sovereign credit ratings in the european union: a model ...sovereign credit ratings has been stable...
TRANSCRIPT
Sovereign Credit Ratings in the
European Union: A
Model-Based Fiscal Analysis
Vito Polito (University of Bath)
Mike Wickens (University of York)
Bratislava, June 1, 2015
Introduction
� Motivation - EU downgrades in 2011; EC 2011 pro-posal; EC 2013 new regulations; EU-based CRA
� Contribution - Black and Scholes model, GBC, modelfor debt forecasts and debt limits
� Findings - earlier downward shift in the credit ratingdistribution, �scal consolidation in the EU
� Outline - historic credit rating, methodology, debtlimit, model-based credit rating
� Previuos work on U.S. and the U.K.
Historic credit rating
Table 1: Rating scales adopted by the three main CRAs.
Moody�s Fitch S&P Credit quality
Investment Aaa AAA AAA Prime
grade (I.G.) Aa1 AA+ AA+ High
Aa2 AA AA grade... ... ...
Baa3 BBB- BBB-
Speculative Ba1 BB+ BB+
grade (S.G.) Ba2 BB BB Speculative
Ba3 BB- BB-... ... ...
Ca CC CCC for recovery
C C
DDD, DD, D D In default
Source: Authors�classi�cation based on Gaillard (2012)
Facts about historic credit ratings of EU countries
1. The sovereign credit ratings of the EU14 countriestaken as a group has been higher than those of othercountries.
2. The cross-section distribution of the EU14 countriessovereign credit ratings has been stable within theinvestment grade at least until 2010.
3. Sharp changes in this distribution have occurred, par-ticularly since 2010.
4. Fluctuations in EU14 sovereign credit ratings haveincreased as the ratings have fallen.
5. Changes in the sovereign credit ratings of several EU14 countries appear to be unrelated to the market�sperception of the probability of sovereign default.
Table 2: Distribution of historic sovereign credit ratingsof EU14 countries at selected dates.
90 95 00 05 06-08 09 10 11 12
Aaa 50% 36% 57% 71% 71% 64% 57% 57% 50%
Aa 36% 43% 36% 21% 21% 29% 21% 7% 14%
A 7% 14% 7% 7% 7% 7% 7% 14% 7%
Baa 7% 7% 0% 0% 0% 0% 7% 0 7%
Ba 0% 0% 0% 0% 0% 0% 7% 14% 14%
B 0% 0% 0% 0% 0% 0% 0% 0 0%
C 0% 0% 0% 0% 0% 0% 0% 7% 7%
Share of investment grade (in percentage)
EU14 100 100 100 100 100 100 93 79 79
ARC 86 78 59 63 n.a. n.a. 61 n.a. 60
Notes: ARC=All Rated Countries in a speci�c year; n.a.=not available.
Source: Moody�s (2012)
1990 95 00 05 10 12:4C
CaaB2
Ba3Ba1
Baa2A3A1
Aa2Aaa
Rat
ing
AUS, GER, NET , UK
1990 95 00 05 10 12:4C
CaaB2
Ba3Ba1
Baa2A3A1
Aa2Aaa
Rat
ing
DEN FIN FRA
1990 95 00 05 10 12:4C
CaaB2
Ba3Ba1
Baa2A3A1
Aa2Aaa
Rat
ing
BEL IT A SWE
1990 95 00 05 10 12:4C
CaaB2
Ba3Ba1
Baa2A3A1
Aa2Aaa
Rat
ing
GRE IRE POR SPA
Historic sovereign credit rating of EU14 countries,1990-2012.
08 09 10 11 12 130
160AU S
08 09 10 11 12 130
350BEL
08 09 10 11 12 130
150D EN
08 09 10 11 12 130
100F IN
08 09 10 11 12 130
180F R A
08 09 10 11 12 130
100GER
08 09 10 11 12 130
15000GR E
08 09 10 11 12 130
1200IR E
08 09 10 11 12 130
500ITA
08 09 10 11 12 130
140N ET
08 09 10 11 12 130
1600POR
08 09 10 11 12 130
500SPA
08 09 10 11 12 130
160SW E
08 09 10 11 12 130
180U K
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
08 09 10 11 12 13C
Baa3
Aaa
C D SR ating
Sovereign credit ratings and 5-year credit default swapprices for EU14 countries, 14/12/2007 - 22/03/2013.
Methodology
� Adapt to GBC Black�Scholes (1973) formula for prob-ability of exercising a European-style call option.
� This gives sovereign default probabilities at di¤erenttime horizons.
� Use CRAs tables to convert into credit rating.
� Debt limit: State-dependent measures from a DSGEmodel of the economy. Time series of the maximumborrowing capacity of an economy.
