Policy irreversibility and interest rate smoothing

Many empirical studies argue that the inertial behavior of the policy rates in industrialized countries can be well explained by a linear partial adjustment version of the Taylor rule. However, the explanatory power of the lagged interest rate has been questioned from various points of view. This paper formally examines a situation in which a central bank has an aversion for frequent policy reversals. Imposing an irreversibility constraint on the control space makes the lagged interest rate a state variable, but the policy function cannot then be expressed as a partial adjustment form even if the original Taylor rule is the correct policy function in the absence of the constraint. The simulation results reveal that the conventional regression tends to falsely support the functionally misspecified partial adjustment model. This implies that the significant role of the lagged interest may simply reflect the central banks’ reversal aversion.