1708 Raport DB PDF

September 2011
Company
Global Research
Global Banking Sector – Credit quality in a deleveraging world
Deutsche Bank Global Markets Research
A
Assiiaa--P
Paacciiffiicc FFooccuuss
TTrraaccyy YYuu,, YYoosshhiinnoobbuu YYaam maaddaa,, JJaam
meess FFrreeeem
maann,,
CChhrriiss EElllleerrttoonn aanndd
GGlloobbaall BBaannkkiinngg S Seeccttoorr TTeeaam
m
Theme
Against a background of 30%-plus falls in bank share prices around the world and growing fears of a severe blow to the European bank sector in the event of a sovereign
debt default, we have developed a danger map and stress testing screens to look at the resilience of loan portfolios within different countries’ banking systems and the
individual banks within them. In this analysis we examine the prospects for post crisis scenarios through the perspective of other severe shocks to international or
domestic banking systems, including the Latin American debt crisis of the 1980's, and the debt deflation crises of Sweden in 1990 and Japan, Thailand and Hong Kong
from 1997 onward.
Deutsche Bank AG/Hong Kong
All prices are those current at the end of the previous trading session unless otherwise indicated. Prices are sourced from local exchanges via Reuters, Bloomberg and
other vendors. Data is sourced from Deutsche Bank and subject companies. Deutsche Bank does and seeks to do business with companies covered in its research
reports. Thus, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report
as only a single factor in making their investment decision. DISCLOSURES AND ANALYST CERTIFICATIONS ARE LOCATED IN APPENDIX 1. MICA(P) 146/04/2011.
26 September 2011 Banks Global FITT - Banking
Research Team Contents

UK/Europe Adam Chaim Jee Hoon Park James Wang
+1-212-250-2966 +82 2 316 8908 +61-2-8258-2054
Table of Contents
Matt Spick
+44(20)754-57895 [email protected] [email protected] [email protected] Summary.............................................................................. 3
[email protected] Noah Gunn Judy Zhang Executive summary ............................................................. 4
+1-212-250-7970 +852 2203 6193 CEEMEA Credit growth and leverage.............................................. 12
Jason Napier Credit quality ..................................................................... 17
[email protected] [email protected] Dan Harverd
+44(20)754-74433 Testing for a recessionary credit cycle ............................ 23
+972-3-710-2019
[email protected] David Ho Clara Jung Crisis . . . what crisis?........................................................ 28
[email protected]
+1-212-250-4424 +82 2 316 8835 Country Sections ............................................................... 36
Carlos Berastain
[email protected] [email protected] Bob Kommers United States ..................................................................... 37
+34-91-3355-971
+7-495-933-9223 Australia ............................................................................. 40
[email protected] Matt Klein Abhishek Jain
[email protected] Japan .................................................................................. 43
+1-212-250-3088 +91 124 4622338
Carolyn Dorrett Hong Kong ......................................................................... 46
[email protected] [email protected] Kazim Andac
+44-20-7547-3133 UK ....................................................................................... 49
+90-212-319-0315
[email protected] Robert Placet Raymond Kosasih Nordics ............................................................................... 53
[email protected] France ................................................................................. 57
+1-212-250-2619 +62 21 318 9525
Alexander Hendricks Germany ............................................................................ 59
[email protected] [email protected] Ryan Ayache
+49-69-9104-5864 Greece ................................................................................ 61
+971-4-428-3781
[email protected] Chris Voie Arinta Harsono Italy..................................................................................... 64
[email protected]
+1-212-250-5085 +62 21 318 9519 Spain .................................................................................. 67
David Lock
[email protected] [email protected] Stefan Swanepoel Israel ................................................................................... 71
+44-20-7541-1521
LATAM +27-11-775-7369 Brazil .................................................................................. 74
[email protected] Sukrit Khatri
[email protected] Mexico................................................................................ 80
Mario Pierry +852 2203 5927
Paola Sabbione Russia ................................................................................. 83
+55-11-2113-5177 [email protected] Rahul Shah
+39-02-8637-9704 India.................................................................................... 87
[email protected] +971 -4 4283-261
[email protected] Manish Shukla China .................................................................................. 90
[email protected]
+91 22 6658 4211 Turkey ................................................................................ 93
Jan Wolter Tito Labarta
[email protected] Marcin Jablczynski Indonesia ........................................................................... 96
+46-8-463-5519 +1-212-250-5944 Malaysia ............................................................................. 99
+44-22-579-8733
[email protected] [email protected] Rafael Garchitorena Thailand ........................................................................... 102
[email protected]
+63 2 894 6644 South Korea ..................................................................... 105
Dimitris Giannoulis Marcelo Cintra
[email protected] Hilal Varol Poland .............................................................................. 108
+302-10-725-6174 +55-11-2113-5095
+90-212-319-0332 Crisis case studies ........................................................... 111
[email protected] [email protected] JAPAN [email protected] The Latin American Debt Crisis ...................................... 112
SebastianYoshida Yoshinobu Yamada Japan: collapse of bubble economy and persistence of
Asia
+44-20-7545-6489 +81-3-5156-6754 deflation ........................................................................... 115
Tracy. Yu
Sebastian [email protected] [email protected] Hong Kong: property bubble .......................................... 117
+852 2203 6186 Argentina’s financial crisis .............................................. 120
USA [email protected] AUSTRALIA Ireland: The Celtic Credit Tiger ...................................... 124
Matthew.O’connor Andrew Hill James Freeman Australian Case Study..................................................... 132
+1(212)250-8489 +65 6423 8507 +61-2-8258-2492 Sweden: Deregulation and macro shocks behind 1990s
bank crisis ........................................................................ 138
[email protected] [email protected] [email protected]
US Sub-Prime Crisis ........................................................ 140
Michael Carrier Sophia Lee Andrew Triggs South Korea: the credit card bubble .............................. 146
+1(212)250-6600 +852 2203 6226 +61-2-8258-2378
[email protected] [email protected] [email protected] The authors of this report wish to acknowledge the
contributions made by Sudhanshu Gaurav of
Head of Global Sector Product: Evalueserve, a third-party provider to Deutsche Bank of
Richard Beaurepaire offshore research support services.
Page 2 Deutsche Bank AG/Hong Kong

Fundamental theme: Charting a danger map for a crisis prone and credit troubled world
Against a background of 30%-plus falls in bank share prices around the world and growing fears of a severe blow to the European bank sector in the event of a
sovereign debt default, we have developed a danger map and stress testing screens to look at the resilience of loan portfolios within different countries’ banking
systems and the individual banks within them. In this analysis we examine the prospects for post crisis scenarios through the perspective of other severe shocks to
international or domestic banking systems, including the Latin American debt crisis of the 1980's, and the debt deflation crises of Sweden in 1990 and Japan,
Thailand and Hong Kong from 1997 onward.
THEMATIC
INDUSTRY Factors/Drivers Key THINKING
WINNERS AND UNDERPERFORMERS
1. Credit growth trends are divergent: Since 1. Downgrades to global growth assumptions are Valuations at least for now discount a gloomy
2007/08 there have been three clear trends: (i) occurring at a time when total debt to GDP ratios although not an extreme outcome. But in this
private sector credit growth within developed are close to record highs. Markets may be environment we expect cost of capital to remain
economies has been grinding down; (ii) overall overestimating loan growth rates in 2012E and very high. The Eurozone crisis poses a double
debt to GDP ratios have risen sharply, courtesy of 2013E. jeopardy for the global bank sector first through
large scale issuance of government debt; (iii) loan potential losses from government bond holdings
2. Bad debt charges drive earnings estimates: in
growth in emerging markets has accelerated. and secondly via the negative feedback loop
developed markets about 90% of 2011E earnings
2. Credit quality in developed markets is fragile: growth and 35% of 2012E growth comes from from financial markets and consumer and
European banks particularly have high levels of lower forecast bad debt charges. Quite modest business confidence onto the key variables of
non-performing loans, and US loan portfolios are adjustments to credit quality assumptions would volume growth and loan quality.
to a large extent built around residential and have a significant impact on estimates. Our danger map suggests Japan, Australia,
commercial real estate. In most emerging Hong Kong and the Nordics are the least risky
3. The US performs better than Europe: On our
markets, credit quality metrics are so good, we countries within the developed markets. The US
screens, if the credit cycle turns down, the US
think they can only move in one direction. and UK come out with slightly below average risk
banking industry outperforms Europe, where a
3. The system can cope with a severe credit severe cycle puts much of the system in loss. scores on macro factors. With the exception of
cycle Because capital ratios are high and pre- Emerging market bank sectors, with higher pre- Germany, which has the most attractive macro
provision profitability is robust, we find that on provision profit margins, remain quite profitable. score within the Eurozone but unattractive
paper at least, bank sectors can cope with a industry fundamentals, the Eurozone countries
4. But there are puzzling variations: While the
normal (i.e. non-Sovereign) but severe credit unsurprisingly have the highest danger scores.
private sectors of the US and UK appear to have
cycle. Emerging markets score a little lower on macro
made an immediate adjustment, private sector
4. At one level it’s a classic post-crisis debt to GDP ratios continue to rise in the Eurozone risk and have more attractive industry
formbook: Looking back at severe banking and from quite a high base. The period of fundamentals although much higher valuations.
crises, we observe 4 common features over the adjustment may be a protracted one: in Japan, Their problems look far more manageable and
4 following years: very slow real GDP growth, Thailand and Hong Kong, the era of mean there are more policy options for them. Bank
an average 150% increase in government debt, reversion stretched out over 10 years. sectors in emerging Asia look least risky.
a contraction in domestic loan markets, and a We prefer lower beta stocks with strong
5. The possibility of renewed crisis is never
reversion towards mean in private sector debt to capital, above-average pre-provision profitability
remote in this environment; we find cross border
GDP ratios. and superior asset quality metrics: WFC, JP
exposures to the periphery public sectors modest
5. But on a different level it’s a “recovery relative to the 1980’s LatAm crisis, but a potential Morgan, Bank of China, Siam Commercial Bank,
environment” like no other, at least since the breakup of the currency adds dimensions to ANZ, Swedbank, Barclays, SMFG, PKO Bank
1930’s: i.e. record low interest rates, record high balance sheet destruction which are difficult to Polski.
private sector plus government debt levels to model.
GDP and a real possibility of serial sovereign
default.
Deutsche Bank AG/Hong Kong Page 3

Executive summary How to make money in this deleveraging environment

It is axiomatic first that it is very difficult to make money in financial stocks
during a period in which the market anticipates a financial crisis; and second
Credit quality in a deleveraging world that such periods can be productive for long-term investors, providing there is
The dramatic decline in bank share prices over the last three months has taken no recapitalization requirement, as stocks tend to price in a very high cost of
place against a background of downward revisions to global growth estimates equity before, during and in the aftermath of a crisis, which subsequently
and deterioration both in the economic reality (lower growth/higher fiscal declines. Our danger map, which we discuss below, suggests there are parts
deficits) of the European Sovereign Debt Crisis and the political consensus on of the world where bank sectors are not particularly risky. These include
how to deal with or contain it. The bank sector is unusually sensitive to the Japan, the Nordic countries, Australia and Germany in developed markets and
economic outlook firstly because the combination of private sector debt to Thailand, Malaysia, Indonesia and Mexico in emerging markets. The danger
GDP and government debt to GDP stands at all time highs, and secondly map also suggests that the Eurozone countries generally are not attractive
because the assumption that credit quality improves accounts for around 90% even in the absence of a crisis (see Matt Spick’s report European banks
of estimated 2011 earnings growth and 35% of 2012 earnings growth in Strategy: Ex-growth and challenged: a bleak outlook for banks 24th August
developed markets. Whatever the ultimate outcome of the Eurozone crisis, 2011), and we find it surprising that the deleveraging process in Europe is so
the bottom line to us is that the problems in the periphery economies, Italy slow relative to the US and UK.
and Spain, are potentially quite to very negative for the GDP outlook, may put
further pressure on normal asset quality measures, and possibly set the scene We think that within developed markets there will be long-term winners in this
for a “super severe” downturn in credit quality with unquantifiable spill-over environment. These include the names that have the combination of financial
effects. strength and strategic/geographic positioning to take market share or extend
their foot print, or which simply have the capital strength to pay dividends and
Figure 1: Private and Public Sector Debt to GDP (%): Developed weather further turbulence without diluting their shareholders. In the US this
Economies category would include Wells Fargo and JP Morgan and in Europe, Barclays
2003 2004 2005 2006 2007 2008 2009 2010 Change Change and BNP Paribas. In Japan we find SMFG attractive and in Australia ANZ. In
from 2003 from 2008 spite of the slightly elevated risk scores in our danger map we believe that
USA 197 202 205 215 222 239 246 245 48 6 Brazil’s Itau Unibanco and China’s China Construction Bank will outperform the
Australia 107 109 114 119 125 132 145 146 39 14 bank sector. In Emerging Europe we like PKO Bank Polski.
Japan 272 281 282 277 278 299 314 320 48 21
UK 152 159 165 171 182 195 216 218 67 23 Crisis . . . what crisis?
Sweden 142 142 154 165 180 196 210 202 59 6
In this report we ask a number of questions including: just what kind of post
France 146 150 155 157 164 171 184 195 49 24
crisis environment are we in? We attempt to answer this by looking back
Germany 189 188 189 182 174 178 189 198 9 20
through the rear view mirror of past crisis environments and conclude that in
Greece 157 166 176 191 200 216 233 254 97 39 developed economies at least we are in an environment like no other: debt
Ireland 140 158 182 198 217 243 271 266 126 23 levels are at all time highs, interest rates are at 200 year lows, and there is a
Italy 171 172 178 184 186 189 195 210 40 21 real risk of developed country sovereign defaults for the first time since 1936.
Portugal 176 183 210 216 223 237 256 261 85 24 In the most severe crises we identify that have taken place against the
Spain 156 164 175 191 201 211 227 247 91 36 background of asset price and debt deflation shocks we find that: government
Median 157 165 177 187 193 203 222 232 54 22 debt rises very sharply, real GDP growth over a four year period is very slow,
the stock of private sector debt contracts or grows very slowly, and the ratio
Source: IMF, various central banks, Deutsche bank estimates of median values.

of credit to private sector credit to GDP declines and in some instances starts
a multiyear journey to mean reversion. The major difference between these
past post crisis environments and this one is that in previous post crisis
periods the expansion in government debt levels started from much lower
levels and that government credit was considered good.
Figure 2: Crises Compared

Sweden 1990/94 Australia 1990/94 Hong Kong 1997/01 Thailand 1997/01 Japan 1997/01 Average UK 2008/12E USA 2008/12E Ireland 2008/12E
Cumulative increase in nominal GDP t-1 to t+4 0.0% 24.1% 5.7% 15.2% -3.0% 8.4% 12.9% 12.9% -15.3%
Cumulative increase in real GDP t-1 to t+4 -0.1% 12.6% 9.8% -1.3% 2.4% 4.7% 2.6% 8.7% -7.4%
Increase in Government debt t-1 to t+4 114% 132% NA 322% 49% 154% 121% 87% 310%
Increase in bank lending t-1 to t+4 0% 10% -14% -13% -9% -5% 10% 4% -42%
Govt debt to GDP Ratio t -1 (%) 28 17 NA 15 100 40 44 62 25
Govt debt to GDP Ratio t +4 (%) 62 31 NA 54 153 75 87 103 114
Bank loans to GDP ratio t (%) 93 81 149 129 106 112 143 174 192
Bank loans to GDP Ratio t+4 (%) 88 76 138 72 100 95 135 159 149
Source: Deutsche Bank
We conclude that we are in a deleveraging and possibly deflationary the major expansion in credit to GDP ratios and the real possibility that
environment, which is likely to be crisis prone and of long duration. We find undervalued currencies and/or hot money inflows are contributing to asset
that in the instances where mean reversion of debt to GDP ratios takes many price and credit bubbles.
years (Japan 1989 to 2003; Thailand 1997 to 2006; Hong Kong 1997 to 2006)
that bank stock performance is poor but when the adjustment in debt to GDP In the UK and Europe, the quantum of non-performing loans is quite high and
is relatively shallow (Australia, Sweden) bank shares perform strongly once the the ratio of provisions to non-performing loans is quite low, making the sector
market starts to focus on earnings power and valuation. vulnerable to the risk of re-provisioning and to new non-performing loan
formation, although so far, credit quality has been protected by surprisingly
Credit quality in 2011 resilient house prices. In the US, credit quality is improving but the system is
highly sensitive to real estate values as well as to employment levels and GDP
We ask what credit quality looks like within the global banking system and we growth.
draw the following broad conclusions. First that it the developed economies it
appears to have been stabilizing in 2010 and 2011 but that it is very fragile. We find that quite modest adjustments to credit quality assumptions have
Second, that in emerging markets, credit quality metrics are so good that it is quite a significant impact on earnings estimates.
doubtful whether they can be sustained and might deteriorate very suddenly
on adverse economic developments give the very rapid pace of credit growth,

Figure 3: Index of House Prices: Developed Markets

2002 2003 2004 2005 2006 2007 2008 2009 2010 % ch from high/low
USA 100 111.3 129.2 149.1 150.2 136.2 110.7 107.2 104.6 -30%
Australia 100 118.2 125.9 127.7 137.7 153.3 160.0 165.5 185.7 186%
Japan 100 93.6 87.8 83.4 81.1 81.4 82.8 79.9 76.2 -24%
Hong Kong 100 88.0 119.7 140.0 137.2 148.2 176.4 172.8 212.6 213%
UK 100 119.5 139.9 147.1 156.5 170.8 159.3 147.5 156.1 -9%
Sweden 100 109.8 120.0 129.0 144.4 158.7 150.6 165.9 174.6 174%
France 100 111.7 128.7 148.4 166.3 177.3 179.4 166.7 177.2 -1%
Germany 100 100.5 100.1 100.5 101.0 102.1 103.3 104.8 106.8 7%
Greece 100 105.4 107.8 119.6 135.1 143.5 145.7 139.4 136.6 -5%
Ireland 100 113.7 123.5 135.0 150.9 139.9 127.1 104.5 93.2 -38%
Italy 100 106.1 112.6 121.3 128.1 134.7 138.2 137.7 137.8 -1%
Spain 100 116.3 132.9 147.2 156.4 159.7 154.6 142.7 133.8 -16%
Danger maps and stress screens Mexico, Thailand, Indonesia, Malaysia and Korea. Although measures of
profitability, loan quality and capital strength within emerging markets are
Turning to the potential resilience of the global banking system to a “normal” generally superior to the banking systems of developed economies and levels
(i.e. non sovereign) credit cycle we ask whether it can cope with a severe of government and external indebtedness are generally low, there are some
credit downturn. flashing warnings signs, particularly in the BRIC banking systems.
In this report we assess credit risk through two screening methodologies. First, private sector loan growth over the last few years has been phenomenal
First, we develop a danger map or score card to measure macro and system and particularly so since 2008; second, the expansion in credit to GDP ratios
risk and second we assess the resilience of national banking systems to a has been very pronounced; third, artificially undervalued currencies and/or hot
major hike in bad debt provisions and to stressed pre-provision profits. money inflows quite often contribute to asset and credit bubbles; fourth,
changes in lending practices (e.g. Brazil payroll loans) or state influence on
Our danger map scores 9 macro factors on a 1 to 5 basis with 5 being most lending policies (China, India) can have a severely adverse impact on credit
risky or dangerous. The score card makes no comment on such industry quality when the cycle turns; last but not least, real trends in credit quality are
fundamentals as profitability or earnings growth but concentrates on often disguised by the velocity of credit growth, by asset price bubbles and by
vulnerability to a downturn in the credit cycle. We find that in the developed high rates of nominal GDP growth, all of which can turn in on themselves very
economies Japan, Australia, Sweden, Germany and Hong Kong are the least rapidly, as witness the US and UK in 2008, Russia in 2009 and, most
risky countries; that Europe’s periphery countries and Spain are the most risky; spectacularly, Ireland in 2008.
and that the US scores slightly below average in terms of risk and the UK is
around average. It is interesting in our view that in the recent hike in bank CDS prices, China
has moved up in tandem with the US and Europe, albeit from a lower base but
Turning to emerging markets we find that their average score is a little below that CDS prices elsewhere in Asia have remained relatively flat.
the average for the developed economies and that the least risky countries are

Figure 4: Bank CDS Prices

350
300
250
200
150
100
50
0
Sep-09 Dec-09 Mar-10 Jun-10 Sep-10 Dec-10 Mar-11 Jun-11 Sep-11
US UK Europe ex UK China Asia

Source: DataStream
Figure 5: Danger Map Scores: Developed Markets

USA Australia Japan Hong UK Sweden France Germany Greece Ireland Italy Portugal Spain Israel Average
Kong
Deregulation of Lending 1 2 1 1 2 1 2 1 2 3 2 1 1 1 2
% of Credit to GDP 4 3 1 3 5 4 3 2 1 5 1 5 5 3 3
Change in % of Credit to GDP 2 1 1 3 3 4 1 2 3 5 2 5 4 2 3
Maturity of Cycle in Years 2 3 2 3 3 1 3 2 5 2 2 3 3 2 3
Credit Mix 3 1 2 3 3 2 2 3 5 4 5 3 3 2 3
Unemployment 4 1 1 1 3 1 4 2 5 4 2 5 5 1 3
Current account position 4 2 1 1 2 1 2 1 5 2 4 4 4 2 3
Level of real interest rates 1 2 2 1 1 1 2 2 2 3 2 2 2 2 2
Exchange rate flexibility 1 1 2 5 1 1 5 5 5 5 5 5 5 2 3
Total Danger Map Score (out of 45) 22 16 13 21 23 16 24 20 33 32 25 33 32 17 23

Figure 6: Danger Map Scores: Emerging Markets

Brazil Mexico Russia India China Turkey Malaysia Thailand Korea Indonesia Poland Average
Deregulation of Lending 1 1 2 3 3 2 1 1 2 2 2 2
% of Credit to GDP 3 1 2 2 5 1 4 2 4 2 1 2
Change in % of Credit to GDP 5 3 3 4 4 5 3 1 3 2 4 3
Maturity of Cycle in Years 5 2 4 3 3 1 2 2 3 3 2 3
Credit Mix 3 2 4 4 4 2 2 2 3 2 2 3
Unemployment 1 3 2 1 1 2 2 2 1 1 3 2
Current account position 3 2 1 3 1 5 1 1 1 2 4 2
Level of real interest rates 2 1 2 3 3 2 2 2 2 2 3 2
Exchange rate flexibility 2 2 3 3 3 2 3 3 3 3 2 3
Total Danger Map Score (out of 45) 25 17 23 26 27 22 20 16 22 19 23 22
We score each factor on a 1(grey) to 5 (blue) basis with 5 denoting the greatest risk/danger
On our stress test, we screen banks and national banking systems on two Figure 7: Summary of Stress Tests (Cont’d)
measures. First we use two years of recessionary loan loss provisions 2 Years of Recessionary 2 years of Recessionary Core tier one to Risk
(generally equivalent to 2% of loans but with variations) as a percentage of losses as % 2012 Forecast losses as % of 2 years of weighted assets 2012E)
Tangible Equity PPP (flexed down by 25%)
tangible book value and second, two years of recessionary loans less two
years of pre-provision profits flexed down by 25% as a percentage of tangible Ireland -90.1% -75.1% 12.4%
book value. We find that the system copes with this quite well. The US Italy -49.8% -14.0% 0.9%
outperforms Europe, the periphery countries are badly hit, and emerging Spain -73.3% -14.1% 9.1%
markets generally remain quite profitable. Israel -18.9% 10.1% 7.9%
Figure 7: Summary of Stress Tests Brazil -67.7% 23.7% 10.8%

2 Years of Recessionary 2 years of Recessionary Core tier one to Risk Mexico -40.9% 0.5% 15.4%
losses as % 2012 Forecast losses as % of 2 years of weighted assets 2012E)
Tangible Equity PPP (flexed down by 25%) Russia -22.7% 22.0% 13.3%
USA -33.3% 4.6% 11.8% India -23.0% 25.4% 8.0%
Australia -61.0% -16.4% 8.5% China -32.0% 11.9% 10.5%
Japan -22.3% -1.2% NA Turkey -18.5% 17.6% 16.1%
Hong Kong -13.0% 16.9% 11.2% Indonesia -73.6% -21.6% 15.9%
UK -32.4% 2.2% 8.2% Malaysia -43.3% -6.0% 6.0%
Sweden -68.1% -39.9% 8.6% Thailand -16.3% 20.9% 10.9%
France -46.0% -3.9% 10.0% Korea -29.0% 0.9% 8.8%
Germany NA NA NA Poland -47.5% -8.9% 15.2%

Greece -82.4% -38.9% 7.4%

The elephant in the room Figure 8: Distribution of European Bank Claims on selected countries
(US$bn) Public sector Banks Private Total Relative to total European
We look at the European sovereign debt crisis and benchmark it against the sector bank loans and Securities
Latin American debt crisis of the 1980’s. We find that this sovereign debt Portugal 32388 40447 121770 194605 0.55%
crisis has at least five dimensions. Ireland 15355 70539 291742 377636 1.07%
Greece 54196 10918 80669 136317 0.39%
The first is the cross border exposures to private sector entities, often
originated and funded by the subsidiaries of the largest European banks. Spain 88054 199269 344866 632189 1.79%
Losses from these exposures have been in the region of US$90bn, with Italy 231216 127261 126891 485368 1.38%
US$50bn of losses in Ireland alone. Further losses from these exposures are Total 421209 448434 965938 1826115 5.18%
baked into analyst estimates; whether these estimates are conservative or not Source: IMF and BIS
remains to be seen.
Bank sector performance in a deleveraging world
The second dimension to the crisis is the sovereign/public sector exposures,
which too are often held in the domestic subsidiaries of European banks. We There are few recent data points to benchmark and measure bank sector
find these exposures generally quite modest when measured against the performance in a deleveraging environment for the good reason that over the
totality of European bank assets and capital. Certainly, they are dimensionally last 30 years private sector debt to GDP ratios have been steadily or rapidly
lower than the exposures of the US money center and European megabanks climbing in most parts of the world.
to Latin America relative to assets and capital in the 1980’s.
We have identified three periods in which there was significant deleveraging
The third dimension is the ownership by domestic banks in the periphery (in terms of private sector debt to GDP ratios) over a long period of time:
economies, Spain and Italy, of their own sovereigns’ bonds. Such holdings are Thailand between 1997 and 2007, Japan between 1989 and 2003, and Hong
typically 150% to 220% of tangible book value and thus substantial write- Kong between 1997 and 2005.
downs could trigger recapitalization requirements.
In Japan, the bank sector underperformed dramatically during the period of
The fourth dimension is the potential spillover effects including runs on
deleverage and then near-quadrupled when the deleveraging period came to
banking systems and the negative feedback loop to consumer and business
an end in 2003. It subsequently underperformed from 2005 reflecting first the
confidence that might develop from a disorderly default, which are impossible
withdrawal of quantitative easing and then the global financial crisis and the
to model.
dilution for Basel 3 related rights issues (for more on this see our report Japan
The fifth dimension is the possibility of an extreme outcome if the Euro was to Redux: After the Reflation trade? 24th May 2011).
break up, which again is not something that readily lends itself to company-
specific modeling.
Possibly the most desirable scenario, but not necessarily the most likely
outcome, is provided by the Latin American 1980’s crisis resolution: i.e. a very
long period of uncertainty and significant write downs of private sector debt
and then a Brady bond type solution to, and write down of, government debt
once the bank sector can afford it.

Figure 9: Japan - Nominal Gross Domestic Product/Bank Loans Figure 11: Thailand bank sector performance in a deleveraging world
(Calendar Year、%)
Note: Bank loans= Domectically Licenced Bank Accounts and Trust Accounts. Do not include for central government.
Source: Cabinet Office, Government of Japan, Bank of Japan
Figure 10: Japan - TSE Bank Index (Year-end) Source: Deutsche Bank
The case of Hong Kong is less conclusive. Bank shares underperformed the
Hong Kong market in the early part of the deleveraging process and then
outperformed from 2003 but were poor investments between 1997 and 2004
and the index is now back below where it started before the 1997 Asian
banking crisis, even though there was no recapitalization requirement.
Source: Tokyo Stock Exchange
In the case of Thailand, the graph below plots the Thai bank sector after the
initial devaluation shock in 1997 through to 2011 against credit to GDP. It can
be seen that the bank sector was dead money in absolute terms and much
worse than that relative to the Thai equity market during the long deleveraging
period. The Thai banking sector started to perform once debt to GDP
troughed, which happened to be when it reverted to pre-crisis levels in 2007.

Figure 12: Hong Kong bank sector performance in a deleveraging world Figure 14: Australia - Debt vs GDP vs share price performance during
1990s
40.0% 52.0%
30.0%
51.0%
20.0%
50.0%
10.0%
0.0% 49.0%
Jun‐1990
Jun‐1991
Jun‐1992
Jun‐1993
Jun‐1994
Oct‐1990
Oct‐1991
Oct‐1992
Oct‐1993
Oct‐1994
Feb‐1991
Feb‐1992
Feb‐1993
Feb‐1994
‐10.0%
48.0%
‐20.0%
47.0%
‐30.0%
‐40.0% 46.0%
Avg share price performance LHS (rolling 6 month) Debt to GDP RHS
Source: Deutsche Bank Source: Deutsche Bank
Unless we are accused of being selective in the data we use, we should point However, in Australia and Sweden the deleveraging period was relatively short
out that after the Swedish crisis, bank stocks quadrupled in 1993 although and shallow and was followed by a sustained period in which debt to GDP
debt to GDP modestly shrank and performed reasonably in 1994 and 1995 rose steadily and in the context of a bull market in housing.
during a more significant deleveraging.
How to navigate this document
Figure 13: Bank sector performance in the early 1990s vs Credit to GDP
This research note divides into three parts which are linked but which can also
Year Swedish Bank Index, YoY Swedish Credit/GDP
stand alone: the first is an overview, which chronicles recent patterns of credit
1989 11.6% 78.2%
growth and changes in private sector and government debt to GDP ratios;
1990 -29.8% 92.9%
which looks in detail at credit quality and loan loss provisioning ratios across
1991 -19.1% 93.0% the countries which we highlight in this report; which stress tests companies
1992 -67.2% 95.8% and banking systems and puts them in the framework of our danger map; and
1993 382.1% 93.3% which assesses this post crisis environment against others. The second part is
1994 -1.5% 87.7% the country section. This provides a perspective on the composition of loan
1995 13.5% 84.0% portfolios in each of the countries covered in this report and some of the
Source: Deutsche Bank credit issues relevant to that country. The third part of the note provides
thumb nail case studies of nine crises ranging from the Latin American Debt
In Australia, we found no statistical correlation over a long period between Crisis of the 1980s to Ireland’s ongoing banking crisis.
bank sector performance and credit to GDP.

Credit growth and

The deleveraging in the US has taken the ratio of private sector debt to GDP
down by over 16 percentage points to 163%, a sudden and large decline but
to a level still well over its long term average. A contraction in the stock of
leverage credit in the US had not occurred previously in any period for at least 60 years.
As we see later, a contraction in private sector credit has become a
commonplace occurrence in Japan since the mid 1990s and has also been a
feature of other post crisis environments. Japan’s debt to GDP ratio reverted
Since the banking crisis of 2008/09 and the accompanying recession there to mean over a ten year period. Thailand delevered for 10 years after the 1997
have been two broad developments in bank and market based credit. currency and banking crisis, with debt ratios again reverting to mean and Hong
Kong for almost as long after its deflationary shock post 1997. Slow credit
Private sector lending growth can mean that system NPL ratios remain high since NPL levels are not
diluted or are diluted very slowly by new flows of performing credit.
First, private sector lending has decelerated or actually contracted in many of
the most indebted developed market economies as households and
Figure 16: Private sector credit to GDP: US
corporations have delevered and as lenders have tightened credit standards.
The median growth in domestic credit between December 2008 and 185.0
175.0
December 2010 for the developed economies was about 2%. The stock of
165.0
credit has declined in Japan and the US, has been flat in the UK but has
155.0
continued to grow in most countries in the Eurozone.
145.0
135.0
Figure 15: Indexed Credit Growth 2003 to 2010: Developed Economies 125.0
2003 2004 2005 2006 2007 2008 2009 2010 Change 115.0
since 08
105.0
USA 100 109 120 133 145 148 145 144 -3%
95.0
Australia 100 112 127 146 169 191 204 212 11% 85.0
Japan 100 96 98 100 103 107 103 102 -4%
Dec 79
Nov 82
Nov 85
Nov 88
Nov 91
Nov 94
Nov 97
Nov 00
Nov 03
Nov 06
Nov 09
Jun 81
May 84
May 87
May 90
May 93
May 96
May 99
May 02
May 05
May 08
Hong Kong 100 106 114 121 146 161 162 208 29%
UK 100 110 120 132 151 161 161 162 0%
Private sector debt to GDP (%) Average
Sweden 100 107 122 139 161 179 183 189 6%
Source: Federal Reserve
France 100 106 116 128 144 153 153 163 7%
Germany 100 99 98 98 99 101 101 104 3%
There have been more modest declines in private sector debt to GDP ratios in
Greece 100 119 141 173 207 240 240 248 3% the UK, Australia, and Sweden amongst others, but the median increase since
Ireland 100 125 164 200 238 234 215 171 -27% 2003 has been around thirty percentage points. Private sector debt to GDP
Italy 100 108 117 130 144 151 152 165 9% ratios, which are heavily influenced by home ownership trends and house
Portugal 100 111 121 131 145 158 162 164 4% prices, are particularly elevated in Spain, Portugal, Ireland and Sweden. As we
Spain 100 117 148 184 215 229 226 229 0% show later, after a severe financial crisis private sector debt to GDP ratios
Median 100 109 120 132 146 161 162 165 2% invariably contract, which may suggest that Europe has yet to adjust since
Source: IMF, various central banks, Deutsche Bank estimates debt to GDP ratios generally have continued to rise in the Eurozone countries
post crisis.
Figure 17: Private Sector Debt to GDP (%), Developed Economies 2003 to In marked contrast to developed economies, the pace of lending growth has
2010 accelerated in many emerging markets since 2008, partly driven by a leakage
2003 2004 2005 2006 2007 2008 2009 2010 Change Change
of US monetary policy. There has been a more than threefold median increase
from from 2008 in the stock of credit of the emerging markets covered in this report since
2003 or high 2003 and a 35% increase since 2008.
USA 148 152 155 164 171 178 174 164 17 -13
Australia 94 98 103 110 116 120 128 123 30 -5 Figure 19: Indexed credit growth 2003 to 2010: Emerging Economies
Japan 92 88 89 89 91 99 99 98 6 -1 2003 2004 2005 2006 2007 2008 2009 2010 Change
UK 113 118 123 128 138 143 148 142 30 -6 since 08
Sweden 101 104 116 126 140 158 170 167 66 -3 Brazil 100 119 145 175 224 293 338 408 39%
France 83 85 89 93 100 103 106 110 28 7 Russia 100 121 170 250 383 515 502 565 10%
Germany 124 120 117 113 109 109 112 111 -13 2 India 100 127 189 242 294 346 404 515 49%
Greece 60 67 75 85 95 105 106 112 52 7 China 100 111 122 140 164 189 251 300 59%
Ireland 109 129 154 173 192 199 206 170 60 -29 Turkey 100 150 226 331 403 555 593 794 43%
Italy 66 69 72 77 83 85 89 94 28 9 Indonesia 100 127 160 183 230 301 332 409 36%
Portugal 120 125 147 152 161 172 180 178 58 -3 Malaysia 100 126 148 165 179 203 219 247 22%
Spain 107 117 137 157 171 177 180 181 74 4 Thailand 100 107 114 121 128 143 142 159 11%
Median 104 111 117 119 127 132 138 133 30 -2 Korea 100 105 114 130 149 170 177 183 8%
Source: Federal Reserve, various central banks, Deutsche Bank estimates Poland 100 103 115 142 185 253 275 298 18%
Private sector deleveraging in the highly indebted economies has been Median 100 120 147 170 204 273 303 354 30%
complemented by deleveraging of banking systems as a response to Source: IMF, various central banks, Deutsche Bank estimates
regulatory and market pressure. According to the Bank of England, global

banks have raised over US$500bn in equity since 2009 and reduced total Rapid credit growth has resulted in a notable expansion in the private sector
assets by over US$3 trn. debt to GDP ratios of emerging economies. These ratios are typically much
lower than in developed markets with a median ratio of 52% against 133%,
Figure 18: : Bank leverage: USA and Europe but interest rates are much higher, suggesting that debt service levels are not
50.0
that different. The consumer debt service in Brazil for instance is amongst the
45.0
highest in the world at 23% although private sector credit to GDP is lower
40.0
35.0
than the average in emerging markets and approximately a third of the median
30.0 level in developed economies.
25.0
20.0
15.0
10.0
5.0
0.0
2007 2008 2009 2010 2011
USA Europe
Source: Deutsche Bank estimates

Figure 20: Private Sector Debt to GDP (%): Emerging Markets Figure 21: Indexed Government Debt Growth: Developed Economies
2003 2004 2005 2006 2007 2008 2009 2010 Change Change from 2003 2004 2005 2006 2007 2008 2009 2010 Increase from
from 2003 2008 or high 2008
Brazil 26 28 29 31 39 43 51 49 23 6 USA 100 108 116 122 130 152 177 199 31%
Russia 21 23 25 30 37 40 42 41 20 1 Australia 100 97 96 94 100 133 204 278 108%
India 30 34 44 49 49 49 50 50 20 1 Japan 100 108 109 108 109 111 115 119 7%
China 125 118 112 110 104 102 125 128 3 26 UK 100 111 120 130 140 169 214 249 47%
Turkey 15 18 23 29 32 39 41 48 33 9 Sweden 100 94 98 107 110 106 103 97 -8%
Indonesia 22 24 25 24 25 27 26 28 6 1 France 100 107 114 115 120 131 148 163 25%
Malaysia 85 95 101 102 99 97 115 115 30 17 Germany 100 106 110 115 116 118 127 133 13%
Thailand 74 72 70 67 65 69 68 69 -5 0 Greece 100 109 117 134 142 156 178 195 25%
Korea 70 68 71 77 82 89 90 84 14 -5 Ireland 100 101 102 101 109 184 241 341 85%
Poland 31 34 29 36 44 40 56 53 23 13 Italy 100 103 108 114 115 116 116 128 10%
Median 30 34 37 42 47 46 53 52 20 4 Portugal 100 107 121 128 132 141 160 180 28%
Source: The IMF, various Central Banks, Deutsche Bank Estimates
Spain 100 105 92 88 84 98 130 186 90%
Median 100 107 110 114 115 132 154 183 27%
Government debt Source: IMF, Deutsche bank Estimates
The second unambiguous development since 2008 has been the notable Relative to GDP, G7 gross public debt levels climbed from 82% in 2007 to
increase in government debt in nearly all developed economies following 112% in 2010. IMF forecasts predicate a further increase to 123% by 2014. If
much lower than expected tax revenues and, in some countries, because of these forecasts are correct, the stock of G7 public debt will have increased by
the cost of bailing out and recapitalising banking systems. The median 83%, from the equivalent of US$30.7trn in 2007 to US$46.0 trn in 2014E. For
increase in public debt levels has been 83% since 2003, 59% since 2007 and the developed countries in this report, the median increase in public sector
27% since 2008. These increases have been more modest than in previous debt to GDP ratios since 2008 has been 19 percentage points.
crisis periods but have started from a higher base.

Figure 22: Public Sector Debt to GDP: Developed Economies Figure 23: Private and Public Sector Debt to GDP (%): Developed
Economies
(%) 2003 2004 2005 2006 2007 2008 2009 2010 Change Change 2003 2004 2005 2006 2007 2008 2009 2010 Change Change
from from from from
2003 2008 2003 2008
USA 49 50 50 51 51 61 72 80 31 19 USA 197 202 205 215 222 239 246 245 48 6
Australia 13 12 11 10 9 12 18 22 9 11 Australia 107 109 114 119 125 132 145 146 39 14
Japan 180 193 193 188 188 200 215 222 42 22 Japan 272 281 282 277 278 299 314 320 48 21
UK 39 41 43 43 44 52 68 76 37 24 UK 152 159 165 171 182 195 216 218 67 23
Sweden 41 37 38 39 39 38 39 35 -6 -3 Sweden 142 142 154 165 180 196 210 202 59 6
France 63 65 66 64 64 68 78 84 21 17 France 146 150 155 157 164 171 184 195 49 24
Germany 65 69 71 69 65 69 76 87 22 18 Germany 189 188 189 182 174 178 189 198 9 20
Greece 97 99 100 106 105 110 127 142 45 32 Greece 157 166 176 191 200 216 233 254 97 39
Ireland 31 29 27 25 25 44 66 96 65 52 Ireland 140 158 182 198 217 243 271 266 126 23
Italy 104 104 106 107 104 104 106 116 11 12 Italy 171 172 178 184 186 189 195 210 40 21
Portugal 56 58 63 64 63 65 76 83 27 18 Portugal 176 183 210 216 223 237 256 261 85 24
Spain 48 47 38 34 30 34 47 66 18 32 Spain 156 164 175 191 201 211 227 247 91 36
Median 52 54 57 57 57 63 74 84 25 19 Median 157 165 177 187 193 203 222 232 54 22
Source: IMF, Deutsche bank Estimates Source: IMF, various central banks, Deutsche bank Estimates
The increase in public sector debt has more than cancelled out whatever Again, there has been a contrasting trend in many emerging markets, where
deleveraging has taken place in the private sector. The combination of public public debt to GDP ratios are not only much lower than in developed
and private sector debt to GDP has risen in every country we follow here since economies but have tended to be stable or in decline since 2003:
2008, with the smallest increases in the US and Sweden. The median ratio of
Figure 24: Public Sector Debt to GDP (%): Emerging Economies
public and private sector debt to GDP has increased from 157% in 2003 to
2003 2004 2005 2006 2007 2008 2009 2010 Change Change
232% in 2010 and this ratio is expected to increase by a further 10 to 15 from from
percentage points by 2014: 2003 2008
Brazil 15 12 9 9 10 9 10 10 -5 1
Russia 27 21 14 8 7 7 8 9 -18 2
India 64 54 64 61 58 56 55 54 -10 -2
China 26 24 24 23 22 20 21 20 -6 0
Turkey 65 59 54 48 42 43 49 42 -23 -1
Indonesia 64 65 61 60 63 68 73 75 11 7
Malaysia 45 46 44 42 42 41 53 53 8 12
Thailand 123 120 116 108 102 106 113 110 -13 4
Korea 22 25 29 31 30 30 34 37 16 7
Poland 6 6 5 5 4 3 4 4 -2 1
Median 36 35 36 37 36 36 41 40 -6 2
Source: IMF, various central banks, Deutsche bank Estimates

The sum of private and public sector debt in emerging markets relative to GDP
has actually declined relative to developed market debt since 2003, although
the pace of private sector debt growth has been significantly higher.
Figure 25: Private and Public Sector Debt to GDP Ratios (%)
2003 2004 2005 2006 2007 2008 2009 2010 Change Change
from from
2003 2008
Brazil 41 40 38 40 49 52 61 59 18 7
Russia 48 43 39 38 44 47 50 49 2 3
India 94 88 108 110 107 105 105 104 10 -1
China 151 142 136 133 126 122 146 148 -3 26
Turkey 80 77 77 77 74 82 90 90 10 8
Indonesia 86 89 87 84 88 95 98 103 17 8
Malaysia 130 140 145 144 141 138 168 168 38 29
Thailand 197 192 186 176 168 174 181 179 -18 4
Korea 92 93 100 108 112 119 123 122 30 2
Poland 37 40 34 41 49 43 60 57 21 14
Median 89 89 93 96 98 100 102 103 13 7
Source: IMF, various central banks, Deutsche bank Estimates
Thus in a nutshell, while leverage in the banking system measured by the

multiple of total assets to net tangible assets has declined since the 2008/09
crisis, debt levels in the global economy have increased in absolute terms and
relative to GDP. Downgrades to GDP growth expectations are thus occurring
either when private sector and public sector debt levels relative to the size of
economies are at or close to record levels, or after a period in which private
sector debt growth in absolute terms and relative to GDP has expanded at a
very rapid pace, particularly in emerging markets.
Large fiscal deficits and fears of government default have weighed very
heavily on bank sectors An IMF analysis of losses from securities and loans
incurred by banks operating in the developed economies between 2007 and
March 2010 suggested that cumulative losses for the banking system were in
the region of US$2.3trn, including US$0.6trn from securities. The only
segment of the securities portfolios which attracted nil losses were the
holding of government bonds, which accounted for about 25% of total
securities.

Credit quality
With the exception of Russia, NPL levels in emerging markets barely changed
in 2008 as any increase in crisis related NPLS was counterbalanced by the
very rapid growth in the stock of credit. CDS prices for Chinese banks have
risen in tandem with US and European banks and to quite similar levels. Few
market observers believe that the published NPL ratios for the Chinese
The loan quality of the global banking system in 2010
banking system provides an accurate assessment of real system loan quality
Loan quality measures are notoriously backward looking and loan quality in that economy and the example of Russia in 2009 and indeed the US in
forecasts are invariably based, by necessity, on consensus economic 2008, shows how dramatically and quickly NPL levels can change. NPLS are
forecasts. As end 2010 the loan quality of the banking systems was fragile. more than 100% covered in many of the emerging market countries.
NPLs/NPA’s were running at around 5% of total loans in the US and Europe
against the position in 2007, when NPL ratios were in the 1 to 2% range. NPls Figure 27: NPL to total loans (%): emerging markets
are forecast to continue to rise at a reasonably subdued rate in the Eurozone in 2007 2008 2009 2010 2011E 2012E
2011 but be close to peak levels at that time but to have peaked already in the Brazil 7.7 8.3 9.7 8.1 9.0 9.4
US. These forecast are consistent with first half results and a scenario of
Russia 0.7 1.8 10.2 8.8 7.1 4.8
moderate GDP growth in a negative real interest rate environment.
India 1.8 1.6 1.4 1.4 1.2 1.2
China 1.8 1.4 1.0 0.7 0.9 0.9
Figure 26: NPL to total loans (%): developed markets
Turkey 4.9 4.6 5.6 3.7 3.4 3.3
2007 2008 2009 2010 2011E 2012E
Asia Average 1.7 1.6 1.5 1.5 1.3 1.2
Australia 0.5 1.1 1.9 2.0 1.6 1.3
Global Emerging 2.3 2.5 2.8 2.5 2.4 2.3
Hong Kong 0.4 0.9 0.8 0.3 0.3 0.3 Market Average
Japan 2.9 2.6 2.3 2.9 2.5 NA Source: Deutsche Bank
USA 0.8 1.8 3.1 2.9 2.1 1.5

France 3.4 3.8 4.6 5.5 5.6 4.9 Impairment charges
Germany 2.8 3.1 4.6 4.9 4.4 4.1
The provision charge was running at around “normal” levels in 2007 and
Greece 4.9 4.6 6.8 9.6 13.5 14.0 thereafter between 2008 and 2010 ran at a multiple of average levels but
Ireland 0.8 4.0 11.2 9.9 11.4 10.8 reached a crisis peak in 2009. The peak was very high relative to history in the
Italy 5.0 5.2 7.2 8.0 8.2 7.5 US and for domestic banking institutions in the UK but was not particularly
Portugal 0.8 1.0 2.0 2.6 3.1 3.4 elevated in Europe. Because revenue lines in the US and Europe are under
Spain 0.9 2.6 4.3 4.8 5.5 5.0 pressure, the assumption of declining bad debt charges is the major driver of
Sweden 0.4 0.8 2.0 1.9 1.7 1.5 projected earnings growth, accounting for over 100% of forecast earnings
UK domestc banks 1.8 3.4 5.8 6.8 7.9 6.6
growth in Europe and the USA in 2011E and about 40% in 2012.
European average 1.5 2.5 4.4 4.8 4.6 4.3
Unweighted average 2.0 2.7 4.4 4.8 5.2 5.1
NPL coverage has fallen in Europe since 2007 from around 60% to around
40% in 2011 but has risen elsewhere.

