Page 1 of 1

Forecasting - missing values

Posted: Thu Oct 14, 2010 8:08 pm
by jag_211085
Hi,
I can get n+1 forecast, but not n+2 or n+3.
I decribe the problem below.

I am trying to forecast with this equation using the Forecast button in the Equation window:

equation reg.ls log(hourly_fees) c @trend @trend^2 @quarter(2) log(hourly_fees(-1))

My sample in the workfile is 2001Q3 to 2010Q4
My sample (adjusted) in my equation is 2001Q4 2010Q1.

That is, I have actuals till 2010Q1 and I want to forecast for 2010Q2-2010Q4

I click Forecast, enter Forecast sample as 2010Q2 2010Q4 and Click OK but only get a value for 2010Q2 and not for 2010Q3 or 2010Q4.

Forecasting - missing values

Posted: Thu Oct 14, 2010 9:14 pm
by EViews Gareth
Are you doing a dynamic, or a static forecast?

Re: Forecasting - missing values

Posted: Thu Oct 14, 2010 10:00 pm
by jag_211085
Dynamic, I think I figured out the answer; using @seas(2) instead of @quarter(2) works.
But I'm not sure why they are different?

Thanks.

Re: Forecasting - missing values

Posted: Fri Oct 15, 2010 3:45 am
by trubador
@seas(2) generates a dummy variable that takes the value of 1 in second quarters and 0 elsewhere. However, @quarter generates a series that shows the quarter of the year in which each observation begins. @quarter(2) is the second lead of this series. The equivalent expressions are @seas(2) and (@quarter=2).

Forecasting - missing values

Posted: Fri Oct 15, 2010 7:23 am
by EViews Gareth
Nice catch.