Mối quan hệ giữa đặc điểm hội đồng quản trị và thông tin bất cân xứng của các công ty niêm yết trên sở giao dịch chứng khoán TP HCM tt tiếng anh

MINISTRY OF EDUCATION AND TRAINING STATE BANK OF VIETNAM BANKING UNIVERSITY HO CHI MINH CITY PHAN BUI GIA THUY RELATIONSHIP BETWEEN BOARD'S CHARACTERISTICS AND ASYMMETRIC INFORMATION OF LISTED FIRMS ON HO CHI MINH STOCK EXCHANGE SUMMARY OF PHD DISSERTATION Major: Finance and Banking CODE: 34 02 01 Academic advisor: DR NGO VI TRONG DR NGUYEN TRAN PHUC HO CHI MINH CITY, 2020 CONTENTS CHAPTER INTRODUCTION 1.1 Research problem 1.2 Research objectives 1.3 Research questions 1.4 Object and scope of this study 1.5 Research methodology 1.6 The significance of study 1.7 The structure of study CHAPTER LITERATURE REVIEW 2.1 Asymmetric information 2.2 Asymmetric information in the stock market 2.2.1 Definition 2.2.2 Basis for measurement 2.2.3 Measurement method 2.3 Review of studies 2.3.1 Research on asymmetric information measurement models 2.3.2 Research on relationship between board's characteristics and asymmetric information 2.3.3 Discussion of research gap 11 CHAPTER RESEARCH METHODOLOGY 13 3.1 Asymmetric information measurement models 13 3.1.1 Glosten and Harris (1988) model 13 3.1.2 George, Kaul and Nimalendran (1991) trade-indicator model 13 3.1.3 George, Kaul and Nimalendran (1991) serial covariance model 14 3.1.4 Kim and Ogden (1996) model 14 3.2 Adoption of asymmetric information measurement model 16 3.3 Research framework 16 3.4 Research hypothesis 17 3.5 Research method 18 3.5.1 Data collection 18 3.5.2 Definition and measure of variables 19 3.5.3 Regression analysis 19 CHAPTER EMPIRICAL RESULTS AND DISCUSSION 22 4.1 Measuring asymmetric information 22 4.1.1 Degree of asymmetric information 22 4.1.2 Degree of asymmetric information for individual stock 25 4.2 Board's characteristics and asymmetric information 29 CHAPTER CONCLUSIONS AND POLICY IMPLICATIONS 35 5.1 Highlights of study 35 5.1.1 Asymmetric information measurement 35 5.1.2 Board's characteristics and asymmetric information 35 5.2 Policy implications 36 5.2.1 Policymakers 36 5.2.2 Listed firms 36 5.2.3 Investors 36 5.3 Limitations and further research 36 5.3.1 Limitations 36 5.3.2 Further research 37 5.4 Conclusion 38 LIST OF AUTHOR'S PUBLICATIONS 39 CHAPTER INTRODUCTION 1.1 Research problem 1.1.1 Academic context Asymmetric information and measuring it on the stock market is the field that attracts a lot of attention from researchers because of the importance and topicality of this research field Firstly, the measure of asymmetric information provides the basis for assessing the information environment of the stock market in a specific nation (Affleck-Graves et al., 1994; Chakravarty et al., 2005; Huang, 2004; Lai et al., 2014) Secondly, the measure of asymmetric information is useful in assessing the effectiveness of policies issued by market regulators (Berkman and Lee, 2002; Chiyachantana et al., 2004; Frijns et al., 2008) Thirdly, this depends on the method of choice used, which is strongly influenced by the characteristics of the trading market or the context of a specific country (Van Ness et al., 2001; De Winne and Majois, 2003) Finally, the measure of asymmetric information has still updated and added to this day For example, Johnson and So (2018) have introduced a new approach, and this approach is thought to be superior to the traditional one On the other hand, asymmetric information is the main cause of the agency problem (Jensen and Meckling, 1976) and the solution to this problem requires a corporate governance mechanism under international practice, namely that board An effective board can reduce the agency cost, increase shareholder value, disclose information transparency, and limit risk of information (Kanagaretnam et al., 2007; Chen et al., 2007; Rutherford and Buchholtz, 2007) According to Zahra and Pearce (1989), Nicholson and Kiel (2004), Hilb (2012), one of the important factors contributing to the effective function of a board is the board's characteristics Many studies focused on the relationship between the characteristics of board and asymmetric information in many different aspects; such as board size, non-executive directors, diversity of women, educational qualification of board, duality, and ownership However, the results of the impact of the board's characteristics on asymmetric information are still controversial and inconsistent due to differences in a specific country, research period, and research method In addition to the inconsistent research results, the studies that mention the diversity of the board, including women on board and the educational qualifications of directors that are likely to impact asymmetric information is still limited Besides, not many studies have examined the effects of independent performance and the knowledge level of board members on asymmetric information under a different type of firm Specifically, the number of studies that compare the effectiveness of independent directors and the educational qualifications of board members in restricting asymmetric information in state-owned firms compared with those in non-state-owned firms is limited Moreover, many researchers are less interested in the nonlinear relationship between directorial ownership and asymmetric information 1.1.2 Practical context The world has witnessed many scandals involving information asymmetry Typically, the fraudulent actions of Enron's company (US) in 2001 through financial reporting fraud led to the company filing bankruptcy, pushing more than seven thousand employees into unemployment Shareholders who invested in Enron's share investment deal suffered losses of $74 billion, resulting in a total loss Recently, the accounting scandal at Toshiba Company (Japan) in 2015 caused a series of losses First, the scandal caused trust among investors