Pith. sign in

REVIEW 2 major objections 4 minor 47 references

Mirror Descent Algorithms for Risk Budgeting Portfolios

T0 review · 2 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Tamed mirror descent gives provable convergence for risk budgeting portfolios.

desk verdict The deterministic MD analysis is genuinely solid, but the advertised stochastic convergence theorem rests on a tamed-gradient monotonicity lemma that is false as stated, so the paper needs repair before its main claim can stand. read the letter →

arxiv 2411.12323 v1 pith:NCU2P7JL submitted 2024-11-19 q-fin.PM math.PRq-fin.RM

Martin Arnaiz Iglesias (UP1 UFR27) , Adil Rengim Cetingoz (UP1 UFR27) , Noufel Frikha (UP1 UFR27) This is my paper · ORCID
classification q-fin.PMmath.PRq-fin.RM MSC 65C0562L2062G3291Gxx
keywords RiskbudgetingMirrordescentmeasuresExpectedShortfallStochasticapproximationTamedgradientDeviationPortfoliooptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Risk budgeting portfolios — long-only portfolios whose asset contributions to total risk match preset budgets — are usually found by solving a system of nonlinear equations that is equivalent to minimizing a strictly convex function with a logarithmic barrier. That barrier makes the gradient explode at the boundary of the positive orthant, so standard mirror descent theory does not apply. This paper shows that multiplying the gradient by the taming factor $\kappa(y)=\min_i y_i\wedge 1$ removes the singularity without moving the minimizer, and proves that deterministic and stochastic mirror descent then converge to the risk budgeting weights, with an explicit non-asymptotic rate for the averaged iterates. The stochastic version covers risk measures expressible as $g(r_\rho(y))=\min_\xi \mathbb{E}[L(\xi,-\langle y,X\rangle)]$, including Expected Shortfall and generalized deviation measures, and the numerical section shows it stays stable and often more accurate than projected SGD as the number of assets grows. If the theorems hold, this is the first scheme to give almost-sure convergence with a quantitative rate for risk budgeting across general risk measures.

What carries the argument

The load-bearing object is the tamed gradient $T(y)=\kappa(y)\nabla\Gamma_g(y)$ with $\kappa(y)=\min_i y_i\wedge 1$, extended by continuity to the boundary by setting components to $-b_i$ where the coordinate is zero. It does two jobs: it stays uniformly bounded on every centred ball, which makes the standard mirror-descent Bregman inequality applicable, and it preserves the strict monotonicity $\langle y-y^\star,T(y)\rangle>0$ for $y\neq y^\star$, which is what forces any limit point to be $y^\star$. Around it, the algorithm uses the negative-entropy Bregman divergence $D_F(y,y')=\sum_i y_i\log(y_i/y'_i)-\sum_i y_i+\sum_i y'_i$ and its explicit proximal map $P^m_y(v)$, the exponential update $y_i e^{-v_i}$ rescaled to the $\ell^1$-ball of radius $m$. In the stochastic case the state is $z=(\xi,y)$, the divergence gains a quadratic $\xi$-term, and the update is given by (3.10), with the same taming factor applied to the $y$-gradient of $H(z,X)=L(\xi,-\langle y,X\rangle)-\sum_i b_i\log y_i$.

What would settle it

Take the three-asset volatility objective $g(r_\rho(y))=y^\top\Sigma y$ with a nonzero mean-return vector and equal budgets, and compute $\langle y-y^\star, \kappa(y)\nabla\Gamma_g(y)\rangle$ at a boundary point $y$ with one coordinate zero; if any such inner product is negative, Lemma 3.2 is false and the proof's exclusion of boundary limit points collapses. Running the SMD recursion (3.10) on the same example with $\gamma_n=n^{-0.55}$ and $m=100$ would then directly show whether the sequence still converges.

