Board Chairman Characteristics and Real Earnings Management

Abstract

This study investigates the influence of board chairman characteristics on the level of real earnings management for listed firms with the lowest positive earnings on the Main Market of Bursa Malaysia. Based on the Ordinary Least Square regression, the findings indicate that board chairman independence and real earnings management have a significant positive association. However, BC’s age, on the other hand, was found to be strongly connected with a lesser degree of real earnings management. Other board chairman characteristics, including tenure, ethnicity, and family membership, did not have a significant influence on the level of real earnings management. In general, the findings are robust and compatible with numerous assumptions, such as incorporating the year dummy variable and eliminating the accruals earnings management control variable. These findings highlight the inconsistent effect of each characteristic of the board chairman. Furthermore, it seems that the board chairman’s characteristics examined in the study are not efficient, except for the board chairman’s age, in reducing the real earnings management where results may be different if the board chairman is a female director. The use of comprehensive characteristics of the board chairman together in one model in this study is novel. However, it can inform policy-makers, firms’ owners, stakeholders, as well as scholars, of the need for improving the board chairman’s role in protecting the firm from real earnings activities, where it has been observed that 97% of the boards of the firms’ are chaired by male directors.

1. Introduction

Last two decades have seen corporate scandals that have led to the demise of large businesses such as Enron and WorldCom. Some Malaysian corporations, such as firms in many other nations, have participated in corporate misbehavior and wrongdoing, including Transmile and Megan Media [1]. Among the reasons for the global series of financial scandals and fraud are the managers’ failures in corporate management. The ultimate goal of any company is to increase the wealth of its shareholders [2]. When managers fail to fulfil their obligations, they sometimes manipulate earnings. Managers of these corporate attempt to show that businesses have been in a good situation for several prior years, but suddenly they become unable to continue managing earnings and, hence, firms go into bankruptcy.

In the discipline of accounting, managing earnings is still a hot topic [3,4,5]. Earnings quality and earnings management (EM) have been the focus of much research, which is steadily growing [1,6] and should be researched more, especially in Malaysia [4,7,8,9,10]. Corporate scandals and the issue of EM increased awareness of the importance of paying close attention to a company’s financial statements among regulators, investors, and the financial community [11].

The agency hypothesis contends that corporate governance (CG) tools such as the board of directors (BOD), ownership, and audit committee (AC), among others, may lessen agency issues where managers may look for their interests rather than the shareholders. CG’s mechanisms are charged with keeping an eye on how businesses operate and have the power to influence how businesses make decisions [12]. Hence, implementing CG’s mechanisms is anticipated to reduce EM [13,14] and improve reporting quality [15]. Most countries have issued a code of CG and continue to develop the code. The goal of issuing the code of CG, globally or in Malaysia, is to safeguard stakeholders by ensuring that managers behave in a way that is consistent with the stakeholders’ interests and that the information they offer accurately represents a firm’s economic performance [16,17,18].

The impact of CG on EM has been studied in the past. On the best type of leadership structure, there is disagreement in the literature [19]. The findings, however, are not always reliable as argued by [20,21], and there are a number of CG and EM-related issues that may be looked at [22], especially in developing countries such Malaysia. There is a need to strengthen CG because the EM prevention procedures in Malaysian enterprises are insufficient [23]. Furthermore, previous studies have not led to clear conclusions about the BOD’s effectiveness in reducing EM as the results of these studies are controversial [21].

Surprisingly, most recently published research studies on CG and EM have mainly focused on developed countries, while the emerging markets have received less attention. As a result, there has been a pressing need for further research on CG in developing countries with distinctive economic environments, such as Malaysia. Importantly, prior studies have not adequately addressed the issue of Board chairman (BC)’s characteristics; so, researching the influence of the BC’s characteristics, e.g., independence, tenure, age, ethnicity, and family membership, on the level of real earnings activities (REM), specifically in Malaysia, fills the gap in the literature. Therefore, more research into the BC’s role in the CG process is necessary [24].

The chairman is given more authority to create effective governance, leadership, and board procedures. Most regulatory bodies demand that companies appoint a non-executive chair or divide the roles of chairman and chief executive officer (CEO). The features of the chairman, however, are unknown [11]. The BC, who is also more likely to be accountable for any failures in financial reporting, has the biggest responsibility for monitoring the firm’s performance.

Some BCs have held their positions for extended stretches of time and are older than others. Therefore, it has become crucial to determine whether a longer BC’s tenure and age are useful in reducing EM, particularly in nations with concentrated ownership, such as Malaysia. Furthermore, although not previously addressed by studies, the personality attributes of the BC (such as ethnicity and gender) may have an impact on his or her attitude toward EM.

Malaysia is of interest for this study because there are certain aspects of the Malaysian business climate that can hinder the effective application of the CG framework. Shareholdings, which have expanded from the conventional family-owned business, are a major concern for Malaysian companies [25,26]. As a result, in 2007, 2012, and 2017, regulators made changes to the Malaysian code of corporate governance (MCCG) to improve the governance standards. This suggests that additional improvements to the CG mechanisms in Malaysian businesses are required to prevent EM.

There are some specific ways in which this study contributes to the literature. It aims to address a gap in the research by conducting an empirical investigation of how the BC’s characteristics affect REM. It advances the body of literature by using resource dependence theory as a supportive theory and agency theory as the underlying theory to clarify the link among the characteristics of the BC and EM. The study is different from other studies as the current study uses comprehensive characteristics of the BC in one model.

The findings may assist regulators and policymakers in comprehending the supervisory functions of each BC’s characteristic in monitoring the EM. This might aid regulatory agencies and policymakers in rethinking about the function of BC’s features. Additionally, this research can assist decision-makers in their efforts to enhance how well the BC’s qualities contribute to a company’s financial reporting quality, namely EM. It also encourages CG scholars to investigate the qualities of the BC in more detail to form a firm conclusion that could be useful to all shareholders, investors, and governments.

