Basel II Risk Weight Functions

CHAPPUIS HALDER & CIE
Basel II IRB Risk Weight

Functions
Demonstration and Analysis
20/02/2013
By
Benoit Genest & Leonard Brie
Global Research & Analytics1
This work was supported by the Global Research & Analytics Dpt. of Chappuis Halder & Cie.
Research Team. Tel.: +33-(0)1-80 18 26 18; fax: +33-(0)1-80 18 26 20
E-mail :[email protected] (London)
Global Research & Analytics Dpt.| 2013 | All rights reserved
Table of contents
Abstract ................................................................................................................................................... 3
1.
2.
Founding Principles of IRB risk weight functions .......................................................................... 4

1.1.
Basics fundamentals ................................................................................................................ 4
1.2.
From Mertons model to the risk weight functions ................................................................. 7
Risk weight functions limits .......................................................................................................... 12

2.1.
Dependence on PD, LGD ...................................................................................................... 12
2.2.
Supervisory estimates of asset correlations ........................................................................... 14
2.3.
From LGD to downturn LGD................................................................................................ 18
2.4.
The maturity adjustment function ......................................................................................... 20
2.5.
A confidence level at 99.9 %................................................................................................. 24
2.6.
The standard normal distribution in the ASRF framework ................................................... 25
2.7.
Processing portfolios default ................................................................................................. 26
Conclusion ............................................................................................................................................. 28
Bibliography .......................................................................................................................................... 29
Abstract
egulatory capital requirements pose a major challenge for financial

institutions today.
As the Asian financial crisis of 1997 and rapid development of credit
risk management revealed many shortcomings and loopholes in measuring
capital charges under Basel I, Basel II was issued in 2004 with the sole intent of improving
international convergence of capital measurement and capital standards.
This paper introduces Basel II, the construction of risk weight functions and their limits in
two sections:
In the first, basic fundamentals are presented to better understand these prerequisites:
the likelihood of losses, expected and unexpected loss, Value at Risk, and regulatory capital.
Then we discuss the founding principles of the regulatory formula for risk weight functions
and how it works.
The latter section is dedicated to studying the different parameters of risk weight
functions, in order to discuss their limits, modifications and impacts on the regulatory capital
charge coefficient.
Key words: Basel II, EL, UL, LGD, PD, Mertons Model, ASRF, Conditional PD, rho, Maturity,
confidence level, PDF.
JEL Classification: C1, G21
1. Founding Principles of IRB risk weight functions

This section covers the basic fundamentals of loss and economic capital in order to
thoroughly explain the founding principles of the IRB regulatory curves.
1.1. Basics fundamentals

Likelihood of losses and required economic capital
We begin with a walkthrough of the basic components in the diagram above: expected loss,
unexpected loss, Value at Risk and Required Economic Capital (Regulatory Capital).
There is no doubt that it is impossible to forecast the exact amount of potential losses a bank
will suffer in any given year. In fact, banks can only forecast the average level of those credit
losses, which are called expected losses and denoted by (EL). In addition, banks are required
to cover these expected losses (EL) with accounting provisions.
Unexpected losses (UL) are losses above expected levels that banks expect to incur in future,
but cannot predict their timing or severity.
Although high levels of interest are charged on credit exposures, a bank cannot absorb all of
its unexpected losses. Therefore, these losses must be covered by Economic Capital (EC)
which has a loss-absorbing function. After all, economic capital protects a bank the same
way a shield protects a fighter, by defending against unexpected losses.
With regards to the Value at Risk (VaR), its nothing more than the maximum potential loss
with a confidence level of a percentage . This loss would be exceeded by only a very small
probability of 1-.
Formula:
=
= ()
Where:
=
: Probability of Default

: Exposure at Default
: Loss given Default
Regulatory Capital Charge measurement approach
Regulatory capital charge can be measured using the standard or Internal Rating Based (IRB)
approach, however there are major differences between the two with respect to the risk
weights used.
In the standard approach, the capital charge formula is:
= !""!# 8%

