Optimal Monetary Policy with Partially Rational Agents

We explore the dynamic behavior of a New Keynesian monetary policy problem with expectations formed, partially, under adaptive learning. We consider two alternative cases: on the first setting, the private economy has the ability to predict rationally real economic conditions (the output gap) but it needs to learn about the future values of the nominal variable (the inflation rate); on the second setup, private agents are fully aware of future inflation rates, however they lack the ability to predict instantly the correct values of the output gap (learning is attached to this variable). In both cases, we find a simple condition indicating the required learning quality that is needed to guarantee local stability. To achieve convergence to the steady-state, the economy does not need to attain full learning efficiency; it just has to secure a minimum learning quality in order to attain the desired long run result.