Asymmetric GARCH model estimation

[sgushak]

Hi,

I'm currently writing my master thesis about GARCH modeling, the problem I have is that I EGARCH model I have very insignificant aysmmetric parameter (p-value indicates 0,64). Since it is insignificant, is there any point to compare EGARCH forecasts with GARCH ? Or I should test the series for asymmetry before specifying the model and choose the model based on that test ?

Thanks in advance!

Sergei

[trubador]

Functional form of EGARCH model is different than that of GARCH. I recommend the GJR/TGARCH model to account for asymmetry. Of course, it would not make much sense if the leverage parameter turns out to be insignificant. Still, you can compare the forecasts to see if the two yield similar results.