Does gender diversity on the board of directors affect the financial performance?

Does gender diversity on the board of

directors affect the financial performance?

A study of Swedish companies registered on Stockholm NASDAQ OMX

Master Thesis, IY2517, MBA program

Authors: Isabelle Wegmann Hachette & Mirza Galijasevic

1

Abstract

This master thesis examines if gender diversity on corporate boards affect the financial performance for large, mid and small cap firms on the Swedish OMX during the years 2005-2013. In corporate governance, the board of directors is an important control mechanism that could enhance the company’s financial performance. Board of directors have the potential to render valuable service to the company in shaping of its strategic development. Traditionally, main parts of the boards have been and are dominated by men, the boards are not balanced in the gender diversity perspective.

In Sweden, there is a debate about introducing quota systems to increase the gender diversity on boards for public companies. There are many motivational factors why the number of women on boards should be increased. A gender diverse board could be considered as a competitive advantage that leads to developed knowledge base, and improved creativity and innovation, increased competition, new perspectives in the decision process or an attractive working environment. However, there are also research that indicates that homogeneous groups are the most efficient in decision making and consensus.

This thesis investigates if the representation of women on the board of directors has any effect on the company’s financial performance and if the effect is positive or negative. Hence, the research question is: “Does gender diversity on the board of directors affect the financial performance?” The majority of the most cited research articles in this research field report from other countries than Sweden. Moreover, the research articles on the Swedish market either lack statistical significance, did not investigate long term effects or did not perform a fundamental regression analysis.

In this thesis we examine the relationship between women on the board of directors and financial performance in public companies listed on the Stockholm NASDAQ OMX stock exchange in Sweden during 2005-2013. The data in this thesis was collected from 250 public companies. The data was processed through a data analysis methodology in order to extract useful information to extract valid data for the regression analysis.

We provide an in-depth statistical analysis in this thesis. We use correlation analysis and regression analysis using data averaged over the following three time periods; 2005-2007, 2008-2010 and 2011-2013. Four different financial performance measures were evaluated for each of the three time periods; Return on Equity, Return on Invested Capital, Return on Sales and Return on Assets. A stepwise backwards elimination model was used to achieve at least a 95% confidence interval for the results in the model. The regression analysis shows that the research question is valid on a 0.05 significance level. Hence, this thesis shows that women on the board of directors affect financial performance positively for all financial performance measures.

The results of this thesis supports the research question that gender diversity on the board of directors affect the financial performance positively.

2

Acknowledgments

We want to take this opportunity to express our gratitude to Marie Aurell and Henrik Sällberg at BTH for the help with completing this thesis. We also want to thank our friends Dr. Jenny Janhager Stier at the Royal Institute of Technology, Dr. Charlotte Immerstrand and Dr. Bertil Wegmann at Linkoping University for important insights after reviewing the report. Equally, we are also grateful to Rachel Remlinger, MBA, and an American colleague for proofreading the thesis.

Finally, we are grateful to our families for their support and encouragement, but most of all, for their patience during all the hours needed to complete this thesis.

3

Abbreviations

CEO = Chief Executive Officer

EBIT = Earnings Before Interest and Tax OMX = Stockholm NASDAQ OMX ROE = Return On Equity

ROA = Return On Assets ROS = Return On Sales

ROIC = Return On Invested Capital

W/B = Women’s representation in percentage of the board of directors

4

List of Figures

Figure 1: The four governance bodies (Corporate governance 2016)

Figure 2 Example of a board recruitment (Spencer Stuart Board Index (2014) Figure 3 History within board of directors

Figure 4 Wish list for new directors (referens?) Figure 5 Life cycle of groups

Figure 6 Dynamics of relations in groups (Kanter 1977) Figure 7: OMXSPI 2005-2014 (Avanza u.d.)

Figure 8: Scatter plots of the dependent data related to the independent data for ROE, ROIC, ROS, ROA. Figure 9: Scatter plots of the dependent data related to the independent data for ROE, ROIC, ROS, ROA, with logarithmic transformations of Total Debt, Total Asset and Turnover.

Figure 10: Probability plots of ROE, ROIC, ROS, ROA Figure 11: Probability plots of ROE, ROIC, ROS, ROA

Figure 12: Scatterplots of the data distribution after removal of outliers

Figure 13: Scatter plots of residuals versus the independent and control variables for ROE, ROIC, ROS, ROA

Figure 14: Scatter plots of residuals versus the fitted value

5

List of Tables

Table 1 Skillset within board of directors ... 16

Table 2: Summary of the literature review ... 31

Table 3: Data source collection ... 41

Table 4: Descriptive statistics of mean W/B, mean W/M, mean W/T ... 44

Table 5: Descriptive statistics of the log transformation of Tot debt, Tot asset, Tot turnover (*_{functions of} LOG). ... 46

Table 6: Descriptive statistics of Total Debt, Total Assets, TurnOver for each dependent variable. ... 47

Table 7: Interval of the mid quartiles for the dependent variables ... 48

Table 8: Descriptive statistics of ROE, ROS, ROIC, ROA ... 49

Table 9: Number of observations and mean value for each group... 53

Table 10 Descriptive statistics for the independent and control variables for ROE ... 54

Table 11 Pearson correlations for correlation with ROE, * p<0.05, ** p<0.01, *** p<0.001 ... 55

Table 12 Descriptive statistics for the independent and control variables for ROIC ... 56

Table 13: Pearson correlations for correlation with ROIC, * p<0.05, ** p<0.01, *** p<0.001. ... 57

Table 14 Descriptive statistics for the independent and control variables for ROS ... 58

Table 15 Pearson correlations for correlation with ROS, * p<0.05, ** p<0.01, *** p<0.001 ... 59

Table 16 Descriptive statistics for the independent and control variables for ROA ... 59

Table 17 Pearson correlations for correlation with ROA, * p<0.05, ** p<0.01, *** p<0.001 ... 61

Table 18 Step wards backwards elimination of independent and control variables. e - indicates that this independent variable has been eliminated due to poor statistical significance. ... 62

Table 19 Coefficients for the final regression with α = 0,05 significance level ... 62

Table 20 Stepwards backwards elimination of independent and control variables for ROIC regression. . e - indicates that this independent variable has been eliminated due to poor statistical significance. ... 64

Table 21 Coefficients for the final regression of ROIC with α = 0,05 significance level ... 65

Table 22 Step wise backwards selection results in the 95%-confidence interval. . e - means that this independent or a control variable has been eliminated due to poor statistical significance. ... 67

Table 23 Coefficients for the final regression of ROS with α = 0,05 significance level ... 67

Table 24 Step wards backwards elimination of independent and control variables for ROA regression. . e - means that this independent variable or control variable has been eliminated due to poor statistical significance. ... 69

Table 25 Coefficients for the final regression of ROA with α = 0,05 significance level ... 70