and customers, overshadowing a brand with 140 years of corporate history Secondly, Toshiba's share value declined by 17% despite the Nikkei 225 index increasing by 7.6% and a series of senior leaders in the Board of Directors resigning (Hass et al., 2018) Vietnam, a country whose stock market is ranked by international organizations as a frontier market (SSC, 2018; HOSE, 2019), therefore the losses in the stock market due to rumors, frauds of financial statements, and non-transparency in disclosing information are very serious Although the authorities are always eager to improve the publicity, transparency of information and compliance with corporate governance of listed firms by issuing Decree 71/2017/ND-CP replaces Circular 121/2012/TT-BTC providing guidance on corporate governance applicable to public companies, and increase the effectiveness of inspection, supervision, and enforcement through the issuance of Decree 145/2016/ ND-CP amending and supplementing a number of articles of Decree 108/2013/ND-CP stipulates the sanctioning of administrative violations in the field of securities, but mistakes in the field of securities Stock always complicated developments According to UBCKNN (2016, 2017, and 2018), the cases of administrative sanctions have increased steadily in recent years For example, in 2016, there were 133 administrative violations In 2017, the number increased by 162% or 217 cases, and in 2018 the number of violations was 397 cases, up 14% From the facts, we can identify the common points of the securities scandals: (i) before the stock market price hit the floor or plunged into panic, there were sharp increases in prices, and (ii) the stock market price dropped significantly in the long run before the authorities issued a warning or had more effective control measures As long as there are some limitations in regulatory oversight, delays in warnings and sanctions, and ineffective implementation of the Board's role and functions, Vietnam's information environment is still poor transparency, hiding the severe degree of asymmetric information among investors Therefore, it is imperative that a preliminary assessment and then a proposal for mechanisms to minimize the information risk of listed firms on Vietnam's stock market be urgent Based on the disagreements about the results of the research, the research gaps, the urgency and current events from the practice, and the quantitative limitations of the relevant studies, measuring asymmetric information and estimating the relationship between the board's characteristics and asymmetric information of listed firms on the Ho Chi Minh Stock Exchange (HOSE) need to be researched and implemented 1.2 Research objectives The overall objective of the thesis is to focus on studying the impact of the board's characteristics on asymmetric information of listed firms in Vietnam, based on which the thesis recommends several policies to limit asymmetric information The thesis aims to achieve specific objectives as follows: - Measuring the level of asymmetric information of listed firms on HOSE - Identifying the suitable asymmetric information measurement model in the context of Vietnam - Identifying and measuring the factors related to the board's characteristics likely to impact on asymmetric information - Recommending many policies to decline the degree of asymmetric information 1.3 Research questions In order to achieve the current research objective, the study seeks to investigate the following questions: - How is the trend of the level of asymmetric information of listed firms on HOSE over the years? - Which is an appropriate econometric model applied to measure asymmetric information in the context of Vietnam? - Are there the factors related to the board's characteristics likely to impact on asymmetric information? - Does the impact of non-executive independent directors and the education level of board members on asymmetric information depend on the type of firm with stateowned firms? - Is there a nonlinear relationship between the shareholding ratio of board members and asymmetric information? 1.4 Object and scope of this study The thesis mainly focuses on the factors related to the board's characteristics likely to affect asymmetric information of listed firms on HOSE from 2009 to 2015 1.5 Research methodology The study uses quantitative methods to achieve the research objectives Research methods are as follows: To answer the first research question, the thesis applies econometric models to measure asymmetric information The measurement result of asymmetric information during the research period will indicate the change in the level of asymmetric information of listed firms on HOSE over the years To answer the second research question, the study checks the similarity of models, estimates the correlation between variables including the levels of asymmetric information using different models and the determinants, such as growth opportunity, liquidity trading, and debt ratio, and tests the change in the level of asymmetric information before and after the period of adjusting price limit range To answer the third research question, the thesis uses the econometric method to perform the regression equation of determinants of the board's characteristics affecting asymmetric information To answer the fourth research question, the study performs the testing of the interaction effects between the type of firm and independent directors and the education level of board members To answer the final research question, the thesis uses the threshold regression model and the piecewise linear regression The research results will be the basis for the implication of a number of policies to limit the degree of asymmetric information of listed firms in Vietnam 1.6 The significance of study Several main contributions of this thesis are as follows: Firstly, the study proposes an appropriate asymmetric information estimation model that can be used to measure asymmetric information in the context of Vietnam Secondly, the study provides evidence that expanding the price limit range from 5% to 7% is likely to increase the level of asymmetric information Finally, the study also adds some outstanding empirical pieces of evidence on the relationship between the board's characteristics and asymmetric information Specifically, the effects of independent directors and the educational qualifications of board members on asymmetric information depend on the moderating of the type of firm Besides, there is a threshold value of 1.74% of the shareholding ratio of the board members at which this ratio tends to impact on asymmetric information in different forms 1.7 The structure of study The structure of the study is organized in five chapters Accordingly, the chapters have the following layout: Chapter 1: Introduction Chapter 2: Literature review Chapter 3: Research methodology Chapter 4: Empirical results and discussion Chapter 5: Conclusions and policy implications CHAPTER LITERATURE REVIEW 2.1 Asymmetric information In general, asymmetric information exists when the amount of information regarding a potential transaction varies from one group of investors to another (Watts and Zimmerman, 1986) 2.2 Asymmetric information in the stock market 2.2.1 Definition Asymmetric information of listed firms on the stock market reflects an object or group of objects that own superior firm-specific information related to future public disclosures not available to uninformed (Chae, 2005) 2.2.2 Basis for measurement The basis for measuring asymmetric information is based on trading gains, expected revenues and losses, and order flow signals 2.2.3 Measurement method 2.2.3.1 Benchmark price method Asymmetric information is measure by taking a difference between the transaction price and the benchmark price of a traded stock (Venkatesh and Chiang, 1986; Lee, 1993; Huang and Stoll, 1996) Benchmark price is the price at which buyers or sellers easily accept transactions At relative levels, the midpoint of the bid-ask spread can represent the benchmark price The larger the deviation between the transaction price and the midpoint equals the higher asymmetric information The advantage of this method is convenient and quick estimating of information problem for individual stock; therefore, Chung et al (2010), Armstrong et al (2014) prefer this method However, the disadvantage of this is that it does not measure the specific degree of asymmetric information, takes a long time to estimate information risk of the overall market, and exaggerates the level of asymmetric information 2.2.3.2 Econometric method The econometric method separates the adverse selection component from the bid-ask spread to measure the level of asymmetric information Some studies that apply this method to 26 observations that the ASCGH satisfies is low The highest value was 24 observations in 2012, and the lowest was in 2010 Meanwhile, classification ASCGKN1 , ASCGKN2 ASCKO hardly reduces the number of observations Next, Table 4.6 below describes the degree of deviation between the measured ASC for the sample and the measured ASC for each stock applied according to research models Table 4.6 ASC for the sample and individual stock Year 2015 2014 2013 2012 2011 2010 2009 Panel A Results of measuring ASC for the sample ASCGH 72.2% 63.3% 79.0% 77.9% 79.1% 78.5% 89.2% ASCGKN1 71.3% 71.7% 75.5% 69.9% 56.7% 61.8% 58.8% ASCGKN2 65.4% 62.4% 75.0% 69.6% 53.7% 63.0% 65.6% ASCKO 65.5% 62.5% 75.0% 70.0% 53.9% 63.1% 65.7% Panel B Results of measuring ASC sorted for individual stock ASCGH 51.6% 50.5% 73.0% 56.8% 61.2% 51.6% 71.2% ASCGKN1 69.5% 73.3% 69.8% 63.5% 52.4% 60.9% 56.4% ASCGKN2 63.6% 68.7% 64.3% 59.7% 49.8% 57.0% 55.2% ASCKO 63.7% 68.7% 64.4% 60.1% 50.1% 57.2% 55.4% Panel C Deviation between ASC for the sample and ASC sorted for individual stock ∆ASCGH 20.6% 12.8% 6.0% 21.1% 17.9% 26.9% 17.9% ∆ASCGKN1 1.8% -1.7% 5.7% 6.5% 4.3% 0.9% 2.4% ∆ASCGKN2 1.7% -6.2% 10.7% 9.9% 3.9% 6.0% 10.4% ∆ASCKO 1.8% -6.3% 10.6% 9.9% 3.7% 5.9% 10.3% Source: Transaction data of listed firms on HOSE The results from Panel C in Table 4.6 show that, by applying the GH model, the deviation between asymmetric information for the sample and asymmetric information for each stock is relatively high; especially in 2010, 2012 and 2015 the deviations are above 20% (26.9%, 21.1% and 20.6%, respectively) The next high deviations were 17.9% in 2009 and 12.8% in 2014, and the lowest accounted for 6.0% in 2013 Meanwhile, by applying the remaining models, these deviations are relatively low Especially, by applying the GKN trade-indicator model, this deviation calculated by absolute value is less than 6.5% and ranges from 0.9% to 6.5% over the years In addition, by using the GKN serial covariance model and KO, the deviations in absolute value are also lower than 6.5% in 2010, 2011, 2014, and 2015 However in 2009, 2012, and 2013, they ranged from 10.3% to 10.7% In other words, by using two the serial covariance models, these deviations calculated by absolute value range from 1.7% to 10.7% 27 In summary, the results from Tables 4.5 and 4.6 show that measure of asymmetric information for individual stock applying GH model is not reliable enough because the number of observations is significantly reduced, and the difference between asymmetric information for the sample and for each stock is relatively high Figure 4.2 below presents the trend of changing in the level of asymmetric information according to GKN trade-indicator model, GKN serial covariance model, and KO model before and after sorted over the years 80% 75% 70% 65% 60% 55% 50% 45% 2009 2010 2011 unsorted GKN trade-indicator unsorted GKN serial covariance unsorted KO 2012 2013 