The Influence of Directors' Diversity and Corporate Sustainability Practices on Firm Performance: Evidence from Malaysia

Abstract

This study aims to examine the relationship between directors' diversity (DIRDIV) and financial performance (FP) with a particular focus on the moderating effect of corporate sustainability practices (CSP). The study analyzes a sample of 104 firms listed on the Bursa Malaysia for the period from 2015 to 2017. Directors' diversity is measured by the Blau index, and Tobin's Q is used as a proxy of FP of the firms while the content analysis method is adopted to measure CSP. The study also employs three control variables, namely, board size, firm size, and leverage. Panel corrected standard errors (PCSE) estimator model has been used to test the hypotheses by STATA software. It is found that directors' diversity in terms of independent and non-independent directors significantly and positively affect the financial performance of the firms. Furthermore, this study reveals that CSP significantly moderates the relationship between directors' diversity and financial performance. This study suggests that the government and regulatory bodies should put more emphasis on diversifying the board and follow up the mandatory CSP to enhance financial performance of the firms, which is likely to ensure their long-term survival and to reduce the risk of collapse in the future.

1. Introduction

Globally, corporate governance (CG) remains a focal issue due to financial scandals, crises and collapses of giant corporations. The most recent scandals of Wells Fargo and Equifax are the examples of long-time scandals linking with the renowned public corporations in the USA (Bhagat & Bolton, 2019). After each and every financial scandal and crisis, the policymakers and regulatory bodies frown at and raise questions regarding the effectiveness of CG mechanisms of the corporations. In general, CG refers to the system for directing and controlling an organization (Cadbury, 1992). The recent global financial crisis that has shaken many economies and led to recession, has explicitly highlighted the critical role of CG in the business world. In recent past, the financial crises of 1997 and 2008 have badly affected the Asian economy in general and Malaysian economy, in particular. Poor CG practices in the region have been identified as the main reason and contributing factor to the crises (Bhagat & Bolton, 2019; Kato, Li, & Skinner, 2017). Evidently, many Malaysian firms, such as, Renong Berhad, Transmile Group Berhad and United Engineers (Malaysia) Berhad shut down due to lack of proper CG practices. Collapses of such corporations negatively affected public confidence on the disclosure of corporate performance in their annual reports (Lins, Servaes, & Tamayo, 2017a). This kind of incidents create demand for more regulation and laws to restrain and regulate corporate activities globally (Bhagat & Bolton, 2019). Dias et al. (2016) revealed that during financial crisis, the financial performance (FP) of firms usually deteriorates. A firm’s FP is extensively recognized as an indicator of management’s performance. It reflects the management’s effectiveness and efficiency in utilizing the firm’s resources (Miles & Van Clieaf, 2017). Thus, firms are mostly concerned with their FP to ensure their long-term survival (Odalo, Achoki, & Njuguna, 2016). Moreover, strong FP gives a greater ability to the firm to undertake higher financial risks in capital budgeting (Gómez-Bezares, Przychodzen, & Przychodzen, 2017). In addition, Laallam, Alom, and Mohamad (2017) identified weak CG practices as the cause of poor FP among Malaysian firms. The current market scenario shows that the FP of listed firms in Bursa Malaysia is more volatile and declining (Schaltegger & Burritt, 2017). Due to corporate financial scandals coupled with financial crisis, there is much concern for improving board effectiveness in the corporate sectors globally (Fidanoski, Simeonovski, & Mateska, 2014; Hasan, Rahman, Sumi, Chowdhury, & Miraz, 2020; Reguera- Alvarado, de Fuentes, & Laffarga, 2017). In this perspective, directors’ diversity (DIRDIV) has been considered as one of the important tools to increase the board effectiveness (Kılıç & Kuzey, 2016). Directors’ diversity refers to the diversity of board members in terms of expertise and socio-biographic characteristics (Abdullah, 2014; Fidanoski et al., 2014).

