The impact of board gender diversity on firm’s risk and firm performance: Evidence of listed firms in vietnam

Nội dung text: The impact of board gender diversity on firm’s risk and firm performance: Evidence of listed firms in vietnam

1. Trường Đại học Kinh tế - Đại học Đà Nẵng THE IMPACT OF BOARD GENDER DIVERSITY ON FIRM’S RISK AND FIRM PERFORMANCE: EVIDENCE OF LISTED FIRMS IN VIETNAM GVHD: PGS.TS. Võ Thị Thúy Anh SVTH: Đỗ Phan Khánh Huyền, Trần Thị Nguyên Thảo Trường Đại học Kinh tế - Đại học Đà Nẵng huyendo.19071997@gmail.com ABSTRACT This paper investigate the effect of board gender diversity on firm performance in Vietnam's listed firms from 2006 to 2017. This paper report a negative and significant connection between board gender diversity and firm performance and show a evidence that there is no relationship between women and firm’s risk. The number of female members of the board always represents a negative relationship with the company's performance, whether it is measured by ROA or Tobin's Q. The results also suggest that there is no statistically significant relationship between gender diversity and firm risk. Board gender diversity includes the proportion of women in boardroom, chairwomen and female CEOs. For the the percentage of female directors, even if the result were significant, its effects on firm risk would be negligible. Having a female CEO or a female chairwoman don’t necessarily change the risk of a firm. Keywords: Board gender diversity, risk, performance, Vietnam. 1. Introduction It is believed that either an efficient firm or a given reasonable policy are influenced by an array of factors. These factors could be considered in many approaches : From the business perspective of firms such as its revenue targets, costs, policies of merger - acquisition (M&A), marketing or business management. In terms of strategic management, there are the future development judgments, short-term and long-term investment decisions, funding strategies. Regarding industry view, these factors are the policies of competitors in the same industry, customer needs or market access. When it comes to members of the Board Directors, the performance of the business is influenced by a lot of factors such as personality, age, education, work experience, gender A specific a male administrator and a female administrator with the same level of education, the same numbers of years of work experience, and the same age but when they deal with a problem, they would be able to offer different ways of recognizing and resolving. In order to manage the performance and the risk of a joint stock company, it is necessary to consider a lot of factors such as firm characteristics, ownership structure and board characteristics. This topic is really hot in Vietnam, especially gender issues in the Board of Directors. The literature about women in the board of directors and the performance, the risk of the firm in the current context of Vietnam is relatively much because this problem increasingly attracts much attention of scholars and economists participating in the research. This study was conducted in more than 300 non-financial companies listed on the Ho Chi Minh Stock Exchange (HOSE). From those reasons, the author chooses the topic “The impact of Board Gender Diversity on Firm’s risk and firm performance: Evidence of listed firms in Vietnam”. 2. Theoretical Basis And Methodology 2.1. Theoretical basis of board gender diversity and risk 16
2. Hội nghị Sinh viên nghiên cứu khoa học năm học 2018-2019 2.1.1. Differences in risk aversion It is found that when the risk aversion grew up, it leads to the differences in investment decisions by women (Hinz, McCarthy, & Turner, 1997). The similar story happens in the research of Watson & Robinson (2003) that enterprises seem to be less volatility in risk with women in board than those run by men. Psychological aspect of Risk aversion among Women. In one of Byrnes et. al (1999) analyses, they pointed out that males are inclined to take more risks even when it appeared clearly a bad idea to do so. A study of Hinz et al. (1997) by examining U.S. investors about their pension choices they concluded that even after had controlled for income and age, women generally chose less risky pension funds. Board attributes affect firm risk. According to previous empirical studies, there is only a limited number of board attributes which examined to influence firm risk. Cheng (2008), which also found that board size has a negative relation with variability of firm performance or firm risk by using US sample over the period 1996-2004. In general, it was implied that these factors are associated to firm risks. Governance gender is considered to be one factor of composition attributes which is affect significantly to firm performance and firm risk. 2.1.2. Do female directors affect firm risk? Perryman et al (2016) which examined how increases in gender diversity for top executives may influence risk-taking behaviors by the firm, performance of the firm, and gender-based compensation differences at the executive level. The findings suggest, paradoxically, that increasing gender diversity both reduces risk and increases performance. Harjoto et al. (2014) which also find that boards with diversity in terms of gender, race, age, experience, tenure and expertise are more risk-averse. Other extant studies have shown that board gender diversity is associated with better earnings quality (Srinidhi, Gul, & Tsui, 2011), lower likelihood of financial restatement (Abbott, Parker, & Presley, 2012), and less aggressive acquisition strategies (Levi, Li, & Zhang, 2013). In this aspect, the prior papers examining the relationship between board room gender and firm risk is still limited. The first could be mentioned is the research of Sila et. Al (2016), in this paper, after conducting a number of estimations, the concluded that there is no evidence illustrates the relationship between gender diversity and firm risks. Mathew et. al (2016) illustrated that if there are more women on the board, it leads to the lower level of risk. Another paper mentioned about the context of board of management, shows that banks run by female CEO are less risky than male CEO (Skala & Weill, 2017). The point is that there is a variety of papers focusing on the relation between the women in boardroom and firm performance, but there is a few number of studies regarding my topic. It could be a reason gave me motivation to find out this relationship between board gender diversity and firm risks. 2.1.3. Background in Vietnam According to Nguyen et al (2015), the development of boardroom system in Vietnam is still in the first steps (Nguyen, Locke, & Reddy, 2015), and the recent circumstance can be described as 5 points follows: (1) the regulations about board of directors are still less improved as reported by World Bank (2006); (2) awareness of corporate governance is still undeveloped (Freeman & Nguyen, 2006); (3) the power of state sector is seemed dominant other economic sectors ; (4) lacking of regulations in order to protect the private property rights; and (5) there are still limited governance mechanisms in both internal and external (Le & Walker, 2008; Nguyen D. C., 2008). After all of Vietnam efforts to improve the gender equality issue, the World Bank (2006 & 2011) assessed Vietnam as one of the countries in the world that had achieved the highest rate of economic participation by women and the highest participation of women in state power structures, such as parliament, in the East-Asian region. In general, both features (undeveloped corporate governance and gender equality issue) mentioned above make Vietnam become a considerable and worthy environment to conduct this examine. Diana et al (2010) paper found that women in Vietnam might be more reluctant to approach cutting-edge devices, write a loans, or conduct economic activities that might offer 17
3. Trường Đại học Kinh tế - Đại học Đà Nẵng higher expected returns but higher risks. They avoid situations that require them to be more competitive or that have less predictable outcomes. Another research (Do.H.L et al, 2016), it is proved by regression results that the emergence of women in board mitigates the risk of enterprises. Furthermore, the ratio of females in board influences business risk by an upside-down U-shape form. 2.2. Methodology of board gender diversity and firm risk The data in this paper is panel data which contains both time series and cross-sectional data, using a pooled OLS will not be appropriate. Therefore, either the random effects (RE) or the fixed effects (FE) model will be used. There is a general consensus that board characteristics are not exogenous variables but are endogenous (Sila, Gonzalez, & Hagendorff, 2016). The endogeneity may be driven by omitted unobserved factors (Sila, Gonzalez, & Hagendorff, 2016). The popular method in the literature to deal with unobservable variables is by using a fixed effects estimator. All the regression will be estimated by using robust standard errors to account for heteroscedasticity and clustering standard errors at the firm level (Petersen, 2009). The baseline model is the following forms: RISKi,t = α + β1WOMENi,t + β2FCEOi,t β3FCHAIRi,t β4Controlsi,t + εi,t 2.2.1. Dependent Variables: Risk-taking measurement Total risk (STDEV) is calculated as the standard deviation of daily stock returns over the last year Systematic risk (BETA) is a measure of a stock's volatility in relation to the market. It measures the exposure of risk a particular stock or sector has in relation to the market. To annualize standard deviations, we multiply total risk by the square root of 250. 2.2.2. Independent Variables: Governance gender diversity (WOMEN) which is defined as the number of female board members on the board divided by the number of all board members over the period of 2007- 2017. Female chair (FCHAIR) When there was a chairwoman appointed, it corresponds with 1 value, and in contrast, receives the value of 0 when company had no chairwoman. Female Chief Executive Officer (FCEO) CEO gender is also identified by hand-collected data from annual reports, detailed in Board of Management information. Control Variables Board and CEO characteristics Board size (boardsize) Board size is defined by the number of total members in boardroom. Wang (2011) findings consistently show that small boards are willing to ratify risky policy choices, take more risky projects, and have higher future firm risk. Board independence (memin) must be considered as the presence of independent directors can result in a more stakeholder-focused board, which could lead to a higher level of firm’s risk CEO tenure (lnceoexe) we also collect the length of CEO tenure and using the natural logarithm of the CEO’s years (Sila et.al, 2016). CEO level (ceolevel) The variable will be valid from 1 to 8 corresponding to the following educational levels: only high school graduate (1), graduated from further education (2), college graduate (3), with a degree university (4), with a master's degree (5), a doctorate degree (6), associate professor (7), associate professor-doctorate (8). Firm-level control variables. ROA is defined as the ratio of earnings before interest and taxes to total assets. Sales Growth (growth) calculated as the annual rate of growth of sales in log form. Research and development expenditure (R_and_D): This variable is measured by the exact value of R&D expenditure based on the balance sheets of listed companies. Capital expenditure (capex) “capex” is capital expenses divided by the total assets of the firm. 18
