The RATS Software Forum

Hi Tom,

How do I estimate a TARCH model, similar to the GJR-GARCH model, but parameterises volatility as a function of lagged absolute value of the residuals?

e.g. TARCH(1,1,1):
sigma(t) = omega + alpha1*abs(resid(t-1)) + gamma1*abs(resid(t-1))*IND(resid(t-1)<0) + beta1*sigma(t-1).

Thanks,
Amarjit

Have you looked at the examples in GARCHUVMAX.RPF? That has a number of examples with asymmetry in different forms.

TomDoan wrote:Have you looked at the examples in GARCHUVMAX.RPF? That has a number of examples with asymmetry in different forms.

Thanks