2. Literature Review and Hypothesis Development

Managers, similar to agents, won’t always operate in the principals’ best interests (shareholders). Isolating the principles from overseeing daily operations and giving management this responsibility may encourage managers to act opportunistically for their personal gain. Due to this circumstance, principals search for mechanisms to better regulate the agent’s decisions and shield them from the agent’s potential impact on the principal’s interests. The use of CG variables in EM-related investigations is theoretically supported by the agency theory [27].

Besides agency theory, the resource dependence theory suggests that the BOD must demonstrate their skill and understanding in internal decision-making, as they have working experience in other organizations, to accomplish more than just reduce uncertainty. As a result, the BOD contributes to different perspectives on internal issues and gives executives the necessary knowledge [28].

Due to differing opinions on the role that CG’s mechanisms play in regulating managers’ actions, regulations are increasingly focusing on a particular orientation for the CG, which typically calls for more external directors with a variety of skills and perspectives. They also often call for increased diversity in terms of gender. Gender reforms are successfully raising the proportion of female directors [29]. Significantly less wrongdoing occurs when there are more women in leadership positions [30]. The results of ref. [31] found that both accruals-based and real earnings management practices are negatively associated with board gender diversity. Furthermore, in countries where gender equality is high, earnings management decreases [32]. Female directors with appropriate financial experience increase earnings quality more than female directors who do not have such expertise [33]. The findings of ref. [34] revealed that being a male director between the ages of 56 and 64 is a major predictor of committing fraud. This indicates that more tenure builds sufficient experience.

As a result, board leadership is a multifaceted concept that encompasses a number of issues that improve strategic decision-making, e.g., the procedures that ensure board members provide the board with relevant knowledge and skills, the manner in which interactions are structured, and the chairman’s leadership abilities [35].

2.1. Board Chairman’s Independence

The major concern of global CG codes is board independence. Many studies in the area of CG examine the composition and effectiveness of boards and contend that a board’s capacity to carry out its crucial duties is influenced by the degree of independence that the board enjoys [19]. The study of ref. [36] claims that the susceptibility of CEO turnover to company performance is significantly reduced when the CEO and chairman jobs are held by the same individual. These results provide credence to the idea that it is challenging for the board to fire poor performance in businesses where the CEO and chairman jobs are merged because there is a lack of independent leadership. According to ref. [35], the link between board leadership and strategy participation is altered by CEO authority. The study found that when the CEO has a lot of influence, the board usually becomes submissive and follows the CEO’s orders.

Independence became a more critical issue in the subject of the BC. Scholars have claimed that a BC who neither serves as founder nor as CEO can enhance the board’s oversight function by thoroughly and impartially investigating management choices, particularly those made by the CEO [11]. The effectiveness of the BC’s leadership has a favorable and considerable impact on board strategy involvement [35]. It has been found that companies with independent BCs have better quality earnings statements than those without independent chairperson [37].

Chairman’s independence has a significant negative impact on the discretionary accruals practices [38,39]. This means that firm with an independent chairman has low discretionary accruals. However, the study of [23] found a statistically significant positive association between chairman independence and discretionary accruals. Meanwhile, Habbash [40] found no connection between EM and the independence of the chairperson. Nonetheless, the analysis anticipates, based on the agency and resource dependency theories, that:

H1. 

The board chairman’s independence negatively affects the real earnings management.

2.2. Board Chairman’s Tenure

In Malaysia and around the world, the majority of authorities decide how long independent directors can serve. For example, MCCG 2012 added a new condition capping their tenure at a total of nine years. In addition, MCCG 2017 asks for shareholders’ consent via a two-tier voting process in the event that an independent director is to be retained after 12 years.

There is conflicting empirical information about a board’s tenure at this time [6,41]. A board with a longer tenure is one that is more effective at carrying out its duties [41], boosting its knowledge and expertise in observing management choices [42]. This supports the idea of stewardship, which states that long-term directors outperform those who hold their positions for a shorter duration of time in terms of performance since the former is more familiar with the company’s operations [43].

The second group of research, on the other hand, demonstrates that longer board’s tenure is detrimental to business success [41], as they may become personal acquaintances, and hence less likely to regulate and control managers, especially in CEO-controlled organizations [44]. This view supports the agency theory where the hypothesis is that long-term directors may become less objective in identifying the company’s interests and may wind up supporting management objectives rather than those of the shareholders.

For board governance and independence, the BC’s tenure is essential, especially in Malaysia, in which the MCCG has made remarks about the dominance of BCs. According to ref. [45], a chairman’s tenure has a considerable impact on lowering discretionary accruals activities and REM practices. Similarly, the study of [46] documented a significant negative influence of the tenure of the BC on discretionary accruals. Likewise, the study of ref. [39] revealed that the BC’s tenure has a significant influence on the two directions of discretionary accruals, income-increasing and income-decreasing earnings management. However, other research has revealed that the term of the BC is unrelated to firm performance [47] and to the level of EM [48]. Nevertheless, the study anticipates that, based on the agency and resource dependency hypotheses.

H2. 

The board chairman’s tenure positively affects the real earnings management.

2.3. Board Chairman’s Age

One of the major aspects that may influence business governance is the age of directors. Age diversity increases the likelihood of CEO dismissal following regulatory sanctions [49]. Based on the resource dependence theory, directors with higher age expect to have more experience and hence participate effectively with company strategies and decisions. According to ref. [50], elder chairmen are more likely than younger chairmen to undertake diversification strategies. According to the report, senior chairmen are risk-averse and include their life cycle into their organisations’ corporate diversification strategy. Similarly, the study of ref. [51] found a substantial positive relationship between the age of the BC and business performance, likely as a result of the older BC’s greater experience and consequently greater risk aversion. Xiong [45] found that the age of the BC is strongly associated with a low level of EM, proposing that the quality of financial reporting might be improved by having an elder chairperson.

