Share repurchase and cash dividend payout policy in vietnamese stock market substitute or complement

中 原 大 學 商 學 博 士 學 位 學 程 博 士 學 位 論 文 股 票 買 回 與 現 金 股 利 之 交 互 影 響 ： 替 代 或 互 補 ？ 越 南 之 實 證 鄧 中 堅 中 華 民 國 109 年 月 中原大學 商學博士學位學程 博士學位論文 股票買回與現金股利之交互影響： 替代或互補？ 越南之實證 Share Repurchase and Cash Dividend Payout Policy in Vietnamese Stock Market: Substitute or Complement? 指導教授：陳怡珮 研究生：鄧中堅 (Dang Trung Kien) 中華民國 109 年 月 摘要 由於過去文獻在股票買回與現金股利政策之替代或互補關係未有一致結論，本文以越南市 場進行驗證，探討股票買回與現金股利支付政策之關係。越南是亞太地區中快速發展的新 興市場，且具有股權集中度高、國家持股比例高等特色，而相較於傳統的現金股利制度， 在越南股票買回是較為新興的現金支付方式，且越為普遍，此議題希冀能利用越南實證的 特殊性豐富公司理財議題之文獻。 實證結果發現，現金股利變動率與股票買回規模存在非線性關係，但其中兩者高度顯著的 正向關係顯示在越南市場中以互補效果為主。接下來利用分量迴歸模型分析此非線性關 係，發現主要導因於股票買回規模的大小，股票買回規模的差異導致了股票買回和現金股 利之間的關係產生變化。亦即，股票買回規模越大，股票買回和現金股利之間的替代效果 越顯著。此外，越南上市公司進行股票買回決策時會以未來成長機會為主要考量。最後， 信號假說和自由現金流假設僅顯示出邊際顯著效果。 關鍵字：越南、股票買回、股利政策 i doi:10.6840/cycu202000209 ABSTRACT Since the share repurchasehas become more common practice in Vietnamese market, this research aims to investigate the relationship between share repurchase and cash dividend payout, the popular cash distribution methods applied by Vietnamese listed firms The substitute or complementary relationship among these financial practices may contribution to the understanding of financial managerial behaviour in Vietnam, a fast development emerging market in Asia Pacific area The empirical results show that there is a nonlinear relationship between the change in dividend and the repurchase yield In addition, a significantly positive correlation reveals the dominant of complementary effect in the whole sample The results are consistent in robustness tests This study employs quantile regression to analyze the nonlinear relationship between share repurchase and cash dividend Empirical results indicate that the substitute or complementary relationship between share repurchase and cash dividend varies with the scale of share repurchase yield The higher the share repurchase yield, the more significant the substitution effect is for the relationship between share repurchase and cash dividend Furthermore, the Vietnamese listed firms seem to more carefully take their future growth into consideration regarding share repurchase activities However, the signaling and free cashflow hypotheses show marginally significances in this case Key words: Vietnam, share repurchase, cash dividend payout policy ii doi:10.6840/cycu202000209 ACKNOWLEDGMENT This study would not have been possible without the generosity; patience and guidance extended by these research oriented individuals who derive great satisfaction in helping others attain success: My advisor, Dr Yi-Pei Chen, the researchers’ adviser, shares her knowledge, shows a greatly concern and support to the researcher; Dr Han-Ching Huang and Dr Yu-Lun Chen (Chung Yuan Christian University), committee members, for all the help, support and assistance; Dr Jung-Hua Hung (National Central University), Dr Tsui-Jung Lin (Chinese Culture University), Dr Chi-Ping Hou (China University of Technology) and Dr Jyun-Ji Tien (Tamkang University), committee members, for giving valuable suggestions to further prove the research study; The Fiinpro, provides the research data bank to accomplish the study; The researchers’ family for their undying support, emotionally, spiritually and financially; The researcher’ friends and classmates who have provided warm-hearted support along the way The Researcher iii doi:10.6840/cycu202000209 CONTENTS 摘 要 i ABSTRACT ii ACKNOWLEDGMENT iii CONTENTS iv TABLE LIST v FIGURE LIST v I INTRODUCTION I LITERATURE REVIEW 1.1 Share repurchases 1.2 Repurchase regulations and Tax policy in Vietnam 1.3 The characteristics of Vietnamese listed firms II DATA AND METHODOLOGY 2.1 Data collection 2.2 Methodology 10 2.3 Dividend change measurement 11 III EMPIRICAL ANALYSIS 13 3.1 General Statistic 13 3.2 Substitution and Complementary effects between share repurchases and dividends 17 3.3 Robustness test 18 IV CONCLUSION 24 REFERENCE 25 iv doi:10.6840/cycu202000209 TABLE LIST Table 1: Tax rate by period in Vietnam Table 2: Variable definitions 12 Table 3: Descriptive Statisic 14 Table 4: Correlation analysis 15 Table 5: Distribution of Share repurchases by the Change in dividend (DDiv) 16 Table 6: Panel regression for the relationship between share repurchase and dividend payout .18 Table 7: Robustness by different conditions for the relationship