The risk weighting coefficient is based on the quality of the loan quantified by external
ratings. However, for some institutions like the BIS, IMF, ECB, EC and MDBs, the risk
weight is always 0%. That means these institutions are considered solvent at all times by
BCBS.
For sovereign banks and corporations, the risk weights are represented in the table below:
Rating
Sovereigns
Banks
Corporates
AAA AA-
0%
20%
20%
A+ A-
20%
50%
50%
BBB+ BBB-
50%
100%
100%
BB+ BB-
100%
100%
100%
B+ B-
100%
100%
150%
Below B-
150%
150%
150%
Unrated
100%
100%
100%
Source: Bank for International Settlements
In the IRB approach, the weighting coefficient is based on internal ratings. Instead of relying
on an outside rating agency, banks are required to estimate their own rating in-house by
effectively using internal rating systems.
In this framework, the regulatory formula for capital charge is not the same as the one in the
standard approach. It is as follows:
! = "#!# &; ; (; )* 8%
(Maturity) and ) (correlation coefficient) and not a fixed coefficient depending only on an
Indeed, the risk weight in the IRB approach is a function depending on PD, LGD, M
external rating.
Furthermore, there are two types of IRB approaches: Foundation IRB and Advanced IRB.
In Foundation IRB:
Banks use their own estimates of PD.
The LGD is fixed at 45% for senior claims on corporates, sovereigns and banks not
secured by recognized collateral.
The LGD is fixed at 75% for all subordinated2 claims on corporates, sovereigns and
banks.
The effective maturity (M) is fixed at 2.5 years for corporate exposures except for
repo-style transactions where the maturity is fixed at 6 months.
In Advanced IRB:
The banks use their own estimates of both PD and LGD.
The effective maturity is not fixed and banks use their own estimates for each
exposure.
Its worth mentioning that banks have to respect minimum guidelines to be authorized to use
an IRB approach.
An IRB approach is better to use than the standard if the regulatory weighting function
calculates a risk weight below the one fixed in the standard approach. As an example, for
corporates, if we found a risk weight below 100%, the IRB approach would allow us to have a
less restrictive capital charge and hence free up resources that can be directed to profitable
investments.
In addition, using an Advanced IRB approach could be more advantageous because of high
LGDs imposed in the foundation of the approach. This could encourage banks to invest in
developing LGD models in order to get regulatory approval.
1.2. From Mertons model to the risk weight functions

The regulatory formula of risk weight functions has been obtained on the basis of simple
assumptions from Mertons model (also known as "Model of the Firm").
The figure below is presented to better understand Mertons Model.
According to the BCBS, a subordinated loan is a facility that is expressly subordinated to another facility.
Mertons Model postulates that the counterparty defaults when the value of its assets is less
than that of its debts. To be more precise, its a firms distance to default, which calculates the
difference between the value of assets and debts of a given firm. The larger the distance to
default, the more solvent the firm is.
The Basel Committee has adopted this principle, adding significant assumptions such as the
infinite granularity of considered portfolios, which means the contribution of an individual
exposure to the portfolio is insignificant and that correlation between assets is not taken into
account. Also, a time horizon of 1 year is assumed by the Basel committee. As a result of
these modifications, the regulatory formula became known as the Asymptotic Single Risk
Factor (ASRF) model.
The ASRF model is used to define the asset value. This model postulates that the value of an
asset depends on two factors:
The systematic factor which models the global environment, generally the overall state
of the economy
Idiosyncratic factor which is specific to the asset
Formally, the ASRF model could be presented as follows:

+ = ,)+ - . ,(1 0 )+ )1+
With:
+ : Value of the asset
)+ : Correlation coefficient
-: Systematic risk factor
1+ : Idiosyncratic or specific risk factor
Where - and 1+ are mutually independent standard normal variables. Consequently + follow
a standard normal distribution.
We define the default variable as a binomial variable 3+ with the following distribution
3+ = 4
1 55 +
0 55 1 0 +
Which means that an asset i default with the probability + .
In the framework of Mertons model, It is assumed that an asset i default if its value goes
below a critical threshold 7+ .
Formally we can write:
From which we deduce:
3+ = 1 " + 7+ (1)
(3+ = 1) = ( 7 )
Knowing the 3+ distribution this is equivalent to:

Since ~ ;(0,1) we have:
+ = ( + 7+ )
+ = (7+ )
Where (. ) is the cumulative normal distribution, therefore:

7+ = ?@ (+ )
Thus from formula (1) above we deduce:
3+ = 1 " + ?@ (+ )
As a result, the critical threshold is nothing else than a function of the default probability of an
asset.
9
The value of an asset + depends on the state of the general economy -. Thus we evaluate the
probability of default conditionally on the realization of the systematic factor y. This can be
interpreted as assuming various scenarios for the economy, determining the probability of
default under each scenario, and then weighting each scenario by its likelihood. The
conditional probability of default is
(3+ = 1|- = ) = ( + ?@ (+ )|- = )
(3+ = 1|- = ) = B,)+ - . ,(1 0 )+ )1+ ?@ (+ )C- = D

(3+ = 1|- = ) = E1+
?@ (+ ) 0 ,)+
,(1 0 )+ )
Finally, knowing that the idiosyncratic risk factor follows a normal standard distribution we
get the following formula:
(3+ = 1|- = ) = E
?@ (+ ) 0 ,)+
,(1 0 )+ )
Consequently, we can affirm that Conditional PD formula is nothing else than a function of
PD and the state of the economy y.
We can describe the Loss of a portfolio as follows
= G + + 3+
systematic risk factor. With these two assumptions fulfilled the -quantile of the loss HI () is
The ASRF framework assumes an infinitely granular portfolio and the existence of only one
almost surely equal:
?@ (+ ) . ,)+ ?@ ()

HI () = G + + E
F
,(1 )+
Formally, we have the worst case scenario when the systematic factor takes the worst
magnitude. Y is a Standard normal variable, so this magnitude is given by ?@ with

some confidence level .
Under this worst-case scenario we have the most serious loss and the capital requirement is
then given by:
10
HK# (, , ), , ) = L 0 M!N

?@ (+ ) . ,)+ ?@ ()
I () = HI () 0 O() = G + + E E
F + F
,(1 )+
P = + E E
?@ + + ,)+ ?@

F + F
,1 )+
If we insert the confidence level = 99,9% we find the regulatory formula for capital charge
coefficient with one year maturity. And risk weight function is directly proportional to the
capital ratio.
11
2. Risk weight functions limits

Now that we explained how risk weight functions are constructed, we will continue to the
core part of our article, which introduces the different limits appearing in the risk weight
functions, their calibrations and impacts on the economic capital required.
In the following analysis, when not specified, we use the BCBS assumptions of the
foundation IRB approach, namely, LGD = 45% and = 99, 90%.
2.1. Dependence on PD, LGD

Since the regulatory capital charge formula is a function of PD and LGD, one cant help but
wonder to what extent does the capital charge depend on these significant input parameters?
In order to respond to this essential question, we analyze the capital charge coefficient
assuming no maturity adjustments function. The maturity adjustments function will be
discussed further. First, as function of PD with a fixed LGD at 45%.
Source: Global Research & Analytics analysis
In general, the capital charge coefficient increases with PD until the inflexion point, which
then causes the capital charge to decrease. This could be explained by the fact that once that
inflexion point is reached, losses are no longer absorbed by UL but by EL, thus lowering the
economic capital required.
12
As a function of LGD, the capital charge coefficient is directly proportional to it. The figure
below shows it with a fixed PD of 0.1%.
Finally, we study the impact of both LGD and PD on capital ratio.
Considering our previous points, the regulatory capital charge evolves as expected. Increasing
with LGD and the presence of an inflexion point for PD. Practical cases are subject to low
13
PD, so the linear dependence on LGD will principally impact capital ratio strongly than PD.
Its shown by the dark blue areas in the figure below.
2.2. Supervisory estimates of asset correlations
In the formula, represents the correlation coefficient. The correlations can be described as
the dependence of the asset value of a borrower on the general state of the economy. Different
asset classes show different degrees of dependency on the overall economy, so its necessary
to adapt the correlation coefficient to these classes.
The asset correlation function is as follows:

1 0 ?STUV
1 0 ?STUV
) = 0.12 E
F
.
0.24
X1
0
E
FY
1 0 ?ST
1 0 ?ST
The asset correlation function is built on two limit correlations of 12% and 24% for very high
and low PDs (100% and 0%, respectively).
14
Source: Bank for International Settlements
However, this formula is incomplete. For Corporate exposures, it has been shown that asset
correlations increase with firm size3. The higher the size of a firm, the higher is its correlation
with the state of the economy. So for SME its necessary to adjust the asset correlation
function to highlight their idiosyncratic risk factor. The asset correlation function is then:
1 0 ?STUV
1 0 ?STUV
) = 0.12 E
F
.
0.24
X1
0
E
FY . 7Z N[K# "#!#
1 0 ?ST
1 0 ?ST
With:
0 " 7 ] 50
705
7Z N[K# "#!# = \0.04 _1 0
` " 7 b5,50c
45
00.04 " 7 d 5
Where S is the firm size.
This size adjustment concerns only borrowers with annual sales below 50 MN.
The following figure shows the regulatory Capital Charge for different firm sizes without
maturity adjustment.
The size here is measured by annual sales.
15
This figure shows that at low PD, capital charge coefficient increases with it. Furthermore, we
observe that at same PD for different firm sizes the capital charge ratio is correctly
distributed. Formally, a small firm will have a lesser or equal capital ratio than a larger one.
At this point, there is a belief amongst specialists and some central bankers that the Basel
methodology unfairly penalises SMEs. The discussion revolves around rules set by the Basel
Committee in 2005 concerning correlations.
These rules, unchanged since then, cause an allocation of capital to SMEs that is much higher
than the actual loss experience incurred by the SME sector historically and up to the present
day. The work carried out by Mr. Dietsch and J. Petey in France is presented hereinafter. It
shows the real discrepancy between regulatory correlations and those actually observed on
conventional SME portfolio:
16
Source : Mesure et gestion du risque de crdit dans les institutions financires M. Dietsch & J. Petey
In the chart below, we show the risk weighted asset calculation for SMEs based on the Basel
rules, and then on actual experience in the sector. The result is a significantly reduced
allocation of capital if actual experience is used instead of the existing rules.
Source : Global Research & Analytics analysis (based on a real French SME portfolio)
Finally, the presence of bounds can be explained by a lack of historical data for large and
small firm, thus calibration is harder to quantify. Hence these bounds represent the global
dependency of small and large firm on the overall economy.
Another asset correlation function has to be built for retail exposures. Indeed retail portfolio is
subject to low correlation because the default of retail customers tends to be more
idiosyncratic and less dependent on the economic cycle than corporate default.
17
N# (: ) 0.15
e"# ff# M: ) 0.04

g M: ) 0.03 E
1 0 ?iSUV
1 0 ?iSUV
F
.
0.16
X1
0
E
FY
1 0 ?iS
1 0 ?iS
This calibration of the asset correlation function has been made over the analysis of historical
banks economic capital data. This new function reflects the relatively high and constant
correlation for residential mortgage exposures, the relatively low and constant correlation for
revolving retail exposures, and, similarly to corporate borrowers, a PD-dependent correlation
in the other retail case.
Finally, the impact of correlation on the capital ratio will be significant for high PD exposure.
2.3. From LGD to downturn LGD

