The relationship between board diversity and firm performance under the moderate effect of ownership structure in Mainland China, Taiwan and Hong Kong

(1)

The relationship between board diversity and firm

performance under the moderate effect of ownership

structure in Mainland China, Taiwan and Hong Kong

Master Thesis

Business Administration – Track International Management

Name: Xinyi Wu

Student number: 11089121 Date: 23-06-2016

Thesis supervisor: dr. Ilir Haxhi Second reader: dr. Niccolo Pisani

(2)

Statement of originality

This document is written by Xinyi Wu 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.

(3)

Abstract

Despite an increase attention in the relationships between diversity of board and firm performance, still limited research that focuses on the influence of how diversity of board affects corporation performance has been done in East Asia, especially in Mainland China, Taiwan and Hong Kong. Previous findings toward this issue are inconsistent and mainly pay attention to developed countries and regions like North America, West Europe and Japan. Therefore, this thesis explores a deeper insight of this relationship in Mainland China, Taiwan and Hong Kong. In order to better reveal the relationship, we introduce the ownership

structure as our moderator. Also, we adopt two aspects of diversity to conduct our study: gender and independence. We support the view that board diversity and ownership structure can affect firm performance. By analyzing 400 listed companies spread over the three regions, we explore how the board diversity (i.e. female directors and independent directors) impacts on firm performance. We find the percentage of female directors has significantly positive influence on firm performance. Furthermore, the percentage of independence directors and ownership show a positive moderating effect on the relationship.

The thesis has several contributions. First, our study extends current literature as it provides a more comprehensive picture of these relationships in the three regions. Second, we

investigate the unexplored moderating effect of ownership. Third, from a managerial perspective, we provide valuable data support that diversity can influence firm performance under ownership structures in these regions.

(5)

1. Introduction

Corporate governance deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment (Shleifer & Vishny, 1997). Daily et al. (2003) defines corporate governance as “the determination of the board uses to which organizational resources will be deployed and the resolution of conflicts among the myriad participants in organizations.” This definition has some differences on former literature, which, for instance, in general consider corporate governance as “the process by which companies are directed and controlled (Cadbury, 1992).” and suggests, “The participants are the shareholders, the management and the board of directors (Monks & Minow, 2011)”. The reason for the emergence of various definitions of corporate governance is simple. These scholars view this issue from different perspectives as well as their distinct backgrounds and analysis methods. In this study, we consider corporate governance as the foundation of the research. From these studies, we can conclude that the mechanisms of governance includes informal governance, regulations, ownership, incentive pay, and boards. Among them, boards diversity is a significant topic in corporate governance.

According to Adams et al. (2015), board diversity can show both challenges and

opportunities. Diversity increase the board’s ability of monitoring through enhancing its independence (Adams & Ferreira, 2009), providing better oversight of firm’s disclosures and reports, and improving the quality of public disclosure (Gul et al., 2011). However, diversity can present negative effects as well. For example, Adams et al. (2015) think diversity may increase decision-making costs in boards, and generate conflicts and factions.

(6)

Among all the dimensions of diversity, gender diversity is one of the most popular topics. But the fact is that women are underrepresented in many occupations and industries (Rubin, 2000). Adams and Ferreira (2009) find that female directors are more likely to attend board meetings than male directors. Some countries have gender quota for female directors. In allusion to this situation, however, Ahern and Dittmar (2012) find that this mandatory quota is negatively associated with firm performance. But we cannot just conclude that diversity will harm firm performance. Triana et al (2014) suggest that diversity is double-edged because it can improve or block strategic change depending on firm performance and female directors’ power. Despite there has been mixed evidence regarding the effect of board

diversity on performance, diversity in board composition is still considered favorable based on these two important reasons (Kang et al., 2007). The second topic in our study is board independence. As Cotter et al. (1997) claim, independent directors play a significant role in controlling and enhancing shareholder wealth. Also, in O’connell & Cramer (2010), Ramdani & Witteloostuijn (2010) and Vafeas & Theodorou (1998) all prove that board independence can positive influence firm performance.

However, the gap in existing literature is they mainly focus on American market or other developed markets like Japan and Australia. Little researches examine the connection

between diversity of board and performance in Asian emerging markets. Also, most research on this issues only researches one country or region (Cheung et al., 2010, Low et al., 2015). Therefore our first research question is:

(7)

in Mainland China, Hong Kong and Taiwan?

Through addressing this question, we are able to build a primary and comprehensive framework to clearly reveal the relationship between diversity of board and company performance in these regions. However, it is not enough to only examine this direct

relationship. According to the literature on diversity, the degree of this relationship will vary from countries with different ownerships (Liu et al., 2014). Thus, to find out if ownership will influence this relationship, we consider ownership structure as our moderator for several reasons.

First, ownership is also a crucial topic in corporate governance area since ownership directly determines who enjoys the most powerful authority to make decisions in the firms (Zattoni, 2011). Empirical studies (Zattoni, 2011; Kang & Kim, 2012) have proven the different ownership structures will have various impacts on firm performance. For example, individual-owned firms in Hong Kong and Taiwan could result in conflicting personal interest and being a burden on financial performance (Khanna & Rivkin, 2001). Second, firms in these three different regions all have unique ownership structures. For instance, in China, board diversity in legal person-controlled firms will have more positive and

significant influence (Liu et al., 2014). As a result, we assume that the different ownership structures may pose positive or negative moderating effects on various types of firms. In our dataset, there are five different ownership structures: state-owned, individual-owned,

corporate-owned, foreign-owned and other types. To better examine the moderating effect, we divide the ownership structure into two parts depending on the percentage of sharing:

(8)

majority structure and minority structure. We will elaborate this division in method part. Here is our second question:

What is the moderating effect of ownership structure on the relationship between board diversity and firm performance in Mainland China, Hong Kong and Taiwan?

Through introducing this moderator, we establish a framework to find out which diversity and ownership structures are vital factors in influencing firm performance by taking a quantitative approach. Thus, the purpose of this study is to empirically focus on how the board diversity relates to the firm performance. Second, we test the moderating effect of ownership structures on the relation between diversity and firm performance. We will analyze the data of top 100 listed companies collecting from exchange markets in China (Shanghai and Shenzhen), Taiwan (Taipei) and Hong Kong. All data is gathered through companies’ annual reports and stock markets and then we recoded the data according to a coding scheme. In this study, we will first test the relationship between female directors and firm performance. Then, we test the relationship between independence directors and firm performance. Finally, we test the effect of ownership structure on the relation between the two diversities and firm performance separately.

The thesis will contribute to the studies on diversity and firm performance in several ways. First, we contribute to the current literature by going more in depth in providing a more comprehensive picture of how gender and independence affect firm performance in Mainland China, Hong Kong and Taiwan. Second, we investigate the unexplored moderating effect of ownership. Third, from a managerial perspective, we provide valuable data support that

(9)

diversity can influence firm performance under ownership structures in these regions. CEOs in these regions should keep these elements in mind when they want to improve the firm performance from the firms inside. Furthermore, the moderating effect of ownership structure provides the multinational perspectives in which factors can affect the firm performance more.

This study starts with a detailed literature review with the main concepts of corporate governance, gender diversity, independence, education backgrounds, firm performance and some other related variables like control variables. Second, the theoretical framework will explain the theoretical backgrounds of our study. Third, the research method is discussed with an overview of all the variables we need to take into account, the data sample we use and the method of analysis. Subsequently, the regression analysis will be presented. Hereafter, we will discuss and interpreter the results as well as the implications for further researches. Finally, the study will make the conclusions with an overview of the whole work.

(10)

2. Literature Review

2.1 Corporate governance

Corporate governance is a set of customs, policies and institutions designed to reduce or eliminate the principal–agent problem (Su & He, 2012). According to Aoki (2001), corporate governance concerns “the structure of rights and responsibilities among the parties with a stake in the firm”. Shleifer and Vishny (1997) think that corporate governance deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment. Daily et al. (2003) define corporate governance as the determination of the broad uses to which organizational resources will be deployed and the resolution of conflicts among the myriad participants in organizations.

Further, Thomsen and Conyon (2012) claim that there are several mechanisms of corporate governance in which some are more important than others. These mechanisms are

stakeholder pressure, regulation, ownership structure, informal governance, board and incentive systems. In this thesis, we will mainly focus on board diversity in Mainland China, Hong Kong and Taiwan.

2.2 Development in China, Hong Kong and Taiwan

Mainland China is an emerging market compared with Western markets. China started its economic reform since 1978. In the early 1990's the transition to a market economy began with the liberalization of business practices, the privatization of small and medium-sized state owned enterprises, and the preparation of large institutions for private ownership (Liu et al., 2011). In 1992 there were only 50 listed companies in China, but in recent years China has

(11)

experienced enormous changes in its capital markets (Shan &Round, 2012). According to the 2012 Yearbook of China Securities and Futures, at 2012 year-end, there were 2484 firms listed on the main boards of the two exchanges, with a total market capitalization of about US$3.8 trillion (Jiang & Kim, 2015). According to official statistics, economic growth has averaged 9.5% over the past two decades and the trend is expected to continue for some time (Cheung et al., 2010).

Hong Kong is another example for several reasons. First of all, Hong Kong enjoys some unique features. As the international financial center of the Asia–Pacific region, Hong Kong combines an Asian family-controlled business environment with Anglo-Saxon legal and corporate governance system (Cheung et al., 2013). Second, the 1997 Asian financial crisis that swept through most of East Asia highlighted the need for corporate governance reform in the region. There are numerous efforts in corporate governance reforms at various levels (Cheung et al., 2010). More importantly, Cheung et al. (2010) think that in the past two decades there has been a rising interest in investing in China, which has demonstrated a huge economic growth potential. International fund management houses have been launching investment products that invest in the Greater China region (including China, Hong Kong, and Taiwan). As we mentioned above, at this time, China was still closed to foreign

investments. Thus, those investments had to invest Chinese enterprises through Hong Kong exchange markets. Last but not the least, Hong Kong has a special position from a history perspective. Due to some historical issues, the United Kingdom had ruled Hong Kong for nearly one hundred years. British set a series of Western corporation rules in Hong Kong that

(12)

made Hong Kong become a financial center for a long time in East Asia. To sum up, Hong Kong has some special places to be chosen as an example that will make the study more logical.

We also use Taiwan as an example for its unique features. Yang et al. (2012) say that public companies in Taiwan typically have family members holding a decisive proportion of company shares. Solomon et al. (2003) support this augment by claiming that founding family members run the majority of companies directly or indirectly, even though these companies may be listed on the stock market. Thus, when we study how gender diversity in BoD will influence corporate performance, Taiwanese corporations can be another a good analysis aspect.

2.3 Board gender diversity and firm performance

There is a large body of literature studies about board gender diversity. Scholars have given different explains towards this issue. Gul et al. (2011) demonstrate that stock prices of firms with gender-diverse boards reflect more firm- specific information after controlling for corporate governance, earnings quality, institutional ownership and acquisition activity. Jurkus et al. (2011) build a model including Fortune 500 firms, finding the connection

between gender diversity in top management and agency costs. From a survey by Carter et al, (2010), the gender diversity of the board and firm financial performance appear to be

endogenous from a sample of major US corporations.

The extant literature offers at least two general conflicting perspectives regarding the relationship between diversity and group performance (Erhardt et al. 2003). Some like

(13)

Francoeur et al. (2008) indicate that firms operating in complex environments do generate positive and significant abnormal returns when they have a high proportion of women officers. Mateos de Cabo et al. (2012) use a large sample from 20 European countries to support the perspective that high level of gender diversity can have a positive effect on corporation performance. Simons and Pelled (1999) think both educational level and cognitive diversity were associated with positive effects on organizational performance. In contrast, some scholars hold the view that board gender diversity could be a disadvantage in terms of firm performance. For instance, Darmadi (2011) say that in Indonesia, female director appointment may be driven more by familial relationships rather than occupational expertise and experiences, thus leading to a decline in firm performance. From Erhardt et al. (2003) we can conclude that heterogeneous teams were slower in their actions and responses and less likely than homogenous teams to respond to competitors’ initiatives. Knight et al. (1999) also find that demographic diversity was negatively related to consensus.

However, most of gender diversity literatures mainly focus on developed regions, like North America and Western Europe, quite a few make good analyses on corporations in Mainland China, Hong Kong and Taiwan. The main reason that accounts for the lack of literature about China can be its special and mixed institutional environment. The strong political connection between the governments and listed firms and the lack of a truly independent judicial system negatively impacts on the efficiency of corporate governance mechanisms (Yang et al., 2011). They argue that there are many problems in Chinese corporate governance. The major

(14)

developed countries are less effective in China. Another reason might be Chinese mixed ownership structure. From Kang and Kim (2012) study, we can find that economy market reform in Mainland China is not clear-cut and thoroughly. That is why Mainland China nowadays can allow state-owned corporations and private-owned corporations to exist simultaneously.

