Forecasting the Taylor rule exchange rate model using directional change tests

Abstract

This study uses the Taylor rule model of exchange rate determination, to analyse how accurately it can predict directional changes in the exchange rate. Using bilateral exchange rate data for the US, UK, Sweden and Australia, we conduct the Pesaran-Timmermann test to determine how accurately this model can forecast changes in direction. The results suggest that although in many studies the standard out-of-sample forecasting ability of this model has been successful, the performance of the change of direction predictions are not consistently accurate over all specifications tested, in which case they may not prove profitable in a trading environment.

Keywords

• forecast| Taylor rule| exchange rate| prediction accuracy