Since we apply the ASRF model, parameters have to be conditioned by the overall state of the
economy.
Though LGD plays a great role in the capital requirement function, Basel II does not provide
an explicit function that transforms average LGDs into conditional LGDs (dependent on the
systematic factor).
18
According to the BCBS, there could be two approaches to derive downturn LGDs:
Using a mapping function like the one used for PDs that would derive downturn LGDs
from average LGDs.
Based on the internal valuations of LGDs during the hostile circumstances to compute
downturn LGDs.
Since the bank practices concerning the LGD quantification are so sophisticated, the BCBS
was forced to affirm that it is unsuitable to use a single supervisory LGD mapping function.
In addition, the BCBS decided that Advanced IRB banks have to estimate their own
downturn LGDs with the obligation that those downturn LGDs during the recession
circumstances must be superior to LGDs during normal circumstances.
In the previous regulatory formula, only the average LGD was used. Capital ratio was directly
proportional to it.
klmn+opq rnmlrsm #N 0
Finally, the downturn LGD was introduced in the regulatory formula substituting the
previous average LGD, in order to keep the proportionality.
tpllmuv woxuvplu #N 0
Knowing the downturn LGD is superior to the average version, we can claim that the new
regulatory formula is conservative. However an intuitive approach could use both LGDs in
the formula. In order to face the downturn LGD to the conditional PD and the average LGD
to the average PD:
+uvp+v+nm woxuvplu #N 0 rnmlrsm
Furthermore, the current version is beneficial to the bank since the current capital ratio is less
than its intuitive version. Its shown below.
tpllmuv 0
+uvp+v+nm woxuvplu #N 0 0 y woxuvplu #N 0 rnmlrsm z
tpllmuv 0
+uvp+v+nm y rnmlrsm 0 woxuvplu z d 0
19
Because
So
rnmlrsm d woxuvplu
tpllmuv d
+uvp+v+nm
The figure below shows our different cases, with an average LGD at 35% and a downturn
LGD at 45%.
At low PD, we notice a similar comportment between the current formula and the intuitive
one. Moreover the current formula is easier to compute since it only depends on downturn
LGD which could explain why this version was chosen by the Basel committee.
2.4. The maturity adjustment function
For two counterparties with the same degree of solvency, credit risk is not necessarily the
same as everything here depends on the loan maturity. With the same degree of solvency, the
higher is the loan maturity, the higher is the credit risk.
Since this maturity affects credit risk, it was necessary for the Basel committee to take it into
account and adjust the capital charge formula.
20
According to the BCBS, there are 3 ways to consider the maturity adjustments:
The first one is considering the maturity adjustments as a consequence of increasing credit
risk. The fact that credit risk increases with increasing maturity has to imply more capital
charge (loan maturity effect).
The second one is considering the maturity adjustment as an anticipation of additional capital
charge due to downgrades. Indeed, downgrades are more likely when it comes to long-term
credits and hence the anticipated capital charge will be higher than for short-term credit
(potential downgrades effect).
The third one is considering the maturity adjustments like a consequence of mark-to-market
(MtM) valuation of credits. The expected losses are taken into account by investors when
pricing the fair value of loans. Those expected losses are increasing with ascending PDs and
decreasing with descending PDs. Therefore, the market value of loans decreases with
increasing PDs and increases with decreasing PDs (marked to market losses effect).
For these reasons the BCBS integrated a maturity adjustments function to the basic regulatory
formula to take into account maturity effects: loan maturity, potential downgrades and marked
to market losses.
The maturity adjustments function is a function of both maturity and PD, and they are higher
(in relative terms) for low PD than for high PD borrowers. The BCBS assumed a standard
maturity of 2.5 years, so they constructed the function around this standard maturity. Its
presented as follows:
(
1 . ( 0 2.55
1 0 1.55
L KN K N[K#:

5 0.11852 0 0.05478 log
The figure below shows the maturity adjustment as a function of PD and maturity, reflecting
the potential downgrade effect and the loan maturity effect.
21
Its worth mentioning for loan maturity less than one year, the adjustment increases with
increasing PD but is still lower than 1. Consequently, the capital charge is lower than the case
without maturity adjustment. For maturity longer than one year, the adjustment decreases with
increasing PD but is still greater than 1, and consequently the capital ratio is also greater than
the case without maturity adjustment.
Considering the maturities adjustment curves, we can wonder how the capital ratio is
impacted by the maturity adjustment. The figure below shows the capital ratio as a function of
PD and maturity thus considering the maturity adjustment.
22
We observe the behaviour of the capital ratio curve remains the same whereas its supposed to
be proportional to the maturity adjustment coefficient. This can be explained because both the
capital and maturity adjustment ratios depend on the PD. Thus the capital ratio curve
maintains its state and is linearly dependent on the maturity only.
23
2.5. A confidence level at 99.9 %