between share repurchase and dividend payout 20 Table 8: Robustness test using Repurchase ratio as explained variable .21 Table 9: Quantile regression for the relationship between share repurchase and dividend payout 23 FIGURE LIST Figure 1: Distribution of Share repurchases by the Change in dividend (DDiv) 17 v doi:10.6840/cycu202000209 I INTRODUCTION For any organization, cash flow management is always one of the most important functions When firms have positive income or available free cash, normally they will be willing to distribute the surplus to their shareholders as a response to their investment For a long time, the world’s listed firms have overwhelmingly preferred to pay dividends in the form of cash, or even stock Dividend policy becomes an important decision concerning whether profit should be distributed to investors or reinvested for future opportunities and growth However, over the last few decades, share repurchase activity has experienced an extraordinary growth and has become a common practice in developed markets such as the US or Europe since the mid-1980s, said Grullon and Michaely (2004) For those strong economies of the Asia-Pacific region, this practice became common later with the approval of share repurchases by Australia in 1989, Hong Kong in 1991, Korea in 1994 and Japan in 1995 Today, share repurchases gradually have become more popular than dividends For a new emerging market like Vietnam, starting with the establishment of the State Securities Commission - the regulator over the securities market from 1997 - repurchases have been allowed from the inital operation of its securities market in 2000: specifically, the opening of the Ho Chi Minh City Stock Exchange in July 2000 - a trading platform for the stock of relatively large corporations, and the Hanoi Stock Exchange in March 2005 for the stock of relatively SMEs (Kien & Chen, 2020) The Vietnamese securities market, as the founders had expected, has worked well to boost the national economy and maintain a high speed of development This is especially due to the transformation of all state-owned enterprises which play a key role in Vietnamese economic sectors into joint-stock companies, under their “equitization”1 process (Webster & Amin, 1998) The stock market has grown significantly — only two stocks were traded in the beginning, compared to nearly 700 listed firms in the current market Securities markets are now becoming the important capital mobilization channel for the Vietnamese economy.2 In general, most Vietnamese listed firms prefer using cash dividends, and a few use stock dividends, to distribute funds to shareholders Share repurchase has been applied only recently but Equitization is a Vietnamese English term that denotes the conversion of a state-owned enterprise in Vietnam into a public limited company or a corporation Mobilized VND 1,000,000 trillion (≈USD 47.6 billion) for the Government; mobilized VND 700 trillion (≈USD 33.3 billion) for the enterprises via auctions for equitization and issuing shares, fund units, make the securities market capitalization reach nearly 40% GDP (as of July 2014) doi:10.6840/cycu202000209 Table 2: Variable definitions Variable Definition Expected sign Ryield Repurchase yield, defined as the total value of conducted repurchase divided by market value of equity in the corresponding year; DDiv Change in dividend = (Current Dividend– Previous Dividend)/Market capitalization in current year Cash (Cash + Cash equivalent + short-term investment)/Total Asset + MB_ratio Market to Book ratio = (Market value of Equity + Total - (-) : substitute (+) : complement Asset – Book value of Shareholder equity) / Total asset Log(TA) Logarithm value of Firm’s total asset - Lev (Short-term debt + Long-term debt)/ Total Asset - Table shows variable definitions In the purpose of testing for the existence of subsitution hypothesis between share repurchase and dividend payout, it is expected for negative correlation between the testing variable Ryield and its main independent variable DDiv On the other words, with the increasing in repurchase activities, it may affect to the lower spending on dividend distribution in the year t when comparing to its of the previous yield Alternatively, a finding of significant positive coefficient may reveal independent relationship between these two variables, or the complementary effect