Re: [Gretl-users] VEC Question

On Fri, 29 May 2009, Ricardo Gonçalves Silva wrote: I'm sorry if this is obvious to you, but since it isn't obvious in general I thought I'd post it to the list. The difficulty with this from an econometric viewpoint is that if you want to estimate a model like \Delta y_t = \alpha \beta' y_{t-1} + \Gamma y_{t-12} + u_t the problem of estimating \beta via ML is not trivial: the VAR representation of the above is a 13-order VAR which not only contains "holes", but also other constraints: y_t = A_1 y_{t-1} + A_2 y_{t-2} + ... + A_{13} y_{t-13} + u_t where (1) A_1 = I + alpha \beta' (2) A_{i} = 0 for i=2..11 (3) A_{12} = -A_{13} The theoretical construction for the Johansen estimator assumes that the only constraint to the VAR representation is (1) (ok, deterministic terms aside), and the code is written on that basis. Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche r.lucchetti(a)univpm.it http://www.econ.univpm.it/lucchetti