Extended Path and NaNs with higher Shock Variance

Stochastic simulations
1

Hello,

I am working with a model that will eventually incorporate a max function in a stochastic setting, so I am opting to use the extended_path command. The model only has one shock, but I have noticed that when I increase the variance of the shock, this leads to the extended path method giving me complex values for the endogenous variables. Anyone have any ideas why this might be the case? I have attached the code for reference.

edynamics_extpath.mod (2.7 KB)

2

To be clear, everything works find when order=0. This problem occurs when I set order to some integer value.

3

Are you using the most recent Dynare version? Older versions may contain bugs that have been fixed.