J_Frait_prednaska_4_TestingResilience_final

advertisement
Testing resilience
of the financial system
Stress testing is simple!
•
the only things one needs are
•
•
a computer to be run by an experienced operator
a couple of friends to discuss assumptions and results
2
Stress testing in the CNB
•
•
•
•
for assessing resilience of the financial system the CNB conducts regular stress
tests of banks since 2003
three (slightly overlapping) stages in development of the stress testing framework
for banks
•
simple static stress testing/sensitivity analysis (2003-2006)
•
static stress testing based on (consistent) macroeconomic scenarios, satellite
models and interbank contagion (2005-2009)
•
dynamic model-based stress testing (2009++)
in paralel, the CNB develops since 2008 a liquidity stress testing model for banks
that is going to be integrated with the main stress testing framework
since 2007, the CNB conducts stress tests of insurance companies (market risk,
insurance-specific risks) and pension funds (market risk)
3
Publication of stress test results
•
•
•
•
the CNB was always very open in publication and communication of
stress test results
traditional means of publication is Financial Stability Report (since
FSR 2004 published in January 2005)
results first published in a special feature/article, since FSR 2007 in
the main text (chapter Financial Sector – part „assessment of the
financial sector‘s resilience“)
Since March 2009, stress tests conducted quarterly (for the CNB
macrofinancial panel) and since February 2010, results are regularly
published in the CNB website
http://www.cnb.cz/en/financial_stability/stress_testing/
4
Stress testing in the CNB
Stage I
Simple static stress testing/sensitivity analysis
(2003-2006)
FSR 2004, FSR 2005, FSR 2006
5
Simple static stress testing
•
•
•
•
•
methodology stems from the IMF FSAP approach, developed in co-operation with
the IMF (Martin Čihák)
for testing credit risk and market risk (interest rate risk and FX risk)
strongly top-down though based on „static“ balance sheets of individual banks and
assumptions how balance sheets would change if (a) interest rates, (b) exchange
rate, (c) NPL changed
impact horizon of 1 year
suitable for simulations of the impact of
• single shocks (sensitivity analysis) like increase of NPLs by 20%,
• ad hoc scenarios defined as combination of risk factors that have direct impact
on banks‘ balance sheets (interest rate, exchange rate, NPL)
6
Mechanics of the stress test I
Transmission of risk factors:
•
•
•
impact of a change (increase) in interest rates: change in net interest income (gap
analysis) plus re-pricing of debt securities (duration analysis)
impact of a change in exchange rate: change in value of FX-denominated assets
and liabilities (using data on net FX position) plus indirect effect on NPL (loans
denominated in FX)
impact of a change in NPL: increase in NPL leads to increase in loan loss provisions
(using information about banks‘ provisioning rate)
7
Mechanics of the stress test II
Other assumptions:
•
•
•
•
in the absence of shocks, banks are assumed to generate profit at the level of the
average of the last 5 years
profit is used to raise capital to the initial level of capital adequacy (if the profit is
sufficient to counterbalance the impact of shocks), the rest (if any) is distributed via
dividends
risk-weighted assets (RWA) after shock are calculated as initial RWA minus 80% of
the overal impact of shocks
results of the stress tests are presented in percentage points of the initial capital
adequacy
8
Ad hoc scenarios in the simple stress test
•
•
•
the CNB used so-called „historical scenarios“ I
and II, i.e. combination of shocks that mimic
past crisis (1997-1998) and past volatility of
variables (see the table taken from FSR 2004)
shocks can be alternatively calibrated for
example as 1 p.p. confidence level (roughly 3
standard deviations)
combination of shocks should be plausible
and reflect possible reaction of authorities and
markets (e.g.. central bank raises interest
rates to defend currency from further
depreciation etc.)
9
Presentation of results of an ad hoc scenario I
(FRS 2005, p. 79)
10
Advantages of simple stress tests I
•
•
can be used to quickly
assess resilience to
specific risks (sensitivity
analysis)
respond to questions like
„how much would the
interest rates have to
increase to get post-test
capital adequacy equal
to the minimum value of
8 % (see chart from FSR
2005)
11
Advantages of simple stress tests II
•
•
•
simple to design and run (xls, no models), in line with FSAP/other IMF missions
(facilitates communication with IMF missions)
can be run repeatedly and the results can be compared over time (see chart from FRS
2005)
serve as a necessary first step in developing more comprehensive framework
12
Stress testing in the CNB
Stage II
Static stress testing based on (consistent) macroeconomic
scenarios, satellite models and interbank contagion
(2005-2009)
FSR 2005, FSR 2006, FSR 2007, FSR 2008/2009
13
Basic building blocks
•
•
based on the static simple stress testing, i.e.
