Paper: arXiv 2409.14510

Authors: Maysam Khodayari Gharanchaei, Reza Babazadeh

Abstract

Forming quantitative portfolios using statistical risk models presents a significant challenge for hedge funds and portfolio managers. This research investigates three distinct statistical risk models to construct quantitative portfolios of 1,000 floating stocks in the US market. Utilizing five different investment strategies, these models are tested across four periods, encompassing the last three major financial crises: The Dot Com Bubble, Global Financial Crisis, and Covid-19 market downturn. Backtests leverage the CRSP dataset from January 1990 through December 2023. The results demonstrate that the proposed models consistently outperformed market excess returns across all periods. These findings suggest that the developed risk models can serve as valuable tools for asset managers, aiding in strategic decision-making and risk management in various economic conditions.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 6.0/10 (capped from 8.0: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 8.0)

Why this score: The paper employs advanced statistical techniques like quadratic optimization, shrinkage methods, and factor risk models, indicating significant mathematical complexity. It demonstrates strong empirical rigor by backtesting on 34 years of real market data (CRSP) across three major crises, evaluating multiple strategies, and presenting performance against benchmarks.

Research Flowchart

  flowchart TD
  A["Research Goal"] -->|"Form Quantitative Portfolios"| B["Statistical Risk Models"]
  B -->|"Apply 5 Strategies"| C["Backtesting Process"]
  D["CRSP Data<br>1990-2023"] --> C
  C -->|"Test in 4 Periods<br>Dot Com, GFC, Covid"| E["Analysis & Comparison"]
  E --> F{"Key Findings"}
  F --> G["Consistent Outperformance"]
  F --> H["Risk Management Tool"]

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 develops statistical risk models for large stock portfolios and applies quadratic optimization and a new shrinkage method to improve risk parity.

Source sentence
Developing statistical risk models using statistical attributes among hundreds of stocks and idiosyncratic characteristics of individual assets; Leveraging quadratic numerical optimization techniques; Utilizing a new shrinkage method to control the covariance risk model's sensitivity; Transforming the risk parity objec

Assumes

No explicit assumptions found in the text.

Data

  • CRSP dataset · CRSP market-prices public
    January 1990 through December 2023 · 1,000 floating US stocks
    Source sentence
    Backtests leverage the CRSP dataset from January 1990 through December 2023.

Recognised sources: CRSP Kenneth French Data Library

Costs

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

Finds 3

  • The proposed models consistently outperformed market excess returns across all tested periods. performance
    Source sentence
    The results demonstrate that the proposed models consistently outperformed market excess returns across all periods.
  • The constant correlation covariance matrix achieved an exceptional return of almost 48% during the Global Financial Crisis. performance
    Source sentence
    An exceptional return of almost 48% during the GFC is proving this.
  • Maximum diversification strategy delivered high returns during market upswings but incurred higher risk, making it less attractive for risk‑averse investors. performance
    Source sentence
    The maximum diversification strategy often delivers high returns during market upswings but comes with significant risk, making it suitable for risk‑tolerant investors.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 27 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.