2.4 Board independence and firm performance

As another hot topic in corporate governance aspect, board independence also attracts numerous attentions. The most acceptable definition of board independence is that an independent director is seen as the one who has never worked in the company or any other subsidiaries, is not related to any employees (Ravina & Sapienza, 2010). However, according to previous literature, if the independent directors can pose significant influence on firm performance is still in debate. There are contradictive opinions towards this issue. On one side, Ghosh and Sirmans (2003) use ROA and ROE to test if board independence will influence firm performance and the result shows that the percentage of outside directors will have positively impacts on firm performance. Mishra and Nielsen (2000) relate the

independent directors, pay-related incentives and firm performance. They use 100 largest commercial bank holding companies as their dataset to successfully prove that board independence has positive relationship with firm performance. Similarly, the percentage of independent directors is associated with the cost of debt financing, i.e. firm’s financial performance after adopting the sample of top 500 firms in the world (Anderson et al., 2004). They provide the market-based evidence that board composition is important with firm’s

(15)

financial reports. Beside this, scholars have conducted numerous researches based on the sample of North-American companies, European companies and Japanese and Australian companies, which are all from developed regions in the world (Bhagat & Black, 2000; Vafeas & Theodorou, 1998; Kang et al., 2007). These regions have sound market regulations, laws and systems. Among these scholars, those who support the view that independence will positively influence performance claim some companies in developed countries tend to have a conservation diversity profile, which means the situation in these companies correspond to the positive hypothesis. In addition, there are also some but very few studies focusing on emerging countries like China, South Korea and Southeast Asian countries. Liu et al. (2015) use a dataset covering nearly all listed companies in Shanghai and Shenzhen stock exchange markets from 1999 to 2012. Thus their result is relative representative. They find that board independence have an overall positive effect on firm operating performance. Also, scholars like Ramdani and Witteloostuijn (2010) show the positive relationship to support the argument based on the sample of Southeast Asian countries like Indonesia, Malaysia and Thailand.

On the other side, there is also some existing literature that supports the argument that

independence will have little relationship even negative relationship with firm performance. Bhagat and Black (2000) conclude that greater board independence will not improve firm performance after conducting the quantitative research in large American companies. Mishra and Nielsen (2000) test board independence in banking industry. They find that there are no significant differences in the effect of board independence on banks’ performance. Moreover,

(16)

Lefort and Urzúa (2008) use a four-year, 160 listed companies to show the conclusion that increasing independence affect company value, but this effect is negative. Fernandes (2008) studies the Portuguese case, casting doubt on the effectiveness of independent board member. He puts forward the opinion that few outside directors will actually cause fewer agency problems and achieve better results in align with shareholder’s interests.

2.5 Moderator-Ownership structure

Firms in Mainland China, Hong Kong and Taiwan have different ownership structures. Some researches before have been done to reveal the relationship between ownership structure and firm performance. As Shleifer and Vishny (1994) state in their study, ownership structure can influence firm’s corporate governance. In China, stated-owned enterprises (SOEs) are the majority of China’s listed firms. State owners control these firms. According to Xu and Wang (1999), a large number of listed companies in China have three main groups of

shareholders-the state, legal persons (institution) and individuals. If these mixed ownership structures can positively affect firm performance in China is still in debate. On one side, Shleifer and Vishny (1996) are clearly against the positive view by saying these mixed ownership will undermine the firm performance. Besides, too much political pressure in firms will influence directors to make business plans. Moreover, shareholders should do their rights in monitoring boards. However, the mixed ownership in China will create difficulty for shareholders (Kang & Kim, 2012). On the other side, some scholars think that

state-ownership can have positive effects. Jefferson (1998) claim state-owned firms can push boards to make quick decisions instead of blocking decision-making process. Also, Stiglitz

(17)

(1996) and Sun et al. (2002) start their researches based on the reality of China and conclude the same results, arguing that state-ownership is able to facilitate the difficulties during the decision-making process. Thus, in Mainland China, the ownership reform attracted much attention.

Similarly, Hong Kong and Taiwan are both representing one typical ownership structure in the listed companies-family owned (Chu, 2009). From previous studies, we find that in

Taiwan, firms with long history are more likely to be family controlled (Stijn & Simen, 2000). Also, as Cheung et al. (2010) mention in their study, almost 90% listed companies in Hong Kong have a major shareholder who by himself or family members own. In this situation, the tradition agency problem will not work in that management is mixed with ownership

(Cheung et al., 2010). Scholars like Demsetz and Lehn (1985), Himmelberg et al. (1999) conclude that family ownership shows no positive and significant relationship with firm performance. This result is in line with Chen et al. (2005), who conduct his research with a sample of 412 listed companies. In contrast, some other scholars like Anderson and Reeb (2003) investigate the relationship between family ownership and firm performance and find out there exists a positive relationship.

Based on the literature above, we conclude that majority ownerships can positive or negative influence on firm performance. Thus, we consider ownership as our moderator to see if majority ownership or minority ownership will have moderating effect on the relationship.

(18)

3. Theoretical framework

3.1 Theory backgrounds

3.1.1 Agency theory

Agency theory is the theoretical framework most often used by investigators in finance and economics to understand the link between board characteristics and firm value (Carter et al., 2003). An agency relationship is a contract under which one or more persons (the principals) engage another person (the agent) to perform some service on their behalf that involves delegating some decision-making authority to the agent (Jensen & Meckling, 1976). Agency problems occur when managers do not have shareholders' best interest in mind when making corporate decisions (Liu et al., 2014). To reduce agency costs, Fama and Jensen (1983) argue that board as a mechanism can control and monitor managers. With this monitoring role, the board will protect shareholders’ benefits. Previous literature shows that diverse boards can be more effective in monitoring (e.g. Adams & Ferreira, 2009). Scholars like Gul et al. (2011) support this view by saying that if firms have more diverse board; they will be able to offset the over-monitoring disadvantage. In our case, we consider two aspects of diversity, female and independence. As Fama and Jensen (1983) say in their study, female directors with doctor are more likely to bring in additional perspectives, which can be useful for firm performance. Thus, considering their studies are based on developed countries with high quality of corporate governance and the unique institution environment in Mainland China, Hong Kong and Taiwan, we tend to believe that board diversity will have positive effects on

(19)

firm performance.

3.1.2 Resource dependence theory

This theory focuses on the interdependence between the organization and the external environment (Zahra & Pearce, 1989). According to Pfeffer and Salancik (1978), businesses depend on the resources in their external environments to survive and these dependencies pose risks to the businesses. Thus, one of the main tasks of the board is to build link between firms and external environment. To our topic, because women have different external

connections to the environment, Carter et al. (2010) predict that they will not have the same effect on board functions and, ultimately, firm performance. In summary, female directors can be better equipped to connect their firms to female customers, women in the labor force and society at large due to their different life experiences and perspectives (Liu et al., 2014). For independent variable, as Pfeffer and Salancik (1978) mention in the study, resources can help reduce dependency between the organization and contingencies. Meanwhile, Williamson (1984) also says resource dependence can help companies to lower the transaction cost. Finally, from the view mentioned in Gaur et al (2015), directors with higher degrees can have more access with external resources. They have more opportunities to get in touch with those who have potentials to help directors during their student careers. Thus, resource dependence theory helps us to explain why our three independent variables can have influence on firm performance.

