To your kind attention:
I am trying to estimate the orthogonal complement of the cointegration
space. I thought of estimating it getting an ML (a la Johansen) estimate of
the cointegration space itself and then computing the orthogonal space of
it, but what are the asymptotic properties of this estimate? Is it
consistent (and at which rate)? Is there any formal proof for this result?
With many thanks in advance
Lorenzo Trapani
|