The agency theory also argues that DIRDIV enhances the independence of the board (Jensen & Meckling, 1976). This increases the strength of monitoring and decision-making, which ultimately reduces the agency cost and increase the firm’s financial performance (Hasan & Rahman, 2017, 2019; Hasan et al., 2020; Kamardin, Latif, Mohd, & Adam, 2014; Ramly, Chan, Mustapha, & Sapiei, 2017). According to the MCCG (2000), it is stipulated that independent directors can bring a wider range of activities to the firm. Moreover, if the BOD is comprised of more independent directors along with the non-independent directors, it is more likely to consider the interest of investors as well as other stakeholders while taking any decision. Studies related to DIRDIV and firms’ FP reveal that they are either positively (Ararat, Aksu, & Tansel Cetin, 2015; Post & Byron, 2015; Terjesen, Couto, & Francisco, 2016) or negatively (Abdullah, 2014; Mahadeo, Soobaroyen, & Hanuman, 2012) correlated. Thus, the link between the two variables is not conclusive till date (Abdullah, 2014; Adams, Haan, Terjesen, & Ees, 2015; Roberson, Holmes, & Perry, 2017). Furthermore, a good number of studies concerning DIRDIV and FP of firms reported an insignificant direct relationship between the two (Galbreath, 2018). In fact, little is known about why and when the DIRDIV would influence FP of firms (Roberson et al., 2017). So, it is plausible to carry out further study on the link between DIRDIV and FP issue in a more holistic way (Hassan, Marimuthu, & Johl, 2015b). As the relationships between DIRDIV and FP have showed mixed results, Roberson et al. (2017), Post and Byron (2015) and Umans (2013) believe that the concerned parties might be benefited from the investigation of critical influences of an interaction variable on that relationship.

Furthermore, Roberson et al. (2017) and Post and Byron (2015) perceive that the variation of the results between DIRDIV and FP must be due to other strategic or contextual factors that are not considered in the previous studies. Accordingly, it is assumed that corporate sustainability practices (CSP) of the firm are one of the possible factors that cause the inconclusive result between the DIRDIV and the FP of firms. The stakeholder theory (Freeman, 1984) also supports the notion that, when a firm is engaged in sustainability activities, the diversified board is able to make better decisions leading to better FP of the firm. However, so far, the researchers have not paid enough attention to the role of CSP as a moderator between DIRDIV and FP of firms. Research in the areas of DIRDIV and CSP are often treated separately with less attention paid to the interaction of both the variables (Fernández-Gago, Cabeza-García, & Nieto, 2016). Therefore, this study attempts to fill the research gap by examining the moderating role of CSP on the relationship between DIRDIV and FP of firms in Malaysia.

2. Literature Review and Hypothesis Development

2.1.Directors’ Diversity and Financial Performance

Usually, the independent director of a firm is not an executive of that firm (Zhang, 2012). According to the MCCG (2000), it is stipulated that independent directors in BOD can bring a broader range of activities to the firm. Moreover, if the board has more independent directors, it considers the interest of all the stakeholders while taking any decision. Adams and Ferreira (2007) stated that the main responsibility of BOD is to formulate strategic decisions for the firm. In this regard, presence of independent directors is helpful in making better strategic choices. Fama and Jensen (1983) also argued that independent non-executive directors can monitor the executive management effectively to ensure the proper utilization of shareholders’ investment and rate of return. Furthermore, Terjesen et al. (2016) states, globally it has been accepted by the researchers, academicians and policy makers that independent directors in the board ensure transparency, effective monitoring and supervision over the top level management of a firm which enhances the FP of the firms.

The resource dependence theory argues that independent directors bring more resources, knowledge, information and justice in the board (Pfeffer & Salancik, 2003). Moreover, independent directors are fairer to disclose and report the company’s activities in their annual report compared to the non-independent directors of BOD. For ensuring better CG practices in the firm, it is essential to appoint independent directors in the board, because they are more conscious about the interests of the stockholders and other stakeholders of the firm (Hasan, Molla, & Khan, 2019; Hasan & Rahman, 2020; Jahid, Rashid, Hossain, Haryono, & Jatmiko, 2020). According to Monks and Minow (2004) both independent and non-independent directors have same duties and responsibilities such as, work for the best interest of shareholders, ensure the best practices of CG, carefully make the strategic decisions, ensure the proper utilization of resources of the firm. A large literature analyzed the effect of independent directors on FP and found inconclusive results. Several studies reveal that (Florackis & Ozkan, 2009; Hasan et al., 2019; Kim & Lim, 2010; Pombo & Gutiérrez, 2011; Tulung & Ramdani, 2018) the percentage of independent directors positively influences firms’ FP. On the contrary, Arosa, Iturralde, and Maseda (2010), Bhagat and Black (2001) find that the existence of independent directors in the board does not increase the value of the firm. Moreover, Kang, Cheng, and Gray (2007) find mixed results between the two variables. Further, the agency theory also postulates that presence of independent directors in the board increases the independence of the board leading to better monitoring and decision making (Jensen & Meckling, 1976). The following hypothesis is therefore posited.

