Re: [Gretl-devel] VECMs with holes

Thanks, Jack and Sven, for the helpful discussion on this. My impressions are that (a) implementing the Eviews-type thing would not be very difficult but would be of doubtful value, while (b) implementing the full-dress version would be a major task, which, if we reckon it's worthwhile, would have to take its place in the TODO queue along with the user-defined Kalman filter, Blundell-Bond and SVARs. So I'm inclined to put this on the back burner for now. Allin.