6

Table of Contents

Abstract ... 1 Acknowledgments ... 2 Abbreviations ... 3 List of Figures ... 4 List of Tables ... 5 Table of Contents ... 6 1. Introduction ... 9 1.1. Problem formulation ... 10 1.2. Research question ... 11 1.3. Research steps ... 12 1.4. Research objective ... 12 1.5. Delimitations ... 12 2. Theory ... 14 2.1. Corporate governance ... 14

2.1.1. The shareholders' meeting ... 14

2.1.2. The board of directors ... 15

2.1.3. Chief Executive Officer (CEO) ... 15

2.2. Board composition ... 15

2.3. Board composition in the US ... 17

2.1. Performance of board of directors ... 19

2.1.1. The legalistic perspective ... 19

2.1.2. Resource dependence ... 20

2.1.3. Class hegemony ... 20

2.1.4. Agency theory ... 20

2.4. Group theory ... 20

2.4.1. Group types ... 21

2.4.2. Life cycle of groups ... 21

2.2. Heterogeneous and homogeneous group composition ... 22

2.3. Group minority perspective ... 22

2.4. Hypothesis ... 24

7

3.1. Positive correlation ... 26

3.2. Missing or negative correlation ... 29

3.3. Discussion regarding literature review... 30

4. Methodological framework ... 33

4.1. Financial performance measures ... 33

4.2. Statistical analysis ... 34

4.2.1. Correlation analysis ... 34

4.2.2. Regression analysis ... 35

4.2.3. Standard assumptions for the linear regression model ... 36

4.2.4. Backwards selection ... 37

4.2.5. Hypothesis testing ... 37

4.2.6. Coefficient of determination ... 38

4.2.7. Normally distributed data ... 38

4.2.8. Data standardization ... 39

4.2.9. Analysis of outliers ... 39

4.3. Test unit of analysis ... 39

4.4. Data ... 40

4.4.1. Data selection and missing data ... 42

4.4.2. Diversity measure ... 42

4.5. Dependent, independent variables and control variables ... 42

4.6. Regression analysis equation ... 42

5. Results ... 44

5.1. Regression validity tests ... 44

5.1.1. Independent variable analysis ... 44

5.1.2. Linear relationship between independent and control variables and dependent variables .. 44

5.1.3. Dependent variable analysis ... 46

5.1.4. Checking the validity of the multiple linear regression analysis ... 49

5.2. Descriptive statistics ... 53

5.2.1. Mean Values and Standard deviations for ROE, ROIC, ROS and ROA ... 53

5.2.2. Descriptive statistics and correlation of variable ROE ... 54

5.2.3. Descriptive statistics and correlation of variable ROIC ... 55

8

5.2.5. Descriptive statistics and correlation of variable ROA ... 59

5.3. Regression analysis ... 61

5.3.1. Regression analysis with dependent variable ROE... 61

5.3.2. Regression analysis with dependent variable ROIC ... 63

5.3.3. Regression analysis with dependent variable ROS ... 66

5.3.4. Regression analysis with dependent variable ROA ... 68

5.3.5. Summary of results of multiple regression analysis ... 72

5.3.6. Summary of hypothesis tests ... 73

9

1. Introduction

This chapter presents the relevant background information for this thesis, such as the problem formulation, research question and the research steps. The purpose of the study and the delimitations are also presented within this chapter.

The monitoring role by the corporate boards is an important corporate control mechanism that could enhance the company’s financial performance. The board of directors and the company executive management team are understood to be the most important leaders of an organization. An assumption underlying the study of leadership is that leaders actually affect organizational performance (Dubrin 2010). Nohria, Joyce and Roberson (2003), concluded that the CEO influenced 15% of the total variance influencing factors in a company’s profitability or total return to shareholders. Kaiser, Hogan and Craig (2008) performed a statistical analysis, which showed that the company leaders are feasibly responsible for somewhere between 15% and 45% of organizational performance. However, as stated by Waldman et al (2001) it is not only the leadership affects but moreover the leadership style, transactional versus charismatic leadership, that is important for performance. Charismatic leadership, a leadership style often possessed by women, showed a slightly positive relationship with performance whereas transactional, a leadership style often possessed by men, did not show a significant correlation to performance. Moreover, Waldman showed that when the environment is uncertain the correlation between charismatic leadership and performance is even stronger.

Over the last 3 decades there has been an increase in the number of women holding seats on company boards in many countries. This phenomenon can be attributed to several of the following factors: a general increase in gender equality in society, expansion of workplace advancement opportunities allowing women to climb the corporate ladder to gain the competence and experience necessary for a board position, as well as political decisions, such as quotas requirements in some countries (Dagens Juridik 2013). Still, there exists an overrepresentation of men on company boards.

10

Researchers have studied the reason why women do not reach entrance to the corporate board room (Furst och Reeves 2008). Women executives consistently explained that the idea of gender-based stereotypes as the major barrier to their achievements (Catalyst 2005). Women tend to adopt socially stereotypical gender roles to be accepted (Eagly och Carli 2003). Men are more prone to emphasize dominance, aggression and achievement. In contrast, women as stereotype emphasize affiliation, nurturance, deference and abasement, roles not traditionally considered as management roles (Rigg och Sparrow 1994).

Kanter (1977) anticipated that the competence of women and their contribution to organizational culture would gradually become more recognized. She defined in 1977 a model based on the premise of social contact theory. Kanter's idea was that the dominant group will create heightened barriers between themselves and the minority group that is not advantageous for the group performance (Kanter 1977). Kanter states that not only do minority members gain from growing in size, but that the majority profits from the resources women bring to the organization. The competition model, derived from ethnic and racial conflict research, holds that discrimination results from the minority’s attempts to appropriate the majority’s resources (Lortie-Lussier och Rinfret 2002). According to Blalock (1967) an increase in the proportionate size of a minority group is perceived by the majority as a threat to their advantages and induces discrimination to secure their dominant position. However, the theory concludes that once a proportionate size has been reached, the minority initiates assertive actions to counter the hostility and discriminatory behaviors of the majority.

1.1. Problem formulation

The question about women on boards and gender diversity was first raised and led by the opinion in USA. During the last decade Europe has overtaken this role and leads the gender diversity progress worldwide. European boards are made up of 20.3% women nearly doubled the average female representation on boards when compared with boards outside of Europe (11.6%) (Prnewswire 2014). European boards are now in line with those in traditionally high-ranking countries like the United States (21.2%) and Australia (22.6 %). The five countries worldwide with the highest percentage of women on large company boards are all European and include two non-Nordic countries for the first time: the United Kingdom (22.6%) and France (28.5 %). In 2013, the European Commission wrote that the number of women on European company boards is increasing but still much lower than the 40% goal of the commission (Europaportalen 2014). There were 16% W/B according to statistics from the European commission in 2013. There is a long way to go in order to reach 40 % in 2020. In 2013, best-in-class were Finland (29%), Latvia (28%) and Sweden (26%). However, Norway which is operating under the effects of a government enforced quota system since 2006 takes the lead with W/B at 44 % in 2013.

