Nationality diversity of the board of directors
and the level of earnings management
Master Thesis Sander van den Berg 10212949
Master Accountancy & Control, variant Accountancy
Final Version 20th of June 2015 17.000 words
First Supervisor: Dr. Réka Felleg
Amsterdam Business School
2
Statement of Originality
This document is written by student Sander van den Berg who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Table of Contents
page
1. Introduction 6
2. Literature and Hypothesis Development 9
2.1 Earnings Management 9
2.1.1 What is Earnings Management? 9
2.1.2 Types and Patterns of Earnings Management 10
2.1.3 Agency Theory and Reasons to Manage Earnings 11
2.2 Corporate Governance and the board 13
2.2.1 Corporate Governance 13
2.2.2 Board of Directors and Characteristics 14
2.2.3 Board of Directors and Earnings Management 15
2.3 Theories about Diversity and Nationality 17
2.3.1 Diversity 17
2.3.2 Culture, Nationalities and Diverse Groups 18
2.4 Hypotheses Development 21 3. Methodology 23 3.1 Sample 23 3.2 Methodology 24 4. Results 29 4.1 Descriptive Statistics 29
4.2 Pearson and Spearman Correlations 32
4.3 Earnings Management via Discretionary Accruals 34
4.4 Earnings Management via Small Profits 37
4.5 Sensitivity Tests 41
5. Discussion and Conclusion 49
4
Tables
page
Table 1 Sample Selection 24
Table 2 Distribution of Observations by SIC Industry 29
Table 3 Descriptive Statistics of Dependent and Independent Variables 30
Table 4 Correlation Matrix 32
Table 5 OLS Regressions of Discretionaryaccruals and 35 Differentnationalitiesproportion
Table 6 OLS Regressions of Discretionaryaccruals and 37 Differentnationalities
Table 7 Logit Regressions of Smallprofit and Differentnationalitiesproportion 39
Table 8 Logit Regressions of Smallprofit and Differentnationalities 40
Table 9 OLS Regression Discretionaryaccruals on 42 Differentnationalitiesproportion of observations with at least three
different nationalities
Table 10 OLS Regression Discretionaryaccruals on Differentnationalities 43 of observations with at least three different nationalities
Table 11 Logit Regression Smallprofit on Differentnationalitiesproportion 44 of observations with at least three different nationalities
Table 12 Logit Regression Smallprofit on Differentnationalities of observations 45 with at least three different nationalities
Table 13 Logit regression Justmeetorbeat on Differentnationalitiesproportion 47
5
Abstract
This study investigates whether firms with more nationality diverse boards of directors have lower levels of earnings management. Based on the human capital theory (Becker, 1964), social identity theory (Tajfel & Turner, 1986), social psychological theory (Westphal & Milton, 2000) and resource dependence theory (Pfeffer & Salancik, 1978), I expect that firms with more nationality diverse boards of directors have lower levels of earnings management. I use two measures to determine the level of earnings management and two measures to determine the nationality diversity of boards of directors. I use the Modified Jones Model by Kothari et al. (2005) and small profits (Leuz et al., 2003) to measure the level of earning management. To determine the nationality diversity of boards of directors, the first measure I use is the proportion of the amount of different nationalities in the board of directors divided by the number of directors in the board of directors. The second measure is the real amount of different nationalities in the board of directors. I use data from the period 2007-2014 and data from the United Kingdom. The results don’t support the hypothesis that firms with more nationality diverse boards of directors have lower levels of earnings management. This implicates that more nationality diverse boards of directors are not appropriate to lower levels of earnings management.
6
1. Introduction
This study researches whether firms that have boards of directors with more nationality diverse directors have lower levels of earnings management. Boards of directors are part of corporate governance mechanisms and are an important alleviation of the principal-agent problem (Jensen & Meckling, 1976). Corporate governance can be described as the collection of control mechanisms that an organization adopts to prevent or dissuade potentially self-interested managers from engaging in activities detrimental to the welfare of shareholders and stakeholders (Scott, 2014, p.8). Managers and people from inside an organization can have incentives to abuse their discretion and their information advantage to manage earnings, which is described as earnings management in the literature. However, earnings management can have (detrimental) negative effects for alignment between the interests of the agent and principal and for efficient contracting, because accounting information is used for these purposes. Besides, accounting information is used for economic decisions. Because of these uses of accounting information, earnings management is not feasible. Boards of directors have to safeguard the interests of the people and organizations outside the firm (Larcker & Tayan, 2011). Therefore boards of directors have to prevent earnings management.
Existing literature has researched the effects of boards of directors and specific characteristics of board of directors on the level of earnings management (Core, Holthausen & Larcker, 1999; Armstrong, Ittner & Larcker, 2012; Larcker & Tayan, 2012). These characteristics of boards of directors are also researched on other accounting or reporting quality measures like conservatism and timely loss recognition. The existing literature of boards of directors researches in particularly the effect of independent or outside directors, board size and tenure of directors and more recently also gender and age diversity. However, the conclusions of the existing literature about boards of directors and the level of earnings management are not (completely) unanimous and the literature doesn’t define a ‘perfect’ or ‘good’ board of directors.
A new part of the literature relates to gender diversity of boards of directors and to race and ethnicity diversity of boards of directors and researches the effects these diversities on the level of earnings management and firm performance (Carter, D’Souza, Simkins & Simpson, 2010; Rahman & Ali, 2006; Gopal, Krishnan & Parsons, 2008). Carter et al. (2010) don’t find a relationship between ethnic and gender diverse boards of directors and firm performance, while Rahman and Ali (2006) for Malaysian firms don’t conclude that more ethnic (race) diverse boards of directors diminish the level of earnings management.
7 However these findings are, especially for ethnicity diversity, in an early stage and are not studied extensively. Future studies can focus on this unexplored and unexplained part of the literature. Related to ethnicity diversity of boards of directors, is nationality diversity of boards of directors. People with a certain nationality have certain characteristics, skills or human capitals that differ from people with other nationalities. These differences in characteristics, skills and human capital between people with different nationalities can apply in a same way to ethnicity diverse groups. For as far as I know, there are no prior studies that investigate the relation between nationality diversity of boards of directors and the level of earnings management. An objective of this study is to investigate this unexplored part of the literature. Besides this objective, this study complements the (related) studies of Carter et al. (2010) and Rahman and Ali (2006), who studied the effect of race and ethnicity diversity on the level of earnings management.
