The diversity of skills on the board of directors and their effect
on firm performance
Master’s Thesis Controlling (EBM870B20)
University of Groningen, Faculty of Economics and Business
Abstract
This study examines the relationship between skill diversity on the board of directors and firm performance. Furthermore, this study also examines the two-way interaction of director and CEO skills on firm performance as well as the three-way interaction of director skills, CEO skills, and CEO power on firm performance. By examining skills as an attribute of board diversity, this study complements research that examines the relationship between board diversity and firm performance. Apart from the fact that this stream of literature provides mixed results, it also focuses mostly on gender as the attribute of diversity. Prior board diversity literature especially lacks the focus on human capital diversity attributes such as skills. The aforementioned relationships are studied by using both director and CEO skills; these are taken from the 2015 proxy disclosures of 100 S&P 500 companies. In total, 18 types of skills are categorized. The findings suggest that board skill diversity is not positively and significantly related to firm performance. Furthermore, CEO skills seem not to moderate this main relationship. In addition, the interaction between director skills, CEO skills, and CEO power tends not to influence firm performance. These findings put the positive effects of board diversity that are discussed in previous literature into perspective.
Keywords: Skills
·
Board of directors·
Board diversity·
Human capital·
CEO power.Name: Matthias de Wit
Student number: S3272869 Supervisor: Dr. R.C. Trapp Co-assessor: Dr. V.A. Porumb
Word count: 11,777
1
Table of content
1. Introduction ... 2
2. Background and hypotheses development ... 6
2.1 Regulatory background ... 6 2.2 Literature review ... 7 2.3 Hypotheses development ... 10 3. Methodology ... 13 3.1 Sample ... 13 3.2 Variable measurement ... 14 3.2.1 Dependent variable ... 15 3.2.2 Independent variables ... 15 3.2.3 Control variables ... 20 3.3 Data analysis ... 20 4. Findings ... 22 4.1 Descriptive analysis ... 22 4.2 Correlation analysis ... 24 4.3 Hypothesis 1 ... 25 4.4 Hypothesis 2 ... 26 4.5 Hypothesis 3 ... 27
4.6 Supplementary analyses and robustness checks ... 28
5. Discussion and conclusions ... 29
5.1 Discussion of findings ... 29
5.2 Implications ... 32
5.3 Limitations ... 33
5.4 Future research ... 34
References ... 36
2
1. Introduction
n 2009, the U.S. Securities and Exchange Commission (SEC) approved a rule that requires companies to improve their annual proxy statements by disclosing information about the skills of their directors.1 In 2015, the SEC announced that it is reviewing these rules, which may even lead to new rules (Petrilla, 2016). The growing interest in the skills of directors is a recognition of their importance since it suggests that information about these skills is important for investors and shareholders. Otherwise, corresponding skill diversity regulation would not emerge through the SEC.2
In line with these developments, social activists, shareholder groups, and government agencies advocate and call for greater board diversity.3 Prior literature about board diversity, however, focuses primarily on the demographic (e.g., Miller and Del Carmen Triana, 2009; Riordan and Shore, 1997) and gender (e.g., Campbell and Mâinguez-Vera, 2008; Post and Byron, 2015) composition of corporate boards. Besides that these studies provide mixed results about the relationship between board diversity and firm performance (Jackson et al., 2003), there is scant academic research or evidence beyond gender diversity that supports the notion that board diversity affects firm performance (Anderson et al., 2011). Consequently, different scholars address the need to examine the effect of other board diversity attributes on firm performance (e.g., Hillman, 2015; Post and Byron, 2015). According to Jackson et al. (2003), skills is an example of an attribute that requires attention in board diversity literature. The authors argue that the weakness in the research on board diversity and firm performance is that studies ignore skills, which form an essential dimension of diversity: different studies mention the importance of skills in determining firm performance (e.g., Krishnan et al., 1997) but ignore the effects of skill diversity. Furthermore, in 2009, Kor and Sundaramurthy have argued that there is still a lack of academic research on how different human capital attributes of board diversity, such as skills, affect firm performance. To the present day, relatively little is known about the actual skills that are present on boards of directors and the effect of these skills on firm performance. In fact, there are no (large) sample empirical studies related to how directors may contribute to corporate boards based on their skills.
1 See file no. S7-13-09 (2009) of the SEC Proxy Disclosure Enhancements. 2 See Rules and Regulations (2013) of the U.S. SEC.
3 See Best Disclosure: Director Qualifications & Skills (2014) of the Council of Institutional Investors.
3 The high priority placed on board diversity can be explained based on the resource dependence theory. According to this perspective, board diversity can benefit companies by providing greater access to resources (Pfeffer and Salancik, 1978). This theory argues that a company’s performance is linked to the opportunities that are available for accessing the required resources from directors’ human capital. Companies that have the disposal of resources from different pools make decisions that are more effective and deliver more creative products than companies that draw from the same pool of resources (Jackson et al., 1995). Therefore, when a board of directors is more diverse based on the human capital of the directors, the board is expected to be better able to provide resources to the company since, in this case, directors bring more knowledge and provide more insights from different perspectives to the board. This could result in better decision-making, which, ultimately, may improve firm performance.
In contrast to skills, expertise is a form of human capital that has received some attention in prior board diversity literature. Both skills and expertise fall into the category of human capital since they are considered as attributes related to the abilities and knowledge needed in the workplace (Jackson et al., 2003). Skills and expertise can be described as, respectively,
“what you can do” (Faingersh, 2012) and “the knowledge about a certain domain” (Sullivan,
1990). Skills tend to be a more specific form of human capital than expertise. Marketing skills are a common type of director skills. Marketing skills reflect, for instance, the abilities of building online communities and increasing revenues from internet advertising. A common type of director expertise is financial expertise. Financial expertise reflects, for instance, the comprehensive knowledge about investment decisions (see, e.g., the 2015 proxy statements of Corning Inc. and Coca-Cola Co.). Prior literature on board diversity that has studied the effect of directors’ human capital, as measured by expertise, on firm performance can be divided into two groups. The first group focuses on particular types of expertise, such as financial (e.g., Güner et al., 2008), industry (e.g., Faleye, 2018), and political (e.g., Hillman, 2005), and the effect of these forms of expertise on firm performance. The second group focuses on expertise diversity and its effect on firm performance. An example of such a study is that of Anderson et al. (2011), who examined the diversity effect of law, consulting, and financial expertise of directors on firm performance in the same setting. These studies, however, provide mixed results with regard to the relationship between board human capital diversity and firm performance.
