variables that will be used within this research are age, gender, nationality and board-experience. Age has been chosen as the age of a person could influence the level of risk being taken and therefore the risk-taking behavior of the board. In this rapid changing society, young people tend to be more innovative in comparison with older people. Therefore younger people might have a different attitude towards risk-taking behavior. Also, the people in the Netherlands are getting older and the pension-age has grown to 67 and is expected to grow even higher the coming decennia. Therefore the trend could be that the average age of a board of directors could change.
Gender diversity is important because both males and females have different characteristics and in general both respond different to risks. Hence, the lobbying in the Netherlands got the percentage of women in the board of directors from 2% to 7,8% between 2006 and 2015. There has been an increase of women in the board of directors but the target of ‘30% women in the board of directors in 2016’ has not been reached (Dekker, 2015).
The variable Nationality is chosen because the Netherlands has a multicultural society with many different nationalities who have different characteristics. Also in boards of directors there are often different nationalities that manage the company. Both cultural and communication aspects could influence the risk-taking behavior of the board of directors and therefore of the company. Company experience is important because experience is something people grow within roles they get. When someone has been with a company for a longer time, this may imply greater managerial power which could influence decisions. Also it could mean that they have less motivation to prove themselves when they have been with the company for a longer time. The aim of this paper is to find out what characteristics and combinations of characteristics of board-members influence the risk-taking behavior of companies in the Netherlands.
How does board diversity influence the risk-taking behavior of the board of directors?
The next section consists of the literature review and the hypotheses. In the third section the methodology is explained and the fourth section contains the data. In the last part (section 5) the results are disclosed and then follow the conclusions, limitations and recommendations for future research in section 6.
Page 8 of 41
2 Literature and hypotheses
In this section literature will be described which relates to board diversity and risk-taking behavior. This will be followed by the discussion about risk-taking behavior, the different characteristics that a board of directors could contain and the hypotheses.
2.1 Resource-based theory
The behavior of someone in terms of risk-taking behavior is highly dependent on the character of that person and also to the internal and external environment he or she is subject to. The same applies for a board of directors. This can be linked to the internal situation of a firm, which is strongly dependent on the resources a firm has. Therefore the resource-based theory applies. The resource-based theory explains that resources are used by a company to develop its strategy (Gallego‐Álvarez, 2011). This is relevant to this research because the strategy, which is influenced by the resources of a company, influences the risk-taking behavior of the board and therefore the company. But the board of directors itself can also be seen as a strategic resource. A strategic resource is a resource by which a company can get access to funds, skills, opportunities and a network (Setiyono et al, 2014). This is also in line with the human capital theory. The human capital theory refers to the knowledge and the characteristics of a person that contributes to his or her productivity (Tan, 2014). Since every person has its own unique characteristics, more diverse resources will lead to more diverse information for the board of directors. Also, Tan found out that more different characteristics lead to various insights and therefore a higher variety of potential solutions and could lead to better decision-making. Brown et al (2002) found out companies with a more diverse board, based on risk management, perform better than a more homogeneous board.
2.2 Upper echelons theory
The upper echelons theory relates to both the resource-based theory and the human capital theory. This theory states that strategic choices are in part predicted by the experience and background-characteristics of a person (Hambrick et al, 1984). The experience and background background-characteristics of a director influence the perception and thus the process of decision-making of the board of directors. More particular to this research, the directors strategic choices on fields of diversification and investments are influenced by this theory and influence the risk-taking behavior of the firm. For example, gender diversity within a board of directors is related to this theory as Post et al (2014) state that female directors have different types of experience and
Page 9 of 41 characteristics. For example, female directors often have more university degrees, are strong in marketing and sales and are more likely to have non-business backgrounds. But it is not limited to gender diversity. The managerial background characteristics (nationality/company experience) and the experience of a person (age/company experience) are important and influence the perception of a director and therefore the risk-taking behavior of the board of directors.
2.3 Agency theory
Another theory that has been written a lot about in board diversity researches is the agency theory. The agency theory was first mentioned by Jensen and Meckling (1976). It has been a much debated item since. The agency theory describes the dilemma between the principal and its agent. In a common company situation it concerns shareholders and a board of directors. The members of the board might have incentives to serve their own benefits instead of serving the benefits of the company’s owners, the shareholders. More particular to this research the board of directors could influence the risk-taking behavior of the company for wrong purposes. According to Adams et al (2010) the board of directors should be supervised by an independent person or team of persons. According to Post et al (2014) a female director contributes in being independent because she is more likely to be considered as an outside director, since they do not belong to ‘the boys’ that dominated the board rooms up until the year 2000.
According to Hillman et al (2003) boards of directors should be appropriate mixed in experience and capabilities to succeed in their roles. There is no optimal board-composition known to make best use of the situation and resources to minimize the agency problem. Hermalin et al (2001) claim that agency theory does affect the relation between risk-taking and board diversity but that the relation is hard to predict.
2.4 Glass cliff theory
As a member of the board there are a lot of decisions you are obliged to make and you are urged to be a leader for the rest of the company. These decisions will contain risks for the company and therefore leadership should be one of the key qualities of the members of the board. Ashby et al (2007) dispute this as good leadership is highly dependent on other factors, for example the relative health of the organization. They find that both women or men are selected for positions within a board based on the health of the organization. The Glass cliff theory is the theory that women are more likely to be selected for board of director roles during periods where the failure-chance is high (Cook et al, 2014).
Page 10 of 41 2.5 Contingency theory
The aspect that this research concerns is that the risk-taking behavior of a company will be affected by the internal and external environment. These will vary for each company and therefore each company will have a different strategy. Therefore the contingency theory is something to take account for in this research. The contingency theory claims there is no best way to lead a company but instead the company has to relate their actions to the internal and external situation (Chenhall, 2003). An organization needs to adapt. The case study from Woods (2008) resulted the risk control system to be contingent in three variables: (I) Central government Policy, (II) Communication Technology, and (III) Organizational size.
