Pairs trading is a market-neutral quantitative strategy that profits from the relative price movement between two historically correlated securities. By simultaneously going long one stock and short another, the position remains largely unaffected by overall market direction. When the price spread between the two stocks deviates from its long-term equilibrium, traders buy the underpriced stock and sell the overpriced one, betting that the spread will revert to its mean.
Research shows that a simple pairs trading rule can yield average annualized excess returns of up to 11% for self-financing portfolios. This strategy has been a cornerstone of hedge fund portfolios for decades and is accessible to retail traders with the right framework.
How the Strategy Works
The core of pairs trading is the spread — the computed price difference between two securities:
Spread = Price of Stock A − Hedge Ratio × Price of Stock B
The hedge ratio determines the appropriate proportional relationship between the two stocks, typically calculated using regression analysis. When the spread deviates more than ±2 standard deviations from its mean (Z-score), a trading signal is generated. This statistical approach removes the need to predict whether the broader market will rise or fall.
Signal Thresholds
| Condition | Action |
|---|---|
| Z-score > +2.0 | Long Stock A, Short Stock B |
| Z-score < -2.0 | Short Stock A, Long Stock B |
| Z-score reverts to ±0.5 | Close positions for profit |
| Z-score exceeds ±3.5 | Stop-loss triggered — close to limit loss |
Pair Selection
Successful pair selection requires rigorous correlation analysis and cointegration testing. The best candidates satisfy these criteria:
- High Correlation: Correlation coefficient ≥ 0.80
- Cointegration: Engle-Granger test p-value < 0.05, confirming a long-run equilibrium relationship between the two assets
- Economic Logic: Both stocks share the same industry or common economic drivers, ensuring the relationship has fundamental justification
The selection process first groups stocks by sector, applies statistical tests, and validates with economic reasoning. Pairs without a fundamental linkage may show spurious historical correlation that will not persist over time.
Risk Management
Pairs trading appears low-risk due to its market-neutral design, but has specific risks that must be managed:
- Spread non-reversion: The expected convergence may not occur if the fundamental relationship between the stocks has structurally changed — such as after a merger, new product launch, or regulatory shift
- Stop-loss levels: Typically set at ±3.5 to ±4.0 standard deviations to cap losses when the spread continues widening beyond expectations
- Position sizing: Equal dollar exposure on both legs, with each trade risking 1-2% of total capital
- Transaction costs: Commissions and slippage can erode small profit margins; accurate cost estimation is essential to ensure net returns remain positive
Algo Lab Application
Our quantitative engine applies pairs trading models to 8,000+ US stocks daily. By computing spreads and Z-scores, we automatically identify stock pairs that meet trading criteria — delivering signals to VIP members via Telegram at 4 PM Hong Kong time.
Internal Links
- Statistical Arbitrage — The statistical foundation of passes trading
- Quantitative Knowledge — Overview of quantitative strategies
Call to Action
Ready to receive daily pairs trading signals? Start your free VIP trial.
FAQ
Q: What is pairs trading?
A: A market-neutral strategy that simultaneously buys one stock and sells a correlated one, profiting from the relative price movement between them.
Q: How are pairs selected?
A: Pairs are screened using high correlation (≥0.80), cointegration testing (p < 0.05), and economic logic to confirm a lasting relationship.
Q: How does Z-score work in pairs trading?
A: Z-score measures how far the spread has deviated from its mean in standard deviations. ±2.0 triggers signals; ±0.5 indicates convergence for exit.
Q: Why are stop-losses important in pairs trading?
A: If the spread fails to revert and continues widening, stop-losses at ±3.5 to ±4.0 standard deviations limit losses when the expected convergence does not materialize.
Q: How do transaction costs affect pairs trading?
A: Commissions and slippage can reduce profits. Accurate cost estimation is essential to ensure the net return remains positive.