AR and smpl

[fmccluer]

Dear Programming,

I am looking for documentation or an explanation on the interrelationship between AR and smpl. Specifically, it appears that when an AR is included the estimation will look back beyond the defined sample period to use pre-sample period values if such data is available.

For example, assume you have data from 2000m1 but smpl 2001m1 to @last
These will not give the same results.

Smpl 2001m1 @last
Eq1.ls depvar c x1 x2 ar(1)

Smpl 2000m12 2000m12
Series x1=@na

Smpl 2001m1 @last
Eq2.ls depvar c x1 x2 ar(1)

My question is: Is there an option that will prevent looking back or outside the smpl period?

Thanks.

[EViews Gareth]

No, there is not.

[EViews Glenn]

But what you can do is to create a new variable for the dependent is the same as the dependent except for missings outside of the sample. Then there is no presample data. You can do this pretty easily using an @recode.