Journal of International Money and Finance 31 (2012) 517533

Contents lists available at SciVerse ScienceDirect

Journal of International Money

and Finance
journal homepage: www.elsevier.com/locate/jimf

Transmission of the nancial and sovereign debt crises to

the EMU: Stock prices, CDS spreads and exchange rates
Theoharry Grammatikos a, Robert Vermeulen b, *
a
University of Luxembourg, Luxembourg School of Finance, 4 Rue Albert Borschette, L-1246 Luxembourg, Luxembourg
b
De Nederlandsche Bank, Economics and Research Division, P.O. Box 98, 1000 AB Amsterdam, The Netherlands

a b s t r a c t

JEL codes: This paper tests for the transmission of the 20072010 nancial
F31 and sovereign debt crises to fteen EMU countries. We use daily
G01 data from 2003 to 2010 on country nancial and non-nancial
G15
stock market indexes to analyze the stock market returns for
Keywords: three country groups within EMU: North, South and Small. The
Financial crisis following results hold for both the North and South European
Euro exchange rate countries, while the smallest countries seem to be relatively iso-
EMU lated from international events. First, we nd strong evidence of
Equity markets crisis transmission to European non-nancials from US non-
Sovereign debt
nancials, but not for nancials. Second, in order to test how the
sovereign debt crisis affects stock market developments we split
the crisis in pre- and post-Lehman sub periods. Results show that
nancials become signicantly more dependent on changes in the
difference between the Greek and German CDS spreads after
Lehmans collapse, compared to the pre-Lehman sub period.
However, this increase is much smaller for non-nancials. Third,
before the crisis euro appreciations coincide with European stock
market decreases, whereas this relationship reverses during the
crisis. Finally, this reversal seems to be triggered by Lehmans
collapse.
2011 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: 31 205243228; fax: 31 205242506.

E-mail addresses: theoharry.grammatikos@uni.lu (T. Grammatikos), r.j.g.vermeulen@dnb.nl (R. Vermeulen).

0261-5606/$ see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jimonn.2011.10.004
518 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

1. Introduction

The global nancial and economic crisis of 20072010, which originated in the United States (US)
subprime mortgage market, triggered a chain reaction in the US and global nancial systems. These
problems reached their zenith with the collapse of Lehman Brothers, when it became clear that there
would be a deep global recession. European nancial markets were not isolated from developments in
the US. To stem the evolving crisis European national governments massively injected capital in
nancial institutions (e.g. Hypo Real Estate, Fortis, Anglo-Irish, etc.) and undertook scal stimulus
measures. The single currency, however, proved vulnerable to the crisis because it was created without
a single supranational government above it to control tax, spending and transfers between the euro
areas richer and poorer economies. Consequently, nancial markets distrust some of the European
governments resulting in a sovereign debt crisis. The country worst hit to date by this sovereign debt
crisis is Greece. It is in a deep economic recession and cannot borrow anymore on international capital
markets. Since early 2010 it receives nancial support from a joint EMU-IMF rescue package.1
This paper investigates the sensitivity of nancials and non-nancials in 15 EMU countries to crisis
developments in the US, to sovereign debt problems in EMU and to exchange rate movements before
and during the nancial crisis. By comparing a tranquil pre-crisis period (1-1-2003 until 26-2-2007)
with a crisis period (27-2-2007 until 31-8-2010) we test explicitly for the transmission of the nancial
crisis. In addition, we distinguish between two sub periods within the nancial crisis, where the
bankruptcy of Lehman Brothers on 15 September 2008 is the break point.
To keep the discussion tractable we split the 15 EMU countries in three groups based on the
nancial markets assessment of default risk: Low default risk, high default risk and small countries.
Low default risk countries are Austria, Belgium, Finland, France, Germany and the Netherlands,
whereas high default risk countries are Greece, Ireland, Italy, Portugal and Spain. We coin the rst
group North and the second South. The smallest EMU countries (Cyprus, Luxembourg, Malta and
Slovenia) have very illiquid stock markets and the crisis transmission process may work differently.
We estimate individual GARCH models for each groups nancials and non-nancials index. The
GARCH model is well known for its ability to deal with time varying stock market volatility, a necessity
when analyzing the turbulent markets during the nancial crisis. We include both the contempora-
neous and lagged US return in the empirical model to capture the dependence on developments in the
US and deal with non-synchronous trading hours. In addition, by including euro-dollar exchange rate
changes we control explicitly for exchange rate movements. Finally, when considering only the crisis
period, Greek CDS spreads are included to investigate the effects of the European sovereign debt crisis
on stock returns. In order to test for the crisis transmission we include an indicator variable to allow the
coefcients to change during the pre-crisis vs. crisis periods and pre-Lehman vs. post-Lehman sub
periods. A signicant coefcient for the indicator variable(s) signals transmission of the crisis.
The results show four key ndings. First there is evidence of crisis transmission to all European non-
nancials from US non-nancials, but not for nancials in the North and South groups. Second, nancials
become signicantly more sensitive to changes in the difference between the Greek and German Credit
Default Swap (CDS) spread after the collapse of Lehman Brothers across all country groups. Non-nancials
exhibit a similar pattern, albeit with smaller magnitude. Third, before the nancial crisis euro appreci-
ations coincide with stock market decreases for both European nancials and non-nancials. However,
during the crisis this association changes: stock prices increase when the euro appreciates for both
nancials and non-nancials. This holds for both the North and South European country groups. However,
stock prices in the smallest EMU countries are far less inuenced by exchange rate movements. Finally, the
reversal in the relationship between euro appreciations and stock prices seems to be triggered by Leh-
mans collapse and the concomitant radical change in default expectations among market participants.
The paper is organized as follows. Section 2 discusses the key dates of the crisis and the country
groups. Section 3 covers the data and empirical methodology. Section 4 presents the empirical ndings
and Section 5 concludes.

