Hi. I am using Dynare for Bayesian estimation, and I’ve chosen posterior_sampling_method = hssmc for posterior sampling. It seems that SMC is restricted to linear models that only solved at order=1, so the likelihood is evaluated with Kalman filter?
I want to estimate the Dynare example model with particle filter for order=2, can I put mh_replic=20000,mh_jscale = 0.55 or use the posterior_sampling_method = hssmc for posterior sampling?
estimation(order=2, datafile='fsdat_simul.m', particle_filter_options=('pruning', true), number_of_particles=2000, nobs=192, nonlinear_filter_initialization=2, mode_compute=9, resampling='generic', resampling_method=kitagawa, resampling_threshold=0.7,mh_replic=20000,mh_jscale = 0.55, bayesian_irf);
In principle that should be possible. Are you encountering issues? mh_jscale is not a relevant option in that context.