4 Possible explanations for these mixed results could be that expertise is a “noisy measure” for studying the effect of the human capital diversity of boards on firm performance or that moderators have been neglected. In contrast to expertise, skills could be a more precise approach for studying the relationship between human capital diversity of boards and firm performance since skills give a more nuanced insight into the human capital of a director. To better substantiate this line of reasoning, the following example might prove helpful. Company X has two directors (A and B) serving on its board. Director A, on the one hand, who has financial expertise, also possesses finance, risk management, and marketing skills. On the other hand, director B, who also has financial expertise, possesses finance, strategy planning, and international business skills. Accordingly, these two directors can be classified as homogeneous with respect to expertise but diverse with respect to skills. In this study, I argue that a director brings more to the board than simply his or her expertise; he or she brings a set of skills, that, in relation to the skills of the other directors, may or may not benefit the company.
To contribute to the literature on the human capital diversity of boards, this study builds upon the relationship between directors’ skills and firm performance. In so doing, this study intends to fill a gap in this research area. With the current focus increasingly on skill disclosure regulation and on skill diversity on corporate boards, it is important to understand the effect of directors’ skills on the companies they serve. Through gathering such insights, practitioners such as investors and shareholders could have actual evidence to analyze the composition of corporate boards in order to make more informed voting decisions and investments. Consequently, the goal of this study is to provide evidence as to whether skill diversity on boards affects firm performance. This leads to the following research question:
“What is the effect of the diversity in directors’ skills on firm performance?”
Furthermore, this study also takes into account the role of the CEO. When the company’s CEO has a broader range of skills, the expectation is that he or she might be better able to take the advice of and cooperate with the directors. Consequently, this may have a positive influence on the implementation of the company’s strategy. Hence, a CEO with a broader range of skills may positively influence the relationship between directors’ skills and firm performance. Because CEOs serve on both the board of directors and top management team (TMT) inside all U.S. listed companies, CEOs have, respectively, major influence on the
5 strategic decisions that companies make (Hambrick and Mason, 1984; Minnick and Noga, 2010) and on the implementation of companies’ strategies (Fama and Jensen, 1983). Therefore, a shared understanding between directors and CEOs can encourage cooperative decision-making and mutual trust, which, ultimately, benefits companies (Sundaramurthy and Lewis, 2003).
However, this potential might not be fully explored in instances when CEOs become more powerful since, from an agency theory perspective, powerful CEOs are more likely to follow their own paths (Jermias and Gani, 2014). This may cause that the board of directors lose the structural power to influence the CEO’s decisions (Westphal and Zajac, 1995), which can, in turn, result in a lower quality of strategic decisions and resource allocation as the CEO does not make full use of the resources provided to him or her (Kor, 2006). Hence, powerful CEOs may negatively influence the moderating effect of the CEO skills on the relationship between directors’ skills and firm performance. Accordingly, this study also examines whether CEO skills positively moderate the relationship between the skills of the directors and firm performance and whether CEO power mitigates this moderating effect.
After the data from 100 S&P 500 companies was analyzed, no relationship was found between skill diversity on boards and firm performance. However, a higher level of international skill intensity on the board of directors is shown to have a positive and significant effect on firm performance. Furthermore, this study does not provide significant evidence for the effect of the interaction between directors’ and CEOs’ skills on firm performance. Moreover, no evidence is found for the effect on firm performance of the interaction between directors’ skills, CEOs’ skills, and CEO power.
Despite the fact that the constructed hypotheses are not supported, this study makes several contributions. First, this study contributes to the existing literature on board diversity in human capital because the results advance our understanding of this subject and help to fill the gap in the literature. To the best of my knowledge, this study is the first to focus on skill diversity on the boards of directors and, therefore, gives more comprehensive insights into board diversity. In this respect, this study responds to those studies that have called for studying attributes of board diversity other than gender and the effect of these attributes on firm performance (e.g., Hillman, 2015; Kor and Sundaramurthy, 2009). Second, the findings should be of interest to companies’ internal and external stakeholders. Both groups of
6 stakeholders can use the results by identifying gaps in the collective skills of the directors on corporate boards. Committees that are responsible for proposing new directors to the board and shareholders can use the results of this study in determining whether or not to select or vote for a new director according to the intensity of international business skills present on the board.
The remainder of this paper is organized as follows. The next section (2), discusses the theoretical framework needed to answer the research question and the formulated hypotheses. In section 3, the methodology of this paper is discussed, which includes the sample selection, the variable measurement, and information about how the data in this study is analyzed. Subsequently, section 4 presents the findings of the data analyses. Section 5 provides concluding comments, a discussion on the implications and limitations of this study, and suggested areas for future research.
2. Background and hypotheses development
This section of the paper discusses the background of this study and the related literature that provides insights into board diversity and firm performance. In addition, a set of hypotheses guided by literature and theory is constructed.
2.1 Regulatory background
In 2009, board diversity received more attention when the SEC introduced new rules for enhancing board diversity disclosure of U.S. listed companies. The argument for the implementation of this rule was that it would be a helpful step forward in providing investors and shareholders with additional information they need to make more informed investment and voting decisions with regard to corporate governance and the election of directors.4 Since then, a large number of U.S. listed companies provide nowadays more explicit and detailed information about the skills of their directors in their annual proxy statements, in the form of matrices, graphs, and other formats (see, e.g., the 2015 proxy statements of Corning Inc., Coca-Cola Co., and General Electric Co.). Indeed, in the last several years, the disclosure of information concerning directors’ skills in proxy statements has risen.5
4 See, e.g., letters from Board of Directors Network, Forum of Executive Women and Integrated Governance Solutions. 5 See Best Disclosure: Director Qualifications & Skills (2014) of the Council of Institutional Investors.
7
2.2 Literature review
Diversity, also termed as “dispersion”, “heterogeneity”, and “dissimilarity” in prior literature, can be defined as “the distribution of differences among the members of a unit with
respect to a common attribute” (Harrison and Klein, 2007, p. 1200). The board of directors is
the unit to be analyzed in this study. The agency theory and resource dependence theory are the two most dominant perspectives used to study the effects of board diversity (Hillman and Dalziel, 2003). From the perspective of both theories, a more diverse board produces more effective board outcomes. Board diversity can, from an agency theory perspective, benefit the company by improving monitoring effectiveness. Agency theorists argue that the monitoring function of the board is needed to constrain the potential opportunistic behavior of executives (Fama and Jensen, 1983; Jensen and Meckling, 1976). According to Jensen and Meckling (1976), a more diverse board strengthens the board’s monitoring role since such boards bring different perspectives to their monitoring duties. From a resource dependence perspective, board diversity benefits companies by providing access to a greater number of relevant resources such as networks (Hillman and Dalziel, 2003) and by giving advice to companies from different perspectives (Pfeffer and Salancik, 1978). While in previous literature the resource dependence theory has more frequently been used to substantiate the positive effects of board diversity on firm performance, the agency theory has more often been used to study the effect of board diversity on different corporate governance practices (e.g., the monitoring function of the board). Since the aim of this study is to examine the effect of board skill diversity on firm performance, the focus is on the resource dependence theory to draw the relationship between the concepts of interest.
