RESET for quantile regression

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This paper proposes a simple specification test for quantile regression models. Our test is based on Ramsey's (J. R. Stat. Soc. B 31:350-371, 1969) RESET (regression specification error test). Comparing to existing nonparametric specification tests, the proposed test does not contain kernel functions and bandwidth parameters and is easy to implement. Although the proposed test is not necessarily consistent against all types of misspecification, simulation results indicate that our test has reasonable size and power properties and can be more powerful than nonparametric specification tests in small samples.

Original languageEnglish
Pages (from-to)381-391
Number of pages11
JournalTest
Volume18
Issue number2
DOIs
Publication statusPublished - 2009 Aug 1
Externally publishedYes

Fingerprint

Quantile Regression
Specification Test
Regression
Specification
Non-parametric test
Misspecification
Kernel Function
Small Sample
Regression Model
Bandwidth
Quantile regression
Specification error
Specification test
Simulation

Keywords

  • Quantile regression
  • RESET

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

RESET for quantile regression. / Otsu, Taisuke.

In: Test, Vol. 18, No. 2, 01.08.2009, p. 381-391.

Research output: Contribution to journalArticle

Otsu, Taisuke. / RESET for quantile regression. In: Test. 2009 ; Vol. 18, No. 2. pp. 381-391.
@article{e73283d0428d438e91ddc54e53731cf5,
title = "RESET for quantile regression",
abstract = "This paper proposes a simple specification test for quantile regression models. Our test is based on Ramsey's (J. R. Stat. Soc. B 31:350-371, 1969) RESET (regression specification error test). Comparing to existing nonparametric specification tests, the proposed test does not contain kernel functions and bandwidth parameters and is easy to implement. Although the proposed test is not necessarily consistent against all types of misspecification, simulation results indicate that our test has reasonable size and power properties and can be more powerful than nonparametric specification tests in small samples.",
keywords = "Quantile regression, RESET",
author = "Taisuke Otsu",
year = "2009",
month = "8",
day = "1",
doi = "10.1007/s11749-008-0097-7",
language = "English",
volume = "18",
pages = "381--391",
journal = "Trabajos de Estadistica Y de Investigacion Operativa",
issn = "0041-0241",
publisher = "Springer New York",
number = "2",

}

TY - JOUR

T1 - RESET for quantile regression

AU - Otsu, Taisuke

PY - 2009/8/1

Y1 - 2009/8/1

N2 - This paper proposes a simple specification test for quantile regression models. Our test is based on Ramsey's (J. R. Stat. Soc. B 31:350-371, 1969) RESET (regression specification error test). Comparing to existing nonparametric specification tests, the proposed test does not contain kernel functions and bandwidth parameters and is easy to implement. Although the proposed test is not necessarily consistent against all types of misspecification, simulation results indicate that our test has reasonable size and power properties and can be more powerful than nonparametric specification tests in small samples.

AB - This paper proposes a simple specification test for quantile regression models. Our test is based on Ramsey's (J. R. Stat. Soc. B 31:350-371, 1969) RESET (regression specification error test). Comparing to existing nonparametric specification tests, the proposed test does not contain kernel functions and bandwidth parameters and is easy to implement. Although the proposed test is not necessarily consistent against all types of misspecification, simulation results indicate that our test has reasonable size and power properties and can be more powerful than nonparametric specification tests in small samples.

KW - Quantile regression

KW - RESET

UR - http://www.scopus.com/inward/record.url?scp=67949111130&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67949111130&partnerID=8YFLogxK

U2 - 10.1007/s11749-008-0097-7

DO - 10.1007/s11749-008-0097-7

M3 - Article

AN - SCOPUS:67949111130

VL - 18

SP - 381

EP - 391

JO - Trabajos de Estadistica Y de Investigacion Operativa

JF - Trabajos de Estadistica Y de Investigacion Operativa

SN - 0041-0241

IS - 2

ER -