Re: [Gretl-users] vecm comparisons
On Fri, 22 Jan 2010, Riccardo (Jack) Lucchetti wrote:
On Thu, 21 Jan 2010, Artur T. wrote:
> Dear Henrique,
>
> yes, you are right. The JMulti gui is a little bit different to gretl --
> that's what confused me a little bit.
>
> The attached file should be correct now.
Which shows identical results to gretl, with the exception of
$vcv/$stderr, which are not adjusted for the degrees of freedom:
<script>
open denmark
vecm 2 1 1 2 3 4 --seas
jMstderr = $stderr * sqrt($df/$T)
print jMstderr
</script>
I tried Eviews 4 on a colleague's box, and its results are identical to
ours: that is, they use the df correction.
Yes, that was one point of interest for me, in asking for output
from various programs. It appears that stata, RATS, R-urca and
eviews all do a df correction on their VECM standard errors, but
JMulTi does not.
Allin.