Signature Methods in Stochastic Portfolio Theory

ArXiv ID: 2310.02322 “View on arXiv”

Authors: Unknown

Abstract

In the context of stochastic portfolio theory we introduce a novel class of portfolios which we call linear path-functional portfolios. These are portfolios which are determined by certain transformations of linear functions of a collections of feature maps that are non-anticipative path functionals of an underlying semimartingale. As main example for such feature maps we consider the signature of the (ranked) market weights. We prove that these portfolios are universal in the sense that every continuous, possibly path-dependent, portfolio function of the market weights can be uniformly approximated by signature portfolios. We also show that signature portfolios can approximate the growth-optimal portfolio in several classes of non-Markovian market models arbitrarily well and illustrate numerically that the trained signature portfolios are remarkably close to the theoretical growth-optimal portfolios. Besides these universality features, the main numerical advantage lies in the fact that several optimization tasks like maximizing (expected) logarithmic wealth or mean-variance optimization within the class of linear path-functional portfolios reduce to a convex quadratic optimization problem, thus making it computationally highly tractable. We apply our method also to real market data based on several indices. Our results point towards out-performance on the considered out-of-sample data, also in the presence of transaction costs.

Keywords: Stochastic Portfolio Theory, Path Signatures, Growth-Optimal Portfolio, Universal Portfolios, Convex Optimization, Equities

Complexity vs Empirical Score

• Math Complexity: 9.5/10
• Empirical Rigor: 8.5/10
• Quadrant: Holy Grail
• Why: The paper uses advanced mathematics including stochastic calculus, signature methods, and rough path theory, with rigorous theoretical proofs (e.g., universal approximation theorems), while also presenting substantial empirical work with simulated data, real market data, and transaction costs, leading to a highly tractable convex quadratic optimization.

  flowchart TD
    A["Research Goal: <br>Develop universal, <br>computationally tractable <br>portfolio construction method"] --> B["Key Methodology: <br>Linear Path-Functional Portfolios <br>using Signature of Market Weights"]
    B --> C{"Data Inputs"}
    C --> D["Simulated Non-Markovian <br>Market Models"]
    C --> E["Real Market Data: <br>Equity Indices"]
    D & E --> F["Computational Process: <br>Convex Quadratic Optimization <br>Maximize Expected Logarithmic Wealth"]
    F --> G["Key Findings & Outcomes"]
    G --> H["Theoretical: <br>Universality - Approximates <br>any continuous portfolio & Growth-Optimal Portfolio"]
    G --> I["Empirical: <br>Out-performance on Out-of-Sample Data <br>Robust even with Transaction Costs"]