3.2 Hypotheses

(20)

discussed above. We will link the hypotheses to our research questions: What is the relationship between diversity within the board of directors and firm performance in Mainland China, Hong Kong and Taiwan and to what extent is the relationship between diversity and firm performance moderated by ownership structure in Mainland China, Hong Kong and Taiwan.

3.2.1 Board gender diversity

The literature has contradicting views towards whether increasing gender diversity can be useful for firm performance. From agency theory, we can see that Carter et al. (2010) claim that gender diversity does not have clear effect on firm performance. On one side, Gul et al. (2011) test firm performance by calculating the stock price and they find the more gender diverse boards can show better firm performance. Moreover, many other scholars also

conclude the positive relationship between female directors and firm performance (Francoeur et al., 2008; Simons & Pelled, 1999; Erhardt et al., 2003; Jurkus et al., 2011; Mateos de Cabo et al., 2012). They all claim the increasing the percentage of female directors can bring more information, points of view and unique thinking. On the other side, Darmadi (2011) takes Indonesia as an example, proving female director appointment will lead to a decline in firm performance. Similarly, some other scholars conclude the negative relationships after using examples in the US, Europe and Japan (Erhardt et al., 2003; Knight et al., 1999).

In our study, we assume that board gender diversity can affect firm performance positively. The resource dependence theory suggests the positive relationship between gender diversity and firm performance (Liu et al., 2014; Hillman et al., 2000). According to Bear et al. (2010)

(21)

gender diversity can provide boards with effective management skills, which can enhance firm performance. First, female directors are more likely to have expert backgrounds outside of business and support specialists and community influential (Hillman et al., 2002). Second, researches also show the firms with more female directors can in fact have more charitable activities, which can improve firms’ reputations, thus improve firm performance (Wang & Coffey, 1992; Williams, 2003; Bear et al, 2010). In addition, increasing female directors can enhance decision-making and stimulate more effective communication among board

members (Bear et al. 2010).

However, as we discussed above, these studies mainly focus on developed countries and some Southeast countries. Though some researches like Jiang et al. (2015); Liu et al. (2011); Liu et al. (2014) have investigated board gender issues in China, they fail to take Hong Kong and Taiwan into account, which are the other two significant economies in East Asia. For instance, Liu et al. (2014) test the degree of the positive effect on the relationship between gender diversity and firm performance in China. They first use regression model to prove that female directors can positively influence the firm performance. Then, they divide the sample into several groups based on the numbers of female directors in boards. The result is still positive. In summary, based on these studies, we can find that board gender diversity does have impacts on firm performance. Thus, we proposes the following hypothesis:

H1: The percentage of female directors will positively influence on firm performance in Mainland China, Hong Kong and Taiwan.

(22)

Apart from this, we also need to consider ownership structures to check the moderating effect on the relation between gender and firm performance. These three regions in our study have different and unique ownership structures. The majority ownership in China is state

ownership (Liu et al, 2014). Xu and Wang (1999) clearly write in their study that a large amount of listed firms in China have three main groups of shareholders. According to Kang and Kim (2012), Mainland China, for example, does not conduct its economy market reform clearly and thoroughly. Thus, Mainland China allows firms with various ownership structures to exist at the same time. As Liu et al. (2014) mention in their research, these structures have great influence on boards when they are making decisions. However, Jefferson (1998) think in state-owned firm, government is able to help boards to make quick and accurate decisions in that government can have a border and more comprehensive understanding of the whole industry. Similarly, the majority ownership in Hong Kong and Taiwan is family ownership (Chu, 2009; Yang et al., 2012). Cheung et al. (2010) think the family ownership can affect firm performance in Hong Kong from the perspective of agency theory. In contrast, Demsetz and Lehn (1985) conduct a research and find no positive relationship.

In this study, we do not consider ownership as an independent variable. Instead, we view ownership has moderating effect on the relationship between gender diversity and firm performance. Based on the literature and theories above, we assume no matter which places, the majority will always have negative or positive influence on firm performance.

(23)

this relation through affecting the decisions from the board.

Combining this assumption above with agency theory and resource dependence theory, we can presume that majority or minority ownership has moderating effects on the relation between gender and firm performance. For example, in firms with majority ownership, tradition agency problem will not exist because there is no separate management and

ownership (Cheung et al., 2010). Firms are able to decide inside how many female directors they need. Also, from the resource dependence perspective, if these firms consider potential female directors will not help them to connect with female clients, firms may even not hire any female directors (Liu et al., 2014). Since we have assumed that female directors can impact on firm performance. Thus, these actions will inevitable influence firm performance. In other words, majority or minority ownership has moderating effect. Therefore:

H3a: Majority ownership structure will influence the relationship between the percentage of female directors and firm performance in Mainland China, Hong Kong and Taiwan. H3b: Minority ownership structure will influence the relationship between the percentage of female directors and firm performance in Mainland China, Hong Kong and Taiwan.

3.2.2 Board independence

Independence is another hot topic in corporate governance. Similarly, there are also two opposite views in previous literature. Ghosh and Sirmans (2003) prove the positive

(24)

get the conclusion that independent directors can reduce the cost of debt in commercial bank holding companies. Anderson et al. (2004) provide market-based evidence to prove the positive relationship. Besides the literature, we can also find evidence in theory backgrounds to support the positive opinion. Hillman and Dalziel (2003) say in the research independent directors can monitor the board effectively. This will help firm to reduce agency inherent in the separation of ownership and control and, in this way, improve firm performance (Fama, 1980). Also, resource dependence can help explain the relationship. Independent directors can show more accesses to outside resources, help firm to lower the transaction cost

(Williamson, 1984) and then improve firm performance. Others like Bhagat & Black (2000); Mishra & Nielsen (2000); Lefort & Urzúa (2008) and Fernandes (2008) argue the board independence has little relationship with firm performance based on the researches in America, West Europe and Japan.

In our study, we also assume the positive relationship between board independence and firm performance. As mentioned above, we make this assumption for several reasons. First, as Liu et al. (2015) suggest in their study, independent directors in China can be easily affected to make decisions by governments due to the unique ownership structure, which is our moderator. Second, independent directors’ access to outside resources help the boards to make more comprehensive decisions. Thus, we take Mainland China, Hong Kong and Taiwan as our samples to investigate the influence of board independence on firm performance. Therefore:

(25)

H2: The percentage of independent directors will positively influence on firm performance in Mainland China, Hong Kong and Taiwan.

