Long Memory Process in Asset Returns with Multivariate GARCH Innovations
Table 4
Diagnostic tests.
(a)
Full BEKK(1,1,1)
Diagonal BEKK(1,1,1)
166.771
270.155
6.996
43.283
9.720
12.224
(0.000)
(0.000)
(0.857)
(0.00202)
(0.640)
(0.427)
5.301
1.393
11.843
20.662
16.073
37.287
(0.947)
(0.999)
(0.458)
(0.055)
(0.187)
(0.021)
6.204
2.776
11.760
7.236
2.939
3.197
(0.905)
(0.996)
(0.465)
(0.654)
(0.965)
(0.805)
1.889
2.926
2.722
0.753
1.006
0.410
6.969
19.891
16.882
4.456
21.923
17.2967
5449.493
3724.512
2725.128
380.193
31340
17700
(b)
CCC-GARCH(1,1)
DCC-GARCH(1,1)
45.318
0.838
1.066
8.735
5.711
5.390
(0.0009)
(0.999)
(0.999)
(0.725)
(0.901)
(0.949)
8.758
0.012
0.013
4.491
1.593
1.690
(0.723)
(1.000)
(1.000)
(0.972)
(0.997)
(0.995)
0.9395
0.0027
0.0035
0.229
1.176
1.197
(0.998)
(1.000)
(1.000)
(1.000)
(0.999)
(0.998)
1.877
2.896
2.717
1.904
3.141
2.936
5.994
18.415
15.870
6.817
24.626
21.209
4333.862
32283.720
24376.48
5283.913
55953.071
41956.369
Notes: is the Kurtosis (exc), is the skewness coefficient, is the Jarque-Bera normality test, and are, respectively, the 12th order Ljung-Box tests for serial correlation in the standardized and squared standardized residuals (null hypothesis: no autocorrelation). LM is the test for ARCH effects (null hypothesis: no ARCH effects).