Conclusion
Dans ce travail nous avons dans une première
étape exploré l'état de l'intégration
financière mondiale en recourant au modèle simple
développé par Akdogan. Puis nous avons procédé
à un test du modèle conditionnel international
d'évaluation des actions. Pour ce faire, nous avons employé une
spécification GARCH multivarié asymétrique
présentant l'avantage de tester concomitamment 6 marchés : 4
développés et 2 émergents. Ainsi, avec cette
spécification les primes de risque et les corrélations
conditionnelles sont autorisées à osciller suivant les dates et
de réagir aux chocs en fonction de leurs importances et leurs signes. Le
prix de risque de covariance est déterminé par un ensemble de
variables économiques et financières. Cette démarche nous
a permis de ressortir des résultats qui appuient l'hypothèse de
l'intégration des marchés nationaux étudiés.
Ensuite, par analogie aux études antérieures nous avons
dérivé une mesure ex ante des gains de la diversification
internationale de portefeuille.
Les résultats finaux montrent que les
corrélations ont effectivement augmenté au cours de cette
dernière décennie. Ceci est tout à fait logique suite aux
mouvements de libéralisation et de déréglementation
entamés par les gouvernements à partir des années 80 d'une
part et d'autre part grâce au développement des nouvelles
technologies de l'information et des télécommunications. De
l'avis de nombreux auteurs l'augmentation des corrélations des
marchés domestiques aurait diminué les gains émanant de la
stratégie de la diversification internationale. Paradoxalement nos
résultats ne soutiennent pas cette vison. En effet, comme cela a
été empiriquement prouvé les gains de la diversification
sont significativement positifs pour tous les marchés. Mieux encore ces
gains ne présentent à première vue aucune tendance
à la baisse.
ANNEXE 1
A-Prix de risque constant
Estimation des paramètres du modèle
BEKK-GARCH
|
System: SYS01
|
|
|
|
|
Estimation Method: ARCH Maximum Likelihood (BHHH)
|
|
Covariance specification: BEKK
|
|
|
|
Date: 03/23/08 Time: 08:02
|
|
|
|
Sample: 2 420
|
|
|
|
|
Included observations: 419
|
|
|
|
Total system (balanced) observations 2933
|
|
|
Presample covariance: backcast (parameter =0.7)
|
|
|
Failure to improve Likelihood after 80 iterations
|
|
|
|
|
|
|
|
|
|
|
|
Coefficient
|
Std. Error
|
z-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
C(1)
|
0.046024
|
0.004991
|
9.221509
|
0.0000
|
C(2)
|
-0.075050
|
0.005967
|
-12.57786
|
0.0000
|
C(3)
|
-0.006498
|
0.005148
|
-1.262357
|
0.2068
|
C(4)
|
0.030996
|
0.003139
|
9.875507
|
0.0000
|
C(5)
|
-0.047768
|
0.003403
|
-14.03854
|
0.0000
|
C(6)
|
0.002852
|
0.002934
|
0.971972
|
0.3311
|
C(7)
|
0.047435
|
0.005874
|
8.074804
|
0.0000
|
C(8)
|
-0.085911
|
0.006557
|
-13.10308
|
0.0000
|
C(9)
|
0.006508
|
0.004440
|
1.465713
|
0.1427
|
C(10)
|
0.042997
|
0.004072
|
10.55867
|