•
same risk factors (interest rate risk, FX risk, credit risk)
•
same transmission channels (impact on net interest income, repricing of
bonds, FX profit/losses, loan loss provisions)
•
same horizon of 1Y
•
same assumptions about profit, CAR etc. (but from FSR 2008/2009, preprovision income instead of profits used)
new features
•
a new risk factor – interbank contagion
•
explicit (consistent, i.e. model-generated) macroeconomic scenarios
•
satellite models to transmit changes in macro variables into risk factors
14
The framework
•
•
QPM model (or since late 2008 G3 DSGE model) generates both baseline forecast
(the official CNB forecast produced quarterly) as well as alternative „adverse“
macroeconomic scenarios
satellite models are credit growth model (ECM model of aggregated credit growth) and
credit risk models (corporate, households)
15
Transmission Channels of Credit Risk
•
•
•
dependent variable of credit risk models: 12M default rate (i.e. new bad loans over
initial portfolio)
12M default rate is also used by commercial banks; the Basel II „PD“ used for IRB
approach in credit risk should be „a long-run average of default rates“
model and explanatory variables
• Corporate Sector
• Merton model
• Macroeconomic shocks (explanatory variables); GDP growth, exchange
rate, inflation, debt
• Households
• Merton model + naive econometric models
• Unemployment rate, real interest rates, GDP
16
Credit Risk Modeling
• Macroeconomic credit risk model for the Czech and Germany
corporates were estimated (Jakubík and Schmieder 2008)
• Czech:
• German:
Variable
df t   (c  b 2 et  2  b 3p t 1  b 4,1 gdpt  b 5 debt t  4  b 6 dumt )
df t   (c  b1irt 1  b 4,1 gdpt  b 4, 2 indprod t 3  b 5 debt t 4 )
Notation
Lag
Czech case
German case
Coefficient Std. error Lag Coefficient Std. error
Constant
-3.060***
0.358
ir
NA
Nominal Interest Rate (b 1 )
e
-2
1.062***
0.323
Real Exchange Rate (b 2 )
-1
-4.850***
0.636
Inflation (b 3 )
p
gdp
0
-4.609***
1.079
GDP (b 4,1 )
indprod
NA
Industry production (b 4,2 )
debt
-4
3.006***
0.246
Credit-to-GDP ratio (b 5 )
dum
0
0.238***
0.043
Dummy variable (b 6 )
Significance level: **: Significant at 5% level; ***: Significant at 1% level;
-1
0
-3
-4
0
-2.6997***
2.2194***
NA
NA
-3.3677***
-0.8215***
1.0871***
0.0400***
0.07141
0.4919
0.328
0.1464
0.1213
0.0125
17
Credit Risk Modeling
•
Macroeconomic credit risk model for the Czech and German households were
estimated (Jakubík and Schmieder 2008)
•
Households models: less successful than for corporates, additional (socio-economic)
indicators may improve modelling
Variable
Notation
Lag
Czech case
German case
Coefficient Std. error Lag Coefficient Std. error
-2.224***
0.071
-5.5656***
0.1072
NA
NA
0
-5.7912***
0.9244
Constant
Household Income (b1)
inc
NA
Credit-to-GDP ratio (b2)
debthouse
NA
NA
NA
-4
5.7186***
0.2788
Unemployment Rate (b3,1)
u
-4
3.695***
0.846
NA
NA
NA
Real Interest Rate (b3,2)
r
-3
1.808**
0.596
NA
NA
NA
Significance level: **: Significant at 5% level; ***: Significant at 1% level;
18
Scenario building
•
•
•
•
•
possible to construct scenarios without a macroeconomic model, but
to achieve the highest possible consistency, using a macro model
(QPM, DSGE, VAR) is of advantage
scenarios should be of a type „low probability – high impact“, but
plausible and have some „story“ behind
should react to risks identified in risk assessment; in case of doublesided risk, opposite scenarios can be built (e.g.
appreciation/depreciation, increase/decrease in interest rates)
the story can be reflected in the name of the scenario (makes it easier
to remember); „sexy“ names are of advantage
use baseline scenario (official forecast) as benchmark; however,
problems with interpreting the results if the stress testing
model/models calibrated conservatively
19
Example FSR 2007: stress test scenarios
•
Three alternative model-consistent scenarios in FSR 2007 (scenarios for the year
2008
• A - safe haven (appreciation of currency)
• B - property market crisis (internal shock with direct impact on banks)
• C - loss of confidence (external shock – increase in risk aversion)
Calibration of baseline and alternative scenarios
(2008 averages)
Baseline Scenario A Scenario B Scenario C
Real GDP growth (%; y-o-y)
4.1
2.4
0.3
2.8
Inflation rate - CPI (%; y-o-y)
6.2
7.0
5.3
8.0
Unemployment rate (%)
6.0
6.3
6.7
6.3
1Y PRIBOR (%)
3.8
2.8
1.5
8.7
1)
...