Figure 28: Provision charge to average loans (%): Developed markets Bank share prices have fallen by about 35% in developed markets and 20% in
2007 2008 2009 2010 2011E 2012E emerging markets so far this year. This compares with a decline in Deutsche
Australia 0.20 0.58 0.49 0.30 0.19 0.21
Bank analysts’ 2012E earnings estimates of less than 20% for European bank
and about 12% for UK banks. US bank earnings estimates have been revised
Hong Kong 0.15 0.51 0.35 0.12 0.21 0.24
up. The decline in bank share prices since August has been accompanied by
Japan 0.26 0.29 0.72 0.47 0.41 NA
very sharp increases in bank CDS prices for banks generally but particularly for
USA 0.62 2.40 4.34 2.22 1.12 0.81
US and European banks
France 0.41 0.97 1.43 1.01 0.75 0.45
Germany 0.25 0.49 0.91 0.59 0.47 0.38
Figure 30: Bank CDS Prices
Greece 0.72 1.33 2.43 2.33 2.22 1.92
350
Ireland 0.18 1.12 6.53 11.24 1.44 1.17
Italy 0.44 0.70 0.96 0.74 0.68 0.62 300
Portugal 0.49 0.69 0.93 0.83 0.87 0.77
250
Spain 0.37 0.94 1.48 1.35 1.11 0.96
Sweden 0.03 0.22 0.79 0.22 0.01 0.07 200
UK domestc banks 0.83 1.34 2.05 1.58 1.31 0.80
150
European average 0.35 0.75 1.05 0.74 0.65 0.55
Unweighted average 0.38 0.88 1.75 1.70 0.82 0.69 100
50
In contrast, emerging markets provision charges rose relatively modestly in
0
2009 (Russia excepted) before falling back to normal or below normal levels in
most countries. Again, growth in the stock of loans as well as robust
economic growth, low or negative real interest rates and upward pressure on US UK Europe ex UK China Asia
asset prices has helped maintain low provision charges. The worry for Source: DataStream
emerging markets is that the undervaluation of currencies is pumping up asset

bubbles and that the very low bad debt charges are a sign post to what almost The fall in share prices may well be connected to investor concern as to a bad
by definition must be a more difficult environment down the line. outcome with the sovereign debt crisis, the Euro, US mortgage related
litigation or all of these, but it may also be the case that the market is
Figure 29: Provision charge to average loans (%): Emerging markets discounting a severe downturn in the credit cycle as GDP growth in the US
and Europe slows down and should growth become negative.
2007 2008 2009 2010 2011E 2012E
Brazil 5.26 5.03 7.16 4.69 4.78 4.83
Russia 1.44 2.98 6.91 2.18 1.09 0.82
India 0.41 0.46 0.45 0.37 0.50 0.46
China 0.33 0.52 0.21 0.23 0.28 0.29
Turkey 1.06 1.44 2.43 1.12 0.83 0.86
Asia Average 0.37 0.57 0.46 0.35 0.34 0.32
Gem average 0.50 0.55 0.51 0.41 0.44 0.41

Figure 31: EPS revisions YTD Figure 32: Bank Share price performance: Local Currency (%)
1w 1m 3m YTD YOY
USA -0.8 -3.5 -23.5 -30.7 -23.4
Australia -3.6 -5.3 -13.4 -22.3 -21.8
Japan -2.8 -6.8 -9.3 -26.2 -22.9
Hong Kong -5.1 -2.3 -18.9 -27.2 -20.3
UK -2.6 -15.8 -36.1 -42.2 -50.5
Sweden -7.5 -17.0 -28.8 -31.9 -26.4
France -19.0 -30.6 -53.7 -51.9 -57.6
Germany -15.8 -31.4 -49.9 -55.5 -43.2
Greece -1.8 -29.6 -52.7 -56.5 -69.4
Ireland -7.0 -33.6 -58.8 -83.9 -92.1
Italy -8.7 -21.3 -42.7 -56.8 -64.4
Spain -5.1 -10.2 -21.2 -24.8 -39.3
Israel -8.6 -8.7 -23.8 -32.9 -23.3
Because NPL levels are high and loan loss reserves against NPLSs are not Brazil -7.4 -0.5 -16.1 -25.6 -17.8
universally so, and because pre-provision profit growth is low or negative, Mexico 0.0 1.8 -18.5 -25.1 -12.4
developed market bank earnings estimates in 2012E and 2013E are highly Russia 13.7 7.7 -25.8 -26.9 -8.0
sensitive to any change in assumptions around asset quality. For instance India -1.5 -7.7 -13.6 -22.2 -21.7
aggregate forecasts for Europe in 2012E assume a decline in the NPL ratio China -1.0 -1.4 -13.4 -16.1 -18.2
from an estimated 5.66% in 2011E to 4.85%, which is consistent with a Turkey 4.2 8.7 -13.5 -22.4 -22.3
decline in the bad debt charge from 0.86% of loans in 201E to 0.70% in
Indonesia -1.1 -2.8 -5.7 -7.7 11.4
2012E.
Malaysia -2.4 -3.6 -5.6 0.6 6.6
Thailand -0.9 -4.5 1.3 -3.9 8.1
Korea -4.7 -9.1 -21.5 -31.2 -19.0
Poland -7.0 -13.4 -30.0 -26.3 -18.4
All other things being equal, a 10% increase in the NPL ratio covered 50% by
new provisioning and a 10% re-provisioning of the existing book of non-
performing loans would arithmetically flow to a doubling of the loan loss
provision charge and a reduction in pre-tax profits of around 40%. The simple
model below shows the sensitivity of bank earnings to changes in NPL and
bad debt assumptions.

Figure 33: Sensitivity of EPS estimates to changes in NPL assumptions Figure 34: Provisions to NPL Coverage ratios: Developed Markets
Sensitivity of profits to NPL assumptions 2006 2007 2008 2009 2010 2011E 2012E
Loans 1000 USA 180.9% 147.5% 100.8% 128.4% 140.8% 179.9%
Pre-provision profit to loans 2.50% Australia 149.5% 133.0% 94.0% 70.3% 60.8% 58.4% 61.8%
NPL Ratio 5.66% Japan 29.1% 32.6% 29.1% 25.8% 24.4% 24.8% 24.9%
Coverage ratio 37% Hong Kong 230.5% 195.8% 335.8% 228.5% 214.8% 208.2% 209.7%
Bad debt charge 0.70% UK 57.9% 47.3% 46.0% 44.7% 47.2% 53.6% 56.2%
Cost of 10% re-provisioning with 50% coverage 0.28% Sweden 65.1% 68.8% 46.6% 51.3% 50.2% 52.8% 55.6%
Cost of 1 percentage point increase in NPLs 50% covered 0.50% France 76.7% 83.2% 77.0% 73.5% 67.3% 68.6% 71.0%
Revised bad debt charge 1.48% Germany 52.2% 55.6% 46.7% 46.4% 47.8% 46.4% 51.6%
Greece 63.1% 57.1% 47.9% 42.9% 41.2% 38.2% 42.7%
Forecast Pre provision profit 25.0 Ireland 63.0% 56.1% 33.5% 43.3% 44.4% 52.8% 58.7%
Forecast provisions at 0.70% 7.0 Italy 59.7% 60.9% 59.0% 48.9% 48.7% 52.5% 54.5%
Forecast pre-tax profit 18.0 Spain 238.6% 196.1% 95.6% 75.2% 73.8% 71.6% 75.4%
New provisions 7.8 Israel 277.7% 328.0% 315.1% 361.9% 378.9% 94.7% 94.0%
Revised pre-tax profit 10.2
Downgrade -44%
Clearly, forecasting error on bank earnings estimates will be greatest for

companies/countries with high NPL ratios (i.e. high betas). Estimates for these
banks will have the greatest downside risk to reprovisioning requirements on
the existing NPL stock and will probably be more vulnerable to further
increases in NPLS, given that the high ratio relative to peers is already flagging
poorer pre-crisis loan underwriting. As loan quality deteriorates, the market
also looks more closely at overall NPL coverage ratios.

Asset quality versus loan quality We discuss this in more detail below, but in brief the Eurozone public debt
market constitutes a US$ 10 trillion asset class and is thus approximately the
The IMF has estimated that about 25% of the estimated US$2.3trn of losses size of the US mortgage market. Sovereign debt and mortgage assets have
from loans and securities during the period 2007 to March 2010 derived from traditionally been at the top of the food chain of bank balance sheets in terms
securities holdings. The table below shows the IMF’s estimate of the of liquidity and capital management. Mortgage related losses cost the banking
distribution of loans and securities across the banking systems of developed system approximately US$ 1trn over the banking crisis. Over the last 12
markets. Two points stand out. First, securities made up 28% of the total of months there has been severe quality degradation within the other low risk
loans and securities; and second, the Euro area, by far the largest component weighted asset class.
(41%) of the bank sector, held US$ 7trn in securities.
Figure 37: Sovereign 10 year bond spreads
Figure 35: Developed Market Banking System March 2010 18
(US$bn) US Banks UK Banks Euro Area Other Mature Developed Total 16

Banks Europe banks Asia Banks
14
Loans 8059 6744 15994 3241 6150 40188 12
Securities 4502 1625 6907 729 1728 15491 10
Total 12561 8369 22901 3970 7878 55679 8
Distribution 23% 15% 41% 7% 14% 100% 6
4
GDP 14119 2182 12476 1755 6346 36878
2
Loan and securities 0.89 3.84 1.84 2.26 1.24 1.51
to GDP (X) 0
Securities as a 36% 19% 30% 18% 22% 28%
percentage of loans Portugal Ireland Greece Italy Spain
and securities Source: DataStream
Source: IMF, Deutsche bank estimates
Figure 38: Key changes in income and expense items Europe

As at March 2010 and again using IMF estimates, approximately 31% of total 2008 2009 2010 2011 2012
European bank securities (US$2.1trn) were government bonds.
Income (€bn) -45 63 34.8 -0.3 27.9
Expenses 32.4 -21.4 24.5 6.2 2.7
Figure 36: Distribution of Euro area securities portfolio March 2010
Pre-provision profits -77.4 84.4 10.3 -6.5 25.2
US$bn %T
Provisions 57.8 53.3 -46.5 -22.9 -14.4
- Residential mortgages 966 14%
Pre-tax profits -135.2 31.1 56.8 16.4 39.6
- Consumer 272 4%
Income (%) -10% 15% 7% 0% 5%
- Commercial Mortgage 264 4%
Expenses 12% -7% 9% 2% 1%
- Corporate 1316 19%
Pre-provision profits -40% 73% 5% -3% 12%
- Governments 2146 31%
Provisons 130% 52% -30% -21% -17%
- Foreign 1943 28%
Pre-tax profits -91% 221% 126% 16% 33%
Total for Securities 6907 100% Source: Deutsche Bank

In a nutshell, because of the sensitivity of the growth outcome to loan loss

provisioning, there is the potential for significant forecasting error even
without factoring in adverse developments in the Eurozone sovereign debt
crisis.
Earnings growth for the US universe is even more dependent that Europe on
the fulfillment of expectations that credit quality will continue to improve and
bad debt charges decline in 2011 and 2012 as revenues are forecast to decline
in 2011 and rise only very modestly (2.2%) in 2012.
Figure 39: Key changes in income and expense items USA

2008 2009 2010 2011 2012
Income (USbn) -15.4 185.4 -32.5 -30.7 10.5
Expenses 26.2 42.3 24.0 -0.5 -14.1
Pre-provision profits -43.7 139.4 -40.4 -30.1 24.5
Provisions 84.1 68.2 -98.5 -56.0 -10.2
Pre-tax profits -121.2 81.7 28.9 26.1 34.7
Income (%) -5% 58% -6% -7% 2%
Expenses 13% 19% 9% 0% -5%
Pre-provision profits -35% 174% -18% -17% 16%
Provisons 158% 50% -48% -52% -20%
Pre-tax profits -151% nm 71% 38% 36%
In emerging markets the story is very different. Forecasts assume that bad
debt provision charges remain fairly constant.

Testing for a
Figure 41: Three year cumulative losses against normal loss rates 2008
to 2010
2008-2010 2008 to 2010 Normal Multiple of normal (X)
recessionary credit cycle USA

Australia
10.8%
1.8%
1.80%
1.80%
6.0
1.0
Japan 1.3% 0.90% 1.5
The traditional framework for valuing or assessing bank shares via normalized UK 5.9% 1.80% 3.3
or peak loan losses (and hence normalized or trough earnings) has a credibility Sweden 1.6% 0.90% 1.8
problem, as indeed do stress tests generally. There are three reasons this may France 3.4% 1.50% 2.3
be the case. Greece 6.1% 1.80% 3.4
Ireland 19.0% 1.50% 12.7
First private and public sector debt is at record levels, suggesting the potential
Italy 2.8% 1.50% 1.9
for extreme outcomes including multiple sovereign defaults. Source: FDIC, Deutsche Bank estimates
Second, the combination of 2008/09 crisis loan loss provisions and securities Furthermore, the bulk of the securities losses and loan losses were taken
losses was so far out of a “normal” range that the market may be recalibrating against assets which the banks and their regulators assumed were the lowest
expectations based upon the last experience rather than on the last few risk. In fact, only government bonds were risk free in 2008/09.
recessions. Three year cumulative loan loss provisions for our US universe
were 10.8% between 2008 and 2010, well over twice the previous peak three Figure 42: Distribution of losses on securities 2007 to 2010 (US$bn)
year loss rate in 1987 to 89, which was when the money center banks USA UK Euro Area Total
provided against around 60% of their Latin American sovereign debt
- Residential mortgages 166 11 104 281
exposures.
- Consumer 0 2 8 10
- Commercial Mortgage 48 8 40 96
Figure 40: Three year Cumulative Loan Provision Rates in Crisis Period
- Corporate 17 7 0 24
Ireland 2008-10 19.0%
- Governments 0 0 0 0
USA 2008-10 10.8%
- Foreign 66 29 72 167
Sweden 1990-92 9.2%
Total losses for Securities 297 57 224 578
Japan 1997-99 6.2% Source: IMF
Australia 1991-93 6.1%

UK 2007-09 5.9% Third, the IMF has counted 63 sovereign debt crises since 1970 but there has
Korea 2001-03 5.1% been no sovereign default by a developed economy since the 1930’s. Large
USA 1987-89 4.5% scale defaults and/or the breakup of the Euro are possible outcomes, but it is
Global average 2007-09 4.1%
impossible to calculate the consequences including the potential loan losses
Source: FDIC, Deutsche Bank estimates and capitalization requirements of those outcomes.
Loan loss provisions in developed markets between 2008 and 2010 ran at In the country sections which follow, we assess loan portfolios on a country
between normal levels and 6 times normal levels. by country basis and stress tests the coverage universe using two screens.

Screen 1: Two years severe recessionary losses as a percentage of 2012E Figure 44: Loan loss charges to average loans: Emerging Markets
tangioble net asset value. 2006 2007 2008 2009 2010 2010E 2012E 2012
Stress
Screen 2: Two years 2012E pre-provision profits, flexed down by 25% less Brazil 6.09% 4.63% 4.64% 6.12% 4.31% 4.41% 4.79% 6.22%
two years severe recessionary losses as a percentage of tangible net asset Mexico 2.35% 4.03% 5.81% 5.87% 3.87% 2.82% 2.51% 5.50%
value. Russia 0.92% 0.74% 2.42% 7.17% 2.55% 0.44% 0.43% 2.12%
India 0.67% 0.78% 0.79% 1.01% 0.93% 0.89% 0.82% 1.30%
Severe recessionary losses is generally taken as four times “normal” or
China 0.74% 0.75% 1.17% 0.46% 0.41% 0.54% 0.51% 2.04%
average loan losses over very long term, although analysts have deviated from
Turkey 0.57% 0.78% 1.12% 1.81% 0.26% 0.42% 0.51% 1.80%
this by exception. The table below summarises the actual and forecast country
Indonesia 2.24% 1.62% 2.33% 2.39% 2.07% 1.72% 1.85% 7.48%
loan loss aggregates from 2007 to 2012E as a percentage of average loans,
Malaysia 0.98% 1.02% 0.67% 0.79% 0.57% 0.40% 0.44% 2.76%
with the final column showing the estimate of severe loan loss charges. For
bank sector in the developed economies the range is from 1% to 3.5% and in Thailand 2.30% 2.76% 1.48% 1.32% 1.12% 0.97% 0.99% 1.17%
the emerging markets the range is from 1.3% to 7.5%. Korea 0.40% 0.54% 1.11% 1.20% 1.26% 0.75% 0.60% 1.76%
Poland 0.04% 0.11% 0.85% 1.68% 1.34% 1.10% 1.01% 4.02%
With the exception of the US, the UK and Ireland the assumption on a severe Source: Deutsche Bank
loan loss charge is above the peak GFC charge of 2009. For the US, the peak
charge is less than 50% of the 2009 charge and for the UK it is a somewhat For the US it could be argued that the loan loss charge of 2009 equivalent to
lower. 9X a normal charge was truly exceptional first in the context of a 30% decline
in house prices over two years and second in the context of exceptional fixed
income trading revenues which allowed banks to appease the markets by
Figure 43: Loan loss charges to average loans: Developed Markets
increasing loan loss reserving ratios.
2006 2007 2008 2009 2010 2010E 2012E 2012
Stress
The assumption of a 25% fall in pre-provision profits sustained over two years
USA NA 1.48% 3.55% 4.65% 2.62% 1.27% 1.01% 2.8%
may or may not be conservative. Large declines in pre-provision profits were
Australia 0.17% 0.19% 0.47% 0.82% 0.45% 0.30% 0.23% 2.12%
recorded by US and European wholesale banks in 2008 as securities losses
Japan 0.31% 0.31% 0.70% 0.38% 0.23% 0.35% 0.38% 1.00%
were taken against the revenue line. Clearly given the uncertainty attached to
Hong Kong -0.13% 0.03% 0.33% 0.21% 0.07% 0.20% 0.22% 1.00%
the value of GIIPS government bonds, there is scope for a repeat of significant
UK 0.82% 0.92% 1.75% 2.58% 1.55% 1.17% 0.83% 2.12% write-downs. As a benchmark a 25% decline in European bank pre-provision
Sweden -0.04% 0.03% 0.33% 0.96% 0.36% 0.18% 0.14% 2.58% profits sustained over two years for our European universe would imply a
France 0.26% 0.46% 1.03% 1.46% 0.90% 0.67% 0.48% 2.12% reduction in income of €110bn or US$1555bn against the expected outcome,
Germany 0.28% 0.44% 0.66% 0.49% NA NA NA NA equivalent to about 1% of loans and 3% of securities. That would be twice as
Greece 0.79% 0.70% 1.20% 1.77% 1.96% 1.91% 1.68% 3.60% bad as the outcome in 2008 and 2009 when compared with the run rate of
Ireland 0.10% 0.16% 1.26% 6.53% 11.24% 1.44% 1.17% 3.51% pre-provision profits in 2007.
Italy 0.50% 0.49% 0.70% 1.14% 0.97% 0.83% 0.71% 2.00%
Spain 0.55% 0.60% 0.99% 1.42% 1.36% 1.29% 1.31% 2.97%
Israel 0.56% 0.31% 0.82% 0.84% 0.49% 0.31% 0.44% 0.92%

Figure 45: Aggregate pre-provision profits by country 2007 to 2012E: Most of the 2008 decline in European and US pre-provision profits was
Developed Markets (local currency) attributable to mortgage related securities losses. A decline in European-pre-
2007 2008 2009 2010 2011E 2012E
provision profits of US$155bn would be equivalent of 37% of the US$ 421bn
of cross border claims by European banks on the GIIPS at end 2010.
USA 123,727 79,989 219,417 178,989 148,849 173,391
Australia 25,882 28,335 35,197 37,111 40,174 43,876
The table below summarises the results at country level. The banking systems
Japan 5,027 3,759 4,686 4,935 4,567 4,671
of most developed economies would be in loss. On average in the developed
Hong Kong 45,645 37,617 33,890 39,943 50,453 59,205 economies this stress tests costs banks in developed economies around 10%
UK 51,629 56,469 56,276 58,959 56,584 63,695 of NAV. The actual outcome would depend on the existing levels of loan lo0ss
Sweden 13,839 13,057 16,176 14,305 14,639 16,088 reserving, tax shields and other factors, but bank sectors would not in
France 19,034 18,635 28,575 34,178 36,700 39,611 aggregate significantly run down tier one ratios.
Germany 44,060 28,460 44,069 NA NA NA
Greece 5,947 6,361 6,443 5,516 5,923 5,927 Figure 47: Summary of Stress Tests
Ireland 6,830 6,739 6,569 4,260 733 859 2 Years of Recessionary 2 years of Recessionary Core tier one to Risk
losses as % 2012 losses as % of 2 years of weighted assets
Italy 29,409 24,202 26,548 23,295 23,873 27,598 Forecast Tangible Equity PPP (flexed down by 25%) 2012E)
Spain 28,805 34,497 41,531 41,029 40,425 44,536 USA -33.3% 4.6% 11.8%
Israel 12,640 4,645 13,695 11,948 12,073 13,488 Australia -61.0% -16.4% 8.5%
Source: Deutsche Bank Japan -22.3% -1.2% NA
Hong Kong -13.0% 16.9% 11.2%
The flex down assumption looks particularly conservative for emerging market UK -32.4% 2.2% 8.2%
banks, whose traditional banking models makes pre-provision profits less Sweden -68.1% -39.9% 8.6%
volatile. France -46.0% -3.9% 10.0%
Germany NA NA NA
Greece -82.4% -38.9% 7.4%
Figure 46: Aggregate pre-provision profits by country 2007 to 2012E:
Ireland -90.1% -75.1% 12.4%
Emerging Markets (local currency)
Italy -49.8% -14.0% 0.9%
2007 2008 2009 2010 2011E 2012E Spain -73.3% -14.1% 9.1%
Brazil 51,238 59,314 75,423 83,042 96,982 114,747 Israel -18.9% 10.1% 7.9%
Mexico 150,386 164,931 183,202 169,226 178,447 208,081
Russia 223,508 318,794 527,570 529,119 571,615 672,001 Brazil -67.7% 23.7% 10.8%
Mexico -40.9% 0.5% 15.4%
India 472,246 614,553 743,420 870,734 1,078,210 1,290,613
Russia -22.7% 22.0% 13.3%
China 531,430 692,684 666,650 857,761 1,080,339 1,247,177
India -23.0% 25.4% 8.0%
Turkey 15,074 15,181 24,380 23,104 21,404 23,537 China -32.0% 11.9% 10.5%
Indonesia 33,565 44,040 53,885 69,061 76,116 90,772 Turkey -18.5% 17.6% 16.1%
Malaysia 17,132 17,862 18,807 22,111 24,515 28,573 Indonesia -73.6% -21.6% 15.9%
Malaysia -43.3% -6.0% 6.0%
Thailand 120,549 154,772 160,155 192,722 230,827 262,143
Thailand -16.3% 20.9% 10.9%
Korea 17,717 18,059 16,387 20,791 25,497 24,587
Korea -29.0% 0.9% 8.8%
Poland 12,309 16,302 15,719 18,013 20,036 23,406 Poland -47.5% -8.9% 15.2%

The outcome in emerging markets is much more benign. The explanation is Figure 49: Country average pre-provision profit margins: Emerging
partly that emerging market banks have much high pre-provision profit Markets
margins. 2006 2007 2008 2009 2010 2011E 2012E
Brazil 13.7% 13.3% 11.5% 11.8% 11.0% 10.9% 11.2%
Figure 48: Country average pre-provision profit margins: Developed
Mexico 10.5% 10.2% 9.8% 10.1% 8.7% 8.1% 7.9%
Markets Russia 6.1% 4.9% 4.7% 6.8% 6.4% 5.8% 5.6%
2006 2007 2008 2009 2010 2011E 2012E
India 3.1% 3.4% 3.6% 3.7% 3.6% 3.6% 3.6%
USA NA 3.4% 2.1% 5.0% 4.4% 3.7% 4.3%
China 3.0% 3.8% 4.3% 3.3% 3.3% 3.6% 3.7%
Australia 2.3% 2.2% 2.1% 2.2% 2.1% 2.2% 2.3%
Turkey 8.7% 9.2% 7.0% 9.7% 7.7% 5.6% 4.9%
Japan 1.3% 1.2% 0.8% 1.0% 1.1% 1.1% 1.1%
Indonesia 7.3% 7.3% 7.5% 7.5% 8.1% 7.3% 7.1%
Hong Kong 4.0% 4.6% 3.4% 2.8% 2.8% 3.0% 3.1%
Malaysia 3.6% 3.6% 3.4% 3.1% 3.2% 3.1% 3.2%
UK 3.2% 3.0% 2.1% 2.3% 2.4% 2.3% 2.6%
Thailand 4.6% 4.1% 4.9% 4.8% 5.3% 5.5% 5.7%
Sweden 1.6% 1.5% 1.3% 1.5% 1.3% 1.3% 1.4%
Korea 2.5% 2.6% 2.3% 1.9% 2.4% 2.8% 2.4%
France 3.1% 1.9% 1.7% 2.2% 2.4% 2.5% 2.6%
Poland 5.4% 5.2% 4.8% 3.9% 4.2% 4.2% 4.4%
Germany 1.0% 0.8% 0.5% 0.8% NA NA NA Source: Deutsche Bank
Greece 3.7% 3.6% 3.0% 2.7% 2.3% 2.5% 2.5%

Ireland 1.7% 1.9% 1.7% 1.7% 1.2% 0.7% 0.8% The screen does show some counter-intuitive results. For instance Nordic
Italy 2.5% 2.4% 1.8% 1.9% 1.7% 1.7% 1.9%
banks screen as the most vulnerable to a severe loan loss scenario. Again, this
principally reflects pre-provision margins, which are low for Nordic banks
Spain 2.7% 2.8% 3.0% 3.3% 3.1% 3.0% 3.2%
relative to European and global peers; but the risk profile of Nordic loan books,
Israel 2.0% 2.2% 0.7% 2.1% 1.8% 1.7% 1.8%
with around 40% of total loans represented by mortgages and 35% by large
corporate is commensurately low.
And partly that analysts make assumptions on severe loan loss provisions
which might prove optimistic in a crisis give the very high rates of credit To provide an overlay to theses simple screens, we have attempted to assess
growth experienced over the last two and five years. the relative riskiness of countries with a heat/danger map score card, which
ranks countries on 9 factors on a scale of 1 to 5 with 5 being the highest risk.
The factors are (i) De-regulation of lending/changes in lending practice; (ii) bank
loans as a percentage of GDP; (iii) Change in the dbt to GDP ratio; (iv) maturity
of the cycle in years;(v) credit mix (vi) unemployment; (vii) current account
position;; (viii) level of real interest rates; (ix) Current account position.
Danger map scores are summarized below. In developed markets we find

Australia, Sweden, Hong Kong and Germany least risky and the GIIPS most
risky. We find emerging markets significantly less risk than developed market
banking systems.

Figure 50: Danger Map Scores: Developed Markets

USA Australia Japan Hong UK Sweden France Germany Greece Ireland Italy Portugal Spain Israel Average
Kong
Deregulation of Lending 1 2 1 1 2 1 2 1 2 3 2 1 1 1 2
% of Credit to GDP 4 3 1 3 5 4 3 2 1 5 1 5 5 3 3
Change in % of Credit to GDP 2 1 1 3 3 4 1 2 3 5 2 5 4 2 3
Maturity of Cycle in Years 2 3 2 3 3 1 3 2 5 2 2 3 3 2 3
Credit Mix 3 1 2 3 3 2 2 3 5 4 5 3 3 2 3
Unemployment 4 1 1 1 3 1 4 2 5 4 2 5 5 1 3
Current account position 4 2 1 1 2 1 2 1 5 2 4 4 4 2 3
Level of real interest rates 1 2 2 1 1 1 2 2 2 3 2 2 2 2 2
Exchange rate flexibility 1 1 2 5 1 1 5 5 5 5 5 5 5 2 3
Total Danger Map Score (out of 45) 22 16 13 21 23 16 24 20 33 32 25 33 32 17 23
Figure 51: Danger Map Scores: Emerging Markets

Brazil Mexico Russia India China Turkey Malaysia Thailand Korea Indonesia Poland Average
Deregulation of Lending 1 1 2 3 3 2 1 1 2 2 2 2
% of Credit to GDP 3 1 2 2 5 1 4 2 4 2 1 2
Change in % of Credit to GDP 5 3 3 4 4 5 3 1 3 2 4 3
Maturity of Cycle in Years 5 2 4 3 3 1 2 2 3 3 2 3
Credit Mix 3 2 4 4 4 2 2 2 3 2 2 3
Unemployment 1 3 2 1 1 2 2 2 1 1 3 2
Current account position 3 2 1 3 1 5 1 1 1 2 4 2
Level of real interest rates 2 1 2 3 3 2 2 2 2 2 3 2
Exchange rate flexibility 2 2 3 3 3 2 3 3 3 3 2 3
Total Danger Map Score (out of 45) 25 17 23 26 27 22 20 16 22 19 23 22
We score each factor on a 1(grey) to 5 (blue) basis with 5 denoting the greatest risk/danger

Crisis . . . what crisis?

domestic economy. In almost all instances for GIIPS banks, exposure to the
domestic sovereign is larger than tangible equity, and domestic loans are a
multiple of tangible equity.
The elephant in the room

The elephant in the room or at least in this report so far is the Eurozone debt
crisis, which is partly a “non crisis” involving elevated commercial bank
NPLS’s in a low growth environment and partly a crisis involving strained
government finances, potential debt default, and potentially, a major currency,
liquidity, and asset quality crisis in the world’s largest banking system by far.
One way of thinking about these exposures is to split them into in-market
(domestic) problems, and cross-border (contagion) problems. The idea that in-
market (domestic) problems are major, and generally we think terminal, should
not be contentious, but we do look at this issue below. Then we look at cross-
border contagion issues.
In-market risks: the domestic sovereign crisis

Before looking at cross-border risk, we should look at why sovereign risk and
banking sector risk become so closely entwined for banks, such that it
becomes a near-iron rule that a failed sovereign will almost always lead to a
failed banking system. The reverse is also often true when debt to GDP and
leverage are high, i.e. a failed banking system will lead to a failed sovereign, as
was the case for Ireland in 2008-10.
The mechanism by which sovereign to bank failure is transmitted is

straightforward, and takes place by two channels. First, a failed or weak
sovereign is unable to meet its “normal” commitments to spend, often
because it is frozen out of the international debt markets. This is effectively
the case in Greece at the moment. Savage retrenchment of government
expenditure is often a contributory factor to economic slowdown, which
drives domestic bad debts. Second, banks will routinely hold for liquidity
purposes large volumes of their parent country sovereign debt. Banks need to
hold large portfolios of liquid assets (and these are growing because of new
requirements such as LCR), and the data tell us that these are routinely met
through the domestic sovereign. Below we summarise data showing
European GIIPS banks’ exposure to their domestic sovereign debt, and to their

Figure 52: Illustration of latest Domestic Sovereign Exposures

Country Bank Shareholder Goodwill Tangible RWAs Tier 1 as Domestic As % T Eq Domestic As % T Eq Loan Book "EU Risk" "EU Risk" Total risk
s Equity Equity forecast GIIPS Loan GIIPS Sov Loans (if bonds (if bonds /
Book Risk known) known) TNAV
Italy Banco Popolare 11,887 5,155 6,732 93,215 7,813 93,661 1391% 11,374 169% 97,726 93,661 11,632 172.8%
Italy Monte dei Paschi 18,461 7,596 10,864 107,053 9,706 157,275 1448% 26,610 245% 161,150 157,275 27,135 249.8%
Italy UBI Banca 12,615 5,475 7,140 93,249 8,224 102,774 1439% 8,240 115% 103,786 102,774 8,242 115.4%
Italy Intesa SanPaolo 59,891 26,168 33,723 333,995 37,894 303,924 901% 64,473 191% 385,185 308,512 66,156 196.2%
Italy Banca Popolare di 4,659 840 3,819 45,432 3,852 32,359 847% 2,206 58% 36,104 32,359 2,343 61.4%
Milano
Italy UniCredito 69,417 25,192 44,225 478,086 48,230 250,400 566% 38,664 87% 564,289 250,400 41,055 92.8%
Italy Credito Emiliano 1,980 356 1,624 17,254 1,487 19,543 1203% 6,729 414% 20,289 19,543 6,729 414.3%
Ireland Bank of Ireland 8,883 0 8,883 75,093 11,085 76,561 862% 4,990 56% 105,000 76,561 0.0%
Iberia Bankinter 3,082 74 3,008 30,578 2,603 41,947 1395% 2,535 84% 41,947 41,947 2,535 84.3%
Iberia Banco Popular 8,712 640 8,072 90,639 9,018 87,902 1089% 8,874 110% 96,619 96,619 9,727 120.5%
Iberia BBVA 39,630 9,722 29,908 331,282 33,629 193,675 648% 53,452 179% 352,411 193,675 54,099 180.9%
Iberia Banco Santander 79,138 26,527 52,611 603,699 63,751 210,430 400% 41,807 79% 737,090 240,174 45,666 86.8%
Iberia Banco de Sabadell 6,009 850 5,159 56,248 5,512 68,847 1334% 7,296 141% 70,976 68,847 7,387 143.2%
Iberia BCP 4,912 472 4,440 57,777 5,365 58,916 1327% 5,829 131% 71,500 63,964 5,829 131.3%
Iberia BES 5,416 121 5,295 69,490 6,138 44,526 841% 2,683 51% 48,000 49,026 3,047 57.5%
Greece EFG EuroBank 4,064 734 3,330 42,617 5,393 41,589 1249% 7,700 231% 49,194 41,589 7,700 231.2%
Greece Alpha Bank 4,117 193 3,924 46,522 5,767 37,826 964% 4,200 107% 46,086 37,826 4,200 107.0%
Greece Piraeus Bank 3,470 384 3,086 36,962 4,137 29,012 940% 8,000 259% 35,790 29,012 8,000 259.3%
Greece ATE Bank 1,198 14 1,184 11,511 1,403 22,320 1885% 4,600 389% 19,164 22,320 4,600 388.5%
Cyprus Bank of Cyprus 2,928 479 2,449 27,342 3,996 10,589 432% 2,000 82% 28,651 10,589 2,000 81.7%
Cyprus Marfin Popular Bank 4,074 1,635 2,439 27,464 3,283 13,710 562% 3,000 123% 26,395 13,710 3,000 123.0%
These exposures show in a very practical way the link between sovereign
restructuring and a domestic banking crisis.

Cross border contagion risks: the sovereign crisis Figure 54: Pie Chart of Eurozone country exposures
Easily the largest component (60%) of bank exposures to Greece, Ireland,

Italy, Portugal, Spain (GIIPS) is in the form of loans to private sector entities,
often booked in and funded from local subsidiaries. Impairments from these
loans are already baked into analyst estimates. Public sector
Since the start of 2008, cumulative UK bank losses alone from non Sovereign Banks
and non-bank Ireland country exposures have exceeded €20bn – over 22% of Non bank private sector
total peak exposures. These losses are greater than the current exposure of all
European banks to Irish public sector debt. If the average loan loss rates
experienced by domestic banking systems are applied to the private sector
exposures in the table below, then cumulative losses on cross border loans to
private sector borrowers in the GIIPS are so far in the region of US$90bn. Source: Deutsche Bank
The cross border exposures to sovereign debt are actually quite modest. On A 2008 IMF study concluded that between 1970 and 2007 there were 124
IMF data the total assets of Europe’s banking systems are around US$28trn. banking crises, 208 currency crises and 63 sovereign debt crises. Carmen
On this basis the aggregate cross border sovereign exposures of GIIPS are Reinhart and Kenneth Rogoff describe banking crises in their book This Time Is
equivalent to about 1.5% of total European bank assets. Greek government Different as an “equal opportunity” event, as commonplace in developed
debt held by non Greek banks represents just 0.12% or so of total European economies as in emerging markets. But a sovereign debt default has not taken
bank assets. place in a developed economy in the 1930’s, shortly after the Great
Figure 53: Distribution of European Bank Claims on selected countries Recession, and then it was relatively minor (Spain suspended interest
payments on external debt between 1936 and 1939).
(US$bn) Public sector Banks Private Total Relative to total European
sector bank loans and Securities
Given that no one can how the sovereign debt crisis or indeed the Euro crisis
Portugal 32388 40447 121770 194605 0.55% will evolve and end and how a disorderly event should it occur would affect
Ireland 15355 70539 291742 377636 1.07% financial markets generally, it is impossible in our view to find a robust
Greece 54196 10918 80669 136317 0.39% framework to stress test losses or incorporate government default
Spain 88054 199269 344866 632189 1.79% assumptions into a normal credit cycle. An alternative but possibly no more
Italy 231216 127261 126891 485368 1.38%
useful approach is to see if other crises and post crisis periods provide a sign
post for crisis resolution, loan quality issues and the like.
Total 421209 448434 965938 1826115 5.18%
Source: IMF and BIS In the appendix to this report we include thumbnail length case studies of 9
Aggregate claims on the GIIPS by European banks including loans to banks banking crises, including the Latin American debt crisis of the 1980’s,
and non bank private sector borrowers of US$1.8trn represents about 5% of Argentina’s banking and currency crisis of 2001/02 and Japan’s banking crisis.
European bank assets. To put this into context, the loan exposures of Some of these crises were extreme boom and bust property cycles within the
systemically important US and UK international banks to defaulting Latin context of an inflationary macro-economic environment: asset bubbles were
American countries in the mid 1980s was equivalent to 10% of their assets blown up by deregulation and popped by monetary tightening. At least in this
and over 15% of their total loans. respect the US sub prime lending crisis was not so different from many
others, although “innovation” in lending practices rather than deregulation
Of course the European exposures in the table above do not include the was the catalyst. Others involved deflationary shocks and a subsequent
domestic banks’ own loan books or their holdings of their governments’ contraction in loan to GDP ratios over a multi year period.
bonds, of which more later.
Figure 55: Summary of Case Study Crises

Country Date Crisis thematic Duration Bank shareholders Policy and political response
Argentina 2001 Deposit run; unpegged currency/75% devaluation; systemic 2 years Survived but 65% down from pre-crisis 2000 Driven by events. Banks kept alive via issuance of
chaos, inflation, high unemployment and severe social levels to 2011 in nominal terms bonds to compensate them for losses of
hardship. Government default exchanging their US$ assets and liabilities for
Pesos at different rates
Japan 1997 A burst asset bubble. Deflationary and systemic pressures 5 years Survivor banks 90% down from 1989 peak and Learn as you go: many mistakes by regulators and
leading to heavy and persistent loan losses, recapitalisation now trading close to 25 year lows. Problems did Central bank but set template for the policy
and consolidation not end with loan clean up response to the 2008/09 crisis
Hong Kong 1997 Property crash and deflation as a consequence of property 2 years No recapitalisation required in spite of 70% Very proactive and wide awake before and after
bubble and fixed exchange rate in context of SE Asian banking decline in property. System never went even crisis. Actions included direct intervention in
crisis and currency devaluations. A period in which bank close to loss. But It took 10 years for bank currency and stock markets and deregulation of
lending contracted significantly in absolute terms and relative shares to reach post crisis highs because of the deposit markets
to GDP shrinkage of the loan book
Mexican default and 1982 Sovereign debt crisis and official debt repayment moratorium 7 years International banks wrote down US$61bn in 1987 Highly accommodative: regulators allowed unusual
Latam Debt Crisis involving 10 countries with total debts to international banks of and 1989. Limited recapitalisation required. Bank accounting practices to flatter balance sheets and
US$191bn. Debts were bank loans not government bonds shares performed surprisingly well The crisis set played for time. Banks effectively bailed out, at a
Loan exposures of US money center banks equivalent to 16% the scene for the consolidation of US money cost, by the IMF and World Bank. Policy making
of total loans and 200% plus of capital center banks driven by events
Sweden 1992 Property bubble following deregulation of system. Collapse 2 years Wiped out if their bank required state assistance Resolute, very effective implementation and clearly
required systemic recapitalisation and selective nationalization. Survivor banks were spectacular medium and thought out plan involving establishment of bad
Government debt to GDP doubled from 30% to 60% in 4 long term investments if bought near the crisis bank and rebuttal of moral hazard through
years and the stock or private sector debt contracted lows shareholder wipe outs
Australia 1992 Multi year property and lending bubble following deregulation 2 years Dividends were cuts. Some large Post crisis the regulatory system was overhauled
of system and entry of foreign banks recapitalisations. Bank shareholders survived and and reformed
then prospered
Ireland 2008 Collapse of property bubble and wholesale deposit run Ongoing Effectively wiped out State guaranteed all liabilities of domestic banking
followed by 20% decline in GDP and pressure on loan quality system and recapitalised it. Created a bad bank to
Loan losses tp GDP of 60% to 70% sets new record for a cleanse system of NPLs. Bank bail out costs almost
developed economy destroyed governments finances
USA 2008 Rising interest rates bursts a housing bubble fostered by the 2 years Bank stocks are down over 60% from pre-crisis Regulatory oversight contributed to the crisis
innovation of AAA sub-prime securitizations and a long period levels ROE remains below 20 year average of Government went into overdrive effectively
of low rates. As house prices declined financial institutions 11% and likely to remain depressed because of nationalizing the GSEs and injected capital into the
were exposed at over leveraged and some as insolvent higher capital and liquidity requirements banks via TARP and regulators effectively increased
capital requirements through SCAP
Korea 2003 The government used the financial sector to jump start the Credit card companies experienced large The government was directly or indirectly
Korean economy after the 1997 Asian banking crisis and then portfolio write downs and were forced into responsible for the boom and bust cycle
failed to contain the after effects of a massive credit mergers and or recapitalization The bank sector
stimulus/bubbles was at its bottom when the govt came out with
measures to stabilize the system

In spite of many differences, possibly the most relevant to the European crisis The pragmatic view, however, has been shaken by apparently hard lined
today was the Latin American debt crisis of the 1980’s. It involved many resistance to bail outs by the German government amongst others. Even
creditor and debtor countries. The debtor nations had borrowed too much without a systemic financial and banking crisis, a rigorous austerity regime
money in the wrong currency. US and UK megabank exposures were suggests a prolonged period of low growth and persistent deflationary
sufficiently large (16% of total loans and 240% of equity) to threaten the pressures on asset quality measures within the indebted economies of the
solvency of the international financial system. The crisis took a long time (7 Eurozone must lie ahead. Just how severe this turns out to be will depend to a
years) to work out. At the time informed observers were fearful that multiple large extent on property values.
country default could trigger a systemic collapse of the international banking
system and a global recession. The policy response was ad hoc and driven by Banking systems around the world are indirect but fairly obvious plays on
events. Politicians, the IMF, the World Bank, central banks, bank regulators, property, which back up 50% or more of loans and 20% to 30% of securities.
suspect accounting principles and financial markets all played their part in So far since 2009, most banking systems have been protected by the better
providing a fix, and ultimately (which coincided with when their balance sheets than expected resilience, at least in nominal terms, of house prices. Even in
could bear the cost) commercial banks took very large (US$60bn) write downs Greece, house prices have held up relatively well and although US house
as a prelude to the resolution of the crisis. prices remain weak, the chances of a further 2008 style decline look very
remote given the relationship between house prices and household income.
Possibly the major difference between the Eurozone crisis and any other over Only in Ireland has there been an outright collapse in residential and
the last 70 years is that the domestic banking systems within the Eurozone commercial real estate values.
hold unlikely weapons of self-destruction within their balance sheets in the
form of their own governments’ bonds, and that they are all inextricably linked. Figure 56: Index of House Prices: Developed Markets
A disorderly default on government debt and withdrawal from the Euro would 2002 2003 2004 2005 2006 2007 2008 2009 2010 % ch from
leave a massive recapitalization requirement even before the huge increase in high/low
private sector foreign currency loan impairments was factored in. USA 100 111.3 129.2 149.1 150.2 136.2 110.7 107.2 104.6 -30%
Australia 100 118.2 125.9 127.7 137.7 153.3 160.0 165.5 185.7 186%
Until recently, the market appeared to have been banking on Benjamin Japan 100 93.6 87.8 83.4 81.1 81.4 82.8 79.9 76.2 -24%
Franklin’s laconically grim political truism on signing the Declaration of Hong Kong 100 88.0 119.7 140.0 137.2 148.2 176.4 172.8 212.6 213%
Independence on July 4th 1776: “We must all hang together, Sir, or most
UK 100 119.5 139.9 147.1 156.5 170.8 159.3 147.5 156.1 -9%
assuredly we will all hang separately” The pragmatic view has been that at the
Sweden 100 109.8 120.0 129.0 144.4 158.7 150.6 165.9 174.6 174%
end of the day the political, social and economic disincentives to disorderly
France 100 111.7 128.7 148.4 166.3 177.3 179.4 166.7 177.2 -1%
debt repudiation or to abandoning the Euro would be greater than the reflation
Germany 100 100.5 100.1 100.5 101.0 102.1 103.3 104.8 106.8 7%
incentive. In 2002 Argentina devalued its currency and defaulted on its
external debt after its banking system suffered a run on its deposits, which Greece 100 105.4 107.8 119.6 135.1 143.5 145.7 139.4 136.6 -5%
remains a possibility within the Eurozone. The Argentinean banking crisis of Ireland 100 113.7 123.5 135.0 150.9 139.9 127.1 104.5 93.2 -38%
2001/02 (which occurred when the ratio of private sector debt to GDP was Italy 100 106.1 112.6 121.3 128.1 134.7 138.2 137.7 137.8 -1%
just 27%) led to the near immediate resignation of the government. The Spain 100 116.3 132.9 147.2 156.4 159.7 154.6 142.7 133.8 -16%
subsequent 75% devaluation impoverished bank depositors, triggered a Source: Deutsche Bank, Case Schiller, Nationwide, IMF
severe recession, an inflationary spiral and a prolonged period of high

unemployment, social hardship and civil unrest. Nevertheless, Argentina’s
economy and bank sectors did subsequently recover and within a relatively
short time span.

The Japanese banking crisis provides the most resonant template for the rose from 15%of GDP pre crisis to 58% in 2000. This level of indebtedness
aftermath of a banking crisis in the context of a low interest rate and low was stratospheric relative to other emerging markets and to developed
demand environment. The Japanese bank loan to GDP ratio moved from markets at that time. Thereafter the ratio of total credit to GDP declined over
about 80% in 1983 to 110% at the peak of the bubble in 1989 and then the next 11 years to 112%, the level pertaining in 1996. This adjustment was
reverted towards mean over the next 15 years. Because there was very little taken initially by a very significant (35%) contraction in the stock of
nominal growth in GDP, the mean reversion was accomplished through a commercial bank credit which fell for 4 consecutive years until recovering
significant reduction in the stock of loans. During this process declining loan strongly from 2002. The Thai banks were forced to recapitalize and their share
quality and high impairments were a persistent issue for the banking system, prices fell 90% plus over the crisis year. After an initial recovery period, Thai
since deflation triggered declines in commercial property values and re- bank shares basically flat lined during the subsequent 9 year deleveraging
provisioning requirements on old NPLs and increased the stock of non- period and began to perform strongly from 2007, when commercial bank
performing loans. As commercial bank loans to GDP shrank, the government loans to GDP troughed at 72%.
stepped up to the plate and issued bonds, which were largely bought by
domestic households and domestic institutions, particularly banks. This Figure 58: Thailand: Banking System and Government debt to GDP 1994
deleveraging process was incredibly painful for the industry..During the period to 2010
of deleveraging loan losses were very high but thereafter fell sharply.
190
180
Figure 57: Japan: government and bank loans to nominal GDP against
loan quality 170
250% 9% 160
8% 150
200% 7% 140
6% 130
150%
5% 120
4% 110
100%
3%
100
50% 2%
1993 1995 1997 1999 2001 2003 2005 2007 2009
1%
%
0% 0% Total sytsem and government loans to GDP
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Provisions to loans(RHS) Risk management loans to total loans(RHS)
The other crisis and post crisis which is possibly relevant to the current period
Government debt (LHS) Banking loans(LHS)
is Hong Kong’s in 1997. Loose US monetary policy was imported into Hong
Kong’s economy via the HK$/US$ peg and real estate values doubled between
In Thailand’s banking crisis a similar reversion to mean took place. The 1992 and 1997. The Asian banking crisis which the devaluation of the Thai
devaluation of the Thai currency in July 1997 contributed to a huge spike in the currency triggered took place just as US monetary policy tightened
loan to GDP ratio, from 110% to 144% and substantial losses within the considerably. Short term interest rates rose from 5.6% in 1996 to 9.5% at end
financial system. If government debt is included the ratio of total debt jumped 1997. Hong Kong was unable to devalue its currency to regain its
from 125% of GDP in 1996 to 185% of GDP, in 1997 as government debt competitiveness against its South East Asian competitors, whose currencies

were significantly devalued in 1997 and 1998. Adjustment therefore had to be

taken through asset values and real wages. Nominal GDP was more or less
static between 1996 and 2002, the stock of bank credit contracted by 47%
from 318% of GDP in 1996 to 163% in 2002 and over that period and there
was a 60% decline in house prices.
Figure 59: HK Macro 1992 to 2001

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
GDP local currency (HKD bn) 805 928 1,047 1,116 1,229 1,365 1,293 1,267 1,318 1,299 1,277
Average short term interest rate Na 3.5% 4.8% 6.2% 5.5% 7.4% 8.5% 6.0% 6.2% 3.7% 1.9%
Period end short term interest rate 4.5% 3.7% 6.3% 6.0% 5.6% 9.5% 5.7% 6.1% 6.0% 2.0% 1.5%
Period end 10 year government bond yield Na na na na 7.3% 8.9% 6.6% 7.7% 6.6% 6.2% 4.6%
CPI 78 84 92 100 106 113 116 111 107 105 102
Current account surplus/deficit (%) Na 5.4% -1.1% na na -4.4% 1.5% 6.3% 4.1% 5.9% 7.6%
House price index 100 111 147 126 142 204 134 120 106 92 82
Banking loans (HKD bn) 2,470 2,857 3,265 3,739 3,915 4,122 3,304 2,813 2,461 2,185 2,076
Hong Kong banks survived this more or less unscathed and for three reasons. So what crisis or post crisis environment is this?
The first is that loan losses from mortgages were very modest because the
banks’ regulator, the HKMA which had been wide awake before and after the The table on the next page compares the post crisis environments on various
crisis, had decreased the maximum loan to value loan progressively as the debt and GDP parameters following five severe banking crises, four of which
bubble built up and therefore loan underwriting standards were very tight; are covered in more detail in a later section of this report. Four trends are
second, bank capital ratios were very high and funding for the large banks was clear. First, real GDP growth was very slow or negative (Sweden and Thailand)
retail; third, the HKMA was extremely proactive in boosting confidence in the over a five year period. Second, there was a very rapid increase in government
system. In mid August 1998 and controversially at the time it used Hong debt in all cases (bar Hong Kong, where there was no government debt) with
Kong’s Exchange Fund to support the currency and buy the stock market in an average increase of 154% and the average ratio of government debt to
large size (US$15bn), with the intention of deterring and damaging short GDP rose by 35 percentage points to 75%. This is a higher growth rate than
sellers, successfully as it turned out. Over the next 18 months the Hang Seng experienced by most developed economies after the GFC but the starting
index, which had fallen over 60% between July 1997 and August 1998, rose point of government debt to GDP ratios was much lower at 40% on average,
three fold, surpassing its pre crisis peak. Hong Kong bank shares, however, including the ratio of 100% for Japan.
more or less flat lined from 1998 to 2006 and did not revert to their pre crisis
peaks until 2007, and then only for a brief period.