0.0000
|
C(11)
|
-0.070826
|
0.004662
|
-15.19320
|
0.0000
|
C(12)
|
0.003505
|
0.003885
|
0.902059
|
0.3670
|
C(13)
|
0.041240
|
0.005053
|
8.161565
|
0.0000
|
C(14)
|
-0.078895
|
0.005150
|
-15.31924
|
0.0000
|
C(15)
|
0.006930
|
0.003491
|
1.984867
|
0.0472
|
C(16)
|
0.025124
|
0.002860
|
8.783271
|
0.0000
|
C(17)
|
-0.037350
|
0.002600
|
-14.36754
|
0.0000
|
C(18)
|
0.002740
|
0.002348
|
1.167062
|
0.2432
|
C(19)
|
0.021992
|
0.002623
|
8.384783
|
0.0000
|
C(20)
|
-0.029376
|
0.001962
|
-14.97105
|
0.0000
|
C(21)
|
0.001225
|
0.001953
|
0.627263
|
0.5305
|
|
|
|
|
|
|
|
|
|
|
|
Variance Equation Coefficients
|
|
|
|
|
|
|
|
|
|
|
|
C(22)
|
|
0.000888
|
0.000389
|
2.282864
|
0.0224
|
|
C(23)
|
0.000136
|
7.26E-05
|
1.878019
|
0.0604
|
C(24)
|
0.000265
|
0.000136
|
1.951052
|
0.0511
|
C(25)
|
0.000283
|
8.35E-05
|
3.389485
|
0.0007
|
C(26)
|
0.000186
|
9.89E-05
|
1.885435
|
0.0594
|
C(27)
|
0.000620
|
0.000110
|
5.656050
|
0.0000
|
C(28)
|
0.000376
|
6.35E-05
|
5.925279
|
0.0000
|
C(29)
|
1.50E-05
|
4.94E-06
|
3.034326
|
0.0024
|
C(30)
|
1.18E-05
|
3.54E-06
|
3.339570
|
0.0008
|
C(31)
|
0.000156
|
6.26E-05
|
2.487610
|
0.0129
|
C(32)
|
1.47E-05
|
3.57E-06
|
4.104440
|
0.0000
|
C(33)
|
0.000756
|
9.18E-05
|
8.242172
|
0.0000
|
C(34)
|
0.000347
|
5.84E-05
|
5.947349
|
0.0000
|
C(35)
|
-2.51E-05
|
4.04E-06
|
-6.208897
|
0.0000
|
C(36)
|
0.000260
|
0.000103
|
2.533522
|
0.0113
|
C(37)
|
7.63E-06
|
3.65E-06
|
2.089688
|
0.0366
|
C(38)
|
0.001063
|
0.000135
|
7.862996
|
0.0000
|
C(39)
|
0.000434
|
8.26E-05
|
5.254514
|
0.0000
|
C(40)
|
0.000871
|
0.000272
|
3.204591
|
0.0014
|
C(41)
|
0.000239
|
8.98E-05
|
2.664710
|
0.0077
|
C(42)
|
0.000498
|
8.54E-05
|
5.831705
|
0.0000
|
C(43)
|
0.000429
|
7.46E-05
|
5.752422
|
0.0000
|
C(44)
|
3.40E-05
|
4.07E-06
|
8.370732
|
0.0000
|
C(45)
|
0.000917
|
0.000114
|
8.016151
|
0.0000
|
C(46)
|
0.000382
|
6.77E-05
|
5.649168
|
0.0000
|
C(47)
|
0.000641
|
0.000152
|
4.206620
|
0.0000
|
C(48)
|
0.000808
|
5.48E-05
|
14.75263
|
0.0000
|
C(49)
|
0.000640
|
7.79E-05
|
8.219700
|
0.0000
|
C(50)
|
0.125869
|
0.051058
|
2.465218
|
0.0137
|
C(51)
|
0.147739
|
0.014775
|
9.999079
|
0.0000
|
C(52)
|
-0.002630
|
0.024784
|
-0.106135
|
0.9155
|
C(53)
|
0.414981
|
0.054764
|
7.577636
|
0.0000
|
C(54)
|
-0.004905
|
0.025140
|
-0.195123
|
0.8453
|
C(55)
|
0.282168
|
0.038123
|
7.401528
|
0.0000
|
C(56)
|
0.389388
|
0.033801
|
11.51990
|
0.0000
|
C(57)
|
0.724636
|
0.140469
|
5.158700
|
0.0000
|
C(58)
|
0.980468
|
0.003414
|
287.1624
|
0.0000
|
C(59)
|
0.999715
|
0.000244
|
4098.390
|
0.0000
|
C(60)
|
0.589198
|
0.135784
|
4.339229
|
0.0000
|
C(61)
|
0.992528
|
0.000627
|
1581.760
|
0.0000
|
C(62)
|
-0.484822
|
0.165108
|
-2.936393
|
0.0033
|
C(63)
|
|
0.441090
|
0.085917
|
5.133895
|
0.0000
|
|
|
|
|
|
|
|
|
|
|
|
|
Log likelihood
|
5843.607
|
Schwarz criterion
|
-26.98527
|
|
Avg. log likelihood
|
1.992365
|
Hannan-Quinn criter.