The risk that is being taken by a company can be influenced by the composition of the board of directors, as above described. In this research board diversity is segmented in four characteristics: age, gender, nationality and company experience.
The effect of board characteristics on risk-taking behavior in itself has little been studied. There have been studies on the effect on company performance in general but the results have not led to a general outcome. One of the few studies that did investigate the board diversity on risk-taking shows us that within the specific banking sector in Indonesia, board diversity increases risk and also reduces performance (Setiyono et al, 2014).
2.6 Risk-taking
Risk has been defined and interpreted in many different ways. The most general and common-used definition is the one from Lowrance (1976): “Risk is a measure of the probability and severity of adverse effects”. According to Furby and Beyth-Marom (1992) something can be defined as risk when the following criteria have been met: (a) the behavior could lead to more than one outcome and (b) some of these outcomes are not desirable or dangerous.
In this research taking behavior is defined as the behavior of the company linked to risk-taking which is influenced by the choices the board of directors make.
According to Croson and Gneezy (2009) risk is a result of human behavior and has been studied several times.
Based on studies within the fields of psychology and finance, it can be concluded that risk-taking is influenced by behavioral differences and thus the characteristics of a person.
Risk is a very broad understanding and can be measured and interpreted in multiple ways. In a similar study of Claessens (2000), the risk of a company is measured by fifteen different
Page 11 of 41 indicators, divided in seven groups. In this research risk-taking will be measured based on four of the more common financial measures used by Claessens: financial risk, liquidity risk, solvency risk and corporate risk policy. In paragraph 3.3.1 these measures will be explained in more detail.
2.7 Characteristics of the board of directors
In this paragraph the researches with respect to board-diversity (independent variables) are described, followed by the concerning hypotheses. The hypotheses focus on the influence of the separate characteristics on the risk-taking behavior.
2.7.1 Age
The age of a person could influence risk-taking behavior. Ali et al (2014) research board gender and age and concludes that a low average age of directors suggests is linked to high market value. They also conclude that a low average age of directors is related to high diversified as most board members are over 50. Even though a low average age is linked to high market value, other research indicates other: Bonn et al (2004) find out that there is no positive relationship between age diversity and the performance or risk of a company.
Hambrick et al (1984) found that young managers are often eager and motivated to be successful in their job. They implement new ideas and keep on learning while their career rises. Older managers mostly have more experience and have probably failed sometimes in their career. As their pension approaches they are fully aware of their own and the company’s reputation the expectation is that age has a negative effect on risk-taking behavior. Also, research shows us that companies are more probable to undergo major changes in their strategy if the company is run by younger directors (Wiersema et al, 1992).
2.7.2 Gender
Existing literature shows us that that women are more risk averse than men are (Jianakoplos and Bernasek, 1998). Sunder and Surette (1998) argue based on their research that risk averse behavior of women is explained by the fact that there are not many women in boards of directors. Though, the glass cliff theory should be taken into consideration which implies that women are more likely to be chosen as a board member in worse financial times where risk often already gets reduced.
According to Stellingwerf (2016) there is no significant relation between the number of female board directors and the risk-taking behavior of a company, which is in line with the results of
Page 12 of 41 Adams and Ragunathan (2013), who found that more female directors in a board of directors of a bank did not lead to more risk-taking behavior during the financial crisis.
But other recent research show different results. Faccio et al (2014) find that companies with a female Chief Executive Officer (hereafter: CEO) experienced less volatile earnings, lower leverage and was more likely to survive when compared to companies which were run by male CEO’s. The study of Berger et al (2014) shows that female directors are not risk averse and Adams and Funk (2012) found that women board members are actually more risk-loving than men.
2.7.3 Nationality
In the board of directors in Dutch companies often more nationalities are present in the board of directors. Research indicates that there are differences between nationalities and their risk-taking behavior. In a cross-cultural manner: Weber et al (1998) conclude that Chinese people are less risk-averse than American people. What more important is for this particular study on Dutch companies is the following: Weber et al found out that different nationalities within one board of directors leads to worse communication. This has been studied diverse times and is known as the social identity theory (Smith et al, 1994 and Lau et al, 1998). The lack of communication leads to less risk-taking because people tend to be more careful.
2.7.4 Company experience
In a board of directors there are different directors with diverse experience. Some have only been within the company and the board of directors for a year while other have lots of years of experience in the company. This could influence the risk-taking behavior of a person and therefore of the board of directors.
Research is not consistent in the influence of company experience on risk-taking behavior. Chen et al (2014) find that CEOs with more company experience have greater managerial power, have more control and are more committed to the strategy of the company. This could lead to less consideration making changes in a company than someone who still needs to prove himself within the company. This would lead to less risk-taking behavior.
Farag et al (2016) find a positive and significant relationship between company experience and risk-taking behavior. This is explained as the CEOs with more experience understand the corporate culture and the group dynamics better and therefore are more likely to make risky decisions.
Page 13 of 41 2.8 Hypothesis development
In this paragraph the above discussed characteristics and literature are used to formulate hypotheses.
Wiersema (1992) and Hambrick (1984) implies that younger managers are more eager and more motivated to implement changes and that companies are more probable to undergo major changes in their strategy if the company is run by younger directors.
Therefore the expectations are a negative association between the average age of the board of directors and risk-taking behavior of the company.
H1: The average age of the board of directors has a negative association with risk-taking behavior
Adams and Funk (2012) and Berger (2014) show that female directors are not risk averse and that female board members are actually more risk-loving than men. Though the research of Jianakoplos and Bernasek (1998), Sunder and Surette (1998), Faccio et al (2014) show different results about the influence of gender within the board of directors, the expectation for this research is that women are more risk averse and therefore that a higher representation of women in the board of directors has a negative association with risk-taking behavior.
H2: A higher representation of female in the board of directors has a negative association with risk-taking behavior
Concerning nationality, the results of Weber et al (1998) are followed, which indicates that, following the social identity theory, more nationalities leads to less communication and therefore more cautious behavior. The expectation is that the higher the number of different nationalities, the less risk is being taken.
