Monetary Policy under Adaptive Learning

The paper studies the conduct of monetary policy, in a simple new Keynesian model, with adaptive learning on the part of the private sector. A key feature is that even though we start out with a linear “structural†model, the system and hence policy responses inherit the non-linear feature of the updating equations for the estimated parameters. In the paper, we contrast two different monetary policy regimes. In the first the central bank follows a simple rule, which comes from the first order conditions, for optimal policy under discretion in the case of rational expectations. In the second, the central bank has full information about the structure of the economy, including the adaptive learning mechanism. It takes the expectations formation mechanism explicitly into account when deriving optimal policy. This framework allows an explicit discussion of the importance of keeping inflation expectations under control. We illustrate with an application to a regime change, where we assume that the incumbent policymaker did not take the learning into account and allowed the expectation formation process to become unhinged. However, before inflation expectations (and actual inflation) spirals out of control, we assume that a sophisticated central banker, who does take the effect of learning into account, takes charge and study how the economy adjusts after the regime change. Under our assumptions the transition is slow. We claim that some features of the transition match important stylised facts associated with the Volcker disinflation in the US. In the end the fully optimal policy delivers less inflation and output gap volatility. It does so by anchoring inflation expectations thereby contributing to the overall stability of the economy. To achieve this result optimal policy is conditional on the degree of perceived persistence. As perceived persistence increases so does inertia in the policy response in the face of inflation shocks. We compare the contrast between the two policy regimes in the paper with the difference between the rational expectations under discretion and commitment.