Simulation issue in endogenous MS DSGE model with ZLB regime #230
Description
Dear Professor Maih,
Sorry to bother you again, and thank you very much for your continued help.
I am currently trying to run simulations for an endogenous Markov-switching DSGE model with a Zero Lower Bound (ZLB). When the model has only two regimes (with different monetary policy coefficients) without ZLB, I can successfully obtain simulation results using
m_sim = simulate(m);
which returns a valid simulation structure.
However, when I introduce the ZLB as a third regime, allowing for endogenous transition probabilities across all three regimes, I encounter an issue. The model is modified as follows (simplified):
parameters(coef,3) phi_pi phi_y zlb_bind
model
i_tylor = phi_pi*infl + phi_y*YD + shock_v;
i = zlb_bind*shock_v + (1 - zlb_bind)*i_tylor;
with the following parameterization:
phi_pi(coef,1), 1.5;
phi_y(coef,1), 0.10;
phi_pi(coef,2), 1.2;
phi_y(coef,2), 0.05;
phi_pi(coef,3), 1.5;
phi_y(coef,3), 0.10;
zlb_bind(coef,1), 0;
zlb_bind(coef,2), 0;
zlb_bind(coef,3), 1;
In this case, both
m = rise('model.rs');
m = solve(m);
run without any errors. However, when I call
m_sim = simulate(m);
the returned object is empty (m_sim = []).
I was wondering whether this behavior is expected when adding a ZLB regime with endogenous switching, or whether there might be additional requirements or constraints for simulation in this setup.
Thank you again for your time and patience.
Best regards,
Simin