H1: Directors’ diversity positively influences the financial performance of firms.

2.2.Impact of Corporate Sustainability Practices (CSP) on the Association between Directors’ Diversity and the Financial Performance of Firms

CSP is the alternative concept of corporate social responsibility (CSR) or sustainable development (Christofi, Christofi, & Sisaye, 2012; Masum, Hasan, Miraz, Tuhin, & Chowdhury, 2020; Teanpitthayamas, Suttipun, & Lakkanwanit, 2021). Before the 1990s, the term ‘sustainability’ was used to mean the ability of a firm to increase its profit gradually. Later, the term ‘CSP’ incorporates three aspects of business activities namely, economic, social, and environmental (Adams, Thornton, & Sepehri, 2012). Many firms who are credited due to their contribution to the technological and economic developments have been criticized for creating environmental and social problems, like water pollution, air pollution, CO₂ emission, waste, production of unhygienic product, and unhealthy environment of the workers (Hussainey & Walker, 2009). To resolve these issues, it is essential for the firms to practice CSP in all sectors (Abd-Mutalib, Jamil, & Wan-Hussin, 2014).

Moreover, the stakeholder theory argues that firms should make a mutual association with its stakeholders by improving the moral, ethical and social standards (e.g., through demonstrating robust CSP) (Freeman, 2004; Jensen, 2001). As diversified directors of the firms are in a position to take care of the stakeholders’ interest over the profit motive of the firm, they are more likely to ensure higher CSP. Therefore, the DIRDIV may affect the FP of firms more with the support of their CSP as a means of responding to stakeholder needs and interests (Lins, Servaes, & Tamayo, 2017b; Rivera, Muñoz, & Moneva, 2017). Hence, whether CSP moderates the relationship between DIRDIV and firms’ FP needs an empirical study. Thus, the following hypothesis is proposed:

H2: CSP moderates the relationship between directors’ diversity and financial performance of firms.

3. Research Framework

The research framework constructed for this study is portrayed in Figure 1. The framework comprises directors’ diversity as explanatory variable, financial performance of the firm is the outcome variable whereas board size, firm size and leverage are the control variables.

Figure 1: The Research Framework

4. Materials and Method

4.1. Population and Sample Selection

All listed firms in Bursa Malaysia main market constitute the population of this study. Among the 805 publicly-listed firms in Bursa Malaysia, this study has narrowed down to top 104 firms on the basis of market capitalization for the years 2015 to 2017 and the total number of firm-year observations are 312. This study has taken all the listed non-financial firms whose market capitalization is more than RM2 billion and above for its target sample. As these firms are obligated to disclose the sustainability report according to Bursa Malaysia Sustainability guide-2015 in their annual report from the year ended 2016. Along with the requirement of Bursa Malaysia Sustainability Reporting Guide-2015, there are some other reasons for selecting the large firms. As the large firms are more visible, they carry out more activities and they have more impact on the society (Hackston & Milne, 1996). Moreover, large firms do believe to get more information which encourage them to do more sustainability practices (Aerts, Cormier, Gordon, & Magnan, 2006; Buniamin, 2010). The years 2015, 2016 and 2017 are chosen because they provide most recent data regarding DIRDIV, CSP and FP of the firms.

4.2. Measurement of the Variables and Data Collection Procedure

This study has measured financial performance using Tobin’s Q. According to Feldman and Montgomery (2015), Tobin’s Q represents not only the evaluation by investors’ evaluation in terms of share price, but also reflects the future prospects of the firm. The value of Tobin’s Q>1 means the firm is worth more than its book value. On the contrary, Tobin’s Q<1 means that the market is expecting the firm to rescind shareholders’ wealth in long run (Terjesen et al., 2016). Tobin’s Q is measured by the summation of the market value of equity and book value of total debts divided by the book value of total assets (Conyon & He, 2017; Gordini & Rancati, 2017; Machmuddah, Sari, & Utomo, 2020; Utami & Hasan, 2021). The present study employs Blau’s heterogeneity index (Blau, 1977) on the basis of proportion of non-executive independent directors and remaining directors (non-independent directors) to measure the directors’ diversity. The Blau’s heterogeneity index is an appropriate measure of heterogeneity (Miller & del Carmen Triana, 2009). It is also the most favorable measure to capture diversification within a group of individual in an organization (Harrison & Klein, 2007). For avoiding biasness of the results, this study uses board size, firm size and leverage as control variables. Board size is measured by the number of board members in the BOD of a firm (Hasan et al., 2019; Tran, Lam, & Luu, 2020). Natural log of total assets is used to measure the firm size (Hasan et al., 2019; Hasan & Rahman, 2020; Jizi, 2017; Odalo et al., 2016). Ratio of total debt to total assets of a firm is applied to measure the leverage of the firm (Chen, Ma, Shi, Tu, & Xu, 2020; Jizi, 2017).

