Dear fellow gretl users, I was wondering whether the impulse response functions for VEC models in gretl are calculated in levels or in first differences? According to the gretl guide, IRF are defined as the partial derivative of the l.h.s. dependent variable divided by the partial derivative of an orthogonalised shock. Since the VEC is represented similar to a VAR in first differences (see eq. 27.2 in the gretl guide), with a vector containing the variables in first differences on the l.h.s. and with an error correction term containing in levels in the r.h.s. of the VEC model equation, one might be lead to think that the impulse response would be given in first differences. Am I correct in assuming that the VEC is transformed to a VAR in levels (see eq. 27.1) prior to the calculation of the IRF, though, so that the responses are in levels also? (all of this is assuming that one puts the variables that enter the model in levels in the GUI dialogue) Thanks for your help and for writing such an amazing piece of software! Stefan