Sovereign Spreads in The Euro Area. Which Prospects For A Eurobond?
Sovereign Spreads in The Euro Area. Which Prospects For A Eurobond?
Sovereign Spreads in The Euro Area. Which Prospects For A Eurobond?
The organisers would like to thank the National Bank of Poland for their support.
The views expressed in this paper are those of the author(s) and not those of the funding organization(s).
1. INTRODUCTION
The dramatic increase in interest rate differentials displayed by the euro sovereign debt
crisis, has led economists and policy makers to reconsider the possibility of a European
government bond jointly issued and guaranteed by euro-area Member States. The idea
that a common Eurobond could offer relief to Member States with weaker fiscal
fundamentals, like Italy and Spain, and help solve the debt crisis has emerged in the
policy debate and in the media in 2010 and, finally, gained strength in the summer of
2011.
Issuance of a common Eurobond, first analyzed in the Giovannini Group Report
(2000), was originally viewed as a strong form of debt management cooperation with the
potential of promoting further market integration and greater liquidity. In the wake of the
US financial crisis, Eurobond proposals have stressed that a common bond would satisfy
the global demand for a risk-free asset and better compete with US Treasuries for the
global financial flows in search of a safe-haven, thereby strengthening the use of the
euro as a reserve currency (see e.g. Gros and Micossi 2009, Mayordomo et al. 2009, De
Grauwe and Moesen 2009). The euro debt crisis has generated sizeable and highly
volatile yield spreads of Member States bonds on German Bunds. In light of this
evidence, a new argument has gained strength: that Eurobonds could provide better
1
We thank the editor, Phillip Lane, two anonymous referees, Bruno Usai and Yian Ma of Mako Investments
for their helpful comments.
market access to weaker Member States, by insulating these countries from financial
contagion, and lower the risk of crisis propagation (see e.g. Delpla and von Weizsacker
2010, Jones 2010, Juncker and Tremonti 2010, ELEC 2010). On the other hand, Issing
(2009) has argued that a common Eurobond is not such a good idea as the medicine of
a common Eurobond would not cure the fiscal problems of its weakest members, but
would instead prolong their reliance on debt, thus taking away the effect of market
discipline on their fiscal policies.
In this paper, we first investigate the determinants of interest-rate differentials between
euro-area Member States in order to provide the relevant stylized facts on the relative
importance of liquidity and default risk. As we find that default risk is the main driver of
yield spreads, suggesting small gains from greater liquidity, in the remaining part of the
paper we focus on arguments for and against Eurobonds based on their effects on default
risk and its redistribution across participating issuers.
A Eurobond jointly guaranteed by euro-area Member States by ensuring market access
(and better borrowing conditions) to weaker Member States does not only reduce their
risk of default but also the risk of crisis propagation to States with relatively better
fundamentals either from contagion or interdependences in the real and financial sectors.
In particular, a Eurobond can insulate more fiscally responsible States against the effect
of contagion, i.e. a rise in yield spreads (and thus in funding costs) due to a shift in
market sentiment. On the other hand, issuance of a common Eurobonds could lower
incentives for fiscal adjustment, since it prevents financial market from exerting their
disciplinary role through higher interest rates. Hence, the economic rationale for a
Eurobond program very much depends on whether yield spreads reflect contagion or the
markets efficient assessment of fiscal fundamentals.
To shed light on the relative role of fiscal fundamentals and contagion in the pricing of
default risk in yield spreads, and thus in the propagation of the euro debt crisis, we
estimate a Global VAR of 10-year yield spreads on Bunds. We find that fiscal
fundamentals matter in the pricing of default risk but only as they interact with other
countries yield spreads; i.e. with the global risk that the market perceives. More
important, the impact of this global risk variable is not constant over time, a clear sign of
contagion driven by shifts in market sentiment. This evidence points to a discontinuity in
the disciplinary role of financial markets. If markets can stay irrational longer than a
country can stay solvent, then the role of yield spreads on national bonds as a fiscal
discipline device is considerably weakened, and issuing Eurobonds can be economically
justified.
In light of this evidence, we then discuss and provide new evidence on the other
benefits and costs of a Eurobond program. Contrary to the safe-haven argument, we
find that German Bunds do not suffer from a lack of liquidity, and thus higher costs,
compared to US Treasuries despite a smaller market size. Moreover, insuring the default
risk of Member States with weaker fundamentals is not without costs for safer Member
States: their expected liabilities will increase and, if the Eurobond did not reach the same
credit quality as German Bunds, an event that we cannot rule out, their borrowing
conditions would also worsen.
This suggests that the political opposition to a common issuance program is well
motivated and that a Eurobond will never be issued without a renewed aim for a stronger
EU political union.
The rest of this paper is organized as follows. In Section 2 we present the relevant
stylized facts. Section 3 evaluates the trade-off between market discipline and crisis
prevention for a Eurobond backed by joint guarantees in the light of the empirical results
of a new nonlinear Global VAR model for yield spreads in the euro area. Section 4 is
then devoted to assess the feasibility of a Eurobond program. Finally, Section 5 presents
the policy conclusions of our analysis.
2. STYLIZED FACTS
To assess the potential for a European government bond issued by euro-area Member
States, we first examine the degree of integration in the European government bond
market and its determinants. The relevant stylized facts are on the relative importance of
liquidity and default risk premia in determining yield spreads in the euro area, and on the
link between fiscal fundamentals and default risk premia priced in yield spreads.
and thus small markets must compensate investors with a liquidity premium. Before the
introduction of the Euro, also expectations of exchange rate fluctuations and different tax
treatment of bonds issued by different countries were relevant. Different tax treatments
were eliminated or reduced to a negligible level during the course of the 90s. The
introduction of the Euro in January 1999 virtually eliminated the expectations on
exchange rate fluctuations, at least until the most recent events that might have induced
some positive probability on the event of the collapse of the EMU.2
The availability of Credit Default Swaps (CDS) for the more recent part of the sample
allows us to measure the default-risk premium component. A CDS is a swap contract in
which the protection buyer of the CDS makes a series of premium payments to the
protection seller and, in exchange, receives a payoff if the bond goes into default. The
difference between a CDS on a Member State bond and the CDS on the German Bund of
the same maturity is a measure of the default risk premium of that State relative to
Germany. Note that, as clearly discussed in Sturzenegger and Zettelmeyer (2006), CDS
is direct measure of the default risk but not of the probability of default, as the price of a
CDS depends both on the probability of default and on the expected recovery value of
the defaulted bond. Moreover, such measure is not perfect; CDS differentials might also
reflect the different liquidity of different sovereign CDSs, as well as counterparty risk
(i.e. the risk that the protection seller of the CDS is not able to honor her obligation when
the bond goes into default). Their imperfections notwithstanding, CDS differentials
provide us with an interesting benchmark to assess what are the main factors driving
yield differentials.
