 # Bayesian prior with inverted gamma distribution

Hi
Following the dynare manual, I am using inverted gamma distribution for setting the prior of the standard deviations and other parameters which do not have an upper bound. I set prior standarad deviation to infinity following the manual. I get decent results. In a recent presentation, a Bayesian econometrician asked me about the prior values of the parameters of inverted gamma distribution that I am using. My question is: if I set the prior mean equal to (say .01) and prior std to infinity, is there any way I can find these two parameters of the inverted gamma distribution from dynare output.

Dynare has a function for computing the hyperparameters:

``````% 1. In the Inverse Gamma parameterization with alpha and beta, we have alpha=nu/2 and beta=2/s, where
%       if X is IG(alpha,beta) then 1/X is Gamma(alpha,1/beta)
par_mean=0.01;
par_variance=Inf;
[s,nu] = inverse_gamma_specification(par_mean,par_variance, 0, 1, false,'param_name');
``````