11

number of W/B increased for the first time in 2014, from 22.3% to 24.7%, and women’s representation in executive management (W/M) increased by one percent, from 17.4% to 18.4 % (Dagens Juridik 2014). Part of the driving factor for these increases is likely due to the fact that Sweden is currently ranked among the top five countries in the world when it comes to political representation for women. In Sweden women make up to 45% of the members of the national parliament (Freidenvall and Hallonsten 2013). As we can see from these numbers, however, the representation of women are far more balanced in politics than in the private sector.

An interesting revelation was presented in the study by AllBright (2013); there were 14 female CEOs on OMX and 237 male CEO’s in 2013 of the board members on Stockholm NASDAQ OMX. Stjerngren indicates that female CEOs have the boards trust for approximately 2-6 years whilst the corresponding figure for male CEOs is 5-6 years (Stjerngren 2012). Organizational change can take time in order to improve performance and CEOs need time to change the organization. Therefore women would have less hours to introduce changes in the organization. Saab, a Swedish defense company, has focused on gender equality for several years and worked systematically with increasing the portion of female managers (Saab 2013). At an early stage the company realized that gender equality has an impact on profitability and therefore a clear key performance indicator was set, just like in all other areas of strategic importance. As a result, in 2007 the management team of Saab decided that the portion of female managers at Saab shall be at least 30% by the year 2015. This is commonly referred to as “30/70” internally within Saab. Back in 2007 Saab only had 14% female managers and the company has now reached 24%. Saab’s President and CEO, Håkan Buskhe, and the rest of the management team are fully committed to achieving the target of 30 %.

In the society, there is a debate about introducing quota systems to increase the number of women in the corporate boards for public corporations. In November 2013, the European Parliament adopted the European Commission's proposal to gender quotas for the people in the largest European company boards (Dagens Juridik 2013). The issue will now proceed to the Council of Ministers for discussion. If the proposal goes through, all listed companies with more than 250 employees should have at least 40% of female directors in 2020. For public companies this should be achieved faster with a completion date of 2018. For those companies that do not meet the requirements or do not show how they try to recruit people of both sexes to their boards, fines and exclusion from public contracts are being discussed as possible sanctions.

1.2. Research question

12

motivation factors for increasing number of women on corporate boards are many. We are looking at motivation factors from the financial perspective: could it be financially beneficial for the company to select a board with gender diversity? In order to investigate this, the research question is formulated as:

x “Does gender diversity on the board of directors affect the financial performance?”

Specifically, we want to investigate if there is a correlation between the company’s financial performance and the number of women on corporate boards. Our unit of analysis are companies registered on the Stockholm OMX during the time period 2005-2013. We do an in-depth statistical review of the data acquired during the period.

1.3. Research steps

We started our research by selecting the research topic of gender diversity related to financial

performance. The second step was to investigate previous research done on the database BTH Summon. The second step was very important to gain knowledge and examine previous work done on the subject to formulate our research question and develop hypotheses to be tested. The next step was to investigate what kind of data could be obtained from the companies to be included in our research. Then we developed our methodology with correlation analysis and multiple regression analysis. In parallel we gathered data from primary sources and secondary sources if necessary. The data is interpreted and presented in tables and figures.

Once the data acquisition was completed we continued with the analysis using a statistical computational program; Minitab 17 Statistical Software (Minitab 2015) to calculate the regression analysis. The analysis then lead us to the results and if necessary a revision of the hypotheses. Finally, from the results we draw conclusions and discussed how our research has contributed to the research society and eventual strengths and flaws with our investigation.

1.4. Research objective

As mentioned above, there is only one research question. The unit of analysis is companies that are registered on the Stockholm OMX. We did start to investigate different industries but did not find any differences in women related to financial performance. This led us to believe that the area of research should not focus on a specific industry but should instead be focused all industries on registry on the Stockholm OMX. The research within this report is exploratory and is described in more detail in the methodology chapter. We have focused on identifying a problem in today’s society.

1.5. Delimitations

13

We decided to only include companies that were registered on Stockholm OMX from 2005-2013. The main reason for only focusing on the Stockholm OMX is that all companies registered there must submit an annual report that contains information such as women on boards. This information is public and available to anyone that has internet access.

14

2. Theory

This chapter presents the relevant theory for this thesis, such as an introduction to corporate governance and group theory.

2.1. Corporate governance

Corporate governance regulates the direction in which a company operates and aims, and is a collective term for describing, formulating, and judging the overarching governance of the companies. It is a wide area and the main purposes are to govern companies so that owners demand for return on invested capital is fulfilled and profit generated. Shareholders, boards of directors, auditors, and management all play an important part in corporate governance. (Blom, Kärreman and Svensson 2012)

The Swedish company act stipulates which corporate bodies a public company must have, the tasks of each of these and the responsibilities of the people within each body. A company must have

three decision making bodies in a hierarchical relationship according to Figure 1; the shareholders'

meeting, the board of directors and the CEO. There must also be a controlling body, the auditor, appointed by the shareholders' meeting. (SCGB 2016)

Figure 1: The four governance bodies (Corporate governance 2016)

2.1.1. The shareholders' meeting

15 2.1.2. The board of directors

The board of directors is responsible for the company’s organization and the management of the company’s business. The board has the outmost liability for the company’s organization and management, as well as the obligation to ensure satisfactory control of the company’s financial position. For Swedish public companies, which are in focus for this thesis, the board must consist of at least three board members. One of these three members must be appointed to the chairman of the board. That person has a particular responsibility for leading the work of the board and ensuring that it fulfils its legal obligations. The election of board members and the auditor is performed during the annual general shareholders’ meeting that is to be held within six months of the end of the financial year. The board is obliged to follow any specific directives approved during the shareholders’ meeting, given that these do not contradict the Swedish company act. (Board of directors 2016)

According to OMX regulations, a maximum of one board member is allowed to be on the executive management team of the company or on one of its subsidiaries. In most cases, this place is taken by the CEO. However, many companies have no member of the executive management team on the board of directors; hence boards listed on OMX are composed almost entirely of non-executive directors. The CEO may be a member of the board but not its chair. The regulation also states that at least two members must also be independent of the company’s major shareholders. (Nasdaq 2016)

2.1.3. Chief Executive Officer (CEO)

The Chief Executive Officer (CEO) is responsible for the company’s day-to-day management, and he or she is subordinate to the board of directors. The CEOs duty is to handle day-to-day management but also to present issues that are outside the scope of day-to-day management to the board of directors who then instruct the CEO in writing on how these are to be handled or decided. According to the Swedish company act and the company’s articles of association, the CEO is obliged to follow the instructions given by the board. The board itself may also decide on matters that are a part of day-to-day management. (CEO 2016)

2.2. Board composition

According to Conger and Lawler (2001) the selection of board members has often been informal. Directors are sometimes appointed because they are personal friends of an existing board member or the CEO. Nowadays, this mindset is changing. More and more companies have started to understand that boards are composed of individuals with different skills, knowledge, information, power, and time to contribute. In order to be able to create an effective board, the needs of the

board must be identified. Table 1 shows typical skillsets that are required for a member of the board

16

to selecting individuals who have different areas of expertise and perspective so that they as a group cover the key knowledge and information areas needed to enable the board to offer guidance on critical issues facing the firm. Power is mainly related to influence in various forms, and finally, opportunity and availability, which is related to availability as this is a critical issue in board effectiveness. For example, active CEOs have managerial, industry, and functional knowledge but the downside is that they are busy. They need to manage their own company and still have time to advise and monitor other companies.

