Pith. sign in

REVIEW 3 major objections 5 minor 46 references

Covariance Matrix Estimation for Positively Correlated Assets

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that rotating the eigenvectors of the sample correlation matrix—rather than only shrinking its eigenvalues—produces a lower-risk covariance estimator for positively correlated assets, cutting out-of-sample…

desk verdict Genuinely new rotation idea and broad empirics, but Algorithm 2 is ambiguous and δ is chosen in-sample, so the headline risk reductions are not yet reproducible. read the letter →

arxiv 2507.01545 v1 pith:CSZCOSOE submitted 2025-07-02 econ.EM

Weilong Liu , Yanchu Liu This is my paper
classification econ.EM MSC 62H1262H2591G10
keywords covariancematrixestimationshrinkageestimatorpositivelycorrelatedassetsglobalminimumvarianceportfolioeigenvectorrotationrotation-equivariantfactor-sortedportfoliosconditionnumber
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

The paper claims that when assets are positively correlated, the eigenvectors of the sample correlation matrix corresponding to weak factors sit very close to the null space of the uniform vector, and that pushing these eigenvectors away from that null space—while preserving orthogonality—yields a better covariance matrix for minimum-variance portfolios. The proposed estimator, ERSE, pairs the eigenvector with the smallest deviation with the one with the largest deviation and rotates them until every eigenvector satisfies a minimum deviation threshold. Each rotation is exactly a linear shrinkage of two eigenvalues toward each other while preserving their sum, so ERSE behaves as a data-driven, multi-pair shrinkage method. On ten factor-sorted portfolio datasets, ERSE reports the lowest out-of-sample variance in every dataset, with average risk reductions of 10.52% versus the best linear shrinkage benchmark and 12.46% versus quadratic-inverse nonlinear shrinkage. If these results hold, ERSE gives practitioners a computationally cheap covariance estimator for decision settings where comovement makes most pairwise correlations positive.

What carries the argument

The central object is the Paired Eigenvector Rotation (PER). PER takes two orthogonal unit eigenvectors $q_1,q_2$ with $T(q_1)<\delta<T(q_2)$, rotates them in their plane via $\hat{q}_1=\cos\theta\,q_1+\sin\theta\,q_2$ and $\hat{q}_2=-\sin\theta\,q_1+\cos\theta\,q_2$, and chooses the smallest rotation angle making $T(\hat{q}_1)=\delta$. This preserves orthogonality and the sum of the two deviation degrees, and it transforms the paired eigenvalues by $\hat\lambda_1=\cos^2\theta\,\lambda_1+\sin^2\theta\,\lambda_2$ and $\hat\lambda_2=\sin^2\theta\,\lambda_1+\cos^2\theta\,\lambda_2$. ERSE iterates PER at most $n-1$ times, always pairing the current smallest-deviation eigenvector with the largest-deviation eigenvector, then rebuilds eigenvalues as $\hat\lambda_i=\hat{q}_i^T R_S\hat{q}_i$ and the covariance estimate as $D_S\hat{Q}\,\text{Diag}(\hat\lambda)\,\hat{Q}^T D_S$. The machinery converts an eigenvector alignment constraint into an eigenvalue-shrinkage scheme without directly modifying the sample covariance or correlation matrix.

What would settle it

Re-run Algorithm 2 but keep the full rotated matrix $\hat{Q}^T R_S\hat{Q}$ instead of keeping only its diagonal $\hat\lambda_i$; if the resulting GMV portfolios have out-of-sample variance similar to or lower than ERSE's, then the claimed improvement is not attributable to the eigenvalue shrinkage that the paper identifies as the mechanism.

