Professional Documents
Culture Documents
Journal of International Money and Finance: Atish R. Ghosh, Jonathan D. Ostry, Mahvash S. Qureshi
Journal of International Money and Finance: Atish R. Ghosh, Jonathan D. Ostry, Mahvash S. Qureshi
a currency union
Atish R. Ghosh
*
, Jonathan D. Ostry, Mahvash S. Qureshi
Research Department, International Monetary Fund, HQ1-09-612, 700 19th Street, NW,
Washington DC 20431, USA
JEL classications:
E62
H62
H63
Keywords:
Fiscal space
Sovereign risk pricing
Eurozone
a b s t r a c t
We examine how membership in a currency union affects public
debt sustainability and market assessments of default risk in
eurozone countries. We argue that there exist offsetting effects:
expectations of bailouts tend to make a given level of debt more
sustainable, lowering bond yields and CDS rates, but constraints on
the use of monetary policy would tend to have the opposite effect,
pushing rates up especially as room for scal maneuver gets
exhausted. We develop a formal concept of scal space (which
takes account of the notion of scal fatigue under which there are
limits to the governments ability to raise the primary surplus in
response to higher debt), and apply it to the eurozone countries,
investigating in particular how currency union membership affects
CDS and bond rates during both quiet and turbulent times for
a given amount of scal space. We nd that in quiet times, CDS and
bond rates for eurozone members were below what would be
expected given their scal space (a bonus from currency union
membership). But when the crisis erupted, CDS and bond yields
rose more sharply for eurozone members than would be predicted
based on their available scal space. Our interpretation is that
sovereign bailouts did not occur with the hoped-for alacrity in
euro-crisis countries, generating sharper penalties for sovereigns
that belong to a currency union.
2012 International Monetary Fund. Published by Elsevier Ltd. All
rights reserved.
* Corresponding author. Tel.: 1 202 623 6288; fax: 1 202 589 6288.
E-mail addresses: aghosh@imf.org (A.R. Ghosh), jostry@imf.org (J.D. Ostry), mqureshi@imf.org (M.S. Qureshi).
Contents lists available at SciVerse ScienceDirect
Journal of International Money
and Finance
j ournal homepage: www. el sevi er. com/ l ocat e/ j i mf
0261-5606/$ see front matter 2012 International Monetary Fund. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jimonn.2012.11.008
Journal of International Money and Finance 34 (2013) 131163
1. Introduction
Events in Europe over the past couple of years have rekindled interest in the public nances of
monetary unions. Although eurozone members with debt sustainability problems are not the only
advanced economies to be facing such concerns (among others are Japan, the United Kingdom, and the
United States), their membership in the monetary union may bring particular challenges (as well as
advantages) in dealing with their unsustainable public nances. In this paper, we examine how
membership in a currency union (CU) affects debt sustainability, and what this implies for market
pricing of default risk of eurozone members, given their available scal space.
There are three key implications of CU membership for the conduct of scal policy and the sustain-
ability of the public nances. First, andmost simply, because CUs typically reect more thanjust economic
considerations, and often represent important historical and political ties as well, members of the union
may expect to receive assistance (nancing or transfers) fromthe rest of the union in times of economic
or nancial distress. In other words, when the budget constraint binds, there is some likelihood that it
will be shifted outwards. Second, because CU members do not have their own independent monetary
policy, macroeconomic stabilization will need to rely more heavily on scal policy. The corollary is not
just that CU members should normally try to keep more scal space than countries with independent
monetary policy, it is also that their risk premiumwill rise more steeply as their scal space is exhausted
because they lack policy instruments to help offset the impact of scal consolidation on economic
activity (and hence on revenues). Third, the scal theory of the price level means that CUmembers must
have money dominant policy regimes since theycannot rely onhigher inationor engineer a change in
the price level in order to erode the real value of their debt. As such, CUmembers must be willing to raise
taxes or reduce spending to ensure that the intertemporal budget constraint is satised.
These three implications go in opposite directions for market assessments of debt sustainability.
The possibility that a CU member will be bailed out by other members of the union pushes out its
budget constraint and makes a given level of debt more likely to be sustainable.
1
Conversely, the
greater need to rely on scal policy for macroeconomic stabilization implies higher government bond
yields or credit default swap (CDS) rates especially when the governments room for maneuver (its
scal space) is low. Likewise, the limited scope for raising the ination tax or eroding the real value of
government debt means that CU members do not have the same safety valve as non-CU members,
and should therefore face correspondingly higher spreads on their public debt.
2
To what extent are these implications reected in markets assessment of eurozone sovereign risk?
To answer this, we compare the behavior of government bond yields and CDS rates for eurozone
countries with those of other advanced economies, controlling for scal space.
3
Intuitively, market
assessment of risk should depend on the governments ability to service its debt that is, its available
scal space and any systematic divergence fromthe scal-space implied rates for eurozone countries
should reect peculiarities associated with membership in the CU.
While several recent studies analyze the market pricing of sovereign risk for the eurozone countries
in the context of the global nancial crisis examining implicitly or explicitly the role of macroeco-
nomic fundamentals, notably, scal space (e.g., Attinasi et al., 2009; Fontana and Scheicher, 2010;
1
More generally, the provision of liquidity by the common central bank could also contribute to lower spreads. In principle, if
the prospect of bailouts reduces spreads for countries expected to receive such bailouts, then it should raise them for countries
expected to provide the bailouts. But one implication of the model discussed below is that risk spreads only rise as debt
approaches the countrys debt limit. Therefore, unless the prospective bailouts threatened the debt sustainability of the
stronger members of the union, it is unlikely that there would be any appreciable increase in their spreads.
2
While the risk of being exposed to higher ination in non-CU member countries could lead investors to require higher
nominal yields than otherwise, the ability to inate away debt lowers the likelihood of a discrete default (which is likely to be
more disruptive and value-destroying than modestly higher ination), thereby implying lower risk premia than otherwise. (For
CDS rates, there is no reason for inationary expectations to lead to higher spreads for non-CU members because ination is not
a credit event.) Moreover, especially during crisis, the absence of a monetary policy tool to stimulate the economy to
counteract the contractionary effects of scal consolidation could weaken growth, making adjustment more difcult, and
increasing the likelihood of default, which would also imply higher spreads.