According to the BCBS, the benchmark for setting the confidence level is BBB+ firms that
have a default probability fixed on average at 0.1%, which amounts to say that the probability
of survival is set at 99.9%. In other words, the BCBS is confident on average at 99.9% that a
BBB+ rated bank would survive in the time horizon of 1 year.
Thats why the confidence level is set at 99.9%. It means a bank can have losses that exceed
its tier 1 and tier 2 capital on average once in a thousand years. Moreover it was also chosen
to protect against estimation errors, which might inevitably occur from banks internal PD,
LGD and EAD estimation, as well as other modeling uncertainties.
Even if a bank is below the BBB+ rating, it has to respect the 99.9% confidence level.
Nowadays most of the banks are above the BBB+ rating so currently the fixed confidence
level is permissive, although this wasnt always the case during the implementation of Basel
agreements.
In the future, we can expect an increase in confidence in order to better match the actual
economy and its evolution.
However, a decrease in confidence could be caused by a critical and massive downgrading of
the banking sector.
The figure below shows capital ratio measurement for different confidence level.
24
Formally, economic capital increases with increasing confidence levels. The higher the
confidence level; the higher the economic capital is and vice-versa.
2.6. The standard normal distribution in the ASRF framework

The statistical properties of the Normal Law allow the Basel Committee to provide a closed
formula for the regulatory RW calculation. Such properties remain a strong hypothesis of the
ASRF model (as it exists on the market risk approach). These are reminded hereafter:
Indeed, let A be the value of a borrower assets, described by the process:
N+ + N . + NM
Where asset value at T can be represented as:

1
log + log + . 0 .
2
And X is a standard normal variable.
We explained in part one that an asset i defaults if its value goes below a critical threshold 7+ .
Such as
+ + 8 + d 7+ 7+
25
Where
1
log + 0 log + 0 .
2
7+
. is a cumulative normal distribution function.
The variable X is standard normal, and can therefore be represented as

,)+ - . ,1 0 )+ 1+ +
Where Y and i are mutually independent standard normal variables, and are the risk factor
presented in part one.
Hence it explain the standard normal distribution used in the ASRF framework.
2.7. Processing portfolios default

Once default is observed, loss distribution is bimodal. In the cases that appear, almost all or
nothing is recovered. The figure below shows the loss distribution in case of default.
For defaulted portfolios PD is equal to 100%, and as a consequence, risk weight functions are
not measurable. Hence expected loss and unexpected loss have to be recalculated together
with LGD on defaulted exposures.
26
Expected losses for defaulted portfolios depend only on expected recuperations on defaulted
exposures. This best estimation of expected loss is called BEEL. Specific provisions are here
to cover the BEEL.
While default LGD is more severe than normal LGD, its supposed to estimate a downturn
effect.
The measurement of unexpected losses relies on the difference between loss recovery and
their best estimation. Its estimated as follows:
maxy0; wmrpv 0 z

If wmrpv : there is no load in economic capital required (e.g. best

estimation of expected loss is conservative)
Else more economic capital is required to cover these unexpected losses (e.g. best
estimation of expected loss is insufficient)
As a conclusion, in order to process defaulted portfolios, banks have to estimate their own
BEEL, default LGD and specific provisions. Methodologies differ from banking
establishment, thus the various ways to estimate these parameters will have different impacts
on the solvency ratio.
27
Conclusion
After reviewing the funding principles behind Basel II and explaining the construction of risk
weight functions, we exhibited input parameters of the function and detailed their limits,
calibrations and impacts over the capital charge coefficient.
The capital ratio is mainly dependent on PD and LGD; when both of them vary at low levels,
the capital ratio is stable. High variations will occur when there are higher levels of these
parameters, especially LGD since the ratio is directly proportional to it.
Others parameters behind the Basel regulatory formula can explain a large number of
adjustments made by financial institutions in the implementation of Basel II.
In this sense, the assumptions and shortcuts decided in the regulatory RWA calibration need
to be known and understood. In return, this would help banks refine their vision of economic
capital and to better measure their actual risk (defaulted portfolio, financing cost of capital for
SME, impact of systemic risk actually found in their own jurisdiction...).
To conclude, having an in-depth understanding of the theoretical foundations behind the Basel
formula will allow for opportunities to better address weaknesses and to help to uncover
unexpected risks of banks.
28
Bibliography
BCBS, 2004 An explanatory note on the Basel II IRB risk weight functions October
BCBS, 2005 An explanatory note on the Basel II IRB Risk Weight Functions July.
BCBS, 2006 International convergence of capital measurement and capital standards. A
revised framework June
C. Bluhm, L. Overbeck and Wagner C. 2003 the introduction of credit risk modeling
Chapman and Hall/CRC (1st edition) June
G. M. Gupton, C. C. Finger, and M. Bhatia. 1997 CreditMetrics Technical Document.

