Forecasting monetary rules in South Africa

Abstract

This paper is the first one to analyze the ability of linear and nonlinear monetary policy rule specifications as well as nonparametric and semiparametric models in forecasting the nominal interest rate setting that describes the South African Reserve Bank (SARB) policy decisions. We augment the traditional Taylor rule with indicators of asset prices in order to account for potential financial stability targets implicitly considered by the SARB. Using an in-sample period of 1986:01 to 2004:12, we compare the out-of-sample forecasting ability of the models over the period 2005:01 to 2008:12. Our results indicate that the semiparametric models perform particularly well relative to the Taylor rule models currently dominating the monetary policy literature, and that nonlinear Taylor rules improve their performance under the new monetary regime.