Assessing sovereign risk the case of rich countries

Description
This paper aims to shed new light on the inability of credit rating agencies (CRAs) to
forecast the recent defaults and so-called quasi-defaults of rich countries. It also describes how Moody’s
sovereign rating methodology has been modified – and could be further improved – to solve this
problem.

Journal of Financial Economic Policy
Assessing sovereign risk: the case of rich countries
Norbert Gaillard
Article information:
To cite this document:
Norbert Gaillard , (2014),"Assessing sovereign risk: the case of rich countries", J ournal of Financial
Economic Policy, Vol. 6 Iss 3 pp. 212 - 225
Permanent link to this document:
http://dx.doi.org/10.1108/J FEP-03-2014-0017
Downloaded on: 24 January 2016, At: 21:49 (PT)
References: this document contains references to 25 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 107 times since 2014*
Users who downloaded this article also downloaded:
Tong Li, (2014),"Shadow banking in China: expanding scale, evolving structure", J ournal of Financial
Economic Policy, Vol. 6 Iss 3 pp. 198-211 http://dx.doi.org/10.1108/J FEP-11-2013-0061
Faten Ben Bouheni, (2014),"Banking regulation and supervision: can it enhance stability in Europe?",
J ournal of Financial Economic Policy, Vol. 6 Iss 3 pp. 244-269 http://dx.doi.org/10.1108/J FEP-11-2013-0059
Galina Hale, J oão A.C. Santos, (2014),"Do banks propagate debt market shocks?", J ournal of Financial
Economic Policy, Vol. 6 Iss 3 pp. 270-310 http://dx.doi.org/10.1108/J FEP-03-2014-0023
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Assessing sovereign risk:
the case of rich countries
Norbert Gaillard
NG Consulting, Corbeil-Essonnes, France
Abstract
Purpose – This paper aims to shed new light on the inability of credit rating agencies (CRAs) to
forecast the recent defaults and so-called quasi-defaults of rich countries. It also describes howMoody’s
sovereign rating methodology has been modifed – and could be further improved – to solve this
problem.
Design/methodology/approach – After converting bond yields into yield-implied ratings, accuracy
ratios are computed to compare the respective performances of CRAs and market participants. Then
Iceland’s and Greece’s ratings at the beginning of the Great Recession are estimated while accounting
for the parameters included in the new methodology implemented by Moody’s in 2013.
Findings – Market participants outperformed Moody’s and Standard &Poor’s in terms of anticipating
the sovereign debt crisis that hit several European countries starting in 2008. However, the new
methodology implemented by Moody’s should lead to more conservative and accurate sovereign
ratings.
Originality/value – The chronic inability of CRAs to anticipate public debt crises in rich countries is
dangerous because the countries affected – which are generally rated in the investment-grade category
– are substantially downgraded, amplifying the sovereign debt crisis. This study is the frst to
demonstrate that Moody’s has learned from its recent failures. In addition, it recommends ways to
detect serious threats to the creditworthiness of high-income countries.
Keywords Debt, Credit rating, Standard & Poor’s, Economic development: Financial markets,
Financial markets and the macroeconomy, Sovereign debt crisis, Moody’s
Paper type Research paper
1. Introduction
The Great Recession had a number of adverse effects on high-income countries. As early
as 2008, most of them were obliged to implement support packages and guarantee
schemes to prevent the bankruptcy of major fnancial institutions. These ad hoc
measures increased sovereign borrowing needs, thus exacerbating fscal imbalances.
For instance, the central government debt of Germany, France, the USA and the UK
grew(respectively) by 14, 26, 44 and 54 per cent, respectively, in the two years from2007
to 2009[1]. Iceland, whose banking sector accounted for an extremely high percentage of
its gross domestic product (GDP), was able to avoid default only by turning to the
International Monetary Fund (IMF) in October 2008. A year later, the Greek
Government’s announcement that its fscal defcit would be much higher than expected
triggered strong distrust of the Hellenic Republic and of other peripheral eurozone
countries (i.e. Cyprus, Ireland, Italy, Portugal and Spain). In May 2010, the European
Union (EU) provided fnancial assistance to members having diffculties by establishing
JEL classifcation – G01, G15, G23, G24, H63
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JFEP
6,3
212
Journal of Financial Economic Policy
Vol. 6 No. 3, 2014
pp. 212-225
©Emerald Group Publishing Limited
1757-6385
DOI 10.1108/JFEP-03-2014-0017
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
the European Financial Stabilisation Mechanism (EFSM) and the European Financial
Stability Facility (EFSF)[2]. This dual approach has been taken with regard to fve
sovereign states so far, but it has not enabled either Greece or Cyprus to remain solvent.
Greece restructured its public debt in March 2012, imposing investor losses that
exceeded 70 per cent (Moody’s, 2013a), and Cyprus made a distressed debt exchange in
July 2013 (Apostolides, 2013; Moody’s, 2013b).
