α
OpenAlpha
Back to strategies
Cross-AssetRisk: Medium

Cross-Asset Pairs

Trade asset pairs connected by real economic relationships

Backtest this strategySee live signals

Risk

Medium

Holding Period

Days to weeks

Best For

When fundamental economic links temporarily break down

How it works

Trade asset pairs connected by real economic relationships

Mathematical Foundation

spread = price_A − β × price_B Z = (spread − μ) / σ

Signal Generation Logic

  1. 1Select pairs that span different asset classes (e.g., AUD/USD vs gold, EUR/USD vs SPX500)
  2. 2Run Engle-Granger cointegration test on each cross-asset pair (p-value < 0.05 required)
  3. 3Estimate hedge ratio β via OLS regression
  4. 4Calculate rolling spread mean and standard deviation over the cointegration window
  5. 5Enter when Z-score crosses ±2.0 — wider threshold reflects higher cross-asset noise
  6. 6Exit when |Z-score| < 0.5 or if cointegration relationship expires
  7. 7Positions are hedged across both legs to be dollar-neutral

Parameters Explained

entry_z

Z-score threshold to enter. Set higher (2.0) than standard pair trading because cross-asset relationships tend to have more noise.

Default

2
exit_z

Z-score threshold to exit. When spread returns to within this band of the historical mean, close the position.

Default

0.5
coint_window

Rolling window for cointegration testing and spread statistics. 120 days (~6 months) balances stability with responsiveness.

Default

120

When It Works

✓

When macroeconomic fundamentals create stable relationships between different asset classes — such as commodity-currency links (AUD/gold), risk-asset correlations (EUR/SPX), or flight-to-safety pairs. Best in stable macro environments with moderate volatility.

When It Fails

✗

During macro regime shifts when cross-asset correlations break down (e.g., risk-off environments where all correlations go to 1). Also fails when commodity prices are driven by supply shocks rather than demand/currency factors.

Risks & Limitations

  • Cross-asset cointegration is less stable than same-asset pairs and can break suddenly
  • Higher transaction costs due to trading different instruments on different venues
  • Currency conversion risk when hedging cross-currency pairs
  • Macro-driven correlation breakdowns can be sharp and persistent
  • Requires broader market knowledge to interpret cross-asset signals correctly

Implementation

Extends the standard pair-trading framework to cross-asset pairs. Uses the same Engle-Granger test and OLS hedge ratio, but with a wider entry threshold and closer monitoring of cointegration stability. Assets from OANDA (forex) and Yahoo Finance (indices, commodities) are combined.

Model parameters

Entry Z-Score

Open when spread Z ≥ 2σ

2.0

Exit Z-Score

Close when spread Z ≤ 0.5σ

0.5

Coint Window

Days for cointegration test

120d

Academic background

Academic Basis

Based on Gatev, Goetzmann & Rouwenhorst (2006), 'Pairs Trading: Performance of a Relative-Value Arbitrage Rule', Review of Financial Studies

Backtest this strategy

Run the exact model on your selected assets and date range. See trade-by-trade performance.

Backtest This →