Also, we need to take moderating effect into account. As mentioned above, the majority the majority ownership in China is state ownership (Liu et al, 2014). Shleifer and Vishny (1996) think Chinese mixed ownership will undermine firm performance. Other scholars like Stiglitz (1996) argue state ownership can facilitate the difficulties during the decision-making process. In the same way, if the majority ownership-family ownership is beneficial to firm

performance is still in debate. Stijn and Simen (2000) support the positive relationship by researching companies in Taiwan. However, some other scholars like Chen et al. (2005) do not agree this positive relationship after researching 412 listed companies.

Like we have explained above, we consider ownership as a moderator. We assume in these three regions, majority and minority ownership can have moderating effect on the

relationship between independent directors and firm performance. Independent directors have some unique advantages. First, they can effectively monitor other board members in case corruption or other problems. Second, from resource dependence theory, independent

directors have ability to reduce resource dependency between organization and contingencies because they have accesses to outside resources (Pfeffer & Salancik, 1978). Thus,

shareholders can affect this relation through impacting the decisions of hiring independent directors.

(26)

ownership has moderating effects. As Cheung et al. (2010) claim, there is no separate management and ownership existing in firms with majority ownership. Also, Li et al. (2008) consider majority ownership as a moderator too. They use 404 companies to find out there is moderating effect on the relationship between independent directors and firm performance. Therefore:

H4a: Majority ownership structure will influence the relationship between the percentage of independent directors and firm performance in Mainland China, Hong Kong and Taiwan. H5b: Minority ownership structure will influence the relationship between the percentage of independent directors and firm performance in Mainland China, Hong Kong and Taiwan.

3.3 Conceptual framework

Majority ownership Firm performance Women Independence Minority ownership H1 H2 H3a H3b H4a H4b

(27)

4. Data &Method

4.1 Sample and data collection

This study uses a quantitative study to analyze the relationship between three aspects of board diversity and firm performance. Then we investigate the influence of board diversity in these three regions under moderation effect of ownership structure. For China, most listed firms are privatized former state-owned enterprises (SOEs) with very unique ownership structures controlled by either state owners or legal person owners (Liu et al., 2014). For Hong Kong, it represents like a well-developed financial market with full infrastructure and regulations. Besides, the majority of most Hong Kong listed firms is family-owned (Cheung et al., 2013). At last, for Taiwan, family groups also control most companies. In our database, the family-owned companies in Taiwan and Hong Kong are mainly foreign-owned and individual-owned. Due to uniqueness ownership structures, we consider it as our moderator. The thesis will contribute to the studies on gender diversity and firm performance in several ways. First, our study extends the current literature by providing the empirical data evidence on board diversity and firm performance based on the data from three regions. Second, we add evidence to the literature that gender diversity, percentage of independent directors and educational level of directors can be beneficial to firm performance in these regions. Third, we use data analysis to examine if board diversity can influence firm performance under the moderation effect of ownership structure.

In this study, we collect data from top 100 companies’ 2014 fiscal reports from four stock exchange markets (Shanghai, Shenzhen, Hong Kong and Taipei) and Orbis database

(28)

respectively. In order to see the degree of moderation effect, we divide the whole sample into two sub-samples depending on their ownership type when testing moderators. We will use Tobin’s Q and ROA (return on assets) for calculating firm performance, which will be elaborated in detail below.

4.2. Variable Construction

4.2.1 Dependent variables

In this study, we will use FirmPerformance as our dependent variable. Two dimensions of firm performance will be used to determine the variable. The first one is ROA, which means the return on assets. ROA is the percentage of how profitable a company’s assets are in generating revenue. The formula of ROA is listed as follows.

ROA=Net Income/Average Total Assets

The second one is Tobin’s Q, which is the ratio of a firm’s market value to its book value. The formula of Tobin’s Q is listed as follows.

Tobin’s Q=Total Market Value/Total Asset Value

4.2.2 Independent variables

Our independent variable is board diversity. In this case, we choose three dimensions to illustrate the relationship between board diversity and firm performance. The first

independent variable is gender diversity. We name this variable “Women” in our study. In our dataset, we find that most listed companies have female directors, which are represented as the percentage. So we convert the percentage into decimal fraction low than 1. The second

(29)

one is the percentage of independent directors in the boards. We define this variable “Independence”. Independent directors are usually those who are not employees of the corporation (Cotter et al., 1997). Lipton and Lorch (1992) view independent directors as those who have no connection with the company. These two viewpoints are nearly the same essentially.

In conclusion, in this study we will use variables-Women, Independence as our independent variables.

4.2.3 Moderator

The moderator in the thesis is the ownership structure. In our dataset, there are four types of ownership structures. Thus we create a dummy moderator with majority and minority ownership (1: > 50%, 0: < 50%). To be clear, if there is a majority ownership in specific company, then the variable ownership structure for this company will equal to 1, interpreting as Majority. On the contrast, the dummy variable will equal to 0, which means the Minority

4.2.4 Control variables

In this study, we use Firmsize. According to Fama and French (1992), firm size is related to firm’s financial performance. Some other studies also show that total asset size has

relationship with Tobins’Q (Chung & Pruitt, 1994; Prevost et al., 2002; Yermack, 1996). Therefore, in this study firm size will be calculated as Log (Total Asset).

The second control variable is Boardsize, which reflects as number of board members. Boardsize was also used as control variables in other studies (Erhardt et al., 2003; Adams &

(30)

Ferreira, 2009). Besides, previous researches like Yermack (1996), Jackling and Johl (2009) and Pearce and Zahra (1992) find that board size has positive relationship with firm financial performance. Therefore, we consider board size as our second control variable with

theoretical support.

The final control variable is Industry. Coombs and Gilley (2005) suggest in their study that industry types can influence firm performance. We choose industry type as our control variables for two reasons. First, the numerator of the Tobin’s Q shows the total market value, which includes the intangible assets. Some knowledge intensive companies therefore own more intangible assets than others. Therefore, industry type needs to be controlled for. Second, we can see that ROA uses total assets. In our dataset, some firms have more assets than other (like petroleum industry). Based on the tow reasons above, we divide all firms into six categories by looking at firms’ NAICS code. These industries are mining (industry 1), manufacturing (industry 2), transportation, communication, electrics and gas (industry 3), wholesale, retail and trade (industry 4), finance, insurance and real estate (industry 5) and all service-related industry (industry 6).

4.3 Method

The relationship between the independent variable board diversity and the dependent variable firm performance with a moderating effect of ownership structure can be explained by the following equation:

! = #$+ #&'∙ )&+ #&'∙ #*'∙ )*+ #&'∙ #+'∙ )++ ,

(31)

ROA and Tobin’s Q. #_$ is the intercept and #_&' stands for our two independent variables, Women and Independence. #&'∙ #*' represents the interaction effect between independent

variables and majority ownership. #&'∙ #+' represents the interaction effect between

independent variables and minority ownership. Finally, the , is the difference between estimated )_&' and actual )_&' (Field, 2009).