eixisted on firm capital payout policy While, the square DDiv will be included into the model, to check if the nonlinear relationship may exist between dividend payout and repurchase practice A positive sign of Cash variable is expected to represent the free cash flow hypothesis While, practicing buyback as a signal to other investors that the firm is undervalued will result in a signifiant negative coefficient of Log(TA) As large firms often provide clearer business information to build shareholder’s confidence, making them are less undervalued by the market The market to book ratio (MB_ratio) stands for the firm future growth If firms retain more cash for other investment plan, spending for payout should lower, which is reflected by a negative relationship from MB_ratio to repurchase In addition, the firm with high debt ratio in financial structure obviously has less resource to buyback its share, reported by an expected negative sign for Lev variable 12 doi:10.6840/cycu202000209 In concerning with the affect of State control and Tax policy to the firms’ capital distribution decision, the government ownership and tax will be use to identify different comparision sub-sample Thus, it may giving a broaden view on research findings In specific, the firms will be seperated into two groups as of whever government contribution as blockholder or not According to Vietnamese security law, shareholders with at least 5% outstanding stock will be classified as blockholders Blockholders are known as activists with more influential voting right Thus, I will create a group calls Gov5, in which government is firm’s blockholder as of it owns more than 5% firm’s share, giving state the right to vote and effect to the business and managerial decision From observation, I recgonize that among 174 completed repurchases, there was over half (92 buyback) belonged to those enterprises with government investment Among 92 practices, 90 or 98% of them were from firms with the state as their blockholder (by holding at least 5% of outstanding share) It is also worth to note that around 90% of state blockholder (82/90) having more than 10% right to vote Thus, group Gov10 which included all of state blockholders with at least 10% outstanding stock was employed, for higher level of robustness test In case of tax, its affect will be observed through the different market reaction between the post and pre-period of tax issue III EMPIRICAL ANALYSIS 3.1 General Statistic The general statistics of all variables are reported in Table It presents an overall view of the listed Vietnamese firms’ characteristic, those practiced the repurchase policy within the observation period General information is summarized in Panel A of Table 3, with high difference between the maximum and the minimum in repurchase yield (Ryield), it seems that firms’ buyback policies are quite fluctuant among firms While there are firms to hold a big volume of repurchases at nearly 27% of theirs market value, some other companies are spending very little amount to buyback their own shares Besides distribution purpose, the managerial intent may be the cause of these practices For example, the firm buys back a small number of their own shares to make it more convenient to manage the amount of outstanding shares or to serve the employee stock option plan Since distribution matter is not the priority, these repurchases may be less related to other payout method, which is related to complementary effect While, the Change in dividends (DDiv) gives another point of view, both of its mean and median are in negative sign That is, on average, firms are going to pay less dividend than expected, the difference in dividend payments is more likely to be shifted to the other kind of payouts, such as share repurchases, resulting in expectation 13 doi:10.6840/cycu202000209 for substitute relationship between payout policies For other explanary variables, Cash, Mb_ratio, Ln_TA, Lev, they all have Median fluctuant around the Mean However, their distance of Maximum and Minium value are high from the Mean, showing a range between the financial ratio of firms, which secure that the observed sample scope may cover the whole market It may become a mark for the accuracy of later finding Table 3: Descriptive Statisic Panel A: General Characteristics Variables Mean Median Maximum Minimum Std Dev Observations Ryield DDiv 