CZK/EUR exchange rate
25.6
27.0
30.5
Source: CNB
Note: 1) In 2008, the baseline expects an correction of the record values initially and then a slight appreciation
20
FSR 2007: Bank Stress Test Scenarios
•
all the scenarios were defined primarily by the evolution (change) of key
macroeconomic indicators such as GDP, inflation, the unemployment rate, short-term
interest rates and the exchange rate
Scenario type and shock size in bank stress test
Scenario type
Baseline Scenario A Scenario B Scenario C
Change in CZK interest rates
-0,2 p.p. 0,1 p.p.
-0,9 p.p. 4,4 p.p.
Change in EUR interest rates
-0,8 p.p. -1,4 p.p. -0,4 p.p. -0,4 p.p.
Change in CZK/EUR exchange
rate (- appreciation)
Loan default rate
Total credit growth
Change in property prices
(+ rise, - fall)
-
-6.7%
-0.4%
20.1%
4.2%
16.4%
5.2%
9.9%
6.9%
14.6%
4.9%
4.9%
15%
0%
-30%
-5%
Note: Changes in parameters represent the difference between 2007
Q4 and 2008 Q1, or, in the case of the baseline, between 2007 Q4 and
the average for 2008.
21
FSR 2007:
Impact of Alternative Scenarios on the Banking Sector
•
Results of bank stress tests
(capital adequacy; % and p.p.)
Scenario type
Capital adequacy (CAR) 1)
Results for chosen scenario type
Overall impact of shocks (p.p. CAR)
Interest rate shock
Exchange rate shock
Credit shock
… households
… non-financial corporations
Interbank contagion2)
CAR before profit allocation
Profit allocation (p.p. CAR)3)
Post-shock CAR
Capital injection (% of GDP)4)
Share of banks with negative capital after shock 5)
Baseline Scenario A Scenario B Scenario C
2007
11.5
2007
11.5
2007
11.5
2007
11.5
-2.1
0.2
-0.1
-2.0
-0.5
-1.0
-0.2
9.4
1.8
11.3
0.0
0.0
-2.8
0.1
-0.2
-2.4
-0.5
-1.5
-0.2
8.7
2.3
11.0
0.1
0.0
-3.0
0.6
0.0
-3.3
-0.5
-2.0
-0.2
8.5
2.2
10.8
0.1
0.0
-6.3
-2.6
0.5
-3.6
-0.5
-0.6
-0.7
5.2
2.8
8.1
1.1
14.9
•
Notes:
1) CAR means the capital adequacy ratio defined in accordance with the relevant CNB regulations (in particular those
governing the capital adequacy of banks and other prudential business rules).
2) Test integrated with interbank contagion and expected level of loss given default (LGD) 100%
and chosen probability of the banks' failure (default) on the basis of the CAR.
Example of presentation of the
results
The results were interpreted as
follows:
• The banking sector seems
to be resilient to a wide
range of risks
• Only an extreme
macroeconomic scenario
would necessitate capital
injections to maintain
sufficient capitalization
3) The scenarios assume that in the absence of shocks each bank would generate profit (or loss) equal to the average for the previous five
years and that it would use any profit (income) as a first line of defence against a declining CAR.
4) The capital needed to ensure that each bank has a post-shock CAR of at least 8%.
5) Market share of banks with negative capital after the impact of the assumed shocks (as a percentage of total assets).
22
FSR 2007:
Impact of Alternative Scenarios on the Banking Sector
• alternative – graphic
– presentation of the
results
• scenario C would
have the strongest
impact on banking
sector
26
24
Baseline
Scenario A
Scenario B
Scenario C
Growth of total loans
(%)
22
20
____
____
____
____
18
16
14
Capital adequacy
(%)
12
10
8
6
4
Share of default loans
(%)
2
0
2006
2007
2008
Source: CNB
Note: Growth in total loans is defined as the average annual rate of growth. The share of new
non-performing loans (NPLs) relates to the estimation of the loan volume at the end of 2007.