Watch

Extended reading notes

Core claim

The paper's central claim is that the unique minimizer $y^\star$ of $\Gamma_g(y)=g(r_\rho(y))-\sum_{i=1}^d b_i\log y_i$ can be computed by deterministic and stochastic Mirror Descent even though $\nabla\Gamma_g$ blows up at the boundary of the positive orthant. The resolution is the tamed gradient $\kappa(y)\nabla\Gamma_g(y)$ with $\kappa(y)=\min_i y_i\wedge 1$, which extends continuously to the closed domain, vanishes only at $y^\star$, and satisfies the strict monotonicity inequality used to rule out boundary limit points. With this tamed gradient and the explicit KL-proximal update, the deterministic scheme converges in the limit and the weighted averages satisfy the rate bound (3.5); the stochastic scheme, built on the representation $g(r_\rho(y))=\min_\xi\mathbb{E}[L(\xi,-\langle y,X\rangle)]$, converges almost surely to $z^\star=(\xi^\star,y^\star)$ and its weighted averages satisfy the a.s. bound (3.12). The normalized limit $u^\star=y^\star/\|y^\star\|_1$ is the unique risk budgeting portfolio. The theorem applies directly to Expected Shortfall and deviation measures with $p=1$; for volatility and variantiles the stated square-integrability condition (3.11) is violated, and the paper asserts that the proof can be adapted.

Load-bearing premise

Everything depends on the tamed gradient pushing away from the boundary at every boundary point, and on the noisy gradients having finite second moments; volatility and variantile examples do not satisfy the second condition as stated, so the paper asserts that the proof can be adapted.

Editorial extensions

If this is right

  • Risk budgeting portfolios for any risk measure of the form $\rho(Z)=\min_\xi \mathbb{E}[L(\xi,Z)]$ can be computed by an explicit, projection-free mirror descent recursion that converges almost surely and comes with a non-asymptotic averaged-iterate rate.
  • The log-barrier singularity no longer blocks stochastic approximation: the factor $\kappa(y)=\min_i y_i\wedge 1$ keeps the update inside the positive orthant while preserving the unique minimizer.
  • For Expected Shortfall and generalized deviation measures the theoretical guarantees apply directly; numerical tests indicate the SMD method stays stable and beats projected SGD as the number of assets grows.
  • The user must know or guess a radius $m\ge\|y^\star\|_1$; choosing $m$ too small provably prevents convergence, while too large $m$ only degrades the explicit rate.
  • The weighted-average estimate $\bar z_n$ satisfies an almost-sure bound of order $(\sum_{k=0}^n\gamma_{k+1})^{-1}$ up to the inverse of the smallest coordinate, which is where the boundary behavior enters.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The taming factor $\min_i y_i\wedge 1$ is a general device for convex programs with logarithmic barriers and explicit exponential proximal maps; it could be tested on other entropy-regularized allocation problems outside risk budgeting.
  • Because the numerics show constant step size converging faster than the provably convergent decreasing-step choices, a natural next conjecture is that simple averaging of later iterates gives the same almost-sure rate under weaker moment assumptions.
  • If the asserted adaptation to volatility and variantiles is written out fully, the same machinery should yield a central limit theorem for the averaged sequence, an extension the paper already flags as future work.
  • The rate bound's dependence on $m\log m$ and dimension $d$, together with the identity $r_\rho(y^\star)=1/g'(r_\rho(y^\star))$, suggests a practical rule for choosing $m$ from the risk level rather than by blind search.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper proposes deterministic and stochastic mirror descent algorithms for computing risk budgeting portfolios under positive homogeneous and sub-additive risk measures, building on the variational characterization of risk budgeting portfolios in [15]. The algorithms are based on a tamed gradient kappa(y) grad Gamma_g(y) with kappa(y) = min_i y_i ^ 1, which is bounded on the domain and intended to vanish only at the minimizer. The deterministic MD scheme is proved to converge with a non-asymptotic rate on the averaged sequence (Theorem 3.1), and a stochastic MD scheme is claimed to converge almost surely with a similar rate (Theorem 3.2) under condition (3.11). Numerical experiments compare the methods with projected SGD for Expected Shortfall, volatility, and deviation measures across portfolio sizes up to 250 assets.