Figures 2 and 3 report interest-rate differentials for euro-area Member States (blue
line) i.e. the spreads of 10-year government bond yields on German Bund yields
along with the associated CDS spreads (red line) and the residual non-default component
(black line). We group the yield spreads on Bunds and the associated CDS into high
yielders (Figure 2) and low yielders (Figure 3).3
Insert Figure 2-3 about here
The following facts emerge from the data:
i)
There is a clear tendency of all spreads on Bunds in the euro-area to comove,
but the nature of the comovement is not constant over time. The CDS spread,
i.e. the default risk component of the yield spread, accounts for virtually the
entire differential (and its variability) in the case of high yielders over the
2
For the pre-EMU period expected exchange rate depreciation and risk can be directly identified via the
difference between the 10-year Fixed Interest Rate Swaps in local currency by the European countries, as these
spreads are immune from sovereign default and liquidity risk. The data show that most of the pre-EMU
fluctuations in spreads of high yielders on Bunds are attributable to this component. Since the inception of the
Euro 10-year Fixed interest rate swaps differential among Euro area member States are equal to zero by
definition.
3
We do not present data for Cyprus, Malta, Slovenia and Slovakia because times series are not long enough
given the short spell of time these countries have spent within the EMU. Data for Luxembourg are also not
reported.
ii)
iii)
iv)
There is an important fact about the comovements of interest-rate spreads in the euro
area: their interdependence is not constant over time. To illustrate this phenomenon we
consider two high-yielders, Greece and Italy, and one low-yielder, Finland. We report in
Figure 4 the Greek, the Italian and the Finnish 10-year spreads on Bunds along with the
spread between yields on US Baa and Aaa corporate bonds, a variable often used to
describe the market attitude toward risk. We consider the full sample 2003-2010 and
three subsamples, the low-volatility period (2003-2007), the financial crisis (May 2007August 2009) and the Greek debt crisis (September 2009-July 2010).
Insert Figure 4 about here
A changing correlation pattern clearly emerges from the data. Over the low-volatility
period, Italy and Greece are placed in the same class of risk by the market and their
spreads on Bunds are very highly correlated, despite the fact that the two markets are
very different in terms of their size and therefore their liquidity. During the financial
crisis the credit risk of the two high-yielders diversify. In fact, both the Italian and the
Greek spread positively react to the increase in the Baa-Aaa spread but their response is
different; the Greek spread widens up to 300 basis points, while the Italian spread peaks
at 150. Interestingly, during the financial crisis also the Finnish spread on Bunds
This type of instrument is the fourth proposal in the Giovannini Group Report (2000). Issuance of a
common bond by the EIB has been considered by Gros and Micossi (2009) for the purpose of financing a
European Financial Stability Fund and by various authors for funding projects envisaged in the Lisbon Strategy
(see e.g. Majocchi 2005).
5
The main difference in the two instruments lies in their legal and political feasibility. Issuance by a
European Institution would encounter lower legal obstacles and could be more politically acceptable by
fiscally sound Member States but would face stronger opposition by non-euro Member States, which would de
facto cross guarantee the debt of the participating States (see Goldschmidt 2009). It is worth noting that the EC
already funds its Balance of Payments Facility by issuing bonds with a AAA rating.
10
in the policy debate and in the media over the course of 2010 and, finally, gained
strength in August 2011. Indeed, a European government bond backed by the joint and
cross guarantees of the participating Member States could ensure market access at better
conditions during crises to weaker sovereign issuers and reduce their risk of default.
Although, it is tempting to think of a Eurobond as providing insurance against credit
risk, the opportunities for risk sharing are slim. First, movements in interest-rate spreads
(a proxy for credit spreads) have a strong common component mainly driven by changes
in international risk factors. Secondly, in the current situation, Eurobonds would imply a
transfer of risk, away from Member States with lower credit standings onto safer issuers,
that is unlikely to be reversed in the near future due to the persistence of relative fiscal
positions.
Absent risk sharing opportunities, what is then the economic rationale for a common
Eurobond? What needs to be argued is that, because a debt crisis in a Member State has
negative spillovers to other States creditworthiness, a common debt backed by joint
guarantees (or issued by an EU Institution) would reduce the risk of crisis propagation. If
a countrys default on its debt increases the probability of default in other countries,
either from contagion or interdependences, then preventing a crisis in a Member State
with weak fiscal fundamentals may improve debt sustainability in States with sounder
finances. If these externalities were significant, the introduction of a Eurobond, by
ensuring market access and better borrowing conditions to weaker Member States, could
also reduce safer States exposure to default risk. Put it simply, providing insurance to
weaker States would work as insurance for all; it would benefit all participating Member
States except, perhaps, the most virtuous ones.
In fact, in the euro area, the probability that a countrys crisis propagates to other
countries is particularly high because of strong real and financial links. A main channel
of transmission is through cross-border holdings of national bonds and increased
vulnerability of the European banking system. Another channel is through trade links.
Finally, a debt crisis may propagate to other countries because of contagion; i.e. through
a rise in yield spreads (and thus higher funding costs) due to a shift in market sentiment
and/or risk awareness.
Why then do economists as Issing (2009) and Stark (2011) oppose the introduction of
Eurobonds? The most forceful argument against a jointly guaranteed Eurobond is that it
would undermine fiscal discipline by removing incentives for sound budgetary policies.
At worst, it could create a moral hazard problem in that a Member State may be tempted
to free ride on other Members legal obligations to assume its debt in case of default. In
particular, a common Eurobond prevents financial markets from exerting their
disciplinary effects through higher interest rates, and undermines the no bailout clause
that prohibits a Member State to be liable for or assume the debt obligations of another
government. Then, with lower risk of default and lower cost of funding, Member States
11
would be encouraged to run lax fiscal policies and take up more debt. This would
weaken the credibility of the euro-zone as an area of stability and fiscal soundness.6
A first argument against Eurobonds is that the cross-default nature of the joint
guarantees would undermine the no bailout clause of the EU Treaty (Article 125 TFEU),
and heighten the risk of moral hazard. However, to assess the impact of Eurobonds on
fiscal discipline one has to ask how effective the no bailout clause is in preventing
irresponsible or even opportunistic behavior. In fact, there has always been skepticism as
to whether governments would adhere to the no bailout clause given the close financial
and economic ties within the euro area and the threat of crisis propagation. After the
rescue of Greece, Ireland and Portugal, these doubts are now stronger than ever, and the
deterrent role of the no bailout clause has lost much of its credibility.7 Summing up,
bailout expectations and moral hazard will always be a problem. It has to be seen
whether it would be wise to further weaken the no bailout principle. For instance, one
could argue that with a jointly guaranteed debt the possibility of imposing strict
conditionality on financial support, as it now happens with the European Financial
Stability Fund (EFSF), would be lost. On the other hand, advocates of Eurobonds as
Delpla and von Weizsacker (2010, 2011) contend that the use of Eurobonds would not
only be limited but also conditional on the implementation of fiscal adjustment and
reforms. Then, the issue is whether to rely on ex-ante or ex-post conditionality to enforce
fiscal consolidation, an alternative that has long characterized the debate over IMF
intervention.
