Least Squares Regression Using a System on a Group

[kszynkar]

Hi, perhaps a silly question, but how do I create a set (System I assume) of linear least square regressions on a group. CSer is a series, and CServGroup is a group. This line of code would get me an equation on on the series:

equation CServEqn.ls CSer CSer(-1)

What I'm trying to do is:

system CServSys.ls CServGroup CServGroup(-1)

So essentially a set of of the operation above. However, when I double-click the resulting system, I don't see anything in it. Is there a way to do this?

Thanks

[EViews Glenn]

Is the latter supposed to be estimating a VAR(1)?

[kszynkar]

Yes

[EViews Glenn]

Why not use the built in VAR tools?

Code: Select all

var var1.ls 1 1 CServEqn CServGroup

[kszynkar]

Apologies, I misunderstood VAR(1).

I'm not interested in capturing the cross terms in the regression (ie. getting a coefficient for the sensitivity of my first series in the group to the second), and the above was just to illustrate what I'm trying to do, but it won't only be an auto-regression, there will be other terms in the regression.

[EViews Glenn]

You'll have to add the specification for each of your equations explicitly (apart form the implied error)

Code: Select all

system a a.add var1= c(1) + c(2)*var2 + c(3)*var1(-1) a.add var2 = c(4) + c(5)*var1(-1) + c(6)*x

etc.
Then

Code: Select all

a.ls

There are some tools for creating common specifications to ease some of the editing burdens. See the documentation on the system object, and the makesystem command proc for a group.

[kszynkar]

Perfect. Thanks for the help.