Paper: arXiv 2402.08108

Authors: Kasper Johansson, Thomas Schmelzer, Stephen Boyd

Abstract

We propose a new method for finding statistical arbitrages that can contain more assets than just the traditional pair. We formulate the problem as seeking a portfolio with the highest volatility, subject to its price remaining in a band and a leverage limit. This optimization problem is not convex, but can be approximately solved using the convex-concave procedure, a specific sequential convex programming method. We show how the method generalizes to finding moving-band statistical arbitrages, where the price band midpoint varies over time.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents advanced mathematical optimization using convex-concave procedures and non-convex formulations (high math), while providing structured numerical experiments with backtest-ready metrics like profit and out-of-sample periods (moderate-high rigor).

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Find multi-asset statistical arbitrages"] --> B{"Data & Inputs"}
  B --> C["Asset Price Data<br>Lev. Constraint Bound<br>Band Width Parameters"]
  
  C --> D["Formulate NCCP<br>Maximize Volatility<br>Subject to Band & Lev."]
  
  D --> E["Sequential Convex<br>Optimization"]
  E --> F{"Convex-Concave<br>Procedure"}
  F -->|"Converges"| G["Optimal Portfolio<br>Moving-Band Arb. Strategy"]
  F -->|"Iterate"| E
  
  G --> H["Outcomes:<br>Generalizes pair trading to N assets<br>Efficient moving-band solver<br>Volatility scaling identified"]

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 a new convex-concave optimization method for finding statistical arbitrage portfolios with high volatility within a price band and leverage limit.

Source sentence
We propose a new method for finding statistical arbitrages that can contain more assets than just the traditional pair.

Assumes 6

  • Asset prices are adjusted for dividends and splits. data
    Source sentence
    (We presume these are adjusted for dividends and splits.)
  • The price band width is fixed to 2 (units). model
    Source sentence
    Here we fix the width of the price band as 2; s and µ can always be scaled so this holds.
  • Leverage is limited by the average total position |s|T ¯P ≤ L. model
    Source sentence
    The constraint |s|T ¯P ≤L is a leverage constraint; it limits the total position of
  • Portfolio value must remain non‑negative (no bust). execution
    Source sentence
    we do not want the portfolio to have negative value, i.e., to go bust.
Show 2 more
  • Short positions are fully collateralized with at least half of initial investment. execution
    Source sentence
    this constraint ensures that the short positions are always fully collateralized by the long positions plus the cash account, with a margin of at least half of our initial investment.
  • Shorting cost is approximated as 0.5% per annum. execution
    Source sentence
    We use 0.5% as a proxy for the annual shorting cost of assets

Data

  • CRSP US Stock Database · Wharton Research Data Services (WRDS) market-prices commercial
    2010-01-04 to 2023-12-30 · 15405 assets
    Source sentence
    ... The data set consists of adjusted prices of 15405 assets from January 4th, 2010, to December 30th, 2023, for a total of 3282 trading days.

Recognised sources: CRSP WRDS

Costs

Unclear - sources disagree The platform classifier and the paper-text extraction give different answers, so no value is asserted.

Platform classifier: Modeled

Paper text extraction: Partial - Transaction costs are modeled in simulation (buy at ask, sell at bid) and a 0.5% annual shorting cost is used, but not in the simple linear trading policy.

Source sentence (the authors' words)
Our numerical experiments take into account transaction costs, i.e., we buy assets at the ask price

Finds 3

  • Moving‑band stat‑arbs outperform fixed‑band stat‑arbs and stay profitable longer out‑of‑sample. performance
    Source sentence
    Our empirical study on historical data shows that moving-band stat-arbs perform better than fixed-band stat-arbs, and remain profitable for longer out-of-sample periods.
  • Among 545 fixed‑band stat‑arbs, 63% were profitable with average annualized return 10%, risk 32% and Sharpe 0.81. performance
    Source sentence
    Of the 545 stat-arbs, 63% were profitable. The average annualized return is 10%, with an average annualized risk of 32%, and an average annualized Sharpe ratio of 0.81.
  • The method discovered 712 unique moving‑band stat‑arbs (vs 545 fixed‑band), with median of 5 assets and median 40 active at a time. performance
    Source sentence
    stat-arbs ... These stat-arbs contained between 1 and 10 assets, with a median value of 5 ... The

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 12 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 (62)

