Luận án Cơ cấu nguồn vốn của các doanh nghiệp thương mại dầu khí tại Việt Nam

G SIZE RISK Durbin-Wu-Hausman chi-sq test: 0.10155 Chi-sq(1) P-value = 0.74998 Wu-Hausman F test: 0.09572 F(1,146) P-value = 0.75747 H0: Regressor is exogenous Tests of endogeneity of: SIZE . ivendog SIZE Instruments: CR GOV GROW PROF TANG RISK L.SIZE Instrumented: SIZE _cons .4642752 .1997888 2.32 0.022 .0694459 .8591044 RISK -.0599679 .2570564 -0.23 0.816 -.5679713 .4480356 TANG .1207111 .0449377 2.69 0.008 .0319037 .2095185 PROF .1595431 .0882252 1.81 0.073 -.0148105 .3338967 GROW -.0103407 .0281005 -0.37 0.713 -.0658739 .0451925 GOV .0460423 .0334059 1.38 0.170 -.0199757 .1120602 CR -.0096598 .0045069 -2.14 0.034 -.0185665 -.0007531 SIZE -.0153296 .0072612 -2.11 0.036 -.0296795 -.0009798 LLEV Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 1.13825023 154 .007391235 Root MSE = .08097 Adj R-squared = 0.1130 Residual .963772604 147 .006556276 R-squared = 0.1533 Model .174477623 7 .024925375 Prob > F = 0.0008 F( 7, 147) = 3.78 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg LLEV CR GOV GROW PROF TANG (SIZE=l.SIZE) RISK + Kết quả kiểm tra với biến độc lập TANG Do P-value = 0,32695 lớn hơn 0,05 nên biến TANG là biến ngoại sinh. + Kết quả kiểm tra với biến độc lập RISK Do P-value = 1,000 lớn hơn 0,05 nên biến ZSCORE là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 0.96093 Chi-sq(1) P-value = 0.32695 Wu-Hausman F test: 0.91078 F(1,146) P-value = 0.34148 H0: Regressor is exogenous Tests of endogeneity of: TANG . ivendog TANG Instruments: CR GOV GROW PROF SIZE RISK L.TANG Instrumented: TANG _cons .493511 .2009112 2.46 0.015 .0964636 .8905584 RISK -.0788931 .2577057 -0.31 0.760 -.5881796 .4303934 SIZE -.0162762 .0072873 -2.23 0.027 -.0306777 -.0018747 PROF .1669984 .0885129 1.89 0.061 -.0079238 .3419206 GROW -.0117185 .0281449 -0.42 0.678 -.0673394 .0439025 GOV .0478409 .0333785 1.43 0.154 -.0181227 .1138045 CR -.0098837 .0045142 -2.19 0.030 -.0188047 -.0009626 TANG .1041232 .0480704 2.17 0.032 .0091249 .1991215 LLEV Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 1.13825023 154 .007391235 Root MSE = .08101 Adj R-squared = 0.1122 Residual .96462838 147 .006562098 R-squared = 0.1525 Model .173621847 7 .024803121 Prob > F = 0.0019 F( 7, 147) = 3.44 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg LLEV CR GOV GROW PROF (TANG=l.TANG) SIZE RISK Durbin-Wu-Hausman chi-sq test: -0.00000 Chi-sq(1) P-value = 1.00000 Wu-Hausman F test: -0.00000 F(1,146) P-value = 1.00000 H0: Regressor is exogenous Tests of endogeneity of: RISK . ivendog RISK Instruments: CR GOV GROW PROF TANG SIZE L.RISK Instrumented: RISK _cons .4690885 .1991854 2.36 0.020 .0754515 .8627254 SIZE -.0155051 .0072392 -2.14 0.034 -.0298114 -.0011988 TANG .1203908 .0449258 2.68 0.008 .031607 .2091746 PROF .1602796 .0881931 1.82 0.071 -.0140106 .3345697 GROW -.0104176 .0280994 -0.37 0.711 -.0659485 .0451133 GOV .046685 .0333416 1.40 0.164 -.0192058 .1125758 CR -.0096722 .0045067 -2.15 0.034 -.0185785 -.0007659 RISK -.0619362 .2569777 -0.24 0.810 -.569784 .4459116 LLEV Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 1.13825023 154 .007391235 Root MSE = .08097 Adj R-squared = 0.1130 Residual .963768752 147 .00655625 R-squared = 0.1533 Model .174481475 7 .024925925 Prob > F = 0.0008 F( 7, 147) = 3.80 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg LLEV CR GOV GROW PROF TANG SIZE (RISK=l.RISK) Phụ lục 13: Kết quả kiểm định tác động tuyến tính của LEV tới tỷ suất sinh lời VCSH (ROE) (mô hình (4)) - Thống kê mô tả các biến trong mô hình - Bảng mô tả hệ số tương quan của mô hình (4) - Mô hình OLS - Kiểm định hiện tượng đa cộng tuyến LLEV2 186 .0108065 .0276444 0 .18 SLEV2 186 .2254839 .1742475 0 .8 LEV2 186 .2845086 .1900569 .0044 .8527 ROE 186 .0793661 .1272789 -.5084 .4152 Variable Obs Mean Std. Dev. Min Max SIZE 0.1729 -0.0595 0.7754 0.0030 -0.2616 1.0000 TANG 0.0213 -0.1405 -0.2193 -0.0766 1.0000 GROW 0.1756 -0.0968 -0.0180 1.0000 GOV 0.0225 0.0612 1.0000 CR -0.7244 1.0000 LEV 1.0000 LEV CR GOV GROW TANG SIZE _cons -1.243517 .2086465 -5.96 0.000 -1.65524 -.8317939 SIZE .0506454 .007741 6.54 0.000 .03537 .0659208 TANG .1371032 .0532896 2.57 0.011 .0319464 .2422599 GROW .1532178 .0344034 4.45 0.000 .0853295 .2211062 GOV -.1354258 .0388589 -3.49 0.001 -.2121064 -.0587453 CR .0013873 .0075689 0.18 0.855 -.0135483 .016323 LEV -.1481903 .0622398 -2.38 0.018 -.2710084 -.0253723 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.99698509 185 .016199919 Root MSE = .11017 Adj R-squared = 0.2508 Residual 2.17265023 179 .012137711 R-squared = 0.2751 Model .82433486 6 .137389143 Prob > F = 0.0000 F( 6, 179) = 11.32 Source SS df MS Number of obs = 186 Mean VIF 1.99 GROW 1.04 0.958980 TANG 1.12 0.895903 CR 2.19 0.456400 LEV 2.26 0.442695 GOV 2.60 0.384540 SIZE 2.74 0.364521 Variable VIF 1/VIF - Kiểm định hiện tượng tự tương quan - Kiểm định hiện tượng phương sai thay đổi - Mô hình FEM Prob > F = 0.0013 F( 1, 30) = 12.672 H0: no first-order autocorrelation Wooldridge test for autocorrelation in panel data Total 102.14 33 0.0000 Kurtosis 2.73 1 0.0982 Skewness 24.51 6 0.0004 Heteroskedasticity 74.90 26 0.0000 Source chi2 df p Cameron & Trivedi's decomposition of IM-test Prob > chi2 = 0.0000 chi2(26) = 74.90 against Ha: unrestricted heteroskedasticity White's test for Ho: homoskedasticity F test that all u_i=0: F(30, 150) = 7.43 Prob > F = 0.0000 rho .61683918 (fraction of variance due to u_i) sigma_e .07888088 sigma_u .10008458 _cons -1.311814 .7293103 -1.80 0.074 -2.752862 .1292342 SIZE .0545053 .0267787 2.04 0.044 .0015932 .1074174 TANG .0981707 .1344986 0.73 0.467 -.1675858 .3639273 GROW .1055284 .0326223 