H3: The number of different nationalities in the board has a negative association with risk-taking behavior
Chen et al (2014) show that more company experience could lead to less considering making changes in a company because they don’t have to prove themselves anymore. Faraq et al (2016) show that more company experience could lead to more risk-taking behavior because the corporate culture and group dynamics are better known.
In this paper the expectations are that fresh blood within the board of directors would lead to making more changes and could lead to more risk. Therefore, in this investigation, the results from Chen et al (2014) are followed. The expectation is that the higher the average company experience within the company, the less risk is being taken.
Page 14 of 41
H4: The average years of company experience within a company has a negative association with risk-taking
behavior
To summarize this paragraph below the hypotheses, including the proposed relations, are displayed in figure 1 below.
The influence of the board diversity on risk-taking behavior
Board characteristics Risk-taking
Age -
Risk-taking
dGender -
Nationality -
Comp_exp -
Figure 1. Hypothesis framework
Risk-taking will be measured based on four financial measures. Risk is segmented in Financial risk, Liquidity risk, Solvency risk and Corporate risk policy. Therefore, per characteristic the hypotheses is divided in four sub-hypotheses though the model and the directions stay the same. In the next section the research method is discussed.
Page 15 of 41
3 Research method
In this section the conceptual model, the methods and the variables for this research are discussed.
3.1 Conceptual model
In figure 2 the conceptual model is shown. The independent variable is board diversity and the dependent variable is risk. The control variables are ROA (Return on assets), Growth and Firmsize.
, g
Figure 2. Conceptual model
3.2 Regression equation for testing hypothesis
In order to test the relationships of the different characteristics on risk-taking behavior it is necessary to create a model that takes in account both the dependent and independent variables and on the other hand the control variables.
In the model the dependent variable risk is segmented in four categories: financial risk, liquidity risk, solvency risk and corporate risk policy. The independent variables are Age, dGender (gender), Nationality, Comp_exp (Company experience) and the control variables ROA (Return on assets), Growth and Firmsize are included.
The model is defined as follows:
Risk, the dependent variable, is measured in four different categories. The character ‘j’ stands for the four different dependent variables: Financial risk, Liquidity risk, Solvency risk and Corporate
Independent variable BOARD DIVERSITY Dependent variable RISK THEORY Control variable ROA (Return on assets) Growth
Page 16 of 41 risk policy. Age, dGender, Nationality and Comp_exp are the independent variables and ROA, Growth and Firmsize are the variables controlled for. is the disturbance term.
Below the variables are defined and further explained.
3.3 Explanation of the variables
3.3.1 Dependent variables
There are different ways of measuring risks. Different researchers use the loan loss provision when studying companies in the banking sector while other use the Z-Score. According to prior research from Faccio et al (2008) and Stellingerwerf (2016) risk is measured by calculating the standard deviation of the return on assets. Other researchers use variations of the standard deviation of the return on assets.
In a more comprehensive study of Claessens (2000), fifteen different variables are used to measure risk. In this paper is chosen to use four different ways to measure risk. Four variables are used and all stand for one risk category. The categories are Fin_risk (financial risk), Liq_risk (liquidity risk), Solv_risk (solvency risk) and Corp_risk_pol (corporate risk policy).
Fin_risk
Financial risk is the risk that the equity owned by the company will not be sufficient in order to meet the long-term liabilities. Financial risk is measured by the debt ratio which is measured by calculating the total debts divided by total assets. According to González et al (2013) risk aversion pushes firms towards lower debt levels and more risk-taking boards will have higher debt levels. The characteristics of a board of directors can influence the percentage of debt within a company. More optimistic managers will have higher debt levels than managers who are less optimistic (Hackbarth, 2008).
Liq_risk
The liquidity risk is the risk of running out of liquidity. The liquidity risk can be measured using the current ratio, the quick ratio or the net working capital set off against the total assets (Claessens et al, 2000). In this paper the quick ratio is used. The quick ratio expresses the company’s ability to meet its short-term debts by measuring the current assets minus stocks divided by the current debts. The difference between the quick ratio and the current ratio is that the current ratio takes in account the stocks which are not included when measuring the quick ratio.
Page 17 of 41 Solv_risk
The solvency risk measures the extent to which the interest payments are covered by the cash flows. Claessens et al (2000) use the Interest coverage ratio to measure the solvency risk, which is a standard measure of credit risk. The interest coverage ratio is measured by dividing the EBIT (Earnings Before Interest and Taxes) by the interest expenses. The higher the ratio, the lower the risk of not being able to pay the interest expenses.
Corp_risk_pol
The corporate risk policy is measured based on the percentage R&D (research and development) expenses. According to Kai et al (2013) R&D investments are often used to measure the riskiness of a corporate policy. The R&D investments are measured by dividing the R&D expenses by the total assets.
3.3.2 Independent variables Age
The first dependent variable is Age. The hypothesis: “The average age of a board member has a negative
association with risk-taking behavior “ is tested. The average age of the board of directors will be
measured by dividing the total age years of the board members by the number of board members:
Age will be treated as a continuous variable. dGender
The second dependent variable is gender. The hypothesis, “A higher representation of female in the
board of directors has a negative association with risk-taking behavior” is tested.
Gender is measured in similar studies in three different ways. Most studies, for example Nguyen et al (2008), obtain for the method of the proportion of female in the board while others use the number of females in total. Another variation is the dummy variable that turns to 1 when there is at least one female in the board of directors (Campbell et al, 2009). Therefore it’s more interesting to check if the presence of women has an association with risk-taking behavior than to check if the percentage of women within a board of directors has an association with risk-taking behavior. The particular interest of this study is to see if not only gender, but also other
Page 18 of 41 characteristics influence risk-taking behavior. Therefore, it was decided to follow Campbell in this study.