Table 1 Skillset within board of directors

Skillset Description

Strategy company’s business model, corporate strategy formulation, competition, global markets

Leadership _{senior executive coaching, CEO development}

Organizational strategy implementation, change management, group effectiveness and organization design

Relationship _{government relations, investors, financial analysts, media, community} Functional

knowledge finance, technical expertise, legal issues, HR, information technology, marketing Ethics _{ethical responsibilities}

A board needs to first identify a particular profile or type of individual they are in need of and then to actively search for directors that fit this profile. Given the limited number of positions on a board and the time demands of board membership, boards must be aggressive in recruiting members. Congler and Lawler (2001) claim that by attracting a few high-profile directors who are leaders in their field can help create a virtuous circle in which new members are attracted because the board has good members already.

A lot of companies have started to utilize so called board recruitment matrices, see Figure 2, in order

17

However, achieving the right mix of membership on a board is a complex process. Selecting directors simply on the basis of their knowledge is not enough. Most of the work of directors is interdependent so board members must be able to work well together. The selection of board members should not be so much about picking talented individuals who fit relatively independent positions as about picking appropriate team members. In this respect, selecting board members should not be all that different from selecting employees for working groups where teamwork and contributing to a team effort is critical to success. The selection process must look carefully at the characteristics of individuals with respect to not only their substantive knowledge and skills but also their ability to work together in a group and their time availability. (Conger and Lawler 2001)

Figure 2 Example of a board recruitment (Spencer Stuart Board Index (2014)

2.3. Board composition in the US

Among public corporations in the US, the members of the boards of directors have a mix of

managerial, functional and specialized backgrounds. As can be seen in Figure 3, the majority of the

members on the board of directors are active or retired CEOs or COOs. Figure 3 originates from the

18

Figure 3 History within board of directors

A recent study from Spencer Stuart (2015) shows that the “wish list” for new director background is contains active CEOs, women, financial expertise and minority. So these are the attributes that should be incorporated in the board recruitment matrix when searching for a new member.

19 2.1. Performance of board of directors

Zahra and Pearce (1989) conclude that boards are guided by four distinct theoretical perspectives. The four perspectives differ in their view of what directors should do and also in the way of influencing company performance.

The four perspectives are defined as: x Legalistic Perspective

x Resource Dependence x Class Hegemony x Agency Theory

There are mainly four board attributes that determine director’s performance in their roles: board composition, characteristics, structure and process. Board composition refers to the size of the board and the mix of different director’s types and minority representation. Characteristics refer to the director’s experience, functional background, independence, stock ownership, and similar variables that influence directors’ interest and performance of their task. Structure refers to board organization, division of labor among standing committees and the efficiency of its operations. Finally, process refers to the decision-making related activities and styles of the board. These four attributes are linked and related to each other.

2.1.1. The legalistic perspective

The legalistic perspective suggests that boards contribute to the performance of their companies by conducting their legally mandated responsibilities. This means that role of the board is to select the CEO, monitor CEO performance, represent shareholders interest and monitor company performance. According to the legalistic perspective, the attributes that determine the director’s performance of their roles are the board’s composition, characteristics, structure and process. Composition refers to the size of the board and the mix of different people, characteristics refers to variables that influence director’s performance such as background. Structure refers to board organization such as division of labor, and finally and process refers to the decision-making related activities (Zahra and Pearce (1989))

20 2.1.2. Resource dependence

The board’s number one task is to extract resources vital to the company’s performance. Boards enhance organizational legitimacy and serve as “boundary spanning role” and make timely information available to executives. The board’s role is to scan the environment, represent the company in the community, and to secure valuable resources. (Zahra and Pearce (1989))

2.1.3. Class hegemony

Boards perpetuate the power and control of the ruling capitalists’ elite over social and economic institutions. The board’s role is to selected recruitment of directors and reduces transaction costs for member firms. This perspective represents the “glass ceiling” perspective; only individuals from selected groups enter into the corporate boards. (Zahra and Pearce (1989))

2.1.4. Agency theory

According to Pfeffer & Salancik (1978) agency theory is based on the relationship between agents and principals. The theory highlights the agency problems that can arise from the separation of ownership and control. The problems can occur in any situation where there is a principal-agent relation; a principal, who wants to have an action performed, and an agent, who is expected to act in the principal’s interest and perform the action. The agency theory states that the major role of boards is to monitor actions of the executives (the agents) and to ensure their efficiency, and to protect the owners (the principals) interests. Furthermore, the role of the boards is to maximize shareholder’s wealth, reduce agency costs, select and reward the CEO, evaluate CEO and company performance, make strategic decisions and control.

The importance of board composition is prominently recognized in each of the four perspectives on board research. For the legalistic and agency principles, composition variables reflect the extent to which management dominates the board. This is often judged by the ratio of outside to inside directors. Outside directors are considered essential for ensuring an effective system of checks and balances. For class hegemony theorists, board composition indicates whether selective recruitment of new directors is done in a way to preserve capitalist interests. The glass ceiling can also prevent other groups, minorities (e.g. women) to get selected as board members. Finally, for resource dependency, board composition mirrors the characteristics of the firm's environment, helping the firm acquire resources vital to survive and grow (Pfeffer & Salancik (1978))

2.4. Group theory

21

members influence one another. Furthermore, the leader should motivate team members so that they have more confidence to perform a task. An underlying assumption according to Granér (1991) is that a group exists only to the extent if it is necessary to fulfil a goal. If there are no goals, the group would not have been created in the first place.

2.4.1. Group types

According to Heap (1980) there are three different group types:

x A natural group, that has been created by the group members’ themselves. This could be a group of friends for instance.

x An organized group, which has been created by an external party but the group members have accepted to belong to the group. This is typically the case within an organization.

x A forced group where the groups have been created by an external party and the group members have been forced to participate in the group.

In order to understand or modify a group, there are various facts that need to be taken into consideration, such as the number of members, number of years within the group, gender, age, educational level etc. To understand how people work and why they behave the way they do need knowledge of the relationships and norms that are in the group are held. Even for understanding leadership, the person must have knowledge of the status and roles working in the group.

According to Granér (1991) the size of a group has a direct impact on the performance. The bigger the group, the more complex is the formal and informal organization. If the group is large, there are more passive group members. There is an upper limit of 8-10 people in order for a group to perform at its top. Most people prefer to be part of a group of 3-4 people as the organizational problems are minimized. This correlates well with OMX regulations, which stipulate that a board of directors must consist of at least three board members.

