The impact of board nationality diversity and board cultural diversity on firm performance - A European inquiry
V.S. Rijpkema s2162490
v.s.rijpkema@student.rug.nl 13-01-2017
MSc International Financial Management Faculty of Economics and Business
University of Groningen
Supervisor: dr. V. Purice Co-Assessor: dr. R.O.S. Zaal Abstract
This study examines the impact of nationality diversity in boards of directors on firm performance.
This research adds to current literature by analysing a unique country balanced panel dataset and by the combined examination of foreign director’s presence in the board, and board director’s foreignness to the firm’s country as measured by cultural distance. Using a sample of 100 firms from the United Kingdom, Spain, Italy, Germany, the Netherlands and Belgium, I find that board nationality diversity and cultural diversity have a positive significant effect on firm performance as measured by Tobin’s q and ROA.
Keywords: Board nationality diversity, firm performance, cultural distance, corporate governance, board of directors
JEL classification: G3
13,002 words
Table of contents
1. INTRODUCTION ---3
2. THEORETICAL PERSPECTIVES ON THE BOARD DIVERSITY-PERFORMANCE RELATIONSHIP ---4
2.1 A
GENCYT
HEORY---4
2.2 R
ESOURCED
EPENDENCET
HEORY---5
2.3 C
ULTURAL DIVERSITY---6
3. EMPIRICAL RESEARCH AND THEORETICAL PREDICTIONS ON THE DIVERSITY- PERFORMANCE RELATIONSHIP ---7
3.1 B
OARD GOVERNANCE AND FROM DIVERSITY TO NATIONALITY DIVERSITY---7
3.2 B
OARD NATIONALITY DIVERSITY COSTS---8
3.4 M
EASURING CULTURAL DIVERSITY--- 10
4. METHODOLOGY --- 12
4.1 S
AMPLE SELECTION--- 12
4.2 V
ARIABLES--- 13
4.2.1 Dependent variable firm performance: Tobin’s q and ROA --- 13
4.2.2 Independent variable --- 13
4.2.3 Control variables --- 14
4.3 R
EGRESSION MODEL--- 15
5. RESULTS --- 17
5.1 D
ESCRIPTIVE STATISTICS--- 17
5.2 C
ORRELATION MATRIX--- 19
5.3 P
OOLEDOLS --- 21
5.4 R
ANDOM AND FIXED EFFECT REGRESSIONS--- 23
6. CONCLUSION --- 27
7. LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH --- 28
APPENDIX --- 30
REFERENCES--- 34
1. Introduction
Corporate boards are essential in forming strategy and investment decisions for firms, and hence forming these corporate boards are key tasks for any firm. The importance of the board of directors also explains why they are among the most academically studied elements of larger sized firms. Globalized firm operations are no new feat, and including foreign directors in corporate boards is not either. However, modern technology is currently evolving cultural, political, and societal relations spanning across national boundaries, which might impact the role of nationality diversity and cultural diversity within firms. Some dynamics of board characteristics and diversity within the board have been researched intensively over the years. Page (2007) argues that diversity of demographics in social groups, including those in firms, is generally beneficial as individuals from different backgrounds are likely to have different perspectives on issues due to their individual life experiences.
While board demographics and effects of board diversity such as age, gender and education have been academically studied extensively over the years, the field focusing on the antecedents and consequences of board nationality diversity is still emerging (Hooghiemstra et al., 2016). However, previous studies focusing on the costs and benefits of diversity within corporate boards seem to be divided into two camps. On the one hand diversity is argued to be beneficial to the firm as the source of knowledge and expertise becomes more heterogeneous and hence the risk of relying on a single or homogenous source of knowledge and expertise are mitigated (Estélyi and Nisar, 2016). On the other hand, studies argue that even if a more heterogeneous input is potentially useful, it also brings new challenges. For example differences in personnel can lead to conflicts and increased informational complexity and thus offset the potentially beneficial more heterogeneous input. However, a counter argument to the increased complexity is that diversity is usually present in corporate boards regardless of foreign national presence. In other words, studies on board nationality diversity are not conclusive.