To test theoretical framework, we conduct different linear regressions. At first, we use a simple regression to examine the relationship between the firm performance and board diversity. The regression is tested for seven models (see Table 1), which were organized in hierarchical order. Model 1 tests the relationship between ROA or Tobin’s Q (firm

performance) and Boardsize, Firmsize and Industry without any independent variables and moderators. The second and third model tests the relationship between independent variables and dependent variables by adding one independent variable at one time. In the fifth, sixth and seventh model, two moderators are tested to examine moderating effect. In the final

(32)

Table 1: Summary of Regression Models 8 1 3 2 3 6 3 2 71 8627 712 27127 2 48 3 378 3 8125 8125 8125 8125 8125 8125 8125 8125 87 85 3 52 712 27127 3 52 8125 812 8

(33)

5. Results and analysis

In this part, we will present the statistical analyses of the data. First, we will start with an overall explanation of the descriptive statistics, corrections between dependent variables, independent variables, moderator and control variables. Second, we will present regression analyses described in the method section. The results will be used to test the proposed hypotheses of this study.

5.1 Descriptive Statistics Analyses and Correction test

The descriptive statistics of variables is shown in Table 2. This table provides us with the mean, standard deviation and correlation. For ROA, the difference between maximum and minimum is not high. This means the data are concentrated. The mean of ROA in our data is 0.0699. More important, the table shows a standard deviation value for degree of ROA of 0.06717. Therefore it can be concluded that in general, the degree of dispersion is not very high. For Tobin’s Q, the difference is not too high as well. The mean of Tobin’s Q is 3.4862. Meanwhile, the standard deviation is 6.60164, which is suitable for normal analyses.

Therefore, in the following data analysis, we can use ROA and Tobin’s Q as our dependent variables to calculate results and test hypotheses.

For Women, the independent variable, the mean of it is 0.0951, which means the percentage of female board members is still low in big listed firms. For Independence, the second

independent variable, the mean of it 0.438, which means there are nearly half board members in big listed firms are independent directors. For control variable, Firmsize, the mean is 98659, implying that the companies in our dataset are all powerful companies. Moreover, the

(34)

mean of Boardsize is 10.45. This means there are enough board members in each company for our analysis.

The mean of Majority, our moderator variable, is 0.591 and the standard deviation is 0.49226, which means in majority sample, there is slightly more companies with majority ownership structures than those with minority structures. Similarly, the mean of Minority, the other moderator, is 0.4115 and the standard deviation is 0.49272. That is to say in minority sample, there is a little bit more companies with minority structures than those with majority

structures. The number of companies with majority ownership is nearly the same with that of companies with minority ownership.

Moreover, Table 2 includes the correction test between dependent variables, independent variables, moderator and controls variables. Normally, there are two methods to test multicollinearity. The first one is the correction test shown in Table 2. There are no

corrections of variables above 0.8. Thus, there is no multicollinerity between variables. The second method is to use Variance Inflation Factor (VIF) test. VIF can show the linear relationship between predictors. If the value of VIF is below 5, as Table 3 and 4 show, it means there is no multicollinearity. Thus, multicollinearity is not a problem in our study. As shown in Table 2, Women has a positive but not significant influence on Tobin’s Q (0.019) and a quite positive influence on ROA at 1% level, which indicates that when the numbers of female directors increases, firms will perform better. Second, Independence has a negative but not significant influence both on ROA and Tobin’s Q. This means in our dataset, when firms have more independent directors will pose some negative impact on firm performance.

(35)

Beside, Majority shows a negative relationship between ROA (-0.040) and Tobin’s Q (-0.047), meaning that companies with majority are not certainly beneficial to firm performance. But Majority has a significant correlation with Firmsize. This indicates that when firm enlarges its size, it is more likely to adopt majority structure. Minority and ROA, Tobin’s Q have a

positive but not significant correlation. This indicates that when a firm has minority ownership, it have more chances to positively influence firm performance. Minority also negatively correlates with Firmsize. This result proves our explanation above from the other side. What’s more, the first two control variables-Boardsize and Firmsize-correlates

negatively significant both with ROA and Tobin’s Q. This means just increasing the number of directors or company scale will not lead to better outcomes for companies.

Table 3: Multicollinearity statistics for dependent variable ROA

Table 4: Multicollinearity statistics for dependent variable Tobin’s Q

Tolerance VIF

Women 1.000 1.000

Independence 1.000 1.000

1

a. Dependent Variable: ROA

Model Collinearity Statistics

Tolerance VIF

Women 1.000 1.000

Independence 1.000 1.000

1

a. Dependent Variable: TobinsQ

(36)

Table 2: Descriptive Statistics and Correction Table

N Mean SD ROA TobinsQ Women Independence Boardsize Firmsize Industry Majority Minority

ROA 400 0.0699 0.06717 TobinsQ 400 3.4862 6.60164 .236** Women 400 0.0951 0.10392 .136** _.019 Independence 400 0.438 1.52641 -.053 -.021 -.014 Boardsize 400 10.45 3.29464 -.121* _-.051 _.011 _.039 Firmsize 400 9.8659 0.61707 -.295** _-.336** _-.109* _.096 _.327** Industry 400 2.945 1.38085 -.022 -.044 .061 .078 .061 .106* Majority 400 0.591 0.49226 -.040 -.047 -.058 .056 .084 .159** _.004 Minority 400 0.4115 0.49272 .037 .044 .062 -.056 -.083 -.156** _.004 _-.995**

*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

(37)

5.2 Regression Analyses

Results of regression analyses between the dependent variables, independent variables and moderator can be found in the following tables. We have two dimensions of dependent variables, ROA and Tobin’s Q. Therefore, we will continue our test based on the order of dependent variables. The moderators are Majority and Minority, which are divided by 50% ownership. To test the moderator effect of ownership structure, an interaction variable between moderator and independent variable (one of the three) is created, by multiplying both variables. Thus, the moderating effect in our case is called “Women

(Independence)·Majority (Minority)”. In order to more clearly show the moderating effect, we introduce the Dawson document, which is used to test two-way interactions. After

running the regression, we need to input the regression coefficients to get the accordingly line charts (Dawson & Richter, 2006).

5.2.1 Dependent Variable-ROA

In this case, the dependent variable is ROA and the independent variables are Women,

Independence. As shown in the Table 5, the analysis was run in a stepwise approach with first only the control variables; afterwards the independent variables were added. We assess whether the hypotheses are accepted or not by viewing the significant values. Hypothesis 1 claims that female directors will influence firm performance positively. In the Table 5 Step 2, the relationship between Women and ROA is significant (! = 0.106, ( < 0.05). Based on the results of the regression analysis, Hypothesis 1 is accepted. Therefore, we can get the conclusion that when a firm owns more female directors in board, the firm performance will

(38)

be better. In the Table 5 Step 3, the relationship between Independence and ROA is not significant (! = −0.024, ( > 0.05). Based on the results of the regression analysis,

Hypothesis 2 is rejected. This means for big listed firms in Mainland China, Hong Kong and Taiwan, it is not a wise method to increase the number of independent directors. Also, we can examine the results through R-square. In this step, R-square shows at nearly the same level (0.088, 0.099, 0.099). This indicates that the models count for 8.8%-9.9% of the variation in ROA. Thus, we can conclude that the variables are good determinants of the variation in ROA.

