I’m comparing two monetary policy rules in a medium‑scale NK model with sticky prices/wages and a banking sector: a standard Taylor Rule for the policy rate vs. a Taylor Rule that targets an interest rate spread. Both rules share the same structure but differ in the instrument.
I’d like to evaluate which rule yields a larger determinacy region (BK‑satisfying equilibria) when varying the policy coefficients over a common domain (e.g., [0,3]). My understanding is that dynare_sensitivity could be used to map determinacy across parameter grids, but the documentation is sparse.
Could anyone provide guidance or an example on using dynare_sensitivity specifically for BK‑condition stability analysis?
Thanks, @jpfeifer (Johannes). I ended up using dynare_sensitivity with roughly 10,000 draws. I included all parameters with well‑defined domains, assigned uniform priors, and specified the corresponding bounds. I then ran two separate sensitivity exercises: in the first, the standard Taylor rule parameters were included in the sampling while the alternative rule was switched off; in the second, the standard rule was suppressed and the spread‑based Taylor rule parameters were sampled instead. I report the results from both exercises in my paper.
Regarding your point about varobs noted in the code you referenced, I agree the command feels somewhat redundant in this context. Since the specific list of observed variables doesn’t materially affect the sensitivity results, it might be useful if Dynare could automatically select a minimal set of endogenous variables when none are explicitly provided.