Pairs trading, or statistical arbitrage, is one of the oldest quantitative trading strategies, and it is still employed today. Over the years, it has expanded from classical distance methods to more sophisticated approaches, and practitioners have increasingly questioned its profitability.
Reference [1] provides a thorough review of the pairs trading literature between 2016 and 2023. The findings are as follows.
- Distance Methods
Distance-based approaches focus on selecting trading pairs using measures such as the sum of squared errors (SSE) or absolute errors (SAE) of normalized price differences. These methods provide simple and intuitive frameworks for identifying co-moving assets and have shown consistent profitability across global markets, including during downturns. Research cited in the document highlights strong market-neutral properties and robustness even after transaction costs. Future work may extend distance methods using richer optimization frameworks, alternative similarity metrics, and broader datasets.
- Cointegration Methods
Cointegration techniques rely on long-run equilibrium relationships between asset prices, providing a theoretically grounded basis for pairs trading. The document notes extensive evidence supporting their validity across equity and bond markets. Advances involve adaptive modeling, regime-switching structures, and incorporating external variables such as macroeconomic or ESG data. Future work aims to strengthen resilience by integrating alternative datasets and improving modeling flexibility under complex market conditions.
- Stochastic Control Methods
Stochastic control frameworks treat pairs trading as a continuous-time optimization problem, dynamically adjusting positions based on spread levels, horizon risk, and divergence risk. These methods extend the classical OU process to include jump-diffusions, regime switching, and stochastic volatility, improving realism and adaptability. The document emphasizes strong empirical performance across various markets, while also noting practical challenges such as transaction costs and liquidity constraints. Future research includes integrating AI/ML for improved adaptability and explicitly modeling trading frictions.
- Time Series Methods
Time series techniques—including GARCH models, OU processes, and fractional OU extensions—focus on short-term dynamics, volatility clustering, and mean reversion. They allow adaptive trading thresholds based on volatility forecasts and have demonstrated improved returns even after accounting for transaction costs. The document highlights opportunities for hybrid models, combining time series techniques with machine learning, copulas, or stochastic control, as well as incorporating slippage, liquidity constraints, and application to emerging markets and high-frequency settings.
- Other Methods (Copula, Hurst Exponent, Entropic Approaches)
The document identifies several alternative approaches designed to capture features that traditional statistical methods miss. Copula methods model complex joint distributions and tail dependencies; Hurst exponent approaches capture long-memory effects; and entropic methods account for model uncertainty. These techniques enhance robustness by addressing nonlinear dependence structures and heavy-tail behavior in spreads. Future research may refine these methods, integrate them with machine learning, and test them across diverse asset classes and market regimes.
This comprehensive review of statistical arbitrage strategies will assist practitioners in research, particularly in adapting these methods to new asset classes such as crypto.
Regarding the profitability, we believe that while simple methods have historically been profitable, increasing competition and market efficiency mean that more sophisticated approaches are often required to maintain or enhance profitability. However, sophistication alone is not sufficient; effectiveness depends on model design, data quality, and market conditions.
Let us know what you think in the comments below or in the discussion forum.
References
[1] Sun, Y. (2025). A Review of Pairs Trading: Methods, Performance, and Future Directions. WNE Working Papers, 19/2025 (482). Faculty of Economic Sciences, University of Warsaw.
Post Source Here: A Recent Review of Pairs Trading and Statistical Arbitrage
source https://harbourfronts.com/recent-review-pairs-trading-statistical-arbitrage/
No comments:
Post a Comment