Hi, I have this code for the Smets Gali paper of 2011. In the calibrated version is all right, but I have some troubles with the estimation block. The BK conditions failed.
@jpfeifer
SOE_data.mat (8.7 KB) uy_SOE_v1.mod (18.2 KB)
Your initial values for the parameter do not satisfy the BK conditions. You should start with the calibrated version of the model. But that requires adjusting your priors, because some of the calibrated values are outside of the prior bounds.
Thank you! I can’t have a fully understand of what condition is the one that brokes BK but, actually changing my priors allowed me to run the model.
Hi, hope you are fine in this new world.
I success at improving all the SOE staff, but I still have some troubles. I have some big initial values explaining my historical descomposition once I’ve asked for the shock_descomposition. Moreover, I don’t know why but in some models as the one I attached, Matlab gave me some Java mistake.
SOE__horas_est_soloMT.mod (23.9 KB) SOE_data_sindesempleo.mat (9.7 KB) SOE_sindesempleo.m (37.1 KB)
Also it calls my attention that to that file, I can not incorporate inversion, and if I do, I have to remove all calvo parameters associated with imports (probpF indpF).
May be you can give me some light on this.
thank you!!
Given the persistence in your data and the short sample, the effect of initial values will stay with you. That is nothing to worry about. Java errors are a common Matlab problem. They are not particularly related to Dynare.