Some other views suggested that directors of higher age may contribute less than the younger ones. Section 129(1) of the Malaysian Companies Act 1965 says that no person beyond the age of seventy may be nominated as a director on the board of a public company. However, the maximum age for a director has been removed by the new Companies Act 2016 (often known as “the 2016 Act”). This means that anyone who is a natural person and at least 18 years old is qualified for director position. According to studies, e.g., ref. [47], age has a detrimental impact on how well somebody performs at work, and an older BC may indicate worse corporate performance. Similarly, the study of ref. [6] found no significant impact of AC chairman’s age on discretionary accruals.

Nevertheless, the study follows resource dependence theory and expects that an older chairman could reduce the level of real earnings management.

H3. 

The board chairman’s age negatively affects the real earnings management.

2.4. Board Chairman’s Ethnicity

Cultural variables have a huge influence on people because they shape their behavior [52]. Hence, it has the potential to dominate director behavior, corporate processes, and corporate morals [52]. For example, the presence of foreign directors increases the likelihood of CEO dismissal following regulatory sanctions [49]. As a result, a BC who is a member of a particular group or ethnicity might have a considerable impact on the firm’s actions and values. In general, Marimuthu [53] determined that expanding the ethnic diversity of the board would improve the company’s financial performance. Similarly, the study of ref. [51] revealed a positive relationship among firm performance and the ethnicity of the BC or CEO. This means that the performance of the company was found to be higher when the chairman was Malay.

Information is more likely to be willingly provided by a board with a high proportion of Malay directors [54]. According to ref. [6], the Malay chair of AC is found to be a significant association with low levels of EM. However, research by ref. [55] found no connection between ethnicity and audit quality, neither for the percentage of Malay directors nor for the Malay chairman. According to several other previous studies, the ethnicity of the board does not significantly affect the quality of financial reporting as determined by EM, e.g., the study by refs. [23,25].

However, in accordance with the agency theory as well as the resource dependence theory, the analysis anticipates that a Malay chairman might bring value to the company and mitigate the real earnings activities. As a result, the study anticipates:

H4. 

The Malay board chairman negatively affects the real earnings management.

2.5. Board Chairman’s Family Membership

A company’s BC holds a key position. The BC position was previously assumed by the CEO, who was typically a founder of a company or a family member, in order to be in the greatest possible position to safeguard the family’s assets and control the decisions made by the entire organization. This situation happened mostly in family-controlled firms. According to ref. [56], family-controlled businesses may think about selecting a family member for the role of BC as the CEO, which would still have a big impact on that person’s opinions.

Scholars have claimed that family shareholders, particularly in developing nations, may exert total managerial control and strive to advance their interests at the expense of minority shareholders [18,57]. According to ref. [58], in small businesses, the family members dominate and abuse the power which means that independent directors are not really autonomous to oversee and supervise the company’s operations. As a result, most regulatory bodies demand that corporations separate the BC and CEO roles. The study of ref. [6] claim that EM practices in a company where a family member serves as chairman are not considerably decreased by CG frameworks. Family businesses could have other goals instead than profit maximizing [59].

Some previous studies fail to get a significant influence of family chairman on firm profitability e.g., ref. [60] or firm performance e.g., refs. [59,61]. This result is in line with agency theory II, where it suggests the existence of conflicts of interest between majority and minority shareholders. As a result, there is a need for an independent chairman to balance the interest of the majority and minority shareholders. According to ref. [6], it is important to employ a BC who does not have a relation to directors or significant shareholders since these connections could limit the ability of CG to rein in EM behaviors. Considering this, the last hypothesis of this study is as follows:

H5. 

The board chairman’s family membership positively affects the real earnings management.

3. Research Design

3.1. Sample Selection

Following earlier research, e.g., refs. [62,63], that chose firms with return on assets (ROA) between 0 and 0.005 to detect EM, this study selects firms with the lowest positive earnings by using the average ROA of listed firms on the Main Market of Bursa Malaysia from 2013 to 2015, following the previous studies [6,64,65,66]. The study excludes firms related to financial services, close-end funds, special purpose acquisition companies (SPAC), and real estate investment trust (REIT), as well as firms with incomplete data on ROA. Furthermore, due to incomplete data and insufficient calculated REM, 12 firms were removed from the sample. As a result, the final number of firms covered by the study is 282 for the three years (846 firms’ observations).

3.2. Measurement of Real Earnings Management

The study uses the three proxies introduced by ref. [63] to calculate REM, the abnormal levels of cash flow from operations (ABCFO), the abnormal levels of production costs (ABPROD) and the abnormal levels of discretionary expenses (ABDISX) by using the following equation to calculate the coefficients for the three proxies, respectively.

$$\frac{{\mathrm{CFO}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}={\beta }_1\left(\frac{1}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_2 \left(\frac{{\mathrm{S}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_3\left(\frac{\Delta {\mathrm{S}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\mathsf{\epsilon }}_{\mathrm{it},}$$

(1)

$$\frac{{\mathrm{PROD}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}={\beta }_1\left(\frac{1}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_2 \left(\frac{{\mathrm{S}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_3 \left(\frac{\Delta {\mathrm{S}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_4\left(\frac{\Delta {\mathrm{S}}_{\mathrm{it}-1}}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(2)

$$\frac{{\mathrm{DISEXP}}_{\mathrm{it}}}{{\mathrm{A}}_{\mathrm{it}-1}}={\beta }_1\left(\frac{1}{{\mathrm{A}}_{\mathrm{it}-1}}\right)+{\beta }_2 \left({\frac{ {\mathrm{S}}_{\mathrm{t}-1}}{\mathrm{A}}}_{\mathrm{t}-1}\right)+{\mathsf{\epsilon }}_{\mathrm{t}}$$

(3)

where:

${\mathrm{CFO}}_{\mathrm{it}}$ is cash flow from operations

${\mathrm{A}}_{\mathrm{it}-1}$ is total assets of previous year

${\mathrm{S}}_{\mathrm{t}}$ is revenues

$\Delta {\mathrm{S}}_{\mathrm{it}}$ is the change in revenues.

