Hi, I want to replicate The Macroeconomic Consequences of Natural Rate Shocks: An Empirical Investigation (nber.org)
To obtain the parameter delta and the natural rate.
Using the specifications on the paper for the calibration outputs a singularity. When changing the parameters, the problem does not seem to change. As the model is small and linear, I dont know what to do to start estimating.
naturalrate.mod (6.3 KB)
Whenever there is a unit root, there will be a singularity in the static Jacobian. That warning message is expected for this type of model.
Hi jpfeifer, I’m replicating the same paper The Macroeconomic Consequences of Natural Rate Shocks: An Empirical Investigation ~
When I run the calib_smoother (calib_smoother(diffuse_filter, datafile=‘data_SGU_r_star_otavio_data.xlsx’, xls_sheet=final_100, filtered_vars);). Smoothed variables are similar to the paper
When I tried to reestimate the model
estimation(diffuse_filter, datafile='data_SGU_r_star_otavio_data.xlsx', xls_sheet=final_100,mode_check, mode_compute=6,xls_range=A1:C122,mh_replic=50000, filtered_vars, lik_init=3);
things are a bit strange. It seems that smoothed variables are in different order. These variables have a unit root process and they are Xm, X, and Xr in the paper. It has something to do with the unit root? Or it is related to the ordering of the variables in the model? (it doesn’t make sense since all the model blocks are practically the same
That is most likely due to the different estimated parameters. Did you check whether
thanks for the quick answer!
I’m rerunning the estimation with estimated_params_init(use_calibration); and tried to calibrate some key parameters and estimate others. I would say that there is sizable difference between the calibration and the posterior mode of the estimation. The strange thing is that Xr (smoothed variable) now looks similar to calib_smoother (it’s not that strange since I’m calibrating delta which is an important parameter related to Xr dynamics)
How do i check if posterior density increases?
