Optimal monetary policy techniques are described from pages 75 to 79 in the Dynare’s manual.
If I understand well,
1/ Dynare is able to compute 3 kind of policy rules: optimal rule (osr), commitment (ramsey_policy), and discretionary (discretionary_policy).
2/ it’s only though simulation features that these policy results are obtained.
Then, if I want to run these optimal policy features in an estimation procedure, how can I proceed ?
(in other words, estimating a model assuming optimal monetary policy that I do not know).
There is maybe some examples (forum, Johannes’s git repository), but not convinced I can.
Thank you in advance for your help
Unfortunately, this is not easily doable with Dynare, but you might be able to do this with Junior Maih’s RISE toolbox.
What you do with optimal policy is specify a subset of parameters and model variables and then compute the remaining parameters (OSR) or variables (instruments in Ramsey) based on some optimality criterion. Once this remainder of the model in the form of optimal policy has been computed, you have a fully specified model for which the likelihood can be computed. In a sense, the additional part resulting from optimal policy is just a bunch of dependent parameters. For that reason, you can in principle evaluate the posterior given your parameters conditional on optimal policy. One paper with corresponding references is “Simple rules versus optimal policy: what fits?” by Ida Wolden Bache, Leif Brubakk and Junior Maih norges-bank.no/Upload/English/Publications/Working%20Papers/2010/Norges_Bank_Working_Paper_2010_3.pdf