2014 2015 sorted GKN trade-indicator sorted GKN serial covariance sorted KO Source: Transaction data of listed firms on HOSE Figure 4.2 Changing in adverse selection component before and after sorted Figure 4.2 shows that, the values of adverse selection component for each stock applying three models, including the GKN trade-indicator model, GKN serial covariance model, and KO model tend to fluctuate together In addition, using the GKN trade-indicator model, the trend line of ASCGKN1 for each stock is very close to that for the sample Meanwhile, using the other two models, the distance between these two trend lines has a wider range Table 4.7 below presents a correlation matrix showing correlation coefficients between variables, including the levels of asymmetric information, such as ASCGKN1 , ASCGKN2 , and ASCKO and the determinants, such as Depth: liquidity trading, TobinQ: growth opportunity, Bank: bank loan ratio, and Debt: total debt ratio, thereby forming the basis for proposing the suitable asymmetric information measurement model in the context of Vietnam 28 Table 4.7 Correlation coefficients between variables Variables n ASCGKN1 ASCGKN2 ASCKO ASCGKN1 1102 ASCGKN2 1102 0.837 *** 1102 0.834 *** 0.999 *** -0.193 *** *** -0.084 *** 0.205 *** 0.203 *** ASCKO Depth 1102 TobinQ 1102 Bank 1102 Debt *** 1102 Depth TobinQ Bank 0.191 *** -0.074 ** -0.123 ** -0.090 -0.043 -0.078 1 -0.135 *** -0.042 *** -0.076 ** 0.077 ** 0.056 * -0.221 *** -0.232 *** 0.724 *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level Source: Transaction data of listed firms on HOSE Depth: liquidity trading; TobinQ: growth opportunity; Bank: bank loan ratio; Debt: total debt ratio Table 4.7 shows that correlation coefficients between ASCGKN1 , ASCGKN2 and ASCKO are strong (over 83.4%), and these are statistically significant at the 1% significance level In particular, the correlation between ASCGKN2 and ASCKO is so high (99.9%), leading ASCGKN2 and ASCKO almost have the same fluctuation trend Besides, the correlation coefficients between Depth, Bank, and Debt and ASCGKN1 are negative and statistically significant at the 1% significance level, while TobinQ is positively correlated with ASCGKN1 and statistically significant at the same level Based on the degree of correlation and the statistical significance level, the determinants, including Depth, Bank, Debt, and TobinQ impact on ASCGKN1 have a better level of interpretation than those impact on ASCGKN2 and ASCKO Next, Table 4.8 below presents the results of comparing and testing the difference in ASCGKN1 , ASCGKN2 , ASCKO and Depth between two periods of adjusting price limit range from 5% to 7% Table 4.8 Degree of asymmetric information under different price limit range ASCGKN1 ASCGKN2 ASCKO Depth DumYear = (n = 432) 56.47% 53.81% 54.04% 0.07% DumYear = (n = 670) 69.01% 64.05% 64.23% 0.04% *** *** *** 4.722*** Variables t-test *** -18.873 -13.378 -13.404 significant at the 1% level; ** significant at the 5% level; * significant at the 10% level Source: Transaction and disclosure data of listed firms on HOSE DumYear: dummy variable specifies the period adjusting the price limit range, where DumYear = 0: period for which the price limit range is 5% and DumYear = 1: period for which the price limit range is 7%; Depth: liquidity of trading, ttest: statistics according to Satterthwaite-Welch test 29 Table 4.8 shows that the policy of widening the price limit range from 5% to 7% not only increases the level of asymmetric information measured by three different models but also reduces the stock liquidity Based on the results in Table 4.6, Table 4.7, Table 4.8, and Figure 4.2, it can be determined that using the GKN trade-indicator model to estimate asymmetric information is more appropriate than using the GKN serial covariance model and KO model 4.2 Board's characteristics and asymmetric information Before estimating the regression equation, this study will perform tests, including testing heterogeneous variance and autocorrelation in the residuals In addition, this study also examined the correlation coefficients among the independent variables and the VIF (Variance Inflation Factor) to ensure that the regression model is not severely multicollinear Table 4.9 below shows the result of the White test with Prob Chi-Square is less than 5%, which shows that the regression model has a problem of heterogeneous variance in the residuals Table 4.9 Heterogeneous variance test White test F-statistic 1.798 Obs*R-squared 189.849 Prob F(115,903) 0.000 Prob Chi-Square(115) 0.000 Source: Transaction and disclosure data of listed firms on HOSE Next, Table 4.10 below shows the result of the Breusch-Godfrey test with Prob ChiSquare is less than 5%, which shows that the regression model has a problem of autocorrelation in the residuals Table 4.10 Autocorrelation test Breusch-Godfrey test F-statistic 51.098 Prob F(2,1002) 0.000 Obs*R-squared 94.312 Prob Chi-Square(2) 0.000 Source: Transaction and disclosure data of listed firms on HOSE Finally, Table 4.14 below shows the correlation matrix among the study variables The results show that, in terms of magnitude, the correlation coefficient between the personal shareholding ratio of directors (Own) and a dummy variable indicating the type of firm (Gov) has a maximum value of 0.592 In addition, VIF index has the highest value of 1.77, leading the multi-collinear problem in the regression model is not significant -0.013 0.072 0.038 0.011 0.075 -0.147 -0.042 0.181 0.076 -0.040 0.091 0.294 0.108 0.058 0.058 0.073 -0.030 -0.082 -0.035 -0.182 -0.083 0.136 -0.057 -0.025 0.244 0.482 (2) 0.160 (1) 0.094 0.027 0.024 -0.154 0.088 0.041 -0.100 -0.029 -0.115 -0.316 -0.001 -0.094 (3) -0.010 0.080 -0.096 0.036 0.072 0.094 -0.098 -0.178 0.111 0.083 0.087 (4) (5) 0.061 0.280 0.077 -0.031 0.107 0.104 -0.152 0.116 -0.103 0.010 -0.097 0.013 -0.107 