1
Recently, Ireland and Portugal also received support from the other EMU countries and the IMF.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 519

2. The key dates of the nancial crisis and a growing divide in Europe

In order to determine the starting point of the nancial crisis we follow the Federal Reserve Bank of
St. Louis crisis timeline.2 This timeline is very detailed and lists all key crisis events. We choose 27
February 2007 as the starting point of the nancial crisis, because on this date the rst signs of
problems in the subprime mortgage market appear with the announcement by Freddie Mac that it will
no longer buy the most risky subprime mortgages and mortgage-related securities. The crisis period
studied ends on 31 August 2010, the last day in our sample. In order to compare the results with
a normal market situation we also examine a non-crisis period which starts on 1 January 2003 and
ends on 26 February 2007. This period marks the stock market recovery after the burst of the IT bubble.
A key milestone event during the nancial crisis is the bankruptcy of Lehman Brothers on 15
September 2008. Using the words of Jean-Claude Trichet: [I]t is clear that since September 2008 we have
been facing the most difcult situation since the Second World War perhaps even since the First World
War. We have experienced and are experiencing truly dramatic times (Spiegel Online, 2010). From
then onwards credit markets freeze and the US subprime mortgage crisis turns into a global nancial
and economic crisis. Consequently, for the latter part of the study we split the crisis period into a pre-
Lehman (27 February 200714 September 2008) and a post-Lehman sub period (15 September 2008
31 August 2010).
The time series in Figs. 14 illustrate the changes before and during the crisis and support the choice
of the periods and sub periods.
Fig. 1 displays the development of the EMU nancials and non-nancials stock indexes. These
indexes are weighted averages of the EMUs individual countrys indexes, with weights based on
market capitalization. Since the general development is similar across countries, we only present the
aggregate EMU index.
Both the nancials and non-nancials indexes bolster a strong increase from 2003 until early 2007.
However, since early 2007 the nancials index starts to decrease shortly after the start of the crisis (27
February 2007). The non-nancials index lags a bit behind and starts to decline as of mid-2007. Up to
September 2008 the decrease is modest, but after the collapse of Lehman Brothers, both nancials and
non-nancials continue their downward trend and decrease further until early 2009, when the situ-
ation seems to stabilize. When assessing how both nancials and non-nancials stand on 31 August
2010, we observe an underperformance of about forty percent for nancials compared to non-nan-
cials over the full sample period.
The development of stock market volatility during 20032010 shows even clearer how the crisis
evolves. Fig. 2 plots the volatility of the EMU nancials index.3 We omit the non-nancials index for
brevity, since it exhibits a very similar volatility pattern.
During 2003 to early 2007 stock market volatility is very low. However, starting early 2007, after the
eruption of the nancial crisis, Fig. 2 shows a gradual increase in volatility. After Lehmans collapse
volatility skyrockets. This is strong evidence of the existence of a break point in the data series around
the key dates chosen for the study. This pattern is similar across all EMU countries.
The euro-dollar exchange rate shows large uctuations during 20032010 (Fig. 3). At the start of
2003 the exchange rate is around parity and appreciates slowly to around 1.3 US dollars per euro early
2007. During the rst part of the crisis, when the problems seem to be contained to the US subprime
mortgage market, the euro appreciates by a further 25% to 1.6 US dollars per euro. However, during the
summer of 2008 it becomes apparent that the nancial crisis is going to spill over to the rest of the
world and affect the real economy worldwide. This shatters all beliefs of a global decoupling from the
US. Even worse, some European economies contract even faster than the US. This results in a sharp
depreciation of the euro and an increasing volatility of the exchange rate. In fact, investors view the US
as a safe haven during the crisis and shift their equity portfolios toward US dollar assets. During 2009

2
The Federal Reserve Bank of St. Louis crisis timeline can be accessed via http://timeline.stlouisfed.org/.
3
We calculate volatility as the standard deviation of the residuals of a univariate GARCH(1,1) model for the specic time
series. The estimation incorporates a lagged return term.
520 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

200 Start crisis Lehman bankruptcy

180

160
index (31-12-2002 = 100)

140 Nonfinancials

120

100 Financials

80

60

40

20

0
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun-
02 03 03 04 04 05 05 06 06 07 07 08 08 09 09 10

Fig. 1. EMU nancials and non-nancials stock market index level. Source: Thomson Datastream.

the euro strengthens, but depreciates again during the European sovereign debt crisis. The exchange
rate is around 1.3 US dollars per euro on 31 August 2010, when the sample period ends.
Finally, to illustrate the effects of the European sovereign debt crisis we focus on two polar cases:
German government debt vs. Greek government debt. Credit Default Swaps (CDS) are a nancial
instrument to insure against the risk of a government default. The price of this insurance, i.e. the CDS
spread, is low for all EMU countries at the outset of the subprime mortgage crisis. Even when this crisis
erupts spreads are low and only start to increase slightly for Greece during 2007. The markets initially
assign a very low probability to the bankruptcy of any EMU country, i.e. markets perceive default as an
extremely unlikely event. However, investors drastically revise their beliefs following the collapse of
Lehman Brothers (Fig. 4). German CDS spreads climb to around 100 basis points in early 2009, while
Greek CDS spreads are already about 250 basis points around that time. Attinasi et al. (2009) nd that
these increasing CDS spreads are associated with bank bailout packages and effective transfer of risk
from the private sector to the government. Sgherri and Zoli (2009) provide supporting evidence by
showing that since October 2008 nancial markets become increasingly worried of the impact of the
bailout packages for government nances. The situation further aggravates when investors start

0.06 Start crisis Lehman bankruptcy

0.05

0.04
volatility

0.03

0.02

0.01

0
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun-
02 03 03 04 04 05 05 06 06 07 07 08 08 09 09 10

Fig. 2. Volatility of EMU nancials index. Source: Thomson Datastream.

 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 521

1.7
Start crisis Lehman bankruptcy

1.5
US$ per Euro

1.3

1.1

0.9
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun-
02 03 03 04 04 05 05 06 06 07 07 08 08 09 09 10

Fig. 3. Euro - US dollar exchange rate. Source: Thomson Datastream.

expressing serious doubts about the Greek governments solvency in late 2009.4 From that point
onwards the German CDS spread remains low until August 2010, but the Greek one skyrockets.
Since late 2009 a division in Europe becomes apparent with some countries following Germany
with decreasing CDS spreads and another group following Greece with increasing CDS spreads.
Table 1 presents the CDS spread in basis points at key dates during the crises for larger EMU
countries. CDS spread data on the smallest EMU countries (Cyprus, Luxembourg, Malta and Slovenia)
are often incomplete and not always reliable. Therefore we focus on the larger EMU countries and split
these into a North and a South group. The North group consists of Austria, Belgium, Finland, France,
Germany and the Netherlands. These countries do experience an increase in their CDS spread, but this
increase is rather small. The South group consists of Greece, Ireland, Italy, Portugal and Spain. These
countries have much higher CDS spreads and investors seriously worry about these countries
government nances. In fact, currently Greece, Ireland and Portugal receive nancial support from the
IMF and other EMU countries.
In the empirical analysis we investigate if stock markets perform differently in the three country
groups, by distinguishing between the returns on nancials and non-nancials indexes. Based on the
evidence in this section we split the crisis period in two sub periods, before Lehman and after Lehman.5
Hence, this study will contrast and compare the market co-movements: 1) pre-crisis vs. crisis and 2)
within the crisis, pre-Lehman vs. post-Lehman.

