[Gretl-users] Re: About Linear Restrictions on the short term coefficients of VECM
Am 16.05.2019 um 09:07 schrieb Reynaldo Senra:
Dear all,
I read the User Guide of Gretl and I only have found explanation on
how to set linear restrictions on alpha and beta (jointly or not).
However, I couldn't find how to set linear restrictions on the
coefficients of the lagged differences of the endogenous variables.
It would be great if you suggest me any pdf or advise about how to do
this.
Hi,
you can restrict certain lags to zero in a 'var' command (see the doc
there), but I'm not sure if it also works in a VECM context.
If you need more general restrictions, you would have to take a two-step
approach AFAIK: First estimate the cointegration coefficients (beta) and
save the error-correction term(s) as one or multiple time series, then
set up a simultaneous equation system in differences and throw in the
ECTs. In that system you should be able to apply restrictions on the
dynamics. Caveat: I guess it will be difficult to get gretl to recognize
the ECTS as lagged endogenous terms, so things like forecasting will not
work correctly out of the box.
cheers
sven
Attachments:
- attachment.html (text/html — 1.9 KB)