I'm using Eviews 10 to estimate and forecast panel data with autoregressive distributed lag (ARDL) mode option.
In my initial database, the dependent variable has no data (NA) for the forecast period. All independent variables have data in the forecast period.
When I generate forecasts, only the first period forecast appears. NA’s appear in the remaining forecast period for the dependent variable that I’m trying to forecast. This happens with either ‘Static’ or ‘Dynamic’ options being used to produce forecasts.
If I fill in the dependent variable missing years with ‘dummy’ values, say, the last year for which data is available historically, I do get both dynamic and static forecasts. However, if I change those ‘dummy’ values for the dependent variable in the forecast period to be able to generate a forecast, I get completely different forecasts for the dependent variable, whether or not I opt to use the option to “insert actuals in out-of-sample observations”.
I know that the estimated dependent variable is in the form of a first difference (change in dependent variable), so it needs the previous year of the dependent variable to produce the next period forecasts. But, in dynamic forecasting, that lagged dependent variable forecasts should have been generated and therefore able to be used for the current year forecast.
If I run the same model with Least Squares, there is no problem and I do get all the forecasts when the dependent variable does not have data in the forecast period.
I believe this is a problem with Eviews and wonder whether the issue has been resolved in later versions, i.e., Eviews 11 or Eviews 12. Can you please let me know? Thank you.
Panel data ARDL forecasting with Eviews 10
Moderators: EViews Gareth, EViews Moderator
-
- Posts: 1
- Joined: Sat Jun 02, 2012 1:06 am
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Panel data ARDL forecasting with Eviews 10
Easiest way to check if there is a difference in EV12 is to download the demo copy and check ;) http://www.eviews.com/demo
Follow us on Twitter @IHSEViews
Who is online
Users browsing this forum: No registered users and 19 guests