4. Hội nghị Sinh viên nghiên cứu khoa học năm học 2018-2019 Price-to-book value (PB) we use the price-to-book ratio as a firm's market to book value by dividing price per share by book value per share. Surplus cash (surplus cash) Surplus cash is net cash flow from operating activities less depreciation and amortization plus research and development expenditure divided by the book value of total assets. Firm size (logsize) which is measured by the the natural log of total assets. Firm age (firmage) This variable is measured by the number of years from the IPO date of each firm Firm leverage (leverage) It is indicated by the proportion of total liabilities on total assets of each separated company.Theoretical basis of board gender diversity and firm performance 2.2.3. Resource dependence theory and board gender diversity (Pfeffer and Salancik, 1978) found resource dependence theory, affirmed that a firms depend on their outermost resource to exist that would lead to potential risk for bussiness. In order to make smaller these endanger, firm should create a association to outward enviroments that supervise these sources. There are three beneficial sides of this connection such as advice and counsel, legitimacy, communication channels. Resource dependence theory are appilied by Hillman et al. (2007) to address the diversity of board room in US listed firms and assits that gender diversity build up these profits. 2.2.4. Agency theory, corporate governance and board gender diversity In corporate finance, the agency problem usually refers to a conflict of benefits between a firm's management and the firm's stockholders. Fama and Jensen (1983) believed that the effectiveness of board instructions and controlling are paramount importance in decreasing these disputes. The gender diversity impact on firm perfomance also rely on firms’ administration standard. In a firms with better manage policy, (Adam and Ferreira, 2009) assited that diversity of gender on board would be detrimental to firm value because of nonessential over-monitoring. By constrast, Gul et al. (2011) assumed that firms could not entirely deal with their incapacitated direction by having gender-diverse board. 2.2.5. Token status theory, sex-role stereotypes and board gender diversity The rareness of women or nonages in peak of governance as “tokens” and in extraordinary situation, “solos,” mentioning to person who is the unique exemplification of a specific demographic category Kanter (1977). Likewise, the previous paper on outlines evoke that individuals improve mental systems of the attributes for definite job containers (Lee and James 2007). (Powell and Butterfield, 2002) explain that since men take a prevelent role of the top governance spots, male occupation candidates are more probable than female candidate to be involed to those ascribes. The historical token status of women on board directors also supplement the cliche that women have inadequate inevitable ensigns for such rooms (Lee and James 2007). The critical mass theory on board gender diversity assits that “one is a token, two is a presence, and three is a voice” (Kristie, 2011). 2.2.6. Different forms of governance style between men and women Aspects of gender roles are related to differences leadership style in association with personality "Communal" and "Agentic". There are 3 pairs Management style involves debates about the relationship between diversity gender and leadership, executive: Style according to assigned tasks and style personal ways (Task-oriented styles and Interpersonally oriented styles); Democratic style and Autocratic style (Democratic styles and Autocratic styles); Transformation style and Resolution style (Transformational styles and Transactional styles). When synthesizing the leadership styles above, it can be assumed that the behavior of the home Female managers, when compared to male managers, will be more inclined to style Interpersonally, 19