2.4.2. Life cycle of groups

22

a resource for the group. Two variables determine the life cycle: productivity and morale. Productivity is the amount and quality of the work accomplished and it depends on the group members’ ability to collaborate, their knowledge and skills, clear goals, and access to needed resources. Morale is the team’s confidence, motivation, and unity in achieving the purpose. The

different stages are shown in Figure 5, productivity being shown on the red line and morale on the

black (dotted) line.

Figure 5 Life cycle of groups

2.2. Heterogeneous and homogeneous group composition

According to Chen and Macmillan (1992) heterogeneity enhances a firm’s action propensity and its actions and response magnitudes, potentially benefiting company performance On the other hand heterogeneity reduces the firm’s speed, both in acting and responding, possibly leading to lower performance as it may initially produce more conflict in the decision making process. The variety of perspectives that emerges can also make decision makers to evaluate more alternatives and more carefully explore the consequences of these alternatives.

Steiner (1972) proposed that heterogeneity has benefits for unstructured, novel tasks, but homogeneity is better for routine tasks. Hambrick and Mason (1984) saw a distinction between stable and turbulent environments. Their view is that heterogeneity benefits turbulent environments and homogeneity is beneficial in stable environments. Eisenhardt (1989) has claimed that heterogeneity possibly leads to lower performance as it might produce more conflicts in the decision making process, as the variety of perspectives causes the board of directors to evaluate more alternatives and more carefully explore the consequences of these alternatives.

First, corporate diversity promotes a better understanding of the marketplace. Second, diversity increases creativity and innovation. Third diversity produces more effective problem-solving. Fourth, diversity enhances the effectiveness of corporate leadership. Homogeneity at the top of a company is believed to result in a narrow perspective while diverse top managers take a broader view. (Robinson and Dechant 1997)

2.3. Group minority perspective

23

when these two are to be integrated and interact in the same business. When women on the board of directors are a minority, Kanter's theory is useful and relevant to this study.

Most studies on critical mass begin with Kanter’s (1977) work on how women behave in corporations. Critical mass theory has gained wide currency over the last twenty years among politicians, the media and international organizations to bring more women into politics. (Lortie-Lussier och Rinfret 2002)

In Figure 6 we find the definitions of a skewed population of a group as token or dominants (Kanter

1977). A token is a category differentiated from the majority that composes less than 20 percent of the total group population. A group is tilted when the category differentiated from the majority composes 20 percent to 40 percent of the total group population. A balanced population of a group is when the categories differentiated compose 40 percent to 60 percent of the total group population. Kanter's idea was that the dominant group will create heightened barriers between themselves and the minority group that is not advantageous for the group performance (Kanter 1977). Kanter’s speculates that not only do minority members gain from growing in size, but that the majority profits from the resources women bring to the organization.

Figure 6 Dynamics of relations in groups (Kanter 1977)

24

total group population. In the gender-workgroup framework women as tokens representation on boards are isolated and subjected to discrimination on the basis of gender stereotypes; once they become a minority (i.e., 20 percent of the group members), they develop networks and can influence the majority; in balanced proportions, opportunities to interact would likely eliminate inequalities (Lortie-Lussier och Rinfret 2002).

Women as tokens representation on boards are isolated and subjected to discrimination on the basis of gender stereotypes; once they become a minority (i.e., approximately a third of the group), they develop networks and can influence the majority; in balanced proportions, opportunities to interact would likely eliminate inequalities (Lortie-Lussier och Rinfret 2002).

The competition model, derived from ethnic and racial conflict research, holds that discrimination results from the minority’s attempts to appropriate the majority’s resources (Lortie-Lussier och Rinfret 2002). According to Blalock (1967), increase in the proportionate size of a minority group is perceived by the majority as a threat to their advantages and induces discrimination to secure their dominant position. However, the theory concludes that once a proportionate size has been reached, the minority initiates assertive actions to counter the hostility and discriminatory behaviors of the majority.

Women are more likely than men to find themselves on a glass cliff – an allusion to the fact that their leadership positions are relatively risky or precarious since they are more likely to involve management of organization units that are in crisis. In this way, compared to men, women who assume leadership positions may be differentially exposed to criticism and in greater danger of being apportioned blame for negative outcomes that were initiated well before they assumed their new roles.

2.4. Hypothesis

25

stated earlier, the board of directors is a group and most of the work is interdependent, which means that members must be able to work well together. Congler and Lawler (2001) stated that the selection of board members should not be so much about picking talented individuals as about picking appropriate team members. Chen and Macmillan (1992) claim that diversity enhances performance due to promoting a better understanding of the marketplace, increasing creativity and innovation, producing more effective problem-solving and enhancing the effectiveness of corporate leadership. Kanter’s theory (1977) is also applicable, claiming that having a dominant group, which is more than 80 percent of the total group population, will create heightened barriers between themselves and the minority group. This is not advantageous for the group performance. Based on the above, we believe that having diversified board of directors, could improve the financial performance as the board of directors should ensure satisfactory control of the company’s financial position. We have derived the following hypothesis:

H1: More women on the board of directors improve the financial performance.

26

3. Literature Review

This chapter provides a summary of previous research related to the representation of women on the board of directors and financial performance. The literature review seeks to identify new ways to interpret prior research, in order to reveal gaps that exist in previously published material. The chapter ends with a table that reviews and summarizes the literature review.

The results of other empirical studies on the relationship between gender diversity and financial performance are not consistent. Some studies have found a positive correlation between women on the board of directors and financial performance, while others have not found any correlation or even a negative correlation between women on the board of directors and financial performance.

3.1. Positive correlation

The studies by the consultancy company McKinsey (2007) and non-profit organization Catalyst (2007) support that more women on the board of directors lead to higher performance. Nevertheless, both studies are fundamentally flawed due to methodological weaknesses. For example, neither study indicates whether or not the differences in the performance measures are statistically significant and the selection of companies in the McKinsey report is based on subjective criteria (which will be described more in detail in the next section), as noticed by (Lortie-Lussier and Rinfret 2002).

The McKinsey study found that companies who ranked in the top quartile of executive-board diversity had Return on Equity (ROE) that was 53% higher. In addition, these same diverse companies had earning margins that were 14% higher than the least diverse companies (McKinsey & Company 2007). The Catalyst study showed that boards that are more diverse also have a higher ROE, higher Return on Sales (ROS), and higher Return on Invested Capital (ROIC) (Catalyst 2007).

Adler (2001) presented an extensive 19-year study of 215 Fortune 500 companies that showed a strong correlation between promoting women into the executive suite and high profitability. Three measures of profitability were used in the Adler study to demonstrate that the 25 Fortune 500 companies with the best record of promoting women to high positions are between 18 and 69 % more profitable than the median Fortune 500 companies in their industry. A breakthrough on the research and opinion came when McKinsey & Company and Catalyst presented their reports that stated that W/B have a positive effect on financial performance in companies. Catalyst (2007) and McKinsley & Co (2007).

