Phân tích sự biến động và xây dựng một danh mục đầu tư ­ cho các cổ phiếu của cùng một ngành

R-squared 0.019372 S.D. dependent var 0.021942 S.E. of regression 0.021728 Akaike info criterion -4.818908 Sum squared resid 0.312066 Schwarz criterion -4.812118 Log likelihood 1596.059 Durbin-Watson stat 1.566997 VËy ta cã m« h×nh CAPM cña cæ phiÕu TRI lµ ; Estimation Equation: ===================== R_TRI = C(1)*R_VNINDEX Substituted Coefficients: ===================== R_TRI = 0.2154451701*R_VNINDEX 5. M« h×nh chØ sè ®¬n-M« h×nh chØ sè thÞ tr­êng (SIM) 5.1.M« h×nh + M« h×nh chØ sè thÞ tr­êng lµ m« h×nh ®¬n gi¶n gièng m« h×nh håi quy ®¬n. + ë ®©y chóng ta sö dông chØ sè VN-INDEX lµm chØ sè thÞ tr­êng, nã ph¶n ¸nh ho¹t ®éng chung cña toµn bé thÞ tr­êng. + Ký hiÖu lîi suÊt cña VN-INDEX lµ RI . + Khi ®ã ta sÏ biÓu diÔn 6 cæ phiÕu theo m« h×nh chØ sè thÞ tr­êng vµ ­íc l­îng chóng. +Ta cã m« h×nh nh­ sau : Ri = + . + Rj = + . + §­êng håi quy nµy lµ ®­êng ®Æc tr­ng cña mçi lo¹i cæ phiÕu. E() = E() = 0 i ,j = COV( ,) = 0 i = COV(,) = 0 i j §iÒu kiÖn ®Ó ­íc l­îng m« h×nh chØ sè thÞ tr­êng theo ph­¬ng ph¸p OLS lµ lîi suÊt cña cæ phiÕu vµ lîi suÊt cña VNINDEX lµ ph©n phèi chuÈn nh­ng do thÞ tr­êng chøng kho¸n ViÖt Nam lµ thÞ tr­êng kh«ng hiÖu qu¶ ,gi¸ c¶ cæ phiÕu ch­a ph¶n ¸nh ®Çy ®ñ c¸c th«ng tin mµ c«ng chóng cã thÓ sö dông ,th«ng tin bÊt c©n x­ng .Trong qu¸ tr×nh tÝnh to¸n chØ sè, bªn c¹nh sù biÕn ®éng vÒ gi¸ cæ phiÕu lµm thay ®æi gi¸ trÞ chØ sè thÞ tr­êng cßn cã mét sè nh©n tè kh¸c lµm thay ®æi c¬ cÊu sè cæ phiÕu niªm yÕt nh­ thªm bít, t¸ch rêi, gép cæ phiÕu. Trong tr­êng hîp nµy sÏ ¶nh h­ëng ®Õn tÝnh kh«ng liªn tôc cña chØ sè, nghÜa lµ chØ sè ngµy b¸o c¸o kh«ng ®ång nhÊt víi ngµy tr­íc ®ã. Do vËy kÕt qu¶ ­íc l­îng cã thÓ kh«ng ph¶n ¸nh ®óng th«ng tin vµ cã thÓ m« h×nh cã nhiÒu khuyÕt tËt. 5.2. ¦íc l­îng m« h×nh chØ sè thÞ tr­êng 5.2.1. Cæ phiÕu BBC *M« h×nh : B¶ng 42 Dependent Variable: R_BBC01 Method: Least Squares Date: 04/15/07 Time: 01:52 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C -0.001109 0.000816 -1.358962 0.1746 R_VNINDEX 0.277686 0.057515 4.828054 0.0000 R-squared 0.034114 Mean dependent var -0.001518 Adjusted R-squared 0.032650 S.D. dependent var 0.021242 S.E. of regression 0.020892 Akaike info criterion -4.895897 Sum squared resid 0.288070 Schwarz criterion -4.882316 Log likelihood 1622.542 F-statistic 23.31010 Durbin-Watson stat 1.660072 Prob(F-statistic) 0.000002 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ EBBC : 5.2.1.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 43 White Heteroskedasticity Test: F-statistic 46.82416 Probability 0.000000 Obs*R-squared 82.36939 Probability 0.000000 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/15/07 Time: 02:22 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.000304 3.33E-05 9.121438 0.0000 R_VNINDEX 0.003948 0.002185 1.807106 0.0712 R_VNINDEX^2 0.680409 0.070312 9.677065 0.0000 R-squared 0.124425 Mean dependent var 0.000435 Adjusted R-squared 0.121768 S.D. dependent var 0.000831 S.E. of regression 0.000779 Akaike info criterion -11.47300 Sum squared resid 0.000400 Schwarz criterion -11.45263 Log likelihood 3800.563 F-statistic 46.82416 Durbin-Watson stat 1.543024 Prob(F-statistic) 0.000000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.0000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi. MÆt kh¸c ta cã P-value cña thèng kª F = 0.0000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi. 5.2.1.2. KiÓm ®Þnh sù t­¬ng quan *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : * Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 44 Breusch-Godfrey Serial Correlation LM Test: F-statistic 19.80128 Probability 0.000010 Obs*R-squared 19.31117 Probability 0.000011 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/15/07 Time: 06:23 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.025056 0.056992 -0.439639 0.6603 C -3.72E-05 0.000805 -0.046191 0.9632 RESID(-1) 0.171635 0.038571 4.449862 0.0000 R-squared 0.029171 Mean dependent var 6.53E-19 Adjusted R-squared 0.026225 S.D. dependent var 0.020876 S.E. of regression 0.020601 Akaike info criterion -4.922481 Sum squared resid 0.279667 Schwarz criterion -4.902110 Log likelihood 1632.341 F-statistic 9.900638 Durbin-Watson stat 1.982331 Prob(F-statistic) 0.000058 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.000010 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1 . MÆt kh¸c ta cã P-value cña thèng kª F = 0.000011 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1 . 5.2.1.3. KiÓm ®Þnh d¹ng hµm *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *Gi¶ thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 45 Ramsey RESET Test: F-statistic 0.444916 Probability 0.504993 Log likelihood ratio 0.446791 Probability 0.503863 Test Equation: Dependent Variable: R_BBC01 Method: Least Squares Date: 04/15/07 Time: 06:54 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.260119 0.063280 4.110609 0.0000 C -0.000847 0.000907 -0.933991 0.3507 FITTED^2 -16.32180 24.46973 -0.667020 0.5050 R-squared 0.034765 Mean dependent var -0.001518 Adjusted R-squared 0.031836 S.D. dependent var 0.021242 S.E. of regression 0.020901 Akaike info criterion -4.893551 Sum squared resid 0.287876 Schwarz criterion -4.873180 Log likelihood 1622.765 F-statistic 11.86771 Durbin-Watson stat 1.664422 Prob(F-statistic) 0.000009 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.503863 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . P-value cña thèng kª F =0.504993 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . 5.2.1.4. Söa m« h×nh a/ Kh¾c phôc tù t­¬ng quan b»ng Durbin hai b­íc : B­íc 1: Ta ®i ­íc l­îng m« h×nh : Theo m« h×nh ban ®Çu ta cã d = 1.660072 vËy =1-d/2 = 0.169964 B­íc 2: ta ®i ­íc l­îng m« h×nh : = + ( - ) + Ta ®Æt : M = (r_bbc - 0.169964*r_bbc(-1)) N = (r_vnindex - 0.169964*r_vnindex(-1)). B¶ng 46 Dependent Variable: M Method: Least Squares Date: 04/15/07 Time: 07:48 Sample(adjusted): 3 663 Included observations: 661 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. N 0.218786 0.057927 3.776890 0.0002 C -0.000995 0.000804 -1.237642 0.2163 R-squared 0.021188 Mean dependent var -0.001263 Adjusted R-squared 0.019702 S.D. dependent var 0.020793 S.E. of regression 0.020587 Akaike info criterion -4.925260 Sum squared resid 0.279310 Schwarz criterion -4.911663 Log likelihood 1629.798 F-statistic 14.26490 Durbin-Watson stat 1.965631 Prob(F-statistic) 0.000173 b/ KiÓm ®Þnh sù tù t­¬ng quan m« h×nh trªn : +Dïng kiÓm ®Þnh BG . +Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 47 Breusch-Godfrey Serial Correlation LM Test: F-statistic 0.158025 Probability 0.691110 Obs*R-squared 0.158708 Probability 0.690349 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/15/07 Time: 08:02 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. N -0.002127 0.058211 -0.036536 0.9709 C -2.66E-06 0.000804 -0.003309 0.9974 RESID(-1) 0.015561 0.039146 0.397524 0.6911 R-squared 0.000240 Mean dependent var -7.41E-19 Adjusted R-squared -0.002799 S.D. dependent var 0.020572 S.E. of regression 0.020601 Akaike info criterion -4.922474 Sum squared resid 0.279242 Schwarz criterion -4.902079 Log likelihood 1629.878 F-statistic 0.079013 Durbin-Watson stat 1.995386 Prob(F-statistic) 0.924037 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.690349 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.691110 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. c/ KiÓm ®Þnh ph­¬ng sai ®ång ®Òu cña m« h×nh trªn dùa vµo kiÓm ®Þnh White B¶ng 48 White Heteroskedasticity Test: F-statistic 0.27921 Probability 0.056096 Obs*R-squared 0.68914 Probability 0.058316 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/15/07 Time: 08:23 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. C 0.000301 3.30E-05 9.139581 0.0000 N 0.003831 0.002210 1.733612 0.0835 N^2 0.654317 0.072013 9.086079 0.0000 R-squared 0.000481 Mean dependent var 0.000423 Adjusted R-squared 0.108781 S.D. dependent var 0.000817 S.E. of regression 0.000772 Akaike info criterion -11.49176 Sum squared resid 0.000392 Schwarz criterion -11.47136 Log likelihood 3801.026 F-statistic 41.27921 Durbin-Watson stat 1.530943 Prob(F-statistic) 0.000000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.058316 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . MÆt kh¸c ta cã P-value cña thèng kª F = 0.056096 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. d/ KiÓm ®Þnh d¹ng hµm cña m« h×nh trªn. B¶ng 49 Ramsey RESET Test: F-statistic 0.044496 Probability 0.833000 Log likelihood ratio 0.044697 Probability 0.832562 Test Equation: Dependent Variable: M Method: Least Squares Date: 04/15/07 Time: 08:34 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. N 0.224796 0.064593 3.480191 0.0005 C -0.001079 0.000897 -1.202535 0.2296 FITTED^2 8.473683 40.17111 0.210940 0.8330 R-squared 0.021254 Mean dependent var -0.001263 Adjusted R-squared 0.018279 S.D. dependent var 0.020793 S.E. of regression 0.020602 Akaike info criterion -4.922302 Sum squared resid 0.279291 Schwarz criterion -4.901907 Log likelihood 1629.821 F-statistic 7.144355 Durbin-Watson stat 1.964392 Prob(F-statistic) 0.000852 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.832562 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . P-value cña thèng kª F =0.833000 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . VËy m« h×nh : M = 0.218786*N - 0.000995 kh«ng cßn khuyÕt tËt n÷a Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ : RBBC = - 0.000995 + 0.218786.RI 5.2.2. Cæ phiÕu CAN *M« h×nh : B¶ng 50 Dependent Variable: R_CAN Method: Least Squares Date: 04/15/07 Time: 19:04 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C -0.000399 0.000759 -0.525706 0.5993 R_VNINDEX 0.244813 0.053449 4.580334 0.0000 R-squared 0.030808 Mean dependent var -0.000759 Adjusted R-squared 0.029339 S.D. dependent var 0.019706 S.E. of regression 0.019415 Akaike info criterion -5.042540 Sum squared resid 0.248778 Schwarz criterion -5.028959 Log likelihood 1671.081 F-statistic 20.97946 Durbin-Watson stat 1.708570 Prob(F-statistic) 0.000006 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ ECAN : 5.2.2.