Transcript TITEL - VBA beleggingsprofessionals

The importance of measuring
credit risk
Beroepsvereniging van Beleggingsprofessionals
21 april 2008
Tom van Zalen
Credit Risk
Agenda
•
•
•
•
What is credit risk?
The importance of credit risk
Modeling credit risk
A link to the recent credit crunch
Credit Risk
2
2
Credit Risk
Agenda
• What is credit risk?
– Definition
– Credit risk drivers
– Systematic versus non-systematic risk
Credit Risk
• The importance of credit risk
• Modeling credit risk
• A link to the recent credit crunch
3
3
• Credit risk, a definition:
Credit risk is the risk of loss due to a debtor's non-payment of
a loan or other line of credit (either the principal or interest
(coupon) or both).
But also:
4
4
Credit Risk
The risk of value losses following from a change in external
credit factors.
Credit Risk
What is credit risk?
• External credit factors:
–Micro
• Individual risk single debt instrument = status quo firm
reflected in rating (and thus the credit spread)
5
Credit Risk
–Macro
• Collective risk fixed income portfolio = status quo
economy reflected in business cycle
Credit Risk
What is credit risk?
5
• Micro: determinants rating:
–Liability risk ~ volume debt versus equity
–Asset risk ~ volume tangible or intangible
–Cash flow risk ~ e.g. profitability, sales, repayment capacity
Credit Risk
What is credit risk?
• Macro: determinants business cycle:
6
Credit Risk
–Inflation and economic growth: Y = C + I + G + T
• Market risk is aggregated liquidity and credit risk
• Market risk is systematic or not diversifiable
6
Credit Risk
What is credit risk?
225
200
175
150
125
100
Credit Risk
75
50
25
0
-25
Apr-99
Oct-99
Apr-00
Oct-00
Apr-01
AAA
Oct-01
Apr-02
AA
Oct-02
A
Apr-03
Oct-03
Apr-04
Oct-04
BBB
7
7
Credit Risk
What is credit risk?
• Yield spreads largely depend upon rating, as a
proxy for credit risk
–Lower ratings face higher yield spreads
• Convex relation: lower ratings face relative higher
spreads
8
8
Credit Risk
–Systematic market risk is significant
• Lower ratings face relative higher non-systematic credit
risk
• Cyclical behaviour credit risk (credit cycles)
• Counter cyclical dependence (higher correlation during
crashes)
Credit Risk
Agenda
• What is credit risk?
• The importance of credit risk
– Credit risk in the Euro-area
– Market participants
Credit Risk
• Modeling credit risk
• A link to the recent credit crunch
9
9
• Market funding offsets bank lending
–Monetary integration = Euro ~ Liquidity
• Deregulated capital of institutional investors goes Europe
• Sovereigns face lower deficits due to disciplinary rules
Brussels
• Corporate entities go public more easy
–Des-intermediation bank ~ Credit risk
Credit Risk
Credit risk in the Euro-area
Credit Risk
• Financial regulation encourages credit risk
management
–Central banking = Basel II / Solvency II
–Accounting = IFRS
10
10
• Conclusion:
Bank-orientated economy with a small financial market focused
on sovereigns … becomes market-orientated economy with a
large financial market focused on corporate entities.
Financial
Government
Industrial
Volume (MM)
400,000
80%
320,000
60%
240,000
40%
160,000
20%
80,000
0%
0
1998
1999
2000
2001
2002
2003
11
11
Credit Risk
100%
1997
Credit Risk
Credit risk in the Euro-area
• Conclusion:
Introduction Euro eliminates foreign exchange risk, which has
caused intensified focus upon credit risk. Diversification over
rating classes has improved, although the average rating
decreased and may explain higher price volatility.