Credit rating agencies and, to a lesser extent, market participants failed to anticipate
either the Icelandic banking collapse or the eurozone debt crisis. These failures were
especially harmful because the countries affected – rated in the single-A, double-A or
triple-Acategory at the dawn of the Great Recession – endured huge downgrades, which
amplifed the sovereign debt crisis. This paper sheds newlight on the inability of CRAs
to forecast the defaults and quasi-defaults of rich countries; it also suggests ways to
improve sovereign rating methodologies with the aimof solving this problem. Section 2
summarizes evidence that the debt of high-income countries has never been a
default-free investment. Section 3 compares the respective ability of investors and CRAs
to anticipate the recent episodes of public debt crisis. Section 4 analyzes the new
sovereign rating techniques developed by Moody’s, “back-testing” them to assess
whether they would have better anticipated the debt turmoil. Section 5 concludes by
drawing lessons from the previous results and making recommendations to enhance
sovereign rating methodologies.
2. Even rich countries can default
Since the late 1980s, the World Bank has classifed as high-income those countries in
which the gross national income (GNI) per capita exceeds a certain threshold: $6,000 in
1988, $9,265 in 2000 and $12,615 in 2013[3]. Similar thresholds may be extrapolated for
years prior to 1988. For the period of 1888-1987, I assume that a sovereign issuer was a
rich country in a given year if its GDP per capita was, on average, ?50 per cent that of
the UK during the three preceding years[4]. The logic of this approach is supported by
three facts. First, GDP per capita fgures for the late nineteenth century and the frst half
of the twentieth century are more reliable than the corresponding GNI per capita fgures.
Second, the UK is an appropriate benchmark country because it was the greatest
economic, monetary and fnancial power until the interwar years. Third, the UK’s GNI
per capita in 1988 was exactly double the threshold at which the World Bank considered
a country to be high-income.
Table I gives details on the 15 rich countries that have defaulted on their foreign
currency (FC) debt since 1888. Nine of these “bankruptcies” occurred in the midst of
major public debt crises (the 1930s, the 1980s and the early 2010s), and two were closely
related to entering a major international confict (Austria and Italy in 1914 and 1940,
respectively); the remaining four defaults were isolated events.
Although there were fewsovereign defaults among high-income countries during the
past 125 years, most of these episodes were striking; in fact, some nearly triggered a
systemic risk. Argentina’s debt crisis in 1890 led to Baring’s bankruptcy and obliged the
Bank of England to launch a bailout to stop panic (Flores, 2010). In 2010, the late reaction
of European authorities to the Greek debacle resulted in a sharp depreciation of the euro
vis-a`-vis the US dollar and threatened the eurozone’s very existence. Other debt
restructurings were abnormally long (Chile, Germany and Austria did not resume
payment until the late 1940s or early 1950s) or imposed heavy losses on creditors (e.g.
213
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Table I.
Chronological listing of
FC debt defaults by rich
countries, 1888-2013
Year of
default Country
Moody’s rating
three years
prior to default Comments
1890 Argentina N.A Argentina experienced a technical default and political
riots in March and July 1890, respectively. A
committee was then appointed by the Bank of
England to work out some plan of readjustment. An
agreement provided for a loan of £14.88 million so
that the Federal Government could continue to meet
its debt service. Previous loans were restructured
(Corporation of Foreign Bondholders, 1904, pp. 32-33)
1891 Uruguay N.A Uruguay was not able to meet the coupons on its
external debt falling due in September and October
1891 (Corporation of Foreign Bondholders, 1904,
p. 423)
1914 Austria N.A At the outbreak of World War I, Austria’s debt service
went into default (Corporation of Foreign Bondholders,
1943, p. 119)
1931 Chile A Seven external loans issued by the Chilean
Government between 1922 and 1930 defaulted during
the second half of 1931 (Foreign Bondholders
Protective Council, 1945, pp. 242-247)
1933 Uruguay A Three external loans issued by the Uruguayan
Government between 1921 and 1930 defaulted during
the second half of 1933 (Foreign Bondholders
Protective Council, 1945, pp. 783-784)
1934 Germany Aa In October and December 1934, Germany defaulted on
its 7 per cent Gold External Loan of 1924 (Dawes
Loan) and also on its 5.5 per cent International Loan of
1930 (Young Loan) (Foreign Bondholders Protective
Council, 1945, pp. 455-456)
1938 Austria Ba Germany occupied Austria in March 1938; two months
later, the 7 per cent International Loan of 1930 (Gold
Bonds) lapsed into default (Foreign Bondholders
Protective Council, 1945, p. 55)
1940 Italy Baa In December 1940, Italy defaulted on its 7 per cent
External Loan of 1925 (Gold Bonds) (Foreign
Bondholders Protective Council, 1945, p. 554)
1959 Czecho-slovakia N.R The communist Czechoslovak Government defaulted
on the October 1959 coupon of the sterling issue of the
Czechoslovak State Loan of 1922. A few months later,
the government announced that it was prepared to
purchase (on or after 31 May 1960) outstanding
sterling bonds at 75 per cent of their nominal face
value (Corporation of Foreign Bondholders, 1966,
p. 123)
(continued)
JFEP
6,3
214
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Greece in 2012). The debt crisis experienced by Argentina and Venezuela in the 1980s
was a milestone in their economic decline and presaged the country’s ill-fated return to
ineffcient Bolivarian policymaking. For Iran, the default was merely symptomatic of a
major institutional and political upheaval. In most cases, the sovereigns’ high ratings
three years prior to their defaults confrm that the bankruptcies were largely
unanticipated by CRAs. Section 3 measures these rating failures and assesses whether
investors performed better.
