Annexe 1:

Stat descriptives
em.ret oil.ret vix.ret gold.ret bonds.ret
nbr.val 2675.00000 2.675e+03 2.675e+03 2675.00000 2675.00000
nbr.null 2.00000 1.600e+01 9.000e+00 5.00000 186.00000
nbr.na 0.00000 0.000e+00 0.000e+00 0.00000 0.00000
min -9.46951 -1.283e+01 -3.506e+01 -7.66781 -18.49741
max 10.95720 1.641e+01 4.960e+01 10.42971 9.62768
range 20.42671 2.924e+01 8.466e+01 18.09752 28.12509
sum 59.65490 -2.063e+01 -9.682e+00 76.38012 -74.33669
median 0.06206 9.643e-03 -5.693e-01 0.05445 0.00000
mean 0.02230 -7.713e-03 -3.619e-03 0.02855 -0.02779
SE.mean 0.02484 4.770e-02 1.417e-01 0.02203 0.04293
CI.mean.0.95 0.04871 9.353e-02 2.779e-01 0.04319 0.08417
var 1.65100 6.087e+00 5.372e+01 1.29791 4.92928
std.dev 1.28491 2.467e+00 7.329e+00 1.13926 2.22020
coef.var 57.61704 -3.198e+02 -2.025e+03 39.89933 -79.89365

Jarque Bera Test

data: y.ret[, j]
X-squared = 12000, df = 2, p-value <2e-16

Jarque Bera Test

data: y.ret[, j]
X-squared = 2300, df = 2, p-value <2e-16

Jarque Bera Test

data: y.ret[, j]
X-squared = 1800, df = 2, p-value <2e-16

Jarque Bera Test

data: y.ret[, j]
X-squared = 5000, df = 2, p-value <2e-16

Jarque Bera Test

data: y.ret[, j]
X-squared = 900, df = 2, p-value <2e-16
ARCH test
ARCH LM-test; Null hypothesis: no ARCH effects

data: y.ret[, j]
Chi-squared = 850, df = 12, p-value <2e-16

ARCH LM-test; Null hypothesis: no ARCH effects

data: y.ret[, j]
Chi-squared = 500, df = 12, p-value <2e-16

ARCH LM-test; Null hypothesis: no ARCH effects

data: y.ret[, j]
Chi-squared = 210, df = 12, p-value <2e-16

ARCH LM-test; Null hypothesis: no ARCH effects

data: y.ret[, j]
Chi-squared = 150, df = 12, p-value <2e-16

ARCH LM-test; Null hypothesis: no ARCH effects

data: y.ret[, j]
Chi-squared = 140, df = 12, p-value <2e-16

Annexe 2:
Corrélations de Pearson entre les rendements mensuels

em.ret oil.ret vix.ret gold.ret bonds.ret

em.ret 0.3520 0.0000 0.0177 0.0934
oil.ret 0.3520 0.9732 0.8666 0.0549
vix.ret 0.0000 0.9732 0.1412 0.1770
gold.ret 0.0177 0.8666 0.1412 0.1749
bonds.ret 0.0934 0.0549 0.1770 0.1749
>
Annexe 3:

Quatre spécifications du modèle DCC:

DCC A
*---------------------------------*
* DCC GARCH Fit *
*---------------------------------*

Distribution : mvt
Model : DCC(1,1)
No. Parameters : 43
[VAR GARCH DCC UncQ] : [0+30+3+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26161
Av.Log-Likelihood : -9.78

Information Criteria
---------------------
Akaike 19.592
Bayes 19.687
Shibata 19.592
Hannan-Quinn 19.626

DCC B
*---------------------------------*
* DCC GARCH Fit *
*---------------------------------*
Distribution : mvt
Model : DCC(1,1)
No. Parameters : 38
[VAR GARCH DCC UncQ] : [0+25+3+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26179
Av.Log-Likelihood : -9.79

Information Criteria
---------------------
Akaike 19.601
Bayes 19.685
Shibata 19.601
Hannan-Quinn 19.632
DCC C
*---------------------------------*
* DCC GARCH Fit *
*---------------------------------*
Distribution : mvnorm
Model : DCC(1,1)
No. Parameters : 37
[VAR GARCH DCC UncQ] : [0+25+2+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26402
Av.Log-Likelihood : -9.87

Information Criteria
---------------------
Akaike 19.767
Bayes 19.849
Shibata 19.767
Hannan-Quinn 19.797

DCC D
*---------------------------------*
* DCC GARCH Fit *
*---------------------------------
Distribution : mvnorm
Model : DCC(1,1)
No. Parameters : 32
[VAR GARCH DCC UncQ] : [0+20+2+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26421
Av.Log-Likelihood : -9.88

Information Criteria
---------------------
Akaike 19.778
Bayes 19.849
Shibata 19.778
Hannan-Quinn 19.804
Annexe 4:
DCC and ADCC parameter estimates.