3.23 0.001 .0410699 .1699869 GOV 0 (omitted) CR -.0016247 .0090242 -0.18 0.857 -.0194557 .0162062 LEV -.2315761 .0927601 -2.50 0.014 -.4148613 -.0482908 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] corr(u_i, Xb) = -0.4058 Prob > F = 0.0044 F(5,150) = 3.58 overall = 0.1860 max = 6 between = 0.2394 avg = 6.0 R-sq: within = 0.1066 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Fixed-effects (within) regression Number of obs = 186 - Mô hình REM - Kiểm định Hausman rho .53122635 (fraction of variance due to u_i) sigma_e .07888088 sigma_u .08397112 _cons -1.239388 .3617273 -3.43 0.001 -1.948361 -.5304155 SIZE .0516253 .0134859 3.83 0.000 .0251934 .0780573 TANG .1248486 .0849323 1.47 0.142 -.0416157 .291313 GROW .1132394 .030173 3.75 0.000 .0541015 .1723773 GOV -.1398447 .0715125 -1.96 0.051 -.2800066 .0003172 CR -.0010048 .007997 -0.13 0.900 -.0166787 .0146691 LEV -.1947213 .0747643 -2.60 0.009 -.3412567 -.048186 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(6) = 32.15 overall = 0.2652 max = 6 between = 0.3511 avg = 6.0 R-sq: within = 0.1046 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Random-effects GLS regression Number of obs = 186 Prob>chi2 = 0.8427 = 2.05 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg SIZE .0545053 .0516253 .0028799 .023135 TANG .0981707 .1248486 -.0266779 .1042899 GROW .1055284 .1132394 -.0077109 .0124018 CR -.0016247 -.0010048 -.0006199 .0041813 LEV -.2315761 -.1947213 -.0368548 .0549067 FEM REM Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients - Kiểm định phương sai thay đổi - Mô hình GLS có hiệu chỉnh Prob > chibar2 = 0.0000 chibar2(01) = 102.62 Test: Var(u) = 0 u .0070511 .0839711 e .0062222 .0788809 ROE .0161999 .1272789 Var sd = sqrt(Var) Estimated results: ROE[MACK,t] = Xb + u[MACK] + e[MACK,t] Breusch and Pagan Lagrangian multiplier test for random effects _cons -.8644895 .1836624 -4.71 0.000 -1.224461 -.5045177 SIZE .0381798 .0069627 5.48 0.000 .0245332 .0518264 TANG .1331902 .0380313 3.50 0.000 .0586504 .2077301 GROW .0380207 .0179589 2.12 0.034 .002822 .0732195 GOV -.0877332 .0307049 -2.86 0.004 -.1479138 -.0275526 CR -.007694 .0059494 -1.29 0.196 -.0193547 .0039667 LEV -.1817025 .0473463 -3.84 0.000 -.2744997 -.0889054 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.0000 Wald chi2(6) = 37.17 Estimated coefficients = 7 Time periods = 6 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 186 Correlation: common AR(1) coefficient for all panels (0.4993) Panels: heteroskedastic Coefficients: generalized least squares Cross-sectional time-series FGLS regression - Mô hình GMM Difference (null H = exogenous): chi2(3) = 5.42 Prob > chi2 = 0.144 Sargan test excluding group: chi2(12) = 18.39 Prob > chi2 = 0.104 iv(CR GOV TANG) Difference (null H = exogenous): chi2(9) = 17.48 Prob > chi2 = 0.042 Sargan test excluding group: chi2(6) = 6.33 Prob > chi2 = 0.387 GMM instruments for levels Difference-in-Sargan tests of exogeneity of instrument subsets: (Not robust, but not weakened by many instruments.) Sargan test of overid. restrictions: chi2(15) = 23.81 Prob > chi2 = 0.068 Arellano-Bond test for AR(2) in first differences: z = -1.02 Pr > z = 0.308 Arellano-Bond test for AR(1) in first differences: z = -2.56 Pr > z = 0.011 DL.(L.LEV L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) _cons CR GOV TANG Standard Instruments for levels equation L2.(L.LEV L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) D.(CR GOV TANG) Standard Instruments for first differences equation _cons -1.508682 .4684399 -3.22 0.001 -2.426807 -.5905567 SIZE .0732086 .0179039 4.09 0.000 .0381177 .1082996 TANG .138094 .0480898 2.87 0.004 .0438397 .2323483 GROW .3054449 .0742103 4.12 0.000 .1599953 .4508945 GOV -.2039809 .0711425 -2.87 0.004 -.3434176 -.0645442 CR -.0416587 .0170653 -2.44 0.015 -.0751061 -.0082114 LEV -.6975006 .2090843 -3.34 0.001 -1.107298 -.2877029 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.000 max = 6 Wald chi2(6) = 48.90 avg = 6.00 Number of instruments = 22 Obs per group: min = 6 Time variable : YEAR Number of groups = 31 Group variable: MACK Number of obs = 186 Dynamic panel-data estimation, one-step system GMM - Kiểm tra hiện tượng nội sinh với biến phụ thuộc ROE + Kết quả kiểm tra với biến độc lập LEV Do P-value = 0,00066 nhỏ hơn 0,05 nên biến LEV là biến nội sinh. + Kết quả kiểm tra với biến độc lập CR Do P-value = 0,05834 lớn hơn 0,05 nên biến CR là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 11.58629 Chi-sq(1) P-value = 0.00066 Wu-Hausman F test: 11.87603 F(1,147) P-value = 0.00074 H0: Regressor is exogenous Tests of endogeneity of: LEV . ivendog LEV Instruments: CR GOV GROW TANG SIZE L.LEV Instrumented: LEV _cons -1.36058 .2283246 -5.96 0.000 -1.811778 -.9093828 SIZE .0520984 .0084861 6.14 0.000 .0353288 .0688679 TANG .1452463 .0574821 2.53 0.013 .0316545 .2588381 GROW .1147038 .0371759 3.09 0.002 .0412396 .188168 GOV -.1410015 .0423846 -3.33 0.001 -.2247588 -.0572443 CR .0086868 .0086549 1.00 0.317 -.0084164 .02579 LEV -.0161248 .0820033 -0.20 0.844 -.1781734 .1459239 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .1088 Adj R-squared = 0.2430 Residual 1.75190838 148 .011837219 R-squared = 0.2725 Model .656103811 6 .109350635 Prob > F = 0.0000 F( 6, 148) = 9.06 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE (LEV =l.LEV) CR GOV GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 3.58399 Chi-sq(1) P-value = 0.05834 Wu-Hausman F test: 3.47947 F(1,147) P-value = 0.06413 H0: Regressor is exogenous Tests of endogeneity of: CR . ivendog CR Instruments: LEV GOV GROW TANG SIZE L.CR Instrumented: CR _cons -1.266362 .230391 -5.50 0.000 -1.721643 -.8110813 