Figure 60: Crises Compared It seems to us that the many of the banking systems in the developed
Sweden Australia Hong Thailand Japan Average UK USA Ireland economies are in a debt deflation environment and that because debt levels
1990/94 1990/94 Kong 1997/01 1997/01 2008/12E 2008/12E 2008/12E
1997/01
are at post war record highs the equilibrium is quite unstable. This suggests a
Cumulative increase 0.0% 24.1% 5.7% 15.2% -3.0% 8.4% 12.9% 12.9% -15.3%
bad and crisis prone background which is likely to be of long duration for
in nominal GDP t-1 to financial institutions.
t+4
Cumulative increase -0.1% 12.6% 9.8% -1.3% 2.4% 4.7% 2.6% 8.7% -7.4% The good news is that the private sectors of the US and the UK started to
in real GDP t-1 to t+4
deleverage immediately and recapitalized their banking institutions early on.
Increase in 114% 132% NA 322% 49% 154% 121% 87% 310%
Government debt t-1
That immediate private sector deleveraging and bank recapitalisation did not
to t+4 occur in Japan is one of the many explanations given for the scale of its
Increase in bank 0% 10% -14% -13% -9% -5% 10% 4% -42% banking crisis and the duration of its subsequent economic problems.
lending t-1 to t+4
The bad news is that the adjustment has not happened in the Eurozone
Govt debt to GDP 28 17 NA 15 100 40 44 62 25
Ratio t -1 (%) countries either at all or to anything like the same extent. A good outturn for
Govt debt to GDP 62 31 NA 54 153 75 87 103 114 loan quality within bank sectors might be for private sector debt to GDP ratios
Ratio t +4 (%)
(and loan to deposit ratios) to revert towards mean without significant
Bank loans to GDP 93 81 149 129 106 112 143 174 192
ratio t (%)
weakness in asset prices over a multi year period. This may be a tall order.
Bank loans to GDP 88 76 138 72 100 95 135 159 149
Prolonged austerity in the Eurozone will most likely equate to prolonged and
Ratio t+4 (%) severe loan quality problems.
If Japan, Thailand, Hong Kong and possibly Ireland provide any signposts the
Third, over the same five year time period of t-1 to t+4 bank lending period in which private sector debt ratios revert to mean is likely to be in the
contracted on average by 5%. Only in Australia after its banking crisis in 1990 region of 7 to 10 years. This would suggest that the post crisis period will be
did the stock of credit increase. The largest contraction was in Hong Kong and of long duration and that there is no quick fix.
Thailand (for Hong Kong we have used mortgage and consumer loans rather
The Latin American debt moratorium episode, insofar as it provides a template
than total loans: the stock of total loans shrank by around 45%). Fourth, in all
for a multi country sovereign debt crisis, also points to a long duration for the
economies, the stock of loans fell relative to GDP.
government debt problems of the periphery countries and other developed
When the experience of the US and UK post crisis is compared with the four economies.
deflationary shocks it looks as if bank sectors and economies are following a
The share price performance of bank sectors in Japan, Thailand and Hong
classic form book: slow or negative real GDP growth for a prolonged period, a
Kong over the 10 years that followed on from the bubble peak might suggest
significant increase in government debt, an abrupt slowdown or contraction in
that post crisis or post bubble, deleveraging and adjustment periods do not
the stock of private sector debt, and, possibly, a reversion to the mean level of
provide a great environment for bank shareholders. Bank sectors in Sweden
private sector debt to GDP.
and Australia struggled and underperformed markets in the early post crisis
The two differences between the post 2009 environment and previous crises period and subsequently performed well, but the era of deleverage in Australia
are first that levels of debt are much higher on average post GFC (this is not and Sweden was a short and shallow and thereafter followed a long period in
universally true as Japan had already a debt to GDP ratio of 253% by 2002); which loan growth outpaced nominal GDP growth and in which housing
and second that government credit was considered good after the previous enjoyed a long bull market.
crises. Arguably, when banks are unable to buy the bonds of their own
sovereigns, the traditional liquidity, capital adequacy and banking models
based on the hierarchies of credit has broken down.
Country Sections

United States Credit Data: Trends

Below we summarize data on revenue, profitability, and asset levels from
2007 to 2010, as well as our forecasts for the US banks under our coverage.
Credit losses likely peaked in 2009 and will likely improve meaningfully over
Overview the next 1-2 years.
In this section we look at the US banks, with a focus on those covered by
Deutsche Bank. Figure 61: Aggregation of quoted bank data, local currency (USD bn), US
2007 2008 2009 2010 2011E 2012E
Since the US credit crisis began in 2007, loan loss provisions have Net Interest Income 180,406 219,782 268,911 252,038 235,427 240,258
averaged 3.1% of average loans. We believe credit losses for US banks Other Operating Income 153,754 98,947 235,225 219,616 199,978 207,981
Total Revenue 334,160 318,729 504,136 471,654 435,404 448,238
peaked in 2009 (at 4.65% of loans) and declined to 2.6% in 2010. Costs 200,166 226,343 268,653 292,664 287,056 272,730
Pre-Provision Profits 123,727 79,989 219,417 178,989 148,349 175,509
Despite the US labor and housing markets slowly recovery, we believe Loan Loss Provisions 53,114 137,188 205,429 106,896 50,339 43,612
credit quality at US banks will continue to improve given most US banks Pre-Tax Profit 80,142 -41,036 40,701 69,595 95,758 129,611
continue to runoff high-risk loans, and benefit from loss mitigation, loan Total Assets 7,369,372 8,800,475 8,987,487 9,050,791 9,094,405 9,135,895
resolution/modification programs. We estimate loan loss provisions Average Assets 7,064,448 7,469,266 9,133,904 9,193,725 9,096,096 9,074,022
declining to about 1% of average loans by 2011 (or $40-45b—half of what Risk weighted assets 5,196,550 6,021,803 6,112,370 5,820,995 5,762,377 5,798,995
Total Loans 3,444,758 4,511,691 4,176,454 4,010,565 3,940,907 3,989,246
it was in 2010). Revenue / Average Loans 9.3% 8.2% 11.4% 11.5% 10.9% 11.1%
Pre-Provision Profit / Average Loans 3.4% 2.1% 5.0% 4.4% 3.7% 4.4%
Leverage ratios remain elevated in the US. Nonfinancial (and Loan Loss Provisions / Average Loans 1.48% 3.55% 4.65% 2.62% 1.26% 1.08%
nongovernment) debt to equity (or net worth) remains well above PPP / Loan Loss Provision Cover 2.3 0.6 1.1 1.7 2.9 4.0
historical levels. In total, US debt/equity was 34% at the end of 2010, Note: Aggregated data from BAC, BBT, C, CMA, FHN, FITB, HBAN, JPM, KEY, MTB, PNC, RF, STI, TCB, USB, WFC, ZION.
Source: Company data and DB estimates.
down from a peak of 38% at 12/31/08, but well above the historical
average of 28% since 1986.
Estimates by Bank
Credit mix has improved for US banks, given many banks have
repositioned loan portfolios by reducing exposure to subprime, option Below we summarize our credit loss forecasts for 2012 for the US banks in
adjustable rate mortgages, home equity, non-owner occupied/multi-family our coverage universe. We then show our estimate of severe recessionary
commercial real estate (CRE). loan losses, which for the purposes of this exercise we have taken as 4x
normalized. We then express this as a percentage of 2012 tangible book value
For the US banks under our coverage, residential real estate (including
(TBV), and also as a percentage of 2012 pre-provision profit (PPP) flexed down
home equity/2nd lien mortgages) represent 30% of loans and CRE
by 25%, as a proxy of a potential contraction during a crisis.
represents 16%. In our view, these categories represent the longer-term
credit risks for the sector.
We estimate annual credit losses of USD $80b for the US banks. Over two
With regards to the Danger Map, we score the US at 22 out of 45. This years, this would be equivalent to 20-25% of 2012 TBV. If we include PPP
puts US in the middle of the pack on the Danger Map. estimates as an offset, most US banks would be breakeven. For those that
wouldn’t be (CMA, FHN, HBAN, KEY, RF, STI, and ZION), losses would be
about 3% of TBV (i.e. these banks would be loss-making).

Figure 62: Summary 2012E data and comparing recessionary loan loss charges as % of opening TBV and of PPP
USD ($M) Loan Loss Provisions Pre-Provision Profit (PPP) 2 Years of Recessionary losses
Less Recessionary As % 2012 Tangible % incl. 2 yrs. of PPP
Ticker Avg Loans Forecast Severe Recession Forecast Losses Forecast Equity (flexed down by 25%)
BAC 884,587 8,399 17,692 32,339 14,647 35,383 25% NM
BBT 112,315 913 2,246 3,597 1,351 4,493 37% NM
C 636,433 11,681 12,729 30,762 18,033 25,457 16% NM
CMA 41,879 266 838 876 38 1,675 28% 6%
FHN 16,100 69 322 321 -1 644 27% 7%
FITB 80,939 540 1,619 2,366 747 3,238 30% NM
HBAN 41,448 307 829 1,052 223 1,658 34% 2%
JPM 697,856 8,076 13,957 44,942 30,985 27,914 19% NM
KEY 54,822 372 1,096 1,361 265 2,193 24% 2%
MTB 55,685 286 1,114 1,499 385 2,227 43% 0%
PNC 153,413 705 3,068 5,494 2,425 6,137 25% NM
RF 85,236 703 1,705 1,996 291 3,409 36% 4%
STI 115,951 976 2,319 2,869 550 4,638 33% 2%
TCB 16,326 139 327 457 131 653 37% NM
USB 209,963 2,054 4,199 9,163 4,964 8,399 35% NM
WFC 781,367 7,902 15,627 35,616 19,989 31,255 28% NM
ZION 37,846 223 757 799 42 1,514 34% 7%
Total 4,022,167 43,612 80,443 175,509 95,065 160,887 23% 3%
Source: Deutsche Bank estimates, (*) flexed down by 25%
United States in the middle of the pack on the Danger Figure 63: Scoring the “danger map”
Danger Factor Score Comment
Map
Deregulation of Lending 1
Below we summarize our Danger Map indicators for the US. We score the US % of Credit to GDP 4 Higher than historical averages
overall at 22 out of 45. This puts US in the middle of the pack on the Danger Change in % of Credit to GDP 2 Improving
Map. This represents the aggregate of a 1 to 5 score across nine risk metrics,
Maturity of Cycle in Years 2
drawn from our analyses of case studies, and our global experiences of bank
Credit Mix 3
lending.
Unemployment 4 Unemployment remains elevated
Current account position 4
Overall, we rate the US as less risky in terms of, deregulation of lending, level
of real interest rate, and exchange rate flexibility. We rate the US as more risky Level of real interest rates 1
in terms of % of credit to GDP, unemployment and current account position. Exchange rate flexibility 1
Total Danger Map Score (out of 45) 22

Understanding the US banks’ loan books

For the US banks under our coverage, residential real estate (including home
equity/2nd lien mortgages) represent 30% of loans and CRE represents 16%.
In our view, these categories represent the longer-term credit risks for the
sector—particularly if the US economy deteriorates further. Write-downs for
residential mortgage and home equity loans for the US as a whole still have a
ways to go, in our view. For example, we estimate that 35-45% of home
equity has little to no equity backing the loan. And for first lien mortgages, we
estimate 30-40% have LTVs of 95%-plus.
Figure 64: US banks under coverage: loan mix (at 3/31/11): Total of USD
4.1b
Credit Card, 1%
Leases, 2% Commercial &
Construction & Industrial, 21%
Development, 4%
All Other, 5%
Mortgage, 17%
Other Consumer,
10%
Home Equity, 14% Commercial Real
Estate, 16%
Note: Data is based on US bank regulatory filings as of 3/31/11. Aggregated data from BAC, BBT, C, CMA, FHN, FITB, HBAN, JPM, KEY,
MTB, PNC, RF, STI, TCB, USB, WFC, ZION.
Source: SNL Financial

Australia Credit Data: Trends

Below we summarise data from 2005 to 2010 for the aggregated Australian
major banks in our coverage universe. We see declining PPP combining with
rising loan losses to reduce PPP/LLP cover from 13.7x to 2.7x. We note that a
Overview majority of the loan loss provision has been used to increase collective
In this section we look at the major Australian banks, with a focus on provision levels which remain elevated.
those covered by DB (ANZ, CBA, NAB and WBC). Over the last five years,
loan loss provisions in the banking book averaged 42bp, and peaked at Figure 65: Aggregation of quoted bank data, local currency (AUD$m),
82bp. Australia
Local Currency 2006 2007 2008 2009 2010
We believe that should there be another economic downturn, the
Australian banking sector is better placed compared with prior to the Net Interest Income 27,785 30,395 34,110 43,712 46,867
global financial crisis to withstand such shocks as: Other Operating Income 18,994 18,790 19,918 20,217 21,479
Total Revenue 46,779 49,185 54,028 63,929 68,346
Banks provision levels have significantly increased, measured on any
Costs 22,291 23,303 25,693 28,732 31,177
metric. For example, the Australian banks are currently holding
Pre-Provision Profits 24,488 25,882 28,335 35,197 37,169
significant levels of collective provisions which can be drawn on
when impaired assets increase. Loan Loss Provisions 1,786 2,273 6,512 13,108 7,906
Pre-Tax Profit 22,702 23,609 21,823 22,089 29,263
Banks have significantly reduced their exposures to high risk sectors
such as commercial property in recent years in response to the GFC.
Total Assets 1,489,237 1,757,207 2,054,943 2,341,064 2,482,298
Australian housing is well insulated from any significant downturn in Average Assets 1,423,027 1,609,856 1,911,065 2,359,868 2,439,524
house prices with low average LVRs, lenders’ mortgage insurance for Risk weighted assets 968,397 1,103,708 1,019,951 1,172,166 1,179,100
LVR>80% and ample room for the central bank to move down rates Core tier one capital 53,945 58,655 59,741 83,248 85,593
(current cash rate is 4.75%).
Risk Cushion above 7% plus 1 24,488 25,882 28,335 35,197 37,169
year PPP
Whilst we cannot see the banks making these losses, even in an
extremely recessionary scenario where losses reach 2.1% of total GLA, Total Loans 1,119,538 1,279,296 1,468,010 1,716,393 1,774,966
or an aggregate credit loss of $40bn for the four majors per annum, the Average Loans 1,059,634 1,199,417 1,373,653 1,592,201 1,745,679
Australian banks will still be profitable with a combined total profit before Revenue / Average Loans 4.4% 4.1% 3.9% 4.0% 3.9%
tax of $3.2bn. We note this loss scenario is significantly higher than the Pre-Provision Profit / Average 2.3% 2.2% 2.1% 2.2% 2.1%
peak loss experienced during the GFC which was 0.82% of GLA. Loans
Loan Loss Provisions / Average 0.17% 0.19% 0.47% 0.82% 0.45%
Based on our analyses of case studies, and our global experiences of Loans
bank lending we put Australia at the low end of the pack on the Danger PPP / Loan Loss Provision Cover 13.7 11.4 4.4 2.7 4.7
Map. Source: Deutsche Bank, Company Data
Capital ratios have improved significantly, with Core Tier 1 ratio increasing
by 2.4% to an average of 7.7%.

Estimates by Bank Figure 67: Scoring the “danger map”

Below we summarise forecast data for 2012 for the aggregated major
Deregulation of Lending 2 No recent deregulation of lending
Australian banks in our coverage universe (ANZ, CBA, NAB and WBC). We
% of Credit to GDP 3 Credit to GDP is 130%
then show our estimate of severe recessionary loan losses, which for the
Change in % of Credit to GDP 1 Only 1% growth in last 3 yrs
purposes of this exercise we have taken as 2.1% of total GLA.
Maturity of Cycle in Years 3 Year 3 of the recovery
We then express this as a % of 2012 BV, and also as a % of 2012 pre- Credit Mix 1 Heavily mortgage led with exposure to
risky lending (e.g. commercial property)
provision profit. These estimates show in our view that (1) the Australian significantly reduced
banks have good pre-provision profit generation, insulating them from Unemployment 1 Historically low unemployment
potential loan losses, and that (2) in a severe credit book downturn, we still Current account position 2
expect a profit. Level of real interest rates 2
Exchange rate flexibility 1
Figure 66: Summary of forecast data and comparing recessionary loan
loss charges as % of opening TBV and of non-flexed PPP Source: Deutsche Bank
Local Currency, 2012E ANZ CBA NAB WBC Total

Forecast Average Loan Balances 413,564 528,140 432,572 505,768 1,880,044 Understanding the Australian banks’ loan books
Forecast Loan Loss Provisions 1,045 1,012 1,491 960 4,508
As discussed above, we believe Australian banks have low risk in their loan
Severe Recession Loan Loss 8,768 11,197 9,171 10,722 39,857
Provisions portfolios which has a high proportion of housing lending. This proportion has
increased over the years.
2012 Pre-Provision Profit 10,160 11,737 10,276 10,849 43,023
2012 PPP less Recessionary Losses 1,393 541 1,106 127 3,166
Figure 68: Housing a major proportion of loan book
Personal
2012 Forecast Tangible Equity 33,668 30,463 33,773 33,314 97,904 2%
2 Years of Recessionary losses 17,535 22,393 18,341 21,445 58,269

2 Years of Recessionary losses as % -52% -74% -54% -64% -60%
2012 Forecast Tangible Equity
Business
2 years of Recessionary losses as % -7% -16% -9% -16% -16% 38%
of 2 years of PPP (flexed down by
25%) Housing
Source: Deutsche Bank, Company Data 60%
Danger Map Indicators for Australia

Source: Deutsche Bank, Company Data
Below we summarise our Danger Map Indicators for Australia. We score
Australia overall at 16 out of 45. This captures a 1 to 5 score across a number
of risk metrics, drawn from our analyses of case studies, and our global
experiences of bank lending. We rate Australia as low on the mix,
unemployment and exchange rate flexibility.

Figure 69: Housing has increased as a proportion over time Figure 71: Write-off history for mortgages
80% 0.15%
70% 0.10%
60% 0.05%
50% 0.00%
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
40% -0.05%
-0.10%
30%
-0.15%
20%
-0.20%
10%
-0.25%
0%
-0.30%
1994 1996 1998 2000 2002 2004 2006 2008 2009 2010 2011
-0.35%
Housing Business
-0.40%
Source: Deutsche Bank, Company Data Source: Deutsche Bank, Company Data
Figure 70: Total provision to RWA improved

1.80%
1.60%
1.40%
1.20%
1.00%
0.80%
0.60%
0.40%
0.20%
0.00%

Japan
Figure 72: Aggregation of quoted bank data, local currency (JPY bn),
Japan
Local Currency 2006 2007 2008 2009 2010
Net Interest Income 8,192 8,313 8,446 8,425 8,104
Overview Other Operating Income 3,442 3,147 1,866 2,637 3,181
Total Revenue 11,634 11,460 10,313 11,062 11,285
The average credit cost ratio for all Japan’s listed banks over the past five
Costs 6,234 6,432 6,554 6,376 6,350
years is very low at 39bp. If we exclude FY3/09, the year of the Lehman
Brothers Shock, the average declines to 30bp.
Loan Loss Provisions 1,290 1,355 3,141 1,721 997
There are two broad reasons for low credit costs. First, the writing down
Pre-Tax Profit 7,109 5,426 -342 2,753 3,920
of bad debts, which ballooned during the late 1980s bubble, has been
completed. Accumulated credit costs from FY3/95 through FY3/04 are
Total Assets 747,757 767,197 794,097 803,458 827,095
equivalent to 20% of FY3/95 lending. Most of the companies whose
Average Assets 750,773 757,477 780,647 798,777 815,276
survival was called into question in the late 1990s have already gone into
legal proceeding. In addition, banks have made ample provisioning for Total Loans 425,465 435,397 455,894 439,706 436,651
loans likely to become bad debts. On the other hand, equity ratios for Average Loans 421,671 430,431 445,645 447,800 438,178
Japanese companies as a whole are improving year by year, so bank Revenue / Average Loans 2.8% 2.7% 2.3% 2.5% 2.6%
borrowing is declining. Lending growth is being sustained by an increase Pre-Provision Profit / Average Loans 1.3% 1.2% 0.8% 1.0% 1.1%
in housing loans, while lending to companies is declining. This is a reason Loan Loss Provisions / Average 0.31% 0.31% 0.70% 0.38% 0.23%
for recently low credit costs. Loans
PPP / Loan Loss Provision Cover 4.2 3.7 1.2 2.7 4.9
Present market consensus, which called for a very moderate rise in credit Source: Deutsche Bank aggregation of company data
costs from 30bp before the failure of Lehman, now expects costs to
contract to 20-25bp. In such circumstances, if stress occurs, a level of
Credit Data: Trends
about 100bp would be appropriate. This is far below the 330bp recorded
in FY3/98 but above the 70bp seen in the immediate aftermath of the Overall NOP (PPP) for Japan’s listed banks is stable at around ¥5trn. NOP
collapse of Lehman. If credit costs were 100bp, they could basically be tends not to fluctuate based on such core businesses as loans and deposits,
absorbed every year by net operating profit (PPP) for the five major banks. but is more sensitive to gains and losses on bond holdings. In Japan as a
whole, the loan-deposit ratio has declined and is currently about 81%, and
We believe bad debt problems are unlikely to re-ignite in Japan, where
bank holdings of yen-denominated assets, such as JGBs, are on the increase.
credit growth is limited. Nevertheless, caution is necessary when it
Even during FY3/09, when Japanese banks reported overall losses, credit
comes to TEPCO loans, with faces nuclear reactor problems. The
costs were within NOP, but impaired losses on shares and foreign bonds
government is considering compensation for regions damaged by the
increased, pushing overall bank earnings into the red. Therefore, the level of
earthquake, so under normal circumstances this would not place any
stress exerted on Japanese banks likely depends on market risk rather than
burden on the banks. That said, if political uncertainty persists, there is a
credit risk. However, after the collapse of Lehman, foreign bond holdings fell
risk not only of delays in making legislation but also of some banks
sharply. The data below are based on parent bank figures, and exclude
shouldering part of the burden.
subsidiaries including securities subsidiaries and non-banks. In Japan, data is
released for parent and consolidated bank operations at the same time.

Estimates by Bank SMFG and Mizuho FG at about 10%. That said, a Cabinet decision has already
been reached regarding government support for TEPCO, and in the absence
Listed Japanese banks can be divided into the five major banking groups and of resistance from opposition parties action based on the decision should
the 83 regional banks. Among the five majors, Mizuho FG and Sumitomo proceed smoothly.
Mitsui Trust HD (SMTH), which have a relatively high proportion of loans to
large companies, have low credit costs. In FY3/11, credit costs for parent Figure 74: Scoring the “danger map”
banks were virtually zero (reverse profit 2.5bp, at Mizuho and 3.5bp cost at Danger Factor Score Comment
SMTH). For MUFG and Sumitomo Mitsui FG (SMFG) credit costs were about
Deregulation of Lending 1 Deregulation had completed.
20bp. We look for these levels to continue in FY3/12. If credit costs of 100bp
% of Credit to GDP 1 90%, down from 106% in 1993
continued for two successive years, we estimate that the impact on tangible
Change in % of Credit to GDP 1 Stable in the last few years
equity would likely be small, with credit costs at Resona HD relatively high
Maturity of Cycle in Years 2 Loans are basically short-term
among the five bank groups.
Credit Mix 2 Well diversified. Loans to Tokyo Electric
Power may be subject to political debates.
Figure 73: Summary of forecast data and comparing recessionary loan Unemployment 1 Unemployment broadly unchanged
loss charges as % of opening TBV and of non-flexed PPP Current account position 1
Local Currency, 2012E MUFG SMFG Mizuho Resona SMTH
Level of real interest rates 2
Forecast Average Loan Balances 76,000 56,850 59,000 27,000 20,200
Forecast Loan Loss Provisions 150 100 120 60 30
Total Danger Map Score (out of 45) 13 Low, due to no major credit growth
Severe Recession Loan Loss Provisions 760 569 590 270 202 Source: Deutsche Bank
2012 Pre-Provision Profit 1,040 760 685 280 245

2012 PPP less Recessionary Losses 280 192 95 10 43 Loan portfolio highlights
The shares of domestic lending by Japanese banks are 26% to large
2012 Forecast Tangible Equity 7,910 5,122 5,112 1,520 1,791
corporations, 43% to SMEs and 25% to individuals. Post-financial crisis, large
2 Years of Recessionary losses 1,520 1,137 1,180 540 404
companies have healthy balance sheets and there have been no large
2 Years of Recessionary losses as % 2012 -19% -22% -23% -36% -23% bankruptcies over the past several years. Most lending to individuals is for
Forecast Tangible Equity
housing loans, for which credit costs are low. Unsecured lending to individuals
2 years of Recessionary losses as % of 2 years 1% 0% -3% -8% -2%
of PPP (flexed down by 25%)
is made by non-banks rather than banks, and the market for such lending
Source: Deutsche Bank estimates shrank dramatically last year following the tightening of regulations on
maximum interest rates and the start of regulations on annual income.
Danger Map Indicators for Japan Accordingly, lending to SMEs presents the greatest risk, but the banks’ loan
portfolios are made up of small loans to numerous borrowers, and even during
As outlined above, credit risk is very low for Japanese banks. Overseas loans the financial crisis loans to SMEs were not a problem. We should also pay
average 11-16% for the top three bank groups (MUFG, SMFG, Mizuho FG) but attention to public guarantees on loans to SMEs. Based on macro-economic
close to half of these loans are to Japanese companies, while the remainder data, about 20% of lending to these borrowers is backed by public
are to major foreign companies or project finance for infrastructure. There is guarantees.
little direct sovereign lending and we do not believe overseas loans are likely
to become a problem. Caution is necessary regarding loans to TEPCO. If all of
the loans were written off, we calculate the impact on tangible equity for

Figure 75: Japanese Banks Domestic Loan Portfolio (Total)
Others,
7.1%
Large
Corporations,
26.2%
Individuals,
24.6%
SMEs,
43.1%
Note: Others include loans to public sector

Source: BoJ
Figure 76: MUFG (parent) Loan Portfolio
Overseas,
15.6%
Large
Corporations,
31.7%
Individuals,
23.8%
SMEs,
28.9%
Source: MUFG

Hong Kong
equity buffer -, and Hong Kong bank’s exposure to the high-end market,
which is the focus of bubble concern, is quite modest
As regards the Outlook and the Danger Map for Hong Kong, we score
Hong Kong overall at 21 out of 45. This captures a 1 to 5 score across a
Overview number of risk metrics, drawn from our analyses of case studies, and our
global experiences of bank lending. This puts Hong Kong in the middle of
In this section we look at the Hong Kong banks, with a focus on those
the pack on the Danger Map – neither in the danger of the red, nor sitting
covered by DB (Hang Seng Bank, Bank of East Asia, Bank of China (Hong
pretty in the green.
Kong), Wing Hang Bank and Dah Sing Banking Group). Over the last five
years, credit costs in Hong Kong have been relatively benign, with
average credit costs at just 10bps. Credit cost peaked in 2008 at 33bps Credit Data: Trends
but then quickly declined to 7bps in 2010. Across the cycle, average Below we summarise data from 2006 to 2010 for the aggregated Hong Kong
annual credit cost is about 25bps which is low relative to other regions. banks in our coverage universe. The data shows that despite downturn in
In the case of severe asset quality deterioration in Hong Kong, we believe 2008, Hong Kong banks’ profitability was robust enough to withstand
that peak provisions could reach 4x their normalised level using previous increases credit costs and key ratios point to a healthy sector overall. During
downturns (1998, 2003, and 2008) as a reference. We estimate combined the period, credit growth was strong at CAGR 15%, reflecting increased loans
credit losses of HKD 19bn per annum for the Hong Kong banks under our to China and rapid recovery in trade volume. Asset quality also remained
coverage. Over the two years, this would be equivalent to 13% of 2012 healthy. Credit cost was already low at 33bps in 2008 relative to other regions
closing tangible book value. If we include PPP estimates as an offset and this further declined to 7bps in 2010. As a result, the PPP/LLP cover
(flexed down by 25% which is consistent with the 2007-2009 experience) quadrupled from 10.1x in 2008 to 41.2x in 2010.
then there would be a gain of 17% of book value, i.e. the banks would be
able to offset the credit losses and still register a profit.
As an international financial centre, Hong Kong has a sizeable exposure to
non-HK loans and trade finance, which accounted for more than 28% total
system loans (Mar 2011). Exposure to the Hong Kong property sector is
sizable,accounting for 52% of domestic loans. As such, recent property
prices increases of 19% and 10% in 2010 and 2011 YTD have increased
concern over potential asset quality deterioration. Recent property price
increases have been driven by historically low mortgage rates (c.1.3%),
increased buying from mainland Chinese, and a shortage of supply. In an
effort to prevent a property bubble, HKMA has introduced various
prudential measures including a maximum LTV of 50% for property priced
above HK$10m (or US$ 1.3m). While a sharp correction in property prices
may well increase credit costs for the sector from a very low base, we
believe Hong Kong banks have robust profitability and sufficient capital to
weather any negative impact. In fact, we estimate the loan-to-value ratio
of system mortgage book to be roughly 40% - suggesting a comfortable

Figure 77: Aggregation of quoted bank data, local currency (HKD mn), in the 2007-2009 credit cycle, when PPP in aggregate fell from HKD 46bn to
Hong Kong HKD 34bn). These are simple numbers, but they do show in our view that (1)
Local Currency 2006 2007 2008 2009 2010
the Hong Kong banks have good pre-provision profit generation, with PPP
expected to reach close to HKD 60bn (by 2012), and that (2) even a severe
credit book downturn, would not have disastrous results on the banks’
Other Operating Income 15,885 22,761 14,244 18,947 21,132
profitability, in aggregate. Overall, the banks seem to be very well cushioned
Total Revenue 52,500 67,337 62,078 62,244 66,525
from incurring huge losses.
Costs 17,520 21,693 24,461 28,354 26,582
Figure 78: Summary of forecast data and comparing recessionary loan
Loan Loss Provisions -1,120 290 3,723 2,501 968
loss charges as % of opening TBV and of non-flexed PPP
Pre-Tax Profit 39,221 49,042 21,538 37,738 45,272
Local Currency, 2012E HSB BEA BOCHK DSB WHB
Total Assets 2,116,406 2,463,212 2,571,453 2,738,297 3,395,620
Forecast Loan Loss Provisions 1,162 964 1,790 236 146
Average Assets 1,980,434 2,289,809 2,517,332 2,654,875 3,066,959
Severe Recession Loan Loss Provisions 5,876 3,980 7,316 901 1,171
Risk weighted assets 1,096,912 1,319,306 1,269,181 1,316,069 1,524,286
2012 Pre-Provision Profit 19,941 7,685 27,453 1,566 2,560
Core tier one capital 137,860 130,014 124,719 147,387 162,756
2012 PPP less Recessionary Losses 14,064 3,705 20,137 666 1,390
Risk cushion above 7% plus 96,055 83,307 73,494 89,153 95,999
1 year PPP
Total Loans 902,787 1,076,523 1,163,205 1,246,874 1,580,419
2 Years of Recessionary losses as % 2012 -13% -17% -11% -12% -14%
Average Loans 868,294 989,655 1,119,864 1,205,040 1,413,646
Revenue / Average Loans 6.0% 6.8% 5.5% 5.2% 4.7%
2 years of Recessionary losses as % of 2 years of 21% 8% 20% 4% 9%
Pre-Provision Profit / 4.0% 4.6% 3.4% 2.8% 2.8% PPP (flexed down by 25%)
Average Loans Source: Deutsche Bank
Loan Loss Provisions / -0.13% 0.03% 0.33% 0.21% 0.07%

Average Loans
Danger Map Indicators for Hong Kong
PPP / Loan Loss Provision -31.2 157.5 10.1 13.6 41.2
Cover Below we summarise our Danger Map Indicators for Hong Kong. We score
Source: Deutsche Bank aggregation of company data
Hong Kong overall at 21 out of 45. This captures a 1 to 5 score across a
number of risk metrics, drawn from our analysis of case studies, and our
Estimates by Bank global experiences of bank lending. We rate Hong Kong as medium on the mix
and stage of cycle issues, and high in terms of % of debt in GDP. This puts
Below we summarise forecast data for 2012 for the aggregated Hong Kong
Hong Kong in the middle of the pack on the Danger Map – neither in the
banks in our coverage universe (Hang Seng Bank, Bank of East Asia, Bank of
danger of the red, nor in the comfort of the green.
China (Hong Kong Holdings), Wing Hang Bank and Dah Sing Banking Group).
We then show our estimate of severe recessionary loan losses, which for the
purposes of this exercise we have taken as 4x normalised.
We then express this as a % of 2012 BV, and also as a % of 2012 pre-

provision profit flexed down by 25% (this is the same PPP experience we saw
Figure 79: Scoring the Hong Kong “danger map” As of March 2011, domestic loans accounted for 71% of system loans.
Danger Factor Score Comments Exposures to property sector are quite high relative to other regions since
Deregulation of Lending 1 Recently introduced maximum LTV for mortgages
mortgage lending accounts for 25% and commercial property lending
(>HK$10m LTV capped at 50%) accounts for 27% of loans for use in Hong Kong. While compared to other
% of Credit to GDP 3 Debt to GDP ratio is high, due to HK being an regions the property exposure may seem high, we believe the risk for
international financial centre, not due to over- mortgage loans is quite low. Due to the recourse nature of mortgage loans in
leverage (Total Debt/GDP = 315%; Total Debt
excluding overseas loans/GDP = 247%)
Hong Kong, actual default is likely to be low, which was the case in 1997
when Hong Kong suffered a severe property price correction (property price
Change in % of Credit to GDP 3 Increased by 83 pp from 2007 to 2010 (seems high
but due to HK being an international financial fell by 52% to October 1998). (For more details, refer to the section on Hong
centre) Kong’s credit crisis case study).
Commercial and industrial loans account for 29% of domestic loans. This
Credit Mix 3 55% of loans for use in HK are exposed to the
residential and commercial property market mainly consists of manufacturing, transport, wholesale & retail trade and loans
Unemployment 1 Unemployment at low level of 3.5% (Apr 2011) and
to financial concerns & brokers. Other commercial loans account for 11%
declining YTD which mainly is for utilities & telecommunications, hotel & catering and loans
Current account position 1 No current account deficit. 2010 current account / to small & medium enterprises. Credit card lending accounted for 2.4% of
GDP = 6.6% loans for use in HK.
Level of real interest rates 1 HK's real interest rate is -3.2% due to high inflation
rate but monetary policy having to follow the US Figure 80: Breakdown of gross loans by geographic usage (Mar 2011)
Exchange rate flexibility 5 HK Dollar exchange rate is pegged to US Dollar
Loans
Loans in outside HK
HK 21%
Understanding the Hong Kong banks’ loan books 71%

Trade
finance
Since Hong Kong is an international financial centre, the system has a sizable Other 7%
exposure to non-Hong Kong loans and trade finance. Loans for use outside Offshore
1%
Hong Kong, which are mostly loans to mainland China, accounted for 21% of Source: Deutsche Bank, HKMA
total loans as of March 2011. When China tightens its credit, Hong Kong’s
overseas loan growth typically benefits and vice versa. For example, in 2010, Figure 81: Breakdown of loan for use in HK by major economic sector
loans for use outside Hong Kong increased by 48% as credit demand in (Mar 2011)
mainland China was channeled to Hong Kong banks. Recovery in Hong Kong Other Major com/ind
individual
export/import volumes as well as the launch of RMB trade settlement in July Other 8% Major
2009 helped to drive trade finance growth. While in 2009 trade finance com/ind B&C
11% com/ind
declined by a 6%, it grew by 57% in 2010. We expect trade finance to 29%
Housing
continue to grow as Hong Kong evolves into the trade settlement centre of
China. Housing
25% Other com/ind
B&C
Other individual
27%
Source: Deutsche Bank, HKMA

UK
LBG is also exposed to a risk of required re-provisioning for existing
problem loans.
As regards the Danger Map and Outlook, we score the UK at 23 out of
45. This captures a 1 to 5 score across a number of risk metrics, drawn
Overview from our analyses of case studies, and our global experiences of bank
lending. We see the UK as vulnerable from a credit/GDP, credit mix, and
This section looks at the UK banks, focussing on those covered by DB
macroeconomic growth and stability perspective. On the Deutsche Bank
(Barclays, Lloyds Banking Group, HSBC, RBS and Standard Chartered).
base case of modest but steady UK economic growth, and expected
The UK banking sector, both listed and unlisted, suffered significantly
resolution of EU sovereign issues in Italy and Spain, we expect loan loss
during the credit crisis, as a combination of rapid credit growth and
normalization to continue, leaving the stocks attractively valued.
significant reliance on short term wholesale funding markets drove
liquidity failures and falling asset prices which later lead to significant
capital destruction. A difficult crisis
Northern Rock, the first UK bank to fail, ran into difficulties in July 2007, The UK banks fared poorly during the crisis, with rapid growth in loans backed
followed by the sale or closure of Bradford & Bingley, Alliance & by higher risk commercial and residential real estate, significant inventories of
Leicester, HBOS and a number of building societies. The remaining five US market-risk assets (especially sub-prime, Alt-A and CDOs held in conduits
listed banks have increased core tier 1 capital from £113bn in 2006 to the and treasury portfolios), and dependence on frequent refinancing of wholesale
£233bn we forecast by end 2011, driving the sector average CT1 ratio to funding combining to fell Northern Rock, Bradford & Bingley, Alliance &
10.6% under Basel 2.5 from a little over 6% in 2006. Leicester, HBOS and RBS. The surviving UK banks have all raised significant
amounts of capital to offset losses, facilitate debt market access, and meet
Stressing the UK banks for a return to recession-level loan losses, we see rising regulatory requirements. Figure 82 shows that the five listed banks have
a fairly resilient picture. Two years of recession losses cost the sector increased core tier 1 capital by £120bn or 106% since end 2006, which we
10% of tangible equity including stressed pre-provision profit, and 33% of expect to result in FY2011 core tier 1 ratios of over 10.5%, including the Basel
TNAV excluding any PPP. Best positioned are Barclays and StanChart 2.5 requirements.
which we would expect to remain profitable. LBG looks worst, driven by
the high level of assumed recession losses; with £65bn in NPLs at 1H11,
Figure 82: UK listed bank capital levels and ratios before and after the crisis (£’m)
Core tier 1 capital Risk-weighted assets Core tier 1 ratio
2006 2010 2011 2012 2011/2006 2006 2010 2011 2012 2011/2006 2006 2010 2011
Barclays 16,776 42,861 44,947 49,357 168% 297,828 398,031 437,365 458,090 47% 5.6% 10.8% 10.3%
LBG 26,664 41,371 38,586 41,271 45% 432,012 406,372 382,373 361,961 -11% 6.2% 10.2% 10.1%
HSBC 42,986 70,802 81,618 91,074 90% 572,365 672,630 706,261 741,574 23% 7.5% 10.5% 11.6%
RBS 20,281 49,309 47,796 52,795 136% 400,257 462,600 507,093 484,770 27% 5.1% 10.7% 9.4%
StanChart 6,446 17,646 19,912 21,911 209% 93,563 149,437 164,518 180,679 76% 6.9% 11.8% 12.1%
Total/Ave 113,153 221,990 232,859 256,407 106% 1,796,025 2,089,070 2,107,609 2,227,074 22% 6.3% 10.6% 10.6%
Source: Deutsche Bank estimates, Company data

All of the banks other than StanChart have also embarked on significant plans Though loan losses are normalising rapidly following a crisis credit cycle which
to reduce the size of their balance sheets in order to improve loan/deposit was materially worse than that of the early 1990’s, loan portfolios remain
ratios, business mix, and overall profitability. At the end of 1H11, LBG and RBS fragile. Consumers have suffered about a c.7% fall in real income in 09/10,
had £125bn each in loans which are considered non-core and in the process of with disposable income further pressured by material increases in taxes (VAT),
being run-off or sold. and gas, electricity, petrol and food prices. LBG estimate that c.2% of their
mortgage borrowers are on government support to meet mortgage interest
Figure 83: Loan losses/average loans – this cycle payments despite record low interest rates.
4.0%
3.5% Figure 85: Lloyds Banking Group impaired loans (£’m)

3.0% 2H08 1H09 2H09 1H10 2H10 1H11 % of HoH % YoY %
loan bk
LLPs / Ave loans
2.5%
Retail 10,106 11,394 11,015 10,464 9,750 9,390 2.6% -4% -10%
2.0%
1.5% o/w Secured 4,756 7,612 7,196 6,861 6,769 6,695 2.0% -1% -2%
1.0% o/w Unsecured 5,350 3,782 3,819 3,603 2,981 2,695 10.4% -10% -25%
0.5% Wholesale 18,470 31,725 35,114 36,779 32,835 30,630 18.7% -7% -17%
0.0% o/w Corporate n/a n/a 9,362 7,906 6,635 5,750 7.8% -13% -27%
2011e
2012e
2013e
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
o/w Commercial n/a n/a 2,695 2,652 2,856 2,993 10.1% 5% 13%
Lloyds Banking Group Barclays RBS HSBC StanChart o/w Corporate Real n/a n/a 16,505 19,624 17,518 16,212 -7% -17%
68.5%
Estate BSU
Source: Company data, Deutsche Bank estimates
Wealth & International 2,728 5,900 12,704 15,632 20,342 23,836 37.2% 17% 52%
Figure 84: Average write-off rates – early 1990’s o/w Int Ireland 1,775 3,995 9,712 11,689 14,445 17,672 64.1% 22% 51%
3.00% Total 32,701 50,921 60,802 64,753 64,606 65,468 10.9% 1% 1%
Source: Company data
2.50%
2.00% Corporate books remain at risk from a double dip in commercial real estate in
Net write-offs
1.50%
particular, as shown by the magnitude of the current CRE NPL portfolios, at
£16bn at LBG, for example (Figure 85).
1.00%
0.50%
Credit Data: Trends
0.00%
We summarise data for the UK banks in our coverage universe in Figure 86.
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Outside of crisis conditions, the (capital-markets-assisted) strong pre-provision

Barclays Lloyds Banking Group RBS Total
profit generation of the banks is evident at 2.4% of loans, comfortably
Source: Company data
sufficient to absorb the 70-80bps of normalised loan losses we expect. We
forecast that the banks will comfortably comply with the Basel 3 capital
requirements, allowing for phasing of the capital definition.

Figure 86: Aggregation of quoted bank data, local currency (£’m), UK shows our standard screen for the proportion of tangible equity placed at risk
2006 2007 2008 2009 2010 2011E 2012E by two years of recession-level loan losses, including and excluding forecast
Net Interest Income 45,664 50,427 75,154 67,414 70,884 68,487 70,316
pre-provision profit.
Other Operating Income 47,980 55,969 55,259 62,034 65,864 64,696 69,230
Total Revenue 93,644 106,396 130,413 129,447 136,748 133,184 139,546
Figure 87: How resilient are the capital bases of the UK banks in a credit
Costs 49,352 54,685 73,822 73,043 77,642 77,812 78,301
downturn
Local Currency, 2012E Barclays RBS.L Lloyds HSBC (£) STAN (£) Total HSBC ($) STAN
Pre-Provision Profits 44,292 51,711 56,591 56,404 59,106 55,372 61,245 ($)
Loan Loss Provisions 11,303 15,672 46,321 63,662 37,688 27,898 19,934 Forecast Average Loan 524,382 526,641 581,653 699,681 183,742 2,516,099 1,111,654 291,929
Balances
Pre-Tax Profit 33,459 36,836 -6,624 -2,662 21,052 26,641 43,484
Forecast Loan Loss 3,731 4,539 5,127 5,828 913 20,138 9,260 1,451
Total Assets 3,293,594 4,829,297 7,441,710 5,663,256 5,832,557 6,179,688 6,204,263 Provisions
Average Assets 3,090,166 4,052,315 6,772,766 6,439,494 5,775,161 5,963,962 6,191,976 Severe Recession Loan 6,401 9,256 10,063 16,672 2,283 60,414 26,488 3,627
Loss Provisions
Risk weighted assets 1,412,101 1,794,983 2,439,385 2,148,350 2,128,109 2,198,433 2,227,946
2012 Pre-Provision Profit 12,586 11,690 10,400 21,696 5,703 62,075 34,471 9,061
Basel 3 additions (if 0 0 0 0 91,000 41,111 41,111
2012 PPP less 6,185 2,434 337 5,024 3,420 17,401 7,983 5,434
known)
Recessionary Losses
Core tier one capital 74,615 91,617 104,793 168,361 172,977 176,452 189,423
Risk Cushion > 7% plus 25,650 28,770 21,527 31,612 34,317 31,124 34,676 2012 Forecast Tangible 51,521 60,790 42,313 91,157 23,159 268,939 144,831 36,794
1 year PPP Equity
Risk cushion to average 1.86% 1.69% 0.81% 1.28% 1.41% 1.27% 1.39% 2 Years of Recessionary 12,802 18,512 20,126 33,343 4,566 89,349 52,976 7,254
loans losses
Total Loans 1,459,265 1,959,020 2,659,526 2,367,061 2,466,079 2,483,136 2,495,521 2 Years of Recessionary -25% -30% -48% -37% -20% -33% -37% -20%
losses as % 2012
Average Loans 1,379,005 1,704,810 2,651,303 2,470,888 2,427,262 2,457,501 2,489,328 Forecast Tangible Equity
Revenue / Average 6.8% 6.2% 4.9% 5.2% 5.6% 5.4% 5.6% 2 years of Recessionary 12% -2% -11% -1% 17% -10% -1% 17%
Loans losses as % of 2 years of
PPP (flexed down by
Pre-Provision Profit / 3.2% 3.0% 2.1% 2.3% 2.4% 2.3% 2.5%
25%)
Average Loans
Source: Deutsche Bank estimates, Company data; Note: Barclays, RBS, LBG, HSBC severe recession losses use repeat of 2009/2010 loan
Loan Loss Provisions / 0.82% 0.92% 1.75% 2.58% 1.55% 1.14% 0.80% losses seen by each bank, including the benefit of fair value reversals in the case of LBG. For StanChart, which hasn’t really seen a credit
cycle in the downturn, we use loan losses of 2.5x our forecast.
Ave Loans
PPP / Loan Loss 3.9 3.3 1.2 0.9 1.6 2.0 3.1
Provision Cover We have used 09/10 actual loan losses for all banks other than StanChart
Source: Deutsche Bank estimates, Company data (which has not had a real credit cycle in years, for which we use 2.5x our 2012
forecast impairment rate) as our recession charge. We haircut 2012 forecast
Estimates by Bank PPP by 25% to reflect likely weak revenue conditions (transaction volumes
and lower investment banking revenues) in a downturn.
High cumulative losses already booked by the UK banks in this cycle provides
some protection to shareholders in the event of a double dip. But the losses In aggregate, we believe these results show a resilient picture for the sector
shown so far also clearly demonstrate how vulnerable certain loan books are. as a whole, with 2 years of recession losses costing the sector 10% of
LBG, for example, has already booked loan losses of over 8% of group loans, tangible equity including PPP, and 33% of TNAV excluding any pre-provision
compared with three year peak losses in the early 1990’s recession of 6% at profit. Best positioned on this basis are Barclays and StanChart, both of which
Barclays (91-93), 5% at LBG (92-94), and 5.2% for RBS (91-93). Figure 87 we would expect to remain profitable. LBG looks worst, driven by the high
level of recession losses we assume, of 173bps of loans, as compared with

normalised LLPs of 55bps. With £65bn in NPLs at 1H11, LBG is also exposed
to a risk of required re-provisioning for existing problem loans, if collateral
values fall.
Danger Map Indicators for UK

Below we summarise our Danger Map for the UK. We score the UK overall at
23 out of 45. This captures a 1 to 5 score across a number of risk metrics,
drawn from our analyses of case studies, and our global experiences of bank
lending. We rate the UK as vulnerable from a credit/GDP, credit mix, and
macroeconomic growth and stability perspective. On the Deutsche Bank base
case of modest but steady economic growth in the UK, and resolution of EU
sovereign issues in Italy and Spain, we expect the normalisation of loan losses
for the UK banks to continue, leaving the stocks apparently attractively valued.
Figure 88: UK danger map

Danger Factor Score Comments
Deregulation of Lending 2 Credit conditions have tightened significantly since the
crisis
% of Credit to GDP 5 Extremely high, exacerbated by London's international
financial centre status
Change in % of Credit to GDP 3 Significant deleveraging and decent nominal GDP
growth seeing some reduction in debt/GDP
Maturity of Cycle in Years 3 Tepid growth, low consumer confidence, negative real
wages, significant government austerity measures
Credit Mix 3 Substantial commercial real estate exposure has lead to
significant losses; remains an area of some vulnerability
Unemployment 3 Stabilised
Level of real interest rates 1 Bank base rate 0.5%, retail price inflation > 5%.
Total Danger Map Score (out 23
of 45)

Nordics
Figure 89: Aggregation of quoted bank data, Nordics (EURbn)
Currency, EURbn 2006 2007 2008 2009 2010 2011E 2012E
Total Revenue 26,591 28,880 29,413 33,256 30,943 31,834 33,980
Costs 13,898 15,041 16,356 17,080 16,638 17,195 17,892
Overview Pre‐Provision Profits 12,693 13,839 13,057 16,176 14,305 14,639 16,088
Loan Loss Provisions ‐325 233 3,336 10,236 3,821 1,992 1,565
In this section we look at the Nordic banks, with a focus on those covered Pre‐Tax Profit 10,181 10,816 7,505 3,562 7,672 9,337 10,793
by Deutsche Bank (Danske Bank, DnB NOR, Nordea, Sv. Handelsbanken, Total Assets
Average Assets 1,366,796 1,545,481 1,750,893 1,831,493 1,846,135 1,859,870 1,855,418
SEB and Swedbank). Over the last five years, we find that loan loss Risk weighted assets 693,640 793,096 806,912 701,005 720,684 744,836 769,926
provisions in the banking book averaged 32bps, and peaked at 100bps Core tier one capital 44,955 17,306 34,952 47,573 52,766 60,181 65,919
Basel 3 deductions (if known)
(2009). The minimum level was reached in 2006, when banks had write- Risk Cushion above 7% plus 1 year PPP 12,693 13,839 13,057 16,176 14,305 14,639 16,088
backs at 3bp of lending. Risk cusion to average loans
Total Loans 840,235 966,506 1,074,674 1,057,432 1,087,048 1,112,243 1,149,034
Average Loans 794,647 903,371 1,020,590 1,066,053 1,072,240 1,099,645 1,130,638
We believe that peak provisions on a further downturn across the Euro Revenue / Average Loans 3.3% 3.2% 2.9% 3.1% 2.9% 2.9% 3.0%
zone would likely reach 4x their normalized level. We estimate credit Pre‐Provision Profit / Average Loans
Loan Loss Provisions / Average Loans
1.6%
‐0.04%
1.5%
0.03%
1.3%
0.33%
1.5%
0.96%
1.3%
0.36%
1.3%
0.18%
1.4%
0.14%
losses of Euro 29bn for the Nordic banks per annum, in the severe PPP / Loan Loss Provision Cover n.m. n.m. 3.9 1.6 3.7 7.3 10.3
scenario. Over two years, this would be equivalent to 68% of 2012
closing tangible book value (TBV). If we include pre-provision profit (PPP)
estimates as an offset (flexed down consistently with the crisis Estimates by Bank
experience), then the loss would be 30% of TBV, i.e. the banks would be Below we summarize forecast data for 2012 for the aggregated Nordic banks
loss-making. in our coverage universe (Danske Bank, DnB NOR, Nordea, Sv.
Despite this severe scenario, we estimate that the sector CT1 would be Handelsbanken, SEB and Swedbank). We then show our estimate of severe
8.0% post the losses. This is well above any regulatory requirement. recessionary loan losses, which for the purpose of this exercise we have
Despite this, if a banking system lose 30% of its equity, it would likely taken as 4x normalized. The normalized LLPs are defined as the 25 year Nordic
trigger capital raisings anyway to keep access to debt capital markets average, which is 65bp.
open. Generally, the banks are domestically focused, and 55-85% of
banks’ lending is concentrated to the Nordics. Compared to EU, the We then express this as a percentage of 2012 book value (BV), and also as a
Nordic banks have a higher proportion of low risk mortgages, percentage of 2012 PPP flexed down by 25%, as a proxy of a potential
representing close to 40% of lending. contraction during a crisis. We estimate credit losses of Euro 29bn for the
Nordic banks per annum. Over two years, this would be equivalent to 68% of
As regards the Danger Map, we score Nordics overall at 16 out of 45. 2012 closing TBV. If we include PPP estimates as an offset, then the loss
This captures a 1 to 5 score across a number of risk metrics. This puts would be 30% of TBV, i.e. the banks would be loss-making.
Nordics at the lower end of the pack on the Danger Map.
Taking a step back, this simple exercise, show in our view that the Nordic
banks’ PPP profit level would not be enough to keep banks in profit if the
Credit Data: Trends region suffered a macro shock and deep credit crisis, similar to the crisis in the
Below we summarize data from 2006 to 2010 for the aggregated Nordic early 1990s.
banks in our coverage universe. We can see clearly that even though the
credit cycle was not that severe, rising loan losses, cut the PPP/LLP ration in
half, from 3.9x to 1.6x.
Figure 90: Summary of forecast data and comparing recessionary loan loss charges as % of opening TBV and of PPP
EURbn, 2012E Danske B. DnB NOR Nordea SHB SEB Swedbank Total
Forecast Average Loan Balances 226,866 146,523 332,313 171,827 120,471 132,638 1,130,638
Forecast Loan Loss Provisions 640 229 521 111 60 3 1,565
Severe Recession Loan Loss Provisions 5,536 3,868 8,507 4,536 3,180 3,502 29,129
2012 Pre‐Provision Profit 3,066 2,399 5,122 1,812 1,807 1,883 16,088
2012 PPP less Recessionary Losses ‐2,470 ‐1,470 ‐3,385 ‐2,724 ‐1,374 ‐1,619 ‐13,041
2012 Forecast Tangible Equity 16,572 13,949 25,157 9,998 10,761 9,092 85,530

2 Years of Recessionary losses 11,071 7,736 17,014 9,072 6,361 7,003 58,259
2 Years of Recessionary losses as % 2012 Forecast Tangible Equity ‐67% ‐55% ‐68% ‐91% ‐59% ‐77% ‐68%
2 years of Recessionary losses as % of 2 years of PPP (flexed down by 25%) ‐39% ‐30% ‐37% ‐64% ‐34% ‐46% ‐40%
Danger Map Indicators for Nordics Understanding the Nordic banks’ loan books
Below we summarize our Danger Map Indicators for Nordic region. We score The Nordic banks’ lending mix is more geared towards low risk mortgages,
Nordics overall at 16 out of 45. This captures a 1 to 5 score across a number compared to rest of Europe. In total, retail lending is 43% of credit exposure,
of risk metrics, drawn from our analyses of case studies, and our global of which 92% is mortgage lending and rest unsecured. Large corporate
experiences of bank lending. This puts Nordics at the lower end of the pack on represent 37% of the average loan books of the Nordic banks under coverage.
the Danger Map. The risk factors in Nordics is the above average This highlights, the Nordic industry structure, with a large proportion of export
indebtedness and high proportion of interest only loans. The mitigating factor driven blue chip companies. The experience in the last two downturns, has
is the regions strong social security systems, the healthy fiscal situation, a been that the SME segment, rather than large corporates, have been the loss
recourse structure in banks’ lending, the current account position, level of real drivers.
interest rates and the exchange rate flexibility.
On the positive, the Nordic banks: 1) do not have large financing businesses
Figure 91: Scoring the Nordic “danger map” within their investment banks, and 2) have a high proportion of domestic retail
Danger Factor Score Comment exposures. As a consequence, the average Nordic loan losses over time, and
peak losses, are below EU averages. Instead, the main loss driver in the
Nordic bank sector, has been the banks’ international operations, primarily in
% of Credit to GDP 4 Higher than the European average
the Baltics and Ireland.
Change in % of Credit to GDP 4 Above the European average
Maturity of Cycle in Years 1 Year 1 of the recovery in the Nordic and Baltic regions
In Figure 93 to Figure 94, we show the lending split for the Nordic bank sector
Credit Mix 2 Traditional model, but high proportion of mortgage
(secured) lending
and for each bank. In General, banks are domestically focused and Nordics
represent between 55% and 85% of total lending.
Unemployment 1 Unemployment falling or stable across Nordics

Figure 92: Nordic banks under coverage: average loan mix (Q4-10) Figure 94: SEB loans (Q4-10)
Other, 4%
RoW, 9%
Germany,
Large
22%
corporates, Retail, 43%
37%
Sweden,
Russia, 52%
1%
Baltics,
12%
Denmark, Norway, Finland,
SME, 16% 1% 2% 1%
Source: Deutsche Bank, companies data Source: Deutsche Bank, company data
Figure 93: Swedbank loans (Q4-10) Figure 95: Nordea loans (Q4-10)
Russia, Ukraine, RoW, Poland,

U.S., 1% Russia,
1% 1% 11% 2%
Baltics, RoW, 1% 1%
14% Baltics, Sweden,
3% 25%
Norway,
1%
Denmark,
1%
Finland,
0% Denmark,
28%
Sweden, Finland,
Norway,
80% 18%
16%
Source: Deutsche Bank, company data Source: Deutsche Bank

Figure 96: Sv. Handelsbanken loans (Q4-10) Figure 98: DnB NOR loans (Q4-10)
Poland, RoW, 6% Poland, RoW, 6%

Baltics, Baltics, Sweden,
4% 1% 5%
0% 5%
Finland,
Denmark, Denmark, 0%
4% 1%
Norway,
13%
Finland,
6%
Sweden,
67%
Norway,
82%
Figure 97: Danske Bank loans (Q4-10)
Poland, RoW, 2%
Baltics, 8% Sweden,
2% 10%
Finland,
9%
Norway,
9%
Denmark,
60%

France
Figure 99: Aggregation of quoted bank data, local currency (Euro bn),
France
Local Currency 2006 2007 2008 2009 2010

Total Revenue 66,547 64,817 65,198 79,863 90,427
In this section we look at the French banks, with a focus on those Costs 41,123 45,783 46,563 51,288 56,249
covered by DB (BNP Paribas, Société Générale and Credit Agricole). Over Pre-Provision Profits 25,424 19,034 18,635 28,575 34,178
the last five years, we find that credit loan loss provisions in the banking Loan Loss Provisions 2,074 4,523 11,571 18,906 12,739
book averaged 82bp, and peaked at 146bp. Pre-Tax Profit 25,513 16,335 8,626 11,052 22,082
We believe that peak provisions on a further downturn across the euro

Total Assets 3,658,480 4,180,439 4,858,774 4,638,741 4,723,759
zone would likely reach 4x their normalised level, given that a slowdown
Average Assets 3,412,506 3,919,460 4,519,607 4,748,758 4,681,250
would not see government (fiscal) intervention, as firepower is exhausted.
Total Loans 904,825 1,052,720 1,198,051 1,385,657 1,439,830
We estimate credit losses of Euro 32bn for the three French majors per
Average Loans 810,401 978,773 1,125,386 1,291,854 1,412,744
annum. Over two years, this would be equivalent to 46% of 2012 closing
tangible book value. If we include PPP estimates as an offset (flexed
Pre-Provision Profit / Average Loans 3.1% 1.9% 1.7% 2.2% 2.4%
down consistent with the 2007-2009 experience) then the loss would be
Loan Loss Provisions / Average Loans 0.3% 0.5% 1.0% 1.5% 0.9%
4% of TBV, i.e. the banks would be slightly loss-making.
These data are somewhat less severe (loss-making) than other countries Source: Deutsche Bank aggregation of company data
in Europe and elsewhere. The French banks’ portfolios still include large
domestic books (where usury laws have prevented the banks moving up Estimates by Bank
the risk curve). But we see more risk in the French banks’ financing
businesses, which survived the downturn surprisingly well. A more Below we summarise forecast data for 2012 for the aggregated French banks
corporate driven downturn would represent the key risk this time out, in in our coverage universe (BNP Paribas, Société Générale and Credit Agricole).
our opinion. We then show our estimate of severe recessionary loan losses, which for the
purposes of this exercise we have taken as 4x normalised.
As regards the Outlook and the Danger Map for France, we score France
overall at 24 out of 45. This captures a 1 to 5 score across a number of We then express this as a % of 2012 BV, and also as a % of 2012 pre-
risk metrics, drawn from our analyses of case studies, and our global provision profit flexed down by 25% (this is the same PPP experience we saw
experiences of bank lending. This puts France at the higher end of the in the 2008 credit cycle, when PPP in aggregate fell from over Euro 25bn to
pack on the Danger Map. under Euro 19bn). These are simple numbers, but they do show in our view
that (1) the French banks have good pre-provision profit generation, but that
Credit Data: Trends (2) in a severe credit book downturn, we would expect a breakeven or worse
result, in aggregate.
Below we summarise data from 2005 to 2010 for the aggregated French
credit cycle was not that severe, declining PPP combines with rising loan
losses to reduce PPP/LLP cover from 12.3x to 1.5x.

Figure 100: Summary of forecast data and comparing recessionary loan Understanding the French banks’ loan books
loss charges as % of opening TBV and of non-flexed PPP As discussed above, we regard French banks as not too high risk in their loan
Local Currency, 2012E CredAg BNPP SocGen portfolios, but with still some danger areas. One of the key points regarding
Forecast Average Loan Balances 412,469 729,875 388,801 the French banks is that their books are in fact very diverse, and not
Forecast Loan Loss Provisions 1,872 3,162 2,346 particularly “French” in nature, as shown below.
Severe Recession Loan Loss Provisions 6,558 10,510 5,599
2012 Pre-Provision Profit 8,554 20,034 11,022
Figure 102: BNP Paribas loan out-standings as at Q1 2011: total of Euro
2012 PPP less Recessionary Losses 1,996 9,524 5,424 685bn
Other BancWest
11% 6%
2012 Forecast Tangible Equity 33,947 66,018 41,054
2 Years of Recessionary losses 13,116 21,020 11,197
CorpFin &
2 Years of Recessionary losses as % 2012 Forecast Tangible Equity 39% 32% 27% Eqp
2 years of Recessionary losses as % 2 years of PPP flexed down by 25% 136% 103% 100% BeLux & 27%
IT
24%
Danger Map Indicators for France Fr M t ge

Cons
9%
Below we summarise our Danger Map Indicators for France. We score France Credit Fr SM E
9%
14%
overall at 24 out of 45. This captures a 1 to 5 score across a number of risk
metrics, drawn from our analyses of case studies, and our global experiences
The low risk segments for the banks are, in our view, the domestic French
of bank lending. We rate France as low on the mix and stage of cycle issues,
retail exposures. These are restricted under French usury laws; it is illegal for
medium in terms of % of debt in GDP and high on exchange rate flexibility and French banks to charge above certain interest rates1, and this has been an
unemployment. This puts France at the higher end of the pack on the Danger effective constraint against banks moving up the risk curve. As a
Map. consequence, we quite happily use rather low estimates of through-cycle loan
losses for residential mortgages and domestic consumer credit.
Figure 101: Scoring the French “danger map” French banks also, however, have large financing businesses within their
Danger Factor Score Comment investment banks, and large corporate and investment banking loan books.
Deregulation of Lending 2 No recent deregulation of lending Although investment grade corporate credit has been remarkably robust over
% of Credit to GDP 3 Close to the European average the last few years, corporate leverage has not reduced, and consequently is as
Change in % of Credit to GDP 1 Close to the European average
likely as any other area to be the “next big risk”. High-yield corporates (leaving
aside leveraged finance / financial sponsor businesses) have also had easy
Maturity of Cycle in Years 3 Year 3 of the recovery in the euro zone.)
access to credit. And we also believe that the French banks have continued to
Credit Mix 2 Heavily mortgage led and ursury influenced
grow specialized lending areas like aircraft leasing or commodities financing.
credit book
Unemployment 4
1
Level of real interest rates 2 The Usury Rate is published quarterly by the Banque de France, as governed by Article 313-3 of the
Exchange rate flexibility 5 French Consumer Code. Usury was originally prohibited in France in January 1886, with the law on
usury later amended in 1966, and and finally reaching its present form in the French Consumer Code.
Total Danger Map Score (out of 45) 24 It covers all loans up to a certain size, beyond which in theory loans would be decided as usurious or
Source: Deutsche Bank not by the Banque de France on a case-by-case basis, although we believe that in practise usury is
not applied at all to large corporate lending.
Germany
about 150% of the 2010 charge and in line with the dynamics of EBA
stress test results for the largest German banks.
Benchmarking these potential losses with the average pre-provision
profits of the last decade of EUR41bn we see limited risk to capital for the
Overview
sector as a whole. However, we point out that we expect individual
In this section we look at the German banks; given the majority of large institutions to report losses; we would refer to historic financial
German banks is not listed we have concentrated on sector data that performance as a reasonable guidance to identify these banks. More
include all German institutions based on HGB accounting rules. Over the generally, we view the German banking sector to be short of capital given
last decade, we find that loan loss provisions peaked at 114bps in 2008; high investments in non-core credit risk, such as debt securities and
another high was reported in the 2002/3 recession with 102bps. interbank lending.
We believe that peak-to-trough cycles for credit risk are somewhat As regards the Danger Map for Germany we result an overall score of 20.
smoothed by HGB accounting rules and the general focus on non-equity This captures a 1 to 5 score across a number of risk metrics, drawn from
investors of the majority of banks in Germany. As evidence we point our analysis of case studies, and our global experience of bank lending.
towards an increase of ‘only’ 2.5x from the lowest reported LLP rate of This puts German within the better half of countries.
46bps in 2005/06 and the above mentioned 115bps peak for 2008. We
doubt that loan provisions exceed EUR40bn in an economic downturn,
Figure 103: Aggregation of German banks (EUR bn)

Local Currency 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Average
Net Interest Income 79.2 85.6 81.7 85.0 88.2 89.1 91.6 90.6 91.5 91.2 87.4
Other Operating Income 34.3 31.0 34.6 30.7 41.1 41.6 34.0 16.6 34.8 34.5 33.3
Total Revenue 113.5 116.5 116.3 115.7 129.3 130.7 125.6 107.2 126.3 125.7 120.7
Costs -81.0 -78.3 -77.3 -75.8 -78.8 -81.5 -81.6 -78.7 -82.2 -83.6 -79.9
Pre-Provision Profits 32.4 38.2 39.0 39.9 50.5 49.2 44.1 28.5 44.1 42.1 40.8
Loan Loss Provisions -19.6 -31.2 -21.8 -17.3 -14.0 -14.0 -23.6 -36.6 -27.0 -26.1 -23.1
Pre-Tax Profit 14.1 10.9 1.8 10.4 33.2 27.6 20.5 -25.0 -2.9 11.6 10.2
Total Assets 6,386 6,452 6,471 6,664 6,903 7,188 7,626 7,956 7,510 8,352 7,151
Average Assets 6,267 6,419 6,462 6,567 6,783 7,045 7,407 7,791 7,733 7,931 7,041
Capital 276 292 294 288 303 337 352 378 379 381 328
Total Loans 3,065 3,037 3,037 3,024 3,035 3,062 3,171 3,245 3,176 3,271 3,112
Average Loans 3,065 3,051 3,037 3,030 3,029 3,048 3,117 3,208 3,211 3,223 3,102
Revenue / Average Loans (%) 3.70 3.82 3.83 3.82 4.27 4.29 4.03 3.34 3.93 3.90 3.89
Pre-Provision Profit / Average Loans (%) 1.06 1.25 1.28 1.32 1.67 1.61 1.41 0.89 1.37 1.30 1.32
Loan Loss Provisions / Average Loans (%) 0.64 1.02 0.72 0.57 0.46 0.46 0.76 1.14 0.84 0.81 0.75
PPP / Loan Loss Provision Cover (x) 1.65 1.22 1.79 2.31 3.60 3.51 1.87 0.78 1.63 1.61 1.76
Source: Deutsche Bank, Deutsche Bundesbank

Danger Map Indicators for Germany Figure 105: Important balance sheet positions
Below we summarize our Danger Map Indicators for Germany. We score Debt securities
20%
Germany overall at 20 out of 45. This captures a 1 to 5 score across a number Interbank loans
31%
of risk metrics, drawn from our analyses of case studies, and our global
experiences of bank lending. This puts German within the better half of
countries.
Figure 104: Scoring the German "heat map"

Score Comment
Deregulation of Lending 1 No recent deregulation of lending Customer loans
49%
% of Credit to GDP 2 Better than European average
Change in % of Credit to GDP 2 Lower than European average
Maturity of Cycle in Years 2 Close to European average
Credit Mix 3 Traditional model, but high non-core credit share
Figure 106: Break-down of customer loans
Unemployment 2 Moderate improvement, good level
Other customer
Level of real interest rates 2 loans
29%
Exchange rate flexibility 5 Domestic
corporate loans
Total Danger Map Score (out of 45) 20 40%
Understanding German banks’ loan books

The 3 most important important balance sheet positions for German banks are
customer loans, followed by interbank assets and debt securities. There Domestic retail
loans
account for 80% of all assets. However, while customer loans are the biggest 31%
position they only make up for less than 50% of important balance sheet
positions. In our view, German banks’ exposure to debt securities and
interbank activities is inflated and needs to be consolidated in the new Customer loans break down in domestic corporate loans with 40%, demestic
regulatory environment. From a risk-reward perspective this should not be retail loans with 31% and other customer loans with 29%. We view this as
negative, in our opinion, as we perceive these activities to be unattractive. quite balance; from a risk perspective we view domestic retail lending as good
quality and corporate lending as average quality. Overall, Germany is a rather
mature market with limited lending growth opportunities.

Greece
11% in Q2 2011 from 2% in 2006. In the same period bad debt charges
(loan loss provisions/loans) increased almost fivefold to 224bps. Neither
NPLs nor cost of risk have peaked yet.
Capital-wise the Greek banks entered the crisis with a CT1 ratio of 8%.
Overview Capital support from the state in 2008 added another 140bps while equity
raisings contributed another 180bps. Nevertheless, these measures have
In this section we look at the Greek banks, with a focus on those covered
been eaten away by the recession and the potential PSI (bond swap of
by Deutsche Bank (Alpha Bank, Piraeus bank, Eurobank and ATE bank).
GGBs) leaving CT1 ratio unchanged at 8% as of Q2 2011.
We also present historical data for National Bank of Greece but refrain
from showing our estimates as we are restricted on the name. In the last As regards the Danger Map, we score Greece overall at 33 out of 45. This
two years Greek banks have been caught in the middle of the Greek captures a 1 to 5 score across a number of risk metrics. We rate Greece
sovereign crisis as they were (and still are) significant owners of Greek as extremely vulnerable from a macroeconomic perspective (fiscal deficit,
public debt. sustainability of debt, unemployment). Greece seems to rank decently on
a credit/GDP basis.
Greek banks have been operating a pure retail model which up to 2009
allowed them to grow lending volumes by 20-25% p.a. and generate
Figure 107: Greek banks’ CT1 ratios 2006-2011
RoEs of 15-20% as leverage in Greece moved towards EU standards.
Core capital (%) RWA (EURm)
They applied the same model to neighboring SEE countries which
contributed up to 25-30% of group pre-tax income. In Greece, however, 2006 2009 2010 Q2 2011 2006 2009 2010 Q2 2011
the loans/GDP ratio did not blow out of proportion and stayed at c.1.5x. National Bank of Greece 9.5% 9.5% 12.0% 10.1% 43,512 67,407 72,306 66,071
Alpha bank 7.7% 8.8% 9.1% 8.0% 33,566 51,100 48,831 47500
Following the outburst of the Greek sovereign crisis Greek banks found
Eurobank 7.7% 8.4% 8.0% 8.5% 34,504 47,827 47,968 44,858
themselves owing c. 20% of Greek public debt or 2.5x their tangible
Piraeus bank 7.5% 7.8% 7.6% 7.2% 21,800 37,394 38,100 36,675
equity. Gradually they saw 15% of their deposit base reduced and
wholesale/interbank markets shutting down. As a result they increased ATE bank 11.4% 3.9% 1.8% 1.8%* 10,261 15,098 12,636 12,841*
Source: company data and Deutsche bank; *Q1 2011
their exposure to ECB funding to c.E100bn or 20% of their total assets.
The Greek banking system (and the sovereign) has entered into a vicious Credit Data: Trends
circle of credit rating downgrades, deposit outflows, loan de-leveraging,
fiscal tightening and recession. 2011 will mark the third recessionary year Below we summarize data from 2006 to 2010 for the aggregated Greek banks
in a row with real GDP growth down c.5% yoy. We expect negative in our coverage universe. Thanks to PPP proving resilient (cheap ECB funding,
growth in 2012 as well (-1.8%). In this context asset quality has foreign operations) we see that the rise in provisions for loan losses is still
deteriorated significantly with the average NPL ratio in Greece rising to covered for albeit the ratio has fallen from 5x to almost 1x (2010).

Figure 108: Aggregation of quoted bank data, local currency (Euro mn), Greece
Local Currency 2006 2007 2008 2009 2010 2011E 2012E
Total Revenue 9,413 11,897 12,974 13,126 12,156 12,263 12,305
Costs 4,728 5,950 6,613 6,683 6,640 6,340 6,378
Pre-Provision Profits 4,685 5,947 6,361 6,443 5,516 5,923 5,927
Loan Loss Provisions 1,006 1,155 2,540 4,227 4,768 4,552 3,922
Pre-Tax Profit 3,732 4,917 3,808 1,999 745 1,462 2,005
Total Assets 232,373 284,159 332,159 354,378 360,837 350,297 351,079

Average Assets 216,514 258,266 308,159 343,268 357,608 355,567 350,688
Risk weighted assets 143,643 183,288 212,224 218,826 219,841 214,106 216,134
Risk Cushion above 7% plus 1 year PPP 4,685 5,947 6,361 6,443 5,516 5,923 5,927
Total Loans 142,832 189,436 235,647 241,588 243,725 233,788 234,253
Average Loans 127,175 166,134 212,542 238,617 242,656 238,756 234,020
Revenue / Average Loans 7.4% 7.2% 6.1% 5.5% 5.0% 5.1% 5.3%
Pre-Provision Profit / Average Loans 3.7% 3.6% 3.0% 2.7% 2.3% 2.5% 2.5%
Loan Loss Provisions / Average Loans 0.79% 0.70% 1.20% 1.77% 1.96% 1.91% 1.68%
PPP / Loan Loss Provision Cover 4.7 5.2 2.5 1.5 1.2 1.3 1.5
Estimates by Bank Figure 109: Summary of forecast data and comparing recessionary loan
loss charges as % of opening TBV and of PPP
Below we summarize forecast data for 2012 for the aggregated Greek banks
Local Currency, 2012E (EURm) AB EFG BOP ATE Total
in our coverage universe (Alpha Bank, Eurobank, Piraeus bank, ATE bank). We
then show our estimate of severe recessionary loan losses, which for the
Forecast Loan Loss Provisions 783 1,257 606 265 2,910
purposes of this exercise we have taken as 6x normalized. We have used this
figure as this, in our view, would reflect the increase in bad debt charges were Severe Recession Loan Loss Provisions 1,614 1,741 1,264 671 5,290
non-performing loans to double vs. 2010 levels. 2012 Pre-Provision Profit 976 1,215 583 361 3,136
2012 PPP less Recessionary Losses -638 -526 -681 -310 -2,155
We then express this as a percentage of 2012 book value (BV), and also as a
percentage of 2012 PPP flexed down by 25%, as a proxy of a potential
contraction during a crisis. We estimate credit losses of Euro 5bn for the
2 Years of Recessionary losses as % 2012E TE -82% -109% -83% -107% -93%
Greek banks (covered) p.a. Over two years, this would be equivalent to 90%
of 2012 closing TBV. If we include PPP estimates as an offset, then the loss 2 years of Recessionary losses as % of 2 years of PPP (*) -45% -52% -55% -64% -39%
would be 50% of TBV.

The results of this exercise show that despite a relatively resilient PPP margin Figure 112: Breakdown of system loans in Greece (E254bn)
Greek banks’ equity would be almost entirely wiped out was recession to 1%
intensify further from here causing NPLs to double in size.
13%
Danger Map Indicators for Greece Corporates
‐mortgages
Below we summarize our Danger Map Indicators for Greece. We score
48% ‐consumer
Greece overall at 33 out of 45. This captures a 1 to 5 score across a number of 31%
risk metrics, drawn from our analyses of case studies, and our global ‐other
experiences of bank lending. We rate Greece as low, in terms of overall
leverage to GDP and high in terms of macroeconomic outlook, loan mix,
Source: Bank of Greece; July data
economy structure and exchange rate flexibility. This places Greece at the
upper end of the Danger Map (red zone).
Figure 113: Breakdown of corporate loans (E122bn)
Figure 110: Greece danger map 2%

‐agricultural
Danger Factor Score Comments ‐industry
16%
Deregulation of Lending 2 No recent deregulation of lending 2% 21% ‐trade
% of Credit to GDP 1 Lower than the European average ‐tourism
5%
Change in % of Credit to GDP 3 Lower than the European average; recession- ‐shipping
driven
9% ‐other financials
Maturity of Cycle in Years 5 4-year recession expected with the maturity cycle ‐construction
prolonged 21%
5%
‐gas
Credit Mix 5 Consumption driven economy with focus on
mortgages/consumer loans 14% ‐transport
6%
‐other
Unemployment 5 Unemployment is rising
Source: Bank of Greece: July data
Current account position 5 Persistent current account deficit
Figure 114: Breakdown of household loans (E116bn)
Exchange rate flexibility 5 Zero
2%
Greek banks’ loan books 29%

‐mortgages
Figure 111: Greek credit to GDP ‐consumer
2006 2007 2008 2009 2010 ‐other
Household & Corporate credit to GDP 78% 87% 91% 97% 112% 69%
Mortgage loans to GDP 25% 28% 27% 34% 35%

Consumer loans to GDP 12% 12% 12% 15% 15%
Source: Hellenic Bank Association
Source: Bank of Greece; July data

Italy
Figure 115: Aggregation of quoted bank data, local currency (Euro bn),
Italy
2006 2007 2008 2009 2010
In this section we look at the Italian banks, with a focus on those covered Total Revenue 62,847 66,468 61,673 61,848 58,383
by Deutsche Bank (Intesa, Unicredit, MPS, Banco Popolare, UBI Banca, Costs 35,758 37,059 37,471 35,301 35,088
Banca Popolare di Milano and Credem). Over the last five years, we find Pre-Provision Profits 27,089 29,409 24,202 26,548 23,295
that loan loss provisions in the banking book averaged 76bps, and peaked Loan Loss Provisions 5,346 6,066 9,351 15,556 12,941
at 114bps (2009); their minimum level was 49bps in 2007. Pre-Tax Profit 13,559 13,130 11,481 6,566 5,441
We believe that peak provisions on a further downturn across the Euro

Total Assets 1,856,996 2,139,441 2,214,036 2,120,334 2,140,436
zone would likely reach 4x their normalized level. We estimate credit
Average Assets 1,677,371 1,998,219 2,176,738 2,165,342 2,136,287
losses of Euro 29bn for the Italian banks per annum. Over two years, this
would be equivalent to 52% of 2012 closing tangible book value (TBV). If Total Loans 1,154,433 1,259,550 1,380,922 1,335,178 1,341,845
we include pre-provision profit (PPP) estimates as an offset (flexed down Average Loans 1,073,784 1,226,544 1,327,390 1,368,707 1,337,877
consistently with the crisis experience), then the loss would be 12% of Revenue / Average Loans 5.9% 5.4% 4.6% 4.5% 4.4%
TBV, i.e. the banks would be loss-making. Pre-Provision Profit / Average Loans 2.5% 2.4% 1.8% 1.9% 1.7%
These data are probably manageable, but not negligible. Italian banks do
not have large investment banking businesses, can count on the support Source: Deutsche Bank aggregation of company data
of retail clients (the private debt to GDP is particularly low, at 38%, and
families’ savings are a valuable source of funding for banks), and generally
Estimates by Bank
their lending portfolios are fragmented. However, the corporate, SMEs,
and small business exposures weight 67% on the Italian banks’ total Below we summarize forecast data for 2012 for the aggregated Italian banks
loans, thus an extensive economic downturn in Italy can hit their asset in our coverage universe (Intesa, Unicredit, MPS, Banco Popolare, UBI Banca,
quality (for example, in 2009, the general drop in exports hurt all sectors Banca Popolare di Milano and Credem). We then show our estimate of severe
and all regions). recessionary loan losses, which for the purposes of this exercise we have
taken as 4x normalized.
As regards the Danger Map, we score Italy overall at 25 out of 45. This
captures a 1 to 5 score across a number of risk metrics. This puts Italy in We then express this as a percentage of 2012 book value (BV), and also as a
the middle of the pack on the Danger Map – neither angry red, nor percentage of 2012 PPP flexed down by 25%, as a proxy of a potential
soothing green. contraction during a crisis. We estimate credit losses of Euro 29bn for the
Italian banks per annum. Over two years, this would be equivalent to 52% of
Credit Data: Trends 2012 closing TBV. If we include PPP estimates as an offset, then the loss
would be 12% of TBV, i.e. the banks would be loss-making. Actually, without
Below we summarize data from 2006 to 2010 for the aggregated Italian banks flexing by 25% PPP, Italian banks would be at breakeven, even suffering of
in our coverage universe. We can see clearly that even though the credit cycle the stressed credit losses.
was not that severe, declining PPP combines with rising loan losses to reduce
PPP/LLP cover from 5.1x to 1.7x.
Figure 116: Summary of forecast data and comparing recessionary loan Figure 117: Scoring the Italian “danger map”
loss charges as % of opening TBV and of PPP
Local Currency, 2012E UCG ISP MPS Banco UBI BPM Credem Total Danger Factor Score Comment
Forecast Average Loan Deregulation of Lending 2 No recent deregulation of lending

Balances 602,518 418,417 168,971 101,640 109,239 40,754 21,785 1,463,322
% of Credit to GDP 1 Lower than the European average
Forecast Loan Loss
Change in % of Credit to GDP 2
Provisions 5,215 2,606 1,041 618 605 228 75 10,389
Severe Recession Loan
Loss Provisions 12,050 8,368 3,379 2,033 2,185 815 436 29,266 Credit Mix 5 Traditional model, but high weight of corporate,
SMEs and Small business
2012 Pre-Provision Profit 13,522 9,416 2,942 1,408 1,576 741 360 29,966
Unemployment 2 Unemployment broadly unchanged
2012 PPP less Recessionary
Losses 1,472 1,047 -437 -624 -609 -74 -75 700 Current account position 4
2012 Forecast Tangible Exchange rate flexibility 5

Equity 43,219 36,690 11,767 6,824 8,423 4,610 1,814 113,347 Total Danger Map Score (out of 45) 25
2 Years of Recessionary Source: Deutsche Bank
losses 24,101 16,737 6,759 4,066 4,370 1,630 871 58,533
2 Years of Recessionary
losses as % 2012E TE -56% -46% -57% -60% -52% -35% -48% -52%
Understanding the Italian banks’ loan books
2 years of Recessionary Small Business, SMEs, and large corporate represent 67% of the average loan
losses as % of 2 years of
books of the Italian banks under coverage: in our view, this highlights one of
PPP (*) -9% -7% -20% -29% -24% -11% -18% -12%
Source: Deutsche Bank estimates, (*) flexed down by 25% the main risks for the sector, in the case of economic downturn. And the
fragmentation of the exposures does not help a lot, if the crisis is spread
These are simple numbers, but they do show in our view that (1) the Italian among all industries and regions.
banks have a decent pre-provision profit generation, but that (2) in a severe
credit book downturn, we would expect a loss, in aggregate (worse for
smaller banks, than for larger banks).
Danger Map Indicators for Italy

Below we summarize our Danger Map Indicators for Italy. We score Italy
of bank lending. We rate Italy as low, for the percentage of debt to GDP and
stage of cycle issues, and high in terms of lending mix and exchange rate
flexibility. This puts Italy in the middle of the pack on the Danger Map – neither
bright red, not soothing green.

Figure 118: Italian banks under coverage: average loan mix (Q1-11): total Figure 120: Intesa loans (Q1-11): total of Euro 369bn
of Euro 1,250bn
Other
Other 7%
6% Households
International Mortgages
8% 20%
Retail
27%
Leasing
9%
Consumer Credit
Large corporate 3%
33%
Small Business
7%
Small Business
11%
Mid/Large SMEs
corporate 12%
34%
SME
Source: Deutsche Bank, company data
23%
Source: Deutsche Bank, companies data (for UCG and ISP, we exclude the exposure to CEE countries)
Figure 119 and Figure 120 refer to the largest Italian banks (Unicredit and
Intesa), whose books partially vary from peers’ for their presence abroad
On the positive, the Italian banks: 1) do not have large financing businesses
(Germany, Austria, CEE for Unicredit, and CEE only for ISP). Despite this
within their investment banks, and 2) have safe Italian retail exposures, due to
diffident geographical mix (which could help in case the economic downturn
the very low debt of families/individuals. As a consequence, we quite happily
regards only Italy), their lending mix is actually similar to the sector average.
use rather low estimates of through-cycle loan losses for residential
mortgages and domestic consumer credit.
Figure 119: Unicredit loans (Q1-11): total of Euro 556bn

Other
8% Households
Mortgages
CEE 18%
12%
Consumer Credit
2%
Leasing
6%
Small Business
8%
SMEs
7%
Mid/Large
corporate
39%

Spain
Stressing the Spanish banks for a return to recession-level loan losses, a
two years of recession losses (which we assume could be 2x higher that
the ones observed at the peak – with the exception of Bankinter at 1.5x
due to its better credit profile) would cost the sector 11% of tangible
equity including stressed pre-provision profit (flexed down by 25%
Overview
consistently with the crisis experience), and 70% of TNAV excluding any
This section looks at the Spanish banks, focussing on those covered by PPP. The best positioned are Santander and BBVA.
DB (Santander, BBVA, Popular, Sabadell, Banesto and Bankinter). Since
As regards the Danger Map and Outlook, we score the Spanish banks at
the start of economic/financial crisis, the profitability, solvency levels and,
32 out of 45. This captures a 1 to 5 score across a number of risk metrics,
in some cases, viability of Spanish financial institutions have been (and
drawn from our analyses of case studies, and our global experiences of
still are) subject to a number of major tensions and challenges. Some of
bank lending. We see the Spanish as vulnerable from a credit/GDP, %
these challenges are related to the peripherals’ sovereign/economic
change of credit to GDP, credit mix and unemployment. Even under the
situation (and their implications for access to funding markets, etc.) and
Deutsche Bank base case, of modest but steady Spain economic growth,
others are more specifically related to Spain and the consequences of the
and expected resolution of EU sovereign issues in Italy and Spain, we
past decade’s housing bubble. Some of these challenges have prompted
expect loan loss to remain higher until mid/late 2012 moment when we
the Spanish Government to launch/promote a profound restructuring
expect the first clear signs of normalization.
process for savings banks, a process that is still ongoing and which still
leaves many questions. For more details see “Spanish Saving Banks: Are
we there yet?” report published on 29 March 2011 and “Interpreting EBA Credit Data: Trends
stress tests results” report published on 16 July 2011. We summarise data for the Spanish in our coverage universe in Figure 121.
As a result of their higher than European peers pre-provision margins (due Outside of crisis conditions, the (capital-markets-assisted) strong pre-provision
to higher NIM and more streamlined cost structures), the stock of generic profit generation of the banks is evident at 3.1% of loans, comfortably
provisions that were built between 2000-2007, the disposal of non-core sufficient to absorb the 130bps of loan losses we currently forecast in 2012
assets, the earnings diversification (in Santander and BBVA’s case only) (2.4x PPP/loan loss provision cover). This rich PPP is mostly explained by a
and admittedly some capital calls (none of them received public capital high NIM and the fact that they enjoy some of the lower cost to income ratios
support), the Spanish listed banks have all managed to remain profitable in Europe.
and solvent since the beginning of the crisis. The six listed banks picture
in this report have increased core tier 1 capital from E57.7bn in 2006 to
the E112bn we forecast by end 2012, driving the sector average CT1 ratio
to 9.1% from a little over 6% in 2006.

Figure 121: Aggregation of quoted bank data, local currency (Euro million), Spain
2006 2007 2008 2009 2010 2011E 2012E
Total Revenue 45,178 52,683 61,850 70,411 72,345 74,194 79,609
Costs 21,448 23,878 27,354 28,881 31,316 33,769 35,073
Pre-Provision Profits 23,729 28,805 34,497 41,531 41,029 40,425 44,536
Loan Loss Provisions 4,833 6,164 11,528 17,706 17,792 17,495 18,277
Pre-Tax Profit 21,390 23,985 21,011 19,162 19,191 18,140 22,305
Total Assets 1,560,799 1,757,005 1,953,690 2,034,473 2,168,870 2,227,550 2,310,092

Average Assets 1,508,776 1,658,902 1,855,347 1,994,082 2,101,672 2,198,210 2,268,821
Risk weighted assets 953,673 1,036,347 1,044,230 1,104,300 1,168,414 1,174,951 1,237,912
Basel 3 additions (if known) 0 0 0 0 0 0 0
Basel 3 deductions (if known) 0 0 0 0 0 0 0
Risk Cushion above 7% plus 1 year PPP 23,729 28,805 34,497 41,531 41,029 40,425 44,536
Risk cusion to average loans
Total Loans 966,234 1,104,434 1,224,103 1,270,545 1,336,880 1,367,925 1,417,928
Average Loans 886,320 1,035,334 1,164,269 1,247,324 1,303,713 1,352,403 1,392,926
Estimates by Bank
Figure 122 shows our standard screen for the proportion of tangible equity
placed at risk by two years of recession-level loan losses, including and
excluding forecast pre-provision profit.
As our recession charge we have used 2x the peak loan losses seen for all
banks in 2009/10 (exception being Bankinter for which we use 1.5x
considering its better relative credit profile). Note that for Santander and
BBVA, which make more than half of their profits outside Europe, mainly
Latam, we have used the same assumption. We haircut 2012 forecast PPP by
25% to reflect likely weak revenue conditions in a downturn.
Figure 122: How resilient are the capital bases of the Spanish banks in a Figure 123: Spain danger map
credit downturn Danger Factor Score Comment
Local Currency, 2012E Santander BBVA Popular Sabadell Banesto Bankinter Total Deregulation of Lending 1 No recent deregulation of lending
Forecast Average Loan % of Credit to GDP 5 One of the highest in Europe
Balances 748,146 364,746 97,041 71,578 69,061 42,355 1,392,926
Change in % of Credit to GDP 4 Grew massively in the last decade
Forecast Loan Loss and deleveraging effort not too
Provisions 11,400 4,224 1,064 774 650 165 18,277 significant yet
Severe Recession Loan Maturity of Cycle in Years 3 Modest GDP growth. Significant
Loss Provisions 22,147 10,234 3,632 2,070 1,194 349 39,627 Government austerity measures.
2012 Pre-Provision Credit Mix 3 Mostly retail oriented, but with high
Profit 27,338 12,137 1,876 1,262 1,505 418 44,536 component of real estate. Large
names (SAN/BBVA) well diversified
2012 PPP less geographically
Recessionary Losses 5,191 1,903 -1,756 -808 311 68 4,909
Unemployment 5 Highest unemployment rate in Europe
2012 Forecast Tangible
Equity 55,987 33,837 8,452 5,388 5,959 3,102 112,724 Level of real interest rates 2
2 Years of Exchange rate flexibility 5

Recessionary losses 44,294 20,467 7,265 4,140 2,388 699 79,253 Total Danger Map Score (out of 45) 32
2 Years of Source: Deutsche Bank estimates
Recessionary losses as
% 2012 Forecast
Tangible Equity -79% -60% -86% -77% -40% -23% -70% Credit quality: Recent trends and outlook
2 years of
Traditional NPL ratio metrics do not represent a true picture of the underlying
Recessionary losses as
% of 2 years of PPP credit quality trends. In addition to loans currently in default, we should
(flexed down by 25%) -6% -7% -53% -42% -2% -2% -11% incorporate in the analysis those loans under surveillance (substandard loans),
the repossessed/acquired real estate assets and those loans that have been
In aggregate, we believe 2 years of recession losses would cost the sector written off already. The system’s NPA stands at c18% of the total lending
11% of tangible equity including PPP, and 70% of TNAV excluding any pre- book, which compares to the most recent stated NPL peak in 1993 of 8.9%
provision profit. Best positioned on this basis are Santander and BBVA, both of and June’s stated NPL ratio of 6.5%.
which we would expect to remain profitable.
Danger Map Indicators for Spain

We score the Spanish banks at 32 out of 45. This captures a 1 to 5 score
across a number of risk metrics, drawn from our analyses of case studies, and
our global experiences of bank lending. We see the Spanish as vulnerable
from a credit/GDP, % change of credit to GDP, credit mix and unemployment.
Even under the Deutsche Bank base case, of modest but steady Spain
economic growth, and expected resolution of EU sovereign issues in Italy and
Spain, we expect loan loss to remain higher until mid/late 2012 moment when
we expect the first clear signs of normalization.
Figure 124: Spanish banks’ credit quality

20%
18%
16%
14%
12%
10%
8%
6%
4%
2%
0%
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Stated NPL
Stated NPL + substandard loans
Stated NPL + substandard loans + repossessed assets
Stated NPL + substandard loans + repossessed assets + written off loans
Source: Deutsche Bank estimates and Bank of Spain
In the coming months a large percentage of loans currently classified as

substandard (mostly real estate-related) will effectively be in default (based in
conversations with the banks, we would estimate 65% of them). We believe
the worst on the real estate asset front is probably behind us. The stock of
repossessed assets may continue heading north but at a much slower pace.
There is lower visibility and clearer risks (on the downside) in the non-real
estate-related credit bucket. Unemployment remains high and is likely to
remain so for some time. Moreover, a sizeable portion of people are running
out of unemployment subsidy and additional austerity measures may add to
the current economic tensions, all of which will contribute to an increase in
residential mortgages/SME lending in arrears.
Linked to the previous point and leaving aside any potential changes by the
Bank of Spain to the provisions requirements, the fact that the stock of
generic provisions is running thinner in most banks, the cost of risk should be
maintained at fairly demanding levels in 2011 and 2012, thus we would be
cautious in calling a normalisation of the cost of risk.

Israel Credit Data: Trends

Below we summarize aggregated data for the Israeli banks under our
coverage for 2006 to 2010. Pre-provision profits showed a high level of
volatility in 2008 due to MBS and Lehman related losses, but recovered
Overview sharply in 2009 on profits from realizing equity holdings. Our current outlook is
The average credit cost ratio of Israel’s banks over the past five years is for stable PPP on limited fee growth and NIMs stability.
55bps which is representative of typical mid-cycle provisioning levels for
the banks. The peak of 84bps in 2009 was significantly lower than the Figure 125: Aggregation of quoted bank data, local currency (ILS bn),
peak of 110bps in the severe downturn of 2002-03. Israel
Local Currency 2006 2007 2008 2009 2010
There are two main reasons for the decline in credit costs in recent years.
After the banks experienced heavy credit losses in the corporate segment Net Interest Income 22,290 23,031 17,890 23,255 24,868
in the 2001-03 recession, there was a multi-year tightening of Other Operating Income 13,751 14,774 12,601 15,996 14,542
underwriting standards and improvement in collateralization levels. The Total Revenue 36,041 37,805 30,491 39,251 39,410
vast majority of growth in corporate credit came from the capital markets, Costs 24,929 25,165 25,846 25,556 27,462
where lending standards were looser. Secondly, there has been an Pre-Provision Profits 11,112 12,640 4,645 13,695 11,948
ongoing change in the credit mix due to an increase in mortgage loans Loan Loss Provisions 3,059 1,814 5,225 5,530 3,256
where credit losses tend to be lower. Pre-Tax Profit 8,053 10,826 -580 8,165 8,692
The banks have reported below normalized levels of credit costs in recent
quarters due to ongoing improvement in asset quality and some large Total Assets 897,281 942,762 982,135 1,006,964 1,020,674
recoveries from corporate lending. This trend reversed somewhat in Average Assets 886,739 920,022 962,449 994,550 1,013,819
recent results with an increased level of general provisioning and as Total Loans 550,613 607,371 668,915 649,309 686,601
current market conditions deteriorate, the outlook is for higher credit loss Average Loans 550,947 578,992 638,143 659,112 667,955
expenses. Revenue / Average Loans 6.5% 6.5% 4.8% 6.0% 5.9%
We assume that peak provisions in a further downturn would reach Pre-Provision Profit / Average Loans 2.0% 2.2% 0.7% 2.1% 1.8%
90bps, which is about double normalized levels, though see the local Loan Loss Provisions / Average Loans 0.56% 0.31% 0.82% 0.84% 0.49%
economy as relatively resilient to global conditions. The main area for PPP / Loan Loss Provision Cover 3.6 7.0 0.9 2.5 3.7
concern would be lending to real estate developers and leveraged holding Source: Deutsche Bank aggregation of company data
companies. With peak provisions at 90bps, the banks would remain mildly
profitable (low single digit ROE) with credit losses more than absorbed by Estimates by Bank
pre-provision profit. We calculate that peak provisions would need to
reach c200bps for the banks to be loss-making, a level of losses that Below we compare our forecasted credit loss expenses to the level we would
would be unprecedented for the system. We score Israel at 17 points on expect in a severe global recession (assumed at 2x normalized levels). The
the danger map (slightly safer than the average), with the main concern main distinction in the credit portfolios of the banks is that Mizrahi is mainly
due high corporate debt from the capital markets. retail focused and so has less cyclical provisioning levels, Hapoalim has more
of a corporate focus, while Leumi and Discount are more evenly balanced
between segments. We estimate that recessionary level loan loss provisions
would be adequately covered by pre-provision profits and that the banks

would remain profitable (with the exception of Discount). On average we Figure 127: Scoring the Israeli banks “danger map”
estimate that two years of recessionary credit losses would be equivalent to Danger Factor Score Comments
19% of 2012 tangible book value, though would expect this to be more than No recent deregulation, some tightening on
compensated for by pre-provision profit, such that equity levels would Deregulation of Lending 1 mortgage lending
increase. Relatively high non-banking credit, decline
% of Credit to GDP 3 in government debt to GDP to 73%
Figure 126: Summary of forecast data and comparing recessionary loan Change in % of Credit to GDP 2 Growth in mortgages
Maturity of Cycle in Years 2 Year 2 of credit growth
loss charges as % of opening TBV and of non-flexed PPP
Local Currency, 2012E Hapoalim Leumi Mizrahi Discount Mostly corporate, but growth from
Credit Mix 2 retail/mortgages
Forecast Average Loan Balances 242,466 241,060 116,866 122,855
Unemployment 1 Unemployment at all time low <6%
Forecast Loan Loss Provisions 1,091 741 467 676
Current account position 2 Moderate surplus though in decline
Severe Recession Loan Loss Provisions 2,425 2,411 584 1,229
Risen 275 bps over past two years. On hold
2012 Pre-Provision Profit 5,225 4,211 2,146 2,026 Level of real interest rates 2 to loosening bias
2012 PPP less Recessionary Losses 2,801 1,800 1,562 797 Central Bank intervened to prevent currency
Exchange rate flexibility 2 appreciation.
2012 Forecast Tangible Equity 25,641 25,478 7,866 11,398 Total Danger Map Score (out of 45) 17
2 Years of Recessionary loan loss provisions 4,849 4,821 1,169 2,457
2 Years of Recessionary losses as % 2012
Forecast Tangible Equity -19% -19% -15% -22% Overview of Israeli banks’ loan portfolios
The Israeli banks had a historic orientation towards corporate lending, though
this has changed considerably in recent years. Due to a process of
Danger Map Indicators for Israel disintermediation by the capital markets, lending to corporate by banks has
Below we summarize our Danger Map Indicators for Israeli banks, with an been stagnant for a number of years. Instead the banks have targeted under-
overall score of 17 points. Overall the quality of the banks’ credit portfolios has penetrated areas such as mortgage and general purpose retail lending. The
improved considerably in recent years, while the macro environment is process has improved the credit mix of the banks in our view and been carried
supportive relative to many other economies since public debt to GDP ratios out in a relatively prudent manner, with the majority of retail lending tied to
have been on the decline and the economy has been running a current payroll accounts. Lending in the corporate segment is concentrated as a result
account surplus. That said, the corporate sector has high levels of leverage of the concentrated nature of corporate ownership in the economy. The six
due to debt raised on the capital markets, a number of high profile leverage largest borrowers account for c10% of credit risk at the largest banks, with
buyouts, and large dividend payouts that have reduced equity levels. The main loans in excess of ILS 400m accounting 18% of total loans at Hapoalim and
source of credit growth in the current cycle has come from mortgages, where 15% at Leumi. Mizrahi is unique amongst the banks for having a high retail
we do not expect significant credit losses (despite elevated real estate prices) orientation, with retail lending accounting for three-quarters of the loan book,
due to low LTVs of c60% on average for new lending. More of concern would most of which is concentrated in mortgages.
be lending to real estate developers who may have overpaid to acquire land
banks as real estate prices were on the rise.