Previous literature that studied the effect of board diversity on firm performance has measured board diversity with different attributes. These attributes can be categorized as “task-related” or “non-task-related” (Adams et al., 2015). Task-related attributes are related to the abilities and knowledge needed in the workplace; non-task-related attributes are the attributes that fall outside this description (Jackson et al., 2003). On the one hand, education (e.g., Jackson et al., 1991) and functional background (e.g., Jehn et al., 1999) are examples of task-related attributes that have been addressed in prior board diversity literature. On the other hand, demographic (e.g., Miller and Del Carmen Triana, 2009) and gender (e.g., Adams and Ferreira, 2009) are examples of non-task-related diversity attributes that have received attention. Of the two, non-task-related attributes of diversity have received the most attention.
8 Gender has, in particular, been widely addressed as a diversity attribute in studying the relationship between board diversity and firm performance (Post and Byron, 2015).
Several of the studies that have examined the effect of gender diversity in the boardroom on firm performance have found evidence of a positive relationship between board diversity and firm performance (e.g., Adams and Ferreira, 2009; Arfken et al., 2004). According to Adams and Ferreira (2009), the presence of women on the board of directors has a significant positive impact on board governance. The authors found evidence that CEOs are more likely to be held accountable for poor stock price performance when boards are more diverse. Accordingly, CEOs are more focused on improving stock price performance. Furthermore, according to Arfken et al. (2004), board gender diversity has a positive effect on firm performance since a more gender-diverse group of directors bring unique perspectives to the board. This results in better problem-solving and formulation of strategies, which benefit both the company and its shareholders (Jensen, 1993). However, there are also studies that have found no relationship between board diversity and firm performance (e.g., Campbell and Mâinguez-Vera, 2008; Farrell and Hersch, 2005). According to Farrell and Hersch (2005), this is because companies merely aim to build a positive public image by including members of minority groups (in this case, female directors) to the board.
Since the major focus of prior board diversity literature is on gender as a diversity attribute and because these studies provide mixed results, different studies call for more research on other board diversity attributes to attain greater insights into the relationship between board diversity and firm performance (e.g., Hillman, 2015; Post and Byron, 2015). In particular, several scholars have requested more research on the task-related attributes of board diversity—and especially those related to the human capital of directors (e.g., Hillman et al., 2000; Jackson et al., 2003). Human capital can be defined as “the set or bundle of knowledge
and perspectives that directors collectively bring to the board” (Kor and Sundaramurthy,
2009, p. 984). Skills are an example of human capital attributes (Pfeffer and Salancik, 1978). According to Hillman et al. (2000), skills, in particular, could provide more precise explanations for the relationship between board diversity and firm performance since human capital is closely linked to one of the most important functions of boards—that is, providing resources to the company (Pfeffer and Salancik’s, 1978). In addition, skills are often invoked as explanations for the effects of board diversity (Jackson et al., 2003).
9 There is only limited research available from the literature stream that focuses on the relationship between the human capital attributes of board diversity and firm performance. Expertise can be characterized as a human capital attribute that has received some attention in previous literature on board diversity. This stream of literature can be categorized into two groups: one that focuses on particular types of expertise, and another that focuses on diverse types of expertise in the same setting. Some examples of the first group are the following scholars: Christensen et al. (2010), who have studied the accounting expertise of audit committee members; Güner et al. (2008), who have studied the financial expertise of directors; and Hillman (2005), who has studied political expertise on the board of directors. These studies, however, provide mixed results. For example, Hillman (2005) has found mixed support for the claim that political expertise on the board of directors improves firm performance. The positive effects of politicians on corporate boards and their financial performance is only observable for companies that operate in heavily regulated industries and that depend more on government agencies. Moreover, Faleye et al. (2018) have found no evidence that industry expertise is beneficial for all companies; rather, the positive effects are mostly limited to R&D companies. With respect to the second group, the study of Anderson et al. (2011) is, to the best of my knowledge, the only study that examined the effect of multiple forms of expertise on firm performance. The authors have examined the diversity effect of legal, consulting, and financial expertise of directors in the same setting and have found a positive effect on firm performance. Their findings indicate that greater board diversity increases firm performance as directors with different kind of knowledge can bring varied perspectives and talents to corporate deliberations.
Although the study of Anderson et al. (2011) contributes to the previous literature on board diversity and firm performance, there is still a critical literature gap. Apart from the mixed results that have been garnered from the previous literature, companies’ environments have changed as they are currently operating in a more complex and a broader range of industries since they have expanded their operations towards a global market (Boone et al., 2007). Therefore, companies have higher information requirements (e.g., Boone et al., 2007; Lehn et al., 2009) and an increased need for advice (e.g., Adams and Mehran, 2003; Coles et al., 2008). On account of this, companies need to continuously adapt themselves to survive and prosper (Danneels, 2002). Consequently, to act appropriately in their changing environment, it is likely that companies need resources from a wider range of director perspectives than only those perspectives related to law, consulting, and finance. This study examines skills as the
10 human capital attribute of diversity since skills give insight into the broader range of knowledge and perspectives that directors bring to boards. Although several previous studies that draw on the resource dependence theory provide empirical evidence of positive effects of board diversity (e.g., Arfken et al., 2004), there are other studies arguing that board diversity can have negative effects on firm performance. Board diversity can, for instance, limit within-unit behavioral and social integration, which diminishes cohesion and morale (Williams and O’Reilly, 1998). Furthermore, relative to homogeneous boards, board diversity can increase costs and the possibility of conflicts as problems arise among directors with differing backgrounds due to greater coordination and communication problems (Lang, 1986). However, directors that bring knowledge from different perspectives to the board are expected to be more important in the current era because of companies’ higher information requirements. This study, therefore, predicts that a more skills-diverse board has positive— instead of negative—effects on firm performance. To the best of my knowledge, there are no previous studies that take into account skills as an attribute of board diversity and examine its effect on firm performance.
2.3 Hypotheses development
In considering the need to study the relationship between board diversity attributes (apart from gender) and firm performance, this study focuses on skills as the attribute of diversity. The need to study the effect of this attribute on firm performance has also been called for by other scholars (Jackson et al., 2003). Skills are an aspect of the human capital of directors since skills are related to the abilities and knowledge needed in the workplace (Jackson et al., 2003).
As I discussed in section 2.2, this study relies on the resource dependence theory to predict a relationship between skills-diverse boards and firm performance. This theory, developed by Pfeffer and Salancik (1978), focuses on how human capital leads to the provision of resources, such as advice, to companies. According to the resource dependence theory, one of a board’s most important functions is to provide resources to the company (Pfeffer and Salancik, 1978). A company’s performance is linked to the opportunities that are available for accessing the required resources. More specifically, the human capital of corporate boards shapes how directors offer advice, govern companies, and affect the ideas and resources that directors provide to companies (Dalziel et al., 2011; Haynes and Hillman, 2010). This study suggests that skills-diverse boards bring multiple, different perspectives in fulfilling their
11 advising role and that these perspectives potentially benefit firm performance through improved problem-solving, resource utilization, and strategy formulation (Jensen, 1993).