� (1) Default probability, (2) mapping credit rating,(3) debt-GDP forecasts and volatility; (4) debt limit.
Methodology (Default probability)
bt+hyt+h
= �hXj=1
"�js=1
�1 + �t+s
� dt+jyt+j
#+�hs=1
�1 + �t+s
� btyt
default threshold (debt limit):bt+hyt+h
pt;t+h = pt+h�1� pt+h�1
� �1� pt+h�2
�::: (1� pt+1)
pt+h = Pr
bt+hyt+h
� bt+hyt+h
j�t!
distance-to-default: DDt+h =Etbt+hyt+h
� bt+hyt+h
��;t+h
Methodology (Credit rating)
Rating Cumulative default probabilityLong-term 1-year 5-year 10-year averageAaa 0.000 0.000 0.000 0.000Aa1 0.008 0.215 0.715 0.265Aa2 0.015 0.430 1.429 0.529Aa3 0.023 0.646 2.144 0.794A1 0.030 0.861 2.858 1.058A2 0.038 1.076 3.573 1.323... ... ... ... ...B3 2.724 11.158 18.541 10.887Caa 11.142 21.083 26.005 19.711Ca 19.561 31.008 33.469 28.534C 27.979 40.933 40.933 37.358
Source: www.moodys.com; and authors�calcs
Table 3: Mapping from cumulative default probabilitiesto sovereign credit rating.
Debt limit
� Ad hoc?
� agnostic (ability/willingness)
� Reinhart, Rogo¤ and Savastano (2003): likely tovary from country to country, as well as over time
� From a structural model of the economy
� Open economy RBC with distortionary taxationon income from capital, labor and consumption
� Similar to Trabandt and Uhlig (2011)
� Financial ability only
� Anticipated or unanticipated �scal policy
Debt limit (model)
� Preferences
U0 = E0
1Xt=0
�tu (ct; 1� nt)
� Households budget constraint
(1� �nt )wtnt +�rkt � �
� �1� �kt
�kt�1
+kt�1 + [1 + (1� �t) rt � �t] bDt�1+zt + (1 + r
�t ) stft�1
= (1 + �ct)ct + kt + bDt + stft
� Consumption
ct =
"��cHt�1�1� + (1� �) �cFt �1�1�
# 1
1�1�
� Production
yt = k�t (Atnt)
1��
� Government budget constraint
gt + (1� �t) rtbt�1 + (1� �t) bt�1 + zt= �ctct + �
nt wtnt + �
kt
�rkt � �
�kt�1 + bt
� BoP
stft�bFt = xt+(1 + r�t ) stft�1�[1+(1� �t) rt��t]bFt�1
� Economy constraint
yt = ct + gt + kt � (1� �) kt�1 + xt
Debt limit (model)
Stationary equilibrium solution
b
y
IGBCL
=1
r
8>><>>:�c�
�1'k � 1
�+ �n (1� �)+
�k�
"1� �
���1�11��k + �
��1#� gy �
zy
9>>=>>;b
y
NDL
=1
r
8>><>>:�c�
�1'k � 1
�+ �n (1� �)+
�k�
"1� �
���1�11��k + �
��1#9>>=>>;
b
y
FL
=1
r
8>><>>:�c�
�1'k � 1
�+ �n;max (1� �)+
�k;max�
"1� �
���1�11��k;max + �
��1#� (g+z)
y
9>>=>>;b
y
MDL
=1
r
8>><>>:�c�
�1'k � 1
�+ �n;max (1� �)+
�k;max�
"1� �
���1�11��k;max + �
��1#9>>=>>;
r =r� + �1� �
� Related literature: Davig, Leeper and Walker (2010);Trabandt and Uhlig (2011)
Debt limit (model)
� Perturbation/projection method?
� Monte Carlo Markov Chain algorithm
� Based on the simulation method for solving ra-tional expectations models
� Judd (1998), Coleman�s (1991)
� Bi (2011) for FL in industrialized countries
� Feed in time-varying mean and volatility of (i) the ra-tio of government expenditures to GDP, (ii) the ratioof transfers to GDP, (iii) the shocks to technologicalprogress and (iv) the actual tax rates
� Country-speci�c �
9 5 0 0 0 5 0 8 1 2
2 0
4 0
A US
9 5 0 0 0 5 0 8 1 2
2 0
4 0
B E L
9 5 0 0 0 5 0 8 1 2
2 0
4 0
DE N
9 5 0 0 0 5 0 8 1 2
2 0
4 0
FIN
9 5 0 0 0 5 0 8 1 2
2 0
4 0
FRA
9 5 0 0 0 5 0 8 1 2
2 0
4 0
G E R
9 5 0 0 0 5 0 8 1 2
2 0
4 0
G RE
9 5 0 0 0 5 0 8 1 2
2 0
4 0
IRE
9 5 0 0 0 5 0 8 1 2
2 0
4 0
IT A
9 5 0 0 0 5 0 8 1 2
2 0
4 0
NE T
9 5 0 0 0 5 0 8 1 2
2 0
4 0
P O R
9 5 0 0 0 5 0 8 1 2
2 0
4 0
S P A
9 5 0 0 0 5 0 8 1 2
2 0
4 0
S W E
9 5 0 0 0 5 0 8 1 2
2 0
4 0
UK .