27

added. A change in the explained variance was used to determine levels of significance. Correlation and regression analyses indicated that board diversity is positively associated with company performance. They claimed that further research is needed to address how gender diversified board of directors operate differently from less gender diversified board of directors.

Krishnan and Park (2005) studied the moderating effects of environmental characteristics in top-management teams in a sample of 679 companies from the 1998 Fortune 1000 list. They tested the following hypotheses; 1) the greater the proportion of women on the top management team, the better the performance of the organization, 2) the positive relationship between women on top management teams and organizational performance will be stronger at higher levels of environmental munificence, instability, and complexity. The methodology used by the researchers was to calculate the proportion of women in the top management teams in percentage. Then they weighted average industry performance for each industry based on the sales in the actual industry the company operated in. The financial measures used in the study were ROA and ROS. They used ROA for each of the industry groups since ROA is viewed by many researchers as a stable variable and more indicative of the efficient use of an organization’s resources. The results from the study were that the first hypothesis was supported both for ROA and ROS and the second hypothesis was not supported. The results show a positive link between the proportion of women on top management teams and organizational performance.

Adams and Ferreira (2009) showed that female board directors have a significant impact on the board of directors and a company’s outcome. The authors studied many issues related to this topic. They performed a regression analysis on ROIC and gender diversity with data from 1,939 companies during the time period 1996-2003. They found that gender diversity has a positive correlation with performance. However, it was not considered to be a statistically significant result. Carter et al. (2003) studied corporate governance, board diversity and company value by mean and

regression analysis having the Tobin Q1 _{as describing the company value. The researchers}

investigated 638 companies in the USA in 1999 and found a significant positive relationship between the quota of women or minorities on the board and company value.

Gavious et al (2012) utilized a univariate and multivariate analysis approach to explore the relation between female directors and earnings management in terms of ROA in high-technology firms. Sixty Israeli firms listed in the USA were investigated during the years of 2002 to 2009 (478 observations). The result was that there is a positive correlation between the proportion of female directors and a firm’s value.

1 _{Tobin’s Q is a financial measure defined by market value of firm capital divided by replacement cost}

28

Campbell and Minguez-Vera (2008) performed a study of 68 companies between the years

1995-2000 that provided 408 observations. The companies in the study were all listed on the public stock market in Madrid and not active in the financial sector. The study investigated the number of women on the board of directors related to the number of stocks and stock value. The authors performed a regression analysis with company value as the dependent variable. Company value

was picked with Tobin’s Q. In order to strengthen the reliability of the study, they used control

variables in terms of the total number of directors, the leverage ratio, ROA and the size of the company. The results showed that greater gender diversity can generate economic benefits for the company, but a clear conclusion that one or more women on the board of directors has an positive effect on the company's value could not be performed. However Campbell and Minguez-Vera

(2008) selected a few conditions that they thought needed to be improved:

• The presence of one or more female on the board of directors • The proportion of female directors

• Two indices of equality and diversity, which are Blau and Shannon Index2

Campell and Vera (2010) investigated the short and long term effects of the appointment of female directors on ROA in Spain during the years 1989-2001. The authors discovered that the stock market reacts positively in the short-term to the announcement of female board members, suggesting that investors on average believe that female directors add value. This belief appears to be confirmed by regression analysis that showed that female board appointments are positively associated with company value over a sustained period.

Lückerath-Rovers (2011), investigated the performance of 117 Dutch companies in relation to W/B. The result showed that companies with women on the board of directors perform financially better than those without women on the board of directors. The authors used the framework of Catalyst (Catalyst 2007) and Mckinsey (McKinsey & Company 2007), and developed the analysis with statistical regression methods for the investigation of women on board of directors. They used the dependent variable ROE in the regression analysis when comparing the companies.

Lönnquist et al (2006) investigated equality related to financial performance for all listed

companies on the A-list and O-list on Stockholm OMX during the years 2002-2005. However, the

authors did not perform a regression analysis. A total of 291 companies were analyzed. The results

of the study clearly show that companies with a larger portion of women on the board of directors

have larger earnings before interest and tax (EBIT), as well as higher revenues. The author’s

selected two control groups, the first was companies with no women on the board of directors and the second was companies with the highest ratio of women on the board of directors to support

their theory that: “the impact of female competence is assumed to be greater the more women and

the larger boards are. The assumption is that with three people or at least 30% of the total number

29

would be the small group to serve as a powerful factor and be able to influence decisions, compared to a group of two people or fewer people could do.” The study used several performance measurements; EBIT, revenue, annual revenue increase, number of employees, total assets, and solidity. Lönnquist et al used independent samples t-test to investigate the relationship. This is a method that is not as reliable as regression analysis, because… Finally, the authors asked for

further research on the issue to gain additional evidence. The most interesting observation that did

come from the study was that over time companies that had a higher portion of women on the board

of directors did have an increase in their EBIT in relation to the companies with no women on the board of directors.

3.2. Missing or negative correlation

The Nordic perspective on board equality was studied and analyzed by (Randøy, Steen och Oxelhiem, A Nordic Perspective on Corporate Board Diversity 2006). The data was acquired during 2005 for the largest 459 publicly traded companies in Scandinavia consisting of 154 companies from Denmark, 144 companies from Norway and 161 companies from Sweden. Due to non-response from some companies on all the study variables, the authors reduced the number of companies studied to 100 companies from Denmark, 86 from Norway and 157 from Sweden. A multivariate regression analysis based on 343 observations was performed. The study results indicated that there are no effects from gender, age, or nationality on the studied company’s stock market performance or ROA. However they did conclude that increased diversity with a maintained board size can be achieved without shareholder value destruction

Rose (2007) investigated whether women on the board of directors and female CEOs could affect the company’s value. The study focused only on the Danish market and did not find any correlation between company performance and female representation. They used regression analysis with the Tobin Q’s as performance indicator for the companies. Rose discussed a plausible reason that board members not originating from the traditional “old boys club” may have decided to assimilate into the traditional circles by suppressing any special feature stemming from the board members’ unconventional background. In other words, there might be a process of socialization where the unconventional board members have adopted the behavior and norms of the conventional board members/business leaders. This could be due to women being tokens rather than a balanced group. Flam and Winberg (2009) studied equality in boards from the investor’s perspective; e.g. should the investor care about gender equality if they want to maximize their stock portfolio. The unit of analysis was thirty randomly selected companies sampled in the years 2000-2008. The variables

were tested using two regression models, the CAPM3 _{and the Fama/French tree-factor model}4_{. The}

models are used to test time-series regressions in order to understand how the effect develops over

3 _{Capital Asset Pricing Model (CAPM), A model that describes the relationship between risk and expected return and}

that is used in the pricing of risky securities (Investopedia 2015).