I argue that diverse skills of directors are important components for boards to effectively perform their resource provision role. When a board is more skills-diverse, it might be better able to provide resources to the company since the directors bring more knowledge and provide more insights from different perspectives to the board. Especially in the current era, where companies are dealing with more complex environments, a skills-diverse board might be necessary to provide the needed resources to guide the company in the right direction. Therefore, I argue that having a diverse mix of skills is paramount for ensuring that the board, as a collective, is equipped to guide the business and strategy of the company. Ultimately, this may have a positive effect on firm performance. The rationale for this assumption is based on the study of Jackson et al. (1995). According to the authors, companies that have the disposal over informational resources from different kind of pools, such as with regard to the experience or knowledge of these pools (Austin, 2003), make more effective decisions and deliver more creative products than companies that draw from the same pool of resources. This can be explained by that companies have in this situation more input from different perspectives, which may help them to make better decisions. Correspondingly, I formulate the following hypothesis:
Hypothesis 1 (H1): Board skill diversity and firm performance are positively related.
As mentioned in section 1, my illumination of the relationship between the diversity of directors’ skills and firm performance also takes into account the influence of the CEO. It is expected that when a company’s CEO has a broader range of skills, he or she might be better able to take the advice of the directors and to cooperate with the directors. This may have a positive influence on firm performance. Consequently, the expectation is that CEOs with a broader range of skills may positively influence the relationship between the skills of the directors and firm performance. Nowadays, all U.S. listed companies maintain a one-tier governance structure. This means that the CEO has a position within both the TMT and the board of directors (see Figure 1 for a visual representation of this governance structure). Since CEOs serve both positions, they have, respectively, major influence on companies’ strategic decisions (Hambrick and Mason, 1984; Minnick and Noga, 2010) and implementation of companies’ strategies (Fama and Jensen, 1983). An amiable relationship between the directors
12 and the CEO may, therefore, be important as, according to Sundaramurthy and Lewis (2003), shared understandings between directors and the CEO can encourage cooperative decision-making and mutual trust that benefits the company. This leads to the following hypothesis:
Hypothesis 2 (H2): The positive relationship between board skill diversity and firm performance is stronger when a CEO has a broader range of skills.
Figure 1: One-tier governance structure of U.S. listed companies
However, as I discussed in section 1, it is expected that the moderating effect of CEO skills mitigates when a CEO has greater power within the company. When a CEO possesses more power, he or she may not fully accept the advice of the directors. The power of a CEO increases when, for instance, he or she has a long tenure and/or chairs the board (Daily and Johnsen, 1997). From an agency theory perspective, a powerful CEO tend to follow his or her own path (Jermias and Gani, 2014) and pursues private benefits at the expense of shareholders (Jensen, 1986). This may cause the board of directors to lose the structural power to influence the CEO’s decisions (Westphal and Zajac, 1995). This can result in a lower quality of strategic decisions and resource allocation as CEOs do not make full use of the resources provided to them (Kor, 2006). Hence, powerful CEOs may negatively influence the moderating effect of the CEO skills on the relationship between directors’ skills and firm performance. Consequently, the following hypothesis is formulated:
Hypothesis 3 (H3): When a CEO has more power, the moderating effect of the CEO skills on the relationship between directors’ skills and firm performance mitigates.
Figure 2 presents a summary of the conceptual model that reflects the expected aforementioned relationships.
13
Figure 2: Summary of the conceptual model
3. Methodology
This section outlines the sample selection and research design, which constitutes the blueprint for the collection, measurement, and analysis of the data. This study can be characterized as a deductive quantitative research that uses archival data. Archival research is deemed most appropriate since this study is focusing on the discovering effect of a wide scale of companies and makes inferences based on a sample. In addition, a content analysis is also applicable with respect to manually collecting different data from proxy statements and transforming the data into word-based variables for subsequent statistical analysis.
3.1 Sample
Table 1 summarizes the sample selection. The sample of this study includes the directors of U.S. S&P 500 companies. This selection is based on the following reason. Due to the 2009 SEC regulations, U.S. listed companies must disclose information about the skills of their directors in their annual proxy statements. As mentioned in section 2.1, the S&P 500 companies are providing nowadays an increasing amount of explicit information about the skills of their directors in these disclosures. Thus, the availability of the relevant data is high, which makes it easier to compare the data concerning director skills across these companies. This offered me the opportunity to classify each director by his or her skills. Moreover, in considering the limited time period of the research and the intensive nature of manually collecting the data of certain variables, I collected and analyzed the data from only one year and from 100 randomly selected S&P 500 companies. The usage of data from one year is not expected to affect the generalizability of the results since the composition of boards does not
H2 (+)
Board skill diversity Firm performance CEO power CEO skills
H3 (-)
14 change greatly over time.6 Furthermore, on the basis of the rationales expressed in some proxy statements, companies tend to replace a departing director with one who has a similar profile. Therefore, the mix of directors’ skills on corporate boards is not expected to change substantially over time. Moreover, a conditional random sampling approach was used to collect the data. However, when a company was randomly selected and its proxy statement did not contain a sufficient level of information concerning the skills of their directors, the company’s data was not considered. This was, however, the case in very few companies.
The data concerning the skills of the directors was obtained from the S&P 500 companies’ 2015 proxy statements. This particular year was selected after a preliminary data collection of the 2014 and 2015 proxy statements from the first 100 S&P 500 companies (according to alphabetical order of the company name) was undertaken. This pre-check indicated that the 2015 proxy statements provide more detailed, explicit, and comprehensive data concerning directors’ skills compared with the 2014 proxy disclosures. The data availability from the 2015 proxy statements is, therefore, higher than that from the 2014 proxy statements.
3.2 Variable measurement
Next, further information about each dependent, independent, and control variable used in this study is provided. These variables are summarized in Table 2.
6 See Board Refreshment Trends at S&P 1500 Firms (2017) of the Harvard Law School Forum on Corporate Governance and
Financial Regulation.
Panel A: Selection of companies
Total numbers of companies listed in S&P 500 500 Less:
Companies (randomly) not selected -400
Total sampled companies 100
Panel B: Sampled companies according to their industry classification:
Mining industry 7
Manufacturing industry 31
Transportation and public utilities industry 12
Wholesale trade industry 2
Retail trade industry 10
Finance, insurance, real estate industry 27
Services industry 10
Public administration industry 1
Total sampled companies 100
Panel C: Sampled companies across years
2015 100
Total sampled companies 100
TABLE 1
15
3.2.1 Dependent variable
In this study, the dependent variable is firm performance (FIRM_PERF). Since financial performance is widely used as a proxy for FIRM_PERF (Campbell and Mâinguez-Vera, 2008; Mahadeo et al., 2012), I used “Tobin’s Q” to measure FIRM_PERF. Tobin’s Q can be characterized as an objective measure of FIRM_PERF since it is beyond the control of management (Jermias and Gani, 2014). Following Cheng (2008), Tobin’s Q is measured as the ratio of the market value of assets to the replacement costs of assets. The data for this variable was derived from the Orbis database and was then compared with data from the Compustat database to ensure accuracy and consistency (Foster, 1986).
