The impact of board structure on financial leverage of Vietnamese listed firms

ed studies focus on finding factors that influence the capital structure of firms listed on the Vietnamese stock markets. Specifically, Đặng and Quách (2014) identify three factors that have a strong impact on the capital structure of a firm, namely, firm size, profitability (positive impact), and taxes (negative impact). Previously, Trương and Võ (2008) affirm that capital structure is positively correlated with company size, industry, and revenue growth and is inversely correlated with profitability. In addition, capital structure is positively correlated with the number of directors. Recently, a series of studies on factors affecting the capital structure of firms in specific industries was also conducted. The studies included firms in the logistics industry (Lương, Phạm, Nguyễn, Nguyễn, Nguyễn, & Phạm, 2020), the food industry (Lê, Bùi, & Lê, 2020), the Vietnam Oil and Gas Group (Vũ & Nguyễn, 2013), and the seafood industry (Nguyễn, 2008). Most studies show that firm size, growth rate, and profitability are positively correlated with capital structure. Some other factors that are negatively correlated are also mentioned, including taxes, liquidity, profits, and business risk. Most of the research on Vietnam markets mentioned above use relatively small datasets over short time periods and focus on factors affecting capital structure. Only in the last few years have some specific studies on the relationship between corporate governance and capital structure been published. Nguyễn, Trần, Nguyễn, Võ, and Nguyễn DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 81 (2016) examine the impact of corporate governance (state ownership, financial institutions, foreign investors, members of the BOD, and the largest shareholders of the firm) and firm characteristics (size, profitability, tangible assets, tax shield, and the gap between optimal leverage and observed leverage) on capital structure decisions. Research results show that corporate capital structure not only depends on the characteristics of the business, but is also influenced by enterprise ownership characteristics. In another study by Phan, Trần, and Trần (2017), the role of CEO duality is examined. Their study confirms that firms with a dual leadership structure performed more effectively. As previous research is still inconclusive, a comprehensive study with a large set of data on companies listed on the Vietnamese stock market would provide significant insights into the relationship between BOD structure and capital structure (financial leverage). Based on the theory and previous research results, this study hypothesizes that: • H1: Board size has an impact on financial leverage. • H2: The number of annual board meetings has an impact on financial leverage. • H3: The number of female directors on the BOD has an impact on financial leverage. • H4: CEO duality has an influence on financial leverage. 3. DATA AND RESEARCH METHOD 3.1. Definitions of variables and data collection method This study examines the relationship between the BOD structure and the financial leverage of companies listed on the Vietnamese stock market. For quantitative analysis, the authors use leverage ratio, which is defined as the ratio of total debt to total assets of the firm, similar to the study by Haque, Arun, and Kirkpatrick (2011). The independent variables used in this study include the size of the board, the number of board meetings per year, the number of female directors on the board, and an indicator variable indicating whether the chairman also holds a CEO position. To increase the effectiveness of the estimate, two variables representing board size and the number of board meetings were converted to logarithms prior to analysis. In addition to the number of female directors on the board, the authors also use two other definitions of gender diversity of the BOD, namely, the percentage of female directors on the board and an indicator variable indicating the presence of female directors. Using different definitions of the gender variable in the analysis will help increase the reliability of the results. To control the impact of other factors on a firm's leverage, the authors use firm size, fixed assets, liquidity, profitability, and business growth rate as control variables, Hoang Mai Phuong and Nguyen Thanh Hong An 82 similar to studies on the effect of firm characteristics on financial leverage according to Bradley, Jarrell, and Kim (1984), Castanias (1983), Long and Malitz (1985), and Titman and Wessels (1988). These studies generally agree that financial leverage has a positive relationship with firm size, fixed assets, and growth rate, and an inverse relationship with returns and liquidity. Detailed definitions of the variables are presented in Table 1. Table 1. Variable definitions Variable Code Definition/Formula Dependent variable Financial leverage Lev Total debt/Total assets Independent variables BOD size Lbsize Logarithm (Number of directors on the board) BOD meeting frequency Lmeet Logarithm (Number of BOD meetings per year) Female directors Female Number of female directors on the BOD CEO duality Ceodual Equals 1 if the CEO is also the BOD chairman, 0 otherwise Control variables Firm size Lfsize Logarithm (Total assets) Fixed assets Fixed_assets (Total assets - Short-term assets)/Total assets Liquidity Liquidity Short-term assets/Short-term liabilities Profitability ROA Net profit/Total assets Growth Salegrowth Annual sales growth The data are collected from audited financial reports, annual reports, and annual executive reports of nonfinancial companies listed on the Ho Chi Minh City Stock Exchange from 2012 to 2019. Companies with insufficient data are excluded from the sample. 