SIZE .0545279 .0083832 6.50 0.000 .0379617 .0710942 TANG .1296265 .0575181 2.25 0.026 .0159636 .2432893 GROW .1263626 .0367734 3.44 0.001 .0536939 .1990312 GOV -.1442231 .0420839 -3.43 0.001 -.227386 -.0610601 LEV -.2476965 .0806151 -3.07 0.003 -.4070017 -.0883912 CR -.0166686 .0104952 -1.59 0.114 -.0374083 .0040712 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10801 Adj R-squared = 0.2539 Residual 1.72671793 148 .011667013 R-squared = 0.2829 Model .681294262 6 .113549044 Prob > F = 0.0000 F( 6, 148) = 10.59 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LEV (CR=l.CR) GOV GROW TANG SIZE + Kết quả kiểm tra với biến độc lập GOV Do P-value = 1,0 lớn hơn 0,05 nên biến GOV là biến ngoại sinh. + Kết quả kiểm tra với biến độc lập GROW Do P-value = 0,01532 nhỏ hơn 0,05 nên biến GROW là biến nội sinh. Durbin-Wu-Hausman chi-sq test: 0.00000 Chi-sq(1) P-value = 1.00000 Wu-Hausman F test: 0.00000 F(1,147) P-value = 1.00000 H0: Regressor is exogenous Tests of endogeneity of: GOV . ivendog GOV Instruments: LEV CR GROW TANG SIZE L.GOV Instrumented: GOV _cons -1.342513 .2244345 -5.98 0.000 -1.786023 -.8990031 SIZE .0549644 .0082993 6.62 0.000 .0385639 .0713649 TANG .1428602 .0565148 2.53 0.013 .0311801 .2545403 GROW .1252568 .0364153 3.44 0.001 .0532957 .1972179 CR -.0036647 .007657 -0.48 0.633 -.0187959 .0114664 LEV -.1666226 .0667371 -2.50 0.014 -.2985033 -.0347419 GOV -.1486345 .0416115 -3.57 0.000 -.2308639 -.0664051 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10698 Adj R-squared = 0.2681 Residual 1.69371098 148 .011443993 R-squared = 0.2966 Model .714301213 6 .119050202 Prob > F = 0.0000 F( 6, 148) = 10.40 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LEV CR (GOV =l.GOV) GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 5.87929 Chi-sq(1) P-value = 0.01532 Wu-Hausman F test: 5.79568 F(1,147) P-value = 0.01731 H0: Regressor is exogenous Tests of endogeneity of: GROW . ivendog GROW Instruments: LEV CR GOV TANG SIZE L.GROW Instrumented: GROW _cons -1.350566 .2795867 -4.83 0.000 -1.903064 -.7980688 SIZE .0561461 .0103563 5.42 0.000 .0356808 .0766113 TANG .1734984 .0722138 2.40 0.018 .0307952 .3162016 GOV -.149723 .0518342 -2.89 0.004 -.2521538 -.0472922 CR -.0049016 .0095596 -0.51 0.609 -.0237926 .0139894 LEV -.2441052 .0925711 -2.64 0.009 -.427037 -.0611734 GROW .4542529 .1788067 2.54 0.012 .100909 .8075968 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .13325 Adj R-squared = . Residual 2.62780733 148 .017755455 R-squared = . Model -.219795141 6 -.036632523 Prob > F = 0.0000 F( 6, 148) = 6.51 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LEV CR GOV (GROW=l.GROW) TANG SIZE + Kết quả kiểm tra với biến độc lập SIZE Do P-value = 0,00119 nhỏ hơn 0,05 nên biến SIZE là biến nội sinh. + Kết quả kiểm tra với biến độc lập TANG Do P-value = 0,29057 lớn hơn 0,05 nên biến TANG là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 10.50164 Chi-sq(1) P-value = 0.00119 Wu-Hausman F test: 10.68345 F(1,147) P-value = 0.00135 H0: Regressor is exogenous Tests of endogeneity of: SIZE . ivendog SIZE Instruments: LEV CR GOV GROW TANG L.SIZE Instrumented: SIZE _cons -1.293263 .2250109 -5.75 0.000 -1.737913 -.8486143 TANG .1408005 .0565281 2.49 0.014 .029094 .2525069 GROW .1251282 .0364215 3.44 0.001 .0531548 .1971015 GOV -.1414161 .0416809 -3.39 0.001 -.2237827 -.0590494 CR -.0037179 .0076583 -0.49 0.628 -.0188516 .0114159 LEV -.1643319 .0667524 -2.46 0.015 -.2962427 -.0324211 SIZE .0531041 .0083215 6.38 0.000 .0366598 .0695484 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10699 Adj R-squared = 0.2679 Residual 1.69428595 148 .011447878 R-squared = 0.2964 Model .713726242 6 .118954374 Prob > F = 0.0000 F( 6, 148) = 9.88 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LEV CR GOV GROW TANG (SIZE=l.SIZE) Durbin-Wu-Hausman chi-sq test: 1.11697 Chi-sq(1) P-value = 0.29057 Wu-Hausman F test: 1.06701 F(1,147) P-value = 0.30332 H0: Regressor is exogenous Tests of endogeneity of: TANG . ivendog TANG Instruments: LEV CR GOV GROW SIZE L.TANG Instrumented: TANG _cons -1.323808 .225272 -5.88 0.000 -1.768973 -.8786426 SIZE .0544563 .0083179 6.55 0.000 .0380191 .0708934 GROW .124434 .0364414 3.41 0.001 .0524213 .1964468 GOV -.1487569 .0416316 -3.57 0.000 -.231026 -.0664879 CR -.0040622 .0076703 -0.53 0.597 -.0192197 .0110952 LEV -.1670931 .0667706 -2.50 0.013 -.29904 -.0351462 TANG .1215807 .0601827 2.02 0.045 .0026523 .2405091 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10703 Adj R-squared = 0.2674 Residual 1.69533344 148 .011454956 R-squared = 0.2960 Model .712678749 6 .118779791 Prob > F = 0.0000 F( 6, 148) = 10.01 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LEV CR GOV GROW (TANG=l.TANG) SIZE Phụ lục 14: Kết quả kiểm định tác động tuyến tính của SLEV tới tỷ suất sinh lời VCSH (ROE) - Mô hình OLS - Kiểm định hiện tượng đa cộng tuyến - Kiểm định hiện tượng phương sai thay đổi Do p-value = 0,0001 nhỏ hơn 0,05 => Có hiện tượng phương sai thay đổi _cons -1.307799 .2034466 -6.43 0.000 -1.709261 -.9063362 SIZE .0546388 .0076362 7.16 0.000 .0395703 .0697074 TANG .1009444 .0529087 1.91 0.058 -.0034607 .2053495 GROW .1578483 .0333746 4.73 0.000 .0919901 .2237066 GOV -.1474352 .0380125 -3.88 0.000 -.2224456 -.0724248 CR -.0040398 .0068984 -0.59 0.559 -.0176526 .0095729 SLEV -.2332051 .0584148 -3.99 0.000 -.3484754 -.1179349 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.99698509 185 .016199919 Root MSE = .10723 Adj R-squared = 0.2902 Residual 2.05819985 179 .011498323 R-squared = 0.3132 Model .938785241 6 .156464207 Prob > F = 0.0000 F( 6, 179) = 13.61 Source SS df MS Number of obs = 186 Mean VIF 1.94 GROW 1.04 0.965334 TANG 1.16 