Ok thanks. I solved the problem by adding the calibration of three variables, like you explained in another reply.
Thanks for your time.
David.
Hey everybody,
please who knows how to obtain the matrix “usmodel_hist_dsge_f19_7_31_mode” and the similar ones in the original code of smets and wouters 2007?
Houda
Thanks @jpfeifer I meant by “obtain” how they are produced, because it seems to be used as input when the code the authors published does not produce these matrix.
Houda
They are the results from their mode-finding given the respective samples. It should be pointed out that Smets/Wouters were not always successful in finding the global modes. See the header of https://github.com/JohannesPfeifer/DSGE_mod/blob/master/Smets_Wouters_2007/Smets_Wouters_2007_45.mod for a reference on this.
Hi @jpfeifer,
As a greenhand, I am trying to implement your code for Smets/Wouters 2007. May I ask you whether this code depends on the estimation part to be running? For instance, if I delete the estimation part, and add the “steady;” or “stoch_simul” command, it always says there is a problem with finding the steady state. Why is that?
Best
Jing
The code should still run in this case. Does it work with estimation? If yes, please provide the files you are using.
Hi Mr. @jpfeifer
I am trying to study your Smets&Wouters replication to understand the mechanism of Dynare. I just wanna ask a base question. I just wonder about that why the stoch_simul(…) command is not run properly? I just tried to capture the IRFs and steady states values… but when ı write this command the model cannot salved.
Thanks for your interest already now
The original SW-model was not fully calibrated. It required estimation to be run first. In my version 4.5 codes I fixed this.
Dear professor,
I replicated SW2007 and run it with dynare well. But I do not know the method to get the prior mode and covariance matrix. ML estimation to get the prior mode? then how do you get the matrix in the mode file, just like usmodel_shock_decomp_mode.mat ?
I look forward to your reply, thank you!
It’s not the prior mode, but the posterior mode. You can get it by running a mode-finder like mode_compute = 5
Got it, thank you, professor!
Hello. I am trying to estimate the model. I want to do forecasting. How can I do it with this model? Also, why do you specify parameters that are not used such as ccs cinvs?
- You would need to add a
forecast-option to the estimation-command. - Only the original authors can answer why those two parameters were contained in their file. They have no purpose. I guess they forgot to delete them.