C R ED ITR ISK

+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R
D ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED IT
SK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R
D ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED IT
SK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R
D ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED IT
SK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R
D ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED IT
SK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R
D ITR ISK
+
C R ED ITR ISK
+
C R ED ITR ISK
+
C R ED IT
+ + +
Regulated by SFA
CREDIT FIRST
SUISSE B O STO N
C R ED ITR ISK
+
A C R ED IT R ISK M AN AG EM EN T FR AM EW O R K
Copyright 1997 Credit Suisse First Boston International. All rights reserved.
CREDITRISK
+
is a trademark of Credit Suisse First Boston International in countries of use.
CREDITRISK
+
as described in this document (CREDITRISK
+
) is a method of credit risk management introduced by Credit Suisse Group.
No representation or warranty, express or implied, is made by Credit Suisse First Boston International or any other Credit Suisse Group
company as to the accuracy, completeness, or fitness for any particular purpose of CREDITRISK
+
. Under no circumstances shall
Credit Suisse First Boston International or any other Credit Suisse Group company have any liability to any other person or any entity
for (a) any loss, damage or other injury in whole or in part caused by, resulting from or relating to, any error (negligent or otherwise), of
Credit Suisse First Boston International or any other Credit Suisse Group company in connection with the compilation, analysis,
interpretation, communication, publication or delivery of CREDITRISK
+
, or (b) any direct, indirect, special, consequential, incidental or
compensatory damages whatsoever (including, without limitation, lost profits), in either case caused by reliance upon or otherwise resulting
from or relating to the use of (including the inability to use) CREDITRISK
+
.
Issued and approved by Credit Suisse First Boston International for the purpose of Section 57, Financial Services Act 1986. Regulated by
the Securities and Futures Authority. The products and services referred to are not available to private customers.
is a leading global investment banking firm, providing comprehensive
financial advisory, capital raising, sales and trading, and financial
products for users and suppliers of capital around the world. It operates in over 60 offices across
more than 30 countries and six continents and has over 15,000 employees.
CREDIT FIRST
SUISSE B O STO N
CREDI TRI S K
+
1
Contents
1. I ntroduction to CREDI TRI SK
+
3
1.1 D evelopm ents in C redit Risk M anagem ent 3
1.2 C om ponents of C R ED ITR ISK
+
3
1.3 The C R ED ITR ISK
+
M odel 4
1.4 Econom ic C apital 4
1.5 Applications of C R ED ITR ISK
+
5
1.6 Exam ple Spreadsheet Im plem entation 5
2. Modelling Credit Risk 6
2.1 Risk M odelling C oncepts 6
2.2 Types of C redit Risk 7
2.3 D efault Rate B ehaviour 8
2.4 M odelling Approach 9
2.5 Tim e H orizon for C redit Risk M odelling 10
2.6 D ata Inputs to C redit Risk M odelling 11
2.7 C orrelation and Incorporating the Effects of B ackground Factors 14
2.8 M easuring C oncentration 16
3. The CREDI TRI SK
+
Model 17
3.1 Stages in the M odelling Process 17
3.2 Frequency of D efault Events 17
3.3 M oving from D efault Events to D efault Losses 18
3.4 C oncentration Risk and Sector Analysis 20
3.5 M ulti-Year Losses for a H old-to-M aturity Tim e H orizon 21
3.6 Sum m ary of the C R ED ITR ISK
+
M odel 22
4. Economic Capital for Credit Risk 23
4.1 Introduction to Econom ic C apital 23
4.2 Econom ic C apital for C redit Risk 23
4.3 Scenario Analysis 24
5. Applications of CREDI TRI SK
+
26
5.1 Introduction 26
5.2 Provisioning for C redit Risk 26
5.3 Risk-B ased C redit Lim its 29
5.4 Portfolio M anagem ent 29
I
t
Appendices
A. The CREDI TRI SK
+
Model 32
A1 O verview of this Appendix 32
A2 D efault Events w ith Fixed D efault Rates 33
A3 D efault Losses w ith Fixed D efault Rates 35
A4 Loss D istribution w ith Fixed D efault Rates 38
A5 Application to M ulti-Year Losses 39
A6 D efault Rate U ncertainty 41
A7 Sector Analysis 41
A8 D efault Events w ith Variable D efault Rates 44
A9 D efault Losses w ith Variable D efault Rates 46
A10 Loss D istribution w ith Variable D efault Rates 47
A11 C onvergence of Variable D efault Rate C ase to Fixed D efault Rate C ase 49
A12 G eneral Sector Analysis 50
A13 Risk C ontributions and Pairw ise C orrelation 52
B . I llustrative Example 58
B 1 Exam ple Spreadsheet-B ased Im plem entation 58
B 2 Exam ple Portfolio and Static D ata 58
B 3 Exam ple U se of the Spreadsheet Im plem entation 60
C . Contacts 66
D . Selected Bibliography 68
List of Tables
Table 1: Representations of the default rate process 9
Table 2: O ne-year default rates (% ) 12
Table 3: D efault rate standard deviations (% ) 13
Table 4: Recovery rates by seniority and security (% ) 14
Table 5: M echanism s for controlling the risk of credit default losses 25
Table 6: Provisioning for different business lines 28
Table 7: Exam ple of credit risk provisioning 28
Table 8: Exam ple portfolio 59
Table 9: Exam ple m apping table of default rate inform ation 59
Table 10: Exam ple 1A - Risk contributions 64
Table 11: Exam ple 1B - Risk analysis of rem oved obligors 65
Table 12: Exam ple 1B - Portfolio m ovem ent analysis 65
List of Figures
Figure 1: C om ponents of C R ED ITR ISK
+
3
Figure 2: D efault rate as a continuous random variable 8
Figure 3: D efault rate as a discrete random variable 9
Figure 4: Rated corporate defaults by num ber of issuers 12
Figure 5: D efaulted bank loan price distribution 13
Figure 6: C R ED ITR ISK
+
M odel - D istribution of default events 18
Figure 7: C R ED ITR ISK
+
M odel - D istribution of default losses 19
Figure 8: Im pact of sectors on the loss distribution 21
Figure 9: Econom ic capital for credit risk 24
Figure 10: Parts of the credit default loss distribution 25
Figure 11: C redit risk provisioning 27
Figure 12: U sing risk contributions 31
Figure 13: Flow chart description of Appendix A 33
2
CREDIT FIRST
SUISSE B O STO N
CREDI TRI S K
+
3
Introdu
to C R E
1.1 Developments in Credit Risk Management
Since the beginning of the 1990s, C redit Suisse First B oston (C SFB ) has been developing and deploying
new risk m anagem ent m ethods. In 1993, C redit Suisse G roup launched, in parallel, a m ajor project aim ed
at m odernising its credit risk m anagem ent and, using C SFB s expertise, at developing a m ore forw ard-
looking m anagem ent tool. In D ecem ber 1996, C redit Suisse G roup introduced C R ED ITR ISK
+
- a C redit Risk
M anagem ent Fram ew ork.
C urrent areas of developm ent in credit risk m anagem ent include: m odelling credit risk on a portfolio basis;
credit risk provisioning; active portfolio m anagem ent; credit derivatives; and sophisticated approaches to capital
allocation that m ore closely reflect econom ic risk than the existing regulatory capital regim e. C R ED ITR ISK
+
addresses all of these areas and the relationships betw een them .
C R ED ITR ISK
+
can be applied to credit exposures arising from all types of products including corporate and retail
loans, derivatives, and traded bonds.
1.2 Components of CREDITRI SK
+
The com ponents of C R ED ITR ISK
+
and the interrelationships betw een them are show n in the follow ing diagram .
Figure 1:
Components of CREDITRI SK
+
CREDITRISK
+
comprises three
main componentsa CREDITRISK
+
Model that uses a portfolio
approach, a methodology for
calculating economic capital
for credit risk, and several
applications of the technology.
I ntroduction to CREDI TRI SK
+
1
CREDITRI SK
+
Credit Risk Measurement
Credit Default
Loss Distribution
Scenario Analysis
Provisioning
Limits
Portfolio Management
Economic Capital Applications
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
A m odern approach to credit risk m anagem ent should address all aspects of credit risk, from quantitative
m odelling to the developm ent of practical techniques for its m anagem ent. In addition to w ell-established credit
risk m anagem ent techniques, such as individual obligor (borrow er, counterparty or issuer) lim its and concentration
lim its, C R ED ITR ISK
+
reflects the requirem ents of a m odern approach to m anaging credit risk and com prises three
m ain com ponents:

The CREDITRISK
+
Model that uses a portfolio approach and analytical techniques applied w idely in the
insurance industry.

A methodology for calculating economic capital for credit risk.

Applications of the credit risk m odelling m ethodology including: (i) a methodology for establishing provisions
on an anticipatory basis, and (ii) a means of measuring diversification and concentration to assist in
portfolio management.
1.3 The CREDITRI SK
+
Model
C R ED ITR ISK
+
is based on a portfolio approach to m odelling credit default risk that takes into account
inform ation relating to size and m aturity of an exposure and the credit quality and system atic risk of an obligor.
The C R ED ITR ISK
+
M odel is a statistical m odel of credit default risk that m akes no assum ptions about the
causes of default. This approach is sim ilar to that taken in m arket risk m anagem ent, w here no attem pt is m ade
to m odel the causes of m arket price m ovem ents. The C R ED ITR ISK
+
M odel considers default rates as continuous
random variables and incorporates the volatility of default rates in order to capture the uncertainty in the level
of default rates. O ften, background factors, such as the state of the econom y, m ay cause the incidence of
defaults to be correlated, even though there is no causal link betw een them . The effects of these background
factors are incorporated into the C R ED ITR ISK
+
M odel through the use of default rate volatilities and sector
analysis rather than using default correlations as explicit inputs into the m odel.
M athem atical techniques applied w idely in the insurance industry are used to m odel the sudden event of an
obligor default. This approach contrasts w ith the m athem atical techniques typically used in finance. In financial
m odelling one is usually concerned w ith m odelling continuous price changes rather than sudden events.
A pplying insurance m odelling techniques, the analytic C R E D ITR ISK
+
M odel captures the essential
characteristics of credit default events and allow s explicit calculation of a full loss distribution for a portfolio of
credit exposures.
1.4 Economic Capital
The output of the C R ED ITR ISK
+
M odel can be used to determ ine the level of econom ic capital required to cover
the risk of unexpected credit default losses.
M easuring the uncertainty or variability of loss and the relative likelihood of the possible levels of unexpected
losses in a portfolio of credit exposures is fundam ental to the effective m anagem ent of credit risk. Econom ic
capital provides a m easure of the risk being taken by a firm and has several benefits: it is a m ore appropriate
risk m easure than that specified under the current regulatory regim e; it m easures econom ic risk on a portfolio
basis and takes account of diversification and concentration; and, since econom ic capital reflects the changing
risk of a portfolio, it can be used for portfolio m anagem ent.
4
CREDIT FIRST
SUISSE B O STO N
CREDI TRI S K
+
5
The C R ED ITR ISK
+
M odel is supplem ented by scenario analysis in order to identify the financial im pact of low
probability but nevertheless plausible events that m ay not be captured by a statistically based m odel.
1.5 Applications of CREDITRI SK
+
C R ED ITR ISK
+
includes several applications of the credit risk m odelling m ethodology, including a forw ard-looking
provisioning m ethodology and quantitative portfolio m anagem ent techniques.
1.6 Example Spreadsheet Implementation
In order to assist the reader of this docum ent, a spreadsheet-based im plem entation that illustrates the range
of possible outputs of the C R ED ITR ISK
+
M odel can be dow nloaded from the Internet (http://w w w.csfb.com ).
1
I ntroduction
M odel
Credit
2.1 Risk Modelling Concepts
2.1.1 Types of Uncertainty Arising in the Modelling Process
A statistically based m odel can describe m any business processes. H ow ever, any m odel is only a
representation of the real w orld. In the m odelling process, there are three types of uncertainty that m ust be
assessed: process risk, param eter uncertainty and m odel error.
Process Risk
Process risk arises because the actual observed results are subject to random fluctuations even w here the
m odel describing the loss process and the param eters used by the m odel are appropriate. Process risk is
usually addressed by expressing the m odel results to an appropriately high level of confidence.
Parameter Uncertainty
Param eter uncertainty arises from the difficulties in obtaining estim ates of the param eters used in the m odel.
The only inform ation that can be obtained about the underlying process is obtained by observing the results
that it has generated in the past. It is possible to assess the im pact of param eter uncertainty by perform ing
sensitivity analysis on the param eter inputs.
Model Error
M odel error arises because the proposed m odel does not correctly reflect the actual process - alternative
m odels could produce different results. M odel error is usually the least tractable of the three types of
uncertainty.
2.1.2 Addressing Modelling Issues
As all of these types of uncertainty enter into the m odelling process, it is im portant to be aw are of them and
to consider how they can be addressed w hen developing a credit risk m odel. Indeed, a realistic assessm ent of
the potential effects of these errors should be m ade before any decisions are m ade based on the outputs of
the m odel.
6
CREDIT FIRST
SUISSE B O STO N
Modelling Credit Risk 2
CREDI TRI S K
+
7
Modelling Credit Risk
The CREDITRISK
+
Model
makes no assumptions
about the causes of default.
This approach is similar to
that taken in market risk
management, where no
assumptions are made
about the causes of market
price movements.
All portfolios of exposures
exhibit credit default risk, as
the default of an obligor
results in a loss.
C R ED ITR ISK
+
addresses these types of uncertainty in several w ays:

N o assum ptions are m ade about the causes of default. This approach is sim ilar to that taken in m arket risk
m anagem ent, w here no assum ptions are m ade about the causes of m arket price m ovem ents. This not only
reduces the potential m odel error but also leads to the developm ent of an analytically tractable m odel.

The data requirem ents for the C R ED ITR ISK

+
M odel have been kept as low as possible, w hich m inim ises the
error from param eter uncertainty. In the credit environm ent, em pirical data is sparse and difficult to obtain.
Even then, the data can be subject to large fluctuations year on year.

C oncerns about param eter uncertainty are addressed using scenario analysis, in w hich the effects of stress
testing each of the input param eters are quantified. For exam ple, increasing default rates or default rate
volatilities can be used to sim ulate dow nturns in the econom y.
2.2 Types of Credit Risk
There are tw o m ain types of credit risk:

Credit spread risk: C redit spread risk is exhibited by portfolios for w hich the credit spread is traded and
m arked-to-m arket. C hanges in observed credit spreads im pact the value of these portfolios.

Credit default risk: All portfolios of exposures exhibit credit default risk, as the default of an obligor results
in a loss.
2.2.1 Credit Spread Risk
C redit spread is the excess return dem anded by the m arket for assum ing a certain credit exposure. C redit
spread risk is the risk of financial loss ow ing to changes in the level of credit spreads used in the m ark-to-
m arket of a product.
C redit spread risk fits m ore naturally w ithin a m arket risk m anagem ent fram ew ork. In order to m anage credit
spread risk, a firm s value-at-risk m odel should take account of value changes caused by the volatility of credit
spreads. Since the distribution of credit spreads m ay not be norm al, a standard variance-covariance approach
to m easuring credit spread risk m ay be inappropriate. H ow ever, the historical sim ulation approach, w hich does
not m ake any assum ptions about the underlying distribution, used in com bination w ith other techniques,
provides a suitable alternative.
C redit spread risk is only exhibited w hen a m ark-to-m arket accounting policy is applied, such as for portfolios
of bonds and credit derivatives. In practice, som e types of products, such as corporate or retail loans, are
typically accounted for on an accruals basis. A m ark-to-m arket accounting policy w ould have to be applied to
these products in order to recognise the credit spread risk.
2.2.2 Credit Default Risk
C redit default risk is the risk that an obligor is unable to m eet its financial obligations. In the event of a default
of an obligor, a firm generally incurs a loss equal to the am ount ow ed by the obligor less a recovery am ount
w hich the firm recovers as a result of foreclosure, liquidation or restructuring of the defaulted obligor.
All portfolios of exposures exhibit credit default risk, as the default of an obligor results in a loss.
2
C redit default risk is typically associated w ith exposures that are m ore likely to be held to m aturity, such as
corporate and retail loans and exposures arising from derivative portfolios. B ond m arkets are generally m ore
liquid than loan m arkets and therefore bond positions can be adjusted over a shorter tim e fram e. H ow ever,
w here the intention is to m aintain a bond portfolio over a longer tim e fram e, even though the individual
constituents of the portfolio m ay change, it is equally im portant to m easure the default risk that is taken by
holding the portfolio.
C R ED ITR ISK
+
focuses on m odelling and m anaging credit default risk.
2.3 Default Rate Behaviour
Equity and bond prices are forw ard-looking in nature and are form ed by investorsview s of the financial
prospects of a particular obligor. H ence, they incorporate both the credit quality and the potential credit quality
changes of that obligor.
Therefore, the default rate of a particular obligor, inferred from m arket prices, w ill vary on a continuous scale
and hence can be view ed as a continuous random variable. In m odelling credit risk, one is concerned w ith
determ ining the possible future outcom es over the chosen tim e horizon.
The process for the default rate can be represented in tw o different w ays:

Continuous variable: W hen treated as a continuous variable, the possible default rate over a given tim e
horizon is described by a distribution, w hich can be specified by a default rate and a volatility of the default
rate. The data requirem ents for m odelling credit default risk are analogous to the data requirem ents for
pricing stock options - a forw ard stock price and the stock price volatility are used to define the forw ard
stock price distribution. The follow ing figure illustrates the path that a default rate m ay take over tim e and
the distribution that it could have over that tim e.