E X P
T RA
M A X RE V
A CT RE V
Components of the theory-based debt limits for EU14countries, 1995:4-2012:4. All variables are as a
proportion to GDP.
95 00 05 08 120
5AU S
95 00 05 08 120
5BEL
95 00 05 08 120
5D EN
95 00 05 08 120
5F IN
95 00 05 08 120
5F R A
95 00 05 08 120
5GER
95 00 05 08 120
5GR E
95 00 05 08 120
5IR E
95 00 05 08 120
5ITA
95 00 05 08 120
5N ET
95 00 05 08 120
5PO R
95 00 05 08 120
5SPA
95 00 05 08 120
5SW E
95 00 05 08 120
5U K.
IGBC LF LD ebt GD P
IGBCL, FL and debt-GDP ratio in EU14 countries,1995:4 - 2012:4.
1 0 1 2 3 4 50
5
1 0
A U S
1 0 1 2 3 4 50
5
1 0
B E L
1 0 1 2 3 4 50
5
1 0
D E N
1 0 1 2 3 4 50
5
1 0
F I N
1 0 1 2 3 4 50
5
1 0
F R A
1 0 1 2 3 4 50
5
1 0
G E R
1 0 1 2 3 4 50
5
1 0
G R E
1 0 1 2 3 4 50
5
1 0
I R E
1 0 1 2 3 4 50
5
1 0
I T A
1 0 1 2 3 4 50
5
1 0
N E T
1 0 1 2 3 4 50
5
1 0
P O R
1 0 1 2 3 4 50
5
1 0
S P A
1 0 1 2 3 4 50
5
1 0
S W E
1 0 1 2 3 4 50
5
1 0
U K .
a v e ra g e 2 0 0 1 2 0 0 72 0 1 0 2 0 1 2
State-dependent probability density function of the FLof EU14 countries at selected dates.
Methodology (debt-GDP forecasts and volatility)
� Reduced-form VAR model
� Time-varying parameters
� Rolling-window estimation
Methodology (debt-GDP forecasts and volatility)
Symbol Description AcronymsCountry-speci�c variables
btyt
Debt-GDP ratio DEBTdtyt
Primary de�cit-GDP ratio DEF t Growth rate real GDP GDP�t In�ation rate INFrst Short-term interest rate IRSrlt Long-term interest rate IRLet Real exchange rate depreciation EXCxtyt
Current account-GDP CAC�ot Oil in�ation rate OIL
Table 4: Variables included in the ROVAR model
� Data: 1975:2 to 2012:4; from 1977:2 for Portugal
� Countries: AUS, BEL, DEN, FIN, FRA, GER, GRE,IRE, ITA, NET, POR, SPA, SWE, UK
9 5 0 0 0 5 1 0 1 20
0 . 1
0 . 2
0 . 3
0 . 4A U S
9 5 0 0 0 5 1 0 1 26 0
7 0
8 0
9 0
1 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1B E L
9 5 0 0 0 5 1 0 1 25 0
1 0 0
1 5 0
9 5 0 0 0 5 1 0 1 20
0 . 20 . 40 . 60 . 8
1D E N
9 5 0 0 0 5 1 0 1 202 04 06 08 01 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1F I N
9 5 0 0 0 5 1 0 1 24 0
6 0
8 0
9 5 0 0 0 5 1 0 1 20
0 . 2
0 . 4F R A
9 5 0 0 0 5 1 0 1 25 0
1 0 0
1 5 0
9 5 0 0 0 5 1 0 1 20
0 . 10 . 2
0 . 30 . 40 . 5
G E R
9 5 0 0 0 5 1 0 1 25 06 07 0
8 09 01 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1G R E
9 5 0 0 0 5 1 0 1 20
1 0 0
2 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1
1 . 5
2I R E
9 5 0 0 0 5 1 0 1 20
5 0
1 0 0
1 5 0
2 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1I TA
9 5 0 0 0 5 1 0 1 21 0 0
1 2 0
1 4 0
9 5 0 0 0 5 1 0 1 20
0 . 20 . 40 . 60 . 8
1N E T
9 5 0 0 0 5 1 0 1 25 06 07 08 09 01 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1P O R
9 5 0 0 0 5 1 0 1 25 0
1 0 0
1 5 0
9 5 0 0 0 5 1 0 1 20
0 . 20 . 40 . 60 . 8
1S P A
9 5 0 0 0 5 1 0 1 202 04 06 08 01 0 0
9 5 0 0 0 5 1 0 1 20
0 . 20 . 40 . 60 . 8
1S W E
9 5 0 0 0 5 1 0 1 202 04 06 08 01 0 0
9 5 0 0 0 5 1 0 1 20
0 . 5
1U K .