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 51 White Heteroskedasticity Test: F-statistic 60.02966 Probability 0.000000 Obs*R-squared 102.0195 Probability 0.000000 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/15/07 Time: 19:10 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.000252 2.72E-05 9.246466 0.0000 R_VNINDEX 0.001304 0.001786 0.730165 0.4655 R_VNINDEX^2 0.624806 0.057482 10.86954 0.0000 R-squared 0.154108 Mean dependent var 0.000376 Adjusted R-squared 0.151541 S.D. dependent var 0.000691 S.E. of regression 0.000637 Akaike info criterion -11.87592 Sum squared resid 0.000267 Schwarz criterion -11.85555 Log likelihood 3933.929 F-statistic 60.02966 Durbin-Watson stat 1.504396 Prob(F-statistic) 0.000000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.0000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi . MÆt kh¸c ta cã P-value cña thèng kª F = 0.0000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi . 5.2.2.2.2 KiÓm ®Þnh sù t­¬ng quan : *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : *GØa thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 52 Breusch-Godfrey Serial Correlation LM Test: F-statistic 14.65332 Probability 0.000142 Obs*R-squared 14.39984 Probability 0.000148 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/15/07 Time: 19:18 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. C -4.92E-05 0.000751 -0.065449 0.9478 R_VNINDEX -0.032376 0.053576 -0.604307 0.5458 RESID(-1) 0.149372 0.039021 3.827966 0.0001 R-squared 0.021752 Mean dependent var -8.15E-19 Adjusted R-squared 0.018783 S.D. dependent var 0.019400 S.E. of regression 0.019217 Akaike info criterion -5.061511 Sum squared resid 0.243367 Schwarz criterion -5.041140 Log likelihood 1678.360 F-statistic 7.326661 Durbin-Watson stat 1.970836 Prob(F-statistic) 0.000713 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.000148 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.000142 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. 5.2.2.3. KiÓm ®Þnh d¹ng hµm : *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *GØa thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 53 Ramsey RESET Test: F-statistic 1.178408 Probability 0.278077 Log likelihood ratio 1.182715 Probability 0.276804 Test Equation: Dependent Variable: R_CAN Method: Least Squares Date: 04/15/07 Time: 19:29 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C -2.72E-05 0.000832 -0.032682 0.9739 R_VNINDEX 0.227263 0.055834 4.070368 0.0001 FITTED^2 -31.74158 29.24023 -1.085545 0.2781 R-squared 0.032538 Mean dependent var -0.000759 Adjusted R-squared 0.029602 S.D. dependent var 0.019706 S.E. of regression 0.019412 Akaike info criterion -5.041306 Sum squared resid 0.248334 Schwarz criterion -5.020934 Log likelihood 1671.672 F-statistic 11.08177 Durbin-Watson stat 1.695455 Prob(F-statistic) 0.000018 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.276804 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . P-value cña thèng kª F =0.278077 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . 5.2.2.4. Söa m« h×nh a/ Kh¾c phôc tù t­¬ng quan b»ng Durbin hai b­íc B­íc 1: ta ®i ­íc l­îng m« h×nh : Theo m« h×nh ban ®Çu ta cã d = 1.708570 vËy =1-d/2 = 0.145715 B­íc 2: ta ®i ­íc l­îng m« h×nh : = + ( - ) + Ta ®Æt : A = (r_can - 0.145715*r_can(-1)) B = (r_vnindex - 0.145715*r_vnindex(-1)) B¶ng 54 Dependent Variable: A Method: Least Squares Date: 04/15/07 Time: 19:56 Sample(adjusted): 3 663 Included observations: 661 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. B 0.198997 0.053985 3.686137 0.0002 C -0.000400 0.000750 -0.533748 0.5937 R-squared 0.020202 Mean dependent var -0.000651 Adjusted R-squared 0.018715 S.D. dependent var 0.019394 S.E. of regression 0.019212 Akaike info criterion -5.063550 Sum squared resid 0.243236 Schwarz criterion -5.049954 Log likelihood 1675.503 F-statistic 13.58761 Durbin-Watson stat 1.960312 Prob(F-statistic) 0.000246 b/ KiÓm ®Þnh sù tù t­¬ng quan m« h×nh trªn +Dïng kiÓm ®Þnh BG . +Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 55 Breusch-Godfrey Serial Correlation LM Test: F-statistic 0.259783 Probability 0.610440 Obs*R-squared 0.260865 Probability 0.609527 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/15/07 Time: 19:58 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. B -0.003813 0.054531 -0.069927 0.9443 C -5.02E-06 0.000751 -0.006690 0.9947 RESID(-1) 0.020058 0.039353 0.509689 0.6104 R-squared 0.000395 Mean dependent var -9.03E-19 Adjusted R-squared -0.002644 S.D. dependent var 0.019197 S.E. of regression 0.019223 Akaike info criterion -5.060919 Sum squared resid 0.243140 Schwarz criterion -5.040524 Log likelihood 1675.634 F-statistic 0.129892 Durbin-Watson stat 1.996339 Prob(F-statistic) 0.878213 Nh×n vµo b¶ng trªn ta thÊy N h­¬ng sai sai sè kh«ng ® P-value cña thèng kª = 0.609527 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.609527 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. c/ KiÓm ®Þnh ph­¬ng sai ®ång ®Òu cña m« h×nh trªn dùa vµo kiÓm ®Þnh White B¶ng 48 White Heteroskedasticity Test: F-statistic 0.073763 Probability 0.928898 Obs*R-squared 0.147995 Probability 0.928672 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/15/07 Time: 20:05 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. C 0.000248 2.66E-05 9.324740 0.0000 B -0.000291 0.001781 -0.163518 0.8702 B^2 0.617702 0.058284 10.59819 0.0000 R-squared 0.000166 Mean dependent var 0.000368 Adjusted R-squared -0.002086 