Page 1 of 1

Forecasting with Models

Posted: Wed Apr 21, 2010 9:38 pm
by lostinregression
I am experiencing a great deal of difficulty with an AR(1) forecast assignment that is due for class tommorow and, I would GREATLY appreciate your help if you are able to give it.

My Workfile Range and Sample are 1994m11 - 2011m02.
My observations span 1994m11-2010m02. I need to forecast 2010m03 - 2011m02.

I need to create and out-of-sample forecast from and estimation with multiple independent variables regressed on a dependent variable. When I create the estimation, I specify the sample range as 1994m11-2010m02. In the attached workfile, projectfinal_eq01 is my specifed equation. When I hit forecast, i alter my sample to 2010m03 - 2011m02 and recieve an error "unable to compute due to missing data."

Following this same process for a univariate estimation, I can get the forecast without issue. This can be seen in revolncbunivariatef if you look at the attached workfile.

What am I doing wrong?

I attached the workfile so you can see this for yourself. I am pretty desperate to get this all worked out!

Re: Forecasting with Models

Posted: Thu Apr 22, 2010 7:04 am
by startz
To do the forecast this way you would need to know the values of the right hand side variables for the forecast period, which you don't