The RATS Software Forum

Hi Tom,

Can I ask for the code of basic EGARCH variance equation? The build-in function cannot serve my aim, since I need do some extra changes to the original equation. I write EGARCH(1,1) as follows, but it seems wrong.

The basic variance equation of EGARCH(1,1) is:
log(σt^2)=ω+α*|ε(t-1)|/σ(t-1) +δ* ε(t-1)/σ(t-1) +β*log(σ(t-1)^2)

The codes I write is:
frml ht = exp(a0+a1*abs(u{1})/v{1}+a2*u{1}/v{1})+a3*(log(v{1}**2))


Many thanks,
Aixia

This is based upon the model used in GARCHUVMAX.RPF but does an asymmetric EGARCH model instead of one of the models shown there. It's easier to keep the variance in variance rather than standard deviation form, so the A and D terms use sqrt(h{1}) in the denominator.