Credit Risk Management. Philippe Jorion. 1. VAR. \var\wbank\4credit.ppt. Credit
Risk. Management. Philippe Jorion. University of California at Irvine. July 2004.
Credit Risk Management
Credit Risk Management VAR
Philippe Jorion University of California at Irvine July 2004 © 2004 P.Jorion E-mail:
[email protected]
Please do not reproduce without author’s permission \var\wbank\4credit.ppt
Credit Risk Management: Plan (1) The nature of credit risk (2) Credit risk measurement (3) The Basel capital charges (4) Procyclicality of credit risk charges (4) Conclusions
Philippe Jorion
1
Credit Risk Management
Credit Risk Management (1) The Nature of Credit Risk
Credit Risk: Definition Risk of economic loss from the failure of a counterparty to fulfill its contractual obligations ! Risk can be defined either: !
» as default mode (only if actual default occurs) » as mark-to-market (accounting for intermediate changes in value due to changes in default probability)
Risk Management - Philippe Jorion
Philippe Jorion
2
Credit Risk Management
Drivers of Credit Risk Default event: discrete variable for default binomial event b = 1 (default) or 0 (no default), with probability p = E(b) ! Credit exposure: amount exposed at default EAD = exposure at default ! Loss given default: fraction of exposure lost LGD = 1 – fractional recovery = 1 – f ! Correlations across variables !
Risk Management - Philippe Jorion
Distribution of Credit Losses Define N as the number of counterparties ! Default-mode credit loss is !
CL =
∑
N i =1
bi × E A D i × LG D i
This is a complex random variable, which depends on the distribution of the risk drivers and all correlations thereof ! This also involves market risk for EAD !
Risk Management - Philippe Jorion
Philippe Jorion
3
Credit Risk Management
Credit Risk: Probability of Default !
Traditional approaches: Based on » accounting data (Altman’s Z-score) » historical default rates across credit rating
Reduced-form models: Estimate PD from credit spreads ! Structural models: Estimate PD from equity prices and economic explanation of default process !
Risk Management - Philippe Jorion
Credit Risk: Exposure at Default Bonds, loans, receivables, guarantees: exposure is close to the notional amount ! Performance letters of credit, commitments have lower exposures ! Swaps and forward contracts: initial exposure is zero but then takes on positive value— potential exposure is stochastic ! Long options: exposure is positive ! Short options: exposure is zero !
Risk Management - Philippe Jorion
Philippe Jorion
4
Credit Risk Management
Credit Risk: Loss Given Default Primary determinant is seniority ! Collateral is relevant ! Stochastic and can be correlated with probability of default !
Risk Management - Philippe Jorion
Credit Risk: Correlations Across Defaults Structural model is required because default correlations cannot be measured directly for obligors ! Approaches: (1) Fixed correlations across all industries (2) Correlations measured from equity prices (3) Correlations from country/industry indices and mapping of obligor on indices !
