Directional accuracy, forecasting error and the profitability of currency trading: model-based evidence

Three models (the flexible-price monetary model, PPP and a univariate ARIMA model) are estimated for 45 currency pairs to find out if the profitability of forecasting-based currency trading is more related to the ability of the underlying model to predict the direction of change than the magnitude of the forecasting error. Theoretical considerations show that a correct prediction of the direction of change is neither a necessary nor a sufficient condition for a profitable trade. The results of the exercise indicate that profitability is more strongly correlated with directional accuracy than with the magnitude of the error.