Matrix is singular in deterministic transition

[mtrktgr]

Hi.

I am working on a two-country model with entry/exit like Ghironi and Melitz (qje, 2005). I'm now trying to compute a deterministic transition path betwee two steady-states. I have already obained two steady-states for different parameter values. In both steady-states, the number of eigenvalues is equal to that of forward looking variables, and the rank condition is verified.

However, when I try to compute a deterministic transition between the two steady states by "simul", I have the following error:

Warning: Matrix is singular to working precision.
> In siml at 75
In simul at 93
In model9 at 664
In dynare at 120

I first thought that the distance between the two steady-states is too large, but even when the two steady-states are very close each other, I have the same error. So my question is what kind of mistakes causes such an error message? I tried many parameter values, but it did not work.

Thanks, Mtr

[SébastienVillemot]

The Blanchard-Kahn conditions are a local property. So even though they are satisfied in the neighborhood of your two steady states, it is still possible that they are not verified on the transition path. I guess (but I am not 100% sure) that this is what happens in your case.