Stoch_simul after Bayesian Estimation - Joint or Marginal Mode?

Estimation ML/Bayesian estimation
1

Hello,

Following Bayesian estimation in Dynare, I understand that using the following allows you to estimate the impulse responses by fixing the parameter values at their posterior mode:

xparam = get_posterior_parameters(‘mode’,M_,estim_params_,oo_,options_);
stoch_simul(order=1,irf=20);

Are the parameter values used the ones that achieve the mode of the joint distribution, or the mode of the marginal distribution for each parameter?

Thank you in advance.

2

The mode if for the joint distribution. The mean would be the mean of the marginal distributions for each parameter.

3

Great, thank you!