4 _{The Fama-French Three-Factor Model is a method for explaining the risk and return of stocks, designed by Nobel}

30

time. The data tested shows that there is a small positive correlation between more women on the board of directors and higher stock revenue. However, these results did not reach a 95% confidence interval.

Norberg and Vernersson (2011) studied the industrial sector of the Stockholm OMX from three year samples (2004, 2007 and 2010). The total number of companies in the study was 68 and the number of samples was 204. The researchers used ROE and operating margin as financial metrics.

The study shows that there is no clear link between gender equality and profitability, as no clear pattern of correlation had been observed between gender dimension and the financial metrics. However, the investigation has shown that equality on company boards has increased over the years and fewer and fewer companies today are without female representation on the board of directors.

Norway is a specifically interesting market to study since the management teams are more diverse than others mainly due to the quota system that has been implemented in Norway. Torchia et al (2011) addressed if an increased number of women (1, 2 or 3 or more women) on the board of directors results in a buildup of critical mass that substantially contributes to company innovation. Tests were conducted on 317 Norwegian companies. The results showed that it was possible to enhance innovation by increasing the number of women. Data was gathered in questionnaires from 2,954 companies in all different categories. The answers were graded on a scale from 1 to 7. Then “Organizational Innovation” became the dependent variable in the regression analysis. The independent variable was women on the board of directors, grouped in four groups: 0, 1, 2, or 3 or more. Board strategic tasks were used as a mediating variable. The control variable was company size and the dummy variable was industrial sector. The analysis used multiple linear regressions.

We have summarized our findings from the literature review in Table 2 below.

3.3. Discussion regarding literature review

We conclude the majority of the previous research has been made in other countries than Sweden. The research that has been done on the Swedish market use limited statistical analysis and weak statistical significance or limited statistical analysis.

Based on the literature review there are a couple of things that can be concluded. First;the majority

31 Ta ble 2 : Sum m ar y of t h e lit e ra tu re r e vie w

Author / Title /Yea

r Pro b le m focu sed /m ar ke t /ye ars / n u m b er of ob se rvation s Sta tistica l m ethod an d Financial me as u re Result o n f inancial perfo rmance v s w om en on corp orate b oard s/R 2 A da m s a nd Ferreira / “ W om en in th e boardr oom an d th ei r i m pact o n g ov ern an ce an d perf or m an ce” / 2009 Fe m

ale board directors im

pact on board in pu ts an d f ir m ou tco m es / US A / 1 99 6-20 03 / 1,93 9 f ir m s 9299 observ ation s Regr es sio n a nal ysi s o n RO C a nd ge nd er di ve rs it y P ositiv e w it h in du str y c lassif ication dum m ies (p<0 .1 , R2=0 ,2 5) . Negativ e with fi rm fi xe d ef fe ct s (p<0.1, R 2=0.1 1). Ca m pbell a nd Merin gu ez -V er a / “Ge nd er D iv ers ity in th e Boardroom and F ir m Fin an cial P erf or m an ce ” /2008 G en der div ers ity in th e boar dr oom a nd fi rm f in an cial perf or m an ce /Spa in /1995 -200 0/ Regr es sio n a nal ysi s o n To bi n’ s Q a nd ge nd er div ers ity P ositiv e relatio ns hi p (p<0 .1, R2 =0 ,16 3) Ca m pell a nd Ver a / “ Fe m ale o n board appoin tm ents a nd f ir m v al uati on : s hort a nd lon g-ter m e ff ect s” / 2 010 A re f em ale board appoin tm ent s positiv el y ass ociated w it h fi rm v al ue? / Spain / 1 989 -2001/6 8 com panies an d 408 o bs erv ation s Regr es sio n a nal ysi s on T obin ’s Q a nd f em ale board a ppoin tm en ts P ositiv e relatio ns hi p both in the sh ort ter m a nd th e long ter m to t he an noun ce m en t o f f em ale board a ppoin tm en ts (p<0 .1, R 2 n ot p resen ted ) C arter, Si m ki ns a nd Si m ps on / “ C orporate go ve rn an

ce, board div

ers ity a nd f ir m v al ue ” / 2003 Corporate g ov ern ance, board div ersit y an d f ir m v al ue / USA / 1997 /638 observ ation s Mean and r egres sion anal ys is, T obin ’s Q a s fi rm va lu e an d ge nd er div ers ity P ositiv e relatio ns hi p and sta tis ticall y sign if ica nt positiv e relatio nship (p<0 .0 5, R2=0 .25) Erh art, J.D an d C .B / “ B oard of director div ersit y an d fir m f in ancial pe rf or m ance” / 2003 Ex am in at ion of f in ancial perform an ce data and th e percen tag e o f w om en a nd m in orities on boards of directors / USA / 1993 & 19 98 / Correlatio n and regress io n anal ys is on RO A and R O I and ge nd er div ers it y P ositiv e co rrelatio n and regres sion relatio nship (p n ot sign if icant, R2=0 .09) Fla m a nd W in berg / “W om en s rep resen tation in boards and th eir eff ect o n prof itabilit y” / 2 009 Equ alit y in boards f ro m th e inv estor’s perspectiv e / S w ed en / 2 000 -2009 / 29 com pa nies 261 obs erv ation s Multi va riate Regr es sio n a nal ysi s o n sto ck va lu e an d ge nd er div ers ity P ositiv e bu t not statis tical si gnif ica nt (p>0 .25 an d p>0.14, R 2=0.4 4) Gavious, Seg ev a nd Yo sef / ” Fe m ale

directors and earn

in gs m ana ge m en t in h ig h-tech nolo gy f ir m s” / 2012 Relatio n bet w ee n f em ale direc tors and ea rn in gs ma na ge me nt in hi gh -techn olo gy f ir m s / Is rael /2002-2009 / 60 f ir m s, 478 obs erv ation s Un iv ariate and m ulti va riate anal ys is on f ir m v al uatio n and w om en o n boards P os itiv e relation sh ip bet w een th e proportion of fe m

