Nonparametric Estimation of Large Covariance Matrices with Conditional Sparsity

Hanchao Wang; Bin Peng; Degui Li; Chenlei Leng

doi:10.1016/j.jeconom.2020.09.002

Nonparametric Estimation of Large Covariance Matrices with Conditional Sparsity

Hanchao Wang, Bin Peng, Degui Li, Chenlei Leng

Mathematics

Research output: Contribution to journal › Article › peer-review

Abstract

This paper studies estimation of covariance matrices with conditional sparse structure. We overcome the challenge of estimating dense matrices using a factor structure, the challenge of estimating large-dimensional matrices by postulating sparsity on covariance of random noises, and the challenge of estimating varying matrices by allowing factor loadings to smoothly change. A kernel-weighted estimation approach combined with generalised shrinkage is proposed. Under some technical conditions, we derive uniform consistency for the developed estimation method and obtain convergence rates. Numerical studies including simulation and an empirical application are presented to examine the finite-sample performance of the developed methodology.

Original language	English
Pages (from-to)	53-72
Number of pages	20
Journal	Journal of Econometrics
Volume	223
Issue number	1
Early online date	23 Oct 2020
DOIs	https://doi.org/10.1016/j.jeconom.2020.09.002
Publication status	Published - Jul 2021

Bibliographical note

Access to Document

10.1016/j.jeconom.2020.09.002

WPLL-JoE-26-Sep-2020
© 2020 Elsevier B.V. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy.
Accepted author manuscript, 320 KBLicence: CC BY-NC-ND

Modelling Large Spot Volatility Structure for High Frequency Data: New Methodology and Practice
Li, D.
THE BRITISH ACADEMY
1/07/20 → 30/08/22
Project: Research project (funded) › Research

Cite this

@article{100e202142854134a2468023887950fa,

title = "Nonparametric Estimation of Large Covariance Matrices with Conditional Sparsity",

abstract = "This paper studies estimation of covariance matrices with conditional sparse structure. We overcome the challenge of estimating dense matrices using a factor structure, the challenge of estimating large-dimensional matrices by postulating sparsity on covariance of random noises, and the challenge of estimating varying matrices by allowing factor loadings to smoothly change. A kernel-weighted estimation approach combined with generalised shrinkage is proposed. Under some technical conditions, we derive uniform consistency for the developed estimation method and obtain convergence rates. Numerical studies including simulation and an empirical application are presented to examine the finite-sample performance of the developed methodology.",

author = "Hanchao Wang and Bin Peng and Degui Li and Chenlei Leng",

note = "{\textcopyright} 2020 Elsevier B.V. This is an author-produced version of the published paper. Uploaded in accordance with the publisher{\textquoteright}s self-archiving policy.",

year = "2021",

month = jul,

doi = "10.1016/j.jeconom.2020.09.002",

language = "English",

volume = "223",

pages = "53--72",

journal = "Journal of Econometrics",

issn = "0304-4076",

publisher = "Elsevier",

number = "1",

}

TY - JOUR

T1 - Nonparametric Estimation of Large Covariance Matrices with Conditional Sparsity

AU - Wang, Hanchao

AU - Peng, Bin

AU - Li, Degui

AU - Leng, Chenlei

PY - 2021/7

Y1 - 2021/7

N2 - This paper studies estimation of covariance matrices with conditional sparse structure. We overcome the challenge of estimating dense matrices using a factor structure, the challenge of estimating large-dimensional matrices by postulating sparsity on covariance of random noises, and the challenge of estimating varying matrices by allowing factor loadings to smoothly change. A kernel-weighted estimation approach combined with generalised shrinkage is proposed. Under some technical conditions, we derive uniform consistency for the developed estimation method and obtain convergence rates. Numerical studies including simulation and an empirical application are presented to examine the finite-sample performance of the developed methodology.

AB - This paper studies estimation of covariance matrices with conditional sparse structure. We overcome the challenge of estimating dense matrices using a factor structure, the challenge of estimating large-dimensional matrices by postulating sparsity on covariance of random noises, and the challenge of estimating varying matrices by allowing factor loadings to smoothly change. A kernel-weighted estimation approach combined with generalised shrinkage is proposed. Under some technical conditions, we derive uniform consistency for the developed estimation method and obtain convergence rates. Numerical studies including simulation and an empirical application are presented to examine the finite-sample performance of the developed methodology.

U2 - 10.1016/j.jeconom.2020.09.002

DO - 10.1016/j.jeconom.2020.09.002

M3 - Article

SN - 0304-4076

VL - 223

SP - 53

EP - 72

JO - Journal of Econometrics

JF - Journal of Econometrics

IS - 1

ER -

Nonparametric Estimation of Large Covariance Matrices with Conditional Sparsity

Abstract

Bibliographical note

Access to Document

Projects

Modelling Large Spot Volatility Structure for High Frequency Data: New Methodology and Practice

Cite this