Existing literature about boards of directors doesn’t have an unanimous definition of a ‘perfect’ or ‘good’ board of directors. It can be that diversity in nationality of boards of directors is of large influence on the quality of the board of directors and it may be that diversity of nationality is of such influence that boards of directors can be ‘perfect’ or ‘good’ based on the diversity of nationality, whereby it can lessen the importance of the characteristics of boards of directors that are already researched in the existing literature. This study can give insights into that issue. This study is furthermore of interest, because the economy and firms operate more and more internationally. Additionally, management/executive teams and boards of directors become more international. These circumstances make nationality diversity (more) relevant in the current and future (business) world.
Based on what is mentioned above, this study explores whether firms that have more nationality diverse boards of directors have lower levels of earnings management. To investigate this, I use a sample of data from public firms in the United Kingdom for the time period 2007-2014. I use two measures to determine the level of earnings management and two measures to determine the nationality diversity of boards of directors. I use the Modified Jones Model by Kothari et al. (2005) and small profits (Leuz et al., 2003) to measure the level of earning management. To determine the nationality diversity of boards of directors, the first measure that I use is the proportion of the amount of different nationalities in the board of directors divided by the amount of directors in the board of directors. The second measure is the real amount of different nationalities in the board of directors. The regression results of this study show no significant negative relation between the earnings management measures and the nationality diversity measures and therefore don’t support the hypothesis that expects that
8 firms with more nationality diverse boards of directors have lower levels of earnings management. This implicates that more nationality diverse boards of directors are not appropriate to lower levels of earnings management.
This study contributes to literature in several ways. The first contribution of this study is the exploration of a new research field in the existing literature. More specifically, this study fills an unexplored part of the literature that investigates the influence of nationality diversity of boards of directors on the level of earnings management. This new research area can give valuable insights. Another contribution is that this study contributes to the studies of Carter et al. (2010) and Rahman and Ali (2006) who research nationality diversity related influences of the board of directors (namely race/ethnicity diversity) on the level of earnings management. The results of this study can give credibility to or weaken the findings and insights of the studies of Carter et al. (2010) and Rahman and Ali (2006). A third contribution to the literature is the expansion of the existing literature about earnings management. This study gives insights to a new factor, nationality diversity of boards of directors, that can influence the level of earnings management. Another related contribution is the expansion of the corporate governance literature. This study demonstrates whether a mechanism of corporate governance, namely the board of directors, can be more effective via more nationality diverse boards of directors.
This study has also societal value. Firstly, this study is of value to policymakers and/or regulators. Policymakers and regulators both internationally and nationally can prefer to have some degree of conservatism, because this reduces the probability of lawsuits, improves efficient contracting and serves the stewardship role, while ‘extreme’ conservatism is not seen as feasible (Scott, 2014). Earnings management is however not feasible for conservatism, therefore policymakers and regulators would like to diminish earnings management. Based on the findings of this study regarding the level of earnings management and its relation with nationality diversity of boards of directors, policymakers and regulators can recommend or obligate more or less nationality diverse boards of directors. Besides, investors can use the findings of this study. Based on the findings and their tolerance for the level of earnings management, investors can decide whether they want to invest in firms with more or less nationality diverse boards of directors. Furthermore, firms themselves also have to consider how shareholders or stakeholders react to the composition of their boards of directors with regards to nationality diversity. In the same way, shareholders and stakeholders can use the findings of this study to consider their connections with the firms they relate to and they might propose a more or less nationality diverse board of directors.
9 This study contains four more sections. Section 2 contains the literature and hypothesis development. Section 3 describes the methodology of this study. Section 4 provides the results. Section 5 discusses and concludes.
2. Literature and Hypothesis Development
2.1 Earnings Management
2.1.1 What is Earnings Management?
The literature describes earnings management in several (overlapping) ways and often in broad definitions. Scott (2014) defines earnings management in the following way, “Earnings management is the choice by a manager of accounting policies, or real actions, affecting earnings so as to achieve some specific reported earnings objectives” (p. 445). Another (partly overlapping) definition that is often used in the earnings management literature is the
definition of Healy and Wahlen (1999):
Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers. (p. 368)
In addition to these definitions, Fischer and Rosenzweig (1995) describe earnings management in a more general way as “Actions of a manager which serve to increase (decrease) current reported earnings of the unit for which the manager is responsible without generating a corresponding increase (decrease) in the long-term economic profitability of the unit” (p. 434). These definitions of earnings management have overlapping parts, while they also are somewhat different. But they have the same main idea. Taking these definitions together, earnings management can be best described as a way in which managers use their own judgment, choices or intentions to change financial reports/financial outcomes or transactions, to misled (some) interested parties with regard to either the underlying economic performance of the firm or to influence the outcomes of contracts that depend on reported accounting numbers. In this paper earnings management is used in the broad sense of the definition, as it is outlined above.
10
2.1.2 Types and Patterns of Earnings Management
There are several ways in which earnings can be managed, but two main ways in which earnings management can be practiced are by discretionary accruals (accrual based earnings management) and by real economic decisions (real earnings management) (Leuz et al., 2003; Barth et al., 2008; Chi et al., 2011; Scott, 2014). Accruals based earnings management is about accounting policy choices and judgments (Scott, 2014, p. 445). This can be further divided into the choice of accounting policy per se, like straight-line versus declining-balance amortization, or policy choices for revenue recognition and via the discretionary accruals itself, like provisions for credit losses, inventory values, warranty costs and timing and amounts of low-persistence special items such as write-offs and provisions for restructuring. Real earnings management is by means of real variables (Scott, 2014, p. 446). Examples of real earnings management are changes in advertising expenses, maintenance, R&D, timing of purchases and disposals of capital assets, stuffing channels and overproduction. Real earnings management can be costly, because it affects the firm’s longer-term interests/performance.
A characteristic of accrual based earnings management is that accruals reverse (Scott, 2014). Therefore, if accruals are increased in period one, they will be lower in a next period, or vice versa. This is different compared to real earnings management. With real earnings management the increase (decrease) in for instance R&D expenses/activities will not (necessarily) lead to a subsequent decrease (increase) in a later period. In this way you can say that accrual based earnings management is more like a temporarily change ‘on paper’ with not (necessarily) an impact on economic value creation, while real earnings management has an real impact on the creation of economic value or firm performance.
