First thank you very much for your kind help, I am grateful.
I have recently asked some people to review my paper, and some of them criticized that since the prior means and posterior means of my DSGE model is quite close, they don’t feel that anything has come from the data and the results are worrying.
However, I believe there are potential reasons for this, I have small prior variances, the number of state variables are much larger than the number of observed variables, potential model misspecification.
Could you please let me know are the results worrying if prior means and posterior means are quite close?
Thank you very much and look forward to hearing from you.
Yes, this is generally a bad sign. If your posterior is close to the prior, the data does indeed not add much information on top of your prior. When you state that your prior precision is high, this seems to be a deliberate design feature. The most crucial aspect is to check whether the reason is really only that the data does not provide much evidence or whether there is a pathological identification problem. Thus, run the identification command and check for proper mixing of the MCMC.