5. Trường Đại học Kinh tế - Đại học Đà Nẵng Democratic and Transformational administration. Opposite, Male administrators tend to follow Task- oriented styles, Autocratic and Transactional. 2.2.7. Experimental studies comparing governance performance between men and women gender Most researchers reject common differences between sex when considering efficiency. In the study of Powell (1993) saw many similarities, he noted that "Men and women are not different in leadership performance, although some governance situations are more suitable for men than women. Bass (1990) find evidence that male managers are better evaluated than female managers but he explained this trend due to bias and stereotyped expectations available. To explain in a basic way, it is easier to understand the effectiveness of governance men and women, Eagly et al. (1995)‘s results shows male and women governance is equally effective, but consistent with the assumption that the combination of governance roles with gender, managers will improve operational efficiency. In general, there is no firm consensus in previous studies about the difference in management effectiveness for the gender difference. Another study came out in 2011 of Zenger and Folkman revealed that women can have a good vision of leadership as men. Research shows that women are better than men because of an important difference in 12/16 characteristics Folkman and Zenger found it most relevant to effective leadership. The benefits of women are not all limited in their field. The set of results shows that in each rank, women are marked better in comprehensive leadership issues compared to the male colleagues. 2.2.8. The research papers show a positive relationship (Huang and Kisgen 2013; Levi et al., 2014) illustrates women in the chief operating officer is more mindful than male in director. Firm with male executives are less likely to accquision and issue debt tools than companies with female executives. Men are more likely to leave the executive board position than women. Board members who are women, are more assiduous controller and in need of more adjusted endeavor than men one (Adams and Ferreira, 2009; Gul et al., 2008). Women have a way to behave differently from men. Female president of the board, female-like behavior executive is described in management theory. (Hillman et al., 2007). (Gul at al., 2011) acknowledged that board gender diversity have capable to compensate not entirely fragile firm management. Carter et al. (2003) found a marked and affirmative correlation between the amount of women administrators and firm performance. Campbell and Minguez-Vera (2008) also admitted the positive link between woman and effcient bussiness activities. Female on board are able to generate stockholder benefits through reducing bids and bid premium (Levi et al., 2014). In addition, the study shows that women are likely to increase the value of shareowner. Yu Liu (2008) also argues that the number of female members in the board directors of firms listed on the stock exchanges in China has a positive relationship with the firm’s performance. However, this relationship is stronger and clearer in private firms than state-owned companies. 2.2.9. The research papers show a negative relationship Ahern and Dittmar (2012) conduct research on the relationship between women in management board and the effectiveness of firms listed on the Norwegian Stock Exchange. The reason may be that, in Norway, executives are women who are younger and have less working experience than men. After increasing the number of women in the executive board increased, the number of projects acquired was much higher but most failed. Triana et al. (2013) demonstrates a positive and negative relationship between gender diversity in governance and the performance of a company. If a company is in a high performance state, the gender diversity in the board will help the company have a lot of changed strategies, so the company's performance is enhanced ans reversal. Since then, the double-edged blade of gender diversity in management can be seen. Bohren and Staubo (2014) argues that the quota imposition makes firms tend to cancel listing on the floor or close themselves to avoid re-structuring the board in accordance with the new rules, which are considered very expensive. The situation that the number of listed companies is less than non-listed fiems will cause the 20
6. Hội nghị Sinh viên nghiên cứu khoa học năm học 2018-2019 remaining companies on the floor to incur a large amount of this new regulation. Lee and James (2007) use the data of top executive announcements from 1990 to 2000 and the result was negative response of investors to CEO announcements is female male. Ryan and Haslam (2005) shows that during the period when the stock market declined, businesses which appoint female members to the board seem to have performance is worse when considering 5 months of activity compared to male member in board. 2.2.10. The research papers show no relationship Some research assumes that there is no impact in the relationship between gender diversity in the Board and business performance. Shrader et al. (1997) found no significant impact can be found in the relationship between rates of senior female management or female member of the board and its performance business (represented by ROA and ROE). Rose (2007) did not find any impact in the relationship between Tobin’s Q and the gender diversity in the board of businesses in Denmark. 2.2.11. The research is done in Vietnam Tran Dang Vinh Hao (2014) conducted researchdone in Vietnam in 2014 on this issue shows that the diversity of the gender of the board has no effect on the performance of firm listed on the stock market performance. Thuy Phan et al (2016) studied the impact of executive characteristics on the performance of firm in Vietnam. The results show that there is no relationship between the number of female members in the executive board and the performance of that company. Nguyen Manh Ha (2016) carried out the research topic "The relationship between corporate governance and the performance of banks in Vietnam”, the results of this study are similar to the study of Tran Dang Vinh Hao (2014), which is no correlation between the number of members in the board and the performance of banks. 