Finance and Economics Discussion Series (FEDS)
September 2025 (Revised September 2026)
Virtue or Mirage? Complexity in Exchange Rate Prediction
Abstract:
Does the virtue of complexity extend to exchange rate forecasting? We test whether Ridge regression with Random Fourier Features (Ridge–RFF) outperforms OLS and the driftless random walk for eight major USD currency pairs across two predictor sets — traditional monetary and Taylor-rule fundamentals — using the Kelly et al. (2024) framework to diagnose why complexity fails and what signal conditions its success would require. Ridge–RFF substantially outperforms OLS at short training windows, reducing MSPE by up to 97% under Taylor-rule fundamentals, but predictive equality tests against the random walk favour Ridge–RFF for only three currencies (AUD, CHF, and JPY) under Taylor-rule fundamentals and one (JPY) under traditional monetary fundamentals. At longer training windows, the Taylor-rule gains persist robustly across intercept specifications while traditional monetary fundamentals offer no reliable random walk outperformance. Taylor-rule market-timing strategies generate portfolio Sharpe ratios approaching 0.40 at the longest training window, but these collapse once an intercept is included, exposing drift capture rather than genuine directional predictability. Tracing the VoC curve across the full regularisation spectrum confirms the failure is structural: the cross-currency mean out-of-sample R2 is bounded above by zero everywhere. The average predictive signal per feature is indistinguishable from zero, implying an optimal regularisation 1012–1013 times above the equity baseline — quantifying both why complexity fails in FX and what signal conditions would be needed for it to succeed.
Keywords: Model complexity, machine learning, Ridge, RFF, FX rate, disconnect puzzle, predictability
DOI: https://doi.org/10.17016/FEDS.2025.089r1
PDF: Full Paper
Original Paper: Accessible materials (.zip) | PDF
Disclaimer: The economic research that is linked from this page represents the views of the authors and does not indicate concurrence either by other members of the Board's staff or by the Board of Governors. The economic research and their conclusions are often preliminary and are circulated to stimulate discussion and critical comment. The Board values having a staff that conducts research on a wide range of economic topics and that explores a diverse array of perspectives on those topics. The resulting conversations in academia, the economic policy community, and the broader public are important to sharpening our collective thinking.