There are several patterns of earnings management (Scott, 2014; Leuz et al., 2003; Burgstahler & Dichev, 1995; Healy, 1985). A first one is known as taking a bath or big bath accounting. This is used when a firm is in stress and/or has (big) losses and the firm has thereby little to lose. The firm can then decide to take more losses by for instance writing off assets or expensing more costs, because the firm is already operating poorly. Because accruals reverse, the current extra losses will reverse in the future. Big bath accounting is also used when there is a new manager, because by using this practice, the new manager can ‘downgrade’ the results of his/her predecessor and can make higher profits in the future (compared to the predecessor). A second pattern is income minimization. This is somewhat the same as taking a bath, but less extreme. A firm can choose this pattern of earnings management for instance during periods of high profitability of politically visible firms or to protect against (foreign) competitors. This pattern can also be used due to tax considerations. Examples of earnings management practices
11 for this pattern are rapid write-offs of capital assets and intangibles or to expend advertising or R&D (more). A third pattern is income maximization. This is used for instance when managers can earn bonuses, therefore from contracting perspectives. Income maximization is also used when firms are close to debt covenant violations and can prevent violation by increasing income. A fourth pattern is income smoothing. Income smoothing is used for instance when managers are risk-averse for their bonuses. When managers smooth earnings, they can receive more constant (bonus) compensation. Besides that, earnings smoothing can also be used for meeting covenants of long-term lending agreements, because if earnings are less volatile due to earnings smoothing, there is less chance of covenant violation. Earnings smoothing can also be used by managers that have a tendency to prevent low earnings or losses, because that can give rise to dismissal. Moreover, earnings smoothing can also be used for external reporting purposes. These different patterns of earning management can change over time for a firm.
Another distinction that is made in the literature is earnings discretion versus earnings smoothing (Leuz et al., 2003; Barth et al., 2008). This distinction is (partly) overlapping with and related to the patterns just described. Earnings discretion is earnings management due to the discretion that managers have in the accrual accounts and therefore can achieve a level of earnings that they wish to reach or to avoid (small) losses. This part of earnings management is about the possibilities that a firm has to achieve the level of earnings that it wants, because it has possibilities due to accruals. Earnings smoothing is the other type of earnings management and is associated with creating an earnings level or pattern that is not (too) volatile and therefore is more constant over time or even slightly increasing. Earnings smoothing is associated with creating constant earnings, while earnings discretion can also be associated with more volatile earnings, because earnings discretion can both increase and decrease earnings in a larger amount.
2.1.3 Agency Theory and Reasons to Manage Earnings
There are several reasons and motivations/incentives to use earnings management and a lot of them are described and/or researched in the literature. To understand these several reasons and motivations/incentives, the agency theory (principal-agent problem) is important to understand (Jensen & Meckling, 1976; Fama & Jensen, 1973). A principal-agent problem arises when an agent pursues its own goals rather than the goal of the principal. Adverse selection and moral hazard are important concepts for the agency theory. In a simplified business setting, there is a contract between a manager(s) (agent) and firm owner(s) (principle) and also between the firm (agent) and external lender(s) (principle). In both situations there is information asymmetry and
12 the agent and principle have contradicting interests in first instance whereby the agent has more information. Adverse selection is a type of information asymmetry whereby one or more parties to a business transaction, or potential transaction, have an information advantage over other parties (Scott, 2014, p. 23). This situation is before a transaction. Moral hazard is a type of information asymmetry whereby one or more parties to a contract can observe their actions in fulfillment of the contract but other parties cannot (Scott, 2014, p. 23). This situation is after the transaction. There will be a contract if the agent and principle want to do business with each other (remuneration contract, debt contract/covenant etcetera). Such a contract between agent and principle will align the interests of the agent and the principle and will make sure that the agent will use the right amount of effort for the principle. Such contracts usually make use of financial accounting information and reported earnings. However, accounting information and reported earnings are not an exact figure and need also judgment and/or accounting choices by the agent. Therefore a possibility to manipulate this accounting information through earnings management is possible. The agent can manipulate for his own gain or benefit, partly because he/she has an information advantage.
The agency problem gives rise to manage earnings. Reasons to manage earnings can apply simultaneously and can be related to each other. A first reasons to use earnings management are related to debt and equity pricing (Francis et al., 2004; Healy & Wahlen, 1999). Accounting information in general and accounting earnings are used as useful information for the pricing of debt and equity. When a firm makes contracts with external lenders, lenders provide more favorable debt conditions when there are (more) positive and/or constant accounting earnings. Therefore firms and managers have an incentive to manage earnings to meet those expectations of positive and constant earnings. When a firm meets those expectations the external lender faces less risk of not receiving money back and interest and therefore the lender can ask a lower interest rate. This reason is contract-based. Related to this reason is that a firm can use earnings management to avoid violation of a debt covenant. It can be less expensive for a firm to manage earnings than to violate a debt covenant. Violating a debt covenant has negative and expensive consequences. Additionally, if a firm plans to issue new equity, the firm will have a tendency to manage earnings upward, because then the firm can raise more money compared to when there is downward or no earnings management before an equity issuance. In a similar way, firms have an incentive to manage earnings upward before an IPO (initial public offering).
A second part of the reasons to manage earnings is related to analysts and the markets (Badertscher, 2011; Healy & Wahlen, 1999). Firms want to meet or beat the expectations of
13 analysts and/or markets. Badertscher (2011) explains that earnings and earnings growth are important for determining firm value. Firms (mostly) prefer to have a high and correct firm value. If analysts and/or markets expect that earnings will be of a certain level and a certain growth, a firm then wants to satisfy these expectations, because otherwise firm value will be lost. Meeting or beating these expectations can be reached by managing earnings.
Another part of the reasons for earnings management is related to bonuses of managers (Healy & Wahlen, 1999; Dechow, Sloan & Sweeney, 1995; Healy, 1985; Gaver, Gaver & Austin, 2000). Firms often use bonuses to align interests of managers to interests of firms, with the purpose that managers will work to create value for the firm. The bonuses are often linked to certain accounting profits, but because accounting earnings are manageable to a certain level, managers can use this possibility to manage earnings to the level of accounting profits that are needed to receive their bonuses.
Another reason that is mentioned in the literature is earnings smoothing, which can be considered as a goal of earnings management (Leuz et al., 2003; Barth et al., 2008; Burgstahler & Dichev, 1997; Dechow et al., 1995). When firms smooth earnings, in periods of high earnings firms’ earnings are lowered and during periods of low earnings firms’ earnings are heightened. This practice gives (more) stable earnings through time. This stability is usually preferred when firms want to create a stable or slightly increasing stock price. The stability can also create a stable growth and capital growth for the firm, which increases firm value. Earnings smoothing can also be used to get broadly used financial ratios, like earnings per share (EPS) or leverage ratios, on a constant or desired level (Berk & DeMarzo, 2011).