  1. Dynamic Conditional Correlation (2002)
  2. Covariance prediction via convex optimization (2022)
  3. The shorting premium and asset pricing anomalies (2014)
  4. Shorting costs and profitability of long–short strategies (2023)
  5. Stocks are special too: An analysis of the equity lending market (2002)
  6. The market for borrowing stock (2002)
  7. A Simple Method for Predicting Covariance Matrices of Financial Returns (2023)
  8. Markowitz Portfolio Construction at Seventy — → 2401.05080
and 54 more
  1. Active portfolio management (2000)
  2. Finding Moving-Band Statistical Arbitrages via Convex-Concave Optimization (2023)
  3. Model predictive control for optimal portfolios with co-integrated pairs of stocks (2012)
  4. Model predictive control for optimal pairs trading portfolio with gross exposure and transaction cost constraints (2018)
  5. Optimal Trading with Cointegrated Pairs of Stocks (2012)
  6. Pairs trading under transaction costs using model predictive control (2018)
  7. Statistical arbitrage and high-frequency data with an application to E urostoxx 50 equities (2010)
  8. Non-linear dependence modelling with bivariate copulas: Statistical arbitrage pairs trading on the S&P 100 (2017)
  9. Statistical arbitrage with vine copulas (2018)
  10. Analytic solutions for optimal statistical arbitrage trading (2010)
  11. Statistical Arbitrage Trading Strategy in Commodity Futures Market with the Use of Nanoseconds Historical Data (2017)
  12. Expectations for statistical arbitrage in energy futures markets (2019)
  13. Statistical arbitrage in cryptocurrency markets (2019)
  14. Large data sets and machine learning: Applications to statistical arbitrage (2019)
  15. Selection of a portfolio of pairs based on co-integration: A statistical arbitrage strategy (2013)
  16. Deep learning statistical arbitrage (2021) — link
  17. Implementing Deep Neural Networks for Financial Market Prediction on the I ntel X eon P hi (2015)
  18. Deep conditional portfolio sorts: The relation between past and future stock returns (2014)
  19. Applying deep learning to enhance momentum trading strategies in stocks (2013)
  20. Pairs trading and outranking: The multi-step-ahead forecasting case (2010)
  21. Pairs selection and outranking: An application to the S&P 100 index (2009)
  22. Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500 (2017)
  23. Testing market efficiency using statistical arbitrage with applications to momentum and value strategies (2004)
  24. Statistical arbitrage: algorithmic trading insights and techniques (2011)
  25. Indexing and statistical arbitrage (2005)
  26. A signal processing perspective on financial engineering (2016)
  27. Evaluation of pairs-trading strategy at the B razilian financial market (2009)
  28. Statistical arbitrage in the US equities market (2010)
  29. A pairs trading strategy based on linear state space models and the K alman filter (2016)
  30. Portfolio Selection (1952)
  31. Price-Bands: A Technical Tool for Stock Trading (2021)
  32. Bollinger on Bollinger bands (2002)
  33. Using bollinger bands (1992)
  34. Wharton Research Data Services (2023)
  35. Refinitiv Financial Data (2023) — link
  36. Most Traded Currencies in the World (2023) — link
  37. The elements of statistical learning: data mining, inference, and prediction (2009)
  38. Dynamic portfolio selection in arbitrage (2007)
  39. Optimal pairs trading: A stochastic control approach (2008)
  40. Pairs trading (2005)
  41. Optimal mean-reverting portfolio with leverage constraint for statistical arbitrage in finance (2019)
  42. Mean-reverting portfolio with budget constraint (2018)
  43. Optimal portfolio design for statistical arbitrage in finance (2018)
  44. Pairs Trading: quantitative methods and analysis (2004)
  45. Co-Integration and Error Correction: Representation, Estimation, and Testing (1987)
  46. Co-integrated variables and error-correcting models (1983)
  47. Modelling of co-integration in the vector autoregressive model (2000)
  48. Statistical arbitrage pairs trading strategies: Review and outlook (2017)
  49. Disciplined convex-concave programming (2016)
  50. Variations and extension of the convex–concave procedure (2016)
  51. Pairs trading: Performance of a relative-value arbitrage rule (2006)
  52. Convex optimization (2004)
  53. A Machine Learning Based Pairs Trading Investment Strategy (2020)
  54. Trading strategies with copulas (2013)

1 of 62 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.