This research adds to current literature by analysing a unique country set and by the combined
examination of foreign director’s presence in the board, and board director’s foreignness to the firm’s
country as measured by cultural distance. Rather than a more common examination of cultural
differences between groups, the constructed cultural diversity measure examines the impact of cultural
diversity within groups, an uncommon and recent concept in finance literature. Using a sample with a
balanced panel dataset of 100 firms from the United Kingdom, Spain, Italy, Germany, the Netherlands
and Belgium, I find that board nationality diversity and cultural diversity have a positive significant
effect on firm performance as measured by Tobin’s q and ROA. These findings contradict recent
research and suggest foreign board presence increases board effectiveness by increasing the diversity
of inputs and experience, outweighing costs as for example lack of knowledge of local (accounting)
rules and friction costs related to difficult coordination and slower communication. A limitation in
researching board diversity is a lack of panel data on board members on firms each year, which limits how well their effect on for example firm financial performance can be measured.
The research question for this study is: “Does a higher share of foreign directors in the board of directors increase firm performance?” The association between the presence of foreign board directors and firm performance is investigated by using a sample of over 700 unique directors, working at 100 firms from six countries.
The remainder of this paper is organized as follows. Section two and three introduce relevant theories, review empirical evidence related to board nationality diversity and cultural diversity, and provide theoretical predictions. The fourth section discusses the sample selection, methodology and descriptive statistics. The fifth section provides an analysis of the empirical findings, which are discussed in the sixth section. Finally, limitations of this research and suggestions for future examinations are provided in the last section.
2. Theoretical perspectives on the board diversity-performance relationship
The board of directors’ responsibilities include the monitoring and control of managers, providing these managers with information and counsel, to ensure compliance with laws and regulations and ensuring the firm fits to the external environment and constituencies (Carter et al, 2010). However, how individual board members perceive and perform their task influences board decisions, which partially determines firm performance (Kiel and Nicholson, 2003).
As a starting point for theorizing on the performance effect of board nationality diversity, we first introduce two theories focusing on how individual board members impact corporate decisions and consequently firm performance. These theories are agency theory and resource dependence theory.
Subsequently, the effect of cultural diversity in boards is discussed by Hofstede’s culture framework.
2.1 Agency Theory
According to agency theory, the interests of owners (principals) and managers (agents) of
firms are not always aligned, due to the separation of ownership and control (Jensen and Meckling,
1976). Corporate governance mechanisms are set up to align the interests of shareholders and
managers and one of such corporate governance mechanism is the board of directors (Jensen and
Meckling, 1976). One of the responsibilities of the board of directors is the monitoring and advising
role to managers. An important aspect for effective board monitoring and control of managers is board
independence. Carter, Simkins and Simpson (2003) argue that board independence is increased through higher board diversity, as more diverse boards result in enhanced monitoring of managers.
Foreign directors can increase board independence, because as new outsiders they have a strong motivation to build up their position as capable managers and have less to lose by challenging the status quo than managers that hold longer tenure and are more dependent (Carter, Simkins and Simpson, 2003). Therefore, a higher proportion of foreign directors may develop a board's monitoring ability and can in turn affect firm performance. However, Carter, Simkins and Simpson (2003) do note that agency theory does not offer an unambiguous prediction on the association between board
diversity and firm financial performance. Additionally, Monks Minow (2004) and Jensen (1993) show that there are more important factors increasing the willingness of directors to monitor than
independence, such as high equity ownership. In conclusion, while agency is often cited in corporate governance literature, it does not offer an unequivocal case for a positive association between board diversity and firm financial performance, but its possibility is not ruled out.
2.2 Resource Dependence Theory
While agency theory is often used in studies that investigate board influence, resource dependency theory can be used to explain individual director influence better (Carter et al, 2010).
Resource dependency theory views the board's role as linking the firm to external organizations in order to address environmental dependencies (Pfeffer and Salanic, 1978). Resource dependency as stated in the original study by Pfeffer and Salancik (1978) suggest linking the firm to its external environment offers several benefits. It (1) creates diverse channels of communication with vital stakeholders to the firm; (2) provides resources like information and expertise; (3) creates legitimacy of the firm within the external environment and (4) provides support from relevant groups in the external environment. Hillman, Cannella and Paetzold (2000) add to these primary benefits a number of director types such as business experts and support specialists which each provide different types of resources to the firm.
Similarly, Broome et al. (2011) suggest that a more diverse board offers many different perspectives and widens the circle of network connections, enhancing innovativeness and creativity.
Moreover, Carter et al (2010) suggest that a board consisting of different groups of people is a positive
signal to the market. In addition to providing unique resources, foreign representation on boards can
provide legitimacy in the view of suppliers, customers and shareholders. It may even provide
legitimacy to the labour market, even though appointing a director only for characteristics such as their
gender or ethnicity rather than resources and skills to stimulate a firm's image and reputation is
referred to as tokenism (Kesner, 1988).