Watch

Extended reading notes

Core claim

The discovery is a structural fact about positively correlated assets plus an estimator built on it. For such assets, the sample correlation matrix has a dominant eigenvector whose deviation degree $T(q)=(\mathbf{1}_n^T q)^2$ is larger than its largest eigenvalue, while the remaining eigenvectors are squeezed toward the null space of the uniform vector; Propositions 1–3 show that the deviation degrees sum to $n$ and that weak-factor eigenvectors have deviation bounded near zero when average correlation is high. The paper argues that moderately amplifying these weak-factor deviations is the right shrinkage direction, and it constructs ERSE to do this. Algorithm 2 pairs the eigenvector with the smallest deviation against the one with the largest deviation and applies a Paired Eigenvector Rotation until every eigenvector satisfies $T(\hat{q}_i)\ge\delta$. Because one rotation sends two eigenvalues to $\cos^2\theta\,\lambda_a+\sin^2\theta\,\lambda_b$ and $\sin^2\theta\,\lambda_a+\cos^2\theta\,\lambda_b$, ERSE is equivalent to repeated linear shrinkage on pairs of eigenvalues, pulling small eigenvalues up and large ones down while preserving their sum. The empirical claim is that plugging ERSE into the global minimum-variance portfolio achieves the lowest out-of-sample variance on all ten datasets, with statistically significant gains in most comparisons and average reductions of 10.52% versus LINC and 12.46% versus QIS.

Load-bearing premise

The paper silently assumes that discarding the off-diagonal entries of the rotated correlation matrix in the eigenvalue reconstruction step does not materially distort the estimate; it never analyzes this truncation, so ERSE's gains could partly reflect an unintended distortion rather than the intended shrinkage.

Editorial extensions

If this is right

  • GMV portfolios built from ERSE should have lower out-of-sample variance than portfolios built from the compared linear and nonlinear shrinkage estimators whenever the sample correlation matrix is positive, with average reductions near 10.5% and 12.5% on factor-sorted portfolios.
  • ERSE covariance estimates have lower condition numbers than the benchmark estimates, which translates into more concentrated, more stable portfolio weights and modest short positions in positively correlated settings.
  • The out-of-sample risk reduction is stable: it appears in both subperiods, across estimation windows from 60 to 240 months, and in all 150 random 200-asset subsamples drawn from the largest dataset.
  • ERSE has low computational cost, requiring at most $n-1$ closed-form rotations per estimation window, making it feasible for portfolios with hundreds of assets.

Reading between the lines

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

  • Our inference: because the positive-correlation assumption is the only structural input, ERSE should transfer to other positively correlated universes such as government-bond portfolios or country-industry indices; the paper does not test these.
  • Our inference: the threshold $\delta=0.25$ is selected using the same datasets that are later used for evaluation, so the headline reductions likely include an in-sample tuning component; a held-out calibration would be needed to confirm the out-of-sample edge.
  • Our inference: Algorithm 2 drops the off-diagonal entries of $\hat{Q}^T R_S\hat{Q}$ after rotation, and no analysis quantifies the resulting information loss; keeping those entries or shrinking them separately is a natural modification that could change the risk comparison.
  • Our inference: since each PER pair is exact two-eigenvalue linear shrinkage, the same rotation geometry could produce adaptive shrinkage by letting $\delta$ vary across eigenvectors or by targeting a different aggregate rather than a uniform floor.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes the Eigenvector Rotation Shrinkage Estimator (ERSE), a rotation-equivariant covariance matrix estimator designed for positively correlated assets. The method begins from the sample correlation matrix, imposes a minimum-deviation constraint on each eigenvector from the null space of the uniform vector, and enforces this constraint through paired eigenvector rotations (PER). The paper proves several properties of the deviation measure and of single PER steps, and then evaluates ERSE in a global minimum-variance (GMV) portfolio framework on ten factor-sorted Ken-French datasets. The headline empirical claim is that ERSE achieves the lowest out-of-sample variance across all datasets, with average risk reductions of 10.52% relative to the best linear shrinkage method (LINC) and 12.46% relative to quadratic-inverse shrinkage (QIS), together with lower condition numbers and more stable portfolio weights.

