Estimate GARCH(0,0) ???

[diego45]

Hi everybody,

I have to estimate a model with restriction and one of the restriction is to estimate

ht = c + VIX_t-1

But when i put 0 and 0 for Garch and Arch parameter eviews send the error : "Specification must include at least one Arch or Garch term"

So i have made a new conditional variance series named garch01 and i have used LSQ method to estimate :

garch01 = c + VIXt-1 but the result doesn't looks like the papers that i'm studying.

If someone have an idea, thanks for help

[EViews Glenn]

What is VIX_t-1?

[startz]

It's a derivative that measures market volatility.

[EViews Glenn]

So it's contextual (the Chicago options volatility index), not GARCH model specific? Then I'm a little unclear what sequence of steps diego45 tried to do which didn't match.

[startz]

I believe the OP is trying to put an exogenous variable in the conditional variance equation. EViews allows this for GARCH, except if you want GARCH(0,0) EViews won't handle it. I suppose a GARCH(0,0) could be done where the LHS variable is the squared error terms, so it's really FGLS.