The shift in US boardrooms over the past decades towards including more women and ethnically diverse directors is shown by Hillman, Cannella, and Harris (2002). However, studying the effect of increased board diversity is important from a resource dependence theory view to understand how firms benefit from added unique human capital, external networks and information. Carter et al.
(2010), pp 397 provide an example by quoting Karen J. Curtin, former executive vice president of Bank of America. Mrs. Curtin argues “There is real debate between those who think we should be more diverse because it is the right thing to do and those who think we should be more diverse because it actually enhances shareholder value”. Firms should benefit from studies on corporate governance diversity by understanding the possible unique competences and resources that foreign directors bring, while understanding the costs. This would also eliminate the practice of tokenism, appointing foreign directors only as a legitimacy signal to the market. Similarly, while gender diversity is of particular interest in some countries due to board regulations stipulating a minimum amount of females to be represented on boards, laws for including a minimum amount of foreign directors are not known. Therefore, it is assumed that foreign directors are more likely recruited for providing the firm with abovementioned resources and skills, best linking the firm to its external environment.
2.3 Cultural diversity
Studying board nationality diversity inevitably includes culture. A fundamental proposition is that even in the contemporary globalized business environment, national cultures influence corporate decisions through managerial decision making and through firm- and country-level characteristics.
Prior investigations have distinguished three levels of cultural influence: (1) country level influence (institutions, laws, regulations, and market development), (2) firm level influence (remuneration practice and stock ownership) and (3) individual level influence (director attitude to risk).
Acknowledging cultural diversity aligns with both agency theory and resource dependency theory. Cultural diversity induces unique information, diverse knowledge and different perspectives (Nederveen Pieterse et al., 2013). Foreign directors could offer firms specific knowledge of their home countries which enhances operations in the market of those directors. This is akin to how Masulis et al. (2012) demonstrated that foreign independent directors positively influence cross-border acquisitions when target firms are from their home country. However, large cultural diversity is linked to frictions such as difficult coordination, slower communication, confusion and a frequent source of misunderstanding (Anderson et al., 2011) and lower intragroup trust (Bjørnskov, 2008).
Hofstede’s culture framework is the result of a very comprehensive study on how values in the
workplace are influenced by culture. Culture is defined by Hofstede (1980) as "the collective
programming of the mind distinguishing the members of one group or category of people from others". The original framework consists of four cultural dimensions and each represents independent preferences for one state of affairs over another, distinguishing countries from each other. Hofstede’s framework therefore incorporates a relative measurement of how individuals from a certain culture have workplace values different to another, which influences how information is interpreted and what decisions are made (Maznevski, 1994).
3. Empirical research and theoretical predictions on the diversity-performance relationship
3.1 Board governance and from diversity to nationality diversity
Corporate governance refers to the system by which the interests of shareholders, board directors, management and other (financial) stakeholders are managed (Shleifer and Vishny, 1997).
Suppliers of capital demand a return on their investment, but need a control mechanism in order to get managers to return some of the firm’s profits to them. Agency theory comprises a large field of academic literature that is concentrated on this. Agency literature often focuses on outside directors to monitor board performance (Dewally and Peck, 2010). As argued by Estélyi and Nisar (2016), this makes a board director's nationality an important issue. In addition to regular credentials that would lead a director to be appointed, appointing a foreign director could signal shareholders and potential investors the firm’s efforts to deal with the complex operational challenges. For instance, more diverse boards with directors linked to markets that the firm is active in could lead to more seized opportunities and so increase firm performance, in addition to enhanced international networking (Oxelheim and Randøy, 2003). For example Masulis et al. (2012) have found foreign independent directors (FIDs) at U.S. firms to make better cross-border acquisitions when target firms are from the respective regions of the FIDs. Furthermore, foreign directors might offer different perspectives due to their different backgrounds and experiences. The more heterogeneous group input can mitigate risks related to relying on a homogenous source of expertise (Estélyi and Nisar, 2016). Furthermore, monitoring effectiveness might be improved if current board members highly appeal to each other and are motivated to keep each other on the board (Forbes and Milliken, 1999). In other words, foreign board members are assumed to be able to provide more critical and independent opinions as they are not yet part of the social network of current board members (Choi et al., 2007).
The combined effect of including a foreign board member is suggested to increase firm
performance by seizing more foreign opportunities, increased board monitoring effectiveness and a
more critical environment. This line of reasoning suggests a positive association between board
nationality diversity and firm performance. To summarize, there are many interdisciplinary sets of
theories with a solid reasoning for the link between board diversity and firm financial performance.
However, there is too little empirical evidence to conclude on a clearly supported positive relationship.