Significance. If the central empirical claim survives scrutiny, ERSE is a useful addition to the covariance-estimation toolbox for high-correlation settings. The paper deserves credit for explicitly targeting the positive-correlation case, for the Perron-Frobenius-based analysis of eigenvector misalignment (Propositions 1-3), and for a broad robustness campaign covering subperiods, estimation windows, and 150 random subsamples. The algebraic proofs of Propositions 1-6 are self-contained. The significance of the result, however, rests on two load-bearing points that are not yet settled: the exact estimator implemented in Algorithm 2, and the use of the same datasets to select the hyperparameter δ that drives the headline performance numbers. The theoretical connection to sequential linear shrinkage also needs strengthening for iterations after the first rotation.

major comments (3)
  1. [Section 4, Algorithm 2, Steps 8-9] The reconstruction in Algorithm 2 is not uniquely defined. Step 8 computes λ̂_i = q̂_i^T R_S q̂_i using the rotated eigenvectors, while Step 9 sets Σ̂ = D_S Q_S Diag(λ̂) Q_S^T D_S with Q_S denoting the original sample eigenvector matrix. If Q_S is the original matrix, the final estimator contains only the original eigenvectors, and the eigenvector rotation enters only through the scalar λ̂_i; the procedure is then an eigenvalue-shrinkage rule, and the paper's 'eigenvector rotation' description is misleading. If Q_S is meant to be the rotated matrix Q̂, then the estimator keeps only the diagonal of Q̂^T R_S Q̂ and discards the off-diagonal entries; for a single paired rotation those entries equal cosθ sinθ(λ_j − λ_i), which are generally nonzero, and Proposition 5 provides no analysis of this truncation. Because the out-of-sample results in Table 2 depend on which implementation is used, the authors must state explicitly which matrix is used in Step 9, make the algorithm consistent with that statement, and ideally report results under both readings if the ambiguity persists.
  2. [Sections 6.1 and 6.3] The key hyperparameter δ is selected on the same data used for the headline evaluation. Section 6.1 states that δ = 0.25 is 'empirically determined', and Section 6.3 recommends the interval [0.15, 0.35] based on out-of-sample variance curves for the same ten datasets. Under this protocol, the reported average risk reductions of 10.52% versus LINC and 12.46% versus QIS are not purely out-of-sample, since the parameter was chosen by inspecting the test results. The authors should either pre-specify δ before seeing the out-of-sample data, select it via a nested procedure within the estimation window, or demonstrate that the rankings in Table 2 are insensitive to δ over a wide interval containing 0.25 for every dataset. The evidence in Figure 4 is suggestive but still relies on the same evaluation period.
  3. [Section 5, Proposition 5] The claimed equivalence of ERSE to sequential linear shrinkage on pairs of eigenvalues is established only for the first PER step. Proposition 5 assumes that q1 and q2 are eigenvectors of R_S, so that cross terms in q̂^T R_S q̂ vanish. After the first iteration of Algorithm 2 the rotated vectors are no longer eigenvectors of R_S; for such vectors q̂1^T R_S q̂2 = cosθ sinθ(λ_j − λ_i) is generally nonzero, and the updated λ̂_i is not a convex combination of two current eigenvalues as written in Eq. (38). The statement that ERSE is 'equivalent to performing multiple linear shrinkage on two distinct eigenvalues' therefore needs either a generalized proof for the iterated procedure or a careful restatement limiting the equivalence to the initial pair.
minor comments (5)
  1. [Section 4, Algorithm 2, line 8] The vector λ̂ is written as (λ̂_1, . . . , n)^T; it should be (λ̂_1, . . . , λ̂_n)^T.
  2. [Tables 2, 4, 5, and 6] The asterisk convention is never defined. The tables should state explicitly that asterisks indicate the significance level of the difference between ERSE and the listed strategy, and should name the test (e.g., the stationary bootstrap of Ledoit and Wolf 2011) and the direction of the alternative.
  3. [Section 3, Proposition 3 proof] The Lagrangian derivation has sign inconsistencies between Eqs. (21) and (24): stationarity gives x = −(1_n + λρ_i)/(2ν), while the text writes x = (1_n − λρ_i)/(2ν). The final bound in Eq. (14) is correct, but the intermediate algebra should be corrected for clarity.
  4. [Figure 3] The caption does not explain what the plotted points and axes represent. Please add a short description of the two-dimensional and three-dimensional illustrations so the reader can interpret the rotation geometrically.
  5. [References] The citation 'Vaart, 1961' appears with the author name 'Vaart, H. R.'; if this refers to H. R. van der Vaart, the spelling should be standardized.