0.026 -0.015 0.007 -0.003 0.269 0.672 0.000 -0.710 0.045 0.126 174 174 Cash(-1) MB_ratio(-1) 0.199 1.346 0.169 0.989 0.679 6.018 0.003 0.396 0.158 0.984 174 174 Ln_TA(-1) Lev(-1) 11.909 0.198 11.793 0.165 13.864 0.741 10.595 0.000 0.664 0.193 174 174 Panel B: Sample Distribution Market HNX HOSE Total Basic Materials Consumer Goods 19 44 25 48 Consumer Services 10 17 Industrials Oil & Gas Technology 22 11 40 2 11 62 Utilities 5 44 130 174 Industry Health Care Total Note: - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset Other information such as industrial and market distribution may be observerd from Panel B of Table It can be seen that buyback are more prefered by listed firms from Ho Chi Minh stock exchange (HOSE) with around three fourth of total share repurchase activities Indetail, there were 130 completed repurchases conducted in HOSE while its numberwas only 44 in HNX With the 14 doi:10.6840/cycu202000209 market characteristic where most of larger firms are listed in HOSE while SMEs go to HNX, we can say that those firms with more advance in capital may have more incentive to conduct the repurchasing in Vietnamese market From the industry classification view, industrial firms are dominant in both market in case of share buyback holding, with total of 62 over 174 observations For the second rank, consumer service enterprises in HNX followed up with 10 completed campaigns, while in HOSE the firms from consumer goods industry were doing more share repurchasing with 44 holded activities Table 4: Correlation analysis Variables RYIELD RYIELD 1.000 CASH(-1) MB_RATIO(-1) LN_TA(-1) LEV(-1) DDIV CASH(-1) MB_RATIO(-1) LN_TA(-1) LEV(-1) DDIV -0.064 1.000 (0.402) - -0.269 0.166 1.000 (0.000) (0.029) - -0.096 -0.138 0.261 1.000 (0.210) (0.069) (0.001) - 0.028 -0.451 -0.299 0.381 1.000 (0.716) (0.000) (0.000) (0.000) - -0.046 0.095 0.080 0.072 -0.096 1.000 (0.549) (0.211) (0.294) (0.346) (0.210) Note: The p-value are reported in parentheses - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset Table presents the correlation coefficients of the dependent variables and key testing variables Cash and Leverage (Lev) share a negative correlation at -0.451, the largest absolute value correlation among variables, which shows a strong relationship between the firm’s capital and its debt ratio The negative correlation makes sense that increasing on leverage level causes higher responsibility of redeeming the debt, thus lower the amount of cash on hand For the relationship between explained and tested variables, the market to book ratio (MB_ratio) shows a strongest 15 doi:10.6840/cycu202000209 connection to explained variable, Ryield, at -0.269 In detail, MB_ratio is negative significantly to repurchase yield (Ryield), while it is insignificant for the main tested variable, DDiv Table reveals the distribution of completed share repurchases under different degree of change in dividend (DDiv) Among 174 completed buybacks, it can be clearly seen that the sample is divided into two groups in two inverted U-shapted terms The left side belongs to the group of change in dividend (DDiv) less than 0, which means firms were paying less dividend than expected While, the right side is the group of increasing dividend payout In detail, firms with lower dividend strategy has highest proportion 55% of repurchasing firms have 44.25% of repurchase activities when dividends are decreasing by zero to 10% than before They slow down at 0% then move to the second inverted U-shape on the other side In wich the firms distribute more dividend from zero to 10% of expected value aslo get peak at 32.76% of total repurchase Table 5: Distribution of Share repurchases by the Change in dividend (DDiv) DDIV No of Repurchase % Rep Average Ryield % firm -30% -30% to -20% -20% to -10% -10% to 0% 13 77 2.87% 1.15% 7.47% 44.25% 0.054 0.003 0.059 0.023 5% 2% 13% 55% 0% 0% to 10% 10% to 20% 20% to 30% >30% 57 10 4.02% 32.76% 5.75% 1.15% 0.57% 0.041 0.020 0.008 0.040 0.201 5% 44% 10% 2% 1% Note: - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization A better imagination may be revealed through Figure as belows, for a clearer overview of their relationship It may be a preliminary evidence of non-linear results 16 doi:10.6840/cycu202000209 Figure 1: Distribution of Share repurchases by the Change in dividend (DDiv) 3.2 Substitution and Complementary effects between share