0.126 0.025 -0.017 0.075 -0.152 0.238 (6) -0.026 0.155 0.001 0.253 -0.040 -0.004 -0.012 -0.592 (7) -0.075 -0.099 0.042 -0.151 0.019 0.006 -0.055 (8) -0.138 -0.145 -0.005 0.102 -0.125 0.015 (9) -0.119 0.158 -0.096 -0.121 0.338 (10) -0.040 0.062 -0.051 -0.178 (11) 0.065 0.107 -0.104 (12) 0.006 0.357 (13) 0.131 (14) (15) 1.10 1.52 1.27 1.20 1.19 1.26 1.12 1.68 1.77 1.20 1.15 1.09 1.18 1.16 VIF ASCGKN: adverse selection component proxies asymmetric information; BoardSize: number of board members; Outd: the proportion of independent non-executive directors; Gender: the proportion of women on the board; Edu: the proportion of board members with postgraduate education; Dual: Chairman also holds the position of CEO; Own: the personal shareholding ratio of board members; Gov: dummy variable indicates state-owned firms; Depth: liquidity trading; Volatility: volatile stock price; Opp: the level of growth opportunity; Bank_St: short-term bank loan ratio; Bank_Lt: long-term bank loan ratio; FirmSize: natural logarithm of total assets; DumYear: dummy variable indicates the period of adjusting price limit range; Industry: dummy variables indicate the industry sectors Source: Transaction and disclosure data of listed firms on HOSE (15) DumYear (14) FirmSize (13) Bank_Lt (12) Bank_St (11) Opp (10) Volatility (9) Depth (8) Gov (7) Own (6) Dual (5) Edu (4) Gender (3) Outd (2) BoardSize (1) ASCGKN Variables Table 4.11 Correlation matrix 30 31 Table 4.12 below presents the regression results on the relationship between the board's characteristics and asymmetric information Table 4.12 Regression results ASCGKN Variables [1] [2] ** Constant 0.203 BoardSize 0.008 ** Outd Gender 0.170 0.008 ** -0.021 -0.021 0.020 0.016 Edu -0.016 Dual -0.003 [3] ** -0.081 *** -0.076 *** 0.013 -0.002 ** -0.098 Gov -0.014 -0.012 -0.062 *** - - 0.089 *** Gov*Edu - - -13.987 ** Depth -0.107 - -15.287 ** -0.101 0.008 ** 0.009 ** -0.020 -0.020 0.022 ** -0.003 ** -0.056 ** 0.081 *** - -14.136 ** 0.180 ** 0.212 -0.066 0.001 *** [6] ** 0.025 -0.011 Own Gov*Outd -0.092 0.008 ** -0.016 0.001 0.205 [5] ** 0.008 ** 0.017 -0.012 ** 0.236 [4] *** -0.098 0.001 ** -0.093 ** -0.029 * -0.026 * - - 0.068 -15.455 ** -0.058 ** ** -14.386 ** 0.064 * -15.812 *** -0.008 *** -0.008 *** -0.008 *** -0.008 *** -0.008 *** -0.008 *** 0.040 *** 0.040 *** 0.040 *** 0.040 *** 0.041 *** 0.041 *** Bank_St -0.078 *** -0.077 *** -0.078 *** -0.078 *** -0.079 *** -0.076 *** Bank_Lt -0.135 *** -0.144 *** -0.141 *** -0.149 *** -0.132 *** -0.142 *** FirmSize 0.027 *** 0.029 *** 0.027 *** 0.029 *** 0.027 *** 0.029 *** DumYear 0.112 *** 0.111 *** 0.111 *** 0.110 *** 0.112 *** 0.111 *** Industry - Yes - Yes - Yes 1019 1019 1019 1019 1019 1019 35.7% 36.5% 35.9% 36.7% 35.8% 36.6% Volatility Opp Observations R adj Durbin-Watson 1.761 Breusch-Pagan 714.207 Hausman *** significant at the 1% level; 14.662 ** 1.768 *** 754.079 1.765 *** 16.718 692.782 14.937 1.773 *** 734.024 1.763 *** 16.743 716.126 14.636 1.771 *** 753.240 *** 16.770 * significant at the 5% level; significant at the 10% level Source: Transaction and disclosure data of listed firms on HOSE ASCGKN: adverse selection component proxies asymmetric information; BoardSize: number of board members; Outd: the proportion of independent non-executive directors; Gender: the proportion of women on the board; Edu: the proportion of board members with postgraduate education; Dual: Chairman also holds the position of CEO; Own: the personal shareholding ratio of board members; Gov: dummy variable indicates state-owned firms; Depth: liquidity trading; Volatility: volatile stock price; Opp: the level of growth opportunity; Bank_St: short-term bank loan ratio; Bank_Lt: long-term bank loan ratio; FirmSize: natural logarithm of total assets; DumYear: dummy variable indicates the period of adjusting price limit range; Industry: dummy variables indicate the industry sectors The regression results in Column [1] and [2] in Table 4.12 show that, there is a positive and significant relationship between number of board members (BoardSize) and asymmetric information, and the directorial shareholding (Own) negatively and significantly impacts on asymmetric information This study does not find the impact of independent directors 32 (Outd), female directors (Gender), education level of board members (Edu), and duality (Dual) on information problem Next, the results in Column [3] and [4] present that independent directors in the state-owned firms reduce asymmetric information less efficiently than those in the non-state-owned firms Finally, the results in Column [5] and [6] show that the type of firm changes the strength of a relationship between the educational qualification of board members and asymmetric information Table 4.13 below presents the regression results of the impacts of board' characteristics on asymmetric information within two different groups, group of state-owned firms and non-state-owned Table 4.13 Regression results for different type of firms ASCGKN  Gov Variables ASCGKN  Non-Gov [1] [2] Constant 0.157 0.140 0.203 ** 0.163 * BoardSize 0.010 * 0.010 ** 0.010 * 0.010 * Outd Gender [3] [4] -0.007 -0.008 -0.060 ** -0.053 ** 0.004 0.004 0.043 0.029 Edu -0.002 Dual -0.012 0.004 -0.064 -0.009 *** Own -0.196 Depth -17.847 * -0.168 ** 0.017 ** -18.018 ** -0.103 -0.051 * 0.018 ** -10.933 * -0.099 ** -12.745 * -0.009 *** -0.009 *** -0.009 *** -0.009 *** 0.045 *** 0.044 *** 0.034 ** 0.034 *** Bank_St -0.115 *** -0.112 *** Bank_Lt *** *** Volatility Opp -0.139 -0.146 -0.023 -0.207 -0.016 ** -0.201 *** FirmSize 0.029 *** 0.030 *** 0.029 *** 0.031 *** DumYear 0.117 *** 0.116 *** 0.094 *** 0.091 *** Industry - Yes - Yes Observations 682 682 337 337 40.1% 41.4% 28.9% 29.4% R adj Durbin-Watson 1.854 Breusch-Pagan 345.077 Hausman *** significant at the 1% level; 1.877 *** 365.446 15.019 ** 1.717 *** 16.404 86.446 12.380 1.739 *** 83.804 *** 13.400 * significant at the 5% level; significant at the 10% level Source: Transaction and disclosure data of listed firms on HOSE ASCGKN  Gov: adverse selection component for state-owned firms; ASCGKN  Non-Gov: adverse selection component for nonstate-owned firms; BoardSize: number of board members; Outd: the proportion of independent non-executive directors; Gender: the proportion of women on the board; Edu: the proportion of board members with postgraduate education; Dual: Chairman also holds the position of CEO; Own: the personal shareholding ratio of board members; Depth: liquidity trading; Volatility: volatile stock price; Opp: the level of growth opportunity; Bank_St: short-term bank loan ratio; Bank_Lt: long-term bank loan ratio; FirmSize: natural logarithm of total assets; DumYear: dummy variable indicates the period of adjusting price limit range; Industry: dummy variables indicate the industry sectors 33 Column [1] and [2] in Table 4.13 show that, regression results for the subsample with state-owned firms are not different from those for the sample (see Column [1] and [2] in Table 4.12) Specifically, the board size has a positive impact on asymmetric information while the shareholding ratio of board members has an opposite effect Regression results in Column [3] and [4] in Table 4.13 show that the board size positively impacts on asymmetric information while the directorial ownership negatively effects Besides, in the subsample with non-state-owned firms, independent directors, and education level of board members negatively impact on asymmetric information These results support the hypothesis that board members performing independently and having a high level of education are likely to limit the asymmetric information Next, Table 4.14 below shows the results of testing the threshold of shareholding of board members affecting asymmetric information Table 4.14 Threshold value of board's personal shareholding Threshold Test Threshold Value vs 1** Regime Impact on ASCGKN Observations Own < 0.0174 ∂ASCGKN = 2.637∗∗∗ ∂Own  Own < 0.0174 388 Own ≥ 0.0174 ∂ASCGKN = -0.081∗∗∗ ∂Own  Own ≥ 0.0174 631 τ = 0.0174 vs *** significant at the 1% level; ** significant at the 5% level; * significant at the 10% level Source: Transaction and disclosure data of listed firms on HOSE ASCGKN: adverse selection component proxies asymmetric information; Own: the personal shareholding ratio of board members; τ: threshold value of the shareholding ratio of board members Table 4.14 shows that there is a threshold value τ = 0.0174 which is statistically significant at the 5% significance level, at which the shareholding of board members effects on asymmetric information in different signs Specifically, when Own < 1.74%, the impact of Own on ASCGKN is positive and statistically significant at the 1% significance level (2.637; p < 0.01) When Own ≥ 1.74%, this impact is negative and statistically significant and statistically significant at the 1% significance level (-0.081; p < 0.01) This result shows that there is a non-linear relationship between the shareholding of board members and asymmetric information To clarify more about this nonlinear relationship, Table 4.15 below shows the piecewise regression results according to the threshold segment of the shareholding of board members impacting on the asymmetric information 34 Table 4.15 Piecewise linear regression ASCGKN Variables [1] [2] ** Const 0.194 BoardSize 0.007 * Outd Gender ASCGKN  Gov 0.145 [3] * 0.007 * [4] [5] 0.138 0.101 0.180 0.009 * 0.009 * 0.008 [6] ** 0.139 * 0.008 * -0.019 -0.016 -0.004 0.000 -0.053 * -0.046 0.020 0.016 0.002 -0.001 0.050 0.037 Edu -0.015 -0.009 Dual -0.004 0.000 0.353 Own_Thr2 -0.104 ** -0.108 *** Gov -0.013 -0.011 -14.062 1.170 * Own_Thr1 Depth ASCGKN  Non-Gov ** -15.178 ** 0.000 0.006 -0.013 -0.009 0.643 1.468 -0.049 * 0.011 ** 3.207 -0.037 0.013 ** 3.069 * -0.225 *** -0.216 *** -0.125 *** -0.119 *** - - - - -18.011 * -17.820 ** -11.603 * -13.158 ** -0.008 *** -0.008 *** -0.009 *** -0.009 *** -0.009 *** -0.009 *** 0.041 *** 0.040 *** 0.045 *** 0.045 *** 0.035 *** 0.035 *** Bank_St -0.080 *** -0.082 *** -0.120 *** -0.124 *** Bank_Lt *** *** *** *** Volatility Opp -0.132 -0.137 -0.134 -0.137 -0.019 -0.186 -0.010 ** -0.184 ** FirmSize 0.028 *** 0.030 *** 0.030 *** 0.032 *** 0.027 *** 0.029 *** DumYear 0.112 *** 0.111 *** 0.117 *** 0.117 *** 0.096 *** 0.092 *** Industry - Yes - Yes - Yes 1019 1019 682 682 337 337 35.6% 36.6% 40.1% 40.7% 29.5% 29.9% Observations R adj Durbin-Watson Breusch-Pagan Hausman *** 1.762 714.205 14.926 significant at the 1% level; ** 1.771 *** 759.391 16.641 1.859 *** 340.504 14.565 1.875 *** 345.292 1.715 *** 17.395 93.521 12.266 1.737 *** 89.614 *** 13.291 * significant at the 5% level; significant at the 10% level Source: Transaction and disclosure data of listed firms on HOSE ASCGKN: adverse selection component proxies asymmetric information; BoardSize: number of board members; Outd: the proportion of independent non-executive directors; Gender: the proportion of women on the board; Edu: the proportion of board members with postgraduate education; Dual: Chairman also holds the position of CEO; Own_Thr1: the personal shareholding ratio of board members is less than 1.74%; Own_Thr2: the personal shareholding ratio of board members is more than 1.74%; Gov: dummy variable indicates state-owned firms; Depth: liquidity trading; Volatility: volatile stock price; Opp: the level of growth opportunity; Bank_St: short-term bank loan ratio; Bank_Lt: long-term bank loan ratio; FirmSize: natural logarithm of total assets; DumYear: dummy variable indicates the period of adjusting price limit range; Industry: dummy variables indicate the industry sectors The regression results in Table 4.15, especially in Column [2], [4], and [6] under controlling the effect of industry sectors, show that there is