Time varying volatility


I am simulating a DSGE model found in this paper mpra.ub.uni-muenchen.de/14977/1/mw33.pdf but am including second moment shocks to TFP. When I do a second third order approximation my IRFs are not smooth. I know this issue has come up before, for instance in this thread: Impulse responses with multiple shocks. However, when I implemented Dr. Juillard’s code, the constructed impulse responses have the variables not reverting to steady state. Is there a way to fix this? I know for higher order approximations Dynare draws a series of shocks for 140 periods. Is there a way to increase the number of periods and perhaps smooth out the IRFs? I’ve attached the code with the IRFs computed both ways. Thanks for the help.
fixed_costs_second_mom.mod (1.73 KB)

The cause of your problem comes from the fact that you did not call “steady”, so oo_.steady_state does not contain the steady state.

Either call “steady” before “stoch_simul”, or use “oo_.dr.ys” instead of “oo_.steady_state” in your calculations at the end of the MOD file. When doing the former, I get an almost flat (up to 1e-5) IRF for consumption, which is reasonable from a computational point of view (I don’t know if this is from an economic point of view, I did not study your model).


Whoops, silly mistake. Thanks for the response!