3.2.2 Independent variables
The first independent variable to be examined in this study is the skills of the directors (DIR_SKILL). DIR_SKILL captures the different kind of skills present on boards of directors; this is needed to measure skill diversity on boards. This variable is part of the primary relationship examined in this study between the skills of the directors and firm
Variable Code Description Source of data
Dependent variable
Firm performance FIRM_PERF
Independent variables
Directors' skills DIR_SKILL
CEO skills CEO_SKILL
CEO power CEO_POW
Control variables
Board size BOARD_SIZE
Gender diversity GENDER_DIV
Return on assets (t-1) ROA
Sales growth SALES_GR Salest minus salest-1.
Leverage LEV The ratio of total debt to total equity. Firm size FIRM_SIZE The total book value of the company's assets.
TABLE 2
Variable definitions
Manually collected from annual 2015 proxy statements (DEF14A) Manually collected from annual 2015 proxy statements (DEF14A) Orbis database; Compustat database
The three variables (1) CEO duality (CEO_DUALITY), a dummy variable with a value of 1 if the CEO is also the chairman of the board and 0 otherwise, (2) the ratio of directors who were appointed after the CEO (RATIO_APP_AFTER_CEO), and (3) the ratio of shares held by the CEO to the director ownership (RATIO_CEO_SHARES) are
standardized and summed to create an index of CEO_POW.
Manually collected from annual 2015 proxy statements (DEF14A) Measured with Tobin's Q (i.e., the ratio of market value of assets to
the replacement costs of assets).
The ratio of the total skill categories of the CEO to the total skill categories present on the company's board of directors. The number of skill categories present on the company's board of directors.
Orbis database; Compustat database Orbis database; Compustat database Last year's income before interest, tax, depreciation, and
amortization (EBITDA) divided by the book value of assets.
The amount of directors serving on the company's board of directors. Manually collected from annual 2015 proxy statements (DEF14A) Manually collected from annual 2015 proxy statements (DEF14A) Orbis database; Compustat database The ratio of the number of women to the number of men on the
board.
Orbis database; Compustat database
16 performance. To examine this relationship, I categorized the skills of each director into several skill categories. In total, I set up 18 skill categories, which are summarized in Table 3. The categories were inductively developed due to a lack of availability of previously used categories. In fact, the categorization is based on the rationales of the companies themselves. This method is expected to be appropriate since the various skill types mentioned in the proxy statements are quite similar among the S&P 500 companies. Furthermore, a great number of these companies provide comprehensive and similar descriptions about the different skill types of their directors. This allowed me to categorize the different skills into different skill categories. During the data collection, I added additional skill categories when I found a new designation for a skill category. Moreover, although companies use the same description for many types of skills, there are some skill types that are more or less the same but are described somewhat differently. These skills were merged into one skill category during the procedure. For example, one of this study’s skill categories contains “financial”, “accounting”, “investment”, and “banking” skills. These skills are quite similar, which may explain why several S&P 500 companies already combined them into one category (i.e., “Financial, Accounting, Investment, and Banking”) in their proxy disclosures (see, e.g., the 2015 proxy statement of Corning Inc.).
No. Skill category
1 Industry and Relevant Industry
2 Financial, Accounting, Investment, and Banking
3 International
4 Governmental, Political, Regulatory, and Law
5 Academic and Education
6 Leadership, Management, and Public Company
7 Sales, Marketing, and Brand Management
8 Technology, R&D, Innovation, and Science
9 Manufacturing, Constructing, and Engineering
10 Risk Management, Risk Oversight, and Crisis
11 Strategy and Operations
12 Supply Chain, Transportation, and Logistics
13 Human Resources, Talent Management, and Compensation
14 Community, Media Relations, and Investor Relations
15 Mergers and Acquisition
16 Information Technology, and E-commerce
17 Legal, Compliance, and Ethics
18 Environmental, Sustainability, Safety, and Health
Table 3
17 On the basis of the study of Anderson et al. (2011), who measured a similar board diversity attribute (i.e., expertise), I assigned a value of 1 if the board of directors have at least one director with skills related to 1 of the 18 skill categories. I then took the sum of the 18 categories, and the total was used as a proxy for skill diversity on boards. Hence, the minimum score for skill diversity of a company’s board is 0, occurring when none of the directors’ skills can be applied to 1 of the 18 skill categories, and the maximum score is 18 otherwise. This type of measuring diversity can be classified as “variety” (Harrison and Klein, 2007; Guevara, 2016). Anderson et al. (2011) have also used this approach to measure expertise diversity on boards of directors.
As aforementioned, the data regarding DIR_SKILL was obtained manually from the 2015 proxy statements filed by the S&P 500 companies in the EDGAR database of the SEC.7 The proxy statements have a chapter named “election of directors”, where the skills of every director are described in texts, matrices, graphs, or other formats. Appendix A provides examples of some formats from where the skill data was obtained. In order to collect the data of the variable DIR_SKILL, I used the form-orientated means of content analysis, which involves routine counting of words (Smith and Taffler, 2000). In particular, I counted the different skills per director mentioned in the section “election of directors” and then classified this data into 1 of the 18 predefined skill categories.8
In this study, the skills of the CEO (CEO_SKILL) is one of the moderators. CEO_SKILL captures the different types of skills that CEOs possess. This measure is used to examine if CEO_SKILL has a positive effect on the relationship between DIR_SKILL and FIRM_PERF. To measure CEO_SKILL, I used the ratio of the total number of skill categories of the CEO to the total number of skill categories of the other directors. The higher the value of this ratio is, the broader range of skills a CEO has relative to the directors. Since none of the CEOs in the sample have different skill types relative to the other directors on the board of the company they serve, a higher ratio also indicates that the skills of the CEO are more in line with those of the other directors. Just as in the case of DIR_SKILL, the data regarding CEO_SKILL was obtained manually from the 2015 proxy statements of the S&P 500
7 See http://www.sec.gov/edgar.shtml.
8 During the data collection procedure, I checked whether the number of directors on the company’s board equals the number
of directors named in the section “election of directors” in the proxy statement since it is possible that the number of directors mentioned in the section “election of directors” is lower than the actual number of directors on the company’s board. This has to do with the fact that in some cases several directors are not suggested for election. When this was the case, I collected the data concerning the skills of these directors from the section of the proxy statement where this data is stated.
18 companies. In order to increase the validity of the data, the data of DIR_SKILL and CEO_SKILL from several companies was coded twice so as to be sure that the data regarding the skill categories present on the boards and the values per skill category were correct.
