Hello Dynare community,
I am currently using Bayesian estimation for my DSGE model and would appreciate your insights regarding the acceptability of my mode check plot and identification results.
Identification: Most of the parameters pass the identification check. However, the relative to prior std for the first parameter muw is close to 0, while other results seem acceptable.
Mode Check Plot:
For rho_c, the posterior sharply declines when greater than 0.9.
For muN, there is a small secondary peak in the first half, resembling a camel-like shape.
For alphaN, the first half is not a smooth curve.
Additionally, some variables (SE_eta_pwg, muA, etc.) show overlapping log posterior and log likelihood kernel.
Prior vs. Posterior: Despite the above issues, the prior and posterior distributions for these parameters look reasonable and do not indicate significant problems.
Questions:
- Based on the above, can these results be considered acceptable?
- If these issues suggest potential problems, what adjustments should I make (e.g., priors, initial values, or estimation settings)?
Thank you for your guidance!
Best,
Yamoi