2.2 Corporate Governance and the Board 2.2.1 Corporate Governance
Corporate governance is a broad definition in the literature and is defined in different (overlapping) ways. Larcker and Tayan (2011) describe corporate governance as “The system of checks and balances meant to prevent abuse by executives” (p. 1). And in more detail as “The collection of control mechanisms that an organization adopts to prevent or dissuade potentially self-interested managers from engaging in activities detrimental to the welfare of shareholders and stakeholders” (p.8). The definition of Merchant and Van der Stede (2012) is more or less the same, “The sets of mechanisms and processes that help ensure that companies are directed and managed to create value for their owners while concurrently fulfilling responsibilities to other stakeholders (e.g. employees, suppliers, society at large)” (p. 553). Additionally to these definitions is the definition of Scott (2014); “Those policies that align the
14 firm’s activities with the interests of its investors and society” (p. 9). Taking these definitions together, corporate governance can be described as the systems, mechanisms and policies that align the interests of the people and organizations outside the firm with the interests of the people inside the firm, usually focused on preventing ‘bad behavior’ from inside managers, like for instance earnings management. The board of directors are representatives of the people and organizations outside the firm, and therefore have to represent the interests of the people and organizations outside the firm.
2.2.2 Board of Directors and Characteristics
The board of directors are important with regard to corporate governance, because the board plays a central role in the corporate governance system (Larcker & Tayan, 2011). For public listed companies in all countries, a board of directors is required, but the characteristics of the board (like terms of mandated structure, stakeholder representation, independence level and other compositional features) can differ between countries. However all the boards of directors have two fundamental responsibilities, namely to advise management and to oversee management’s activities (Larcker & Tayan, 2011). In this way the board of directors represents the interests of the shareholders, investors, debtholders or other people or organizations from outside the firm and should safeguard their interests. The Organization for Economic Cooperation and Development (OECD) describes its vision of the responsibilities of the board as follows (OECD, 2004); “The corporate governance framework should ensure the strategic guidance of the company, the effective monitoring of management by the board, and the board’s accountability to the company and the shareholders”. This vision expects both advise and oversight from the board of directors. To fulfill the advisory function, the board consults with management regarding the strategic and operational direction of the firm. To fulfill the oversight function, the board has to monitor management and has to ensure that management acts diligently in the interests of the shareholders (Larcker & Tayan, 2011).
Literature treats and investigates several variables that are characteristics of a board. Examples of these variables that characterize board of directors are the percentage of outside directors, the percentage of old board members, percentage of busy board members, whether there is an outside chairman, percentage of outside board members appointed by the CEO, whether the CEO is the board chair, the board size, percentage of inside directors, percentage of interlocked outside directors, the CEO percentage stock ownership, whether the non-CEO insider owns 5%, percentage stock ownership per outside director, whether an outside block holder owns 5% (Armstrong et al., 2012; Core et al., 1999; Klein, 2002).
15 An important and fundamental characteristic and distinction of boards is whether the board is an one-tier board or a two-tier board (Maassen, 2002). Usually Anglo-Saxon countries have adopted an one-tier board model (Maassen, 2002; Larcker & Tayan, 2011). In an one-tier board model, the executive directors and non-executive directors operate together in one organizational layer. Some boards of this type are dominated by executive directors, while other boards can be dominated by non-executive directors. It could be the case that the CEO is also chairman of the board. One-tier boards usually also make use of board committees like audit, remuneration or nomination committees (as do two-tier boards). The one-tier board structure is used for example in the US, UK and Canada. The two-tier board has an extra organizational layer, that has been designed to separate the executive function of the board from the monitoring function (Maassen, 2002; Larcker & Tayan, 2011). The supervisory board (the extra upper layer) has only non-executive directors and is responsible for appointing members to the management board, approving financial statements, making decisions regarding major capital investment, mergers and acquisitions and the payment of dividends. The management board (the lower layer) is usually composed of executive managing directors and is responsible for making decisions on such matters as strategy, product development, manufacturing, finance, marketing, distribution and supply chain. Usually the corporation laws or statues don’t allow the CEO to be the chairman of the supervisory board too. This makes the supervisory board more independent. The two-tier board is used for example in the Netherlands, Germany and Finland. Therefore the most explicit difference between the one- and two-tier boards is that in the one-tier board the functions of the board are more jointly, while the two-tier board separates the board functions more formally.
2.2.3 Board of Directors and Earnings Management
Research has been done in a lot of ways to determine the effect of the board of directors on the level of earnings management. The board of directors has to represent the people and organizations from outside the firm and has to safeguard their interests. These people and organizations outside the firm, make use of financial/accounting information and use this information to make decisions with regard to the firm and also use this information for contracts. This financial information can facilitate efficient contracting (Dechow, 1994). However, to make good decisions and to facilitate efficient contracting, the needed information has to be relevant and a faithful representation of the underlying economics (IASPlus, 2015). Earnings management diminishes these relevance and faithful representation principles. Therefore from the viewpoint of the people and organizations outside the firm, there should be
16 no earnings management, and therefore the board of directors has to prevent earnings management. Another reason why earnings management needs to be prevented by the board of directors is that otherwise managers or CEOs/CFOs can use earnings management for more favorable compensation. This will lower the benefits of efficient contracting and lower the (financial) benefits for the people and organizations outside the firm. Another reason is that through earnings management assets (and liabilities and equity) can be misstated, this can make contracting with for instance debt holders more difficult to conduct. Overall lower earnings management gives better financial transparency, which can benefit the people and organizations outside the firm. The board of directors has to take this interest into account.
As described above, the quality of a board, in the sense of how ‘good’ a board is, can be determined in several ways. Therefore no uniform standard exists of a good or strong board. However, most findings in this research field conclude that ‘better’ boards of directors prevent or diminish earnings management more than worse boards of directors do.
