Contemporaneous Spillover Effects between the U.S. and the U.K. Equity Markets

• Author: Marinela Adriana Finta,Alireza Tourani‐Rad,Bart Frijns
• Date: 01 February 2017
• DOI: http://doi.org/10.1111/fire.12116
• Published date: 01 February 2017

The Financial Review 52 (2017) 145–166

Contemporaneous Spillover Effects

between the U.S. and the U.K. Equity

Markets

Marinela Adriana Finta∗, Bart Frijns, and Alireza Tourani-Rad

Auckland University of Technology

Abstract

We use high frequency data and the “identificationthrough heteroskedasticity” approach

of Rigobon (2003) to capture the contemporaneous volatility spillover effects between the

U.S. and U.K. equity markets. We demonstrate the relevance of taking into account the in-

formation present during simultaneous trading hours by comparing the results generated by

our structural vector autoregression with those of a traditional reduced-form vector autoregres-

sion. Our findings clearly demonstrate that contemporaneous relations matter and that ignoring

them leads to inappropriate conclusions regarding the magnitude and direction of volatility

spillover.

Keywords: contemporaneous spillovers, identification through heteroskedasticity, volatility

spillovers

JEL Classifications: C32, C58, G1

∗Corresponding author: Department of Finance, Auckland University of Technology, Private Bag 92006,

1020 Auckland, New Zealand; Phone: +64 9 921 9999, ext. 7151; Fax: +64 9 921 9940; E-mails:

marinela.finta@aut.ac.nz, marinela.finta@gmail.com.

We thank participants at the 2015 European Financial Management Association, the 5th International

Conference of the Financial Engineering and Banking Society, the 2nd International Workshopon Finan-

cial Markets and Nonlinear Dynamics, the 2015 Annual Scientific Conference of Romanian Academic

Economists from Abroad, the 2015 workshop at Fern Universityin Hagen, the 2014 New Zealand Finance

Colloquium, the 2014 Conference on High Frequency Data and DerivativeMarkets and the 2014 Auckland

Finance Meeting for helpful comments and suggestions.

C2017 The Eastern Finance Association 145

146 M. A. Finta et al./The Financial Review 52 (2017) 145–166

1. Introduction

The global financial crisis, which started with Lehman Brothers’ collapse on

September 2008 and rapidly spread to other countries, triggered a resurgence in

research on the international transmission of volatility (Diebold and Yilmaz, 2009;

Singh, Kumar and Pandey, 2010; Dimpfl and Jung, 2012; Louzis, 2015). This crisis

caused a period of high volatility and instability in equity markets, and had a strong

negative impact in terms of economic growth for many economies around the world.

The crisis once again highlighted that economic shocks originating in one market not

only affect that particular market, but are transmitted to other markets with serious

implications for financial markets. Understanding these “spillover effects” among

equity markets is therefore of great importance.

While many studies consider the dynamic aspect of volatility spillover, the total

volatility spillover between markets can be explained by both dynamic and contem-

poraneous effects. The dynamic effects refer to spillovers that occur over time, for

example, when information from one market can only affect the other market in the

next trading period. Contemporaneous spillovers are the spillovers that take place at

the same time, for example, when markets have overlapping trading hours and infor-

mation can be transmitted immediately. Given that many markets have overlapping

trading hours, capturing these contemporaneous spillovers is important. However,

contemporaneous effects cannot be identified using traditional vector autoregression

(VAR) and generalized autoregressive conditional heteroskedasticity (GARCH) mod-

els (Savva, Osborn and Gill, 2009; Hakim and McAleer, 2010), hence many studies

have not considered these effects.

As a way to identify the contemporaneous spillovers, Rigobon (2003) proposes

the “identification through heteroskedasticity” approach. This technique relies on the

existence of nonproportional changes in volatility to identify the contemporaneous in-

teractions without any a priori assumptions regarding the direction of causality. This

approach has recently been employed by Andersen, Bollerslev, Diebold and Vega

(2007) and Ehrmann, Fratzscher and Rigobon (2011) who show that there are con-

temporaneous spillover effects among different marketsand assets at the return level.

In this study, we analyze the contemporaneous spillovers in volatility, by using

a structural VAR (SVAR) and Rigobon’s (2003) approach. Specifically, we focus on

the U.S. and the U.K. stock markets as part of their trading hours overlap.1We us e

high frequency data and split the trading period in three: the period before the U.S.

market opens, the overlapping trading period, and the period when the U.K. market

1These markets share up to three hours of overlapping trading. In contrast to Dimpfl and Jung (2012) and

of Clements, Hurn and Volkov(2015), we exclude Japan, as this market does not share any overlapping

trading hours with the United States and the United Kingdom. The United States and the United Kingdom

are highly integrated (Morana and Beltratti, 2008) and related through trade and investments,so that news

in one market most likely has implications for the other. For instance, Andersen, Bollerslev, Diebold and

Vega(2007) show that news is transmitted fast and there are direct spillovers among the U.S. and European

equity markets.