EViews.com

This thread is about the BPTest Add-in that calculates the Breusch-Pagan LM test for random effects.

Change Log

• 2010/04/16 - Initial Release
• 2010/04/28 - Fixed documentation file name
• 2010/08/16 - Fix for unbalanced panels / equations that don't use every cross-section in the panel
• 2010/11/02 - Added an error message to indicate that the equation should be estimated without cross-section effects prior to running the test.
• 2010/11/24 - Added Honda (1985), King and Wu (1997), SLM (Moulton and Randolph, 1989), GHM (Gourierouxet al., 1982) LM tests.
• 2012/05/10 - Edited the BPtest documentation; Added a command for exclusion of tests.
• 2012/09/26 - Minor changes made.

The documentation file in this add-in had the wrong name. That has now been fixed.

Hi guys,

I've just installed the LM test for Random Effects add-in .
When I run this add-in from equation object (Proc->Add-ins->Breusch-Pagan Random Effects LM test) it returns an error saying: Log or square root of negative number.
I do not understand where the problem may come from, any advice?

Thanks.

Btw, I'm using 7th version of Eviews

Could you post your workfile with the equation in question?

Hi,

Please find attached the workfile with equation.

Thanks.

There's something odd about that equation. Did you copy it from another workfile, or add data to the workfile after estimating it?

It is no longer estimatable.

I resampled it and included less periods due to large size of the file.

This is the workfile for the full sample, I used 7zip to compress, hope this works.

Btw, thx for the fast reply.

Hi, I have the same problem, is there already a solution? Thanks, Ralf

Working on it...

The way it is currently calculating things will only work on panel equations that use every cross-section in the workfile. I'm trying to figure out how to get it to work on equations that only use some cross-sections.

Try re-downloading the add-in. I think this one should work (for you both?)

Thanks for your help in resolving this problem Gareth.
Everything works fine.

Yes, it's working. Thanks!

Great.

[Updates in the BPtest add-in]

(1) Added Honda (1985), King and Wu (1997), SLM (Moulton and Randolph, 1989), GHM (Gourierouxet al., 1982) LM tests.
(2) Considered either joint test for both cross-section and period effects or one-sided test for one effect.