3
The earlier literature on this topic focused on emerging market economies, which were more susceptible to debt problems;
for references to that literature, see Attinasi et al. (2009).
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 132
Aizenman et al., 2011; Bernoth and Erdogan, 2012; Maltritz, 2012) our analysis differs from them in
two key respects. First, we explicitly take into account the implications of CUmembership for eurozone
countries, and how they have varied before and during the crisis, arguing that any deviations of
sovereign risk pricing from macroeconomic fundamentals may partly reect rational behavior by
market participants (e.g., anticipation of bailouts before the crisis, and realization of lack of safety
valves and existence of policy constraints during the crisis). Second, most commonly in the existing
literature, public debt and/or the scal decit are used as proxies for scal space. While these variables
are certainly reasonable to use, and are likely to be correlated with scal space, they are inherently
backward looking. Yet solvency (and hence the likelihood of default) depends on the governments
ability or willingness to raise revenues or cut expenditures over the (possibly innite) future.
4
For our concept of scal space, therefore, we go beyond the usual debt or decit proxies and follow
Ghosh et al. (2011) to model how primary adjustments are undertaken. Consistent with empirical
evidence (Bohn, 1998, 2008; Mendoza and Ostry, 2008), we posit that governments typically act
responsibly, raising the primary balance in response to rising public debt. But there is a limit to this
process: as the requisite primary balance rises, it becomes increasingly difcult to keep cutting primary
expenditures or raising taxes. As long as (eventually) the increase in the primary balance does not keep
pace with the rising interest burden as debt increases, there will be a point the debt limit at which,
barring an extraordinary scal effort, debt dynamics become explosive and the government becomes
unable to fully meet its obligations. The distance between current (or projected) debt levels and this
debt limit constitutes the countrys scal space.
The advantage of our scal space measure over other proxies such as the debt ratio is evident from
the model itself: two governments with the same public debt might face very different spreads
depending on the distance to their debt limit. Also, the model predicts that the response of spreads to
scal space will be highly non-linear: essentially zero when there is ample scal space but rising to
innity (complete shut out from the markets) as scal space is exhausted. And, relatedly, markets will
tend to react very late and governments will be able to borrow at close to the risk-free rate until they
have almost run out of scal space.
Applying our concept of scal space to a sample of 23 advanced economies over 19702007, we
estimate scal space in the eurozone both individually and for the union in aggregate. We nd a wide
range of available scal space across advanced economies in general, but also among the eurozone
countries. Our estimates suggest that Greece, Italy, Japan and Portugal have the least scal space, while
Belgium, France, Ireland, Spain, and the United States are also constrained in their degree of scal
maneuvre. Australia, Korea, New Zealand, Norway, and Sweden generally have the most scal space to
deal with unexpected shocks.
Equippedwiththis concept of scal space, we turnnext to market assessments of debt sustainability in
eurozone countries, and ndthree main results. First, consistent with the model, we nd that our concept
of scal space is a better predictor of government bond yields or of CDS rates than just the level of public
debt, and that the relationship is nonlinear (government bond yields and CDS spreads rise more sharply
as scal space is exhausted). Second, from the time of the introduction of the euro till the 2008 global
nancial crisis, rates for eurozone members were lower than they should have been, given their scal
space. This suggests that the rst effect mentioned above dominated, and markets expected bailouts in
the event that a eurozone member ran into debt sustainability problems.
5
Third, when the crisis erupted,
the opposite happened: rates for eurozone members became higher than they otherwise should have
been on the basis of their available scal space. This implies that as the expected bailouts of countries
running into debt difculties did not occur (or at least did not happen with the hoped-for alacrity),
markets recognized that the constraints of CU membership made it more difcult to maintain debt
sustainability and this is what is (at least partly) reected in the wider rates for eurozone members.
4
Recognizing this, some of the studies (e.g., Aizenman et al., 2011) use debt divided by the average tax revenue over the past
ve years (as an indicator of the governments ability to collect taxes in the future), while others (e.g., Heppke-Falk and Hfner,
2004; and Attinasi et al., 2009) use projected scal decits.
5
Of course, another possibility is that markets were irrationally exuberant at the introduction of the euro, underpricing
default risk for members of the union.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 133
The rest of this paper is organizedas follows. Section2 provides a brief snapshot of howthe eurozone
got into its current scal situation. Section 3 discusses debt sustainability dynamics in currency unions,
and their implications for primary balance adjustment and scal space. Section 4 presents our scal
space estimates for the eurozone countries, and for several other advanced economies for comparison.
Section 5 analyzes the markets pricing of risk in the eurozone countries both before and after the global
nancial crisis, controlling for their available scal space. Section 6 concludes.
2. How did it all happen?
One of the major accomplishments of the European single currency, much heralded at the time, was
the convergence of interest rates across the monetary union. During the 1980s and much of the 1990s,
government bond yields had diverged widely across Europe, with sovereign issues by traditionally-
high ination countries such as Greece, Italy, Portugal, Spain priced at 400 to 1000 basis points
(bps) above corresponding German government bunds. With the advent of the single currency in 1999,
however, these spreads narrowed dramatically, never exceeding 50 bps until the onset of the global
nancial crisis in 2008 (Fig. 1). This compression of spreads reected the removal of any currency risk,
since countries in the eurozone would no longer be able to erode the real value of their debt by
devaluing their currency. Implicitly, however, markets were also assuming that outright default was
unthinkable. Indeed, the ip side to the removal of any currency risk was the loss of the safety valve of
partial default through ination and currency depreciation that had hitherto been available.
Henceforth, a government unable to fully honor its obligations would either need a bailout or would
have to outright default on the face value of its debt.
Eurozone convergence of key economic indicators ination, interest rates, output growth, decits
and public debt ratios continued through much of the 2000s (albeit at a slowing pace) until, on the
eve of global nancial crisis (end-2007), public debt in proportion of GDP ranged from 25 percent in
Ireland, 35 percent in Finland and Spain, 45 percent in the Netherlands, 65 percent in France and
Germany, 70 percent in Portugal, to just over 100 percent in Italy and Greece (Table 1).