S.D. dependent var 0.000675 S.E. of regression 0.000623 Akaike info criterion -11.92047 Sum squared resid 0.000255 Schwarz criterion -11.90007 Log likelihood 3942.715 F-statistic 58.55391 Durbin-Watson stat 1.696595 Prob(F-statistic) 0.000000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.928672 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . MÆt kh¸c ta cã P-value cña thèng kª F = 0.928898 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . d/ kiÓm ®Þnh d¹ng hµm cña m« h×nh trªn B¶ng 56 Ramsey RESET Test: F-statistic 0.368945 Probability 0.543789 Log likelihood ratio 0.370523 Probability 0.542719 Test Equation: Dependent Variable: A Method: Least Squares Date: 04/15/07 Time: 20:12 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. B 0.188342 0.056788 3.316548 0.0010 C -0.000193 0.000825 -0.233782 0.8152 FITTED^2 -27.59505 45.43078 -0.607409 0.5438 Ta thÊy P-value cña thèng kª Log likelihood ratio cã= 0.542719 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F = 0.543789 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . VËy m« h×nh : A = 0.198997*B - 0.0004005. kh«ng cßn khuyÕt tËt n÷a. Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ : RCAN = - 0.0004005 + 0.198997.RI 5.2.3 Cæ phiÕu KDC *M« h×nh : B¶ng 57 Dependent Variable: R_KDC Method: Least Squares Date: 04/17/07 Time: 08:55 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.017484 0.082430 -0.212113 0.8321 C 0.000213 0.001170 0.182090 0.8556 R-squared 0.000068 Mean dependent var 0.000239 Adjusted R-squared -0.001447 S.D. dependent var 0.029920 S.E. of regression 0.029942 Akaike info criterion -4.176093 Sum squared resid 0.591705 Schwarz criterion -4.162513 Log likelihood 1384.287 F-statistic 0.044992 Durbin-Watson stat 1.820078 Prob(F-statistic) 0.832085 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ EKDC : 5.2.3.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 58 White Heteroskedasticity Test: F-statistic 0.187624 Probability 0.828971 Obs*R-squared 0.376741 Probability 0.828308 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/17/07 Time: 09:01 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.000938 0.000174 5.384410 0.0000 R_VNINDEX -0.001074 0.011419 -0.094012 0.9251 R_VNINDEX^2 -0.224948 0.367500 -0.612105 0.5407 R-squared 0.000569 Mean dependent var 0.000894 Adjusted R-squared -0.002464 S.D. dependent var 0.004066 S.E. of regression 0.004071 Akaike info criterion -8.165427 Sum squared resid 0.010921 Schwarz criterion -8.145056 Log likelihood 2705.756 F-statistic 0.187624 Durbin-Watson stat 1.973371 Prob(F-statistic) 0.828971 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.828308 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . MÆt kh¸c ta cã P-value cña thèng kª F = 0.828971 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu 5.2.3.2. KiÓm ®Þnh sù t­¬ng quan *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : *GØa thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 59 Breusch-Godfrey Serial Correlation LM Test: F-statistic 5.296714 Probability 0.021677 Obs*R-squared 5.278402 Probability 0.021592 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/17/07 Time: 09:10 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.004872 0.082190 0.059280 0.9527 C 3.10E-06 0.001166 0.002662 0.9979 RESID(-1) 0.089392 0.038842 2.301459 0.0217 R-squared 0.007973 Mean dependent var 5.97E-19 Adjusted R-squared 0.004963 S.D. dependent var 0.029919 S.E. of regression 0.029845 Akaike info criterion -4.181078 Sum squared resid 0.586987 Schwarz criterion -4.160706 Log likelihood 1386.937 F-statistic 2.648357 Durbin-Watson stat 1.983417 Prob(F-statistic) 0.071521 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.021592 < 0.05 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.021677 < 0.05 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. 5.2.3.3. KiÓm ®Þnh d¹ng hµm *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *GØa thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 60 Ramsey RESET Test: F-statistic 0.038783 Probability 0.843940 Log likelihood ratio 0.038958 Probability 0.843531 Test Equation: Dependent Variable: R_KDC Method: Least Squares Date: 04/17/07 Time: 09:17 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.001323 0.116358 -0.011372 0.9909 C 3.13E-05 0.001491 0.020995 0.9833 FITTED^2 1742.577 8848.522 0.196934 0.8439 R-squared 0.000127 Mean dependent var 0.000239 Adjusted R-squared -0.002908 S.D. dependent var 0.029920 S.E. of regression 0.029964 Akaike info criterion -4.173131 Sum squared resid 0.591670 Schwarz criterion -4.152760 Log likelihood 1384.306 F-statistic 0.041855 Durbin-Watson stat 1.820379 Prob(F-statistic) 0.959012 Ta thÊy P-value cña thèng kª Log likelihood ratio cã= 0.843531 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F = 0.843940 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. 