a threshold value of the shareholding of board members (coefficient of Own_Thr1 and Own_Thr2) at which the shareholding of board members tends to impact on asymmetric information in different forms This result supports the hypothesis that there is a non-linear relationship between the shareholding of board members and asymmetric information 35 CHAPTER CONCLUSIONS AND POLICY IMPLICATIONS 5.1 Highlights of study 5.1.1 Asymmetric information measurement Applying Glosten and Harris (1988) model to measure asymmetric information has certain limitations due to the assumptions of this model Besides, in general, the level of asymmetric information is applied according to three models: George, Kaul and Nimalendran (1991) trade-indicator model; George, Kaul and Nimalendran (1991) serial covariance model; and Kim and Ogden (1996) model should pay attention and tend to increase in the period from 2011 to 2015, the period of widening the price limit range, increasing from 5% to 7% Finally, based on the results of asymmetric information measurement; the correlation between the degree of asymmetric information and signaling determinants, including liquidity trading, debt ratio, and growth opportunity; and testing the change in the level of asymmetric information at two stages of adjusting the price limit range, it can be assumed that George, Kaul and Nimalendran (1991) trade-indicator model is the suitable model that can be applied for measuring asymmetric information in the context of Vietnam 5.1.2 Board's characteristics and asymmetric information There is a positive impact of the board size on asymmetric information Meanwhile, the independent and non-executive directors and the education level of board members have a negative impact on asymmetric information; however, this negative effect is only reflected in the group of state-owned firms In addition, the type of firm with state-owned firms changes the strength of the impact of the independence and educational qualification of board members on asymmetric information Moreover, there is a nonlinear relationship between the shareholding of directors and asymmetric information, which expressed through the threshold value of 1.74% of the shareholding ratio of board members affecting asymmetric information However, this study does not find the impact of female directors and duality on the level of asymmetric information, even considering the type of firm with state-owned and nonstate-owned firms 36 5.2 Policy implications 5.2.1 Policymakers Policymakers need to take more active steps in limiting asymmetric information through close supervision and sanctions that are sufficient to deter the listed firms from disclosing sufficient information Besides, if the solutions to ensure the stability of the stock market are not synchronized, the expansion of the price limit range should be considered 5.2.2 Listed firms Listed firms should not build a board of directors with more members because experience further shows that smaller boards allow for playing the better strategic role and are less likely to become rubberstamping entities In addition, the company not only aligns the interests between the company and directors through directorial ownership but ensures powers or incentives are suitable for responsibilities that board members have to take Moreover, shareholders in non-state-owned firms should appoint and/or reassign independent directors because these characteristics are important resources of the company in limiting information risks among investors In addition, state-owned firms need to facilitate favorable conditions for board members to fully take advantage of their capabilities and expertise in reducing asymmetric information 5.2.3 Investors Based on the results of this research, investors need to observe the characteristics of the stock market, firms, and, importantly, board of directors This observation is very useful because it helps investors to assess the overall level of information risk of listed firms and then re-allocate the proportion of stock investment in the investment portfolio effectively 5.3 Limitations and further research 5.3.1 Limitations There are some limitations to this study as follows: - This study uses closing data at the end of the day to measure asymmetric information due to the convenience of collecting data - This study has not yet mentioned the ability to apply other asymmetric information measurement models, such as the variance decomposition model (Hasbrouck, 1991) and the probability model of information-based trading (Easley et al., 1996) - The sample is not large enough, and the dataset has an unbalanced panel data format 37 - This study has not yet mentioned other characteristics of board, such as board's meetings (Becker-Blease and Irani, 2008; Elbadry et al., 2015), remuneration (Abad et al., 2017), and tenure (Yunos et al., 2012; Handajani et al., 2014) - This study has not yet mentioned that state-owned firms have different equity ownership structures Likely, the state-owned shareholding with the controlling rights (over 51%), the right to veto the General Meeting of Shareholders resolutions (over 36%), or the right to nominate candidates to the Board of Directors and Supervisory Board (over 10%) is likely to affect information risk or the relationship between the board's characteristics and asymmetric information - And finally, this study has not considered the endogenous factors that can occur in independent variables, such as the ownership ratio of board members (Demsetz and Lehn, 1985; Cho, 1998) and the independent directors (Armstrong et al., 2014; Goh et al., 2016) Specifically, there is an endogeneity problem because asymmetric information and liquidity trading are simultaneously determined (Kavajecz, 1999; Libby et al., 2002; Barakat et al., 2014) 5.3.2 Further research Here are some suggestions