repurchases and dividends Applying the Hausman test, the P-value is insignificant at 10%, which means the null hypothesis of “individual specific effects are random” can not be rejected On the other word, random effect should be appropriate for panel regression Furthermore, since the observation period (10 years) is much smaller than the cross section (with around 174 repurchase observations), statistically, this is insufficient for applying the Period Random Effect Thus make Cross-section Random Effects are suitable for testing the coefficient between the explanatory variables in this study The main testing with nonlinear relationship by panel regression will be shown on Table It can be clearly seen from Table 6, there is a significantly positive-positive nonlinear relationship between the change in dividend (DDiv) and repurchase In which the coefficient of DDiv and DDiv^2 are significantly positive (0.051 and 0.212, respectively) at the significant levels of 5% and 1% This result supports for the complementary hypothesis, which means the dividend and share repurchase are independent payout methods Firms may distribute their earnings after tax and other liability obligations for whatever dividend payout or buyback the share The P-value of F-statistic is less than 1%, it confims that the explanatory variables joinly can influent the dependent variable Most of explanatory variables share the expected signs, except Cash when its sign was negative rather than positive However, among them, only MB_ratio is significantly negative as expected, which supports the idea that firm may pay out less cash in order to maintain their future growth The coefficient of MB_ratio is -0.015 at the 1% significant level 17 doi:10.6840/cycu202000209 Table 6: Panel regression for the relationship between share repurchase and dividend payout Dependent Variable: RYIELD Variable Coefficient Std Error t-Statistic Prob C DDIV DDIV^2 CASH(-1) MB_RATIO(-1) LN_TA(-1) LEV(-1) 0.083 0.051** 0.212*** -0.013 -0.015*** -0.003 -0.012 0.078 0.020 0.047 0.022 0.005 0.007 0.022 1.056 2.484 4.527 -0.597 -3.097 -0.407 -0.528 0.292 0.014 0.000 0.551 0.002 0.685 0.598 Chi-Sq Statistic 10.61116 Chi-Sq d.f Prob 0.1012 Cross-section random R-squared Adj R Yes 0.161 0.131 F-Stastistic Prob (F-Stastistic) Obs 5.344 0.000 174 Hausman Test Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.1 levels, respectively - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset The result then be rechecked by several robustness as follow 3.3 Robustness test For more accuracy for the result found on panel regression test, several conditions have been included for robustness in Table For the first option (1), a subsample is created from group of dummy dividend equal to (Dummy_Div = 1), which mean testing for the group of firms that practices both dividend payout and share repurchase at the year The condition of different tax levels (Tax_0 =1 and Tax_0=0) were added together with Dummy_Div=1 for model (2) In which during some special period of economic recession, a zero tax policy was applied by Vietnamese government to support individual investors, as they are often the vulnerable group with the lack of voting right within the firms when comparing to other dominant or blockholder investors The effect of government control to firm’s payout policy is checked via the group of state blockholder 18 doi:10.6840/cycu202000209 at above 5% and 10% outstanding share (Gov and Gov10 group alternatively) in test (3) and (4) While, Tobit regression is applied in test (5) to check for the study’s consistency as the range of observation for the dependent variable is censored in some way The results of those robustness tests in Table show a highly identical on the sign of coefficient and qualitatively similar to those of panel regression test in most of conditions Interestingly, the significant positive – positive relationship of DDiv and DDiv^2 to repurchase yield, in model (3) when tax was at higher rate and in model (4), (5) when state is blockholder of the firm, may related to the result of my previous study In which we found that the Vietnamese state enterprises still prefer for high (dividend) payout policy even in higher tax issue period, in order to build the confidence of their shareholders (Kien and Chen, 2020) It can be used as a signal by the firm to shareholders that its business is still