${\mathrm{PROD}}_{\mathrm{t}}$ reflects the cost of goods sold (COGS) plus the change in inventory ($\Delta \mathrm{INV}$)

$\Delta {\mathrm{S}}_{\mathrm{it}-1}$ is the one-year lag of $\Delta {\mathrm{S}}_{\mathrm{it}}$.

${\mathrm{DISEXP}}_{\mathrm{it}}$ is the sum of R&D, advertising, selling, general and administrative costs.

Ordinary Least Squares (OLS) is used to estimate the coefficients of ${\beta }_1,{\beta }_2,{\beta }_3,{\beta }_4{\mathrm{and}\mathsf{\epsilon }}_{\mathrm{it}}$ for each equation. To calculate ABCFO, ABPROD and ABDISEXP, the study used the tool available in STATA software presented by [67], after running Equations (1)–(3). Furthermore, research has shown that companies that manipulate earnings tend to have low ABCFO, low ABDISEXP, and/or high ABPROD in case of increased earnings or vice versa in case of decrease-earning manipulation. To ensure consistency between variables, the values of ABCFO and ABDISEXP are multiplied by −1 [68,69,70]. The values of ABCFO, ABDISEXP, and ABPROD were then combined using the following equation to show the whole value of the abnormal real earnings management (ABREM).

$$\mathrm{ABREM}=\left(\mathrm{ABCFO}\ast -1\right)+\mathrm{ABPROD}+\left(\mathrm{ABDISEXP}\ast -1\right)$$

(4)

Based on recent ABREM studies [3,71,72], the study uses the absolute value of ABREM to reflect the value of REM, meaning that the value sign of ABREM was ignored.

3.3. Regression Models

To ascertain the extent to which the characteristics of the BC (independence, tenure, age, ethnicity, female director, and family membership) had an impact on REM, the following regression model was used.

Table 1. Variables Definitions and Measurements.

Acronym	Measurement
REM	Three proxies introduced by [63]
BCIND	“1” if BC is independent, and “0”, otherwise.
BCTEN	Number of years that the BC has been a director of the company.
BCAGE	BC’s age.
BCETH	“1” if BC is Bumiputra (Malay) director, and “0”, otherwise.
BCFM	“1” if BC has a family relationship with another director or major shareholder, and “0”, otherwise.
BSIZE	Number of board directors.
BMEET	Number of board meetings per year.
ACSIZE	Number of AC directors.
ACMEET	Number of AC meetings per year.
OC	Percentage of outstanding shares held by largest five shareholders.
Big4	“1” if firms were audited by Big4 firms, and “0”, otherwise.
FSIZE	Natural log of total assets
LEV	Total debt to total assets.
ROA	Net income/total assets.
NCFO	“1” if cash flow from operations is negative, and “0”, otherwise
INDUS	“1” for manufacturing industry observation, and “0”, otherwise
AEM_K	Absolute value of discretionary accruals using MJM by [73].

provides details about these variables used in the regression Model.

REM = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ BCIND + ${\mathsf{\beta }}_2$BCTEN + ${\mathsf{\beta }}_3$BCAGE + ${\mathsf{\beta }}_4$BCETH + ${\mathsf{\beta }}_5$BCFM + ${\mathsf{\beta }}_6$BSIZE + ${\mathsf{\beta }}_7$BMEET + ${\mathsf{\beta }}_8$ACSIZE + ${\mathsf{\beta }}_9$ACMEET + ${\mathsf{\beta }}_{10}$OC + ${\mathsf{\beta }}_{11}$Big4 + ${\mathsf{\beta }}_{12}$FSIZE + ${\mathsf{\beta }}_{13}$LEV + ${\mathsf{\beta }}_{14}$ROA + ${\mathsf{\beta }}_{15}$ NCFO + ${\mathsf{\beta }}_{16}$INDUS + ${\mathsf{\beta }}_{17}$ AEM_K +$\mathsf{\epsilon }$. $\left(\mathrm{Model}1\right)$

The study incorporated several groups of control variables to reduce the likelihood of endogeneity and error in determining the relationship between BC’s characteristics and REM which is needed and used by previous studies in EM. Some of these variables are related to the BOD’ characteristics, e.g., board size and meetings. Other are related to AC’s characteristics, e.g., size and meetings. Other variables of CG’s mechanisms, e.g., ownership concentration, and biggest four audit firms also used. Furthermore, variables related to firm-specific characteristics, e.g., leverage, firm size, ROA, and cash flow from operation, are also used in the regression to control the relationship.

The important contribution of this model is that other techniques of EM, namely accrual earnings management of the Modified Jones Model (MJM) by ref. [73] (hereafter named AEM_K), were added as a control variable in the regression to control the relationship between BC’s characteristics and REM. It is confirmed that managers could choose a single EM tool to gain their desired earnings reporting (substitute hypothesis) or may use both (complement hypothesis).