In the aftermath of the global nancial crisis, public debt rose across the board, reecting various
combinations of nancial sector restructuring costs, discretionary stimulus spending, automatic
stabilizers, and the collapse of asset-price related revenues (Table 2). Germany saw its public debt
increase by some 18 percent of GDP, roughly the same as France. But the most dramatic increase was in
0
2
4
6
8
10
12
14
16
18
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Italy
Greece
Portugal
Ireland
Belgium
Spain
France
Netherlands
Finland
Austria
Introduction of Euro
in non-physical form
Beginning of
financial crisis
IMF
approves
Greece
program
European Council
announces Euro
arrives in 2002
Introduction of
Euro notes and
coins
IMF
approves
Ireland
program
IMF approves
Portugal
program
Note: Long-run government bond yield spreads over German bonds.
(In percentage points)
Fig. 1. Long-term Govt. Bond yields in eurozone countries, 19952011. Source: IFS database.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 134
Ireland, whose public debt ratio rose more than four-fold to about 105 percent, mostly corresponding
to the costs of bailing out its over-sized banking sector and the loss of revenue when the property
bubble punctured. The second-largest increase was experienced by Greece, where the ratio of public
debt to GDP rose to 163 percent (the highest in the eurozone), on the back of higher expenditure, lower
revenues, and the discovery of scal skeletons that implied discrete revisions to the public debt
Table 1
Gross public debt in eurozone countries, 20002011 (In percent of GDP).
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Austria 66.2 66.8 66.2 65.3 64.7 64.2 62.3 60.2 63.8 69.5 71.8 72.2
Belgium 107.8 106.5 103.4 98.4 94.0 92.0 88.0 84.1 89.3 95.9 96.2 98.5
Finland 43.8 42.5 41.5 44.5 44.4 41.7 39.6 35.2 33.9 43.5 48.4 48.6
France 57.3 56.9 59.0 63.2 65.1 66.7 63.9 64.2 68.3 79.0 82.4 86.3
Germany 60.2 59.1 60.7 64.4 66.2 68.5 67.9 65.2 66.7 74.4 83.2 81.5
Greece 103.4 103.7 101.5 97.3 98.8 100.3 106.1 105.4 110.7 127.1 142.8 163.3
Ireland 37.5 35.2 31.9 30.7 29.1 27.1 24.7 24.8 44.2 65.2 92.5 105.0
Italy 108.5 108.2 105.1 103.9 103.4 105.4 106.1 103.1 105.8 116.1 118.7 120.1
Netherlands 53.8 50.7 50.5 52.0 52.4 51.8 47.4 45.3 58.5 60.8 62.9 66.2
Portugal 48.4 51.1 53.7 55.7 57.5 62.5 63.7 68.3 71.6 83.1 93.4 106.8
Spain 59.3 55.6 52.6 48.8 46.3 43.2 39.7 36.3 40.2 53.9 61.2 68.5
Source: IMFs WEO database.
Table 2
Public debt and decit in selected eurozone countries, 20072011 (In percent of GDP).
2007 2008 2009 2010 2011
Greece Gross debt 105.4 110.7 127.1 142.8 163.3
Revenue 40.0 40.0 37.5 39.0 40.5
Expenditure 41.9 44.2 48.1 45.5 45.7
of which: Interest expenditure 4.7 5.0 5.1 5.7 6.8
Residual 1.1 5.7 9.1 15.4
Ireland Gross debt 24.8 44.2 65.2 92.5 105.0
Revenue 36.3 35.0 33.7 34.3 34.3
Expenditure 31.6 36.9 43.9 61.9 41.5
of which: Interest expenditure 1.0 1.4 2.0 3.1 3.3
Residual 17.4 10.8 0.2 5.3
Italy Gross debt 103.1 105.8 116.1 118.7 120.1
Revenue 46.1 45.9 46.5 46.0 46.0
Expenditure 47.1 48.2 51.3 50.4 49.9
of which: Interest expenditure 4.9 5.1 4.6 4.5 4.9
Residual 0.5 5.4 1.8 2.4
Portugal Gross debt 68.3 71.6 83.1 93.4 106.8
Revenue 41.1 41.1 39.7 41.6 44.7
Expenditure 44.0 45.5 48.9 49.8 48.2
of which: Interest expenditure 3.0 3.1 2.9 3.0 4.1
Residual 1.1 2.3 2.2 9.9
Spain Gross debt 36.3 40.2 53.9 61.2 68.5
Revenue 41.1 37.1 34.9 36.1 35.1
Expenditure 36.8 39.0 43.3 43.4 42.6
of which: Interest expenditure 1.6 1.6 1.8 1.9 2.4
Residual 2.1 5.3 0.0 0.2
France Gross debt 64.2 68.3 79.0 82.4 86.3
Revenue 49.8 50.0 49.2 49.6 51.0
Expenditure 51.7 52.5 55.9 55.9 55.6
of which: Interest expenditure 2.7 2.9 2.4 2.4 2.7
Residual 1.5 4.0 2.9 0.7
Germany Gross debt 65.2 66.7 74.4 83.2 81.5
Revenue 43.7 44.0 44.9 43.6 44.6
Expenditure 43.7 44.2 48.2 47.9 45.6
of which: Interest expenditure 2.8 2.7 2.6 2.5 2.7
Residual 1.3 4.4 4.5 2.8
Source: IMFs WEO database.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 135
series. Both Portugal and Spain sawtheir debt ratios rise by some 3040 percent of GDP, but given their
very different starting points, Portugals debt, at about 107 percent of GDP, now stands considerably
higher than that of Spain (68.5 percent of GDP). Moreover, while Spains decits reect a roughly equal
split of lost revenue and greater expenditure (each about 6 percent of GDP), for Portugal it was mainly
higher expenditure (4 percent of GDP).
The overall result has been an increase in the public debt ratios of the eurozone countries from an
average of 63 percent at end-2007 to 92 percent as of end-2011, with the minimum ratio in the region
rising from 25 to 50 percent of GDP, and the maximum rising from 105 to 163 percent of GDP. The
difference between the highest and the lowest debt ratios in the union has thus widened from 80 to
about 113 percent of GDP.
Even more dramatic has been the widening dispersion of government bond yields across members
of the union. With German bund yields themselves falling because of accommodative monetary policy
and safe haven effects, spreads over bunds have reached 1300 bps for Greece, 700 bps for Ireland,
283 bps for Spain, and 763 bps for Portugal (Fig. 1). Curiously, however, the rise in these countries
yields occurred with a delay through 2009 and even early 2010, spreads barely reacted (even though
actual and prospective public debt ratios were clearly higher by end-2009). Moreover, even though the
countries that need to pay higher yields are generally those with higher debt ratios, the correlation is
far fromperfect: for instance, Spains public debt is only 68 percent of GDP almost exactly the same as
Germanys debt ratio prior to the onset of the global nancial crisis. As discussed below, these
observations are no coincidence, but rather follow from the rational behavior of markets in a world
where governments may face limited scal space.
