Paper: arXiv 2502.11701

Authors: Hyunglip Bae, Haeun Jeon, Minsu Park, Yongjae Lee, Woo Chang Kim

Abstract

In practice, including large number of assets in mean-variance portfolios can lead to higher transaction costs and management fees. To address this, one common approach is to select a smaller subset of assets from the larger pool, constructing more efficient portfolios. As a solution, we propose a new asset selection heuristic which generates a pre-defined list of asset candidates using a surrogate formulation and re-optimizes the cardinality-constrained tangent portfolio with these selected assets. This method enables faster optimization and effectively constructs portfolios with fewer assets, as demonstrated by numerical analyses on historical stock returns. Finally, we discuss a quantitative metric that can provide a initial assessment of the performance of the proposed heuristic based on asset covariance.

Complexity vs Empirical Score

  • Math Complexity: 7.0/10
  • Empirical Rigor: 6.0/10 (capped from 6.5: transaction costs not modelled; no out-of-sample test; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 6.5)

Why this score: The paper introduces a novel heuristic (OSCAR) with detailed mathematical derivations, including Cholesky decomposition and angle-based optimization, leading to a high math score. It validates the method with numerical experiments on historical stock returns, providing implementation and performance analysis, though it lacks full backtest-ready details like transaction costs.

Research Flowchart

  flowchart TD
  A["Research Goal: Sparse Tangent Portfolio<br>Minimize transaction costs by selecting assets"] --> B{"Methodology"}
  B --> C["Data: Historical Stock Returns<br>Covariance Matrix Σ"]
  C --> D["Cholesky Decomposition-based<br>Heuristic: Surrogate Formulation"]
  D --> E["Select k Assets from Candidate List"]
  E --> F["Re-optimize Cardinality-constrained<br>Tangent Portfolio"]
  F --> G{"Numerical Analysis"}
  G --> H["Findings: Faster Optimization<br>& Effective Asset Reduction"]
  H --> I["Outcome: Quantitative Metric<br>for Heuristic Assessment"]

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 OSCAR, a new asset selection heuristic that generates a candidate list via a surrogate formulation and re‑optimizes the cardinality‑constrained tangent portfolio with the selected assets.

Source sentence
we propose a new asset selection heuristic which generates a pre-defined list of asset candidates using a surrogate formulation and re-optimizes the cardinality-constrained tangent portfolio with these selected assets.

Assumes 6

  • The expected return vector µ and covariance matrix Σ of asset returns are known. model
    Source sentence
    given the expected return vector and covariance matrix of returns (µ, Σ) ∈(Rn, Rn×n).
  • The covariance matrix Σ is positive‑definite so that a Cholesky decomposition exists. model
    Source sentence
    Decompose Σ = LTΣLΣ by Cholesky decomposition.
  • The Sharpe ratio is scale‑invariant with respect to portfolio weights. statistical
    Source sentence
    For any λ > 0, the Sharpe ratio is scale-invariant with respect to w, i.e. SR(w|µ, Σ) = SR(λw|µ, Σ)
  • Cardinality constraint k is a fixed integer specifying the maximum number of assets to hold. model
    Source sentence
    where k ∈{1, 2, . . . , n} is the number of assets that one can or want to hold
Show 2 more
  • Asset selection can be approximated by choosing the k largest absolute components of |LTΣ ˆw|. model
    Source sentence
    of |LTΣ ˆw| where | · | is element‑wise absolute value operator.
  • Historical daily closing prices from Yahoo Finance represent the true return process for the assets. data
    Source sentence
    For our experiment, we use six 10‑year daily closing price datasets from January 2014 to December 2023, obtained from Yahoo Finance.

Data

Recognised sources: Yahoo Finance

Costs

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

Finds 2

  • OSCAR achieves higher efficacy and efficiency than other heuristics in numerical experiments. performance
    Source sentence
    Our numerical experiments demonstrate that OSCAR performs with higher efficacy and efficiency compared to other heuristics.
  • CPLEX often fails to find optimal solutions within the one‑day time limit for larger datasets and higher cardinalities. negative
    Source sentence
    Notably, the CPLEX was unable to derive an optimal solution within a day solving for the S&P100 dataset (97 assets) with k = 0.10×N

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 24 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 (25)

  1. Portfolio Selection (1952)
  2. A uniformly distributed random portfolio (2016)
  3. Sparse tangent portfolio selection via semi-definite relaxation (2016)
  4. Computational study of a family of mixed-integer quadratic programming problems (1996)
  5. Optimal lot solution to cardinality constrained mean–variance formulation for portfolio selection (2006)
  6. Sparse and stable Markowitz portfolios (2009)
  7. Sparse portfolio selection via quasi-norm regularization (2013) — link
  8. Optimization of cardinality constrained portfolios with a hybrid local search algorithm (2003)
and 17 more
  1. Optimal cardinality constrained portfolio selection (2013)
  2. Index tracking with cardinality constraints: A stochastic neural networks approach (2020)
  3. A new method for mean-variance portfolio optimization with cardinality constraints (2013)
  4. Mean–variance efficient large portfolios: A simple machine learning heuristic technique based on the two-fund separation theorem (2022)
  5. Heuristics for cardinality constrained portfolio optimisation (2000)
  6. Algorithm for cardinality-constrained quadratic optimization (2009)
  7. Computational aspects of alternative portfolio selection models in the presence of discrete asset choice constraints (2001)
  8. A generalized approach to portfolio optimization: Improving performance by constraining portfolio norms (2009)
  9. Portfolio optimization with groupwise selection (2014)
  10. Mutual fund performance (1966)
  11. Computers and intractability (1979)
  12. Portfolio optimisation with transaction cost (2011)
  13. Sparse portfolio selection via the sorted L1-Norm (2020)
  14. Adaptive l 1-regularization for short-selling control in portfolio selection (2019)
  15. Mean-variance optimization for asset allocation (2021)
  16. Sparse and robust portfolio selection via semi-definite relaxation (2020)
  17. Index Tracking with Cardinality Constraints: A Stochastic Neural Networks Approach (2019) — link

0 of 25 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.