Importantly, as there are different measurements of the control variables of accrual earnings management, the study re-estimates Model 1 by using other measurements of the accrual earnings management where the study uses Modified Jones Model (MJM) by ref. [74] (hereafter named AEM_D) in Model 2 and ref. [75]’s Model (hereafter named AEM_J) in Model 3, as listed below:

REM = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ BCIND + ${\mathsf{\beta }}_2$BCTEN + ${\mathsf{\beta }}_3$BCAGE + ${\mathsf{\beta }}_4$BCETH + ${\mathsf{\beta }}_5$BCFM + ${\mathsf{\beta }}_6$BSIZE + ${\mathsf{\beta }}_7$BMEET + ${\mathsf{\beta }}_8$ACSIZE + ${\mathsf{\beta }}_9$ACMEET + ${\mathsf{\beta }}_{10}$OC + ${\mathsf{\beta }}_{11}$Big4 + ${\mathsf{\beta }}_{12}$FSIZE + ${\mathsf{\beta }}_{13}$LEV + ${\mathsf{\beta }}_{14}$ROA + ${\mathsf{\beta }}_{15}$ NCFO + ${\mathsf{\beta }}_{16}$INDUS + ${\mathsf{\beta }}_{17}$ AEM_D +$\mathsf{\epsilon }$. $\left(\mathrm{Model}2\right)$

REM = ${\mathsf{\beta }}_0$ + ${\mathsf{\beta }}_1$ BCIND + ${\mathsf{\beta }}_2$BCTEN + ${\mathsf{\beta }}_3$BCAGE + ${\mathsf{\beta }}_4$BCETH + ${\mathsf{\beta }}_5$BCFM + ${\mathsf{\beta }}_6$BSIZE + ${\mathsf{\beta }}_7$BMEET + ${\mathsf{\beta }}_8$ACSIZE + ${\mathsf{\beta }}_9$ACMEET + ${\mathsf{\beta }}_{10}$OC + ${\mathsf{\beta }}_{11}$Big4 + ${\mathsf{\beta }}_{12}$FSIZE + ${\mathsf{\beta }}_{13}$LEV + ${\mathsf{\beta }}_{14}$ROA + ${\mathsf{\beta }}_{15}$ NCFO + ${\mathsf{\beta }}_{16}$INDUS + ${\mathsf{\beta }}_{17}$ AEM_J +$\mathsf{\epsilon }$. $\left(\mathrm{Model}3\right)$.

4. Results and Discussion

4.1. Descriptive Statistics

Table 2. Descriptive Statistics of Variables.

Variable	Obs.	Mean	Min.	Max.
REM	864	0.129593	0.0016986	0.7042756
BCTEN	864	12.410	0.330	45.330
BCAGE	864	64.297	30.000	88.000
BSIZE	864	7.418	4.000	17.000
BMEET	864	5.459	3.000	10.000
ACSIZE	864	3.244	3.000	6.000
ACMEET	864	5.039	3.000	10.000
LFSIZE	864	13.485	10.098	18.579
LEV	864	20.775	0.000	68.560
ROA	864	4.412	0.010	15.160
Conc5	864	0.546	0.141	0.948
AEM_K	864	0.0361019	0.000034	0.2421655
B. Dummy Variables	Yes (1)		No (0)
	Freq.	Percent	Freq.	Percent
BCIND	326	37.73	538	62.27
BCETH	446	51.62	418	48.38
BCFM	335	38.77	529	61.23
Big4	459	53.13	405	46.88
NCFO	198	22.92	666	77.08
INDUS	366	42.36	498	57.64

Definitions of the acronym were presented in Table 1. Numbers of REM, BMEET and ACMEET presented after winsorizing extreme observations.

exhibitions the descriptive statistics value of variables used in the study. The mean value of REM is 12.96% with a minimum value of 0.17% and a maximum value of 70.43%. Regarding the BC’s characteristics, Table 2 shows that 326 of firms’ observation (37.73%) have an independent BC while 538 of firms’ observation (62.27) have not appointed an independence chairman. This indicates that the majority of firms’ observation are less likely to appoint an independence chairman. In terms of BC’s tenure, the average is 12.41 years, according to Table 2, which is less than the findings of ref. [47] in Switzerland, ref. [48] in Taiwan and ref. [43] in Saudi Arabia, which were 14 years, 14.71 years and 8.014 years, respectively. Table 2 reveals that the BC’s age on average is 64.28 years, which is in line with the findings of ref. [51], who found that the majority of Malaysian BCs are between the ages of 50 and 71 years.

Regarding BC’s ethnicity, results show that 446 firms ‘observation (51.62%) appointed a Malay chairman while 418 firms’ observation appointed non-Malay chairman. Concerning the BC’s family membership, Table 2 shows that 335 firms’ observation (38.77%) appointed a chairman for who has a family relationship with another director or major shareholder while 529 firms’ observation do not.

The results indicate that the mean value of board size is 7.418 members, and its meeting is 5.459 per year. Regarding the AC’s size and meeting, the mean values is 3.244 members in the AC and 5.039 times of meeting annually for AC’s members, respectively. Furthermore, the result reveals that the top five shareholders possess 54.60% of the company’s ownership. In terms of biggest audit firms, 459 firm-year observations (53.13%) were audited by the biggest audit firms.

Regarding the firm-specific features, the findings show that the average value of leverage, ROA, and total assets (Natural log) are 20.76%, 4.41% and 13.46, respectively. The findings reveal that cash flow from operations for 198 firm-year observations (22.92%) was a negative value. Furthermore, 366 (42.36%) firm-year observations belonged to the manufacturing business. Finally, the results show that the mean value of AEM-K is 3.61% with a minimum value of 0.003% and a maximum value of 24.22%.

4.2. Diagnostic Tests

The study performed many diagnostic tests prior to running the regression of the empirical models. For example, the outlier test found issues with some variables, namely REM, BMEET, ACMEET. Hence, the top and bottom extreme observations of these variables were winsorizing by using 1% for REM and ACIZE while using 5% for BMEET to resolve the outlier problem. Furthermore, the study used Skewness and Kurtosis, as descriptive numerical methods to test the normality of each variable.

Table 3. Skewness and Kurtosis of Variables.