Discrete variable: B y treating the default rate as a discrete variable, a sim plification of the continuous
process described above is m ade. A convenient w ay of m aking default rates discrete is by assigning credit
ratings to obligors and m apping default rates to credit ratings. U sing this approach, additional inform ation
is required in order to m odel the possible future outcom es of the default rate. This can be achieved via a
rating transition m atrix that specifies the probability of keeping the sam e credit rating, and hence the sam e
value for the default rate, and the probabilities of m oving to different credit ratings and hence to different
values for the default rate. This is illustrated in the follow ing figure.
8
CREDIT FIRST
SUISSE B O STO N
Possible path of
default rate
Frequency of default
rate outcomes
Figure 2:
Default rate as a
continuous random
variable
D
e
f
a
u
l
t

r
a
t
e
Tim e horizon
The discrete approach w ith rating m igrations and the continuous approach w ith a default rate volatility are
different representations of the behaviour of default rates. B oth approaches achieve the desired end result of
producing a distribution for the default rate.
The above tw o representations of default rate behaviour are sum m arised in the follow ing table:
Treatment of default rate Data requirements
C ontinuous variable D efault rates
Volatility of default rates
D iscrete variable C redit ratings
Rating transition m atrix
The C R ED ITR ISK
+
M odel is a statistical m odel of credit default risk that m odels default rates as continuous
random variables and incorporates the volatility of the default rate in order to capture the uncertainty in the
level of the default rate. A m apping from credit ratings to a set of default rates provides a convenient approach
to setting the level of the default rate.
2.4 Modelling Approach
2.4.1 Risk Measures
W hen m anaging credit risk, there are several m easures of risk that are of interest, including the follow ing:

Distribution of loss: The risk m anager is interested in obtaining distributions of loss that m ay arise from the
current portfolio. The risk m anager needs to answ er questions such as W hat is the size of loss for a given
confidence level?.

Identifying extreme outcomes: The risk m anager is also concerned w ith identifying extrem e or catastrophic
outcom es. These outcom es are usually difficult to m odel statistically but can be addressed through the use
of scenario analysis and concentration lim its.
Table 1:
Representations of the
default rate process
CREDI TRI S K
+
9
Modelling Credit Risk
Figure 3:
Default rate as a discrete
random variable
2
Possible path of
default rate
B
AAA
AA
BBB
BB
A
Default
Frequency of default
rate outcomes
D
e
f
a
u
l
t

r
a
t
e
Tim e horizon
The C R ED ITR ISK
+
M odel
treats default rates as
continuous random variables
and incorporates default
rate volatility to capture the
uncertainty in the level of
the default rate.
2.4.2 A Portfolio Approach to Managing Credit Risk
C redit risk can be m anaged through diversification because the num ber of individual risks in a portfolio of
exposures is usually large. C urrently, the prim ary technique for controlling credit risk is the use of lim it system s,
including individual obligor lim its to control the size of exposure, tenor lim its to control the m axim um m aturity
of exposures to obligors, rating exposure lim its to control the am ount of exposure to obligors of certain credit
ratings, and concentration lim its to control concentrations w ithin countries and industry sectors.
The portfolio risk of a particular exposure is determ ined by four factors: (i) the size of the exposure, (ii) the
m aturity of the exposure, (iii) the probability of default of the obligor, and (iv) the system atic or concentration
risk of the obligor. C redit lim its aim to control risk arising from each of these factors individually. The general
effect of this approach, w hen applied in a w ell-structured and consistent m anner, is to create reasonably w ell-
diversified portfolios. H ow ever, these lim its do not provide a m easure of the diversification and concentration
of a portfolio.
In order to m anage effectively a portfolio of exposures, a m eans of m easuring diversification and concentration
has to be developed. An approach that incorporates size, m aturity, credit quality and system atic risk into a single
portfolio m easure is required. C R ED ITR ISK
+
takes such an approach.
2.4.3 Modelling Techniques Used in the CREDITRI SK
+
Model
The econom ic risk of a portfolio of credit exposures is analogous to the econom ic risk of a portfolio of
insurance exposures. In both cases, losses can be suffered from a portfolio containing a large num ber of
individual risks, each w ith a low probability of occurring. The risk m anager is concerned w ith assessing the
frequency of the unexpected events as w ell as the severity of the losses.
In order to keep m odel error to a m inim um , no assum ptions are m ade about the causes of default.
M athem atical techniques applied w idely in the insurance industry are used to m odel the sudden event of an
obligor default. In m odelling credit default losses one is concerned w ith sudden events rather than continuous
changes. The essential characteristics of credit default events are captured by applying these insurance
m odelling techniques. This has the additional benefit that it leads to a credit risk m odel that is analytically
tractable and hence not subject to the problem s of precision that can arise w hen using a sim ulation-based
approach. The analytic C R ED ITR ISK
+
M odel allow s rapid and explicit calculation of a full loss distribution for a
portfolio of credit exposures.
2.5 Time Horizon for Credit Risk Modelling
A key decision that has to be m ade w hen m odelling credit risk is the choice of tim e horizon. G enerally, the tim e
horizon chosen should not be shorter than the tim e fram e over w hich risk-m itigating actions can be taken.
C R ED ITR ISK
+
does not prescribe any one particular tim e horizon but suggests tw o possible tim e horizons that
can provide m anagem ent inform ation relevant for credit risk m anagem ent:

A constant tim e horizon, such as one year.

A hold-to-m aturity or run-off tim e horizon.

10
CREDIT FIRST
SUISSE B O STO N
CREDITRISK
+
is based
on a portfolio approach -
summarising information
about size, maturity, credit
quality and systematic risk
into a single measure.
CREDI TRI S K
+
11
2.5.1 Constant Time Horizon
A constant tim e horizon is relevant, as it allow s all exposures to be considered at the sam e future date.
For various reasons, one year is often taken as a suitable tim e horizon: credit risk m itigating actions can
norm ally be executed w ithin one year, new capital can be raised to replenish capital eroded by actual credit
losses during the period, and, furtherm ore, one year m atches the norm al accounting period. G iven these
factors, C R ED ITR ISK
+
suggests a tim e horizon of one year for credit risk econom ic capital.
2.5.2 Hold-to-Maturity Time Horizon
Alternatively, a hold-to-m aturity tim e horizon allow s the full term structure of default rates over the lifetim e of the
exposures to be recognised. This view of the portfolio enables the risk m anager to com pare exposures of
different m aturity and credit quality and is an appropriate tool, in addition to the constant tim e horizon, for
portfolio m anagem ent. The role that the C R ED ITR ISK
+
M odel plays in active portfolio m anagem ent is discussed
later in this docum ent.
A benchm ark tim e horizon of one year can be used for portfolios w here there is an intention to m aintain
exposures for longer than the term of the booked transactions (e.g. traded bond portfolios).
2.6 Data Inputs to Credit Risk Modelling
2.6.1 Data Inputs
Any m odelling of credit risk is dependent on certain data requirem ents being m et. The quality of this data w ill
directly affect the accuracy of the m easurem ent of credit risk and therefore any decision to be m ade using the
results should be m ade only having fully assessed the potential error from uncertainties in the data used.
The inputs used by the C R ED ITR ISK
+
M odel are:

C redit exposures

O bligor default rates

O bligor default rate volatilities and

Recovery rates.
The C R ED ITR ISK
+
M odel presented in this docum ent does not prescribe the use of any one particular data set
over another. O ne of the key lim itations in m odelling credit risk is the lack of com prehensive default data.
W here a firm has its ow n inform ation that is judged to be relevant to its portfolio, this can be used as the input
into the m odel. Alternatively, conservative assum ptions can be used w hile default data quality is being im proved.
2.6.2 Credit Exposures
The exposures arising from separate transactions w ith an obligor should be aggregated according to the legal
corporate structure and taking into account any rights of set-off.
The C R ED ITR ISK
+
M odel is capable of handling all types of instrum ents that give rise to credit exposure,
including bonds, loans, com m itm ents, financial letters of credit and derivative exposures. For som e of these
transaction types, it is necessary to m ake an assum ption about the level of exposure in the event of a default:
for exam ple, a financial letter of credit w ill usually be draw n dow n prior to default and therefore the exposure
at risk should be assum ed to be the full nom inal am ount.
In addition, if a m ulti-year tim e horizon is being used, it is im portant that the changing exposures over tim e are
accurately captured.
2
Modelling Credit Risk
Credit Risk Measurement
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
Figure 4:
Rated corporate defaults
by number of issuers
O ne-year default rates
show significant fluctuations
from year to year.
12
CREDIT FIRST
SUISSE B O STO N
Credit Risk Measurement
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
Table 2:
One-year default rates (%)
2.6.3 Default Rates
A default rate, w hich represents the likelihood of a default event occurring, should be assigned to each obligor.
This can be achieved in a num ber of w ays, including:

O bserved credit spreads from traded instrum ents can be used to provide m arket-assessed probabilities of
default.

Alternatively, obligor credit ratings, together w ith a m apping of default rates to credit ratings, provide a
convenient w ay of assigning probabilities of default to obligors. The rating agencies publish historic default
statistics by rating category for the population of obligors that they have rated.
C redit rating O ne-year default rate
Aaa 0.00
Aa 0.03
A 0.01
B aa 0.12
B a 1.36
B 7.27
Source: C arty & Lieberm an, 1997, M oodys Investors Service G lobal C redit Research
A credit rating is an opinion of an obligors overall financial capacity to m eet its financial obligations (i.e. its
creditw orthiness). This opinion focuses on the obligors capacity and w illingness to m eet its financial
com m itm ents as they fall due. An assessm ent of the nature of a particular obligation, including its seniority in
bankruptcy or liquidation, should be perform ed w hen considering the recovery rate for an obligor.
It should be noted that one-year default rates show significant variation year on year, as can be seen in the
follow ing figure. D uring periods of econom ic recession, the num ber of defaults can be m any tim es the level
observed at other tim es.
Source: Standard & Poors Ratings Perform ance 1996 (February 1997)

Another approach is to calculate default probabilities on a continuous scale, w hich can be used as a
substitute for the com bination of credit ratings and assigned default rates.
F
r
e
q
u
e
n
c
y
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
7 0
6 0
5 0
4 0
3 0
2 0
1 0
0
CREDI TRI S K
+
13
2
Modelling Credit Risk
Credit Risk Measurement
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
Credit Risk Measurement
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
Table 3:
Default rate standard
deviations (%)
Figure 5:
Defaulted bank loan
price distribution
F
r
e
q
u
e
n
c
y
2.6.4 Default Rate Volatilities
Published default statistics include average default rates over m any years. As show n previously, actual
observed default rates vary from these averages. The am ount of variation in default rates about these averages
can be described by the volatility (standard deviation) of default rates. As can be seen in the follow ing table,
the standard deviation of default rates can be significant com pared to actual default rates, reflecting the high
fluctuations observed during econom ic cycles.
One-year default rate (%)
Credit rating Average Standard deviation
Aaa 0.00 0.0
Aa 0.03 0.1
A 0.01 0.0
B aa 0.12 0.3
B a 1.36 1.3
B 7.27 5.1
Source: C arty & Lieberm an, 1996, M oodys Investors Service G lobal C redit Research
The default rate standard deviations in the above table w ere calculated over the period from 1970 to 1996
and therefore include the effect of econom ic cycles.
2.6.5 Recovery Rates
In the event of a default of an obligor, a firm generally incurs a loss equal to the am ount ow ed by the obligor
less a recovery am ount, w hich the firm recovers as a result of foreclosure, liquidation or restructuring of the
defaulted obligor or the sale of the claim . Recovery rates should take account of the seniority of the obligation
and any collateral or security held.
Recovery rates are subject to significant variation. For exam ple, the figure below show s the price distribution
of defaulted bank loans and illustrates that there is a large degree of dispersion.
Source: D efaulted B ank Loan Recoveries (N ovem ber 1996) , M oodys Investors Service G lobal C redit Research
$
0
-
$
1
0
$
1
1
-
$
2
0
$
2
1
-
$
3
0
$
3
1
-
$
4
0
$
4
1
-
$
5
0
$
5
1
-
$
6
0
$
6
1
-
$
7
0
$
7
1
-
$
8
0
$
8
1
-
$
9
0
$
9
1
-
$
1
0
0
1 4
1 2
1 0
8
6
4
2
0
There is also considerable variation for obligations of differing seniority, as can be seen from the standard
deviation of the corporate bond and bank loan recovery rates in the table below.
Seniority and security Average Standard deviation
Senior secured bank loans 71.18 21.09
Senior secured public debt 63.45 26.21
Senior unsecured public debt 47.54 26.29
Senior subordinated public debt 38.28 24.74
Subordinated public debt 28.29 20.09
Junior subordinated public debt 14.66 8.67
Source: H istorical D efault Rates of C orporate B ond Issuers, 1920-1996 (January 1997) M oodys Investors Service G lobal C redit Research
Publicly available recovery rate data indicates that there can be significant variation in the level of loss, given
the default of an obligor. Therefore, a careful assessm ent of recovery rate assum ptions is required. G iven this
uncertainty, stress testing should be perform ed on the recovery rates in order to calculate the potential loss
distributions under different scenarios.
2.7 Correlation and Incorporating the Effects of Background Factors
D efault correlation im pacts the variability of default losses from a portfolio of credit exposures. The C R ED ITR ISK
+
M odel incorporates the effects of default correlations by using default rate volatilities and sector analysis.
2.7.1 The Random Nature of Defaults and the Appearance of Correlation
C redit defaults occur as a sequence of events in such a w ay that it is not possible to forecast the exact tim e
of occurrence of any one default or the exact total num ber of defaults. Often, there are background factors
that may cause the incidence of default events to be correlated, even though there is no causal link between
them. For exam ple, if there is an unusually large num ber of defaults in one particular m onth, this m ight be due
to the econom y being in recession, w hich has increased the rates of default above their average level. In this
econom ic situation, it is quite likely that the num ber of defaults in the follow ing m onth w ill also be high.
C onversely, if there are few er defaults than on average in one m onth, because the econom y is grow ing, it is
also likely that there w ill be few er defaults than on average in the follow ing m onth. The defaults are correlated
but there is no causal link betw een them - the correlation effect observed is due to a background factor, the
state of the econom y, w hich changes the rates of default.
2.7.2 Impact of the Economy on Default Rates
There is general agreem ent that the state of the econom y in a country has a direct im pact on observed default
rates. A recent report by Standard and Poors stated that A healthy economy in 1996 contributed to a
significant decline in the total number of corporate defaults. Compared to 1995, defaults were reduced by
one-half.
1
Another report by M oodys Investors Service stated that The sources of [default rate volatility]
are many, but macroeconomic trends are certainly the most influential factors.
2
As the above quotations indicate and as can be seen in Figure 4 above, there is significant variation in the
num ber of defaults from year to year. Furtherm ore, for each year, different industry sectors w ill be affected to
different degrees by the state of the econom y. The m agnitude of the im pact w ill be dependent on how sensitive
an obligors earnings are to various econom ic factors, such as the grow th rate of the econom y and the level of
interest rates.
14
CREDIT FIRST
SUISSE B O STO N
Table 4:
Recovery rates by
seniority and security (%)
Often, there are
background factors that
may cause the incidence
of defaults to be correlated,
even though there is no
causal link between them.
1
Standard and Poors
Ratings Performance
1996, February 1997
2
Moodys Investors
Service, Corporate Bond
Defaults and Default
Rates, January 1996
CREDI TRI S K
+
15
Econom ic m odels that attem pt to capture the effect of changes in the econom y on default rates can be
developed in order to specify the default rates for subsequent use in a credit risk m odel. H ow ever, this
approach can have several w eaknesses, including the follow ing:

Since there are lim ited publicly available default rate statistics by country or by industry sector, it is difficult
to verify the accuracy of an econom ic m odel used to derive default rates.

Even if a causal relationship could be established relating default rates to certain econom ic variables, it is
questionable w hether such relationships w ould be stable over several years.
Therefore, alternative approaches that attem pt to capture the observed variability of default rates have to be
sought.
2.7.3 Incorporating the Effects of Background Factors
It is possible to incorporate the effects of background factors into the specification of default rates by allow ing
the default rate itself to have a probability distribution. This is accom plished by incorporating default rate
volatilities into the m odel.
The C R ED ITR ISK
+
M odel m odels the effects of background factors by using default rate volatilities that result
in increased defaults rather than by using default correlations as a direct input. B oth approaches, the use of
default rate volatilities and default correlations, give rise to loss distributions w ith fat tails.
Section 3 of this docum ent describes in detail how the C R ED ITR ISK
+
M odel uses default rate volatilities in the
m odelling of credit default risk.
The C R ED ITR ISK
+
M odel does not attem pt to m odel correlations explicitly but captures the sam e concentration
effects through the use of default rate volatilities and sector analysis
3
. There are various reasons w hy this
approach has been taken, including the follow ing:

Instability of default correlations: G enerally, correlations calculated from financial data show a high degree
of instability. In addition, a calculated correlation can be very dependent on the underlying tim e period of
the data. A sim ilar instability problem m ay arise w ith default rate volatilities: how ever, it is m uch easier to
perform scenario analysis on default rate volatilities, ow ing to the analytically tractable nature of a m odel
that uses volatilities rather than correlations.