A dummy variable is created (dGender), which will turn to 1 when there is at least one female in the board of directors.
Nationality
The third characteristic is nationality. The hypothesis “The number of different nationalities in the board
has a negative association with risk-taking behavior” is tested. This variable is treated as a continuous
variable. The percentage of nationalities in the board and the number of board members will be used as indicator of how diverse the board is. The formula to measure this variable is as follows:
#NAT is the number of different nationalities on the board and N is the number of directors in the board. This formula is used to obtain a value of 0 if all members have the same nationality and a value of 1 if all members have a different nationality.
Comp_exp
The fourth characteristic is comp_exp (company experience). The hypothesis “The average years of
company experience within a company has a negative association with risk-taking behavior” is tested.
Comp_exp is defined as the average number of years the board of directors has been with the company:
Comp_exp is treated as a continuous variable. When the average Comp_exp is higher, the less risk is being taken.
3.3.3 Control variables
There are different control variables added to the model that have been shown to have impact on risk-taking behavior. Boubakri et al (2013) is being followed on three of his firm-specific controls (2000): Return on assets (ROA), Growth and Firmsize. As the sample consists of two different indexes where Firmsize is diversified between the companies Firmsize is taken into account.
Page 19 of 41 ROA
Return on assets is the first control variable. This variable will be treated as a proxy for the financial health of a company. Companies that are not profitable will not likely to take more risk and are therefore controlled for.
Growth
Asset growth is used as a proxy for the growth that a company makes. If it concerns a strong growing company in a growing market it is probable that they invest a lot and therefore take more risk.
Firmsize
The sample consists of two different indexes where the size of the firm is diversified between the companies. Firmsize is included as bigger firms usually show less risk-taking behavior (Stellingwerf, 2016). Firmsize is computed as the natural log of total assets.
Page 20 of 41
4 Data
In this section the data of the research is discussed. First the general information is shown where after follow the descriptive statistics which are discussed.
4.1 General
For this paper the financial data of the companies in the Amsterdam Exchange Index and the Amsterdam Midcap Index per December 2016. These are the two biggest stock indexes in the Netherlands. Following prior literature, financial services are not taken in account in the sample. With financial services banks and insurance companies are meant. A full list of companies is attached in Appendix 1.
For the companies, the annual data from 2006 – 2015 is retrieved. Unfortunately, not all observations and data necessary for this research are available in public databases. The quality of the board data (characteristics of the board) in on-line databases as Datastream and Amadeus (Bureau van Dijk) via WRDS were not complete and/or reliable.
Therefore the board data (characteristics of the board) is manually retrieved from the annual reports. The financial data is retrieved via Datastream.
4.2 Selection
The board and financial data is retrieved of 340 firm years for 38 companies. In table 1 the sample selection is illustrated and below is the process described.
Both the AMX and the AEX include 25 companies. As the period contains 10 years (2006 – 2015) the starting point is a total of 500 observations. As financial services are not included in this investigation nine companies are not included (minus 90 observations). From four companies the annual data for 2015 was not (publically) available yet (minus four observations). From other companies not all annual reports where available. This decreases the number of observations with 66. The total decrease is 160 (32%) which leaves 390 observations for the statistical analysis.
Page 21 of 41
# Observations Description
500 Number of companies * number of years -90 Financial services not included
-4 Data over 2015 not available yet
-66 Board characteristics unavailable due to missing annual reports 340 Total number of observations
Table 1. Sample selection process
Table 2 is a list with the different types of companies illustrated with the frequency and percentage per type of company.
Technology firms are the most represented in this research, followed by Industry and Real Estate Firms. Pharmacy, Materials and Employment Agency are types of companies that are not much represented in this research, once.
Type of company Frequency Percentage
Industry 5 13%
Retail and wholesale 2 5%
Transport and logistics 2 5%
Chemistry 2 5% Telecommunications 2 5% Construction 2 5% Technology 6 16% Consumer goods 2 5% Real Estate 4 11%
Oil and Gas 3 8%
Pharmacy 1 3%
Food and Drink 3 8%
Materials 1 3%
Employment agency 1 3%
Consumer services 2 5%
Grand total 38 100%
Table 2. Type of companies
4.3 Descriptive statistics
In this paragraph the core statistical variables and the correlation matrix is shown and discussed. Also, a multicollinearity test is performed in 4.3.3.
4.3.1 Core statistical variables
Page 22 of 41 The number of observations for the dependent variables varies between 300 and 339. The average number of Financial risk is 0.27. The minimum score is 0, which indicates no debts, while the highest debt-ratio is 0.74 (high risk).
The mean of Liquidity risk is 1.14 which indicates that the average company-observation in the sample has the 14% more current assets to meet their current liabilities. The minimum score is 0.38 and the maximum score 5.17. The lower the score the higher the indication of risk-taking. For the variable Solvency risk the standard deviation is 143.41 with an average of 1.69. Also the distance between the minimum score and the maximum score is high: 2.905. This can be explained by the fact that the calculation of the solvency risk (EBIT / Interest expenses) can vary a lot due to their character. With a small amount of debt, the interest expenses will be low and no matter if the EBIT is positive or negative, the outcome can be very large if the EBIT number is big. The average of corporate risk policy is 0.03, which indicates that 3% of the value of the total assets is invested for research and development purposes. In 55% of the observations there was no investment in research and development (the minimal score). The maximum amount that was invested in research and development was 0.51 which. A higher value results in a more risky corporate risk policy.
For all independent variables the number of observations is 340. For the variable age the average of the executive board in the sample is 52 (with a minimum age of 41 and a maximum age of 63) with a standard deviation of 4.03. In approximately one-third of the executive boards is at least one female present. The average of nationality is 0.33. In 103 observations there was only one nationality present in the board of directors (minimum score of 0). In 14 observations all the board members had a different nationality (maximum score of 1). The average company experience is 10.52 years with a standard deviation of 6.38. Seven directors have joined an executive board without any experience within the company while the maximum experience of an executive director is 32.25 years.