Xie, Davidson and DaDalt (2003) find that boards of directors that contain more independent outside directors and contain more financially and corporate experienced directors are less likely to be involved and are less involved with earnings management. Xie et al. (2003) also conclude that there is a lower level of earnings management when boards of directors and audit committees meet more frequently. Chtourou, Bédard and Courteau (2001) find that effective boards in general constrain earnings management. More specifically, there is less earnings management when there are more outside board members and when those members have more experience with the firm itself and also with other firms (Chtourou et al., 2001). Cornett, Marcus and Tehranian (2008) conclude the same as Xie et al. (2003) and Chtourou et al. (2001) with regard to independent outside directors and the level of earnings management and additionally find a negative relation between earnings management and institutional ownership of shares and institutional investor representation. They also find that these board characteristics are also effective to diminish earnings management when there are strong incentives for managers and or CEOs to use earnings management for more favorable compensation. Liu and Lu (2007) find that there is less earnings management when there are higher levels of corporate governance in general. They conclude that good corporate governance mitigates agency problems, especially agency conflicts between the largest shareholders and the minority shareholders. Larcker, Richardson and Tuna (2007) set up their own renewed corporate governance measures and search for relations between their measures and the level of earnings management. They have mixed findings about the relations between these measures and the level of earnings management. This can give reason to believe that there
17 is no relation between how good a board is and the level of earnings management. Cornett, McNutt and Tehranian (2009) conclude that the level of earnings, board independence and the amount of capital that a firm has, are negatively related to earnings management. They also conclude that more independent boards appear to constrain earnings management when there is a higher CEO pay-for-performance sensitivity (PPS). Therefore more independent boards can prevent earnings management even when there is a strong incentive by managers to use earnings management. Klein (2002) finds a negative relation between board independence and abnormal accruals and finds additionally that reductions in board (or audit committee) independence are accompanied by large increases in abnormal accruals. The most pronounced effects seem to occur when the board (or audit committee) is comprised of a minority of outside directors (Klein, 2002). Hence outside directors are important to diminish earnings management. Klein (2002) concludes suggests that boards that are more independent of the CEO are more effective in monitoring the corporate financial accounting process (Klein, 2002).
Although literature does not seem to have an uniform ‘good’ board and determining the quality of a board is difficult, it can be concluded that boards that have ‘better’ characteristics prevent or diminish earnings management more. Outside and independent directors seem to be of particular importance to prevent or diminish earnings management.
2.3 Theories about Diversity and Nationality 2.3.1 Diversity
Diversity can be seen as heterogeneity. Homogeneous, the opposite of heterogeneous, is defined as 1) “composed of parts or elements that are all of the same kind; not heterogeneous” or 2) “of the same kind or nature; essentially alike” (Dictionary, 2015a). Heterogeneous is defined as 1) “different in kind; unlike; incongruous” or 2) “composed of parts of different kinds; having widely dissimilar elements or constituents” (Dictionary, 2015b). Hence homogeneity has things that are of the same, while heterogeneity has things that are not the same. This same or different things can be everything, but in the context of boards of directors and board members aspects like age, education, experience, race, ethnicity etcetera are important and in this paper specifically the nationality of board members.
Both homogeneous and heterogeneous groups, and hence boards of directors, have their advantages and disadvantages. People that differ (more) from each other have more conflicts and less cohesion, because they are less socially integrated as a group (Brodbeck, 2011). This suggests that homogeneous groups can be better and can work more efficiently. Homogeneous groups also have more group-feeling, because homogeneity makes generalization from the
18 behavior of one group member to the group as a whole easier. More group-feeling also tends to provide that people pursue shared goals and work (Brodbeck, 2011). Group members also share more information and insights, experiment more collectively, seek and give more feedback and ask and support each other more (Brodbeck, 2011). Group members who behave inconsistently with other group members, have a negative influence on group dynamic (Brodbeck, 2011). Homogeneous groups also tend to have more in common, share the same opinion, communicate more frequently and trust each other more (Early and Mosakowski, 2000). Homogeneous groups tend also to be more cooperative and to have less conflicts (Williams & O’Reilly, 1998).
Heterogeneous groups and heterogeneous group members are more different from each other and these differences could potentially lead to more conflicting views. These differences could lead to ‘problems’, but differences can also enhance the individual group members’ creativity and learning. This can lead to more divergent perspectives and opinions, which can give a broader range of knowledge and better understanding of the context in which a group operates (Brodbeck, 2011). These differences can give heterogeneous groups more knowledge, judgement, creativity and perspectives (Francoeur et al., 2008). More specifically, differences in ethnicity or culture, give more creativity and flexibility in organizations (McLeod, Lobel and Cox Jr., 1996). Another advantage of diverse groups is that members who are different from the group characteristics, tend to ask more questions that would not be asked by members who have the group characteristics (Campbell and Mínguez-Vera, 2008). This can be the case for different nationalities in a board of directors.
Both groups have advantages and disadvantages. The biggest advantage of homogeneous groups is that they are more efficient, collaborative and are more one group. On the other hand heterogeneous groups are more creative, stimulate group-thinking and have more (diverse) knowledge and understanding.
2.3.2 Culture, Nationalities and Diverse Groups
Culture can be defined as “the collective programming of the mind distinguishing the members of one group or category of people from the others” (Hofstede, 1980). Hofstede (1980) states that people carry ‘mental programs’ with them, which are developed in the early youth within the family atmosphere and are strengthened during school and in organizations and these mental programs contain a component of the national culture. These mental programs and different components are most clearly expressed via the different social values that different people carry with them within different countries. Hofstede (1980, 1984, 2001, 2015) researched these values. Hofstede (1980, 1984, 2001, 2015) defines the following values; (1) (high vs low) power
19 distance, (2) individualism vs collectivism, (3) masculinity vs femininity, (4) (high vs low) uncertainty avoidance, (5) long term orientation vs short term normative orientation and (6) indulgence vs restraint. The GLOBE project also researched and defined cultural characteristics (House, Hanges, Javidan, Dorfman & Gupta, 2004). Both Hofstede and GLOBE give scores on these cultural characteristics to countries from all over the world, whereby the scores differ substantially between countries (Hofstede, 1980, 1984, 2001, 2015; House et al., 2004). These characteristics that differ per country, go ‘into’ the nationality of a country. Therefore, different nationalities and its characteristics that differ, can also go into the board of directors when there are boards that are diverse in nationality. These circumstances make boards of directors more diverse and can potentially make boards of directors more valuable and ‘better’, because there can be more diverse values. Board of directors members with a different nationality can also be seen as more outside and more independent members.
Several theories about diversity can apply to nationality diverse boards of directors. These theories originate from other fields than economics and/or accounting, namely social and psychological fields of research. These theories can possibly give insight on the effects of diversity in nationality of boards of directors and possibly on the level of earnings management.
