Exchange Rate Forecasting with Advanced Machine Learning Methods

Historically, exchange rate forecasting models have exhibited poor out-of-sample performances and were inferior to the random walk model. Monthly panel data from 1973 2014 for ten currency pairs of OECD countries are used make out-of sample forecasts with artificial neural networks XGBoost models. Most approaches show significant substantial predictive power in directional forecasts. Moreover, evidence suggests that information regarding prediction timing is a key component performance.