Significance. If the stochastic convergence theorem were correct, the paper would make a useful contribution: it would provide the first a.s. convergence guarantee with a quantitative rate for risk budgeting portfolios across general risk measures, going beyond the unanalyzed projected SGD of [15]. The deterministic convergence proof is coherent, the explicit proximal updates in (3.4) and (3.10) are a practical strength, and the numerical study is extensive, including stability comparisons with classical and tamed SGD. However, the central stochastic claim rests on a false monotonicity lemma, and one of the paper's headline examples, volatility, does not satisfy the square-integrability condition required by Theorem 3.2. The advertised main result is therefore not established as it stands.

major comments (2)
  1. [Appendix A.3, Step 1] Lemma 3.2 is false as stated, and Theorem 3.2 relies on it in an essential way. The proof is omitted with the explanation that the result follows from convexity of h, but the tamed expected gradient is not grad h(z): the y-components are multiplied by the common scalar kappa(y) = min_i y_i ^ 1. Convexity gives <z - z*, grad h(z)> > 0, which does not imply the displayed inequality for the tamed vector. A concrete counterexample is the two-asset volatility model with X ~ N((10,0), I_2), b = (1,1), g(x) = x^2, and L(xi,x) = (x - xi)^2. Here z* = (xi*, y*) with xi* = -10/sqrt(2) and y* = (1/sqrt(2), 1/sqrt(2)). Take y = (3/2, 1/sqrt(2)), which is interior and lies in B_m for m = 2.5 >= ||y*||_1, and set kappa = 1/sqrt(2), a = (1-kappa)*10*(3/2 - 1/sqrt(2))/2, and xi = a - 15. A direct computation gives <E[(d_xi H, kappa grad_y H)](z), z - z*> approximately -1.4 < 0. This is not a boundary artifact. Inequality (A.8) in Appendix A.3 requires this inner product to be nonnegative for the Robbins-Siegmund argument, and Step 1 of the proof uses the liminf of this inner product to conclude that z_infinity = z*. A negative value invalidates both the a.s. convergence claim and the rate bound (3.12).
  2. [Section 3.3.3] Theorem 3.2 does not cover two of the paper's headline examples, and the asserted adaptation is not provided. The text explicitly states that condition (3.11) is not met for volatility (g(x) = x^2) and for deviation measures with p > 1, including variantiles, and then says that the proof of Theorem 3.2 can be readily adapted, without giving details. This is a load-bearing gap: the proof in Appendix A.3 uses (3.11) to control E[Y_{k+1} | F_k] <= N_*^2 and to prove sup_n E[M_n^2] < infinity. For the quadratic and polynomial losses, the squared derivatives grow in xi, so those controls fail. Since the abstract and introduction advertise volatility and variantiles among the covered risk measures, the paper must either prove a theorem that covers these cases or substantially restrict the claims.
minor comments (4)
  1. [Notations] In the Notations paragraph, 'weletlet' should read 'we let'.
  2. [Remark 5] In Remark 5, 'upper-boudn' should be 'upper bound'.
  3. [Section 3.3.2] In the sentence introducing the proximal mapping, 'proxymal' should be 'proximal'.
  4. [Section 4.2.2, footnote 3] The footnote references 'Theorem 3.5', but no Theorem 3.5 appears in the paper; the intended reference is presumably Theorem 3.2.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity: the convergence analysis is self-contained given the externally published characterization in [15], and the main gaps (omitted proof of Lemma 3.2, unverified condition (3.11)) are correctness risks rather than circular reductions.

full rationale

The derivation chain is not circular. The optimization target y* is imported from Theorem 2.1, quoted from [15, Theorems 1 and 2], a published prior characterization rather than a restatement of the present paper's algorithmic result; the paper does not claim to prove that equivalence, it uses it as a premise. The DMD and SMD convergence proofs then proceed from explicit Bregman-proximal inequalities (A.1)-(A.13), the Robbins-Siegmund lemma, and the tamed-gradient construction kappa(y) = min_i y_i ^ 1, whose zero-set identity is derived from the first-order conditions of h and Gamma_g rather than assumed as the conclusion. The rate bounds (3.5) and (3.12) are explicit a.s. upper bounds with constants M* and N*, not fitted to data, and they depend on stated assumptions such as (3.11). The numerical benchmark uses L-BFGS-B with semi-analytic VaR/ES expressions from [15] to build reference portfolios, but this is an independent optimizer on the same objective, not a prediction forced by the algorithm's own outputs. The paper does contain self-citations to [15], including a co-authored prior work, and these are load-bearing in the sense that the target formulation comes from there; however, the cited result is externally published and not equivalent to the present convergence claim. Two genuine rigor issues are flagged: Lemma 3.2 is asserted to follow from convexity of h without a proof, and condition (3.11) is conceded to fail for volatility and variantiles with only a claim that the proof can be adapted. These are correctness and completeness concerns, not circularity: no equation in the paper reduces a predicted quantity to an input by construction, and no fitted parameter is renamed as a prediction. Therefore the circularity score is 0.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central convergence analysis is largely self-contained, but it inherits the characterization of risk budgeting via Gamma_g from the authors' own prior work [15], and it assumes a stochastic representation and a square-integrability condition that fail for two of the numerical examples as stated. The algorithmic parameter m is selected without knowledge of the true optimizer.

free parameters (3)
  • m (proximal domain radius) = 100 in the main experiments; must satisfy m >= ||y*||_1
    Chosen by hand. The convergence theorems hold only when m bounds the unknown norm of the optimizer; Figure 4 shows convergence fails for m < ||y*||_1.
  • step-size schedule (gamma_n) = gamma_n = n^{-0.55} for DMD, n^{-0.75} for SMD; gamma_n = 1 in some DMD runs
    Hand selected. The theory only requires sum gamma_n = infinity and sum gamma_n^2 < infinity, but the observed convergence speed depends strongly on the choice.
  • Initial point y0 = y0_i = 1/(d sigma_i^2), with sigma_i^2 the variance of asset i under the first Student-t component
    Hand chosen; affects the initialization error bound D_F(y*, y0) and finite-time behavior.
assumptions (7)
  • domain assumption Theorem 2.1 from Cetingoz, Fermanian and Gueant [15]: for RB-compatible risk measures, the unique risk budgeting portfolio is the normalized minimizer of Gamma_g(y) = g(r_rho(y)) - sum_i b_i log y_i
    This is the foundation of the algorithm; it is cited but not proved here, and one present author is a co-author of [15].
  • domain assumption r_rho is continuous on R_+^d, continuously differentiable on the positive orthant, positive, positive homogeneous and sub-additive
    Needed for Euler's decomposition (2.1), for strict convexity of Gamma_g, and for the gradient formula.
  • domain assumption The stochastic representation (3.6) with a unique minimizer xi*(y) for every y
    Used to define the stochastic problem (3.7) and Proposition 3.1; it holds for ES and deviation measures only under continuity and increasing cdf conditions.
  • domain assumption Uniform square-integrability condition (3.11)
    Assumed in Theorem 3.2; it is violated by volatility and p>1 deviation measures, where only an asserted adaptation is given.
  • standard math The prox mapping is explicit and the Bregman inequality (A.1) holds
    Taken from Nemirovski et al. [35]; used in the proof of Theorems 3.1 and 3.2.
  • ad hoc to paper The tamed gradient vanishes only at the minimizer y*
    Lemma 3.1 proves this in the deterministic case; the stochastic analogue Lemma 3.2 is stated for boundary points but is false there in general, so the proof needs an interior-limit argument.
  • ad hoc to paper User-selected domain radius m satisfies m >= ||y*||_1
    Assumed in Theorems 3.1 and 3.2; y* is unknown, so this is a blind choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mirror Descent Algorithms for Risk Budgeting Portfolios." pith.science (2026). https://pith.science/paper/NCU2P7JL

@misc{pith2026241112323,
  author       = {Pith},
  title        = {Pith review of: Mirror Descent Algorithms for Risk Budgeting Portfolios},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NCU2P7JL}},
  note         = {Machine review of arXiv:2411.12323}
}
read the original abstract

This paper introduces and examines numerical approximation schemes for computing risk budgeting portfolios associated to positive homogeneous and sub-additive risk measures. We employ Mirror Descent algorithms to determine the optimal risk budgeting weights in both deterministic and stochastic settings, establishing convergence along with an explicit non-asymptotic quantitative rate for the averaged algorithm. A comprehensive numerical analysis follows, illustrating our theoretical findings across various risk measures -- including standard deviation, Expected Shortfall, deviation measures, and Variantiles -- and comparing the performance with that of the standard stochastic gradient descent method recently proposed in the literature.