Circularity Check

1 steps flagged · score 2.0 of 10

Theoretical derivation is self-contained algebra and benchmarks are external, but the ERSE hyperparameter δ = 0.25 is 'empirically determined' on the same datasets used for the out-of-sample evaluation, so the headline 10.52%/12.46% risk reductions are partly in-sample-selected.

  1. fitted input called prediction [Section 6.1 (δ = 0.25 choice), Section 6.2 (Table 2 headline claims), Section 6.3 (post-hoc robustness defense)]
    "In this study, we set δ = 0.25, which is empirically determined and has proven effective in most cases. ... The ERSE method achieves an average risk reduction of 10.52% relative to LINC across all 10 datasets"

    ERSE's hyperparameter δ controls the rotation intensity and therefore the estimated eigenvalues and the resulting GMV risk. The paper says δ = 0.25 was 'empirically determined' and does not state that the selection used data disjoint from the ten factor-sorted datasets on which Table 2's out-of-sample evaluation runs. The headline reductions (10.52% vs LINC, 12.46% vs QIS) and the 'lowest variance across all data sets' claim are reported at this single fitted value, so the reported out-of-sample gains are partly in-sample-selected.

full rationale

The theoretical derivation chain (Sections 3-5) is self-contained. Proposition 1 is the identity Tr(R_S) = n combined with 1^T Q Q^T 1 = n; Proposition 2 uses the classical Perron-Frobenius theorem and elementary inequalities. The load-bearing Proposition 5, that a PER step is equivalent to linear shrinkage λ̂1 = cos²θ λ1 + sin²θ λ2, follows directly from eigenvector orthogonality (q1^T R_S q2 = 0); it is algebra, not an input-output reduction. Proposition 6 restates the definition of T and the ℓ2 norm. There are no self-citations: the reference list contains no work by the authors, and no uniqueness theorem is imported from prior author work. The empirical benchmarks (LINC, QIS, GIS, LIS, LIN1P, etc.) are external standard methods implemented via Ledoit's public covShrinkage repository with default parameters, so the comparisons are against independent baselines and the out-of-sample results are falsifiable. The only selection-on-evaluation concern is the hyperparameter δ = 0.25, which the paper says was 'empirically determined' on what appears to be the same ten factor-sorted portfolios used for the out-of-sample evaluation, making the specific reported reductions partly in-sample-tuned. The Section 6.3 sweep shows the qualitative outperformance holds for a wide range of δ, which prevents the central claim from reducing to the fit; however, the recommended interval [0.15, 0.35] is itself derived from the evaluation data. Separately, a correctness risk rather than circularity: Algorithm 2's Step 9 is ambiguous about whether Q_S is the original or the rotated eigenvector matrix; under the rotated reading, nonzero off-diagonal entries cosθ sinθ(λ_j − λ_i) of Q̂^T R_S Q̂ are silently discarded, and Proposition 5 covers only the diagonal entries. This should be resolved to pin down the exact estimator behind Table 2, but it does not constitute a circular reduction. Overall verdict: no structural circularity; the single mild selection issue yields a score of 2.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central estimator relies on one fitted threshold (δ), a domain restriction that all sample correlations are positive, and the rotation-equivariant form. The theoretical motivation for rotating weak-factor eigenvectors relies on a bound (Prop. 3, Eq. 15) whose proof has an inequality-direction gap. No new physical or economic entities are introduced.