Variable	Skewness	Kurtosis
REM	1.972	8.058
BCIND	0.506	1.256
BCAGE	−0.407	3.290
BCETH	−0.065	1.004
BCTEN	1.156	4.258
BCFM	0.461	1.212
BSIZE	0.985	4.840
BMEET	1.455	4.661
ACSIZE	2.173	8.194
ACMEET	1.838	7.793
LFSIZE	0.796	3.497
LEV	0.422	2.475
ROA	0.657	3.574
Big4	−0.125	1.016
Conc5	−0.083	2.318
NCFO	1.289	2.661
AEM_K	1.763	7.013

Definitions of the acronym were presented in Table 1.

shows that individual variables dataset has no severe violation of the normality assumption as the Skewness and Kurtosis are not exceeding the threshold of ±3 and ±10 [76], respectively.

Moreover, the correlation matrix test is used to examine for collinearity or multicollinearity problems. As a result, there are no issues with multicollinearity or collinearity in the dataset, according to

Table 4. Pearson Correlation.

	REM	BCIND	BCTEN	BCAGE	BCETH	BCFM	BSIZE	BMEET	ACSIZE	ACMEET	Conc5	Big4	LFSIZE	LEV	ROA	NCFO	AEM_K
REM	1
BCIND	0.0285	1
	0.4031
BCTEN	−0.0334	−0.1954 ***	1
	0.3274	0.0000
BCAGE	−0.0798 **	0.1759 ***	0.3047 ***	1
	0.0190	0.0000	0.0000
BCETH	−0.0748 **	0.4382 ***	−0.3043 ***	0.1128 ***	1
	0.0279	0.0000	0.0000	0.0009
BCFM	0.0427	−0.5950 ***	0.3868 ***	−0.0462	−0.6509 ***	1
	0.2103	0.0000	0.0000	0.1753	0.0000
BSIZE	−0.0850 **	−0.0516	0.0622 *	0.0928 ***	0.1040 ***	0.0100	1
	0.0124	0.1298	0.0675	0.0063	0.0022	0.7692
BMEET	−0.0436	−0.0588 *	−0.2479 ***	−0.0350	0.2768 ***	−0.2572 ***	0.2002 ***	1
	0.2002	0.0840	0.0000	0.3038	0.0000	0.0000	0.0000
ACSIZE	−0.0893 ***	0.0304	−0.0450	0.0809 **	0.2173 ***	−0.1554 ***	0.3058 ***	0.2714 ***	1
	0.0087	0.3722	0.1865	0.0173	0.0000	0.0000	0.0000	0.0000
ACMEET	−0.0628 *	−0.1352 ***	−0.1116 ***	−0.0362	0.1304 ***	−0.1373 ***	0.1424 ***	0.5860 ***	0.1364 ***	1
	0.0649	0.0001	0.0010	0.2879	0.0001	0.0001	0.0000	0.0000	0.0001
Conc5	0.0745 **	−0.1125 ***	−0.1244 ***	−0.0350	0.0036	−0.0482	0.0481	0.1445 ***	0.0855 **	0.0711 **	1
	0.0286	0.0009	0.0002	0.3036	0.9167	0.1573	0.1576	0.0000	0.0120	0.0366
Big4	−0.0450	−0.0535	0.0544	−0.0132	0.0885 ***	−0.0284	0.1218 ***	0.1552 ***	0.2030 ***	0.1093 ***	0.0933 ***	1
	0.1860	0.1159	0.1103	0.6976	0.0093	0.4042	0.0003	0.0000	0.0000	0.0013	0.0061
LFSIZE	−0.2004 ***	−0.0313	0.1121 ***	0.1296 ***	0.1980 ***	−0.1421 ***	0.3634 ***	0.3486 ***	0.2725 ***	0.2986 ***	0.0941 ***	0.4677 ***	1
	0.0000	0.3578	0.0010	0.0001	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0056	0.0000
LEV	−0.0456	−0.0248	−0.0765 **	0.1243 ***	0.1140 ***	−0.0064	0.1407 ***	0.0796 **	−0.0089	0.1283 ***	−0.0807 **	0.1255 ***	0.3433 ***	1
	0.1810	0.4668	0.0245	0.0003	0.0008	0.8513	0.0000	0.0193	0.7947	0.0002	0.0177	0.0002	0.0000
ROA	0.0551	−0.0298	0.0059	−0.0128	−0.0336	0.0009	0.0815 **	0.0041	−0.0022	−0.0492	0.0023	0.0211	0.0447	−0.0958 ***	1
	0.1057	0.3821	0.8616	0.7073	0.3244	0.9788	0.0166	0.9042	0.9477	0.1489	0.9460	0.5359	0.1894	0.0048
NCFO	0.1606 ***	0.0755 **	−0.1733 ***	−0.0579 *	0.0925 ***	−0.0722 **	−0.0545	0.0409	−0.0349	−0.0115	−0.0819 **	−0.1224 ***	−0.1118 ***	0.1244 ***	1	1
	0.0000	0.0264	0.0000	0.0890	0.0065	0.0339	0.1096	0.2292	0.3058	0.7352	0.0161	0.0003	0.0010	0.0002	0.0154
AEM_K	0.2082 ***	0.0342	−0.1925 ***	−0.0472	0.0624 *	−0.0761 **	−0.0848 **	0.0773 **	−0.0273	−0.0251	−0.0406	−0.0766 **	−0.1165 ***	0.0576 *	0.0514	0.2451 ***	1
	0.0000	0.3147	0.0000	0.1659	0.0666	0.0253	0.0127	0.0231	0.4228	0.4612	0.2334	0.0244	0.0006	0.0906	0.1311	0.0000

Definitions of the acronym were presented in Table 1. *, **, and *** are significant at 10%, 5%, and 1% levels, respectively.

of the correlation matrix. The Breusch-Pagan/Cook-Weisberg test provides proof of heteroscedasticity problems. Additionally, Wooldridge’s test demonstrates that the autocorrelation issue is not present in the study’s data. As a result, the regression with a robust option should be utilized in the study to address the heteroscedasticity issue.

