Kalman Filter Pairs Trading in Chinese Futures Markets
Summary
This project describes a statistical arbitrage strategy for Chinese futures. It screens contract pairs with an Augmented Dickey-Fuller test for stationary spreads, estimates a dynamic hedge ratio with a Kalman filter, and uses the spread’s half-life to set a rolling window for z-score signals. The stated rules enter short positions when the z-score crosses an upper threshold and long positions when it crosses a lower threshold, then exit at zero. Pairs receive equal market-value allocations in the portfolio.
The study uses daily main-contract data from four Chinese exchanges, splitting the sample into in-sample and out-of-sample periods. It reports portfolio performance statistics and says the out-of-sample portfolio had a higher expected daily return and CAGR, but a lower daily Sharpe ratio and longer average drawdown duration. These findings come from a limited historical backtest. The author notes that fees and slippage are excluded, and main-contract data must be mapped to tradable contracts for live use. The sample is short, and the reported results do not establish future profitability.
Key ideas
- The strategy selects futures pairs whose spreads pass a stationarity test.
- A Kalman filter updates the hedge ratio over time.
- Spread half-life informs the rolling period used to calculate z-scores.
- Entry thresholds trigger contrarian positions, with exits at a zero z-score.
- The reported backtest omits transaction costs and slippage.
Tags
Cited by
- Strategies Gold Miners vs Gold Beta-Hedged Residual Reversion: fade a >=2-sigma 10-session GDX-vs-GLD idiosyncratic move, dollar/beta-neutral long-short, 15-session time stop (USEQ 1-DAY)
- Hypotheses Gold Miners vs Gold Beta-Hedged Residual Reversion: fade a >=2-sigma 10-session GDX-vs-GLD idiosyncratic move, dollar/beta-neutral long-short, 15-session time stop (USEQ 1-DAY)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.