I have just started learning Bayesian estimation with Dynare, and got confused about the series generated by the “estimate” command.
From what I understand, the series stored in oo_.SmoothedVariables should be exactly equal to the data I’ve fed into Dynare. However, what I get instead is something that looks like deviations from the steady state. Is my understanding correct?
I then proceed to simulate the model using the estimated shocks. I use simult_.m file with oo_.steady_state and all of the the smoothed shocks as the inputs. The generated series I get are pretty close to the actual data I’ve used, but don’t quite replicate it. Obviously, I’d like to get this step down before proceeding to decompose the contributions of each individual shock to the behavior of the observable data.
In all these exercises, I am using the .mod files from Dynare manual Chapter 5 (RCB_Est.mod with data generated by RCB_DataGen.mod).
I really appreciate any comments you might have to help me figure this out!
I am attaching a graph of the original (generated) data vs. the series obtained through estimation and simulation. They look very close - I am not sure what I should expect, so this might be great, or not - but it seems that there is a one-period gap between the two series.
Gen vs Sim.pdf (13 KB)