Quantile regression models with factor-augmented predictors and information criterion

Tomohiro Ando, Ruey S. Tsay

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

For situations with a large number of series,N, each withTobservations and each containing a certain amount of information for prediction of the variable of interest, we propose a new statistical modelling methodology that first estimates the common factors from a panel of data using principal component analysis and then employs the estimated factors in a standard quantile regression. A crucial step in the model-building process is the selection of a good model among many possible candidates. Taking into account the effect of estimated regressors, we develop an information-theoretic criterion. We also investigate the criterion when there is no estimated regressors. Results of Monte Carlo simulations demonstrate that the proposed criterion performs well in a wide range of situations.

Original languageEnglish
Pages (from-to)1-24
Number of pages24
JournalEconometrics Journal
Volume14
Issue number1
DOIs
Publication statusPublished - 2011 Feb

Fingerprint

Predictors
Regression model
Quantile regression
Factors
Information criterion
Common factors
Prediction
Monte Carlo simulation
Principal component analysis
Modeling methodology

Keywords

  • Approximate factor models
  • Generated regressors
  • Information-theoretic approach
  • Panel data
  • Quantiles

ASJC Scopus subject areas

  • Economics and Econometrics

Cite this

Quantile regression models with factor-augmented predictors and information criterion. / Ando, Tomohiro; Tsay, Ruey S.

In: Econometrics Journal, Vol. 14, No. 1, 02.2011, p. 1-24.

Research output: Contribution to journalArticle

Ando, Tomohiro ; Tsay, Ruey S. / Quantile regression models with factor-augmented predictors and information criterion. In: Econometrics Journal. 2011 ; Vol. 14, No. 1. pp. 1-24.
@article{179ea709108f4afa93797b0dadc1e5f0,
title = "Quantile regression models with factor-augmented predictors and information criterion",
abstract = "For situations with a large number of series,N, each withTobservations and each containing a certain amount of information for prediction of the variable of interest, we propose a new statistical modelling methodology that first estimates the common factors from a panel of data using principal component analysis and then employs the estimated factors in a standard quantile regression. A crucial step in the model-building process is the selection of a good model among many possible candidates. Taking into account the effect of estimated regressors, we develop an information-theoretic criterion. We also investigate the criterion when there is no estimated regressors. Results of Monte Carlo simulations demonstrate that the proposed criterion performs well in a wide range of situations.",
keywords = "Approximate factor models, Generated regressors, Information-theoretic approach, Panel data, Quantiles",
author = "Tomohiro Ando and Tsay, {Ruey S.}",
year = "2011",
month = "2",
doi = "10.1111/j.1368-423X.2010.00320.x",
language = "English",
volume = "14",
pages = "1--24",
journal = "Econometrics Journal",
issn = "1368-4221",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - Quantile regression models with factor-augmented predictors and information criterion

AU - Ando, Tomohiro

AU - Tsay, Ruey S.

PY - 2011/2

Y1 - 2011/2

N2 - For situations with a large number of series,N, each withTobservations and each containing a certain amount of information for prediction of the variable of interest, we propose a new statistical modelling methodology that first estimates the common factors from a panel of data using principal component analysis and then employs the estimated factors in a standard quantile regression. A crucial step in the model-building process is the selection of a good model among many possible candidates. Taking into account the effect of estimated regressors, we develop an information-theoretic criterion. We also investigate the criterion when there is no estimated regressors. Results of Monte Carlo simulations demonstrate that the proposed criterion performs well in a wide range of situations.

AB - For situations with a large number of series,N, each withTobservations and each containing a certain amount of information for prediction of the variable of interest, we propose a new statistical modelling methodology that first estimates the common factors from a panel of data using principal component analysis and then employs the estimated factors in a standard quantile regression. A crucial step in the model-building process is the selection of a good model among many possible candidates. Taking into account the effect of estimated regressors, we develop an information-theoretic criterion. We also investigate the criterion when there is no estimated regressors. Results of Monte Carlo simulations demonstrate that the proposed criterion performs well in a wide range of situations.

KW - Approximate factor models

KW - Generated regressors

KW - Information-theoretic approach

KW - Panel data

KW - Quantiles

UR - http://www.scopus.com/inward/record.url?scp=79951740418&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79951740418&partnerID=8YFLogxK

U2 - 10.1111/j.1368-423X.2010.00320.x

DO - 10.1111/j.1368-423X.2010.00320.x

M3 - Article

AN - SCOPUS:79951740418

VL - 14

SP - 1

EP - 24

JO - Econometrics Journal

JF - Econometrics Journal

SN - 1368-4221

IS - 1

ER -