0
$\begingroup$

I am new to researching modeling and forecasting using the GARCH model. So I am still confused about the result that I get. I forecast stock return volatility using Eviews. The best ARIMA model in my case is ARIMA(1, 2) based on the highest adjusted R squared value, and the lowest AIC and SIC. Then, I continue to model using GARCH. I get GARCH(2,1) as the best model. When I run forecasting using dynamic forecast in Eviews I get constant result forecasting. Is it appropriate? Can anyone explain why I get consistent results?

$\endgroup$
1
  • 1
    $\begingroup$ Greetings! It may help to provide output or plots of your data so others may understand whats going on. $\endgroup$ Jun 21, 2023 at 11:05

1 Answer 1

0
$\begingroup$

If the model specifies ARMA(1,2)* for the conditional mean and GARCH(2,1) for the conditional variance, the point forecast that you would typically obtain is the estimated conditional mean for the time period of interest (e.g. one time period ahead). This is determined by the conditional mean model; the conditional variance model has nothing to do with this. Under ARMA(1,2), the conditional mean and thus the point forecast will be time varying. If you observe something else, you might have coded the model incorrectly in EVIews.

*You wrote ARIMA(1,2) but ARIMA needs 3 parameters. I assume you meant ARIMA(1,0,2) which is the same as ARMA(1,2).

$\endgroup$
1
  • $\begingroup$ Thanks for your reply, @Richard Hardy. Ya you're right, I mean ARIMA (1,0,2) $\endgroup$
    – LIN_Nisa
    Jun 21, 2023 at 15:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.