Table 5: ROA regression

Dependent Variable: ROA

Beta Sig. Beta Sig. Beta Sig.

Boardsize -.028 .584 -.033 .513 -.033 .516 Firmsize -.287 .000 -.273 .000 -.270 .000 Industry .010 .842 .002 .967 .004 .939 Women .106 .028 .106 .029 Independence -.024 .612 R-Square .088 .099 .099 F-Statistics 12.678 10.813 8.686 F-Change 12.678 4.850 .258

(39)

After the analyses above, we can see that the dependent variable ROA shows positive relationship with female directors (0.028) while shows no significant relationship with independence. Therefore, we will test the moderating effect on the relationship between independent variables and dependent variable ROA with the following tables. Before the test, we first transform the independent variables and moderators into Z score separately. This is a way to avoid the collinearity in the results. Then, we make the interaction term by

multiplying two Z scores together; afterwards we can use the interaction term to check the moderating effect.

This regression analysis was also done in a stepwise approach. First, the control variables were tested. Second, independent variables and moderating effect were added. Third, we add the interaction term. The results are shown in the following tables. In the Table 6, it shows the relationship between independent variables and ROA under the moderating effect of Majority. In the Table 7, it shows the relationship between independent variables and ROA under the moderating effect of Minority. Both tables are organized as follows. In the first section, control variables are put for all the steps in these models. Second, independent variables are added hierarchically as Z score. Then, two moderators are transformed into Z scores as well and put in two tables. Finally, the interaction term is used.

(40)

Table 6: Moderating Effect between independent variables and ROA-Majority

Step 1 Step 2 Step 3 Step 4 Step 1 Step 2 Step 3 Step 4

Control

Boardsize .584 .513 .507 .480 .584 .586 .582 .568

Firmsize .000 .000 .000 .000 .000 .000 .000 .000

Industry .842 .967 .965 .964 .842 .815 .813 .821

Independent Variables (Zscore)

Women .028 .028 .023

Independence .600 .595 .626

Moderator (Zscore)

Majority .787 .793 .843 .706

Interaction term (Zscore)

Women·Maj .466

Indep· Maj .676

(41)

Table 7: Moderating Effect between independent variables and ROA-Minority

Control

Boardsize .584 .513 .506 .478 .584 .586 .582 .567

Firmsize .000 .000 .000 .000 .000 .000 .000 .000

Industry .842 .967 .963 .959 .842 .815 .811 .819

Women .028 .028 .023

Moderator (Zscore)

Minority .756 .762 .817 .684

Women·Min .446

Indep· Min .673

(42)

Hypothesis 3a states that Majority will have positive moderating effect. From Table 6, we can see that majority ownership structure does not have a significant positive moderating effect on the relationship between Women and ROA (! > 0.05). Besides, by using Dawson two-way interactions worksheet, we can get the line chart that reflects the degree of

interaction effect. From Figure 1, we can see that no matter Majority is at high level or low level, the slopes of the two lines are nearly the same. This indicates that the change of moderator will not have obvious influence on Women and firm performance. Thus, based on the results above, hypothesis 3a is rejected.

Figure 1: Dawson worksheet for H3a-ROA.

Hypothesis 3b states that Minority will have positive moderating effect. From Table 7, we can see that minority ownership structure does not have a significant positive moderating effect on the relationship between Women and ROA (! > 0.05). We can also use Dawson

(43)

worksheet to further illustrate the result. From Figure 2, we can see that no matter Minority is at high level or low level, the slopes of the two lines are still nearly the same. This indicates that the change of moderator will not have obvious influence on Women and firm

performance. Thus, based on the results above, hypothesis 3b is rejected.

Figure 2: Dawson worksheet for H3b-ROA.

In the Table 6, it shows the relationship between Independence and ROA under the moderating effect of Majority. However, Majority ownership structure does not have a significant positive moderating effect on the relationship (! > 0.05). Besides, by using Dawson two-way interactions worksheet, we can get the line chart that reflects the degree of interaction effect. From Figure 3, we can see that low majority and high majority has nearly no influence on ROA since the absolute value of the slopes is almost the same. This indicates there is no moderating effect between Independent and ROA. Thus, Hypothesis 4a is rejected.

(44)

Figure 3: Dawson worksheet for H4a-ROA.

In the Table 7, it also shows the relationship between Independence and ROA under the moderating effect of Minority. It is clearly to see that Minority ownership structure does not have a significant positive moderating effect on the relationship (! > 0.05). Besides, by using Dawson two-way interactions worksheet, we can get the line chart that reflects the degree of interaction effect. From Figure 4, we can see that low majority and high majority has nearly no influence on ROA. The reason is the same as mentioned above. The absolute value of the slopes is almost the same. This indicates there is no moderating effect between Independent and ROA. Thus, Hypothesis 4b is rejected.

(45)

5.2.2 Dependent Variable: Tobin’s Q

In this case, the dependent variable is Tobin’s Q and the independent variables are Women, Independence. As we mentioned above, Tobin’s Q in our dataset is another dependent variable that can help us to make a more comprehensive understanding. As shown in the Table 9, the analysis was run in a stepwise approach with first only the control variables; afterwards the independent variables were added. Hypothesis 1 claims that female directors will influence firm performance positively. In the Table 8 Step 2, the relationship between Women and Tobin’s Q is significant (' = −0.020, ! < 0.05). Based on the results of the regression analysis, Hypothesis 1 is accepted. Therefore, we can get the conclusion that when a firm owns more female directors in board, the firm performance will be better. In the Table 8 Step 3, the relationship between Independence and Tobin’s Q is not significant (' =

(46)

This means for big listed firms in Mainland China, Hong Kong and Taiwan, it is not a wise method to increase the number of independent directors. Also, we can examine the results through R-square. In this step, R-square shows at the same level (.117, .117, .117). This indicates that the models count for 11.7% of the variation in Tobin’s Q. Thus, we can conclude that the variables are good determinants of the variation in Tobin’s Q

Table 8: Tobin’s Q regression

After the analyses above, we can see that the dependent variable Tobin’s Q shows positive relationship with female directors (-.415) while shows no significant relationship with independence. Therefore, we will test the moderating effect on the relationship between independent variables and dependent variable Tobin’s Q with the following tables. Before the test, we first transform the independent variables and moderators into Z score separately like we did in the former part. We will make this regression analysis in a stepwise approach. First,

Beta Sig. Beta Sig. Beta Sig.