2.2.12. The reason for the mixed opinions in previous studies The empractical evidence may partly be explained by other estimation methods and different from: data format, sample selection, research method, definition, methods of measuring and other quantitative methods. Some studies there is no control over factors. In addition, previous studies performed spread across different countries, different legal systems, so the ways it effects on the gender is different and the research stage is different, this is a significant reason for the opposite of results. 2.3. Methodology of board gender diversity and firm performance The author used fixed effect regression model, it is an extension of classical linear regression model. Estimated model used: Firm_Perfomancei,t = Board_Gender_ + Board_ + CEO_ + Firm_ + + The research use Fixed effect model (FEM) regression controlling for effect of year. The study also includes year fixed effects (λt) so as to control the fluctuations of economics on a yearly basis. 2.3.1. Percentage of the number of members who is female Measuring gender diversity by percentage of female members of the board (Adam and Ferreira, 2009; Ahern and Dittmar, 2012), Yu Liu (2014), Adams and Ferreira (2008). 2.3.2. Firm Performance This study evaluate the firm performace by using Return on Asset Beside, author use Tobin’s Q (Market cap and totoal liabilities divided by the book value of total assets) to show efficency operational firm performance. 21
7. Trường Đại học Kinh tế - Đại học Đà Nẵng 2.3.3. Variable group describes board characteristics Ceo_Chair (sign as Duality): This is the dummy variable that describes the cumbersome, if the executive and the chairman of the board are the same person, then the value is 1, if it is otherwise, the value is 0. Independent Members (sign as Indepent): Independent members of the board are the number of members who only participate in the board of directors who do not hold any executive office position. Female is the chairman of the board (sign as F_Board): This variable will have a value of 1 for companies with female board chairpersons and a value of 0 for companies with a male chairman. Board education level (sign as B_Level) The variable will be valid from 1 to 8 corresponding to the following educational levels: only high school graduate (1), graduated from further education (2), college graduate (3), with a degree university (4), with a master's degree (5), a doctorate degree (6), associate professor (7), associate professor-doctorate (8). Board experience (sign as B_exper): This variable is measured based on the number of years working and operating the company. 2.3.4. Variable group describes firm characteristics Firm size (sign as Size): The value of the variable is measured by the natural logarithm of the book value of the company's total assets. Firm age (sign as Age): It is measured by the number of years established by the firm, indicating how long the company has existed in the market. Leverage (sign as Lev): Debt ratio is determined by the total liabilities divided by the total assets of the firm. 2.3.5. Variable group describes CEO characteristics Female Chief Executive (sign as F_CEO): This variable describes the gender of the executive director, the variable will be equal to 1 if the executive is female and if they are the male, variable will have a value of 0. CEO experience (sign as C_Exper): This variable represents the working experience of the CEO. This variable is measured based on the number of years working and operating the company. CEO educational level (sign as C_Level): This is a measure of CEO's education level, which will be valid from 1 to 8, similary to Chair educational level. 3. Results And Discussions 3.1. Board gender diversity and firm’risk 3.1.1. Descriptive Statistics The average total risk measured by the standard deviation of daily stock return is 2.70% with a standard deviation of 0.86%. Moreover, there are only 3 observations with a standard deviation of daily stock return higher than 10% which are SVT, TIX, and TNC; and all of them are in 2017. The mean beta is 0.68 and the median is 0.64 which is fairly close to the mean. The mean proportion of women on board is 14.1%. The median is close to the mean at 14.3%. It can be seen from the data that the gender diversity on the board of directors is still low. For female CEO, the mean is 9.1% which implies that only 9.1% of all the firm in the data over the period 2007-2017 has a female CEO. The mean for female chair is a little higher at 12.6%. The first group of independent variables controls for the board characteristics of the firm including board size, the number of independent board members, CEO tenure, and the CEO’s education level. The size of the board of directors ranges from 3 to 11 in my sample. The average size of the board is 5.7, and the median is 5. 22
