Paper: arXiv 2510.05377

Authors: Bibhas Adhikari

Abstract

In this work, we consider weighted signed network representations of financial markets derived from raw or denoised correlation matrices, and examine how negative edges can be exploited to reduce portfolio risk. We then propose a discrete optimization scheme that reduces the asset selection problem to a desired size by building a time series of signed networks based on asset returns. To benchmark our approach, we consider two standard allocation strategies: Markowitz’s mean-variance optimization and the 1/N equally weighted portfolio. Both methods are applied on the reduced universe as well as on the full universe, using two datasets: (i) the Market Champions dataset, consisting of 21 major S&P500 companies over the 2020-2024 period, and (ii) a dataset of 199 assets comprising all S&P500 constituents with stock prices available and aligned with Google’s data. Empirical results show that portfolios constructed via our signed network selection perform as good as those from classical Markowitz model and the equal-weight benchmark in most occasions.

Complexity vs Empirical Score

  • Math Complexity: 7.0/10
  • Empirical Rigor: 5.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper involves moderately advanced mathematical concepts like weighted signed graphs, structural balance theory, and correlation matrix estimations, resulting in a high math complexity. It also includes empirical backtesting on two real-world datasets and compares against standard benchmarks, showing substantial implementation and data handling, placing it in the high rigor category.

Research Flowchart

  flowchart TD
  A["Research Goal: Exploit Negative Edges<br>in Financial Networks to Reduce Portfolio Risk"] --> B["Methodology: Build Signed Networks from<br>Correlation Matrices of Asset Returns"]
  
  B --> C{"Data Inputs"}
  C --> D1["Market Champions Dataset<br>21 S&P500 Companies 2020-2024"]
  C --> D2["S&P500 Constituents Dataset<br>199 Assets Aligned with Google Data"]
  
  D1 & D2 --> E["Discrete Optimization Scheme<br>Asset Selection via Signed Networks"]
  D1 & D2 --> F["Benchmarks<br>Markowitz Mean-Variance & 1/N Equal Weight"]
  
  E & F --> G["Computational Process<br>Apply Strategies on Full & Reduced Universes"]
  
  G --> H["Key Findings<br>Signed Network portfolios perform as well as<br>Markowitz and Equal-Weight benchmarks in most cases"]

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 discrete optimization scheme that reduces the asset selection problem by building a time series of signed networks and then applying standard allocation methods on the reduced universe.

Source sentence
We then propose a discrete optimization scheme that reduces the asset selection problem to a desired size by building a time series of signed networks based on asset returns.

Assumes 5

  • Asset dependencies are captured by raw or denoised correlation matrices. model
    Source sentence
    In this work, we consider weighted signed network representations of financial markets derived from raw or denoised correlation matrices
  • Negative edges (negative correlations) can be used to identify hedge‑protected assets. model
    Source sentence
    examine how negative edges can be exploited to reduce portfolio risk
  • A rolling window of length T is used to compute sample means for each asset. data
    Source sentence
    by comparing each pair’s returns to their own sample means over a rolling window of length T
  • Backtesting is performed using historical daily prices without accounting for trading frictions. execution
    Source sentence
    We employ the backtesting method for analyzing the performance of the proposed method
Show 1 more
  • Short selling may be allowed or disallowed in the Markowitz optimization variants (MP and MPNS). model
    Source sentence
    We determine the optimized portfolio allocation vector solving the Markowitz’s with short selling and no short selling optimization problems

Data

  • Market Champions dataset · Kaggle market-prices public
    2020-2024 · 21 major S&P500 companies
    Source sentence
    The first is the Market Champions dataset from Kaggle , which contains daily stock prices for 21 prominent S&P 500 companies
  • 199‑asset S&P500 dataset · Kaggle market-prices public
    199 S&P 500 constituents
    Source sentence
    The second dataset is from Kaggle and consists of 199 S&P 500 stocks with price data aligned with Google’s dataset

Recognised sources: Kaggle

Costs

Not modeled A trading or execution result is reported without costs.

Finds 2

  • Portfolios constructed via the proposed signed‑network selection perform comparably to classical Markowitz and equal‑weight benchmarks on most occasions. performance
    Source sentence
    Empirical results show that portfolios constructed via our signed network selection perform as good as those from the classical Markowitz model and the equal-weight benchmark in most occasions.
  • For the Market Champions dataset, the hedge‑score based reduction reproduces the same portfolio as the naive Markowitz no‑short‑selling solution (MPNS). theoretical
    Source sentence
    the values of all the statistics obtained for Markowitz’s no short selling method (MPNS) match with the corresponding statistics obtained by employing our proposed method combining hedge scores of the assets and the MPNS

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 11 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive (1 by independent authors): Local and Global Balance in Financial Correlation Networks: … (uses its method or data)

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.