Analysing the return distributions of Australian stocks: the CAPM, factor models and quantile regressions

David E. Allen, Abhay Kumar Singh, Robert Powell

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Traditionally, ordinary least square (OLS) regression methods are used to test asset pricing models. This study focuses on the use of quantile regression as an alternative approach to the analysis of risk and return distributions in quantitative finance. It empirically examines the behaviour of two widely used asset pricing factors, beta and book to market ratios, but the focus is on minimising absolute deviations around the median rather than minimising squared deviations around the mean of their distributions, as we apply quantile regressions as opposed to OLS. We show how OLS is less able to capture the extreme values or the adverse losses in the return distribution, which on the other hand are captured by quantile regressions. The study not only shows that the factors do not necessarily follow a linear relationship but also shows that the traditional use of OLS becomes less effective when it comes to analysing the extremes within a distribution, which are often a source of keen interest for investors and risk managers.

LanguageEnglish
Pages88-109
Number of pages22
JournalGlobal Business and Economics Review
Volume15
Issue number1
DOIs
Publication statusPublished - 2013
Externally publishedYes

Fingerprint

Capital asset pricing model
Ordinary least squares
Quantile regression
Return distribution
Factors
Deviation
Asset pricing
Median
Book-to-market ratio
Quantitative finance
Risk and return
Asset pricing models
Extreme values
Regression method
Investors
Managers

Keywords

  • book to market ratios
  • capital asset pricing model
  • CAPM
  • quantile regression
  • S&P/ASX-50 stocks

Cite this

@article{2d98ac318759482c8344627f75920fc7,
title = "Analysing the return distributions of Australian stocks: the CAPM, factor models and quantile regressions",
abstract = "Traditionally, ordinary least square (OLS) regression methods are used to test asset pricing models. This study focuses on the use of quantile regression as an alternative approach to the analysis of risk and return distributions in quantitative finance. It empirically examines the behaviour of two widely used asset pricing factors, beta and book to market ratios, but the focus is on minimising absolute deviations around the median rather than minimising squared deviations around the mean of their distributions, as we apply quantile regressions as opposed to OLS. We show how OLS is less able to capture the extreme values or the adverse losses in the return distribution, which on the other hand are captured by quantile regressions. The study not only shows that the factors do not necessarily follow a linear relationship but also shows that the traditional use of OLS becomes less effective when it comes to analysing the extremes within a distribution, which are often a source of keen interest for investors and risk managers.",
keywords = "book to market ratios, capital asset pricing model, CAPM, quantile regression, S&P/ASX-50 stocks",
author = "Allen, {David E.} and Singh, {Abhay Kumar} and Robert Powell",
year = "2013",
doi = "10.1504/GBER.2013.050670",
language = "English",
volume = "15",
pages = "88--109",
journal = "Global Business and Economics Review",
issn = "1097-4954",
publisher = "Inderscience Enterprises Ltd.",
number = "1",

}

Analysing the return distributions of Australian stocks : the CAPM, factor models and quantile regressions. / Allen, David E.; Singh, Abhay Kumar; Powell, Robert.

In: Global Business and Economics Review, Vol. 15, No. 1, 2013, p. 88-109.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Analysing the return distributions of Australian stocks

T2 - Global Business and Economics Review

AU - Allen, David E.

AU - Singh, Abhay Kumar

AU - Powell, Robert

PY - 2013

Y1 - 2013

N2 - Traditionally, ordinary least square (OLS) regression methods are used to test asset pricing models. This study focuses on the use of quantile regression as an alternative approach to the analysis of risk and return distributions in quantitative finance. It empirically examines the behaviour of two widely used asset pricing factors, beta and book to market ratios, but the focus is on minimising absolute deviations around the median rather than minimising squared deviations around the mean of their distributions, as we apply quantile regressions as opposed to OLS. We show how OLS is less able to capture the extreme values or the adverse losses in the return distribution, which on the other hand are captured by quantile regressions. The study not only shows that the factors do not necessarily follow a linear relationship but also shows that the traditional use of OLS becomes less effective when it comes to analysing the extremes within a distribution, which are often a source of keen interest for investors and risk managers.

AB - Traditionally, ordinary least square (OLS) regression methods are used to test asset pricing models. This study focuses on the use of quantile regression as an alternative approach to the analysis of risk and return distributions in quantitative finance. It empirically examines the behaviour of two widely used asset pricing factors, beta and book to market ratios, but the focus is on minimising absolute deviations around the median rather than minimising squared deviations around the mean of their distributions, as we apply quantile regressions as opposed to OLS. We show how OLS is less able to capture the extreme values or the adverse losses in the return distribution, which on the other hand are captured by quantile regressions. The study not only shows that the factors do not necessarily follow a linear relationship but also shows that the traditional use of OLS becomes less effective when it comes to analysing the extremes within a distribution, which are often a source of keen interest for investors and risk managers.

KW - book to market ratios

KW - capital asset pricing model

KW - CAPM

KW - quantile regression

KW - S&P/ASX-50 stocks

UR - http://www.scopus.com/inward/record.url?scp=84870676800&partnerID=8YFLogxK

U2 - 10.1504/GBER.2013.050670

DO - 10.1504/GBER.2013.050670

M3 - Article

VL - 15

SP - 88

EP - 109

JO - Global Business and Economics Review

JF - Global Business and Economics Review

SN - 1097-4954

IS - 1

ER -