3. Data and methodology

3.1. Data

The database consists of daily Datastream stock market index prices for the United States and euro
area countries. We consider both a nancials sector index and a non-nancials market index (i.e. total

4
The announcement by Dubai World regarding its debt problems on 25 November 2009 is also linked to a reassessment of
sovereign debts worldwide. Unreported results using this break date show smaller differences in the estimated coefcients
indicating that Lehmans bankruptcy had a more dramatic impact on market expectations for sovereign debt at least for EMU
stock markets.
5
Dooley and Hutchison (2009) consider three sub periods of the nancial crisis by splitting the pre-Lehman crisis period into
two sub periods. Since they study developing countries their focus is slightly different.
522 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

1200 Lehman bankruptcy

1000
Greece

800
basis points

600

400

200
Germany
0
Feb-07 May-07 Aug-07 Nov-07 Feb-08 May-08 Aug-08 Nov-08 Feb-09 May-09 Aug-09 Nov-09 Feb-10 May-10 Aug-10

Fig. 4. Greek and German 5 year government bond CDS spreads. Source: Thomson Datastream.

market index excluding nancials). These data are available for all EMU countries in euros and for the
United States in US dollars. The data exhibit the familiar properties of nancial time series, in particular
non-normality and time varying volatility. We compute daily returns, calculated as log price changes,
for the North group, South group and Small countries using daily weights based on stock market
capitalization values.6
The euro-dollar exchange rate is measured in US dollars per euro. Therefore, an appreciation of the
euro is captured by an increase in the exchange rate. In the regressions we include the daily exchange
rate change, measured as the change between two trading days in US dollars per euro. For the crisis
period we also include the daily change in the difference between the 5 year Greek government bond
CDS spread and the 5 year German government bond CDS spread. This difference captures the
investors aversion against Greek government bonds relative to German government bonds.
Table 2 summarizes the daily mean return and volatility of both nancials and non-nancials
indexes during the two periods: 1) Pre-crisis (1/1/200326/2/2007) and 2) Crisis (27/2/200731/8/
2010), and the two sub periods within the nancial crisis 2a) Pre-Lehman (27/2/200714/9/2008) and
2b) Post-Lehman (15/9/200831/8/2010).
The average daily mean return exhibits a boom-bust pattern during the sample period 20032010
(see also the graphical representation in Fig. 1). It is positive for all country groups during the pre-crisis
period and negative during the crisis period. However, there are some differences between groups. The
small countries have the least negative return for both nancials and non-nancials during the crisis,
whereas these indexes decrease fastest for the North group. During the crisis, this pattern is identical
for both the pre-Lehman and post-Lehman crisis sub periods.
The striking differences in volatility between the pre-crisis period and the different crisis sub
periods are illustrative of the increasing uncertainty during the nancial crisis. Low volatility charac-
terizes the tranquil markets pre-crisis. During the pre-Lehman crisis period volatility doubles or triples
for all groups compared to the pre-crisis period. The increase in volatility is generally stronger for
nancials, compared to non-nancials. Tension in stock markets increases further after the collapse of
Lehman Brothers. In particular, volatility skyrockets for nancials in the North and South groups. These
large swings in time varying volatilities need to be taken into account when deciding on the empirical
methodology.

6
These weights and a full set of empirical results for individual EMU countries are available upon request.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 523

Table 1
5 year government bond CDS spreads on key dates of the nancial crises.

CDS spread (basis points)

North
Date Finland Germany Netherlands France Austria Belgium
27 February 2007 N/A 3 2 2 2 2
15 September 2008 13 10 12 14 13 23
31 August 2010 30 44 49 81 93 132
South
Date Italy Spain Portugal Ireland Greece
27 February 2007 7 3 4 11 6
15 September 2008 47 44 45 34 57
31 August 2010 231 246 346 352 926

Source: Thomson Datastream.

3.2. Empirical methodology

There exist several theoretical denitions and empirical methodologies to test for the spreading or
transmission of nancial shocks. In this paper we follow a denition employed in the contagion
literature according to which: (Shift-) contagion occurs when the transmission channel intensies or,
more generally, changes after a shock in one market. (Pericoli and Sbracia, 2003).7 This implies that we
model the transmission channel during tranquil times, i.e. during the non-crisis period, and test if
shocks are transmitted more strongly during the 20072010 crises.
Scholars use several methodologies when dealing with daily data, e.g. GARCH models (Hamao et al.,
1990; Engle et al., 1990), correlations (Forbes and Rigobon, 2002), logit models (Bae et al., 2003) and
factor models (Dungey and Martin, 2007).8 The large swings in stock market volatility before and
during the crisis, but also within the crisis sub periods, warrant an empirical strategy able to accom-
modate the time varying nature of volatility (see Fig. 2). Since GARCH models are able to capture these
changes successfully we follow the empirical strategy in the spirit of GARCH and factor models.
During tranquil times we postulate the following model:

ri;k;t ai;k li;k ri;k;t1 bi;k rus;k;t gi;k rus;k;t1 di;k DEt i;k;t (1)

where ri,k,t represents the day t equity return of country group i (North, South or Small) for k being
either nancials or non-nancials, ri,k,t1 the one day lagged return of group i and index k, rus,k,t the
contemporaneous equity return in the US (either nancials or non-nancials), rus,k,t1 the one day
lagged return in the US and DEt the change in the euro-dollar exchange rate. Eq. (1) allows for a lagged
effect of returns to capture the persistence in stock markets. In contrast to latent factor models, we
assume that the factors are observed and exogenous for all three groups stock market indexes. In
addition, due to the non-synchronous trading hours we allow for a lagged effect from the US market.
The residual captures the specic news relevant for country group i, market k. It is allowed to be of
the GARCH(1,1) type:

s2i;k;t ui;k aVi;k 2i;k;t bVi;k s2i;k;t1 :

V
where aVi;k and bi;k are the coefcients of the volatility dynamics.
Eq. (1) captures the movements of country group i, market ks stock market returns during tranquil
times. However, during the crisis we allow the transmission channel to intensify. Therefore, we
introduce an indicator variable that is equal to zero in tranquil times (I(c 0) 0) and equal to one
during the crisis period (I(c 1) 1).

