Cointegration-Based Mean-Reversion Pairs Trading in Indian Equities

Summary

This project describes a market-neutral pairs-trading approach using NSE-listed stocks from different sectors. It selects candidate pairs within a sector, estimates a hedge ratio with ordinary least squares, forms a price spread, and applies an Augmented Dickey-Fuller test to assess stationarity. Pairs meeting the stated significance threshold are considered for trading. The strategy takes a long position when the spread falls below a threshold and a short position when it rises above one, closing positions as the spread returns toward its mean. The project compares Bollinger Band and Z-score signal approaches.

The study uses adjusted closing prices over a historical period and reports that some pairs showed cointegration, while statistical significance alone did not ensure profitability. It says the Bollinger Band approach performed better than the Z-score approach in its test, but provides no detailed performance figures in the text. The author notes sensitivity to volatility and external conditions, execution costs, and differences between companies as limitations. Results are specific to the tested pairs and period; further testing is needed before drawing conclusions about live trading.

Key ideas

• The strategy seeks pairs whose spread is stationary and tends to revert toward its mean.
• Ordinary least squares is used to estimate the hedge ratio between two stocks.
• The project applies an Augmented Dickey-Fuller test and uses spread thresholds to generate trades.
• A significant cointegration test does not guarantee a profitable strategy.
• Transaction costs, execution risk, and differences between companies can weaken results.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.