The data has been collected from the audited annual reports of the selected firms as they are most accessible and acceptable sources of information in Malaysia (Sadou, Alom, & Laluddin, 2017). The information of proxies of FP of firms such as Tobin’s Q and other control variable information namely leverage and firm size have been collected from the financial database, namely, Thomson Reuters Data Stream. The data of the moderating variable, CSP has been collected from the annual reports of the selected firms by content analysis. Amran (2012) reveals that most of the Malaysian firms use the annual report to disclose their sustainability information. Moreover, Deegan and Rankin (1996) find that annual reports are more reliable than any other sources of sustainability information. The content analysis method has been widely used for collecting the CSP data by many previous studies (Jahid et al., 2020; Saleh, Zulkifli, & Muhamad, 2011; Uadiale & Fagbemi, 2012; Uwuigbe & Egbide, 2012; Zahari, Esa, Rajadurai, Azizan, & Muhamad Tamyez, 2020).

4.3. Model Specification

To investigate the effects of directors’ diversity on financial performance the following analytical model is specified, with variable code names and descriptions.

TQ_it = α + β₁DIRDIV_it + β₂BRDSIZE_it+ β₃FRMSIZE_it + β₄LEVRGE_it + ε_it       (1)

The hierarchical moderated multiple regression model is developed following Baron and Kenny (1986) and Hair, Black, Babin, and Anderson (2010) to examine the moderation effect of CSP according to the hypothesis of this study.

TQ_it = α + β₁DIRDIV_it + β₂BRDSIZE_it + β₃FRMSIZE_it + β₄LEVRGE_it + β₅CSP_it + ε_it       (2)

TQ_it = α + β₁DIRDIV_it + β₂BRDSIZE_it + β₃FRMSIZE_it + β₄LEVRGE_it + β₅CSP_it + β₆DIRDIV × CSP_it + ε_it       (3)

Where,

TQ = Tobin’s Q (Market based financial performance measure of firm)

DIRDIV = Independent-Non independent directors’ diversity (measured by Blau Index)

CSP = Corporate sustainability practices (measured by content analysis)

BRDSIZE = Board size (measured by the total number of directors in the board of a firm)

FRMSIZE = Firm size (measured by natural log of total assets of a firm)

LEVRGE = Leverage (measured by total debt divided by total assets of a firm)

DIRDIV × CSP = Interaction terms

i = Observation and t = Year of observation

5. Results

5.1. Descriptive Analysis

Table 1 shows the descriptive statistics of the dependent variable, the independent variables, the control variables and the moderating variable used in the study. The dependent variable is TQ while the independent variable is DIRDIV, CSP is the moderating variable and BRDSIZE, FRMSIZE, and LEVRGE are the control variables. Firm performance, as measured by TQ, varies from as low as 0.21 to a maximum of 13.87 with an average of 1.90. The mean score is similar to those reported by Abdullah and Ismail (2013); Hassan, Marimuthu, and Johl (2015a). The mean score, minimum and maximum value of Directors’ diversity are 0.4660, 0.0 and 0.5 respectively. Based on the Blau index (Blau, 1977) the range of minimum to maximum is 0.00 to 0.50 for the DIRDIV of a firm as in this study there are two categories of directors in the board are considered. The study finds a higher level of DIRDIV in the BODs in Malaysia. The mean score, minimum and maximum value of board size are 9.0577, 5 and 17 respectively. The results find that the minimum number of directors is 5 and maximum 17.

Table 1: Descriptive Statistics

The mean score, minimum and maximum value of firm size are 6.7798, 5.2769 and 8.1590 respectively. The mean score, minimum and maximum value of leverage are 0.2538, 0.00 and 0.6851 respectively. It reveals that some firms have no debt and other has 68.51% debt of their total assets. The mean score, minimum and maximum value of CSP are 164.9583, 0.00 and 1098 respectively. It indicates that some firms have no CSP and some have high level of CSP disclosed in their annual reports.