running well and able to create the return for investments In addition, in concern to the approaching method to identify the feature of repurchase activity, another variable calls Repurchase ratio (Rep_ratio) will be replaced to the dependent variable in the main model, for broaden view of the connection between repurchase and dividend payout applied Repurchase ratio is calculated by using the total volume of share repurchase divided to the outstanding share of firm before buyback This simple ratio is one of common methods that applied by many investors or researchers to have the first and most directly understanding about the influence of this financial announcement to firm situation The test result is shown on Table as bellows 19 doi:10.6840/cycu202000209 Table 7: Robustness by different conditions for the relationship between share repurchase and dividend payout Dependent Variable: RYIELD Conditions applied Variable C DDIV DDIV^2 CASH(-1) MB_RATIO(-1) LN_TA(-1) LEV(-1) Cross-section random Adj R2 Obs (2) Dummy_Div=1 (3) Dummy_Div=1 (4) Dummy_Div=1 (5) Dummy_Div=1 Tax_0=1 Tax_0=0 Gov5=1 Gov10=1 0.074 0.053*** 0.213*** 0.009 -0.012** -0.003 -0.002 0.133 -0.04 0.585*** -0.013 -0.016** -0.008 -0.004 0.011 0.072*** 0.135*** 0.002 -0.010** 0.004 -0.011 0.042 0.059** 0.203** 0.015 -0.005 -0.002 0.036 0.059 0.056** 0.223*** -0.003 -0.004 -0.003 0.018 Yes Yes Yes Yes Yes 0.349 158 0.358 54 0.217 104 0.031 85 0.043 77 (1) Dummy Div=1 (6) TOBIT 0.080 0.023 0.217*** -0.022 -0.011* -0.003 -0.020 174 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.1 levels, respectively - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset 20 doi:10.6840/cycu202000209 Table 8: Robustness test using Repurchase ratio as explained variable Dependent Variable: REP_RATIO (1) Raw Sample group (2) Dummy_Div = Variables Coef t-value Sig (P) Coef t-value Sig (P) C 0.017 0.298 0.766 0.007 0.121 0.904 DDIV 0.034** 2.186 0.030 0.037** 2.306 0.023 DDIV^2 0.063* 1.815 0.071 0.065* 1.892 0.061 CASH(-1) -0.017 -1.033 0.303 -0.005 -0.273 0.785 -0.01*** -2.940 0.004 -0.008*** -2.291 0.023 0.002 0.498 0.619 0.003 0.510 0.611 -0.031* -1.888 0.061 -0.024 -1.406 0.162 MB_RATIO(-1) LN_TA(-1) LEV(-1) Cross-section random Adj R2 Observations Yes Yes 0.091 0.051 163 148 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.1 levels, respectively - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset It is clearly seen from Table that the result is consistent whever the repurchase yield (Ryield) or the repurchase ratio (Rep_ratio) is used as dependent variable on the panel regression test In detail, it is positive significant for DDiv and DDiv^2 (0.034 and 0.063 at 5% and 10% significant level) and negative significant for MB_ratio (-0.01 at 1% level) The test of repurchase ratio also show a negative significant coefficient of -0.031 at 10% significant level in case of Leverage (Lev) variable The negative coefficient of Lev variable supports the idea that firm may spend less in repurchasing when they have more liability obligations to deal with In addition, a restricted condition of firms practice both dividend and repurchase at the year (Dummy_Div =1) has been applied, it reveals a quantitatively similar result to the original group These results create 21 doi:10.6840/cycu202000209 confidence to the found result about the complementary relationship between firm's share repurchase and dividend payout Furthermore, since the nonlinear effect has been discovered in the model, it raise the question of is there any different treatment in different group or range of obsereved sample Thus, the quantile regression is applied to investigate for any futher hidden movement behind the scene From Table 9, quantile regression gives an interesting result, the relation of DDiv variable to repurchase yield is changed to negative significant in higher level at 10% level for 65th quantile and at 1% significance level for upper levels from 70th quantile It means the group of high repurchase yield, there is opposite relationship between repurchase yield and dividend payout, listed firm seems to be more interested in interchanged between the two fund distribution policies It may become the evidence for the existence of substitution relationship between share repurchase and dividend payout applied by Vietnamese firms In addition, when comparing for the original result, positive-positive