3. Public nances in a currency union
Regardless of its exchange rate regime or monetary arrangement, the government will (at least
eventually) be forced to satisfy its intertemporal budget constraint (IBC).
6
This constraint requires that
the present value of primary (i.e., non-interest) expenditure be no greater than the present value of
revenue (including seignorage) in real terms, net of initial obligations (the real value of base money and
the outstanding stock of government debt):
X
N
j 0
s
tj
s
tj
T
tj
g
tj
1 r
j
M
t1
D
t1
P
t
(1)
where g
t
is primary expenditure, s
t
is tax and non-tax revenues, s
t
is seignorage, T
t
is the real value of any
unrequited transfers received by the government, M
t
is the stock of high-powered money, and D
t
is
nominal stock of outstanding government debt. In addition to this solvency constraint, the government
might face rollover or liquidity risk i.e., the possibility that, even though the government could satisfy
its IBC, the market is unwilling to lend (perhaps because it fears that the government will choose not to
satisfy its IBC due to political or other costs of adjustment). Liquidity risk could then force
the government to default because it can neither meet all of its maturing obligations immediately nor
roll themover. Inwhat follows, however, we will largely abstract fromliquidity risk, andfocus onthe IBC.
3.1. Implications of currency union membership
What are the implications of membership in a currency union? These are threefold: rst, given the
enormous economic cost and nancial disruption that a disorderly break up of a monetary unionwould
imply, there is signicant likelihood that members encountering debt servicing difculties would
receive some form of assistance fromother members of the union. Indeed, in the case of Europe, break
6
Bohn (2008) argues that debt will be sustainable provided that the primary balance always reacts positively to the lagged
level of debt (regardless of the size of the response as long as it is strictly positive), which ensures that debt grows slower than
the rate at which it is discounted. This concept of sustainability, however, does not rule out an innite debt-to-GDP ratio, which
we consider to be implausible.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 136
up of the eurozone would also entail signicant political costs, with the post-war dream of a united
Europe set back by decades, possibly forever. All the more reason, therefore, why markets would expect
some form of bailout either outright transfers or ofcial nancing in the event that a member runs
into problems.
Transfers of course make insolvency less likely and, in practice, given that illiquidity problems can
readily turn to solvency problems, so does the availability of nancing especially at below-market
interest rates. Current eurozone crisis countries have received both forms of support; European
Union (EU)-level transfers (EU federal spending net of the national contribution to the budget) and
exceptional nancing.
7
Moreover, regardless of whether such transfers are in fact received, market
perceptions about the possibility of such transfers may delay the point at which the market cuts off
funding the government when there are doubts about its ability to otherwise satisfy the IBC.
Second, unless business cycles happen to be synchronized across members of the monetary union,
national scal policy will have to shoulder a greater burden for macroeconomic stabilization.
8
In the
run up to euro adoption, there was a large body of literature comparing the eurozone countries to the
United States with the main conclusion that the correlation of shocks and de facto mobility of labor are
lower in Europe (Bertola and Ichino, 1995; Clark and van Wincoop, 2001). While the average corre-
lation of GDP growth in eurozone countries did increase (from about 0.50 over 19901999 to 0.65 over
20002008), it is still considerably lower than that in the United States, including because the federal
budget is much smaller in Europe. The upshot is that members of a CU will want to keep more scal
space in order to be able to use scal policy for macroeconomic stabilization as monetary policy will
necessarily be geared to the average cycle of the union. But the corollary is that, as scal space is
exhausted, countries in the CU will see steeper increases in spreads because they do not have scope for
expansionary monetary policy to offset the effects of the scal contraction on output, and thus on
revenues.
Third, while in general the IBC can be satised by adjusting taxes and expenditures or by increasing
seignorage and eroding the real value of outstanding obligations, members of a monetary union do not
have the latter option.
9
In a money dominant regime, the government restores solvency by reducing
primary expenditures or raising revenues. In a scal dominant regime, the government restores
solvency by increasing seignorage or by engineering a discrete change in the price level through a burst
of ination that erodes the real value of its nominal liabilities (base money and government debt). The
union as a whole could of course decide to tolerate higher ination, but barring that, individual
countries must conduct policy within a money dominant rather than a scal dominant regime.
Accordingly, CU members must be willing to undertake the primary balance adjustments necessary to
ensure the sustainability of public debt if membership in the union is to be viable over the longer run.
3.2. Primary balance adjustment and scal space
SinceCUmembers must relyonprimarybalanceadjustment toensuredebt sustainability, it is natural
to gauge the sovereigns solvency according to the behavior of its primary balance. Here we assume that
governments typically act responsibly, raising the primary balance in response to rising debt in order to
meet the mounting interest payments, and ensure long-term solvency. But this process of raising the
primary balance to meet the higher interest payments cannot continue indenitely (because eventually
the balance would need to exceed GDP), and in practice it becomes increasingly difcult to keep cutting
7
In addition, budget nancing has also been provided indirectly through ECB Emergency Liquidity Assistance (ELA), Long-
term Renancing Operation (LTRO), and Securities Market Program (SMP). While these programs are intended to provide
liquidity to the nancial system, banks in turn may increase their purchase of government bonds, and thus channel the liquidity
provision into budgetary nance.
8
In practice, governments typically respond asymmetrically to cycles, with expansions during downturns not being fully
offset by commensurate contractions during booms, leading to a positive debt bias and correspondingly less scal space.
9
The constraint on the ination tax could lead to lower nominal and possibly real interest rates as markets do not have to
worry about ination eroding the real value of their claims (CDS rates should be unaffected by this as ination is not a credit
event). But to the extent that this constraint implies a higher probability of outright default, it could imply higher real interest
rates and higher CDS rates in CU member countries.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 137
primary expenditures or raising taxes. In other words, beyond a certain point, scal fatigue sets in
whereby the primary balance maystill be anincreasing functionof laggeddebt but at a decreasingrate.
