On Fri, 4 Nov 2011, Andreas Noack Jensen wrote:
I am fitting a VECM with exogenous variables i Gretl but I
think something is wrong because I have fitted (what I
believe to be) the exact same model i Ox and and almost
identical model i CATS and the results differ. Please take a
look at the script [...]
Thanks, Andreas. I've taken a first look, and I see that the
differences with Ox are confined to the case of the restricted
exogenous variable. That is, if I omit the restricted broken
trend, Tr76, from the analysis but leave everything else the
same, the results agree perfectly. So I'll try to track down
the source of the difference with Tr76 added.
Allin Cottrell