Before the other variables are discussed, some additional information about the data collection of these upcoming variables must be first addressed. As aforementioned, the data for DIR_SKILL and CEO_SKILL was obtained manually from the 2015 proxy statements. Since these statements mainly refer to 2014, my intention was to collect the data for the other variables from the year 2014 out of online databases. However, collecting data from the year 2014 for all the other variables out of online databases was for several variables not the optimal option since the 2015 proxy statements disclose data not only from 2014 but also from the moment the various companies published them. Because most companies published their proxy statements at the end of the first quarter of 2015, the 2015 proxy statements also contain data from the year 2015. Accordingly, the data from the 2015 proxy statements is not fully comparable with the data from 2014 out of the online databases. For instance, according to the 2015 proxy statement of ADT Corp., its board contained nine directors; according to the BoardEx Database, however, ADT Corp. had eight directors in 2014. Hence, between December 31, 2014 and the date when ADT Corp. published its 2015 proxy statement, a new director had been appointed to the board of directors. Therefore, to ensure data accuracy and consistency, I decided to manually collect the data from the 2015 proxy statements for the independent variable “CEO power” and for the control variables “board size” and “gender diversity”.
CEO power (CEO_POW) is another moderator in this study. CEO_POW is included to examine if a powerful CEO mitigates the moderating effect of CEO_SKILL on the relationship between DIR_SKILL and FIRM_PERF. Following Haynes and Hillman (2010), I used three out of their four CEO_POW items to create an index of CEO_POW: CEO duality (CEO_DUALITY); the ratio of directors who were appointed after the CEO (RATIO_APP_AFTER_CEO); and the ratio of shares held by the CEO to the director ownership, in other words, the total shares held by all the directors on the board (RATIO_CEO_SHARES). I did not take their fourth CEO_POW item, which is the ratio of the number of non-independent directors to the total number of directors, into account because the CEO is the only dependent director within most of the sample companies. In fact, within 60 companies, the CEO is the only dependent director on the board of directors.
19 Moreover, within 20 of the 100 sample companies there are only two dependent directors. And within only 10 of these companies, there are three or more dependent directors on the board. This low ratio of dependent directors might be due to the current SEC regulation, which requires that independent directors must comprise the majority of U.S. listed companies’ board of directors.9 Therefore, by taking into account the ratio of non-independent directors to the total number of directors, the reliability of the variable CEO_POW decreases. Next, further information about the three items of CEO_POW are discussed.
According to Zajac and Westphal (1996), CEO_DUALITY is an indicator of the authority that CEOs have over the directors and allows them to exert greater influence over the boards. In this study, CEO duality is measured as a dummy variable: with a value of 1 if the CEO is also the chairman of the board, and 0 if not (Haynes and Hillman, 2010; Krause and Semadeni, 2013). Although different corporate governance guidelines recommend avoiding CEO duality (see, e.g., Principles of Corporate Governance, 2012; Sarbanes-Oxley Act, 2002), Jermias and Gani (2014) have found that CEO duality is still observable within a substantial number of U.S. corporate boards. This study substantiates this since CEO duality occurs within 43 out of the 100 sample companies.
Directors who are appointed after the CEO may feel that they should be loyal to the CEO since the CEO is responsible for the appointment of new directors (Boeker, 1992). This increases the personal ties between the CEO and the directors appointed after him or her (Finkelstein, 1992; Daily and Johnson, 1997). Accordingly, boards of directors comprising a greater proportion of directors who are appointed after the CEO may increase the powerbase of the CEO.
CEOs with ownership positions in the companies they serve have more power since they represent not only the management of the company but also its shareholders (Daily and Johnson, 1997). Zald (1969) has also found that CEOs with ownership positions have more power and, therefore, are better able to influence strategic decisions, as compared to CEOs with no or minor ownership positions.
20 Following Chen (2014), the three items CEO_DUALITY, RATIO_APP_AFTER_CEO, and RATIO_CEO_SHARES were standardized and summed to create an index of CEO_POW. The higher the value of this variable is, the higher is the power of the CEO. Consequently, the expectation is that, in this case, the board of directors is less likely to have influence on the CEO. When the CEO has more power, the expectation is that he or she tends to follow his or her own path and is less willing to take the advice of the other directors. Furthermore, a reliability analysis was performed to measure the degree to which the three items of CEO_POW are consistent with each other. The results of this analysis reveal that the Cronbach’s alpha for the three items of CEO_POW is 0.6. This is higher than the Cronbach’s alpha when taking all the four items of CEO_POW from the study of Haynes and Hillman (2010) into consideration (i.e., 0.5).
3.2.3 Control variables
I introduced six variables into the analysis to control for. First, board size (BOARD_SIZE) was measured to control for whether larger boards are related to firm performance. Second, because I wanted to examine if board skill diversity positively affects firm performance, I controlled for gender diversity (GENDER_DIV) since this attribute of board diversity has provided the most significant evidence of a positive relationship between board diversity and firm performance in previous literature. As I explained earlier, the data for both variables BOARD_SIZE and GENDER_DIV was manually collected from the 2015 proxy statements. Third, prior year’s (i.e., 2013) return on assets (ROA) was selected as the variable to control for the effects of past performance on current performance. Fourth, growth opportunities are found to be a significant determinant of board composition (Lehn et al., 2009). Therefore, sales growth (SALES_GR) was included to control for this effect. Finally, leverage (LEV) and firm size (FIRM_SIZE) are the fifth and sixth control variables in this study, both of which are often included in previous and comparable studies and are shown to have a significant effect on firm performance (Campbell and Mâinguez-Vera, 2008; Mahadeo et al., 2012). The data for the variables ROA, SALES_GR, LEV, and FIRM_SIZE was collected from both the Orbis and Compustat database to avoid the loss of observations. Subsequently, the data obtained from both databases was compared to ensure the accuracy and consistency of the data.
3.3 Data analysis
Since most of the dependent and independent variables in this study consist of continuous data and are normally distributed, the data was analyzed with multiple multivariate analyses.