5.2.3.4. Söa m« h×nh a/ Kh¾c phôc tù t­¬ng quan b»ng Durbin hai b­íc : B­íc 1: ta ®i ­íc l­îng m« h×nh : Theo m« h×nh ban ®Çu ta cã d = 1.820078 vËy =1-d/2 = 0.089961 B­íc 2: ta ®i ­íc l­îng m« h×nh : = + ( - ) + Ta ®Æt : M = (r_kdc - 0.089961*r_kdc(-1)) N = (r_vnindex - 0.089961*r_vnindex(-1)) B¶ng 61 Dependent Variable: M Method: Least Squares Date: 04/17/07 Time: 09:36 Sample(adjusted): 3 663 Included observations: 661 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. N -0.015658 0.083416 -0.187710 0.8512 C 0.000193 0.001166 0.165116 0.8689 R-squared 0.000053 Mean dependent var 0.000214 Adjusted R-squared -0.001464 S.D. dependent var 0.029823 S.E. of regression 0.029845 Akaike info criterion -4.182573 Sum squared resid 0.586990 Schwarz criterion -4.168976 Log likelihood 1384.340 F-statistic 0.035235 Durbin-Watson stat 1.981653 Prob(F-statistic) 0.851162 b/ KiÓm ®Þnh sù tù t­¬ng quan m« h×nh trªn : +Dïng kiÓm ®Þnh BG . +Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 62 Breusch-Godfrey Serial Correlation LM Test: F-statistic 0.033341 Probability 0.855172 Obs*R-squared 0.033491 Probability 0.854793 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/17/07 Time: 09:41 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. N 0.000348 0.083499 0.004171 0.9967 C 1.73E-07 0.001167 0.000148 0.9999 RESID(-1) 0.007125 0.039018 0.182595 0.8552 R-squared 0.000051 Mean dependent var 2.52E-19 Adjusted R-squared -0.002989 S.D. dependent var 0.029822 S.E. of regression 0.029867 Akaike info criterion -4.179598 Sum squared resid 0.586960 Schwarz criterion -4.159203 Log likelihood 1384.357 F-statistic 0.016671 Durbin-Watson stat 1.994755 Prob(F-statistic) 0.983468 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.854793 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.855172 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. c/ KiÓm ®Þnh ph­¬ng sai ®ång ®Òu cña m« h×nh trªn dùa vµo kiÓm ®Þnh White B¶ng 63 White Heteroskedasticity Test: F-statistic 0.194938 Probability 0.822933 Obs*R-squared 0.391422 Probability 0.822250 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/17/07 Time: 09:44 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. C 0.000932 0.000173 5.398011 0.0000 N -0.001591 0.011474 -0.138692 0.8897 N^2 -0.234694 0.376084 -0.624047 0.5328 R-squared 0.000592 Mean dependent var 0.000888 Adjusted R-squared -0.002446 S.D. dependent var 0.004026 S.E. of regression 0.004031 Akaike info criterion -8.185045 Sum squared resid 0.010692 Schwarz criterion -8.164649 Log likelihood 2708.157 F-statistic 0.194938 Durbin-Watson stat 1.942522 Prob(F-statistic) 0.822933 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.822250 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. MÆt kh¸c ta cã P-value cña thèng kª F = 0.822933 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. d/ KiÓm ®Þnh d¹ng hµm cña m« h×nh trªn B¶ng 64 Ramsey RESET Test: F-statistic 0.000109 Probability 0.991657 Log likelihood ratio 0.000110 Probability 0.991635 Test Equation: Dependent Variable: M Method: Least Squares Date: 04/17/07 Time: 09:48 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. N -0.014773 0.118868 -0.124280 0.9011 C 0.000183 0.001501 0.121672 0.9032 FITTED^2 118.8854 11365.71 0.010460 0.9917 R-squared 0.000054 Mean dependent var 0.000214 Adjusted R-squared -0.002986 S.D. dependent var 0.029823 S.E. of regression 0.029868 Akaike info criterion -4.179547 Sum squared resid 0.586990 Schwarz criterion -4.159152 Log likelihood 1384.340 F-statistic 0.017645 Durbin-Watson stat 1.981671 Prob(F-statistic) 0.982510 Ta thÊy P-value cña thèng kª Log likelihood ratio cã= 0.991635 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F = 0.991657 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. VËy m« h×nh : M = -0.01566*N + 0.000193 kh«ng cßn khuyÕt tËt n÷a Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ : RKDC = 0.000193 - 0.01566.RI 5.2.4. Cæ phiÕu LAF *M« h×nh : B¶ng 65 Dependent Variable: R_LAF Method: Least Squares Date: 04/17/07 Time: 10:05 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.373183 0.121015 3.083768 0.0021 C 0.001226 0.001718 0.714003 0.4755 R-squared 0.014204 Mean dependent var 0.000677 Adjusted R-squared 0.012710 S.D. dependent var 0.044240 S.E. of regression 0.043958 Akaike info criterion -3.408158 Sum squared resid 1.275308 Schwarz criterion -3.394578 Log likelihood 1130.100 F-statistic 9.509625 Durbin-Watson stat 1.900258 Prob(F-statistic) 0.002129 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ ELAF : 5.2.4.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 66 White Heteroskedasticity Test: F-statistic 0.149643 Probability 0.861045 Obs*R-squared 0.300511 Probability 0.860488 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/17/07 Time: 10:11 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.002099 0.001653 1.270336 0.2044 R_VNINDEX 0.053635 0.108384 0.494858 0.6209 R_VNINDEX^2 -0.466936 3.488024 -0.133868 0.8935 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.860488 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. MÆt kh¸c ta cã P-value cña thèng kª F = 0.861045 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. 