Risk Management - Philippe Jorion
Philippe Jorion
5
Credit Risk Management
Measuring Credit Risk: Example 1 credit of $100 million
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% -$100
N=1, E(Loss)=$1 million, SD=$10 million
-$90
-$80
-$70
-$60
-$50
-$40
-$30
-$20
-$10
$0
10 independent credits of $10 million 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% -$100
N=10, E(Loss)=$1 million, SD=$3 million
-$90
-$80
-$70
-$60
-$50
-$40
-$30
-$20
-$10
$0
Risk Management - Philippe Jorion
cr2a.wmf at 80%
Measuring Credit Risk: Example 100 independent credits of $1 million
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% -$100
N=100, E(Loss)=$1 million, SD=$1 million
-$90
-$80
-$70
-$60
-$50
-$40
-$30
-$20
-$10
$0
1000 independent credits of $100,000 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% -$100
N=1000, E(Loss)=$1 million, SD=$0.3 million
-$90
Risk Management - Philippe Jorion
Philippe Jorion
-$80
-$70
-$60
-$50
-$40
-$30
-$20
-$10
$0 cr2b.wmf at 80%
6
Credit Risk Management
Credit Risk Management (2) Credit Risk Measurement
Credit Risk Models: Taxonomy
Default mode: considers default only
Bottom-up: models each credit individually (e.g. corporate, sovereign) Mark-to-market: considers changes in market values
Conditional model: accounts for changing parameters (PD), function of macro conditions Structural correlation model: explains correlations by joint asset/stock prices
Unconditional models: uses fixed parameters (but can be changed) Reduced-form correlation model: relates default to common macro factors
Top-down: aggregates risk (e.g. retail)
Risk Management - Philippe Jorion
Philippe Jorion
7
Credit Risk Management
Measuring Credit Risk: CreditMetrics Bottom-up, mark-to-market, unconditional, structural model of portfolio credit risk ! Credit risk driver is migration of credit rating, from transition matrix (unconditional model) ! Correlations in credit rating changes inferred from equity market correlations (structural model) ! Bond/loan values based on credit ratings and credit spreads ! Expected exposures from market risk model; no integration of market and credit risk !
Risk Management - Philippe Jorion
Exposures User portfolio
Market volatilities
Expected exposure
Credit VAR Credit rating
Credit spreads
Correlations
Equities Seniority correlations
Bond Recovery Correlation Rating model rate migration valuation
Distribution of values for a single credit
Joint rating changes
Portfolio Value at Risk due to credit Source: CreditMetrics
Philippe Jorion
crm.wmf
8
Credit Risk Management
Distribution of Losses
Frequency
Unexpected loss at 99% level
Expected loss
Loss Risk Management - Philippe Jorion
Credit Loss Distribution Skewed to the left ! Similar to a short position in an option !
» Merton model decomposes a position in a credit -sensitive bond as a long position in a risk-free bond plus a short position in a put on the firm value
Risk Management - Philippe Jorion
Philippe Jorion
9
Credit Risk Management
Option Approach to Corporate Debt Value of the firm
Equity
Face value K of debt
Debt
0
K
Value of the firm Risk Management - Philippe Jorion
merton.wmf at 80%
Option Approach to Corporate Debt Equity
0 Debt
K
0
K Value of the firm
Risk Management - Philippe Jorion
Philippe Jorion
merton2.wmf at 80%
10
Credit Risk Management
Option Distribution Option value
Distribution of option values
Spot price
Distribution of spot prices Risk Management - Philippe Jorion
dist-put.wmf at 80%
Option Distribution
Risk Management - Philippe Jorion
Philippe Jorion
ERM-risk.swf
11
Credit Risk Management
Using the Loss Distribution (1): Expected Losses Pricing of products: bid-ask spreads, or corporate bond yields should be sufficient to cover expected credit losses (plus remuneration of economic capital) ! Provisions: charge-offs against income to account for expected losses !
Risk Management - Philippe Jorion
Expected Credit Loss !
Over the target horizon (say 1 year), this is
ECL = ∫ (b × EAD × LGD)f (b,EAD,LGD)db dEAD dLGD !
When the variables are independent, this is
ECL = [∫ bf (b)db][∫ EADf (EAD)dEAD][∫ LGDf (LGD)dLGD]
ECL = Pr ob[default] × E[EAD] × E[LGD] Risk Management - Philippe Jorion
Philippe Jorion
12
Credit Risk Management
Expected Credit Loss: Example !
Compute the ECL on a BBB-rated $100 million 5-year bond » the expected cumulative default rate is 2.28% » the exposure is assumed constant » the expected recovery rate is 47%
ECL = Pr ob[default] × E[EAD] × (1 − f ) ECL = 2.28%× $100m × (1 − 47%) = $1.2m !
Note that this does not depend on correlations
Risk Management - Philippe Jorion
Using the Loss Distribution (2): Unexpected Losses Covered by equity capital: institution should set aside enough capital to cover worst loss ! Can be lowered by diversification or credit derivatives ! Remuneration of economic capital should also go into the pricing of the product !
