Error in Bayesian Estimation: mcmc options: you should reduce mh_init_scale

Dear Everyone;

When I try to estimate model, with estimation(datafile=datas,mh_jscale=0.2,mode_compute=4,mh_replic=5000);

command, I get this error:

Log data density [Laplace approximation] is NaN.

Estimation::mcmc: Multiple chains mode.
Warning: File not found or permission denied

In posterior_sampler_initialization at 102
In posterior_sampler at 59
In dynare_estimation_1 at 448
In dynare_estimation at 105
In qeUKdynare_bayesreviz at 470
In dynare at 223
Estimation::mcmc: Old metropolis.log file successfully erased!
Estimation::mcmc: Creation of a new metropolis.log file.
Estimation::mcmc: Searching for initial values…
Estimation::mcmc: I couldn’t get a valid initial value in 100 trials.
Estimation::mcmc: You should reduce mh_init_scale…
Estimation::mcmc: Parameter mh_init_scale is equal to 0.400000.
Estimation::mcmc: Enter a new value…

Changing the mh_init_scale or mode_compute or initial values for estimated parameters, don’t change the result.

I don’t know that why I am getting this error. All responses highly appreciate.

Thank you.

My impression is that you first have a problem with the estimated mode, which has to be dealt with before running the MCMC. The Laplace approximation of the marginal density of the sample, which is computed at the estimated mode, should not be NaN.

Best,
Stéphane.