Page 23 of 41
Variable N Mean Deviation Standard Min Score Max Score
Fin_risk 339 0.27 0.15 0 0.74 Liq_risk 300 1.14 0.68 0.38 5.17 Solv_risk 335 1.69 143.41 -2,544.26 361.32 Corp_risk_pol 339 0.03 0.06 0 0.51 Age 340 52 4.03 41 63 dGender 340 0.32 0.47 0 1.00 Nationality 340 0.33 0.28 0 1.00 Comp_experience 340 10.52 6.38 0 32.25 ROA 339 0.07 0.08 -0.27 0.69 Growth 301 0.06 0.30 -0.76 2.70 Firmsize 340 15,592 40,901 119 303,123
Table 3. Core statistic values
4.3.2 Correlation matrix
In this paragraph the correlation matrix is shown and discussed. Correlation measures the linear relation between two variables and has a range between -1 (fully negative correlation) and 1 (fully positive correlation). When two variables relate too much to each other, there could be a sign for multicollinearity. According to Brooks (2008) there is no indication for multicollinearity as long as the correlation is between -0.70 and 0.70 percent. Multicollinearity is explained in 4.3.3. Table 4 presents the correlation between the variables for the total sample. There are no (significant) high correlations between the independent variables. The expectations were that there might be a high correlation between age and company experience but the correlation is 0.195.
As expected, there is correlation between the dependent variables financial risk, liquidity risk, solvency risk and corporate risk policy. The correlation varies between 0.282 and 0.490.
When taking the rule of thumb of 70 percent in account (Brooks, 2008), on the first sight there is no indication of multicollinearity between independent variables.
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (1) Fin_risk 1.000 (2) Liq_risk -0.444*** 1.000 (3) Solv_risk 0.056 -0.291*** 1.000 (4) Corp_risk_pol -0.469*** -0.490*** -0.282*** 1.000 (5) Age -0.198*** -0.016 -0.013 -0.018 1.000 (6) dGender 0.037 -0.248*** 0.032 -0.128*** 0.096** 1.000 (7) Nationality -0.014 0.027 0.057 0.032 0.194*** 0.174*** 1.000 (8) Comp_exp -0.017 -0.257*** 0.024 -0.187*** 0.195*** -0.196*** 0.033 1.000 (9) ROA -0.072* -0.065 0.359*** -0.292*** 0.080 0.069 -0.0292** 0.031 1.000 (10) Growth 0.062 -0.057 0.081* 0.001 -0.145*** -0.116** 0.004 -0.096** 0.073 1.000 (11) Firmsize -0.110** -0.142*** 0.040 -0.103** 0.399*** 0.321*** 0.037 -0.021 0.104** 0.016 1.000
4.3.3 Multicollinearity
Multicollinearity indicates a (too) strong coherence between the predictor variables which could lead to interpretation troubles. When two or more variables in a regression show a high multicollinearity, indicates that one of the independent variables can be considered as a linear combination of other independent variables. This test is done to see if there are independent variables that are redundant. The results from the correlation matrix indicates no sign for multicollinearity but to be sure multicollinearity will be tested for by using the Variance Inflation Factor (hereafter: VIF).
In table 4 the results of this test are shown. For all independent variables the VIF is between 1.0 and 1.4 in the model. According to Rogerson (2001) the recommended maximum VIF value is 5.0. According to Pan & Jackson (2008) the recommended maximum is 4.0. Following this literature there is no sign of multicollinearity in the model.
Fin_risk Liq_risk Solv_risk Corp_risk_pol
Age 1.348 1.302 1.376 1.348 dGender 1.247 1.271 1.254 1.247 Nationality 1.127 1.126 1.143 1.127 Comp_exp 1.101 1.041 1.103 1.101 ROA 1.055 1.062 1.067 1.055 Growth 1.075 1.093 1.076 1.075 Firmsize 1.345 1.350 1.359 1.345
Page 26 of 41
5 Results
In this section the results of the sample are shown and described. In paragraph 5.1 the main results are shown and in paragraph 5.2 the hypotheses and the results are discussed. Paragraph 5.3 contains a summary of the test results and in paragraph 5.4 the results of the control variables are discussed.
5.1 Main results
The relationship between risk taking and board diversity is explored by analyzing the four different regressions. In table 6 the results of the four models are displayed.
The dependent variable Financial risk (in the first model), measures the debt ratio to determine the financial risk of a company. A higher coefficient leads to a higher financial risk and therefore higher risk-taking behavior (and vice versa). The variable Liquidity risk (regression 2) measures risk taking based on the liquidity risk a company takes. The liquidity risk is measured by the quick ratio. A higher coefficient means a higher quick ratio, which means a lower liquidity risk and therefore lower risk-taking behavior (and vice versa). The third regression measures risk taking by looking at the Solvency risk. The interest coverage ratio is used to calculate the solvency risk, which is measured by dividing the EBIT (Earnings Before Interest and Taxes) by the interest expenses. A higher coefficient leads to a higher solvency risk and therefore higher risk-taking behavior. The variable Corporate risk policy is measured by looking at the research and development expenses, divided by the total assets. A higher coefficient leads to a higher corporate risk policy and therefore higher risk-taking behavior.
The relationship between the characteristics of the board and some risk aspects are significant in multiple ways. In the first regression, where Financial risk is measured, Age is significant at a 1% significance level and dGender (gender) is significant at a 10% level. For Liquidity risk both dGender (gender) and Company experience are significant at a 1% significance level while age is significant at a 10% level. For Solvency risk, Age is significant at a level of 10% and nationality is significant at a 5% level. Corporate risk policy is significant for dGender (gender) and Company experience at a 1% level.
The control variable ROA (Return on assets) is significant at a 1% level for models three (Solvency risk) and four (Corporate risk policy). The control variable Growth is significant at a level of 10% for model 2: Liquidity risk. The control variable Firmsize is not significant in any of the four models.
