Influence of Initial Values on Historical Decomposition

Hi everydody, please see if someone could help me on this:

After using shock_decomposition, I can see that the influence of initial values on some of my endogenous variables are very persistent, even after a long period. (Also, some of IRF take a long time to converge to long run values).

I am confused about those results. Is there anything I should worry about? It seems I dont have unit root problems, as Blanchard and Kahn conditions are satisfied and numerical steady-state convergence is achieved.

Best Regards,


This just means that your model has a lot of persistence. For example, you may have set a shock with a high autocorrelation coefficient.