Machine Learning and the Random Walk Puzzle: Forecasting the CAD/USD Exchange Rate with Expanding Window Evaluation and SHAP Interpretability
arXiv:2606.15058v1 Announce Type: new Abstract: This study examines whether machine learning (ML) models can outperform the naive random walk benchmark in forecasting the monthly USD/CAD exchange rate. Using daily data from the Bank of Canada spanning January 2017 to May 2026, resampled into 113 monthly observations, five ML models are evaluated: linear regression, random forest, gradient boosting, XGBoost, and AdaBoost. These models are benchmarked against the naive random walk model and exponential smoothing with Holt-Winters seasonality (ETS). All models are evaluated using an expanding-win
The proliferation of advanced machine learning techniques and increased computational power makes it timely to rigorously test their applicability to complex financial forecasting. Continuous advancements in ML algorithms necessitate ongoing evaluation against traditional benchmarks.
Improving currency exchange rate forecasting beyond naive benchmarks has significant implications for global trade, investment strategies, and central bank policy. Understanding these capabilities helps sophisticated readers assess the evolving landscape of financial markets.
The study validates the potential for ML models, particularly XGBoost, to offer more accurate short-term currency forecasts, potentially shifting how financial institutions approach predictive analytics. This enhances the toolkit available for managing currency exposure and making speculative bets.
- · Quantitative hedge funds
- · Forex traders using ML models
- · Financial data science platforms
- · Bank of Canada (if they adopt these models)
- · Traditional forecasting methodologies
- · Financial institutions slow to adopt ML
- · Investors relying solely on macroeconomic intuition
Financial institutions may allocate more resources to developing and integrating sophisticated ML models for currency forecasting.
Improved forecasting accuracy could lead to more efficient currency markets, potentially reducing arbitrage opportunities and volatility.
The demonstrated robustness of ML in currency forecasting could accelerate its adoption across other complex financial predictions, leading to a broader transformation of financial analysis.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG