High-dimensional covariance matrix estimation in approximate factor models1

Jianqing Fan, Yuan Liao, Martina Mincheva

Research output: Contribution to journalArticle

98 Citations (Scopus)

Abstract

The variance-covariance matrix plays a central role in the inferential theories of high-dimensional factor models in finance and economics. Popular regularization methods of directly exploiting sparsity are not directly applicable to many financial problems. Classical methods of estimating the covariance matrices are based on the strict factor models, assuming independent idiosyncratic components. This assumption, however, is restrictive in practical applications. By assuming sparse error covariance matrix, we allow the presence of the cross-sectional correlation even after taking out common factors, and it enables us to combine the merits of both methods.We estimate the sparse covariance using the adaptive thresholding technique as in Cai and Liu [J. Amer. Statist. Assoc. 106 (2011) 672-684], taking into account the fact that direct observations of the idiosyncratic components are unavailable. The impact of high dimensionality on the covariance matrix estimation based on the factor structure is then studied.

Original languageEnglish (US)
Pages (from-to)3320-3356
Number of pages37
JournalAnnals of Statistics
Volume39
Issue number6
DOIs
StatePublished - Jan 1 2011

Fingerprint

Covariance Matrix Estimation
Factor Models
Covariance matrix
High-dimensional
Adaptive Thresholding
Factor Structure
Variance-covariance Matrix
Common factor
Regularization Method
Sparsity
Finance
Dimensionality
Economics
Estimate
Covariance matrix estimation
Factors

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Keywords

  • Common factors
  • Cross-sectional correlation
  • Idiosyncratic
  • Seemingly unrelated regression
  • Sparse estimation
  • Thresholding

Cite this

Fan, Jianqing ; Liao, Yuan ; Mincheva, Martina. / High-dimensional covariance matrix estimation in approximate factor models1. In: Annals of Statistics. 2011 ; Vol. 39, No. 6. pp. 3320-3356.
@article{01882c1d13924100b715f1777863de1a,
title = "High-dimensional covariance matrix estimation in approximate factor models1",
abstract = "The variance-covariance matrix plays a central role in the inferential theories of high-dimensional factor models in finance and economics. Popular regularization methods of directly exploiting sparsity are not directly applicable to many financial problems. Classical methods of estimating the covariance matrices are based on the strict factor models, assuming independent idiosyncratic components. This assumption, however, is restrictive in practical applications. By assuming sparse error covariance matrix, we allow the presence of the cross-sectional correlation even after taking out common factors, and it enables us to combine the merits of both methods.We estimate the sparse covariance using the adaptive thresholding technique as in Cai and Liu [J. Amer. Statist. Assoc. 106 (2011) 672-684], taking into account the fact that direct observations of the idiosyncratic components are unavailable. The impact of high dimensionality on the covariance matrix estimation based on the factor structure is then studied.",
keywords = "Common factors, Cross-sectional correlation, Idiosyncratic, Seemingly unrelated regression, Sparse estimation, Thresholding",
author = "Jianqing Fan and Yuan Liao and Martina Mincheva",
year = "2011",
month = "1",
day = "1",
doi = "10.1214/11-AOS944",
language = "English (US)",
volume = "39",
pages = "3320--3356",
journal = "Annals of Statistics",
issn = "0090-5364",
publisher = "Institute of Mathematical Statistics",
number = "6",

}

High-dimensional covariance matrix estimation in approximate factor models1. / Fan, Jianqing; Liao, Yuan; Mincheva, Martina.

In: Annals of Statistics, Vol. 39, No. 6, 01.01.2011, p. 3320-3356.

Research output: Contribution to journalArticle

TY - JOUR

T1 - High-dimensional covariance matrix estimation in approximate factor models1

AU - Fan, Jianqing

AU - Liao, Yuan

AU - Mincheva, Martina

PY - 2011/1/1

Y1 - 2011/1/1

N2 - The variance-covariance matrix plays a central role in the inferential theories of high-dimensional factor models in finance and economics. Popular regularization methods of directly exploiting sparsity are not directly applicable to many financial problems. Classical methods of estimating the covariance matrices are based on the strict factor models, assuming independent idiosyncratic components. This assumption, however, is restrictive in practical applications. By assuming sparse error covariance matrix, we allow the presence of the cross-sectional correlation even after taking out common factors, and it enables us to combine the merits of both methods.We estimate the sparse covariance using the adaptive thresholding technique as in Cai and Liu [J. Amer. Statist. Assoc. 106 (2011) 672-684], taking into account the fact that direct observations of the idiosyncratic components are unavailable. The impact of high dimensionality on the covariance matrix estimation based on the factor structure is then studied.

AB - The variance-covariance matrix plays a central role in the inferential theories of high-dimensional factor models in finance and economics. Popular regularization methods of directly exploiting sparsity are not directly applicable to many financial problems. Classical methods of estimating the covariance matrices are based on the strict factor models, assuming independent idiosyncratic components. This assumption, however, is restrictive in practical applications. By assuming sparse error covariance matrix, we allow the presence of the cross-sectional correlation even after taking out common factors, and it enables us to combine the merits of both methods.We estimate the sparse covariance using the adaptive thresholding technique as in Cai and Liu [J. Amer. Statist. Assoc. 106 (2011) 672-684], taking into account the fact that direct observations of the idiosyncratic components are unavailable. The impact of high dimensionality on the covariance matrix estimation based on the factor structure is then studied.

KW - Common factors

KW - Cross-sectional correlation

KW - Idiosyncratic

KW - Seemingly unrelated regression

KW - Sparse estimation

KW - Thresholding

UR - http://www.scopus.com/inward/record.url?scp=84906280924&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906280924&partnerID=8YFLogxK

U2 - 10.1214/11-AOS944

DO - 10.1214/11-AOS944

M3 - Article

VL - 39

SP - 3320

EP - 3356

JO - Annals of Statistics

JF - Annals of Statistics

SN - 0090-5364

IS - 6

ER -