The Optimality and Controllability of Discretionary Monetary Policy

This paper addresses two issues -- the time-inconsistency of optimal policy and the controllability of target variables within new-classical and new-Keynesian model structures. We can resolve both issues by delegation. That is, we design central bank loss functions by determining the two target values and the weight between the two targets. With a single decision maker, the time-inconsistency issue does not exist; the target controllability issue does. Delegating the long-run target values (target variables’ equilibriums under the Ramsey optimal policy) and the same weight as society to the central bank can achieve Ramsey optimality and path controllability. With multiple decision makers (game), both issues of time-inconsistency and target controllability exist and the delegation becomes more complicated. The long-run target values can only achieve asymptotic, not path, controllability. Path controllability requires the delegation of short-run target values, which commits or binds the central bank to follow exactly the Ramsey optimal paths. The short-run inflation target value conforms to the macroeconomic structure (i.e., Phillips curve). With path controllability, the constant average and state-contingent inflation biases are removed. To eliminate the stabilization bias, the delegated weight must differ from society in a dynamic game. When the Phillips curve exhibits output (inflation) persistence, the central bank must place more weight on output (inflation) stabilization. When the Phillips curve exhibits principally forward-looking behavior, the delegated weight can require a conservative or liberal central bank. In sum, delegating certain short-run target values and a different weight can cause discretionary monetary policy to prove Ramsey optimal and path controllable in a dynamic game.