The following paragraph presents a reasoning and empirical evidence for a negative association between board nationality diversity and firm performance.
3.2 Board nationality diversity costs
While there are sound arguments and findings showing how firms can benefit from increasing board internationalization, there are also studies that find no relation or even costs to increasing board nationality diversity. While a foreign director could enhance firm performance by for example new or different input and international expertise, these positive effects could be offset by negative effects.
Adding a foreign director could result in increased informational complexity and conflicts due to for example language and cultural differences, resulting in financial distress (Santen and Donker, 2009).
Similarly, Anderson et al. (2011) report that friction in culturally diverse boards is related to more difficult coordination, slower communication, more confusion and a source of misunderstanding. In a study on cultural diversity within corporate boards, Frijns et al. (2016) find that cultural diversity negatively affects firm performance. Their study suggests that potential benefits are not outweighed by frictions due to cultural diversity. Farrell and Hersch (2005) suggest firms seek members of minority groups to build an image of inclusiveness after finding no major impact on firm performance by including female board members. In other words, appointing directors more for what they are (foreign nationals) – than who they are (experts) could then negatively influence firm performance.
Furthermore, local laws, regulations and governance standards and accounting rules might be different than in the foreign director's home country (Masulis et al., 2012). This is argued to negatively affect their task to evaluate managerial performance or challenge managerial decisions. However, this effect might be stronger for directors originated from outside the EU, as EU laws have a converging effect on regulations within EU countries, as can be seen by for example IFRS accounting regulations (Barth et al., 2008).
The combined effect of these negative relations between foreign directors and effective board governance suggest a negative association between board nationality diversity and firm performance.
However, due to a complex interaction of both benefits and costs, either a positive or a negative association is a realistic possibility based both on theory and empirical evidence. In conclusion, because the influence of board nationality diversity on firm performance is ambiguous, the first hypothesis is stated in null format and testing is done two-tailed.
Hypothesis 1: Board nationality diversity has an effect on firm financial performance.
3.3 Performance difference due to foreignness
While the previous proposition examines whether nationality diversity influences firm performance as a result of increased board effectiveness, it is important to acknowledge that some countries are more similar or dissimilar. The national context of the firm may play an important role in the board nationality diversity-firm performance relationship. Similarly, the national origin of individual directors reflects the institutional environment of the country they have spent the majority of their formative years in (Hambrick et al., 1998). The combination of legal, political and cultural institutions partially determines how individuals and organizations deal with uncertainty, how tasks are perceived and appropriate actions are taken (Crossland and Hambrick, 2007).
The complex interactions of individual, firm and national characteristics are reflected in empirical findings on this matter. As previously indicated, Masulis et al. (2012) have found foreign independent directors (FIDs) to make better cross-border acquisitions for their home country’s target firms. However, they have also found poorer firm performance for firms with FIDs when the FID’s home region becomes less important to the firm. Miletkov et al. (2013) also recognize both benefits and costs associated with employing foreign directors, but have examined their impact on firm performance outside the USA in a cross-country study. They show that foreign director’s effect on firm performance is more positive in countries with lower levels of investor protection for cross- border protection, but poorer performance in countries with better legal institutions. Similarly Oxelheim and Randøy (2003) find an increase in firm value by an outsider Anglo-American board presence in Norwegian and Swedish firms. Their study suggests enhanced international orientation and a stimulus for further globalization of firms due to the firm's commitment to the Anglo-American corporate governance model, seen as stiffer, more demanding and transparent in corporate monitoring (Oxelheim and Randøy, 2003).
These results suggest that corporate governance could be differently affected by how foreign a
director is to the firm's country’s corporate governance system and culture. However, the focus of this
study is an examination of how board nationality diversity rather than cultural diversity influences
firm performance. Including an extensive examination of how the national culture and its institutions
of each foreign director’s home country moderates the director’s influence on board effectiveness and
in turn firm performance is out of the scope for this research. Instead, I will draw upon previous
cultural diversity research and construct a measure of cultural distance within a firm’s board (Kogut
and Singh, 1988; Beugelsdijk and Frijns, 2010; Frijns, Dodd, and Cimerova, 2016). By doing so, the
third proposition still incorporates the suggestion that nationality diversity of directors influences
board governance effectiveness, which in turn affects firm performance. However, firm performance is
influenced differently depending on how foreign a director is to the country's culture and corporate governance style. This proposition is examined by classifying directors as culturally similar or dissimilar the firm's country. It allows running a regression that includes the percentage of people from a different cultural background in a manner similar to how Carter et al (2010) have examined director's ethnicity effects on firm performance. Moreover, it is in line with previous research on cultural diversity in boards of Frijns, Dodd and Cimerova (2016), who find cultural diversity within boards negatively impacts firm performance measured by Tobin's q and ROA.
