Re: [Gretl-users] VECM and restrictions: gretl crashes
On Thu, 4 Mar 2010, Artur T. wrote:
for the same file and session as yesterday I just wanted to do some
further work. Gretl does not calculate estimations for a restricted VECM
and gives this error after the restriction block:
"Matrices not conformable for operation
Error executing script: halting
> end restrict"
[ But a new VECM icon appeared, and trying to use it caused a
crash.]
Both of these issues are now fixed. The substantive point (the
wrongly sized matrices) was something specific to the case where
you have exogenous variables confined to the cointegration space
but no unrestricted exogenous variables, and that's fixed in CVS.
But in addition we now ensure that if "restrict --full" fails for
some reason, you don't get a non-functional VECM icon appearing.
Allin