|
-27.35241
|
|
Akaike info criterion
|
-27.59240
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(FRANCE)=C(1)+C(2)*SF+C(3)*TF
|
|
|
R-squared
|
0.518285
|
Mean dependent var
|
0.007204
|
|
Adjusted R-squared
|
0.515969
|
S.D. dependent var
|
0.065663
|
|
S.E. of regression
|
0.045684
|
Sum squared resid
|
0.868187
|
|
Prob(F-statistic)
|
1.948219
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(GB)=C(4)+C(5)*SG+C(6)*TG
|
|
|
R-squared
|
0.418904
|
Mean dependent var
|
0.006003
|
|
Adjusted R-squared
|
0.416110
|
S.D. dependent var
|
0.059061
|
|
S.E. of regression
|
0.045130
|
Sum squared resid
|
0.847281
|
|
Prob(F-statistic)
|
2.033734
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(H_KONG )=C(7)+C(8)*SH+C(9)*TH
|
|
|
R-squared
|
0.412699
|
Mean dependent var
|
0.007196
|
|
Adjusted R-squared
|
0.409876
|
S.D. dependent var
|
0.095514
|
|
S.E. of regression
|
0.073374
|
Sum squared resid
|
2.239612
|
|
Prob(F-statistic)
|
1.846122
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(JAPAN )=C(10)+C(11)*SJ+C(12)*TJ
|
|
|
R-squared
|
0.551221
|
Mean dependent var
|
0.005680
|
|
Adjusted R-squared
|
0.549064
|
S.D. dependent var
|
0.062824
|
|
S.E. of regression
|
0.042188
|
Sum squared resid
|
0.740395
|
|
Prob(F-statistic)
|
1.921386
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(SINGAPOUR )=C(13)+C(14)*SSIG+C(15)*TSIG
|
|
R-squared
|
0.446412
|
Mean dependent var
|
0.004888
|
|
Adjusted R-squared
|
0.443750
|
S.D. dependent var
|
0.081329
|
|
S.E. of regression
|
0.060657
|
Sum squared resid
|
1.530591
|
|
Prob(F-statistic)
|
1.943161
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(USA )=C(16)+C(17)*SU+C(18)*TU
|
|
|
R-squared
|
0.425057
|
Mean dependent var
|
0.005830
|
|
Adjusted R-squared
|
0.422293
|
S.D. dependent var
|
0.045459
|
|
S.E. of regression
|
0.034552
|
Sum squared resid
|
0.496638
|
|
Prob(F-statistic)
|
2.028161
|
|
|
|
|
|
|
|
|
|
Equation: DLOG(MONDE )=C(19)+C(20)*SM+C(21)*TM
|
|
|
R-squared
|
0.340022
|
Mean dependent var
|
0.005970
|
|
Adjusted R-squared
|
0.336849
|
S.D. dependent var
|
0.045652
|
|
S.E. of regression
|
0.037176
|
Sum squared resid
|
0.574945
|
|
Prob(F-statistic)
|
2.256448
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Covariance specification: BEKK
|
|
|
|
GARCH = M + A1*RESID(-1)*RESID(-1)'*A1 +
B1*GARCH(-1)*B1
|
|
M is an indefinite matrix
|
|
|
|
A1 is diagonal matrix
|
|
|
|
B1 is diagonal matrix
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Tranformed Variance Coefficients
|
|
|
|
|
|
|
|
|
|
|
|
Coefficient
|
Std. Error
|
z-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
M(1,1)
|
0.000888
|
0.000389
|
2.282864
|
0.0224
|
|
M(1,2)
|
0.000136
|
7.26E-05
|
1.878019
|
0.0604
|
|
M(1,3)
|
0.000265
|
0.000136
|
1.951052
|
0.0511
|
|
M(1,4)
|
0.000283
|
8.35E-05
|
3.389485
|
0.0007
|
|
M(1,5)
|
0.000186
|
9.89E-05
|
1.885435
|
0.0594
|
|
M(1,6)
|
0.000620
|
0.000110
|
5.656050
|
0.0000
|
|
M(1,7)