Lack of empirical data: There is little em pirical data on default correlations. D efaults them selves are
infrequent events and so there is insufficient data on m ultiple defaults w ith w hich to calculate explicit
default correlations. Since default correlations are difficult to calculate directly, som e approaches use asset
price correlations to derive default correlations, but this can only be considered a proxy. This technique
relies upon additional assum ptions about the relationship betw een asset prices and probabilities of default.
Furtherm ore, it is questionable how stable any relationship, that m ay be inferred or observed during a period
of norm al trading, w ould be in the event of default of a particular obligor. In addition, w here there is no asset
price for the obligor, for exam ple in a retail portfolio, there is no obvious w ay of deriving default correlations.
2
Modelling Credit Risk
The CREDITRISK
+
Model
captures concentration risk
through the use of default
rate volatilities and sector
analysis.
3
Sector analysis is
discussed in Sections
2.8 and 3.4
2.8 Measuring Concentration
The above discussion has highlighted the fact that there are background factors that affect the level of default
rates. The state of the econom y of each different country w ill vary over tim e and, w ithin each country, different
industry sectors w ill be affected to differing degrees. A portfolio of exposures can have concentrations in
particular countries or industry sectors. Therefore, it is im portant to be able to capture the effect of
concentration risk in a credit risk m odel.
The C R ED ITR ISK
+
M odel described in this docum ent allow s concentration risk to be captured using sector
analysis. An exposure can be broken dow n into an obligor-specific elem ent, w hich is independent of all other
exposures, and non-specific or system atic elem ents that are sensitive to particular driving factors, such as
countries or industry sectors.
16
CREDIT FIRST
SUISSE B O STO N
C
M
CREDI TRI S K
+
17
CR ED IT
M odel
The CREDI TRI SK
+
Model
3
Credit Risk Measurement
Exposures D efault Rates
CREDITRISK
+
Model
Recovery
Rates
D efault Rate
Volatilities
3.1 Stages in the Modelling Process
The m odelling of credit risk is a tw o stage process, as show n in the follow ing diagram :
B y calculating the distribution of default events, the risk m anager is able to assess w hether the overall credit
quality of the portfolio is either im proving or deteriorating. The distribution of losses allow s the risk m anager to
assess the financial im pact of the potential losses as w ell as m easuring the am ount of diversification and
concentration w ithin the portfolio.
3.2 Frequency of Default Events
3.2.1 The Default Process
The C R ED ITR ISK
+
M odel m akes no assum ption about the causes of default - credit defaults occur as a
sequence of events in such a w ay that it is neither possible to forecast the exact tim e of occurrence of any
one default nor the exact total num ber of defaults. There is exposure to default losses from a large num ber of
obligors and the probability of default by any particular obligor is sm all. This situation is w ell represented by the
Poisson distribution.
What is the
FREQUENCY
of defaults?
What is the
SEVERITY
of the losses ?
Stage 1
Stage 2
Distribution of
default losses
W e consider first the distribution of the num ber of default events in a tim e period, such as one year, w ithin a
portfolio of obligors having a range of different annual probabilities of default. The annual probability of default
of each obligor can be conveniently determ ined by its credit rating and a m apping betw een default rates and
credit ratings. If w e do not incorporate the volatility of the default rate, the distribution of the num ber of default
events w ill be closely approxim ated by the Poisson distribution. This is regardless of the individual default rate
for a particular obligor.
H ow ever, default rates are not constant over tim e and, as w e have seen in the previous section, exhibit a high
degree of variation. H ence, default rate variability needs to be incorporated into the m odel.
3.2.2 Distribution of the Number of Default Events
The C R ED ITR ISK
+
M odel m odels the underlying default rates by specifying a default rate and a default rate
volatility. This aim s to take account of the variation in default rates in a pragm atic m anner, w ithout introducing
significant m odel error.
The effect of using default rate volatilities can be clearly seen in the follow ing figure, w hich show s the
distribution of the num ber of default events generated by the C R ED ITR ISK
+
M odel w hen default rate volatility
is varied. Although the expected num ber of default events is the sam e, the distribution becom es significantly
skew ed to the right w hen default rate volatility is increased. This represents a significantly increased risk of an
extrem e num ber of default events.
3.3 Moving from Default Events to Default Losses
3.3.1 Distribution of Default Losses
G iven the num ber of default events, w e now consider the distribution of losses in the portfolio. The distribution
of losses differs from the distribution of default events because the am ount lost in a given default depends on
the exposure to the individual obligors. U nlike the variation of default probability betw een obligors, w hich does
not influence the distribution of the total num ber of defaults, the variation in exposure m agnitude results in a
loss distribution that is not Poisson in general. M oreover, inform ation about the distribution of different
exposures is essential to the overall distribution. H ow ever, it is possible to describe the overall distribution of
losses because its probability generating function has a sim ple closed form am enable to com putation.
18
CREDIT FIRST
SUISSE B O STO N
Including default rate volatility
Excluding default rate volatility
Figure 6:
CREDITRI SK
+
Model -
Distribution of default
events
P
r
o
b
a
b
i
l
i
t
y
N um ber of defaults
CREDI TRI S K
+
19
In the event of a default of an obligor, a firm generally incurs a loss equal to the am ount ow ed by the obligor
less a recovery am ount, w hich the firm obtains as a result of foreclosure, liquidation or restructuring of the
defaulted obligor. A recovery rate is used to quantify the am ount received from this process. Recovery rates
should take account of the seniority of the obligation and any collateral or security held.
In order to reduce the am ount of data to be processed, tw o steps are first follow ed:

The exposures are adjusted by anticipated recovery rates in order to calculate the loss in a given default.

The exposures, net of the above recovery adjustm ent, are divided into bands of exposure w ith the level of
exposure in each band being approxim ated by a com m on average.
The C R ED ITR ISK
+
M odel calculates the probability that a loss of a certain m ultiple of the chosen unit of
exposure w ill occur. This allow s a full loss distribution to be generated, as show n in the figure below.
3.3.2 Impact of Incorporating Default Rate Volatilities
Figure 7 com pares the default loss distributions calculated w ithout default rate volatility and w ith default rate
volatility. The key features and differences are:

Same expected loss: B oth default loss distributions have the sam e level of expected losses.

Fatter tail: The key change is the level of losses at the higher percentiles; for exam ple, the 99th percentile
is significantly higher w hen the im pact of the variability of default rates is m odelled. There is now
considerably m ore chance of experiencing extrem e losses.
Since the tail of the distribution has becom e fatter, w hile the expected loss has rem ained unchanged, it can be
concluded that the variance of the default loss distribution has increased. This increase in the variance is due
to the pairwise default correlations between obligors. These pairwise default correlations are incorporated
into the CREDITRISK
+
Model through the default rate volatilities and sector analysis. It should be noted that
w hen the default rate volatilities are set to zero, the default events are independent and hence the pairw ise
default correlations are also zero.
In Appendix A, w e give an explicit form ula for the pairw ise default correlations im plied by the C R ED ITR ISK
+
M odel w hen default rate volatilities are incorporated into the m odel.
3
CREDI TRI SK
+
Model
Excluding default rate volatility
Including default rate volatility
Figure 7:
CREDITRI SK
+
Model -
Distribution of default
losses
The C R ED ITR ISK
+
M odel
allow s explicit calculation of
the loss distribution of a
portfolio of credit exposures.
S ize of loss
P
r
o
b
a
b
i
l
i
t
y
3.4 Concentration Risk and Sector Analysis
The C R ED ITR ISK
+
M odel m easures the benefit of portfolio diversification and the im pact of concentrations
through the use of sector analysis.
3.4.1 Concentration Risk
D iversification arises naturally because the num ber of individual risks in a portfolio of exposures is usually large.
H ow ever, even in a portfolio containing a large num ber of exposures, there m ay be an opposing effect ow ing
to concentration risk. C oncentration risk results from having in the portfolio a num ber of obligors w hose
fortunes are affected by a com m on factor. In order to quantify concentration risk, the concepts of system atic
factors and specific factors are necessary.
Systematic factors
System atic factors are background factors that affect the fortunes of a proportion of the obligors in the
portfolio, for exam ple all those obligors having their dom icile in a particular country. The fortunes of any one
obligor can be affected by a num ber of system atic factors.
Specific factors
In general, the fortunes of an obligor are affected to som e extent by specific factors unique to the obligor.
System atic factors im pact the risk of extrem e losses from a portfolio of credit exposures, w hile diversification
largely elim inates the im pact of the specific factors.
C oncentration risk is dependent on the system atic factors affecting the portfolio. The technique for m easuring
concentration risk is sector analysis.
3.4.2 Sector Analysis - Allocating all Obligors to a Single Sector
The m ost straightforw ard application of the C R ED ITR ISK
+
M odel is to allocate all obligors to a single sector.
This approach assum es that a single system atic factor affects the individual default rate volatility of each
obligor. Furtherm ore, this use of the m odel captures all of the concentration risk w ithin the portfolio and
excludes the diversification benefit of the fortunes of individual obligors being subject to a num ber of
independent system atic factors.
Therefore, the m ost straightforw ard application of the C R ED ITR ISK
+
M odel, in w hich all obligors are allocated
to a single sector, generates a prudent estim ate of extrem e losses.
3.4.3 Sector Analysis - Allocating Obligors to one of Several Sectors
In order to recognise som e of the diversification benefit of obligors w hose fortunes are affected by a num ber
of independent system atic factors, it can be assum ed that each obligor is subject to only one system atic factor,
w hich is responsible for all of the uncertainty of the obligors default rate. For exam ple, obligors could be
allocated to sectors according to their country of dom icile. O nce allocated to a sector, the obligors default rate
and default rate volatility are set individually. In this case, a sector can be thought of as a collection of obligors
having the com m on property that they are influenced by the sam e single system atic factor.
3.4.4 Sector Analysis - Apportioning Obligors across Several Sectors
A m ore generalised approach is to assum e that the fortunes of an obligor are affected by a num ber of
system atic factors. The C R ED ITR ISK
+
M odel handles this situation by apportioning an obligor across several
sectors rather than allocating an obligor to a single sector.
20
CREDIT FIRST
SUISSE B O STO N
Concentration risk is
dependent on the systematic
factors affecting the portfolio.
The technique for measuring
concentration risk is sector
analysis.
So far it has been assum ed that all risk in the portfolio is system atic and allocable to one of the system atic
factors. In addition to the effects of system atic factors, it is likely that the fortunes of an obligor are affected
by factors specific to the obligor. Potentially specific risk requires an additional sector to m odel each obligor,
since the factor driving specific risk for a given obligor affects that obligor only. H ow ever, the C R ED ITR ISK
+
M odel handles specific risk w ithout recourse to a large num ber of sectors by apportioning all obligorsspecific
risk to a single Specific Risk Sector.
3.4.5 The Impact of Sectors on the Loss Distribution
As stated above, the C R ED ITR ISK
+
M odel allow s the portfolio of exposures to be allocated to sectors to reflect
the degree of diversification and concentration present. The m ost diversified portfolio is obtained w hen
each exposure is in its ow n sector and the m ost concentrated is obtained w hen the portfolio consists of a
single sector.
The figure below show s the im pact of sectors on the loss distribution. As the num ber of sectors is increased,
the im pact of concentration risk is reduced. The graph illustrates this by plotting the ratio of the 99th percentile
of the credit default loss distribution for a given num ber of sectors to the 99th percentile of the credit default
loss distribution w hen the portfolio is considered to be a single sector.
3.5 Multi-Year Losses for a Hold-to-Maturity Time Horizon
As discussed in Section 2.5, the C R ED ITR ISK
+
M odel allow s risk of the portfolio to be view ed on a hold-to-
m aturity tim e horizon in order to capture any default losses that could occur until m aturity of the credit
exposure.
Analysing credit exposures on a m ulti-year basis enables the risk m anager to com pare exposures of different
size, credit quality, and m aturity. The loss distribution produced provides, for any chosen level of confidence, an
indication of the possible cum ulative losses that could be suffered until all the exposures have m atured.
The benefits of looking at portfolio credit risk from this view point include the follow ing:

The full term structure of default probabilities is taken into account.

The full uncertainty of default losses over the life of the portfolio is also captured.
For exam ple, because the one-year average default rates for investm ent grade obligors are relatively sm all but
the corresponding exposures m ay be large, a one-year tim e horizon m ay not be the best m easure for active
portfolio m anagem ent. H ow ever, a m ulti-year view w ill reflect the fact that defaults follow a decline in credit
quality over m any years.
CREDI TRI S K
+
21
3
CREDI TRI SK
+
Model
The CREDITRISK
+
Model
allows the portfolio of
exposures to be decomposed
into sectors to reflect the
degree of diversification and
concentration present.
Figure 8:
Impact of sectors on the
loss distribution
The CREDITRISK
+
Model
allows the risk of the portfolio
to be viewed on a hold-to-
maturity time horizon in order
to capture any default losses
that could occur until maturity
of the credit exposure.
N um ber of sectors
9
9
t
h

p
e
r
c
e
n
t
i
l
e

r
a
t
i
o
0. 50
0. 55
0. 60
0. 65
0. 70
0. 75
0. 80
0. 85
0. 90
0. 95
1. 00
0 1 2 3 4 5 6 7 8 9
3.5.1 Using the CREDITRI SK
+
Model to Calculate Multi-Year Loss Distributions
The C R ED ITR ISK
+
M odel can be used to calculate m ulti-year loss distributions by decom posing the exposure
profile over tim e into separate elem ents of discrete tim e, w ith the present value of the rem aining exposure in
each tim e period being assigned a m arginal default probability relevant to the m aturity and credit quality. These
decom posed exposure elem ents can then be used by the C R ED ITR ISK
+
M odel to generate a loss distribution
on a hold-to-m aturity basis.
3.6 Summary of the CREDITRI SK
+
Model
The key features of the C R ED ITR ISK
+
M odel are:

The CREDIT RISK

+
Model captures the essential characteristics of credit default events. C redit default
events are rare and occur in a random m anner w ith observed default rates varying significantly from year
to year. The approach adopted reflects these characteristics by m aking no assum ptions about the tim ing
or causes of default events and by incorporating the default rate volatility. B y taking a portfolio approach,
the benefits of diversification that arise from a large num ber of individual risks are fully captured.
C oncentration risk, resulting from groups of obligors that are affected by com m on factors, is m easured
using sector analysis.

The CREDIT RISK

+
Model is scaleable and computationally efficient. The C R ED ITR ISK
+
M odel is highly
scaleable and hence is capable of handling portfolios containing large num bers of exposures. The low data
requirem ents and m inim um of assum ptions m ake the C R ED ITR ISK
+
M odel easy to im plem ent for a w ide
range of credit risk portfolios, regardless of the specific nature of the obligors. Furtherm ore, the efficiency
of the m odel allow s com prehensive sensitivity analysis to be perform ed on a continuing basis, w hich is a
key requirem ent for the ability to quantify the effects of param eter uncertainty.
22
CREDIT FIRST
SUISSE B O STO N
E
C
CREDI TRI S K
+
23
4.1 Introduction to Economic Capital
4.1.1 The Role of Economic Capital
The analysis of uncertainty is the essence of risk m anagem ent. Therefore, m easuring the uncertainty or
variability of loss and the related likelihood of the possible levels of unexpected losses in a portfolio of
exposures is fundam ental to the effective m anagem ent of credit risk. Sufficient earnings should be generated
through adequate pricing and provisioning to absorb any expected loss. The expected loss is one of the costs
of transacting business w hich gives rise to credit risk. H ow ever, economic capital is required as a cushion for
a firms risk of unexpected credit default losses, because the actual level of credit losses suffered in any one
period could be significantly higher than the expected level.
4.2 Economic Capital for Credit Risk
4.2.1 Credit Default Loss Distribution
Know ledge of the credit default loss distribution arising from a portfolio of exposures provides a firm w ith
m anagem ent inform ation on the am ount of capital that the firm is putting at risk by holding the credit portfolio.
G iven that econom ic capital is necessary as a cushion for a firm s risk of unexpected credit default losses, a
percentile level provides a m eans of determ ining the level of econom ic capital for a required level of
confidence. In order to capture a significant proportion of the tail of the credit default loss distribution, the 99th
percentile unexpected loss level over a one-year tim e horizon is a suitable definition for credit risk econom ic
capital. This can be seen in the follow ing figure.
Econo
Capita
Economic Capital for Credit Risk 4
Economic Capital
Credit Default
Loss Distribution
Scenario Analysis
4.2.2 Benefits and Features of Economic Capital
Econom ic capital as a m easure of risk being taken by a firm has several features and benefits including the
follow ing:

It is a m ore appropriate m easure of the econom ic risk than that specified under the current regulatory regim e.

It m easures econom ic risk on a portfolio basis and hence takes account of the benefits of diversification.

It is a m easure that objectively differentiates betw een portfolios by taking account of credit quality and size
of exposure.