Morgan Guaranty Trust Co.
A. Hamerle, T. Liebig and Rosch D. Discussion Paper Credit Risk Factor Modeling and the
Basel II IRB Approach Deutsch Bank series 2: Banking and Financial Supervision No
02/2003.
M. Crouhy and al. 2000 A comparative analysis of current credit risk models Journal of
Banking & Finance 24 59-117
P. Crosbie. 1999 Modeling default risk. KMV Corporation,

.
Altman, E. I. (1968). Financial Ratios, Discriminant Analysis and the Prediction of Corporate
Bankruptcy. Journal of Finance 189-209.
Altman, E. I. (2000). Predicting Financial Distress of Companies: Revisiting the z-score and
Zeta models. Retrieved 19 August 2009.
Allen and Powell, Credit risk measurement methodologies Bank for international Settlements.
(2011). Long-term Rating Scales Comparison. Retrieved 1 June 2011.
29
.
Bharath, S. T., & Shumway, T. (2008). Forecasting Default with the Merton Distance-toDefault Model. TheReview of Financial Studies, 21(3), 1339-1369.
Crouhy, M., Galai, D., & Mark, R. (2000). A comparative analysis of current credit risk
models. Journal of
.
D'Vari, R., Yalamanchili, K., & Bai, D. (2003). Application of Quantitative Credit Risk
Models in Fixed Income Portfolio Management. Retrieved 16 August 2009.
Du, Y., & Suo, W. (2007). Assessing Credit Quality from the Equity Market: Can a Structural
Approach Forecast Credit Ratings? Canadian Journal of Administrative Sciences, 24(3), 212228.
Grice, J., & Dugan, M. (2001). The Limitations of Bankruptcy Prediction Models: Some
Cautions for the Researcher. Review of Quantitative Finance and Accounting, 17(2), 151-166.
Gutzeit, G., & Yozzo, J. (2011). Z-Score Performance Amid Great Recession American
Bankruptcy Insitute Journal, 30(2), 44-46.
Huang, M., & Huang, J. (2003). How Much of the Corporate-Treasury Yield Spread is Due to
Credit Risk?Unpublished Manuscripy, Stanford University.
Jarrow, R. A. (2001). Default Parameter Estimation Using Market Prices. Financial Analysts
Journal, 57(5),75.
Jarrow, R. A., Lando, D., & Turnbull, S. (1997). A Markov Model for the Term Structure of
Credit Spreads. Review of Financial Studies, 10, 481-523.
Kealhofer, S., & Bohn, J. R. (1993). Portfolio Management of Default Risk. Retrieved 11 June
2009.
Lechner, A., & Ovaert, T. (2010). Techniques to Account for Leptokurtosis and Assymetric
Behaviour inReturns Distributions. Journal of Risk Finance, 11(5), 464-480. Moody's KMV
Company. (2003). RiskCalc Australia Fact Sheet. Retrieved 11 June 2009.
30
Ohlson, J. A. (1980). Financial Ratios and the Probabilistic Prediction of Bankruptcy. Journal
of Accounting Research, Spring, 109-131.
Pesaran, M. H., Schuermann, T., Treutler, B. J., & Weiner, S. M. (2003). Macroeconomic
Dynamics and Credit Risk: A Global Perspective. Retrieved 1 June 2011.
Sy, W. (2007). A Causal Framework for Credit Default Theory: APRA, wp08-03.
Sy, W. (2008). Credit Risk Models: Why They Failed in the Financial Crisis: APRA, wp0803.
31

