Bayesian estimation - Accuracy of results

Hello All,

I was able to successfully run the estimation for my model however I am not sure if these results are accurate. I was only able to compute the mode using mode_compute = 6. I also tried mode_compute = 4, 7, 8, 9 with no luck. I have checked the forum and read all relevant threads and have implemented mode_check which seems to look alright although I am not sure about the standard deviation of shocks because the blue line does not necessarily intersect at the peak. The MCMC diagnostics look right for some parameters in the sense that the blue and red lines are converging, however again when it comes to the standard deviation of shocks there seems to be lack of convergence. Is this just an issue with poor priors? Another issue of concern is the graphs of the smoothed shocks, they do not look correct to me… some are flat lines. Any advice on what may be the problem?

Many Thanks,
leeper_est.mod (15.5 KB)

The data file is missing

Hello jpfeifer,

My apologies I have attached the data file.

Many Thanks
pc_data.m (36.4 KB)

Hello,

Also I am unable to reproduce estimated irfs when using the option bayesian_irf. I get the following error message in the Dynare command window

PosteriorIRF: Dynare cannot solve the model (unstable trajectory)

Many Thanks,