7
This denition has also been used among others by Forbes and Rigobon (2002).
8
For a detailed review we refer to Pericoli and Sbracia (2003) and Dungey and Martin (2007).
524 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

Table 2
Descriptive statistics of daily nancials and non-nancials index returns.

Period Mean Variance

Pre-crisis Crisis Crisis Crisis Pre-crisis Crisis Crisis Crisis

Sub period Pre-Lehman Post-Lehman Pre-Lehman Post-Lehman

Financials
North 0.09% 0.11% 0.14% 0.08% 0.68% 4.53% 1.86% 6.65%
South 0.08% 0.09% 0.12% 0.06% 1.03% 4.45% 2.30% 6.16%
Small 0.11% 0.07% 0.10% 0.05% 0.43% 1.78% 1.44% 2.05%
Non-Financials
North 0.08% 0.06% 0.08% 0.04% 0.49% 2.18% 1.20% 2.97%
South 0.07% 0.03% 0.05% 0.02% 0.79% 2.27% 1.22% 3.11%
Small 0.09% 0.02% 0.02% 0.01% 0.60% 1.63% 0.94% 2.18%

Note: The North, South and Small return series are calculated as a weighted average of the respective group countries nancials
or non-nancials indexes and the volatility is directly calculated from this generated return series. The North group consists of
Austria, Belgium, Finland, France, Germany and the Netherlands; the South group consists of Greece, Ireland, Italy, Portugal and
Spain; and the Small group of Cyprus, Luxembourg, Malta and Slovenia. Source: Thomson Datastream.

Put differently, we allow the coefcient on the factors to differ during the crisis period. Eq. (2)
species this model:
c
ri;k;t ai;k li;k ri;k;t1 bi;k rus;k;t bi;k Ic 1rus;k;t gi;k rus;k;t1 gci;k Ic 1rus;k;t1
di;k DEt di;k Ic 1DEt i;k;t
c
(2)
c c
where, I(c 1) is the indicator variable and bi;k , gci;k and di;k
are the crisis transmission coefcients. Note
that these coefcients need to be added on top of the normal coefcients to quantify the full crisis
effect. Therefore, crisis transmission is present whenever the crisis coefcients are signicantly
different from zero.9
Finally, we capture sovereign debt problems by log differences of the gap between Greek and
German CDS spreads. The Greek CDS spread is of prime concern since most attention is being directed
toward Greece during the sovereign debt crisis, at least during the time period we study. In Eq. (3) we
distinguish between the pre-Lehman and post-Lehman crisis sub periods.

aLeh
ri;k;t ai;k li;k ri;k;t1 bi;k rus;k;t bi;k IaLeh 1rus;k;t gi;k rus;k;t1
aLeh
i;k IaLeh 1rus;k;t1 di;k DEt di;k IaLeh 1DEt hi;k Dln CDSGRC;t
gaLeh



i;k IaLeh 1Dln CDSGRC;t  CDSDEU;t i;k;t
 CDSDEU;t haLeh (3)

The coefcient hi,k captures the coefcient on Greek CDS spreads during the crisis and haLeh i;k
the
additional effect during the post-Lehman sub period (i.e. it captures the sovereign debt risk post-
Lehman). The indicator variable is zero before the collapse of Lehman (I(aLeh) 0) 0) and one after
the collapse of Lehman Brothers (I(aLeh 1) 1). The pre-Lehman sub period is from 27-2-2007 until
14-9-2008, while the post-Lehman sub period runs from 15-9-2008 until 31-8-2010. Hence, in this set-
up we treat the pre-Lehman sub period as baseline and sovereign debt crisis transmission is present
when the post-Lehman coefcient of the respective variable is signicantly different from zero.
One important assumption in the above equations is the exogeneity of US returns. In order for this
to be valid we need to establish that US returns cause European returns, but that this relationship does
not hold in the other direction. Several arguments justify this choice. The US stock market is by far the
largest stock market measured in terms of capitalization. It accounts for about 40% of world stock

9
Didier et al. (2010) present a similar model set-up, but use both dummies for both the pre-crisis and crisis periods.
Quantitatively this is identical to our approach, but requires an F-test to test for coefcient differences. By using a dummy only
for the crisis period, we can directly infer coefcient differences.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 525

market capitalization. Individual EMU stock markets are quite a bit smaller than the US market. For
example, the largest EMU market is the French one and corresponds to only 5% of global market
capitalization. Empirical ndings reported by Ehrmann et al. (Forthcoming) indeed show that Euro-
pean returns only have a marginal impact on US returns, whereas US returns have a strong impact on
European results. Earlier research by Masih and Masih (2001) also nd similar results. In addition,
linear causality tests presented by Beine et al. (2008) conrm the leading impact of the US in global
stock markets. Finally, unreported Granger causality tests using the data in this paper provide further
supporting evidence for the exogeneity of US stock market returns.

4. Empirical ndings

4.1. The transmission of the nancial crisis

This section studies the dependence of EMU nancials and non-nancials stock indexes on
developments in the US before and during the 20072010 nancial crisis, while Section 4.2 discusses
the Greek sovereign debt crisis. Finally in Section 4.3 we analyze the role of the euro-dollar exchange
rate during both crises.
Table 3 presents the results when estimating Eq. (2) for the nancials (left panel) and non-nancials
(right panel) stock indexes in the three EMU country groups on US returns and the euro-dollar
exchange rate. For the nancials indexes the coefcient on the lagged return of the dependent vari-
able shows negative persistence for both North and South, but not for small countries. A negative sign
indicates that a good day is followed by a poorer one and vice versa.
When assessing the contemporaneous (day t) impact of US nancials on European nancials we
nd a large and signicant coefcient for North and South. However, the small EMU countries seem to
be relatively isolated from US events during tranquil times. This is in line with the results of Didier et al.
(2010) who nd that stock market illiquidity can explain a low degree of comovement with the US. The
North group seems to have the closest contemporaneous ties to the US, where a 1% increase in the US
nancials index results in a 0.5% increase in the North nancials index. During the crisis, links intensify
for small countries, but not for the other groups.
The one day lagged US return shows a similar picture. Dependence is signicant for all groups, but
small in magnitude for the small country group. During the crisis we observe only an increase for the
small country group, but this increase is rather modest.
The right panel in Table 3 presents the dependence patterns for non-nancials before and during
the crisis in the same fashion. The coefcients on the lagged return are more negative for all groups,
indicating a stronger mean reversion tendency for non-nancials. Perhaps surprising, the contempo-
raneous dependence of European non-nancials on US non-nancials is even larger than the depen-
dence of European nancials on US nancials across all three groups. In contrast to nancials, the
dependence of non-nancials increases signicantly across all groups during the crisis. Hence, for non-
nancials there is strong evidence in favor of crisis transmission.
The lagged dependence on the US is also positive and highly signicant for most European coun-
tries. However, we only observe an increase for small countries during the crisis. Hence, the crisis
transmission mechanism is strongest for contemporaneous dependence on the US. Overall, total
dependence of EMU non-nancials on US developments increases sharply during the crisis.10