Basel II Risk Weight Functions

Uploaded by

Copyright:

Available Formats

Basel II Risk Weight Functions

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basel II Risk Weight Functions

Uploaded by

Copyright:

Available Formats

CHAPPUIS HALDER & CIE

Basel II IRB Risk Weight

Global Research & Analytics1

Global Research & Analytics Dpt.| 2013 | All rights reserved

Founding Principles of IRB risk weight functions .......................................................................... 4

Basics fundamentals ................................................................................................................ 4

From Mertons model to the risk weight functions ................................................................. 7

Risk weight functions limits .......................................................................................................... 12

Dependence on PD, LGD ...................................................................................................... 12

Supervisory estimates of asset correlations ........................................................................... 14

From LGD to downturn LGD................................................................................................ 18

The maturity adjustment function ......................................................................................... 20

A confidence level at 99.9 %................................................................................................. 24

The standard normal distribution in the ASRF framework ................................................... 25

Processing portfolios default ................................................................................................. 26

Global Research & Analytics Dpt.| 2013 | All rights reserved

egulatory capital requirements pose a major challenge for financial

Global Research & Analytics Dpt.| 2013 | All rights reserved

1. Founding Principles of IRB risk weight functions

1.1. Basics fundamentals

Global Research & Analytics Dpt.| 2013 | All rights reserved

: Probability of Default

Regulatory Capital Charge measurement approach

     =      !""!# 8%

Global Research & Analytics Dpt.| 2013 | All rights reserved

Source: Bank for International Settlements

Banks use their own estimates of PD.

Global Research & Analytics Dpt.| 2013 | All rights reserved

The banks use their own estimates of both PD and LGD.

1.2. From Mertons model to the risk weight functions

Global Research & Analytics Dpt.| 2013 | All rights reserved

Idiosyncratic factor which is specific to the asset

Formally, the ASRF model could be presented as follows:

+ : Value of the asset

Global Research & Analytics Dpt.| 2013 | All rights reserved

-: Systematic risk factor

1+ : Idiosyncratic or specific risk factor

Which means that an asset i default with the probability + .

From which we deduce:

Knowing the 3+ distribution this is equivalent to:

Where (. ) is the cumulative normal distribution, therefore:

Thus from formula (1) above we deduce:

Global Research & Analytics Dpt.| 2013 | All rights reserved

(3+ = 1|- = ) = B,)+ - . ,(1 0 )+ )1+ ?@ (+ )C- = D

almost surely equal:

?@ (+ ) . ,)+ ?@ ()

magnitude. Y is a Standard normal variable, so this magnitude is given by ?@  with

Global Research & Analytics Dpt.| 2013 | All rights reserved

   HK# (, , ), , ) = L  0 M!N 

?@ +  + ,)+ ?@ 

Global Research & Analytics Dpt.| 2013 | All rights reserved

2. Risk weight functions limits

2.1. Dependence on PD, LGD

Source: Global Research & Analytics analysis

Global Research & Analytics Dpt.| 2013 | All rights reserved

Source: Global Research & Analytics analysis

Finally, we study the impact of both LGD and PD on capital ratio.

Source: Global Research & Analytics analysis

Global Research & Analytics Dpt.| 2013 | All rights reserved

Source: Global Research & Analytics analysis

: Probability of Default

= !""!# 8%

Which means that an asset i default with the probability + .

(3+ = 1|- = ) = B,)+ - . ,(1 0 )+ )1+ ?@ (+ )C- = D

?@ (+ ) . ,)+ ?@ ()

magnitude. Y is a Standard normal variable, so this magnitude is given by ?@ with

HK# (, , ), , ) = L 0 M!N

?@ + + ,)+ ?@

N# (: ) 0.15

e"# ff# M: ) 0.04

klmn+opq rnmlrsm #N 0

tpllmuv woxuvplu #N 0

+uvp+v+nm woxuvplu #N 0 rnmlrsm

L KN K N[K#:

. is a cumulative normal distribution function.

maxy0; wmrpv 0 z

If wmrpv : there is no load in economic capital required (e.g. best