Canonical Correlation-based Model Selection for the Multilevel Factors

In Choi; Rui Lin; Yongcheol Shin

doi:10.1016/j.jeconom.2021.09.008

Canonical Correlation-based Model Selection for the Multilevel Factors

In Choi, Rui Lin, Yongcheol Shin

Economics

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

Abstract

We develop a novel approach based on the canonical correlation analysis to identify the number of the global factors in the multilevel factor model. We propose the two consistent selection criteria, the canonical correlations difference (CCD) and the modified canonical correlations (MCC). Via Monte Carlo simulations, we show that CCD and MCC select the number of global factors correctly even in small samples, and they are robust to the presence of serially correlated and weakly cross-sectionally correlated idiosyncratic errors as well as the correlated local factors. Finally, we demonstrate the utility of our approach with an application to the multilevel asset pricing model for the stock return data in 12 industries in the U.S.

Original language	English
Pages (from-to)	22-44
Number of pages	23
Journal	Journal of Econometrics
Volume	233
Issue number	1
DOIs	https://doi.org/10.1016/j.jeconom.2021.09.008
Publication status	Published - 1 Mar 2023

Bibliographical note

Keywords

Multilevel Factor Models
Principal Components
Canonical Correlation Difference
Modified Canonical Correlations
Multilevel Asset Pricing Models

Access to Document

10.1016/j.jeconom.2021.09.008

CLS
© 2021 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, 590 KBLicence: CC BY-NC-ND

New Cross-Sectionally Dependent Panel Data Methods for the Analysis of Macroeconomic and Financial Networks
Shin, Y. (Principal investigator), Chen, J. (Co-investigator) & Wang, W. (Co-investigator)
ECONOMIC AND SOCIAL RESEARCH COUNCIL (ESRC)
1/12/20 → 31/07/24
Project: Research project (funded) › Research

Cite this

@article{4696516647914869825a042db47e66ad,

title = "Canonical Correlation-based Model Selection for the Multilevel Factors",

abstract = "We develop a novel approach based on the canonical correlation analysis to identify the number of the global factors in the multilevel factor model. We propose the two consistent selection criteria, the canonical correlations difference (CCD) and the modified canonical correlations (MCC). Via Monte Carlo simulations, we show that CCD and MCC select the number of global factors correctly even in small samples, and they are robust to the presence of serially correlated and weakly cross-sectionally correlated idiosyncratic errors as well as the correlated local factors. Finally, we demonstrate the utility of our approach with an application to the multilevel asset pricing model for the stock return data in 12 industries in the U.S.",

keywords = "Multilevel Factor Models, Principal Components, Canonical Correlation Difference, Modified Canonical Correlations, Multilevel Asset Pricing Models",

author = "In Choi and Rui Lin and Yongcheol Shin",

note = "{\textcopyright} 2021 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 = "2023",

month = mar,

day = "1",

doi = "10.1016/j.jeconom.2021.09.008",

language = "English",

volume = "233",

pages = "22--44",

journal = "Journal of Econometrics",

issn = "0304-4076",

publisher = "Elsevier",

number = "1",

}

TY - JOUR

T1 - Canonical Correlation-based Model Selection for the Multilevel Factors

AU - Choi, In

AU - Lin, Rui

AU - Shin, Yongcheol

PY - 2023/3/1

Y1 - 2023/3/1

N2 - We develop a novel approach based on the canonical correlation analysis to identify the number of the global factors in the multilevel factor model. We propose the two consistent selection criteria, the canonical correlations difference (CCD) and the modified canonical correlations (MCC). Via Monte Carlo simulations, we show that CCD and MCC select the number of global factors correctly even in small samples, and they are robust to the presence of serially correlated and weakly cross-sectionally correlated idiosyncratic errors as well as the correlated local factors. Finally, we demonstrate the utility of our approach with an application to the multilevel asset pricing model for the stock return data in 12 industries in the U.S.

AB - We develop a novel approach based on the canonical correlation analysis to identify the number of the global factors in the multilevel factor model. We propose the two consistent selection criteria, the canonical correlations difference (CCD) and the modified canonical correlations (MCC). Via Monte Carlo simulations, we show that CCD and MCC select the number of global factors correctly even in small samples, and they are robust to the presence of serially correlated and weakly cross-sectionally correlated idiosyncratic errors as well as the correlated local factors. Finally, we demonstrate the utility of our approach with an application to the multilevel asset pricing model for the stock return data in 12 industries in the U.S.

KW - Multilevel Factor Models

KW - Principal Components

KW - Canonical Correlation Difference

KW - Modified Canonical Correlations

KW - Multilevel Asset Pricing Models

U2 - 10.1016/j.jeconom.2021.09.008

DO - 10.1016/j.jeconom.2021.09.008

M3 - Article

SN - 0304-4076

VL - 233

SP - 22

EP - 44

JO - Journal of Econometrics

JF - Journal of Econometrics

IS - 1

ER -

Canonical Correlation-based Model Selection for the Multilevel Factors

Abstract

Bibliographical note

Keywords

Access to Document

Projects

New Cross-Sectionally Dependent Panel Data Methods for the Analysis of Macroeconomic and Financial Networks

Cite this