Hi again, I did it ! I managed to make an asymmetric ECM with Gretl on the basis of help got through this email-list. Thanks for everybody who answered to my messages ! Now I am wondering, if there is in Gretl some special features created for trend-stationar variables ? In some other econometrical programmes there are automatic differences in modelling systems, if variables turn out to be trend-stationar, but from Gretl I have not managed to notice them. Of course user has to know what kind of models are suitable for his/her data, but while working frequently with trend-stationar series I have begun to think, if Gretl had something "extra" to offer in this kind of situations. Sincerely, Hanne