The RATS Software Forum

Hi,

I was just wondering if it is possible to adapt the code for the GARCH estimation of DCC or can you only do it with two step likelihood?

I have used the code:
garch(p=1,q=1,i=nodrift,model=ar1,mv=dcc,asymmetric,pmethod=simplex,piters=100,method=bfgs,iter=1000,trace)

with ar1 representing my system of mean equations.
In my understanding the i term and asymmetric commands relate only to the volatility and not to the correlation.
However the sum of my a and b DCC coefficients is 0.99 suggesting the need for an integrated DCC GARCH and I was wondering how to perform this.

Thanks for your help. The RATS support services are excellent.

* That meant to read: I was just wondering if it is possible to adapt the code for the GARCH estimation of DCC for integrated DCC or can you only do it with two step likelihood?

gemmamc88 wrote:* That meant to read: I was just wondering if it is possible to adapt the code for the GARCH estimation of DCC for integrated DCC or can you only do it with two step likelihood?

The GARCH instruction won't do DCC with an integrated DCC model. You'll have to modify one of the examples on http://www.estima.com/forum/viewtopic.php?f=11&t=792 to incorporate the restriction.