sign of DDiv and DDiv^2 for the full sample and the negative-positive sign of those in upper level quantile, we can say that the complementary effect is very strong within those firms preferring lower repurchase yield This phenomenon makes complementary approach dominant in case of full sample observation about share repurchase in Vietnamese stock market The market to book ratio (Mb_ratio) continuously has negative significant to repurchase yield in all quantile, results as a strong confirmation for the opinion that Vietnamese firms in take caring of their future investment plan, they may lower the spending for buyback activity While, Cash variable shows a negative significant at 10% in 70th quantile, it is opposite to the expected positive correlation, making the free cash flow hypothesis unsupported in this study 22 doi:10.6840/cycu202000209 Table 9: Quantile regression for the relationship between share repurchase and dividend payout Dependent Variable: RYIELD Variable C DDIV DDIV^2 CASH(-1) MB_RATIO(-1) LN_TA(-1) LEV(-1) Pseudo R2 Adj R2 Random effect 0.083 (0.292) 0.051** (0.014) 0.212*** (0.000) -0.013 (0.551) -0.015*** (0.002) -0.003 (0.685) -0.012 (0.593) 0.161 0.131 Quantile regressions 10th quant 20th quant 30th quant 40th quant 50th quant 60th quant 65th quant 70th quant 80th quant 90th quant 0.001 (0977) 0.000 (0.985) 0.001 (0.981) 0.001 (0.938) 0.000 (0.969) 0.000 (0.981) 0.000 (0.996) 0.002 -0.034 0.003 (0.877) 0.005 (0.863) 0.005 (0.939) 0.002 (0.817) -0.001 (0.596) 0.000 (0.955) -0.002 (0.818) 0.008 -0.027 0.015 (0.556) 0.010 (0.757) 0.009 (0.905) 0.000 (0.963) -0.001 (0.360) -0.001 (0.703) -0.003 (0.797) 0.021 -0.015 0.013 (0.650) 0.025 (0.509) 0.030 (0.710) -0.002 (0.875) -0.002 (0.236) 0.000 (0.863) -0.001 (0.917) 0.030 -0.005 0.020 (0.492) 0.035 (0.316) 0.050 (0.552) -0.012 (0.252) -0.003* (0.073) 0.000 (0.897) -0.006 (0.619) 0.041 0.007 0.048 (0.177) 0.014 (0.887) 0.378* (0.072) -0.012 (0.285) -0.004** (0.023) -0.002 (0.499) -0.018 (0.210) 0.059 0.025 0.054 (0.314) -0.095*** (0.009) 0.551*** (0.000) -0.028 (0.124) -0.005** (0.026) -0.002 (0.729) -0.013 (0.589) 0.084 0.051 0.022 (0.748) -0.109*** (0.000) 0.556*** (0.000) -0.037** (0.045) -0.008*** (0.007) 0.002 (0.735) -0.028 (0.290) 0.119 0.088 0.014 (0.884) -0.103*** (0.000) 0.514*** (0.000) -0.038 (0.273) -0.010** (0.013) 0.004 (0.666) -0.011 (0.729) 0.169 0.140 0.135 (0.430) -0.079* (0.062) 0.451*** (0.000) -0.053 (0.520) -0.012** (0.019) -0.004 (0.770) 0.024 (0.714) 0.222 0.194 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.1 levels, respectively - The dependent variable Ryield denotes Dividend Yield - The key testing variable DDiv denotes Change in dividend = (Dividend t – Expected Dividend t)/Market capitalization - The controlling variable Cash: (Cash + Cash equivalent + short-term investment)/Total asset; MB_ratio denotes Market to Book ratio = (Market value of Equity + Total Asset – Book value of Shareholder equity) / Total asset; Log(TA) denotes Log value of Firm’s total asset; Lev: (Short-term debt + Long-term debt)/ Total Asset 23 doi:10.6840/cycu202000209 IV CONCLUSION In conclusion, this study has fulfilled the gap of lack understanding about non-linear connection among different capital distribution policies applied by Vietnamese listed firms In detail, the panel regression proves that there is non-linear relationship with the complementary effect seems to be dominant between share repurchase and dividend payout policy of the firm This finding have been supported by robustness tests by different conditions However, by applying the quantile regression, it reveals that the group of upper level (high repurchase yield) have more interested in substitution relationship In which, the dividend and repurchase are interchangeable While, the complementary approach seems to have strong influence for firms from group of lower quantile In addition, the signaling and free cash flow hypothesis are less likely having major influence to the relationship among payout methods However, the study proves that Vietnamese firms may have serious consideration about their future growth when considering about share repurchase decision In the final, for sure there is scope of improvement remains and it is part of future study 24 doi:10.6840/cycu202000209 REFERENCE Alphonse, P., & Quoc Trung, T (2014) A Two-Step Approach to Investigate Dividend Policy: Evidence