It is a dynam ic m easure, w hich reflects the changing risk of a portfolio and hence can be used as a tool
for portfolio optim isation.
4.3 Scenario Analysis
4.3.1 The Role of Scenario Analysis
The purpose of scenario analysis is to identify the financial im pact of low probability but nevertheless plausible
events that m ay not be captured by a statistically based m odel. Therefore, the use of a credit risk m odel should
be supplem ented by a program m e of stress testing of the assum ptions used.
There are tw o types of stress tests that should be perform ed: (i) scenario analysis w ithin the C R ED ITR ISK
+
M odel, and (ii) scenario analysis outside the C R ED ITR ISK
+
M odel.
4.3.2 Scenario Analysis within the CREDITRI SK
+
Model
The inputs into the C R ED ITR ISK
+
M odel can be stressed individually or in com bination. For exam ple, it is
possible to sim ulate dow nturns in the econom y by increasing default rates and default rate volatilities - sectors
of the portfolio can be stressed to varying degrees reflecting the fact that each sector could be affected to a
different extent. Sim ilarly, the financial im pact of rating dow ngrades can be assessed by increasing the default
rate assigned to an obligor. For a derivatives portfolio, this can be extended to include the effects of m ovem ents
in m arket rates on credit exposures.
G iven the efficient m anner in w hich the default loss distribution can be calculated, it is possible to calculate
the im pact of changing param eter inputs used by the m odel across a w ide range of values.
24
CREDIT FIRST
SUISSE B O STO N
Expected Loss
99th Percentile
Loss Level
Economic Capital
Figure 9:
Economic capital for
credit risk
Loss
P
r
o
b
a
b
i
l
i
t
y
Economic Capital
Credit Default
Loss Distribution
Scenario Analysis
4.3.3 Scenario Analysis outside the CREDITRI SK
+
Model
C ertain stress tests can be difficult to perform w ithin the C R ED ITR ISK
+
M odel: for exam ple, the im pact of
political or financial uncertainty w ithin a country. For these types of scenarios, analysis that is conducted
w ithout reference to the outputs of the C R ED ITR ISK
+
M odel, such as looking at the exposure at risk for a given
scenario, provides a realistic m eans of quantifying the financial im pact.
A firm should control the risk of catastrophic losses through the use of obligor and concentration lim its,
keeping any one of these lim its w ithin the loss for the percentile level used to determ ine the econom ic capital
given by the C R ED ITR ISK
+
M odel.
The figure below illustrates the w ay in w hich the distribution of losses can be considered to be divided into
three parts.
It is possible to control the risk of losses that fall w ithin each of the three parts of the loss distribution in the
follow ing w ays:
Part of loss distribution Control mechanism
U p to Expected Loss Adequate pricing and provisioning
Expected Loss - 99th Percentile Loss Econom ic capital and/or provisioning
G reater than 99th Percentile Loss Q uantified using scenario analysis and
controlled w ith concentration lim its
Scenario analysis deals w ith quantifying and controlling the risk of extrem e losses. Losses up to a certain
confidence level, such as the 99th percentile level, are controlled by the use of adequate pricing, provisioning
and econom ic capital. Provisioning for credit risk is discussed in detail in Section 5.2.
CREDI TRI S K
+
25
4
Economic Capital
Scenario analysis
provides a means of
quantifying catastrophic
losses - potential losses
can be controlled through
concentration limits.
Figure 10:
Parts of the credit default
loss distribution
Table 5:
Mechanisms for
controlling the risk of
credit default losses
Expected Loss
99th Percentile
Loss Level
Economic Capital
Loss
P
r
o
b
a
b
i
l
i
t
y
C overed by
pricing and
provisioning
C overed by capital
and/or provisions
Q uantified using scenario
analysis and controlled
w ith concentration lim its
Applic
CR ED IT
5.1 Introduction
C R ED ITR ISK
+
includes several applications of the credit risk m odelling technology in the areas of provisioning,
setting risk-based credit lim its, and portfolio m anagem ent.
5.2 Provisioning for Credit Risk
O ne application of C R ED ITR ISK
+
is in defining an appropriate credit risk provisioning m ethodology that reflects
the credit losses of the portfolio over several years and hence that m ore accurately presents the true earnings
of the business by m atching incom e w ith losses.
5.2.1 The Need for Credit Provisions
G enerally, current accounting and provisioning policies recognise credit incom e and credit losses at different
tim es, even though the tw o events are related. U sually, credit loss provisions are m ade only w hen exposures
have been identified as non-perform ing. These provisions are often supplem ented w ith other specific and
general credit provisions.
In relation to any portfolio of credit exposures, there is a statistical likelihood that credit default losses w ill occur,
even though the obligors are currently perform ing and it is not possible to identify specifically w hich obligors
w ill default. The level of expected loss reflects the continuing credit risk associated w ith the firm s existing
perform ing portfolio and is one of the costs of doing credit-related business. This level of expected loss should
be taken account of w hen executing any business that has a credit risk im pact.
W hen default losses are m odelled, it can be observed that the m ost frequent loss am ount w ill be m uch low er
than the average, because, occasionally, extrem ely large losses are suffered, w hich have the effect of increasing
the average loss. Therefore, a credit provision is required as a m eans of protecting against distributing excess
profits earned during the below average loss years.
26
CREDIT FIRST
SUISSE B O STO N
Applications of CREDI TRI SK
+
5
The CREDITRISK
+
credit
risk provisioning methodology
more accurately reflects the
true earnings of the business
by matching income with
losses.
Applications
Provisioning
Limits
Portfolio Management
c
T
CREDI TRI S K
+
27
5
Applications
The Annual Credit Provision
reflects the continuing credit
risk associated with the
portfolio and is one of the costs
of doing business that creates
credit risk.
Figure 11:
Credit risk provisioning
The Increm ental C redit
Reserve protects against
unexpected credit losses
and is used to absorb losses
that are higher than the
expected level.
5.2.2 Annual Credit Provision (ACP)
The starting point for provisioning is to separate the existing portfolio into a non-perform ing and a perform ing
portfolio. The non-perform ing portfolio should be fully provisioned to the expected recovery level available
through foreclosure, adm inistration or liquidation. O nce fully provisioned, the non-perform ing portfolio should
then be separated out and passed to a specialist team for ongoing m anagem ent.
As for the perform ing portfolio, since no default has occurred, one needs a forw ard-looking provisioning
m ethodology. U nder C R ED ITR ISK
+
, the Annual C redit Provision (AC P) represents the future expected credit
loss on the perform ing portfolio, w hich is calculated as follow s:
ACP = Exposure x Default Rate x (100% - Recover Rate)
The AC P should be calculated frequently in order to reflect the changing credit quality of the portfolio. The AC P
is the first elem ent of the credit provisioning m ethodology.
5.2.3 Incremental Credit Reserve (ICR)
The AC P represents only the expected or average level of credit losses. As experience show s, actual losses
that occur in any one year m ay be higher or low er than this am ount, depending on the econom ic environm ent,
interest rates, etc. In fact, a better w ay of view ing the annual credit loss of the portfolio is as a distribution of
possible losses (outcom es), w hose average equals the AC P but has a sm all probability of m uch larger losses.
In order to absorb these variations in credit losses from year to year, a second elem ent of the provisioning
m ethodology, the Increm ental C redit Reserve (IC R), can be established.
The C R ED ITR ISK
+
M odel provides inform ation on the distribution of possible losses in the perform ing credit
portfolio. The IC R provides protection against unexpected credit losses (i.e. in excess of the AC P) and is
subject to a cap derived from the C R ED ITR ISK
+
M odel (the IC R C ap). The IC R C ap represents an extrem e
case of possible credit losses (e.g. the 99th percentile loss level) on the perform ing portfolio.
ACP=Average level of credit losses
ICR Cap =99th Percentile
Loss Level
Typical
ICR
Loss
P
r
o
b
a
b
i
l
i
t
y
5.2.4 Provisioning for Different Business Lines
The credit risk provisioning m ethodology described above relates to credit risk arising from a loans business
w here the incom e is accounted for on an accruals basis rather than by m arking-to-m arket.
A credit risk provision can also be established for other credit business lines, such as traded bond portfolios
and derivatives portfolios. In each case, the C R ED ITR ISK
+
M odel provides the inform ation required in order to
establish the provision that ensures that the accounting principle of m atching incom e w ith losses is m aintained.
For exam ple, for a portfolio of bonds, part of the expected loss is incorporated w ithin the m arket price and
hence only the increm ental credit reserve is required. This is described in the follow ing table.
Portfolio type Accounting treatment Provision
Loan Accrual AC P (1 year) charge to P&L each year
(C ounterparty risk) (credit neutral) IC R (1 year)
Derivatives M ark-to-m arket AC P (full m aturity) held as m ark-to-m arket adjustm ent
(C ounterparty risk) (credit neutral) IC R (1 year)
Bond M ark-to-m arket IC R (1 year) to support business and protect against
(Issuer risk) (credit inclusive) distribution of profits
5.2.5 Managing the Credit Risk Provision
As credit defaults occur, loans or exposures are m oved from the perform ing to the non-perform ing portfolio
and hence provisioned to the expected recovery level. This increase in provision is then charged first against
the AC P and then, to the extent necessary, against the IC R. To the extent that actual credit losses are less than
the AC P w ithin any given year, the balance is credited to the IC R up to the IC R C ap, beyond w hich the balance
is taken into P&L. This ensures that the IC R is replenished during low loss years follow ing a large unexpected
loss, but that the IC R never exceeds the IC R C ap.
A w orked exam ple can be seen in the table below :
Year 1 2 3 4 5
Assumptions
Actual loan losses 500 600 300 300 650
AC P 500 525 550 610 625
IC R - Initial level 1,900 - - - -
IC R C ap 2,000 2,100 2,200 2,250 2,300
Income Statement
O perating profit 2,100 2,100 2,205 2,315 2,430
Less: AC P (500) (525) (550) (610) (625)
Add: excess unutilised provision over IC R C ap 0 0 0 135 0
Pre-tax profit 1,600 1,575 1,655 1,840 1,805
IC R (pre cap) 1,900 1,825 2,075 2,385 2,225
IC R C ap (as above) 2,000 2,100 2,200 2,250 2,300
Excess unutilised provision over IC R C ap 0 0 0 135 0
ICR (with cap applied) 1,900 1,825 2,075 2,250 2,225
28
CREDIT FIRST
SUISSE B O STO N
Table 6:
Provisioning for different
business lines
Table 7:
Example of credit risk
provisioning
CREDI TRI S K
+
29
5.3 Risk-Based Credit Limits
A system of individual credit lim its is a w ell-established m eans of m anaging credit risk. M onitoring exposures
against lim its provides a trigger m echanism for identifying potentially unw anted exposures that require active
m anagem ent.
5.3.1 Standard Credit Limits
The system of credit lim its m ay be view ed from a different perspective, if applying the m ethodologies described
w ithin this docum ent.
In particular, in order to equalise a firm s risk appetite betw een obligors as a m eans of diversifying its portfolio,
a credit lim it system could aim to have a large num ber of exposures w ith equal expected losses. The expected
loss for each obligor can be calculated as the default rate m ultiplied by the exposure am ount less the expected
recovery. This m eans that individual credit lim its should be set at levels that are inversely proportional to the
default rate corresponding to the obligor rating.
As m ight be expected, this m ethodology gives larger lim its for better ratings and shorter m aturities, but has the
benefit of allow ing a firm to relate the size and tenor of lim its for different rating categories to each other.
This approach can be extended to base lim its on equalising the portfolio risk contribution for each obligor.
A discussion on risk contributions and their use in portfolio m anagem ent is provided later in this section.
5.3.2 Concentration Limits
Any excess country or industry sector concentration can have a negative effect on portfolio diversification and
increase the riskiness of the portfolio. As a result, a com prehensive set of country and industry sector lim its is
required to address concentration issues in the portfolio. C oncentration lim its have the effect of lim iting the loss
from identified scenarios and is a pow erful technique for m anaging tailrisk and controlling catastrophic losses.
5.4 Portfolio Management
The C R ED ITR ISK
+
M odel m akes the process of controlling and m anaging credit risk m ore objective by
incorporating into a single m easure all of the factors that determ ine the am ount of risk.
5.4.1 Introduction
C urrently, the prim ary technique for controlling credit risk is through the use of lim it system s, including:

Individual obligor lim its to control the size of exposure

Tenor lim its to control the m axim um m aturity of transactions w ith obligors

Rating exposure lim its to control the am ount of exposure to obligors of certain credit ratings and

C oncentration lim its to control concentrations w ithin countries and industry sectors.
5
Applications
A system of standard
credit limits can be supple-
mented with portfolio level
risk information to manage
credit risk.
Applications
Provisioning
Limits
Portfolio Management
5.4.2 Identifying Risky Exposures
The risk of a particular exposure is determ ined by four factors: (i) the size of exposure, (ii) the m aturity of the
exposure, (iii) the probability of default, and (iv) the system atic or concentration risk of the obligor. C redit lim its
aim to control risk arising from each of these factors individually. H ow ever, for m anaging risks on a portfolio
basis, w ith the aim of creating a diversified portfolio, a different m easurem ent, w hich incorporates size, m aturity,
credit quality and system atic risk into a single m easure, is required.
5.4.3 Measuring Diversification
The loss distribution and the level of econom ic capital required to support a portfolio are m easures of portfolio
diversification that take account of the size, m aturity, credit quality and system atic risk of each exposure.
If the portfolio w ere less diversified, the spread of the distribution curve w ould be w ider and a higher level of
econom ic capital w ould be required. C onversely, if the portfolio w ere m ore diversified, a low er level of econom ic
capital w ould be required. These m easures can be used in m anaging a portfolio of exposures.
5.4.4 Portfolio Management using Risk Contributions
The risk contribution of an exposure is defined as the increm ental effect on a chosen percentile level of the
loss distribution w hen the exposure is rem oved from the existing portfolio. If the percentile level chosen is the
sam e as that used for calculating econom ic capital, the risk contribution is the increm ental effect on the
am ount of econom ic capital required to support the portfolio.
Risk contributions have several features including the follow ing:

The total of the risk contributions for the individual obligors is approxim ately equal to the risk of the entire
portfolio.

Risk contributions allow the effect of a potential change in the portfolio (e.g. the rem oval of an exposure)
to be m easured.

In general, a portfolio can be effectively m anaged by focusing on a relatively few obligors that represent a
significant proportion of the risk but constitute a relatively sm all proportion of the absolute portfolio
exposures.
Therefore, risk contributions can be used in portfolio m anagem ent. B y ranking obligors in decreasing order of
risk contribution, the obligors that require the m ost econom ic capital can easily be identified.
This is illustrated in the follow ing exam ple. A portfolio w as created from w hich a sm all num ber of exposures
w ith the highest risk contributions w ere rem oved. The effect on the loss distribution and the levels for the
expected loss and the econom ic capital can be seen in the figure opposite.
30
CREDIT FIRST
SUISSE B O STO N
For managing credit
risks on a portfolio basis,
with the aim of creating
a diversified portfolio, a
different measure that
incorporates the magnitude
of the exposure, maturity,
credit quality and systematic
risk into a single measure
is required.
In general, a portfolio
can be effectively managed
by focusing on a relatively
few obligors that represent
a significant proportion of
the risk but constitute a
relatively small proportion
of the absolute portfolio
exposures.
Applications
Provisioning
Limits
Portfolio Management
The reduction in the 99th percentile loss level is larger than the reduction in the expected loss level, w hich
leads to an overall reduction in the econom ic capital required to support the portfolio.
5.4.5 Techniques for Distributing Credit Risk
O nce obligors representing a significant proportion of the risk have been identified, there are several
techniques for distributing credit risk that can be applied. These include the follow ing:

C ollateralisation: In the context of the C R ED ITR ISK

+
M odel, taking collateral has the effect of reducing the
severity of the loss given that the obligor has defaulted.

Asset securitisations: Asset securitisations involve the packaging of assets into a bond, w hich is then sold
to investors.

C redit derivatives: C redit derivatives are a m eans of transferring credit risk from one obligor to another,
w hile allow ing client relationships to be m aintained.
CREDI TRI S K
+
31
5
Applications
Original 99th Percentile
Loss Level
New 99th Percentile
Loss Level
New Expected Loss
Original Expected Loss
Figure 12:
Using risk contributions
Inform ation about risk
contributions can be used
to facilitate risk m anagem ent
and efficient use of econom ic
capital.
New Economic Capital
Original Economic Capital
Loss
P
r
o
b
a
b
i
l
i
t
y
A1 Overview of this Appendix
This appendix presents an analytic technique for generating the full distribution of losses from a portfolio of
credit exposures. The technique is valid for any portfolio w here the default rate for each obligor is sm all and
generates both one-year and m ulti-year loss distributions.
The appendix applies the concepts discussed in Sections 2 and 3 of this docum ent. The key concepts are:

D efault rates are stochastic.

The level of default rates affects the incidence of default events but there is no causal relationship betw een
the events.
In order to facilitate the explanation of the C R ED ITR ISK
+
M odel, w e first consider the case in w hich the m ean
default rate for each obligor in the portfolio is fixed. W e then generalise the technique to the case in w hich the
m ean default rate is stochastic. The m odelling stages of the C R ED ITR ISK
+
M odel and the relationships betw een
the different sections of this appendix are show n in the figure opposite.
32
CREDIT FIRST
SUISSE B O STO N
CR ED IT
M odel
Appendix A - The CREDI TRI SK
+
Model A
A2 Default Events with Fixed Default Rates
In Sections A2 to A5 w e develop the theory of the distribution of credit default losses under the assum ption
that the default rate is fixed for each obligor. G iven this assum ption and the fact that there is no causal
relationship betw een default events, w e interpret default events to be independent. In Sections A6 onw ards,
the assum ption of fixed default rates is relaxed, w hich introduces dependence betw een default events.
In Section A13 this dependence is quantified by calculating the correlation betw een default events im plied by
the C R ED ITR ISK
+
M odel.
CREDI TRI S K
+
33
A
The CREDI TRI SK
+
Model
Figure 13:
Flowchart description of
Appendix A
T
D efault events w ith
fixed default rates
D efault losses w ith
fixed default rates
C alculation procedure
for loss distribution w ith
fixed default rates
C onvergence of variable
default rate case to fixed
default rate case
Application to
m ulti-year losses
D efault rate uncertainty Sector analysis
D efault events w ith
variable default rates
D efault losses w ith
variable default rates
C alculation procedure for
loss distribution w ith
variable default rates
G eneral sector analysis
Risk contributions and
pairw ise correlations
A2
A3
A4
A6 A7
A8
A9
A10
A12 A13
A11
A5
A2.1 Default Events
C redit defaults occur as a sequence of events in such a w ay that it is not possible to forecast the exact tim e
of occurrence of any one default or the exact total num ber of defaults. In this section w e derive the basic
statistical theory of such processes in the context of credit default risk.
C onsider a portfolio consisting of N obligors. In line w ith the above assum ptions, it is assum ed that each
exposure has a definite know n probability of defaulting over a one-year tim e horizon. Thus
(1)
To analyse the distribution of losses arising from the w hole portfolio, w e introduce the probability generating
function defined in term s of an auxiliary variable z by
(2)
An individual obligor either defaults or does not default. The probability generating function for a single obligor
is easy to com pute explicitly as
(3)
As a consequence of independence betw een default events, the probability generating function for the w hole
portfolio is the product of the individual probability generating functions. Therefore
(4)
It is convenient to w rite this in the form
(5)
Suppose next that the individual probabilities of default are uniform ly sm all. This is characteristic of portfolios
of credit exposures. G iven that the probabilities of default are sm all, pow ers of those probabilities can be
ignored. Thus, the logarithm s can be replaced using the expression
4
(6)
and, in the lim it, equation (5) becom es
(7)
w here w e w rite
(8)
for the expected num ber of default events in one year from the w hole portfolio.
To identify the distribution corresponding to this probability generating function, w e expand F(z) in its
Taylor series:
(9)
34
CREDIT FIRST
SUISSE B O STO N
A obligor for default of y probabilit Annual
A
p

0
defaults) n ( ) (
n
n
z p z F
) 1 ( 1 1 ) ( + + z p z p p z F
A A A A

+
A
A
A
A
z p z F z F )) 1 ( 1 ( ) ( ) (

+
A
A
z p z F )) 1 ( 1 log( ) ( log
) 1 ( )) 1 ( 1 log( + z p z p
A A
4
This approximation
ignores terms of
degree 2 and higher in
the default probabilities.
The expressions derived
from this approximation
are exact in the limit as
the probabilities of default
tend to zero, and give
good approximations
in practice.
) 1 (
) 1 (
) (

z
z p
e e z F
A
A

A
A
p
n
n
n
z z
z
n
e
e e e z F

0
) 1 (
!
) (

CREDI TRI S K
+
35
Thus if the probabilities of individual default are sm all, although not necessarily equal, then from equation (9)
w e deduce that the probability of realising n default events in the portfolio in one year is given by
(10)
A2.2 Summary
In equation (10) w e have obtained the w ell-know n Poisson distribution for the distribution of the num ber of
defaults under our initial assum ptions. The follow ing should be noted:

The distribution has only one param eter, the expected num ber of defaults . The distribution does not
depend on the num ber of exposures in the portfolio or the individual probabilities of default provided that
they are uniform ly sm all.

