Random Walk Metropolis algorithm

Estimation ML/Bayesian estimation
1

Dear all,
I am trying to estimate a DSGE model and I would like to generate posterior statisctics of parameters( mean and standard deviation) by using random walk metropolis in Dynare 4.5.7.
I would like to ask
1/ what is the difference between the estimations bellow?
2/ what is the estimation to run to get the posterior estimation by random walking Metropolis?
3/ if not, what is the modification should i insert to get the posterior estimation by random walking Metropolis?.

  • estimation(datafile=data_nfbsop_54q307q2, first_obs=1, prefilter=0,lik_init=3,mode_check,mh_replic=0);

  • estimation(datafile=data_nfbsop_54q307q2, first_obs=1 ,prefilter=0,mode_compute=0,mode_file=cn_fsample_mode,lik_init=2,conf_sig=0.90,bayesian_irf,smoother,mh_nblocks=2,mh_jscale=0.25,mh_drop=0.5,mh_replic=500000);

  • estimation(datafile=data_nfbsop_54q307q2,first_obs=1,prefilter=0,mode_compute=0,mode_file=cn_amend_7obs_mode,lik_init=1,presample=0,conf_sig=0.90,mh_nblocks=1,smoother,mh_jscale=0.2,mh_drop=0.5,mh_replic=40000);

Thank you
SANDRA

2

See the official Dynare documentation regarding estimation at https://www.dynare.org/manual/Estimation.html.

3

The main difference is whether mode-finding is executed and which Kalman filter initialization is used. With mh_replic=0 no Metropolis-Hastings is run.

4

I find your data file have cn;
do you model china economy?

i am a chinise ,
Shanghai,China

5

thank you for your reply and well noted

6

Hello, no sorry it is model of usa economy.

7

thank you for your reply, well noted.

8

i see
thanks for your reply