0.860973 CR 1.92 0.520477 SLEV 2.05 0.488232 GOV 2.63 0.380686 SIZE 2.82 0.354861 Variable VIF 1/VIF Total 99.42 33 0.0000 Kurtosis 2.37 1 0.1234 Skewness 28.19 6 0.0001 Heteroskedasticity 68.85 26 0.0000 Source chi2 df p Cameron & Trivedi's decomposition of IM-test Prob > chi2 = 0.0000 chi2(26) = 68.85 against Ha: unrestricted heteroskedasticity White's test for Ho: homoskedasticity - Kiểm định hiện tượng tự tương quan Do p-value = 0,0014 nhỏ hơn 0,05 => Có hiện tượng tự tương quan - Mô hình FEM - Mô hình REM Prob > F = 0.0014 F( 1, 30) = 12.345 H0: no first-order autocorrelation Wooldridge test for autocorrelation in panel data F test that all u_i=0: F(30, 150) = 7.58 Prob > F = 0.0000 rho .66234078 (fraction of variance due to u_i) sigma_e .07689913 sigma_u .10770177 _cons -1.505584 .7138739 -2.11 0.037 -2.916131 -.0950368 SIZE .0644429 .0262871 2.45 0.015 .0125021 .1163837 TANG -.0084231 .1358991 -0.06 0.951 -.2769469 .2601006 GROW .1195542 .0320525 3.73 0.000 .0562215 .1828869 GOV 0 (omitted) CR -.0095351 .0091316 -1.04 0.298 -.0275782 .0085081 SLEV -.3697677 .0974804 -3.79 0.000 -.5623798 -.1771557 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] corr(u_i, Xb) = -0.5327 Prob > F = 0.0002 F(5,150) = 5.33 overall = 0.2051 max = 6 between = 0.2553 avg = 6.0 R-sq: within = 0.1509 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Fixed-effects (within) regression Number of obs = 186 rho .52997541 (fraction of variance due to u_i) sigma_e .07689913 sigma_u .08165616 _cons -1.338968 .3541025 -3.78 0.000 -2.032997 -.6449402 SIZE .0574115 .0132793 4.32 0.000 .0313846 .0834385 TANG .0638489 .0848942 0.75 0.452 -.1025407 .2302386 GROW .1250699 .0295484 4.23 0.000 .0671561 .1829838 GOV -.1594924 .0699865 -2.28 0.023 -.2966634 -.0223215 CR -.0073305 .0078254 -0.94 0.349 -.0226679 .008007 SLEV -.3023243 .0755584 -4.00 0.000 -.450416 -.1542326 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(6) = 42.68 overall = 0.3016 max = 6 between = 0.3850 avg = 6.0 R-sq: within = 0.1468 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Random-effects GLS regression Number of obs = 186 - Kiểm định Hausman Do p-value = 0,7467 lớn hơn 0,05 => Lựa chọn là mô hình REM. - Kiểm định phương sai thay đổi Do p-value = 0,0000 nhỏ hơn 0,05 => Có hiện tượng phương sai thay đổi - Mô hình GLS Prob>chi2 = 0.7467 = 2.70 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg SIZE .0644429 .0574115 .0070314 .0226864 TANG -.0084231 .0638489 -.0722721 .1061204 GROW .1195542 .1250699 -.0055158 .0124198 CR -.0095351 -.0073305 -.0022046 .0047064 SLEV -.3697677 -.3023243 -.0674434 .0615902 FEM REM Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients Prob > chibar2 = 0.0000 chibar2(01) = 100.80 Test: Var(u) = 0 u .0066677 .0816562 e .0059135 .0768991 ROE .0161999 .1272789 Var sd = sqrt(Var) Estimated results: ROE[MACK,t] = Xb + u[MACK] + e[MACK,t] Breusch and Pagan Lagrangian multiplier test for random effects _cons -.9263552 .1724005 -5.37 0.000 -1.264254 -.5884563 SIZE .0407286 .0066496 6.12 0.000 .0276956 .0537617 TANG .0701348 .0349898 2.00 0.045 .001556 .1387136 GROW .0421343 .018268 2.31 0.021 .0063297 .0779388 GOV -.0994337 .0305826 -3.25 0.001 -.1593744 -.039493 CR -.0071516 .0046994 -1.52 0.128 -.0163623 .002059 SLEV -.2016365 .0409922 -4.92 0.000 -.2819797 -.1212933 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.0000 Wald chi2(6) = 48.56 Estimated coefficients = 7 Time periods = 6 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 186 Correlation: common AR(1) coefficient for all panels (0.4734) Panels: heteroskedastic Coefficients: generalized least squares Cross-sectional time-series FGLS regression - Mô hình GMM Difference (null H = exogenous): chi2(3) = 12.80 Prob > chi2 = 0.005 Sargan test excluding group: chi2(18) = 29.56 Prob > chi2 = 0.042 iv(CR GOV TANG) Difference (null H = exogenous): chi2(9) = 31.63 Prob > chi2 = 0.000 Sargan test excluding group: chi2(12) = 10.73 Prob > chi2 = 0.552 GMM instruments for levels Difference-in-Sargan tests of exogeneity of instrument subsets: (Not robust, but not weakened by many instruments.) Sargan test of overid. restrictions: chi2(21) = 42.37 Prob > chi2 = 0.004 Arellano-Bond test for AR(2) in first differences: z = -1.05 Pr > z = 0.293 Arellano-Bond test for AR(1) in first differences: z = -2.13 Pr > z = 0.033 DL.(L.SLEV L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) _cons CR GOV TANG Standard Instruments for levels equation L(2/3).(L.SLEV L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) D.(CR GOV TANG) Standard Instruments for first differences equation _cons -.7963793 .3274357 -2.43 0.015 -1.438141 -.1546171 SIZE .043875 .0116175 3.78 0.000 .021105 .0666449 TANG -.0016645 .0505796 -0.03 0.974 -.1007987 .0974698 GROW .2290531 .0553453 4.14 0.000 .1205783 .337528 GOV -.0929749 .048125 -1.93 0.053 -.1872983 .0013484 CR -.0349962 .0113192 -3.09 0.002 -.0571814 -.012811 SLEV -.5903064 .1396856 -4.23 0.000 -.8640852 -.3165276 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.000 max = 6 Wald chi2(6) = 59.78 avg = 6.00 Number of instruments = 28 Obs per group: min = 6 Time variable : YEAR Number of groups = 31 Group variable: MACK Number of obs = 186 Dynamic panel-data estimation, one-step system GMM - Kiểm tra hiện tượng nội sinh với biến phụ thuộc ROE + Kết quả kiểm tra với biến độc lập SLEV Do P-value = 0,00537 nhỏ hơn 0,05 nên biến SLEV là biến nội sinh. + Kết quả kiểm tra với biến độc lập CR Do P-value = 0,05803 lớn hơn 0,05 nên biến CR là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 7.75177 Chi-sq(1) P-value = 0.00537 Wu-Hausman F test: 7.73870 F(1,147) P-value = 0.00611 H0: Regressor