EGARCH vs TARCH

[shane_neo]

Hi,

I am a new user of EVIEWS. I am having a couple of confusions.

1. Which one is the best criteria (AIC, SIC, Hannan-Quinn) to choose a model and why.

2. The second one is regarding the leverage effect as per the EGARCH and TARCH model. I'm getting opposite results regarding the presence of leverage effect in the market from EGARCH and TARCH. I'm pasting the result below. Please take a look and tell me what inference can be drawn about leverage effect.

TARCH

Dependent Variable: SER01
Method: ML - ARCH (Marquardt) - Normal distribution
Date: 06/26/10 Time: 16:15
Sample: 1 234
Included observations: 234
Convergence achieved after 21 iterations
Bollerslev-Wooldridge robust standard errors & covariance
Presample variance: backcast (parameter = 0.7)
GARCH = C(2) + C(3)*RESID(-1)^2 + C(4)*RESID(-1)^2*(RESID(-1)<0) +
C(5)*GARCH(-1) + C(6)*GARCH(-2)

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000132 5.51E-05 2.403725 0.0162

Variance Equation

C 3.39E-07 1.74E-07 1.953363 0.0508
RESID(-1)^2 0.570786 0.184325 3.096637 0.002
RESID(-1)^2*(RESID(-1)<0) -0.515874 0.183355 -2.813522 0.0049
GARCH(-1) -0.081199 0.016686 -4.866279 0
GARCH(-2) 0.688836 0.030623 22.49377 0

R-squared -0.000253 Mean dependent var 0.000157
Adjusted R-squared -0.000253 S.D. dependent var 0.001558
S.E. of regression 0.001558 Akaike info criterion -10.35742
Sum squared resid 0.000566 Schwarz criterion -10.26882
Log likelihood 1217.818 Hannan-Quinn criter. -10.3217
Durbin-Watson stat 1.81388


EGARCH

Dependent Variable: SER01
Method: ML - ARCH (Marquardt) - Generalized error distribution (GED)
Date: 06/26/10 Time: 16:17
Sample: 1 234
Included observations: 234
Convergence achieved after 21 iterations
Presample variance: backcast (parameter = 0.7)
LOG(GARCH) = C(2) + C(3)*ABS(RESID(-1)/@SQRT(GARCH(-1))) + C(4)
*ABS(RESID(-2)/@SQRT(GARCH(-2))) + C(5)*RESID(-1)
/@SQRT(GARCH(-1)) + C(6)*LOG(GARCH(-1))

Variable Coefficient Std. Error z-Statistic Prob.

C 3.45E-08 2.15E-06 0.016028 0.9872

Variance Equation

C(2) -1.373639 0.613257 -2.239908 0.0251
C(3) 0.61561 0.231667 2.657302 0.0079
C(4) -0.236349 0.247088 -0.956542 0.3388
C(5) 0.166417 0.121326 1.371648 0.1702
C(6) 0.915421 0.041057 22.29641 0

GED PARAMETER 0.561158 0.062314 9.005361 0

R-squared -0.010205 Mean dependent var 0.000157
Adjusted R-squared -0.010205 S.D. dependent var 0.001558
S.E. of regression 0.001566 Akaike info criterion -11.07394
Sum squared resid 0.000571 Schwarz criterion -10.97057
Log likelihood 1302.651 Hannan-Quinn criter. -11.03226
Durbin-Watson stat 1.796011