Dynare forums

Dear Johannes,

Could you please look at my mode check plot and priors and posteriors plot?

I use mode_compute 4,6,9 to find mode until the log posterior density does not improve(keep unchanged) , however, rhoa (persistence of transitory technology shock) always reaches that large (0.997-0.998) after those several rounds of mode finding. And measurement error just a straight upwards line.

After I run the MCMC, rhoa and measurement error both hit the upper bond. And when I use mode_compute=9 to run MCMC, rhoa and measurement error do not converge, however when I use mode_compute= 4 or 6, all converge.

Could you give me any suggestions? Thanks a lot!

Catherine

Given that both rhoa and rhop go to 1 I would think your observation equation is still wrong, usually a constant term that is not accounted for.

Regarding measurement error: try what happens when you lift the upper bound.