is exogenous Tests of endogeneity of: SLEV . ivendog SLEV Instruments: CR GOV GROW TANG SIZE L.SLEV Instrumented: SLEV _cons -1.385781 .2208881 -6.27 0.000 -1.822283 -.9492789 SIZE .055348 .0083654 6.62 0.000 .038817 .071879 TANG .1256133 .0569231 2.21 0.029 .0131263 .2381003 GROW .1221647 .0358979 3.40 0.001 .051226 .1931035 GOV -.1503101 .0412512 -3.64 0.000 -.2318276 -.0687927 CR .0008038 .007832 0.10 0.918 -.0146732 .0162808 SLEV -.1249991 .0788141 -1.59 0.115 -.2807454 .0307472 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10512 Adj R-squared = 0.2933 Residual 1.63545599 148 .011050378 R-squared = 0.3208 Model .772556203 6 .128759367 Prob > F = 0.0000 F( 6, 148) = 10.12 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE (SLEV =l.SLEV) CR GOV GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 3.59275 Chi-sq(1) P-value = 0.05803 Wu-Hausman F test: 3.48817 F(1,147) P-value = 0.06380 H0: Regressor is exogenous Tests of endogeneity of: CR . ivendog CR Instruments: SLEV GOV GROW TANG SIZE L.CR Instrumented: CR _cons -1.359226 .2211286 -6.15 0.000 -1.796204 -.9222492 SIZE .0588834 .008211 7.17 0.000 .0426574 .0751095 TANG .0844509 .0572762 1.47 0.142 -.0287339 .1976357 GROW .1308951 .0355482 3.68 0.000 .0606475 .2011426 GOV -.1576952 .0408692 -3.86 0.000 -.2384577 -.0769327 SLEV -.3139232 .071747 -4.38 0.000 -.455704 -.1721424 CR -.0190113 .0090443 -2.10 0.037 -.036884 -.0011386 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10452 Adj R-squared = 0.3014 Residual 1.61676154 148 .010924064 R-squared = 0.3286 Model .79125065 6 .131875108 Prob > F = 0.0000 F( 6, 148) = 12.94 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE SLEV (CR=l.CR) GOV GROW TANG SIZE + Kết quả kiểm tra với biến độc lập GOV Do P-value = 1,0 lớn hơn 0,05 nên biến GOV là biến ngoại sinh. + Kết quả kiểm tra với biến độc lập GROW Do P-value = 0,02175 nhỏ hơn 0,05 nên biến GROW là biến nội sinh. Durbin-Wu-Hausman chi-sq test: 0.00000 Chi-sq(1) P-value = 1.00000 Wu-Hausman F test: 0.00000 F(1,147) P-value = 1.00000 H0: Regressor is exogenous Tests of endogeneity of: GOV . ivendog GOV Instruments: SLEV CR GROW TANG SIZE L.GOV Instrumented: GOV _cons -1.409198 .2177555 -6.47 0.000 -1.839509 -.9788861 SIZE .058928 .0081476 7.23 0.000 .0428273 .0750286 TANG .1055946 .0556755 1.90 0.060 -.0044271 .2156162 GROW .1308125 .0352736 3.71 0.000 .0611076 .2005174 CR -.0084628 .0069358 -1.22 0.224 -.0221687 .0052431 SLEV -.2508157 .0625099 -4.01 0.000 -.3743429 -.1272885 GOV -.1605028 .0405252 -3.96 0.000 -.2405855 -.0804201 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10371 Adj R-squared = 0.3121 Residual 1.59188198 148 .010755959 R-squared = 0.3389 Model .816130207 6 .136021701 Prob > F = 0.0000 F( 6, 148) = 12.65 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE SLEV CR (GOV =l.GOV) GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 5.26553 Chi-sq(1) P-value = 0.02175 Wu-Hausman F test: 5.16937 F(1,147) P-value = 0.02444 H0: Regressor is exogenous Tests of endogeneity of: GROW . ivendog GROW Instruments: SLEV CR GOV TANG SIZE L.GROW Instrumented: GROW _cons -1.436452 .2640859 -5.44 0.000 -1.958318 -.9145862 SIZE .0604654 .0099006 6.11 0.000 .0409006 .0800303 TANG .1239604 .0681419 1.82 0.071 -.0106964 .2586172 GOV -.1630919 .049091 -3.32 0.001 -.2601016 -.0660821 CR -.0085514 .0083985 -1.02 0.310 -.0251479 .0080451 SLEV -.3140634 .0830412 -3.78 0.000 -.4781631 -.1499637 GROW .4238197 .1638947 2.59 0.011 .0999437 .7476958 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .12558 Adj R-squared = -0.0086 Residual 2.33405803 148 .015770662 R-squared = 0.0307 Model .073954163 6 .012325694 Prob > F = 0.0000 F( 6, 148) = 8.18 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE SLEV CR GOV (GROW=l.GROW) TANG SIZE + Kết quả kiểm tra với biến độc lập SIZE Do P-value = 0,00145 nhỏ hơn 0,05 nên biến SIZE là biến nội sinh. + Kết quả kiểm tra với biến độc lập TANG Do P-value = 0,44767 lớn hơn 0,05 nên biến TANG là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 10.13659 Chi-sq(1) P-value = 0.00145 Wu-Hausman F test: 10.28609 F(1,147) P-value = 0.00165 H0: Regressor is exogenous Tests of endogeneity of: SIZE . ivendog SIZE Instruments: SLEV CR GOV GROW TANG L.SIZE Instrumented: SIZE _cons -1.361467 .218332 -6.24 0.000 -1.792917 -.9300157 TANG .1040825 .055687 1.87 0.064 -.0059618 .2141269 GROW .1306339 .0352795 3.70 0.000 .0609172 .2003505 GOV -.1534925 .0405946 -3.78 0.000 -.2337123 -.0732726 CR -.0084683 .0069369 -1.22 0.224 -.0221766 .0052399 SLEV -.247773 .062528 -3.96 0.000 -.3713361 -.1242099 SIZE .0571113 .0081699 6.99 0.000 .0409667 .0732559 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10373 Adj R-squared = 0.3119 Residual 1.59241672 148 .010759572 R-squared = 0.3387 Model .815595468 6 .135932578 Prob > F = 0.0000 F( 6, 148) = 12.07 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE SLEV CR GOV GROW TANG (SIZE=l.SIZE) Durbin-Wu-Hausman chi-sq test: 0.57653 Chi-sq(1) P-value = 0.44767 Wu-Hausman F test: 0.54882 F(1,147) P-value = 0.45998 H0: Regressor is exogenous Tests of endogeneity of: TANG . ivendog TANG Instruments: SLEV CR GOV GROW SIZE L.TANG Instrumented: TANG _cons -1.396952 .2184343 -6.40 0.000 -1.828605 -.9652993 SIZE .058659 .0081577 7.19 0.000 .0425384 .0747795 GROW .1304328 .035286 3.70 0.000 .0607033 .2001623 GOV -.1608157 .0405374 -3.97 0.000 -.2409226 -.0807088 CR -.0089322 .0069663 -1.28 0.202 -.0226985 .004834 SLEV -.2538427 .0626584 -4.05 0.000 -.3776635 -.130022 TANG .0905025 .0592897 1.53 0.129 -.0266612 .2076661 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10374 Adj R-squared = 0.3118 Residual 1.59267234 148 .0107613 