free parameters (1)
  • δ (minimum deviation threshold) = 0.25
    Controls how far weak-factor eigenvectors are rotated toward the uniform vector. The authors state in Section 6.2 that δ=0.25 'is empirically determined and has proven effective in most cases,' meaning it was chosen using the same datasets that are later used for evaluation.
assumptions (5)
  • domain assumption Returns follow a multivariate normal distribution
    Section 3 opens with 'We assume that the return vector for these n assets follows a multivariate normal distribution with the covariance matrix Σ.' This underlies the sample covariance and correlation setup but is not used in the empirical robustness checks.
  • domain assumption Assumption 1: all elements of the sample correlation matrix are positive
    Section 2 states 'we assume that R_S is a positive matrix, with each sample correlation coefficient ρ_i,j > 0.' This holds for the factor-sorted portfolios studied but restricts the scope of the method.
  • domain assumption Assumption 2: estimator is of the rotation-equivariant form Σ̂ = D_S Q_S Diag(λ̂) Q_S^T D_S
    Section 3, equation (3). The paper restricts attention to estimators that share the sample eigenvectors or a rotation of them.
  • standard math Perron-Frobenius theorem for irreducible positive matrices
    Used in Proposition 2 to assert that the largest eigenvector of a positive correlation matrix is strictly positive and simple.
  • ad hoc to paper Bound in Proposition 3, Eq. (15): (1^T ρ_i)^2 / (ρ_i^T ρ_i) ≥ (1+(n-1)b)^2 / (1+(n-1)b^2)
    The derivation from (14) to (15) appears to use ρ_i^T ρ_i ≤ 1+(n-1)b^2, which has the wrong direction unless additional structure is assumed. The bound may be true, but the proof as written is incomplete. This motivates the claim that weak-factor eigenvectors are near the null space.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Covariance Matrix Estimation for Positively Correlated Assets." pith.science (2026). https://pith.science/paper/CSZCOSOE

@misc{pith2026250701545,
  author       = {Pith},
  title        = {Pith review of: Covariance Matrix Estimation for Positively Correlated Assets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CSZCOSOE}},
  note         = {Machine review of arXiv:2507.01545}
}
read the original abstract

The comovement phenomenon in financial markets creates decision scenarios with positively correlated asset returns. This paper addresses covariance matrix estimation under such conditions, motivated by observations of significant positive correlations in factor-sorted portfolio monthly returns. We demonstrate that fine-tuning eigenvectors linked to weak factors within rotation-equivariant frameworks produces well-conditioned covariance matrix estimates. Our Eigenvector Rotation Shrinkage Estimator (ERSE) pairwise rotates eigenvectors while preserving orthogonality, equivalent to performing multiple linear shrinkage on two distinct eigenvalues. Empirical results on factor-sorted portfolios from the Ken French data library demonstrate that ERSE outperforms existing rotation-equivariant estimators in reducing out-of-sample portfolio variance, achieving average risk reductions of 10.52\% versus linear shrinkage methods and 12.46\% versus nonlinear shrinkage methods. Further checks indicate that ERSE yields covariance matrices with lower condition numbers, produces more concentrated and stable portfolio weights, and provides consistent improvements across different subperiods and estimation windows.

Figures

Figures reproduced from arXiv: 2507.01545 by the authors.

Figure 1
Figure 1. Dynamic Evolution of the Average and Minimum Corre [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Values of λ S n and T (q S n ) in the Ten Factor-Sorted Portfolio Datasets a numerical comparison of λ S n and T(q S n ) using the ten factor-sorted portfolio datasets tested in Section 2. The dominance of λ S n and T(q S n ) in positively correlated assets is clearly demonstrated in [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Visualization of the Paired Eigenvector Rotation [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Out-of-Sample Variance of Strategies across Vary [PITH_FULL_IMAGE:figures/full_fig_p025_4.png]
Figure 5
Figure 5. Figure 5: Distribution of Portfolio Weights 6.5. Sharpe Ratio The tradeoff between risk and return remains a central consideration in portfolio management. The Sharpe ratio, calculated as the ratio of mean excess returns to their standard deviation, provides insight into how eff…
Figure 6
Figure 6. Figure 6: Portfolio risks with Different Estimation Windows [PITH_FULL_IMAGE:figures/full_fig_p030_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

46 extracted references · 45 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    M., Distaso, W., and Žikeš, F

    Abadir, K. M., Distaso, W., and Žikeš, F. (2014). Design-free estimation of variance matrices. Journal of Econometrics , 181(2):165--180

    work page 2014

  3. [3]

    Barberis, N., Shleifer, A., and Wurgler, J. (2005). Comovement. Journal of Financial Economics , 75(2):283--317

    work page 2005

  4. [4]

    and Saxena, K

    Barroso, P. and Saxena, K. (2022). Lest we forget: Learn from out-of-sample forecast errors when optimizing portfolios. The Review of Financial Studies , 35(3):1222--1278

    work page 2022

  5. [5]

