I am somewhat confused. The trace test and the max eigenvalue test often
lead to different inferences about the number of cointegrating vectors. The
trace test may be more appropriate to the sequential testing methodology
used in the Johansen methodology. I would also think about the economics of
the problem. You will have to identify the cointegrating vectors if you
wish to make progress. May I dare to say that the economics of the problem
are just as important as the statistical methodology.
On 26/10/06, Javier García <javigarcia83(a)yahoo.es> wrote:
Then how is it that the nule hypothesis r = 0 and r = 1 are both accepted
to the two contrasts? the only consistent hypothesis with the results is
that both nule hypothesis is that the rank (pi) minor or equal to "r",
because the rank can't be 0 and 1 at the same time.
Gretl-users mailing list
John C Frain
Trinity College Dublin