5.2. Test of Multicollinearity

Multicollinearity is the issue of having high correlation among independent variables, which could inflate the regression results (Pallant, 2007). This study presents correlation matrix in Table 2, which represent correlation among the variables of the study. From Table 2, this study documented highest correlation between TQ and FRMSIZE, which is 51.58%, and significant at 1% level whereas lowest correlation exist between DIRDIV and FRMSIZE, which is 0.18% and insignificant. Most of the variables in this study have lower level of correlation. Hair, Black, Babin, Anderson, and Tatham (2006) and Tabachnick, Fidell, and Ullman (2007) recognize the problem of multicollinearity if the correlation between variables is more than 0.9. Since the highest value is less than 0.9, there is no evidence of multicollinearity problem among variables in the model.

Table 2: Correlation Matrix

Correlation is significant at the 0.01 level***, 0.05 level** and 0.1* level

In addition to correlation matrix, this study also considers VIF and tolerance value to examine the issue of multicollinearity further, which is presented in Table 3. Hair et al. (2006) note that multicollinearity problems exist when VIF values are above 10 (or Tolerance value is less than 0.10). As shown in Table 3, there appeared to be no evidence of multicollinearity problem in the model as all variables’ VIF are less than 10 and tolerance value is more than 0.10.

Table 3: VIF and Tolerance Value

5.3. Test of Heteroscedasticity and Autocorrelation

In multiple regression model, for analyzing the panel data, heteroscedasticity problem is a major concern as it can invalidate the efficiency of statistical results (Brooks, 2014; Hair et al., 2010). It is argued that ignoring the presence of heteroscedasticity can result in inefficient coefficient estimations and biased standard errors (Baltagi, 2008). To detect heteroscedasticity, the formal statistical test Breusch and Pagan (1979) has been used in this study. According to Brooks (2014) the null hypothesis of the Breusch-Pagan test is homoscedasticity; if p-value < 0.05, it is a case of heteroscedasticity. From Table 4, test reports the value of Chi² statistics is 144.44 and the corresponding p-value < 0.05. As the null hypothesis is rejected, the heteroscedasticity problem is found in the model.

Table 4: Results of Heteroscedasticity and Autocorrelation Tests

Autocorrelation is the issue of error components being correlated across time due to high similarities. The regression model assumes that the error term of units is not correlated and not influenced by other units. Although this is a violation of the ordinary assumption, it is a common issue in panel or time-series analysis (Wooldridge, 2010). Gujarati and Porter (2009) suggested that the Wooldridge test is most suitable for serial correlation and to detect first-order autocorrelation in panel data. Further, test for autocorrelation in panel data is used to detect serial or first-order autocorrelation. The result of the test presented in Table 4 shows that this model is found to be not significant at p > 0.05. The result failed to reject the null hypothesis and concluded that the data for TQ model have no first-order autocorrelations. However, this can only be accurate if the panel data is free from cross sectional dependence as explained by Petersen (2009).

5.4. Cross Sectional Dependence and Hausman Test

Cross-sectional dependence is also known as contemporaneous correlation, refers to correlation of the residuals across entities. Pesaran’s test is the appropriate test to explore whether the data has cross-sectional dependence problem. It is the most appropriate test for the panel data that has large cross-sectional units and small time-series (Hoyos & Sarafidis, 2006). The test is applied to the model and confirmed the existing of cross-sectional dependence in the model, which is presented in the Table 5.

Table 5: Results of Cross Sectional Dependence and Hausman Tests

Accordingly, the presence of the problem has to be corrected. According to Gujarati and Porter (2009), the Hausman test is employed to decide whether fixed or random effects model is suitable for this study. Based on the Table 5, the Hausman test shows that random effect model is more appropriate to analyze the panel data for this study.

