Pairs Trading FTSE-DAX

Pairs trading involves taking opposite positions in two correlated assets (e.g., long FTSE, short DAX). You profit from the relative performance, not the absolute direction. When the spread between them deviates from historical norms, you bet it will return to normal.

FTSE and DAX are major European indices with high historical correlation (~0.85-0.95). They're both affected by European economic conditions but have different sector compositions. When they diverge, they tend to converge again, creating trading opportunities.

The hedge ratio determines how much of each asset to trade to create a market-neutral position. It's calculated from regression. For example, if hedge ratio is 0.5, for every $1 of FTSE, you trade $0.50 of DAX (or adjust position sizes accordingly).

Singapore traders can access FTSE-DAX via: CFDs (IG, CMC, Saxo), ETFs (iShares FTSE 100, Xtrackers DAX), or Futures (through Interactive Brokers). Trading hours are 3 PM - 11:30 PM SGT.

The primary target is Z-Score = 0, meaning the spread has returned to its historical mean. This captures the mean reversion. Some traders exit at Z = ±1 for partial profits or trail to capture more.

Cointegration is a statistical property where two non-stationary series combine to form a stationary (mean-reverting) spread. It's stronger than correlation - two series can be highly correlated but not cointegrated. Test using Engle-Granger or Johansen.

Half-life is the expected time for the spread deviation to reduce by 50%. Calculated from mean reversion speed. If half-life is 10 days, expect spread to be 50% closer to mean after 10 days. Use for setting realistic holding periods.

FTSE is GBP-based, DAX is EUR-based. Options: 1) Ignore if holding period is short, 2) Use currency-hedged ETFs, 3) Hedge net currency exposure with FX forwards. Monitor GBP/EUR for major moves during trade.

If rolling correlation drops below 0.75-0.80, the relationship may be breaking down. Consider exiting or tightening stops. Monitor for structural changes (policy divergence, Brexit effects) that could cause permanent divergence.

Weekly recalculation is standard. If hedge ratio changes more than 10-20%, consider rebalancing positions. More frequent in volatile periods. Balance accuracy against transaction costs.

OU process models mean-reverting behavior: dS = θ(μ - S)dt + σdW. θ = mean reversion speed, μ = long-term mean, σ = volatility. Used to estimate half-life (-ln(2)/θ) and model spread dynamics for optimal entry/exit.

Kalman filter provides time-varying hedge ratio that updates as the relationship evolves. Unlike fixed OLS, it adapts to regime changes. State-space model treats hedge ratio as hidden state that changes over time.

Spread behavior differs across regimes - high correlation (strong mean reversion) vs low correlation (breakdown). Markov-switching models identify current regime. Adapt strategy: normal trading in high-corr regime, avoid in low-corr.

ML can: 1) Random forest to classify good vs bad trade setups, 2) LSTM/GRU to forecast spread direction, 3) Gradient boosting for reversion probability. Features: Z-Score, correlation, RSI, sector spreads, macro factors.

Long spread: Buy FTSE call + DAX put. Short spread: Buy FTSE put + DAX call. Risk reversals for zero-cost exposure. Vol arbitrage if relative IV is mispriced. Options provide defined risk and leverage.