    M., Griffith, T

    Blau, B. M., Griffith, T. G., and Whitby, R. J. (2023). Industry regulation and the comovement of stock returns. Journal of Empirical Finance , 73:206--219

    work page 2023

  6. [6]

    Bodnar, T., Mazur, S., and Okhrin, Y. (2017). Bayesian estimation of the global minimum variance portfolio. European Journal of Operational Research , 256(1):292--307

    work page 2017

  7. [7]

    Bodnar, T., Parolya, N., and Schmid, W. (2018). Estimation of the global minimum variance portfolio in high dimensions. European Journal of Operational Research , 266(1):371--390

    work page 2018

  8. [8]

    T., Hu, J., Li, Y., and Zheng, X

    Cai, T. T., Hu, J., Li, Y., and Zheng, X. (2020). High-dimensional minimum variance portfolio estimation based on high-frequency data. Journal of Econometrics , 214(2):482--494

    work page 2020

Show all 46 references
  1. [9]

    Conlon, T., Cotter, J., and Kynigakis, I. (2025). Asset allocation with factor-based covariance matrices. European Journal of Operational Research , 325(1):189--203

    2025

  2. [10]

    Cui, L., Hong, Y., Li, Y., and Wang, J. (2024). A regularized high-dimensional positive definite covariance estimator with high-frequency data. Management Science , 70(10):7242--7264

    2024

  3. [11]

    Dai, C., Lu, K., and Xiu, D. (2019). Knowing factors or factor loadings, or neither? Evaluating estimators of large covariance matrices with noisy and asynchronous data. Journal of Econometrics , 208(1):43--79

    2019

  4. [12]

    Dai, R., Uematsu, Y., and Matsuda, Y. (2024). Estimation of large covariance matrices with mixed factor structures. The Econometrics Journal , 27(1):62--83

    2024

  5. [13]

    J., and Uppal, R

    DeMiguel, V., Garlappi, L., Nogales, F. J., and Uppal, R. (2009a). A generalized approach to portfolio optimization: Improving performance by constraining portfolio norms. Management Science , 55(5):798--812

    2009

  6. [14]

    DeMiguel, V., Garlappi, L., and Uppal, R. (2009b). Optimal versus naive diversification: How inefficient is the 1/N portfolio strategy? The Review of Financial Studies , 22(5):1915--1953

    2009

  7. [15]

    Ding, Y., Li, Y., and Zheng, X. (2021). High dimensional minimum variance portfolio estimation under statistical factor models. Journal of Econometrics , 222(1, Part B):502--515

    2021

  8. [16]

    Fan, J., Fan, Y., and Lv, J. (2008). High dimensional covariance matrix estimation using a factor model. Journal of Econometrics , 147(1):186--197

    2008

  9. [17]

    Fan, J., Li, Y., and Yu, K. (2012). Vast volatility matrix estimation using high-frequency data for portfolio selection. Journal of the American Statistical Association , 107(497):412--428

    2012

  10. [18]

    Fan, J., Liao, Y., and Mincheva, M. (2013). Large covariance estimation by thresholding principal orthogonal complements. Journal of the Royal Statistical Society: Series B (Statistical Methodology) , 75(4):711--730

    2013

  11. [19]

    Fan, J., Liu, H., and Wang, W. (2018). Large covariance estimation through elliptical factor models . The Annals of Statistics , 46(4):1383 -- 1414

    2018

  12. [20]

    and Xu, Y

    Goto, S. and Xu, Y. (2015). Improving mean variance optimization through sparse hedging restrictions. Journal of Financial and Quantitative Analysis , 50(6):1415--1441

    2015

  13. [21]

    Green, T. C. and Hwang, B.-H. (2009). Price-based return comovement. Journal of Financial Economics , 93(1):37--50

    2009

  14. [22]

    and Fan, J

    Kim, D. and Fan, J. (2019). Factor GARCH-Itô models for high-frequency data with application to large volatility matrix prediction. Journal of Econometrics , 208(2):395--417

    2019

  15. [23]

    Lam, C. (2016). Nonparametric eigenvalue-regularized precision or covariance matrix estimator. The Annals of Statistics , 44(3):928--953, 26

    2016

  16. [24]