4.3. Regression Results

The study uses the OLS regression to investigate the relationship between BC’s characteristics and REM as presented in

Table 5. OLS Regression Results.

Variables	Model 1	Model 2	Model 3
BCIND	0.0266 **	0.0256 **	0.0238 **
	(0.0113)	(0.0112)	(0.0111)
BCTEN	0.000861	0.000894	0.000959
	(0.000597)	(0.000589)	(0.000584)
BCAGE	−0.000848 *	−0.000771	−0.000840 *
	(0.000475)	(0.000472)	(0.000475)
BCETH	−0.0118	−0.0134	−0.0132
	(0.0111)	(0.0111)	(0.0112)
BCFM	0.0117	0.00826	0.00700
	(0.0130)	(0.0128)	(0.0128)
BSIZE	7.15 × 10−5	9.87 × 10−5	0.000125
	(0.00251)	(0.00243)	(0.00242)
BMEET	0.00322	0.00398	0.00354
	(0.00375)	(0.00371)	(0.00370)
ACSIZE	−0.00849	−0.00709	−0.00638
	(0.00755)	(0.00783)	(0.00781)
ACMEET	0.000331	−0.00122	−0.000812
	(0.00429)	(0.00424)	(0.00427)
Conc5	0.0929 ***	0.0937 ***	0.0944 ***
	(0.0287)	(0.0285)	(0.0287)
Big4	0.0174 *	0.0177 *	0.0170 *
	(0.00962)	(0.00949)	(0.00948)
LFSIZE	−0.0146 ***	−0.0143 ***	−0.0149 ***
	(0.00349)	(0.00346)	(0.00348)
LEV	0.000147	0.000135	0.000181
	(0.000271)	(0.000267)	(0.000268)
ROA	0.00310	0.00257	0.00283
	(0.00190)	(0.00190)	(0.00188)
NCFO	0.0372 ***	0.0286 ***	0.0286 ***
	(0.0102)	(0.00986)	(0.00997)
AEM_K	0.597 ***
	(0.150)
AEM_D		0.630 ***
		(0.123)
AEM_J			0.607 ***
			(0.121)
Constant	0.255 ***	0.244 ***	0.254 ***
	(0.0528)	(0.0525)	(0.0529)
INDUS	Included	Included	Included
F-value	5.69	6.06	6.29
Sig.	0.0000	0.0000	0.0000
Observations	864	864	864
R-squared	0.122	0.140	0.139

Definitions of the acronym were presented in Table 1. Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

. All the Models shown in Table 5 fit at the percentage of 1%. Furthermore, all Models have an accepted R² (higher than 12%). It shows that selected variables are connected to and have an impact on the REM. Model 1 is considered as the main Model where AEM_K, besides other variables related to CG and firm-specific characteristics, added as control variables in the regression to investigate the relationship between BC’s characteristics and REM. Furthermore, to ensure vigorous results of the relationship, the study uses different measurements of the control variables of accrual earnings management. The study re-estimate Model 1 by using AEM_D in Model 2 and AEM_J in Model 3.

All Models in Table 5 show that there is a significant positive relationship between BC’s independence and the level of REM. The results reflect that BC who is independent are associated with a higher level of REM. This finding suggests that an independent chairman still has challenges safeguarding the interests of all parties, as they could still be overpowered by insider members of directors who possess a disproportionate number of shares and have other power over the makeup and decisions of the board. The results are inconsistent with the theories of agency and resource dependence.

Furthermore, the findings of the present study are incongruent with those of other studies, which found a strong relationship between the BC’s independence and a decline in discretionary accruals, e.g., refs. [38,39]. It also contradicts earlier research that concluded that the BC’s independence has no effect on the firm’s performance, e.g., ref. [77] and that it has no impact on the level of discretionary accruals, e.g., ref. [40] or REM, e.g., ref. [46].

Regarding BC’s age, it was found to be significantly associated with a lower level of REM. This finding indicates that a BC of higher age is more likely to protect the interest of shareholders by mitigating earnings manipulation. The findings imply that the chairman’s lengthy life makes him more effective at managing managers and less likely to take risks. This result is congruent with those of refs. [45,51], who indicated that BC’s age positively influences the firm’s performance and increases the financial reporting quality by reducing the practice of EM, respectively.

In terms of BC’s tenure, the study found that BC’s tenure has no discernible relationship with mitigating REM. The finding conflicts with results from previous studies. For instance, the study of ref. [78,79] found that the duration of the independent directors considerably reduces the possibility of misleading financial reporting and significantly mitigates EM for the company, respectively. Furthermore, it is inconsistent with ref. [45], who found that the chairman’s tenure has a considerable impact on lowering discretionary accruals and managing real earnings through operational activities. Furthermore, it is also inconsistent with ref. [46] who documented a significant negative influence of the tenure of the BC on discretionary accruals.

According to Table 5, there is no discernible relationship between a BC’s ethnicity (Malay chairman) and REM. Malay chairmen appear to struggle with using their power or perhaps do not participate in decision-making enough. The outcome is in line with research that found no association between the board’s ethnic composition or the representative of the Malay chairman and the quality auditing, e.g., ref. [55]; or between the board’s ethnicity and discretionary accruals, e.g., refs. [23,25].

In terms of BC’s family membership, it is found to have no relationship with the activities of REM. The finding implies that rather than choosing a chairman based on their experience and knowledge, company and/or shareholder, particularly controlled shareholders, choose a chairman who is in family relationship with either other director or major and controlled shareholders. As a result, the family chairman seems to have difficulties in mitigating the activities of REM. This result is consistent with other previous studies that provide evidence that a family chairman has no influence on a firm’s performance and profitability [60,61].

