Paper: arXiv 2406.06706
Authors: Qiqin Zhou
Abstract
\begin{“abstract”} In this paper, we integrated the statistical arbitrage strategy, pairs trading, into the Black-Litterman model and constructed efficient mean-variance portfolios. Typically, pairs trading underperforms under volatile or distressed market condition because the selected asset pairs fail to revert to equilibrium within the investment horizon. By enhancing this strategy with the Black-Litterman portfolio optimization, we achieved superior performance compared to the S&P 500 market index under both normal and extreme market conditions. Furthermore, this research presents an innovative idea of incorporating traditional pairs trading strategies into the portfolio optimization framework in a scalable and systematic manner.
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper integrates advanced Bayesian methods (Black-Litterman) with co-integration analysis, involving rigorous statistical formulas, but is heavily backtest-driven with specific date ranges, prescreening logic, and reported out-of-sample performance, indicating strong implementation focus.
Research Flowchart
flowchart TD A["Research Goal<br>Develop scalable arbitrage strategy<br>using Black-Litterman & Pairs Trading"] --> B["Data Collection<br>Historical Equity Data<br>Market Regimes (Normal/Extreme)"] B --> C["Statistical Arbitrage<br>Pair Selection &<br>Spread Deviation Analysis"] C --> D["Black-Litterman Input<br>Implied Equilibrium Returns<br>Trusted Investor Views"] D --> E["Bayesian Integration<br>Posterior Expected Returns<br>Portfolio Optimization"] E --> F["Outcome<br>Efficient Mean-Variance Portfolio<br>Outperforms S&P 500<br>Robust to Market Conditions"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes integrating a pairs‑trading statistical arbitrage strategy into the Black‑Litterman portfolio optimization framework to construct efficient mean‑variance portfolios.
Source sentence
In this paper, we integrated the statistical arbitrage strategy, pairs trading, into the Black-Litterman model and constructed efficient mean-variance portfolios.
Assumes 6
- A fixed transaction cost of 5 basis points per trade is applied to both opening and closing positions. execution
Source sentence
We assume a fixed transaction cost of 5 bps per trade executed, for both opening or closing of a long or short position.
- A short‑sale cost of 1% per annum is deducted daily from the profit of short positions. execution
Source sentence
The annualized short-sale cost is assumed to be 1%.
- Co‑integration tests are performed on a rolling window of 60 days. data
Source sentence
We chose the window size w to be 60 days as it is approximately the length of one earnings cycle.
- Selection frequency must be greater than 10 trading days to capture mean‑reversion profits. model
Source sentence
greater than 10 days to ensure that the profit generated by the pairs can be fully captured.
Show 2 more
- Asset returns are normally distributed with mean m and covariance matrix S. statistical
Source sentence
Assuming there are N assets in the market with returns distributed normally with mean m and covariance matrix S, denoted as R ∼N(m, S).
- The prior view on returns is expressed via a linear system with a picking matrix P and view vector Q. model
Source sentence
We express our views on the expected returns using a linear system of equations. Let K be the total number of views, P be a K × N picking matrix, Q be a K-dimensional view vector.
Data
2016 – 2018 · S&P 500 index · equities · daily
- Historical stock price data (training set) market-prices unknown
Source sentence
We used 2006 to 2016 as the training set and 2016 to 2018 as the out-of-sample test set.
- Historical stock price data (out‑of‑sample test) market-prices unknown
Source sentence
We used 2006 to 2016 as the training set and 2016 to 2018 as the out-of-sample test set.
Costs
Modeled
Fixed transaction cost of 5 bps per trade and an annualized short‑sale cost of 1% are subtracted from daily profit.
Source sentence
We assume a fixed transaction cost of 5 bps per trade executed, for both opening or closing of a long or short position. The
Finds 1
Sharpe 1.78 (benchmark 1.41) · max drawdown 8.4%
- During the 2007‑2009 financial crisis the strategy showed higher returns than the S&P 500 in the later period while maintaining similar patterns early on. performance
Source sentence
The return exhibits a similar pattern to the S&P 500 index, especially in the first 100 days. However, our portfolio has a higher return over the S&P 500 during the later period.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 8 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (14)
- Testing market efficiency using statistical arbitrage with applications to momentum and value strategies (2004)
- Statistical arbitrage in the US equities market (2010)
- Pairs trading and selection methods: is cointegration superior? (2015)
- Pairs Trading: Quantitative Methods and Analysis (2011)
- Portfolio Optimization with Robust Covariance and Conditional Value-at-Risk Constraints — → 2406.00610
- Cointegration and error correction: Representation, estimation, and testing (1987)
- Asset Allocation: Combining Investor Views with Market Equilibrium (1990)
- Portfolio selection (1952)
and 6 more
- A step-by-step guide to the Black-Litterman model (2007)
- Indexing and statistical arbitrage (2005)
- The high sensitivity of pairs trading returns (2013)
- Real-time segmentation of on-line handwritten arabic script (2014)
- Fast classification of handwritten on-line Arabic characters (2014) — link
- Estimate and Replace: A Novel Approach to Integrating Deep Neural Networks with Existing Applications (2018) — link
1 of 14 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.