Figure 128: Israeli banks average loan mix Figure 130: Mizrahi – retail oriented loan mix
Corporate
Retail (incl small 18%
Mortgages business)
25% 25%
Commercial
4%
Retail/small
business Mortgages
Commercial 21% 57%
13%
Corporate
37%
Source: Deutsche Bank, company data Source: Deutsche Bank, company data
Figure 129: Hapoalim – corporate oriented loan mix
Mortgages
21%
Corporate
43%
Retail/small
business
26%
Commercial
10%

Brazil Credit data: Trends

Below we summarize aggregate data from 2006 to 2010 for the three largest
Brazilian banks. We note that the deterioration in credit quality in 2009 led
PPP/LLP to 1.9x, which improved to 2.6x (a normalized level) in 2010; Figure
Overview 131.
In this section we look at the Brazilian financial system, with a focus on
Figure 131: Brazilian financial system data
Banco do Brasil, Bradesco and Itau Unibanco. Over the last five years,
R$mn 2006 2007 2008 2009 2010
loan loss provisions averaged 5.2% of average loans, peaking at 6.1% in
2009; their minimum level was 4.3% in 2010.
We assume that peak provisions could reach 6.2% (roughly in line with Total Revenue 88,199 103,016 117,019 141,045 159,308
2009’s peak) of average loans in the event of a significant downturn in the
Costs 40,191 52,406 57,705 65,622 76,266
Brazilian economy. (We note that the Brazilian economy contracted 0.6%
in 2009). In such case, we estimate credit losses of R$63.8bn for the
Loan Loss Provisions 18,252 17,898 23,975 38,965 32,470
three Brazilian banks. Over two years, this would be equivalent to 68% of
2012 tangible book value (TBV). If we include pre-provision profit Pre-Tax Profit 22,978 33,244 35,549 36,868 50,814
estimates as an offset (reduced by 25% for a crisis scenario), then there

would be a 24% increase in TBV, meaning that system would remain Total Assets 875,371 1,143,407 1,594,489 1,823,045 2,203,769
profitable. Average Assets 790,052 1,009,390 1,368,948 1,708,767 2,013,407
Total Loans 290,813 438,998 595,366 678,787 828,136
Brazilian banks are well reserved and well capitalized, and the Brazilian
Central Bank (BCB) has instruments to provide liquidity in case of another Average Loans 299,559 386,253 517,182 637,077 753,461
credit crunch, especially by reducing its high reserve requirements on Revenue / Average Loans 29.4% 26.7% 22.6% 22.1% 21.1%
bank deposits. In addition, the BCB cautiously requires that banks Pre-Provision Profit / Average Loans 13.7% 13.3% 11.5% 11.8% 11.0%
maintain a capital ratio of at least 11%. According to the BCB, the Loan Loss Provisions / Average Loans 6.09% 4.63% 4.64% 6.12% 4.31%
Brazilian banks’ overall capital ratio exceeded 17% in mid-2011. We note PPP / Loan Loss Provision Cover 2.2 2.9 2.5 1.9 2.6
that the current capital ratio is inflated by a large volume of deferred tax Source: Deutsche Bank aggregation of Company data.
credits, which will have to be gradually removed from the capital

calculations to adjust to Basel III requirements. However, the BCB plans Estimates by bank
to initiate a migration to Basel III in July 2012 and expects most banks to
adjust relatively easily to the new rules mainly by retaining profits. Below we summarize forecast data for 2012 for the aggregate banks in our
analysis, Banco do Brasil, Bradesco and Itau Unibanco. We estimate loan loss
In the danger map, we score Brazil overall at 25 out of 45. This captures a provision of 6.2% for the system, roughly in line with the peak of 2009; Figure
1 to 5 score across a number of risk metrics. This puts Brazil into a 132.
moderate to high zone on the danger map (in context to emerging
markets). We estimate that the potential impact in such scenario represents c.70% of
tangible book (68%) if provisions remain at such level for two years.
However, if we offset such provision with two years of pre-provision profit
(flexed down by 25%), the system would remain profitable, as post-provision
profits would represent 24% of the system tangible equity.
Figure 132: Brazilian banks – impact of recessionary loan loss charges on Figure 133: Brazilian banks: Danger Map
TBV and PPP
R$mn, 2012E BBAS3 BBD ITUB Total Danger Factor Score Comments
Forecast Average Loan Balances 419,838 267,854 338,286 1,025,978 Deregulation of Lending 1 No meaningful changes in regulation recently
Forecast Loan Loss Provisions 16,403 12,645 20,077 49,125 % of Credit to GDP 3 Ratio has doubled in the past 7 years, but from a
low level and remains below the level of developed
Severe Recession Loan Loss Provisions 20,992 17,411 25,371 63,774
markets
2012 Pre-Provision Profit 38,528 31,962 44,257 114,747
Change in % of Credit to GDP 5 Credit has grown at a CAGR of 20% since 2003
2012 PPP less Recessionary Losses 17,536 14,551 18,886 50,973
Maturity of Cycle in Years 5 We expect loan growth to start to decelerate in the
Loan Losses to Total Loans 5.0% 6.5% 7.5% 6.2% coming years
2012 Forecast Tangible Equity 51,147 58,161 79,123 188,430 Credit Mix 3 Credit growth driven by individual loans; mortgage
penetration remains very low
2 Years of Recessionary losses 41,984 34,821 50,743 127,548
Unemployment 1 Unemployment at historically low levels
2 Years of Recessionary losses as % 2012 Forecast TE -82% -60% -64% -68%
Current account position 3 Current account represents 2.3% of GDP
2 years of Recessionary losses as % of 2 years of PPP* 31% 23% 20% 24%
Source: Deutsche Bank estimates. * Flexed down by 25%. Level of real interest rates 2 Rates have been rising for the past 1.5 years, but
are low compared to historical levels and have
recently started to trend down
Danger Map Indicators for Brazil Exchange rate flexibility 2 Floating exchange rate system
Below we summarize our Danger Map Indicators for the Brazilian banks. We Total Danger Map Score (out of 45) 25
Source: Deutsche Bank estimates.
score Brazil at 25 out of 45. This captures a 1 to 5 score across a number of
risk metrics, drawn from our analyses of case studies, and our global
experiences of bank lending. We believe that unemployment should not be a Brazilian bank loan book
major concern in Brazil, as it is in historical lows. However, recent growth in System loans have been growing at a double digit pace since 2004 (at a CAGR
credit and maturity credit cycle should be on investor’s radar screen, as credit of 20%), led by a declining interest rate environment, economic stabilization,
has grown at a CAGR of 20% in the past 7 years. Finally, we believe that such and changes in the regulatory environment, which essentially made it easier
growth should decelerate in the coming years; see Figure 133. for banks to foreclose on mortgages; Figure 134. As such, loan penetration
has improved to 47% of GDP from 22% in 2003; Figure 135. Such growth
has been driven primarily by consumer loans (especially payroll loans, credit
cards, and autos), including mortgages, although the mortgage market
remains fairly underpenetrated at less than 10% of total credit (or less than
5% of GDP).

Figure 134: System credit growth (YoY) Figure 136: Loan breakdown
(YoY growth) Agribusiness
8%
40%
34.4%
35%
Individuals
37%
30%
Large Corporate
32%
25%
19.8%
20%
15%
14.9%
10% SME
Dec-04 Sep-05 Jun-06 Mar-07 Dec-07 Sep-08 Jun-09 Mar-10 Dec-10 23%
Source: Deutsche Bank estimates and Company reports. * Reflects the loan portfolios of Banco do Brasil, Bradesco, Itau Unibanco, and
Source: Deutsche Bank estimates and Brazilian Central Bank. Santander Brasil on June 30, 2011.
Figure 135: Loan penetration (%of GDP) Figure 137: Individual loan breakdown
(percent of GDP) Other
50% 6%
Mortgages
47.3% 6%
45%
Auto
33%
40%
Credit Cards
17%
35%
30%
24.7%
25%
20% 21.8%
Personal
17%
Payrolls
15% 20%
Dec-01 Dec-02 Dec-03 Dec-04 Dec-05 Dec-06 Dec-07 Dec-08 Dec-09 Dec-10
Source: Deutsche Bank estimates and Company reports. * Reflects the loan portfolios of Banco do Brasil, Bradesco, Itau Unibanco, and
Source: Deutsche Bank estimates and Brazilian Central Bank. Santander Brasil on June 30, 2011.

Loan spreads are notoriously very high in Brazil, which is usually explained by We do not believe there is a credit bubble in Brazil
the overall high level of interest rates, high reserve requirements on bank
deposits, heavy tax burden on the financial sector, and high bank Investors are increasingly concerned about the consequences of the fast
concentration. The natural consequence is a high non-performing loan ratio; expansion in credit that Brazil has experienced over the past few years,
Figure 138 and Figure 139. especially the possibility of a banking crisis. These concerns have been
heightened by the heavy debt burden on consumers, strong inflow of foreign
Figure 138: System credit spreads capital, and the rapid increase in real estate prices.
(credit spreads, % annual)
55
However, we do not believe that Brazil is heading towards a banking crisis, as
50.8 1) credit as a share of GDP was extremely low before it started growing in
50
45.2 2004, and remains relatively low today when compared to other countries, 2)
45
financial bubbles such as the US mortgage bubble usually thrive under low
40
interest rates, which is clearly not the case in Brazil. Moreover, in contrast to
35
33.1 the US, the mortgage market amounts to less than 5% of GDP in Brazil, and 3)
30.7
30
30.0
Brazilian banks are well reserved and well capitalized, and the Central Bank has
27.4
25 instruments to provide liquidity in case of another credit crunch, especially by
20 18.4 reducing its high reserve requirements on bank deposits.
19.3
14.4
15 In addition to high reserve requirements on bank deposits, the Brazilian Central
10 Bank cautiously requires that banks maintain a capital ratio of at least 11% –
Dec-03 Nov-04 Nov-05 Nov-06 Nov-07 Nov-08 Nov-09 Nov-10
above international standards. According to the Central Bank, the Brazilian
Corporates Individuals System
Source: Deutsche Bank estimates and Brazilian Central Bank.
banks’ overall capital ratio exceeded 17% in mid-2011. The current capital
ratio is inflated by a large volume of deferred tax credits, which will have to be
gradually removed from the capital calculations to adjust to Basel III
Figure 139: System NPLs
(loans overdue more than 90 days as percent of total loans) requirements. However, the Central Bank plans to initiate a migration to Basel
9.0
8.6
III in July 2012 and expects most banks to adjust relatively easily to the new
8.0 7.7
rules mainly by retaining profits.
7.0
6.6
But loan growth is likely to decelerate in the coming
5.9
6.0
5.9 5.1
years
5.0 5.2
Brazilian consumers face a very high debt service burden of 26% (perhaps the
4.0
4.0 3.8
highest in the World), reflecting both high interest rates and average
3.0
3.5 2.8 consumer loan duration of less than two years. However, such ratio is not
new and has remained relatively stable since 2008 despite the strong increase
2.0
1.7 1.6
in consumer debt, due to a decline in interest rates and increase in loan
1.0
Dec-03 Nov-04 Nov-05 Nov-06 Nov-07 Nov-08 Nov-09 Nov-10
duration; Figure 140 and Figure 141.
Corporates Individuals System

This does not mean that one should not worry about the heavy consumer debt Figure 141: Interest rate vs. loan maturity – Individuals
burden. In the near term, the introduction of macro-prudential measures at the 600 91%
end of 2010 and monetary tightening are slowing consumer credit expansion.
550 83%
In the medium term, however, the heavy debt burden could restrain further
financial development. We believe that further credit penetration will depend 500 75%
essentially on a decline in interest rates and an increase in loan maturities,

450 67%
especially through further development of the mortgage market. While the
mortgage market in Brazil is still very small and has plenty of room to grow, 400 59%
we believe that further expansion will depend on critical changes in the 350 51%
mortgage system, primarily in the development of alternative funding
instruments, as currently mortgages are primarily funded by savings accounts. 300 43%
250 35%
Figure 140: Household debt and debt service Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11
Maturity (days), lhs Interest rate, rhs
40% Source: Deutsche Bank estimates and Brazilian Central Bank.
35%
Asset quality should deteriorate, but less than in
30% previous cycle
25%
Non-performing loans increased sharply in 2009 in the aftermath of the
Lehman Brothers crisis and domestic recession. As the economy recovered,
20% delinquency of consumer loans declined steadily and bottomed at 5.7% in
January 2011 (very close to the series’ nadir of 5.3% registered in October
15% 2000). Non-performing consumer loans are rising again, but remain relatively
low by historical standards. Corporate loan delinquency, however, failed to
10%
Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 decline alongside the economic recovery. A possible explanation is that the
Debt service/income Debt/income market is facing an “adverse selection problem,” as the better credits pursue
cheaper loans from the BNDES, and riskier borrowers have to seek financing
from commercial banks.
We expect NPLs to increase further in the coming months, both in the

consumer and corporate segments, due to rising interest rates and expected
deceleration in economic activity. However, we do not expect NPLs to rise
more than they did in previous cycles, given a historically low unemployment
rate, and rising labor earnings.

Figure 142: Unemployment

13%
13.0%
12%
11%
10.5%
10%
9%
8.6%
8%
7% 7.0%
6.5%
6%
5.9%
5%
Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12
Base case Worst case Actual data

Figure 143: Consumer NPL scenarios

9.0%
8.6%
8.5%
8.3%
8.0%
7.5%
7.2%
7.0%
6.5%
6.4%
6.0%
5.7%
5.5%
5.0%
Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12
Base case Worst case Actual data


Mexico
Figure 144: Mexico financial system data
Pesos mn 2006 2007 2008 2009 2010
Overview Total Revenue 248,155 288,707 317,584 342,077 342,043
Costs 123,246 143,155 155,437 161,815 177,518
In this section we look at the Mexican financial system and Banorte,
which is the fourth largest bank in Mexico and the only one under our
coverage. Most banks in Mexico are owned by global banks, such as Loan Loss Provisions 28,678 59,321 98,194 106,203 74,947
Citigroup, BBVA, Santander, HSBC, and Scotia Bank. Over the last five Pre-Tax Profit 99,215 91,065 66,737 76,999 94,279
years, loan loss provisions in the banking system averaged 4.1% of
average loans, peaking at 5.4% in 2009; their minimum level was 2.3% in Total Assets 3,651,317 4,244,874 4,943,505 4,828,977 5,357,606
2006. Average Assets 3,434,826 3,948,095 4,594,189 4,886,241 5,093,291
We assume that peak provisions reach 5.5% of average loans (similar to Total Loans 1,371,970 1,717,512 1,910,513 1,993,291 2,166,634
2009) in the event of a significant downturn in the Mexican economy. In Average Loans 1,259,122 1,544,741 1,814,013 1,951,902 2,079,963
such case, we estimate credit losses of Ps154bn. Over two years, this Revenue / Average Loans 19.7% 18.7% 17.5% 17.5% 16.4%
would be equivalent to 41% of 2012 tangible book value (TBV). If we Pre-Provision Profit / Average Loans 10.2% 9.7% 9.1% 9.4% 8.1%
include pre-provision profit (PPP) estimates as an offset (reduced by 25% Loan Loss Provisions / Average Loans 2.28% 3.84% 5.41% 5.44% 3.60%
for a crisis scenario), then there would still be a 1% increase in TBV. PPP / Loan Loss Provision Cover 4.5 2.5 1.7 1.7 2.3
Source: Deutsche Bank, CNBV.
We think the banks in Mexico are well protected, given an average capital
ratio of 17.4% for the system, well above Mexico’s minimum requirement
of 8%, and an average coverage ratio of 200% of NPLs as of December Estimates for Banorte and System
2010. Furthermore, we view the above scenario as very extreme, as GDP Below we summarize forecast data for 2012 for the Mexican financial system
fell 6.5% in 2009, but recovered 5.5% in 2010 and our economists do not and Banorte. We estimate loan loss provisions of 5.5% of system loans and
expect a recession in Mexico next year. DB economists forecast GDP 3.5% of Banorte loans, similar to the levels reached in 2009 when GDP in
growth of 4.0% in 2011 and 3.5% in 2012, even with a potential Mexico fell 6.5%.
slowdown in the US. Nonetheless, profitability could be under pressure
from higher provisions. The potential impact of such a scenario would represent more than 40% of
In the Danger Map, we score Mexico overall at 17 out of 45. This tangible equity for both the system and Banorte if provisions remained at such
captures a 1 to 5 score across a number of risk metrics. This puts it at a levels for two years. Nonetheless, if we offset such provisions with two years
moderate to low risk in the Danger Map, but not at an extreme case. of pre-provision profits (flexed down by 25%), the system and Banorte would
remain profitable, as post-provision profits would represent 12% of Banorte’s
tangible equity and 1% of the system’s. Nonetheless, profitability would be
Credit Data: Trends severely pressured, which could significantly impact valuation as Banorte
Below we summarize data from 2006 to 2010 for the aggregated Mexican trades at 2.1x trailing tangible book value.
financial system. The deterioration in the credit cycle peaked in 2009 and has
been improving. The coverage of loan loss provisions (LLP) by PPP improved
to 2.3x in 2010, near the level seen in 2007.
Figure 145: Mexico banks – impact of recessionary loan loss charges on Mexican bank loan books
TBV and PPP
Consumer loans are still a relatively minor percentage of Mexican bank loans.
Pesos mn, 2012E Banorte System
Indeed, consumer loans represent 19% of total loans for the system as of
Forecast Average Loan Balances 390,546 2,802,916 December 2010 and only 11% for Banorte. Meanwhile, commercial loans
Forecast Loan Loss Provisions 6,600 65,962 represent 64% of system loans and 68% of Banorte’s loans, and mortgages
Severe Recession Loan Loss Provisions 13,669 154,160 represent 17% of system loans and 21% of Banorte’s loans. As such, we
2012 Pre-Provision Profit 23,400 208,081
think significant asset quality deterioration can be contained, as consumer
loans are the riskiest types of loans. For example, the consumer NPL ratio for
2012 PPP less Recessionary Losses 9,731 53,920
the system reached 8.0% in 2008 from 5.9% in 2007 and 4.4% in 2006.
2012 Forecast Tangible Equity (TE) 63,787 753,292 However, the total NPL ratio only increased to 3.2% in 2008 from 2.5% in
2007 and 2.0% in 2006. The consumer NPL ratio has since improved to 4.4%
2 Years of Recessionary losses 27,338 308,321
and the total NPL ratio has improved to 2.3% at the end of 2010.
2 Years of Recessionary losses as % 2012 Forecast TE -43% -41%
2 years of Recessionary losses as % of 2 years of PPP* 12% 1% Figure 147: System loans as of December 2010
Source: Deutsche Bank estimates. * Flexed down by 25%.
Mortgages
17%
Danger Map Indicators for Mexico
Below we summarize our Danger Map Indicators for Mexican banks. We
score Mexico overall at 17 out of 45. This captures a 1 (least risk) to 5 (most
risk) score across a number of risk metrics, drawn from our experience of past
crises. We rate Mexico low in credit penetration and the level of interest
rates, which are near historical lows. We think unemployment is the biggest
concern, as it is near a ten-year high, but it is improving. This puts Mexico at a Consumer
19%
moderate to low level of risk.
Figure 146: Mexican banks: Danger Map Commercial
64%
Deregulation of Lending 1 No meaningful changes in regulation recently
One of the lowest ratios in Latam, and still below pre-
% of Credit to GDP 1
Tequila crisis levels
Credit has grown at a CAGR of less than 15% in the past Source: Deutsche Bank, CNBV.
Change in % of Credit to GDP 3
seven years
Credit growth is accelerating after single digit growth in
2009
Consumer loan penetration is one of the lowest in the
Credit Mix 2
region
Unemployment 3 Near a ten-year high, but improving
Current account position 2 Current account represents 0.9% of GDP

Interest rates at historically low levels and unlikely to rise in
the near term
Exchange rate flexibility 2 Floating exchange rate system
Source: Deutsche Bank estimates.

Figure 148: Banorte loans as of December 2010
Mortgages
21%
Consumer
11%
Commercial
68%
Source: Deutsche Bank, company data.

Russia
First, while the state was almost debt-free, Russian companies and banks
were heavily dependent on international capital markets. In total they had
around USD500bn in foreign debt in 2008, or over 30% of GDP and
almost equal Russia’s FX reserves. In 2009, around USD120bn in foreign
debt was coming due and this caused imminent problems when capital
Overview
markets closed.
In this section we look at the Russian banks, with a focus on those
Second, the slowdown in the global economy caused by the financial
covered by Deutsche Bank (Sberbank, VTB, Nomos, Bank Saint
crisis led to an unprecedented collapse in commodity prices. In a matter
Petersburg and Vozrozhdenie). Over the 2006-10 period, we find that loan
of six months, the oil price slumped from USD145 to just over USD30 per
loss provisions averaged 2.8%, and peaked at 7.3% in 2009. The
barrel. Two-thirds of Russia’s exports and half of budget revenues are
minimum level was reached in 2007, when banks had risk costs of 0.8%
coming from oil. Our economists estimate that each USD5 change in the
of loans.
oil price has a 0.3ppt impact on real GDP growth.
We think that Russian banks are highly exposed to the region's economy
Third, the slumping commodity prices wreaked havoc on the RUR
(=oil price). In fact, a crisis now would probably more damaging than the
exchange rate. Between 2Q08 and 1Q09, the RUR shed 35% vis-à-vis the
2008/09 crisis: asset quality is worse, IB exposure bigger and state
USD. Due to the dominance of oil and gas for the Russian economy, the
support less likely due to fiscal constraints. Liquidity in the sector has
ruble is closely correlated to the oil price.
deteriorated in recent months although is still abundant compared to the
2008/09 crisis. Currency weakness is a much bigger problem in Russia than in most large
emerging markets, because the economy is not diversified: Russia exports
While Russian banks’ earnings are highly geared to changes in asset
commodities and imports finished goods and consumer products. There is
quality, we see their capital as resilient. High profit margins (before
hardly a domestic producers’ base that competes with imported goods or
impairment charges are taken into account) allows the banks to
exports to developed markets’ consumers.
accumulate bad debt provisions without making a loss, while low
leverage ratios mitigate the impact of rising bad loans on earnings and
In 2009, Russia’s real GDP shrank 7.8% and inflation surged to 15.5% from
capital. This conclusion is underpinned by the experience of the 2008/09
single-digit in 2008, making it one of the worst performing economies in the
crisis when most banks (with the exception of VTB) remained profitable in
industrialized world in 2009. Instead of stabilizing the economy through import
spite of surging NPLs.
substitution and rising exports, the ruble devaluation merely led to imported
As regards the Danger Map, we score Russia overall at 23 out of 45, inflation and depleted consumption.
which puts Russia at the higher end on the Danger Map for Emerging
markets. Russia’s main weaknesses are the poor current asset quality, Credit Data trends
the one-dimensional exposure of the economy / currency to change in the
oil price and the limited ability of the authorities to mitigate its impact. The crisis had an immediate impact on credit quality. The payment system
almost came to a halt as companies and consumers stopped paying ruble bills
The 2008/09 crisis experience in Russia in anticipation of further devaluation and the economy moved into cash
hoarding mode. Credit quality soured and the sector’s NPL ratio surged from
The 2008/09 crisis fully exposed Russia’s dependence on the global economy,
1.8% in mid-2008 to a 10.0% peak in February 2010.
international capital markets and the limited diversification of the domestic
economy.

Russian banks’ capital proved resilient to the crisis. Exceptionally high profit What is different now?
margins (before credit costs are taken into account) allowed the banks to
accumulate bad debt provisions without making a loss. This is demonstrated While the probability of a repeat of the 2008/09 crisis is, perhaps, limited, we
by aggregated data of the Russian banks in our coverage universe. Russian think a crisis now would be more damaging to Russian banks for several
banks’ pre-provision profit margin reached 7.0% in 2009, and this allowed reasons:
them to absorb 7.3% risk costs.
Bad assets are weighing on the sector: Current NPL ratios are three to
Russian banks also benefit from low leverage ratios, which mitigate the five times higher than before the 2008/09 crisis, as most Russian banks
impact of rising bad loans on earnings and capital. Russian banks assets total have not dealt with the bad loans from the crisis. At the same time,
6-8x Tier 1 capital, which compares to 12-15x for banks such as Bank of coverage ratios are much lower. At first glance the banks look well-
America and JPM and up to 25-30x for some European banks (Barclays, provisioned with NPL coverage of over 100%, but coverage drops to
SocGen, BNP). 70% if restructured loans - which account for 5-10% of loan portfolios -
are also taken into consideration.
Figure 149: Aggregation of quoted bank data, Russia) Increased IB exposure: Since the 2008/09 crisis, VTB, and to a lesser
In RUR bn 2006 2007 2008 2009 2010 2011E 2012E extent Sberbank, have built up a significant IB presence. For Sberbank a
Net Interest Income 253 336 523 691 699 824 989 loss in IB is unlikely to have more than a 5% P&L impact, but at VTB -
Other Operating Income 10 8 13 -7 2 34 39 where IB accounted for 19% of 2010 revenues and 36% of profits - a loss
Total Revenue 372 488 638 868 939 1,107 1,324 at IB could also easily halve profits and knock 1% off the forecast 10%
Costs 191 258 305 322 397 519 615 Tier 1 ratio at YE11, bringing VTB close to a new capital increase.
Pre-Provision Profits 181 231 332 546 542 588 709
On the positive, balance sheets have improved compared to 2008. Two years
Loan Loss Provisions 27 35 170 571 216 30 33
of strong deposit growth - spurred by positive real deposit rates
Pre-Tax Profit 153 196 160 -30 317 556 666 (unprecedented in Russian modern history) and significant social spending by
Total Assets 5,090 7,639 11,069 11,374 13,888 16,843 19,652 the government – has brought the sector’s LDR down from 140% in 2008 to
Average Assets 4,413 6,365 9,354 11,221 12,631 15,366 18,248 just over 100%. Even VTB, which heavily relied on capital markets’ funding
Risk weighted assets 4,641 6,869 9,915 10,173 12,859 15,566 18,101 before the 2008/09 crisis, now derive over 60% of funding from deposits,
Basel 3 additions (if known) albeit with significant support / pressure from state and state companies.
Core tier one capital 518 1,094 1,209 1,362 1,654 2,054 2,462
Basel 3 deductions (if known)
Risk Cushion above 7% + 1 year PPP 182 231 333 546 542 588 709
Risk cushion to average loans
Total Loans 3,474 5,661 8,061 7,580 8,901 11,137 13,547
Average Loans 2,688 4,567 6,861 7,820 8,190 10,019 12,342
Source: Deutsche Bank’s aggregation of company data

Figure 150: NPL ratio Figure 152: Loans-to-Deposits ratio

25% 250
20% 200
15% 150
10% 100
5% 50
0% 0
Halyk KKB VZRZ VTB SBER Nomos BPSB VTB KKB Nomos Halyk SBER VZRZ BPSB
YE08 YE10 YE08 YE10
Source: Companies Source: Companies
Figure 151: NPL coverage Figure 153: Deposits as % of IBL

4.0x 100%
3.0x 80%
60%
2.0x
40%
1.0x
20%
0.0x 0%
Halyk VZRZ VTB KKB SBER Nomos BPSB VTB KKB Nomos Halyk BPSB VZRZ SBER
YE08 YE10 YE08 YE10
Source: Companies Source: Companies

Estimates by Bank Danger Map Indicators for Russia

Below we summarize forecast data for 2012 for the aggregated Russian banks Below we summarize our Danger Map Indicators for Russia. This captures a 1
in our coverage universe. We then show our estimate of severe recessionary to 5 score across a number of risk metrics, drawn from our analyses of case
loan losses, which for the purpose of this exercise we have taken as 4x studies, and our global experiences of bank lending. We score Russia overall
normalized. We estimate the normalized LLPs at 1.5% of loans, which is at 23 out of 45, which puts Russia at the upper end on the Danger Map for
based on medium term sector growth of around 20%, a mid-cycle provision Emerging Markets. The main weaknesses in Russia are the poor credit quality
level of 5% of loans and 50bp write-offs. at present, as Russian banks have yet to fully digest the legacy of bad loans
from the previous crisis, the difficulty in repossessing collateral and the one-
The table below highlights that the impact of a recessionary scenario on dimensional exposure of the economy and the currency to the oil price, which
Russian banks’ capital would be minimal. We estimate the accumulated severely limits policy makers’ ability to counter the impact of a global fallout.
Recessionary losses of the five Russian banks under coverage at only
RUR32bn (on a pre-tax basis). Two years of recessionary losses would thus Figure 155: Scoring Russia on the “danger map
lead to an aggregated net loss (using the 20% statutory tax rate in Russia) of Danger Factor Score Comments
only RUR52bn, or only 2% of tangible equity. Deregulation of Lending 2 No recent deregulation of lending. State banks
dominate the sector
Figure 154: Summary of forecast data, Russian banks % of Credit to GDP 2 Banking sector penetration is low, mitigating the
impact of credit growth on the economy
In RUR bn, 2012E SBER VTB Nomos VZRZ STBK Total
Change in % of Credit to GDP 3 Credit penetration broadly unchanged since
Forecast Average Loan Balances 7,436 4,034 480 153 239 12,342 2007
Forecast Loan Loss Provisions 8 17 6 1 17 50 Maturity of Cycle in Years 4 Bad loans from the 08/09 crisis still weighing on
Severe Recession Loan Loss Provisions 446 242 29 9 14 741 balance sheets
2012 Pre-Provision Profit 482 182 26 6 13 709 Credit Mix 4 High % of large corporate loans. Difficulties in
repossessing collateral
2012 PPP less Recessionary Losses 36 -60 -3 -3 -2 -32
Unemployment 2 Unemployment broadly unchanged
2012 Forecast Tangible Equity 1,575 740 77 22 48 2,392
Current account position 1 Surplus
2 Years of Recessionary losses 892 484 58 18 29 1,434
Level of real interest rates 2 Real rates negative, but less than before the
2 Years of Recessionary losses as % 2012 -57% -65% -74% -84% -59% -60% 2008/09 crisis
Exchange rate flexibility 3 Currency largely driven by the oil price
2 years of Recessionary losses as % of 2 years of -11% -28% -24% -44% -20% -17%
PPP (flexed down by 25%) Total Danger Map Score (out of 45) 23

India
Figure 156: Aggregation of quoted bank data, local currency, (INR m)
Local Currency 2006 2007 2008 2009 2010
Net Interest Income 524,319 608,975 773,836 886,112 1,189,297
Overview Total Revenue 811,835 1,057,705 1,260,207 1,593,150 1,865,096

Costs 477,093 585,459 645,654 849,730 994,361
In this section we look at the Indian banks, with a focus on those covered Pre-Provision Profits 334,742 472,246 614,553 743,420 870,734
by DB (six government owned banks (PSU banks) and five private banks). Loan Loss Provisions 72,341 107,794 136,748 205,543 227,184
Over the last five years, we find that loan loss provision (as % of average Pre-Tax Profit 264,434 366,544 477,701 540,807 645,812
loans) averaged 86bp, with maximum of 115bps (FY2010) and minimum
of 69bps (FY2007). Total Assets 20,833,528 25,897,684 31,584,242 36,185,766 43,606,867
Average Assets 18,849,475 23,365,606 29,115,971 33,885,004 39,896,317
The average through cycle credit cost of Indian Banks varies between 25
Total Loans 12,148,814 15,046,066 18,596,312 21,515,733 26,467,671
to 100bps. We believe that peak provisions in case of a recession would
Average Loans 10,851,025 13,771,012 17,289,199 20,312,562 24,311,089
likely reach 2x their normalised level. We estimate credit losses of
INR 1.08 trn per annum for the eleven Indian banks. Over two years, this
would be equivalent to 38% of 2012 closing tangible book value. As we
had seen in the period 2007 to 2009 the PPP at Indian banks went up. In
case of another recession we believe that Indian banks will still continue Source: Deutsche Bank aggregation of company data
to report profits, but may be a lower level of profitability.
The actual impact on earnings by way of credit costs could be lower than Estimates by Bank
above if RBI allows any one-time relaxation to the banks as it did in 2008.
Below we summarise forecast data for 2012 (FY13E or year ending 31 March
In end 2008 RBI allowed banks to restructure real estate loans once and
2013) for the aggregated Indian banks in our coverage universe. We then
other loans twice and yet continue to treat them as standard assets. In
show our estimate of severe recessionary loan losses, which for the purposes
these cases banks were required to make only NPV provisions on
of this exercise we have taken as 2x normalised.
restructured loans, which were generally lower than the case compared
to situation had these loans turned NPL
As regards the Outlook and the Danger Map for India, we score India provision profit flexed down by 25% (this is the same PPP experience we saw
overall at 26 out of 45. This captures a 1 to 5 score across a number of in the 2008 credit cycle). These are simple numbers, but they do show in our
risk metrics, drawn from our analyses of case studies, and our global view that (1) the Indian banks have good pre-provision profit generation, but
experiences of bank lending. This puts India at the upper end on the that (2) even in a severe credit book downturn we would still expect Indian
Danger Map for Emerging Markets. banks to continue to make profits, in aggregate.
Credit Data: Trends

Below we summarise data from 2005 to 2010 for the aggregated Indian banks
in our coverage universe. We see that while pre-provision profits (PPP) have
risen over the period, loan loss provisions (LLP) have risen as well resulting in
decline in PPP/LLP cover from 4.6x to 3.8x.
Figure 157: Summary of forecast data and comparing recessionary loan loss charges as % of opening TBV and of non-flexed PPP
Local Currency, 2012E Axis HDFC Bank ICICI Kotak Yes Bank of Bank of India Canara SBI PNB Union Total
Baroda
Forecast Average Loan Balances 1,637,040 1,812,874 2,454,575 461,721 412,495 2,592,216 2,391,376 2,369,452 11,004,607 2,740,900 1,697,801 29,575,057
Forecast Loan Loss Provisions 15,423 18,869 24,574 2,958 905 12,826 14,192 11,792 125,486 23,049 11,732 261,805
Severe Recession Loan Loss Provisions 24,556 36,257 49,092 9,234 2,062 20,738 28,697 28,433 132,055 32,891 20,374 384,389
2012 Pre-Provision Profit 76,448 95,251 109,882 27,668 15,498 80,060 64,375 74,850 380,146 101,505 52,527 1,078,210
2012 PPP less Recessionary Losses 51,892 58,994 60,790 18,434 13,436 59,323 35,679 46,417 248,091 68,614 32,154 693,822
2012 Forecast Tangible Equity 219,406 294,575 601,765 123,043 44,610 250,740 200,931 240,904 958,049 257,127 150,975 3,342,125
2 Years of Recessionary losses 69,410 76,866 104,074 19,577 17,490 109,910 101,394 100,465 466,595 116,214 71,987 1,253,982
2 Years of Recessionary losses as % 2012 -32% -26% -17% -16% -39% -44% -50% -42% -49% -45% -48% -38%
2 years of Recessionary losses as % of 2 years 21% 22% 10% 18% 13% 4% -2% 5% 11% 14% 5% 25%
of PPP (flexed down by 25%)
Danger Map Indicators for India Understanding the Indian banks’ loan books
Below we summarise our Danger Map Indicators for India. We score India As discussed above, we regard Indian banks as not too high risk in their loan
portfolios. One of the key points regarding the Indian banks is that their books
are in fact very diverse, with retail loans accounting for ~20% of total loans.
of bank lending. We rate India as medium in terms of % of debt in GDP and
high on change in leverage and mix. This puts India at the upper end on the
Danger Map for Emerging Markets Figure 159: Loan mix of Indian Banking system – total INR37.09trn (June
2011)
Figure 158: Scoring the India “danger map”
Danger Factor Score Coment
Deregulation of Lending 3 Indian banks are required to lend 40% of loans to

priority sector - agriculture & weaker sections. This is
besides holding 24% in government bonds Industry
% of Credit to GDP 2 Debt to GDP tends to range between 50 to 60% as
fiscal deficit continues to remain high and both central
Services
and state governments borrow to plug the gap
Change in % of Credit to GDP 4 20 percentage point expansion since 2003
Retail Loans
Credit Mix 4 Some state influenced lending policy
Agriculture & Allied
Unemployment 1 Activities
Total Danger Map Score (out of 45) 26 Source: Reserve Bank of India, Deutsche Bank

Retail loans in India are largely secured loans – mortgages, auto loans. Credit
card loans are only 0.5% of the banking system loans as Indian banks have
been very conservative in giving unsecured retail loans. Due to the significant
investments being made in infrastructure in India it is the largest sector within
industry group followed by metals.
Fee income business for Indian banks is still largely related to lending
business. Fee income from non-lending activities like asset management,
advisory, investment banking, etc. is still very small. However, Indian Banks
are required to invest 24%of their liabilities in to government securities –
statutory liquidity ratio (SLR). The available for sale (AFS) part of this
investment needs to be MTM every quarter and it makes them vulnerable to
interest rates.

China
collateral, (4) Rmb46.4bn of loans with invalid guarantee by the local
government, (5) Rmb35bn of loans used for non-production activities, (6)
Rmb8bn of overdue loans. We believe the asset quality of the Listed
Banks is much stronger than the sector as a whole given their tighter
underwriting standards. Even if we were to assume a NPL ratio of 12%
Overview
for the total loans (Rmb2.83trn) extended to the local government, the
In this section we look at the Chinese banks, with a focus on those banks impact on pre-tax profit was 7% after incorporating the excess provisions
covered by DB (ICBC, CCB, ABC, BOC, BoCom, CMB, CNCB, MSB, of Rmb265bn as an offset. Hence, we believe the asset quality risks for
CRCB, SPDB, Industrial, CEB, SZDB, BOBJ, BONJ and BONB). Over the the Listed Banks are manageable.
last five years, we find the average credit cost for listed China banks
As regards the Outlook and the Danger Map for China, we score China
amounted to 71bps. Credit cost peaked in 2008 at 117bps but since then
overall at 27 out of 45. This captures a 1 to 5 score across a number of
quickly declined to 46bps and 41bps in 2009 and 2010, respectively.
risk metrics, drawn from our analyses of case studies, and our global
In the case of severe asset quality deterioration in China, we believe that experiences of bank lending. This puts China at the upper end on the
peak provisions could reach 4x normalized level or 2x the distress level Danger Map for Emerging Markets.
experienced in 2008. We estimate credit losses of Rmb815bn per annum
for the Chinese banks under our coverage. Over the two years, this would Credit Data: Trends
equivalent to 32% of 2012 closing tangible book value. If we include PPP Below we summarize data from 2006 to 2010 for the aggregated Chinese
estimates as an offset (flexed down by 25% which is consistent with the banks in our coverage universe. The data shows that despite economic
2007-2009 experience in the western markets) then there would be a gain downturn in 2008, Chinese banks’ profitability was robust enough to
of 12% of book value, i.e. the banks would be able to offset the credit withstand rise on credit cost and key ratios point to healthy sector overall.
losses and still register a profit. Credit cost peaked at 117bps in 2008 and then declined to a relatively low
Chinese economy is driven by fixed capital investment (in 2010, fixed level at 46bps in 2009 and 41bps in 2010. As a result, PPP/LLP ratio recovered
asset investment amounted to 61% of total GDP), whereas banks need to from 3.7x in 2008 to 8.1x in 2010.
comply with the national credit policy. In 2008-9, China government had
significantly loosened both its monetary and fiscal policy by launching the
RMB4trn investment plan issued in late 2009 to stimulate the economy,
given global financial crisis. According to the estimates by the National
Audit Office (NAO), the local government had raised Rmb5.22trn (or
15.3% of China’s 2009 GDP) of new debt since 2008 and the bulk of the
debt was bank loans to government investment corporations set up by
local government (or what are known as local government financing
vehicles, LGFV). Hence, Chinese banks grew loans by 33% yoy in 2009
and this raised market concerns about system asset quality, especially for
loans extended to the real estate sectors and LGFV. According to NAO
and our estimates, as of end 2010, weak quality LGFV loans amounted to
Rmb560bn or 11.2% of the total LGFV loans of Rmb4.97trn. This included
(1) Rmb244bn of loans to LGFV of improper corporate registration, (2)
Rmb150bn of evergreen loans, (3) Rmb73.1bn of loans with overstated

Figure 160: Aggregation of quoted bank data, local currency (Rmb mn), Estimates by Bank
China
Rmb mn 2006 2007 2008 2009 2010
Below we summarize forecast data for 2012 for the aggregated Chinese
banks in our coverage universe (ICBC, CCB, ABC, BOC, BoCom, CMB, CNCB,
Net Interest Income 578,586 789,665 940,607 902,844 1,136,956
MSB, CRCB, SPDB, Industrial, CEB, SZDB, BOBJ, BONJ and BONB). We then
show our estimate of severe recessionary loan losses, which for the purposes
Total Revenue 648,808 920,331 1,141,316 1,155,567 1,436,732
of this exercise we have taken as 4x normalized. We then express this as a %
Costs 292,433 388,900 448,632 488,917 578,971 of 2012 BV, and also as a % of 2012 pre-provision profit flexed down by 25%
Pre-Provision Profits 356,375 531,430 692,684 666,650 857,761 (this is the same PPP experience we saw in the 2007-2009 credit cycle in the
Loan Loss Provisions 88,685 104,486 189,243 93,760 105,820 western market). In China, PPOP fell by 3.5% only from Rmb693bn to
Pre-Tax Profit 267,716 422,138 491,475 574,713 755,852 Rmb667bn in 2009 for the banks under our coverage.
Total Assets 24,921,508 29,988,375 35,157,042 44,325,586 52,616,564

These are simple numbers, but they do show in our view that (1) the Chinese
Average Assets 23,060,126 27,454,942 32,572,709 39,741,314 48,471,075
banks have good pre-provision profit generation, with PPP expected to reach
Risk weighted assets 14,125,608 16,677,758 19,008,077 24,484,866 29,884,024
to Rmb 1494bn (by 2012), and that (2) even a severe credit book downturn,
Core tier one capital 1,408,478 1,719,038 1,874,841 2,194,276 2,921,514 would not have dramatic results on the banks’ profitability, in aggregate.
Risk Cushion above 7% plus 1 year Overall, the banks seem to be well cushioned from incurring huge losses.
PPP 776,060 1,083,025 1,236,960 1,146,985 1,687,394
Total Loans 12,865,398 14,912,561 17,481,457 23,492,779 27,734,706
Average Loans 12,034,698 13,888,980 16,197,009 20,487,118 25,613,743
Note: Due to the inaccessibility of historical data, we don’t include the data of ABC and CRCB.
Figure 161: Summary of forecast data and comparing recessionary loan loss charges as % of opening TBV and of non-flexed PPP
Rmb mn ICBC CCB ABC BOC BoCom CMB CNCB MSB CRCB SPDB INDB CEB SZDB BOBJ BONJ BONB
Forecast Average Loan Balances 8,150,416 6,882,660 5,994,681 6,679,937 2,740,118 1,750,445 1,632,531 1,329,834 157,468 1,395,463 1,076,992 986,059 507,877 442,722 110,650 152,736
Forecast Loan Loss Provisions 42,661 24,682 30,564 29,353 18,931 8,454 13,438 9,723 531 5,811 5,348 4,968 3,370 2,396 843 1,313
Severe Recession Loan Loss Provisions 172,789 145,912 127,087 141,615 58,091 37,109 34,610 28,192 3,338 29,584 22,832 20,904 10,767 9,386 2,346 3,238
2012 Pre-Provision Profit 353,994 269,261 239,440 225,413 89,660 56,565 52,584 43,785 7,059 44,243 39,448 32,884 14,095 14,320 5,485 5,440
2012 PPP less Recessionary Losses 181,205 123,349 112,353 83,799 31,569 19,456 17,974 15,592 3,720 14,659 16,616 11,979 3,328 4,935 3,139 2,202
2012 Forecast Tangible Equity 1,118,941 898,945 755,683 807,606 288,475 224,396 186,852 166,724 31,350 170,612 133,073 152,342 48,989 68,621 24,510 20,339
2 Years of Recessionary losses 345,578 291,825 254,174 283,229 116,181 74,219 69,219 56,385 6,677 59,168 45,664 41,809 21,534 18,771 4,692 6,476
2 Years of Recessionary losses as % 2012 Forecast Tangible Equity -31% -32% -34% -35% -40% -33% -37% -34% -21% -35% -34% -27% -44% -27% -19% -32%
2 years of Recessionary losses as % of 2 years of PPP (flexed down by 25%) 17% 12% 14% 7% 6% 5% 5% 6% 12% 4% 10% 5% -1% 4% 14% 8%

Danger Map Indicators for China As discussed above, the total outstanding bank loans to local governments
and LGFVs amounted to RMB8.5tn by the end of 2010, or 16.6% of total
Below we summarise our Danger Map Indicators for mainland China. We system loan outstanding, according to NAO’s report. We believe the impact
score China overall at 27 out of 45. This captures a 1 to 5 score across a on banks’ earnings is manageable when the LGFV loans hit maturity (24.5% in
number of risk metrics, drawn from our analysis of case studies, and our 2011, 17.2% in 2012, 28.1% in 2013-15, 30.2% after 2016). Even if we were
global experiences of bank lending. We rate China as medium on the maturity to assume a 12% NPL ratio for loans to the local government, the pre-tax
of cycle in years, change in % of credit to GDP and deregulation of lending, profit impact on the banks that under our coverage was 7% only after
and low in terms of % of unemployment, current account position and incorporating the excess provision.
exchange rate flexibility. This puts China at the upper end on the Danger Map
for Emerging Markets. In addition, around 18.8% of total loan was granted to the real estate sector as
end of March 2011, including 12.3% to mortgage finance and 6.5% to real
Figure 162: Scoring the China “danger map” estate developers. With a minimum down-payment of 30% and an estimated
Danger Factor Score Comments loan-to-valuation (LTV) ratio of less than 55%, we believe the asset quality risk
Deregulation of Lending 3 Price competition for loans are restricted, with for mortgage loans is low. Additionally, we think the risk of loan to real estate
floor lending rates for different loans. developers is controllable as the Listed Banks lend to large developers with
% of Credit to GDP 5 Rapid increase in 2009 (In 2010, Total Rmb & strong track record and adequate collateral.
forex loan/GDP = 128% and Total Rmb loan
/GDP = 120%)
In combination with loans extended to real estate developers, corporate loans
Change in % of Credit to GDP 4 Rapid incresae in 2009, around 100% to
120% during last 5 years accounted for 75% of total loans by the end of March 2011, which mainly
consist of loans to manufacturing, transportation, utilities, wholesale & retail
Maturity of Cycle in Years 3 Around 24.5% and 17.2% of the local
government bank loans of Rmb8.9trn will trade sectors. And as for the retail loans, which accounted for 23% of total
mature in 2011 and 2012 loans, is mainly for housing mortgage, personal consumer loans, auto loans
Credit Mix 4 Some state influenced lending policy and credit cards. The remaining 2% of total loans is for overseas purpose.
Unemployment 1 Stable around 4%
Current account position 1 No current account deficit. Current A/C Figure 163: Breakdown of gross Figure 164: Breakdown of loan for by
surplus = 5.1% of GDP in 2010
loans by term, China (Mar 2011) major sector, China (Mar 2011)
Level of real interest rates 3 Real interest rate is negative, but is rising due
to higher lending rates and falling inflation Others ST Retail Others
Other Corp Housing
2% 7% 2%
expectation 1% 12%
Mid & LT Retail
Exchange rate flexibility 3 Undervalued currency leading to potential 16% Other Retail
credit bubble 10%
Total Danger Map Score (out of 45) 27 Mid & LT Corp

43% Real Estate
Source: Deutsche Bank 6%
Understanding the Chinese banks’ loan books ST Corp

Other Corporate
70%
31%
China’s economic growth has been driven by fixed asset investment (FAI Source: Deutsche Bank, PBOC Source: Deutsche Bank, PBOC
accounted for 61% of total GDP in 2010) and as a result, around 59% of
banks’ loans are medium to long term loans. We expect this ratio to fall as
China becomes increasingly consumption dependent. Overall, China’s asset
quality remains strong, with NPL ratio falling to 1.1% as of end March 2011.
Turkey
Figure 165: Turkish banks financial summary (2006-2010), Turkey*
TRYm 2006 2007 2008 2009 2010
Total Revenue 18,760 24,390 26,964 36,238 36,010
Overview Costs 8,898 10,044 12,756 13,235 15,049

In this section, we analyze the asset quality evolution of Turkish banks we Loan Loss Provisions** 697 1,278 2,417 4,539 773
cover (Akbank, Garanti, Halkbank, Isbank, Vakifbank, Yapi Kredi, Bank Asya, Pre-Tax Profit 8,016 11,234 10,031 15,823 18,079
Albaraka Turk) during the last global financial crises (2008 and 2009) and Total Assets 309,728 358,910 452,058 522,027 623,542
possible impact of a further global downturn on Turkish banks. While averaged Average Assets 286,979 334,319 405,484 487,043 572,785
Total Loans 142,297 185,327 246,708 255,503 340,953
at 95bps over the last five years, the cost of risk adjusted for NPL collections
Average Loans 121,569 163,812 216,017 251,106 298,228
(CoR) bottomed at 57bps in 2006 and peaked at 181bps in 2009, when the Revenue / Average Loans 15.4% 14.9% 12.5% 14.4% 12.1%
negative influence of global crisis felt intensely. Pre-Provision Profit / Average Loans 8.7% 9.2% 7.0% 9.7% 7.7%
Loan Loss Provisions / Average Loans** 0.57% 0.8% 1.1% 1.81% 0.3%
Taking into consideration the trends seen between 2006 – 2011; on a global
PPP / Loan Loss Provision Cover** 15.2 11.8 6.3 5.4 29.9
downturn specific loan loss provisions would increase 3.5x from their * An aggregate of Akbank, Garanti, Halkbank, Isbank, Vakifbank, Yapi Kredi, Albaraka Turk and Bank Asya
** Loan loss provisions includes collections from previous periods
normalized levels, in our view. On this basis, we estimate that the loan loss Source: Deutsche Bank
provisions of Turkish banks we cover would increase to TRY8bn (cumulative)

per annum (CoR: +180bps). In two-year horizon; this would therefore be Estimates by Bank
equivalent to 18% of 2012e closing tangible book value. Having said that the
potential hike in specific provisions is considerable in Turkish banks standards, In this section, we have analyzed the possible impact of global financial crises
we believe that the system would remain profitable. Assuming that all the on the individual banks we cover (Akbank, Garanti, Halkbank, Isbank,
remaining P/L items remain unchanged as in our base assumptions, the pre- Vakifbank, Yapi Kredi, Albaraka Turk and Bank Asya). In order to remain on the
provision profit (PPP) should more than offset the deterioration in the asset conservative side, we assumed that loan losses of each bank would increase
quality outlook. We expect net interest income and fee revenues to be the 3.5x and remain at such levels for two years.
major drivers of PPP in the underlying time frame. As regards the Danger
As can be seen from Figure 166; the potential impact of a hike in loan loss
Map, we score Turkish banks overall at 22 out of 45. This captures 1 to 5
provisions on tangible equity would be 18% on average. However, this still
score across a number of risk metrics. This put Turkey in-line with the average
implies positive bottom-line earnings for Turkish banks, thanks to the expected
for Emerging markets.
strength at top-line revenues that are to be driven primarily by asset repricing
and sustained credit volume expansion.
Credit Data: Trends
Below, we summarize data between 2006 and 2010 for the cumulated Turkish
banks in our coverage universe. It is evident that even at the time of the global
crisis, Turkish banks managed to cover 5.4x of their loan losses in 2009 and
the ratio increased to 29.9x in 2010.

