Estimation methods when using Bianchi Nicolo method for indeterminacy


I’ve programmed a small scale NK model that permits indeterminant solutions using the method proposed by Bianchi and Nicolo. But based on my current estimation specification I seem to have trouble generating posterior distributions that have coverage in both the determinant and indeterminant regions. I’m currently using the Monte-Carlo based optimization routine for mode finding and RWMH. Is there a more appropriate method? Additionally does Dynare have any future plans to incorporate the hybrid MH algorithm proposed in BN or the sequential monte carlo algorithm being used by many in the indeterminacy literature?


The problem is that these model are often bimodal and therefore require particular samplers. At the moment, there is no concrete plan of adding such samplers, but that might change if there is demand.

Thanks, Johannes. Obviously demand =1 here but I realize that this may be a small subset of researchers interested in these issues.