Dear dynare creators and users,
I have a loglinearized model in dynare. When I run model_diagnostics and check command, everything is fine ( no unit root). However, when I stoch_simul the system with a preference shock, I can see for the consumption of Ricardian HH zero mean but also zero variance in theoretical moments section. Moreover, dynare tells me, that this consumption is either nonstationary or constant. Given equation is in accordance with the literature. The IRF then seem weird. How this could happen? Thank you very much for your comments.