23
Example of sensitivity analysis:
the role of real estate prices (FSR 2007)
•
•
•
•
Simple sensitivity analysis within a scenario stress testing possible; in FSR 2007, we
looked at sensitivity of banks to real estate prices
Radical assumption: increase in the share of NPL leads to a fall in real estate prices of
the same extent
Simple test for mortgage loans
Test demonstrated the banking
sector’s high resilience to a
mortgage loan portfolio shock
This is due to very conservative
LTV ratio around 50 %
(%; 2007)
12
8
4
0
0
10%
20%
30%
40%
Growth of default mortgages and decline in real estate prices
Pre-shock CAR
CAR after credit shock
CAR after credit shock and sale of collateral at a loss
CAR after credit shock and sale of collateral (incl. profit allocation)
Source: CNB
Note: Scenarios of additional defaults as 10-40% of mortgage loans
becoming default loans. Banks or clients would sell collateral at 90-60%
value.
24
Another sensitivity test within a scenario: the role of
interest rate risk (FSR 2007)
•
The sensitivity analysis - capital adequacy of the banking sector would fall below the
regulatory minimum if short-term interest rates rose by more than 4.4 percentage
points
Smooth change in interest rate for selected scenario
(capital adequacy; %; 2007)
16
14
12
10
8
6
4
2
0
-2
-4
0%
0.5%
1%
Banks
1.5%
2%
Insurers
2.5%
3%
3.5%
4%
4.5%
Pension funds
Source: CNB
Note: Calculation for scenario C.
25
Example FSR 2008/2009: macroeconomic scenarios
• Three scenarios reflecting the risks from the global financial crisis
• Europe in recession (= baseline prediction)
• Nervousness of the markets (a la „loss of confidence“, i.e. increase in
risk aversion)
• Economic depression (very large decline in GDP)
Stress-Test Scenarios
in CNB's Financial Stability Report (June 2009)
Europe in
recession
Development of key macroeconomic variables in 2009
Real GDP (%, y-o-y)
-2.4
Inflation (%, y-o-y)
1.2
Interest rate 1Y PRIBOR (%)
2.4
Exchange rate CZK/EUR
26.6
Nervousness of
markets
-3.9
1.7
4.6
28.8
Economic
depression
-6.2
1.3
2.6
27.8
26
FSR 2008/2009: capital adequacy looks satisfactory
even in large depression
Results of stress testing scenarios
(%; banking sector)
16
14
Capital adequacy development
(%)
12
10
8
6
Share of non-performing loans (%)
4
2
0
06/2007
12/2007
06/2008
12/2008
Europa in recession
Nervousness of markets
Economic depression
Source: CNB
06/2009
12/2009
• Horizon of stress tests
is just one year.
• In a longer horizon,
the NPL share
continues to grow and
capital adequacy
deteriorates further.
• Still, unless recession
is very long and very
deep, the banks
should manage
without public funds.
27
FSR 2008/2009: presentation of the results
Results of bank stress tests
(capital adequacy; % and p.p.)
Scenario type
Europe in
recession
Capital adequacy (CAR) at the end 2008
Overall impact of shocks (p.p. CAR)
Interest rate shock
Exchange rate shock
Credit shock
… households
… non-financial corporations
Interbank contagion
Income allocation 2/
Post-shock CAR
Capital injection (CZK billions) 3/
Capital injection (% of GDP) 3/
Number of banks (CAR below 8 %) 4/
Share of banks (CAR below 8 %) 4/
Number of banks with negative capital
Share of banks with negative capital 5/
5/
1/
Nervousness of
markets
Economic
depression
12.3
12.3
12.3
-3.2
1.3
0.0
-4.4
-1.3
-3.0
-0.1
2.2
11.3
8.0
0.2
4
8.2
0
0.0
-5.4
0.0
0.1
-5.4
-1.5
-3.1
-0.1
3.1
10.0
15.7
0.4
8
21.8
0
0.0
-5.0
1.2
0.1
-6.2
-1.8
-3.9
-0.1
2.4
9.7
15.5
0.4
4
5.0
1
5.2
1/ CAR means the capital adequacy ratio defined in accordance with the relevant CNB regulations
(in particular those governing the capital adequacy of banks and other prudential business rules).
2/ We assume that banks would generate income in all adverse scenarios which would be used to
strengthen the capital. The level of income for individual banks is estimated using the past
development of income and parameters of the scenario. Every bank allocates the income to reach
the starting CAR level.
3/ The capital needed to ensure that each bank has a post-shock CAR of at least 8%.
4/ Banks with post-shock CAR 0 - 8 %.
5/ Banks with post-shock negative capital.