Another argument against Eurobonds is that they lower the credit risk premium and
thus the interest rate that weaker Member States have to pay on their debts. By
preventing financial markets from exerting their disciplinary role, Eurobonds will further
reduce the incentives for fiscal adjustment. Interestingly, the argument applies even if
issuance of national bonds continued to remain substantial because less default risk
would translate into lower interest rates on national bond issues (assuming that their
seniority would be the same as Eurobonds). In other words, Eurobonds would crosssubsidize the national bonds of weaker Member States.
The case for relying on the disciplinary effects of widening yield spreads depends on
whether financial markets efficiently price risk. In fact, experience shows that market
signals, i.e. yield spreads, can be dominated by swings in market sentiment and, more
importantly, can remain weak for a long time and then change violently when it is too
late to prompt fiscal adjustment.
Hence, the relevant issue to address is whether yield spreads reflect the markets
assessment of fiscal fundamentals or contagion, that is, a shift in market sentiment
following the emergence of other countries fiscal distress. The euro-area sovereign debt
crisis triggered by the insolvency of Greece, Ireland and Portugal provides an interesting
case study. To the extent that the rise in Italian and Spanish spreads just reflects poor
6
See Issing (2009) for a stand against Eurobonds, Becker (2010) and Berrigan (2010) for a discussion.
The debate on strengthening the Stability and Growth Pact offers further evidence that fiscal discipline cannot
rely on the no bailout clause.
7
12
fiscal fundamentals, i.e. high debts, low growth and expected budget deficits, the
economic rationale for introducing Eurobonds would be weak. A strong case for
Eurobonds could instead be made if high yield spreads stemmed from a sudden shift in
market sentiment. To the extent that market irrationality and contagion play a greater
role than fiscal fundamentals in the pricing of risk, a common Eurobond is a useful
instrument to halt the crisis transmission.
To shed light on the determinants of yield spreads and thus on the relative role of fiscal
fundamentals and contagion in the propagation of the euro debt crisis, in the next section
we estimate a Global VAR of 10-year yield spreads on Bunds.
(Y
Yt GER = 0 + 1 Yt i1 Yt GER
+ 2 Et (bi bGER ) + 3 Et (d i d GER )
1
(Y
Yt GER
) = w (Y
*
j i
ij
Yt GER
1
if dist ij < 1, otherwise wij* = 0
dist ij
wij*
w
j i
*
ij
13
The model relates yield spreads on Bunds to local fundamentals and common factors.
Local fundamentals are chosen to capture default risk8. Following Attinasi et al. (2010),
we include the average for a 2-year period of the expected budget balance to GDP ratio
(di) and debt to GDP ratio (bi). The expected variables are the European Commission
Forecasts, that are released on a bi-annual basis. We include in the model the difference
between each countrys forecast and the forecast of the same variables for Germany.
Our specification is completed by the inclusion of two global risk factors. The first one is
an international factor, the US corporate Baa-Aaa spread, computed on the basis of the
data made available in the FRED database of the Federal Reserve of St. Louis. The
second factor is constructed to deliver country-specific stochastic trends for the impact
of other countries yield spreads, in which the levels of the spreads of all other countries
are mapped into the factor by taking into account their distance from the country
considered, as measured by differences in fiscal fundamentals. We call this factor global
spread. For each country, the global spread is determined by a weighted average of the
yield spreads in all other countries where the weights are constructed to make the global
spread more dependent on the spreads of those countries that are more similar in terms
of fiscal fundamentals. The global spread variable is inspired by the construction of
global variables in the GVAR modeling approach (see, e.g., Pesaran, Schuermann,
Weiner 2004, and Dees, Di Mauro, Pesaran, Smith 2007), where global macro variables
are constructed for each country by using trade weights. Using the distance in terms of
fiscal fundamentals makes the global variable country-specific and the weights more
volatile than in standard GVAR based on trade weights. The time-varying weights,
related to the changing forecasts for fiscal fundamentals, have the potential of explaining
the changing correlation of spreads discussed in the descriptive data analysis. To
illustrate the point we report in Figure 5 the global spreads for a typical low-yielder, the
Netherlands, and a typical high yielder, Ireland.
Insert Figure 5 about here
Note that, in the no-crisis period, the global spread variables for the Netherlands and
Ireland are very strongly correlated with a very similar mean, while in the wake of a
crisis the two global variables diverge as the higher distance of the Netherlands from the
high-yielders generates a lower mean and a lower volatility for its global spread.
Our measure of the distance in terms of fiscal fundamentals includes both projections
on debt to GDP and deficit to GDP ratios. We have assessed the performance of this
specification of distance against a range of alternative specifications. In particular, we
have considered alternative measures based on debt only, as some recent proposals on
Eurobonds (see Delpla and von Weiszacker 2010) concentrate on debt as the only
criterion to identify the credit quality of different bonds. The evidence is that using debt
as the only indicator of fiscal fundamentals delivers a global spread variable that
8
In a baseline version of our model we have also included the Amihud (2002) measure of (il)liquidity used by Acharya and
Pedersen (2005) that labels a bond as illiquid if the bond prices move a lot in response to little volumes. However, we excluded
this variable from the final specification as it turned out to be always non-significant.
14
performs much worse than that based on debt and deficit. Figure 6 illustrates this point
for the case of Italy by reporting the domestic yield spread together with two alternative
measures of the global spread: the one adopted in our model and an alternative one in
which the distance is measured only by using the debt-to-GDP ratio.
Insert Figure 6 about here
Figure 6 shows that the global spread based on debt and deficit does a much better job
in capturing fluctuations of the domestic yield spread than the global spread based on
debt only; the debt measure of the fiscal distance would have put Italy much closer to
Greece, Portugal and Ireland in the euro debt crisis than the market has actually done.
The debt-to-GDP ratio is one of the pricing variables of bonds but clearly not the only
one. Default risk premia depend on fiscal sustainability and fiscal sustainability depends
on the level of debt, on how it is financed, and on future primary surpluses.
We have estimated the model over the period 2005-2011 to have a sample that
includes three different periods of equal length: the calm period (2005-2007), the
financial crisis period (2007-2009) and the euro-debt crisis period (2009-2011) 9.