Risk Management - Philippe Jorion
Philippe Jorion
13
Credit Risk Management
Comparison of Capital Charges Portfolio A
Portfolio B
Portfolio C
High credit quality, diversified (500)
High credit quality, concentrated (100)
Low credit quality, diversified (500)
Assuming Zero Correlations CreditMetrics
777
2,093
1,989
CreditRisk+
789
2,020
2,074
Internal Model 1
767
1,967
1,907
Internal Model 2
724
1,906
1,756
Basel Rules
5,304
5,304
5,304
CreditMetrics
2,264
2,941
11,436
CreditRisk+
1,638
2,574
10,000
Internal Model 1
1,373
2,366
9,654
Basel Rules
5,304
5,304
5,304
Assessing Correlations
ISDA-Portfolio of $66.3 billion notional, 99%1-year VAR Risk Management - Philippe Jorion
Using the Loss Distribution (3): Marginal Contribution to Risk Individual credits should be evaluated not only on the basis of their stand-alone risk, but also their contribution to portfolio risk ! Economic capital should be allocated to credit positions in relation to their contribution to risk: risk premium ! Banks that use economic capital efficiently, or diversify, can charge lower risk premiums, and be more competitive !
Risk Management - Philippe Jorion
Philippe Jorion
14
Credit Risk Management
Portfolio Credit Risk Models: Issues Ideally, internal credit risk models give best estimate of economic capital, like internal model approach for market risk ! In practice, models need to be validated ! Validation difficult: !
» credit events more rare, probabilities difficult to estimate » validation needs to be performed over many years, across economic cycles » correlation across borrowers is an important driver of credit risk and is difficult to measure Risk Management - Philippe Jorion
Risk Management (3) The Basel Capital Charges
Philippe Jorion
15
Credit Risk Management
Basel Capital Adequacy: Credit Risk Charges !
Standardized Approach: » assign external ratings into new risk weights
!
Foundation Internal Ratings Based Approach: » map internal ratings into probability of default (PD), risk weights; supervisors supply other inputs
!
Advanced Internal Ratings Based Approach: » banks supply internal ratings, loss given default (LGD), exposure at default (EAD), maturity (M) CRC =
∑
i
K i × EAD i = 8%
(∑
i
RWi × EAD i
)
Risk Management - Philippe Jorion
Standardized Approach New Basel Risk Weights: Standardized Approach Credit Rating Claim
AAA/AA-
A+/A-
BBB+/BBB-
BB+/B-
Below B-
Unrated
Sovereign
0%
20%
50%
100%
150%
100%
Banks-option 1
20%
50%
100%
100%
150%
100%
Banks-option 2
20%
50%
50%
100%
150%
50%
Short-term
20%
20%
20%
50%
150%
20%
Claim Corporates
AAA/AA20%
A+/A50%
BBB+/BB100%
Below BB-
Unrated
150%
100%
Notes: • Under option 1, the bank rating is based on the sovereign country in which it is incorporated • Under option 2, the bank rating is based on an external credit assessment • Short-term is below three month Risk Management - Philippe Jorion
Philippe Jorion
16
Credit Risk Management
Basel Committee Goals Maintain the 8% minimum ratio for the banking system as a whole ! Establish risk-sensitive risk charge to minimize regulatory arbitrage ! Provide incentives for further development of good risk management systems ! Calibrate capital charge to unexpected loss over one year horizon at 99.9% confidence level !
Risk Management - Philippe Jorion
Basel Committee Philosophy “[T]he new framework is intended to align regulatory capital requirements more closely with underlying risks, and to provide banks and their supervisors with several options for the assessment of capital adequacy.” -- William McDonough, January 2001
Risk Management - Philippe Jorion
Philippe Jorion
17
Credit Risk Management
Internal Rating Based (IRB) Approach !