There is no necessity for the exposures to have equal probabilities of default; indeed, the probability of
default can be individually specified for each exposure if sufficient inform ation is available.
The Poisson distribution w ith m ean can be show n to have standard deviation given by . H istorical evidence
of the standard deviation of default event frequencies exists in the form of year-on-year default rate tables.
Such data suggests that the actual standard deviation is invariably m uch larger than . Thus, our initial
assum ption of fixed default rates cannot account for observed data. B efore addressing this in Section A6, w e
first consider the derivation of the credit loss distribution from the results on default events above, retaining
our initial assum ptions for now.
A3 Default Losses with Fixed Default Rates
A3.1 Introduction
U nder our initial assum ptions, the distribution of num bers of defaults in a portfolio of exposures in one year
has been obtained. H ow ever, our m ain objective is to understand the likelihood of suffering given levels of loss
from the portfolio, rather than given num bers of defaults. The distributions are different because the sam e level
of default loss could arise equally from a single large default or from a num ber of sm aller defaults in the sam e
year. U nlike the variation of default probability betw een exposures, w hich does not influence the distribution of
the total num ber of defaults, differing exposure am ounts result in a loss distribution that is not Poisson in
general. M oreover, inform ation about the distribution of different exposures is essential to the overall
distribution. H ow ever, it is possible to describe the overall distribution because its probability generating
function has a sim ple closed form am enable to com putation.
A3.2 Using Exposure Bands
The first step in obtaining the distribution of losses from the portfolio in an am enable form is to group the
exposures in the portfolio into bands. This has the effect of significantly reducing the am ount of data that m ust
be incorporated into the calculation.
B anding introduces an approxim ation into the calculation. H ow ever, provided the num ber of exposures is large
and the w idth of the bands is sm all com pared w ith the average exposure size characteristic of the portfolio, the
approxim ation is insignificant. Intuitively, this corresponds to the fact that the precise am ounts of exposures in
a portfolio cannot be critical in determ ining the overall risk.
A
The CREDI TRI SK
+
Model
!
defaults) (n y Probabilit
n
e
n

O nce the appropriate notation has been set up, an estim ate of the effect of banding on the m ean and standard
deviation of the portfolio is given below.
A3.3 Notation
In this section, the notation used for the exposure banding described above is detailed.
Reference Symbol
O bligor A
Exposure L
A
Probability of default P
A
Expected Loss
A
In order to perform the calculations, a unit am ount of exposure L, denom inated in a base currency, is chosen.
For each obligor A, define num bers
A
and
A
by w riting
and (11)
Thus,
A
and
A
are the exposure and expected loss, respectively, of the obligor, expressed as m ultiples of
the unit.
The key step is to round each exposure size
A
to the nearest w hole num ber. This step replaces each exposure
am ount L
A
by the nearest integer m ultiple of L. If a suitable size for the unit L is chosen, then, after the rounding
has been perform ed for a large portfolio, there w ill be a relatively sm all num ber of possible values for
A
each
shared by several obligors.
The portfolio can then be divided into m exposure bands, indexed by j, w here 1 j m . W ith respect to the
exposure bands, w e m ake the follow ing definitions
Reference Symbol
C om m on exposure in Exposure B and j in units of L
j
Expected loss in Exposure B and j in units of L
j
Expected num ber of defaults in Exposure B and j
j
The follow ing relations hold, expressing the expected loss in term s of the probability of default events
;hence (12)
N ote that, because w e have rounded the
j
to m ake them w hole num bers, the expected loss
A
w ill be affected,
by equation (12) unless a com pensating rounding adjustm ent is m ade to the expected num ber of default
events
j
. If no adjustm ent is m ade, the rounding process w ill result in a sm all rounding up of the expected loss.
U nder the assum ption stated above, that the unit size is sm all relative to the typical exposure size of the
portfolio, these approaches each have an im m aterial effect on the loss distribution. In the rest of this Appendix
it is assum ed that an adjustm ent to the default probabilities to preserve the expected losses is m ade. Provided
the exposure sizes are rounded up, then, as w ill be show n in Section A4.2, the rounding leads to a sm all
overstatem ent of the standard deviation.
36
CREDIT FIRST
SUISSE B O STO N
A A
L L
A A
L
j j j

j A
A
A
A
j
j
j

:
CREDI TRI S K
+
37
As in equation (8), let stand for the total expected num ber of default events in the portfolio in one year. Since
is the sum of the expected num ber of default events in each exposure band, w e have
(13)
A3.4 The Distribution of Default Losses
W e have analysed the distribution of default events under our initial assum ptions. W e now proceed to derive
the distribution of default losses.
Intuitively, the default loss analysis involves a second elem ent of random ness, because som e defaults lead to
larger losses than others through the variation in exposure am ounts over the portfolio. As w ith default events,
the second random effect is best described m athem atically through its probability generating function. Thus,
let G (z) be the probability generating function for losses expressed in m ultiples of the unit L of exposure:
(14)
The exposures in the portfolio are assum ed to be independent. Therefore, the exposure bands are independent,
and the probability generating function can be w ritten as a product over the exposure bands
(15)
H ow ever, by treating each exposure band as a portfolio and using equation (9), w e obtain
(16)
Therefore
(17)
This is the desired form ula for the probability generating function for losses from the portfolio as a w hole.
In the next section, w e show how to use the probability generating function to derive the actual distribution
of losses under our initial assum ptions.
For later reference, equation (17) can be restated in a slightly different form . First, define a polynom ial P(z)
as follow s
(18)
w here w e have used equations (12) and (13) for the total num ber of defaults in the portfolio. The probability
generating function in equation (17) can now be expressed as
(19)
This functional form for G (z) expresses m athem atically the com pounding of tw o sources of uncertainty arising,
respectively, from the Poisson random ness of the incidence of default events and the variability of exposure
am ounts w ithin the portfolio.
A
The CREDI TRI SK
+
Model



m
j
j
j
m
j
j
1 1

0
L) n losses aggregate ( ) (
n
n
z p z G
) ( ) (
1
z G z G
m
i
i

j
j j j
j
j
z n
n n
n
j n
j
e z
n
e
z p z G

0 0
!
defaults) n ( ) (



+

m
j
m
j
j
j j
j
j j
z m
j
z
e e z G
1 1
1
) (

,
_

,
_

m
j
j
j
m
j
j
j
m
j
j
j
j
z
z
z P
1
1
1
) (

)) ( ( ) (
) 1 ) ( (
z P F e z G
z P


N ote that G (z) depends only on the data and . Therefore, to obtain the distribution of losses for a large
portfolio of credit risks, all that is needed is know ledge of the different sizes of exposures w ithin the portfolio,
together w ith the share of expected loss arising from each exposure size. This is typically a very sm all am ount
of data, even for a large portfolio.
A4 Loss Distribution with Fixed Default Rates
A4.1 Calculation Procedure
In this section, a com putationally efficient m eans of deriving the actual distribution of credit losses is derived
from the probability generating function given by equation (17). In Section A10, this approach w ill be
generalised to com pute the distribution for the C R ED ITR ISK
+
M odel.
For n an integer let A
n
be the probability of a loss of nxL, or n units from the portfolio. W e w ish to com pute A
n
efficiently. C om paring the definition in equation (14) w ith the Taylor series expansion for G (z), w e have
(20)
In our case G (z) is given in closed form by equation (17). U sing Leibnitzs form ula w e have
(21)
H ow ever
(22)
and by definition
(23)
Therefore
(24)
U sing the relation
j
=
j
x
j
from equation (12), the follow ing recurrence relationship is obtained
(25)
This recurrence relationship allow s very quick com putation of the distribution. In order to com m ence the
com putation, w e have the follow ing form ula for the first term , w hich expresses the probability of no loss arising
from the portfolio
(26)
38
CREDIT FIRST
SUISSE B O STO N
n
z
n
n
A
dz
z G d
n
p
0
) (
!
1
) nL of loss (
0
1
1
1
0
). (
!
1 ) (
!
1

,
_

z
m
j
v
j
n
n
z
n
n
j
z
dz
d
z G
dz
d
n
dz
z G d
n

0
1
0 1
1
1
1
1
) (
1
!
1

+
+

,
_

,
_

z
n
k
m
j
j
k
k
k n
k n
j
z
dz
d
z G
dz
d
k
n
n

'
+

,
_

+
+

otherwise 0
some for 1 k if )! 1 (
0
1
1
1
j v k
z
dz
d
j j
z
m
j
v
j
k
k
j

1
0
1
1
)! 1 ( ) (

k n
z
k n
k n
A k n z G
dz
d
j
j
j
v n
n v j
j j
k n j
j v k
n k
n
A
n
v
A k n k
k
n
n
A

,
_

:
1
some for 1
1
)! 1 ( )! 1 (
1
!
1

j
j
v n
n v j
j
n
A
n
A

m
j j
j
e e P F G A
1
)) 0 ( ( ) 0 (
0

CREDI TRI S K
+
39
Again, it is w orthw hile to note that the calculation depends only on know ledge of and . In practice, these
represent a very sm all am ount of data even for a large portfolio consisting of m any exposures.
A4.2 Precision Using Exposure Bands
The banding process described in Section A3 introduces a sm all degree of approxim ation into the data. In this
section, w e show that the approxim ation error is norm ally not m aterial by considering the effect on the portfolio
m ean and standard deviation.
In term s of the notation above, the total portfolio expected loss and total portfolio standard deviation are
; (27)
w here the expected loss and standard deviation are expressed in the chosen unit L.
In order to represent the banding, suppose that the above are expressions for the truem ean and standard
deviation, but that now the are rounded to integer m ultiples of the unit as explained above. This process
introduces an error; how ever, w rite
w here (28)
Each
j
is at m ost of absolute size one. It is assum ed that the exposures are rounded up, so that each
j
is positive.
The expected loss is unaffected by the m ethod of rounding chosen, because its expression is independent of
the banded exposure am ounts. This w as noted above.
For the standard deviation, w e have
(29)
w here is the expected loss for the portfolio. Taking square roots and neglecting higher-order term s in the
Taylor series w e obtain
(30)
For a real portfolio, the expected loss and the quantity 2 are of the sam e order. W e conclude that:

The expected loss calculated by the m odel is unaffected by the banding process.

The standard deviation is overstated by an am ount com parable w ith the chosen unit size.
A5 Application to Multi-Year Losses
A5.1 Introduction
The recurrence relation above w as derived on the basis of a one-year loss distribution. In this section it is
show n how the initial m odel can be applied over a m ulti-year tim e horizon.
As in the discussion over a one-year horizon, consider a portfolio of obligors w ith sm all probabilities of default.
For sim plicity it is assum ed that the future of the portfolio is divided into years. The exposures are perm itted to
vary from year to year. In particular, each exposure has an individual m aturity corresponding to the norm al
m aturity of bonds, loans or other instrum ents.
A
The CREDI TRI SK
+
Model

m
j
j
1

m
j
j j
1
2

j j j
+

1 0
j

+ + +


2
1
2
1
2
1
2 2

m
j
j
m
j
j j
m
j
j j

2
2
1
2
+

,
_

+
A5.2 Term Structure of Default
In order to address a m ulti-year horizon, m arginal rates of default m ust be specified for each future year for
each obligor in the portfolio. C ollectively, such m arginal default rates give the term structure of default for
the portfolio.
A5.3 Notation
Fix the follow ing notation:
Reference Symbol
(t)
Probability of default of exposure j in year t p
j
(t) (t)
Am ount of exposure j in year t L
j
= L
j
(t) (t)
Expected loss in exposure j in year t
j
= L
j
As for the one-year discussion, L is the unit of exposure and the
j
(t)
are dim ensionless w hole num bers. U nder
the natural assum ption that defaults by the sam e exposure in different years are m utually exclusive, the
probability generating function for m ulti-year losses from a single exposure is given by
(31)
For the generating function of total losses, w e have
(32)
In the lim it of sm all probabilities of default w e argue as for equation (6) to obtain
(33)
and w e obtain
(34)
and
(35)
The probability generating function for the loss distribution is therefore given by
(36)
This has the sam e form as the one year probability generating function (17). Therefore, the recurrence relation
given by equation (25) for the distribution of losses over one year is also applicable to the calculation of the
m ulti-year distribution of losses
(37)
40
CREDIT FIRST
SUISSE B O STO N


+ +
T
t
t
j
T
t
t
j
T
t
t
j j
t
j
t
j
z p z p p z G
0
) (
0
) (
0
) (
) 1 ( 1 1 ) (
) ( ) (

,
_

+
j
T
t
t
j
t
j
z p z G
1
) (
) 1 ( 1 log ) ( log
) (

,
_

+
T
t
t
j
T
t
t
j
t
j
t
j
z p z p
1
) (
1
) (
) 1 ( ) 1 ( 1 log
) ( ) (

,
_

) (
) (
: ,
) (
1
) (
) 1 ( ) ( log
t
j
t
j
t j
t
j
j
T
t
t
j
p z z p z G

T
t
j
t
j
t
j
T
t
j
t
j
t j
t
j
t
j
t
j
t
j
p p
1
: ,
) (
) (
1
: ,
) (
: ,
) (
) ( ) ( ) (

t j
t
j
t
j
t
j
z
e z G
,
) (
) (
) (
) 1 (
) (

n v t j
v n
t
j
n
t
j
t
j
A
n
A
) (
) (
; ,
) (

CREDI TRI S K
+
41
A6 Default Rate Uncertainty
The previous sections developed the theory of the loss distribution from a portfolio of obligors, each of w hich
has a fixed probability of default. In the follow ing sections, the C R ED ITR ISK
+
M odel w ill be developed from this
theory by incorporating default rate uncertainty and sector analysis. These concepts are introduced in this
section and Section A7 respectively.
Published statistics on the incidence of default events, for exam ple am ong rated com panies in a given country,
show that the num ber of default events, and therefore the average probability of default for such entities,
exhibits w ide variation from year to year
5
.
Such year-on-year statistics m ay be thought of as sam ples from a random variable w hose expected value
represents an average rate of default over m any years. The appearance of random ness is due to the incidence
of factors, such as the state of the econom y, that influence the fortunes of obligors. The standard deviation of
the variable m easures our uncertainty as to the actual default rate that w ill be exhibited over a given year.
O w ing to default rate uncertainty, there is a chance that default rates w ill turn out to be higher over, for exam ple,
the next year than their average over m any years suggests. This in turn leads to a higher chance of
experiencing extrem e losses.
The situation m ay be sum m arised by the follow ing three intuitive facts about default rate uncertainty:

O bserved default probabilities are volatile over tim e, even for obligors having com parable credit quality.

The variability of default probabilities can be related to underlying variability in a relatively sm all num ber of
background factors, such as the state of the econom y, w hich affect the fortunes of obligors. For exam ple,
a dow nw ard trend in the state of the econom y m ay m ake m ost obligors in a portfolio m ore likely to default.

