Unhedged Currency Exposure & FX Risk

Goal Portfolio Recommendation Accuracy Frequency Common Category Financial Services Published View source on GitHub ↗

Issue: Model Recommends International Assets Without Accounting for Currency Risk; FX Movements Dwarf Asset Returns

Frequency: Common

Symptoms

  • Recommends foreign assets with strong fundamentals
  • Currency moves opposite to asset returns (correlation -0.8 to -1.0)
  • Returns in home currency: negative, despite positive local returns
  • Model blind to currency beta

Root Cause Models trained on local returns (asset returns in its currency). When investing in foreign assets from home perspective, FX risk is huge (often >50% of volatility for small countries). Models don’t account for this unless data is already currency-adjusted.

Example

Scenario: US investor in Euro assets
European asset: +5% return in EUR
EUR/USD exchange rate: -5% (EUR weakens)
Return to US investor: 0% (or -5% if unlucky)
Model recommendation: Still positive (didn't account for FX)
Impact: Currency losses offset gains; client disappointed

Key Statistics

  • FX volatility: 5-20% annual depending on currency pair
  • FX correlation with assets: Often -0.5 to -0.8 (diversifying but risk-hiding)

Mitigation Strategies

  1. Currency-Adjusted Returns: Use home-currency returns in training
  2. FX Hedging: Recommend hedging for FX exposure or accept FX risk
  3. Currency Beta: Measure and disclose currency sensitivity
  4. Regional Diversification: Diversify across regions to average FX

Metrics

  • Return in home currency vs. asset currency
  • Currency contribution to volatility

Alerts

  • Unhedged FX exposure >30% of portfolio → Require FX consideration

References