5.2.4.2. KiÓm ®Þnh sù t­¬ng quan  *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : *GØa thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 67 Breusch-Godfrey Serial Correlation LM Test: F-statistic 1.648899 Probability 0.199560 Obs*R-squared 1.652271 Probability 0.198650 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/17/07 Time: 10:19 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.009264 0.121171 -0.076455 0.9391 C -1.37E-05 0.001717 -0.007970 0.9936 RESID(-1) 0.050048 0.038975 1.284095 0.1996 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.198650 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.199560 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1 5.2.4.3. KiÓm ®Þnh d¹ng hµm *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *GØa thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 68 Ramsey RESET Test: F-statistic 0.013284 Probability 0.908277 Log likelihood ratio 0.013344 Probability 0.908035 Test Equation: Dependent Variable: R_LAF Method: Least Squares Date: 04/17/07 Time: 10:28 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.372905 0.121130 3.078565 0.0022 C 0.001133 0.001899 0.596636 0.5510 FITTED^2 3.286674 28.51634 0.115256 0.9083 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.908035 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. -value cña thèng kª F =0.908277 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. VËy m« h×nh : R_LAF = 0.3732*R_VNINDEX + 0.00123 kh«ng cßn khuyÕt tËt n÷a Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ : RLAF = 0.3732*R_VNINDEX + 0.00123 5.2.5. Cæ phiÕu NKD *M« h×nh : B¶ng 69 Dependent Variable: R_NKD Method: Least Squares Date: 04/17/07 Time: 10:39 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.084336 0.049992 -1.686968 0.0921 C -0.001674 0.000710 -2.358844 0.0186 R-squared 0.004293 Mean dependent var -0.001550 Adjusted R-squared 0.002785 S.D. dependent var 0.018185 S.E. of regression 0.018159 Akaike info criterion -5.176244 Sum squared resid 0.217643 Schwarz criterion -5.162663 Log likelihood 1715.337 F-statistic 2.845862 Durbin-Watson stat 1.661876 Prob(F-statistic) 0.092082 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ ENKD : 5.2.5.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 70 White Heteroskedasticity Test: F-statistic 1.503173 Probability 0.223185 Obs*R-squared 3.006317 Probability 0.222426 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/17/07 Time: 10:43 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.000276 7.32E-05 3.765912 0.0002 R_VNINDEX 0.000110 0.004803 0.022949 0.9817 R_VNINDEX^2 0.263678 0.154571 1.705869 0.0885 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.222426 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. MÆt kh¸c ta cã P-value cña thèng kª F = 0.223185 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu. 5.2.5.2. KiÓm ®Þnh sù t­¬ng quan : *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : *GØa thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 71 Breusch-Godfrey Serial Correlation LM Test: F-statistic 19.36694 Probability 0.000013 Obs*R-squared 18.89968 Probability 0.000014 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/17/07 Time: 10:55 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.009757 0.049361 -0.197660 0.8434 C -1.73E-05 0.000700 -0.024772 0.9802 RESID(-1) 0.169189 0.038445 4.400789 0.0000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.000013 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.000014 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1. 5.2.5.3. KiÓm ®Þnh d¹ng hµm *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *GØa thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 72 Ramsey RESET Test: F-statistic 0.617234 Probability 0.432359 Log likelihood ratio 0.619753 Probability 0.431139 Test Equation: Dependent Variable: R_NKD Method: Least Squares Date: 04/17/07 Time: 10:54 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.127788 0.074563 -1.713820 0.0870 C -0.002430 0.001195 -2.032304 0.0425 FITTED^2 181.1357 230.5575 0.785642 0.4324 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.431139 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F =0.432359 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. 5.2.5.4. Söa m« h×nh a/ Kh¾c phôc tù t­¬ng quan b»ng Durbin hai b­íc B­íc 1: ta ®i ­íc l­îng m« h×nh : Theo m« h×nh ban ®Çu ta cã d = 1.661876 vËy =1-d/2 = 0.169062 B­íc 2: ta ®i ­íc l­îng m« h×nh : = + ( - ) + Ta ®Æt : P = (r_nkd - 0.169062*r_nkd(-1)) Q = (r_vnindex - 0.169062*r_vnindex(-1)). B¶ng 73 Dependent Variable: P Method: Least Squares Date: 04/22/07 Time: 09:51 Sample(adjusted): 3 663 Included observations: 661 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. Q -0.126918 0.050371 -2.519655 0.0120 C -0.001449 0.000699 -2.072125 0.0386 R-squared 0.009542 Mean dependent var -0.001293 Adjusted R-squared 0.008039 S.D. dependent var 0.017975 S.E. of regression 0.017903 Akaike info criterion -5.204726 Sum squared resid 0.211210 Schwarz criterion -5.191129 Log likelihood 1722.162 F-statistic 6.348661 Durbin-Watson stat 1.972084 Prob(F-statistic) 0.011982 b/ KiÓm ®Þnh sù tù t­¬ng quan m« h×nh trªn +Dïng kiÓm ®Þnh BG . +Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 74 Breusch-Godfrey Serial Correlation LM Test: F-statistic 0.121589 Probability 0.727429 Obs*R-squared 0.122120 Probability 0.726746 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/22/07 Time: 09:57 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. Q -0.000792 0.050456 -0.015689 0.9875 C -1.23E-06 0.000700 -0.001760 0.9986 RESID(-1) 0.013611 0.039035 0.348696 0.7274 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.726746 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.727429 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. c/ KiÓm ®Þnh ph­¬ng sai ®ång ®Òu cña m« h×nh trªn dùa vµo kiÓm ®Þnh White B¶ng 75 White Heteroskedasticity Test: F-statistic 3.073287 Probability 0.066934 Obs*R-squared 6.117452 Probability 0.066947 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/22/07 Time: 10:00 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. C 0.000248 7.10E-05 3.493316 0.0005 Q -0.002321 0.004760 -0.487606 0.6260 Q^2 0.356427 0.155157 2.297203 0.0219 6 4 Vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0. 0.066947> 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . MÆt kh¸c ta cã P-value cña thèng kª F = 0.066934 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . d/ KiÓm ®Þnh d¹ng hµm cña m« h×nh trªn B¶ng 76 Ramsey RESET Test: F-statistic 1.221307 Probability 0.269508 Log likelihood ratio 1.225738 Probability 0.268238 Test Equation: Dependent Variable: P Method: Least Squares Date: 04/22/07 Time: 10:03 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. Q -0.158501 0.057906 -2.737198 0.0064 C -0.002032 0.000876 -2.319878 0.0207 FITTED^2 114.6329 103.7282 1.105127 0.2695 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.268238 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F =0.269508 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . VËy m« h×nh : P = -0.12692*Q - 0.00145 kh«ng cßn khuyÕt tËt n÷a Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ: RNKD =-0.12692*RI - 0.00145 5.2.6. Cæ phiÕu TRI *M« h×nh : B¶ng 77 Dependent Variable: R_TRI Method: Least Squares Date: 04/22/07 Time: 20:23 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.216315 0.059862 3.613584 0.0003 C 0.000119 0.000850 0.140232 0.8885 R-squared 0.019401 Mean dependent var -0.000199 Adjusted R-squared 0.017915 S.D. dependent var 0.021942 S.E. of regression 0.021744 Akaike info criterion -4.815917 Sum squared resid 0.312057 Schwarz criterion -4.802336 Log likelihood 1596.068 F-statistic 13.05799 Durbin-Watson stat 1.567166 Prob(F-statistic) 0.000325 Ghi l¹i phÇn d­ cña m« h×nh nµy lµ ETRI : 5.2.6.1. KiÓm tra c¸c khuyÕt tËt cña m« h×nh KiÓm ®Þnh ph­¬ng sai ®ång ®Òu dùa vµo kiÓm ®Þnh White. ¦íc l­îng m« h×nh : Gi¶ thiÕt : ( Ph­¬ng sai sai sè ®ång ®Òu ) ( Ph­¬ng sai sai sè thay ®æi ) B¶ng 78 White Heteroskedasticity Test: F-statistic 12.79603 Probability 0.000004 Obs*R-squared 24.74749 Probability 0.000004 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/22/07 Time: 20:25 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. C 0.000332 6.83E-05 4.860367 0.0000 R_VNINDEX 0.004981 0.004478 1.112269 0.2664 R_VNINDEX^2 0.728769 0.144116 5.056813 0.0000 6.2.1011 phiÕu BB- 0.001ThÊy P-value cña thèng kª = 0.00004< 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi . MÆt kh¸c ta cã P-value cña thèng kª F = 0.00004 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ ph­¬ng sai sai sè thay ®æi . 5.2.6.2. KiÓm ®Þnh sù t­¬ng quan *Dïng kiÓm ®Þnh BG ®Ó chóng ta ­íc l­îng m« h×nh sau ®©y : *GØa thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 79 Breusch-Godfrey Serial Correlation LM Test: F-statistic 32.56334 Probability 0.000000 Obs*R-squared 31.17130 Probability 0.000000 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/22/07 Time: 20:29 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX -0.024395 0.058636 -0.416041 0.6775 C -3.58E-05 0.000830 -0.043165 0.9656 RESID(-1) 0.217573 0.038128 5.706430 0.0000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.00000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1 . MÆt kh¸c ta cã P-value cña thèng kª F = 0.000000 < 0.05 vµ 0.01 nªn gi¶ thiÕt bÞ b¸c bá tøc lµ tån t¹i sù tù t­¬ng quan bËc 1 . 5.2.6.3. KiÓm ®Þnh d¹ng hµm *Dïng kiÓm ®Þnh Ramsey ®Ó ­íc l­îng m« h×nh sau : M« h×nh : *GØa thiÕt :  : (d¹ng hµm ®óng).  : (d¹ng hµm sai ) . B¶ng 80 Ramsey RESET Test: F-statistic 0.048668 Probability 0.825467 Log likelihood ratio 0.048887 Probability 0.825011 Test Equation: Dependent Variable: R_TRI Method: Least Squares Date: 04/22/07 Time: 20:32 Sample: 2 663 Included observations: 662 Variable Coefficient Std. Error t-Statistic Prob. R_VNINDEX 0.214207 0.060663 3.531119 0.0004 C 0.000203 0.000931 0.217824 0.8276 FITTED^2 -9.261469 41.98169 -0.220607 0.8255 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.825011 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. P-value cña thèng kª F =0.825467 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng. 6.2.1011 phiÕu BB- 0.001445.2.6.4. Söa m« h×nh a/ Kh¾c phôc tù t­¬ng quan b»ng Durbin hai b­íc B­íc 1: ta ®i ­íc l­îng m« h×nh : Theo m« h×nh ban ®Çu ta cã d = 1.567166 vËy =1-d/2 = 0.216417 B­íc 2: ta ®i ­íc l­îng m« h×nh : = + ( - ) + Ta ®Æt : M = (r_tri - 0.216417*r_tri(-1)) N = (r_vnindex - 0.216417 *r_vnindex(-1)). B¶ng 81 Dependent Variable: M Method: Least Squares Date: 04/23/07 Time: 03:40 Sample(adjusted): 3 663 Included observations: 661 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. N 0.178574 0.059811 2.985654 0.0029 C 5.00E-05 0.000829 0.060334 0.9519 R-squared 0.013346 Mean dependent var -0.000156 Adjusted R-squared 0.011849 S.D. dependent var 0.021366 S.E. of regression 0.021239 Akaike info criterion -4.862969 Sum squared resid 0.297261 Schwarz criterion -4.849372 Log likelihood 1609.211 F-statistic 8.914128 Durbin-Watson stat 1.966262 Prob(F-statistic) 0.002935 b/ KiÓm ®Þnh sù tù t­¬ng quan m« h×nh trªn +Dïng kiÓm ®Þnh BG . +Gi¶ thiÕt :  : (kh«ng cã sù tù t­¬ng quan bËc 1).  : ( cã sù tù t­¬ng quan bËc 1 ). B¶ng 82 Breusch-Godfrey Serial Correlation LM Test: F-statistic 0.176116 Probability 0.674869 Obs*R-squared 0.176872 Probability 0.674075 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/23/07 Time: 03:43 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. N -0.001608 0.059971 -0.026813 0.9786 C -1.89E-06 0.000830 -0.002277 0.9982 RESID(-1) 0.016391 0.039059 0.419662 0.6749 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.674075 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. MÆt kh¸c ta cã P-value cña thèng kª F = 0.674869 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ kh«ng tån t¹i sù tù t­¬ng quan bËc 1. c/ KiÓm ®Þnh ph­¬ng sai ®ång ®Òu cña m« h×nh trªn dùa vµo kiÓm ®Þnh White B¶ng 83 White Heteroskedasticity Test: F-statistic 0.446070 Probability 0.056489 Obs*R-squared 0.544859 Probability 0.056579 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 04/23/07 Time: 07:18 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. C 0.000332 6.73E-05 4.932767 0.0000 N 0.003511 0.004528 0.775336 0.4384 N^2 0.633129 0.145667 4.346419 0.0000 Nh×n vµo b¶ng trªn ta thÊy P-value cña thèng kª = 0.056579 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . MÆt kh¸c ta cã P-value cña thèng kª F = 0.056489 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ ph­¬ng sai sai sè ®ång ®Òu . d/ kiÓm ®Þnh d¹ng hµm cña m« h×nh trªn B¶ng 84 Ramsey RESET Test: F-statistic 0.004478 Probability 0.946669 Log likelihood ratio 0.004498 Probability 0.946528 Test Equation: Dependent Variable: M Method: Least Squares Date: 04/23/07 Time: 07:24 Sample: 3 663 Included observations: 661 Variable Coefficient Std. Error t-Statistic Prob. N 0.177894 0.060711 2.930196 0.0035 C 7.43E-05 0.000906 0.082077 0.9346 FITTED^2 -4.111433 61.44229 -0.066915 0.9467 Ta thÊy P-value cña thèng kª Log likelihood ratio cã=0.946528 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . P-value cña thèng kª F =0.946669 > 0.05 vµ 0.01 nªn kh«ng cã c¬ së b¸c bá gi¶ thiÕt tøc lµ m« h×nh cã d¹ng hµm ®óng . VËy m« h×nh : M = 0.1785736329*N + 5.00148671e-05 kh«ng cßn khuyÕt tËt n÷a Khi ®ã ­íc l­îng cña m« h×nh ban ®Çu lµ : RTRI =5.00148671e-05 +0.1785736329.RI NhËn xÐt : Trong c¶ 6 m« h×nh chØ sè thÞ tr­êng trªn ta thÊy tÊt c¶ c¸c hÖ sè nªn 6 lo¹i cæ phiÕu nµy ®Òu lµ c¸c tµi s¶n thô ®éng ( kÐm n¨ng ®éng ). 6. øng dông kÕt qu¶ cña m« h×nh chØ sè ®¬n (SIM) ®Ó ph©n tÝch ®a d¹ng ho¸ danh môc trong ®Çu t­ Ta cã : Trong ®ã : * : Tæng rñi ro cña cæ phiÕu i. *  : rñi ro hÖ thèng. * : rñi ro phi hÖ thèng ( rñi ro riªng cña mçi cæ phiÕu). IV/ C¸c kÕt luËn rót ra Qu¸ tr×nh ph©n t¸n vµ tèi thiÓu ho¸ rñi ro lµ mét h×nh thøc ®a d¹ng hãa. Trong ®ã nhµ ®Çu t­ nªn ®Çu t­ nªn ®Çu t­ vµo nhiÒu lo¹i chøng kho¸n kh¸c nhau ®Ó t¹o thµnh mét danh môc ®Çu t­ sao cho tæng rñi ro trªn toµn bé danh môc sÏ ®­îc giíi h¹n ë møc nhá nhÊt. §iÒu nµy cã nghÜa lµ “ kh«ng nªn bá trøng vµo mét giá ”. Tõ c¸c kÕt qu¶ tÝnh to¸n ë trªn cho ta biÕt ®­îc sù ¶nh h­ëng cña lîi suÊt thêi kú trÔ ®Õn lîi suÊt thêi kú hiÖn t¹i vµ ¶nh h­ëng cña c¸c nhiÔu ®Õn lîi suÊt cña c¸c cæ phiÕu thêi kú hiÖn t¹i (Th«ng qua m« h×nh ARIMA). Bªn c¹nh ®ã gióp chóng ta cã thÓ t¹o lËp ®­îc tËp danh môc hiÖu qu¶ vµ danh môc rñi ro thÊp nhÊt (MVP). §èi víi c¸c cæ phiÕu nãi riªng còng nh­ c¸c chøng kho¸n riªng lÎ nãi chung, rñi ro tæng thÓ cña danh môc ®Çu t­ ®­îc ®Þnh l­îng b»ng ®é chªnh lÖch chuÈn () cña danh môc ®Çu t­. Tuy nhiªn, trong danh môc ®Çu t­, c¸c cæ phiÕu kh¸c nhau cã thÓ triÖt tiªu rñi ro kh«ng hÖ thèng lÉn nhau chØ cßn l¹i phÇn rñi ro hÖ thèng. V× vËy, rñi ro cã hÖ thèng lµ mét trong nh÷ng yÕu tè quyÕt ®Þnh møc doanh lîi dù kiÕn cña mét tµi s¶n, v× thÕ chóng ta cÇn biÕt c¸ch ®o ®é rñi ro cã hÖ thèng cho c¸c nhµ ®Çu t­ kh¸c nhau. Th­íc ®o mµ chóng ta sÏ sö dông lµ hÖ sè (®­îc x¸c ®Þnh trong m« h×nh CAPM). §©y lµ lý do gi¶i thÝch t¹i sao khi ph©n tÝch rñi ro nhµ ®Çu t­ ph¶i quan t©m nhiÒu ®Õn phÇn rñi ro hÖ thèng () h¬n lµ tæng rñi ro cña danh môc (). Tµi liÖu tham kh¶o 1. Nguyªn lý tµi chÝnh – to¸n cña thÞ tr­êng chøng kho¸n, V­¬ng Qu©n Hoµng – Ng« Ph­¬ng chÝ, NXB chÝnh trÞ quèc gia. 2. Gi¸o tr×nh ph©n tÝch vµ ®Çu t­ chøng kho¸n, Uû ban chøng kho¸n Nhµ n­íc, trung t©m nghiªn cøu vµ båi d­ìng nghiÖp vô chøng kho¸n, Th.S Lª ThÞ Mai Linh chñ biªn, NXB ChÝnh trÞ quèc gia. 3. B¸o ®Çu t­ chøng kho¸n c¸c n¨m 2006, 2007. 4. Gi¸o tr×nh kinh tÕ l­îng vµ bµi tËp kinh tÕ l­îng – Tr­êng §¹i häc kinh tÕ Quèc d©n – khoa To¸n kinh tÕ, Bé m«n ®iÒu khiÓn kinh tÕ, NXB Khoa häc vµ kü thuËt. 5. Bµi gi¶ng m«n ph©n tÝch vµ ®Þnh gi¸ tµi s¶n tµi chÝnh, PGS . TS. Hoµng §×nh TuÊn, Khoa to¸n kinh tÕ Tr­êng §¹i häc kinh tÕ Quèc d©n. 6. C¸c Website vÒ chøng kho¸n : www.bvsc.com.vn, www.vietstock.com,vn, www.bsc.com.vn, ....