AAA
AA
A
BBB
Issuance (#)
400
80%
320
60%
240
40%
160
20%
80
0%
0
1998
1999
2000
2001
2002
2003
12
12
Credit Risk
100%
1997
Credit Risk
Credit risk in the Euro-area
• Banks (traditionally)
–Pricers of risk (loan originations)
–Sellers of risk (securitization = credit risk transfer of higher
rated bonds)
• High-rated homogeneous asset-backed securities (e.g.
mortgages)
• Medium-rated heterogeneous collateralized debt
obligations (=tranching & structured)
Credit Risk
Market participants
Credit Risk
• Asset managers
–Traders of risk
–Buyers of risk (investment management)
13
13
Credit Risk
Market participants
• Hedge funds
–Traders of risk (zero-position = arbitrage = long/short
strategies)
–Sellers & buyers of risk
• Private equity
14
14
Credit Risk
–Pricers of risk
–Sellers of risk (funding using lower-rated by issuing highyield bonds)
Credit Risk
Market participants
Buyers
Traders
Sellers
Hedge funds
Investment
grade
$ Premiums
Institutional
Instituional
Investors
Banks Banks
$ Savings
$
Investors
Credit Risk
Non-investment
grade
Private equity
Private Equity
Accounting
“Hold-to-maturity”
Trading portfolio
15
15
Credit Risk
Agenda
• What is credit risk?
• The importance of credit risk
• Modeling credit risk
– Expected loss
– Unexpected loss
Credit Risk
• A link to the recent credit crunch
16
16
• Credit risk is the probability of default (PD) of a loss
given default (LGD) due to changes in external
credit factors
• Follows from credit loss distribution
Credit Risk
• Measured by:
Credit Risk
Modeling credit risk
– Expected loss (μ)
– Unexpected loss (σ)
17
17
Credit Risk
Modeling credit risk- Expected loss
• Credit loss distribution function
Frequency
UL
Credit Risk
SD
EL
Risk Capital (unexpected loss)
3 bp
Economic capital
Portfolio Value
Expected
Promised
μ =EL
0
Value
Value
18
18
• Expected loss
– Measures the expected loss on a (portfolio of) loans given
the characteristics of the counterparty and the loan
conditions and the presence of collateral.
– Is the μ of the credit loss distribution.
– Credit spread = E[L] + liquidity spread
19
19
Credit Risk
Expected loss = probability of default x loss given default
E[L] = PD x LGD = % x % = %
Credit Risk
Modeling credit risk- Expected loss
• Probability of default
– Probability that a firm will default on its payment obligations
(e.g. coupon payments, principal repayment) within one
year.
– Often follows from rating
Credit Risk
Modeling credit risk – Expected loss
Credit Risk
20
20
14
5
GDP grow th
4
10
3
8
2
6
1
4
0
2
0
1970
-1
1975
1980
1985
1990
1995
2000
GDP Growth
Unemployment Rate
• Number of defaults
depends on the state
of the economy.
12
-2
2005
Year
8000
Besloten Venootschap
One-man Business
Individuals
Total
Credit Risk
10000
Bankruptcies
• Number of defaults
varies over time;
6
Unemployment Rate
Credit Risk
Modeling credit risk – Expected loss
6000
4000
2000
0
1970
1975
1980
1985
1990
1995
2000
Year
21
21
2005
Credit Risk
Modeling credit risk – Expected loss
• Loss given default
– The fraction of the outstanding loan that will not be
recovered once default occurred.
– Influenced by:
• Collateral
• Guarantees
Credit Risk
• Value of collateral may be correlated with the
occurrence of default:
– Example: commercial real estate mortgages
– “Haircuts” provide a correction for this issue
22
22
• Unexpected loss
23
23
Credit Risk
– If the realized credit loss would always equal its expected
value, then there would be no risk.
– In practice however, the credit loss is stochastic in nature
and thus risk arises.
– The possible deviation from the expectation is risk and is
measured by the standard deviation of the loss distribution.
– Unexpected loss is the σ of the credit loss distribution
Credit Risk
Modeling credit risk – Unexpected loss
• Default occurrence
– Occurrence of default follows a binomial distribution:
• With probability PD a default will occur
• With probability 1 – PD no default will occur
Credit Risk
Modeling credit risk – Unexpected loss
– For a portfolio with n loans, all having the same PD, the
total number of defaults is distributed as follows:
Credit Risk
# defaults ~ Binomial(n, PD):
µ = n x PD
σ2 = n x PD x (1 – PD)
24
24
• Binomial distribution for different values of n:
25
25
Credit Risk
• According to the central limit theorem, for large n,
the binomial distribution will converge to a normal
distribution.
Credit Risk
Modeling credit risk – Unexpected loss
• Loss given default:
– For a long time assumed constant due to:
• Complexity reasons
• Little effect to loss distribution compared to uncertainty
in the default event.