Table I.
Year of
default Country
Moody’s rating
three years
prior to default Comments
1978 Iran N.R During the second half of 1978, strikes and
demonstrations paralyzed Iran; the Shah exiled
himself in January 1979 and the new Islamic
Government refused to repay Iran’s foreign debt (“Iran
May Face Severe Economic Damage If It Refuses to
Pay $10 Billion of Loans”, Wall Street Journal, 26
November 1979)
1982 Argentina N.R The early 1980s were characterized by an increase in
US interest rates and a global economic slowdown. In
this context, most Latin American countries registered
higher ratios of public debt to exports and greater
concentrations of debt in short maturities; even
wealthy countries (e.g., Argentina, Uruguay and
Venezuela) had to restructure sovereign debts from
1982. The IMF played a prominent role in the
restructuring of Argentina’s and Uruguay’s debt
(Sachs 1985, pp. 532-535; Rieffel 2003, p. 159)
1983 Uruguay N.R
1983 Venezuela Aaa
2012 Greece A1 After revising its fscal defcit and public debt fgures
in 2009, Greece found it increasingly diffcult to secure
low-cost fnancing. Despite an EU/IMF rescue
program in May 2010, the country was forced to
restructure 54 sovereign bonds in March 2012; a
second debt restructuring occurred in December 2012
(Moody’s 2013a, pp. 34-36)
2013 Cyprus Aa3 The exposure of Cypriot banks to Greek debt,
combined with deteriorating domestic macroeconomic
conditions, constrained Cyprus to conclude a
distressed bond exchange in July 2013 (Moody’s
2013b)
Notes: N.A. denotes Not Applicable (Moody’s did not issue sovereign ratings until 1918); N.R. denotes
Not Rated. A country is “rich” if: for the period of 1888-1987, its GDP per capita is greater than 50 per
cent of the UKvalue; or for the periodof 1988-2013, it is rankedinthe high-income categorybythe World
Bank analytical classifcations. Here “rich” status during default year y depends on the country’s
classifcation during the period y–3 through y–1
Sources: www.ggdc.net/MADDISON/oriindex.htm for per capita GDP; World Bank (2013) for the
World Bank analytical classifcations; Standard and Poor’s (2006) for default events; author’s ratings
database for Moody’s ratings
215
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
3. Extent of CRA failures to anticipate defaults of high-income countries
Several studies research the performances of sovereign ratings and market
indicators. Investigating the interwar sovereign defaults, Flandreau et al. (2011) fnd
that credit ratings outperformed yield-implied ratings (YIRs) at the fve-year
horizon[5]. In contrast, CRA ratings had less predictive power at the one- and
three-year horizons. These results are consistent with Gaillard’s (2011, 2014a)
analysis of the eurozone debt crisis, which shows that credit default swap-implied
ratings of peripheral countries were (on average) lower than their corresponding
CRAratings as early as January 2009 – that is, 16 months prior to the EU/IMF rescue
package for Greece.
This section compares the ability of Moody’s and S&P ratings and bond yields to
anticipate not only Greece’s default in 2012 but also the quasi-defaults that hit
high-income countries between January 1, 2008 and January 1, 2013. These
quasi-defaults refer to the IMF standby loan granted to Iceland in 2008 and the
EFSF/EFSM/ESM fnancial assistance programs requested by Ireland, Portugal
and Spain in 2010, 2011 and 2012, respectively[6].
The frst step in making this comparison is the conversion of bond yields into YIRs.
The countries under study are the sovereigns (both high-income and non-high-income
countries in all regions) that are assigned a credit rating by Moody’s and S&P and for
which a ten-year bond yield is available during the period from January 1, 2008 to
January 1, 2010. All ratings below Baa3/BBB? were relabeled “SG” (i.e. speculative
grade). The objective is to assign the SGclassifcation to those countries whose YIRs are
currently BB?/Ba1, but could be even lower. The comparison of Moody’s and S&P
ratings with their respective YIRs is made for three dates: January 1, 2008, January 1,
2009 and January 1, 2010. Hence, there are six samples.
Estimates of the YIRs were derived using the method developed by Breger et al.
(2003) and refned by Kou and Varotto (2008). For each rating category, a penalty
function p(b) is computed that depends on the value of bond yield boundaries. This
penalty value increases when the bond yield is outside the upper or lower boundaries
that correspond to its rating. The penalty function is defned as follows:
P(b) ?
1
m
?
i?1
m
max (y
i,R
a?1
? b, 0) ?
1
n
?
j?1
n
max (y ? c
j,R
a
, 0)
Where,
R
a
and R
a?1
?two adjacent rating categories, with R
a?1
one notich higher than R
a
;
y
i, R
a?1
?the bond yield of country i with the rating R
a?1
;
y
i, R
a
?the bond yield of country j with the rating R
a
;
m ?total number of countries rated R
a?1
;
n ?total number of countries rated R
a
;
b ?bond yield boundry between R
a
and R
a?1
;
The optimum boundaries are those that minimize the penalty function for each pair of
adjacent ratings. Once the optimum boundaries are obtained, it is possible to derive
YIRs.
JFEP
6,3
216
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Tables II-IV present the Moody’s and S&P ratings and the YIRs derived for
high-income countries that are included in the six samples for the three dates.