Similar to how increased nationality diversity may offer both benefits and costs, the net effect of cultural diversity on firm performance depends on whether the positive aspects outweigh the negative aspects. It is noted that the negative aspects may be mitigated better when firms have particular benefits of having culturally diverse boards. Directors from different national cultural backgrounds may be an important information source to firms with higher percentages of foreign sales (Masulis et al., 2012). Therefore, the ratio of foreign sales to total sales is included as a control variable. However, the main proposition is that firms are more likely to perceive negative effects on firm performance when cultural distance between foreign director’s national culture and the firm’s national culture is high.
Hypothesis 2: There is a negative association between foreignness of directors and firm financial performance.
3.4 Measuring cultural diversity
To test the second hypothesis, a measure of cultural diversity on boards is constructed. This construct is based on four cultural dimensions that characterize different cultural traits of a nation, stemming from Hofstede’s original study in 1980. The dimensions are power distance (PDI), individualism-collectivism (IDV), masculinity-femininity (MAS), and uncertainty avoidance (UAI).
The values of countries scored for each dimension were obtained from a respondent group of global IBM managers from 70 countries from 1967-1973 (Hofstede, 1980).
The power distance dimension indicates the degree to which less powerful members of society accept and expect power to be distributed unequally. A high PDI score reflects a hierarchical society and a low PDI a society where people strive for equal distribution of power and justification for inequalities in that distribution.
The individualism-collectivism dimension indicates whether individuals prefer a loosely-knit
social framework where people are likely to mostly care for themselves and immediate families (high
individualism) versus a tightly-knit social framework where people are expected to look after extended relatives as well (high collectivism). A low IND dimension score reflects high individualism and a sense of “I”, whereas a high IND dimension score reflects high collectivism and a sense of “we”.
A high score on the MAS index reflects a masculine and competitive society with a preference for achievement, heroism, assertiveness and material rewards for success. A lower score represents a more feminine and consensus-oriented society with a preference for cooperation, modesty, caring for the weak, and quality of life.
The uncertainty avoidance dimension indicates the degree to which individuals feel uncomfortable with uncertainty and ambiguity. A low score reflects more relaxed attitudes where practice counts more than principles, while a high score reflects societies with rigid codes of belief and behaviour, and less tolerance to uncommon behaviour and ideas.
The measure of cultural diversity of firm boards is constructed by computing the cultural distance between individual cultural dimensions
1of the firm’s directors and the firm’s country, following Kogut and Singh (1988):
where CD
ijis the cultural distance between each two countries (i, j), I
kiis the culture score on dimension k for a country i, I
kjis the cultural score on dimension k for country j, and Vk is the in- sample variance of the score for the specific cultural dimension. Consequently, the cultural diversity scores of each firm's board is constructed per firm year and normalized for the size of the board. The larger CD
ij, the more two different cultures differ from each other.
1
The scores on Hofstede’s cultural dimensions are extracted from https://geert-hofstede.com/countries.html,
Retrieved December 1, 20164. Methodology
4.1 Sample selection
The sample of this study consists of 100 firms from a population of publicly traded firms from Germany, the United Kingdom, France, Italy, Spain and the Netherlands in the period 2014- 2015. The included countries are selected as they are the six largest EU countries in terms of GDP and have well developed board governance institutes, albeit somewhat different in terms of board structure and independence of directors (Hopt and Leyens, 2004) and because financial data is readily available.
Firm data is obtained from Orbis, as it is a quite comprehensive database which includes both firm and director information, it can handle user-specific proxy’s and lets users construct, save and export samples. However, it is prone to errors such as double entries, for example some directors are shown twice for the same company with slightly different first, middle or family names. Hence, the sample will require double-checking in addition to coding for national versus foreign directors for the dependent variable board nationality diversity. The initial sample of 125 firms was a random selection of 4588 listed companies for which Orbis has data on directors active in companies from the included countries Germany, the United Kingdom, France, Italy, Spain and the Netherlands between 2014 and 2015.