Figure 166: Turkish banks – impact of recessionary loan loss charges on opening TBV and PPP
TRYm, 2012E Akbank Garanti Halkbank Isbank Vakifbank Yapi Kredi Bank Asya Albaraka Turk Total
Forecast Average Loan Balances 76,058 93,071 61,976 88,638 63,494 72,415 14,599 8,863 479,112
Forecast Loan Loss Provisions* 425 474 203 437 395 398 87 48 2,467
Severe Recession Loan Loss Provisions (3.5x of the estimate) 1,487 1,659 709 1,530 1,382 1,395 304 167 8,634
2012 Pre-Provision Profit 4,581 4,592 2,866 4,508 2,446 3,612 559 372 23,537
2011 Pre-Provision Profit 4,050 4,131 2,720 4,169 2,209 3,307 498 317 21,404
2012 PPP less Recessionary Losses 3,093 2,933 2,157 2,978 1,064 2,217 255 206 14,903
2012 Forecast Tangible Equity 18,624 20,409 10,416 16,449 10,785 13,445 2,273 1,138 93,538
2 Years of Recessionary losses 2,975 3,317 1,418 3,061 2,765 2,789 609 333 17,267
2 Years of Recessionary losses as % 2012e TE -16% -16% -14% -19% -26% -21% -27% -29% -18%
2 years of Recessionary losses as % of 2 years of PPP** 19% 16% 27% 21% 7% 18% 8% 16% 18%
* Loan loss provisions includes collections from previous periods
** Flexed down by 25%
Danger Map Indicators for Turkey Figure 167: Turkish banks : Danger Map
In Figure 167, we have summarized the Danger Map Indicators for Turkish
banks. We score nine indicator from 1 (less risky) to 5 (high risky), drawn from
% of Credit to GDP 1 Well below the European average
our expertise from the past crises.
Change in % of Credit to GDP 5 Credit to GDP has tripled since 2003, albeit
from a low level
On this basis, we rate Turkey at 22 out of 45, which located Turkish banks in Maturity of Cycle in Years 1 Year 1 of the recovery in the Turkish banks.
line with average for Emerging markets. Loan to GDP levels of Turkish banks
Credit Mix 2 A simple model still high share corporate loan
are well-below European banks and on this basis, increase in loan/GDP ratios with limited CoR
are above European averages (credit to GDP has tripled since 2003, albeit Unemployment 2 Unemployment broadly unchanged
from a low level). However, the recovery cycle of NPL cycle of Turkish banks Current account position 5
start just after the recovery starts as NPL additions declines and NPL Level of real interest rates 2
collections increases.

Turkish banks: The structure of the loan book Figure 169: Turkish banks – Loan breakdown by currency (1Q11)
The main drivers of NPL additions were SME and unsecured retail loans
(mainly credit card and general purpose loans), when asset quality records of 30%
38% 32% 31% 35% 36% 32% 35%
the corporate segment remained strong in 2008 and 2009. That is being said; 42%
system loans expanded at CAGR of 23% between 2008 and 1Q11, and were
driven by retail and SME segments. In the meantime, household debt/GDP
increased to 17.3% from 13.6% in 2008. As a result, 52% of total loans were 70% 68% 69% 65% 64% 68% 65%
62% 58%
composed of SME, credit card and consumer loans as of 1Q11 (2008: 48%).
On this basis, we think that those banks that are more exposed to these
segments could feel more pressure at the asset quality level in the event of
Akbank Garanti Halkbank Isbank Vakifbank Yapi Albaraka Bank DB
deterioration in global economic conditions. Kredi Turk Asya coverage
TRY FX
Figure 168: Turkish banks – Loan breakdown by segment (1Q11)
Source: Company data, Deutsche Bank
35% 39%
49% 53% 53% 53% 49% 48%
56%
24%
24% 9%
21% 9% 22%
19%
21% 30%
30% 36%
36% 23% 14% 38% 22%
12%
14% 14%
11% 11% 7% 6% 8%
2% 3%
Akbank Garanti Halkbank Isbank Vakifbank Yapi Kredi Albaraka Bank Asya DB
Turk coverage
Credit card SME Consumer Corporate
Source: Company data, Deutsche Bank
On the other hand, relative to its global counterparts, we think Turkish banks
would remain immune to another potential global crisis, on the back of; 1)
less-sophisticated balance sheets (i.e. no embedded derivatives), 2) Corporate
loans ・less risky relative to SME/consumer loans ・still capture 48% share in
the overall loan mix, and 3) Turkish banks are not allowed to lend in FX to
retailers and that partially limits the hike in NPL additions at times of
depreciation of TRY against foreign currencies. It should be noted that only
35% of loans are in FX in Turkey, which mainly represent loans extended to
the corporate segment (see Figure 169).
Indonesia Credit data trends

The table below summarises data from 2007-2010 and our projections for 2011-
12F for banks in our coverage. During the last financial crisis in 2008/2009, we
saw pick-up in Indonesian banks’ LLP from a low of 1.62% in 2007 to a high of
Overview 2.33% in 2008 and 2.39% in 2009. Some of these increases were reflecting
In this section, we look at Indonesian banks, with a key focus on major some loan write-off, which were more visible in some banks due to legacy loans
banks such as Mandiri, BCA, BRI and BNI. These represent our top picks (eg BNI and Mandiri). Given the economy’s strong recovery and performance,
in the Indonesian banking sector. Over the past four years, blended credit banks too benefit from rising income. Hence, despite generally still a high LLP
costs for the industry reached its peak of 239bps of average loans during charge of approx 2.1% of average loans, Indonesian banks’ PPP/LLP ratio
the recent global financial crisis in 2008/09. Still, we can argue that the improved to 3.9x from a low of 3.1-3.2x in 2008-09.
actual credit costs could be lower than stated numbers as some of these
Figure 170: Aggregation of quoted bank data (Rpbn), Indonesia
numbers were related to legacy loans (eg BNI and Mandiri).
Local Currency 2007 2008 2009 2010 2011E 2012E
We believe that peak credit costs to reflect recessionary risks could reach Net Interest Income 57,238 70,186 80,353 93,889 104,635 124,199
double than our existing forecasts. While this may appear to be low Other Operating Income 13,655 16,680 20,896 33,241 36,892 40,669
compared to some countries in the regions, we argue that this is Total Revenue 70,894 86,866 101,250 127,130 141,528 164,868
conservative considering that 1) actual new NPL formation is merely 1.5%
Costs 37,383 42,672 47,160 58,126 65,477 74,283
of average loans (which below our existing provision charges of 1.8% of
Pre-Provision Profits 33,565 44,040 53,885 69,061 76,116 90,772
loans); and 2) to begin with, Indonesian banks have coverage ratio of over
Loan Loss Provisions 7,413 13,731 17,177 17,640 17,945 23,737
100% (in some cases reaching as high as 321% for BCA) which can be
used to mitigate rising NPL formations. Pre-Tax Profit 26,464 30,705 37,190 51,491 58,460 67,230
Based on these analyses, we concluded that Indonesia banks stand to Total Assets 1,073,314 1,230,128 1,410,977 1,657,098 1,904,269 2,158,157
withstand rising recessionary risks (and in turn credit risks). We estimate Average Assets 988,855 1,151,721 1,320,553 1,534,037 1,780,683 2,031,213
that assuming recessionary losses, this would only reduce PPOP by 50% Risk weighted assets 528,265 706,698 844,300 1,057,118 1,204,819 1,538,460
- in other words Indonesian banks would still book ROAE of approx 10%. Basel 3 additions (if known) 0 0 0 0 0 0
Also, the country’s economy is largely under leveraged with loan to GDP Core tier one capital 96,150 103,449 131,004 162,773 206,092 244,255
of 28%. Basel 3 deductions (if known) 0 0 0 0 0 0
In “Danger Map” indicators for Indonesian banking sector; the country Risk Cushion above 7% plus 1 year 33,565 44,040 53,886 69,061 76,117 90,773
PPP
scores 19 out of 45; which indicates abilities to absorb risks of
Risk cushion to average loans 0 0 0 0 0 0
recessionary. Based on this, this indicates that Indonesian banking sector
is still attractive. Generally, we still see a number of years of robust loan Total Loans 509,050 670,110 768,337 932,465 1,148,348 1,422,936
growths. Average Loans 457,466 589,580 719,224 850,401 1,040,406 1,285,642

Revenue / Average Loans 15.5% 14.7% 14.1% 14.9% 13.6% 12.8%
Pre-Provision Profit / Average 7.3% 7.5% 7.5% 8.1% 7.3% 7.1%
Loans
Loan Loss Provisions / Average 1.62% 2.33% 2.39% 2.07% 1.72% 1.85%
Loans
PPP / Loan Loss Provision Cover 4.5 3.2 3.1 3.9 4.2 3.8
Source: Deutsche Bank; aggregation of company data

Estimates by banks Figure 172: Scoring the "Indonesian "heat map"

Table below summarises our forecast data for major banks that are in our top
Deregulation of Lending 2
picks. We show how severe recessionary loan losses (which for this purpose,
% of Credit to GDP 2 Loan to GDP of 28% is the lowest in the region
we assume a 2 time of “normal” provision charges) affect banks’ profitability. (vs Phils of 32%, Malaysia of 117% and SG of
115%)
We conclude that Indonesian banks have capacity to absorb higher provision
Change in % of Credit to GDP 2 Current loan/GDP at 28% may seem like a
charges in the case of recessionary losses. Amongst major banks; Mandiri and sluggish pace looking from 2005 loan/GDP of
BCA stand out in terms of ability to absorb higher provision charges. We also 25%. However, loans have almost tripled in size
since then. And the industry is still poised for
argue that Indonesian banks’ high coverage ratios suggest their ability to
loan growth of 20-25% per annum for the
mitigate rising credit costs. medium term
Figure 171: Summary of forecast data and computing recessionary loan Credit Mix 2 Banking system credit mix: 32% consumer loans
loss charges as % of equity and PPP and 68% corporate loans (48% investment loans
and 21% WC loans)
Local currency, 2012F Rpbn Mandiri BCA BRI BNI Total
Forecasts ave Loans 343,445 232,719 325,993 179,991 1,082,148 Unemployment 1 Current unemployment level at 7%, down from a
Forecast Loan Loss Provisions 4,293 2,327 8,802 4,050 19,472 high of 11% in 2005
as % of ave loans 1.25 1.00 2.70 2.25 1.80
Severe Recession Loan Loss Provisions (2x) 8,586 4,654 17,604 8,100 38,944 Current account position 2 Current account to GDP is below 1%. Current
2012F PPOP 20,541 16,086 28,895 12,702 78,224 account deficit is at USD1.9bn
2012F PPOP less Recessecionary Losses 11,955 11,432 11,291 4,602 39,280 Level of real interest rates 2 ID's real interest rate is 2.2%, July inflation has
eased to 4.6% YoY (vs 5.5% in Jun) and Central
2012F Total equity 71,041 48,267 61,456 44,300 225,065
Bank has maintained the BI rate at 6.75% since
2 years of Recessionary losses 8,586 4,654 17,604 8,100 38,944
2 years of Recessionary losses as % of 2012F total equity 12.1 9.6 28.6 18.3 17
February
2 years of Recessionary losses as % of 2012F PPOP 41.8 28.9 60.9 63.8 49.8 Exchange rate flexibility 3 IDR to USD has strengthen 5.8% YTD
For Indonesian banks; given their already high credit cost (hence high coverage); we only assume double in credit costs
Danger Map Indicators for Indonesia Understanding Indonesian banks’ loan books
Below we summarise our “Danger Map” indicators for Indonesian banking
Generally, Indonesia banking sector has a relatively young credit history.
sector. Overall, we score 19 out of 45; which is derived based on a score
Overall loan portfolios are still tilted towards corporate segments (including
between 1 to 5 across a number of risk metrics, drawn from our analyses of
SME and/or commercials) – accounting for approx 70% of total loan books. In
case studies and our global experiences of bank lending. Based on this, this
recent cases, investments had had fast growth of approx 27% yoy; largely
indicates that Indonesian banking sector is still attractive. Generally, we still
reflecting the country’s rising investments – necessary for supply side
see a number of years of robust loan growths. The government’s fiscal
expansions. This suggests low inflations risks arising from surging consumer
discipline would also likely keep overall loan to GDP at below 30%.
loans. The two pie charts below, we show major banks’ typical loan
breakdown. It is consistent that consumer loans representing abt 25% for
BCA and 14% for Mandiri.

Figure 173: Indonesian banking sector – System loan breakdown

100%
90%
WC Invt Cons
80%
70%
60%
50%
40%
30%
20%
10%
0%
Sep-01
Sep-02
Sep-03
Sep-04
Sep-05
Sep-06
Sep-07
Sep-08
Sep-09
Sep-10
Mar-01
Mar-02
Mar-03
Mar-04
Mar-05
Mar-06
Mar-07
Mar-08
Mar-09
Mar-10
Mar-11
Source: Deutsche Bank and Bank Indonesia
Figure 174: Mandiri – Loan mix – Tilted towards corps Figure 175: BCA – Loan mix – Tilted towards corps
Consumer,
34.3, 14%
Consumer
, 39.9 , Corporate
Corporate, 25% , 54.2 ,
103.1, 43% 34%
2Q11 BBCA 2Q11
Commercial
(incl
SME/Micro),
105.1, 43%
SME,
65.7 , 41%
Source: Deutsche Bank and Mandiri
Source: Deutsche Bank and BCA

Malaysia
Figure 176: Aggregation of quoted bank data, local currency (RMm),
Malaysia
Local Currency 2006 2007 2008 2009 2010
Total Revenue 26,183 30,452 31,926 36,248 41,206
In this section we look at the Malaysian banks, with a focus on those
Costs 11,223 13,320 14,065 17,441 19,096
covered by DB (AMMB, CIMB, Hong Leong Bank, Maybank, Public Bank
and RHB Capital). Over the last five years, we find that credit loan loss Pre-Provision Profits 14,960 17,132 17,862 18,807 22,111
provisions in the banking book averaged 81bp, and peaked at 102bp. Loan Loss Provisions 4,125 4,780 3,538 4,771 3,919
Pre-Tax Profit 10,758 12,962 13,808 12,696 18,565
Assuming peak provisions on a further macro downturn would reach 4x
their normalised level, we estimate credit losses of RM24.7 for the five
Total Assets 765,660 869,185 937,195 1,052,562 1,142,905
Malaysian banks per annum. Over two years, this would be equivalent to
43% of 2012 closing tangible book value. If we include PPP estimates as Average Assets 696,249 817,422 903,190 994,878 1,097,734
an offset (flexed down consistent with the 2007-2009 experience) then Total Loans 450,226 489,555 567,675 643,880 725,875
the loss would be 7% of TBV, i.e. the banks would be slightly loss- Average Loans 419,188 469,890 528,615 605,778 684,878
making. Revenue / Average Loans 6.2% 6.5% 6.0% 6.0% 6.0%
As regards the Outlook and the Danger Map for Malaysia, we score
Malaysia overall at 20 out of 45. This captures a 1 to 5 score across a Loan Loss Provisions / Average Loans 0.98% 1.02% 0.67% 0.79% 0.57%
number of risk metrics, drawn from our analyses of case studies, and our PPP / Loan Loss Provision Cover 3.6 3.6 5.0 3.9 5.6
global experiences of bank lending.
Estimates by Bank
Credit Data: Trends
Below we summarise forecast data for 2012 for the aggregated Malaysia
Below we summarise data from 2005 to 2010 for the aggregated French
banks in our coverage universe (AMMB, CIMB, Hong Leong Bank, Maybank,
Public Bank, RHB Capital). We then show our estimate of severe recessionary
credit cycle was not that severe, declining PPP combines with rising loan
loan losses, which for the purposes of this exercise we have taken as 4x
losses to reduce PPP/LLP cover from 12.3x to 1.5x.
normalised.

provision profit flexed down by 25%. These are simple numbers, but they do
show in our view that (1) the Malaysian banks have good pre-provision profit
generation, but that (2) in a severe credit book downturn, we would expect a
breakeven or worse result, in aggregate.

Figure 177: Summary of forecast data and comparing recessionary loan Figure 178: Scoring the Malaysia “danger map”
loss charges as % of opening TBV and of non-flexed PPP Danger Factor Score Comments
Local Currency, 2012E AMMB CIMB Hong Malayan Public RHBC Total Deregulation of Lending 1 No recent deregulation of lending
Holdings Group Leong Banking Bank
% of Credit to GDP 4 Low by global standards, although high relative to
Bank
other emerging markets
Forecast Average Loan Balances 75,480 196,179 46,551 279,462 191,845106,021 895,537
Change in % of Credit to GDP 3 Leverage rising rapidly again
Forecast Loan Loss Provisions 409 453 252 1,013 902 576 3,605
Maturity of Cycle in Years 2 Year 3 of the GDP growth recovery. Inflation the key
Severe Recession Loan Loss 2,083 5,415 1,285 7,713 5,295 2,926 24,717 concern although currently benign
Provisions
Credit Mix 2 Dominated by mortgages, auto financing and working
2012 Pre-Provision Profit 2,595 6,336 1,538 8,111 6,169 3,207 27,955 capital funding
2012 PPP less Recessionary Losses 512 921 253 398 874 281 3,238 Unemployment 2 Unemployment broadly unchanged
Current account position 1 Strong surplus
2012 Forecast Tangible Equity 11,423 27,548 8,119 35,791 16,764 14,151 113,795 Level of real interest rates 2 Low
2 Years of Recessionary losses 4,166 10,829 2,570 15,426 10,590 5,852 49,434 Exchange rate flexibility 3 Undervalued currency
2 Years of Recessionary losses as % -36% -39% -32% -43% -63% -41% -43% Total Danger Map Score (out of 30) 20
2012 Forecast Tangible Equity Source: Deutsche Bank
2 years of Recessionary losses as % -2% -5% -3% -9% -8% -7% -7%
of 2 years of PPP (flexed down by Household debt levels in Malaysia are at the top end of the range relative to
25%)
regional peers (Figure 179), and mortgage loan to valuation ratios are also at
the high end (Figure 180). However, we note that:
Danger Map Indicators for Malaysia Like-for-like household debt comparisons are not easy to make and in
Below we summarise our Danger Map Indicators for Malaysia. We score Malaysia’s case, substantial pension fund savings provide a material
Malaysia overall at 20 out of 45. This captures a 1 to 5 score across a number offset. If we net EPF balances from Malaysian household debt, Malaysia
of risk metrics, drawn from our analyses of case studies, and our global would be at the other end of the range in Figure 179.
experiences of bank lending. We rate Malaysia as lower risk on each of the High mortgage LTV’s are a little more concerning, although from a
key categories. The one area of risk we note is credit to GDP, where we are nationwide perspective, percentage house price appreciation has been in
mindful of consumer leverage and mortgage LTV ratios, as discussed below. the mid/high single digit levels for many years, which is not suggestive of
bubble conditions.
So there are some offsets which suggest a situation less negative than the
headlines indicate. But clearly given weakening macro conditions the potential
for credit quality deterioration is an increasingly relevant issue.

Figure 179: Household debt to GDP Understanding the Malaysian banks’ loan books
80 %
As discussed above, we regard Malaysian banks as having a moderate level of
70 risk in their loan portfolios. Balances are dominated by three key categories;
60 residential lending, auto financing and working capital funding.
50
In terms of residential property (24% aggregate loans) and auto financing
40 (17%), the recent challenge for banks has been competitive pricing which has
30 driven down product profitability. Clearly the problem would be exacerbated
by a credit cycle. As noted above there are reasons to be cautious on this
20
front although at this stage there are no signs of deterioration.
10
0 Working capital funding (18% aggregate loans) is largely driven by the SME
sector which is likely to see more substantial credit quality deterioration in the
MY SK TW HK SG CH TH ID IN PH
event of a macro slowdown. Given Malaysia’s economy is highly export
Source: Deutsche Bank, CEIC. Data is last available.
dependent the SME sector is vulnerable to shifting global trends.
Figure 180: Mortgage LTV ratios Figure 181: Aggregate loan balances of RM655bn
90
100%
80 90%
System mortgage LTV ratio (%)
70 80%
60 70%
50 60%
40 50%
30
40%
30%
20
20%
10
10%
0
0%
Singapore
Thailand
Australia
China
Malaysia
Indonesia
Korea
HK
Taiwan
AMMB CIMB Hong Leong Maybank Public Bank RHB
Residential property Auto financing Working Capital

Source: Deutsche Bank, Company data
Non-residential property Other personal Other
Source: Deutsche Bank, Company data. Data is latest available

Thailand Credit Data: Trends

Below we summarized data from 2006 to 2010 for Thai banks under our
coverage. Credit costs surged during 2006-2007 but this was mainly due to
regulatory changes that required banks to book 100% of uncovered NPLs.
Overview During the 2009 recession, the asset quality of Thai banks remained strong,
In this section we look at Thai banks, focusing on the seven banks with NPL ratios and credit costs continuing to decline while LLR/NPL ratios
covered by DB (Bangkok Bank, Bank of Ayudhya, Kasikornbank, Krung strengthened. This reflects improved risk management at banks as well as the
Thai Bank, Siam Commercial Bank, TMB Bank and Thanachart Capital). financial strength of Thai corporates. Moreover, pre-provisioning profit rose,
Over the past 10 years, credit loan loss provisions in the Thai banking led by a solid recovery in non-interest income amid improved cost efficiency.
system averaged 84bp, with a peak of around 170bp. The PPP/ loan loss provision coverage ratio jumped to 4.7x in 2010 from only
1.5x in 2007.
Despite the recent slowdown in the global economy, we believe that the
asset quality of Thai banks will remain resilient due to improved risk
Figure 182: Aggregation of quoted bank data, local currency
management systems that allow banks to detect problem loans before
Local Currency 2006 2007 2008 2009 2010
they become NPLs and help clients. Another key factor is the financial
strength of Thai corporates following a prolonged de-leveraging cycle
after the 1997 crisis (the industry’s net D/E dropped to 1.0 x in 2010 vs.
more than 3x in 1997). Even during the 2009 economic contraction, Thai Total Revenue 259,962 275,668 305,014 319,246 377,945
banks were able to reduce their NPL ratios and reduce credit costs. Costs 130,711 155,119 150,242 159,091 185,223
To gauge the potential impact from a new domestic recession, we
Loan Loss Provisions 64,167 80,476 46,733 44,222 40,918
assumed the credit costs of Thai banks increased 3x from our current
Pre-Tax Profit 65,085 40,073 108,039 115,933 151,804
estimate (i.e. around the same level as their peak in 2007). We estimate
two years of high provisioning levels could cut 25% of the industry’s
closing tangible book value in 2012F. Based on our scenario analysis, Thai Total Assets 6,367,671 6,561,537 7,294,279 7,760,496 9,081,442
banks’ equity is still sufficient to cover a severe recession. Average Assets 6,084,432 6,464,604 6,927,908 7,527,387 8,420,969
Total Loans 2,863,286 2,969,773 3,356,232 3,354,787 3,965,959
We believe a new credit cycle has already started in Thailand, led by
Average Loans 2,791,280 2,916,530 3,163,003 3,355,509 3,660,373
strong domestic demand, a recovery in private investment and solid
consumption. This in turn has been supported by government stimulus Revenue / Average Loans 9.3% 9.5% 9.6% 9.5% 10.3%
programs and rising farm commodity prices. Retail loans have led the Pre-Provision Profit / Average Loans 4.6% 4.1% 4.9% 4.8% 5.3%
recovery since 2008, especially for loans with collateral backing, i.e. Loan Loss Provisions / Average Loans 2.30% 2.76% 1.48% 1.32% 1.12%
mortgage or car loans. Loans for business began picking up in mid-2010 PPP / Loan Loss Provision Cover 2.0 1.5 3.3 3.6 4.7
and we expect this trend to continue over the next few years. Source: Deutsche Bank aggregation of company data
As regards the Outlook and Danger Map for Thailand, we gave Thai banks
an overall score of 16 out of 45. This captures a 1 to 5 score across a
number of risk metrics, drawn from our analyses of case studies and our
global experiences of bank lending. This puts Thai banks at the lower end
of the pack on the Danger Map.

Estimates by Bank Danger Map Indicators for Thailand

To gauge the potential impact from a severe recession scenario, we assumed Below we summarize our Danger Map Indicators for Thailand. We gave Thai
Thai banks’ credit costs would increase 3x from our current estimate (around banks an overall score of 16 out of 45 which captures a 1 to 5 score across a
the same level as the peak cycle in 2007). We estimate two years of high number of risk metrics, drawn from our analyses of case studies and our
provisioning levels could cut 25% of the industry’s closing tangible book value global experiences of bank lending. We rate Thai banks as low on deregulation
in 2012. Based on our scenario analysis, Thai banks’ equity is still sufficient to and current account position.
cover a severe recession.
Figure 184: Scoring the Thailand “danger map”
We then expressed this as a percentage of 2012 BV and also as a percentage Danger Factor Score Comment
of 2012 pre-provision profit. In this assumed recession scenario where credit Deregulation of Lending 1
costs tripled from our normalized estimate, Thai banks’ pre-provision operating
% of Credit to GDP 2
profit should be sufficient to cover the shortfall even if we assume another
Change in % of Credit to GDP 1 Increased by 8ppts from 2007 to 101% in 2010
25% downside in our PPP forecast (assuming flat loan growth and 20-30bps
decline in NIM as seen during 2009 given other factors remaining constant).
Credit Mix 2
Although these are simple numbers (see Fig. 2), we believe they demonstrate
Unemployment 2
that Thai banks can generate solid pre-provision profit that can mitigate two
years of high provisioning levels. Current account position 1
Figure 183: Summary of forecast data and comparison of loan loss costs Exchange rate flexibility 3
as % of opening TBV and non-flexed PPP Total Danger Map Score (out of 45) 16
Local Currency, 2012E BBL BAY KBANK KTB TMB TCAP SCB Total
Forecast Average Loan 1,635,099 812,725 1,406,982 1,596,207 438,348 725,752 1,454,578 8,069,691
Balances Understanding Thai banks’ loan books
Forecast Loan Loss 6,200 11,040 8,000 6,800 2,000 3,266 4,400 41,706
Provisions Thailand is a manufacturing-led economy (85% of GDP) with exports
Severe Recession Loan Loss 18,600 33,120 24,000 20,400 6,000 9,798 13,200 125,118 accounting for 60-70% of GDP. As a result, the loan books of Thai commercial
Provisions
banks have the biggest exposure to the manufacturing sector (40% of total
2012 Pre-Provision Profit 54,136 33,187 56,838 40,517 8,029 16,940 58,386 268,033
loans). Loans for public utilities and housing are each about 16-17% while
2012 PPP less Recessionary 35,536 67 32,838 20,117 2,029 7,142 45,186 142,915
Losses loans for agriculture accounted for the smallest portion at only 2-3% of total
loans. In terms of customer exposure, corporate and SMEs loans accounted
2012 Forecast Tangible 262,460 114,413 174,254 145,973 55,636 41,037 210,105 1,003,876 for more than 60% of total loans by commercial banks. The remainder was
Equity
retail loans (15% housing, 11% auto and 6% personal loans).
2 Years of Recessionary 37,200 66,240 48,000 40,800 12,000 19,595 26,400 250,235
losses
2 Years of Recessionary 14% 58% 28% 28% 22% 48% 13% 25%
losses as % 2012 Forecast
Tangible Equity
2 years of Recessionary 61% 177% 75% 90% 133% 103% 40% 83%
losses as % of 2 years of PPP
flexed down by 25%

Figure 185: Loan breakdown by business Government banks play an important role in Thailand’s economic growth as
Constructi Agriculture their loan books account for more than 25% of total system lending.
2%
on Customers of state-owned banks are mainly from lower-income groups and
8%
therefore are not targeted by commercial banks. Over the past few years the
Housing Manufactu
loans ring
government has encouraged state-owned banks to lend more to such groups
16% 39% in order to boost the economy, particularly in upcountry areas.
During the global financial crisis in 2008-09, loans for the retail segment,
Public
Others mainly for mortgages and auto loans, remained resilient, driven by healthy
utilities
16%
19% domestic demand and government stimulus programs. This helped mitigate a
slowdown in demand by corporates. If the global economy enters a new
recession, asset quality for Thai retail customers should be manageable given
that banks were quite stringent in credit screening criteria and mostly lent
Figure 186: Loan breakdown by customer
against collateral (houses, vehicles, etc).
Other Gov
Consumer 4% 4%
6% Growth in corporate and SME loan has seen an improvement since mid-2010.
Corporate Momentum remains strong in this segment driven by the sharp recovery in
Auto 36% the export sector, solid growth in private investment and government
11%
economic stimulus programs. We believe the credit cycle has already started
in Thailand and this new investment phase should allow healthy loan growth
from the corporate and SME segments to continue over the next few years. In
Housing
15%
SME our view, the risk profile of corporate and SME clients has seen a drastic
24%
improvement thanks to healthy balance sheets as a result of a prolonged de-
Source: Deutsche Bank aggregation of company data leveraging cycle following the 1997 crisis. Also, banks have heeded the
lessons of 1997 and now monitor concentration risk very closely.
Figure 187: Loan market share for banks in Thailand
2 0 1 0 L oan Mark e t Sh are , %
Thai commercial banks SFIs (State-owned banks) Foreign bank
7 6 5
22 25 27
72 69 68
2008 2009 2010

Source: Bank of Thailand

South Korea
the GFC at only 0.5% in 2009 and 3.3% in 2010 as bank management’s
focus remained on corporate restructuring and removing potential NPLs
rather than growth, and 4) sector credit cost looks to have stabilized at the
moment as we estimate it to be 77bp in 2011, a big decline from 121bp
Overview
in 2010.
In this section we look at Korean banks, focusing on the top eight listed As for the Danger Map for South Korea, we score South Korea overall at
banks covered by DB (KB, Shinhan, Woori, Hana, IBK, KEB, BS, DGB). 22 out of 45. This captures a 1 to 5 score across a number of risk metrics,
Over the last five years (2006-2010), the average credit cost for the banks drawn from our analyses of case studies, and our global experiences of
averaged 88bp, peaking at 121bp in 2010. bank lending. This puts South Korea in-line with the average for Emerging
Assuming we face a worse than expected global macro impacting the markets.
South Korean economy, we believe credit cost could increase to about 2x
the average level of the past five years. We estimate credit losses of Credit Data: Trends
KRW18tr for our coverage banks per annum under this scenario, and over Below we summarize data from 2005 to 2010 for the aggregated Korean
two years this would be equivalent to 29% of 2012 closing tangible book banks in our coverage universe. During the last crisis the PPP/LLP cover
value. Including PPP estimates as an offset (flexed down consistent with declined rapidly as PPP declined on lower market rates given the asset
the 2007-2009 experience) then the PPP should actually be enough to sensitive nature of the Korean banks following big interest rate cuts by the
cover all the credit losses. central bank, while LLP increased due to rising bankruptcies.
Our assumptions may look optimistic to some given the South Korean
Figure 188: Aggregation of quoted bank data, local currency (KRW bn),
economy’s large dependence on global trade, but our key reasoning lies
on the corporate restructuring activity that was implemented in 2008 and South Korea
has been repeated every year since, most recently in 2Q 2011. When the Local Currency 2006 2007 2008 2009 2010
GFC came, the Korean government and the banks worked together to Net Interest Income 23,361 26,708 30,159 28,034 32,009
reduce not only actual but potential NPLs in the system since late 2008 by Other Operating Income 3,444 6,099 3,664 4,069 5,550
launching an extensive corporate restructuring program, with the goal of Total Revenue 26,804 32,807 33,823 32,103 37,559
preemptively defining corporate borrowers that may face financial Costs 12,716 15,090 15,764 15,716 16,768
troubles in the following years. Additional financial support was provided Pre-Provision Profits 14,089 17,717 18,059 16,387 20,791
for companies experiencing short-term liquidity problems but with good Loan Loss Provisions 2,302 3,608 8,665 10,061 10,836
business models, problem companies were pushed to go through
Pre-Tax Profit 13,551 16,488 10,124 6,267 9,710
corporate restructuring programs, and finally companies unlikely to
survive in the next few years were liquidated. As a result, credit costs
Total Assets 895,762 1,040,724 1,248,894 1,226,199 1,267,734
remained relatively high during 2008-2010 at average 116bps, but this
was relatively well contained when compared to the 218bp seen during Average Assets 833,395 968,243 1,144,809 1,237,547 1,246,966
the Credit Card Crisis. Total Loans 623,799 722,468 839,503 843,700 871,658
Average Loans 574,687 673,134 780,986 841,602 857,679
Other reasons for our relatively low projection for recession type
provisioning are; 1) the government’s policy reaction and the banks’ risk Revenue / Average Loans 4.7% 4.9% 4.3% 3.8% 4.4%
management has improved after two large financial crises – the Asian Pre-Provision Profit / Average Loans 2.5% 2.6% 2.3% 1.9% 2.4%
Financial Crisis (1998) and the Credit Card Crisis (2003), 2) government Loan Loss Provisions / Average Loans 0.40% 0.54% 1.11% 1.20% 1.26%
has ample room to provide fiscal support should it feel the need to PPP / Loan Loss Provision Cover 6.1 4.9 2.1 1.6 1.9
intervene in the markets, 3) system loan growth has been very slow since Source: Deutsche Bank aggregation of company data

Estimates by bank global experiences of bank lending. We rate South Korea as relatively high in
terms of % of debt to GDP and medium for the credit mix and stage of cycle.
Below we summarize forecast data for 2012 for the Korean banks in our This would put South Korea in-line with the average for Emerging markets.
coverage universe. We then show our estimate of severe recessionary loan
losses, which for this exercise we have assumed as 2x the average of past Figure 190: Scoring the Korean “danger map”
five years. We then express this as a % of 2012 BV, and also as a % of 2012 Danger Factor Score Comment
pre-provision profit flexed down by 25% (we saw PPP fall by 9% during the
2008 credit cycle but assume a more conservative scenario). If we flexed it
% of Credit to GDP 4 Has started coming down since 2009 but
down by 9% this would result in average +7% of 2012 forecast tangible still high vs. Asian average
equity for the banks listed in the table below. Change in % of Credit to GDP 3 Rapid increase during 2005-2008 but has
started to come down since
Figure 189: Summary of forecast data and comparing recessionary loan Maturity of Cycle in Years 3 Year 3 of the recovery after GFC
loss charges as % of opening TBV and of non-flexed PPP Credit Mix 3 Balanced loan portfolio between household
and corporate
Local Currency, 2012E BS DGB Hana IBK KB KEB Shinhan Woori
Unemployment 1 Remains low at 3.3% and employment
Forecast Average Loan numbers on the uptrend
27,682 24,683 126,760 142,581 216,156 70,832 196,592 239,614
Balances
Forecast Loan Loss
120 132 560 1,120 1,300 520 920 1,600 Level of real interest rates 2
Provisions
Severe Recession Loan Loss Exchange rate flexibility 3
487 434 2,231 2,509 3,804 1,247 3,460 4,217
Provisions Total Danger Map Score (out of 45) 22
2012 Pre-Provision Profit 692 638 2,651 3,687 5,326 1,847 5,656 4,782 Source: Deutsche Bank
2012 PPP less Recessionary

205 204 420 1,178 1,521 600 2,196 565
Losses Understanding the Korean banks’ loan books
As mentioned above we believe credit risks in the Korean banks’ loan books
2012 Forecast Tangible
3,104 2,590 16,502 14,592 26,312 10,710 29,869 23,010 are not too high. Household loan delinquency ratio has remained low at below
Equity
2 Years of Recessionary 1% since 2006 and has remained steady even during the GFC, while corporate
974 869 4,462 5,019 7,609 2,493 6,920 8,434
losses restructuring and bad asset disposals and better lending discipline have
2 Years of Recessionary helped improve asset quality. About 50% of the loans are collateralized, with
losses as % 2012 Forecast -31% -34% -27% -34% -29% -23% -23% -37% most of the collateral being real estate – given that house price is only up by
Tangible Equity
7% during the past five years, we believe the collaterals are relatively solid.
2 years of Recessionary
losses as % of 2 years of 2% 3% -3% 4% 1% 3% 5% -5%
PPP (flexed down by 25%) The weakness of the loan books would be the 44% SME loans, as the
segment remains naturally sensitive to macro slowdowns. In addition,
although household loan asset quality looks stable and shows no signs of
Danger Map Indicators for South Korea deterioration, 70-80% of the loans are floating rate loans and credit risks may
rise if interest rates remain on the uptrend.
Below we summarize our Danger Map Indicators for South Korea. We score
South Korea overall at 22 out of 45. This captures a 1 to 5 score across a
number of risk metrics, drawn from our analyses of case studies, and our

Figure 191: Loan breakdown

Other Large corp
3% 10%
Household
43%
SME
44%
Source: Deutsche Bank, FSS
Figure 192: Loan breakdown by collateral
Credit
43% Collateral
49%
Guarantee
8%
Source: Deutsche Bank, FSS

Poland
Figure 193: Aggregation of quoted bank data, local currency (PLN m),
Poland
Local Currency (PLNm) 2006 2007 2008 2009 2010 2011E 2012E
Overview Other Operating Income 11,100 13,361 14,648 14,930 14,462 14,691 15,792
Total Revenue 22,803 27,494 33,676 32,590 35,168 38,511 43,021
In this section we look at the Polish banks, with a focus on those covered Costs 13,720 15,185 17,374 16,871 17,155 18,474 19,615
by Deutsche Bank. Over the last five years, we find that loan loss Pre-Provision Profits 9,083 12,309 16,302 15,719 18,013 20,036 23,406
provisions in the banking book averaged 80bps, and peaked at 170bps Loan Loss Provisions 60 260 2,862 6,725 5,711 5,212 5,396
(2009). The minimum level was reached in 2006, when banks had Pre-Tax Profit 9,169 12,247 13,741 9,113 12,415 15,727 18,183
experienced balanced additions and write-backs (4 bps net charge).
Total Assets 387,785 497,561 636,804 639,237 686,194 741,380 804,288
We use the assumption that peak provisions on a further downturn would
Average Assets 353,638 429,109 547,817 618,482 641,029 690,622 748,546
likely reach 4x their normalized level. We estimate credit losses of PLN
Total Loans 188,242 282,739 390,724 409,156 446,050 505,034 567,474
21.6bn for the Polish banks per annum, in the severe scenario, i.e. 4x
Average Loans 168,125 235,490 336,732 399,940 427,603 475,542 536,254
more than our 2012E current credit losses of PLN 5.4bn estimate. Over
two years, this would be equivalent to 48% of 2012 closing tangible book
value (TBV). If we include pre-provision profit (PPP) estimates as an offset
(flexed down consistently with the crisis experience), then the loss would
be 9% of TBV, i.e. the banks would be in the loss-making territory. Source: Deutsche Bank aggregation of company data
Despite this severe scenario, we estimate that the sector now possesses
PLN 23.4bn risk cushion above CT1 7% ratio if one year PPP is included. Estimates by Bank
Moreover, we see the whole sector enjoying backstops from parent
companies from Western Europe or Polish State (with balance sheet big Below we summarize forecast data for 2012 for the aggregated Polish banks
enough to accommodate potential recapitalization needs). in our coverage universe (PKO BP, Pekao and 5 smaller players). We then
show our estimate of severe recessionary loan losses, which for the purpose
As regards the Danger Map, we score Poland overall at 23 out of 45. This of this exercise we have taken as 4x normalized.
captures a 1 to 5 score across a number of risk metrics. This puts Polish
banks in the middle of the pack on the Danger Map – neither angry red, We then express this as a percentage of 2012 book value (BV), and also as a
nor soothing green. percentage of 2012 PPP flexed down by 25% (very severe stress test for
Polish market as we never saw 25% contraction of PPP in the last 15 years),
Credit Data: Trends as a proxy of a potential contraction during a crisis. We estimate credit losses
Below we summarize data from 2006 to 2010 for the aggregated Polish banks of PLN 21.6bn for Polish banks per annum. Over two years, this would be
in our coverage universe. PPP/LLP cover ratio deteriorated to 2.3x in the worst equivalent to 48% of 2012 closing TBV. If we compare severe loan losses to
2009 period, before improving to 3.2x last year and what we forecast to be PPP flexed down by 25%, we see 123%, i.e. loss-making banking sector on a
3.8x this year. All those ratios mark a deterioration from extremely healthy pre-tax level; however, we would point to differentiation as 4 banks out of 8
coverage in 2008 (5.7x) – not even mentioning even better period 2006-7 – but would still retain profitability.
we are overall satisfied with fairly manageable development of the recent
crisis in Poland.

Figure 194: Summary of forecast data and comparing recessionary loan Figure 195: Scoring the Polish “danger map”
loss charges as % of opening TBV and of flexed PPP Danger Factor Score Comment
PLNm, 2012E PKO PEO BHW BZW BRE GTN ING KRB Total Deregulation of Lending 2 No recent deregulation of lending
Forecast Average Loan Balances 156,097 95,401 15,935 39,051 70,909 44,987 40,601 30,428 536,254 % of Credit to GDP 1 Much below the European average
Forecast Loan Loss Provisions 1,932 703 140 366 630 770 252 317 5,396 Change in % of Credit to GDP 4 Above the European average
Severe Recession Loan Loss 7,727 2,811 558 1,464 2,520 3,079 1,009 1,267 21,582
Maturity of Cycle in Years 2 Year 2 of the recovery in the euro zone.
Provisions
2012 Pre-Provision Profit 7,896 4,616 1,278 2,362 2,002 1,945 1,598 777 23,406 Credit Mix 2 Mortgage led
2012 PPP less Recessionary Losses 169 1,805 720 898 -518 -1,134 589 -490 1,824 Unemployment 3 Unemployment broadly unchanged but high
2012 Forecast Tangible Equity 26,530 22,503 5,422 8,045 8,439 4,998 6,845 3,272 90,831 Level of real interest rates 3
2 Years of Recessionary losses 15,453 5,622 1,116 2,927 5,040 6,157 2,017 2,534 43,165 Exchange rate flexibility 2
2 Years of Recessionary losses as % Total Danger Map Score (out of 45) 23
2012 Forecast Tangible Equity 58% 25% 21% 36% 60% 123% 29% 77% 48% Source: Deutsche Bank
2 years of Recessionary losses as %
of 2 years of PPP (flexed down by
25%)
130% 81% 58% 83% 168% 211% 84% 217% 123% Understanding the Polish banks’ loan books
In Poland, we observed fairly strong credit extension before the end-2008
Danger Map Indicators for Poland crisis. FX adjusted retail loan growth peaked at c. 40% during late 2007 / early
2008, mainly on the back of mortgage boom. One risky element in Poland is
Below we summarize our Danger Map Indicators for the Polish market. We that significant part of mortgage production happened in CHF, and even
score Poland overall at 23 out of 45. This captures a 1 to 5 score across a though the sector switched predominantly to PLN mortgage since post-
number of risk metrics, drawn from our analyses of case studies, and our Lehman, the FX part of the mortgage book still makes up 62% of total (1H11).
global experiences of bank lending. This puts Poland in the middle of the pack
on the Danger Map – neither bright red, not soothing green. The risk factors in In the last months retail lending stabilized at c. 10% pa growth runrate, while
Poland are faster than average pace of credit extension heading into the crisis corporate business turned from deleveraging into growth early this year.
of 2009 as well as persistently high unemployment. The mitigating factor is Polish corporate lending is extremely underdeveloped as corporate loan to
very low credit penetration, especially for corporate and mortgage loans, as GDP ratio of 15% is the lowest in the EU. We continue to find strong
well as collateralized nature of lending (mortgage-led), tight market regulation fundamental case for further developments in Polish corporate and mortgage
and resilient macroeconomic backdrop in Poland. lending (penetration for mortgage lending is c. 18% of GDP) and fairly limited
prospects for further consumer loan sales. Consumer lending book overall is
quite sizeable and historically well developed and in this segment we see shift
from POS lending types into card / overdraft lending, typical for the market
moving from nascent into gradually maturing. Barring the current stress in the
marketplace from potential slowdown fears, we see good prospects for Polish
lending to grow at 10%+ rate in the years to come. Moreover, Polish banks
L/D ratio is 110% with the bulk of non-deposit funding gathered from parent
banks and spent for mortgage funding. The withdrawal of parent funding since
end-2008 was not observed.

Figure 196: Polish corporate and retail loan growth
50%
Corporate
40%
Retail
30%
20%
10%
0% Nov-01
Nov-02
Nov-03
Nov-04
Nov-05
Nov-06
Nov-07
Nov-08
Nov-09
Nov-10
Dec-00
-10%
Source: NBP, Deutsche Bank
Figure 197: Polish sector loan split
corporate housing
32% 39%
consumer
29%
Source: NBP, Deutsche Bank

Crisis case studies

The Latin American Debt

The reason the large US and UK banks did not reserve against the problem in
the early years of the crisis is that they could not afford to. Problem country
exposure was over two times the capital and loan loss reserves of the money
Crisis center banks and of two of the UK banks. A 40% provision would have wiped
out their capital. US and UK regulators were pragmatic and did not require loan
loss provisions either against restructured loans or against interest arrears.
“Countries don’t go out of business . . .” Regulatory forbearance gave the banks time to build up capital and loan loss
reserves. It also allowed them to finance the debtor countries’ interest
The IMF has recorded 63 sovereign debt crises since 1970. The last major and payments, so throughout the crisis there was no technical default, since the
protracted sovereign debt crisis which directly threatened the solvency of banks simply accrued the interest by granting “new money”. During this
major European and US financial institutions was the Latin American debt period the IMF and other supranational agencies provided more new money
crisis in the 1980’s. The rough chronology to the onset of the crisis was that and thus the percentage of total lending to indebted countries by private
Mexico defaulted in 1982 on an unmanageable debt burden built up in the mid sector lending institutions shrank. A similar pattern is evident today as new
to late 1970s as international banks “recycled” the OPEC surpluses; other money bailouts dilute commercial banks exposures.
Latin American countries defaulted soon after and by 1983 there were 10
countries in default with outstanding loans of around US$ 200bn. The Latin American debt crisis was resolved ultimately from 1989, through the
plan devised by the US Treasury Secretary Nicholas Brady. The Brady Plan
US and European banks initially denied there was a problem. Citicorp’s sought permanent reductions in principal and interest payments for the debtor
Chairman Walter Wriston explained: “Countries don’t go out of business . . . nations and in return the countries which signed up to the plan had to agree to
The infrastructure doesn’t go away, the productivity of the people doesn’t go introduce economic reforms to promote growth. The Brady Plan could not
away, the natural resources don’t go away. And so their assets always exceed have been implemented had the banks not written down their exposures. The
their liabilities, which is the technical reason for bankruptcy. And that’s very banks swapped their exposures for instruments collateralized by US
different from a company.“ Citicorp was one of the largest lenders to Latin government zero coupon bonds, which could be carried at par with certainty
American countries and had substantial local operations in Latin American by the creditor banks.
countries.
An FDIC study of the crisis suggested that forgiveness of existing debts via
In contradiction to Walter Wriston’s assertion, a secondary market in Latin debt and debt for equity swaps by private lenders between 1989 and 1994
American debt developed and the market discount to balance sheet par values amounted to approximately US$61bn (out of US$191bn of outstanding loans).
showed the requirement for very significant provisions against sovereign These losses were mostly taken by bank shareholders. As a benchmark, the
loans. A similar tension exists today between the market value of some total capital of the 8 largest US money center banks as December 1989 (by far
sovereign bonds and the carrying value in held to maturity books. the most exposed US lenders) was US$26.4bn and the capital of the entire
FDIC banking system was US$ 204bn.
Citicorp and the other heavily exposed US and UK banks provided against their
exposures in a meaningful way only in 1987 (with provisions to exposure of
around 30%) and again (a further 30% to 30% of gross exposure) in 1989 –
i.e. 5 and 7 years after the problem became explicit. Loan loss provisions of
the FDIC insured banks over the 3 year period were 4.5% of average loans,
the highest three year cumulative charge since the FDIC was formed and a
record which was unbeaten until 2008 to 2010.
Figure 198: Average financial data for Eight Money-Center Banks: 1977 Figure 199: Average financial ratios for eight money-center banks 1977-
to 1989 1988
Year (US$m) Total Total Net Total LDC Total Provisions Total loan Year LDC Loans to LDC Loans Reserves to LDC Loans to Net income Net income to
Assets Capital Income Loans Loans Reserves for Loans charge offs total loans to capital LDC loans Cap& Reserves to assets Capital
1977 347495 14282 1554 192571 32554 1538 905 829 1977 16.9% 227.9% 4.7% 205.8% 0.45% 10.9%
1978 392575 15437 1911 217269 35811 1814 866 598 1978 16.5% 232.0% 5.1% 207.6% 0.49% 12.4%
1979 451834 17166 2320 246468 43999 2123 751 447 1979 17.9% 256.3% 4.8% 228.1% 0.51% 13.5%
1980 490753 18918 2614 274920 47614 2310 873 667 1980 17.3% 251.7% 4.9% 224.3% 0.53% 13.8%
1981 519436 20348 2629 312275 53703 2736 1065 654 1981 17.2% 263.9% 5.1% 232.6% 0.51% 12.9%
1982 546729 22115 2764 332799 54655 3036 1583 1254 1982 16.4% 247.1% 5.6% 217.3% 0.51% 12.5%
1983 541968 24211 2853 337542 55704 3416 1933 1518 1983 16.5% 230.1% 6.1% 201.6% 0.53% 11.8%
1984 560921 26655 2835 359018 58518 4107 2575 1957 1984 16.3% 219.5% 7.0% 190.2% 0.51% 10.6%
1985 590235 28233 2550 361849 56595 5451 4301 3003 1985 15.6% 200.5% 9.6% 168.0% 0.43% 9.0%
1986 605566 30343 2659 362495 54387 6988 4779 3426 1986 15.0% 179.2% 12.8% 145.7% 0.44% 8.8%
1987 593584 24954 -5529 338617 52720 17107 13065 2875 1987 15.6% 211.3% 32.4% 125.3% -0.93% -22.2%
1988 577589 29937 6268 332452 49146 16390 2270 2793 1988 14.8% 164.2% 33.3% 106.1% 1.09% 20.9%
1989 584847 26438 -2616 344130 43543 20284 9535 5544 1989 12.7% 164.7% 46.6% 93.2% -0.45% -9.9%
Source: Source: FDIC History of the Eighties – Lessons for the Future Source: FDIC History of the Eighties – Lessons for the Future
At the time of the crisis, some observers viewed the Latin American debt The market discounted money centre bank earnings from 1985 onwards
crisis either as the most serious to threaten the financial system since the figuring that bank capital would be written down as some stage. Major bank
Great Depression or so serious as likely to be a be a catalyst for another Great stocks traded at steep discount to market averages over 1987 to 1990 as they
Depression. But that did not stop that decade from being good for equity built up sufficient capital to absorb the losses although, perhaps surprisingly,
markets and good for bank shares too. For the US Money Center banks, loans they performed broadly in line with the S&P 500 until Citicorp “bit the bullet”
to lesser developed countries (LDC loans) peaked at 18% of total loans and in1987.
230% of capital just a year or two before the Mexico default. Increased
provisioning requirements significantly reduced profitability of exposed banks The price that the Latin American debtor nations paid from first over
from 1985 onwards and in two of the years the money center banks reported borrowing and then going into a debt moratorium was a lengthy period – a lost
an aggregate loss. decade – of economic stagnation.
The price the US Money Center banks paid was that their businesses stopped Looking back at the crisis and its resolution it is possible to take away the
growing; loan books shrank as they retrenched from international loan markets impression that the IMF, World Bank, bank regulators and banks tackled the
they lost share within their domestic markets. Capital levels and net asset problem in a resolute and clearly though out manner. In fact, policy was made
value per share fell slightly between 1984 and 1989 and total loans also up and driven by events and all participants were effectively playing for extra
declined. A round of mergers followed in the early 1990’s and the roll call of time to build up bank reserves to accommodate shock.
US money centre banks declined from eight to three. At some stage after the
crisis is resolved, further consolidation looks a probable outcome in the
Eurozone.