As noted in Ostry et al. (2010) and Ghosh et al. (2011), a reasonable characterization of the primary
balance function, pb
t
f(d
t1
), therefore, may be a cubic function (which indeed turns out to be the case
empirically): at lowlevels of debt, there may be littleor evena negative relationshipbetweentheprimary
balance and debt; as debt increases, the relationship turns positive, but at sufciently high levels of debt,
thecurve attens out. Once theincrease inthe primary balance (inresponsetohigher debt) does not keep
pace with the mounting interest payment (given by the interest rate-growth rate differential multiplied
by the stock of outstanding debt), there will be a point the debt limit, d at whichthe primarysurplus is
insufcient to meet the interest payment; thus, debt increases further, and in the absence of an
extraordinary scal effort, thedebt dynamics becomeexplosive andthegovernment necessarily defaults.
If the primary balance response to debt is purely deterministic, then the government will be able to
borrow from the market at the risk-free rate up to d, but will be completely shut out of the market
thereafter. If theprimarybalanceis subject toshocks, thenthemarket interest ratewill begintorisebefore
d, again becoming innite at the debt limit.
The deterministic casecanbereadilyillustrated. Fig. 2shows thetwocomponents of the standarddebt
dynamics equation, d
t
d
t1
r gd
t1
pb
t
: the primary balance function (with its posited cubic
shape), andtheinterest payment schedule, r gd
t1
. As shown, therearetwointersections of thesetwo
schedules. The lower intersection, labeledd
, is thelong-runequilibriumdebt ratiotowhichtheeconomy
will conditionally converge. This equilibrium is dynamically stable: below d
, while lowering the debt limit, d, and the available scal space. Structural reforms (or other
extraordinary efforts) that improve the primary balance shift the f(d
t1
) schedule upwards, lower the
long-run debt level, d
(or d) is reported under the projected interest rates. This is because, given
these countries estimated primary balance reaction function and assumed interest rate-growth rate
differential, public debt would not be expected to converge to a nite steady-state debt ratio (it follows
that there is no maximum debt level below which convergence occurs, so it is not meaningful to
calculate d either).
15
For Greece and Italy, however, d
1
6
3
1
5
2
Table 9 (continued)
Specication Lagged FS FE FE/TE Crisis years Ted spread Financial contagion
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Ted spread (log) 0.009
(0.061)
0.015
(0.066)
Bank lending
to EMU
0.426
(0.676)
1.175
(0.958)
Bank to
EMU*Crisis 0.210
(0.613)
0.223
(0.668)
Observations 595 595 595 595 595 595
595 595 595 595 482 482
R-squared 0.763 0.806 0.828 0.851 0.888 0.916 0.816 0.863 0.749 0.772 0.767 0.798
Country xed
effects
No No Yes Yes Yes Yes No No No No No No
Year effects No No No No Yes Yes
Yes Yes No No No No
Notes: Dependent variable is (log) CDS spread averaged over the quarter. Bank lending to EMU is the share of foreign claims on EMU countries in the sample in total foreign claims
(immediate borrower basis; quarterly observations as reported by BIS). Ted spread is average of daily observations over the quarter. All other variables are dened as in Table 7. Constant is
included in all specications. Clustered standard errors (by country) reported in parentheses. *,**,*** indicate signicance at 10, 5 and 1 percent levels, respectively.
A
.
R
.
G
h
o
s
h
e
t
a
l
.
/
J
o
u
r
n
a
l
o
f
I
n
t
e
r
n
a
t
i
o
n
a
l
M
o
n
e
y
a
n
d
F
i
n
a
n
c
e
3
4
(
2
0
1
3
)
1
3
1
1
6
3
1
5
3
Table 10
Robustness: Real bond yield, scal space and Euro membership, 2000Q12011Q4.
Specication Lagged FS FE FE/TE Crisis years Ted spread Financial contagion
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Fiscal space (log) 0.239*
(0.125)
0.173
(0.124)
0.829***
(0.160)
0.781***
(0.167)
0.825***
(0.170)
0.755***
(0.182)
0.344*
(0.176)
0.356**
(0.151)
0.361*
(0.177)
0.372**
(0.158)
0.315
(0.195)
0.338**
(0.140)
Eurozone 0.743**
(0.306)
0.953**
(0.356)
0.771**
(0.320)
0.947**
(0.363)
0.691**
(0.315)
0.877**
(0.357)
0.572*
(0.325)
1.140***
(0.372)
Crisis 1.379***
(0.347)
1.719***
(0.356)
1.140***
(0.308)
1.364***
(0.354)
1.322***
(0.351)
1.532***
(0.349)
0.492
(0.543)
0.959
(0.581)
Eurozone Crisis 2.299***
(0.672)
2.459***
(0.617)
0.962*
(0.516)
1.116**
(0.517)
0.825*
(0.462)
1.067**
(0.464)
1.925***
(0.641)
1.991***
(0.587)
1.742**
(0.725)
1.694**
(0.646)
y2008 1.447***
(0.394)
1.803***
(0.486)
1.899***
(0.262)
1.899***
(0.262)
y2009 1.425***
(0.383)
2.027***
(0.478)
0.955
(0.569)
1.299**
(0.588)
y2010 1.025**
(0.468)
2.009***
(0.570)
1.183**
(0.443)
1.442***
(0.439)
y2011 1.257
(0.762)
2.161**
(0.886)
2.136***
(0.558)
2.457***
(0.452)
Eurozone y2008 1.036**
(0.377)
1.120***
(0.360)
Eurozone y2009 1.693**
(0.718)
1.732**
(0.730)
Eurozone y2010 1.478**
(0.638)
1.494**
(0.547)
Eurozone y2011 2.958**
(1.222)
3.156**
(1.133)
Real GDP per capita
(log)
0.823**
(0.370)
1.319**
(0.505)
5.045***
(1.292)
4.987**
(2.066)
5.651**
(2.313)
3.392
(3.060)
0.940**
(0.402)
1.472**
(0.533)
0.790*
(0.416)
1.326**
(0.538)
0.553
(0.533)
1.596***
(0.455)
Trade openness 0.144
(1.753)
2.590
(4.456)
2.293
(4.960)
0.765
(1.663)
0.680
(1.814)
2.528*
(1.310)
Ination 0.140
(0.130)
0.122
(0.176)
0.126
(0.170)
0.164
(0.133)
0.140
(0.137)
0.145
(0.254)
CAB/GDP 0.095***
(0.030)
0.066*
(0.035)
0.069**
(0.030)
0.089**
(0.036)
0.082**
(0.034)
0.069**
(0.030)
Fiscal balance/GDP 0.076*
(0.037)
0.055
(0.042)
0.072
(0.046)
0.040
(0.036)
0.040
(0.033)
0.066
(0.047)
VIX index (log) 0.012
(0.198)
0.265
(0.185)
0.520***
(0.166)
0.191
(0.214)
0.239
(0.204)
Ted spread (log) 0.350***
(0.060)
0.235***
(0.067)
A
.
