Discussion of “Is Optimal Monetary Policy Always Optimal?”∗

A burgeoning part of the monetary policy design literature posits that optimal monetary policy is the solution to a constrained optimization problem where the monetary authority has the discretion to minimize a social welfare function, taking the economy and other policies as given. In their paper “Is Optimal Monetary Policy Always Optimal?” Troy Davig and Refet Gurkaynak (henceforth, DG) argue convincingly that the solution thus attained may fail to deliver the desired outcome. The key insight is that any economy is characterized by multiple inefficiencies and multiple policymakers with possibly conflicting objectives, not necessarily coinciding with the social welfare function. As a result, designing policy under the false assumption that the central bank is the “only game in town” has undesirable side effects that induce inefficiencies. For example, political considerations may introduce elements other than the social welfare function to the objectives of fiscal policy. In this environment, optimality cannot be attained unless the different policies in question are jointly examined and formulated. Designing “optimal policy” is a question of the optimal mix of various policies (e.g., monetary, fiscal, regulatory, structural) for which different policymakers are responsible. The problem of “optimal policy” is not the same as the problem of “optimal monetary policy,” and equating the two yields flawed conclusions both for what monetary policy can do and for what it should do.