There are 26 pairwise observations of high-income countries that fell into default
or quasi-default by January 1, 2013. In 19 cases, CRAs’ ratings were higher than
YIRs: major rating gaps appear for Iceland, Ireland and Spain. In four other cases,
the two risk indicators were about equal; for the remaining three countries, CRAs’
ratings (more precisely, S&P ratings) were actually lower.
Table II.
Moody’s and S&P ratings
and the corresponding
YIRs for high-income
countries as of 1 January
2008
Country Ten-year yield (bps)
Moody’s
rating
YIR for Moody’s
rating
S&P
rating
YIR for S&P
rating
Japan 1.502 Aaa Aaa AA AAA
Switzerland 2.339 Aaa Aaa AAA AAA
Taiwan 2.563 Aa3 Aaa AA? AAA
Singapore 2.670 Aaa Aaa AAA AAA
Hong Kong 3.462 Aa2 Aaa AA AAA
Canada 3.990 Aaa Aa1 AAA AA?
USA 4.032 Aaa Aa1 AAA AA?
Germany 4.324 Aaa Aa1 AAA AA?
Sweden 4.361 Aaa Aa1 AAA AA?
Spain 4.408 Aaa Aa1 AAA AA?
Finland 4.413 Aaa Aa1 AAA AA?
The Netherlands 4.418 Aaa Aa1 AAA AA?
Austria 4.438 Aaa Aa1 AAA AA?
France 4.438 Aaa Aa1 AAA AA?
Denmark 4.443 Aaa Aa1 AAA AA?
Belgium 4.469 Aa1 Aa1 AA? AA?
Slovenia 4.482 Aa2 A1 AA AA?
Ireland 4.509 Aaa A1 AAA AA?
Portugal 4.533 Aa2 A1 AA? AA?
UK 4.566 Aaa A1 AAA AA?
Greece 4.638 A1 A1 A AA?
Italy 4.657 Aa2 A1 A? AA?
Norway 4.693 Aaa A1 AAA A?
The Czech Republic 4.719 A1 A1 A A?
South Korea 5.740 A2 Baa1 A A
Israel 6.150 A2 Baa1 A BBB?
Australia 6.327 Aaa Baa3 AAA BBB?
New Zealand 6.415 Aaa Baa3 AA? BBB?
Iceland 10.225 Aaa SG A? SG
Notes: Countries are listed in order of increasing ten-year yield; “bps” ?basis points; Hungary was
not classifed as a high-income country until 2007, so its ratings and YIRs are not reported in this table;
the samples of Moody’s and S&P ratings consist of 48 and 50 observations, respectively; rows printed
in italics correspond to countries that fell into default or quasi-default during the period fromJanuary 1,
2008 to January 1, 2013
Sources: Author’ s computati ons, www. moodys. com; www. standardandpoors. com;
www.thomsonone.com
217
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The second step of my investigation consists of computing accuracy ratios (ARs). These
ratios are used to assess:
• whether low CRAs ratings and YIRs were assigned to issuers that subsequently
defaulted or quasi-defaulted; and
• whether high CRAs ratings and YIRs were assigned to issuers that did not.
The ARvalues range between ?1 and 1, where 1 represents the maximumaccuracy (i.e.
all defaulters are assigned the lowest rating) and ?1 represents the worst performance
(i.e. all defaulters are assigned the highest rating). The formula for calculating ARs is:
Table III.
Moody’s and S&P ratings
and the corresponding
YIRs for high-income
countries as of 1 January
2009
Country 10-year yield (bps)
Moody’s
rating
YIR for Moody’s
rating
S&P
rating
YIR for S&P
rating
Japan 1.175 Aaa Aaa AA AAA
Hong Kong 1.349 Aa2 Aaa AA? AAA
Taiwan 1.375 Aa3 Aaa AA? AAA
Singapore 2.048 Aaa Aa1 AAA AAA
USA 2.219 Aaa Aa1 AAA AAA
Switzerland 2.228 Aaa Aa1 AAA AAA
Sweden 2.430 Aaa Aa1 AAA AAA
Canada 2.688 Aaa Aa1 AAA AAA
Germany 2.944 Aaa Aa2 AAA AAA
UK 3.017 Aaa Aa2 AAA AAA
Denmark 3.363 Aaa Aa2 AAA AAA
France 3.414 Aaa Aa2 AAA AAA
Finland 3.478 Aaa Aa2 AAA AAA
The Netherlands 3.554 Aaa Aa2 AAA AAA
Belgium 3.778 Aa1 Aa2 AA? AA?
Spain 3.822 Aaa A1 AAA AA
Austria 3.871 Aaa A1 AAA AA
Norway 3.885 Aaa A1 AAA AA
Portugal 3.962 Aa2 A1 AA? AA
Australia 3.990 Aaa A1 AAA A?
South Korea 4.220 A2 A1 A A?
The Czech Republic 4.292 A1 A1 A A?
Italy 4.379 Aa2 A1 A? A?
Ireland 4.457 Aaa A1 AAA A
New Zealand 4.650 Aaa A1 AA? A
Israel 4.680 A1 A1 A A
Slovenia 5.149 Aa2 A1 AA A?
Greece 5.201 A1 A1 A A?