Boardsize .066 .187 .067 1.339 .067 1.335 Firmsize -.357 .000 -.359 -7.095 -.360 -7.084 Industry -.010 .837 -.008 -.175 -.009 -.191 Women -.020 -.415 -.020 -.413 Independence .012 .246 R-Square .117 .117 .117 F-Statistics 17.47 13.118 10.482 F-Change 17.470 0.173 .060

(47)

the control variables were tested. Second, independent variables and moderating effect were added. Third, we add the interaction term. The results are shown in the following tables. In the Table 9, it shows the relationship between independent variables and Tobin’s Q under the moderating effect of Majority. In the Table 10, it shows the relationship between independent variables and Tobin’s Q under the moderating effect of Minority. Both tables are organized as follows. In the first section, control variables are put for all the steps in these models. Second, independent variables are added hierarchically as Z score. Then, two moderators are

(48)

Table 9: Moderating Effect between independent variables and Tobin’s Q-Majority

Control

Boardsize .187 .181 .183 .212 .187 .188 .190 .154

Firmsize .000 .000 .000 .000 .000 .000 .000 .000

Industry .837 .861 .862 .863 .837 .824 .826 .865

Women .678 .681 .798

Moderator (Zscore)

Majority .938 .949 .933 .354

Women·Maj .184

Indep· Maj .042

Dependent Variable:Tobin's Q

(49)

Table 10: Moderating Effect between independent variables and Tobin’s Q-Minority

Control

Boardsize .187 .181 .184 .213 .187 .188 .190 .154

Firmsize .000 .000 .000 .000 .000 .000 .000 .000

Industry .837 .861 .863 .870 .837 .824 .827 .865

Women .678 .682 .799

Moderator (Zscore)

Minority .915 .927 .908 .370

Women·Min .176

Indep· Min .042

Dependent Variable:Tobin's Q

(50)

The method of analyzing Tobin’s Q is similar to what is used to test ROA. This regression analysis was also done in a stepwise approach. First, the control variables were tested. Second, independent variables and moderating effect were added. The results are shown in the following tables. Hypothesis 3a states that Majority will have positive moderating effect. From Table 9, we can see that majority ownership structure does not have a significant positive moderating effect on the relationship between Women and Tobin’s Q (p>0.05). Besides, by using Dawson two-way interactions worksheet, we can get the line chart that reflects the degree of interaction effect. From Figure 5, we can see that no matter Majority is at high level or low level, the slopes of the two lines are nearly the same. This indicates that the change of moderator will not have obvious influence on Women and firm performance. Thus, based on the results above, hypothesis 3a is rejected.

(51)

Hypothesis 3b states that Minority will have positive moderating effect. From Table 10, we can see that minority ownership structure does not have a significant positive moderating effect on the relationship between Women and Tobin’s Q (! > 0.05). We can also use Dawson worksheet to further illustrate the result. From Figure 6, we can see that no matter Minority is at high level or low level, the slopes of the two lines are still nearly the same. This indicates that the change of moderator will not have obvious influence on Women and firm performance. Thus, based on the results above, hypothesis 3b is rejected

Figure 6: Dawson worksheet for H3b-Tobin’s Q

In the Table 9, it shows the relationship between Independence and Tobin’s Q under the moderating effect of Majority. However, Majority ownership structure this time shows a positive correlation on the relationship (! < 0.05). To further examine the result, we again use Dawson two-way interactions worksheet to get the line chart that reflects the degree of

(52)

interaction effect. From Figure 7, we can see that low majority and high majority has clearly impacts on Tobin’s Q since the absolute value of the slopes is different this time. This indicates the Majority has moderating effect on the relationship between Independence and Tobin’s Q. Thus, Hypothesis 4a is accepted.

Figure 7: Dawson worksheet for H4a-Tobin’s Q

In the Table 10, it also shows the relationship between Independence and Tobin’s Q under the moderating effect of Minority. It is still clearly to see that Minority ownership structure does have a significant positive moderating effect on the relationship (! < 0.05). From Figure 8, we can see that similar results with the one above. The reason is the same as mentioned above. The absolute values of two slopes are different. This means Minority can influence the

(53)

Figure 8: Dawson worksheet for H4b-Tobin’s Q

(54)

6. Discussion

This study tests the relationship between gender diversity, independence directors, ownership structure and firm performance. Ownership structure, as a moderator, is divided into two dimensions: Majority and Minority. Firms with one certain ownership structure over 50% will be marked as Majority. If firms’ ownership structures are dispersed, they will be

considered as Minority in our case. Firm performance is measured in ROA and Tobin’s Q. We assumed in the hypotheses that the female directors and independent director can positively influence firm performance. However, as shown by the statistics analysis nearly all

hypotheses of our research are not supported. We will explain in depth in this chapter.

6.1 Findings

The first hypothesis claims that female directors can lead to better firm performance. That is to say, if these listed firms in Mainland China, Hong Kong and Taiwan increase the

percentage of female directors, they will achieve greater success. This hypothesis is

supported by the statistical analysis and correction matrix and then in the regression model. The research shows that female directors have a positive and significant influence on ROA, which corrects with Liu et al. (2014) findings. This was predicted because increasing the percentage of female directors will bring more information, points of view and unique

thinking (Simons & Pelled, 1999; Erhardt et al., 2003; Jurkus et al., 2011). According to Bear et al. (2010), gender diversity can offer boards with more effective management skills. In conclusion, it is proved that female directors will have positive influence on firm

(55)

H1: The percentage of female directors will positively influence on firm performance in Mainland China, Hong Kong and Taiwan.

The second hypothesis focuses on the relationship between independent directors and firm performance. In Liu et al. (2015) findings, they find that board independence have an overall positive influence on firm performance in China. However, this finding is contrast to

Fernandes (2008) and Lefort & Urzúa (2008). From testing ROA and Tobin’s Q, we both find that the relationship between independence and firm performance is not significant. Our result is in line with Bhagat and Black (2000), who also use ROA and Tobin’s Q to test. This result indicates only increasing the number of independent directors will not lead to better firm performance. As Mishra and Nielsen (2000) mention in their research, too many outside directors sometime block the process of making decisions in board. Our result supports this view effectively. Hence, in our case, it cannot prove that there are any relationships.

H2: The percentage of independent directors will positively influence on firm performance in Mainland China, Hong Kong and Taiwan.

The Hypothesis 3a and 3b assumes the majority and minority ownership structure will have impacts on the relationship between female directors and firm performance. Many

The relationship between board diversity and firm performance under the moderate effect of ownership structure in Mainland China, Taiwan and Hong Kong