9 5 0 0 0 5 1 0 1 20
1 0 0
2 0 0
S t . d e v . 1 q u a rt e r a h e a d f o re c a s t e rro r
A c t u a l d a t a
Debt-GDP ratio in EU14 countries, 1995:1-2012:4:actual observation (solid line) and standard deviation of1-period ahead forecast error (dotted line) from ROVAR
model.
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
A U S
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
B E L
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
D E N
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
F IN
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
F R A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
G E R
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
G R E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
IR E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
IT A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
N E T
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
P O R
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
S P A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
S W E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
U K .
IG B C L
F L
M D L
h i s to r i c
Model-based (5-year horizon) and historic credit ratingsin EU14 countries, 1995:4-2012:4.
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
A U S
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
B E L
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
D E N
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
F I N
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
F R A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
G E R
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
G R E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
I R E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
I T A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
N E T
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
P O R
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
S P A
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
S W E
9 5 0 0 0 5 0 8 1 2C
B a a 3
A a a
U K .
m o d e l
h i s t o r i c
Model-based (5-year ahead) and historic credit ratingsfor EU14 countries, 1995:4 - 2012:4. Debt limit rangesfrom IGBCL to FL. Dotted lines denote con�dence
bands.
0 8 0 9 1 0 1 1 1 2 1 30
1 6 0
A U S
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
3 5 0
B E L
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 5 0
D E N
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 0 0
F I N
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 8 0
F R A
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 0 0
G E R
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 5 0 0 0
G R E
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 2 0 0
I R E
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
5 0 0
I T A
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 4 0
N E T
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 6 0 0
P O R
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
5 0 0
S P A
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 6 0
S W E
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
0 8 0 9 1 0 1 1 1 2 1 30
1 8 0
U K
0 8 0 9 1 0 1 1 1 2 1 3C
B a a 3
A a a
C D S
C r e d i t r a t i n g ( H i s t o r i c )
C r e d i t r a t i n g ( M o d e l )
Sovereign credit ratings (historic and model-based) and5-year credit default swap prices of EU14 countries,
14/12/2007 - 22/03/2013.
Table 5: Model-based and historic sovereign credit ratingof EU14 countries, summary statistics.
Average credit rating
Model Historic
1995-2012 2008-2012 1995-2012 2008-2012
AUS Aaa Aaa Aaa Aaa
BEL Aaa Aaa Aa1 Aa1
DEN Aaa Aaa Aaa Aaa
FIN Aaa Aaa Aaa Aaa
FRA Aaa Aaa Aaa Aaa
GER Aaa Aaa Aaa Aaa
GRE C C Baa1 Ba1
IRE Aa1 Aa2 Aa1 A1
ITA Ba2 Ba3 Aa3 Aa3
NET Aaa Aaa Aaa Aaa
POR Aa3 Baa2 Aa3 A3
SPA A1 Aa2 Aa1 Aa2
SWE Aaa Aaa Aa1 Aaa
UK Aa1 Aa3 Aaa Aaa
Credit rating changes
Model Historic
1995-2012 2008-2012 1995-2012 2008-2012
Total 168 98 40 24
Downgrades 82 61 24 24
Notes: Authors�calculations based on data in Figure 12.
Table 6: Distribution of the model-based sovereign creditrating of EU14 countries at selected dates based on themedian value rating when the debt limit ranges betweenFL and IGBCL.
00 05 06 07 08 09 10 11 2012
Aaa 79% 86% 86% 79% 50% 64% 50% 50% 50%
Aa 7% 0% 0% 14% 36% 7% 21% 21% 14%
A 0% 0% 0% 0% 0% 7% 7% 7% 14%
Baa 7% 7% 7% 0% 0% 0% 0% 0% 0%
Ba 0% 0% 0% 0% 7% 14% 7% 7% 7%
B 0% 0% 0% 0% 0% 0% 7% 0% 7%
C 7% 7% 7% 7% 7% 7% 7% 14% 7%
IG 93% 93% 93% 93% 86% 79% 79% 79% 79%
Notes: IG=Investment grade. Source: Authors�calculations.