The first theory that can be used is the human capital theory of Becker (1964). This theory states that each individual has a (different) capacity of education, experience and skills. Individuals’ capacities enhance individuals’ cognitive and productive capabilities which can benefit both the individual and the organization. It can be argued that persons with different backgrounds, which apply to persons with different nationalities, have different and unique human capitals. Therefore more diversified boards of directors, through different nationalities, can have a broader set of human capitals and can lead to better boards. Terjesen et al. (2009) and Carter et al. (2010) for instance research the influence of gender (women) and ethnicity diversity on boards and they argue that more diversified boards can have more human capital. But it is also possible that due to diversity, a lack of the ‘right’ human capital arise. A counter argument for this lack of ‘right’ human capital is that only ‘skilled’ diverse people will be hired for the board of directors (Dunn, 2012). It can be argued that these ideas in the literature also apply for nationality diverse boards of directors, because boards of directors consist, just as other groups, out of individuals that have specific capacities and therefore human capitals. Argued on unique human capitals, that every individual with a certain nationality has, nationality diverse boards positively affect the effectiveness of boards and that can make boards better (Carter et al., 2010).
20 Another theory that can be applicable is the social identity theory (Tajfel & Turner, 1986). This theory states that people try to have people around them with similarities like similar demographic profiles (nationalities), values and perspectives. With this (dis)similarities people see themselves as (out-) in-group members. It can be argued that the same (different) nationality is a requirement to be an in- (out-)group member. Usually in-group members receive better evaluations from their fellow in-group members and out-group members are more problematic with being in the group or identifying themselves with the group (Terjesen et al., 2009). It could be that when out-group nationalities exist and hence out-group members, that these out-group members are not part of the group, but this is difficult to predict. It may be that there is a threshold and that when this threshold is reached due to enough different nationalities, that the group of different nationalities is one group, with only group members instead of in-group and out-in-group members. Literature about gender diversity states that at least three women are needed to break through the threshold (Asch, 1951). It can be argued that this minimum of three can also be applicable for nationalities, whereby in a board of directors at least three persons with other nationalities are needed, to let the out-group nationality members become in-group members. However the threshold of exactly three people and in this paper three different nationalities is questionable, because the total amount of group members is probably also of influence. The amount of different nationalities to the amount of board members, might be a better indication to determine whether group members are in-group members.
A third theory is the social psychological theory (Westphal & Milton, 2000). This theory focuses on the influence that majority and minority groups or members have in group dynamics. This theory states that members who have a majority status, have the potential to exert (disproportionately) more influence in group decisions than members that have a minority status (Westphal & Milton, 2000). The minority status can apply to different or ‘outside’ nationalities. This can apply because this theory further suggests that demographic (national) differences between members lower social cohesion between groups and that the social barriers reduce the probability that minority viewpoints will influence group decisions (Westphal & Milton, 2000). Hence, it might be that diverse nationalities will have too low power. But it can also be argued that when there are a lot more different nationalities, that these nationalities become a ‘majority’ instead of a ‘minority’ and therefore have power. Westphal & Milton (2000) also state that minority group members may encourage divergent thinking in decision-making processes, while Campbell & Mínguez-Vera (2008) state that more diversity among members generates more diverse opinions and critical thinking that makes decision-making more time-consuming and less effective, but possibly more considered.
21 Another theory that can apply is the resource dependence theory (Pfeffer & Salancik, 1978). This theory considers the links and connections that various people have and in this case about the links to external parties for the company that the board of directors has. Pfeffer & Salancik (1978) suggest 4 benefits of these links or connections; (1) provision of resources such as information and expertise; (2) creation of channels of communication with constituents of importance to the firm; (3) provision of commitments of support from important organizations or groups in the external environment; and (4) creation of legitimacy for the firm in the external environment. Different members of the board of directors could provide different beneficial resources and information to the firm due to different work experience and background (Hillman, Cannella & Paetzold, 2000). Boards of directors with different nationalities and hence more diverse boards of directors will provide more valuable resources and links, which can make boards of directors better and could give better firm performances (Hillman et al., 2000). The different nationalities and their links can signal positive signs to the labor market and product market and can therefore attract talent, that otherwise could pass by. In this way it can also be argued that different nationalities give valuable resources, links and information, which makes the board better and more valuable.
2.4 Hypothesis Development
The theories of diversity just mentioned explain advantages and disadvantages of groups that have diverse group members. The theories can also apply to boards of directors, because boards of directors have group members who can differ (substantially) from each other. The directors in the board of directors differ from each other if they have different nationalities, because Hofstede (1980, 1984, 2001, 2015) and GLOBE (House et al., 2004) state that countries, and hence nationalities, have different scores for social values and have therefore different characteristics per country and per nationality. The human capital theory of Becker (1964) suggests that different individuals have their own (different) capabilities and human capital. This human capital will be ‘richer’ in total when groups have more diverse group members, this applies for boards of directors that are more nationality diverse. The social identity theory of Tajfel and Turner (1986) suggests, in the context of boards of directors, that when ‘enough’ outside nationality group members are a member of the board of directors, the board of directors becomes an good integrated board of directors (all in-group members). A nationality diverse board of directors therefore don’t have to be an obstacle for an integrated board of directors and could possibly even have extra (as the human capital theory suggests) knowledge, capabilities, capitals etcetera, which a non-nationality diverse board of directors doesn’t have. In more or
22 less the same way can be argued that the social psychological theory of Westphal and Milton (2000) suggests that directors of the board of directors who have a nationality that belongs to the minority, can become majority if enough diverse nationalities are on the board of directors. These nationality diverse board of directors members have potentially extra capabilities, knowledge, experience etcetera. The resource dependence theory of Pfeffer and Salancik (1978) suggests furthermore that more different, and hence more nationality diverse boards of directors, have valuable links and connections.
These four diversity theories together suggest that nationality diverse boards of directors have valuable attributes that less nationality diverse boards of directors don’t have. Furthermore heterogeneous groups, and hence more nationality diverse groups, tend to be more conflicting, but give more diverse ideas and suggestions, and give as well more understanding and knowledge. Heterogeneous groups also tend to ask more questions and think in a more considered way. These attributes are valuable for boards of directors for both their advisory role as well as their oversight role. The empirical research about boards of directors that is described earlier in this paper, suggests that independent and more outside directors are characteristics of a ‘better’ board of directors. Directors that have different nationalities than other directors are probably seen as more outside or independent. And these boards of directors that have these better characteristics seem to prevent or diminish earnings management more (Klein, 2002; Armstrong et al., 2012; Larcker & Tayan, 2012). Overall more nationality diverse boards of directors seem to be ‘better’.
People and organizations outside the firm (principles) use accounting information to make economic decisions and use accounting information for efficient contracting with firms. Besides, accounting information is also used for compensation and to value assets. Earnings management is not feasible for these purposes. Managers or people inside the firm (agents) can abuse their discretion in accounting for (own) benefits, and hence possibly manage earnings. Boards of directors have to prevent the abuse managers and people inside the firm, because they have to safeguard the interests of the people and organizations outside the firm. Based on the foregoing the hypothesis is as follows:
Hypothesis: Firms that have boards of directors with more nationality diverse directors have lower levels of earnings management.