Hungary 8.410 A2 Baa3 BBB BBB
Notes: Countries are listed in order of increasing ten-year yield; “bps” ?basis points; Iceland fell into
quasi-default in 2008 and so was removed from the January 1, 2009 samples; the samples of Moody’s
and S&P ratings consist of 47 and 49 observations, respectively; rows printed in italics correspond to
countries that fell into default or quasi-default during the period fromJanuary 1, 2009 to January 1, 2013
Sources: Author’ s computati ons, www. moodys. com; www. standardandpoors. com;
www.thomsonone.com
JFEP
6,3
218
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
AR ? 2
??
?
R
i
?R
1
,…,R
max
(D
R
? D
R
i?1
)(N
R
? N
R
i?1
)
2DN ?
? 0.5)
?
Where,
D ?total number of defaults and quasi-defaults;
N ?total number of issuers;
R
i
?rating of a given agency;
D
R
i
?total number of defaults and quasi-defaults rated R
i
and less;
N
R
i
?total number of issuers rated R
i
and less;
Table IV.
Moody’s and S&P ratings
and the corresponding
YIRs for high-income
countries as of January 1,
2010
Country Ten-year yield (bps)
Moody’s
rating
YIR for Moody’s
rating
S&P
rating
YIR for S&P
rating
Japan 1.295 Aa2 Aaa AA AAA
Taiwan 1.555 Aa3 Aaa AA? AAA
Switzerland 2.025 Aaa Aa1 AAA AA?
Singapore 2.651 Aaa Aa1 AA AA?
Hong Kong 2.779 Aa2 Aa1 AA? AA?
Germany 3.388 Aaa Aa1 AAA AA?
Sweden 3.398 Aaa Aa1 AAA AA?
The Netherlands 3.565 Aaa Aa1 AAA AA?
Finland 3.587 Aaa Aa1 AAA AA?
France 3.610 Aaa Aa1 AAA AA?
Canada 3.611 Aaa Aa1 AAA AA?
Denmark 3.680 Aaa Aa2 AAA AA
Belgium 3.716 Aa1 Aa2 AA? AA
USA 3.837 Aaa Aa3 AAA AA?
Austria 3.880 Aaa Aa3 AAA AA?
Spain 3.998 Aaa Aa3 AA? AA?
Czech Rep 4.006 A1 Aa3 A A?
UK 4.010 Aaa Aa3 AAA A?
Portugal 4.075 Aa2 Aa3 A? A?
Italy 4.139 Aa2 A1 A? A?
Norway 4.169 Aaa A1 AAA A
Slovenia 4.872 Aa2 A2 AA A
Ireland 4.880 Aa1 A2 AA A
Israel 5.110 A1 A2 A A
South Korea 5.400 A2 A2 A A?
Australia 5.725 Aaa A3 AAA A?
Greece 5.764 A2 A3 BBB? A?
New Zealand 6.125 Aaa Baa1 AA? BBB?
Hungary 8.210 Baa1 Baa3 BBB? BBB?
Notes: Countries are listed in order of increasing ten-year yield; “bps” ?basis points; Iceland fell into
quasi-default in2008 andso was removedfromthe January1, 2010 samples; the samples of Moody’s and
S&P ratings consist of 47 and 49 observations, respectively; rows printed in italics correspond to
countries that fell into default or quasi-default during the period fromJanuary 1, 2010 to January 1, 2013
Sources: Author’ s computati ons, www. moodys. com; www. standardandpoors. com;
www.thomsonone.com
219
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
D
0
?0;
N
0
?0.
Table V compares, for three different dates, the ARs computed for CRA ratings and
those computed for YIRs; these values document that the former are signifcantly lower
than the latter. Not surprisingly, the accuracy of CRAs suffers from the higher ratings
they assigned to Iceland in 2008 and to Ireland and Spain until 2010. Moody’s poor
performance is driven by its belated downgrades of peripheral eurozone countries
(Gaillard, 2014a, 2014b). These results are embarrassing for CRAs because they cast
doubt on howreliably the agencies can identify medium-termrisk. The reported values
also raise the question of whether CRAs have managed to improve their sovereign
rating methodologies since 2010.
4. An updated sovereign rating methodology: the case of Moody’s
The poor performance exhibited during the eurozone debt crisis and the new EU
regulation enacted in 2013[7] have obliged CRAs to revise their assessment of sovereign
credit risk (Moody’s, 2013c; Standard and Poor’s, 2013). This section examines the new
sovereign rating methodology implemented by Moody’s and evaluates whether it would
have led to more accurate ratings if applied at the beginning of the Great Recession.
Moody’s updated methodology is based on four broad rating factors: a country’s
economic, institutional and fscal strengths as well as its susceptibility to event risk.
Each of these factors consists of several subfactors (Table VI). After estimating or
calculating each subfactor, the outcomes for each indicator are mapped to one of 15
ranking categories: very high plus (VH?), very high (VH), very high minus (VH?), high
plus (H?), high (H), high minus (H?), medium plus (M?), medium (M), medium minus
(M?), low plus (L?), low (L), low minus (L?), very low plus (VL?), very low (VL) and
very low minus (VL?)[8]. These mappings are used to determine the score for the
subfactors and also for the broader rating factors.