The overall lesson of the 1980’s sovereign debt crisis for European banks
today could be first that even very large risk concentrations within the global
financial system can be restructured, given time and supportive institutional
and regulatory structures and attitudes; and second that the stronger bank
balance sheets become the more likely is that losses will be recognized. The
less hopeful message is that debt forgiveness did not immediately restore the
fortunes or growth of the debtor countries.
There are many differences between the Latin American sovereign and the
Eurozone sovereign debt crises. Latin American countries defaulted on cross
border loans by foreign banks rather than on government bonds or locally
funded commercial banking loans, and thus default did not blow up the
balance sheets of their domestic banking institutions. The banks most at risk
in the 1980’s were global and systemically important megabanks whilst
arguably those most at risk today are small and large domestic banks which
are natural holders of their sovereign’s debt. Latin American countries had
borrowed too much in the wrong currency but were not burdened by a fixed
exchange rate. Last but not least, the magnitude of sovereign cross border
exposures was dimensionally larger in the 1980’s relative to capital and total
loans unless the view is taken that Italy and Spain are classified as
problem/defaulting countries.

Japan: collapse of
borrowing demand did not expand despite monetary easing. Banks were
concerned about further narrowing of their already narrow loan-deposit
spreads, and this drove them to seek new borrowers who would borrow at
bubble economy and relatively wide spreads. They found these new borrowers in the commercial
real estate sphere. Land prices had risen substantially due to the real estate
boom accompanying economic expansion, with real estate and construction
persistence of deflation companies settling into a cycle of new investment using land on which the
price had risen as collateral. During this period, bank lending as a percentage
of nominal GDP grew rapidly from 83% in 1983 to 106% in 1989. The Nikkei
Average also peaked at 38,900 at the end of 1989.
Figure 200: Bank Loans/Nominal GDP (Calendar Year,%)
110
However, by 1990 housing prices in Tokyo and other urban centers were
around five times their level in 1984, resulting in increased criticism that
housing acquisition had become difficult for individuals. In 1990, the MoF
100
introduced an aggregate ceiling on bank lending for three sectors: real estate,
construction, and nonbanks. The BoJ also reversed direction toward monetary
90 tightening, and the economic bubble proceeded to collapse.
2. Financial crisis outline

80
Bank lending as a percentage of GDP declined after peaking in 1989, but then
climbed again to 106% in 1993. This indicates that the bubble did not collapse
70 dramatically, but gradually. Although the BoJ had switched direction toward
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
Note : Loans exclude those to the Central Government
monetary tightening, there were still surplus funds in the market. Real estate
Source: BOJ, Deutsche Securities prices also fell gradually, with the result that banks continued to support via
new lending borrowers whose earnings had deteriorated. However, the sense
1. Bubble creation and collapse of alarm at corporations with weakened earnings suddenly strengthened when
a crisis emerged in 1993 among one type of nonbank, namely specialized
The 1985 Plaza Accord effectively resulted in yen appreciation against the housing loan companies known as Jusen. Agricultural financial institutions had
dollar, in response to the large scale of the US trade deficit with Japan. In one incurred substantial losses from lending to Jusen, but injections of public
year following this accord, the yen's rate against the dollar strengthened funds into them were greatly delayed by the sense of injustice among
dramatically from ¥235 to the ¥150 level. The BoJ was concerned over a taxpayers. This prompted the financial sector to look for ways to dispose of
strong yen recession caused by a slowdown in export industry, and it bad debts without recourse to public funds, or in other words time-consuming
maintained its low interest rate policy despite the signs of inflation in 1985. ways. Bank lending as a percentage of GDP eventually fell to 100% in 1998.
The government also reversed its position from fiscal austerity to expansion of
public works. Japan suffered a severe credit crunch nonetheless, with Hokkaido Takushoku
Bank and Yamaichi Securities failing in 1997, and Long-term Credit Bank of
1985 also marked a major turning point for the bank sector. Deposit interest Japan and Nippon Credit Bank failing in 1998. The financial crisis deepened,
rates were liberalized in stages from 1985. Corporations had greatly improved due to the failure of three out of the 21 major banks as of 1990. Bank lending
their balance sheet following the two oil shocks in the 1970s, and their also fell sharply from 1999. Unrealized gains on marketable securities were
included in Tier II capital from the end of 1992, when Japan introduced Basel 4. Persistence of deflation
1, but banks' lending capacity also declined in this regard, partly due to lower
share prices. ¥12.7trn in public funds was injected into Japanese banks, which is equivalent
to approximately 3.2% of GDP as of 2004. Thus far ¥12.3trn or 96.9% of this
Banks received injections of public funds, mainly as Tier II capital in 1997, and has been returned, with the Deposit Insurance Corporation generating a profit
as Tier I capital in 1998, but they were unable to halt their deterioration in of more than ¥1trn. Resona Holdings has outstanding public funds, but even it
asset quality. Credit costs at the major banks persistently exceeded their net plans to repay them over the next five years or so. Major banks deemed to
operating profit (PPP) in FY3/94–3/03. We believe that Japan's financial crisis have insufficient reserves during the process of bad debt disposal provisioned
was lengthened as a result of delays in injecting public funds, and the low cautiously even for loans in special mention category, in between bad debt
level of net operating profit relative to credit costs. and normal loans, and no banks received public funds in FY08 (FY3/09), when
the Lehman Brothers shock occurred. Some banks have applied for public
fund injections this year in connection with the Great East Japan Earthquake,
3. End of the financial crisis
but we believe this can be discussed separately from the credit cycle.
Rapid reorganization of the remaining 18 major banks occurred during 2000–
01 in response to adverse operating conditions. By 2001 the 18 banks were Bank lending as a percentage of GDP started to rise in 2008–09, but we
concentrated into seven groups: MTFG, UFJ, SMFG, Mizuho Holdings, believe this was because major corporations switched to bank borrowing
Resona Holdings, Sumitomo Trust, and Chuo Mitsui Trust (subsequently when the commercial paper market was not functioning normally for a while
MTFG merged with UFJ to form MUFG, and the two trust banks merged to after the Lehman shock, not because lending demand recovered. This
form SMTH, resulting in the current five groups). weighting certainly started to fall again in 2010. The opposition of taxpayers to
public fund injections meant that it took 20 years for the completion of one
The Financial Services Agency (FSA) tightened up its inspection manual, and credit cycle in Japan, and following the Lehman shock we believe that the US
requested major banks to halve their bad debt ratio over two years, after it and European countries put Japan's experience to good use. They now
reached 8.5% in FY3/02. As a result, an injection of public funds into the understand that the burden on the taxpayer is lessened in the end by rapid
Resona group was decided in May 2003. Resona Holdings maintained its injections of public funds, as we see it. We also believe that the likes of China
listing despite this injection of public funds, however, resulting in greatly have learnt from how rapid currency appreciation caused a bubble in Japan.
increased confidence in the financial system, with bank share prices more or
less quadrupling over three years until 2006. Earnings at the major banks also What is the current situation in Japan, now that it has left behind the bad debt
improved markedly in FY3/05–3/06, due to substantial reversals of provisions problems from the bubble era? The economic supply-demand gap that arose
against losses on loans. during the bubble era has not yet gone away, and a zero interest rate policy is
continuing. Deflation is consequently continuing in Japan, in contrast to the
We regard it as significant that as of 2004 bank lending as a percentage of inflationary trend around the world. Deposit growth and weak lending are
GDP finally returned to its pre-bubble level. Cumulative credit costs at the continuing, and banks' loan-deposit ratio has recently fallen to 81% after
major banks over the 10 years spanning FY3/95–3/04 exceeded 20% of their dropping below 100% in 2000. Japan is now a low interest rate spread, low
lending in FY3/95. During the 1990s, financial crises also ended in Korea and growth market, and the major Japanese banks have changed course toward
Sweden when a level of 20% was exceeded in this way. Cumulative credit expansion of operations overseas. We now expect earnings at Japanese
costs at the major Japanese banks reached approximately ¥66.9trn, which is banks to diverge according to their success in harnessing growth in Asia.
equivalent to 13% of GDP in 1995 and 17% in 2004.

Hong Kong: property

Figure 201: Property price index
120
`
bubble 100
80
60
Executive summary 40
The bubble formation and burst from Sept 1995 to Oct 1999 happened at a 20
very rapid pace in Hong Kong. Property price rose by 90% from Sept 1996 to
0
October 1997, and then fell by 52% until Oct 1998. The housing boom was
92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11
not driven by leverage but rather due to strong economic growth, high
Source: Deutsche Bank, CEIC
negative real interest rate (which avg -3% pre-crisis and peaked at -6.4%),
shortage of supply and speculative capital flowing into the property market.
Property bubble formation driven by negative real
The convergence of factors such as rising interest rate, expected increase in interest rates and shortage of supply
housing supply and strengthening US Dollar has caused the burst. The impact
Hong Kong uses the pegged currency regime and since 1983, Hong Kong
to the banking sector and the economy was manageable as Hong Kong banks’
Dollar has been pegged to the US Dollar. As an international trading hub and
profitability was robust enough to weather the rise in NPL and credit cost. We
financial center, pegged currency offers exchange rate stability to Hong Kong,
estimate the total cost to the banking sector to be 11%, measured as
but downside of it is in times of strong economic growth, the risk of asset
percentage of total system credit cost to 1997 GDP. Due to the recourse
price bubble increases as raising interest rate is not an available tool to fight
nature of Hong Kong mortgage loans and relatively low loan-to-value ratio,
inflation.
actual default on mortgage loans remained low. Equity to asset ratio of Hong
Kong banks remained healthy at 9.1% in 1998, which helped to avoid credit
The buildup of property bubble in early 1990s leading up to the bubble
crunch. In fact, loans for use in Hong Kong only contracted by 3.8% in 1998.
bursting in October 1997 was attributed to various factors. First, Hong Kong’s
inflation rate surged to average 8.5% driven by strong real GDP growth which
averaged 5.30% from 1992 to 1997. Second, Hong Kong’s real interest rate
fell to negative territory as Hong Kong’s pegged currency regime prevented
interest rate hikes. From 1992 to 1997, real interest rate in Hong Kong
averaged -3 %, peaking at -6.40%, and this caused money inflow into property
and stock markets. Third, strong stock market performance and wealth
created flowed into the property market. Hang Seng Index increased by 288%
from 1992 to Aug 1997. Fourth, strong property price increase brought
speculative capital inflow into the property market. Lastly, mortgage
competition intensified which brought down mortgage rate from 10.75% in
Aug 1995 to as low as 9% in May 1996.

As a result, property price index rose by 81% from 1992 to April. Concerned Figure 203: Real interest rate and property index
about property bubble and bank’s increasing exposure to the property market,
Real interest rate(LHS) Property price index(RHS)
HKMA introduced prudential measures, mainly limiting banks’ property
14% 120
exposures to 40% of loan book and capping maximum loan-to-value ratio to
12% 110
70%. This as well as rise in mortgage rates helped to stabilize property prices
10% 100
from mid-1994 to late 1995. During this period, property price declined by
8% 90
39%. However, prices and transaction volumes picked up again in 1996 due to
6% 80
shortage of supply as a result of restricted land supplies in earlier years. The
4% 70
rate of price increase accelerated with property price increasing by 13% and 2% 60
43 in 1996 and 1997 respectively. 0% 50
-2% 92 93 94 95 96 97 98 99 00 01 02 40
Hong Kong’s housing bubble was not caused by leverage. We estimate loan- -4% 30
to-value ratio remained relatively low at c. 60~70%. HKMA introduced more -6% 20
prudential measures, reducing maximum LTV for properties above HKD12m to -8% 10
higher of 60% of value of HKD 8.4m in Dec 1996: From 1992 to 1997, system Source: Deutsche Bank, CEIC
mortgage loans increased by CAGR 16%, property loans as a percentage to

total loans increased from 15% to 24%. Bubble burst but impact to banking sector was
manageable
Figure 202: Real GDP growth and CPI
The bursting of bubble was caused by many factors. First, interest rate started
Real GDP growth CPI (%)
12%
to rise. The Hong Kong prime rate rose by 100bps between January 1997 and
November 1997. Since mortgage loans in Hong Kong are floating rate tied to
9% prime rate, mortgage rates increased in tandem. Three months HIBOR also
rose by 510bps from Jan 1997 to Oct 1997. Second, US Dollar started to
6% strengthen, causing a capital outflow from Hong Kong. Third, the real
economy started to deteriorate with the global economic slowdown that
3% began in July 1997. Lastly, housing supply was expected to increase with the
government announcing in Oct 1997 its plan to increase the number of flats to
0%
at least 85,000 flats a year. Other government measures included plans to
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
-3%
increase home ownership rate from 50% to 70% by 2007 and to reduce the
average waiting time for public rental housing from 7 to 3 years by 2005.
-6%
Source: Deutsche Bank, CEIC
The bursting of property bubble occurred very rapidly. From October 1997 to
October 1998, property prices fell by 52%. Hong Kong GDP contracted by
5.1% and exports declined by 7% in 1998. Consumer and business sentiment
also weakened. Hong Kong’s fiscal deficit reached to HKD 59.8bn (or 4.4% of
its nominal GDP) and unemployment rate surged from 2.4% in Jan 1998 to
6.4% in Jan 1999 (a level unseen since 1983).

Total cost to banking system was 11% of 1997 GDP System loans contract and deflation lasts for eight
The impact to the banking system was negative but manageable. Specifically, years
asset quality deterioration caused a surge in classified loans which more than Post-bubble bursting, banks in Hong Kong saw fall in both domestic and
tripled from 2.7% in 1996 to 10.1% in 1999. NPL ratio was as high as 7.6% in offshore lending. System loans contracted by 51% from 1998 to 2003, with
1999. As a result, Hong Kong banks’ credit cost also surged to 1.29% in Dec the major driver being overseas loans (down 88%) as post-Asian crisis,
1998. This is about five times higher than historical average credit cost of regional banks reduced their short-term overseas bank borrowings for which
0.23% (from Mar 1997 to Mar 2011). The total cost of Hong Kong property Hong Kong banks played a significant role as an international financial center.
bubble bursting to the banking sector is estimated to be roughly 11% of 1997 Loans for use in Hong Kong also contracted from 1997 to 2003 but at a much
GDP . more manageable level of 16%. Mortgage loans recorded positive growth of
7.3% and 3.3% in 1998 and 1999 respectively but remained relatively flat until
Figure 204: Hong Kong banks’ credit cost/GDP 2007.
4.0%
3.5% It was under these conditions that the complete interest rates deregulation
3.5%
took place in 2001 which increased competition. The resulting reduction in
3.0%
2.6%
mortgage interest rates has helped to improve the demand for properties,
2.5% helping to halt the falling property prices. To facilitate recovery, HKMA and the
Hong Kong government also undertook such measures: (1) From June 1998,
2.0% all land sales in Hong Kong were suspended until April next year, the first time
1.3%
1.5% 1.1% 1.2%
1.3%
1.2% the government has suspended such sales since 1953. (2) HKD3.88bn rate
1.0% 0.8% 0.9% rebate and increased funding to first time home buyers. (3) Fixed rate
0.6%
0.4% 0.4%
mortgages was cut from 10.5% to 10.25% with Hong Kong Mortgage Corp
0.5% 0.3%
0.1% 0.1% 0.2% committing to buy HKD3.5bn of fixed rate mortgage loans from participating
0.0% banks.
95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
The bursting of bubble was followed by a quick recovery in the real economy.
Source: Deutsche Bank, HKMA
Hong Kong’s real GDP turned positive in 1999, registering 3.4% growth. Fiscal
balance also turned surplus of HKD 19.4bn (or 1.5% of its nominal GDP) in
Despite the severity of the property price correction, Hong Kong banks’
1998. Banking sector ROE improved to 11.6% in 1999 then to 14.7% in 2000.
operating profit was strong enough to withstand the spike in credit costs. In
However, Hong Kong experienced deflation until 2005 which was an
fact none of the banks under our coverage reported a loss. Although system
adjustment mechanism since the option of currency devaluation was not
wide ROE dropped significantly, ROE remained at 9.5% in 1998 (vs 15.7% in
available due to pegged exchange rate system. The return of monetary and
1997). Banks’ profitability was robust enough to weather the bubble burst. The
financial stability has provided a sound foundation for economic activity to
bubble burst also affected caused Hang Seng Index to drop by 20% in 1997
recover. Expansionary fiscal policy, especially an increase in government
and fall by another 6% in 1998. The market collapse fortunately did not cause
infrastructure spending, continued in 2002, which contributed to maintain
significant write down of banks’ capital, enough to cause credit crunch. In fact,
growth in domestic demand. Fast-growing mainland Chinese economy and
system equity to asset ratio remained robust at 8.9% in 1999 (vs 9.9% in
booming trade with China has also helped the recovery of Hong Kong
1997). Strong capital position prevented capital crunch, helping the quick
economy.
recovery of the overall economy.

Argentina’s financial
CEDRO’s to affected customers, representing the interest in the
underlying, rescheduled deposits.
crisis 5. Pesification: All bank accounts and loans to the private sector
denominated in dollars would be converted to pesos. The main problem
for the banks was that this conversion was to be made at different
exchange rates, assets were converted at 1.1 ARS/US$ and liabilities at
Banking crisis of 2001/2002 1.4 ARS/US$.
Beginning in 2001, after two years of economic 6. Compensation bonds: The Argentine government issued bonds (Boden
A number of measures by recession, private depositors began to withdraw 2007 & Boden 2012) in order to compensate the banks for losses related
the Argentine government funds from the banking system due to growing to the asymmetric pesification.
failed to slow the economic
concerns about the solvency of the system. In
response to this run on deposits, the Argentine 7. Amparos: Many depositors claimed that the pesification was
crisis in 2001-2002
government instituted a number of measures to unconstitutional and sued the banks for compensation (with many
stem the flow of money out of the system. gaining favorable rulings). The Central Bank rules permit the losses
related to amparos to be accounted for as an intangible asset and
1. Corralito: In December 2001, following a massive run on deposits, amortized over five years. At the end of December 2006, the Supreme
including a 3% decrease (US$1.8bn) on Court ruled that the pesification was constitutional and ordered banks to
November 30th alone, the government pay depositors at the 1.4 ARS/US$ exchange rate, adjusted for inflation
GDP declined 11% in 2002,
enforced limits on the amount of cash that and 4% annual interest (this works out to around 3.08 ARS/US$).
the fourth straight year of
could be withdrawn from banks within
decline, but has rebounded
specified time periods ($1,000 per month). Following these events the economy went into a deep recession. Indeed,
strongly since
However, this measure increased public GDP declined 4.4% in 2001 and 10.9% in 2002. Furthermore, deposits fell
concern about the solvency of the banking 23% in 2001 and private sector loans fell 19% in 2001 and another 25% in
system and contributed to a further decrease in deposit levels, as many 2002; Figure 205. As such, loans to GDP fell sharply in 2003 to 18% from
depositors withdrew funds up to the permitted limit. 27% in 2002. While GDP growth and loan growth have both rebounded
strongly since 2003, this has been offset by high rates of inflation, which has
2. Default: In December 2001, following a downgrade to its sovereign kept loans to GDP relatively stable between 15-16%; Figure 206.
credit rating, the suspension of loans from the IMF, and the resignation
of President De la Rua, the interim government, led by Adolfo Rodriguez
Saa formally defaulted on Argentina’s debt of about US$150bn.
3. Devaluation: In January 2002, Argentina abandoned the fixed 1-to-1

parity between the peso and the dollar. The peso depreciated
significantly, reaching a low of 3.86 ARS/US$ in June 2002.
4. Corralon: In January 2002, the government extended the maturity for

time deposits, freezing them further. Banks were required to issue

Figure 205: GDP growth vs. loan and deposit growth Inflation has risen above
The crisis triggered very high rates of inflation,
50% 15% as the devaluation of the peso created extreme
20% again, but still below
pressures on domestic prices. Consumer
40% the peak of 41% in 2002
10% inflation reached a high of 41% in 2002, while
30% wholesale price rose over 100%. The Central
20%
5% Bank introduced a new short-term financial instrument know as a LEBAC
(Letras del Banco Central) or Central Bank external bills in order to absorb the
10% 0%
excess liquidity generated through its temporary advances to the financial
0% sector. This helped to reduce inflation, and consumer prices quickly stabilized,
-5%
increasing only 3.6% in 2003. Since 2003, inflation has increased again,
-10%
reaching 25% in 2010 as the government has pursued a policy high economic
-10%
-20% growth at the expense of high inflation; Figure
Interest rates have also risen 207.
-30% -15%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 again, but not enough to
Private sector loan growth (left axis) Deposit growth (left axis)
Real GDP growth (right axis) stave off inflation Interest rates rose dramatically during the crisis
Source: Deutsche Bank, Central Bank of Argentina. as well, reaching levels over 100% in 2001. The
Central bank began to reduce interest rates after April 2002, reaching a low of
Figure 206: Loans to GDP 1.5% in 2003. As the economy recovered and continued to grow, interest
rates have climbed again reaching 11.5% at the end of 2008, but stabilizing
30% 29%
near 10% since then; Figure 208. The reluctance by the government to
29%
27% significantly raise interest rates again has led to negative real rates, which has
been positive for the banks as they have benefited from relatively low funding
costs and strong volume growth. However, the high levels of inflation present
a major risk to the development of the long-term credit market.
18%
16% 16% 16%
16% 16%
15% 15%
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Source: Deutsche Bank, Central Bank of Argentina.

Figure 207: Inflation Rate (yoy) and resulted in many Argentine companies defaulting on their foreign currency
41.2% debt obligations. However, the devaluation of the currency did make imported
goods more expensive, which eventually led to increased demand for
25.2%
domestic products that helped the economy recover. The currency eventually
23.0%
strengthened to 2.93 ARS/US$ by the end of 2003, but has been depreciating
at a much more steady pace since then; Figure 209.
17.0%
14.8%
12.3% One of the major impacts of the crisis was the

10.0%
Unemployment has
tremendous social hardship. Unemployment
reached a record high of 19.7% during the crisis.
6.1% improved beyond pre-crisis
3.6% The quality of life of the average Argentine was
levels
drastically lowered; many businesses closed or
-0.2% went bankrupt, many imported products
-1.1%
-2.4% became virtually inaccessible, and salaries remained stagnant at the levels
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 before the crisis. However, the unemployment rate has steadily improved as
Source: Deutsche Bank estimates, Central Bank of Argentina. the economy recovered, reaching levels even better than before the crisis.
The improvement in the labor market has allowed households to finance a
Figure 208: Overnight interest rate (Period-end) larger portion of their consumption through consumer loans and wages have
113.2%
increased with inflation. The unemployment rate improved to 7.3% in 2010;
Figure 210.
11.5%
10.5% Figure 209: FX Rate (Period-end)

9.5% 9.7%
9.3%
8.2% 3.98
7.9%
3.80
6.8%
6.3% 3.45
3.38
3.15
3.03 3.06
2.93 2.97
2.3%
1.5%
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 1.00 1.00 1.00
Source: Deutsche Bank estimates, Central Bank of Argentina.
In January 2002, the Argentine government

The currency dropped 70% abolished the fixed parity between the peso and 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Source: Deutsche Bank, Bloomberg Finance LP

in 2002 and has depreciated the US dollar, which had been in effect for ten
gradually since 2003 years, and the currency depreciated
dramatically, reaching 3.38 APs/US$ by the end
of 2002. This led to very high rates of inflation
Figure 210: Unemployment rate (Period-end) Figure 211: System ROE

19.7%
24.4%
17.3%
19.2%
15.1% 14.3% 13.4%

13.8% 14.0% 11.0%
7.0%
12.1%
1.7%
10.1% 0.0%
-0.2%
8.7% 8.4%
8.2% 8.0% -4.2%
7.3%
-22.7%
-59.2%
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Source: Deutsche Bank, Central Bank of Argentina.
Source: Deutsche Bank, INDEC
ROE for the system was low before the crisis, Figure 212: System NPLs as % of loans
45.0%
but fell to -59.2% in 2002 and -22.7% in 2003,
ROE for the system due to negative margins and high provision 40.0%
collapsed in 2002, but has charges, as asset quality deteriorated 35.0%

recovered to well above pre- significantly. However, ROE has rebounded to
30.0%
crisis levels well above pre-crisis levels reaching 24% in
2010, as the banks have benefited from strong 25.0%
loan growth, improved asset quality, and much better NIMs; Figure 211. 20.0%
System NPLs spiked to 39% of private sector loans in 2003 and remained at 15.0%
34% in 2004. Given the strong economic growth we have seen since then,
10.0%
asset quality has improved dramatically with private sector NPLs reaching a
low of 2.1% in 2010, just above the total NPL ratio of 1.8%; Figure 212. 5.0%
0.0%
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Total NPL ratio Private sector NPL ratio

Source: Deutsche Bank, Central Bank of Argentina

Ireland: The Celtic Credit

Figure 214: The rise in credit since 2002
2,000,000
1,800,000
Tiger
1,600,000
1,400,000
EUR millions
1,200,000
1,000,000
800,000
Background to the Crisis 600,000

400,000
200,000
Wider macroeconomic factors: 0
Rapid economic growth and easier liquidity driven in part by greater risk
Nov-03
Nov-08
Jan-03
Jun-03
Jan-08
Jun-08
Jul-05
Jul-10
May-06
Mar-02
Feb-05
Mar-07
Feb-10
Apr-04
Oct-06
Apr-09
Aug-02
Sep-04
Dec-05
Aug-07
Sep-09
Dec-10
appetite and falling inflation and interest rates, characterised the global
economic environment during the late 1990s. This lead to strong credit Non-financial corps Financial corps Govt. Household
growth, higher leverage and significant rises in asset prices. Ireland was no
Source: Haver, Deutsche Bank estiamtes
exception, and these conditions would combine to have a significant impact
on the Irish economy and banking system. Originally, Irish growth was driven by supply-side changes, but the late 1990s
saw a shift to demand-driven growth as wages increased, bank lending
Joining the European Economic Community in 1973 as the lowest GDP/capita expanded and government expenditure accelerated. Euro membership and
country amongst the seven members at the time, Ireland experienced low interest rates facilitated cheaper international borrowing, fuelling growth
significant growth in GDP from the late 1980s, driven by more stable despite relatively uncompetitive domestic inflation and wages. Ireland
economic conditions, foreign investment and fairly high population growth achieved fiscal surpluses for a decade up to 2006, masking an increasing
(Figure 213). Economic growth dipped in the early 1990s, but from 1995 to dependence in tax revenues on asset prices linked to corporation tax, stamp
2002 Ireland’s GDP nominal growth rate was above 10% per annum, with real duty and capital gains tax. Deductibility of mortgage interest payments from
GDP growth above 7% from 1996 to 2000. income tax also served to encourage property investment.
Figure 213: Irish GDP over time Irish Banking Sector

250,000 25% Closer integration with Europe increased the availability of wholesale funding
200,000 20% to Irish banks without foreign exchange risk, whilst the entry of foreign banks
150,000 15% in the domestic lending market compressed margins and increased debt
GDP (EUR millions)
100,000 10% availability. The result was a huge expansion in property lending. Credit growth
YoY GDP %
50,000 5% was in excess of 10% every year from 2001 until 2008, with 2004-2006 seeing
0 0%
credit growing ~30% a year (Figure 215). Some institutions were more
-50,000 -5%
aggressive than others: for example Anglo Irish had a CAGR of c.42% loan
growth from 2003-2006, whilst BKIR was c.27%. Nevertheless the rapid
-100,000 -10%
quantum of lending expansion was a systemic phenomenon (Figure 216).
-150,000 -15%
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
Nominal GDP Nominal GDP growth Real GDP growth
Source: Haver, Deutsche Bank estimates

Figure 215: Private Sector Credit Growth 2001 onwards Failure of risk management & regulation
450,000 45.0% The recent government-commissioned report into the banking crisis2 identified
4 key failings of bank management: 1) concentration of lending in property &
350,000 35.0%
CRE in particular; 2) poor lending criteria and “a tidal wave of uncritical
250,000 25.0% enthusiasm” to take part in the property boom; 3) remuneration incentives in
top and middle management; 4) breaches of accounting and company law at
YoY growth
EUR m
150,000 15.0%
specific institutions. The report cites failures within the regulatory framework
50,000 5.0%
where supervisors did not take sufficient action, did not have the resources
available, and underestimated the systematic vulnerabilities.
-50,000 -5.0%
-150,000 -15.0% Figure 217: House prices

Jan-01
Jan-02
Jan-03
Jan-04
Jan-05
Jan-06
Jan-07
Jan-08
Jan-09
Jan-10
Jan-11
Jul-01
Jul-02
Jul-03
Jul-04
Jul-05
Jul-06
Jul-07
Jul-08
Jul-09
Jul-10
450000 40%
400000 30%
House prices (YoY %)

350000
Source: Haver, Deutsche Bank estiamtes 20%
House price (E)

300000
250000 10%
Figure 216: BKIR & AIB loan balances 200000 0%

150000
160,000 -10%
100000
140,000 50000 -20%
120,000 0 -30%
Q1 1991
Q1 1992
Q1 1993
Q1 1994
Q1 1995
Q1 1996
Q1 1997
Q1 1998
Q1 1999
Q1 2000
Q1 2001
Q1 2002
Q1 2003
Q1 2004
Q1 2005
Q1 2006
Q1 2007
Q1 2008
Q1 2009
Q1 2010
100,000
80,000
60,000
New houses (E) Second hand houses (E)
40,000
Source: Department for Environment, Community and Local Government
20,000
0
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
AIB BKIR
Source: Company data, Deutsche Bank estimates
2
Klaus Regling & Max Watson, ‘A Preliminary Report on The Sources of Ireland’s Banking Crisis’
Figure 218: Household completions Figure 219: Sector composition of debt across peripheral countries
100,000 30% 393% 382%
400%
90,000
Housing com pletions (U YoY %)

20%
Housing com pletions (units)

80,000 350% 57% 332%
10% 103% 325%
70,000
0% 300% 63% 69% 264%
60,000
-10% 250% 138%
50,000 67% 54% 52%
% GDP
-20% 139%
40,000 200%
-30% 50% 55%
30,000 88%
150%
-40% 96% 38%
20,000
100% 86%
10,000 -50%
152%
0 -60% 50% 114% 120%
91%
64%
Aug-98
Aug-00
Aug-02
Aug-04
Aug-06
Aug-08
Aug-10
Dec-97
Dec-99
Dec-01
Dec-03
Dec-05
Dec-07
Dec-09
Apr-99
Apr-01
Apr-03
Apr-05
Apr-07
Apr-09
Apr-11
0%
Spain Portugal Ireland Greece Italy
Completions Completions YoY % Financial Non-Financial Business Households Government
Source: Department for Environment, Community and Local Government Source: Bank of Spain (Data as of June 2010), Bank of Portugal (Data as of June 2010), Bank of England (Data as of September 2010), Central
Bank of Ireland (Data as of October 2010), Bank of Greece (Data as of October 2010), Bank of Italy (Data as of October 2010), IMF (data as of
October 2010), Eurostat.
Catalysts: house price falls, economic slowdown & liquidity dry up,
house price falls Figure 220: Progression of Irish liability guarantee schemes
Irish Bank shares peaked in 1H07, with house prices beginning to fall in 1Q07
for the first time since 2002 (Figure 217). GDP growth turned negative in the 30 Sept 2008: CIFS announced. 2‐yr blanket guarantee of all
deposits & covered debt securities (c.E400bn)
3Q07, though Irish banks reported a combined E6.5bn of pre-tax profits in
9 Dec 2009: ELG Scheme announced. Covered new deposits &
2007 compared with E5.3bn in 2006, with Anglo Irish reporting a 46% rise
eligible debt securities for up to 5 yrs
(Sept 2007 vs Sept 2006). However, in 1H08, amidst further deteriorating
7 Sept 2010: ELG scheme issuance window extended
economic conditions, the banks began to warn of falling profitability.
to Dec 2010
29 Sept 2010: CIFS ends
On 25 Sept, 10 days after the Lehman bankruptcy, Ireland became the first
19 Nov 2010: ELG issuance window extended until June 2011
Euro country to enter recession. The subsequent liquidity freeze prompted the
government to issue a guarantee 5 days later covering all liabilities (deposits 31 Mar 2011: ELG scheme covered E111bn of
liabilities, of which E81bn were deposits
and wholesale funding) at 6 Irish financial institutions (AIB, BKIR, Anglo Irish,
Irish Nationwide, EBS, Irish Life & Permanent) totalling E400bn in obligations. 1 June 2011: ELG issuance window extended until Dec 2011
At the time this guarantee was billed by the government as “the cheapest 31 Dec 2011: ELG issuance window due to close
bailout in the world”, but it was to prove controversial as the scale of banking
sector problems became apparent, and led to questions around the state’s Source: BKIR, NTMA.
ability to underwrite such a large guarantee (E400bn of liabilities vs GDP of

E160bn in 2009). The scale of debt in the public and private sectors across the
peripheral countries is shown in Figure 219.

Cost & Impact of the crisis Construction industry collapse

The construction sector, which employed > 12% of the workforce during the
GDP & unemployment impact boom, has shrunk by 60% and continues to fall. The scale of housing
During 2008-2010 Ireland suffered one of the most severe recessions in the oversupply means it is unlikely, in our view, that the construction industry or
history of the modern developed world. The peak-to-trough fall in GDP (from house prices will recover in the near term. A recent government review of
March 2007 to December 2010) was 23% (Figure 221), whilst unemployment 2,846 housing developments in Ireland1 found that, of the 121,275 dwellings
rose from a long-run average of c.4% during the first half of the decade to planned for these developments, 23,250 are complete and vacant, with
over 14%, remaining stubbornly high today (Figure 222). another 19,830 in various stages of completion. Combined, these amount to
c.4 years of current new house builds, and are likely to cap significant near-
Figure 221: GDP & construction component term gains in house prices, in our view. House prices in Ireland have already
60,000 3,000 fallen 37% peak-to-trough (Figure 223).
Construction industry quarterly

50,000 2,500
Figure 223: House prices peak to trough
GDP quarterly (EUR)
40,000 2,000
450,000
30,000 1,500 400,000
20,000 1,000 350,000

37%
Average house price

300,000
10,000 500
250,000
0 0 200,000
Nov-97
Nov-99
Nov-01
Nov-03
Nov-05
Nov-07
Nov-09
Jul-98
Jul-00
Jul-02
Jul-04
Jul-06
Jul-08
Jul-10
Mar-97
Mar-99
Mar-01
Mar-03
Mar-05
Mar-07
Mar-09
Mar-11
150,000
100,000
Irish GDP Construction industry
50,000
Source: Haver 0
Dec-95
Sep-96
Dec-98
Sep-99
Dec-01
Sep-02
Dec-04
Sep-05
Dec-07
Sep-08
Dec-10
Jun-97
Jun-00
Jun-03
Jun-06
Jun-09
Mar-95
Mar-98
Mar-01
Mar-04
Mar-07
Mar-10
Figure 222: Unemployment remains high
300 16 Source: Haver, Deutsche Bank estimates
Employees in Consturction industry
14
250
Unenmploment rate (%)
12
200
10
150 8
6
100
4
50
2
0 0
Mar-00
Mar-01
Mar-02
Mar-03
Mar-04
Mar-05
Mar-06
Mar-07
Mar-08
Mar-09
Mar-10
Mar-11
Sep-00
Sep-01
Sep-02
Sep-03
Sep-04
Sep-05
Sep-06
Sep-07
Sep-08
Sep-09
Sep-10
Employees in construction (LHS) Unemployment rate (RHS)
Source: Haver

Figure 224: Irish Bank shares since peak Attempts at resolving the crisis
120
State guarantee, bailouts and NAMA

100
As mentioned above, the initial government response to the crisis facing
80 Ireland in Sept 2008 was to guarantee E400bn of financial sector liabilities in
60 order to keep the sector solvent. This was followed in December by the first
of several recapitalisation rounds (E5.5bn contributed in this case). In January
40
2009, Anglo Irish Bank was nationalised, and a further E7bn was injected into
20 AIB & BKIR. As property prices fell further, the scale of non-performing
0
property development loans became more apparent. The Irish government’s
solution for this was to create a ‘bad bank’ (similar to that used by Sweden in
Jun-07
Jun-08
Jun-09
Jun-10
Jun-11
Feb-08
Feb-09
Feb-10
Feb-11
Oct-07
Apr-08
Oct-08
Apr-09
Oct-09
Apr-10
Oct-10
Apr-11
Aug-07
Dec-07
Aug-08
Dec-08
Aug-09
Dec-09
Aug-10
Dec-10
the 1990s) called the National Asset Management Agency (NAMA),
BKIR AIB ANGLO IRISH IRISH LIFE & PERMANENT announced in April 2009. Since then NAMA has bought E72.3bn of loans from
Source: Datastream the 5 main Irish banks, at an average discount of 58% - significantly higher
than was initially expected. These crystallised bank losses, and combined with
Impact on shareholders provisions taken against other portfolios retained by the banks have resulted in
From their highs in 1H07, Irish bank shares had already fallen by 70% before impairments of c.E70bn in the three years 2008-2010, of which Anglo Irish
the Lehman Brothers bankruptcy (Figure 224). A recovery during 2009 on represented over half. Adding in losses taken by RBS’s Ulster Bank Subsidiary
hopes that the government policies to deal with the crisis would be effective and Lloyds Banking Group’s HBOS Irish operations (both of which did not
was not sustained. participate in NAMA) raises total impairments taken up until 2010 to c.E90bn,
c.60% of GDP. NAMA has subsequently written down the value of its
Anglo Irish and Irish Nationwide are now nationalised entities in run-down. AIB property loans even further, though it still budgets a net gain to the taxpayer of
is now 99.8% state owned and the building society EBS (also state owned) E1bn over its ten year life.
has become a subsidiary of AIB BKIR recently underwent a E4.3bn capital
raise using a combination of LME, rights issue and institutional placing, Figure 225: Provisions through the crisis
resulting in 15% state ownership, 19% bondholder ownership and the balance 80,000
private investors. The banks still listed: BKIR, AIB, Irish L&P are currently 1%, 70,000
0.5% and 0.25% of their peak share prices respectively (Figure 224). Fianna 60,000
Fail, the ruling party which had been in power since 1987, was defeated by a
50,000
EUR millions
Fine Gael & Labour coalition in Irish national elections in early 2011.
40,000
30,000
20,000
10,000
0
2008 2009 2010 Cumulative
AIB BKIR Anglo Irish Irish Life & Permanent EBS Irish Nationwide
Source: Deutsche Bank estimates, company data

Figure 226: Irish 5yr CDS Figure 227: Irish Debt to GDP
1,400
120%
1,200
100% 95% 96% 95% 94%
92% 90%
1,000 82%
80% 74%
64% 66%
800
60% 54%
49%
44%
600 38% 36%
40% 32% 31% 30%
28%
25% 25%
400
20%
200
0%
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
0
Jun-09
Jun-10
Jun-11
Feb-09
Feb-10
Feb-11
Oct-08
Apr-09
Oct-09
Apr-10
Oct-10
Apr-11
Dec-08
Aug-09
Dec-09
Aug-10
Dec-10
General government debt/GDP ratio
Source: Datastream Source: National Treasury Management Agency
…leading to IMF/EU bailout Figure 228: ECB financing to Irish Banks

The scale of property losses saw AIB and BKIR raise additional capital during 80,000
2010, with BKIR completing a E3bn capital raise in which the state 70,000
participated. The liability guarantee was due to expire at the end of September
60,000
2010, but with Irish banks struggling to fund in the unsecured markets (Figure
50,000
228), this was extended and further capital raisings were ordered (Figure 220).
EUR m
40,000
In November 2010, amidst increasing concern over the effectiveness of
implemented austerity measures, size of public debt (94% of GDP at the end 30,000
of 2010, Figure 227), further likely losses from the banking system, and rising 20,000
CDS premia (Figure 226), the Irish government formally requested support EU 10,000
and IMF. The result was an E85bn package, comprised of E50bn for budgetary -
financing needs and E25bn to strengthen bank capital bases and for future
Nov-10
Jun-10
Jan-11
Jun-11
Jul-10
May-11
Feb-11
Mar-11
Oct-10
Apr-11
Aug-10
Sep-10
Dec-10
contingencies. The Irish state was also required to implement a package of
austerity reforms, and conduct a fresh round of stress testing of the banking Source: Central Bank of Ireland
sector.

IMF/EU-required recapitalisation of the banks

The stress tests, announced at the end of March 2011, sought to establish
recapitalisation additional capital requirements for the remaining 4 Irish
institutions on the basis of future expected loan losses, plus a buffer designed
to cover unforeseen losses and bolster market confidence. As shown in Figure
229 below, the report estimated E20bn of further losses over the 2011-2013
on the base scenario, rising to E27bn in the adverse case. This is on top of the
E70bn losses taken by Irish banks since 2008, and assumes no further
impairment provisioning requirements at Anglo Irish. If these loss expectations
are correct, then almost a quarter of the E400bn of loans on Irish bank balance
sheets in 2008 will have been written down over the 6 year period. By way of
comparison, the cumulative three year peak losses in the early 1990’s
recession in UK saw losses of 6.0% at Barclays (91-93), 4.9% at LBG (92-94),
and 5.15% for RBS (91-93), further evidence of the scale of the Irish credit
crisis and the poor quality of lending which took place in the prior decade.
Figure 229: Central Bank expected loss assumptions

AIB BKIR Irish Life & Permament EBS Total
Base Stress Base Stress Base Stress Base Stress Base Stress
Residential Mortgages 2,005 3,066 1,361 2,366 1,624 2,679 848 1,380 5,838 9,491
Corporate 564 972 799 1,179 0 0 0 0 1,363 2,151
SME 2,157 2,674 1,445 1,837 0 0 0 0 3,602 4,511
CRE 3,653 4,490 3,148 3,847 231 400 127 197 7,159 8,934
Non-mortgage corporate and other 1,167 1,403 627 891 259 342 0 0 2,053 2,636
Total 9,546 12,605 7,380 10,120 2,114 3,421 975 1,577 20,015 27,723
Recap requirement 13,300 5,200 4,000 1,500 24,000
Source: Central Bank of Ireland Financial Measures Programme Report

The E24bn additional capital raise for the sector is ongoing. The Credit
Stabilisation Act (passed by the last Irish government in Dec 2010) gives the
Minister of Finance extraordinary powers to inject capital where required,
overriding normal equity and bondholder rights and potentially forcing cram
down on debtholders. So far these powers have been used to inject capital
into AIB, and to generate equity from debtholders in Anglo & AIB. The BKIR
capital raise has been completed, though there remains scope for these
powers to be used against subordinated debtholders.3
Though lossmaking at present, Bank of Ireland management aim to produce a

low teens ROE in 2014 driven by an improvement in net interest margin (up by
around a third), normalisation of loan losses, and substantial cost cutting. If
achieved, this could see the bank producing E800-900m a year in attributable
earnings at that point, compared with a current market cap of about E3bn. It is
also interesting to note that the company expects to produce a >15% core tier
1 ratio at that date, despite the intent to repay E1.8bn in government
preference shares by then. If we assume forward RWAs of, say E62bn, and a
normalised core tier 1 requirement of 12%, this would imply surplus capital of
about E1.8bn, 60% of current market cap. This plan requires a return to far
lower funding costs and easier market access to wholesale debt markets for
the Irish banks, as well as loan loss normalisation, prompting a sharp
improvement in profitability. We note the purchase of 35% of Bank of Ireland
from the Irish government in July 2011 by a group of investors including
Fairfax Financial Holdings, WL Ross & Co, Fidelity and Capital Group.
3
See our note on BKIR’s capital raise: ‘One step at a time’ (July 2011)
Australian Case Study

Figure 230: Australia’s nominal and Real GDP growth
16%
14%
12%
10%
Australian Banking Crisis – 1991 – 1993 8%

6%
In this section we review the major Australian Banking crisis of 1991–1993 4%

2%
which resulted in 2 of the largest banks in the country taking significant
0%
writedowns and requiring substantial capital injections in order to remain -2%
solvent. The crisis was driven by a combination of: i) deterioration in the -4%
Jan-85
Jan-86
Jan-87
Jan-88
Jan-89
Jan-90
Jan-91
Jan-92
Jan-93
Jan-94
Jan-95
macroeconomic environment; ii) deregulation of the banking industry leading
to strong competition and poor lending practices.
Nominal GDP growth Real GDP
We will review these issues in the sections below, coupled with the cost of Source: Deutsche Bank, RBA
the crisis and how the crisis was resolved.