ale directors and th

32

Author / Title /Yea

r Pro b le m focu sed / m ar ke t /ye ars / n u m b er of ob se rvation s Sta tistica l m ethod and Financial me as u re Result o n f inancial perfo rmance v s w om en on corp orate b oard s/p-va lue, R2 ( w hen available) Krishn an and P ark / ”A f ew good w om en : O n top m an ag em en t tea m s” / 2005 Ef fe ct s of e nvironm en tal c har acteristics in top -m ana ge m en t tea m s / U S A / 1998 /679 f irm s Regr es sio n a nal ysi s o n R O A a nd wo me n o n boards P os itiv e relation sh ip bet w een th e proportion of w om en o n T M T s and org an iz ation al perf orm an ce (p<0.001, R 2=0. 28) Lü ck erat h an d R ov ers / ” Wom en o n boards an d f ir m per fo rm an ce” / 2011 ) Fin an cial perf or m an ce i n relat ion to th e nu m ber of w om en on th e board / T he N et he rl an ds / 2005 -2007 / 9 9 fi rm s Regr es sio n a nal ysi s o n ROE , ROS an d R O IC an d w om en on boards F ir m s w ith w om en on boards perf orm better th an f ir m s w ith ou t w om en on boards (R 2=p<0.01 ,R 2=0.2 4) No rberg and Ver nersso n / ”Do Gender eq ual co m pa nies ha ve h ig he r prof itabilit y? ” / 2011 Relatio n bet w ee n g en der eq ua lit y a nd profitabilit y / S w ed en / 2004, 2007, 2010 / 68 f irm s an d 204 obs er vations Regr es sio n a nal ysi s o n R O E an d operatin g ma rg in an d g en de r div ers ity No clear lin k bet w ee n g en der eq ualit y a nd prof itabilit y, w ea k stati stical si gn ificance (R 2=0.17) Rose / ”Do es f em ale board re presen tation inf lu en ce fi rm perf or m an ce? T he D an ish evidence” / 2 007 C an w om en o n boards an d f em ale CE O s affe ct the f ir ms ’ va lu e / D enm ar k / 1998-2001 / 4 43 observ ation s Regr es sio n a nal ysi s o n To bi n Q ’s an d w om en on boards No lin k bet w ee n fir m per fo rm ance (T obin Q’s) and f em

ale board represen

33

4. Methodological framework

This chapter presents the methodology and statistical theory used to test the hypothesis

The financial performance measures chosen have been selected as an outcome of available information from the annual reports as well as being used in various in the literature review within the research area. . We present the statistical analysis theory that we will use to analyze and retrieve the results. We perform a quantitative study to examine the relationship between diversity and performance in public companies in Sweden during the years of 2005-2013. We use contemporary data in the period of 2005 to 2013. We don’t have any control over the events that we are researching as it is archival data and it is a contemporary event which clearly leads to that exploratory research method is the one most suitable for this work according to Yin (2009).

4.1. Financial performance measures

Return on Equity (ROE) is the amount of net income returned as a percentage of shareholders equity (Damodaran 2007). ROE measures a corporation's profitability by revealing how much profit a company generates with the money shareholders have invested and it is an important measure of the profitability of a company. Higher values are generally favorable meaning that the company is efficient in generating income on new investment. ROE is calculated as:

ܴܱܧ ൌ ݊݁ݐ݅݊ܿ݋݉݁

ݏ݄ܽݎ݄݁݋݈݀ݏ݁ݍݑ݅ݐݕ

Return on Invested Capital (ROIC) is used to assess a company's efficiency at allocating the capital under its control to profitable investments (Damodaran 2007). It is calculated as net income minus dividends, and then divided by total capital. The ROIC measure gives a sense of how well a company is using its money to generate returns. ROIC is calculated as:

ܴܱܫܥ ൌ݊݁ݐ݅݊ܿ݋݉݁ െ ݀݅ݒ݅݀݁݊݀ݏ

ݐ݋ݐ݈ܽܿܽ݌݅ݐ݈ܽ

Return on Sales (ROS) is net profit as a percentage of sales revenue (Damodaran 2007). It is the return achieved from standard operations and does not include unique or one off transactions. ROS, sometimes referred to as operating margin, can be used both as a tool to analyze a company's performance against its past performance, and to compare similar companies' performances against one another. The ratio varies widely by industry but is useful for comparing different companies in the same business. ROS is calculated as:

ܴܱܵ ൌ ݊݁ݐ݌ݎ݋݂݅ݐ

ݏ݈ܽ݁ݏݎ݁ݒ݁݊ݑ݁

34

expensive equipment to generate income compared to others. Hence, the ROA of such companies will be lower than the ROA of companies which are low asset-intense. An increasing trend of ROA indicates that the profitability of the company is improving. Conversely, a decreasing trend means that profitability is deteriorating. ROA is calculated as:

ܴܱܣ ൌ ݊݁ݐ݌ݎ݋݂݅ݐ

ݐ݋ݐ݈ܽܽݏݏ݁ݐݏ

According to the literature review summary in Table 2, previous articles and studies within the

research area use mainly Tobin’s Q and ROE, ROA, ROS and ROIC as financial performance indicators.

4.2. Statistical analysis

We use statistical analysis to investigate if there is a correlation between the number of women on the board of directors and a company’s financial performance by studying the linear relationship between variables. We study the effect from independent variable defined as the percentage of women on the board of directors, with a set of control variables on different types of dependent variables (such as financial performance measures) by applying regression analysis.

4.2.1. Correlation analysis

Covariance is a measure of the linear relationship between two variables. A positive value indicates an increasing linear relationship. A negative value indicates a decreasing linear relationship (Newbold, Carlson and Thorne 2006).

The sample covariance is calculated as:

ܥ݋ݒሺݔǡ ݕሻ ൌσሺݔ௜െ ݔҧሻሺݕ௜െ ݕതሻ

݊ െ ͳ

Where xi and yi are the observed values and ݔҧ and ݕത are the sample means, and n is the sample

size.

Correlation is a measure of linear association between two variables. Values of the correlation coefficient are between -1 and +1. A correlation coefficient of either 1,0 or -1.0 indicates that two variables are perfectly linearly related, and a correlation coefficient of 0 indicates that there is no linear relationship between two variables. A correlation coefficient higher than 0.3 indicates a moderate to strong relationship, a correlation coefficient between 0.2 to 0.3 indicates a weak to moderat relationship and a correlation coefficient between 0.1 and 0.2 indicates a negligible to weak relationship and less than 0.1 indicates negligible relationship. (Newbold, Carlson and Thorne 2006)

35

ߩ ൌܥ݋ݒሺݔǡ ݕሻ

ߪ௫ߪ௬

Where ߪ௫ and ߪ௬are the variances of the two variables, x and y respectively.

4.2.2. Regression analysis

We study the effect from one or more independent variables on one dependent variable by applying regression analysis as described by Newbold et al (2006). We hypothesize a model of the relationship, and check if the parameters can be used to estimate the regression equation. We will use tests to determine the validity of the model as described below. If the model is deemed satisfactory, the estimated regression equation can be used to predict the value of the dependent variable given values for the independent variables.

Regression analysis is related to correlation analysis as follows: For simple linear regression, the sample correlation coefficient is the square root of the coefficient of determination, with the sign of the correlation coefficient being the same as the sign in the estimated regression equation. The linear regression model provides the expected value of the random variable Y when observed X variables take specific values. The actual observed value of Y for a given value of X is modeled as being equal to the expected values or population mean plus a random error, ε, which has a mean

0 and variance σ2_.

We will use multiple dimensional linear regression analysis; the estimated regression model is then given by the equation:

ݕ௜ ൌ ߚ଴൅ ߚଵݔଵ൅ ߚଶݔଶ൅ ڮ ൅ ߚ௡ݔ௡൅ ɂ௜

Where ߚ଴ and ߚଵ are the estimated values of the coefficients and ɂ௜ is the difference between the

predicted value of Y on the regression line.