*---------------------------------*
* DCC GARCH Fit *
*---------------------------------*

Distribution : mvt
Model : DCC(1,1)
No. Parameters : 43
[VAR GARCH DCC UncQ] : [0+30+3+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26161
Av.Log-Likelihood : -9.78

Optimal Parameters
-----------------------------------
Estimate Std. Error t value Pr(>|t|)
[em.ret].mu 0.082408 0.012870 6.402892 0.000000
[em.ret].ar1 -0.066759 0.017261 -3.867703 0.000110
[em.ret].omega 0.017298 0.004934 3.505874 0.000455
[em.ret].alpha1 0.132261 0.017098 7.735514 0.000000
[em.ret].beta1 0.866531 0.015755 55.000984 0.000000
[em.ret].shape 5.076698 0.508937 9.975097 0.000000
[oil.ret].mu 0.037027 0.031749 1.166246 0.243515
[oil.ret].ar1 -0.019828 0.020488 -0.967821 0.333134
[oil.ret].omega 0.026254 0.014465 1.815067 0.069514
[oil.ret].alpha1 0.064708 0.014834 4.362307 0.000013
[oil.ret].beta1 0.932149 0.015473 60.242915 0.000000
[oil.ret].shape 8.041626 1.218865 6.597636 0.000000
[vix.ret].mu -0.348725 0.105948 -3.291464 0.000997
[vix.ret].ar1 -0.081486 0.020549 -3.965541 0.000073
[vix.ret].omega 5.404775 1.301503 4.152717 0.000033
[vix.ret].alpha1 0.150119 0.026747 5.612499 0.000000
[vix.ret].beta1 0.757994 0.041817 18.126377 0.000000
[vix.ret].shape 4.865022 0.470363 10.343122 0.000000
[gold.ret].mu 0.046686 0.017825 2.619070 0.008817
[gold.ret].ar1 0.024601 0.018086 1.360231 0.173757
[gold.ret].omega 0.012679 0.004209 3.012372 0.002592
[gold.ret].alpha1 0.043182 0.007834 5.511968 0.000000
[gold.ret].beta1 0.946431 0.008604 110.002600 0.000000
[gold.ret].shape 5.879910 0.657746 8.939490 0.000000
[bonds.ret].mu -0.033837 0.033235 -1.018112 0.308625
[bonds.ret].ar1 -0.008462 0.020232 -0.418236 0.675774
[bonds.ret].omega 0.022601 0.012524 1.804674 0.071126
[bonds.ret].alpha1 0.060050 0.008135 7.381566 0.000000
[bonds.ret].beta1 0.938061 0.008148 115.123003 0.000000
[bonds.ret].shape 9.483015 1.667643 5.686477 0.000000
[Joint]dcca1 0.009601 0.135166 0.071029 0.943375
[Joint]dccb1 0.938974 1.629322 0.576298 0.564414
[Joint]mshape 10.294565 0.990668 10.391537 0.000000

Information Criteria
---------------------
Akaike 19.592
Bayes 19.687
Shibata 19.592
Hannan-Quinn 19.626

*---------------------------------*
* ADCC GARCH Fit *
*---------------------------------*
Distribution : mvt
Model : aDCC(1,1)
No. Parameters : 49
[VAR GARCH DCC UncQ] : [0+35+4+10]
No. Series : 5
No. Obs. : 2675
Log-Likelihood : -26133
Av.Log-Likelihood : -9.77

Optimal Parameters
-----------------------------------
Estimate Std. Error t value Pr(>|t|)
[em.ret].mu 0.056763 0.013402 4.235331 0.000023
[em.ret].ar1 -0.062715 0.017995 -3.485245 0.000492
[em.ret].omega 0.020342 0.004893 4.157145 0.000032
[em.ret].alpha1 0.000000 0.017028 0.000001 0.999999
[em.ret].beta1 0.868881 0.019392 44.806907 0.000000
[em.ret].gamma1 0.232947 0.035380 6.584147 0.000000
[em.ret].shape 5.597762 0.621992 8.999736 0.000000
[oil.ret].mu 0.016340 0.032539 0.502181 0.615540
[oil.ret].ar1 -0.021558 0.020884 -1.032306 0.301929
[oil.ret].omega 0.021860 0.011184 1.954604 0.050630
[oil.ret].alpha1 0.023935 0.010044 2.382954 0.017174
[oil.ret].beta1 0.940084 0.011867 79.215149 0.000000
[oil.ret].gamma1 0.065893 0.014796 4.453426 0.000008
[oil.ret].shape 8.633389 1.409647 6.124503 0.000000
[vix.ret].mu -0.220027 0.103124 -2.133608 0.032875
[vix.ret].ar1 -0.078164 0.019954 -3.917150 0.000090
[vix.ret].omega 3.876921 0.235583 16.456675 0.000000
[vix.ret].alpha1 0.197828 0.001229 160.962807 0.000000
[vix.ret].beta1 0.830534 0.004535 183.143871 0.000000
[vix.ret].gamma1 -0.231003 0.001204 -191.909611 0.000000
[vix.ret].shape 5.153947 0.493377 10.446269 0.000000
[gold.ret].mu 0.052450 0.017899 2.930265 0.003387
[gold.ret].ar1 0.019524 0.018885 1.033857 0.301203
[gold.ret].omega 0.011795 0.003524 3.346844 0.000817
[gold.ret].alpha1 0.056026 0.008879 6.309923 0.000000
[gold.ret].beta1 0.948428 0.006362 149.085181 0.000000
[gold.ret].gamma1 -0.025717 0.011223 -2.291418 0.021939
[gold.ret].shape 5.823969 0.643685 9.047850 0.000000
[bonds.ret].mu -0.051591 0.033908 -1.521493 0.128136
[bonds.ret].ar1 -0.008092 0.020197 -0.400651 0.688677
[bonds.ret].omega 0.019026 0.010676 1.782076 0.074737
[bonds.ret].alpha1 0.032438 0.008175 3.967873 0.000073
[bonds.ret].beta1 0.944762 0.007131 132.485704 0.000000
[bonds.ret].gamma1 0.043600 0.011454 3.806474 0.000141
[bonds.ret].shape 9.937791 1.850991 5.368903 0.000000
[Joint]dcca1 0.015018 0.012188 1.232172 0.217885
[Joint]dccb1 0.922149 0.107798 8.554410 0.000000
[Joint]dccg1 0.000000 0.004111 0.000000 1.000000
[Joint]mshape 10.647588 0.824277 12.917488 0.000000