R
.
G
h
o
s
h
e
t
a
l
.
/
J
o
u
r
n
a
l
o
f
I
n
t
e
r
n
a
t
i
o
n
a
l
M
o
n
e
y
a
n
d
F
i
n
a
n
c
e
3
4
(
2
0
1
3
)
1
3
1
1
6
3
1
5
4
Table 10 (continued)
Specication Lagged FS FE FE/TE Crisis years Ted spread Financial contagion
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Bank lending to EMU 0.158
(1.249)
0.521
(1.326)
Bank lending to EMU*Crisis 2.481
(2.227)
2.030
(2.238)
Observations 1009 1009 1101 1097 1101 1097 1101 1097 1009 1009 816 812
R-squared 0.150 0.204 0.373 0.390 0.402 0.420 0.235 0.286 0.208 0.249 0.220 0.284
Country xed effects No No Yes Yes Yes Yes No No No No No No
Year effects No No No No Yes Yes Yes Yes No No No No
Notes: Real long-term government bond yield is averaged over the quarter. Bank lending to EMU is the share of foreign claims on EMU countries in the sample in total foreign claims
(immediate borrower basis; quarterly observations as reported by BIS). Ted spread is average of daily observations over the quarter. All other variables are dened as in Table 7. Constant is
included in all specications. Clustered standard errors (by country) reported in parentheses. *,**,*** indicate signicance at 10, 5 and 1 percent levels, respectively.
A
.
R
.
G
h
o
s
h
e
t
a
l
.
/
J
o
u
r
n
a
l
o
f
I
n
t
e
r
n
a
t
i
o
n
a
l
M
o
n
e
y
a
n
d
F
i
n
a
n
c
e
3
4
(
2
0
1
3
)
1
3
1
1
6
3
1
5
5
possible nancial contagion from the eurozone countries during the crisis by including the countrys
exposure through cross-border bank loans to (other) eurozone countries, and its interaction with the
crisis variable. Although the contagion variable (pre- or during the crisis) is insignicant, its inclusion
statistically weakens the signicance of eurozone membership (most likely because of high correlation
between the eurozone membership and bank exposure to EMU variables in this smaller sample).
33
In Table 10, we repeat the above robustness analysis for real bond yields, and nd the results to be
very similar to those reported in Table 7. Specically, in all estimations (Table 10, cols. [1][12]), we nd
that eurozone countries had lower government bond yields relative to other advanced economies
before the crisis. But this pattern changed dramatically during the crisis when eurozone countries on
average experienced higher yields than their other advanced counterparts.
6. Conclusion
This paper examines howmembership in a currency union affects debt sustainability, and what this
implies for market pricing of default risk of eurozone members given their available scal space. We
argue that CU membership carries three key implications: rst, when the budget constraint of a CU
member binds, it may be expected to receive assistance fromother members of the union, thus making
a given level of debt more sustainable and lowering the likelihood of default. Conversely, however, the
greater need to rely on scal policy for macroeconomic stabilization implies higher government bond
yields (or credit default swaps (CDS) rates) especially when the government has almost exhausted its
scal space. Likewise, the limited scope for raising the ination tax or eroding the real value of
government debt means that CU members do not have the same safety valve as non-CU members,
and should therefore face correspondingly higher spreads on their public debt.
In examining how the market assesses default risk of CU members, it is essential to control for
countries scal space. Applying a forward-looking concept of scal space, as developed in Ghosh et al.
(2011), to a sample of 23 advanced economies, our estimates suggest that Greece, Italy, Japan and
Portugal have the least scal space, while Belgium, France, Ireland, Spain, and the United States are also
constrained in their degree of scal maneuvre. Australia, Korea, New Zealand, Norway, and Sweden
generally have the most scal space to deal with unexpected shocks.
We nd that our concept of scal space is a better predictor of government bond yields or of CDS rates
than just the level of public debt, and the relationship is nonlinear, rising more sharply as scal space is
exhausted. Controlling for scal space, we nd that, during the pre-crisis period (19992008), CDS and
bond rates for eurozone members were systematically and signicantly lower than would be predicted
given their scal space perhaps because markets anticipated possible bailouts (or suffered from irra-
tional exuberance over the introductionof the euro). But whenthe crisis erupted, the opposite happened:
rates for eurozone countries rose more sharply than for other advanced economies. This suggests that as
the expected bailouts of countries running into debt difculties did not occur with the hoped-for alacrity,
markets began to recognize that the policy constraints of CU membership made it more difcult to
maintain debt sustainability and this is what is reected in the wider rates for eurozone members.
It does not follow, however, that observed spreads are necessarily rational; even allowing for post-
2008 eurozone spreads to be different (i.e., once markets recognized that bailouts would not be
automatic and that currency union membership imposed constraints on adjustment), there are sizable
residuals for some eurozone countries. While these are presumably justiable in some cases, in others
it is hard to rule out the possibility that they reect multiple equilibria whereby pessimistic expec-
tations lead to excessively wide spreads that are then justied by the greater default risk given the high
spreads. If that is the case, then there is a clear role for ofcial nancing to provide liquidity to help
avoid an illiquidity problem from becoming an insolvency problem. But if that is not the case, then
provision of ofcial nancing might simply delay necessary adjustment. While this paper has gone
some way toward understanding the determinants of spreads for currency union countries, future
research will need to try to pin down more precisely how much of the observed spreads is justied on
33
In addition to a countrys nancial exposure to all eurozone countries, we also analyze the effect of nancial exposure to
Greece only, but nd the results to be very similar.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 156
the basis of fundamentals, including the countrys scal space and its monetary and exchange rate
arrangements, and how much must be attributed to multiple equilibria or market skittishness.
