Fixing parameters for IRF after calibrating a DSGE

Stochastic simulations
1

I calibrated a DSGE of a Neo keynesian model and got the IRF. Now I one to change just one parameter to estimate new IRF’s but every time I run the stochastic simulation the parameters come back to the calibrated. How can I get it Fixed and do IRF’s? Someone can help me?
Thank you

2

You should perhaps post the file and explain what you are trying to change.

3

Not sure I understand, if you have beta = 0.99 in your first model, you want to change it to beta = 0.7, for example?

4

the calibration changed the initial beta .99 to beta .6. Iran IRF wuth .6 but iwant to run IRF with beta .4 all the other parameters equal.

5

i have tried to do with this code but the gamma_parm come back to the value got in the calibration
*/
// load previous estimation results
basecase = load(‘IPF_MEX_cali_step3_endo_Kim_RERPC_results.mat’);
parnames = fieldnames(basecase.oo_.posterior_mean.parameters);
// parvalues = struct2cell(basecase.oo_.posterior_mean.parameters);
parvalues = basecase.oo_.posterior.metropolis.mean;
for ii = 1:numel(parnames)
set_param_value(parnames{ii},parvalues(ii));
end;

*/

// change the value of specific parameters
pname = ‘gamma_parm’;
i = strmatch(pname,M_.param_names,‘exact’);
M_.params(i) = .9;

estimation(order=1, presample=5, datafile=‘MEX_obs_V3.csv’, lik_init=2, first_obs =1, mh_replic=100000, mh_nblocks=1, mh_jscale=0.35, mode_compute =1, bayesian_irf, optim=(‘MaxIter’, 100000),irf=40);

// Compute IRFs and simulate model for 250,000 periods with posterior mean parameters
stoch_simul(order=1, irf=40, nodecomposition, contemporaneous_correlation, nograph, periods = 100000);
//stoch_simul(order=1, irf=40, irf_shocks = (EPS_FXI,EPS_RERPQ,EPS_RS), periods = 100000) DPCobs GDP_GAP RERPC FXI RSobs RNX;
//stoch_simul(order=1, irf=40, irf_shocks = (EPS_FXI,EPS_RERPQ,EPS_RS), periods = 100000) DPCobs GDP_GAP RERPC FXI RSobs RNX;

//shock_decomposition(parameter_set=posterior_mode,datafile=‘MEX_obs_V3.csv’,nograph);
//generate_trace_plots(1);

6

Sorry I cannot help with the code. But I think you can write the parameters manually, maybe. Say you have only two parameters in the model which you have estimated, beta = 0.95, delta = 0.5.

If you want to change the value of beta, for example, you can write the model in another mod file.

var C Y;

varexo e;

parameters beta delta;

beta = 0.5 //(new value)
delta = 0.5

model(linear);
7

If parameters are calibrated to match a target, they cannot be independently set. You would need to remove the calibration first.

8

thank you , how can i blocke the calibration?

9

thank yoy. How can I take the model after calibration (take model and their parameters son I can change one ) and do IRF

10

Without having the files, it is impossible to tell what is going on.