Monetary policy is often analysed in terms of simple rules. Such rules may be useful for many purposes, even when they do not describe the actual monetary policy strategy exactly. This paper compares monetary policy in Sweden during the inflation-targeting regime 19932002 with the policies implied by certain simple instrument rules. Calibrated rules that are commonly used in theoretical analyses do not provide good approximations of Sveriges Riksbanks (the central bank of Sweden) policy, whereas rules with reaction coefficients that have been estimated using the banks own (realtime forecasts do capture policy behaviour quite well. There are different forecastbased rules including different arguments and forecast horizons that describe monetary policy about equally well. A close reading of various policy documents, e.g., Inflation Reports, minutes from the banks Executive Board meetings, and speeches, shows that large deviations from the simple rules are associated with factors that are usually neglected in theoretical models of monetary policy. Examples of such factors are concerns for credibility and uncertainties about various economic relationships.
monetary policy inflation target Taylor rule forecast-based policy rules transparency