BANKING ON CURRENCY FORECASTS - HOW PREDICTABLE IS CHANGE IN MONEY

The paper examines the predictive performance of four structural exchange rate models using both parametric and nonparametric techniques. Error correction versions of the models are fit so that plausible long-run elasticities can be imposed on the fundamental variables of each model. A variety of model evaluation statistics are reported. Our findings confirm that fundamental exchange rate models forecast no better than a random walk model for short-term prediction horizons. For longer horizons, error correction terms can explain exchange rate movements significantly better than a no change forecast for a subset of the models and currencies we consider.