The results of the estimation reported in Table 1 can be summarized as follows.
i)
All spreads are very persistent.
ii)
The global spread variables that use non-linearly the fiscal fundamentals are
always significant with different impact coefficients. We have assessed the
robustness of this result by computing different measures of distance based
respectively on debt-only and deficit-only forecasts. The distance based on
the weighted average of debts and deficits dominates all alternatives, and the
specification based on a measure of distance that depends only on debt is the
worst performing one.
iii)
The Baa-Aaa spread is in general significant for the low yielders but it is not
for the high yielders, with the exception of Ireland.
iv)
The linear effect of the fiscal fundamentals is rather weak. Debt and deficit
are simultaneously significant only in the case of Greece, while debt has also
some marginally significant effect in the case of Portugal and Spain.
v)
Panel restrictions cannot be imposed on the system as the coefficients differ
importantly across countries.
The evidence reported shows that there is a relationship between yield spreads and
fiscal fundamentals, but this is non-linear. Fiscal fundamentals do not matter per se but
because they determine the sensitivity of the domestic yield spread to other countries
spreads. Fiscal fundamentals are important in the determination of the domestic spread
as they define the distance between countries and therefore select the reference group
relevant to determine the global spread variable. Countries with sound fiscal
fundamentals are immune from the global risk priced in the yield spreads of countries in
9
The sample period was also chosen to conduct a robustness check using CDS spreads, that are available from our full sample
only from 2006 onward.
15
distress, while weaker Member States are affected by global risk to the extent that they
are fiscally closer to troubled countries. In such a framework, markets do have a role as a
discipline device as the interdependence among different countries might very well
change over time but in a way related to fundamentals. On the other hand, when global
risk factors (captured by global spreads) are muted, poor fiscal fundamentals are not
priced in domestic yield spreads, which points to a discontinuity in the disciplinary role
of financial markets.
The structural stability of the coefficients on the global spread variables is an issue of
some relevance: in fact, instability of the impact of the global variable on domestic
spreads would imply that episodes of contagion dominate the fundamentals driven
interdependence across countries. To investigate this issue we consider first the results of
subsample estimation over three periods: the calm period (2005-2007), the financial
crisis period (2007-2009) and the euro-debt crisis period (2009-2011). The results,
reported in Table 2, show that the coefficients on the global spreads are always
statistically significant, but also that there is some evidence of instability, in particular
during the euro debt crisis. Using the terminology introduced by Forbes and Rigobon
(2002), this evidence shows the presence of contagion in the sense that the
interdependence captured by the global trend, i.e. the impact of other countries spreads
through fiscal fundamentals, might not be constant over time. If our specification
captures correctly the fundamentals driving the spreads, then the effect of contagion can
be used to measure the impact of market sentiment in driving yield differentials away
from the path consistent with fundamentals.
To measure the effect of contagion we consider a case study for Italy and Spain and
estimate a Multivariate GARCH model for the yield spread of each country and the
associated global spread. This specification allows for a time-varying conditional
variance-covariance between the yield spread of domestic bonds on Bunds and the
global spread relevant for each country, and it can be used to generate a time-varying
estimate of the impact of the global spread on the domestic spread. In practice, we
estimate the following specification:
(
(
)
)
(
(
)
)
Yt i Yt GER
Yt i1 Yt GER
FFt i1
1
= 0 + 1 i
+ 2 Et * +
i
GER *
GER *
FFt 1
Yt Yt
Yt 1 Yt 1
uti
+ 3 (Baat Aaat ) + H t1 / 2 GR
ut
vech(H t ) = M + Avech ut 1ut' 1 + Bvech(H t 1 )
i = ITA, ESP
16
This specification models the joint process of the yield spread of Italian (or Spanish)
bonds on German Bunds and the global spread variable relevant to Italy (or Spain) as a
persistent process with a mean determined by the expected fiscal fundamentals (FFIT are
the same fundamentals adopted in the system specification while FF* are weighted
average of the fiscal fundamentals of other countries with weights determined by the
distance from Italy (or Spain) previously defined). The time-varying variance-covariance
matrix of residuals, Ht, is modelled as a diagonal BEKK (Engle and Kroner 1995)
system. Therefore, the conditional variances, covariances and correlation are allowed to
vary over time.
The model provides us with a natural measure of contagion: the dynamic conditional
beta in the terminology of Bali and Engle (2010), which is the coefficient determining
the effect of a shock in the global spread on the Italian (or Spanish) spread.
E u ti u t* = t u t* , t = h12 ,t h 22 1,t
Variations in the coefficient t reflect a time varying interdependence between the
domestic spread and the global spread and they therefore illustrate how contagion affects
Italy (and Spain) following a shock to the global spread. The time varying estimate of
ti are reported in Figure 7 along with the estimates obtained from the SUR model.
Insert Figure 7 about here
The Figure displays some evidence of contagion during the global financial crisis that
becomes very strong during the euro debt crisis. During both crises the exposure of Italy
and Spain to their global spread variables becomes much higher than that predicted on
the basis of the distance from other euro-area Member States, as measured by fiscal
fundamentals. Note also that the impact of contagion on the domestic yield spreads is
very strongly correlated across the two countries. In Figure 8 we provide an estimate of
the cost of contagion as measured by the difference between the impact effect of the
global spread as estimated in the Multivariate GARCH model and in the constant
parameter SUR system.
Insert Figure 8 about here
The estimates show that the effect of contagion in the euro debt crisis is very sizable as
it stands at two-hundred basis points in the case of Italy, and at an even higher level in
case of Spain.
What does our evidence say on the relative role of fiscal fundamentals and contagion
and thus on the trade-off between setting incentives for fiscal discipline and reducing
risk of crisis propagation through shifts in market sentiment? Domestic yield spreads do
depend on fundamentals, non-linearly, but there is contagion on the top of
interdependence. The long-run fluctuations in yield spreads are related to fundamentals,
17
but such a relation is not constant over time and episodes of contagion can be traced in
the data of our sample. This evidence tells us that it is well possible that markets can stay
away from the fundamentals driven equilibrium for longer than a country can stay
solvent.
Financial markets do set incentives for fiscal discipline but they do so discontinuously;
their overreaction to global risk variables is itself an important source of instability and
crisis propagation. The efficiency of international financial markets and the role for
supranational policy intervention in crisis prevention is a highly debated topic in
economics, as witnessed by the huge literature on contagion and the role of the IMF. Our
evidence suggests that relying only on the disciplinary effects of financial markets to halt
a crisis may not be enough as yield spreads are significantly driven by market sentiment.
4. FEASIBILITY
Will a Eurobond ever be issued? The introduction of a common Eurobond backed by
the several and joint guarantees of the participating euro-area Member States faces both
legal obstacles and strong opposition by safer Member States. The cross-default nature
of such guarantees are not only against the spirit but most likely violates the no bailout
clause, i.e. the letter of Art 125 of the Treaty on the Functioning of the European Union
(TFEU), and may thus require changes in the EU legal infrastructure, either in the
Treaty, or in EU legislation. A common Eurobond also faces strong political opposition
by Germany and other Member States with sound budgetary policies and low debts. In
what follows we examine the costs of Eurobonds for safer issuers that motivate
resistance to a common issuance program and the possible benefits that could mitigate
such position.