H ow ever, a change in the econom y or another factor w ill not cause obligors to default w ith certainty.
W hatever the state of the econom y, actual defaults should still be relatively rare events. Therefore the
analysis above w hich considered rare events is relevant in a suitably m odified form .
The second point m ade above is that uncertainty arises from factors that m ay affect a large num ber of obligors
in the sam e w ay. In order to m easure this effect and hence quantify the im pact of individual default rate
volatilities at the portfolio level, the concept of sector analysis is necessary. This concept is introduced in the
next section.
A7 Sector Analysis
A7.1 Introduction
It w as noted above that the variability of default rates can be related to the influence of a relatively sm all
num ber of background factors on the obligors w ithin a portfolio. In order to m easure the effect of these factors,
it is necessary to quantify the extent to w hich each factor has an influence on a given portfolio of obligors.
A factor such as the econom y of a particular country m ay be considered to have a uniform influence on obligors
w hose dom icile is w ithin that country, but relatively little influence on other obligors in a m ultinational portfolio.
In this section, the m easurem ent of background factors is addressed by dividing the obligors am ong different
sectors, w here each sector is a collection of obligors under the com m on influence of a m ajor factor affecting
default rates. An initial exam ple m ight be a division of the portfolio according to the country of dom icile of each
obligor. In Section A12, a m ore general sector analysis, w hich allow s for the fact that, in reality, obligors m ay
be under the sim ultaneous influence of a num ber of m ajor factors, is presented.
A
The CREDI TRI SK
+
Model
5
If the default rates of obligors
were fixed, default events
would still have a non-zero
standard deviation arising
from the randomness of the
default events themselves.
However, as remarked in
Section A2.2, comparison
with historic data shows
that observed volatility is
far higher than can be
accounted for in this way.
A7.2 Further Notation
N ew notation is needed to keep track of the division of the portfolio into sectors and to record the volatility of
the default rate for each sector. W rite S
k
: 1 k n for the sectors, each of w hich should be thought of for now
as a subset of the collection of obligors.
The C R ED ITR ISK
+
M odel regards each sector as driven by a single underlying factor, w hich explains the
variability over tim e in the average total default rate m easured for that sector. The underlying factor influences
the sector through the total expected rate of defaults in the sector, w hich is then m odelled as a random variable
w ith m ean
k
and standard deviation
k
specified for each sector. The standard deviation w ill reflect the degree
to w hich the probabilities of default of the obligors in the portfolio are liable to all be m ore or less than their
average levels. For exam ple, in a sector consisting of a large num ber of obligors of low credit quality, the m ean
default rate m ight be 5% per annum and the standard deviation of the actual default rate m ight be a sim ilar
quantity. Then there w ill be a substantial chance in any year of the actual average probability of default in the
sector being, say, 10% instead of 5% . In turn it is m uch m ore likely that, say, 12% of the obligors w ill actually
default in that year. H ad the standard deviation been zero, reflecting that w e w ere certain about the probability
of each obligor defaulting, then a year in w hich as m any as 12% of the obligors default w ould have been a
m uch m ore rem ote possibility.
The table below sum m arises the new notation to specify the sector decom position of the portfolio. In particular,
for each sector w e introduce a random variable x
k
representing the average default rate over the sector.
The m ean of x
k
is
k
and the standard deviation is
k
.
Sector S
k
: 1 k n
Random variable representing the m ean num ber of defaults x
k
Long-term annual average num ber of defaults - m ean of x
k

k
Standard deviation of x
k

k
For each sector, the data requirem ents are set out below. O ur original notation set up in Section A3.3 is also
repeated for com parison
Exposure Data within Sector Previous New
Notation Notation
B ase unit of exposure L L
(k) (k)
Exposure sizes in units L
j
= L
j
L
j
= L
j
1 j m 1 k n ;1 j m(k)
(k) (k)
Expected loss in each exposure band in units
j
= L
j

j
= L
j
1 j m 1 k n ;1 j m(k)
The m ean
k
is related to the expected loss data by the follow ing relation w hich is the analogue of
equation (13):
(38)
42
CREDIT FIRST
SUISSE B O STO N

) (
1
) (
) (
k m
j
k
j
k
j
k

CREDI TRI S K
+
43
A7.3 Estimating the Variability of the Default Rate
For each sector, in addition to the expected total rate of default
k
over the sector given by equation (38),
w e m ust specify a standard deviation
k
of the total expected rate of default. W e discuss a convenient w ay
to estim ate the standard deviation by reference to equation (38) for the m ean. Although equation (38)
is expressed in term s of exposure bands, it can equivalently be expanded as a sum over all the obligors in
the sector
(39)
w here the sum m ation extends over all obligors A belonging to sector k, and the relation
(40)
expresses the average probability of default of the obligor over the tim e period. To obtain an estim ate of the
standard deviation of each sector, w e assum e that, together w ith a probability of default p
A
, a standard deviation

A
has been assigned for the default rate for each obligor w ithin the sector. A convenient w ay to do this is to
assum e that the standard deviation depends on the credit quality of the obligor. This pragm atic m ethod
assum es that the credit quality of the obligors w ithin a sector is a m ore significant influence on the volatility of
the expected default frequency than the nature of the sector.
W e obtain an estim ate of
k
from the set
A
of obligor standard deviations by an averaging process. Recall
that only one random variable, x
k
is held to account for the uniform variability of each of the probabilities
of default. That is, the actual default probability for each obligor in the sector w ill be m odelled as a random
variable proportional to x
k
, w hose m ean is equal to the specified m ean default rate for that obligor. To express
this dependence w rite x
A
for the random default probability of the single obligor A. O ur assum ption can then
be w ritten
(41)
N ote that the m ean of x
A
is correctly specified as p
A
by this equation. Assum ing equation (41), in particular,
w e have
(42)
w here w e have used equation (39). The sum runs over all obligors in the sector. W e estim ate the standard
deviation of this sector so as to ensure that this condition holds. Thus the standard deviation of the m ean
default rate for a sector is estim ated as the sum of the estim ated standard deviations for each obligor in the
sector. An alternative and m ore intuitive description of the standard deviation
k
determ ined in this w ay is that
the ratio of
k
to the m ean
k
is an average of the ratio of standard deviation to m ean for each obligor,
w eighted by their contribution to the default rate. This is easily seen as follow s. B y equation (42)
(43)
A
The CREDI TRI SK
+
Model

A
A
A
k

A
A
A
p

k
k
A
A
A
x
x

k
A
A
A
k
k
A
k
k
A
A
A
A

,
_

A
A
A
A
A
A
A
A
A
A
k
k
p
p
p
p

According to historical experience, the ratio

A
/p
A
is typically of the order of one, so that the standard deviation
of the num ber of defaults observed year on year am ong obligors of sim ilar credit quality is typically of the
sam e order as the average annual num ber of defaults. Equation (43) show s that the sam e is true for
each sector, as one w ould expect. In the absence of detailed data, the obligor specific estim ates of the ratio

A
/p
A
can be replaced by a single flat ratio. Then, w riting
k
for this uniform ratio, equation (43) reduces to
the sim ple form
(44)
If the nature of the sector m ade it m ore appropriate to estim ate the standard deviation
k
directly, this w ould
be equivalent to estim ating the flat ratio
k
directly.
A8 Default Events with Variable Default Rates
A8.1 Conditional Default Rate
In this section, the distribution of default events for the C R ED ITR ISK
+
M odel is obtained. This is achieved by
calculation of the probability generating function. M ost of the w ork has been done already in the calculation of
the probability generating function in equation (7) w hen the default rate is fixed. As in equation (2), the
probability generating function for default events is w ritten
(45)
B ecause the sectors are independent, F(z) can be w ritten as a product over the sectors
(46)
W e therefore focus on the determ ination of F(z) for a single sector. In the notation of Section A7, the average
default rate in sector k is a random variable, w ritten x
k
, w ith m ean
k
and standard deviation
k
. C onditional
on the value of x
k
, w e can w rite dow n the probability generating function for the distribution of default events
as follow s
(47)
w here equation (7) has been used. Suppose that x
k
has probability density function
k
(x), so that
(48)
Then, the probability generating function for default events in one sector is the average of the conditional
probability generating function given by equation (47) over all possible values of the m ean default rate, as the
follow ing com putation show s:
(49)
In order to obtain an explicit form ula for the probability generating function, an appropriate distribution for X
k
m ust be chosen. W e m ake the key assum ption that x
k
has the G am m a distribution w ith m ean
k
and standard
deviation
k
.
The G am m a distribution is chosen as an analytically tractable tw o-param eter distribution. B efore proceeding to
evaluate equation (49) explicitly, the basic properties of the G am m a distribution are stated.
44
CREDIT FIRST
SUISSE B O STO N
k k k

0
defaults) n ( ) (
n
n
z p z F

n
k
k
z F z F
1
) ( ) (
[ ]
) 1 (
) (


z x
k k
e x x z F
dx x f dx x x x P
k k
) ( ) ( +

0
) 1 (
0
0
0
) ( ) ( ) defaults n ( ) defaults n ( ) (
x
z x
n
x
n
n
n
k
dx x f e dx x f x P z z P z F
CREDI TRI S K
+
45
A8.2 Properties of the Gamma Distribution
The G am m a distribution, w ritten (, ), is a skew distribution, w hich approxim ates to the N orm al distribution
w hen its m ean is large. The probability density function for a (, ) - distributed random variable X is
given by
(50)

w here () = e
-x
x
-1
dx is the G am m a function.
x = 0
The G am m a distribution (, ) is a tw o param eter distribution, fully described by its m ean and standard
deviation. These are related to the defining param eters as follow s
and (51)
H ence, for sector k, the param eters of the related G am m a distribution are given by
and (52)
A8.3 Distribution of Default Events in a Single Sector
W ith the choice of G am m a distribution for the function (x), the expression for the probability generating
function
(53)
given by equation (49), can be directly evaluated. B y substitution, change of variable and definition of the
G am m a integral
(54)
U pon rearrangem ent this becom es, for sector k
(55)
This is the probability generating function of the distribution of default events arising from sector k.
It is possible to identify the distribution of default events underlying this probability generating function.
B y expanding F
k
(z) in its Taylor series
(56)
the follow ing explicit form ula is obtained
(57)
This can be identified as the probability density of the N egative B inom ial distribution.
A
The CREDI TRI SK
+
Model
dx x e dx x f dx x X x P
x
1
) (
1
) ( ) (

2 2

2 2
/
k k k

k k k
/
2

0
) 1 (
) ( ) (
x
z x
k
dx x f e z F

) 1 (
1
) 1 )( (
) (
1 1 ) (
1
) (
) (
1 1
0
1
1
1
0
1
) 1 (
z z
z
dy
e
z
y
dx
x e
e z F
y
y
x
x
z x
k
+

,
_

k
k
k
k
k
k
p
z p
p
z F
k

,
_

1
where
1
1
) (

,
_

+

1
1
) 1 ( ) (
n
n n
k
k
k k
z p
n
n
p z F
k

n
k
k
k
p
n
n
p P
k

,
_

+

1
) 1 ( ) defaults n (

A8.4 Summary
The portfolio has been divided into n sectors w ith annual default rates distributed according to
(58)
The probability generating function for default events from the w hole portfolio is given by
(59)
w here the param eters
k
,
k
and p
k
are given by
; and ; (60)
The default event distribution for each sector is N egative B inom ial. The default event distribution for the w hole
portfolio is not N egative B inom ial in general but is an independent sum of the N egative B inom ial sector
distributions. The corresponding product decom position of the probability generating function is given by
equation (59).
A9 Default Losses with Variable Default Rates
A9.1 Introduction
The probability generating function in equation (59) gives full inform ation about the occurrence of default
events in the portfolio. In order to pass from default events to default losses, this distribution m ust be
com pounded w ith the inform ation about the distribution of exposures. In Section A3.4, w e perform ed this
process conditional on a fixed m ean default rate. W e now generalise this process to incorporate the volatility
of default rates.
A9.2 The Distribution of Default Losses
B y analogy w ith equation (14), w e introduce a second probability generating function G(z), the probability
generating function for losses from the portfolio. Thus let
(61)
be the probability generating function of the distribution of loss am ounts. W e seek a closed form expression
for G (z) and a m eans of efficiently com puting G (z).
As for the distribution of default events, sector independence gives a product decom position of the probability
generating function
(62)
w here G
k
(z) is the loss probability generating function for sector k, 1 k n.
B y analogy w ith equation (18), w e define polynom ials P
k
(z), 1 k n, by
(63)
46
CREDIT FIRST
SUISSE B O STO N
) , (
k k

,
_

n
k
k
k
n
k
k
k
z p
p
z F z F
1 1
1
1
) ( ) (

2 2
/
k k k

k k k
/
2
) 1 /(
k k k
p +

0
L) n losses aggregate ( ) (
n
n
z p z G

n
k
k
z G z G
1
) ( ) (

,
_

,
_

,
_

) (
1
) (
) (
) (
1
) (
) (
) (
1
) (
) (
) (
) (
1
) (
k m
j
k
j
k
j
k
k m
j
k
j
k
j
k m
j
k
j
k
j
k
k
j
k
j
z
z
z P

CREDI TRI S K
+
47
w here the expression for
k
in equation (38) has been used. The P
k
(z) provide the link betw een default events
and losses, because the follow ing relation holds
(64)
This is directly analogous to the form ula G(z)=F(P(z)) obtained in equation (19) for a fixed m ean default rate,
except that there is now one such relation for each sector. In order to see that the relation continues to hold
in the present case, w e expand equation (63) as a sum over individual obligors belonging to sector k. Thus
(65)
B y equation (41) w e have
(66)
The left hand side of equation (66) is the probability generating function of the distribution of losses w here
each obligor A has default rate x
A
. This can be seen by com paring equation (17) - the expressions are the
sam e, except that in equation (17) term s w ith the sam e exposure am ount have been collected.
Just as in equation (53), w hich expresses F
k
(z) as an integral of the Poisson probability generating function
over the space of possible values of the random variable x
k
, a conditional probability argum ent show s that G
k
(z)
is the integral of the left hand side of equation (66) over the sam e space. Thus
(67)
W here the last step follow s from equation (66). B y substitution into equation (55) and taking the product over
each sector, w e obtain
(68)
This is a closed form expression for the probability generating function. In the next section a recurrence relation
for com puting the distribution of losses from this expression is derived.
A10 Loss Distribution with Variable Default Rates
In this section, a recurrence relation, suitable for explicitly calculating the distribution of losses from equation
(68), is presented. The relation is a form of the recurrence relation in Section A4, derived for a w ider class of
distributions.
A
The CREDI TRI SK
+
Model
)) ( ( ) ( z P F z G
k k k

A
A
A
k
A
A
A
A
A
A
k
A
A
z
z
z P

1
) (
( )
( ) 1 ) (
) 1 (
1

+

z P x
z
x
z x z x x
k k
A
A
A
A
k
k
A
A
A
A A
A
A A
e e e e

0
) 1 ) ( (
0
) 1 (
0
0
) (
) ( ) ( ) nL of Loss ( ) (
k
k k
k
A
v
A
A
k
x
k k k
z P x
x
k k k
z x
n
x
k k k k
n
k
dx x f e
dx x f e dx x f x P z z G

,
_

n
k
k m
j
k
j
k
j
k
k
k
n
k
k
k
k
j
z
p
p
z G z G
1
) (
1
) (
) (
1
) (
1
1
) ( ) (

A10.1 General Recurrence Relation

Suppose, in general, a pow er series expansion
(69)
defines a function G(z) w hich satisfies the differential equation
(70)
w here A and B are polynom ials given respectively by
(71)
In other w ords, w e require that the logarithm ic derivative of G (z) be a rational function. Then, the term s
of the pow er series expansion in equation (69) satisfy the follow ing recurrence relation
(72)
To see this, rearrange the differential equation (70) as follow s
(73)
B y differentiating G term by term , this leads to
(74)
For n 0 the term s in z
n
on the left hand and right hand side respectively are
and (75)
Equating these expressions and rearranging w e obtain
(76)
or equivalently
(77)
as required.
A10.2 Application
In equation (68), the probability generating function of losses w as derived in the form
(78)
48
CREDIT FIRST
SUISSE B O STO N

0
) (
n
n
n
z A z G
) (
) ( ) (
) (
1
)) ( log (
z B
z A
dz
z dG
z G
z G
dz
d

s
s
r
r
z b b z B
z a a z A
+ +
+ +
... ) (
... ) (
0
0

,
_

+
) 1 , 1 min(
0
1
) , min(
0 0
1
) (
) 1 (
1
n s
j
j n j
n r
i
i n i n
A j n b A a
n b
A
G z A
dz
dG
z B ) ( ) (

,
_

,
_

,
_

,
_

+
0 0 0
1
0
) 1 (
n
n
n
r
i
i
i
n
n
n
s
j
j
j
z A z a z A n z b

+
+
) , min(
0
1
) 1 (
n s
j
j n j
A j n b

) , min(
0
n r
i
i n i
A a

+
+ +
) , min(
1
1
) , min(
0
1 0
) 1 ( ) 1 (
n s
j
j n j
n r
i
i n i n
A j n b A a A n b

+
+
) 1 , 1 min(
0
1
) , min(
0
1 0
) ( ) 1 (
n s
j
j n j
n r
i
i n i n
A j n b A a A n b

,
_

n
k
k m
j
k
j
k
j
k
k
k
n
k
k
k
k
j
z
p
p
z G z G
1
) (
1
) (
) (
1
) (
1
1
) ( ) (

CREDI TRI S K
+
49
Taking logarithm ic derivatives w ith respect to z, it follow s that
(79)
This expresses G (z)/G (z) as a rational function. Accordingly, after calculation of polynom ials A(z) and B (z)
such that
(80)
the calculation in Section A10.1 is applicable and leads to a recurrence relation for the loss am ount distribution.
N ote that the sum m ation described in equation (80) m ust be perform ed directly by adding the rational
sum m ands. Provided the unit size is chosen so that the exposures
j
and therefore the degrees of num erators
and denom inators of the rational sum m ands are not too large, this com putation can be perform ed quickly.
A11 Convergence of Variable Default Rate Case to Fixed Default Rate Case
Although the C R ED ITR ISK
+
M odel is designed to incorporate the effects of variability in the average rates of
default, there are tw o circum stances in w hich the C R ED ITR ISK
+
M odel behaves as if default rates w ere fixed.
These are w here:

The standard deviation of the m ean default rate for each sector tends to zero.

The num ber of sectors tends to infinity.