5. Robustness Test

5.1. Including Year Dummy Variable

This study re-estimated Models by using a dummy variable to represent the years’ effects [80,81,82]. It is has been documented by the previous studies that the business cycle could affect the results, e.g., [83]. The findings seen in

Table 6. OLS Regression Results Including Years Dummy Variable.

Variables	Model 1	Model 2	Model 3
BCIND	0.0267 **	0.0256 **	0.0238 **
	(0.0113)	(0.0112)	(0.0111)
BCTEN	0.000853	0.000885	0.000950
	(0.000597)	(0.000588)	(0.000584)
BCAGE	−0.000855 *	−0.000778 *	−0.000849 *
	(0.000474)	(0.000471)	(0.000474)
BCETH	−0.0117	−0.0132	−0.0130
	(0.0112)	(0.0111)	(0.0112)
BCFM	0.0119	0.00847	0.00722
	(0.0131)	(0.0128)	(0.0128)
BSIZE	0.000103	0.000131	0.000160
	(0.00252)	(0.00244)	(0.00243)
BMEET	0.00324	0.00401	0.00356
	(0.00374)	(0.00370)	(0.00369)
ACSIZE	−0.00856	−0.00716	−0.00644
	(0.00757)	(0.00785)	(0.00783)
ACMEET	0.000236	−0.00133	−0.000928
	(0.00428)	(0.00423)	(0.00426)
Conc5	0.0929 ***	0.0937 ***	0.0945 ***
	(0.0287)	(0.0284)	(0.0287)
Big4	0.0176 *	0.0178 *	0.0171 *
	(0.00958)	(0.00945)	(0.00944)
LFSIZE	−0.0147 ***	−0.0144 ***	−0.0151 ***
	(0.00351)	(0.00347)	(0.00349)
LEV	0.000152	0.000140	0.000188
	(0.000272)	(0.000268)	(0.000269)
ROA	0.00317	0.00265	0.00293
	(0.00194)	(0.00193)	(0.00190)
NCFO	0.0366 ***	0.0280 ***	0.0280 ***
	(0.0103)	(0.00990)	(0.0100)
AEM_K	0.603 ***
	(0.149)
AEM_D		0.634 ***
		(0.122)
AEM_J			0.612 ***
			(0.121)
Constant	0.257 ***	0.245 ***	0.255 ***
	(0.0531)	(0.0528)	(0.0532)
Industry dummy	Included	Included	Included
Years dummy	Included	Included	Included
F-value	5.20	5.55	5.75
Sig.	0.0000	0.0000	0.0000
Observations	864	864	864
R-squared	0.123	0.142	0.140

Definitions of the acronym were presented in Table 1. Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

are similar to the prior released in Table 5. This means that the results are robust and can generalized.

5.2. Excluding the Control Variables of AEM Proxies

This study re-estimated Models by excluding the control variables of AEM proxies. The findings seen in

Table 7. OLS Regression Results by Excluding the Control Variables of AEM Proxies.

Variables	Model 1
BCIND	0.0251 **
	(0.0113)
BCTEN	0.000568
	(0.000597)
BCAGE	−0.000817 *
	(0.000485)
BCETH	−0.0116
	(0.0116)
BCFM	0.0106
	(0.0134)
BSIZE	−0.000545
	(0.00252)
BMEET	0.00494
	(0.00379)
ACSIZE	−0.00860
	(0.00773)
ACMEET	−0.00124
	(0.00431)
Conc5	0.0878 ***
	(0.0291)
Big4	0.0171 *
	(0.00968)
LFSIZE	−0.0160 ***
	(0.00356)
LEV	0.000248
	(0.000274)
ROA	0.00381 **
	(0.00192)
NCFO	0.0467 ***
	(0.0109)
Constant	0.298 ***
	(0.0534)
Industry dummy	Included
Years dummy	Included
F-value	4.94
Sig.	0.0000
Observations	864
R-squared	0.098

Definitions of the acronym were presented in Table 1. Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

are similar to the prior released in Table 5. This means that the results are robust and can be generalized.

6. Conclusions

The separation of the chairman and CEO roles has received the majority of attention at this point, while other characteristics of BC have been given less attention. This discrepancy could have a significant impact on the board’s ability to safeguard the company from EM practices, particularly in Malaysia, where the Finance Committee on Corporate Governance (FCCG) commented in the lead-up to the MCCG on the influence of strong BCs who seek to control board composition.

This study examines the role of BC’s characteristics on mitigating REM by firms listed on the Malaysian Main Market with the lowest earnings for the period of 2013 to 2015. Results show that there is a significant positive relationship between BC’s independence and the activities of REM. However, BC’s age was found to be significantly associated with lower level of REM. Other characteristics, e.g., BC’s tenure, BC’s ethnicity and BC’s family membership did not play a significant role on mitigating REM. In general, the findings are robust and consistent with various assumptions, such as including the year dummy variable and excluding the control variable of accruals earnings management.

These results have significant implications for Malaysian CG policy by identifying the characteristics of the BC that are related with a lower or higher level of activities related to REM. However, the current study fails to investigate the role of BC’s female representative in mitigating the level of REM due to the low number of female chairmen on the board of listed firms, which need for more investigation in future studies. Data shows that only 24 firms’ observation (2.78) have appointed a female chairman while 840 firms’ observation (97.22) do not prefer a female chairman, hence the dataset of this variable suffers from outliers.

As a result, this study calls the attention of regulators and researchers of CG to further investigate the same variable used in this study, as well as overcome the difficulties raised in this study by covering recent years in order to produce a solid conclusion that may assist policymakers, investors, creditors, and all shareholders. Regulators and practitioners should be interested in research on the role of BC’s characteristics in mitigating the level of REM.