Inferences in panel data with interactive effects using large covariance matrices

Jushan Bai, Yuan Liao

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We consider efficient estimation of panel data models with interactive effects, which relies on a high-dimensional inverse covariance matrix estimator. By using a consistent estimator of the error covariance matrix, we can take into account both cross-sectional correlations and heteroskedasticity. In the presence of cross-sectional correlations, the proposed estimator eliminates the cross-sectional correlation bias, and is more efficient than the existing methods. The rate of convergence is also improved. In addition, we find that when the statistical inference involves estimating a high-dimensional inverse covariance matrix, the minimax convergence rate on large covariance estimations is not sufficient for inferences. To address this issue, a new “doubly weighted convergence” result is developed. The proposed method is applied to the US divorce rate data. We find that our more efficient estimator identifies the significant effects of divorce-law reforms on the divorce rate, and provides tighter confidence intervals than existing methods. This provides a confirmation for the empirical findings of Wolfers (2006) under more general unobserved heterogeneity.

Original languageEnglish (US)
Pages (from-to)59-78
Number of pages20
JournalJournal of Econometrics
Volume200
Issue number1
DOIs
StatePublished - Sep 2017

Fingerprint

Panel data
Covariance matrix
Estimator
Inference
Divorce
Law reform
Statistical inference
Minimax
Convergence rate
Unobserved heterogeneity
Confidence interval
Rate of convergence
Efficient estimation
Heteroskedasticity

All Science Journal Classification (ASJC) codes

  • Economics and Econometrics

Keywords

  • Conditional sparsity
  • Cross-sectional correlation
  • Heteroskedasticity
  • High dimensionality
  • Optimal weight matrix
  • Thresholding
  • Unknown factors

Cite this

@article{b65c82694660442b8995e920038b89ca,
title = "Inferences in panel data with interactive effects using large covariance matrices",
abstract = "We consider efficient estimation of panel data models with interactive effects, which relies on a high-dimensional inverse covariance matrix estimator. By using a consistent estimator of the error covariance matrix, we can take into account both cross-sectional correlations and heteroskedasticity. In the presence of cross-sectional correlations, the proposed estimator eliminates the cross-sectional correlation bias, and is more efficient than the existing methods. The rate of convergence is also improved. In addition, we find that when the statistical inference involves estimating a high-dimensional inverse covariance matrix, the minimax convergence rate on large covariance estimations is not sufficient for inferences. To address this issue, a new “doubly weighted convergence” result is developed. The proposed method is applied to the US divorce rate data. We find that our more efficient estimator identifies the significant effects of divorce-law reforms on the divorce rate, and provides tighter confidence intervals than existing methods. This provides a confirmation for the empirical findings of Wolfers (2006) under more general unobserved heterogeneity.",
keywords = "Conditional sparsity, Cross-sectional correlation, Heteroskedasticity, High dimensionality, Optimal weight matrix, Thresholding, Unknown factors",
author = "Jushan Bai and Yuan Liao",
year = "2017",
month = "9",
doi = "10.1016/j.jeconom.2017.05.014",
language = "English (US)",
volume = "200",
pages = "59--78",
journal = "Journal of Econometrics",
issn = "0304-4076",
publisher = "Elsevier BV",
number = "1",

}

Inferences in panel data with interactive effects using large covariance matrices. / Bai, Jushan; Liao, Yuan.

In: Journal of Econometrics, Vol. 200, No. 1, 09.2017, p. 59-78.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Inferences in panel data with interactive effects using large covariance matrices

AU - Bai, Jushan

AU - Liao, Yuan

PY - 2017/9

Y1 - 2017/9

N2 - We consider efficient estimation of panel data models with interactive effects, which relies on a high-dimensional inverse covariance matrix estimator. By using a consistent estimator of the error covariance matrix, we can take into account both cross-sectional correlations and heteroskedasticity. In the presence of cross-sectional correlations, the proposed estimator eliminates the cross-sectional correlation bias, and is more efficient than the existing methods. The rate of convergence is also improved. In addition, we find that when the statistical inference involves estimating a high-dimensional inverse covariance matrix, the minimax convergence rate on large covariance estimations is not sufficient for inferences. To address this issue, a new “doubly weighted convergence” result is developed. The proposed method is applied to the US divorce rate data. We find that our more efficient estimator identifies the significant effects of divorce-law reforms on the divorce rate, and provides tighter confidence intervals than existing methods. This provides a confirmation for the empirical findings of Wolfers (2006) under more general unobserved heterogeneity.

AB - We consider efficient estimation of panel data models with interactive effects, which relies on a high-dimensional inverse covariance matrix estimator. By using a consistent estimator of the error covariance matrix, we can take into account both cross-sectional correlations and heteroskedasticity. In the presence of cross-sectional correlations, the proposed estimator eliminates the cross-sectional correlation bias, and is more efficient than the existing methods. The rate of convergence is also improved. In addition, we find that when the statistical inference involves estimating a high-dimensional inverse covariance matrix, the minimax convergence rate on large covariance estimations is not sufficient for inferences. To address this issue, a new “doubly weighted convergence” result is developed. The proposed method is applied to the US divorce rate data. We find that our more efficient estimator identifies the significant effects of divorce-law reforms on the divorce rate, and provides tighter confidence intervals than existing methods. This provides a confirmation for the empirical findings of Wolfers (2006) under more general unobserved heterogeneity.

KW - Conditional sparsity

KW - Cross-sectional correlation

KW - Heteroskedasticity

KW - High dimensionality

KW - Optimal weight matrix

KW - Thresholding

KW - Unknown factors

UR - http://www.scopus.com/inward/record.url?scp=85025446992&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85025446992&partnerID=8YFLogxK

U2 - 10.1016/j.jeconom.2017.05.014

DO - 10.1016/j.jeconom.2017.05.014

M3 - Article

AN - SCOPUS:85025446992

VL - 200

SP - 59

EP - 78

JO - Journal of Econometrics

JF - Journal of Econometrics

SN - 0304-4076

IS - 1

ER -