R-squared = 0.3386 Model .815339855 6 .135889976 Prob > F = 0.0000 F( 6, 148) = 12.43 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE SLEV CR GOV GROW (TANG=l.TANG) SIZE Phụ lục 15: Kết quả kiểm định tác động tuyến tính của LLEV tới tỷ suất sinh lời VCSH (ROE) (mô hình 6) - Mô hình OLS - Kiểm định hiện tượng đa cộng tuyến - Kiểm định hiện tượng phương sai thay đổi Do p-value = 0,0000 nhỏ hơn 0,05 => Có hiện tượng phương sai thay đổi. _cons -1.339528 .209912 -6.38 0.000 -1.753749 -.9253079 SIZE .0501077 .0076522 6.55 0.000 .0350076 .0652079 TANG .1061781 .0547735 1.94 0.054 -.0019066 .2142629 GROW .1379675 .0338686 4.07 0.000 .0711344 .2048006 GOV -.1370819 .0386967 -3.54 0.001 -.2134423 -.0607216 CR .0165871 .0053404 3.11 0.002 .0060488 .0271254 LLEV .2770091 .1017879 2.72 0.007 .0761505 .4778678 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.99698509 185 .016199919 Root MSE = .10966 Adj R-squared = 0.2577 Residual 2.15240194 179 .012024592 R-squared = 0.2818 Model .844583149 6 .140763858 Prob > F = 0.0000 F( 6, 179) = 11.71 Source SS df MS Number of obs = 186 Mean VIF 1.63 GROW 1.02 0.980280 CR 1.10 0.908207 LLEV 1.14 0.875225 TANG 1.19 0.840117 GOV 2.60 0.384158 SIZE 2.71 0.369554 Variable VIF 1/VIF Total 96.15 33 0.0000 Kurtosis 2.03 1 0.1538 Skewness 26.87 6 0.0002 Heteroskedasticity 67.25 26 0.0000 Source chi2 df p Cameron & Trivedi's decomposition of IM-test Prob > chi2 = 0.0000 chi2(26) = 67.25 against Ha: unrestricted heteroskedasticity White's test for Ho: homoskedasticity - Kiểm định hiện tượng tự tương quan Do p-value = 0,0037 nhỏ hơn 0,05 => Có hiện tượng tự tương quan - Mô hình FEM - Mô hình REM Prob > F = 0.0004 F( 1, 30) = 15.584 H0: no first-order autocorrelation Wooldridge test for autocorrelation in panel data F test that all u_i=0: F(30, 150) = 7.06 Prob > F = 0.0000 rho .56297139 (fraction of variance due to u_i) sigma_e .07977965 sigma_u .09054832 _cons -1.090947 .7307255 -1.49 0.138 -2.534791 .3528973 SIZE .0410363 .0264501 1.55 0.123 -.0112265 .0932991 TANG .0809629 .1397827 0.58 0.563 -.1952344 .3571602 GROW .0838297 .0317397 2.64 0.009 .021115 .1465443 GOV 0 (omitted) CR .0091387 .007854 1.16 0.246 -.0063801 .0246576 LLEV .2776268 .1679595 1.65 0.100 -.0542452 .6094989 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] corr(u_i, Xb) = -0.1681 Prob > F = 0.0181 F(5,150) = 2.83 overall = 0.2029 max = 6 between = 0.2676 avg = 6.0 R-sq: within = 0.0861 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Fixed-effects (within) regression Number of obs = 186 rho .51381824 (fraction of variance due to u_i) sigma_e .07977965 sigma_u .08201581 _cons -1.254507 .3586086 -3.50 0.000 -1.957367 -.5516468 SIZE .0473334 .01315 3.60 0.000 .0215599 .0731068 TANG .0983094 .0868978 1.13 0.258 -.0720073 .268626 GROW .093262 .029352 3.18 0.001 .0357332 .1507908 GOV -.1261546 .0701075 -1.80 0.072 -.2635629 .0112536 CR .0119654 .0064503 1.86 0.064 -.000677 .0246077 LLEV .2725342 .132303 2.06 0.039 .0132251 .5318433 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000 Wald chi2(6) = 29.87 overall = 0.2751 max = 6 between = 0.3795 avg = 6.0 R-sq: within = 0.0856 Obs per group: min = 6 Group variable: MACK Number of groups = 31 Random-effects GLS regression Number of obs = 186 - Kiểm định Hausman Do p-value = 0,8381 lớn hơn 0,05 => lựa chọn là mô hình REM. - Kiểm định phương sai thay đổi của mô hình Do p-value = 0,000 nhỏ hơn 0,05 => Có hiện tượng phương sai thay đổi - Mô hình GLS Prob>chi2 = 0.8381 = 2.08 chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg SIZE .0410363 .0473334 -.0062971 .0229496 TANG .0809629 .0983094 -.0173465 .1094895 GROW .0838297 .093262 -.0094323 .0120777 CR .0091387 .0119654 -.0028266 .004481 LLEV .2776268 .2725342 .0050927 .1034713 FEM REM Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients Prob > chibar2 = 0.0000 chibar2(01) = 95.28 Test: Var(u) = 0 u .0067266 .0820158 e .0063648 .0797797 ROE .0161999 .1272789 Var sd = sqrt(Var) Estimated results: ROE[MACK,t] = Xb + u[MACK] + e[MACK,t] Breusch and Pagan Lagrangian multiplier test for random effects _cons -.9551869 .1854934 -5.15 0.000 -1.318747 -.5916265 SIZE .0368312 .0066866 5.51 0.000 .0237257 .0499366 TANG .1186686 .0346368 3.43 0.001 .0507817 .1865555 GROW .018178 .0172348 1.05 0.292 -.0156015 .0519576 GOV -.1073717 .028351 -3.79 0.000 -.1629386 -.0518047 CR .0101613 .0047794 2.13 0.033 .0007938 .0195287 LLEV .1580683 .0640931 2.47 0.014 .0324481 .2836885 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.0000 Wald chi2(6) = 41.97 Estimated coefficients = 7 Time periods = 6 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 186 Correlation: common AR(1) coefficient for all panels (0.4584) Panels: heteroskedastic Coefficients: generalized least squares Cross-sectional time-series FGLS regression - Mô hình GMM Difference (null H = exogenous): chi2(4) = 6.97 Prob > chi2 = 0.137 Sargan test excluding group: chi2(6) = 11.70 Prob > chi2 = 0.069 iv(LLEV CR GOV TANG) Difference (null H = exogenous): chi2(6) = 15.84 Prob > chi2 = 0.015 Sargan test excluding group: chi2(4) = 2.83 Prob > chi2 = 0.586 GMM instruments for levels Difference-in-Sargan tests of exogeneity of instrument subsets: (Not robust, but not weakened by many instruments.) Sargan test of overid. restrictions: chi2(10) = 18.67 Prob > chi2 = 0.045 Arellano-Bond test for AR(2) in first differences: z = -1.74 Pr > z = 0.082 Arellano-Bond test for AR(1) in first differences: z = -2.34 Pr > z = 0.019 DL.(L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) _cons LLEV CR GOV TANG Standard Instruments for levels equation L2.(L.GROW L.SIZE) GMM-type (missing=0, separate instruments for each period unless collapsed) D.