5.5. Regression Analysis

From the diagnostic tests, this study finds that random effect model is more appropriate to run the multiple regressions of this study. However, the potential econometric problems of heteroscedasticity, and cross sectional dependence are found in the data. Random effects models with heteroscedasticity cannot be efficiently estimated with OLS. To solve the above issues, this study can use a feasible generalized least square (FGLS, or xtgls command in Stata)to correct the standard error (Wooldridge, 2010). However, Beck and Katz (1995) suggested that for cross sectional time-series data, researchers should use OLS with heteroscedastic panels corrected standard errors (OLS-PCSE, or xtpcse command in Stata), because the standard errors of the estimated coefficients based on FGLS may underestimate the true sampling variability. Their Monte-Carlo analysis shows that OLS-PCSE performs better than FGLS in estimating the standard errors (Moundigbaye, Messemer, Parks, & Reed, 2019; Nithithanatchinnapat & Joshi, 2019). Moreover, the FGLS estimator is more appropriate for panels with T > N and PCSE is more suitable for the panel with T < N (Miao, Gu, Zhang, Zhen, & Wang, 2019). Another advantage of this technique is that it allows for disturbances that are heteroscedastic and contemporaneously corrected across panels (Reed & Webb, 2010). The PCSE estimate is robust not only to unit heteroscedasticity, but also robust against possible contemporaneous correlation across the units (Bailey & Katz, 2011; Hasan et al., 2019). Thus, this study assumes that PCSE is the most suitable estimator for analyzing the panel data for this study. The results of the multiple regressions of DIRDIV and FP and the moderating role of CSP on the relationship between the above two variables are presented in the table 6. The results of Model 1 shows that there is a positive and significant relationship between directors’ diversity and financial performance, which suggests that if the BOD is comprised of the same proportion of both independent and non-independent directors, it increases the financial performance of firms. The result supports the hypothesis H₁. The findings of this study are consistent with the study of Duchin, Matsusaka, and Ozbas (2010), Kweh, Ahmad, Ting, Zhang, and Hassan (2019), Mahadeo et al. (2012), but inconsistent with the results of Zabri, Ahmad, and Wah (2016). To measure the directors’ diversity and financial performance of firms the hierarchical moderated multiple regression model has been used. This regression model is more appropriate to evaluate the effect of a moderating variable in a study (Han, Yoon, Suh, Li, & Chae, 2019; Li, Sharp, Bergh, & Vandenberg, 2019; Ruiz-Jiménez, del Mar Fuentes-Fuentes, & Ruiz-Arroyo, 2016; Tran & Pham, 2019). The hierarchical regression result in the Table 6 shows that the interaction term DIRDIV * CSP is found to have negative and statistically significant impact on TQ. Hence, the results in Model 3 indicate that CSP negatively and significantly moderates the relationship between DIRDIV and FP of firms in Malaysia. The result supports the hypothesis H2.

Table 6: Results of Regression Model Using Panel Corrected Standard Errors (PCSE) Estimator

Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.

5.6. Robustness Check

This study also conducted additional analysis to ensure the sensitivity and robustness of the main results reported earlier. The diagnostic tests confirmed the problems of heteroscedasticity and cross-sectional dependence are found in the study model. Therefore, this study corrected the issues by employing Driscoll and Kraay’s standard errors (xtscc command in STATA) estimator based on Hoechle (2007), which is also robust to heteroscedasticity and cross-sectional dependence in panel data (Table 7). After analyzing the robustness tests it has been found that the results are almost similar between the XTPCSE and XTSCC estimations. Therefore, it can be said that the estimation of the study model is robust and free from misappropriation, which lead to a reliable conclusion.

Table 7: Results of Regression Model Using Driscoll and Kraay’s Standard Errors Estimator moderating impact of CSP on the relationship between

Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.

6. Conclusion

This study has examined and analyzed the relationship between directors’ diversity and financial performance of firms in Malaysia. It also investigates the moderating effect of corporate sustainability practices on the relationship between the above two variables. The motivation for studying the corporate governance mechanisms originates from the inconclusive findings of the relationship between directors’ diversity and financial performance of the firms. After analysis of the data, it is found that CSP has a strong effect to moderate the relationship between board diversity and financial performance of the firms in Malaysia. Hopefully, it would provide contributing evidence to explain the mechanisms behind the link between directors’ diversity and financial performance. The findings have policy implications that the government and the regulatory bodies should put more emphasis on diversifying the board and following up the mandatory corporate sustainability practices to enhance financial performance of the firms in Malaysia and to ensure their long term survival as well as to reduce the risk of collapse in the future. Although this study makes a definite empirical contribution to the existing literature, there are some limitations that need to be taken into account for further research. Firstly, this study has focused only one characteristic, namely, directors’ diversity of BOD for their impact on financial performance of the firms. While the other characteristics of BOD, like political connection, multi directorship, CEO duality, etc., are also important factors that might be considered in future research. Secondly, in this study, the Tobin’s Q, a market-based financial measure has been used as a proxy of the financial performance of firms. Future studies may consider the book value measure like ROA, ROE and some other proxies for measuring the financial performance of the firms.