The capital multiplier for credit i is
N−1 ( PD) + ρ( PD) ⋅ N−1 ( 0.999) 1+( M−2.5) ⋅b( PD) −PD⋅ LGD]⋅ K =[LGD⋅ N 1−1.5⋅b( PD) 1 ρ PD − ( )
N(.) is the normal cumulative distribution fct., M is the effective maturity, ρ(PD) a correlation function, b(PD) a maturity function ! The total capital charge is CRC =
∑
i
K i × EAD i
Risk Management - Philippe Jorion
IRB Approach Regulatory capital adds up individual charges ! Economic capital charge, however, should reflect the contribution of obligor i to the portfolio risk; generally depends on portfolio ! This is “portfolio-invariant” with a one-factor model and sufficiently diversified portfolio ! The capital charge should be proportional to the probability that the obligor return is less than a cutoff, conditional on common factor m !
Pr ( Ri ≤ Z i | m ) Risk Management - Philippe Jorion
Philippe Jorion
18
Credit Risk Management
IRB Approach: Rationale (1) Obligor i has return Ri on its asset value and defaults if return less than threshold Zi ! Assume Ri∼N(0,1) ! Since Prob(Ri≤Zi)=PDi, we have Zi=N-1(PDi) ! Decompose the return into a market return, with correlation √ρ, and idiosyncratic return !
ρ i ⋅ M + 1 − ρi ⋅ ε i
Ri =
assuming M and ε∼N(0,1) Risk Management - Philippe Jorion
IRB Approach: Rationale (2) Pr ( Ri ≤ Zi | m ) = Pr
(
ρ i ⋅ M + 1 − ρi ⋅ ε i ≤ N −1 ( PDi ) | m
)
N −1 ( PDi ) + ρ i ⋅ (− M ) Pr ( Ri ≤ Zi | m ) = Pr ε i ≤ | m 1 ρ − i
Taking the extreme value such Pr(-M≤m) = 0.999, N −1 ( PDi ) + ρ i ⋅ N −1 (0.999) Pr ( Ri ≤ Zi | m ) = N − 1 ρ i
With an adjustment for expected loss, maturity, and LGD, this is the Basel IRB model Risk Management - Philippe Jorion
Philippe Jorion
19
Credit Risk Management
IRB Approach: Correlation Function Correlation Function
!
0.25
Corporate Mortgages Other Retail
0.20
!
0.15
0.10
! !
0.05
0.00 0%
2%
4%
6%
8% 10% 12% 14% 16% 18% 20% Probability of Default
Parameter ρ measures sensitivity of obligor to business cycle Values around 0.1-0.2 for rated corporates (but imprecisely estimated) Values lower for retail Also, lower values for smaller obligors, proxied by higher PD Risk Management - Philippe Jorion
IRB Approach: Maturity Adjustment Maturity Function for Corporates
!
3.00
PD= 1.30% PD= 0.25% PD=10.00%
2.50
!
2.00
!
1.50
! 1.00
Shorter maturities give banks more control over deteriorating loan Baseline for 1 year Increase function for longer maturities Effect is weaker for lower credits, however; maturity function depends on PD
0.50 0
1
2
3
4
5
6
Maturity (years)
Philippe Jorion
7
8
9
10 Risk Management - Philippe Jorion
20
Credit Risk Management
IRB Capital Requirements 20% 18% 16%
Corporates Mortgages Other Retail
14% 12% 10% 8% 6% 4% 2% 0% 0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
Probability of Default Risk Management - Philippe Jorion
IRB Approach: Diversification Issues Basel capital charges add up credit risk charges for each individual obligor ! Some allowance is made for diversification effects by having different asset classes ! Within each class, “portfolio invariance” is achieved with a single-factor model and perfect diversification; capital has been calibrated to “typical” portfolio correlations ! However, this penalizes banks with greater than average diversification !
Risk Management - Philippe Jorion
Philippe Jorion
21
Credit Risk Management
Risk Management (4) Procyclicality of Credit Risk Charges
Issues Concern is that new capital standards will exacerbate business cycles ! In a downturn, !