Figures

Figures reproduced from arXiv: 2411.12323 by the authors.

Figure 1
Figure 1. Evolution of the three components of (u k )k≥0 of the DMD algorithm for different step sequence (γn)n≥1. Left: γn ≡ 1. Center: γn = n −0.55. Right: γn = n −0.75. Dashed lines are the asset weights of the reference portfolio in all subplots. We now run the SMD algorithm with N = 250, 000 iterations, comprising 25,000 samples and 10 epochs. We initialize (y 0 , ξ0 ) and set m as before, with γn = n −0.55 1While implem… view at source ↗
Figure 2
Figure 2. A rapid convergence to the weights and VaR of the reference portfolio is achieved, with the convergence of (ξ k )k≥0 appearing faster than that of (u k )k≥0 [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. Convergence of the SMD algorithm. Left: evolution of the three compo￾nents of (y k )k≥0. Center: evolution of the three components of (u k )k≥0’s — dashed lines are the asset weights of the reference portfolio. Right: evolution of (ξ k )k≥0 — dashed line is the VaR of the reference portfolio. We proceed with our analysis by assessing the robustness of both MD algorithms with respect to the dimension d, corresponding… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: From left to right: evolution of the sequence (u k )k≥0 for m = 10, 35, 100 and 1000 respectively. the mean deviation error (MDE) in asset weights compared to those of the reference portfolio, defined as MDE = 1 d X d i=1 |u N i − u ⋆ i |. For the DMD algorithm, we use…
Figure 5
Figure 5. Figure 5: Errors in the objective function values for the unnormalized portfolio weights obtained via c-SGD, calculated across 100 samples for each portfolio size. The error Γ(y k ) − Γ(y ∗ ) is measured at various iterations k = 3 × 105 , 6 × 105 , and 9 × 105 to illustrate the…
Figure 6
Figure 6. Figure 6: The evolution of error for the SMD and t-SGD algorithms across itera￾tions. An initial and illustrative assessment can be conducted by tracking the evolution of error throughout the iterations. Specifically, we can simulate data from a fixed model and run the SMD and t…
Figure 7
Figure 7. Figure 7: offers further insights into the algorithms’ performance. The distribu￾tion of the error in the objective function value, Γ(y k ) − Γ(y ⋆ ), is shown for the SMD and t-SGD algorithms, computed over 100 samples across various portfolio sizes: d = 10, 25, 50, 100 and 250…
Figure 8
Figure 8. Figure 8: Distribution of the MDEs obtained by the SMD and t-SGD algorithms, computed over 100 samples for each portfolio size and iteration n. 4.3 Convergence using different risk measures Previous numerical analyses have focused on a specific risk measure – ES – to have a comp…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 45 canonical work pages

  1. [15]

    A. R. Cetingoz, J.-D. Fermanian, and O. Guéant. Risk budgeting portfolios: Existence and computation.Mathematical Finance, 2023

    work page 2023

  2. [1]

    C. Acerbi. Spectral measures of risk: A coherent representation of subjective risk aversion.Journal of Banking & Finance, 26(7):1505–1518, 2002

    work page 2002

  3. [2]

    Acerbi and D

    C. Acerbi and D. Tasche. On the coherence of Expected Shortfall.Journal of Banking & Finance, 26(7):1487–1503, 2002

    work page 2002

  4. [3]

    Ararat, F

    Ç. Ararat, F. Cesarone, M. Çelebi Pınar, and J. M. Ricci. Mad risk parity portfolios. Annals of Operations Research, pages 1–26, 2024

    work page 2024

  5. [4]

    Asimit, L

    V. Asimit, L. Peng, R. Tunaru, and F. Zhou. Risk budgeting under general risk measures. Submitted

    work page

  6. [5]

    Y. F. Atchadé, G. Fort, and R. Moulines. On perturbed proximal gradient algorithms. Journal of Machine Learning Research, 18(10):1–33, 2017

    work page 2017

  7. [6]

    X. Bai, K. Scheinberg, and R. Tutuncu. Least-squares approach to risk parity in portfolio selection.Quantitative Finance, 16(3):357–376, 2016

    work page 2016

  8. [7]

