Stock Pairs Trading

Pairs trading requires holding two positions simultaneously, so capital requirements are higher than single-stock trading. If you trade the legs as CFDs or spread bets, you post margin on both legs, but hedged (long one, short the other) positions usually get a margin offset, so the combined requirement is lower than two outright positions. For a basic banking pair like HSBC-Barclays, expect roughly GBP 4,000-5,000 of margin. Starting capital of GBP 10,000-15,000 allows proper position sizing and multiple pair opportunities. If you hold the long leg as cash shares, you need the full value of that leg plus margin on the short CFD/spread bet, so requirements are higher.

No, pairs trading is NOT risk-free. While it's market-neutral (immune to broad market direction), it has specific risks: (1) Spread blowout - the divergence continues instead of reverting, (2) Correlation breakdown - the historical relationship stops working, (3) Company-specific events - earnings or news affecting one stock permanently, (4) Execution risk - slippage when entering/exiting both legs. You can lose 10-20% on a pairs trade if the spread moves against you significantly, and leveraged CFDs/spread bets can lose more than your initial margin.

Look for brokers offering single-stock CFDs or spread bets so you can short either leg easily. Useful features: (1) A margin offset on hedged positions (IG, Interactive Brokers, Saxo, CMC Markets), (2) Basket or OCO orders for simultaneous execution, (3) Spread charting or pairs analysis tools (some platforms build this in), (4) API access if you want to automate (IG and Interactive Brokers offer APIs). For the long leg you can also use cash shares on a mainstream platform. AlgoKing integrates with multiple brokers and provides pairs analysis tools for simulation practice.

In UK markets, intraday pairs trades can work for highly liquid banking pairs. Most pairs trades last 2-10 days. Banking pairs typically revert in 3-7 days, data/IT pairs in 8-15 days. Set maximum holding periods based on the pair's historical half-life. If a trade hasn't begun reverting within 1.5x the expected half-life, exit regardless of P&L - the original thesis may no longer apply. Avoid holding pairs through weekends when unexpected news can occur.

Yes, and you'll usually need to. The key is matching pound exposure, not share count. With CFDs and spread bets there are no fixed lot sizes, so you can size each leg precisely. For example, to put about GBP 8,750 on each leg: HSBC at 1,750p means GBP 8,750 / GBP 17.50 = 500 shares; Barclays at 1,250p means GBP 8,750 / GBP 12.50 = 700 shares. If you also apply a hedge ratio (beta or regression), scale one leg by that ratio. The flexibility of CFDs/spread bets makes fine-tuning exposure straightforward.

Dividends create an artificial spread change. When a stock goes ex-dividend, its price drops by roughly the dividend amount, widening or narrowing the spread depending on which leg goes ex-dividend. Two approaches: (1) Avoid pairs with an imminent ex-dividend date on only one leg (check the corporate actions calendar), (2) Adjust your spread calculation for the expected dividend drop. With CFDs and spread bets a dividend adjustment is applied to your account (credited if long, debited if short), so the economic effect largely washes out; with cash shares the price simply drops on the ex-date. Either way, account for it in your spread calculations.

Distance method simply tracks the normalized price difference (spread) and trades when it exceeds historical thresholds - it's simpler but makes no assumption about mean reversion. Cointegration method uses statistical tests to confirm the spread is actually mean-reverting, providing stronger theoretical foundation. In practice, cointegration method has higher win rate but fewer signals. Distance method generates more trades but with lower conviction. Most professionals use cointegration as primary filter, then distance/Z-score for entry timing.

Hedge ratios aren't static - recalculate at least weekly. The relationship between stocks changes due to earnings impact, market cap changes, and evolving business fundamentals. Use rolling regression (60 days for active trading) to update hedge ratios. If hedge ratio changes dramatically (>20% shift), the pair may be entering an unstable period - consider reducing position or avoiding the pair temporarily. For existing positions, you can adjust by adding/reducing shares in one leg to maintain neutrality, but frequent adjustments increase transaction costs.