    Ledoit, O. (1995). Essays on Risk and Return in the Stock Market . PhD thesis, Massachusetts Institute of Technology, Sloan School of Management, Cambridge, MA, USA

    1995

  17. [25]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2003). Improved estimation of the covariance matrix of stock returns with an application to portfolio selection. Journal of Empirical Finance , 10(5):603--621

    2003

  18. [26]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2004a). Honey, I shrunk the sample covariance matrix. Journal of Portfolio Management , 30(4):110--119

    2004

  19. [27]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2004b). A well-conditioned estimator for large-dimensional covariance matrices. Journal of Multivariate Analysis , 88(2):365--411

    2004

  20. [28]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2008). Robust performance hypothesis testing with the sharpe ratio. Journal of Empirical Finance , 15(5):850--859

    2008

  21. [29]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2011). Robust performances hypothesis testing with the variance. Wilmott , 2011(55):86--89

    2011

  22. [30]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2012). Nonlinear shrinkage estimation of large-dimensional covariance matrices. The Annals of Statistics , 40(2):1024--1060

    2012

  23. [31]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2017). Nonlinear shrinkage of the covariance matrix for portfolio selection: Markowitz meets goldilocks. The Review of Financial Studies , 30(12):4349--4388

    2017

  24. [32]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2020). Analytical nonlinear shrinkage of large-dimensional covariance matrices. Annals of Statistics , 48:3043--3065

    2020

  25. [33]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2022a). The power of (non-)linear shrinking: A review and guide to covariance matrix estimation. Journal of Financial Econometrics , 20(1):187--218

    2022

  26. [34]

    and Wolf, M

    Ledoit, O. and Wolf, M. (2022b). Quadratic shrinkage for large covariance matrices. Bernoulli , 28(3):1519--1547

    2022

  27. [35]

    Malceniece, L., Malcenieks, K., and Putniņš, T. J. (2019). High frequency trading and comovement in financial markets. Journal of Financial Economics , 134(2):381--399

    2019

  28. [36]

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

    1952

  29. [37]

    M\"orstedt, T., Lutz, B., and Neumann, D. (2024). Cross validation based transfer learning for cross-sectional non-linear shrinkage: A data-driven approach in portfolio optimization. European Journal of Operational Research , 318(2):670--685

    2024

  30. [38]

    A., Kuhn, D., and Esfahani, P

    Nguyen, V. A., Kuhn, D., and Esfahani, P. M. (2022). Distributionally robust inverse covariance estimation: The Wasserstein shrinkage estimator. Operations Research , 70(1):490--515

    2022

  31. [39]

    and Wang, Q

    Pirinsky, C. and Wang, Q. (2006). Does corporate headquarters location matter for stock returns? The Journal of Finance , 61(4):1991--2015

    2006

  32. [40]

    Reh, L., Krüger, F., and Liesenfeld, R. (2023). Predicting the global minimum variance portfolio. Journal of Business and Economic Statistics , 41(2):440--452

    2023

  33. [41]

    Shi, F., Shu, L., Yang, A., and He, F. (2020). Improving minimum-variance portfolios by alleviating overdispersion of eigenvalues. Journal of Financial and Quantitative Analysis , 55(8):2700--2731

    2020

  34. [42]

    Stein, C. (1986). Lectures on the theory of estimation of many parameters. Journal of Soviet Mathematics , 34(1):1373--1403

    1986

  35. [43]

    Touloumis, A. (2015). Nonparametric stein-type shrinkage covariance matrix estimators in high-dimensional settings. Computational Statistics and Data Analysis , 83:251--261

    2015

  36. [44]

    Vaart, H. R. (1961). On certain characteristics of the distribution of the latent roots of a symmetric random matrix under general conditions. Annals of Mathematical Statistics , 32:864--873

    1961

  37. [45]

    Wang, H., Peng, B., Li, D., and Leng, C. (2021). Nonparametric estimation of large covariance matrices with conditional sparsity. Journal of Econometrics , 223(1):53--72

    2021

  38. [46]

    Won, J.-H., Lim, J., Kim, S.-J., and Rajaratnam, B. (2013). Condition-number-regularized covariance estimation. Journal of the Royal Statistical Society Series B: Statistical Methodology , 75(3):427--450

    2013

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.