18
would see their risk exposure increased.10 Even if their borrowing costs did not change,
issuers with no default risk would see their expected liabilities, and thus debt burdens,
increase, as they will have to pay in the case a risky issuer defaulted on its obligations.
The cost of a Eurobond issuance program, and its implications for the interest rates to
be paid on national bonds, deserve further scrutiny. This is what we do in what follows.
19
20
ranges of ratings in Figure 9, pointing to the critical share of Italian (S&P A+) and
Spanish (S&P AA+) debt, is suggestive of this segmentation.
If the problem is one of potential fiscal capacity, then Eurobonds will hardly reach a
better credit quality than the average of the national bonds of participating States. If
instead fiscal autarky is the explanation, they could be even safer than US Treasuries, but
this would require that a number of problems, from bailout redistribution rules to greater
fiscal unity, be effectively addressed.
Further insight on the credit standing that Eurobonds would eventually achieve can be
obtained by looking at the bonds issued by the European Investment Bank (EIB) to
finance investment projects. EIB bonds are backed by the capital subscribed by EU
Member States, have a Aaa rating, and are the debt instruments which come closest to
what could be an Eurobond backed by several and joint guarantees.
21
22
The second issue is whether liquidity gains can partly compensate safer Member
States, in particular Germany, for assuming the risk of weaker Member States. Indeed, it
is often argued that a common Eurobond would better satisfy the global demand for a
safe-haven asset than German Bunds, if its market size approached that of US
Treasuries. The idea is that, since safe German Bunds are in scarce supply, a common
Eurobond would attract the demand by international investors and strengthen the use of
the euro as international reserve currency. In turn, this would reduce the borrowing costs
for all euro-area sovereign issuers.
However, for a Eurobond to reach the international benchmark status of US Treasuries
two conditions must be satisfied: i) Eurobonds should be of the same credit quality of
German Bunds, and; ii) their market should reach a similar size of the US market, which
requires large outstanding volumes to the point of replacing national bonds markets.
Nothing ensures that both conditions will be satisfied.
The claim that German Bunds suffer from a lack of liquidity compared to US
Treasuries, to which we now turn, also deserves further investigation.
Data are taken from the BIS database and refer to the volumes of bonds outstanding with a maturity longer than one year.
23
Interest Rate Swaps); the US Baa-Aaa spread; the differential between CDS on 10-year
US and German bonds.
As shown in Figure 11, in the period preceding the financial crisis the asset swap
spread points to a sizeable liquidity premium, around a mean of 40 basis points, paid by
German Bunds over US Treasuries. Hence, in the pre-crisis period, a common Eurobond
market would have enabled the euro-area market to better compete with the US market
as the most liquid market globally.
However, when the financial crisis hits the markets and the Baa-Aaa spread starts
fluctuating away from its low-risk period mean, the liquidity premium paid by German
Bunds on US Treasuries disappears, and in fact is reversed, shadowing the US Baa-Aaa
spread to reach a discount of over 50 basis point at the peak of the crisis. Interestingly, as
the Baa-Aaa spread reverts toward its pre-crisis level, the asset swap spread initially
seems to converge back toward a small premium but, with the onset of the Greek-debt
crisis, it goes back to a discount level, fluctuating initially in the range of 20-30 basis
points, to then increase slightly above 50 basis points in the heat of the August 2011
crisis. During the euro debt crisis the US-German Asset Swap spread cannot be
explained by the default risk component as the CDS spread indicates that US Treasuries
are perceived as less risky than German Bunds. While expected exchange rate
fluctuations play a relevant role in the explanation of the yield spread during the
financial crisis, when yield differentials and asset swap spreads have the opposite sign,
and are negatively correlated, they are remarkably stable in the euro debt crisis.
Hence, while in the pre-crisis period US Treasuries did enjoy the status of the most
liquid and safest benchmark globally, they appear to have lost their safe-haven appeal
thereafter.14 In fact, the data tell us that in the two different crises international portfolio
shifts generated a discount on German Bunds with respect to US Treasuries, despite the
smaller size of the German Bund market.
How should this evidence be interpreted? A first possible explanation is that the
financial crisis has generated a flight to safety towards German Bunds that has reversed
the pre-crisis situation; interestingly this reversion has lasted over the euro debt crisis,
probably due to a portfolio shifts towards German Bunds away from bonds issued by
riskier Member States. However, this interpretation is not consistent with the behavior of
CDS spreads that would instead justify a flight to quality from German to US bonds.
A second explanation lies in the relative supply increase of US Treasuries that could
have required an increase in their proper risk premium to be accommodated in
international investors portfolios.
A third explanation is that, with the outbreak of the crisis, the liquidity of German
Bunds has increased, closing the gap with US Treasuries. If this were the case, then there
would be no liquidity gain for Germany from a common Eurobond market.
14
While it would worth looking at traded volumes to better understand the apparent reversal in international
benchmark status, data are difficult to find because US Treasuries and German Bunds are mostly traded over
the counter.
24
Whatever is the explanation, what matters for our purposes is that the argument that
Germany may still be penalized by a market size much smaller than the US market, and
that it could benefit from a common Eurobond market is not supported by the data.15
5.
POLICY CONCLUSIONS
Should euro-area Governments issue a Eurobond? In this paper we have provided an
evaluation of the potential benefits and costs of a Eurobond backed by the several and
joint guarantees of euro-area Member States. As default risk is the main driver of
interest-rate differentials, the efficiency gains from greater liquidity appear to be small
and appealing only to small safe issuers. Contrary to the safe-haven argument, German
Bunds do not seem to suffer from a lack of liquidity, and thus higher costs, compared to
US Treasuries despite a smaller market size. Moreover, insuring the default risk of
Member States with weaker fundamentals is not without costs for safer Member States:
their expected liabilities will increase and, if the Eurobond did not reach the same credit
quality as German Bunds, an event that we cannot rule out, their borrowing conditions
would also worsen. More important, fiscal fundamentals seem to matter in the pricing of
credit risk as they determine the sensitivity of domestic yield spreads to other countries
spreads. Countries with weak fiscal fundamentals are affected by the global risk priced
in the yield spreads of countries in distress to the extent that they are fiscally closer to
the latter. This suggests that financial markets, though discontinuously, have a role as a
discipline device. Getting rid of their disciplinary role by allowing market access with a
jointly guaranteed Eurobond may not be wise. Italy is a case in point: would fiscal
adjustment and budget balance reforms ever be implemented in August 2011 without the
pressure of a high BTP-Bund spread?
