First, I build a very simple RBC model with technology uncertainty shock.I could derive desired result that precautionary saving effect reduce a consumption and invest more.
Unfortunately, I try to imprement Andreasen(2018)'s Dynare pruning package and this result doesn’t match my first result.
I try to show that uncertainty shocks decrease consumption,wage,working hour,investment,capital, and output under sticky price.(posted code was more simple one)
Although model block ware same, these two IRF did not macth(i.e. some variable’s sign ware opposite ).
While I could derive consistent result in simulaton based IRF, model should be correct.
I can not distinguish this is my coding issue or some theoritical reason. Must these two IRF are same?
I am using the pruning toolbox of Andreasen on estimating the impact of uncertainty shock to the TFP on a basic RBC with home production sector. They way I do stoch_simul I do it in the same way as in their pruning documentation.
However, when I put the calibrated values for the shocks based on the values I obtained from the particle filter, I get explosive Impulse Response Function as shown in the figures.
Can someone please tell me what is the problem and how to fix it?