Recursive Thick Modeling and the Choice of Monetary Policy in Mexico

The choice of monetary policy is the most important concern of central banks. However, this choice is always confronted, inter alia, with two relevant aspects of economic policy: parameter instability and model uncertainty. This paper deals with both types of uncertainty using a very specific class of models in an optimal control framework. For optimal policy rates series featuring the first two moments similar to those of the actual nominal interest rates in Mexico, we show that recursive thick modeling gives a better approximation than recursive thin modeling. We complement previous work by evaluating the usefulness of both recursive thick modeling and recursive thin modeling in terms of direction-of-change forecastability.