21 Before I set up the regression analyses, a bivariate analysis was first conducted to indicate whether there is a relationship between the concepts of interest. Subsequently, to test the relationships between the concepts of interests further, several regression analyses were conducted. An ordinary least-squares regression was conducted to test the relationship between DIR_SKILL and FIRM_PERF (H1). The corresponding equation is as follows:
FIRM_PERF = β0 + β1 DIR_SKILL + β2 BOARD_SIZE + β3 GENDER_DIV + β4 ROA + β5 SALES_GR + β6 LEV + β7 FIRM_SIZE + ε
Subsequently, a moderated regression was conducted to test the two-way interaction of the expected moderating effect of CEO_SKILL on the relationship between DIR_SKILL and FIRM_PERF (H2). The corresponding equation is as follows:
FIRM_PERF = β0 + β1 DIR_SKILL + β2 CEO_SKILL + β3 DIR_SKILL*CEO_SKILL + β4 BOARD_SIZE + β5 GENDER_DIV + β6 ROA + β7 SALES_GR + β8 LEV + β9 FIRM_SIZE + ε
Finally, a moderated regression was conducted to examine the three-way interaction of DIR_SKILL, CEO_SKILL, and CEO_POW on FIRM_PERF (H3). The corresponding equation is as follows:
FIRM_PERF = β0 + β1 DIR_SKILL + β2 CEO_SKILL + β3 CEO_POW + β4 DIR_SKILL*CEO_SKILL + β5 DIR_SKILL*CEO_POW + β6
CEO_SKILL*CEO_POW + β7 DIR_SKILL*CEO_SKILL*CEO_POW + β8 BOARD_SIZE + β9 GENDER_DIV + β10 ROA + β11 SALES_GR + β12 LEV + β13 FIRM_SIZE + ε
Because the independent variables in the two-way and three-way interactions (i.e., DIR_SKILL, CEO_SKILL, and CEO_POW) consist of different measurement levels, it was impossible to draw correct interpretations from the interaction variables. Therefore, the independent variables were first standardized before performing the regression analyses since a unit of measurement into which any scale of measurement can be converted was needed.
22
4. Findings
This section presents the findings for the data analysis of the proposed relationships. First, the descriptive analysis summarizes the data of the sample. Second, a bivariate analysis draws the first indications of the possible relationships between the concepts of interest and identifies whether there are multicollinearity issues. Finally, multivariate analyses provide further information about the results of possible relationships between the concepts of interest.
4.1 Descriptive analysis
Table 4 reveals the descriptive statistics of the variables used in this study. The average Tobin’s Q (FIRM_PERF) is 1.36, with a minimum of 0.00 and a maximum of 8.73. With respect to the skills, the average number of skill categories present on the board of directors is 8.72, with minimum and maximum values of, respectively, 4 and 15. Concerning the skills of the directors, the sample includes 1,143 directorships, of which 100 also serve as CEO inside the same company; this means that every CEO from the 100 sample companies also has a place on the company’s board of directors. The average CEO_SKILL is 0.57, which means that the number of skill categories of the CEO is, on average, marginally higher than that of the other directors. In addition, the average skill categories of CEOs do not deviate substantially from those of the other directors; the average number of skill categories per director and CEO are, respectively, 4.4 and 4.8. This indicates that the average level of human capital, as measured by skills, between CEOs and other directors is quite similar. Furthermore, the board of directors with the highest level of skill diversity has directors with skills across 15 skill categories; in contrast, the board of directors with the lowest level of skill diversity has directors with skills across 4 skill categories. Moreover, since CEO_POW is the sum of CEO_DUALITY, RATIO_APP_AFTER_CEO, and RATIO_CEO_SHARES, its possible minimum and maximum values are, respectively, 0 and 3. The highest value of CEO_POW found in the sample is 2.99. Since the CEO corresponding to the highest value of CEO_POW in the sample is also the founder of the company he serves (see the 2015 proxy statement of Fedex Corp.), high values of CEO_POW could indicate that the CEO is also the founder of the company. Low values of CEO_POW may indicate that the CEO is not a long time active in the company where he or she serves.
23 Table 5 reveals the presence and average intensity per skill category on the sample companies’ board of directors. The skill categories “Financial, Accounting, Investment, and Banking” (with a score of 100%) and “Leadership” (with a score of 99%) are the two most common skill categories present on the boards of directors of the sample companies. In contrast, the skill categories “Supply Chain, Transportation, and Logistics” (with a score of 9%) and “Community, Media Relations, and Investor Relations” (with a score of 15%) are the two least common skills categories present. With respect to the average intensity per skill category (i.e., the number of directors on the company’s board who have the respective skill to the total number of directors), the skill categories “Leadership” (with an intensity score of 87%) and “Supply Chain, Transportation, and Logistics” (with an intensity score of 3%) have the highest and lowest score. Since the average board size is 11, this means that on average 10 directors (87% of 11) on a board have leadership skills.
Variable N Mean Median Std. Dev. Minimum Maximum Range
Dependent variable FIRM_PERF 100 1.36 0.98 1.46 0.00 8.73 8.73 Independent variables DIR_SKILL 100 8.72 9.00 2.26 4.00 15.00 11.00 CEO_SKILL 100 0.57 0.56 0.23 0.17 1.00 0.83 CEO_DUALITY 100 0.49 0.00 0.50 0.00 1.00 1.00 RATIO_APP_AFTER_CEO 100 0.37 0.36 0.28 0.00 1.00 1.00 RATIO_CEO_SHARES 100 0.50 0.55 0.31 0.00 0.99 0.99 CEO_POW 100 1.50 1.46 0.83 0.08 2.99 2.91 Control variables BOARD_SIZE 100 11.43 11.00 2.29 8.00 24.00 16.00 GENDER_DIV 100 0.21 0.21 0.09 0.00 0.47 0.47 ROA 100 0.13 0.11 0.08 0.01 0.43 0.42 SALES_GR 100 0.03 0.00 0.29 -0.42 2.46 2.88 LEV 100 3.25 1.81 3.99 -3.79 20.51 24.30 ASSETS 100 123,686 24,031 343,423 3,663 2,572,773 2,569,110 Table 4 Descriptive statistics
24
4.2 Correlation analysis
Table 6 reveals the Pearson’s correlations containing all the variables used in this study. Since CEO_DUALITY, RATIO_APP_AFTER_CEO, and RATIO_CEO_SHARES are considered as the three items for the variable CEO_POW in this study, it is not surprising that all these three items correlate highly with CEO_POW. In addition to this, the correlations of the three items and CEO_POW suffer from multi-collinearity. However, the variance inflation factors are not higher than 10, and I did not consider them as separate variables but rather relied on the sum of these variables. Therefore, the exclusion of one or more of these items of CEO_POW for further analysis did not seem necessary.
No. Skill category Presence Intensity
1 Industry and Relevant Industry 0.93 0.55
2 Financial, Accounting, Investment, and Banking 1.00 0.60
3 International 0.80 0.48
4 Governmental, Political, Regulatory, and Law 0.57 0.19
5 Academic and Education 0.28 0.06
6 Leadership, Management, and Public Company 0.99 0.87
7 Sales, Marketing, and Brand Management 0.55 0.21
8 Technology, R&D, Innovation, and Science 0.56 0.19
9 Manufacturing, Constructing, and Engineering 0.18 0.04
10 Risk Management, Risk Oversight, and Crisis 0.56 0.26
11 Strategy and Operations 0.82 0.53
12 Supply Chain, Transportation, and Logistics 0.09 0.03
13 Human Resources, Talent Management, and Compensation 0.31 0.11
14 Community, Media Relations, and Investor Relations 0.15 0.05
15 Mergers and Acquisition 0.24 0.10
16 Information Technology, and E-commerce 0.23 0.05
17 Legal, Compliance, and Ethics 0.27 0.06
18 Environmental, Sustainability, Safety, and Health 0.19 0.05
Table 5
25 Next, the findings of the hypotheses are discussed. Based on the study of Jermias and Gani (2014), I performed collinearity diagnostic tests simultaneously with the regression analyses. The stability of the models decreases when the variance inflation factor (VIF) increases. All the variables have VIFs between 1 and 2. This indicates that multi-collinearity does not hinder the interpretation of the results of the regression analyses.