4.2. The transmission of the sovereign debt crisis

The economic and nancial crisis does not only affect the European nancial sector, but has far
reaching consequences for scal policy. First, governments inject large amounts of equity capital in
banks and other nancial institutions in order to stabilize the nancial system, resulting in rising
sovereign debt levels. In addition, many governments provide implicit guarantees on banks assets,

10
Adding additional control variables such as t-bill rate changes or oil price changes does not alter these results. These
robustness checks are available upon request.
 526
T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533
Table 3
European nancials and non-nancials before and during the crisis.

Financials Non-nancials

North South Small North South Small

Lagged index return 0.079*** (0.025) 0.079*** (0.025) 0.026 (0.025) 0.194*** (0.022) 0.182*** (0.024) 0.121*** (0.024)
US return 0.484*** (0.034) 0.369*** (0.030) 0.015 (0.018) 0.543*** (0.031) 0.410*** (0.026) 0.091*** (0.030)
US return*crisis dummy 0.027 (0.043) 0.047 (0.040) 0.057** (0.027) 0.145*** (0.043) 0.181*** (0.041) 0.122*** (0.047)
Lagged US return 0.334*** (0.031) 0.235*** (0.029) 0.084*** (0.020) 0.382*** (0.029) 0.279*** (0.026) 0.258*** (0.032)
Lagged US return*crisis dummy 0.026 (0.038) 0.037 (0.039) 0.081*** (0.030) 0.020 (0.040) 0.057 (0.040) 0.002 (0.045)
DEuro-dollar exchange rate 0.238*** (0.030) 0.187*** (0.028) 0.053** (0.021) 0.219*** (0.028) 0.147*** (0.025) 0.123*** (0.031)
DEuro-dollar exchange rate*crisis dummy 0.503*** (0.065) 0.466*** (0.076) 0.192*** (0.048) 0.328*** (0.046) 0.293*** (0.048) 0.108* (0.058)

Note: Equation (2) is estimated using a GARCH (1,1) model on stock market nancials and non-nancials indexes for the pre-crisis and crisis periods (1/1/200331/8/2010).
*, ** and *** denote signicance at the 10%, 5% and 1%, respectively. The independent variables include the lagged return of the respective groups nancials(non-nancials) index, and the
US nancials (non-nancials) index return and its lagged value. Finally, we include the change in the euro-dollar exchange rate. Note that the full crisis coefcient is the coefcient during
the tranquil period plus the crisis dummy coefcient, e.g. the full crisis coefcient on US contemporaneous returns for North nancials is 0.4840.027 0.457.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 527

which increase the risk prole of the governments. Second, government decits increase as well due to
the countercyclical scal policy measures induced by the global economic recession.
Table 4 presents the results when estimating Eq. (3) for the nancials (left panel) and non-nancials
(right panel) indexes in the three country groups. Let us concentrate rst on the nancials. The coef-
cients regarding the contemporaneous and lagged dependence on US nancials are similar to those
reported in Table 3. It is also worth noting that the contemporaneous dependence remains the same for
all groups during the entire crisis (before and after the collapse of Lehman). However, small countries
face smaller lagged dependence on US nancials after the collapse of Lehman.
Before Lehman, changes in the difference between Greek and German CDS spreads affect only small
country nancials. However, the magnitude is economically negligible. Fig. 4 shows that Greek and
German CDS spreads move quite closely in the pre-Lehman sub period. The correlation between the
two variables is 0.87 during this sub period. Hence, in the pre-Lehman sub period the Greek CDS spread
seems to follow a common European default premium increase. However, during the post-Lehman sub
period the correlation between Greek and German CDS spreads drops to 0.31 and the Greek CDS spread
captures more strongly the pure Greek debt problems. An increase in Greek CDS spreads has a negative
effect on stock prices, since rising CDS spreads signal economic hardship. Therefore, we expect
a negative sign on the coefcients of hi,k and haLeh
i;k
.
The mechanism at work is linked to banks exposure to Greek government debt and may be
described as follows. When the Greek CDS spread increases, banks bonds lose value, which in turn
justies a price decrease of their stock prices. In fact, new investors in the bank demand a higher
expected return for the increased riskiness of the bank, at the expense of a loss for existing
shareholders.
The results in Table 4 conrm this mechanism. After the collapse of Lehman all three country groups
have a signicantly negative loading on changes in Greek CDS spreads. The South group faces the
largest absolute coefcients, which can be explained by the higher perceived sovereign default risk for
these countries following Lehmans collapse. However, a strong increase is also visible for the North
group. This increase is probably not so much linked to the North governments default perceptions but
rather to their nancial sectors exposure to Greek debt. Indeed, foreign investors hold over 75 percent
of Greek government debt. In fact, the EMU countries most exposed are from all three groups: Belgium,
Cyprus, France, Germany, Greece, Portugal and the Netherlands (Blundell-Wignall and Slovik, 2010). It
is no surprise then that developments in Greece strongly affect nancials across the EMU. In sum, even
though the sovereign debt crisis affects the group of EMU countries most exposed to a default risk,
nancials across other EMU countries are also affected due to the many nancial interlinkages.
Next we discuss the result for the non-nancials presented in the right panel of Table 4. Again,
similar to the nancials, we observe an even larger degree of contemporaneous and lagged dependence
on US developments. Before the collapse of Lehman we document that both North and South non-
nancials have a signicantly negative loading on changes in the difference between the Greek and
German CDS spreads, although its magnitude is small. This can be explained by the high correlation of
Greek CDS spreads and German CDS spreads before the collapse of Lehman. It is most likely that we
capture the non-nancials sensitivity to aggregate EMU sovereign debt developments, which mainly
involve rescue packages for the nancial sector. Similar to the nancials we do observe increases in
dependence on Greek CDS spreads post-Lehman for both the North and South country groups.
However, the magnitude of the increase is much smaller compared to the nancials.
Table 5 provides key gures on the scal stance of EMU governments. Gross debt to GDP ratios
increase across all EMU countries during 20072011, but with different magnitudes. The increases in
Ireland (62.3%), Greece (38.2%) and Spain (36.3%) are the largest, resulting in several debt rating
downgrades and, in the case of Greece, Ireland and Portugal even an inability to borrow on the nancial
markets.11
When considering debt sustainability, we need to distinguish between the different factors
increasing the debt. First, the debt can increase due to the primary balance, which is fully controlled by

11
Ferreira and Gama (2007) provide evidence on the transmission of sovereign debt rating downgrades to international stock
markets.
 528
T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533
Table 4
European nancials and non-nancials before and after the collapse of Lehman Brothers.