for further research: - The researchers could pay attention to the use of intraday data to measure asymmetric information - Besides, further research need to consider the use of other asymmetric information measurement models, such as the variance decomposition model (Hasbrouck, 1991), the probability model of information-based trading (Easley et al., 1996), and measuring asymmetric information based on event study method (Kanagaretnam et al., 2007; Barakat et al., 2014; Borisava and Yadav, 2015) - Further research should collect more data, and the dataset should have a balanced panel data format - In addition, the researchers need to study other characteristics of board, such as board's meetings, remuneration, and tenure that are likely to impact on asymmetric information - The relationship between the board's characteristics and asymmetric information should be studied within the scope of firms with different state-owned shareholding structures 38 - Finally, the researchers should consider the endogenous factors that can occur in independent variables, such as the shareholding ratio of board members and independent directors Specifically, there is an endogeneity problem involving the phenomenon that liquidity trading could jointly determine with asymmetric information 5.4 Conclusion The research results of the thesis reflect the level of asymmetric information of listed firms on HOSE, and the study has shown the George, Kaul, and Nimalendran (1991) tradeindicator model is suitable It is applied to measure asymmetric information in the context of Vietnam Moreover, the research results have reflected relatively clear about the relationship between the board's characteristics and asymmetric information It is useful and necessary for policymakers, listed firms, and investors to refer to these research results 39 LIST OF AUTHOR'S PUBLICATIONS BEING A PHD STUDENT Jounal of science: Phan Bui Gia Thuy, Nguyen Tran Phuc Ngo Vi Trong (2020) Recommending a model of asymmetric information measurement of listed firms on the Ho Chi Minh Stock Exchange Journal of Economics & Development (accepted on 29/06/2020) Phan Bui Gia Thuy, Nguyen Tran Phuc Ngo Vi Trong (2019) Independence and education of board of director and asymmetric information: The moderating effect of state-owned firms Journal of Economics & Development, No 268, Oct 2019, 33-42 Phan Bui Gia Thuy, Tran Đuc Tai Tran Thi Tu Anh (2017) The impact of Chief Executive Officer characteristics on firm performance Journal of Science Ho Chi Minh City Open University, 55(4), 58-66 Nguyen Van Ngai, Tran Thi Tu Anh Phan Bui Gia Thuy (2016) The level of asymmetric information: Evidence from listed firms on Ho Chi Minh stock exchange Journal of Science Ho Chi Minh City Open University, 1(46), 58-66 Conference proceeding: Phan-Bui, T.G., Tran, A.T.T., Nguyen, T.P & Ngo, T.V (2019) Identifying The Effect Of Board's Characteristics On Asymmetric Information The 6th Vietnam International Conference in Finance (VICIF-2019), 4-5 July 2019, Danang, Vietnam Phan, T.B.G., Nguyen, T.P & Ngo, T.V (2018) Signals of market and firm characteristics and asymmetric information Asia Conference on Business and Economic Studies, 8-9 September 2018, University of Economics Ho Chi Minh City, Vietnam ISBN: 978-604922-660-1 Phan, T.B.G & Ngo, T.V (2015) Effects of Non-Executive Directors on Firm Performance in Financial Crisis: Empirical Evidence in Vietnam The 2nd Vietnam International Conference in Finance (VICIF-2015), 4-5 June 2015, Ho Chi Minh City, Vietnam Funded project: Project Research on the effect of Chief Executive Officer charateristics on firm's performance Institution Position Year Status Nguyen Tat Thanh University Leader 2013 Accepted 40 BEFORE BEING A PHD STUDENT Jounal of science: Vo Hong Duc, Hoang Đinh Son Phan Bui Gia Thuy (2014) Identify factors affecting the remuneration of the Board of Director: Evidence from listed firms on Ho Chi Minh stock exchange Journal of Science Ho Chi Minh City Open University, 1(34), 13-26 Vo Hong Duc and Phan Bui Gia Thuy (2013) Corporate Governance and Firm’s Performance: Empirical Evidence from Vietnam Journal of Economic Development, No 218, 62-78 Vo Hong Duc and Phan Bui Gia Thuy (2013) The effect of duality, board's working experience, and the role of moderating of growth opportunity on firm's performance Journal of Science Ho Chi Minh City Open University, 3(31), 52-65 Vo Hong Duc and Phan Bui Gia Thuy (2013) The Impact of Female Board Members on Firms' Performance Banking Technology Review, No 85, 21-30 Vo Hong Duc and Phan Bui Gia Thuy (2013) The effects of board's characteristics on the firm performance: Evidence in Vietnam Journal of Economics & Development, No 188(II), 68-75 Conference proceeding: Vo, D.H and Phan, T (2013) Corporate Governance and Firm Performance: Empirical Evidence from Vietnam 42nd Australian Conference of Economists, July 7-10, 2013, Murdoch University, Perth, Western Australia ISBN: 978-921877-12-4 Book: Vo Hong Duc and Phan Bui Gia Thuy (2014) Corporate Governance: Theory and Control Mechanism Ho Chi Minh, Vietnam: Youth Funded project: Project Relationship between corporate governance and firm's performance Institution Position Year Status Ho Chi Minh City Open University Member 2013 Accepted ... relation to stock liquidity  Duality Duality, the Chairman holding the position of Chief Executive Officer, may affect the independence of board members Therefore, separating duality will limit the... The effect of duality, board's working experience, and the role of moderating of growth opportunity on firm's performance Journal of Science Ho Chi Minh City Open University, 3(31), 52-65 Vo... the type of firms, namely state-owned firms Control variables, including factors related to characteristics of market, such as liquidity trading, stock price volatility, growth opportunity, period