[Gretl-users] Questions re: Chow tests

[Allin Cottrell]

I wrote:

> I just ran an arbitrary example with the same setup of 
> regressors as you mention, and "sd_" variables were added for 
> all 5 independent variables as well as the lagged dependent 
> variable.  If you could send me whatever's needed to replicate 
> your model, I'll be happy to give it a try.  (One possibility 
> that occurs to me is that, somehow, there might be perfect 
> collinearity among some of the created variables.)

Ah, but that possibility reveals a bug: if any collinear 
auxiliary regressors are dropped in the test regression, the 
numerator degrees of freedom for the test will be reported 
wrongly.  I'm committing a fix for this to gretl CVS.

Allin Cottrell