• same style of
presentation
• information about the
capital injections
needed
28
Stress testing in the CNB
Stage III
Dynamic model-based stress testing
(2009++)
FSR 2008/2009; FSR 2009/2010
29
Problems with static stress testing
•
The 2005-2009 framework limited as regards its ability to
•
capture the effects of credit, interest and currency shocks over time in a
more dynamic way,
•
analyze the impact of shocks in a longer horizon than a one-year
horizon (up to two to three years),
•
estimate the pre-provision income as a function of both the
macroeconomic development and a bank’s business model,
•
be expressed in the variables used in current regulatory framework (PD,
LGD) and thus mimick the stress testing done by individual banks within
Pillar II of Basel II
•
capture pro-cyclical nature of current Basel II regulation,
•
integrate fully the funding liquidity shock within the macroeconomic
stress testing framework,
•
link the interbank contagion and second-round liquidity shocks to
development of the individual bank’s capital and liquidity conditions in a
non-linear way, and
•
capture potential two-way interaction between the banking system and
the macroeconomic environment (feedback effect).
30
Example of the „time“ problem: market vs credit risk
•
difference in time horizon between the effects of market and
credit risks
•
•
•
•
impact of a change in interest rates or other market variables (the
exchange rate or stock prices) on the balance sheets of financial
institutions is virtually immediate (revaluation of securities)
credit risk accumulates over a longer time frame (one to three
years) as loans gradually shift into the NPL category
the „Phase II“ CNB stress testing framework addressed this
discrepancy with a compromise assuming an impact horizon of
one year
macro variables of the projected year were averaged to
produce the „shock“ as the difference between initial and
average future value = underestimates peaks in possible crisis
(Lehman September 2008)
31
Example of the evolution of the impact of shocks
•
•
scenario „nervousness of markets“ from the FSR 2008/2009 assumed
losses due to unfavourable interest rate changes in some quarters, but
these losses are fully reversed in the following periods
this dynamics of the directional changes in the shocks over time generates
stress situations in the financial sector that cannot be captured by the
standard stress tests using averages for the entire test period.
32
Solution: move to „dynamic stress testing“
1.
2.
3.
modelling of the banking sector (see Aikman et al. 2009)
•
banks’ balance sheets would be modelled dynamically, for example
for each quarter, as they are hit by the individual shocks
•
this would allow the shock impact horizon to be extended, for
example to six to eight quarters
•
losses would accumulate gradually
further satellite models needed (for etc.)
•
pre-provision income
•
other risk parameters (property prices, LGD, yield curve)
threshold model for integration of liquidity shock and interbank contagion
•
if any of the key variables (e.g. the capital adequacy ratio)
overstepped a pre-defined threshold, other shocks would be
generated (e.g. interbank contagion, outflow of liquidity)
33
Scheme of dynamic stress testing
Network model
(interbank contagion)
34
Current framework of the dynamic stress tests
• CNB now performs stress tests with every new quarterly
macroeconomic forecasts (i.e. 4 times a year – February, May,
August and November)
• alternative macro scenarios: one scenario reflects actual CNB‘s
macroeconomist forecast, one or two adverse scenarios run in DSGE
model are outlined by the financial stability team together with
modelling division experts (14 variables used),
• the horizon is set to 8 quarters – for example August 2010 stress tests
performed on mid-2010 portfolios with August 2010 forecasts focused
on horizon 3Q2010 – 2Q2012.
35
Dynamic features of CNB‘s stress tests
• Tests are set as dynamic – for every item in assets, liabilities,
income and costs there is an initial state to which the impact of
shocks is added in one quarter and the results serve as the initial
state for following quarter
• this is repeated in next 8 quarters for which the prediction is
generated.
• Four risks are tested: credit risk, interest rate risk, currency (FX) risk
and interbank contagion
• Conservative calibration of stress test parameters (slight
overestimation of risks, slight underestimation of buffers)
36
Bringing the stress tests in line with Basel II
•
Pillar I: change in credit risk terminology/risk factors
•
•
•
•
explicit PD (default rates), LGD, EL (expected loss)
loan segments very close to Basel II segments (corporate, retail,
other)
for banks in IRB approach, application of Basel II formula to
determine capital requirements
Pillar II: exchange of views with banks on stress testing
methodology
•
•
adjustments in interest rate impact (use of derivatives, interest
rate sensitivity of current accounts etc.)
