Serving the Quantitative Finance Community

 
User avatar
sunflowerhx
Topic Author
Posts: 0
Joined: July 27th, 2005, 8:13 pm

GARCH(1,1) with constraints

July 28th, 2005, 10:27 am

Hi,I am using Eviews to estimate a GARCH(1,1) with an additonal exogenous variable in the variance eqn.sigma(t)^2 = w+alpha*resid(t-1)^2+beta*sigma(t-1)^2 + gamma*new_variable.I want to perform a likelihood ratio test of the above against the same model with the constraints alpha=beta=0iesigma(t)^2 = w+ gamma*new_variable.However, Eviews does not allow me to set both alpha & beta to zero, only one or the other.Can anyone recommend another package or work around this.Rgds
 
User avatar
Rez
Posts: 24
Joined: May 28th, 2003, 9:27 pm

GARCH(1,1) with constraints

July 28th, 2005, 10:56 am

Matlab, RATS
 
User avatar
meteor
Posts: 0
Joined: September 22nd, 2004, 5:20 pm

GARCH(1,1) with constraints

July 28th, 2005, 12:33 pm

Splus
 
User avatar
sunflowerhx
Topic Author
Posts: 0
Joined: July 27th, 2005, 8:13 pm

GARCH(1,1) with constraints

July 29th, 2005, 3:44 pm

Rez,Manage to have a look at Matlab + Garch Toolbox.How do you add an extra variable to the conditional variance eqn?Thanks
 
User avatar
quantstudent19
Posts: 0
Joined: January 5th, 2004, 2:29 pm

GARCH(1,1) with constraints

July 31st, 2005, 1:56 pm

SAS is quite flexible wrt garch modelsgoogle the program for the egarch model and replace the asymetry term by your new variable
 
User avatar
buckeye
Posts: 0
Joined: August 9th, 2005, 4:21 am

GARCH(1,1) with constraints

August 9th, 2005, 1:07 pm

Sunflowerhx or any of the GARCH gurus,You mentioned that you have added an additional exogenous variable to the variance equation using Eviews. Can you please tell me how you added that variable (either in Eviews, R, or Matlab).Thanks a lot.
 
User avatar
gregh
Posts: 0
Joined: August 4th, 2005, 10:55 am

GARCH(1,1) with constraints

August 9th, 2005, 6:42 pm

I would suggest you look at R. It gave me some flaxibility in estimationg garches.best regards
 
User avatar
buckeye
Posts: 0
Joined: August 9th, 2005, 4:21 am

GARCH(1,1) with constraints

August 9th, 2005, 10:11 pm

I did take a look in R earlier before I posted on this list. Maybe I am missing something. Can anyone guide me on this (either in R or EViews).Thanks.
 
User avatar
sunflowerhx
Topic Author
Posts: 0
Joined: July 27th, 2005, 8:13 pm

GARCH(1,1) with constraints

August 11th, 2005, 7:08 pm

In Eviews, simply add the exogenous variable(s) in the Variance Regressors box, in the Equation Estimation dialog box.You can also do this in S-Plus which I have started using. But coming back to my original problem when i try to fit a garch(0,0)model with the exogenous variable usng the command below, I get a "Trivial model specified" error message. BP.Garch = garch(BP.ts~1, formula.var=~garch(0,0)+BP.ev)Can anyone help??Regards