Information Criteria
---------------------
Akaike 19.575
Bayes 19.683
Shibata 19.575
Hannan-Quinn 19.614
Annexe 5:

GOGARCH
*------------------------------*
* GO-GARCH Fit *
*------------------------------*

Mean Model : AR
(Lag) : 1
GARCH Model : sGARCH
Distribution : manig
ICA Method : fastica
No. Factors : 5
No. Periods : 2675
Log-Likelihood : -26166.59
------------------------------------

U (rotation matrix) :

[,1] [,2] [,3] [,4] [,5]

[1,] 0.3492 -0.0209 0.00624 -0.93676 0.00914
[2,] -0.0373 0.1515 -0.01261 -0.00772 0.98764
[3,] -0.0165 0.9873 -0.02875 -0.02988 -0.15270
[4,] 0.0487 -0.0289 -0.99830 0.01210 -0.00638
[5,] 0.9349 0.0328 0.04868 0.34841 0.03360

A (mixing matrix) :

[,1] [,2] [,3] [,4] [,5]

[1,] 1.1204 -0.00587 -0.0259 -0.61364 0.0132
[2,] -0.0768 -0.00899 -0.0256 -0.00574 2.4646
[3,] -2.4425 0.13729 -0.0539 6.87081 -0.0665
[4,] -0.0137 -0.00634 1.1372 0.03408 0.0193
[5,] -0.0670 2.21643 -0.0473 -0.01366 0.0906

> coef(gogarch.fit)
F_1 F_2 F_3 F_4 F_5
omega 0.006865 0.004974 0.01014 0.1259 0.004468
alpha1 0.106279 0.058883 0.04540 0.1614 0.065874
beta1 0.884899 0.938481 0.94348 0.7133 0.930820
skew -0.019634 0.020293 -0.13514 0.2655 -0.083297
shape 1.725530 3.605020 1.76831 1.4076 2.711686
Annexe 6:
Corrélations conditionnelles dynamiques
Annexe 7:
Compare correlations
>
> cor(rcor(dcc1.fit)[1,2,],rcor(dcc2.fit)[1,2,])
[1] 0.9796
> cor(rcor(dcc1.fit)[1,2,],rcor(gogarch.fit)[1,2,])
[1] 0.1448
> cor(rcor(dcc2.fit)[1,2,],rcor(gogarch.fit)[1,2,])
[1] 0.1114
>
>
> cor(rcor(dcc1.fit)[1,3,],rcor(dcc2.fit)[1,3,])
[1] 0.9741
> cor(rcor(dcc1.fit)[1,3,],rcor(gogarch.fit)[1,3,])
[1] 0.2512
> cor(rcor(dcc2.fit)[1,3,],rcor(gogarch.fit)[1,3,])
[1] 0.2805
>
>
> cor(rcor(dcc1.fit)[1,4,],rcor(dcc2.fit)[1,4,])
[1] 0.9874
> cor(rcor(dcc1.fit)[1,4,],rcor(gogarch.fit)[1,4,])
[1] -0.04495
> cor(rcor(dcc2.fit)[1,4,],rcor(gogarch.fit)[1,4,])
[1] -0.01791
>
>
> cor(rcor(dcc1.fit)[1,5,],rcor(dcc2.fit)[1,5,])
[1] 0.9812
> cor(rcor(dcc1.fit)[1,5,],rcor(gogarch.fit)[1,5,])
[1] 0.07137
> cor(rcor(dcc2.fit)[1,5,],rcor(gogarch.fit)[1,5,])
[1] 0.05241
Annexe 8:
Les surfaces de corrélation d'impact des actualités
Annexe 9:
Ratios de covertures