Decoupling and Contagion: The Special Case of the Eurozone Sovereign Debt Crisis

• Author: Dirk G. Baur
• Date: 01 March 2020
• DOI: http://doi.org/10.1111/irfi.12220
• Published date: 01 March 2020

Decoupling and Contagion: The

Special Case of the Eurozone

Sovereign Debt Crisis

DIRK G. BAUR

Accounting and Finance, UWA Business School, Perth, Western Australia, Australia

ABSTRACT

This paper demonstrates that the Eurozone sovereign debt crisis constitutes a

special case in the contagion literature with general implications. Perfectly

correlated bond markets imply that contagion can only occur if there is a

decoupling to lower correlation levels with increased idiosyncratic shocks

leading to more severe but less systemic spillovers. This theoretical prediction

is fully supported by the empirical analysis. We also show that dynamic

coexceedance estimates provide a more robust and more general picture of

contagion than correlation-based tests. Coexceedances identify only one

major incidence of contagion that affected five periphery Eurozone countries

in May 2010 and coincided with flight to quality from the periphery to the

core and the 2010 “flash crash”in US equity markets.

JEL Codes: G010; G140; G180; G320

Accepted: 10 June 2018

I. INTRODUCTION

This paper analyzes financial contagion in sovereign bond markets with a focus

on the European debt crisis or “Eurocrisis”between 2009 and 2011. The analy-

sis is special since it is the first crisis that affected the countries of a currency

union with highly integrated bond markets.

1

The high and quasi-perfect level

of integration implies that correlation-based tests cannot identify financial con-

tagion.

2

Since correlation-based tests define contagion as an increase in correla-

tions in a crisis period relative to a precrisis period, if the precrisis correlation

level is one, a positive change is not possible ruling out contagion by definition.

The perfect correlation levels of the Eurozone bond markets thus provide an

1 Lehkonen (2014) studies stock market integration and financial crises for a large sample of

industrial and emerging markets.

2 The first contagion tests (Baig and Goldfajn 1999; Forbes and Rigobon 2002; Bekaert

et al. 2005) were based on a change in cross-country correlations in a crisis period relative to

a tranquil precrisis period. A significant increase in correlations was interpreted as evidence

for contagion. These tests were designed for equity markets with average return correlations

well below one. Dungey et al. (2005) provide an overview of existing methodologies.

© 2018 International Review of Finance Ltd. 2018

International Review of Finance, 20:1, 2020: pp. 133–154

DOI: 10.1111/irfi.12220

interesting case study how high levels of integration

3

affect contagion or more

generally what role the degree of market integration and diversification play for

the transmission of shocks and the severity of contagion.

We use a simple model to highlight the inverse relationship between integra-

tion and contagion. The model predicts that full integration minimizes the

magnitude of contagion as shocks are diversified and shared while full segmen-

tation maximizes the potential magnitude of contagion as shocks are not diver-

sified and not shared. The model and empirical analysis adds to the recent

literature on the relationship between integration, financial crisis, and conta-

gion (e.g., Berger and Pukthuanthong 2012; Bekaert et al. 2014; Lehko-

nen 2014).

We are not the first to analyze financial contagion in the Eurozone debt cri-

sis. The literature is growing rapidly and it is impossible to cite and discuss all

of them. The recent literature includes Blatt et al. (2015), Caporin et al. (2013),

Claeys and Vasicek (2012), De Santis (2012), Mink and DeHaan (2012), Missio

and Watzka (2011), and Pragidis et al. (2015) among many others. One of the

first studies on bond market contagion is Dungey et al. (2006a) who analyze

the Russian and LTCM crisis. Favero and Giavazzi (2000) analyze interest rate

spreads on German rates between 1988 and 1992.

The first studies on financial contagion focused on stock market contagion.

Baig and Goldfajn (1999) and Forbes and Rigobon (2002) used correlation-based

tests with the latter introducing an adjustment for heteroskedasticity. Chiang

et al. (2007) estimate dynamic conditional correlations and model heteroske-

dasticity within a multivariate GARCH framework. Bae et al. (2003) used coex-

ceedances to measure financial contagion and Jondeau and Rockinger (2006)

and Rodriguez (2007) employ copula-based models. Hong et al. (2009) model

extreme risk spillovers between financial markets using a Granger causality

framework. The importance of methodological and econometric issues in the

analysis of contagion is examined in Corsetti et al. (2005), Dungey et al. (2005),

Pericoli and Sbracia (2003), and Pesaran and Pick (2007). These studies indicate

that the analysis of financial contagion is not straightforward.

We contribute to the growing literature by emphasizing that contagion in

the Eurozone bond markets represents a special case and by theoretically and

empirically analyzing the role of market integration and correlations for the

incidence of contagion. More specifically, we show that the Eurozone bond

markets partially disintegrated thereby increasing the potential magnitude of

financial contagion. In other words, the disintegration was a necessary condi-

tion for strong contagion effects. We also demonstrate that correlation-based

measures are too narrow to identify contagious effects in periods of decoupling

in contrast to coexceedances which identify contagion in such conditions and

3 Pukthuanthong and Roll (2009) argue that correlations are a poor measure of integration for

correlation levels smaller than one, for example, correlations can be small even when two

countries are perfectly integrated. However, if the correlations are equal to one for two coun-

tries, they are perfectly integrated.

© 2018 International Review of Finance Ltd. 2018134

International Review of Finance