Page 27 of 41
Variables 1. Fin_ risk 2. Liq_risk 3. Solv_risk 4. Corp_risk_policy
β P β P β P β P Constant 0.614 0.000 *** 0.491 0.409 146,427 0.230 0.002 0.972 Age -0.007 0.006 *** 0.022 0.055 * -4,383 0.071 * 0.001 0.169 dGender 0.037 0.102 -0.487 0.000 *** -1,710 0.929 -0.022 0.007 *** Nationality -0.010 0.771 0.193 0.213 74,704 0.022 ** 0.007 0.629 Comp_exp 0.001 0.363 -0.033 0.000 *** 0,625 0.642 -0.002 0.000 *** ROA -0.070 0.512 -0.281 0.553 737,039 0.000 *** -0.236 0.000 *** Growth 0.029 0.319 -0.234 0.075 * 19,295 0.491 -0.001 0.912 Firmsize 0.000 0.328 0.000 0.169 0.000 0.560 0.000 0.302 F-value 2.556 0.140 8.417 0.000 *** 7.971 0.000 *** 8.159 0.000 *** Adj. R-squared 0.035 0.164 0.141 0.143
Table 6. Main results. * significant at 10%, ** significant at 5%, *** significant at 1%
5.2 Discussion of test results
Based on the results presented in table 6, the results of the independent variables are discussed per model. The control variables are discussed in 5.3.
5.2.1 Model 1: association with financial risk Goodness of fit
In model 1 the association of board characteristics with financial risk is measured. The adj. R-squared is 0.035 which indicates that 3.5% of the influence on financial risk is explained by the model. This means there is a small linear relationship between the dependent and independent variables. This is the lowest value in comparison with the other models.
The F-value for model 1 is 2.556 but is not significant. The p-value is 14%. Based on this result it model 1 is not statistically reliable.
For model 1 a positive association leads to lower risk-taking behavior and vice versa. The statistics of model 1 are discussed below.
Sign of β’s
Independent variables
For the association of age with risk taking behavior, it is hypothesized that the average age of the board of directors has a negative association with risk-taking behavior. This is based on Wiersema (1992) and Hambrick (1984) which imply that younger managers are more eager and more motivated to implement changes and therefore more probable to undergo major changes
Page 28 of 41 in the strategy of the company. The beta is -0.007 which means that if the average age of the board of directors is 1 year older 0.007 less risk is taken by the board of directors. As hypothesized there is a negative association of Age on Fin_risk.
For the association of gender on risk-taking behavior it is hypothesized that women are more risk averse than men and that therefore the representation of women in the board of directors has a negative association with risk-taking behavior. This is based on Jianakoplos and Bernasek (1998), Sunder and Surette (1998) and Faccio et al (2014). The beta of dGender in model 1 is 0.037, which indicates that the presence of women in the board of directors leads to more financial risk taken. The above result is in contrast with the expectations, but is in line with Berger et al (2014) and Adams and Funk (2014), who state that female directors are not risk averse and that female board members are actually more risk-loving than men.
Concerning the relationship between the number of different nationalities on risk-taking behavior it is hypothesized that the higher the number of different nationalities, the less risk is being taken. The above hypothesis is based on Weber et al (1998) who follows the social identity theory and indicates that more nationalities lead to less communication and therefore more cautious behavior. The beta of Nationality in model 1 is -0.010 which indicates that a higher number of different nationalities leads to less risk-taking behavior which is in line with the expectations.
The number of average company years is hypothesized to lead to a decrease in risk-taking behavior. This is based on Chen et al (2014) who show that more company experience could lead to less considering making changes in a company because they don’t have to prove themselves anymore. For Comp_exp the beta is 0.001 for model 1. This indicates that a higher average company experience of the board of directors (in years) leads to more liquidity risk. This is in contrast to the hypothesis. However, the literature was divided on this subject. The result is in line with Faraq et al (2016) who show that more company experience could lead to more risk-taking behavior because the corporate culture and group dynamics are better known.
Control variables
For the control variable ROA it was expected that companies that are not profitable and thus have a negative return on assets, will not likely to take unnecessary risks. For model 1 the beta of ROA is -0.070, which indicates that companies with a lower return on assets take more risk. The expectations were that more growing companies invest more than non-growing companies. For model 1 the beta is 0.029 which indicates that more risk is being taken by growing
Page 29 of 41 companies. This confirms the expectations that more growing companies invest more and therefore take more risk.
It was expected that Firmsize would influence the risk-taking behavior as bigger firms usually show less risk-taking behavior (Stellingwerf, 2016). For model 1 the beta is 0.000. This indicates that there is zero to barely influence of Firmsize on financial risk.
None of the values of the control variables in model 1 were significant. Significance
Of the independent variables, Age is significant on a 1% level. The results of the other independent variables are not significant.
Test results hypotheses
For the independent variable Age a significant value of -0.007 resulted on a 1% level. This is in line with the hypothesis but the F-value of model 1 is not significant. Therefore hypothesis 1 for model 1 is rejected, the other independent variables show no significant results in model 1 so the hypotheses 2, 3 and 4 are also rejected.
5.2.2 Model 2: association with liquidity risk Goodness of fit
The association of board characteristics with liquidity risk is measured in model 2. The adjusted R-squared is 0.164 which indicates that 16.4% of the influence on liquidity risk is explained by the model and therefore there is a linear relationship between the dependent and the independent variables. The F-value for model 2 is 8.417 on a significance level of 1%.
In contrast to models 1, 3 and 4, for model 2 a positive association leads to higher risk-taking behavior and vice versa.
Sign of β’s
Independent variables
For the variable Age, model shows a beta of 0.022 positive (significance level 10%) which means that if the average of the board of directors is 1 year older, 0.022 less liquidity risk is taken by the board of directors. This is in line with the hypothesis.
The beta of dGender in model 2 is -0.487, which indicates that the presence of women in the board of directors leads to more liquidity risk taken.