Yes, index vs constituent pairs trading (basis trading) is a valid strategy. Long a UK banks ETF or FTSE 350 Banks exposure, short individual bank stocks (or vice versa) exploits deviations in the index vs component relationship. Key considerations: (1) Calculate proper weights - HSBC is a large weight in UK bank indices, so short a proportional amount of the index exposure in HSBC, (2) This is essentially betting on a stock's relative performance vs the index, (3) Track basis (spot vs futures/ETF) separately from the pair spread. This strategy works well around index rebalancing events when weights change.

For active individual traders, 3-5 pairs is optimal. This provides diversification without overwhelming monitoring capacity. Each pair requires attention - watching both legs, monitoring correlation, tracking spread. More than 5 pairs splits focus too much and increases execution complexity. Capital efficiency also matters: with hedged margin offsets, 3-5 pairs typically utilizes 50-70% of available capital. Systematic/algorithmic traders can handle 20-30 pairs with automation, but discretionary traders should stay focused.

Kalman filter provides an adaptive hedge ratio that responds to recent price action while smoothing noise. Implementation: (1) State equation: hedge ratio follows a random walk plus noise, (2) Observation equation: Stock A return = hedge ratio × Stock B return + noise, (3) Update the hedge ratio each period using the Kalman gain. Advantage over rolling regression: Kalman adapts faster to regime changes while being more stable than a short-lookback regression. Python's pykalman or a manual implementation works. The key parameter is process variance - higher values make the filter more responsive but noisier.

Joint hypothesis problem: when backtesting pairs, you're simultaneously testing if (1) your pair selection criteria work, (2) your entry/exit rules work, (3) your position sizing works. If results are bad, you don't know which component failed. Solution: test components separately. First, test if selected pairs actually mean-revert out-of-sample (regardless of trading rules). Then, given mean-reverting pairs, test different entry thresholds. Finally, test sizing rules. Use walk-forward optimization with separate in-sample selection and out-of-sample validation periods to avoid overfitting.

PCA identifies common factors driving returns across stocks. In stat arb, decompose universe returns into principal components (PC1 might be the market factor, PC2 sector rotation, etc.). Calculate each stock's loadings on the PCs. A stock's residual return (after removing PC influences) represents its idiosyncratic movement. Pairs trade on residual spreads rather than raw price spreads - this isolates true relative value from common factor moves. PCA-based stat arb typically shows lower volatility and more consistent returns because you're trading a purer spread after removing market and sector effects.

Microstructure matters significantly for pairs profitability. Key effects: (1) Bid-ask spread - you pay the spread on both legs, eating into narrow convergence profits. Calculate round-trip cost as a % of expected profit. (2) Market impact - larger orders move prices; stagger entry for large positions. (3) Lead-lag effects - one stock may lead another by seconds/minutes; the leader provides signals, trade the lagger. (4) Order book imbalance - check depth on both sides before entering. Solutions: use limit orders, break up large trades, avoid illiquid times (the open and the last minutes before the close). Microstructure costs can consume 30-50% of gross profit if not managed.

Cost-robust design principles: (1) Widen entry thresholds beyond the theoretical optimal - enter at Z-score ±2.5 instead of ±2.0 to ensure the move is significant enough to cover costs, (2) Include realistic costs in the backtest - 0.1% round-trip minimum for liquid CFDs/shares, higher for illiquid stocks, plus CFD/spread-bet financing on overnight holds, (3) Filter by expected holding period - very quick reversion means more turnover and more costs; prefer pairs with moderate half-life (5-10 days), (4) Scale position with spread extremity - larger positions at Z-score ±3.0 where expected return is higher, (5) Partial exits reduce impact cost vs a full position close, (6) Avoid pairs with wide bid-ask spreads. Net Sharpe after costs should exceed 1.0 for a viable strategy.