explicit (expert) modelling of yield curve
37
Credit risk I
•
the methodology is being continuously improved; the tests work with four separate
loan portfolios: non-financial corporations, households – consumer,
households – mortgages, other loans
Two impacts o credit risk:
1. Expected loss (EL)
•
PDxLGDxEAD
•
PD is a result of satellite models (dependent variable; smoothed default rate
df), LGD set expertly (or via simple models)
•
EAD is non-defaulted stock of exposures; total exposure modelled via credit
growth model(s)
2. Risk-weighted assets (RWA)
•
IRB formula using PD, LGD and EAD
•
not precise (non-linearity, not all banks have IRB approach for credit risk
management), but close to how banks behave
38
NPLs
NPL ratio - the ratio of non-performing loans to total loans
• product of PD/df, existing NPLs, stock of loans (L) and outflow of NPLs outof the
balance sheets
NPL(2)/L(2) = approx. [NPL(1) + L(1)*df - a*NPL(1)]/L(2)
• expert judgment/assumptions about NPL outflow (parameter a of around 15% in
a quarter):
• parameter a may change during bad times, very difficult to model
39
Illustrative example of credit shock impact: expected
loss/provisions, NPL and RWA
Initial state
Parameters
Exposure in bil. CZK
Default loans
Non-default
portfolio (NP)
50
1000
Loss (PD x LGD)
PD
NPL ratio
(quarterly)
4.8%
Impact calculation
LGD
3%
45%
in % NP
bil. CZK
1.4%
14
Calculation of
credit losses
Note: quarterly PDs, yearly PDs = 4 x 3% = 12%
Final state
Impact
on RWA
Exposure in bil. CZK
Default loans
Non-default
portfolio (NP)
NPL ratio
Capital requirements
(KP)
RWA
50 + 30 - 0,15x50
= 72,5
1000 - 30 = 970
6.9%
function (970; PD; LGD)
12,5 x KP
For simplicity: 0% credit
growth assumed
New NPLs (0,03 x 1000)
NPL outflow (assumed 15%
each quarter)
Parameter LGD
• CNB‘s LGDs: first expert estimate in July 2009 versus adjustment in 2010
"baseline"
45%
Corporates
Households
housing credit
consumer credit
Other clients
protracted recession
55%
10%
45%
45%
20%
70%
45%
LGDs for baseline
Corporates
Households
housing credit
20%
consumer loans
55%
Other clients
Property price index and LGD for house purchase loans
(loss of confidence scenario, FSR 2009/2010)
(2007 Q4 = 100; LGD in %)
120
60
110
50
100
40
90
30
80
20
70
10
60
03/08
09/08
03/09
09/09
03/10
Property price index
Source: CNB, CZSO, CNB calculation
09/10
03/11
09/11
0
03/12
LGD (right-hand scale)
45%
45%
Since May 2010 (FSR
2009/2010), simple models for
„elevated“ LGDs (role of GDP,
property prices and
unemployment)
Example of IRB formula
• Impact of macro stress tests on IRB minimum capital requirements
(CR) for a hypothetical portfolio (CR are measured in % of exposure)
• Taken from Jakubík and Schmieder (2008)
Stress
Scenario
HS 10%
Czech
Republic
HS 20%
HS 10%
Germany
HS 20%
Parameter
Corporate-PD (%)
Household-PD (%)
LGD (%)
Capital Requirements (%)
Corporate-PD (%)
Household-PD (%)
LGD (%)
Capital Requirements (%)
Corporate-PD (%)
Household-PD (%)
LGD (%)
Capital Requirements (%)
Corporate-PD (%)
Household-PD (%)
LGD (%)
Capital Requirements (%)
End 2006
portfolio
(unstressed)
3.5
2.59
45
7.82
3.5
2.59
45
7.82
1.43
0.115
45
4.66
1.43
0.115
45
4.66
Forecasted 2007 stress portfolio
PD stress only
(case 1)
PD and LGD stress
(case 2)
5.5
2.69
45
54
8.66 (+10.7%)
10.39 (+32.9%)
10.6
2.75
45
54
10.37 (32.6%)
12.45 (59.2%)
1.63
0.128
45
54
4.87 (+4.5%)
5.84 (25.3%)
1.85
0.148
45
54
5.09 (+9.2%)
6.11 (31.1%)
•
The quantiles of all
macroeconomic
variables change
by 10 percentage
points (moderate
stress scenario,
HS 10%) and 20
percentage points
(severe stress
scenario, HS
20%), respectively,
in the unfavourable
direction.