Page 30 of 41 The above result is in contrast with the expectations which were based on Jianakoplos and Bernasek (1998), Sunder and Surette (1998) and Faccio et al (2014), but is in line with Berger et al (2014) and Adams and Funk (2014), who state that female directors are not risk averse and that female board members are actually more risk-loving than men.
The beta of Nationality in model 2 is 0.193 which indicates that a higher number of different nationalities leads to less risk-taking behavior. This is in line with the expectations.
For Comp_exp the beta is -0.033 in model 2. This indicates that a higher average company experience of the board of directors (in years) leads to more liquidity risk. This is in contrast to the expectations but in line with the results from model 1, where the same situation occurred. The result is in line with Faraq et al (2016) who show that more company experience could lead to more risk-taking behavior because the corporate culture and group dynamics are better known.
Control variables
The control variable ROA has a beta of -0.281 which is contrary to the expectations that non-profitable companies take less risks.
For Growth the beta is -0.234 which is in line with the expectations that more risk is being taken by growing companies.
For Firmsize the beta is 0.000 which indicates no association between firmsize and liquidity risk. The results for ROA and Firmsize are not significant.
Significance
Age is significant on a 10% level, while the results of dGender and Comp_exp are significant on a 1% level. Nationality is not significant in model 2.
Test results hypotheses
For the independent variable Age a significant value on a 10% level resulted. This is in line with the hypothesis for model 2 and therefore hypothesis 1 for model 2 is not rejected.
Both dGender and Comp_exp show a significant value on a 1% level. These results are the opposite of the expectations of hypotheses 2 and 4 and are therefore rejected.
The result of Nationality in model 2 is not significant and as a result the hypothesis 3 for model 2 is rejected.
Page 31 of 41 5.2.3 Model 3: association with solvency risk
Goodness of fit
Model 3 measures the association of board characteristics with solvency risk. The adj. R-squared is 0.141 which indicates that 14.1% of the influence on solvency risk is explained by the independent variables in the model. This means there is a linear relationship between the dependent and independent variables. The F-value for the model is 7.971 on a significance level of 1%.
For model 3 a positive association leads to lower risk-taking behavior and vice versa. Sign of β’s
Independent variables
Independent variable Age in model 3 model shows a beta of -4.383 which means that the higher the average age of the board of directors, the less solvency risk is taken by the board of directors. This is in line with the hypothesis.
For dGender the beta is -1.710 which indicates that the presence of women in the board of directors leads to less solvency risk. This is in line with the hypothesis.
For Nationality, the beta is 74.704 in model 3 (solvency risk), which indicates that a higher number of different nationalities leads to a higher risk-taking behavior which is the opposite of the expectations and the aforementioned literature.
The beta of Comp_exp in model 3 is 0.625 which indicates a more risk-taking behavior when the average number of company years is higher. This is contrary to the expectations.
Control variables
The control variable ROA has a beta of 737,039 which is contrary to the expectations that non-profitable companies take less risks. This result is significant on a 1% level.
For Growth the beta is 19,295 which is in line with the expectations that more risk is being taken by growing companies.
The beta for Firmsize is 0.000 for model 3 also which indicates no association between firmsize and liquidity risk. The results for Growth and Firmsize are not significant.
Significance
For the independent variable Age a significant value resulted and Nationality is significant on a 5% significance level in model 3.
Page 32 of 41 Both dGender and Comp_exp show no significant values.
Test results hypotheses
For the independent variable Age a negative significant value resulted. This is in line with the expectations for model 3 and therefore hypothesis 1 for model 3 is not rejected.
Nationality is significant on a 5% significance level. This result is contrary to the expectations of hypothesis 3 in model 3 and is therefore rejected.
Both dGender and Comp_exp show no significant values. The hypotheses 2 and 4 for model 3 are rejected.
5.2.4 Model 4: association with corporate risk policy. Goodness of fit
In model 4 the association of board characteristics with corporate risk policy is measured. The adj. R-squared is 0.143. This means that more than 14.3% of the influence on the corporate risk policy is explained by the model and that there is a linear relationship between the dependent and independent variables. The F-value for model 4 is 7.159 on a significance level of 1%. Sign of β’s
Independent variables
In model 4 the beta of the independent variable age is 0.001 which indicates a higher average age of the board leads to a small increase of risk-taking (on corporate risk policy).
The beta of dGender in model 4 is -0.022 on a significance level of 1% which is in line with the expectations. This is in line with the expectations that the presence of women leads to less risk-taking behavior.
For Nationality, the beta is 0.007 which indicates that when the average number of company years is higher, more risk-taking behavior is taken. This is contrary to the expectations.
The beta of Comp_exp on Corp_risk_policy is -0.002 which is in line with the hypothesis. When the average years of experience of a board of directors is higher, the less risk is being taken.
Control variables
The control variable ROA has a beta of -0.236. This is in line with the expectations that non-profitable companies take less risks. Based on this result it can be concluded that non-non-profitable organizations invest less in research and development than profitable organizations.
Page 33 of 41 For Growth the beta is -0.001 which is contrary to the expectations. The result indicates that less risk is being taken by growing companies.
For Firmsize the beta is 0.000 in model 4. This indicates no association between firmsize and liquidity risk. The results for Growth and Firmsize are not significant while ROA is significant on a 1% level.
Significance
Both Age and Nationality have no significant values, while dGender and Comp_exp show significant results on a 1% level in model 4.
Test results hypotheses
For the independent variables Age and Nationality no significant value resulted. Hypotheses 1 and 3 are rejected.
For dGender and Comp_exp significant negative values resulted. This is conform the expectation of hypothesis 2 and 4 for model 4 and are thus not rejected.
5.2.5 Summary of results
Below in table 7 the results of the hypotheses and their directions are shown.