from Vietnamese Stock Market International Journal of Economics and Finance, 6(3) Bradford, W., Chao,C & Song, Z (2013) Cash Dividend Policy, Corporate Pyramids, and Ownership Structure: Evidence from China International Review of Economics & Finance, 27, 445–64 Brennan, M.J & Thakor, A.V (1990) Shareholder Preferences and Dividend Policy The Journal of Finance, 45(4), 993-1018 Brown, C., Handley, J & O'Day, J (2015) The Dividend Substitution Hypothesis: Australian Evidence Abacus, 51, 37-62 Chen, D., Ming, J., & Ming, X (2009) Dividends for Tunneling in a Regulated Economy: The Case of China Pacific-Basin Finance Journal, 17, 209–23 Dittmar, A (2000) Why Do Firms Repurchase Stock? Journal of Business, 73(3), 331–55 Geraci, M (2018) Modelling and estimation of nonlinear quantile regression with clustered data Computational Statistics & Data Analysis, 136, 30-46 Ginglinger, E & L’her, J-F (2011) Ownership structure and open market stock repurchases in France European Journal of Finance Grullon, G & Michaely, R (2002) Dividends, Share Repurchases, and the Substitution Hypothesis The Journal of Finance 10 Jacob, M., & Jacob, M (2013) Taxation, Dividends, and Share Repurchases: Taking Evidence Global The Journal of Financial and Quantitative Analysis, 48(4), 1241-1269 11 Jagannathan, M & Stephens, C (2003) Motives for Multiple Open-Market Repurchase Programs Financial Management, 32 12 Jagannathan, M.; Stephens, C & Weisbach, M (2000) Financial Flexibility and The Choice Between Dividends and Stock Repurchases Journal of Financial Economics, 57, 355-384 13 Jong-Cook, B & Bao Trung, P (2016) Share Repurchases in Vietnam: Why Do Firms Repurchase Shares? Journal of International Trade & Commerce 14 Kanas, A (2005) Non-Linearity in the Stock Price-Dividend Relation Journal of International Money and Finance, 24, 583-606 25 doi:10.6840/cycu202000209 15 Kien, D & Chen, Y (2020) Ownership Structure Impact on Dividend Policy of Listed Companies on Vietnamese Securities Market Journal of Mathematical Finance, 10, 223241 16 Koenker, R & Bassett, G (1978) Regression quantiles Econometrica, 46(1), 33-50 17 Lee, B., & Rui, O (2007) Time-Series Behavior of Share Repurchases and Dividends The Journal of Financial and Quantitative Analysis, 42(1), 119-142 18 Lee, Y.T., Liu, Y.J., Roll, R & Subrahmanyam, A (2006) Taxes and Dividend Clientele: Evidence from Trading and Ownership Structure Journal of Banking and Finance, 30, 229– 246 19 Lie, E (2002) Do Firms Undertake Self-Tender Offers to Maximize Capital Structure? The Journal of Business, 75 (4), 609–639 20 Lin, G (2009) Quantile Regression with Censored Data Ph.D Dissertation, University of Illinois at Urbana-Champaign 21 Litzenberger, R & Ramaswamy, K (1982) The effects of dividends on common stock prices tax effects or information effects? The Journal of Finance, 37(2), 429-443 22 Mork, R., Shleifer, A & Vishny, R.W (1988) Management ownership and market valuation Journal of Financial Economics, 20, 293–315 23 Nohel, T & Tarhan, V (1997) Share repurchases and firm performance: New evidence on the agency costs of free cash flow Journal of Financial Economics 24 Rau, P R., & Vermaelen, T (2002) Regulation, Taxes, and Share Repurchases in the United Kingdom Journal of Business, 75, 245–282 25 Sáez, M & Gutiérrez, M (2015) Dividend Policy with Controlling Shareholders Theoretical Inquiries in Law, 16 (1) 26 Sarig, O (2004) A Time-Series Analysis of Corporate Payout Policies Review of Finance, 8, 515–536 27 Vu, C.D (2019) Thị trường chứng khốn Việt Nam mơi trường thực thi FTA hệ Financial Magazine, Ministry of Finance of Vietnam 28 Webster, L & Amin, M.R (1998) Equitization of state enterprises in Vietnam Private sector discussions, Washington, D.C., World Bank Group 26 doi:10.6840/cycu202000209 ... further in investigating the contribution of share repurchase activities in the Vietnamese market It is important because it will enhance the understanding of corporate payout policy in the Vietnamese. .. between share repurchase and cash dividend Empirical results indicate that the substitute or complementary relationship between share repurchase and cash dividend varies with the scale of share repurchase. .. of share repurchases by Australia in 1989, Hong Kong in 1991, Korea in 1994 and Japan in 1995 Today, share repurchases gradually have become more popular than dividends For a new emerging market