Abstract:
We study for a benchmark small open emerging economy an optimal robust monetary policy à la Hansen and Sargent (2003) considering additive model uncertainty. The robust control approach supposes that economic agents are not able to assign probabilities to a set of all plausible models and rather focuses on the worst possible misspecification from a benchmark model. Our findings are threefold. First, conducting a global robust optimal monetary policy can be limited since the departure from the benchmark model leads to multiple equilibria. Second, when model uncertainty arises only from the IS curve or the UIP condition, the space of unique solutions is expanded. In fact, when the central bank has a preference for robustness on the IS curve only, it should be more aggressive to demand and real exchange rate shocks but more conservative to cost-push shocks. On the other hand, when it has a preference for robustness only for the UIP, the central bank should be more aggressive to demand and cost-push shocks. Third, a sensitivity analysis suggests that conducting a global robust optimal monetary policy with the same misspecification in all equations is limited due to the persistence of inflation, the low exchange-rate pass-through and the need to anchor inflation expectations. Finally, we propose a Bayesian estimation for the Sidaoui and Ramos-Francia's model over the period 2001-2019.