Acknowledgments
We are grateful to Peter Srensen, IMF colleagues, and participants at the Danmarks Nationalbanks
conference on the European Sovereign Debt Crisis and the European Central Banks Sovereign Debt
Workshop for helpful comments and suggestions, and to Hyeon Ji Lee for excellent research assistance.
The views expressed herein are those of the authors and should not be attributed to the IMF, its
Executive Board, or its management.
Appendix A
Fig. A1. Actual (solid line) and tted (dashed line) CDS spreads (in logs) for selected eurozone countries: annual data. Note: Based on
estimates reported in Table 5, col. [4].
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 157
0
2
4
6
8
2003q2 2005q2 2007q2 2009q2 2011q2
Austria
0
2
4
6
8
2003q12005q12007q12009q12011q1
Belgium
0
2
4
6
8
2003q12005q12007q12009q12011q1
France
0
2
4
6
8
2003q12005q12007q12009q12011q1
Germany
0
2
4
6
8
2003q12005q12007q12009q12011q1
Greece
0
2
4
6
8
2007q4 2008q4 2009q4 2010q4 2011q4
Ireland
0
2
4
6
8
2003q12005q12007q12009q12011q1
Portugal
0
2
4
6
8
2004q2 2006q12007q42009q32011q2
Spain
0
2
4
6
8
2003q12005q12007q12009q12011q1
Italy
Fig. A2. Actual (solid line) and tted (dashed line) CDS spreads (in logs) for selected eurozone countries: quarterly data. Note: Based on
estimates reported in Table 7, col. [4].
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 158
Table A1
Estimation results for the scal reaction function.
Sample specication 19702007 19852007
(1) (2) (3) (4)
Lagged debt 0.208*** 0.225*** 0.081 0.086
(0.059) (0.061) (0.076) (0.070)
Lagged debt_square 0.003*** 0.003*** 0.002* 0.002*
(0.001) (0.001) (0.001) (0.001)
Lagged debt_cubic 0.00001*** 0.00001*** 0.00001* 0.00001**
(3.0e-06) (3.0e-06) (3.0e-06) (3.0e-06)
Output gap 0.497*** 0.491*** 0.485*** 0.441***
(0.047) (0.046) (0.053) (0.053)
Government expenditure gap 0.185*** 0.184*** 0.183*** 0.183***
(0.047) (0.045) (0.052) (0.047)
Trade openness 0.091* 0.146***
(0.050) (0.054)
Ination 3.400 4.620**
(2.519) (2.008)
Oil price
a
8.775*** 9.529***
(3.216) (3.244)
Age dependency 0.072
(0.101)
Future dependency 0.015
(0.067)
Nonfuel price
a
3.005
(8.362)
Political stability 0.068**
(0.030)
IMF arrangement 1.142
(0.999)
Fiscal rules 0.300
(0.347)
Observations 642 642 496 491
Number of countries 23 23 23 23
R-squared 0.282 0.316 0.304 0.405
AR (1) coefcient 0.791 0.760 0.819 0.749
Notes: Dependent variable is general government primary balance to GDP (in percent); in all specications, country-specic
xed effects included, and error term assumed to follow an AR (1) process; robust standard errors reported in parentheses;
***, **, and * denote signicance at 1, 5, and 10 percent levels, respectively.
a
Applies to oil and nonoil commodities exporters only.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 159
Table A2
Estimation results for the scal reaction function with eurozone.
Sample specication 19702007 19852007
(1) (2) (3) (4)
Lagged debt 0.226*** 0.253*** 0.102 0.111
(0.060) (0.061) (0.077) (0.071)
Lagged debt_square 0.003*** 0.003*** 0.002* 0.002**
(0.001) (0.001) (0.001) (0.001)
Lagged debt_cubic 1.2E-05*** 1.2E-05*** 6.0E-06* 6.2E-06**
(0.000) (0.000) (0.000) (0.000)
Output gap 0.499*** 0.487*** 0.491*** 0.453***
(0.050) (0.049) (0.058) (0.056)
Government expenditure gap 0.176*** 0.161*** 0.169*** 0.136**
(0.056) (0.056) (0.061) (0.056)
Trade openness 0.093* 0.119**
(0.054) (0.057)
Ination 7.293*** 0.67
(2.683) (5.728)
Oil price
a
8.912*** 9.676***
(3.244) (3.359)
Age dependency 0.012 0.072
(0.094) (0.113)
Nonfuel commodity price
a
0.403
(7.689)
Political stability 0.082**
(0.035)
Fiscal rules 0.215
(0.352)
Observations 551 551 406 401
R-squared 0.285 0.329 0.311 0.407
AR (1) coefcient 0.796 0.751 0.832 0.755
Notes: Dependent variable is general government primary balance to GDP. Eurozone countries are treated as a single entity over
19992007, and (nominal GDP) weighted average is taken for primary balance, lagged debt, output gap and government gap
(simple average is taken for trade openness, ination rate, age dependency ratio, political stability index, and scal rules index).
In all specications, country-specic xed effects are included, and error term is assumed to follow an AR (1) process. Robust
standard errors in parentheses. *, **, *** indicate signicance at 10, 5 and 1 percent levels, respectively.
a
Applies to oil and nonoil commodities exporters only.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 160
Table A3
Variable denitions and data sources.
Variable Description Source
Age dependency ratio Ratio of the dependent to working
age (1564 years) population.
UNs database. Available online at: http://data.un.
org/Data.aspx?qdependencyratio(percent)
&dPopDiv&fvariableID%3a42
CDS spread In logs Bloomberg
Crisis Binary variable equal to 1 for 200811
(2008Q3:2011Q4 in quarterly data)
Currency union Binary variable equal to 1 for eurozone
countries in the sample
Current account
balance to GDP
In percent Authors calculations based on IMFs World
Economic Outlook (WEO) database.
Fiscal rules Binary variable equal to 1 if a country
has any type (expenditure, revenue,
balance budget, and debt) of scal
rule in a given year, and 0 otherwise.
IMFs Fiscal Rules database, 19852009
Future age dependency
ratio
Projected age dependency ratio 20
years ahead.
