[Gretl-users] Re: Chow test p-value
On Thu, 27 Jun 2019, Sven Schreiber wrote:
Am 27.06.2019 um 11:17 schrieb Marcelo Duarte:
>
> I'm sorry but I can't seem to find a database somewhat similar to mine
> on data file examples.
> I'm attaching the *.gdt file were the problem occurs.
> The regression is as follows:
>
> panel FDIstock 0 lnGDP(-1) XR(-1) BIT Border PT --pooled --robust
>
> Chow tests:
>
> chow Eurozone --dummy
Thanks for the example. If you observe the Chow regression output, there
was a collinearity issue with the interaction term Eu_Border. I'm
betting that you have all zeros or all ones in one of the subsamples
there, and I'm pretty sure that's the cause.
Yes, all observations with Border = 1 also have Eurozone = 1. In
addition the interaction Eurozone * PT is all zeros. This shouldn't
really be a problem, but the code for the --robust version of the Chow
test was not handling properly the dropping of extraneous regressors.
That's now fixed in git, snapshots will follow.
thanks for the report
Me too, it's good to have this fixed!
Allin