23
3. Methodology
3.1 Sample
I use data from two databases to conduct this study. The first database is the ‘Bureau van Dijk Amadeus Managers’ database and the second database is the ‘Compustat Global Fundamentals Annuals’. The Bureau van Dijk database is an European private and public firms database that contains both financial and non-financial data and contains the necessary data to determine the nationality diversity of boards of directors, while the Compustat database contains the necessary financial data to determine the level of earnings management. I use only public firms in my sample, because earnings management incentives for private firms differs from public firms (Burgstahler, Hail & Leuz, 2006; Coppens & Peek, 2005; Beatty, Ke & Petroni, 2002).
I use data of firms listed in the United Kingdom. I use this country, because the Bureau van Dijk database has much data about firms that are listed in the United Kingdom. Furthermore the United Kingdom has an one-tier board structure, which is used widely in the world. Moreover, the United Kingdom is a developed and sophisticated country, whereby stock exchanges and the financial markets in the United Kingdom are seen worldwide as proper markets. Furthermore the United Kingdom can be seen as an international market, which will probably have more nationality diverse employees. These facts give me reason to use data from the United Kingdom and the results of this study can probably be generalized because of the (financial market related) characteristics of the United Kingdom.
I would have preferred to use data from firms in the United States of America, because a lot of accounting and economic research is based on U.S. data and this would make generalization and applicability of findings of this study better. However, the available databases don’t have the data of directors with regard to the nationalities of board of directors for U.S. firms. Because of this, I decided to use European data, because countries in Europe are also researched in accounting and economics. I decided to choose U.K. data for the reasons just mentioned above. The above mentioned circumstances in the U.K. apply also (in a similar way) to the U.S., therefore this study could be also applicable for the U.S.
The earnings management measures that I use in this study are often used in U.S.-based research and are of U.S. origin. Despite, I use these earnings management measures on U.K. data. As just described above, the U.S. and the U.K. are in many respects (almost) similar to each other. This makes applicability of the earnings management measures in the U.K. possible. Literature that investigates the level of earnings management in Europe also uses U.S. origin
24 earnings management measures (e.g. Burgstahler, Hail & Leuz, 2006; Maijoor & Vanstraelen, 2006; Van Tendeloo & Vanstraelen, 2008).
I use the period 2007 – 2014, which is an 8 year time frame. I use this timeframe to have enough data and data variability in the measures of nationality diversity of boards of directors and to have a recent time frame that represent recent findings. I also use this time frame to have boards of directors that are more complete.
Table 1 shows the sample selection procedure and the final sample contains 4.482 firm-year observations. There are 13.426 observations when the data from the two databases are merged. Hereafter I drop observations of financial firms, because they have different accruals. Thereafter observations are dropped due to missing data to measure the level of earnings management and for the control variables.
Table 1 Sample Selection
Firm-year observations Merge of Bureau van Dijk Amadeus Managers and Compustat
Global Fundamentals Annuals
13.426
Financial firms (SIC 6021-6669) (3.812)
9.614 Missing data to calculate discretionary accruals and small profits
and losses (EM measures) (3.802)
5.812
Missing data to calculate control variables (1.330)
Final sample 4.482
3.2 Measures of nationality diversity of boards of directors and earnings management
The independent variable
The independent variable in this study is the nationality diversity of boards of directors. I use data from the Bureau van Dijk Amadeus Managers database to determine nationality diversity. From this database I retrieve data about the function of managers, the nationality of managers and when managers started their function. I keep data about managers who have a function as director and therefore remove data about managers who are not directors. I therefore can determine the board of directors for a certain firm.
For every firm for every year, I determine the number of directors of the board of directors and the amount of different nationalities in the board of directors. I then divide the
25 amount of different nationalities in the board of directors by the number of directors in the board of directors to get a proportion of different nationalities in the board of directors. I use this proportion as a measure for nationality diversity of the board of directors. A higher value for this proportion means a higher amount of different nationalities relative to the amount of directors in the board of directors and represents more nationality diversity. Besides this proportional measure of nationality diversity of boards of directors, I use the real amount of different nationalities that are present in the board of directors as a measure for the nationality diversity. I therefore have two measures for the nationality diversity of boards of directors: (1) the proportional measure (Differentnationalitiesproportion) and (2) the real amount of different nationalities (Differentnationalities).
The dependent variable
The dependent variable of this study is the level of earnings management. Literature uses different models to determine the level of earnings management. Examples of well-known models are the Healy Model (Healy, 1985), DeAngelo Model (DeAngelo, 1986), Jones Model (Jones, 1991), Industry Model (Dechow & Sloan, 1991) and Modified Jones Model (Dechow, Sloan & Sweeney, 1995). For this study, I use the Modified Jones Model, because this model is an often used and an established model in the literature to determine the level of earnings management. I use a more detailed version of the Modified Jones Model, the version of the Modified Jones Model that also incorporates return on assets (Kothari, Leone & Wasley, 2005; Cahan, Zhang & Veenman, 2011). Kothari et al. (2005) argue and conclude that adding return on assets to the Modified Jones Model declines the chance on misspecification, because it corrects for performance.
The earnings management model that I use in this study is as follows: TACC i, t TA(i, t − 1)= β0 + β1 1 TA(i, t − 1)+ β2 (∆SALES(i, t) − ∆REC(i, t)) TA(i, t − 1) + β3 PPE(i, t) TA(i, t − 1) + β4 ROA(i, t) + ε(i, t)
TACC is the total amount of accruals of a firm. TA is the total amount of assets of a firm. ∆SALES is the change in sales revenue compared to sales revenue of the previous year. ∆REC is the change in receivables compared to receivables of the previous year. PPE is the gross amount of plant, property and equipment. ROA is the return on assets. Total accruals is the difference between income before extraordinary items and net cash flow from operating activities. This difference is due to discretionary accruals and non-discretionary accruals. To determine the level of earnings management, via ε, the model first estimates the amount of
non-26 discretionary accruals. This is estimated via the sales revenue, the receivables, the PPE and the ROA of a firm. These items give rise to non-discretionary accruals and these items give a representation of the firm’s characteristics and operating conditions. These items are scaled by total assets. To determine the expected amount of non-discretionary accruals for a certain firm in a certain year, the model sorts firms by SICs. If possible the model uses 3-digit SICs, thereafter (if no 3-digit SIC is possible) 2-digit SICs and thereafter 1-digit SICs (if no 2-digit SIC is possible), whereby at least 20 firms per SIC group are necessary to have reliable estimates of (non-discretionary) accruals. The level of earnings management (via the ε) for a firm in a certain year is determined by the difference between the total accruals and the estimated (non-discretionary) accruals. In this study the absolute value of ε is used to determine the level of earnings management, because this study doesn’t need to have a distinction between income-decreasing and income-increasing accruals. Therefore both income-decreasing and income-increasing earnings management are considered to be earnings management for this study.