Next, factors 1 and 2 (economic and institutional strengths) are combined with equal
weight to yield the economic resiliency factor. The combination of this new factor with
factor 3 (fscal strength) yields an assessment of the government’s fnancial strength,
which is fnally combined with factor 4 (susceptibility to event risk) to obtain an
indicative FC credit rating (Table VII). It is worth noting that factor 4 follows a
maximum function. Moody’s explains that “as soon as one area of risk warrants an
assessment of elevated risk, the country’s overall susceptibility to event risk is scored at
that specifc, elevated level”. This approach is innovative in that – unlike previous
methodologies – it does not rely exclusively on a weighting system. Moody’s also insists
Table V.
Accuracy ratios for YIRs
and for ratings assigned
by Moody’s and S&P
January 1, 2008
(fve-year AR)
January 1, 2009
(four-year AR)
January 1, 2010
(three-year AR)
Moody’s ratings ?0.358 ?0.308 ?0.202
YIRs 0.054 0.043 ?0.053
S&P ratings ?0.268 ?0.327 ?0.061
YIRs ?0.028 0.005 ?0.112
Source: Author’s computations
JFEP
6,3
220
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Table VI.
Sovereign rating
methodology implemented
by Moody’s in 2013
Broad rating
factors Rating subfactor
Subfactor
weighting within
factor Subfactor indicators
Factor 1:
economic
strength
Growth dynamics
Scale of the economy
National income
Adjustment factors
50%
25%
25%
1-6 scores
Average Real GDP Growth
t ? 4 to t ? 5
Volatility in Real GDP Growth
t ? 9 to t
WEF Global Competitiveness Index
t
Nominal GDP (US$)
t ? 1
GDP per capita (PPP, US$)
t ? 1
Diversifcation
Credit Boom
Factor 2:
institutional
strength
Institutional
framework and
effectiveness
Policy credibility
and effectiveness
Adjustment factor
75%
25%
1-6 scores
World Bank Government
Effectiveness Index
World Bank Rule of Law Index
World Bank Control of Corruption
Index
Infation Level
t ?4 to t ?5
Infation Volatility
t ?9 to t
Track Record of Default
Factor 3:
Fiscal
strength
Debt burden
Debt affordability
Adjustment factors
50%
50%
1-6 scores
General Government Debt/GDP
t
General Government Debt/Revenue
t
General Government Interest
Payments/Revenue
t
General Government Interest
Payments/GDP
t
Debt Trend
t – 4 to t ?1
General Government FC
Debt/Geneneral Gov ernment Debt
t
Other Public Sector Debt/GDP
t
Public Sector Financial Assets or
SWFs/GDP
t
Factor 4:
susceptibility
to event risk
Political risk
Government
liquidity risk
Banking sector risk
External
vulnerability risk
Maximum function
Maximum function
Maximum function
Maximum function
Domestic Political Risk
Geopolitical Risk
Fundamental Metrics
Market Funding Stress
Strength of Banking System
Size of Banking System
Funding Vulnerabilities
(Current Account Balance ?FDI)/
GDP
t
External Vulnerability Indicator
(EVI)
t ?1
Net International Investment
Position/GDP
t
Notes: FDI ?foreign direct investment; PPP ?purchasing power parity; SWF ?sovereign wealth
fund; WEF ?World Economic Forum
Source: Moody’s (2013c)
221
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
that it may “consider additional factors that are diffcult to measure or that have a
meaningful effect in differentiating credit quality only in some, but not all cases”[9].
Moreover, ratings may “refect circumstances in which the weighting of a particular
factor will be substantially different from the weighting suggested by the scorecard”.
These caveats notwithstanding, the meticulous nature of Moody’s methodology makes
it easier to run back-tests and to estimate what its sovereign ratings would have been
under this new risk assessment regime.
As an example of such back-testing, I can estimate Iceland’s and Greece’s ratings for
January 2008 and January 2009, respectively. These two high-income countries were the
frst to request fnancial assistance (Iceland in October 2008 and Greece in May 2010),
and Greece was the frst sovereign to default in March 2012. The reference years for
Iceland and Greece are, respectively, 2007 and 2008. The primary sources used for this
exercise are Moody’s Statistical Handbook – Country Credit reports published in
November 2007 and November 2008; additional sources are the IMF and World Bank
databases and the World Economic Forum Global Competitiveness reports released in
2007 and 2008, respectively[10].
The results of these back-tests are as follows. For Iceland, the government’s fnancial
strength and susceptibility to event risk are estimated at H? and VH, respectively,
which yields an indicative FC rating of Baa2 – that is, eight notches below the actual
rating in January 2008 (Aaa). The weakness of the Icelandic banking sector is the key
subfactor accounting for the reduced rating[11]. For Greece, the government’s fnancial
strength and susceptibility to event risk are estimated at Hand M?, respectively, which
yields an indicative FC rating of Baa1 – that is, three notches below the actual rating in
January 2009 (A1). The Achilles’ heel of the Greek economy is its current account defcit,
which refects the country’s inadequate net international investment position.
These fndings suggest that the new methodology implemented by Moody’s would
have performed much better in 2008-2009 than did the one it actually applied (Gaillard,
2014b). However, I cannot draw incontrovertible conclusions because all ratings are
Table VII.
Moody’s scorecard
Government’s fnancial strength (Combination of factors 1, 2 and 3)
VH? VH VH? H? H H? M? M M? L? L L? VL? VL VL?