(LLEV CR GOV TANG) Standard Instruments for first differences equation _cons -1.496699 .5863232 -2.55 0.011 -2.645871 -.3475264 SIZE .0554342 .0215363 2.57 0.010 .0132239 .0976446 TANG .1558134 .0557725 2.79 0.005 .0465014 .2651255 GROW .3203796 .0909025 3.52 0.000 .142214 .4985453 GOV -.1514176 .0867085 -1.75 0.081 -.3213631 .0185278 CR .0173756 .0054357 3.20 0.001 .0067218 .0280294 LLEV .2600849 .0893844 2.91 0.004 .0848948 .4352751 ROE Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.000 max = 6 Wald chi2(6) = 35.62 avg = 6.00 Number of instruments = 17 Obs per group: min = 6 Time variable : YEAR Number of groups = 31 Group variable: MACK Number of obs = 186 Dynamic panel-data estimation, one-step system GMM - Kiểm tra hiện tượng nội sinh với biến phụ thuộc ROE + Kết quả kiểm tra với biến độc lập LLEV Do P-value = 0,37605 lớn hơn 0,05 nên biến LLEV là biến ngoại sinh. + Kết quả kiểm tra với biến độc lập CR Do P-value = 0,23510 lớn hơn 0,05 nên biến CR là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 0.78357 Chi-sq(1) P-value = 0.37605 Wu-Hausman F test: 0.74690 F(1,147) P-value = 0.38887 H0: Regressor is exogenous Tests of endogeneity of: LLEV . ivendog LLEV Instruments: CR GOV GROW TANG SIZE L.LLEV Instrumented: LLEV _cons -1.458269 .2275748 -6.41 0.000 -1.907985 -1.008554 SIZE .054734 .0082892 6.60 0.000 .0383536 .0711143 TANG .1006796 .0591602 1.70 0.091 -.0162283 .2175874 GROW .1131403 .0361551 3.13 0.002 .0416933 .1845873 GOV -.1496557 .041682 -3.59 0.000 -.2320245 -.0672869 CR .0126444 .0054526 2.32 0.022 .0018694 .0234194 LLEV .3129275 .1208844 2.59 0.011 .0740451 .55181 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .1071 Adj R-squared = 0.2664 Residual 1.69759681 148 .011470249 R-squared = 0.2950 Model .710415386 6 .118402564 Prob > F = 0.0000 F( 6, 148) = 10.46 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE (LLEV =l.LLEV) CR GOV GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 1.40974 Chi-sq(1) P-value = 0.23510 Wu-Hausman F test: 1.34925 F(1,147) P-value = 0.24729 H0: Regressor is exogenous Tests of endogeneity of: CR . ivendog CR Instruments: LLEV GOV GROW TANG SIZE L.CR Instrumented: CR _cons -1.401705 .2300639 -6.09 0.000 -1.856339 -.9470701 SIZE .0530306 .0083521 6.35 0.000 .0365258 .0695354 TANG .1023603 .0588597 1.74 0.084 -.0139537 .2186743 GROW .1105846 .0362608 3.05 0.003 .0389289 .1822403 GOV -.1429349 .0419387 -3.41 0.001 -.2258109 -.060059 LLEV .2531314 .1084132 2.33 0.021 .0388937 .4673691 CR .0084824 .0063351 1.34 0.183 -.0040365 .0210012 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .1072 Adj R-squared = 0.2650 Residual 1.7008887 148 .011492491 R-squared = 0.2937 Model .707123486 6 .117853914 Prob > F = 0.0000 F( 6, 148) = 9.79 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LLEV (CR=l.CR) GOV GROW TANG SIZE + Kết quả kiểm tra với biến độc lập GOV Do P-value = 1,0 lớn hơn 0,05 nên biến GOV là biến ngoại sinh. + Kết quả kiểm tra với biến độc lập GROW Do P-value = 0,03359 lớn hơn 0,05 nên biến GROW là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 0.00000 Chi-sq(1) P-value = 1.00000 Wu-Hausman F test: 0.00000 F(1,147) P-value = 1.00000 H0: Regressor is exogenous Tests of endogeneity of: GOV . ivendog GOV Instruments: LLEV CR GROW TANG SIZE L.GOV Instrumented: GOV _cons -1.443787 .2268093 -6.37 0.000 -1.89199 -.9955839 SIZE .0542889 .0082678 6.57 0.000 .0379508 .070627 TANG .1074589 .0585999 1.83 0.069 -.0083416 .2232594 GROW .1132057 .0361315 3.13 0.002 .0418055 .184606 CR .012246 .0054296 2.26 0.026 .0015165 .0229755 LLEV .2655978 .107703 2.47 0.015 .0527636 .4784321 GOV -.1482231 .0416219 -3.56 0.000 -.2304731 -.0659731 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10703 Adj R-squared = 0.2674 Residual 1.69538463 148 .011455302 R-squared = 0.2959 Model .712627559 6 .11877126 Prob > F = 0.0000 F( 6, 148) = 10.37 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LLEV CR (GOV =l.GOV) GROW TANG SIZE Durbin-Wu-Hausman chi-sq test: 4.51541 Chi-sq(1) P-value = 0.03359 Wu-Hausman F test: 4.41085 F(1,147) P-value = 0.03742 H0: Regressor is exogenous Tests of endogeneity of: GROW . ivendog GROW Instruments: LLEV CR GOV TANG SIZE L.GROW Instrumented: GROW _cons -1.457728 .2689029 -5.42 0.000 -1.989113 -.9263431 SIZE .0540027 .0097992 5.51 0.000 .0346382 .0733672 TANG .1354799 .0712627 1.90 0.059 -.0053439 .2763036 GOV -.1456646 .0493466 -2.95 0.004 -.2431795 -.0481498 CR .016642 .0069063 2.41 0.017 .0029944 .0302896 LLEV .2621624 .1276507 2.05 0.042 .009909 .5144159 GROW .3927295 .1651789 2.38 0.019 .0663157 .7191432 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .12684 Adj R-squared = -0.0289 Residual 2.38098664 148 .016087748 R-squared = 0.0112 Model .027025552 6 .004504259 Prob > F = 0.0000 F( 6, 148) = 7.16 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LLEV CR GOV (GROW=l.GROW) TANG SIZE + Kết quả kiểm tra với biến độc lập SIZE Do P-value = 0,00065 nhỏ hơn 0,05 nên biến SIZE là biến nội sinh. + Kết quả kiểm tra với biến độc lập TANG Do P-value = 0,30036 lớn hơn 0,05 nên biến TANG là biến ngoại sinh. Durbin-Wu-Hausman chi-sq test: 11.63334 Chi-sq(1) P-value = 0.00065 Wu-Hausman F test: 11.92816 F(1,147) P-value = 0.00072 H0: Regressor is exogenous Tests of endogeneity of: SIZE . ivendog SIZE Instruments: LLEV CR GOV GROW TANG L.SIZE Instrumented: SIZE _cons -1.390528 .2274142 -6.11 0.000 -1.839926 -.9411296 TANG .1056903 .0586133 1.80 0.073 -.0101367 .2215173 GROW .1132438 .0361383 3.13 0.002 .0418302 .1846575 GOV -.1406277 .0416921 -3.37 0.001 -.2230163 -.058239 CR .0119651 .0054312 2.20 0.029 .0012323 .0226979 LLEV .2624893 .1077272 