Bayesian simult_ ?

[spartnoy]

Dear forum,

I computed the probability that my variable of interest falls below a certain threshold (e.g. capital-to-asset ratio < 0.09) using Monte Carlo simulations for all exogenous shocks. (You can think of a stress test.) This was an easy exercise using the "simult_" command when all shocks are normally distributed.
However, more realistically when using a Bayesian approach (my shocks are inv_gamma_pdf), I do not know how to draw from the posterior distributions of the exogenous shocks.

Is there a built-in function in Dynare?

Regards,
Samuel

[MichelJuillard]

There must be a confusion somewhere. The solution of the linear(ized) rational expectation model rests on the assumption of zero mean shocks. If your shocks are distributed according to an inverse gamma pdf, they are not zero mean.

Maybe you are confusing the PRIOR on the STANDARD DEVIATION of the shocks with the shocks themselves. The likelihood based method used in Dynare assumes that the shocks are normally distributed.

Best

Michel