Factor 4:
susceptibility
to event risk
VL? Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3
VL Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3
VL? Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3
L? Aa1 Aa2 Aa3 A1 A2 A3 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1
L Aa1 Aa2 Aa3 A1 A2 A3 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1
L? Aa1 Aa2 Aa3 A1 A2 A3 Baa2 Baa3 Ba1 Ba3 B1 B2 B3 Caa1 Caa2
M? Aa2 Aa3 A1 A2 A3 Baa1 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Caa2
M Aa2 Aa3 A1 A2 A3 Baa1 Baa3 Ba1 Ba2 B1 B2 B3 Caa1 Caa2 Caa3
M? Aa3 A1 A2 A3 Baa1 Baa2 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Caa2 Caa3
H? Aa3 A1 A2 A3 Baa1 Baa2 Ba1 Ba2 Ba3 B2 B3 Caa1 Caa2 Caa3 Caa3
H A1 A2 A3 Baa1 Baa2 Baa3 Ba2 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 Caa3
H? A1 A2 A3 Baa1 Baa2 Baa3 Ba2 Ba3 B1 B3 Caa1 Caa2 Caa3 Caa3 Caa3
VH? A2 A3 Baa1 Baa2 Baa3 Ba1 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 Caa3 Caa3
VH A2 A3 Baa1 Baa2 Baa3 Ba1 Ba3 B1 B2 Caa1 Caa2 Caa3 Caa3 Caa3 Caa3
VH? A3 Baa1 Baa2 Baa3 Ba1 Ba2 B1 B2 B3 Caa1 Caa2 Caa3 Caa3 Caa3 Caa3
Note: Outcomes displayed are the midpoint of the three-rating range (e.g. “Aa1” is used to denote the range Aaa-Aa2)
Source: Moody’s (2013c)
JFEP
6,3
222
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
based, in part, on subjective judgments. With regard to Iceland, for example, Moody’s
(2008) stated in April 2008 that “[(although)] the government has a low direct debt,
contingent liabilities are enormous. The banking sector in Iceland represents the most
burdensome contingent liability”. The agency was doubtless aware of the Icelandic
economy’s fundamental weakness, yet its Aaa rating was maintained until May 2008.
5. Recommendations to enhance sovereign risk assessment
CRAs have consistently failed to anticipate the public debt crises in rich countries. None
of the fve sovereign debt issuers that defaulted or quasi-defaulted during the period
from January 2008 through January 2013 were assigned a rating lower than A1 or A?
by Moody’s and S&P 12 months prior to their request for fnancial assistance. The new
methodologies adopted by each of these CRAs in 2013 addressed important issues that
have burdened European economies – in particular, contingent liabilities arising from
the banking system, declining competitiveness and risk stemming from external
vulnerability. Even so, there are several threats to the creditworthiness of high-income
countries that merit closer examination.
Rapid growth may refect an “overheated” economy and lead to the creation of
asset bubbles; it is therefore crucial to assess the sustainability of growth. For
instance, sovereign risk analysts should be concerned if a country’s main stock
market indices or housing prices grow 2-3 times as fast as GDP for several
consecutive years. More fundamentally, a persistent increase in the ratio of national
wealth to national income may be an evidence of speculation on certain types of
assets (Piketty and Zucman, 2013).
In addition to monitoring the traditional indicators of governance effectiveness, it is
critical to evaluate the extent to which a country’s institutional and political framework
may compromise the consistency of its economic policy and hinder the implementation
of adequate and timely measures to preserve creditworthiness. Governance fexibility
depends on citizens viewing such institutions as legitimate. It depends also on how
effciently the electoral/political system enables the government to secure clear
majorities for implementing its economic policy, on the government’s promptness in
taking fscal and tax measures to address economic problems, on the likelihood that
brutal changes in economic policy will become necessary and on transparency in
policymaking.
The sustainability of exchange rates also merits further investigation. Countries
characterized by a structural current account defcit and weak nonprice
competitiveness can hardly afford a substantial appreciation in their currency. The
fnancial diffculties experienced by peripheral eurozone countries are partly related
to the euro’s high valuation, and meeting this challenge remains an important task
(Deutsche Bank, 2013).
Finally, CRAs should unceasingly monitor economic and fscal dynamics in times of
low risk aversion. During 2003-2007, despite an average annual GDP growth rate
exceeding 4 per cent, Greece was unable to reach a primary fscal surplus. In the
meantime, the country slipped from 35th to 65th in the World Economic Forum Global
Competitiveness rankings. This poor performance should have induced CRAs to lower
Greece’s rating well before the start of the Great Recession. Such a monitoring strategy
is required to achieve more accurate ratings and to avoid major rating reversals in
response to an economic slowdown.
223
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Notes
1. These percentages are based on fgures published by World Development Indicators.
2. Since July 2013, the EFSF has no longer engaged in new fnancing programs. Thereafter, the
European Stability Mechanism (ESM) became the permanent venue responding to new
requests for fnancial assistance.
3. Values are given in US dollars (siteresources.worldbank.org/DATASTATISTICS/
Resources/OGHIST.xls).
4. Although this paper studies not high-income but rather rich countries (as just defned),
hereafter the terms “high-income”, “rich” and “wealthy” are used interchangeably.
