Foreign bias in Australia's international equity holdings

• Author: Anil V. Mishra
• Published date: 01 April 2017
• Date: 01 April 2017
• DOI: http://doi.org/10.1016/j.rfe.2016.09.002

Foreign bias in Australia's international equity holdings

Anil V. Mishra

Schoolof Business, Western SydneyUniversity, Australia

abstractarticle info

Articlehistory:

Received5 March 2016

Receivedin revised form 26 July 2016

Accepted16 September 2016

Availableonline 22 September2016

The paperuses various approaches:capital asset pricing, mean-variance, globalminimum-variance,Bayes-Stein,

Bayesianand multi-prior to develop foreignequity bias measures for Australia’sinternationalequity holdings in

41 countries,over the period 2001to 2012. Bayesian modelsallow for various degreesof mis-trust in the ICAPM

model. Multi-Prior restrictsthe expected return for each assetto lie within specified confidenceinterval around

its estimatedvalue. Mean-Variancecomputes optimal weightsby sample estimates of meanand covariance ma-

trix of samplereturn. Bayes-Stein shrinkseach asset’s historicalmean return toward thereturn of the minimum

varianceportfolio andimproves precisionassociated with estimatingthe expectedreturn of each asset.The plau-

sible sourcesof foreign equity bias are trade, GDP per capita,real GDP growth rate, exchange ratevolatility, tax

credit, stock market development, familiarity and institution variables. The paper finds that economic cost of

theobserved foreign bias is low.The paper analyses correlationeffect on the foreignbias and finds that economic

loss decreaseswith an increase in correlation.

© 2016 ElsevierInc. All rights reserved.

1. Introduction

Australia'sequity holdings as a percentage of total international eq-

uity holdingsare primarily concentrated in UnitedStates (43%), United

Kingdom (10%) and Japan (5%).

1

Australian investors are found to ex-

hibit bias towards these three foreign countries in their international

equityholdings. The paperfocuses on the extentto which Australian in-

vestors' underweight or overweight foreign markets in their interna-

tional equityholdings.

There are no papers which specifically st udy foreign bias in

Australia's international equity holdings across a range of developed

anddeveloping countries.

2

This is firstpaper to develop measuresof for-

eign equity bias for Australia that takesinto account the scepticism of

investors in the CAPM model.

3

This paper also develops foreign bias

measures for Australia's intern ational equity holdings based on

Garlappi,Uppal, and Wang (2007)Multi-Prior model's vola tility correc-

tion technique.

4

The paper develops foreign equity bias measuresfor Australia's in-

ternationalequity holdingsbased on a Bayes–Stein shrinkage estimator

that minimizesthe impact of estimation error by shrinking the sample

mean towardsa minimum-varianceportfolio, thus improvingprecision

in estimatingthe expected returnfrom each asset. The improvedability

to estimate expected returns result s in improved out-of-sample

performance.

5

Thepaper develops measuresof foreign equitybias for Australia'sin-

ternational equity holdings in41 countries, using capital asset pricing

(CAPM), mean–variance, global minim um–variance, Bayes–Stein,

Bayesianand multi-priorapproaches. The paperalso identifies the plau-

sible sources of foreign equity bias in Australia's international equity

holdings. In a dynamicpanel setting over the period 2001–2012,it re-

latesthe measures of foreignbias to various categoriesof variables:eco-

nomic development (trade, gros s domestic product per capita , real

gross domesticproduct growth rate), stockmarket development (size,

turnover, foreign listing),familiarity (language, trade), institution(in-

stitutionalquality, legal)and other variables(real exchange rate volatil-

ity, reward risk, tax credit, global financial c risis, distance). The

empiricalestimation employsArellano–Bover/Blundell–Bondlinear dy-

namicpanel-data methodsto account for countryspecificheterogeneity

and to controlfor simultaneity biascaused by the possibilitythat some

of the explanatory variablesare endogenous.

6

The papercontributes to the existingliterature by assessingthe eco-

nomic cost of the observed foreign bias in the Australian context. The

paper analyses the effect of corre lation on foreign bias. Levy (2013)

Reviewof Financial Economics 33 (2017)41–54

E-mailaddress: a.mishra@westernsydney.edu.au.

1

Author's own calculations based on 2012 Coordinated Portfolio Investment Survey

(CPIS)dataset.

2

Mishra (2011)studies home bias that relies on the ICAPMapproach. Warren (2010)

examinesequity home biasfor Australia superannuationfunds usinga model that reflects

observeddecision processes.

3

Thereare few papers in theglobal context thatemploy Bayesian approachto take into

accountinvestors' scepticism in the ICAPMmodel (Pastor (200 0),Pastorand Stambaugh

(2000),Li(2 004),Asgharianand Hansson (2006),Baeleet al. (2007) and others).

4

Knight(1921) states that the Bayesian decisionmaker is neutral to uncertainty.

5

See Stein (1955),Berger(1974),Gorman and Jorgensen(2002),Herold and Maurer

(2003),Ledoitand Wolf (2003),Wang(2005),Zellner (2010)f or shrinkage approach .

6

Ahearneet al. (2004) and Chan et al. (2005)use pure cross-sectionalanalysis.

http://dx.doi.org/10.1016/j.rfe.2016.09.002

1058-3300/©2016 Elsevier Inc. All rightsreserved.

Contents listsavailable at ScienceDirect

Review of Financial Economics

journal homepage: www.elsevier.com/locate/rfe