UNs database. Available online at: http://data.un.
org/Data.aspx?qdependencyratio(percent)
&dPopDiv&fvariableID%3a42
Government bond yield
(long-term)
In percent IMFs WEO database
Government
expenditure gap
Difference between actual and potential
(calculated using the Hodrick-Prescott
lter) real government consumption
spending
Authors calculations based on WEO database
IMF arrangement Binary variable equal to one if a
country has an IMF support program
in a given year, and zero otherwise.
IMFs History of Lending Arrangements database
Available online at http://www.imf.org/external/
np/n/tad/extarr1.aspx.
Ination Three-year lagged moving average of
CPI ination
Authors calculations based on WEO database
Lagged debt to GDP ratio In percent IMFs WEO database
Nonfuel commodity price Log of (trend) nonfuel commodity price
index applied to nonfuel commodity
exporters only.
Authors calculations based on WEO database
Oil price Log of (trend) oil price applied to oil
exporters only.
Authors calculations based on WEO database
Output gap Difference between actual and potential
(calculated using the HodrickPrescott
lter) real GDP.
Authors calculations based on WEO database
Political stability index Smaller (larger) values indicating
higher (lower) risk.
International Country Risk Guide (ICRG) dataset
Primary balance to
GDP ratio
In percent IMFs WEO and OECD databases
Real GDP growth rate In percent IMFs WEO database
Real GDP per capita In logs World Banks World Development Indicators
Trade openness Sum of exports and imports to
GDP (in percent)
Authors calculations based on WEO database
VIX index In logs Bloomberg
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 161
References
Abiad, A., Ostry, J., 2005. Primary Surpluses and Sustainable Debt Levels in Emerging Market Countries. IMF Policy Discussion
Paper 05/6.
Afonso, A., Furceri, A., Gomes, P., 2012. Sovereign credit ratings and nancial market linkages: application to European data.
Journal of International Money and Finance 31 (3), 606638.
Aizenman, J., Hutchison, M., Jinjarak, Y., 2011. What Is the Risk of European Sovereign Debt Defaults?. In: Fiscal Space, CDS
Spreads and Market Pricing of Risk. NBER Working Paper No. 17407.
Attinasi, M., Checherita, C., Nickel, C., 2009. What Explains the Surge in Eurozone Sovereign Spreads during the Financial Crisis
of 200709?. ECB Working Paper no. 1131/2009.
Barrios, S., Iversen, P., Lewandowska, M., Setzer, R., 2009. Determinants of Intra-euro-area Government Bond Spreads during the
Financial Crisis. European Economy Economic Papers No. 388.
Bernoth, K., Erdogan, B., 2012. Sovereign bond yield spreads: a time-varying coefcient approach. Journal of International
Money and Finance 31 (3), 639656.
Bertola, G., Ichino, A., 1995. Wage Inequality and Unemployment: United States vs. Europe. In: NBER Macroeconomics Annual,
vol. 10. University of Chicago Press, Illinois.
Bohn, H., 1998. The behavior of U.S. public debt and decits. Quarterly Journal of Economics 113 (3), 949963.
Bohn, H., 2008. The sustainability of scal policy in the United States. In: Neck, R., Sturm, J. (Eds.), Sustainability of Public Debt.
MIT Press, Cambridge, pp. 1549.
Clark, T., van Wincoop, E., 2001. Borders and business cycles. Journal of International Economics 55 (1), 5985.
De Grauwe, P., 2011. The Governance of a Fragile Eurozone. CEPS Working Paper No. 346.
Fontana, A., Scheicher, M., 2010. An Analysis of Euro Area Sovereign CDS and Their Relation with Government Bonds. ECB
Working Paper no. 1271/2010.
Gande, A., Parsley, D., 2005. News spillovers in the sovereign debt market. Journal of Financial Economics 75 (3), 691734.
Ghosh, A., Kim, J., Mendoza, E., Ostry, J., Qureshi, M., 2011. Fiscal Fatigue, Fiscal Space and Debt Sustainability in Advanced
Economies. NBER Working Paper No. 16782. Also forthcoming in the Economic Journal.
Gros, D., 2011. External versus Domestic Debt in the Euro Crisis. Available online. http://www.voxeu.org/index.php?qnode/
6550.
Heppke-Falk, K., Hfner, F., 2004. Expected Budget Decits and Interest Rate Swap Spreads Evidence for France, Germany and
Italy. Discussion Paper Series 1: Studies of the Economic Research Center No. 40/2004.
Hull, J., Predescu, M., White, A., 2004. The relationship between credit default swap spreads, bond yields, and credit rating
announcements. Journal of Banking and Finance 28 (11), 27892811.
IMF, 2010. Global Financial Stability Report: April 2010. International Monetary Fund, Washington DC.
IMF, 2012. Fiscal Monitor: April 2012. International Monetary Fund, Washington DC.
Islamilescu, J., Kazemi, H., 2010. The reaction of emerging market credit default swap spreads to sovereign credit rating changes.
Journal of Banking and Finance 34 (12), 28612873.
Maltritz, D., 2012. Determinants of sovereign yield spreads in the Eurozone: a Bayesian approach. Journal of International
Money and Finance 31 (3), 639656.
Table A4
Coding of Moody and S&P ratings.
Credit rating Moody S&P Transformation
Investment grade Aaa AAA 20
Aa1 AA 19
Aa2 AA 18
Aa3 AA 17
A1 A 16
A2 A 15
A3 A 14
Baa1 BBB 13
Baa2 BBB 12
Baa3 BBB 11
Speculative grade Ba1 BB 10
Ba2 BB 9
Ba3 BB 8
B1 B 7
B2 B 6
B3 B 5
Caa1 CCC 4
Caa2 CCC 3
Caa3 CCC 2
Ca CC 1
C C
SD
D
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 162
Mendoza, E., Ostry, J., 2008. International evidence on scal solvency: is scal policy responsible? Journal of Monetary
Economics 55 (6), 10811093.
Norden, L., Weber, M., 2004. Informational efciency of credit default swap and stock markets: the impact of credit rating
announcements. Journal of Banking and Finance 28 (11), 28132843.
Ostry, J., Ghosh, A., Kim, J., Qureshi, M., 2010. Fiscal Space. IMF Staff Position Note SPN/10/11.
Roubini, N., Sachs, J., 1988. Political and Economic Determinants of Budget Decits in the Industrial Democracies. NBER Working
Paper No. 2682.
A.R. Ghosh et al. / Journal of International Money and Finance 34 (2013) 131163 163