Figure 231: Credit growth prior to the crisis
Background to the crisis 35.0
%
30.0
In the early 1980’s Australia embarked on a series of financial system reforms,
25.0
primarily in response to the report handed down by the Campbell Committee.
20.0
These reforms included the removal of interest rate ceilings on bank deposits,
15.0
the introduction of the tender system for the sale of Treasury notes and
10.0
bonds, and the floating of the AUD in December 1983.
5.0
But the single most important catalyst for the banking crisis was the 0.0
deregulation of the banking market, which saw the entry of 15 foreign banks in
1985. In the words of former Reserve Bank of Australia (RBA) Governor Ian
Macfarlane, “they were eager to gain a foothold in Australia, and this meant Total Credit Housing Business
lending to where it was easiest to do so, which was lending to businesses. Source: Deutsche Bank, RBA
Foreign banks everywhere have always found it difficult to break into the
household lending market”4. Annual credit growth averaged 17.9% in the ten years ended 31 December
1989, and total credit in the economy rose from 55% of GDP as at 30 June
Deregulation led to a boom in asset prices, an associated company takeover 1985 to 81% of GDP as at 30 June 1990. Rolling 12 month credit growth
frenzy, and a property bubble. In turn, this drove a multi-year boom in GDP peaked in June 1988 at 31.5%, just a couple of years prior to the start of the
growth and credit growth in Australia, as shown in the chart below. crisis.
4
“The Recession of 1990 and its Legacy”, Lecture 3 in the Boyer Lectures, delivered by Ian
Macfarlane, 2006
The banking sector reacted to deregulation by chasing growth. Most of this In the 5.5 years from December 1984 to June 1990, commercial property
growth was concentrated in the commercial property market and largely prices increased at a compound annual growth rate (CAGR) of 19% p.a., with
reflected speculative investment as shown in the chart below which highlights growth peaking in CY88 and CY89 at ~30% p.a. This was followed by a peak
business credit growing above business CAPEX before the crisis. to trough fall in office property prices of -30% over the subsequent three
Interestingly, as highlighted in the previous chart, housing was not as big a years.
driver of total credit growth over this time.
Figure 232: Business credit outpaced business capex Figure 234: Inflation was a persistent problem
40 % 10.00%
9.00%
30
8.00%
20 7.00%
6.00%
10
5.00%
0 4.00%
3.00%
-10
2.00%
-20 1.00%
1987/88 1988/89 1989/90 1990/91 1991/92 1992/93 1993/94 1994/95
0.00%
Business lending growth Business capex growth 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
Source: Deutsche Bank, ABS, RBA Source: Deutsche Bank, ABS
Figure 233: Commercial property prices spiked Figure 235: World GDP growth stalled sharply
500.00 5
%
450.00 4.5
400.00
4
350.00
300.00 3.5
250.00 3
200.00
2.5
150.00
100.00 2
50.00 1.5
0.00
1
Apr‐87
Apr‐94
Feb‐86
Sep‐86
Feb‐93
Sep‐93
Dec‐84
Jul‐85
Jun‐88
Mar‐90
Dec‐91
Jul‐92
Jun‐95
Oct‐90
Aug‐89
Nov‐87
Jan‐89
Nov‐94
May‐91
0.5
0
All Property Retail Office Industrial 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
Source: Deutsche Bank, IPD. Index base 100 at December 1984 quarter Source: Deutsche Bank, World Bank

The strong period of Australian economic growth was brought to a rather Figure 236: Non-accruals/GLAs peaked at 8% in FY92…
sudden end in 1990, driven by: 10%
9%
The RBA aggressively raising rates in order to fight inflation, which had 8%
increased to 7.5% in the second half of 1989. The official cash rate 7%
6%
peaked at over 18% in the second half of 1989, and was accompanied by
5%
mortgage lending rates of 16%-17%; and
4%
The stalling of global economic growth, which flowed through to 3%

2%
Australia. Of the 18 OECD countries of reasonable size and development,
1%
17 experienced a recession in the early 1990s5. 0%
ANZ CBA NAB WBC Average
Impact of the crisis
FY89 FY90 FY91 FY92 FY93 FY94
The combination of the factors identified above resulted in: Source: Deutsche Bank, Company Data
Australian GDP growth falling to 0% in 1991 and -3% in 1992; Figure 237: …while the bank sector ROE fell to -3%
Rising unemployment (with the unemployment rate rising to 12% in the 20%
years following the recession and banking crisis); and 15%
10%
An unwinding of the asset price bubble.
5%
The worsening conditions in the economy soon flowed through to the banking
0%
sector. This resulted in a significant increase in non-accrual loans during the ANZ CBA NAB WBC Average
crisis, as shown below. -5%
-10%
-15%
-20%
FY89 FY90 FY91 FY92 FY93 FY94 FY95
The bank sector average ROE fell to -3% in FY92 as a result, and bank sector
share prices suffered, most notably ANZ and WBC, which were hit the hardest
by the crisis. Eventually, these banks had to be recapitalized. Meanwhile, the
broader market index recovered much more quickly.
5
Macfarlane, 2006
Figure 238: ANZ and NAB share prices fell considerably The State Bank of South Australia, which was bailed out by the South
250 Australian government when it lost $3.3bn;
The Pyramid Building Society (the second largest at the time), which failed
200
due to the risky commercial loans it made and a run on deposits in early
150 1990. The Victorian Government eventually provided financial assistance
equal to 25 cents in the dollar to depositors, which cost over $900m;
100
The Western Australian Teacher’s Credit Society, which sought
50 government assistance and was bought by R&I bank; and
0 A number of other financial institutions, including merchant banks such as

2‐Jan‐89 2‐Jan‐90 2‐Jan‐91 2‐Jan‐92 2‐Jan‐93 Rothwell’s and Spedley’s, a mortgage trust (Estate Mortgage), and a
friendly society (the Order of the Sons of Temperance)7.
ANZ NAB WBC
Source: Deutsche Bank, IRESS. Note: Base index of 100 as at 2 January 1989. CBA excluded given it was not listed until September 1991
Measuring the crisis
Figure 239: The All Ordinaries recovered more quickly The banking crisis/recession was a fairly serious one, with the recession
2500
starting in the September quarter of 1990 and lasting until the September
quarter of 1991. During the recession, GDP fell by 1.7%, employment by
2000 3.4% and the unemployment rate rose to 10.8%. It took the economy 5
quarters to reach the level of nominal GDP achieved in June 1990.
1500
Nonetheless, the recession was not as deep as the 1982 recession, when
1000 GDP fell by 3.8%.
500 In the banking sector, problem loans to gross loans and advances peaked at
7% on average for the major banks. The role of deregulation in the crisis
0
cannot be over-emphasised. Treasury notes that the share of non-performing
1‐Dec‐88
1‐Dec‐93
1‐Dec‐89
1‐Dec‐90
1‐Dec‐91
1‐Dec‐92
1‐Sep‐89
1‐Sep‐90
1‐Sep‐91
1‐Sep‐92
1‐Sep‐93
1‐Jun‐89
1‐Jun‐90
1‐Mar‐89
1‐Jun‐91
1‐Mar‐90
1‐Jun‐92
1‐Mar‐91
1‐Jun‐93
1‐Mar‐92
1‐Mar‐93
loans to total assets peaked at 12% for the foreign bank sector in the early
1990s, twice the peak in the broader system8.
Source: Deutsche Bank, IRESS
The crisis has a clear impact on credit growth. Although remaining positive in
The problems were not limited to the major banks. During this period, there the early stages of the recession, growth in total credit then turned negative in
were a number of Australian financial institutions outside the majors which mid-1991, and it wasn’t until the June quarter of 1993 that credit growth
failed, including6: exceeded nominal GDP growth.
The State Bank of Victoria, which collapsed largely due to problem loans
held by its subsidiary Tricontinental. State Bank of Victoria was sold to
CBA in 1990; 7
Macfarlane, 2006
8
Australian Treasury, Economic Roundup Summer 2003-04, “Globalisation: the role of institution
6
Australian Government Treasury, Study of Financial System Guarantees, 2004 building in the financial sector”
Figure 240: Credit lagged GDP in the recovery Resolution of the crisis
2.5%
2.0%
The crisis saw recapitalization of both ANZ and WBC. WBC raised $1.2bn of
1.5%
equity capital in September 1992 in a 3-for-10 rights issue. Meanwhile ANZ
1.0%
raised $779m in a 1-for-5 rights issue, which was conducted in two
0.5%
instalments (the first in March 1992 and the second in March 1993).
0.0%
‐0.5%
Prudential regulation also picked up significantly after the crisis. Tougher
‐1.0%
capital requirements were applied to Australian banks, partly reflecting
‐1.5%
Australia’s adoption of the Basel accord, and the RBA also introduced
increasingly sophisticated risk management guidelines which led to
differentiated capital requirements for specific risks of institutions. Capital
Nominal GDP growth Total credit growth adequacy requirements were also introduced and tightened for non-bank
Source: Deutsche Bank, ABS deposit taking institutions over the same period. 9
Figure 241: ROE recovered to pre-crisis levels in FY95 While the economy took some time to recover, the recession had the effect of
25%
permanently lowering inflation and inflationary expectations in the economy.
The Australian economy went from a situation in the 1980s of inflationary
20% expectations above 10% to a low inflation environment. This was in part
15%
assisted by the adoption of inflation targeting by the RBA in 1993.
10% This focus on permanently lowering the inflation rate, along with continued
emphasis on financial deregulation and the opening up of the economy to
5%
international competition, has seen an unbroken period of continuous
0% economic expansion from 1991 to today. As shown below, this environment
was also assisted by the depreciation of the AUD.
‐5%
Returns for the majors also suffered as a result. It was not until FY95 that the
average ROE of the major banks reached pre-crisis levels, but the industry’s
attractive dynamics has seen a sustained period of strong returns since then.
9
Australian Government, March 1997, “Financial System Inquiry Final Report”.
Figure 242: Inflation fell dramatically post crisis Some years after the crisis, the Wallis Committee recommended in 1997 that
the Australian Prudential Regulation Authority (APRA) should be set up to
assume the responsibility for prudential regulation of the industry, replacing
the Reserve Bank of Australia.
Source: Deutsche Bank, RBA
Figure 243: AUD depreciation assisted the real economy

0.85
0.80
0.75
0.70
0.65
0.60
5‐Jan‐90 5‐Jan‐91 5‐Jan‐92 5‐Jan‐93 5‐Jan‐94 5‐Jan‐95
Source: Deutsche Bank, IRESS
This long period of economic expansion was achieved despite the Asian
financial crisis of the late 1990s and the global financial crisis which began in
2008. The expansive environment has been very supportive of the Australian
banking sector. In addition, continued strong profitability in the sector, and the
vigilance of the prudential regulator, has meant that we have not seen a repeat
of the 1991/92 crisis.

Sweden: Deregulation
promoted households, and housing co-corporations, to borrow as interest
payments were fully tax deductible. The combination of high inflation and fully
deductable interest, implied very low or negative real interest rates, all trough
and macro shocks the 1980s.
2. Financial crisis outline

behind 1990s bank crisis From the mid-1980s, overall lending growth virtually doubled from 14%
annually to 29%, in the years from 1985 to 1990. This credit expansion, drove
asset inflation and prices on Swedish Commercial Real Estate increased in
1. Bubble creation and collapse total by almost 140% in just 5 years, until they peaked in 1990. The Swedish
banking crisis started in 1991, and banks’ loan losses rapidly increased from
The Swedish, Norwegian and Finnish, financial crises in the late 1980s and
just 47bp in the preceding five years to a peak of 360bp in 1993.
early 1990s have strong similarities and patterns. In particular, most empirical
resarch agree that the deregulation of the Swedish credit market in the early
The crisis was triggered by a combination of domestic and international
and mid-1980s, as well as a negative real interest rates are key explanations
factors and events. First, interest rates started to rise, both internationally and
for the rapid credit growth, which drove asset prices to un-sustainable levels.
in Sweden as the Swedish Riksbank wanted to tame inflation. Sweden had a
In addition, to a bursting real esate bubble, the severity of the banking crises in
peg to the Deutsche Mark at the time, and as the pressure on the Swedish
both Sweden and Norway, was excarbated by a currency- and macro shock.
krona increased, due to high inflation and a deteriorating fiscal position, the
Riksbank was forced to at one point raise over-night interest rates to 500% to
The deregulation. The Swedish financial sector was heavliy regulated in the
defend the currency peg. The defence was unsuccessful and in November
late 1970s and early 1980s. Most notably, banks were subject to a) lending
1992, the peg was abandoned and in the next two months the Swedish krona
celings b) liquidity requirements c) caps on lending rates. In addition, Sweden
fell by 20%. Real estate companies had borrowed in Deutsche Marks and
had a currency regulation, which restriced foreigners and Swedish residents
invested in Swedish real estate, the devaluation triggered a number of
investments in foreign currency denominated financial securities and real
defaults, and significant loan losses, in the real estate sector. Another trigger
estate. The combined effect of these regulations and requirements fostered
point for the crisis was a Swedish tax reform in 1990/1991, which cut the tax-
inefficient banks, and crippled their basic function as suppliers of credit in the
shield on interest payments from 50% to 30%. Households debt service
economy. The deregulation started in early 1980s, and during the short period
ability deteriorated and the reform was a negative for house prices.
of only three years, between 1983 and a1985, the financial system was
transformed from being heavily regulated to a much more market oriented
system. By 1985, the strict liquidity ratios, lending caps and ceilings on lending 3. End of the financial crisis
rates were all gone. An explanation for the deregulation was the
In the autumn 1992, the Swedish state had nationalised Nordea and Gota
internationalisation of capital markets as well as the Swedish state’s need to
Bank, and issued capital guarantees to Första Sparbanken. Moreover, in
get access to a more developed fixed income market in order to fund the
September 1992, the government issued a blanket guarantee on all bank debt
growing budget deficit.
(not including equity), which covered all banks. Furthermore, Föreningsbanken
got in 1993, a capital guarantee from the state, should its capital ratios fall
In addition to deregulation the dynamics of the Swedish economy, set the
below 9%.
stage for rapid credit expansion. In the 1980s, the Swedish economy was
characterised by high inflation, weak fiscal disciplined and a number of
currency devaluations, which kept inflation expectations high. The tax system

A vehicle for bank resolution was created in May 1993, called Figure 245: Swedish credit losses increased 7x and peaked at 350bp in
Bankstödsnämnden (Banking support authority) and all in all, policy actions and 1992
the resolution regimes proved successful in stabilising the system. The
blanket guarantee on all bank debt, kept funding markets open for banks and Swedish banks' credit losses, yearly
there was a political consensus around the policy actions, which gave them 4.00%
credibility. However, the banks and the macro economy in Sweden, was
naturally helped by the 20% depreciation of the currency. After contracting by 3.50%
in total 5.2% in 1991-1993, the Swedish GDP expanded by 3.9% in 1994 and
3.00%
4.2% in 1995.
2.50%
The direct support to the banking system was SEK66bn (P. Englund, 1999,
2.00%
Oxford Review, Vol 15) and the final cost is estimated to SEK35bn, equal to
2.1% of Swedish GDP (Jennergren, Näslund 1997, Bankriser och dess 1.50%
hantering).
1.00%
Figure 244: Rapid credit expansion driving asset prices to unsustainable 0.50%
levels
0.00%

1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Source: Deutsche Bank, company data, SCB, Swedish Riksbank
Source: Deutsche Bank, Urban Bäckström, What Lessons Can Be Learned from Recent Financial Crisis? The Swedish Experience

US Sub-Prime Crisis
banks were undercapitalized when adjusting for all of the on and off-balance
sheet exposure they had to the mortgage market. Certain customers began
pulling funds from some banks that were viewed as less safe and questions
started to arise regarding bank liquidity. Many banks had been depending on
overnight borrowings (which had to be renewed every night) that were backed
Background
by the same MBS they were facing substantial losses on. As this market dried
The financial crisis was largely caused by excessive risk-taking and poor up, the solvency of certain banks came into question. The lack of transparency
regulatory oversight. When you get down to the details, the bursting of the of banks’ balance sheets and the interconnectedness among them magnified
housing bubble in 2007 was the primary catalyst behind the economic the problems. In total, FDIC insured banks incurred almost $600b of loan
collapse. losses during the crisis (through 1Q11), equal to the previous 18 years of
losses combined. Over 300 banks have failed since the start of the crisis.
A brief background on the financial crisis
Steadily decreasing interest rates from the early 1980s onward made buying a Other contributors to the crisis
home more affordable for many. Additionally, easy and available credit Credit ratings agencies: Triple-A ratings were given to many MBS which in
provided by the banks starting in the early/mid 2000s made this even easier. In hindsight, did not deserve them given the poor quality of the underlying
the early 2000s, banks began to hold onto a large percentage of the MBS they mortgages. Having a triple-A rating allowed for a different class of investor to
originated to capture the spread. With the demand for MBS rising in the US invest (some institutions required this rating to invest), providing a continuous
and abroad and supply of conforming loans limited, banks began to securitize flow of capital to underwrite new loans and MBS.
subprime loans. Additionally, the underwriting of low- or no-documentation
loans and adjustable rate mortgages (ARMs) began to increase, further Limited oversight: Regulators continued to rate firms they oversaw as safe in
lowering the quality of mortgages and MBS. Partly enabling the growth in the face of mounting troubles. Firms began taking on too much risk with too
MBS originations were Fannie Mae and Freddie Mac who also began little capital and too much reliance on short term borrowings. Helped by
increasing their holdings of risky loans. deregulation of the OTC derivatives market in 2000, regulators permitted the
growth of a massive shadow banking system (including repo lending, off-
The risk involved with MBS was largely overlooked due to the historical track balance-sheet entities, and use of OTC derivatives).
record of real estate, which for the most part increased in value over the
longer term and was at the time, in a very strong uptrend (property values in Derivatives: The signing of The Commodity Futures Modernization Act of
the US more than doubled from 2000 to 2007 based on the Case Schiller 2000 deregulated the over the counter (OTC) derivatives market. Without any
Index). Mortgage related debt also more than doubled over this period as a oversight, the OTC derivatives market rapidly expanded, growing to $673
result. The increase in demand for housing led to inflated prices and trillion in notional amount. The Financial Crisis Inquiry Commission noted that
homeowners became increasingly leveraged, many relying on continued OTC derivatives fueled the crisis in several ways including:
home price increases and future mortgage refinancing to avoid future default.
This did not occur and in 2007 when the mortgage market started to unravel, Credit default swaps (CDS) fueled the mortgage securitization pipeline.
housing prices began to plummet. The sharp housing downturn led to a deep CDS were sold to investors to protect against the default or decline in
global recession. value of mortgage securities backed by risky loans. CDS helped to expand
the market, further fueling the housing bubble.
The direct impact on banks
As housing prices declined, banks faced substantial losses on their MBS and
derivative exposure to this market. As losses continued, it became clear that

CDS were essential to the creation of synthetic CDOs (bets on the Figure 246: Historical Charge-off and NPA Ratios
performance of actual mortgage-related securities). The use of CDOs 4.0%
amplified the losses from the collapse of the housing bubble by allowing 3.5%
multiple bets on the same securities and helped spread them throughout 3.0%
the financial system.
2.5%
After the bubble burst, certain institutions, which had not been required to 2.0%
hold meaningful capital against credit protection positions, were bailed 1.5%
out for the US gov’t. Additionally, the existence of millions of derivatives 1.0%
contracts between systemically important financial institutions added to 0.5%
uncertainty and escalated panic, leading to the government assistance of 0.0%
these institutions. 1984 1987 1990 1993 1996 1999 2002 2005 2008 1Q11
Cost and Impact of Crisis NPL Ratio Net Charge‐Off Ratio
Source: FDIC
Nonperforming loans (NPLs) and net charge-offs (NCOs)

As a result of the financial crisis, banking sector NPLs (held at all FDIC insured Figure 247: S&P 500 vs. BKX
institutions) increased from 0.40% in 2006 to a peak of 3.7% in 4Q09 before 20%
falling to 3% in 1Q11. While NPLs increased in all loan categories before
0%
peaking in 4Q09, construction and development and other construction loans
‐8%
contributed dramatically to this growth. NCOs followed the same trajectory, ‐20%
increasing from 0.60% in 2007 to a peak of 3.0% in 4Q09. NCOs have fallen ‐40%
substantially since then to 1.8% in 1Q11, but remain elevated by historical ‐60% ‐51%
standards. See Figure 246. ‐59%
‐80%
‐84%
Bank stock performance ‐100%
As of 8/12, bank stocks (measured by the BKX) were down almost 60% since
the start of 2007 vs. down only 8% for the S&P 500. At the bottom, bank
stocks were down 84% from pre crisis levels vs. a 51% decline for the S&P. BKX S&P 500
See Figure 247. Source: Capital IQ
Return on assets
Return on assets (ROA) for the banking industry fell to 0.15%-0.20% in 2008-
2009 vs. 1.3% pre-crisis and 1% on average over the last 20 years given large
credit losses. Since then, ROAs have improved (to 0.7% in 2010), given
improving credit. While ROAs averaged 1% over the last 20 years, they were
much less than this in previous periods. ROAs could be lower than the recent
trend going forward as banks somewhat revert to more traditional banking
models given regulatory pressures. See Figure 248.

Return on common equity: Measuring the Crisis and Resolution

Return on common equity (ROCE) remains below its 20 year average of 11%
and pre-crisis levels of 12.5% reached in 2006. While we have seen much GDP and loan growth: Loan growth exceeded GDP growth from 2002
improvement in late 2010 and early 2011, we may not see longer term through 1Q08, but with the start of the financial crisis, new loan originations
ROCE’s return to pre-crisis levels given higher capital requirements, fell dramatically and banks began to shrink their balances sheets. Loan growth
specifically the likely requirement for banks to hold meaningfully higher at FDIC insured banks has gone from a low of -10% annualized in 3Q09 to
common equity (close to 10% for some of the larger banks vs. ~4% pre- +2.6% in 2Q11 (slightly exceeding GDP of 1.3% based on Fed data). Note the
crisis). Additionally, increased regulation has led to fee revenue pressures and spike in loan growth in 1Q10 was due to banks consolidating off-balance sheet
more of a shift to the traditional banking model, which could further reduce assets due to accounting changes. Loan growth was fairly close to GDP
returns. See Figure 249. growth in 4Q10 (2.9% vs. 2.3% for GDP) and slightly exceeded it in 2Q11
(2.6% vs. 1.3% for GDP). In dollar terms, GDP has yet to exceed its pre-crisis
Figure 248: Historical ROA peak reached in1Q08. See Figure 250 and Figure 251.
1.4%
2006: 1.3%
1.2% Figure 250: Loan vs. GDP growth
1.0% Average: 1.0% 20%
1Q10: Asset
0.8% 2010: 0.66% 15% 3Q07: 11% Consolidation
0.6% 10%
0.4%
5%
0.2%
0%
0.0% 2001 2003 2005 2007 2009 2Q11
1990 1994 1998 2002 2006 2010 ‐5%
Source: FDIC
‐10%
‐15%
Figure 249: Historical ROCE
16.0% Loan Growth GDP Growth
14.0% Source: FDIC, Federal Reserve, and Bureau of Economic Analysis
2006: 12.5%
12.0%
10.0% Average: 11.2%
8.0%
2010: 5.9%
6.0%
4.0%
2.0%
0.0%
1990 1994 1998 2002 2006 2010
Source: FDIC

Figure 251: Annualized US GDP (in chained 2005 dollars) Figure 252: Fed Funds Rate
13.5 6%
Trillions
13.0
5%
12.5
4%
12.0
3%
11.5
11.0 2%
10.5 1%
10.0
0%
1Q01 1Q02 1Q03 1Q04 1Q05 1Q06 1Q07 1Q08 1Q09 1Q10 1Q11
Source: Bureau of Economic Analysis Source: Deutsche Bank
Monetary and fiscal policy: The Federal Reserve has lowered the fed funds Figure 253: Inflation Adjusted 10 Yr Rate
rate to near zero at the end of July 2011 from a high of over 5% before the 4%
start of the crisis. Inflation adjusted 10-yr treasury rates have fallen by 210bps
to 0.30% over this same period (See Figures 252 and 253 below). Additionally,
3%
in late 2008, the Federal Reserve announced its program to purchase $1.25
trillion of agency MBS through the end of 1Q10 to provide support to the
mortgage and housing markets and to improve conditions in the financial 2%
markets. This program supplemented a similar program by the Treasury

Department. The purchasing of these securities increased market liquidity and 1%
also acted to keep mortgage rates low (by driving up prices/preventing their
decline). To further stabilize the economy, the Congress passed a stimulus bill 0%
with almost $1 trillion of spending increases and tax cuts.

Resolution the end of 2010 (based on projections). The results of the test were that the
total capital needed for the 19 BHCs to reach the SCAP capital buffer targets
Bank recapitalization and Basel 3 implications under the more adverse scenario was $185 billion, the majority needed to be
Emergency Economic Stabilization Act of 2008. In response to crisis, the in the form of Tier 1 common.
Congress passed this Act to improve the strength of financial institutions, and
enhance market liquidity. Under the Troubled Asset Relief Program or TARP The impact of TARP and SCAP on capital. While total capital levels were
(part of the Act), the Treasury was allocated $700b to purchase mortgage fairly stable from 1990-2008, common equity had accounted for a decreasing
backed securities from banks to increase banks’ liquidity positions. However, portion over this period and hybrid and preferred stock (included in Tier 1 and
these funds were never used for their intended purpose, and on October 14, Tier 2 Capital) became an increasingly large component before the 2008-2009
2008, $250b of these funds was allocated to purchasing senior preferred downturn (see Figure 254). Tier 1 and Tier 2 capital levels increased following
shares of banks. capital raises at the banks via the government TARP program starting in
November 2008 and Tier 1 common began to increase in 2009 given common
Starting in November 2008, banks in aggregate received $245b in funding raises to repay TARP and positive earnings.
through sale of preferred securities to the government (under TARP). Before
the results of the SCAP test in March 2009, Tier 1 capital was still the focus of Figure 254: Regulatory capital levels
regulators, so the preferred shares significantly helped banks’ capital (as well
as liquidity) positions. The Act also authorized a temporary increase in the 18%
FDIC deposit insurance limit to $250K until December 2009 to increase 16%
confidence in the system. The limit increase was subsequently extended 14%
through December 31, 2013 and made permanent under Dodd-Frank. 12%
10%
Supervisory Capital Assessment Program (SCAP). On May 7th, 2009 the
Federal Reserve released the results the SCAP test, which was an 8%
assessment of the adequacy of bank capital levels in response to the 6%
economic down-turn. The Fed determined that the 19 largest U.S. bank 4%
holding companies should have a capital buffer sufficient to withstand losses 2%
and allow them to meet the credit needs of their customers under conditions 0%
of a more severe recession than was anticipated at the time. 1991 1994 1997 2000 2003 2006 2010
The SCAP focused not only on the amount of capital banks held, but also on
Total Capital Ratio Tier 1 Ratio Tier 1 Common
its composition. The SCAP’s emphasis on Tier 1 common reflected the fact
that common equity is the first level of the capital structure to absorb losses, Source: SNL
Note: Data based on banks under coverageNew capital and liquidity requirements
offering protection to depositors and more senior parts of the capital
structure, lowering the risk of insolvency.
A SCAP capital buffer was determined for each bank and targeted a Tier 1
ratio of at least 6% and a Tier 1 common ratio of at least 4% at the end of
2010 under the more adverse macroeconomic scenario discussed below. A
BHC was required to have an additional SCAP buffer if it’s pro forma Tier 1
ratio was below 6% or if its pro forma Tier 1 common ratio was below 4% at

Basel 3 capital and liquidity. Final rules on capital and liquidity have not been Regulation of hedge funds: The Act requires hedge funds and private
written in the US, they will likely be guided by Basel 3 proposals which focus equity firms to register with the SEC and requires them to disclose trade
on the following: information to determine whether they pose systemic risk to the system.
Financial Stability Oversight Council: The Act also creates the Financial
Quality and level of capital: The Basel Committee recommended raising
Stability Oversight Council, made up of 10 federal regulators, which has
the quality of capital. Recommendations include an increase in the
the responsibility of identifying and responding to new risks in the
minimum common equity ratio from 2% to 4.5% and a capital
financial system and setting new minimum capital requirements. The
conservation buffer of 2.5%, bringing the total common equity
council also has the authority to break up or require financial companies to
requirement to 7%. Under the recommendations, capital requirements
divest assets if they pose danger to the financial system.
would be phased in from 2013 to January 2019. The Committee also
stated systemically important banks should hold an additional common Derivatives reform: The Act gives the SEC and Commodity Futures
equity buffer (we think 100-250bps), potentially bringing minimum Tier 1 Trading Commission (CFTC) the authority to regulate OTC derivatives and
common levels close to 10% for large banks. creates a central clearinghouse for them. Credit exposure from derivatives
transactions is now considered when determining banks lending limits.
Liquidity standards: The committee introduced minimum global liquidity
standards consisting of both a short term liquidity coverage ratio and a Durbin Amendment and Interchange Fees: The Act calls for the Federal
longer term, structural net stable funding ratio. Reserve to place limits on interchange fees, requiring fees to be
reasonable and proportional to the cost of processing transactions. The
Other regulatory changes
final cap was set at $0.22, which will reduce bank’s debti card
Dodd-Frank Wall Street Reform and Consumer Protection Act. The Act
interchange revenue by ~50% starting in 4Q11.
which passed in July 2010 was intended to prevent another financial crisis and
future bailouts, and provide protection to consumers from bank practices. The Securitizations: The Act requires companies that securitize mortgages to
Act creates/includes: retain 5% of the credit risk unless the underlying loans meet certain
criteria. This rule is currently being debated—the final criteria for
Consumer Financial Protection Bureau (CFPB): Housed at the Federal exemption are unknown.
Reserve, the CFPB is expected to protect consumers from deceptive
Deposit insurance: Dodd-Frank permanently increased the insured
bank practices.
deposit limit to $250k which puts an additional burden on banks through
Too big to fail: In attempt to end too big to fail, the Act requires banks to higher assessments (the increase to $250k was previously temporary)—
write living wills which will be used to safely liquidate failing banks. The this will affect all banks. Along with these changes, the FDIC changed its
Act also places limitations on how large banks could grow through assessment calculations on April 1, 2011, which shifts the burden of
acquisitions (with liability caps to complement deposit limits that are replenishing the insurance fund to large banks.
already in place).
Volker Rule: Prohibits proprietary trading and owning/investing in hedge
funds and private equity firms, which will likely lead to the sale of these
assets. While the details of the Volker rule have not yet been determined,
banks will have two years to comply with the rule once finalized. As
nontraditional businesses are sold off, banks will likely look for ways to
maintain their balance sheets which could lead to acquisitions.

South Korea: the credit card bubble

Figure 255: Datapoints before and after the crisis (t=2003)
t-5 t-4 t-3 t-2 t-1 t t+1 t+2 t+3 t+4 t+5
Macro
GDP local currency (W tr) 501 549 603 651 721 767 827 865 909 975 1,026
Average short term interest rate 15.2 6.8 7.1 5.3 4.8 4.3 3.8 3.7 4.5 5.2 5.5
Period end short term interest rate 7.7 7.3 6.9 4.9 4.9 4.4 3.4 4.1 4.9 5.8 3.9
Period end 10 year government bond yield NA NA 7.2 7.0 5.8 5.4 3.9 5.6 5.0 5.8 4.9
CPI 7.5 0.8 2.3 4.1 2.8 3.5 3.6 2.8 2.2 2.5 4.7
Current account surplus/deficit (%) 11.9 5.3 2.8 1.7 1.3 2.4 4.5 2.2 1.5 2.1 0.3
Credit Structure
Government debt (W tr) 80 99 111 122 134 166 204 248 283 299 309
Total credit in economy NA NA NA 883 1,025 1,050 1,064 1,145 1,317 1,475 1,733
Banking loans 200 250 311 357 472 538 566 614 699 804 917
- of which:
Mortgage loans NA NA NA NA NA 153 169 190 216 221 239
Retail/consumer loans ex mortgages NA NA NA NA NA 101 107 116 130 143 150
Other loans NA NA NA NA NA 285 290 308 353 440 528
Industry data
Top 5 banks share of banking sector assets 47.4% 50.1% 57.3% 57.3% 57.6% 56.3% 56.0% 60.7% 60.9% 60.1%
Non performing loans 12.9% 8.0% 3.4% 2.3% 2.6% 1.9% 1.2% 0.8% 0.7% 1.1%
Bad debt charge to loans 3.4% 2.8% 1.6% 1.5% 2.2% 1.3% 0.5% 0.4% 0.4% 0.9%
Bank sector equity to assets ratio 5.0 4.0 4.8 4.7 4.6 5.6 6.8 7.0 6.9 5.8
Bank sector ROE -16.7% -13.4% 13.6% 12.2% 3.8% 16.9% 20.3% 16.2% 16.2% 7.9%
Asset prices
Stock market index year end 562 1,028 505 694 628 811 896 1,379 1,434 1,897 1,124
Year on year change in average house prices -12.4% 3.4% 0.4% 9.9% 16.4% 5.8% -2.1% 4.0% 11.6% 3.1% 3.1%
Source: Deutsche Bank, BOK, FSS, Kookmin Bank

Background and chronology Surprise window dressing in large corporate group rattles the Korean
credit markets, increases awareness of rapid deterioration in credit card
The credit card bubble and bust in Korea, in summary, was the result of a operations
failure to contain the after-effects of a set of policies designed to stimulate the In February 2003, the Korean prosecutors found window dressing in a large
economy by jump-starting consumption. The Korean government was conglomerate group affiliate named SK Global. Money market funds with SK
suffering the aftermath of the Asian Financial Crisis in 1998 and needed to Group corporate debt in its portfolio faced immediate redemption pressure
prevent the rising unemployment and weak household income growth leading and the incident left the corporate debt market in Korea paralyzed. The
to a vicious cycle of weaker consumer spending and lower corporate profits. uncertainties in the financial markets sparked awareness about the
Hence, in September 1998, the Korean government announced a set of deteriorating asset quality of the credit card companies, and it became almost
consumer finance boosting measures, which included fiscal stimulus for the impossible for credit card companies to refinance maturing debt.
housing market and financial support for the credit card business. The
measures also included aggressive lowering of property transaction, In March 2003, to stabilize the financial markets the Korean government came
acquisition and capital gain tax in order to boost the property market. out with measures to curb concerns about credit card companies, mainly by
capital raisings by the major shareholders. The government followed suit with
In addition to the accommodative policy for consumer spending, financial another set of measures in April 2003 to stabilize the financial markets, this
institutions in Korea quickly changed their focus to retail banking from wholesale time deciding to support rollover of credit card company debt and increase the
banking after large corporate failures following the Asian Financial Crisis, leading size of capital raisings by the credit card companies. Such measures had a
to eased lending attitudes to households. The financial institution behaviour temporary effect of calming the financial markets.
change with the backdrop of an accommodative consumer spending policy led
to a big increase in household credit to W439tr as of end-2002 from W184tr as Asset quality deterioration intensifies, turmoil at credit card companies
of end-1998, with a CAGR of 24.3%. The household credit / disposable income and financial markets once again
ratio during the period surged to 107% from 56%, setting the stage for a big The stability in the financial markets didn’t last long, as the credit card
surge in credit card asset delinquencies ahead. companies started to see a surge in delinquencies and loan loss provisioning,
posted accumulated loss of W5.1tr during from January to September 2003 vs.
Figure 256: Korean household’s debt servicing capability and credit card accumulated profits of W1.4tr during the same period in 2002. Refinancing of
asset quality credit card debt became almost impossible and the companies started to
encounter serious liquidity problems. Money market funds faced severe
redemption pressure, causing trouble for the asset management industry.
140% 21%
120% 18%
100% 15%
Household credit Resolution of the crisis by recapitalization of the credit card sector and
/ disposable
income (LHS) credit recovery programs for low credit individuals
80% 12%
Credit card Given pressure from the government to recapitalize, mono-line credit card
companies turned to their major shareholder for fresh capital or pursued
delinquency
60% 9% (RHS)
Credit card NPL mergers with other affiliates, while bank owned credit card companies were
40% 6%
merged into the mother banks. As a result, major credit card companies such
ratio (RHS)
20% 3% as Kookmin Credit Card were merged by Kookmin Bank, KEB Credit Card was
0% 0%
merged by KEB and Woori Card was merged by Woori Bank. Samsung Card,
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
being a mono-line credit card company received financial support from group
Source: Deutsche Bank, BOK, FSS
affiliates to raise W938bn.

The government also urged the creditor financial institutions to provide credit withstand the credit losses, 3) overall household loan quality remained resilient
recovery programs for financially troubled individuals, while also setting up amid increasing household income and property prices, and 4) credit card
funds to provide government-run credit recovery programs. issuance requirements were made stricter to fully reflect repaying capability as
represented by the big decline in the number of credit card issued during 2003
Cost and impact of the credit card crisis and 2004.
The banking system’s NPL ratios were only marginally impacted as credit card Figure 258: Credit card usage and no. of credit cards issued
assets only accounted for only 3.4% of total bank loans. However, loan losses
and the resulting ROE saw visible impact as credit cost for the sector rose to
(m)
2.2% in 2003 from 1.5% in 2002 and ROE fell to 3.8% from 12.2% during the (W tr)
700 140
period. The banking sector share prices were actually at bottom when the 600 120
government came out with measures to stabilize the financial markets in April 500 100 Credit card
2003 and started to rebound. 400 80
usage (LHS)
300 60
No. of credit
Figure 257: Banking sector profitability before, during and after the crisis 200 40 cards issued
(RHS)
25% 4.5% 100 20
20% 4.0% ‐ ‐
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
15% 3.5%
10% 3.0%
Source: Deutsche Bank, BOK
5% 2.5%
ROE (LHS)
0% 2.0%
‐5% 1.5% Credit cost
(RHS)
‐10% 1.0%
‐15% 0.5%
‐20% 0.0%
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Source: Deutsche Bank, BOK, FSS
Measuring the crisis

Although the credit card bubble burst in Korea led to the highest credit cost
for the banking sector since the Asian Financial Crisis at 2.2%, the impact on
the overall banking system only lasted a year, with sector ROE, credit cost and
NPLs all returning to pre-crisis levels. The impact on the overall economy
seemed equally manageable, as nominal GDP growth rate picked up towards
pre-crisis levels a year after the crisis. We believe this was because 1) credit
card assets only accounted for less than 4% of total bank loans, 2) banks had
enough funds to recapitalize most of the largest credit card companies and

Appendix 1
Important Disclosures
Additional information available upon request
For disclosures pertaining to recommendations or estimates made on securities other than the primary subject of this research, please see the most
recently published company report or visit our global disclosure look-up page on our website at http://gm.db.com/ger/disclosure/DisclosureDirectory.eqsr.
Analyst Certification
The views expressed in this report accurately reflect the personal views of the undersigned lead analyst about the subject issuers and the securities of those issuers.
In addition, the undersigned lead analyst has not and will not receive any compensation for providing a specific recommendation or view in this report.
Tracy Yu/Yoshinobu Yamada/James Freeman/Chris Ellerton
Equity rating key Equity rating dispersion and banking relationships
Buy: Based on a current 12- month view of total share-holder return

(TSR = percentage change in share price from current price to 400 59 %
projected target price plus pro-jected dividend yield ) , we recommend 350
that investors buy the stock. 300
Sell: Based on a current 12-month view of total share-holder return, we 250 36 %
recommend that investors sell the stock 200
Hold: We take a neutral view on the stock 12-months out and, based 150
on this time horizon, do not recommend either a Buy or Sell. 100 16 % 5%
50 12 % 3%
Notes:
0
1. Newly issued research recommendations and target prices always
supersede previously published research. Buy Hold Sell
2. Ratings definitions prior to 27 January, 2007 were:
Buy: Expected total return (including dividends) of 10% or more over a Companies Covered Cos. w/ Banking Relationship
12-month period
Asia-Pacific Universe
Hold: Expected total return (including dividends) between -10% and
10% over a 12-month period
Sell: Expected total return (including dividends) of -10% or worse over
a 12-month period

Regulatory Disclosures
1. Important Additional Conflict Disclosures
Aside from within this report, important conflict disclosures can also be found at https://gm.db.com/equities under the "Disclosures Lookup" and "Legal" tabs.
Investors are strongly encouraged to review this information before investing.
2. Short-Term Trade Ideas

Deutsche Bank equity research analysts sometimes have shorter-term trade ideas (known as SOLAR ideas) that are consistent or inconsistent with Deutsche Bank's
existing longer term ratings. These trade ideas can be found at the SOLAR link at http://gm.db.com.
3. Country-Specific Disclosures
Australia and New Zealand: This research, and any access to it, is intended only for "wholesale clients" within the meaning of the Australian Corporations Act and
New Zealand Financial Advisors Act respectively.
Brazil: The views expressed above accurately reflect personal views of the authors about the subject company(ies) and its(their) securities, including in relation to
Deutsche Bank. The compensation of the equity research analyst(s) is indirectly affected by revenues deriving from the business and financial transactions of
Deutsche Bank.
EU countries: Disclosures relating to our obligations under MiFiD can be found at http://www.globalmarkets.db.com/riskdisclosures.
Japan: Disclosures under the Financial Instruments and Exchange Law: Company name - Deutsche Securities Inc. Registration number - Registered as a financial
instruments dealer by the Head of the Kanto Local Finance Bureau (Kinsho) No. 117. Member of associations: JSDA, Type II Financial Instruments Firms Association,
The Financial Futures Association of Japan. Commissions and risks involved in stock transactions - for stock transactions, we charge stock commissions and
consumption tax by multiplying the transaction amount by the commission rate agreed with each customer. Stock transactions can lead to losses as a result of share
price fluctuations and other factors. Transactions in foreign stocks can lead to additional losses stemming from foreign exchange fluctuations. "Moody's", "Standard &
Poor's", and "Fitch" mentioned in this report are not registered credit rating agencies in Japan unless “Japan” is specifically designated in the name of the entity.
Russia: This information, interpretation and opinions submitted herein are not in the context of, and do not constitute, any appraisal or evaluation activity requiring a
license in the Russian Federation.

Deutsche Bank AG/Hong Kong
Asia-Pacific locations
Deutsche Bank AG Deutsche Bank AG Deutsche Equities India Pte Ltd Deutsche Securities Inc.
Deutsche Bank Place Filiale Hongkong 3rd Floor, Kodak House 2-11-1 Nagatacho
Level 16 International Commerce Centre, 222, Dr D.N. Road Sanno Park Tower
Corner of Hunter & Phillip Streets 1 Austin Road West,Kowloon, Fort, Mumbai 400 001 Chiyoda-ku, Tokyo 100-6171
Sydney, NSW 2000 Hong Kong SEBI Nos: INB231196834 Japan
Australia tel: (852) 2203 8888 INB011196830, INF231196834 Tel: (81) 3 5156 6770
Tel: (61) 2 8258 1234 Tel: (91) 22 6658 4600
Deutsche Bank (Malaysia) In association with Deutsche Securities Korea Co. Deutsche Bank AG
Berhad Deutsche Regis Partners, Inc. 17th Floor, YoungPoong Bldg., Singapore
Level 18-20 Level 23, Tower One 33 SeoRin-Dong, One Raffles Quay
Menara IMC Ayala Triangle, Ayala Avenue Chongro-Ku, Seoul (110-752) South Tower
8 Jalan Sultan Ismail Makati City, Philippines Republic of Korea Singapore 048583
Kuala Lumpur 50250 Tel: (63) 2 894 6600 Tel: (82) 2 316 8888 Tel: (65) 6423 8001
Malaysia
Tel: (60) 3 2053 6760
Deutsche Securities Asia Ltd In association with In association with
Taiwan Branch TISCO Securities Co., Ltd PT Deutsche Verdhana Indonesia
Level 6 TISCO Tower Deutsche Bank Building,
296 Jen-Ai Road, Sec 4 48/8 North Sathorn Road 6th Floor, Jl. Imam Bonjol No.80,
Taipei 106 Bangkok 10500 Central Jakarta,
Taiwan Thailand Indonesia
Tel: (886) 2 2192 2888 Tel: (66) 2 633 6470 Tel: (62 21) 318 9541
International locations
Deutsche Bank Securities Inc. Deutsche Bank AG London Deutsche Bank AG Deutsche Bank AG
60 Wall Street 1 Great Winchester Street Große Gallusstraße 10-14 Deutsche Bank Place
New York, NY 10005 London EC2N 2EQ 60272 Frankfurt am Main Level 16
United States of America United Kingdom Germany Corner of Hunter & Phillip Streets
Tel: (1) 212 250 2500 Tel: (44) 20 7545 8000 Tel: (49) 69 910 00 Sydney, NSW 2000
Australia
Tel: (61) 2 8258 1234
Deutsche Bank AG Deutsche Securities Inc.
Filiale Hongkong 2-11-1 Nagatacho
International Commerce Centre, Sanno Park Tower
1 Austin Road West,Kowloon, Chiyoda-ku, Tokyo 100-6171
Hong Kong Japan
Tel: (852) 2203 8888 Tel: (81) 3 5156 6770
Deutsche Bank Global Research
Global Banking Report
September 2011
Global Disclaimer
The information and opinions in this report were prepared by Deutsche Bank AG or one of its affiliates (collectively "Deutsche Bank"). The information herein is believed to be reliable and has been obtained from public sources believed to be reliable. Deutsche Bank makes no representation as to the accuracy or
completeness of such information.
Deutsche Bank may engage in securities transactions, on a proprietary basis or otherwise, in a manner inconsistent with the view taken in this research report. In addition, others within Deutsche Bank, including strategists and sales staff, may take a view that is inconsistent with that taken in this research
report.
Opinions, estimates and projections in this report constitute the current judgement of the author as of the date of this report. They do not necessarily reflect the opinions of Deutsche Bank and are subject to change without notice. Deutsche Bank has no obligation to update, modify or amend this report or to
otherwise notify a recipient thereof in the event that any opinion, forecast or estimate set forth herein, changes or subsequently becomes inaccurate. Prices and availability of financial instruments are subject to change without notice. This report is provided for informational purposes only. It is not an offer or a
solicitation of an offer to buy or sell any financial instruments or to participate in any particular trading strategy. Target prices are inherently imprecise and a product of the analyst judgement.
As a result of Deutsche Bank’s March 2010 acquisition of BHF-Bank AG, a security may be covered by more than one analyst within the Deutsche Bank group. Each of these analysts may use differing methodologies to value the security; as a result, the recommendations may differ and the price targets and
estimates of each may vary widely.
In August 2009, Deutsche Bank instituted a new policy whereby analysts may choose not to set or maintain a target price of certain issuers under coverage with a Hold rating. In particular, this will typically occur for "Hold" rated stocks having a market cap smaller than most other companies in its sector or
region. We believe that such policy will allow us to make best use of our resources. Please visit our website at http://gm.db.com to determine the target price of any stock.
The financial instruments discussed in this report may not be suitable for all investors and investors must make their own informed investment decisions. Stock transactions can lead to losses as a result of price fluctuations and other factors. If a financial instrument is denominated in a currency other than an
investor's currency, a change in exchange rates may adversely affect the investment. Past performance is not necessarily indicative of future results. Deutsche Bank may with respect to securities covered by this report, sell to or buy from customers on a principal basis, and consider this report in deciding to
trade on a proprietary basis.
Unless governing law provides otherwise, all transactions should be executed through the Deutsche Bank entity in the investor's home jurisdiction. In the U.S. this report is approved and/or distributed by Deutsche Bank Securities Inc., a member of the NYSE, the NASD, NFA and SIPC. In Germany this report is
approved and/or communicated by Deutsche Bank AG Frankfurt authorized by the BaFin. In the United Kingdom this report is approved and/or communicated by Deutsche Bank AG London, a member of the London Stock Exchange and regulated by the Financial Services Authority for the conduct of investment
business in the UK and authorized by the BaFin. This report is distributed in Hong Kong by Deutsche Bank AG, Hong Kong Branch, in Korea by Deutsche Securities Korea Co. This report is distributed in Singapore by Deutsche Bank AG, Singapore Branch, and recipients in Singapore of this report are to contact
Deutsche Bank AG, Singapore Branch in respect of any matters arising from, or in connection with, this report. Where this report is issued or promulgated in Singapore to a person who is not an accredited investor, expert investor or institutional investor (as defined in the applicable Singapore laws and
regulations), Deutsche Bank AG, Singapore Branch accepts legal responsibility to such person for the contents of this report. In Japan this report is approved and/or distributed by Deutsche Securities Inc. The information contained in this report does not constitute the provision of investment advice. In
Australia, retail clients should obtain a copy of a Product Disclosure Statement (PDS) relating to any financial product referred to in this report and consider the PDS before making any decision about whether to acquire the product. Deutsche Bank AG Johannesburg is incorporated in the Federal Republic of
Germany (Branch Register Number in South Africa: 1998/003298/10). Additional information relative to securities, other financial products or issuers discussed in this report is available upon request. This report may not be reproduced, distributed or published by any person for any purpose without Deutsche
Bank's prior written consent. Please cite source when quoting.
Copyright © 2011 Deutsche Bank AG

1708 Raport DB PDF

Uploaded by

Copyright:

Available Formats

1708 Raport DB PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1708 Raport DB PDF

Uploaded by

Copyright:

Available Formats

September 2011

Research Team Contents

Page 2 Deutsche Bank AG/Hong Kong

Deutsche Bank AG/Hong Kong Page 3

Executive summary How to make money in this deleveraging environment

Page 4 Deutsche Bank AG/Hong Kong

Figure 2: Crises Compared

Deutsche Bank AG/Hong Kong Page 5

Figure 3: Index of House Prices: Developed Markets

Page 6 Deutsche Bank AG/Hong Kong

Figure 4: Bank CDS Prices

US UK Europe ex UK China Asia

Figure 5: Danger Map Scores: Developed Markets

Deutsche Bank AG/Hong Kong Page 7

Figure 6: Danger Map Scores: Emerging Markets

Figure 7: Summary of Stress Tests Brazil -67.7% 23.7% 10.8%

USA -33.3% 4.6% 11.8% India -23.0% 25.4% 8.0%

Australia -61.0% -16.4% 8.5% China -32.0% 11.9% 10.5%

Japan -22.3% -1.2% NA Turkey -18.5% 17.6% 16.1%

Hong Kong -13.0% 16.9% 11.2% Indonesia -73.6% -21.6% 15.9%

UK -32.4% 2.2% 8.2% Malaysia -43.3% -6.0% 6.0%

Sweden -68.1% -39.9% 8.6% Thailand -16.3% 20.9% 10.9%

France -46.0% -3.9% 10.0% Korea -29.0% 0.9% 8.8%

Germany NA NA NA Poland -47.5% -8.9% 15.2%

Page 8 Deutsche Bank AG/Hong Kong

Deutsche Bank AG/Hong Kong Page 9

Source: Tokyo Stock Exchange

Page 10 Deutsche Bank AG/Hong Kong

Source: Deutsche Bank Source: Deutsche Bank

Deutsche Bank AG/Hong Kong Page 11

Credit growth and

regulatory and market pressure. According to the Bank of England, global

Source: Deutsche Bank estimates

Deutsche Bank AG/Hong Kong Page 13

Page 14 Deutsche Bank AG/Hong Kong

Deutsche Bank AG/Hong Kong Page 15

Thus in a nutshell, while leverage in the banking system measured by the

Page 16 Deutsche Bank AG/Hong Kong

USA 0.8 1.8 3.1 2.9 2.1 1.5

Deutsche Bank AG/Hong Kong Page 17

emerging markets is that the undervaluation of currencies is pumping up asset

Page 18 Deutsche Bank AG/Hong Kong

Deutsche Bank AG/Hong Kong Page 19

Clearly, forecasting error on bank earnings estimates will be greatest for

Page 20 Deutsche Bank AG/Hong Kong

(US$bn) US Banks UK Banks Euro Area Other Mature Developed Total 16

Securities 4502 1625 6907 729 1728 15491 10

Total 12561 8369 22901 3970 7878 55679 8

Distribution 23% 15% 41% 7% 14% 100% 6

Source: IMF, Deutsche bank estimates

Figure 38: Key changes in income and expense items Europe

Deutsche Bank AG/Hong Kong Page 21

In a nutshell, because of the sensitivity of the growth outcome to loan loss

Figure 39: Key changes in income and expense items USA

Page 22 Deutsche Bank AG/Hong Kong

recessionary credit cycle USA

Australia 1991-93 6.1%

Deutsche Bank AG/Hong Kong Page 23

Page 24 Deutsche Bank AG/Hong Kong

We believe that peak provisions on a further downturn across the euro