I have two vodels GARCH(1,1) and TARCH and I need to choose between them basing on their forecasting perfomance.
First of all i'm not getting completely the static forecasting. Since it uses the last-step fata to make a forecast, it appears that in GARCH model forecasted values of variance are equal to the actual one's, aren't they? (since all we need is to get y = c + e - c and e (epsilon) from the last stage). So static forecasting doesn't make any sense in GARCH models, does it?
Then, when i try to make a forecast (dynamic one), the statistics i get is about forecasted dependent value (y), not variance. So i need to compute RMSE, TIC for variance on my own or there's any other way around it?
I've read abot diebold mariano test but not completely sure how to make it in eviews. Is there any built-in function?
Can i just compute MSE for two models and choose the one with the lowest MSE?
My teacher has given me an excel file with forecasted variances there. And it uses instead of epsilon (for out-of sample forecast) sigma squared. Does eviews also just takes the sigma squared forecasted on the previous step and pretends it is an actual epsilon?
GARCH forecasting, comparing models
Moderators: EViews Gareth, EViews Moderator
Return to “Econometric Discussions”
Who is online
Users browsing this forum: No registered users and 2 guests