Financials Non-nancials

North South Small North South Small

Lagged index return 0.079** (0.033) 0.079** (0.033) 0.006 (0.038) 0.216*** (0.033) 0.192*** (0.034) 0.193*** (0.038)
US return 0.419*** (0.041) 0.340*** (0.036) 0.067 (0.043) 0.625*** (0.042) 0.548*** (0.045) 0.154*** (0.057)
US return*post-Lehman 0.014 (0.052) 0.060 (0.051) 0.015 (0.047) 0.064 (0.056) 0.011 (0.057) 0.044 (0.064)
Lagged US return 0.316*** (0.039) 0.273*** (0.033) 0.304*** (0.038) 0.453*** (0.048) 0.409*** (0.048) 0.265*** (0.062)
Lagged US return*post-Lehman 0.028 (0.047) 0.035 (0.045) 0.197*** (0.043) 0.119** (0.057) 0.111* (0.057) 0.020 (0.073)
DEuro-dollar exchange rate 0.028 (0.083) 0.061 (0.077) 0.098 (0.101) 0.062 (0.061) 0.027 (0.062) 0.120 (0.075)
DEuro-dollar exchange rate*post-Lehman 0.358*** (0.102) 0.476*** (0.111) 0.000 (0.112) 0.187** (0.073) 0.207*** (0.077) 0.120 (0.099)
Dlog(CDS spread Greece Germany) 0.002 (0.002) 0.003 (0.002) 0.004** (0.002) 0.004** (0.002) 0.006** (0.002) 0.002 (0.002)
Dlog(CDS spread Greece Germany)*post-Lehman 0.066*** (0.017) 0.108*** (0.022) 0.053*** (0.010) 0.025*** (0.008) 0.037*** (0.010) 0.007 (0.011)

Note: Equation (3) is estimated using a GARCH (1,1) model on stock market nancials and non-nancials indexes for the pre-Lehman (27/2/200714/9/2008) and post-Lehman sub periods
(15/9/200831/8/2010). *, ** and *** denote signicance at the 10%, 5% and 1%, respectively. The independent variables include the lagged return of the respective groups nancials(non-
nancials) index, and the US nancials (non-nancials) index return and its lagged value. We include as well the change in the euro-dollar exchange rate and the change in the log of Greek
minus German CDS spreads. Note that the full post-Lehman coefcient is the coefcient during the crisis plus the Lehman dummy coefcient, e.g. the full post-Lehman coefcient on US
contemporaneous returns for North nancials is 0.419 0.014 0.433.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 529

Table 5
Composition of changes in the government debt ratio in EMU member states (% of GDP) and post-Lehman sovereign debt risk
haLeh
i;k coefcients for individual EMU countries.

Country Gross debt ratio Change in Source of change in debt Post Lehman sovereign
debt ratio ratio 20072011: debt risk haLeh
i;k

2007 2011 20072011 Primary Interest & Stock-ow Financials Non-nancials

balance growth contribution adjustment

Austria 59.5 72.9 13.4 2.2 6.2 5 0.08 0.03

Belgium 84.2 100.9 16.7 2.2 8.5 6 0.08 0.03
Cyprus 58.3 67.6 9.3 9.1 4 3.8 0.17 0.03
Finland 35.2 54.9 19.7 0.3 3.4 16.5 0.08 0.06
France 63.8 88.6 24.8 15.6 5.9 3.4 0.08 0.03
Germany 65 81.6 16.5 2.2 8.4 5.9 0.05 0.02
Greece 95.7 133.9 38.2 19.7 15 3.5 0.26 0.09
Ireland 25 87.3 62.3 35.5 14.6 12.2 0.16 0.03
Italy 103.5 118.9 15.5 1 14.8 1.7 0.09 0.03
Luxembourg 6.7 23.6 16.9 3.4 0.5 13.1 0.02 0.00
Malta 61.9 72.5 10.6 3.2 5.7 1.7 0.01 0.00
Netherlands 45.5 69.6 24.1 7 5.9 11.2 0.07 0.03
Portugal 63.6 91.1 27.5 16.4 8.9 2.2 0.12 0.05
Slovenia 23.4 45.4 22 12.2 4.3 5.5 0.02 0.01
Spain 36.2 72.5 36.3 25.7 7.2 3.4 0.11 0.04

Note: The post-Lehman sovereign debt risk haLeh i;k

coefcients result from estimating Equation (3) for each individual EMU
county. Detailed estimation results for the individual EMU countries are available upon request. Source: European Commission
Spring 2010 Economic Forecast and authors own calculations.

the government. The second factor is due to interest and growth contributions. This is not directly
controlled by the government since it depends on previous governments expenses and current
economic conditions. Finally, we consider stock-ow adjustments, which are not direct expenses, but
can be viewed more as investments since they increase government assets. These stock-ow adjust-
ments are very important during the current crisis due to the bank bailouts.12 The countries with the
largest increase in debt due to the primary balance and interest & growth contribution are Portugal,
Ireland, Greece and Spain.
Table 5 also reports the haLeh
i;k
coefcients when estimating Eq. (3) for individual EMU countries. A
full set of individual country results is available upon request. We investigate if these coefcients are
related to the debt situation in the specic country. The scatter plots in Fig. 5 show a negative rela-
tionship between the total debt change and the individual country haLehi;k
coefcients for both nancials
and non-nancials.
Next, Table 6 presents Spearman rank correlation coefcients for the haLeh
i;k
coefcients vis--vis the
three debt components mentioned in Table 5. This will help to shed light on which source of the
government debt increase is associated with the increased dependence on the Greek debt problems.
First, for nancials the rank correlation coefcient is quite large and negative for the primary balance
suggesting that nancials stock prices react signicantly negative to a large increase in government
spending. While the rank correlation is negative as well for non-nancials, it is not signicant.
Second, for both nancials and non-nancials the rank correlation coefcient is signicantly
negative for interest and growth contributions. In fact, stock markets perceive a large government debt,
with high interest expenses, as a drag on economic growth. Moreover, slower economic growth hurts