8. Hội nghị Sinh viên nghiên cứu khoa học năm học 2018-2019 The average number of independent members is 1.8, and the median is 2. This indicates that half of the firms in the sample have only 2 or fewer independent board members. Also, about 20% of the observation has zero independent board member. The tenure of CEO is also included in our data the average tenure is 1.445 which is equivalent to 4.24 years. The median tenure is 5 years. The longest tenure in the data is 27 years which is the case of TCM in 2008. The median education level is 4 which is equivalent to university degree level. Less than 5% of the CEO has an education lower than university level. On the other hand, about 30% has an education level of 5 or higher. It means that about 30% of CEOs have master or higher degrees. Over 67% of CEOs has only university degree. ROA, is used as a measure of the profitability of companies. The average ROA is 7% and the median is 5.3%. The standard deviation is 8.3% which is quite large given the mean is only 7%. For measuring growth, we use yearly sale growth. The average sale growth rate in the sample 27.6%, but the median is only 10.3%. 3.1.2. OLS Regression In all four models, model (iv) produces a statistically significantly negative relationship at 5% significance level between the proportion of women on board and firm risk, specifically systematic risk. Similarly, female CEO also shows a positive significant relationship with systematic risk at 5% significance level. However, the two significantly negative relationships disappear in model (3) after controlling for the year fixed effects. This suggests that the above significant relationship is potentially driven by unobserved factors which will be further investigated in FEMs below. In addition, female chair shows a significantly positive relationship with systematic risk in model (iii) at 1% significance level. Nevertheless, it’s entirely possible that this is just a spurious relationship. 3.1.3. Fixed Effects Regression While WOMEN and FCEO variables represent a negative relationship with Total risk and positive with Systematic risk, the other independent variable (FCHAIR) has shown the opposite result. However, unfortunately, all of the results in terms of relationship between gender diversity in board and both types of risk are insignificant. For the first group of control variables (board and management characteristic), there are only three significant relationships with 16 examined relationships by 4 models. Firstly, board size is negatively associated, at 10% of statistical significance, with STDEV in model (1) which included time effect control. Secondly, the variable named lnceoexe which indicating CEO experience, has a negative association with total risk in model (1) similarly. And the last significant results in this group is the positive relationship between ceolevel (illustrates the educational standard of CEO) and systematic risk in model (4). The coefficients of other variables in this group are statistically insignificant at any level of p-value. Interestingly, the variable of independent board members has no significant associations with both total risk and systematic risk. A comparison between OLS and FE models. The results from the FEMs suggest no relationship between board gender diversity and firm risk. Also, most of the coefficients in the model come up insignificant. The results may be driven by endogeneity between the independent variables in the models or unobserved firm heterogeneity. I perform a comparison between the FE models and the OLS models to see if this is the case. The only model that produce a significant relationship between the proportion of women and firm risk is OLS model (ii) and (iv) for systematic risk. Therefore, we will investigate between FE models and OLS models for systematic risk only. Table 4 show the 4 models on systematic risk from table 2 and table 3. Robustness tests 23
9. Trường Đại học Kinh tế - Đại học Đà Nẵng The reason for not finding a meaningful relationship between gender diversity and firm risk could be due to the misidentifications of the model. In this section, I’ll show that the results of my model are robust to various identifications of the models as well as alternative measures of risk. First of all, there could be endogeneity problems within the set of gender diversity (proportion of women, female CEO, female chair), and including all three of them could cause the regression estimator to give inaccurate results. Therefore, for each risk measure, we estimate the models again using a single variable or two of the three variables for gender diversity. However, when I conduct the robustness checks by examining the dependent variables one-by-one, there is no significant relationship found. Second, as in Sila et. al (2016), we include the lags of the dependent variables to see if a dynamic model can explain the link between gender diversity and firm risk. However, it is examined that even though I add these variables, the relationships between gender diversity variables and risk measures are still insignificant. Finally, we investigate the relationship between gender diversity and alternative measures of firm risk (R&D spending, and leverage). Sila et. al (2016) suggests that the differences in risk-taking behaviour due to gender may not always be reflected in a firm’s stock volatility Overall, although there is one significant relationship found, this test is still lack of evidence to prove that there is an existed relationship between gender diversity in board and firm risks. 