All in all, issuing a Eurobond does not seem a good idea. Can it be rescued? The
answer is yes, and the reason lies in the evidence of substantial contagion effects. We
find that fiscal fundamentals per se are not significant determinants of yield spreads but
only as they interact with other countries spreads; i.e. with the global risk that the
market perceives. When global risk factors are muted, poor fiscal fundamentals are not
priced in domestic yield spreads, which points to a discontinuity in the disciplinary role
of financial markets. More important, the interdependence captured by the global spread
variable, i.e. the impact of other countries spreads through fiscal fundamentals, is not
constant over time, a clear sign of contagion in the definition of Forbes and Rigobon
(2002). Then, a shift in market sentiment following the emergence of a country fiscal
distress may propagate a debt crisis to relatively safer countries through a rise in yield
15
A referee made us notice that the fact that the German Bund future was the dominant contract before the
Euro and it has quickly become virtually the only future contract on Government Bonds in the Euro plays an
important role in determining the very high liquidity of 10-year bunds.
25
spreads that worsen their borrowing conditions. The effect of contagion for Italy is
estimated to be as high as 200 basis points during the August 2011 crisis, while it is even
greater in the case of Spain.
Financial markets do set incentives for fiscal discipline but they do so discontinuously;
evidence of an overreaction to global risk variables points to shifts in market sentiment
as an important source of instability and crisis propagation. If markets can stay irrational
longer than a country can stay solvent, then their disciplinary role is considerably
weakened, and issuing a Eurobond can be economically justified.
Furthermore, the introduction of a Eurobond, by ensuring market access and better
borrowing conditions, would not only reduce the risk of default in weaker Member
States but could also avoid the propagation of the crisis to more fiscally responsible
States, either from contagion or interdependences in real and financial sectors. Put it
simply, providing insurance to States with weaker fiscal fundamentals would work as
insurance for all; it would benefit all participating Member States except, perhaps, the
most virtuous ones.
However, one may wonder why the Dutch, the Finnish and the German taxpayers
should insure the risk of fiscally reckless Member States and even pay for it. In fact,
political opposition to a Eurobond is easily understood but fails to realize that
alternatives could be worst. Financial assistance to Greece, Ireland and Portugal under
the EFSF may turn out to be costlier to euro-area taxpayers in the case these economies
will not recover, and EFSFs resources are not enough to rescue Italy and Spain if they
defaulted on their debts. The costs of crisis resolution may well be higher than those to
be paid for crisis prevention with a Eurobond program. In fact, the consequences of an
Italian and Spanish default are unthinkable. Such an event will lead to the collapse of the
euro area financial system, a deep recession and a disruption of trade with huge costs for
all euro area Member States. Thinking that the taxpayers of fiscally responsible States
will not be hit by a crisis of such proportions is a dangerous illusion. The emergence of
non-negligible CDSs on the French and German debts is evidence that the market
already assigns a probability that the Italian and Spanish debt crisis will spread to
countries with better fundamentals.
No doubt, participation to a Eurobond program should be subject to ex ante strict
conditionality and binding fiscal rules to enforce fiscal discipline and mitigate moral
hazard problems, making up for the weakened role of financial markets. Likewise, the
problem of an equitable sharing of the costs and benefits of a Eurobond program cannot
be dismissed. In fact, advocates of Eurobonds have made various proposals to reinforce
fiscal discipline, as well as to redistribute the benefits and costs of Eurobonds. Although
such solutions have considerable merit, they miss the main point: efforts to design a
perfect Eurobond program will never be enough to convince its opponents that it is a free
lunch.
What Europe and the Euro need is a renewed aim for a stronger EU political union.
Indeed, it is hard to think of solutions of the euro debt crisis without further steps
towards political integration. Sustainability requires institutional changes; for instance, a
26
new entity in charge of crisis management with independent decision power and some
form of fiscal union to address macroeconomic imbalances with a common policy.
Indeed, it is easier to think of a common Eurobond as a valuable instrument and the
natural byproduct of a common fiscal policy. Unfortunately, a reform of EU fiscal
governance takes time while time to halt the crisis is running out very fast. Why then not
to think of a Eurobond program as the first step towards greater fiscal integration? No
doubt, the introduction of a Eurobond could signal a political will for greater fiscal unity,
paving the way for deeper reforms of EU fiscal governance.
27
REFERENCES
Acharya, V.V. and L.H. Pedersen (2005). Asset Pricing with Liquidity Risk, Journal
of Financial Economics, 77, 375-410.
Adrian, T. and M. Brunnemaier (2010). CoVar, mimeo, http://www.princeton.edu/
~markus/research/papers/CoVaR
Attinasi, M.G., C. Checherita and C. Nickel (2009). What explains the surge in euro
area sovereign spreads during the Financial Crisis 2007-2009?, ECB Working Paper
1131.
Bali, T.G. and R.F. Engle (2010). Resurrecting the Conditional CAPM with Dynamic
Conditional Correlations, http://www.bus.wisc.edu/finance/workshops/documents/BaliEngle.pdf
Beber, A., M.W. Brandt and K.A. Kavajecz (2009). Flight to Quality or Flight to
Liquidity? Evidence from the Euro area Bond Market, Review of Financial Studies, 20,
527-552
Becker, W. (2010). The creation of a common European government bond.
Arguments against and alternatives, in The creation of a common European bond
market, Cahier Comte, Boel, No.14, ELEC, March.
Berrigan, J. (2010). Joint issuance of euro-denominated government bonds, in The
creation of a common European bond market, Cahier Comte, Boel, No.14, ELEC,
March.
Boonstra, W. (2010). How EMU can be strengthened by central funding of public
deficits, in The creation of a common European bond market, Cahier Comte, Boel,
No.14, ELEC, March.
Dees, S., Di Mauro, F. Pesaran, M. H. and L.V. Smith, (2007). "Exploring the
international linkages of the euro area: a global VAR analysis", Journal of Applied
Econometrics, 22(1), 1-38.
De Grauwe, P. and W. Moesen (2009). Gains for all: A proposal for a common euro
bond, CEPS Commentary, April 3, and Intereconomics, May/June.
