Hello and thanks for your help in advance!
I am trying to carry out forecasts using a basic NKM model. My approach has been to use the “forecast” command inside the “estimation” one, so that point and mean forecasts are calculated by sampling parameter sets from the posterior distribution. According to other posts in the forum and the user guide point forecasts should reflect both parameter and shock uncertainty (by sampling values from the normal distribution shocks follow and including them in the forecast draws), while mean forecasts omit shock uncertainty (expected value of 0) and only reflect parameter uncertainty.
However, my forecasting results show identical values in all aspects (mean, variance, median…) for the point and mean forecasts. After investigating a bit I have found out that this is happening because no values for shocks are being sampled into the draws of the point forecasts. I have seen this because one of my variables (v) is just the value of the monetary shock (eps_v), and it is constant at 0 over all the point forecast draws (seen in _forc_point1.mat inside the metropolis folder).
Does anyone know why this is happening? Is it a bug or is it something wrong with my code? (I attach the .mod file and the dataset below).
NKMmodel.zip (2.8 KB)
Thank you very much again!