Dynare forums

Dear Mr. Pfeiffer,

I have a question regarding detrending the data with a break. There is a huge drop in the US employment rate series during the financial crisis and it seems that HP filter does not work precisely in this case. Is there any way to correct this break and fix the problem? As my other time series are de-trended using one-sided HP filter, i would prefer to not use first-difference filter for employment rate . I would appreciate your help.

Kind regards,

All of the filters mentioned are dealing with trend removal, not with breaks or outliers. For that reason, even first differences would not help. If you consider this as a structural break, you need to remove this break first, e.g. by regressing the series on a dummy.

Thanks for the answer!