After correcting for missing values and removing financial firms, the data is based on a random sample of 100 traded companies. The final sample includes 54 firms from the United Kingdom, 24 from France, 11 from Italy, 8 from Spain, 2 from the Netherlands and 1 from Germany and contains 200 firm-year observations in the period 2014-2015. Financial firms were removed as they are subject to different regulations and accounting standards (Jackling and Johl, 2009). Missing data on board director’s appointment date, age, gender and nationality was supplemented by information from annual reports and Bloomberg. Additionally, Orbis does not provide panel data on board characteristics per firm year, so to construct a balanced panel dataset all included firm’s annual reports for the included years have been checked for board composition. Director tenure was calculated by calculating the difference between director appointment dates and the respective year- ends. The panel data have both a cross-sectional and a time series dimension, so while the selection of firms is random, the same firms are investigated at subsequent points in time.
All used variables in this research are provided in Table 1.A in the appendix, and described in
more detail in the following section.
4.2 Variables
4.2.1 Dependent variable firm performance: Tobin’s q and ROA
The dependent variable in the performed regressions is the firm's financial performance. To measure this performance two alternative performance indicators are used. Financial performance is measured by return on assets (ROA) as a proxy for accounting-based performance and Tobin's q (TobinsQ) as a proxy for market-based performance at the year-ends of 2014 and 2015. Following corporate governance literature, Tobin’s q is used to capture firm value (Estélyi and Nisar, 2016; Lang and Stulz, 1994; Masulis et al., 2012; Yermack, 1996). The board of directors likely influences firm’s investment decisions. Including Tobin’s q allows to measure how firm value is created by help of the board. Tobin's q is computed as the sum of market value of the firm’s assets (market value of equity, book value of debt and book value of preferred stock), divided by the replacement value of the firm’s assets (book value of total assets). If Tobin’s q is one, then the market value of the shareholders and creditors investment in the firm is equal to the amortized historical cost of the firm’s assets (Carter et al, 2010). However, it is a market based measure because it is also a forecast of future cash flows and an assessment by investors on the opportunities of the firm (Chung and Pruitt, 1994).
Return on assets (ROA) is included as an additional, accounting based, measure of profitability that is commonly used to indicate firm performance. ROA is calculated as net income divided by the book value of total assets at the end of the year. It is a more traditional indication of firm performance, indicating how well the firm is able to produce accounting based revenues in excess of expenses. The net income based ROA is used as it is of more value to shareholders than the ROA measure based on income before interest, taxes, depreciation and amortization (EBITDA) (Bhagat and Bolton (2008).
4.2.2 Independent variable
The independent variable in this research is board nationality diversity. It is defined as the
proportion of foreign directors (NationalityDiversity) to the total size of the board. For the second
hypothesis, the cultural diversity of boards is measured as the average of cultural distances between
foreign director's on a board and the firm's country (CDBoard).
4.2.3 Control variables
Average board tenure (Tenure) is included, because tenure affects how efficiently information is processed, and nationality diversity has been shown to be more positively related to performance in longer tenured teams (Nielsen and Nielsen, 2009). It is defined as the average number of years each board director has been on the firm's board.
Gender of boards is included following popular gender based corporate governance studies.
These studies often focus on the gender diversity effect by including the percentage of female board directors, in addition to explaining that women have different skills and resources than men. While some studies argue that gender diversity is paired with more diverse opinions and critical thinking which is more time consuming and less effective (Lau and Murnighan, 1998), others add that it may produce both more conflict and creativity and innovation (Williams and O'Reilly, 1998). However, rather than to predict on gender diversity's positive or negative influence on firm performance, it is included on the note that each feat of board demography diversity likely has many complex and conflicting effects affecting board performance (Forbes and Milliken, 1999). It is measured by a ratio of female directors to total board size (Gender). On that same note, age (Age) is another board demographic feat that could reflect experience and preferences for risk taking of directors (Carter, Simkins and Simpson, 2003). Additionally, age is often included in board nationality diversity studies.
It is measured as the average age of all board directors per firm.
A number of firm specific variables are used to account for firm-specific and industry characteristics that potentially correlate. Firm’s foreign sales (ForeignSales) are considered to capture the degree of a firm’s foreign operations. Firms with higher degrees of foreign orientation are likely to have a more nationality diverse board (Masulis et al, 2012; Oxelheim and Randøy, 2003), while the amount of foreign sales might be related to firm performance. The firm’s age (FirmAge) is included because it is shown to affect firm dynamics in several ways. Firm growth, growth variability, and job creation and destruction decreases with firm age and size, while the probability of a firm’s survival increases with age and size (Cooley and Quadrini, 2001). Furthermore, older firms have had more time to expand their professional network outside domestic boundaries, especially as firms internationalize their operations. The value of natural logarithm of total assets (TotalAssetsLn) is included to control for the size of the firm, as this is often shown to be related to market returns and Tobin’s q (Morck, Shleifer and Vishny, 1988; Yermack, 1996).
