Dynare forums

Dear all,

I am working on Smets and Wouter(2003) paper with updated data. However, after I transferred the data to zero mean stationary, the mode shows close to zero standard deviations. The data is already in percentage scale, I am not sure this is the problem. After I times 100, the mode works fine with this but one variable (crhols) always gives 0.99 with zero standard deviation. It is very strange since that the model's identification is also fine.
I have been stuck by this issue for quite long time, could anyone help me with this please? Thank you very much for help!


All the best,
Michael

Which data transformation did you actually use. The data you uploaded is not mean 0, particularly for pinf_obs and r_obs. That will give estimation a hard time. Also, it seems to mod-file is designed for data multiplied by 100 (the prior distributions for the standard errors look too large if not scaled), but your data is not.