3.2. Research method To analyze the relationship between the variables representing the characteristics of the BOD and the leverage ratio, the authors propose the following research Model: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽3𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽4𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽5𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽6𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽7𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽8𝑅𝑂𝐴𝑖𝑡 + 𝛽9𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽10𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜀𝑖𝑡 (1) To estimate Model (1), the authors use the ordinary least squares method with panel data (POLS). However, estimation by this method does not guarantee consistency for two reasons. First, the omitted factors are likely to interact with the independent DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 83 variables in the Model, causing the estimate to be biased and inconsistent. Second, since this Model does not take advantage of information from the differences between firms, the estimates may be less accurate (Wooldridge, 2002). As a remedy, the authors restructure Model (1) to incorporate the differences among companies (representing by  in the new Model) in the dataset: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽3𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽4𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽5𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽6𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽7𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽8𝑅𝑂𝐴𝑖𝑡 + 𝛽9𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽10𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (2) Model (2) is estimated by the regression method with random and fixed effects, respectively. The LM test is used to choose between the POLS regression Model and the regression Model with random effects. Then, the Hausman test is used to choose between the regression Model with random effects and the regression Model with fixed effects. However, the recent research of Liao, Mukherjee, and Wang (2015) indicates that firms tend to adjust their leverage toward an optimal value over time. As discussed, if the BOD actually impacts financial leverage decisions, the adjustment effect implies that the BOD would refer to the past leverage level when deciding the future leverage ratio. In other words, Levit and Levit-1 are correlated. The fact that Model (2) omits this important variable (i.e., Levit-1) reduces the accuracy of the estimates. Furthermore, if Levit-1 is correlated with the present structure and operation of the BOD, a case which is raised in previous research by Berger et al. (1997), the omission of Levit in Model (2) would render the estimates inefficient and inconsistent. As a remedy, Model (2) is restructured as follows: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿. 𝐿𝑒𝑣𝑖𝑡 + 𝛽3𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽4𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽5𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽6𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽7𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽8𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽9𝑅𝑂𝐴𝑖𝑡 + 𝛽10𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽11𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (3) Model (3) cannot be consistently estimated by the methods used for Models (1) and (2) because the endogeneity problem caused by the inclusion of the variable Levit-1. Instead, Model (3) is estimated using the system GMM method by Blundell and Bond (1998). This method provides a plausible solution for endogeneity when valid instruments are not available, which is a common problem encountered by researchers in the field of corporate governance (Nguyen, Locke, & Reddy, 2015). In particular, this method utilizes a set of instrument variables derived from the available dataset, namely, lagged and differenced variables of the endogenous variables. Together with the estimation method using moments, the system GMM produces consistent estimates. Hoang Mai Phuong and Nguyen Thanh Hong An 84 4. RESULTS AND DISCUSSION 4.1. Descriptive statistics The authors collected data based on the variable definitions and methodology presented in Section 3. The final dataset included 1,592 observations collected from 199 firms from 2012 to 2019. Summary statistics are presented in Table 2. Table 2. Descriptive statistics Number of observations Mean Median Std. dev. Min Max Lev 1,592 0.49 0.50 0.21 0.03 1.29 Bsize 1,592 5.76 5.00 1.33 1.00 11.00 Meet 1,592 10.97 7.00 11.61 1.00 170.00 Female 1,592 0.90 1.00 1.04 0.00 6.00 Female_% 1,592 0.15 0.14 0.17 0.00 1.00 Female_dummy 1,592 0.55 1.00 0.50 0.00 1.00 Ceodual 1,592 0.28 0.00 0.45 0.00 1.00 Lfsize 1,592 21.17 21.08 1.26 18.46 26.72 Fixed_assets 1,592 0.41 0.37 0.23 0.01 0.98 ROA 1,592 0.06 0.05 0.08 -0.85 0.78 Liquidity 1,592 2.28 1.58 2.46 0.05 39.37 Salegrowth 1,592 0.18 0.07 1.53 -24.16 29.56 Notes: The analysis is performed based on 1,592 observations collected from 199 nonfinancial companies listed on the Ho Chi Minh City Stock Exchange from 2012 to 2019; Bsize is the number of board members; Lev is the ratio of total debt to total assets of the business; Meet is the number of board meetings per year; Female is the number of female members on the Board of Directors. Female_% is the percentage of female directors; Female_dummy is an indicator variable, with the value equal to 1 if the board has female members and 0 if not; Ceodual is an indicator variable with value equal to 1 if the chairman of the BOD also holds the position of CEO and 0 if not. Lfsize is the logarithm of the firm's total book value; Fixed_assets is the ratio of fixed assets to total assets; ROA is the ratio of after-tax profit to total assets; Liquidity is the ratio of short-term assets to short-term liabilities. Salegrowth is the rate of sales growth. Based on Tables 2 and 3, we can see that the average leverage of the surveyed enterprises is about 49.0%. This figure is significantly higher than the figure of 29.0% in the study of Nguyen, Locke, and Reddy (2015) on the companies listed on the Ho Chi Minh City stock market from 2008 to 2012. The average leverage ratio of businesses was kept quite stable over