    Bardou, N

    O. Bardou, N. Frikha, and G. Pagès. Recursive computation of value-at-risk and conditional value-at-risk using mc and qmc. In Pierre L’Ecuyer and Art B. Owen, editors,Monte Carlo and Quasi-Monte Carlo Methods 2008, pages 193– 208, Berlin, Heidelberg, 2009. Springer Berlin Heidelberg

    work page 2008

Show all 47 references
  1. [8]

    Bardou, N

    O. Bardou, N. Frikha, and G. Pagès. CVaR hedging using quantization-based stochastic approximation algorithm. Mathematical Finance, 26(1):184–229, 2016

    2016

  2. [9]

    Bardou, N

    O. Bardou, N. Frikha, and G. Pagès. Computing var and cvar using stochastic approximation and adaptive unconstrained importance sampling.Monte Carlo Methods and Applications, 15(3):173–210, 2009

    2009

  3. [10]

    Bellini, F

    F. Bellini, F. Cesarone, C. Colombo, and F. Tardella. Risk parity with expec- tiles. European journal of operational research, 291(3):1149–1163, 2021

    2021

  4. [11]

    Ben-Tal and M

    A. Ben-Tal and M. Teboulle. An old-new concept of convex risk measures: the optimized certainty equivalent.Mathematical Finance, 17(3):449–476, 2007

    2007

  5. [12]

    M. J. Best and R. R. Grauer. On the sensitivity of mean-variance-efficient portfoliostochangesinassetmeans: someanalyticalandcomputationalresults. The Review of Financial Studies, 4(2):315–342, 1991

    1991

  6. [13]

    S. Boyd, K. Johansson, R. Kahn, P. Schiele, and T. Schmelzer. Markowitz portfolio construction at seventy.arXiv preprint arXiv:2401.05080, 2024

    2024 arXiv

  7. [14]

    Bruder and T

    B. Bruder and T. Roncalli. Managing Risk Exposures Using the Risk Budgeting Approach. Available at SSRN: https://ssrn.com/abstract=2009778, 2012. 26

    2012

  8. [16]

    A. R. Cetingoz and O. Guéant. Factor risk budgeting and beyond. arXiv preprint arXiv:2312.11132, 2023

    2023 arXiv

  9. [17]

    Chaves, J

    D. Chaves, J. Hsu, F. Li, and O. Shakernia. Efficient algorithms for computing risk parity portfolio weights.Journal of Investing, 21(3):150, 2012

    2012

  10. [18]

    Costa, L

    M. Costa, L. Huang, and S. Gadat. Cv@r penalized portfolio optimization with biased stochastic mirror descent.forthcoming in Finance & Stochastics, 2024

    2024

  11. [19]

    Crépey, N

    S. Crépey, N. Frikha, and A. Louzi. A Multilevel Stochastic Approximation Algorithm for Value-at-Risk and Expected Shortfall Estimation.forthcoming for Finance & Stochastics, 2024

    2024

  12. [20]

    Crépey, N

    S. Crépey, N. Frikha, A. Louzi, and G. Pagès. Asymptotic Error Analysis of Multilevel Stochastic Approximations for the Value-at-Risk and Expected Shortfall. forthcoming for Electronic Journal of Probability, 2024

    2024

  13. [21]

    Crépey, N

    S. Crépey, N. Frikha, A. Louzi, and J. Spence. Adaptive Multilevel Stochastic Approximation of the Value-at-Risk.arXiv:2408.06531, 2024

    2024 arXiv

  14. [22]

    B. F. P. da Costa, S. M. Pesenti, and R. S. Targino. Risk budgeting portfolios from simulations.European Journal of Operational Research, 311(3):1040–1056, 2023

    2023

  15. [23]

    Embrechts, T

    P. Embrechts, T. Mao, Q. Wang, and R. Wang. Bayes risk, elicitability, and the expected shortfall.Mathematical Finance, 31(4):1190–1217, 2021

    2021

  16. [24]

    N. Frikha. Shortfall risk minimization in discrete time financial market models. SIAM Journal on Financial Mathematics, 5(1):384–414, 2014

    2014

  17. [25]

    N. Frikha. Multi-level stochastic approximation algorithms. The Annals of Applied Probability, 26(2):933 – 985, 2016

    2016

  18. [26]