28
29
BG
ESP
0.059
(0.020)
0.919
-0.193 0.118
0.011
0.096
-0.050 -0.149 -0.028 -0.065 -0.211 -0.203 -0.521 -0.324 -0.386 -1.598
Adj R - squared
Mean Dep. Variable
SE of Regression
(0.018)
(0.060)
(0.008)
(0.005)
(0.019)
IRE
ITA
0.057
0.348
(0.032)
(0.021)
(0.004)
(0.117)
(0.063)
(0.091)
(0.007)
(0.006)
(0.042)
0.961
0.888
0.867
0.969
0.972
0.897
0.874
0.926
0.979
(0.013)
(0.079)
(0.010)
FIN
(0.020)
0.185
(0.031)
GRE
(0.021)
(0.013)
-0.093 0.032
(0.045)
(0.004)
-0.002 0.016
(0.009)
FRA
(0.003)
0.059
(0.017)
0.326
(0.029)
(0.022)
0.131
(0.178)
-0.000 0.079
(0.005)
(0.131)
0.039
(0.003)
(0.024)
0.014
(0.033)
0.079
0.222
(0.008)
(0.019)
(0.018)
(0.029)
0.207
(0.059)
(0.044)
OE
(0.014)
PT
(0.015)
(0.024)
-0.005 0.003
(0.035)
0.016
NL
0.009
(0.014)
(0.006)
0.002
0.024
(0.006)
(0.004)
0.620
0.340
(0.036)
(0.018)
(0.065)
0.000
(0.009)
0.023
(0.004)
0.403
(0.027)
(0.245)
-0.047
(0.052)
-0.008
(0.016)
1.612
(0.100)
(0.019)
(0.031)
(0.017)
(0.008)
(0.058)
(0.047)
(0.038)
(0.019)
(0.028)
(0.102)
0.98
0.43
0.059
0.99
0.70
0.097
0.98
0.17
0.027
0.98
0.21
0.027
0.99
2.80
0.329
0.99
1.60
0.220
0.99
0.76
0.075
0.99
0.19
0.021
0.98
0.28
0.034
0.99
1.15
0.148
(Y
Yt GER = 0 + 1 Yt i1 Yt GER
+ 2 Et (bi bGER ) + 3 Et (d i d GER )
1
Table2SpreadsonBunds, SeeminglyUnrelatedRegression(Spain,Greece,Ireland,Italy,Portugal,),subsampleevidence
ESP
GRE
0.267
0.222
(0.028)
(0.029)
(0.019)
(0.167)
(0.615)
(0.333)
2
3
4
5
(0.251)
2005:08
2007:08
2007:08
2009:09
(0.101)
2009:09
2011:08
(0.781)
2005:08
2007:08
2009:09
2011:08
2005:08
2007:08
(0.076)
2007:08
2009:09
(0.569)
2009:09
2011:08
(0.216)
2005:08
2007:08
(0.039)
2007:08
2009:09
(0.027)
2009:09
2011:08
0.801
0.746
0.870
0.908
0.797
0.937
0.938
0.785
0.930
0.856
0.807
0.820
0.843
0.897
0.938
(0.046)
(0.044)
(0.033)
(0.029)
(0.040)
(0.031)
(0.036)
(0.037)
(0.048)
(0.034)
(0.046)
(0.039)
(0.038)
(0.046)
(0.036)
0.013
1.184
(0.026)
(0.026)
(0.306)
(0.806)
(0.647)
(0.445)
(0.206)
(0.020)
(0.202)
(0.402)
(0.269)
(0.124)
(0.327)
(0.120)
(0.389)
0.010
0.104 -0.336 -0.019 -0.062 -0.061 -0.001 -0.077 -0.334 -0.008 -0.058 -0.128 0.011 -0.013 -0.081
(0.007)
(0.032)
(0.137)
(0.006)
(0.035)
(0.097)
(0.003)
(0.059)
(0.120)
(0.015)
(0.044)
(0.105)
(0.013)
(0.033)
(0.142)
0.023
0.246
(0.024)
(0.010)
(0.535)
(0.001)
(0.029)
(0.015)
(0.113)
(0.113)
(0.024)
(0.020)
(0.336)
(0.018)
(0.039)
(0.01)
(0.271)
1.192
0.791
0.180
1.106
1.990
0.219
0.807
2.008
0.191
0.891
0.842
0.610
1.459
1.275
1.755
(0.143)
(0.055)
(0.053)
(0.101)
(0.108)
(0.102)
(0.055)
(0.156)
(0.071)
(0.076)
(0.061)
(0.04)
(0.207)
(0.074)
(0.185)
-1.075 -0.623 -0.092 -0.948 -1.453 -0.177 -0.736 -1.156 -0.191 -0.631 -0.754 -0.427 -1.182 -1.070 -1.626
(0.015)
(0.052)
(0.052)
(0.106)
(0.147)
(0.097)
(0.063)
(0.187)
(0.077)
(0.094)
(0.059)
(0.069)
(0.215)
(0.100)
(0.185)
Adj R - squared
Mean Dep. Variable
SE of Regression
0.83
0.03
0.006
0.99
0.45
0.034
0.96
1.62
0.162
0.96
0.25
0.010
0.99
1.11
0.064
0.98
7.06
0.574
0.95
-0.02
0.005
0.99
0.87
0.083
0.98
3.99
0.351
0.96
0.23
0.008
0.99
0.72
0.044
0.96
1.35
0.122
0.89
0.12
0.017
0.99
0.62
0.039
0.99
3.38
0.252
(Y
2007:08
2009:09
POR
2005:08
2007:08
(0.203)
2009:09
2011:08
ITA
Sample
2007:08
2009:09
IRL
Yt GER = 0 + 1 Yt i1 Yt GER
+ 2 Et (bi bGER ) + 3 Et (d i d GER )
1
Table2SpreadsonBunds,SeeminglyUnrelatedRegression(Belgium,Finland,France,NetherlandsandAustria),subsampleevidence
BEL
FIN
Sample
2005:08
2007:08
-0.001 0.058
(0.014)
(0.090)
1
2
2007:08
2009:09
(0.07)
2009:09
2011:08
2005:08
2007:08
(0.015)
2007:08
2009:09
(0.065)
FRA
2009:09
2011:08
2005:08
2007:08
2007:08
2009:09
5
6
0.020
0.014
0.021
(0.058)
(0.012)
(0.009)
(0.053)
(0.010)
(0.012)
2007:08
2009:09
(0.065)
2009:09
2011:08
(0.026)
2005:08
2007:08
2007:08
2009:09
(0.018)
2009:09
2011:08
(0.047)
0.894
0.904
0.820
0.922
0.732
0.788
0.861
0.691
0.777
0.925
0.704
0.867
0.884
0.824
0.715
(0.041)
(0.041)
(0.035)
(0.047)
(0.047)
(0.035)
(0.052)
(0.052)
(0.031)
(0.053)
(0.040)
(0.038)
(0.046)
(0.057)
(0.252)
(0.176)
(0.019)
(0.145)
(0.093)
(0.043)
(0.064)
(0.069)
(0.011)
(0.224)
(0.053)
(0.024)
(0.284)
(0.141)
0.003
0.004 -0.026 0.011 -0.033 -0.025 -0.010 0.003 -0.025 0.009 -0.009 -0.046
(0.003)
(0.041)
(0.015)
(0.055)
(0.004)
(0.013)
(0.004)
(0.013)
(0.029)
(0.005)
(0.032)
(0.023)
(0.004)
(0.032)
(0.039)
0.001 -0.047 -0.013 0.012 -0.013 -0.003 0.011 -0.024 0.000 -0.020 0.023
(0.0119)
(0.007)
(0.010)
(0.076)
(0.013)
(0.024)
(0.013)
(0.006)
(0.033)
(0.012)
(0.006)
(0.018)
(0.010)
(0.012)
(0.042)
0.859
0.791
0.098
0.842
0.543
0.037