42
Credit growth, RWA & capital adequacy (CAR)
Capital ratio: model-based (negative) credit growth
(in %)
16
15
14
13
12
11
10
9
8
06/07
12/07
06/08
12/08
Baseline
06/09
12/09
06/10
12/10
06/11
Protracted recession
• PD, LGD – push RWA upwards
• Stock of exposures – push
RWA downwards
Capital ratio: 8% credit growth assumption
(in %)
16
15
14
13
12
11
10
9
8
06/07
12/07
06/08
12/08
Baseline
06/09
12/09
06/10
Protracted recession
• Potential „deleveraging“ leads
to higher CAR in worse
scenario (protracted recession
in July 2009 tests).
• Thus, in bad times, there are
two competiting drivers of RWA
12/10
06/11
• For comparison a scenario with
positive credit growth (and
higher PD, LGD): negative
impact on CAR confirmed (via
higher RWA)
How to work with pre-provision income, profits and capital
•
•
•
•
•
until June 2010 (FSR 2009/2010), pre-provision income was
expertly set at x % of average of past 2 years (x < 100%, thus
additional stress applied in the sense of lower intermediation
activity)
during 1H2010, a simple model of pre-provision income was
estimated (the main determinants: nominal GDP, yield curve,
NPLs and capital adequacy)
profit/loss is generated using the pre-provision income and the
impact of shocks
regulatory capital is adjusted every 2Q to get back to initial CAR
thus, a P/L account and balance sheet of all banks generated
every quarter = possible to cross-check with reality later on
44
Modelling pre-provision income
•
•
comparison of model estimation versus expert setting of pre-provision income
conservative estimation – estimate of the model minus 1 stdev of growth
Outturn versus model estimate of adjusted operating profit on the past
(quarterly values in CZK billions; seasonally adjusted)
Estimate of adjusted operating profit for each scenario
(CZK billions; seasonally adjusted)
30
25
25
20
20
15
15
10
10
5
0
06/08
5
0
12/03
12/04
12/05
12/06
Model estimate on past
Source: CNB, CNB calculation
12/07
12/08
Outturn
12/09
Baseline Scenario
12/11
06/11
12/10
06/10
12/09
06/09
12/08
Return of Recession
Loss of Confidence
Original estimate for Baseline Scenario: 90% of average
Original estimate for Return of Recession: 80% of average
Original estimate for Loss of Confidence: 70% of average
Source: CNB, CNB calculation
45
Net income, P/L and capital adequacy: an example
•
For final evaluation of banks‘ resilience capital adequacy is
estimated.
Link between shocks impact and capital adequacy must reflect
•
•
•
•
(net) income generated by banks even under stress,
asymmetric treatment of profits in calculation of regulatory capital,
topping up of regulatory capital (set for 2nd calender quarter every year).
Initial state
Estimate of P&L over quarter
Final state
Regulatory
capital
RWA
CAR
Loss from
shock impact
Net
income
P/L
Regulatory
capital
RWA
CAR
Example 1
100
1000
10.0%
20
30
+10
100
1020
9.8%
Example 2
100
1000
10.0%
40
30
-10
90
1020
8.8%
Regular cross-check of the stress testing framework I
•
•
regular consultations with commercial banks on stress testing methodology
project of „joint stress tests with selected banks“
•
basically bottom-up stress tests – CNB gives the increase in risk
parameter PD, banks themselves calculate the impact
•
since summer 2009, currently third round completed
•
aggregate results published in the FSR 2009/2010
Approximate rise in PD for individual portfolios
(EAD weighted; %)
Actual situation as of
31 Dec. 2009
Baseline
scenario
Adverse
scenario
PD
(%)
LGD
(%)
PD
(%)
PD
(%)
Corporate exposure categories
2,65
41,34
3,62
5,62
Large enterprises
1,77
41,19
2,42
3,76
Small and medium-sized enterprises (SMEs)
Specialised lending
Retail exposure categories
Retail-assessed SMEs
Loans for house purchase
Other loans to individuals
3,54
2,95
3,00
3,49
2,15
4,48
40,66
44,00
33,08
45,78
19,76
53,93
4,81
4,12
3,54
4,89
2,36
5,29
7,43
6,48
4,48
7,60
2,80
6,48
Source: CNB
47
Regular cross-check of the stress testing framework II
•
verification of the models and assumptions (over time, banking sector changes thus the stress
testing framework should react as well - Basel II, use of derivatives etc.)