2.44 0.016 .0496071 .4753714 SIZE .0523409 .0082899 6.31 0.000 .035959 .0687229 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10705 Adj R-squared = 0.2671 Residual 1.69602054 148 .011459598 R-squared = 0.2957 Model .711991648 6 .118665275 Prob > F = 0.0000 F( 6, 148) = 9.82 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LLEV CR GOV GROW TANG (SIZE=l.SIZE) Durbin-Wu-Hausman chi-sq test: 1.07259 Chi-sq(1) P-value = 0.30036 Wu-Hausman F test: 1.02432 F(1,147) P-value = 0.31316 H0: Regressor is exogenous Tests of endogeneity of: TANG . ivendog TANG Instruments: LLEV CR GOV GROW SIZE L.TANG Instrumented: TANG _cons -1.42896 .2273922 -6.28 0.000 -1.878315 -.9796053 SIZE .0538877 .0082813 6.51 0.000 .0375229 .0702525 GROW .1123597 .0361587 3.11 0.002 .0409057 .1838137 GOV -.148644 .0416442 -3.57 0.000 -.230938 -.06635 CR .0119878 .0054382 2.20 0.029 .0012413 .0227343 LLEV .2763392 .1082772 2.55 0.012 .0623703 .4903082 TANG .0852591 .0626015 1.36 0.175 -.0384491 .2089674 ROE Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.40801219 154 .015636443 Root MSE = .10708 Adj R-squared = 0.2667 Residual 1.69702866 148 .01146641 R-squared = 0.2953 Model .710983531 6 .118497255 Prob > F = 0.0000 F( 6, 148) = 10.11 Source SS df MS Number of obs = 155 Instrumental variables (2SLS) regression . ivreg ROE LLEV CR GOV GROW (TANG=l.TANG) SIZE Phụ lục 16: Kết quả kiểm định tác động tuyến tính của DR tới WACC (mô hình 7) - Mô hình OLS - Kiểm định hiện tượng đa cộng tuyến - Kiểm định hiện tượng phương sai thay đổi Do p-value = 0,0000 nhỏ hơn 0,05 => Có hiện tượng phương sai thay đổi - Kiểm định hiện tượng tự tương quan _cons .0198625 .0035355 5.62 0.000 .0128862 .0268388 RISK .0960591 .0449033 2.14 0.034 .0074545 .1846637 GROW -.0004234 .0052691 -0.08 0.936 -.0108205 .0099737 TANG .0150826 .0079262 1.90 0.059 -.0005577 .0307229 CR .0070981 .0009205 7.71 0.000 .0052817 .0089144 DR .0137375 .0069385 1.98 0.049 .0000463 .0274288 WACC Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .076809141 185 .000415185 Root MSE = .01687 Adj R-squared = 0.3146 Residual .051224112 180 .000284578 R-squared = 0.3331 Model .025585029 5 .005117006 Prob > F = 0.0000 F( 5, 180) = 17.98 Source SS df MS Number of obs = 186 Mean VIF 1.22 GROW 1.04 0.958537 TANG 1.05 0.949464 DR 1.28 0.782759 RISK 1.33 0.750546 CR 1.38 0.723483 Variable VIF 1/VIF Total 68.74 26 0.0000 Kurtosis 3.01 1 0.0830 Skewness 8.42 5 0.1346 Heteroskedasticity 57.31 20 0.0000 Source chi2 df p Cameron & Trivedi's decomposition of IM-test Prob > chi2 = 0.0000 chi2(20) = 57.31 against Ha: unrestricted heteroskedasticity White's test for Ho: homoskedasticity Prob > F = 0.0306 F( 1, 30) = 5.153 H0: no first-order autocorrelation Wooldridge test for autocorrelation in panel data Do p-value = 0,0306 nhỏ hơn 0,05 => Có hiện tượng tự tương quan - Mô hình FEM - Mô hình REM - Kiểm định Hausman Do p-value = 0,0187 nhỏ hơn 0,05 => Lựa chọn mô hình FEM - Kiểm định phương sai thay đổi của mô hình Do p-value = 0,00 nhỏ hơn 0,05 => Có hiện tượng phương sai sai số thay đổi - Mô hình GLS Prob>chi2 = 0.0187 = 11.82 chi2(4) = (b-B)'[(V_b-V_B)^(-1)](b-B) Test: Ho: difference in coefficients not systematic B = inconsistent under Ha, efficient under Ho; obtained from xtreg b = consistent under Ho and Ha; obtained from xtreg GROW -.0053177 -.0050655 -.0002522 .0011327 TANG .0274943 .0198168 .0076775 .0136797 CR .0028267 .0038993 -.0010726 .0004513 DR -.0064066 .0012576 -.0076642 .0090016 FEM REM Difference S.E. (b) (B) (b-B) sqrt(diag(V_b-V_B)) Coefficients . hausman FEM REM Prob>chi2 = 0.0000 chi2 (31) = 1445.93 H0: sigma(i)^2 = sigma^2 for all i in fixed effect regression model Modified Wald test for groupwise heteroskedasticity _cons .020757 .0032378 6.41 0.000 .014411 .0271031 RISK .1209279 .0547299 2.21 0.027 .0136593 .2281965 GROW -.0048506 .0022325 -2.17 0.030 -.0092263 -.000475 TANG .0085904 .0066287 1.30 0.195 -.0044016 .0215824 CR .0050646 .0010746 4.71 0.000 .0029584 .0071708 DR .014762 .0055899 2.64 0.008 .003806 .0257179 WACC Coef. Std. Err. z P>|z| [95% Conf. Interval] Prob > chi2 = 0.0000 Wald chi2(5) = 39.73 Estimated coefficients = 6 Time periods = 6 Estimated autocorrelations = 1 Number of groups = 31 Estimated covariances = 31 Number of obs = 186 Correlation: common AR(1) coefficient for all panels (0.5192) Panels: heteroskedastic Coefficients: generalized least squares Cross-sectional time-series FGLS regression Phụ lục 17: Hệ số Z’’ của các DN thương mại dầu khí trong mẫu nghiên cứu STT Mã CK 2014 2015 2016 2017 2018 2019 1 ASP 7,72 5,88 3,78 0,71 0,97 0,51 2 CNG 12,74 10,21 15,91 449,03 1158,12 - 3 COM - - - - - 36,40 4 GAS 11,91 14,40 10,88 10,30 16,46 26,70 5 HFC 1,63 1,09 0,35 0,11 0,28 0,01 6 HTC - - - 3,49 4,62 6,86 7 MTG -0,59 0,31 1,43 -1,84 - - 8 OIL 3,62 5,34 5,64 9,18 6,90 4,32 9 PCG - - - - - 13,18 10 PCN - - 16,93 - - - 11 PCT 31,16 - - - - - 12 PDT -0,62 -0,76 -0,59 -0,26 0,23 0,00 13 PEG - - - - - - 14 PGC 1,55 -0,02 2,13 2,22 3,27 3,74 15 PGD - 9,11 11,13 - - - 16 PGS 2,04 1,85 4,61 3,83 52,32 4,25 17 PIT 1,57 1,03 1,55 -0,13 -0,37 1,75 18 PLX 1,34 4,11 11,72 4,22 4,78 4,56 19 PMG 2,47 1,31 3,27 2,46 0,75 1,38 20 PND 8,06 10,41 2,93 2,45 26,07 0,84 21 POV 6,59 5,72 6,14 4,39 3,78 8,22 22 PPY - - - - - 14,89 23 PSC 2,15 2,62 2,68 2,56 2,14 1,77 24 PTX -0,42 -0,20 -0,57 -0,79 -1,64 -1,01 25 PVG 1,17 0,72 -0,11 0,00 0,27 3,92 26 PVO 5,80 - 15,49 17,04 - - 27 SFC 3,25 3,23 3,38 3,11 - 19,26 28 TDG 1,91 6,77 6,00 5,22 4,31 2,40 29 TLP 0,24 0,49 1,54 1,11 1,17 0,58 30 TMC 28,46 6,76 77,64 4,84 5,05 6,48 31 VMG 13,40 31,41 - -1,08 -0,30 -1,38 Ghi chú: “-” thể hiện các DN thương mại dầu khí không sử dụng nợ vay