12
Consider Luxembourg which has the smallest debt to GDP ratio at the start of the crisis. The main reason for its debt
increase is stock-ow adjustments, associated with the support to Fortis and Dexia. Ireland also started with a low debt level in
2007 and its stock-ow adjustments are similar in size to those of Luxembourg due to the bailout of Anglo-Irish and other
nancial institutions. However, contrary to Luxembourg the Irish economy contracted strongly and primary decits increased
fast. Hence, the part of Irish debt which is spending instead of investment is much larger than in Luxembourg. Therefore, this
debt will have a permanent nature and has to be renanced when governments do not run primary surpluses.
530 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

0.05 0.02
Post Lehman sovereign debt risk

0.00

Post Lehman sovereign debt risk

0.00
0 10 20 30 40 50 60 70
0 10 20 30 40 50 60 70
-0.05
-0.02

-0.10
-0.04
-0.15

-0.06
-0.20

-0.25 -0.08

-0.30 -0.10
Change in gross debt ratio 2007-2011 (%) Change in gross debt ratio 2007-2011 (%)

Fig. 5. Individual EMU countries post-Lehman sovereign debt risk (haLeh

i;k
coefcients) for nancials (left gure) and non-nancials
(right gure). Source: Authors own calculations. See Table 5.

company prots, since taxes need to be raised in the future. Hence, since stock prices reect future
expectations, stock markets in countries with high interest expenses and slow growth perform worse.
Finally, the stock-ow adjustments do not exhibit a signicant relationship with the coefcients on
the change of the difference between Greek and German CDS spreads. There are even some signs of
a positive effect for nancials, most likely because governments bailed out nancials. These bailouts are
naturally good news for the nancial sectors equity holders. However, future prospects of the economy
will be harmed when these stock-ow adjustments create large losses for governments.
Summarizing the results, the transmission of the nancial crisis to the EMU nancials sector in the
early stages of the crisis is largely contained possibly due to (expectations of) decisive European actions to
protect the vulnerable institutions and to the inherent insulation of other institutions to the mortgage
crisis. Expectations of a strong EMU action were apparently shattered after the collapse of Lehman
resulting in a serious transmission of the Greek sovereign debt crisis to the nancials of most EMU
countries. Less attention or ability to protecting the non-nancial sector following the mortgage crisis, on
the other hand, can probably explain why that crisis was transmitted rather strongly to this sector in EMU.
On the positive side the non-nancial sector seems to be less affected by the sovereign debt crisis so far.

4.3. The effects of the euro-dollar exchange rate during the crises

The theoretical literature does not agree on the sign of the relationship between stock prices and
exchange rates. One part of the literature argues that developments in the current account determine
exchange rate changes (See e.g. Dornbusch and Fisher (1980)). According to the harmful to exports
hypothesis exchange rate movements have an effect on rm competitiveness, which affects future
protability and consequently stock prices. A more expensive currency makes European products more
expensive, results in decreasing exports and, consequently, protability.13 So, this theory predicts
a negative relationship between stock prices and exchange rates (i.e., exchange rate appreciations
coincide with negative stock market returns).
The second strand starts from portfolio balance models (Branson (1983) and Frankel (1983)). These
models view exchange rate determination as a variable equating the supply and demand of nancial
assets. Hence, portfolio balance models predict a positive relationship between stock prices and
exchange rates (exchange rate appreciations are correlated with positive stock market returns). A stock
price increase increases the value of the equity market, which is associated with an exchange rate
appreciation. This view can be termed the signal of economic strength hypothesis.

13
This also holds for multinationals. A subsidiary of a European rm operating in the US generates its prots in US dollars, but
then converted in euros; the protability of the European multinational is smaller.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 531

Table 6
Correlations between changes in government debt ratios and the post-Lehman sovereign risk coefcients.

Financials Non-nancials

Source of change in debt ratio 20072011 Spearman p-value Spearman p-value

Primary balance 0.49 0.07 0.29 0.29

Interest & growth contribution 0.56 0.03 0.51 0.05
Stock-ow adjustment 0.31 0.27 0.00 0.99

Note: The Spearman rank correlation coefcients are calculated from the numbers in Table 5.

During tranquil times the coefcient of European nancials on euro-dollar exchange rate changes is
in line with the predictions from the harmful to exports hypothesis, as shown by the signicantly
negative coefcient for all groups (see Table 3). The strongest effect is for the North countries and the
weakest effect for small countries. For the North and South country groups the effect is also
economically large. A strengthening of the euro by 1 US dollar cent is associated with a 0.2% lower
return on the nancials index. These results are in line with the ndings of Dunne et al. (2010), who
nd a negative coefcient on exchange rates for the French stock market index during 19992006.
During the crisis, however, this relationship is fundamentally different. All country groups have
a positive crisis coefcient on exchange rate changes and this coefcient is large enough to turn the
overall effect of the exchange rate from negative to positive. For example, for North countries, a 1 US
dollar cent appreciation of the euro-dollar exchange rate results in an increase of about 0.3% in stock
prices during the crisis. In sum, the role of a strong euro during the crisis period appears to have
changed from harmful to exports to a signal of economic strength.
For non-nancials the effect of the euro-dollar exchange rate between tranquil times and the crisis
is as striking as it is for the nancials (see right panel Table 3). All country groups have signicantly
negative coefcients on exchange rate changes before the crisis, whereas the crisis coefcient on
exchange rate changes is positive and signicant for virtually all countries. Again, the sign reverses
during the crisis, implying that euro appreciations lift European non-nancial equity prices. This shows
that the relationship between exchange rates and stock prices is time varying and depends on what the
markets perceive as normal or crisis periods.
Table 4 shows that for both nancials and non-nancials the reversal of the exchange rate effect
described above takes place during the crisis period. First, the coefcient on exchange rate changes is
insignicant in the sub period up to the collapse of Lehman for all country groups. However, the big
turn takes place after the collapse of Lehman, where the signs of both the North and South country
groups turn signicantly positive. However, small countries are rather insensitive to exchange rate
movements.
The trigger point for the reversal of the exchange rate effect can thus be identied with Lehmans
collapse. This single event underlies the dramatic impacts when a too big to fail nancial institution
was allowed to fail. It might be expected that Lehmans collapse would radically change markets
default expectations. Perhaps it is not surprising then that in such more pessimistic expectations
a signal of economic strength from a strong euro would override any concerns that it might be
harmful to exports.
Summarizing the results, we nd that the relationship between exchange rate movements and
stock market returns changes from negative before the crisis, neutral during the rst part of the crisis
and positive after the collapse of Lehman. This holds for both North and South countries, but is not the
case for small countries which appear to be less integrated with the international nancial system. As
the depth of the nancial crisis and subsequent recession becomes obvious, an appreciation of the euro
is more and more a sign of condence in the economic prospects of the euro area.