3.2. Board gender diversity and firm performance 3.2.1. Summary statistics The results of descriptive statistics for the variables used in this study in the period 2010-2017 are presented in Table 3.1. Whereby: ROA and Tobin’s Q are two variables that measure company performance, , 3173 observations of Tobin’s Q and 3195 observing ROA. The average ROA of HOSE is 7.2%, while the company with the highest ROA is 7.83% and the lowest is -158.7%. The average number of female members in the management board of listed companies in Ho Chi Minh Stock Exchange is about 14%. In terms of Board Characteristics, on average, the management of each company listed on HOSE has 27.1% independent members, which is roughly the same as the research done by Yu Liu in China in 2014. The educational background of board members in Vietnam ranges from just graduated from high school to associate professor. However, most of the board members have bachelor's and master's degrees. In addition, the average chairman of the board of companies listed on the HOSE has about 6 years of experience operating the company, but there are also some companies with the chairman of the board with experience to do up to 35 years, and obviously there are companies too, its board chairman has not had any years of experience. And can be seen, about 36%, companies on the HOSE have duality situation. Regarding Firm Characteristics, most firms have natural logarithms of assets of about 27,525. The average financial leverage of companies is about 40%. The age of the firm on HOSE fluctuates from newly established to 17 years listed on the floor, of which the average of listed companies is from 5 to 6 years. When it comes to CEO characteristics, about 9.5% of firms have female CEOs, a big number. And, this figure is double the number of companies with women CEOs in the country, according to Yu Liu (2010). According to the data collected by the author, most CEOs in companies listed on HOSE are those with bachelor's and master's degrees. Up to now, the CEO with the longest working experience in Vietnamese joint stock companies is about 30 years. Currently, the number of companies with state-owned capital remains very high and the state is planning to reduce the rate of ownership of state capital. Therefore, at the time of the study, the total average state owned capital in enterprises was about 24.1%. In particular, there are businesses with almost no state capital contribution, and there are enterprises holding 96% of the state. 24
10. Hội nghị Sinh viên nghiên cứu khoa học năm học 2018-2019 3.2.2. Pearson correlation coefficients The absolute value of the correlation coefficients between independent variables lies in the range of from [-0,355 ; 0,46] shows that the variables are not highly correlated. Moreover the absolute value of the correlation coefficients <0.55. Therefore, according to gujarati's study (2008), it is difficult for multicollinearity to occur when we perform regression analyzes. 3.2.3. Regression result of board gender diversity and firm performance The performance of the firm and the number of female members in board of director have a significantly negative relationship. This result is consistent with the research results of Ahern and Dittmar (2012), Bohren and Staubo (2014) and Ryan and Haslam (2005). Huang and Kisgen 2013; Levi et al., 2014 considered that women make prudent decisions more than men. Women limit business risks, maximize shareholders' assets, reduce agency costs and improve operational efficiency. However, in Vietnam, women may be too cautious and require many adjustment, which is costly and time-consuming. Sometimes, this over-caution also removes good business opportunities that are waiting for businesses and can reduce the company's business performance. Besides, in Vietnam, women often do not like adventure, they do not like too much risk. However, in business, people have to trade between risk and profit, the higher the risk, the higher the profit. Risk avoidance, will make female managers ignore potential business projects and chance to maximize firm value. Independent members of the board of directors, will reduce the efficiency of the company from the perspective of market evaluation (Q). Normally, in Vietnam, the public often believes that those who only work on the board of director, do not have a comprehensive view of the case of the firm, so they will likely make decisions affect firm performance. Mike W. Peng (2004) also believes that independent members will increase agency costs, and they negatively affect the performance of the company. From an accounting perspective, the female chairman of the board of directors and the president's working experience will be able to calculate to buy properties at reasonable prices but still bring good quality of operation, improve product quality, and bring high revenue for businesses. The duality will make the power of the company focus on a single person, the CEOs will become arbitrary and make wrong decisions. The larger the size of the company, the less effective the company will be if it is in a market perspective (Q). When calculating the Tobin's Q index, the denominator is too large and lead to reduce this index However, from an accounting perspective (ROA), the company will gain economies of scale. This result is similar to that of (Lauterbach and Vaninsky, 1999). The formula for Tobin Q is (market capitalization + total debt) / total assets, a higher debt, it will make that firm value in the market increase. However, ROA, using too much debt, make the company reduce profits because part of the revenue must offset interest expenses. The longer the number of companies listed in the market, the older the firm's assets are, the firm has to spend a lot of money to re-invest in equipment, and its assets increase, but revenue still has to