the years surveyed, ranging from 47.7% to 49.6%. However, there was a clear difference in leverage between different industries. The Arts and Entertainment sector had the lowest leverage, at around 13.1%, while the Wholesale and Construction–Real Estate sectors had the highest leverages, exceeding 58.0%. Figure 1 also shows that the leverage ratios of industries also had different fluctuations over the years. This information not only gives us a general picture of the leverage ratio of the businesses listed on the Ho Chi Minh City stock exchange but is also very useful in DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 85 estimating the regression in the following section. As for the independent variables, the surveyed firms' boards of directors have an average of about six members, with the majority of the firms having a BOD consisting of five members. This number is almost equivalent to that found by Nguyen et al. (2015) for 2008 to 2011, showing that the size of the BOD of listed companies in Vietnam remained quite stable through the years. Boards usually hold an average of 11 meetings per year, but most of them only hold about seven meetings. Regarding the proportion of women, the BODs of the surveyed companies average 15% female members (equivalent to about one person) and more than 55% of the companies surveyed have female directors. This figure is higher than in the study of Nguyen et al. (2015), in which the proportion of women directors was 12% and about 51% of surveyed firms had female board members. This is a sign that there has been an improvement in gender diversity on the boards of Vietnamese-listed companies in recent years. Finally, about 28% of the surveyed businesses have a chairman of the board who concurrently holds the position of CEO. This figure has decreased by 4% compared with the corresponding study of Nguyen et al. (2015) for 2008 to 2012. Table 4 presents the correlation coefficients among the variables. The results show that the correlation coefficients between the independent and control variables are below 0.5, implying that multicollinearity is not a serious problem in the regression analysis. This conclusion is supported by the calculated VIF for the independent and control variables of less than 2. In addition, the correlation analysis results show that the number of board meetings is positively correlated, while board size is negatively correlated, with firm leverage. However, this result is not reliable because the correlation analysis does not take into account the impact of other covariates on the relationship between the two considered variables. To analyze the relationship between the independent variables and a firm's leverage ratio consistently and effectively, regression analysis should be performed. Figure 1. Financial leverage by industries Notes: Companies are classified by their first registered type of business. Industries are classified in accordance with NAICS (North American Industry Classification System) 2007. Hoang Mai Phuong and Nguyen Thanh Hong An 86 Table 3. Financial leverage by various industries 2012 2013 2014 2015 2016 2017 2018 2019 Mean Wholesale 0.6462 0.6185 0.6081 0.5859 0.5629 0.5573 0.5487 0.5538 0.5852 Retail 0.5491 0.5644 0.5196 0.5360 0.5484 0.5732 0.5040 0.5209 0.5394 Information Technology 0.4706 0.4939 0.5120 0.4918 0.4852 0.4343 0.4619 0.4766 0.4783 Accommodation and Catering 0.3022 0.2541 0.2810 0.5585 0.5690 0.3811 0.4367 0.5068 0.4112 Mining 0.3388 0.3002 0.2861 0.3183 0.2498 0.2512 0.2940 0.2571 0.2869 Arts and Entertainment 0.1309 0.1498 0.1558 0.1304 0.1313 0.1153 0.1214 0.1107 0.1307 Production 0.4431 0.4583 0.4496 0.4423 0.4502 0.4606 0.4665 0.4582 0.4536 Agriculture 0.4282 0.3625 0.3295 0.3733 0.2975 0.3361 0.3333 0.4119 0.3590 Utilities 0.4936 0.4735 0.4576 0.4550 0.4566 0.4606 0.4430 0.4493 0.4611 Transportation and Warehousing 0.4058 0.4002 0.3994 0.4076 0.3926 0.3777 0.3668 0.3408 0.3864 Construction and Real Estate 0.5845 0.5856 0.5855 0.5859 0.5839 0.5862 0.5876 0.5658 0.5831 Mean 0.4953 0.4959 0.4880 0.4874 0.4839 0.4872 0.4859 0.4774 0.4876 Notes: Companies are classified by their first registered type of business. Industries are classified in accordance with NAICS (North American Industry Classification System) 2007. Table 4. Correlation coefficients of research variables Lev Lbsize Lmeet Female Ceodual Lfsize Fixasset ROA Liquidity Sale- growth Lev 1.00 Lbsize -0.05** 1.00 Lmeet 0.12*** 0.09*** 1.00 Female -0.03 0.25*** 0.02 1.00 Ceodual 0.02 -0.01 0.01 0.10*** 1.00 Lfsize 0.28*** 0.29*** 0.20*** 0.17*** -0.01 1.00 Fixasset -0.16*** 0.17*** -0.11*** -0.05* -0.11*** 0.10*** 1.00 ROA -0.46*** 0.10*** -0.04 0.04 -0.06** -0.04* -0.05* 1.00 Liquidity -0.46*** -0.02 -0.03 -0.03 0.07*** -0.14*** -0.15*** 0.29*** 1.00 Salegrowth 0.01 0.02 0.05** 0.04* 0.05* 0.07*** -0.05** -0.01 0.02 1 Notes: *, **, and *** correspond to the 10%, 5%, and 1% levels of significance, respectively. The analysis is based on 1,592 observations collected from 199 nonfinancial companies listed on the Ho Chi Minh City Stock Exchange from 2012 to 2019. Bsize is the number of board members. Lev is the ratio of total debt to total assets of the business. Meet is the number of board meetings per year. Female is the number of female members of the Board of Directors. Female_% is the percentage of female directors. Female_dummy is an indicator variable equal to 1 if the board has female members, and 0 if not. Ceodual is an indicator variable equal to 1 if the chairman of the BOD also holds the position of CEO, and 0 if not. Lfsize is the logarithm of the firm's total book value. Fixed_assets is the ratio of fixed assets to total assets. ROA is the ratio of after-tax profit to total assets. Liquidity is the ratio of short-term assets to short-term liabilities. Salegrowth is the rate of sales growth. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 87 4.2. Regression analysis To analyze the relationship between the characteristics of the BOD and financial leverage ratio, the authors conducted a preliminary empirical analysis using a static Model structured as in Models (1) and (2) of Section 3. The results are shown in Table 5. Before interpreting the results, a number of standard tests were performed to select the optimal Model. First, Breusch-Pagan’s LM test was performed to choose between the POLS Model (Column 1) and the regression Model with random effects (Column 2). The results show that the test statistic, Chibar-square (1) = 22.800, corresponds to the value Prob (chibar-square) = 0.000, indicating that the Model with random effects (Column 2) was more effective than the POLS Model (Column 1). To choose between the random effects Model (Column 2) and the fixed effects Model (Column 3), the Hausman test was performed. The results show that the test statistic Chi-square (16) = 1,055.300, which corresponds to Prob (chi-square) = 0.000, and that the fixed effects Model (Column 3) is more consistent and efficient than the random effects Model (Column 2). In conclusion, the two tests indicate that the Model with fixed effects should be used. First, the regression results with the POLS Model (Column 1) suggest that the makeup of the BOD has an impact on the firm's leverage decisions. Specifically, the variable Female has a negative coefficient and is statistically significant at 1%, implying that having more female directors is associated with a lower rate of financial leverage. The remaining three characteristics of the board, namely, size of the board (Lbsize), number of meetings per year (Lmeet), and a CEO who is also the chairman of the board (Ceodual), have no impact on a firm’s leverage ratio. However, the results are not reliable because estimation by the POLS method can be biased and inconsistent due to the omission of firm-level characteristics. To take into account unobserved factors at the firm level, the tests show that the regression Model with fixed effects (Column 3) is optimal. The regression results in Column (3) show that, after taking into account other unobservable characteristics at the firm level, the variables represent the performance of the BOD, including the size of the board (Lbsize), the number of meetings per year (Lmeet), the number of female members (Female), and whether the CEO is also the chairman of the BOD (Ceodual), have no impact on the leverage ratio of the firms. In Columns (4) and (5), the authors re-estimate Model (2) with two other common definitions of board gender diversity, namely, the proportion of female directors (Female_%) and presence of female directors (Female_dummy). In summary, the results of the static Model suggest that the BOD does not have an impact on the financial leverage ratio of businesses listed on the Ho Chi Minh City Stock Exchange. This result is contrary to the findings of Nguyen et al. (2015) and Phan et al. (2017). Hoang Mai Phuong and Nguyen Thanh Hong An 88 4.3. Robust tests Some recent studies have shown that firms tend to keep their leverage at a target level they consider optimal. This means that when the leverage is higher than the target, the business will adjust it downward and when the leverage is lower than the target, they will adjust it upward (Liao et al., 2015). This adjustment mechanism implies that leverage in the present (Levit) is correlated with leverage in the past (L.Levit). To the extent that this is true, Models (1) and (2) have omitted an important factor (L.Levit), and this would lead to inefficiencies in the estimates presented in Table 5. More seriously, if the omitted variable and the independent variables (in this case, the quality of the board's performance) are correlated (the so-called endogenous phenomenon), estimates for the relationship between the BOD’s quality and the firm’s financial leverage would be biased and inconsistent, implying that the conclusions based on the static Model in Table 5 do not reflect reality. This possibility is even more evident as some past studies, such as Kyereboah-Coleman and Biekpe (2006), Abor (2007), Bokpin and Arko (2009), and Rose et al. (2013), have shown that the BOD really has a voice in deciding a firm’s financial leverage. Correlation analysis between L.Levit and the independent variables in Model (2) also shows that the correlation coefficient between L.Levit and Lbsizeit is -0.05 and statistically significant at 10%, the correlation coefficient between L.Levit and Lmeetit is 0.13 and statistically significant at 1%, the correlation coefficient between L.Levit and Lfsizeit is 0.25 and statistically significant at 1%, the correlation coefficient between L.Levit and Fixed_assetsit is -0.16 and statistically significant at 1%, the correlation coefficient between L.Levit and ROAit is -0.39 and statistically significant at 1%, and the correlation coefficient between L.Levit and Liquidityit is -0.42 and statistically significant at 1%. Based on the results of previous studies and preliminary empirical evidence, we find that endogeneity is very likely to exist, and the hypothesis testing results based on the estimates in Table 5 may not be correct. Therefore, the authors reconstructed Model (2) by adding the one-step lag variable of the dependent variable (L.Levit) to the list of independent variables. The modified Model is shown in Model (3). Model (3) was re-estimated using the POLS method (Column (1)) and the regression method with fixed effects (Column (2)). However, estimates using POLS or fixed effects regression cannot consistently estimate the regression coefficients in the presence of endogeneity due to dynamic Model structures (Blundell & Bond, 1998). Therefore, Model (3) was re-estimated by the system GMM method, which is capable of estimating the regression coefficients consistently in the presence of endogeneity. The results are presented in Column (3). Columns (4) and (5) present the estimation results of Model (3) by the system GMM method with the variable Female replaced by Female_% and Female_dummy. Estimated results in all Columns in Table 6 show that L.Levit has a positive and statistically significant correlation of 1% with Levit. This means that the dynamic Model is valid. However, with dynamic Model structure, estimates by the POLS and fixed effects regression methods for Model (3) presented in Columns (1) and (2) are inconsistent. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 89 Columns (3), (4), and (5) present the estimates for Model (3) by the system GMM method with the variables Female, Female_%, and Female_dummy, respectively. The Arellano- Bond test statistic for series correlation shows that the second-order series correlation, AR (2), does not exist. Therefore, the authors chose lag variables of order 3 onwards as instrumental variables. This eliminates the correlation between the instrumental variables and the error and remedies the endogeneity problem completely. The Hansen J-test statistic in all three Columns is not statistically significant, i.e., the instrumental variables are not correlated with the error term, suggesting that the endogeneity problem is fixed (Wintoki, Linck, & Netter, 2012; Roodman, 2009). The results in Columns (3), (4), and (5) show that, in general, the quality of the board of directors has an impact on the firm's financial leverage decisions, with two out of four variables reflecting the characteristics of the board of directors having statistically significant coefficients. Specifically, the number of board meetings during the year is negatively correlated and statistically significant in the Model at the 5% and 10% levels with the variables Female and Female_dummy. This implies that enterprises with a large number of BOD meetings per year often have low financial leverage. Conversely, leverage ratios are higher in businesses where BODs are less active. Vafeas (1999) suggests that the number of board meetings demonstrates a positive effect on the performance of the board and is a good representation of managerial oversight. Therefore, the research results show that companies with more-active BODs often control their leverage at a lower level than firms with less-active boards. The frequency of board meetings reduces the debt ratio, indicating the effort of the BOD in actively monitoring financial operations (Anderson et al., 2004). This result is similar to that of Stephanus et al. (2014) and Vafeas (1999). The gender diversity of the BOD, whether defined by the number of female directors, the proportion of female directors, or the presence of female directors, is negatively correlated and remains statistically significant at the 5% level in all three specifications. These results are consistent with the studies of Harris (2014) and Abobakr and Elgiziry (2015). The negative correlation between board gender diversity and a firm’s capital structure can be explained by the characteristics and role of women directors. Specifically, female directors are often believed to have a different way of thinking and a different attitude toward risk than their male counterparts (Adams & Funk, 2012). Female directors are considered stricter in supervisory work (Abobakr, & Elgiziry, 2015; Adam, & Ferreira, 2009), have a more conservative attitude in risk assessment (Jiankoplos, & Bernasek, 1998), and typically have behavior and roles similar to independent directors (Adams, de Haan, Terjesen, & van Ees, 2015). Based on the characteristics that previous studies have pointed out, female directors are more likely to be stricter and more cautious in decisions regarding capital structure than their male counterparts. This may, in part, explain why firms with female members on the board of directors usually have lower leverage ratios. Contrary to previous studies, the size of the BOD and CEO duality were found to have no impact on a firm’s leverage ratio. This result differs from those of Berger et al. (1997), Anderson et al. (2004), Abor (2007), Bokpin and Arko (2009), Kyereboah- Coleman and Biekpe (2006), Rose et al. (2013) and Phan et al. (2017). Hoang Mai Phuong and Nguyen Thanh Hong An 90 Table 5. Regression results for static Models Variable POLS_static (1) Random_static (2) Fixed_static 1 (3) Fixed_static 2 (4) Fixed_static 3 (5) Lbsize -0.0090 [0.650] -0.0040 [0.820] -0.0010 [0.960] -0.0040 [0.860] -0.0020 [0.930] Lmeet 0.0050 [0.420] -0.0030 [0.530] -0.0060 [0.400] -0.0060 [0.410] -0.0060 [0.400] Female -0.0120*** [0.000] -0.0050 [0.210] -0.0020 [0.730] Female_% -0.0250 [0.530] Female_dummy -0.0050 [0.710] Ceodual 0.0080 [0.370] 0.0000 [0.955] -0.0020 [0.880] -0.0020 [0.890] -0.0020 [0.890] Lfsize 0.0420*** [0.000] 0.0780*** [0.000] 0.1010*** [0.000] 0.1010*** [0.000] 0.1010*** [0.000] Fixed_assets -0.2150*** [0.000] -0.1300*** [0.000] -0.1040** [0.030] -0.1040** [0.030] -0.1040** [0.030] ROA -0.7720*** [0.000] -0.4970*** [0.000] -0.4720*** [0.000] -0.4720*** [0.000] -0.4720*** [0.000] Liquidity -0.0310*** [0.000] -0.0190 [0.000] -0.0170*** [0.000] -0.0170*** [0.000] -0.0170*** [0.000] Salegrowth -0.0030 [0.100] 0.0020* [0.060] 0.0020** [0.040] 0.0020** [0.040] 0.0020** [0.050] Constant -0.1330* [0.090] -0.9120*** [0.000] -1.4910*** [0.000] -1.4920*** [0.000] -1.4880*** [0.000] Year dummies Yes Yes Yes Yes Yes Industry dummies Yes Yes No No No Number of observations 1,592 1,592 1,592 1,592 1,592 R2 0.4950 0.3853 0.3330 0.3330 0.3330 F statistic (or Wald statistic) 49.7260 811.6100 9.2830 9.4630 9.2930 P-value 0.0000 0.0000 0.0000 0.000 0.0000 Notes: *, **, and *** correspond to the 10%, 5%, and 1% levels of significance, respectively. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 91 The analysis is performed on 1,592 observations collected from 199 nonfinancial companies listed on the Ho Chi Minh City Stock Exchange from 2012 to 2019. Lev is the dependent variable, defined as the ratio of total debt to total assets of the business. Bsize is the number of board members. Meet is the number of board meetings per year. Female is the number of female members of the Board of Directors. Female_% is the percentage of female directors. Female_dummy is an indicator variable equal to 1 if the board has female members, and 0 if not. Ceodual is an indicator variable equal to 1 if the chairman of the BOD also holds the position of CEO, and 0 if not. Lfsize is the logarithm of the firm's total book value. Fixed_assets is the ratio of fixed assets to total assets. ROA is the ratio of after-tax profit to total assets. Liquidity is the ratio of short-term assets to short- term liabilities. Salegrowth is the rate of sales growth. Column (1) presents the regression results for Model (1), using the POLS regression method. Column (2) presents the regression results for Model (2) using the regression method with random effects. Column (3) presents the regression results for Model (2), using the regression method with fixed effects. Column (4) presents the regression results for Model (2), using the regression method with fixed effects and variable Female_% replacing variable Female. Column (5) presents the regression results for Model (2), using the regression method with fixed effects and variable Female_dummy replacing variable Female. The test statistic of the LM test for random effects is Chibar-square(1) = 22.8 (Prob(chibar-square)=0.000), indicating that the Model with random effects (Column 2) is more efficient than the POLS Model (Column 1). The test statistic for the Hausman test is Chi-square(16)=1055.3 (Prob(chi-square)=0.000), indicating that the Model with fixed effects (Column 3) is consistent and more efficient than the Model with random effects (Column 2). In general, the Model used for analysis is the fixed effects Model. Model (1) is structured as follows: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽3𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽4𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽5𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽6𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽7𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽8𝑅𝑂𝐴𝑖𝑡 + 𝛽9𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽10𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜀𝑖𝑡 (4) Model (2) is structured as follows: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽3𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽4𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽5𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽6𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽7𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽8𝑅𝑂𝐴𝑖𝑡 + 𝛽9𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽10𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (5) Table 6. Regression results for dynamic Models Variable POLS-dynamic (1) Fixed-dynamic (2) GMM-Dynamic 1 (3) GMM-Dynamic 2 (4) GMM-Dynamic 3 (5) L.Lev 0.820*** [0.000] 0.410*** [0.000] 0.752*** [0.000] 0.759*** [0.000] 0.705*** [0.000] Lbsize -0.010 [0.330] -0.012 [0.480] -0.006 [0.820] -0.021 [0.430] 0.001 [0.980] Notes: *, **, and *** correspond to the 10%, 5%, and 1% levels of significance, respectively. Hoang Mai Phuong and Nguyen Thanh Hong An 92 Table 6. Regression results for dynamic Models (cont.) Variable POLS-dynamic (1) Fixed-dynamic (2) GMM-Dynamic 1 (3) GMM-Dynamic 2 (4) GMM-Dynamic 3 (5) Lmeet -0.002 [0.520] -0.002 [0.660] -0.019** [0.040] -0.015 [0.012] -0.014* [0.090] Female -0.003 [0.130] -0.001 [0.840] -0.018** [0.020] Female_% -0.103** [0.020] Female_dummy -0.032** [0.030] Ceodual 0.005 [0.270] -0.003 [0.710] -0.010 [0.340] -0.008 [0.420] -0.009 [0.410] Lfsize 0.012*** [0.000] 0.078*** [0.000] 0.044*** [0.000] 0.045*** [0.000] 0.036*** [0.000] Fixed_assets -0.069*** [0.000] -0.120*** [0.000] -0.080* [0.070] -0.076* [0.090] -0.081** [0.050] ROA -0.283*** [0.000] -0.415*** [0.000] -0.218** [0.050] -0.226** [0.050] -0.347*** [0.000] Liquidity -0.008*** [0.000] -0.013*** [0.000] -0.001 [0.510] -0.001 [0.680] -0.002 [0.340] Salegrowth -0.001 [0.420] 0.001 [0.330] 0.003 [0.190] 0.002 [0.240] 0.003 [0.160] Constant -0.083** [0.040] -1.221*** [0.000] 0.000 [.] -0.709*** [0.000] 0.000 [.] Year dummies Yes Yes Yes Yes Yes Industry dummies Yes No Yes Yes Yes Number of observations 1,393 1,393 1,393 1,393 1,393 R2 0.875 0.448 - - - F statistic (or Wald statistic) 453.598 34.288 29,205.64 15,832.75 30,388.74 P-value 0.000 0.000 0.000 0.000 0.000 Number of instruments 87.000 87.000 87.000 Arellano-Bond AR(1) (P-value) 0.000 0.000 0.000 Arellano-Bond AR(2) (P-value) 0.630 0.562 0.562 Hansen J (P- value) 0.288 0.312 0.312 Notes: *, **, and *** correspond to the 10%, 5%, and 1% levels of significance, respectively. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 93 The analysis is performed on 1,592 observations collected from 199 nonfinancial companies listed on the Ho Chi Minh City Stock Exchange from 2012 to 2019. The use of lagged variables (1 period) reduces the number of observations to 1393. Lev is the dependent variable, defined as the ratio of total debt to total assets of the business. L.Lev is lagged one period from Lev. Bsize is the number of board members. Meet is the number of board meetings per year. Female is the number of female members of the Board of Directors. Female_% is the percentage of female directors. Female_dummy is an indicator variable equal to 1 if the board has female members, and 0 if not. Ceodual is an indicator variable equal to 1 if the chairman of the BOD also holds the position of CEO, and 0 if not. Lfsize is the logarithm of the firm's total book value. Fixed_assets is the ratio of fixed assets to total assets. ROA is the ratio of after-tax profit to total assets. Liquidity is the ratio of short-term assets to short-term liabilities. Salegrowth is the rate of sales growth. Column (1) presents the regression results for Model (3), using the POLS method. Column (2) presents the regression results for Model (3) using the regression method with fixed effects. Column (3) presents the regression results for Model (3) using the system GMM method. Column (4) presents the regression results for Model (3) using the system GMM method and variable Female_% replacing variable Female. Column (5) presents the regression results for Model (3) using the system GMM method and variable Female_dummy replacing variable Female. Model (3) is structured as follows: 𝐿𝑒𝑣𝑖𝑡 = 𝛽1 + 𝛽2𝐿. 