I am trying to study the transitional dynamics of a deterministic growth model that I suspect to be AK type, that is, it should not have a transitional dynamics. When I run the model in Dynare in fact I get that all eigenvalues are positive, but when I try to simulate the effect of a permanent shock the model does not jump to the a steady state immediately. What happens is that before going to the next steady state the model shows a weird spike in the opposite direction which doesn’t make any sense. It really looks like a bug in the program. To check that I then ran the basic AK model and the same thing happens.
So my question is: can it be the case that dynare is not designed to deal with models that do not show transitional dynamics? Or you there is some trick to run those models that I ma not aware of?
Thank you,
g