Several additional variables relate to various aspects of corporate governance mechanisms,
which have been previously shown to be related to firm performance. Board size (BoardSize) is
included as several studies show that Tobin’s q and board size are related, however some find an
inverse relation (Yermack, 1996) while others find a strong positive relation (Dalton et al, 1998;
Jackling and Johl, 2009). Yermack (1996) refers to agency theory for the negative relationship. Dalton et al (1998) argue larger board enhance firm decision making due to better information stemming from a greater and more diverse knowledge pool, following resource dependency theory. Lastly, industry and country dummies are included.
4.3 Regression model
To test the relationship between board nationality diversity and firm financial performance, and cultural diversity within boards and firm financial performance, ordinary least squares (OLS) regression is first used. However, this method brings concerns of endogeneity and omitted unobserved factors (Hermalin and Weisbach, 1988), and
Finding a causal relationship between board nationality diversity and firm financial performance is challenging. Hermalin and Weisbach (2003) argue that board characteristics are often not exogenous random variables. Adams and Feirreira (2007) point out that board characteristics are partly endogenously chosen to fit to firms characteristics and its operational environment (Adams and Ferreira, 2007). Omitted unobserved variables that vary over time and across firms can influence the selection of diverse directors and firm performance (Adams and Feirrera, 2007). To mitigate the endogeneity issue, a panel regression with firm fixed-effects could be employed (Adams and Ferreira, 2009). A Durbin–Wu–Hausman or Hausman specification test is performed. The null hypothesis in this test states that a random effects (RE) panel regression is preferred due to higher efficiency, while the alternative states that a fixed effects (FE) regression is preferred as neither is (highly) efficient but at least the fixed effects model is consistent. Both models assign an intercept to each firm that is constant over time, and both assume equal relations between the dependent variable and independent variables across sections (i) and time (t). The main difference between the models is that an added random variable in the RE model is assigned to each intercept that varies across sections (firms) but is constant over time (Brooks, 2008). The test returns an insignificant result (p-value 0.1429), suggesting a random effects regression should be run. The result of this Hausman test is likely explained by a large cross section (n=100) relative to the time series (time span of two years), making a random effects model more like fixed effects model when the timespan increases. The Hausman test essentially computes whether there is correlation between the idiosyncratic errors (a
i) and explanatory variables across the time periods (Wooldridge, 2010).
The panel data allows for repeated measures on firms (i) over time (t), as data is collected on a
random selection of firms at a given point in time, and those same firms are then re-examined at
subsequent point(s) in time. The implications of this data structure are discussed after introducing the
regression models. The independent variable board nationality diversity (NationalityDiversity) is regressed on firm performance variables Tobins q and ROA alternatively. The subsequent random effects model should more correctly reflect standard errors compared to cross-sections data, as each additional year of data is not independent of previous years (Cameron, 2007). The random effects model assigns an intercept to each firm which is constant over time, and assumes an equal relationship between the dependent and independent variables cross-sectionally and over time.
The fundamental regression model tested for the first hypothesis with a pooled OLS method (base year 2014) is:
Firm performance measure
i,t= α
0+ δ
0y2015
t+ β
1NationalityDiversity
i,t+ β
2ForeignSales
i,tβ
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ β
9- 18IndustryDummy
+ β
19-23CountryDummy +
εi(u
i=0) (t= 1,2)
(1) And for the second hypothesis:
Firm performance measure
i,t= α
0+ δ
0y2015
t+ β
1CDBoard
i,t+ β
2ForeignSales
i,tβ
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ β
9-18IndustryDummy
+ β
19- 23CountryDummy +
εi(u
i=0) (t= 1,2)
(2)
However, it is likely that individual effect u
iis not zero in this panel data, as effects similar to for example experience (tenure) like personality that are not captured in regressors can influence OLS assumptions of endogeneity, and homoskedasticity and autocorrelation. For that, and reasons mentioned both above and in subsequent sections a random effects regression is run. Furthermore, despite an insignificant Hausman test, the FE models are also run as robustness checks because I am interested in capturing firm-specific diversity variation effects on the performance variables.
RE model hypothesis 1:
Firm performance measure
i,t= a
0+ δ
0y2015
t+ β
1NationalityDiversity
i,t+ β
2ForeignSales
i,t+ β
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ β
9- 18IndustryDummy
+ β
19-23CountryDummy
+ ω
it(t= 1,2) (3) (ω
it= ε
i +v
it, where ε
imeasures the random deviation of each entity’s intercept term from the ‘global’
intercept term a, and v
itis the remainder disturbance that varies over time and entities, capturing
anything left unexplained about firm performance).