» bank capital is eroded due to increased losses, » existing obligors are downgraded; higher PD leads to greater capital charge » raising capital may be more difficult !
This could lead banks to tighten lending activity during a downturn, making it worse
Risk Management - Philippe Jorion
Philippe Jorion
22
Credit Risk Management
Example A bank starts with 100 loans and $8 capital ! 10 loans default, with LGD=45% (loss=$4.5) ! Under Basel I, capital drops from $8 to $3.5 and remaining charge is 90x8%=$7.2: shortfall is 7.2-3.5=$3.7 ! Under Basel II, remaining 90 loans become more risky with charge of 90x10%=$9.0: shortfall is 9.0-3.5=$5.5 ! Capital charge would have increased by 25% =(10-8)/8 !
Risk Management - Philippe Jorion
Numerical Exercise Consider typical bank portfolio over 1998-02 ! Compute new capital charge under IRB ! Using S&P ratings, charge would have increased by 30-45%: ratings are more stable and supposed to be “through the cycle” ! Using KMV's EDFs based on stock prices, charge would have increased by 3-83%: EDFs are “point-in-time” and more responsive !
Source: Kashyap and Stein (2004) Risk Management - Philippe Jorion
Philippe Jorion
23
Credit Risk Management
Solutions Bank regulators can force capital to be higher than Basel minimum: not optimal ! Procyclical effects can be dampened by: (1) smoothing PD, (2) flattening the IRB curve, or (3) changing the aggregate capital charge (e.g. !
using a moving average or a countercyclical index)
!
The latter solution is preferred because it maintains informativeness in reported capital charges, does not invite regulatory arbitrage
Risk Management - Philippe Jorion
Risk Management (5) Conclusions
Philippe Jorion
24
Credit Risk Management
Conclusions (1) Credit risk models are much more complex than market risk models but are still based on Value-at-Risk methodology ! Like VAR models, diversification effects and correlations are essential ! Credit risk models are still difficult to validate and not trusted by bank regulators !
» no backtesting framework » validation needs to economic cycles Risk Management - Philippe Jorion
Conclusions (2) The Basel 2 charges represent a considerable improvement over the Basel 1 rules, introducing risk-sensitive capital charges ! The charges are calibrated to a one-factor model with perfectly diversified remaining risk, so that total capital is still about 8% ! Risk-sensitive credit charges, however, may introduce procyclicality, inducing banks to tighten lending during a recession ! Basel 2 will spread risk management methods !
Risk Management - Philippe Jorion
Philippe Jorion
25
Credit Risk Management
Bibliography !
!
Basel Committee on Banking Supervision, 2004, International Convergence of Capital Measurement and Capital Standards: A Revised Framework Gordy, Michael, 2003, “A Risk-Factor Model Foundation for Ratings-Based Bank Capital Rules,” Journal of Financial Intermediation 12 (July) » mgordy.tripod.com/research.html#pubs
!
Anil K Kashyap and Jeremy C. Stein, 2004, “Cyclical implications of the Basel II capital standards,” FRB Chicago Economic Perspectives 28 (1st quarter) » www.chicagofed.org/publications/economicperspectives/ ep_1qtr2004_part2_kashyap_stein.pdf
References !
!
!
!
Philippe Jorion is Professor of Finance at the Graduate School of Management at the University of California at Irvine Author of “Value at Risk,” published by McGraw-Hill in 1997, which has become an “industry standard,” translated into 7 other languages; revised in 2000 Author of the “Financial Risk Manager Handbook,” published by Wiley and exclusive text for the FRM exam; revised in 2003 Editor of the “Journal of Risk”
Phone: (949) 824-5245 FAX: (949) 824-8469
Philippe Jorion
E-Mail:
[email protected] Web: www.gsm.uci.edu/~jorion
26
Credit Risk Management
Philippe Jorion
27