DSGE - VAR: harmonic mean estimator

Greetings dear all,
I estimated a DSGE-VAR, now I’m trying to calculate the marginal likelihood associated with my DSGE-VAR, I’m not using Dynare, the problem that I have is that while I get values of log(posterior) near to -600 and I get log(density) values truncated proposed by Geweke between -30 and -16, when I combine these values I get an harmonic mean estimator “very unexpected” close to 10 ^ -200.

Could you give me some advices to fix this problem, please?

I would appreciate, please


Aldo Yehand

This sounds like a convergence issue of the modified harmonic mean estimator. As you are not using Dynare code, we cannot really provide help with this.

Thank you very much for your attention, I have an additional question about the marginal density reported in Dynare:
¿The reported marginal density is expressed in logarithms: log (marginal density) or not?

Thanks a lot



Yes, Dynare reports the log data density.