    Frikha and L

    N. Frikha and L. Huang. A multi-step richardson–romberg extrapolation method for stochastic approximation. Stochastic Processes and their Appli- cations, 125(11):4066–4101, 2015

    2015

  19. [27]

    Föllmer and A

    H. Föllmer and A. Schied. Convex risk measures. John Wiley & Sons, Ltd, 2010

    2010

  20. [28]

    Griveau-Billion, J.-C

    T. Griveau-Billion, J.-C. Richard, and T. Roncalli. A fast algorithm for com- puting high-dimensional risk parity portfolios. 2013

    2013

  21. [29]

    G. Lan, A. Nemirovski, and A. Shapiro. Validation analysis of mirror descent stochasticapproximationmethod. Mathematical Programming, 134(2):425–458, 2012

    2012

  22. [30]

    Maillard, T

    S. Maillard, T. Roncalli, and J. Teiletche. The properties of equally weighted riskcontributionportfolios. The Journal of Portfolio Management, 36(4):60–70, 2010

    2010

  23. [31]

    Markowitz

    H. Markowitz. Portfolio selection.The Journal of Finance, 7(1):77–91, 1952. 27

    1952

  24. [32]

    Markowitz

    H. Markowitz. The optimization of a quadratic function subject to linear con- straints. Naval Research Logistics Quarterly, 3(1-2):111–133, 1956

    1956

  25. [33]

    Mausser and O

    H. Mausser and O. Romanko. Long-only equal risk contribution portfolios for cvar under discrete distributions.Quantitative Finance, 18(11):1927–1945, 2018

    1927

  26. [34]

    R. O. Michaud. The markowitz optimization enigma: Is ‘optimized’optimal? Financial Analysts Journal, 45(1):31–42, 1989

    1989

  27. [35]

    Nemirovski, A

    A. Nemirovski, A. Juditsky, G. Lan, and A. Shapiro. Robust stochastic approx- imation approach to stochastic programming.SIAM Journal on Optimization, 19(4):1574–1609, 2009

    2009

  28. [36]

    A. S. Nemirovskij and D. B. Yudin.Problem complexity and method efficiency in optimization. John Wiley and Sons, 1983

    1983

  29. [37]

    Nesterov

    Y. Nesterov. Primal-dual subgradient methods for convex problems.Mathe- matical Programming, 120(1):221–259, 2009

    2009

  30. [38]

    Qian et al

    E. Qian et al. Risk parity portfolios: Efficient portfolios through true diversi- fication. Panagora Asset Management, 2005

    2005

  31. [39]

    R. T. Rockafellar and S. Uryasev. Optimization of Conditional Value-at-Risk. Journal of risk, 2(3):21–41, 2000

    2000

  32. [40]

    R. T. Rockafellar and S. Uryasev. Conditional value-at-risk for general loss distributions. Journal of banking & finance, 26(7):1443–1471, 2002

    2002

  33. [41]

    R. T. Rockafellar and S. Uryasev. The fundamental risk quadrangle in risk management, optimization and statistical estimation. Surveys in Operations Research and Management Science, 18(1):33–53, 2013

    2013

  34. [42]

    R. T. Rockafellar, S. Uryasev, and M. Zabarankin. Generalized deviations in risk analysis. Finance Stochast., 10:51–74, 2006

    2006

  35. [43]

    R. T. Rockafellar, S. Uryasev, and M. Zabarankin. Risk tuning with generalized linear regression. Mathematics of Operations Research, 33:712–729, 2008

    2008

  36. [44]

    Roncalli

    T. Roncalli. Introduction to Risk Parity and Budgeting. CRC Press, 2013

    2013

  37. [45]

    F. Spinu. An algorithm for computing risk parity weights. Technical report, 2013

    2013

  38. [46]

    P. Wolfe. The simplex method for quadratic programming. Econometrica: Journal of the Econometric Society, pages 382–398, 1959

    1959

  39. [47]

    Z. Zhou, P. Mertikopoulos, N. Bambos, S. Boyd, and P. W. Glynn. Stochastic mirror descent in variationally coherent optimization problems. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors,Advances in Neural Information Pro...

    2017

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.