|
0.000376
|
6.35E-05
|
5.925279
|
0.0000
|
|
M(2,2)
|
1.50E-05
|
4.94E-06
|
3.034326
|
0.0024
|
|
M(2,3)
|
1.18E-05
|
3.54E-06
|
3.339570
|
0.0008
|
|
M(2,4)
|
0.000156
|
6.26E-05
|
2.487610
|
0.0129
|
|
M(2,5)
|
1.47E-05
|
3.57E-06
|
4.104440
|
0.0000
|
|
M(2,6)
|
0.000756
|
9.18E-05
|
8.242172
|
0.0000
|
|
M(2,7)
|
0.000347
|
5.84E-05
|
5.947349
|
0.0000
|
|
M(3,3)
|
-2.51E-05
|
4.04E-06
|
-6.208897
|
0.0000
|
|
M(3,4)
|
0.000260
|
0.000103
|
2.533522
|
0.0113
|
|
M(3,5)
|
7.63E-06
|
3.65E-06
|
2.089688
|
0.0366
|
|
M(3,6)
|
0.001063
|
0.000135
|
7.862996
|
0.0000
|
|
M(3,7)
|
0.000434
|
8.26E-05
|
5.254514
|
0.0000
|
|
M(4,4)
|
0.000871
|
0.000272
|
3.204591
|
0.0014
|
|
M(4,5)
|
0.000239
|
8.98E-05
|
2.664710
|
0.0077
|
|
M(4,6)
|
0.000498
|
8.54E-05
|
5.831705
|
0.0000
|
|
M(4,7)
|
0.000429
|
7.46E-05
|
5.752422
|
0.0000
|
|
M(5,5)
|
3.40E-05
|
4.07E-06
|
8.370732
|
0.0000
|
|
M(5,6)
|
0.000917
|
0.000114
|
8.016151
|
0.0000
|
|
M(5,7)
|
0.000382
|
6.77E-05
|
5.649168
|
0.0000
|
|
M(6,6)
|
0.000641
|
0.000152
|
4.206620
|
0.0000
|
|
M(6,7)
|
0.000808
|
5.48E-05
|
14.75263
|
0.0000
|
|
M(7,7)
|
0.000640
|
7.79E-05
|
8.219700
|
0.0000
|
|
A1(1,1)
|
0.125869
|
0.051058
|
2.465218
|
0.0137
|
|
A1(2,2)
|
0.147739
|
0.014775
|
9.999079
|
0.0000
|
|
A1(3,3)
|
-0.002630
|
0.024784
|
-0.106135
|
0.9155
|
|
A1(4,4)
|
0.414981
|
0.054764
|
7.577636
|
0.0000
|
|
A1(5,5)
|
-0.004905
|
0.025140
|
-0.195123
|
0.8453
|
|
A1(6,6)
|
0.282168
|
0.038123
|
7.401528
|
0.0000
|
|
A1(7,7)
|
0.389388
|
0.033801
|
11.51990
|
0.0000
|
|
B1(1,1)
|
0.724636
|
0.140469
|
5.158700
|
0.0000
|
|
B1(2,2)
|
0.980468
|
0.003414
|
287.1624
|
0.0000
|
|
B1(3,3)
|
0.999715
|
0.000244
|
4098.390
|
0.0000
|
|
B1(4,4)
|
0.589198
|
0.135784
|
4.339229
|
0.0000
|
|
B1(5,5)
|
0.992528
|
0.000627
|
1581.760
|
0.0000
|
|
B1(6,6)
|
-0.484822
|
0.165108
|
-2.936393
|
0.0033
|
|
B1(7,7)
|
0.441090
|
0.085917
|
5.133895
|
0.0000
|
|
|
|
|
|
|
|
|
|
|
Tests de spécification du MEDAF à prix
du risque constant
Test de wald Variante 1
|
Wald Test:
|
|
|
|
System: SYS4
|
|
|
|
|
|
|
|
|
|
|
|
Test Statistic
|
Value
|
df
|
Probability
|
|
|
|
|
|
|
|
|
|
Chi-square
|
7.637098
|
6
|
0.2659
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Null Hypothesis Summary:
|
|
|
|
|
|
|
|
|
|
|
Normalized Restriction (= 0)
|
Value
|
Std. Err.
|
|
|
|
|
|
|
|
|
|
C(1)
|
-0.040657
|
0.029731
|
|
C(4)
|
-0.014589
|
0.009651
|
|
C(7)
|
0.203577
|
0.214285
|
|
C(10)
|
0.001894
|
0.006214
|
|
C(13)
|
0.160557
|
0.159290
|
|
C(16)
|
0.098468
|
0.105580
|
|
|
|
|
|
|
|
|
|
Restrictions are linear in coefficients.
|
Test de wald Variante 2
Wald Test: System: SYS4
Test Statistic Value df Probability
Chi-square 3.696326 6 0.5939
Null Hypothesis Summary:
Normalized Restriction (= 0) Value Std. Err.
C(3) - C(18) 21.23442 18.20960
C(6) - C(18) -3.773720 7.029277
C(9) - C(18) -5.684867 6.343819
C(12) - C(18) -2.628210 8.972324
C(15) - C(18) -5.538799 6.559892
|
|
