Rules of Percentage Rate in Conditions of Uncertainty
The paper is a review of research on the subject of the effectiveness of monetary policy implemented in conditions of uncertainty, when the policy itself is identified with the rule of percentage rate most often the simple rule both with the traditional interpretation of the nature of the uncertainty, i.e. Bayesian uncertainty (effective rules) and uncertainty in Knight’s sense (effectively resistant rules).
In successive steps of the analysis the changes are investigated which should be introduced to the effective rule in order to take into consideration the risk/uncertainty located in different points of the decision process: inaccuracy of evaluations of the strength of the relationship between instruments and goals, inaccuracy of evaluations of inertial effects, incomplete knowledge about the current state of the economy, etc. Also investigated are the consequences of uncertainty of a more fundamental character: for example the existing but not explicit error of model specification (cause-effect description of the economy) and even uncertainty of the paradigm. The results of theoretical and empirical works show that Brainard’s principle of conservatism, for many years being an unwritten but obliging indicator how to conduct monetary policy, has not got the value of generality it is an effective method of facing uncertainty only in a particular situation (uncertainty of evaluation of the direct multiplier or imprecise data and the Bayesian loss function). With uncertainty located in other places of the model and the search for resistant rules an energetic policy of percentage rates may be desirable. Comparison of the rules themselves and the effects of their implementation with different location of the uncertainty and different methods of estimating the costs of possible errors leads to the somewhat banal conclusion: an effective monetary policy can be conducted when the uncertainty has not got a total (astructural) character and when it is known where the errors can appear when the decision-maker is aware of what he really does not know.