The other measure to determine the level of earnings management, is the presence of small profits. Firms have incentives to use accounting discretion to avoid losses and to avoid reporting losses and prefer to report profits (Burghstahler & Dichev, 1997; Degeorge, Patel & Zeckhauser, 1999; Leuz, Nanda & Wysocki, 2003). If firms have large losses, it is probably impossible to report (large) profits. However, it might be possible to report small profits within the discretion that firms have with respect to accounting. Therefore earnings management can additionally be determined via small profits (Leuz et al., 2003). To determine whether firms manage earnings via small profits, I use a dummy variable for small profits. To determine whether firms have small profits, I use net income and divide net income by total assets to control for the size of the firm (Leuz et al., 2003). The dummy for small profits has the value of 1 when net income divided by total assets is in the range of 0% to 1%.
I have therefore two measures to determine the level of earnings management, (1) the Modified Jones Model of Kothari et al. (2005) via discretionary accruals earnings management (Discretionaryaccruals) and (2) earnings management via small profits (Smallprofit) (Leuz et al., 2003).
Control variables
I use several control variables in this study. The first control variable that I use is leverage, which I define as total debt divided by total assets. Firms with higher levels of leverage may have a greater incentives to manage earnings to prevent themselves from not meeting debt
27 covenants, because missing for instance requirements of debt covenants can be very costly for firms (Becker, DeFond, Jiambalvo & Subramanyam, 1998; Press & Weintrop, 1990; DeFond & Jiambalvo, 1994). A second control variable that I use is the size of the firm, measured by the ln of total assets. Firms that are larger and have more assets, have in general more possibilities to manage earnings via non-discretionary and discretionary accruals which may or may not be included in the Modified Jones Model, however these firms are also more scrutinized by regulators, which makes earnings management more difficult (Xie et al., 2003; Rahman & Ali, 2006). Another control variable that I use is sales growth, measured as the percentage of annual sales growth. The accruals of firms are probably also affected when firms have sales growth. Sales growth can also change the expectations of market participants of future performance. Accruals may be used to meet these expectations (Ahmed, Billings, Morton & Stanford-Harris, 2002). A fourth control variable is the size of the board, measured by the number of directors in the board of directors. It can be possible that boards of directors that are bigger in size, have less possibilities to manage earnings, because literature describes that bigger boards are ‘better’ and that it is more difficult for bigger boards to permit earnings management collectively (Core et al., 1999; Armstrong et al., 2012). The last control variable that I use is whether a firm has a loss. Firms that have a negative net income have greater incentives to use earnings management (discretionary accruals) to prevent a loss (Burgstahler & Dichev, 1997; Leuz et al., 2003). I use a dummy variable for this control variable which equals 1 if the firm has a loss and 0 otherwise.
Other control variables that are important in the context of this study are characteristics of the board of directors. Examples of board of directors characteristics are mentioned before in the literature section of this study, but existing literature found that firms with boards of directors that are more independent or outside and that have CEO separation from the board of directors are of particular importance for ‘better’ boards and lower levels of earnings management (Klein, 2002; Armstrong et al., 2012; Rahman & Ali, 2006). Other board of directors characteristics can also influence the level of earnings management. However, these board of directors characteristics are not available in the databases and therefore I can’t control for these board characteristics, although they can influence the level of earnings management in firms.
28 Empirical model
The model that I use in this study is as follows;
Level of EM = α0 + α1 nationality diversity of board of directors + α2 leverage +
α3 firm size + α4 sales growth + α5 board size + α6 loss
The measures for the level of earnings management (EM) are the Modified Jones Model (Discretionaryaccruals) (Kothari et al., 2005) and the small profit dummy (Smallprofit) (Leuz et al., 2003). The measures for the nationality diversity of the board of directors are (1) the proportion of different nationalities in the board of directors to the number of directors in the board of directors (Differentnationalitiesproportion) and (2) the real amount of different nationalities in the board of directors (Differentnationalities).
To research the relations between the variables I use OLS regressions when I use Discretionaryaccruals and use logit regressions when I use Smallprofit. I checked the distribution of the variables and the distribution of the regression residuals and the distributions look normally distributed. The values of the VIFs are all lower than 2, so there is no multicollinearity. I winsorize variables at the 1st and 99th percentiles where possible and necessary and I use robust tests in the regressions and control for clusters by gvkey.
The hypothesis expects that firms with boards of directors that are more nationality diverse have lower levels of earnings management and which implicates that more nationality diverse boards are appropriate to lower levels of earnings management. To support this hypothesis, a negative coefficient (α1) for the association between the earnings management
measure and the nationality diversity is needed. Therefore, I expect that the coefficient for the relation between the level of earnings management and nationality diversity will be negative. This negative coefficient supports the hypothesis. With respect to the control variables, I expect that leverage, losses, sales growth and firm size increase the level of earning management and therefore will have a positive coefficient, while I expect that board size decreases the level of earnings management and will therefore have a negative coefficient.
29
4. Results
4.1 Descriptive Statistics
Table 2 shows the distribution of the observations per industry according to the SIC industry coding. The manufacturing industry represents the largest percentage of observations in the sample, namely 37.64% of the observations. Service firms represent the second largest percentage of observations, namely 28.11%, followed by mining firms, which represent 10.75% of the observations. The representation of the other industries ranges from 2.63% to 9.77%, while the agriculture, forestry, fishing and public administration industries don’t have observations in the sample.
Table 2 Distribution of Observations by SIC Industry
Industry SIC Observations %
Agriculture, Forestry, Fishing 0100-0999 0 0.00
Mining 1000-1519 482 10.75
Construction 1520-1999 118 2.63
Manufacturing 2000-4010 1687 37.64
Transportation & Public Utilities 4011-4999 438 9.77
Wholesale Trade 5000-5199 167 3.73
Retail Trade 5200-6020 330 7.36
Services 7000-9720 1260 28.11
Public Administration 9721-9999 0 0.00
Total Observations 4482 100.00
Table 3 presents the descriptive statistics of both the dependent and the independent variables. These descriptive statistics give preliminary insights to explore the hypothesis of this study.