5. AYIRamounts to the transformation of a bond yield into a credit rating, as described in what
follows.
6. The fnancial assistance agreement with Spain was designed to cover capital shortfalls in
Spanish banks.
7. Regulation (EU) No. 462/2013 of the European Parliament and of the Council of May 21, 2013
amending Regulation (EC) No. 1060/2009 on credit rating agencies require a greater
transparency of sovereign rating methodologies and a detailed evaluation of any changes
made to the quantitative and qualitative assumptions (and their relative weight) that underlie
a revised rating.
8. The move from one ranking category to another is contingent upon precise thresholds
stipulated by Moody’s.
9. For instance, Avendaño et al. (2011) show that CRAs take workers’ remittance fows into
account when rating low- and middle-income countries that are small (e.g. Central American
sovereigns).
10. Details are available from the author upon request.
11. For an examination of the roles of capital rules and bank business models in determining the
safety of banks, see Blundell-Wignall and Roulet (2013).
References
Apostolides, A. (2013), “Beware of German gifts near elections: howCyprus got here and why it is
currently more out than in the Eurozone”, Capital Markets Law Journal, Vol. 8, No. 3.
Avendaño, R., Gaillard, N. and Nieto-Parra, S. (2011), “Are workers’ remittances relevant for credit
rating agencies?”, Review of Development Finance, Vol. 1 No. 1.
Blundell-Wignall, A. and Roulet, C. (2013), “Bank business models, capital rules and structural
separation policies”, Journal of Financial Economic Policy, Vol. 5 No. 4.
Breger, L., Goldberg, L. and Cheyette, O. (2003), “Market implied ratings”, Barra Credit Series,
Research Insights.
Corporation of Foreign Bondholders (1904), “Thirty-frst Annual Report of the Council of the
Corporation of Foreign Bondholders”, London.
Corporation of Foreign Bondholders (1943), “Seventeenth Annual Report of the Council of the
Corporation of Foreign Bondholders”, London.
Corporation of Foreign Bondholders (1966), “Ninety-third Annual Report of the Council of the
Corporation of Foreign Bondholders”, London.
Deutsche Bank (2013), “Where is the FX ‘pain threshold’?”, Global Markets Research – Macro.
JFEP
6,3
224
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Flandreau, M., Gaillard, N. and Packer, F. (2011), “To err is human: US rating agencies and the
interwar foreign government debt crisis”, European Review of Economic History, Vol. 15
No. 3.
Flores, J.H. (2010), “Competition in the underwriting markets of sovereign debt: the baring crisis
revisited”, Law and Contemporary Problems, Vol. 73 No. 4.
Foreign Bondholders Protective Council (1945), “Report 1941 through 1944”, New York, NY.
Gaillard, N. (2011), A Century of Sovereign Ratings, Springer, New York, NY.
Gaillard, N. (2014a), “What is the value of sovereign ratings?”, German Economic Review, Vol. 15
No. 1.
Gaillard, N. (2014b), “How and why credit rating agencies missed the Eurozone debt crisis”,
Capital Markets Law Journal, Vol. 9 No. 2.
Kou, J. and Varotto, S. (2008), “Timeliness of spread implied ratings”, European Financial
Management, Vol. 18 No. 3.
Moody’s Investors Service (2008), “Iceland”.
Moody’s Investors Service (2013a), “Sovereign default and recovery rates, 1983-2012”.
Moody’s Investors Service (2013b), “Moody’s says Cypriot debt exchange amounts to a default”.
Moody’s Investors Service (2013c), “Sovereign bond ratings”.
Piketty, T. and Zucman, G. (2013), “Capital is back: wealth-income ratios in rich countries
1700-2010”, Working paper, Paris School of Economics, Paris.
Rieffel, L. (2003), Restructuring Sovereign Debt – The Case for Ad Hoc Machinery, Brookings
Institution Press, Washington, DC.
Sachs, J. (1985), “External debt and macroeconomic performance in Latin America and East Asia”,
Brookings Papers on Economic Activity, Vol. 16, No. 2, pp. 523-573.
Standard & Poor’s (2006), “Sovereign defaults at 26-year low, to show little change in 2007”.
Standard & Poor’s (2013), “Sovereign government rating methodology and assumptions”.
World Bank (2013), “World bank analytical classifcations”.
About the author
Norbert Gaillard is a French Economist and an Independent Consultant (norbertgaillard.com). He
graduated from Sciences Po Paris and Princeton University. His PhD dissertation, completed in
2008, dealt with sovereign rating methodologies. He has served as consultant to the International
Finance Corporation, the World Bank, the State of Sonora (Mexico), the OECD and the European
Parliament. He has also served as a visiting professor at the Graduate Institute (Geneva) and as a
Euromoney Country Risk expert. His main areas of expertise are public debt and sovereign risk,
local government debt and subnational risk and credit rating agencies. He is among the youngest
researchers to have his biography included in the Who’s Who in the World. He has written several
research articles and book chapters on public debt and credit rating agencies. He has published
two books: Les Agences de Notation (La Découverte, Paris, 2010) and A Century of Sovereign
Ratings Springer, New York, 2011. Norbert Gaillard can be contacted at:
[email protected]
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
225
Assessing
sovereign risk
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
4
9

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_140085407.pdf
 

Attachments

Back
Top