RE model hypothesis 2:
Firm performance measure
i,t= a
0+ δ
0y2015
t+ β
1CDBoard
i,t+ β
2ForeignSales
i,t+ β
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ β
9-18IndustryDummy
+ β
19-23
CountryDummy
+ ω
it(t= 1,2) (4)
FE model hypothesis 1:
Firm performance measure
i,t= a
0+ δ
0y2015
t+ β
1NationalityDiversity
i,t-1+ β
2ForeignSales
i,t+ β
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ λ
t+ v
it(t= 1,2)
(5) λ
tis a time-varying intercept that captures the variables that affect firm performance and vary over time but are constant cross-sectionally (firms). Additionally, the FE models have no industry and country dummies because the model estimates for the same firms across time, and no firm changes its industry or country within the selected time period.
FE model hypothesis 2:
Firm performance measure
i,t= a
0+ δ
0y2015
t+ β
1CDBoard
i,t-1+ β
2ForeignSales
i,t+ β
3BoardSize
i,t+ β
4Tenure
i,t+ β
5Age
i,t+ β
6Gender
i,t+ β
7TotalAssetsLN
i,t+ β
8FirmAge
i,t+ λ
t+ v
it(t= 1,2) (6)
in which the coefficient of primary interest is β
1and H
0: β
1= 0 and H
a: β
1≠ 0 for the RE and FE regressions. Firm performance measure is either Tobin’s q or ROA alternatively. The definitions of all variables are described in detail in the previous section and a summary list is provided in the appendix in table 1.A.
5. Results
5.1 Descriptive statistics
Table 1 presents sample summary statistics based on 100 listed firms located in the United Kingdom, France, Spain, Italy, The Netherlands and Germany within the period 2014-2015. The average board examined in firm-years includes 15.18% foreign directors, while there are also fully domestic (0%) and fully foreign director boards (100%).
The measures of financial performance indicate that the sample firms were not financially
successful on average during the sample period, however some variation is noted. The mean Tobin’s q
was 0.97 and suggests the market value of firms were on average less than the book value of assets.
However, the variation in the sample is significant with a minimum Tobin’s q of 0.13 and a maximum of 4.49. The mean ROA of 0.2% with a minimum of -30.13% and a maximum of 31.62% show a similar high variation.
The gender diversity on the boards is reflected by an average proportion of 13.4% female board members, while the most diverse board has 50% female directors. The average number of years directors have been board members is 6.47 years, their average age 57.31 years and the mean board size is 6.78. The firms are on average 35.38 years old, with 8 centenary firms of over 100 years old and maxima of 271 and 272 years. The cultural distances of boards (CDBoard) show a minimum of 0 (all directors have the same nationality) and a maximum cultural diversity score of 2.21, as the measure reflects average board cultural diversity. It is noted however, that the largest individual cultural distance score was observed for a firm from the United Kingdom with a Greek foreign director and a cultural distance score of 4,248. The smallest individual cultural distance score was included multiple times. This cultural distance score of 0,081 is seen for firms from the United Kingdom with a director from the United States. This confirms many corporate governance investigations that classify the countries the United Kingdom and the United States as close in terms of culture and institutions, often classified as ‘Anglo-Saxon’.
2Table 1
Descriptive statistics
NationalityDiversity is the ratio of directors with a foreign nationality to total board size (i.e. he/she is from a different country than where the firm is located). CDBoard is calculated as the average of cultural distances between foreign director's home countries and the firm country per firm. BoardSize is the number of directors per board. Gender is the ratio of female directors to total board size per firm. Age and FirmAge reflect the average of board member’s age per firm and the firm's age respectively. Tenure offers the average number of years a firm's board members have served on the board. TobinsQ is calculated by dividing the firm’s market value of assets by the book value of its assets. ROA is the return on assets as calculated by annual net income divided by the book value of total assets. ForeignSales is the ratio of foreign sales to total sales, and TotalAssetsLn is the natural logarithm of the book value of year-end total assets.
Variable Mean Standard deviation Min Max Obs
Panel A: Board characteristics
NationalityDiversity CDBoard
.152 .178
.234 .378
0 0
1 2.21
200 200
BoardSize 6.78 3.29 1 19 200
Gender .134 .136 0 .5 200
Age 57.31 6.46 31 75 200
2