In particular, the effect of either a large num ber of sectors or a low standard deviation of default rates on the
portfolio is the sam e; the behaviour in either case is as if default rates w ere fixed. In the section on generalised
sector analysis, this fact w ill be used to facilitate the analysis of specific risk w ithin a portfolio. In this section,
a proof is given of the first convergence fact. The proof of the second convergence fact is sim ilar.
The proof proceeds by show ing that the probability generating function for the C R ED ITR ISK
+
M odel converges
to the form
(81)
w hich is the probability generating function in equations (17) for losses conditional on a fixed m ean default
rate. The C R ED ITR ISK
+
M odel has the follow ing probability generating function for default losses, given at
equation (68)
(82)
w here , , and
A
The CREDI TRI SK
+
Model

n
k
k m
j
k
j
k
j
k
k
k m
j
k
j
k
k k
n
k
k
k
k
j
k
j
z
p
z
p
z G
z G
z G
z G
1
) (
1
) (
) (
) (
1
1
) (
1
) (
) (
1
) (
) (
) (
) (

n
k
k m
j
k
j
k
j
k
k
k m
j
k
j
k
k k
k
j
k
j
z
p
z
p
z B
z A
1
) (
1
) (
) (
) (
1
1
) (
) (
) (
1
) (
) (

,
_

m
j
j
j
j
z
e z G
1
) 1 (
) (

,
_

n
k
k m
j
k
j
k
j
k
k
k
n
k
k
k
k
j
z
p
p
z G z G
1
) (
1
) (
) (
1
) (
1
1
) ( ) (

) (
1
) (
) (
k m
k
k
j
k
j
k

) 1 /(
k k k
p +
k k k
/
2

2 2
/
k k k

W e consider the lim it w here
k
tends to zero. Then

k
0; p
k
=
k
/(1+
k
) 0 and
k
=
k
/
k

k
/p
k
Therefore
(83)
In the lim it
(84)
O n collecting term s in the exponent having com m on values n across different values of k, the sum m ation over
k is elim inated
(85)
as required.
A12 General Sector Analysis
A12.1 Introduction
In the derivation of the C R ED ITR ISK
+
M odel probability generating function for the distribution of losses in
Section A9, it w as assum ed throughout that the portfolio is divided into sectors, each of w hich is a subset of
the set of obligors. This corresponds to a situation in w hich obligors fall into classes, each of w hich is driven
by one factor but all of w hich are m utually independent.
W e now consider a m ore generalised situation in w hich, as before, a relatively sm all num ber of factors explain
the system atic volatility of default rates in the portfolio, but it is not necessarily the case that the default rate
of an individual obligor depends on only one of the factors. In these m ore general circum stances, it is not
possible to describe the portfolio w ith sectors consisting of groupings of the obligors, but the C R ED ITR ISK
+
M odel incorporates this situation in the sam e w ay as before, replacing the concept of a sector w ith that of
a system atic factor.
To understand how to generalise the sector analysis already presented, w e re-exam ine the derivation of the
probability generating function for the C R ED ITR ISK
+
M odel. In equation (68), the probability generating function
w as derived by expressing it as a product over the sectors and then integrating w ith respect to the distribution
of default rates for each sector:
(86)
H ow ever, this expression can also be view ed as a m ultiple integral
(87)
W e regard the integrand as the probability density function of a com pound Poisson distribution for any
given set of values of the m eans x
k
, 1 k n. H ow ever, w e are sim ultaneously uncertain about all these values.
Therefore, the probability density function is then integrated over the space of all possible states represented
by the values of the x
k
and w eighted by their associated probability density functions.
50
CREDIT FIRST
SUISSE B O STO N

,
_

,
_

n
k
p
k m
j
k
j
k
j
k
k
k
n
k
k m
j
k
j
k
j
k
k
k
n
k
k
k
k
k
j
k
k
j
z
p
p
z
p
p
z G z G
1
) (
1
) (
) (
1
) (
1
) (
) (
1
) ( ) (
1
1
1
1
) ( ) (

k j
k
j
k
j
k
j
k j
k
j
k
j
k m
j
k
j
k
j
k
j
k
z
n
k
z
e e e z G
,
) (
) (
) (
,
) (
) (
) (
1
) (
) (
) (
1
) (

j
j
j
j
z
e z G
) 1 (
) (

n
k
x
k k k
z P x
n
k
k
dx x f e z G z G
k k
1
0
) 1 ) ( (
1
) ( ) ( ) (

0
1
) 1 ) ( (
0
) ( ... ) (
1
1 n
n
k
k k
x
k
n
k
k k
z P x
x
dx x f e z G
CREDI TRI S K
+
51
U sing equation (66), w e can exam ine the exponent in the integrand in its equivalent form
(88)
w here w e have used the delta notation
(89)
To generalise the concept of a sector in these circum stances, allow ing each obligor to be influenced by m ore
than one factor x
k
, w e replace the delta function w ith an allocation of the obligors am ong sectors by choosing,
for each obligor A, num bers
(90)
The allocation
Ak
represents the extent to w hich the default probability of obligor A is affected by the factor
underlying sector k. The sector analysis discussed in Section A7 corresponds to the special case
(91)
In the general case the expression in equation (88) is replaced by
(92)
w here each obligor contributes a term
w here (93)
Equation (65) is replaced by
w here (94)
A12.2 Performing the Sector Decomposition
In this section, w e show how to assim ilate the data for the C R ED ITR ISK
+
M odel for generalised sector analysis.
W e assum e that for each obligor in the portfolio an estim ate has been m ade of the extent to w hich the volatility
of the obligors default rate is explained by the factor k. As explained in Section A12.1, this is expressed by a
choice of num ber
(95)
for each sector k and obligor A in the portfolio. The num ber
Ak
represents our judgem ent of the extent to
w hich the state of sector k influences the fortunes of obligor A.
As in the special case discussed in Section A7, w e m ust also provide estim ates of the m ean and standard
deviation for each sector. W e indicate a m ethod of estim ating these param eters, assum ing again that estim ates
have been obtained of both quantities for each obligor by reference to obligor credit quality.
A
The CREDI TRI SK
+
Model



n
k k A
A
A
k
k
Ak
k A
A
A
k
k
n
k
k k
A A
z
x
z
x
z P x
1 , 1
) 1 ( ) 1 ( ) 1 ) ( (

'

k A
k A
Ak
1
0

n
k
Ak Ak
1
1 :
Ak Ak



k A
A
A
k
k
Ak
n
k
k k
A
z
x
z P x
, 1
) 1 ( ) 1 ) ( (

) 1 (
A
z x
A

n
k
k
k
Ak
A
A
A
x
x
1

A
A
A
Ak
k
k
A
z z P

1
) (

A
A
A
Ak k

n
k
Ak Ak
1
1 :
The m ean for each sector is the sum of contributions from each obligor, but now w eighted by the allocations

Ak
. Thus
(96)
Then, by analogy w ith equation (43), w e express the ratio
k
/
k
as a w eighted average of contributions from
each obligor
;hence (97)
This estim ates the standard deviation for each factor. The discussion in Section A7 is recaptured w hen

Ak
=
Ak
as discussed above.
A12.3 Incorporating Specific Factors
So far w e have assum ed all variability in default rates in the portfolio to be system atic. Potentially, w e require
an additional sector to m odel factors specific to each obligor.
H ow ever, specific factors can be m odelled w ithout recourse to a large num ber of sectors. It w as rem arked in
Section A11 that assigning a zero variance to a sector is equivalent to assum ing that the sector is itself a
portfolio com posed of a large num ber of sub-sectors. H ence, for a portfolio containing a large num ber of
obligors, only one sector is necessary in order to incorporate specific factors. Let the specific factor sector be
sector 1. Then, for each obligor A, the proportion of the variance of the expected default frequency for that
obligor that is explained by specific risk is
A1
. Sector 1 w ould be assigned a total standard deviation given by
equation (97). H ow ever, for the specific factor sector only, this standard deviation can be set to zero.
The specific factor sector then behaves as the lim it of a large num ber of sectors, one for each obligor in the
portfolio, w ith independent variability of their default rate. The lost standard deviation represented by
1
is a
m easure of the benefit of the presence of specific factors in the portfolio.
A13 Risk Contributions and Pairwise Correlation
A13.1 Introduction
In this section, w e derive form ulae for tw o useful m easures connected w ith the default loss distribution, as
follow s:

Risk contributions are defined as the contributions m ade by each obligor to the unexpected loss of the
portfolio, m easured either by a chosen percentile level or the standard deviation.

Pairw ise correlations betw een default events give a m easure of the extent to w hich concentration risk is
present in the portfolio.
A13.2 Risk Contributions
In this section, w e derive a form ula for the contribution of an individual obligor to the standard deviation of the
loss distribution in the C R ED ITR ISK
+
M odel.
For a portfolio of obligors A having exposure E
A
, the risk contribution for obligor A can be defined as the
m arginal effect of the presence of E
A
on the standard deviation of the distribution of credit losses. Alternatively,
the risk contribution can be defined as the m arginal effect of the presence of E
A
on som e other m easure of
portfolio aggregate risk, such as a given loss percentile.
52
CREDIT FIRST
SUISSE B O STO N

A
A Ak k

A
A Ak
A
A
A
A Ak
k
k

A
A Ak k

CREDI TRI S K
+
53
In the first case, an analytic form ula for the risk contribution is possible. The risk contribution can be w ritten
, or equivalently (98)
M oreover, for m ost m odels including the C R ED ITR ISK
+
M odel, the risk contributions defined by equation (98)
add up to the standard deviation. This is because of the variance-covariance form ula
(99)
w here
A
and
B
are the standard deviations of the default event indicator for each obligor. Provided the m odel
is such that the correlation coefficients are independent of the exposures, equation (99) expresses the
variance as a hom ogeneous quadratic polynom ial in the exposures. H ence, by a general property of
hom ogeneous polynom ials w e have
(100)
If the m arginal effect on a given percentile is chosen as the definition of risk contributions, then an analytic
form ula w ill not be possible. Instead, one can use the approxim ation described next.
Let , and X be the expected loss, the standard deviation of losses and the loss at a given percentile level
from the distribution. D efine the m ultiplier to the given percentile as w here
(101)
Then, w e can define risk contributions to the percentile in term s of the contributions to the standard deviation
by w riting
(102)
Then, in view of equations (100) and (101), w e have
(103)
In the analysis below, w e w ill concentrate on the determ ination of the contributions to the standard deviation.
In order to evaluate the right hand side of equation (98), w e derive analytic form ulae for m ean and variance of
the distributions of default events and default losses in the C R ED ITR ISK
+
M odel. W e use the follow ing
definitions, referring to a sector k, w hich are consistent w ith the notation used previously. Since the m ean and
variance of the distribution of losses in the C R ED ITR ISK
+
M odel are both additive across sectors, w e can w ork
w ith one sector for m ost of the analysis. For ease of notation, w e have therefore suppressed the reference to
sector k w here it is not necessary.
Reference Symbol Mean Variance
Loss severity polynom ial (equation 94) P(z)
2
D efault event probability generating function conditional E(z,x)
E

E
on m ean x
2
Probability density function for m ean x f(x)
f

f
2
D efault event probability generating function F(z)
F
=
k

F
2
C R ED ITR ISK
+
M odel probability generating function G(z)
G
=
k

G
A
The CREDI TRI SK
+
Model
A
A A
E
E RC

A
A
A
E
E
RC

2
2

B A
B A B A AB
E E
,
2

2
2
2
1
2 2
A
A
A
A
A
E
E RC
X +
A A A
RC C R +

( ) X RC C R
A
A A
A
A
+ +

H ere
k
and
k
are the m ean num ber of default events in sector k and the expected loss from sector k
respectively. W e have
(104)
This is m erely a restatem ent of equation (64). Also, by equation (53)
(105)
For the probability generating functions E, F and G , w e have, by general properties of probability generating
functions
, and (106)
, and
(107)
B y definition of x, w e have
(108)
B ecause E(z, x) is the probability generating function of a Poisson distribution, w e also have
(109)
B y equations (105) and (106), bringing the differentiation by the auxiliary variable z under the integration sign,
w e obtain
(110)
Sim ilarly, using equations (105) and (107)
(111)
H ence
(112)
Equations (110) and (112) are the m ean and variance of the distribution of default events. To provide the link
to the m om ents of the loss distribution, w e use equation (104), w hich yields, by the chain rule
; (113)
H ence
(114)
54
CREDIT FIRST
SUISSE B O STO N
)) ( ( ) ( z P F z G

x
dx x f x z E z F ) ( ) , ( ) (
) 1 ( ) 1 (
2
2
2 2
dz
dG
dz
G d
G G
+ +
) 1 (
dz
dE
E
) 1 (
dz
dF
F
) 1 (
dz
dG
G

) 1 ( ) 1 (
2
2
2 2
dz
dE
dz
E d
E E
+ + ) 1 ( ) 1 (
2
2
2 2
dz
dF
dz
F d
F F
+ +
x x
E
) (
E E

2
f
x x
E F
dx x xf dx x f x

) ( ) ( ) (
( ) ( )
2 2 2 2 2 2 2
) ( ) (
f f f
x
E E
x
E E F F
dx x f x d x f + + + + +

2 2
f f F
+
dz
dP
z P
dz
dF
z
dz
dG
)) ( ( ) (
2
2
2
2
2
2
2
) (
)) ( ( )) ( ( ) (
dz
z P d
z P
dz
dF
dz
dP
z P
dz
F d
z
dz
G d
+

,
_

2
2
2
2
2
2
2
) 1 ( )). 1 ( ( ) 1 ( )). 1 ( ( ) 1 ( )) 1 ( ( ) 1 ( )) 1 ( (

,
_

+ +
dz
dP
P
dz
dF
dz
dP
P
dz
dF
dz
P d
P
dz
dF
dz
dP
P
dz
F d
G

CREDI TRI S K
+
55
Successive differentiation of equation (94) yields
; and (115)
O n substituting equations (112) and (115) into equation (114), w e obtain
(116)
Substituting for
k
, w e obtain
(117)
Finally, sum m ing over sectors gives the standard deviation of the C R ED ITR ISK
+
M odel Loss D istribution for the
w hole portfolio
(118)
N ote that this is the standard deviation of the actual distribution of losses. As in the earlier sections, the

k
denote the standard deviations of the factors driving the default rates in each sector.
Risk contributions can now be derived directly by differentiating equation (118). Thus, by equation (98)
(119)
H ence
(120)
w here w e have interchanged E
A
and
A
for notational convenience. H ence
(121)
This is the required form ula for risk contributions to the standard deviation. As rem arked above, it can be show n
explicitly that the risk contributions add up to the standard deviation of the portfolio loss distribution. Thus, from
equation (121),
(122)
H ence, using equation (118)
(123)
as required.
A
The CREDI TRI SK
+
Model
k
k
A
A Ak
k
dz
dP

1
) 1 (


A
A A Ak
k
dz
P d
) 1 (
1
) 1 (
2
2

1 ) 1 ( P
( )
2 2
2 2 2
1

,
_

,
_

+

A
A Ak
A
A A Ak
A
A Ak
k
k k F G

( )

+

,
_

,
_

+
A
A A Ak
k
k
k k
A
A A Ak
k
k
k k G

2
2 2
2
2 2 2

+

,
_

A
A A
n
k
k
k
k

1
2
2 2
A
A
A
A A
E
E
E
E RC

2
2

,
_

,
_

,
_

,
_

k
A Ak k
k
k
A A
A
k
k
k
k
B
B B
A
A
A
E
E
E
E
RC

2 2
2 2
2
2
2

,
_

,
_

+

k
Ak k
k
k
A
A A
A
E
E
RC

,
_

,
_

,
_

+
A k
k Ak
A A
k
k
A
A A
A k
k Ak
k
k
A
A A
A
A
E E
E
E
RC

2
2
2

,
_

,
_

,
_

+

2
2
2 2
1
k
k
k
k
A
A A
k A
A A
Ak k
k
k
A
A A
A
A
E
RC
A13.3 Pairwise Correlation
In this section, w e derive a form ula for the pairw ise correlation betw een default events in the C R ED ITR ISK
+
M odel.
To define carefully the pairw ise correlation over a tim e period t, w e associate to each obligor its indicator
function I
A
, w hich is the random variable having the values
(124)
Then, the correlation betw een default of tw o obligors A and B in the tim e period t is defined by
(125)
That is, the statistical correlation betw een the indicator functions of A and B in the tim e period. If the expected
values of I
A
, I
B
and the product I
AB
are
A
,
B
and
AB
, respectively, then
A
,
B
and
AB
are, respectively, the
expected num ber of defaults of A, B and of both obligors in the tim e period. Then, because the indicator functions
can only take on the values 0 or 1, the standard expression for correlation reduces to the follow ing form :
(126)
W e seek an expression for the right hand side of equation (126) in the context of the C R ED ITR ISK
+
M odel.
W e take tw o distinct obligors A and B and m ake the follow ing definitions, w here the general sector decom position
is used w ith n sectors.
Reference Obligor A Obligor B
Tim e period t
Instantaneous default probability PA PB
Expected num ber of defaults
A
= 1 - e
-p
A
t
p
A
t
B
= 1 - e
-p
B
t
p
B
t
Sector decom position
AK
; 1 k n
BK
; 1 k n
The unknow n term in equation (126) is the expected joint default expectation
AB
. Since A and B are distinct,
for any realised values of the sector m eans x
k
, 1 k n the events of default are independent, w e have, w riting
x
A
and x
B
as in equation (93)
(127)
w here, as show n in the table, w e have approxim ated the integrand, ignoring higher pow ers of the default
expectations and using the follow ing approxim ation
6
(128)
In view of the sector decom position, w e have, by equation (93)
and (129)
For convenience, define coefficients
kk
by w riting
(130)
56
CREDIT FIRST
SUISSE B O STO N

'

otherwise 0
period time the in defaults A obligor if 1
A
I
) , (
B A AB
I I
2 1 2 2 1 2
) ( ) (
B B A A
B A AB
AB

1
1
) ( ...
x x
n
k
k k k B A AB
n
dx x f x x
B A
t x t x
x x e e
B A


) 1 )( 1 (
6
The integration in (127) can
be performed without using
the approximation (128) to
give an exact value for the
correlation. It is interesting
to note that, while the exact
integration, which has a
similar form to equation (54),
depends on knowledge of
the probability generating
function f, only the mean
and standard deviation of
f are required to estimate
the approximate correlation
given by equation (138).
In this sense the choice
of gamma distribution for
the variability of the mean
default rate is irrelevant to
correlation considerations.

n
k
A Ak
k
k
A
x
x
1

n
k
B Bk
k
k
B
x
x
1

B A
k k
k B Ak
k k

CREDI TRI S K
+
57
Then
(131)
and w e deduce that
(132)
H ence
(133)
or
(134)
H ow ever
(135)
Thus
(136)
Substituting for
kk
, w e obtain
(137)
This sim plifies to
(138)
Equation (138) is the form ula for default event correlation betw een distinct obligors A and B in the
C R ED ITR ISK
+
M odel. Equation (138) is valid w herever the likely probabilities of default over the tim e period in
question are sm all, taking into account their standard deviation. It is not universally valid. N ote, in particular, that
the value of
AB
given by the form ula can be m ore than one if too large values of the m eans and standard
deviations are chosen. This corresponds to the approxim ation used at equation (128) - the left-hand side is
clearly alw ays less than unity w hile the approxim ating function is unbounded.
W e note tw o salient features of equation (138):

If the obligors A and B have no sector in com m on then the correlation betw een them w ill be zero. This is
because no system atic factor affects them both.