𝐿𝑒𝑣𝑖𝑡 + 𝛽3𝐿𝑏𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽4𝐿𝑚𝑒𝑒𝑡𝑖𝑡 + 𝛽5𝐹𝑒𝑚𝑎𝑙𝑒𝑖𝑡 + 𝛽6𝐶𝑒𝑜𝑑𝑢𝑎𝑙𝑖𝑡 + 𝛽7𝐿𝑓𝑠𝑖𝑧𝑒𝑖𝑡 + 𝛽8𝐹𝑖𝑥𝑒𝑑_𝑎𝑠𝑠𝑒𝑡𝑠𝑖𝑡 + 𝛽9𝑅𝑂𝐴𝑖𝑡 + 𝛽10𝐿𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦𝑖𝑡 + 𝛽11𝑆𝑎𝑙𝑒𝑔𝑟𝑜𝑤𝑡ℎ𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (6) 5. CONCLUSIONS This study examines the impact of BOD characteristics on the financial leverage ratio of nonfinancial companies listed on the Ho Chi Minh City Stock Exchange for eight years from 2012 to 2019. The dataset includes 199 companies with 1,592 observations. The system GMM used to analyze the data helps avoid correlation between the instrumental variables and the errors, overcomes the endogeneity problem, and estimates the regression coefficients consistently. The research results show that, in general, the characteristics of the BOD have an impact on a firm’s financial leverage decisions, with two out of four variables reflecting the characteristics of the BOD being statistically significant. Specifically, the number of meetings of the BOD during the year has a negative and statistically significant coefficient. This shows that companies with more active BODs, as indicated by the number of meetings per year, often control their leverage at a lower level than firms with less active boards. In addition, board gender diversity, whether defined by the number of female directors, the proportion of female directors, or the presence of female directors, is negatively correlated with financial leverage and remains statistically significant at 5% in all three specifications. This can be explained by a higher level of risk aversion among women, which influences board decisions regarding the level of borrowing, resulting in Hoang Mai Phuong and Nguyen Thanh Hong An 94 lower debt ratios compared to boards without female directors. The size of the BOD and CEOs with dual roles were found to have no impact on the leverage ratio. This result complements and follows previous studies on the influence of governance structure on corporate capital structure, which is the basis for providing suggestions to managers on operating and managing their businesses. ACKNOWLEDGMENTS This research is funded by a Dalat University research grant. REFERENCES Abobakr, M., & Elgiziry, K. (2015). The effects of board characteristics and ownership structure on the corporate financial leverage. Accounting and Finance Research. 5(1), 1-14. Abor, J. (2007). Corporate governance and financing decisions of Ghanaian listed firms. Corporate Governance, 7(1), 83-92. Adam, R. B., & Ferreira, D. (2009). Women in the boardroom and their impact on governance and performance. Journal of Financial Economics, 94(2), 291-309. Adams, R., & Mehran, H. (2003). Is corporate governance different for bank holding companies? Economic Policy Review, 9(1), 123-142. Adams, R., & Funk, P. (2012). Beyond the glass ceiling: Does gender matter? Management Science, 58(2), 219-235. Adams, R., de Haan, J., Terjesen, S., & van Ees, H. (2015). Board diversity: Moving the field forward. Corporate Governance: An International Review, 23(2), 77-82. Al-Najjar, B. (2011). The determinants of the frequency of board meetings: Evidence from categorical analysis. Journal of Applied Accounting Research, 13(2) 178-190. Anderson, R. C., Mansi, S. A., & Reeb, D. M. (2004). Board characteristics, accounting report integrity, and the cost of debt. Journal of Accounting and Economics, 37(3), 315-342. Berger, P. G., Ofek, E., & Yermack, D. L. (1997). Managerial entrenchment and capital structure. Journal of Finance, 4, 1411-1438. Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data Models. Journal of Econometrics, 87, 115-143. Bokpin, G. A. & Arko, A. C. (2009). Ownership structure, corporate governance, and capital structure decisions of firms: Empirical evidence from Ghana. Studies in Economics and Finance, 26(4), 246-256. Boyd, B. (1995). CEO duality and firm performance: A contingency Model. Strategic Management Journal, 16(4), 301-312. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 95 Bradley, M., Jarrell, G., & Kim, E. H. (1984). On the existence of an optimal capital structure: Theory and evidence. Journal of Finance, 39, 857-878. Brick, I., & Chidambaran, N. K. (2007). Board meetings, committee structure, and firm value. Journal of Corporate Finance, 16(4), 533-553. Buchdadi, A. D., Ulupui, I. G. K. A, Dalimunthe, S., Pamungkas, B. G., & Fauziyyah, Y. (2019). Board of director meeting and firm performance. Academy of Accounting and Financial Studies Journal, 23(2), 1-7. Castanias, R. (1983). Bankruptcy risk and optimal capital structure. Journal of Finance 38, 1617-1635. Đặng, T. Q. A., & Quách, T. H. Y. (2014). Các nhân tố ảnh hưởng tới cấu trúc vốn của các doanh nghiệp niêm yết tại Sở giao dịch chứng khoán Thành phố Hồ Chí Minh. Tạp chí Phát triển và hội nhập, 18(28), 34-39. Demsetz, H. (1983). The structure of ownership and the theory of the firm. Journal of Law & Economics, 26(2), 375-390. Eluyela, D. F., Akintimehin, O. O., Okere, W., Ozordi, E., Osuma, G. E., Ilogho, S. O., & Oladipo, O. A. (2018). Board meeting frequency and firm performance: Examining the nexus in Nigerian deposit money banks. Heliyon, 4(10), 1-14. Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control. Journal of Law & Economics, 26(2), 301-325. Fosberg, R. (2004). Agency problems and debt financing: Leadership structure effects. Corporate Governance, 4, 31-38. Francis, B., Hasan, I., & Wu, Q. (2015). Professors in the boardroom and their impact on corporate governance and firm performance. Financial Management, 44(3), 547-581. Friend, I., & Hasbrouck, J. (1985). Determinants of capital structure. Philadelphia, USA: Rodney L. White Center for Financial Research Publishing. Friend, I., & Lang, L. H. P. (1988). An empirical test of the impact of managerial self- interest on corporate capital structure. Journal of Finance, 43(2), 271-281. Haque, F., Arun, T. G., & Kirkpatrick, C. (2011). Corporate governance and capital structure in developing countries: A case study of Bangladesh. Applied Economics, 43, 673-681. Harris, C. (2014). Women directors on public company boards: Does a critical mass affect leverage? Business and Economics Faculty Publications, 29, 1-18. Jensen, M. (1986). Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review, 76(2), 323-329. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 306-360. Jiankoplos, N. A., & Bernasek, A. (1998). Are women more risk averse? Economic Inquiry, 36, 620-630. Hoang Mai Phuong and Nguyen Thanh Hong An 96 Johl, S. K., Kaur, S., & Cooper B. J. (2015). Board characteristics and firm performance: Evidence from Malaysian public listed firms. Journal of Economics, Business and Management, 3(2), 239-243. Kajananthan, R. (2012). Effect of corporate governance on capital structure: Case of the Sri Lankan listed manufacturing companies. Researchers World, 3(4), 63. Kyereboah-Coleman, A., & Biekpe, N. (2006). Corporate governance and financing choices of firms: A panel data analysis. South African Journal of Economics, 74(4), 670-681. Lê, T. D., Bùi, Đ. T., & Lê, T. H. (2020). Nhân tố ảnh hưởng đến cấu trúc vốn của doanh nghiệp ngành thực phẩm trên thị trường chứng khoán. Tạp chí Tài chính. Retrieved from cau-truc-von-cua-doanh-nghiep-nganh-thuc-pham-tren-thi-truong-chung-khoan- 325679.html. Liao, L. K., Mukherjee, T., & Wang, W. (2015). Corporate governance and capital structure dynamics: An empirical study. The Journal of Financial Research, 38(2), 169-191. Lipton, M., & Lorsch, J. (1992). A modest proposal for improved corporate governance. The Business Lawyer, 48(1), 59-77. Long, M., & Malitz, I. (1985). The investment-financing nexus: Some empirical evidence. Midland Corporate Finance Journal, 3, 53-59. Lương, T. T. D., Phạm, N. U. N., Nguyễn, T. T., Nguyễn, T. K. L, Nguyễn, T. T. T., & Phạm, B. T. (2020). Nhân tố ảnh hưởng đến cấu trúc vốn của doanh nghiệp logistics trên thị trường chứng khoán Việt Nam. Tạp chí Tài chính. Retrieved from doanh-nghiep-logistics-tren-thi-truong-chung-khoan-viet-nam-325694.html. Marinova, J., Plantenga, J., & Remery, C. (2015). Gender diversity and firm performance: Evidence from Dutch and Danish boardrooms. The International Journal of Human Resource Management, 27(15), 1777-1790. Modiglani, F., & Miller, M. H. (1958). The cost of capital, corporate finance and theory of investment. The American Economic Review, 48(3), 261-297. Modiglani, F., & Miller, M. H. (1963). Corporate income taxes and the cost of capital: A correction. The American Economic Review, 53(3), 433-443. Morrison, A. M., White, R. P., & Velsor, E. V. (1987). The narrow band. Leadership in Action, 7(2), 1-7. Nguyễn, T. (2008). Các nhân tố ảnh hưởng đến cấu trúc tài chính của các doanh nghiệp chế biến thủy sản Khánh Hòa. Tạp chí Khoa học-Công nghệ Thủy sản, 3, 54-59. Nguyen, T., Locke, S., & Reddy, K. (2015). Ownership concentration and corporate performance from a dynamic perspective: Does national governance quality matter? International Review of Financial Analysis, 41, 148-161. DALAT UNIVERSITY JOURNAL OF SCIENCE [ECONOMICS AND MANAGEMENT] 97 Nguyễn, T. H., Trần, D. T., Nguyễn, H. N. H., Võ, T. T. N., & Nguyễn, T. T. (2016). Quản trị công ty và quá trình điều chỉnh động cấu trúc vốn - quan sát từ các doanh nghiệp niêm yết trên thị trường chứng khoán Việt Nam. Tạp chí Khoa học Đại học Mở, 50(5), 25-40. Ntim, C. G., & Osei, K. A. (2011). The impact of corporate board meetings on corporate performance in South Africa. African Review of Economics and Finance, 2(2), 83-103. Phan, B. G. T., Trần. Đ. T., & Trần, T. T. A. (2017). Ảnh hưởng của đặc điểm Tổng giám đốc điều hành đến hiệu quả hoạt động của doanh nghiệp. Tạp chí Khoa học Đại học Mở, 55(4), 51-63. Roodman, D. (2009). How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal, 9(1), 86-136. Rose, C., Munch-Madsen, P., & Funch, M. (2013). Does board diversity really matter? Gender does not, but citizenship does. International Journal of Business Science and Applied Management, 8(1), 15-27. Schwartz-Ziv, M. (2017). Gender and board activeness: The role of a critical mass. Cambridge University Press, 52(2), 751-780. Sheikh, N. A., & Wang, Z. (2012). Effects on corporate governance on capital structure: Empirical evidence from Pakistan. Corporate Governance, 12(5), 629-641. Simpson, W. G., & Gleason, A. E. (1999). Board structure, ownership, and financial distress in banking firms. International Review of Economics & Finance, 8(3), 281-292. Stephanus, R. W., Anastasia, K. N. F. T., & Toto, R. (2014). The effect of corporate governance on capital structure in public listed companies in Indonesia. Abstract of Emerging Trends in Scientific Research, 1, 1-16. Tarus, D. K., & Ayabei, E. (2016). Board composition and capital structure: Evidence from Kenya. Management Research Review, 39(9), 1056-1079. Titman, S., & Wessels, R. (1988). The determinants of capital structure choice. The Journal of Finance, 43(1), 1-19. Tran, T. T. T., Hoang, H. L., & Tran, T. H. Y. (2015). Relationship between gender diversity on boards and firm's performance - Case study about ASEAN banking sector. International Journal of Financial Research, 6(2), 150-159. Trương, Đ. L., & Võ, K. T. (2008). Các yếu tố ảnh hưởng đến cấu trúc vốn của các công ty cổ phần niêm yết trên thị trường chứng khoán Việt Nam. Tạp chí Nghiên cứu kinh tế, 361, 21-26. Vafeas, N. (1999). Board meeting frequency and firm performance. Journal of Financial Economics, 53(1), 113-142. Vũ, T. N. L., & Nguyễn, T. D. (2013). Các nhân tố ảnh hưởng tới cấu trúc vốn của các doanh nghiệp tại Tập đoàn Dầu khí Việt Nam. Tạp chí Kinh tế & Phát triển, 193, 23-28. Hoang Mai Phuong and Nguyen Thanh Hong An 98 Wen, Y., Rwegasira, K., & Bilderbeek, J. (2002). Corporate governance and capital structure decisions of the Chinese listed firms. Corporate Governance: An International Review, 10(2), 75-83. Wintoki, M. B., Linck, J. S., & Netter, J. M. (2012). Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics, 105(3), 581-606. Wiwattanakantang, Y. (1999). An empirical study on the determinants of the capital structure of Thai firms. Pacific-Basin Finance Journal, 7, 371-403. Wooldridge, J. M. (2002). Econometric analysis of cross section and panel data. Cambridge, UK: MIT Press Publishing.