•
Geršl, A. – Seidler, J.: Stress test verification as part of an advanced stress-testing
framework. CNB, FSR 2009/2010
•
use baselines, but assymetric assessment needed (better to overestimate risks than
underestimate)
•
conservative calibration of models needed
Verification of CAR estimate
(CAR in %; estimate for 1-year horizon)
15
14
13
12
11
10
9
8
04/Q4
05/Q3
Reality
06/Q2
Prediction
07/Q1
07/Q4
08/Q3
09/Q2
Prediction – known macro
48
Presentation of results:
FSR 2009/2010, aggregate results
•
A slight change in the way of presentation
•
as a simplified profit/loss account
•
capital injection needs expressed verbally in the text
Impact of the alternative scenarios on the banking sector
Baseline Scenario
2010
2011
Return of Recession
2010
2011
Loss of Confidence
2010
2011
CZK billions
-40,6
-26,1
-56,1
-48,2
-48,1
-68,8
% of assets
-1,0
-0,6
-1,3
-1,1
-1,1
-1,7
4,8
0,1
-1,8
0,0
0,6
0,0
-1,5
0,0
-17,9
-0,4
2,3
0,1
-0,1
0,0
0,0
0,0
-0,4
0,0
-1,3
0,0
-0,3
0,0
-1,6
0,0
CZK billions
% of assets
83,3
2,0
90,6
1,9
75,3
1,8
59,2
1,4
85,5
2,0
70,8
1,7
CZK billions
% of assets
47,3
1,1
62,6
1,3
19,4
0,5
8,3
0,2
19,1
0,5
2,6
0,1
Expected credit losses
Profit/loss from market risks
CZK billions
% of assets
Interbank contagion
CZK billions
% of assets
Earnings for covering losses
(adjusted operating profit)
Pre-tax profit/loss
Source: CNB, CNB calculation
49
Presentation of results:
FSR 2009/2010, capital adequacy and NPLs
•
In quarterly publication, we present (see the CNB website)
•
charts on main macro variables (GDP, inflation, exchange rate and 3M interbank
rates),
•
charts on NPLs development and
•
chart on capital adequacy development
In FSR, further suplementary charts available (such as provisioning etc.)
•
Risk costs of the banking sector in each scenario
(provisioning as % of gross loans for given year)
Capital adequacy ratios in each scenario
(%)
16
3,5
15
3,0
14
2,5
13
2,0
12
1,5
11
1,0
10
0,5
0,0
9
8
03/08
2009
09/08
03/09
09/09
Baseline Scenario
03/10
09/10
03/11
09/11
Return of Recession
Baseline
Scenario
Return of
Recession
Loss of
Confidence
03/12
2010
2011
Loss of Confidence
Source: CNB, CNB calculation
Source: CNB,CNB calculation
50
Ad-hoc tests in FSR 2009/2010
•
Test of concentration of credit portfolios
•
collapse of three largest borrowers in each bank
Test of decline in value of certain asset class
•
decline in the value of exposures to Greece, Spain, Portugal and Italy by 50 % or
100 %
•
Impact of the ad-hoc "propagation of Greek crisis" test
(in the Loss of Confidence scenario)
(%)
Impact of the collapse of the three largest debtors of each bank
(in the Loss of Confidence scenario)
(%)
3
6
14
12
10
8
6
4
2
0
5
4
3
2
1
0
Without collapse of With collapse of three With collapse of three
three largest debtors
largest debtors
largest debtors
(LGD=45%)
(LGD=100%)
12
3
11
2
2
10
1
9
1
0
8
Original Loss of
Confidence
With losses from
With losses from
exposures to risky
exposures to risky
countries (LGD = 50%) countries (LGD =
100%)
Credit and market losses in 2010 (% of assets)
Loan losses in 2010 (%)
End-2010 CAR (%; right-hand scale)
Source: CNB
End-2010 CAR (%; right-hand scale)
Source: CNB, CNB calculation
51
Reverse stress test
•
•
Reverse logic: we are looking for values of parameters/macro developments that would
bring the banking system as a whole to the limit of 8 % CAR
In the Czech Republic, the GDP growth would have to be around -6.5% in 2011 to
cause a serious problem for banks
Results of reverse stress test
(%; for Return of Recession scenario)
20
15
10
5
0
-5
-10
03/08
09/08
03/09
09/09
03/10
09/10
03/11
Original GDP path
Adjusted GDP path
CAR: original
CAR: according to adjusted GDP path
CAR regulatory minimum
Source: CNB, CNB calculation
09/11
03/12
52
References
•
http://www.cnb.cz/cs/financni_stabilita/zatezove_testy/
•
ZFS 2009/2010
• Verifikace zátěžových testů jako součást pokročilého rámce zátěžového
testování
• Procykličnost finančního systému a simulace „feedback“ efektu
•
ZFS 2006,
• Vývoj kreditního rizika a zátěžové testování bankovního sektoru v ČR
53
Download