5. Conclusion

This study uses daily data on stock market indexes for the United States and 15 euro area countries
to test for the presence of transmission of the 20072010 nancial and sovereign debt crises. We
consider both a nancials sector index and a non-nancials market index (i.e. total market index
532 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533

excluding nancials). After splitting the EMU into three country groups, North, South and Small, we
explain daily stock market returns in each group using a GARCH model. We include both the
contemporaneous and lagged US return to capture the dependence of EMU markets on developments
in the US. Indicator variables allow the coefcients to change during the pre-crisis vs. crisis periods and
pre-Lehman vs. post-Lehman sub periods.
There is strong evidence of crisis transmission from US non-nancials to European non-nancials,
whereas the increase in dependence of European nancials on US nancials is rather limited. Following
the collapse of Lehman Brothers nancials become much more dependent on changes in Greek CDS
spreads compared to the pre-Lehman sub period. However, the increase is modest for non-nancials.
The relationship between euro-dollar exchange rate changes and stock price returns changes from
negative before the crisis, neutral during the rst part of the crisis to positive after the collapse of
Lehman. These results hold both for the North and South country groups, whereas there is no relation
for the smallest EMU countries.
These ndings have important implications for the ongoing debate about how to reform the
nancial system so as to mitigate systemic risk in the future. First, the importance of the non-nancial
crisis transmission channels need to be taken into account and further investigated. Second, exchange
rates and stock market returns are correlated but this relationship seems to depend crucially on the
overall market sentiment. Research on the time varying nature of this relationship is a fruitful area for
further research. Finally, this study shows the serious impact on market risk perceptions from a single
but dramatic default event. Hence, policies to reduce the failure possibility of institutions deemed too
big to fail become increasingly important. This studys ndings are not only relevant for EMU coun-
tries. For example, future research may very well devote its attention to the role of California state debt
for the Federal US governments debt position.

Acknowledgements

We are grateful to Viral Acharya, Ian Cooper, Robert G. King, Sergio Mayordomo, Deyan Radev,
seminar participants at De Nederlandsche Bank, the XVI Spring Meeting of Young Economists in
Groningen, the Netherlands, the 15th Annual International Conference on Macroeconomic Analysis
and International Finance in Rethymnon, Greece, the 9th INFINITI Conference on International Finance
in Dublin, Ireland, an anonymous referee and the Editors, George Kouretas and Athanasios Papado-
poulos for useful comments and constructive discussions. This paper presents the authors personal
views and opinions. They do not necessarily coincide with those of de Nederlandsche Bank. All errors
are our own.

References

Attinasi, M.G., Checherita, C., Nickel, C., 2009. What explains the surge in Euro Area sovereign spreads during the nancial crisis
of 200709? ECB Working Paper 1131.
Bae, K.-H., Karolyi, G.A., Stulz, R.M., 2003. A new approach to measuring nancial contagion. The Review of Financial Studies 16
(3), 717763.
Beine, M., Capelle-Blancard, G., Raymond, H., 2008. International nonlinear causality between stock markets. The European
Journal of Finance 14 (8), 663686.
Blundell-Wignall, A., Slovik, P., 2010. The EU stress test and sovereign debt exposures. OECD Working Papers on Finance,
Insurance and Private Pensions, No. 4.
Branson, W.H., 1983. Macroeconomic Determinants of Real Exchange Rate Risk. In: Herring., R.J. (Ed.), Managing Foreign
Exchange Risk. Cambridge University Press, Cambridge.
Didier, T., Love, I., Martinez Peria, M. S., 2010. What explains stock markets vulnerability to the 20072008 crisis? World Bank
Policy Research Working Paper 5224.
Dooley, M., Hutchison, M., 2009. Transmission of the U.S. subprime crisis to emerging markets: evidence on the decoupling
recoupling hypothesis. Journal of International Money and Finance 28 (8), 13311349.
Dornbusch, R., Fisher, S., 1980. Exchange rates and the current account. American Economic Review 70, 960971.
Dungey, M., Martin, V.L., 2007. Unraveling nancial market linkages during crises. Journal of Applied Econometrics 22, 89119.
Dunne, P., Hau, H., Moore, M., 2010. International order ows: explaining equity and exchange rate returns. Journal of Inter-
national Money and Finance 29, 358386.
Ehrmann, M. Fratzscher, M. and Rigobon, R. Stocks, bonds, money markets and exchange rates: measuring international
nancial transmission. Journal of Applied Econometrics, Forthcoming.
Engle, R.F., Ito, T., Lin, W., 1990. Meteor showers or heat waves? Heteroskedastic intra-daily volatility in the foreign exchange
market. Econometrica 58, 525542.
 T. Grammatikos, R. Vermeulen / Journal of International Money and Finance 31 (2012) 517533 533

Ferreira, M.A., Gama, P.M., 2007. Does sovereign debt ratings news spill over to international stock markets? Journal of Banking
& Finance 31, 31623182.
Forbes, K.J., Rigobon, R., 2002. No contagion, only interdependence: measuring stock market comovements. The Journal of
Finance 57 (5), 22232261.
Frankel, J., 1983. Monetary and portfolio balance models of exchange rate determination. In: Bhandari, J., Putnam, B. (Eds.),
Economic Interdependence and Flexible Exchange Rates. MIT Press, Cambridge, MA, pp. 84114.
Hamao, Y., Masulis, R.W., Ng, V., 1990. Correlations in price changes and volatility across international stock markets. Review of
Financial Studies 3, 281307.
Masih, R., Masih, A.M.M., 2001. Long and short term dynamic causal transmission amongst international stock markets. Journal
of International Money and Finance 20 (4), 563587.
Pericoli, M., Sbracia, M., 2003. A primer on nancial contagion. Journal of Economic Surveys 17 (4), 571608.
Sgherri, S., Zoli, E., 2009. Euro Area sovereign risk during the crisis. IMF Working Paper WP/09/222.
Spiegel Online. 2010. European Central Bank president Jean-Claude Trichet: A quantum leap in governance of the Euro Zone is
needed. Spiegel Online, 15 May 2010.