4.3 Hypothesis 1
H1 predicts that skill diversity on boards and firm performance are positively related. This hypothesis concerns the incremental effect of DIR_SKILL on FIRM_PERF. Although the relationship between DIR_SKILL and FIRM_PERF has not been examined before, the control variables used in this model are consistent with previous studies (Anderson et al., 2011). Table 7 summarizes the findings of the regression analysis for H1 by presenting two models. Model 1 contains all the control variables and the dependent variable FIRM_PERF. Model 2 includes in addition the independent variable DIR_SKILL. The R-squared(0.301) of Model 1 reveals that the control variables account for 30.1% of the variation in FIRM_PERF. The R-squared(0.302) of Model 2 reveals that the combination of the control variables and DIR_SKILL accounts for 30.2% of the variation in FIRM_PERF. This indicates that DIR_SKILL does not explain a large part of the variation in FIRM_PERF. With respect to the coefficient for DIR_SKILL, Table 7 reveals that the level of skill diversity on the various boards of directors is not positively and significantly related to FIRM_PERF (β = -0.026, p = 0.671). Thus, these findings do not support H1.
1 2 3 4 5 6 7 8 9 10 11 12 13 1 FIRM_PERF 1 2 DIR_SKILL -0.113 1 3 CEO_SKILL 0.122 0.187 1 4 CEO_DUALITY -.285** 0.149 0.081 1 5 RATIO_APP_AFTER_CEO 0.008 0.054 -0.030 .351** 1 6 RATIO_CEO_SHARES -0.113 -0.053 0.050 .396**.338** 1 7 CEO_POW -0.196 0.078 0.051 .858**.677**.703** 1 8 BOARD_SIZE -0.116 .221*-0.065 0.087 0.099 -0.114 0.035 1 9 GENDER_DIV -0.010 .198* .203* 0.138 -0.037 0.178 0.121 0.082 1 10 ROA .619**-0.112 0.051 -.247*-0.166 -0.133 -.239*-0.067 -0.047 1 11 SALES_GR 0.078 0.023 -0.068 -0.125 -0.027 0.076 -0.049 -0.005 -0.035 -0.085 1 12 LEV -.233* .255* 0.005 .290**-0.070 0.133 0.188 0.151 .211*-.298**-0.028 1 13 FIRM_SIZE -.238* .206*-0.033 .258** 0.117 0.073 .213* .218* 0.137-.324**-0.030 .509** 1 Table 6 Correlation analysis
**. Si gni fi ca nt correl a tion a t the 1% l evel (2-tai l ed). *. Si gni fi ca nt correl a tion a t the 5% l evel (2-tai l ed).
26
4.4 Hypothesis 2
H2 expects that the positive relationship between board skill diversity and firm performance is stronger when a CEO has a broader range of skills. This hypothesis relates to the incremental effect of CEO_SKILL on the relationship between DIR_SKILL and FIRM_PERF. Table 8 reveals the results of the regression analysis for H2. Contrary to my prediction, the coefficient for DIR_SKILL*CEO_SKILL is negative and not significant (β = -0.055, p = 0.679). Therefore, these findings do not support H2.
Variable Exp. Sign Coefficient Est. (β) Std. Error Coefficient Est. (β) Std. Error
Constant 0.427 0.757 0.573 0.835 Independent variable DIR_SKILL + -0.026 0.060 Control variables BOARD_SIZE + -0.130 0.570 -0.008 0.058 GENDER_DIV + -0.220 1.486 0.072 1.509 ROA + 9.076 *** 1.692 9.067 *** 1.699 SALES_GR + 0.625 0.447 0.632 0.449 LEV + -0.019 0.038 -0.160 0.039 FIRM_SIZE + 0.000 0.000 0.000 0.000 n F-statistic Sig. R2 Highest VIF 1.472
*. Si gni fi ca nce a t the 5% l evel (two-tai l ed). ***. Si gni fi ca nce a t the 0,1% l evel (two-tai l ed).
Model 1: Regression
(dependent variable: FIRM_PERF)
Model 2: Regression
(dependent variable: FIRM_PERF)
Table 7
Regression results of DIR_SKILL on FIRM_PERF (H1)
100 100 0.302 5.686 0.301 6.662
**. Si gni fi ca nce a t the 1% l evel (two-tai l ed).
0.000 0.000
27
4.5 Hypothesis 3
H3 posits that when a CEO has more power, the moderating effect of the CEO skills on the relationship between directors’ skills and firm performance is mitigated. This hypothesis concerns the incremental effect of the three-way interaction between DIR_SKILL, CEO_SKILL, and CEO_POW on FIRM_PERF. Table 9 reveals the findings of the regression analysis for H3. Contrary to my prediction, the coefficient for DIR_SKILL*CEO_SKILL*CEO_POW is positive and not significant (β = 0.036, p = 0.810). Hence, these findings do not support H3.
Variable Exp. Sign Coefficient Est. (β) Std. Error Coefficient Est. (β) Std. Error Coefficient Est. (β) Std. Error
Constant 0.427 0.757 0.360 0.780 0.310 0.793 Independent variables DIR_SKILL + 0.030 0.154 0.026 0.155 CEO_SKILL + 0.182 0.151 0.174 0.153 DIR_SKILL * CEO_SKILL + -0.055 0.133 Control variables BOARD_SIZE + -0.130 0.570 -0.001 0.058 0.000 0.059 GENDER_DIV + -0.220 1.486 -0.326 1.541 -0.350 1.549 ROA + 9.076 *** 1.692 8.926 *** 1.699 9.043 *** 1.730 SALES_GR + 0.625 0.447 0.665 0.449 0.663 0.451 LEV + -0.019 0.038 -0.014 0.039 -0.014 0.039 FIRM_SIZE + 0.000 0.000 0.000 0.000 0.000 0.000 n F-statistic Sig. R2 Highest VIF
*. Si gni fi ca nce a t the 5% l evel (two-tai l ed).
Model 3: Moderated regression
(dependent variable: FIRM_PERF)
100 4.583 0.314
***. Si gni fi ca nce a t the 0,1% l evel (two-tai l ed). **. Si gni fi ca nce a t the 1% l evel (two-tai l ed).
0.000 0.000
0.000
1.445 1.474 1.517
Table 8
Regression results of the moderating effect of CEO_SKILL on the relationship between DIR_SKILL and FIRM_PERF (H2)
6.662 5.182
0.301 0.313
Model 1: Regression
(dependent variable: FIRM_PERF)
Model 2: Regression
(dependent variable: FIRM_PERF)
