Abstract:
The monetary policy literature has recently devoted considerable attention to Taylor-type rules, in which the interest rate set by the central bank depends on measures of inflation and aggregate output. We show that if policy-makers attempt to choose the optimal rule within a Taylor-type class they may be led to rules that generate indeterminacy and/or instability under learning. This problem is compounded by uncertainty about structural parameters. We advocate a procedure in which policy-makers restrict attention to rules that lie in the determinate stable region for all plausible calibrations, and which minimize the expected loss, computed using structural parameter priors, subject to this constraint