0.693
0.389
0.050
0.803
0.710
0.196
0.607
0.841
0.270
(0.078)
(0.055)
(0.029)
(0.071)
(0.065)
(0.020)
(0.072)
(0.026)
(0.012)
(0.076)
(0.037)
(0.021)
(0.064)
(0.063)
(0.037)
-0.755 -0.623 -0.018 -0.794 -0.327 -0.014 -0.710 -0.286 -0.032 -0.644 -0.567 -0.188 -0.578 -0.612 -0.224
(0.086)
(0.052)
(0.028)
(0.077)
(0.072)
(0.018)
(0.073)
(0.029)
(0.013)
(0.080)
(0.044)
(0.022)
(0.071)
(0.074)
(0.040)
Adj R - squared
Mean Dep. Variable
0.96
0.05
0.99
0.48
0.95
0.78
0.96
-0.03
0.98
0.29
0.83
0.26
0.92
0.03
0.99
0.25
0.83
0.35
0.95
0.02
0.99
0.30
0.89
0.25
0.96
0.03
0.99
0.41
0.82
0.40
SE of Regression
0.006
0.034
0.085
0.006
0.029
0.029
0.006
0.019
0.039
0.005
0.019
0.022
0.005
0.035
0.170
2005:08
2007:08
0.020 -0.117 -0.473 -0.037 0.230 -0.128 -0.022 -0.054 -0.021 -0.000 -0.001 -0.062 -0.028 0.020 -0.119
2009:09
2011:08
OE
(0.041)
(0.021)
(Y
NL
Yt GER = 0 + 1 Yt i1 Yt GER
+ 2 Et (bi bGER ) + 3 Et (d i d GER )
1
10-Y Bunds
10
9
8
7
6
5
4
3
2
90
92
94
96
98
00
02
04
06
08
10
Spreads on Bunds
16
12
Italy
Spain
Portugal
France
Finland
Austria
Netherland
Belgium
Greece
Ireland
-4
90
92
94
96
98
00
02
04
06
08
10
Figure 1. 10-Y Government bond yield spreads in the Euro area Spreads are differences between 10-year
yields in % annual terms.
Source: Datastream/Thomson Financial
12
Portugal
Italy
Ireland
Greece
Spain
-4
2005
2006
2007
2008
2009
2010
2011
2.4
2.0
1.6
Belgium
Finland
France
Netherlands
Austria
1.2
0.8
0.4
0.0
-0.4
2005
2006
2007
2008
2009
2010
2011
Figure 2. Post-EMU Spreads of Euro area vs. German 10-year bond yields Yield differentials are
presented in % annual terms and refer to the 10-year maturity of the term structure of interest rates, the most actively traded
maturity in the Eurozone government securities market. German bond yields are taken as the reference
Source: Datastream/Thomson Financial
Yields Componenents BG
Yields Componenents F N
2.5
Yields Componenents F R
.8
2.0
.8
.6
.6
1.5
.4
.4
1.0
.2
0.5
.2
.0
0.0
.0
-.2
- 0 .5
- 1 .0
-.4
2005
2006
2007
2008
2009
2010
2011
- .2
2005
2006
2007
yield spread vs G ER
cds spread vs G ER
non default component vs GER
2008
2009
2010
2011
yield spread vs G ER
cds spread vs G ER
non default component vs G ER
Yields Componenents NL
2 .0
1 .6
.6
1 .2
.4
0 .8
0 .4
.2
0 .0
.0
- 0 .4
-.2
- 0 .8
2005
2006
2007
2008
2009
yield spread vs G ER
cds spread vs G ER
non default component vs GER
2010
2011
2005
2006
2007
2008
2009
yield spread vs G ER
cds spread vs G ER
non default component vs G ER
Figure 3a. The default and non default component in yields spreads Low Yielders
2006
2007
2008
2009
yield spread vs G ER
cds spread vs GER
non default component vs G ER
Yields Componenents OE
.8
2005
2010
2011
2010
2011
Yields Componenents ES
Yields Componenents GR
Yields Componenents IR
25
12
20
10
15
10
-5
-1
-10
2005
2006
2007
2008
2009
2010
2011
-2
2005
2006
2007
2008
2009
2010
2011
Yields Componenents IT
12
10
3
8
2
6
4
2
0
0
-1
-2
2005
2006
2007
2008
2009
2010
2011
2005
2006
2007
2008
2009
Figure 3b. The default and non default component in yields spreads High Yielders
2006
2007
2008
2009
Yields Componenents PT
2005
2010
2011
2010
2011
Changing Correlations
The low-risk period
2005-2011
16
1.2
1.0
12
0.8
8
0.6
0.4
0.2
0
0.0
-4
-0.2
2005
2006
2007
2008
2009
2010
2011
III
IV
II
2005
III
IV
2006
II
2007
3.5
16
3.0
12
4
2.5
2.0
1.5
1.0
0
1
0.5
0.0
0
II
III
2007
IV
II
III
IV
2008
II
2009
III
III
IV
2009
II
III
2010
IV
II
III
2011
IV
12
10
8
6
4
2
0
-2
2005
2006
2007
2008
2009
2010
2011
Figure 5. Global Spreads for an high-yielder and for a low-yielder Sources: Authors Calculation.
8
7
6
5
4
3
2
1
0
2008
2009
2010
2011
Figure 6. Global Spreads based on different indicators of distance. Sources: Authors Calculation.
-1
-2
II III IV
2006
II
III IV
2007
II
III IV
2008
II
III IV
2009
II
III IV
2010
II
III IV
2011
Figure 7. Interdependence and Contagion between the Local spreads and the Global Spreads
Sources: Authors Calculation.
5
4
3
2
1
0
-1
-2
II III IV
2006
II
III IV
2007
II
III IV
2008
II
III IV
2009
II
III IV
2010
II
III IV
2011
EIB
6
5
4
3
2
1
0
-1
2005
2006
2007
2008
2009
2010
2011
300%
DEBT/GDP
200%
EU17AAA
EU17NotFullAAA
87.37%
88.55%
223,00%
100%
60%
0%
47,16%
32,45%
2010
2012(exp)
EU17BBB+andBelow
JapanAA
UnitedStatesAA
48,37%
31,17%
7,04%
242,10%
92,00%
102,40%
2010
2012(exp)
7,73%
EU17
UnitedStates
2010
2012(exp)
Japan
Figure 10. The composition by rating range of euro-area and non-euro area government debt
Germany vs US
4
-1
2005
2006
2007
2008
2009
2010
2011