Paper: arXiv 2312.11132

Authors: Adil Rengim Cetingoz, Olivier Guéant

Abstract

Portfolio optimization methods have evolved significantly since Markowitz introduced the mean-variance framework in 1952. While the theoretical appeal of this approach is undeniable, its practical implementation poses important challenges, primarily revolving around the intricate task of estimating expected returns. As a result, practitioners and scholars have explored alternative methods that prioritize risk management and diversification. One such approach is Risk Budgeting, where portfolio risk is allocated among assets according to predefined risk budgets. The effectiveness of Risk Budgeting in achieving true diversification can, however, be questioned, given that asset returns are often influenced by a small number of risk factors. From this perspective, one question arises: is it possible to allocate risk at the factor level using the Risk Budgeting approach? First, we introduce a comprehensive framework to address this question by introducing risk measures directly associated with risk factor exposures and demonstrating the desirable mathematical properties of these risk measures, making them suitable for optimization. Then, we propose a novel framework to find portfolios that effectively balance the risk contributions from both assets and factors. Leveraging standard stochastic algorithms, our framework enables the use of a wide range of risk measures to construct diversified portfolios.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 4.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper introduces advanced mathematical constructs like custom risk measures with proofs of convexity and differentiability, and proposes stochastic optimization algorithms; while it includes two case studies, the summary and excerpt focus on theoretical framework development rather than extensive backtesting data or performance metrics.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Can Risk Budgeting be applied<br>to factor allocations?"] --> B["Methodology<br>Introduce Factor Risk Measures<br>with desirable mathematical properties"]
  B --> C["Input Data<br>Asset returns & Factor exposures"]
  C --> D["Computational Process<br>Stochastic Optimization Algorithm<br>to balance asset & factor risk"]
  D --> E["Outcome<br>Balanced Risk Budgeting Framework<br>for true diversification"]

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 novel Asset‑Factor Risk Budgeting framework that balances risk contributions from both assets and factors.

Source sentence
we propose a novel framework to find portfolios that effectively balance the risk contributions from both assets and factors.

Assumes 6

  • (1, X1, …, Xd) is linearly independent model
    Source sentence
    where (1, X1, . . . , Xd) is assumed to be linearly independent.
  • R is continuous on Rd and continuously differentiable on Rd\{0} model
    Source sentence
    We assume that R is continuous on Rd and continuously differentiable on Rd\{0}.
  • Risk of any non‑zero position is positive (R(y) > 0 for y ≠ 0) model
    Source sentence
    We also assume that ∀y ∈Rd\{0}, R(y) > 0, i.e. the risk of a non-zero financial position is always positive.
  • A 5‑year look‑back window is used for each weekly rebalancing data
    Source sentence
    We consider a 5-year look-back window, which implies that at each rebalancing, we take into account data from the last 5 years.
Show 2 more
  • Weekly rebalancing of portfolios execution
    Source sentence
    we rebalance our portfolio on a weekly basis to obtain the desired portfolios
  • Transaction costs are modeled via bid/ask spreads of 0.02% equities, 0.01% bonds, 0.05% commodities execution
    Source sentence
    We account for transaction costs by factoring in bid/ask spreads of 0.02% for equities, 0.01% for bonds, and 0.05% for commodities.

Data

  • Daily returns (Jan 2015 – Nov 2023) market-prices unknown
    2015-01 to 2023-11
    Source sentence
    Our database consists of daily returns between January 2015 and November 2023.
  • BARRA risk model (equity factor model) macro unknown
    Source sentence
    we construct equity portfolios under the BARRA risk model

Costs

Modeled Costs are part of the reported headline results.

Bid/ask spreads of 0.02% for equities, 0.01% for bonds, 0.05% for commodities; total cost 0.31% in backtest.

Source sentence
We account for transaction costs by factoring in bid/ask spreads of 0.02% for equities, 0.01% for bonds, and 0.05% for commodities.

Finds 5

  • Asset weights are stable and asset risk contributions align with budgets, but Real Rates factor can dominate up to ~70% of risk. performance
    Source sentence
    we observe that the asset weights are stable over time and that the risk contributions of the assets are very well aligned with the given vector of risk budgets
  • Long‑only FRB and AFRB strategies reduce losses by avoiding high exposure to Real Rates and increasing exposure to commodity factor. performance
    Source sentence
    the long-only FRB and AFRB strategies mitigate these losses by avoiding high exposures to this factor and also by providing significant exposure to the commodity factor
  • FRB has high turnover (17.25% avg) leading to 0.31% total transaction cost, while AFRB has low turnover similar to RB (2.43% avg). performance
    Source sentence
    resulting in high transaction costs, which, in total, amounted to 0.31% in our backtest (with 0.02% for the RB and 0.05% for the AFRB).
  • All three strategies show negative annualized returns, driven by poor performance of government bonds and Real Rates after 2021. performance
    Source sentence
    We observe negative annualized returns from the three strategies, primarily due to their high exposure to government bonds asset class and real rates factor, which performed poorly after the end of 2021
  • Table 3 reports sample mean, volatility, Expected Shortfall, max drawdown and turnover for RB, FRB (long‑only) and AFRB. statistical
    Source sentence
    Table 3 demonstrates the main statistics associated with the different strategies computed on daily returns after transaction costs.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 28 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.