– Random variable with values: 0% < LGD < 100%
Credit Risk
• Is modeled using a Beta distribution:
Credit Risk
Modeling credit risk – Unexpected loss
– Distribution can be bound between two points
– Distribution can have a wide range of shapes
26
26
• Beta distribution:
– Two shape parameters: α and β
– B(α,β)
Credit Risk
Modeling credit risk – Unexpected loss
Credit Risk
27
27
• Credit loss distribution: from a single loan to a
portfolio of loans.
– E[L] is additive
– U[L] is not! Correlations need to be taken into account.
Credit Risk
Modeling credit risk – Unexpected loss
– Consider a portfolio that contains two loans, x & y with
corresponding portfolio weights wx and wy:
Credit Risk
2
σportfolio
  x2σ2x  y2σ2y  2 x yσ xσ yρxy
28
28
Credit Risk
Agenda
• What is credit risk?
• The importance of credit risk
• Modeling credit risk
• A link to the recent credit crunch
Credit Risk
– Structured finance
– Correlations
29
29
• Structured finance
– Pooling of assets and the subsequent sale to investors of
tranched claims on the cash flows backed by these pools.
– Key aspect of tranching:
• Create one or more classes of securities whose rating is
higher than the average rating of the underlying asset
pool.
30
30
Credit Risk
– Characterized by:
• Pooling of assets
• De-linking of credit risk
• Tranching of liabilities
Credit Risk
A link to the recent credit crunch
• A structured finance transaction in figure:
Senior
Funds
SPV
Funds
Originator
Credit Risk
A link to the recent credit crunch
Mezzanine
Traded assets
Assets
Liabilities
Tranches
Investors
• The original credit risk is distributed in the economy
and crops up everywhere.
31
31
Credit Risk
Junior
• Tranching is made possible by imperfect correlation
between the assets in the original asset pool.
• A diversified pool of risky assets is expected to have
a relatively predictable return pattern.
– E[L] of original asset pool = E[L] of total tranched pool
– U[L] of original asset pool = U[L] of total tranched pool
• E[L] and U[L] are portioned and attributed to the
different classes in the tranched pool.
32
32
Credit Risk
• Tranched pool is structured in such a way that:
Credit Risk
A link to the recent credit crunch
Last loss
Lower
Lowest risk expected
yield
First loss
Higher
Highest risk expected
yield
BB
BB
BB
BB
BB
BB
BB
BB
BB
AAA / AA
BB
BB
BB
BB
BB
BB
BB
Pool of
mortgage
loans
AA / A
BBB/ BB
BB / B
Unrated
Source: CSMA
33
33
Credit Risk
BB
BB
Credit Risk
A link to the recent credit crunch
Credit Risk
A link to the recent credit crunch
• It’s all about correlations!!
I
Credit Risk
34
34
• Yield spreads (Jan ‘04 – Jan ‘08):
Credit crunch
1.500
Madrid
bombings
1.250
1.000
Credit Risk
A link to the recent credit crunch
0.750
Credit Risk
0.500
0.250
0.000
-0.250
sprBBB
35
Jan-08
sprA
Jul-07
Jan-07
sprAA
Jul-06
Jan-06
Jul-05
Jan-05
Jul-04
Jan-04
sprAAA
35
• Yield return correlations:
0.300
1.000
0.900
0.250
0.800
0.700
Credit Risk
A link to the recent credit crunch
0.200
0.600
0.150
0.500
0.100
0.300
0.200
0.050
0.100
0.000
0.000
Nov-07
(AA,BBB)
May-07
Nov-06
May-06
Nov-05
May-05
(AA,A)
vol (DJ Eurostoxx 50)
36
36
Credit Risk
0.400
• What happened?
– US economy tightened and housing prices declined
– Correlation between high rating yield returns and the
market volatility is always close to one ~ AAA/AA can serve
as a proxy for the riskiness of the market.
– Correlation between individual loans must then also
increase.
– Credit risk in pool based on assumed low correlations ~
credit risk is underestimated!
37
37
Credit Risk
– Correlation between high rated (= market) and lower rated
was low but started to increase.
Credit Risk
A link to the recent credit crunch
• Could the recent credit crunch have been prevented
with adequate credit risk management?
• Lessons learned:
– Don’t trust on historical correlations only
– Use dynamic / stress correlations
Credit Risk
A link to the recent credit crunch
Credit Risk
38
38