H Independent variable Dependent variable Direction Result
1. Age *** 1. Fin_risk Negative Not rejected 1. Age * 2. Liq_risk Positive Not rejected 1. Age * 3. Solv_risk Negative Not rejected 1. Age 4. Corp_risk_pol Positive Rejected 2 dGender 1. Fin_risk Positive Rejected 2. dGender *** 2. Liq_risk Negative Rejected ‡ 2. dGender 3. Solv_risk Negative Rejected 2. dGender *** 4. Corp_risk_pol Negative Not rejected 3. Nationality 1. Fin_risk Negative Rejected 3. Nationality 2. Liq_risk Positive Rejected 3. Nationality ** 3. Solv_risk Positive Rejected ‡ 3. Nationality 4. Corp_risk_pol Positive Rejected 4. Comp_exp 1. Fin_risk Positive Rejected 4. Comp_exp *** 2. Liq_risk Negative Rejected ‡ 4. Comp_exp 3. Solv_risk Positive Rejected 4. Comp_exp *** 4. Corp_risk_pol Negative Not rejected
Table 7. Hypothesis results.
Page 34 of 41 For the independent variable Age, three of the four (sub) hypotheses are not rejected. For dGender, hypothesis 2, model 4 is not rejected and three are rejected, whereof one is significant contrary the expectation. For independent variable Nationality, all hypotheses are rejected, whereof hypothesis 3 for model 3 shows a significant result (Solv_risk), but the opposite direction than expected.
For Comp_exp (Company experience), three of the four (sub) hypotheses are rejected. Hypothesis 4 (model 4), the negative association with Corp_risk_pol (Corporate risk policy) is not rejected. Hypothesis 4 (model 2) shows a significant result for Liq_risk (liquidity risk) but contrary to the expectations.
So none of the hypotheses showed uniform test results for the four different measures of risk-taking.
Page 35 of 41
6 Conclusions and limitations
In this section the summary with the conclusions are discussed. Hereafter follow the limitations and the further research suggestions.
6.1 Conclusions
The objective of this study is to examine if the characteristics of a board of directors, the board diversity, influences the risk-taking behavior of a company. The research question is: “How does board diversity influence the risk-taking behavior of the board of directors?”
The board diversity variables used are Age, dGender (gender), Nationality and Company_exp (company experience). Risk-taking behavior, the dependent variable, is measured on four different ways: financial risk is based on the debt ratio, the quick ratio is used to measure liquidity risk, solvency risk is based on the interest coverage ratio and the percentage of R&D investments is used to measure corporate risk policy. Control variables added to the model are ROA (return on assets), Firmsize and Growth. The sample consists of the companies of the Amsterdam Exchange Index and the Amsterdam Midcap Index in the period of 2006 to 2015. Financial institutions, banks and insurance companies, are excluded of the sample.
Based on literature four hypotheses are formed. It is hypothesized that Age, dGender, Nationality and Company_exp have negative associations with financial risk, liquidity risk, solvency risk and corporate risk policy.
The financial data is retrieved via Datastream. The board data is manually retrieved from the yearly annual reports, as the quality of the data in databases was not sufficient. After excluding banks and insurance companies and the fact that not all data was available leaves a total remainder of 340 observations for the statistical regression.
The results present evidence that multiple characteristics of the board of directors have a significant effect on risk-taking. The diversity variable age has a negative effect on financial risk, liquidity risk and solvency risk, though the effect on financial risk cannot be trusted 100% as the model is not 100 percent reliable. The presence of a female in the board of directors has a positive effect on liquidity risk but a negative effect on corporate risk policy. The number of different nationalities within a board of directors has a positive effect on solvency risk. For company experience, the results indicate that a high company experience has a positive effect on liquidity risk and a negative effect on corporate risk policy.
Page 36 of 41 The result of the regression of board diversity on risk are presented in section 5. The results fail to indicate a clear relationship between board diversity and risk-taking behavior. More particular, there is for the variables dGender, Nationality and Comp_exp no clear evidence that they influence the different types of risks in any way. The results are diverse and varied where the variables sometimes indicate a negative relationship and other results indicate the opposite. Only for the variable Age the relation with risk is rather clear as three of the four results are significant in a negative way, meaning that older board-members are more risk averse than younger board-members.
6.2 Limitations and further research suggestions
The research has diverse limitations that will be described in this paragraph. The most important limitation is the fact that the risk-taking behavior of a company is not only influenced by the board of directors, but also by many other aspects like laws and regulations, the economic climate, the influence of shareholders and the influence of the supervisory board. Laws and regulations can be different for a country, but also for a specific industry. The economic climate in the period of 2006-2015, used in this research, has been volatile: in the period of 2007-2011 the world was in a recession and from 2012-2015 the markets slowly started to grow again. These different times ask for different strategies where risk is one of the key-items for.
Another important limitation is the fact that the influence on risk-taking behavior is measured by looking at the characteristics of the complete board of directors, while the responsibilities and powers of the members differ. For example, the CEO (Chief Executive Officer) and CFO (Chief Financial Officer) have a lot more power and responsibilities than other representatives of the board. Therefore the influence of these persons is probable a lot more but the research aims on the characteristics of the complete board, without focusing on the specific members.
In this research the number of different nationalities is one of the characteristics but the cultural background is not. Cultural background could definitely have influence on the risk-taking behavior of a board of directors but is not taken into account.
6.3 Contribution and further research
The intention of this research was to investigate if not only gender, but also other characteristics of the board of directors would influence the risk-taking behavior of a company. The contribution of this study is that it has shown that other characteristics like age, nationality and company experience also influence the risk-taking behavior of the company. Also, this research
Page 37 of 41 aims on a sample with a lot of Dutch companies while other researches within this field most of the times aimed on companies of the S&P 500, which contains the 500 biggest American companies.
Further research could focus on both the characteristics of the executive board and the supervisory board. Another suggestion is to aim on the characteristics of one or two of the executive functions within companies, for example the CEO and the CFO, because of their power and responsibilities.
This study could be a start for more studies which focus more on the different functions within the executive board. More profound research on this could help Dutch companies to select their members of the executive board.