offset the loan interest, ROA is reduced. Women who work in board of director will help the company grow and access more customers and potential investors. Therefore, the market often values women as CEOs more than men. The results show that the older the CEO and the higher the education level members will reduce the performance of the company. Older CEOs tend to work in a traditional style, having more limited visions than young CEOs. Their experience comes from the newly established market, the characteristics of business operations also change, making them difficult to adapt and make the right decisions about the future direction for the company. There are many previous studies on the relationship between state ownership and company performance. Mei Yu (2013) provides evidence of a positive relationship between state ownership and operational efficiency. In Vietnam, the market often evaluates companies with state owned firms that are developed stablly and sustainably. In addition, social activities of the community also attract potential customers for the company, from which revenue growth and operational efficiency improvement. 25
11. Trường Đại học Kinh tế - Đại học Đà Nẵng 3.2.4. Robustness check From the results of the regression table, in any company with different ownership structure, the relationship between operational efficiency and women in the board of director is still a negative relationship. This relationship is always sustainable, although in companies with higher state ownership structure, the different the operating mechanism, the company has more concerns than focusing on improving profits such as social justice, equal rights, human rights, social responsibility, community responsibility. 3.3. Discussions The results in this paper illustrates that the negative effect of female directors on firm performance is stronger in state-controlled firms than in low person-controllded firms. This may be explained by the fact that the greater the percentage of state-owned companies, the more concentrated of power is on shareholders who represents the state. Therefore, the gender diversity in management, especially the more women, requires more care, adjustment and consideration many sides of issues. In a company with little state ownership, it seems that women are less effective than the women in high-owned company. Logically reasoning, the female directors in these companies are forced to spend a portion of their profits for other purposes, so the negative effects of female on this case is stronger than that of less state owned companies. The results of this study support the results of a series of previous studies on the relationship between gender diversity and company performance. Ahern and Dittmar (2012) , Triana et al. (2013) , Bohren and Staubo (2014), Thuy Anh Vo and Nha Khanh Bui (2017). In addition to gender diversity, the article also finds empirical evidence of the relationship between leverage, company size and the company's age to its performance. The proportion of women doesn’t affect firm risk; and even if it does, the effects would be negligible. Second, neither the appearance of a female CEO nor a female chair affects firm risk. The coefficients for both dummy variables are quite large for firm total risk compared to the proportion of women, unfortunately they aren’t significant. If all these relationships were significant, it may be argued that having a woman on board might not be as influential as having a woman holding an important position like a chair or an executive. Moreover, it would be puzzling that a female chair would increase the total risk while a female CEO would reduce it. Unfortunately, all coefficients for the two variables are insignificant, so we can’t make any concrete inference about the relationship between gender diversity and a firm risk here. 4. Conclusions This study enlarge the current literature on board diversity by bestowing empirical evidence of the effect of women directors on firm performance in Vietnam. The results show that the percent of women directors has a significant and negative impact on firm performance measured by Tobin's Q and Return on Assets. The negative impact essentially originates from the high rate of state-owned firms rather than from firm with a low rate of the state. The number of women directors is also important. We further document that the negative impact of board gender diversity is significant in firms controlled by low state owners and mainly controlled by state owners. The proportion of female directors doesn’t affect firm risk, and the economic impact would have been very small even if the results were significant. Having either a female CEO or female chair would have affected the risk profile of the firm much more than having more female directors if these relationships had been significant. The conclusion is that there is no evidence suggesting a link between the gender diversity in boardrooms and the Vietnamese firms’ risk, this is also consistent with the findings in Sila et. al (2016). The implication is that having a more gender diverse board is not necessarily associated with improvement in firm risk. Therefore, public policies that push for a greater gender diversity may not be beneficial for businesses in terms of risk. More studies should be done to examine this issue in greater detail before any regulations that forces more gender diversity in companies like gender quota are enacted 26