If it is accepted that, as suggested by historical data, the ratios

k
/
k
are of the order of unity, then
depending on the sector decom position the correlation has the sam e order as the term (
A

B
) in the
equation. This is the geom etric m ean of the tw o default probabilities. Therefore, in general one w ould
expect default correlations to typically be of the sam e order of m agnitude as default probabilities
them selves.
A
The CREDI TRI SK
+
Model

1
1 ,
) ( ...
x x
n
k
k k k
k k
k k k k AB
n
dx x f x x

n
k
k j x
n
k j j
j j j
x
k k k k kk
k k j x
n
k k j j
j j j
k k
x x
k k k k k k k k k k AB
j k
j k k
dx x f dx x f x
dx x f dx dx x f x f x x
1
;
; 1
2
, :
, ; 1
,
) ( ) (
) ( ) ( ) (

( )

+ +
n
k
k k kk
n
k k k k
k k k k AB
1
2 2
1 ,
'

( )

+ +
n
k
k k kk
n
k k k k
k k k k AB
1
2 2
1 ,
'

B A
n
k
k
k
B Bk
n
k
k
k
A Ak
n
k k
k k k k

,
_

,
_

1 1 1 ,
'

+
n
k
k kk B A AB
1
2

( )

,
_

n
k
k
k
B A Bk Ak B A
B B A A
B A AB
AB
1
2
2 1
2 1 2 2 1 2
) ( ) (

( )

,
_

n
k
k
k
Bk Ak B A AB
1
2
2 1

B1 Example Spreadsheet-Based Implementation
The purpose of this appendix is to illustrate the application of the C R ED ITR ISK
+
M odel to an exam ple portfolio
of exposures w ith the use of a spreadsheet-based im plem entation of the m odel.
The im plem entation, consisting of a single spreadsheet together w ith an addin, can be dow nloaded from the
Internet (http://w w w.csfb.com ) to reproduce the results show n in this appendix. The spreadsheet contains
three exam ples of the use of the C R ED ITR ISK
+
M odel. In addition, the spreadsheet can be used on a user-
defined portfolio.
For illustrative purposes, w e have lim ited the exam ple portfolio size to only 25 obligors. H ow ever, the
spreadsheet im plem entation has been designed to allow analysis of portfolios of realistic size. U p to 4,000
individual obligors and up to 8 sectors can be handled by the spreadsheet im plem entation. H ow ever, there is
no lim it, in principle, to the num ber of obligors that can be handled by the C R ED ITR ISK
+
M odel. Increasing the
num ber of obligors has only a lim ited im pact on the processing tim e.
B2 Example Portfolio and Static Data
The three exam ples are based on a portfolio consisting of 25 obligors of varying credit quality and size of
exposure. The exposure am ounts are net of recovery. D etails of this portfolio are given in Table 8 opposite.
58
CREDIT FIRST
SUISSE B O STO N
Illustra
Exam p
Appendix B - I llustrative Example B
CREDI TRI S K
+
59
B
I llustrative Example
a
p
Table 8:
Example portfolio
Table 9:
Example mapping table of
default rate information
Credit
Name Exposure Rating
1 358,475 H
2 1,089,819 H
3 1,799,710 F
4 1,933,116 G
5 2,317,327 G
6 2,410,929 G
7 2,652,184 H
8 2,957,685 G
9 3,137,989 D
10 3,204,044 D
11 4,727,724 A
12 4,830,517 D
13 4,912,097 D
14 4,928,989 H
15 5,042,312 F
16 5,320,364 E
17 5,435,457 D
18 5,517,586 C
19 5,764,596 E
20 5,847,845 C
21 6,466,533 H
22 6,480,322 H
23 7,727,651 B
24 15,410,906 F
25 20,238,895 E
The exam ple uses a credit rating scale to assign default rates and default rate volatilities to each obligor.
A table giving an exam ple m apping from credit ratings to a set of default rates and default rate volatilities is
given. The table is show n as Table 9 below. The credit rating scale and other data in the table are designed for
the purposes of the exam ple only.
Credit Mean Standard
Rating Default rate Deviation
A 1.50% 0.75%
B 1.60% 0.80%
C 3.00% 1.50%
D 5.00% 2.50%
E 7.50% 3.75%
F 10.00% 5.00%
G 15.00% 7.50%
H 30.00% 15.00%
B3 Example use of the Spreadsheet Implementation
Three exam ples are given, each based on the sam e portfolio, as follow s:

All obligors are allocated to a single sector.

Each obligor is allocated to only one sector. In this exam ple, countries are the sectors. This assum es that
each obligor is subject to only one system atic factor, w hich is responsible for all of the uncertainty of the
obligors default rate.

Each obligor is apportioned to a num ber of sectors. Again, countries are the sectors. This reflects the
situation in w hich the fortunes of an obligor are affected by a num ber of system atic factors.
The exam ples are installed on the spreadsheet im plem entation, together w ith the results generated by the
m odel. For each exam ple, the inputs to the m odel have been set to generate the follow ing:

Percentiles of loss.

Full loss distribution.

C redit risk provision.

Risk contributions.
In this section, the steps to reproducing these results using the m odel im plem entation are described.
B3.1 Activating the CREDITRI SK
+
Model
C hoose one of the three exam ple w orksheets to reproduce the results. Each w orksheet is equipped w ith
a m acro button. Press the button to activate the m odel im plem entation.
B3.2 Data Input Screen
O n activation, the m odel w ill show the D ata Input Screen. This screen is used to set the w orksheet ranges of
data to be read in to the m odel and to specify the form of output data required. The D ata Input Screen has
been preset to the correct ranges corresponding to the layout of each exam ple w orksheet.
60
CREDIT FIRST
SUISSE B O STO N
Activate Model
CREDI TRI S K
+
61
Press the Proceed button on the D ata Entry Screen to proceed to the next step.
B3.3 Input Data Check
The m odel im plem entation has been preset to identify errors in the data read in before the calculation
com m ences. The user is given the option of sw itching off this facility via the D ata Input Screen. The m odel
im plem entation ensures that the data satisfies the follow ing three criteria:

The sector allocation table contains only num eric data.

The decom position of each obligor to the various sectors adds up to 100% .

A sector m ust contain at least one allocation entry.

The Input D ata C heck screen indicates the location of an error in the sector allocation table.
B
I llustrative Example
Data Entry Screen
Press the Proceed button
on the Data Entry Screen
to proceed to the next step.
Credit Suisse First Boston
Example 1A
Data Input Ranges
Obligor Name $ B$ 11:$ B$ 35
Exposure $ C$ 11:$ C$ 35
Mean Default Rate $ E$ 11:$ E$ 35
Standard Deviation $ F$ 11:$ F$ 35
Range of Sectors $ G$ 11:$ G$ 35
Confirm Number of Sectors 1
Optional Settings
Use Sector 1 for Specific Risk
X Check Input Data
Display Options
X Preliminary Statistical Data
X Percentile Losses
Data Output Ranges
Percentiles Output Range $ M$ 11:$ N$ 11
Distribution Output Range $ P$ 11:$ Q$ 11
Risk Contributions Output Range $ I$ 11:$ K$ 11
Percentile for Risk Contributions 99
X Print Percentiles to Worksheet
X Print Distribution to Worksheet
X Print Risk Contributions to Worksheet
Proceed Alter Percentiles
Cancel
Press the Proceed button on the Input D ata C heck Screen to proceed w ith the calculation.
B3.4 Portfolio Loss Distribution Summary Statistics
The m odel im plem entation has been preset to display sum m ary statistics of the portfolio loss distribution.
The user is given the option to sw itch off this facility via the D ata Input Screen.
Press the Proceed button to proceed to the next step. The m odel im plem entation now calculates the full
distribution of losses.
B3.5 Percentile Losses, Loss Distribution and Credit Risk Provision
The m odel im plem entation has been preset to display sum m ary statistics of the portfolio loss distribution.
The user can sw itch off this facility via the D ata Input Screen.
62
CREDIT FIRST
SUISSE B O STO N
Input Data Check Screen
Press the Proceed button
on the Input Data Check
Screen to proceed with
the calculation.
Input Data Check Screen
Press the Proceed button
to proceed to the next step.
The model implementation
now calculates the full
distribution of losses.
Credit Suisse First Boston
No Errors Were Detected In The Input Data
Data Input Error Trapping
Error Type
Sector allocation cells with non-numeric entries
None
Obligors whose sector decomposition does not sum to 100%
None
Sectors having zero expected loss
None
Proceed
Return to Input Screen
Credit Suisse First Boston
Portfolio Loss Distribution Summary Statistics
Portfolio Aggregate Exposure 130,513,072
Portfolio Expected Loss 14,221,863
Portfolio Standard Deviation 12,668,742
Amounts are stated in the input units of currency
Proceed
Return to Input Screen
CREDI TRI S K
+
63
Press the Proceed button to output the loss percentiles, the loss distribution, and the risk contributions to
the w orksheet.
Loss Distribution
A graph of the loss distribution has been set up on each w orksheet using the results generated from the
step above.
Credit Risk Provision
From the sum m ary statistic data above, the Annual C redit Provision (AC P) is given by the Expected Loss, i.e.
14,221,863. If the 99th percentile level is chosen as the determ ining level for the Increm ental C redit Reserve
C ap (IC R C ap), then the IC R C ap is 55,311,503.
B3.6 Risk Contributions
In each exam ple, the m odel im plem entation has been preset to output risk contributions for each obligor in the
exam ple portfolio. The risk contribution calculated by the m odel is defined as the m arginal im pact of the obligor
on a chosen percentile of the loss distribution. The m odel im plem entation has been preset to calculate risk
contributions by reference to the 99th percentile loss. This setting can be altered to a different percentile via
the D ata Entry Screen.
B
I llustrative Example
Input Data Check Screen
Press the Proceed button
to output the loss percentiles,
the loss distribution, and
the risk contributions to the
worksheet.
Credit Loss Distribution
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
0 10, 000, 000 20, 000, 000 30, 000, 000 40, 000, 000 50, 000, 000 60, 000, 000 70, 000, 000
Loss
M
a
r
g
i
n
a
l

P
r
o
b
a
b
i
l
i
t
y
Credit Suisse First Boston
Summary Statistical Data
Aggregate Exposure 130,513,072
Expected Loss 14,221,863
Standard Deviation 12,668,742
Loss Percentiles
50.00 11,089,455
75.00 20,498,062
95.00 38,908,486
97.50 46,152,128
99.00 55,311,503
99.50 62,033,181
99.75 68,612,540
99.90 77,133,478
Proceed
Return to Input Screen
B3.7 Using Risk Contributions For Portfolio Management
Exam ple 1 has been split into tw o exam ples, 1A and 1B , to illustrate the use of risk contributions in portfolio
m anagem ent as follow s:

In exam ple 1A, all 25 obligors are included in the portfolio. Table 10 show s that obligors 24 and 25 have
the largest risk contributions.

In exam ple 1B , obligors 24 and 25 have been rem oved from the portfolio. The other portfolio data is
unchanged.
The risk contribution output from exam ple 1A is repeated in the table below.
Expected Risk
Name Loss Contribution
1 107,543 228,711
2 326,946 764,758
3 179,971 426,743
4 289,967 716,735
5 347,599 896,874
6 361,639 910,914
7 795,655 2,163,988
8 443,653 1,199,910
9 156,899 434,047
10 160,202 437,350
11 70,916 225,356
12 241,526 756,325
13 245,605 794,754
14 1,478,697 4,773,594
15 504,231 1,602,530
16 399,027 1,330,448
17 271,773 892,720
18 165,528 560,564
19 432,345 1,477,654
20 175,435 593,559
21 1,939,960 6,850,969
22 1,944,097 7,110,748
23 123,642 487,938
24 1,541,091 9,056,197
25 1,517,917 10,618,120
The effect on the test portfolio of rem oving obligors 24 and 25 is show n in table 12 below. Rem oving these
obligors in exam ple 1B has tw o effects on the portfolio:

The expected loss of the portfolio has been reduced by 3,059,008 from 14,221,863 to 11,162,856. The
am ount of expected loss rem oved is exactly equal to the expected losses from the tw o rem oved obligors
because expected loss is additive across the portfolio. Thus the AC P provision in respect of the portfolio
can be reduced by 3,059,008.

The 99th percentile loss from the portfolio has declined by 15,364,646 from 55,311,503 to 39,946,857.
This is approxim ately predicted by the total risk contributions of 19,674,317 from the tw o obligors rem oved.
The risk contributions give an estim ate of the effect of rem oving the obligors. Thus, if the 99th percentile
loss is used as the IC R C ap for the portfolio, then the IC R C ap can be reduced by 15,364,646. Furtherm ore,
if the sam e percentile is used as the benchm ark confidence level for determ ining econom ic capital, then the
am ount of econom ic capital required to support the portfolio is reduced by the sam e am ount.
64
CREDIT FIRST
SUISSE B O STO N
Table 10:
Example 1A Risk
Contributions
CREDI TRI S K
+
65
The tables below sum m arise the portfolio m ovem ent and the risk details of the rem oved obligors.
Expected Risk
Name Exposure Loss Contribution
24 15,410,906 1,541,091 9,056,197
25 20,238,895 1,517,917 10,618,120
Total 35,649,801 3,059,008 19,674,317
Absolute %
Example 1A Example 1B Movement Movement
Exposure 130,513,072 94,863,271 (35,649,801) 27.3%
M ean 14,221,863 11,162,856 (3,059,007) 21.5%
99th Percentile 55,311,503 39,946,857 (15,364,646) 27.8%
Although the exam ple incorporates unrealistic levels of default rates, the percentage m ovem ents in Table 12
illustrate a general feature of portfolio risk m anagem ent. The rem oval of the obligors w ith the largest risk
contributions from a portfolio has a greater im pact on the portfolio risk, as m easured by the 99th or percentile
loss, than on the expected loss of the portfolio. Therefore, a significant reduction in the econom ic capital
required to support a portfolio of credit exposures can be achieved by focusing on the m anagem ent of a sm all
num ber of obligors w ith large risk contributions.
Thus in this case, rem oval of tw o obligors, representing 21.5% of the expected loss of the portfolio, has
elim inated 27.8% of the total risk as m easured by the 99th percentile loss.
B3.8 Setting the Percentile Levels
The m odel im plem entation provides a facility to change the percentile loss levels calculated and output by the
m odel. This facility is accessed from the D ata Entry Screen.
B
I llustrative Example
Table 11:
Example 1B - Risk
analysis of removed
obligors
Table 12:
Example 1B - Portfolio
movement analysis
Data Entry Screen
Credit Suisse First Boston
Percentiles must be
chosen as numbers
between 0 and 99.9
OK
Cancel

Percentile Levels Setting

50.0
75.0
95.0
97.5
99.0
99.5
99.75
99.9
Number of points required 8
For m ore inform ation about C R ED ITR ISK
+
, please contact any of the follow ing:
Portfolio Management and Credit Derivatives
London John C hrystal 44-171-888-3235
john.crystal@ csfb.com
M ark Venn 44-171-888-4279
m ark.venn@ csfb.com
New York Jean-Francois D reyfus 1-212-325-5919
jean-francois.dreyfus@ csfb.com
Stephen Lazarus 1-212-325-5911
stephen.lazarus@ csfb.com
66
CREDIT FIRST
SUISSE B O STO N
Conta
Appendix C - Contacts
CREDI TRI S K
+
67
Risk Management
London M ark H olm es 44-171-888-2426
m ark.holm es@ csfb.com
Andrew C ross 44-171-888-3839
andrew.cross@ csfb.com
Tom W ilde 44-171-888-2235
tom .w ilde@ csfb.com
C
Contacts
cts
C
1. C arty, Lea V., Lieberm an, D ana. (January 1997) Historical Default Rates of Corporate Bond Issuers,
1920-1996, M oodys Investor Services, G lobal C redit Research.
2. C arty, Lea V., Lieberm an, D ana. (N ovem ber 1996) Defaulted Bank Loan Recoveries, M oodys Investor
Services, G lobal C redit Research.
3. (April 1997) CreditMetrics. J.P.M organ & C o. Incorporated.
4. D aykin, C .D ., Pentikinen, T., Pesonen, M . (1994) Practical Risk Theory for Actuaries. C hapm an & H all
5. Kealhofer, Stephen. (1995) M anaging D efault Risk in Portfolios of D erivatives. Derivative Credit Risk:
Advances in Measurement and Management. Renaissance Risk Publications
6. M arvin, Sally G . (D ecem ber 1996) Capital Allocation: A Study of Current and Evolving Practices in
Selected Banks. O ffice of the C om ptroller of the C urrency
7. Panjer, H arry H ., W illm ot, G ordon E. (1992) Insurance Risk Models. Society of Actuaries.
8. (February 1997) Ratings Performance 1996 Stability & Transition. Standard & Poors.
9. Theodore, Sam uel S., M adelain, M . (M arch 1997) Modern Credit-Risk Management And The Use of
Credit Derivatives: European Banks Brave New World (And Its Limits). M oodys Investor Services,
G lobal C redit Research.
68
CREDIT FIRST
SUISSE B O STO N
Select
Bibliog
Appendix D - Selected Bibliography D
t
g
C
D
S
C
D
S
C
D
S
C
D
S
C
D
CREDIT FIRST
SUISSE B O STO N
One Cabot Square, London E14 4QJ
Regulated by SFA