Overview of Credit Ratings & Credit Rating Models
Philip Delbridge – Vice President Credit Risk Management Credit Suisse
– Provide an overview of credit ratings and credit rating models.
– Cover the following topics:
Definition of credit ratings
How credit ratings are used by banks and other investors
How credit ratings are derived (fundamental analysis versus models)
Role of Credit Rating Agencies
How credit rating models are used in the rating process
– Focus of traditional credit ratings:
– Will ignore structured finance ratings; different rating process applies to structured
products (CDO’s, CLO’s, ABS, RMBS etc.).
Date: 26/01/2012, Slide 2
Definition of Credit Ratings
– S&P defines credit ratings as:
“Credit ratings are forward-looking opinions about credit risk. Standard & Poor’s credit ratings
express the agency’s opinion about the ability and willingness of an issuer, such as a corporation
or state or city government, to meet its financial obligations in full and on time”
– Credit ratings are a measure of an entities default risk.
– Two broad ways credit ratings can be presented as a measure of default risk:
1.Ordinalranking of default risk (letter rating scale), or
2.Cardinal measure of default risk (Probability of Default % or PD)
Date: 26/01/2012, Slide 3
Uses of Credit Ratings
– Banks use credit ratings in 4 broad ways:
1. To calculate Risk Weighted Assets (RWA) for Basel II/III capital requirements.
“The derivation of risk-weighted assets is dependent on estimates of the PD, LGD, and EAD”
Source: Basel II
2. Price Risk – bank pricing models require a credit rating or PD to determine the
appropriate margin on a loan.
3. Assess Collateral – credit ratings are a key driver of haircuts on collateral.
4. Determine Risk Appetite – credit ratings determine the amount of exposure a bank will
incur to a particular borrower.
Date: 26/01/2012, Slide 4
Deriving a Credit Rating
– Traditional credit ratings (corporates, banks, insurance, municipals, sovereigns etc.)
are derived using fundamental analysis covering:
1. Business risk (or direct macroeconomic risk for sovereigns);
2. Financial risk
Source: Standard & Poor’s
Date: 26/01/2012, Slide 5
[slide shown in class]
Date: 26/01/2012, Slide 6
Role of the Credit Rating Agencies (“CRA’s”)
– The CRA business is dominated by S&P, Moody’s and Fitch.
– Rating produced by CRA’s have two broad roles in financial markets:
1.Disseminate information about repayment capacity of borrowers (issuers)
2.Used in private contracts between financial markets participants (banks, investors, regulators
– Rating produced by CRA’s are attractive to fixed income investors:
–CRA’s have access to non-public information
–Ratings may include additional information not available to fixed income investors
– Banks use ratings produced by CRA’s to supplement their own analysis:
–For example banks typically streamline internal due diligence for borrowers rated investment
grade by the CRA’s
Date: 26/01/2012, Slide 7
Measures of Credit Risk
– Letter rating scales (e.g. AAA to D) used by the CRA’s and banks are an ordinal
measure of default risk.
– Letter rating scales rank borrowers from lowest to highest risk.
– Banks, investors, regulators etc. require a cardinal measure of default risk (i.e.
observed PD), which can be obtained one of two ways:
1. Letter ratings can be mapped to a PD using observed default rates
2. Credit rating model calibrated to derive an ex-ante PD
Letter Rating Scale Probability of Default %
AAA 0.000% ≤ PD ≤ 0.022%
AA+ 0.022% ≤ PD ≤ 0.027%
AA 0.027% ≤ PD ≤ 0.035%
AA- 0.035% ≤ PD ≤ 0.045%
A+ 0.045% ≤ PD ≤ 0.058%
A 0.058% ≤ PD ≤ 0.071%
A- 0.071% ≤ PD ≤ 0.099%
BBB+ 0.099% ≤ PD ≤ 0.170%
BBB 0.170% ≤ PD ≤ 0.289%
BBB- 0.289% ≤ PD ≤ 0.501%
BB+ 0.501% ≤ PD ≤ 0.864%
BB 0.864% ≤ PD ≤ 1.477%
BB- 1.477% ≤ PD ≤ 2.528%
B+ 2.528% ≤ PD ≤ 4.332%
B 4.332% ≤ PD ≤ 7.424%
B- 7.424% ≤ PD ≤ 12.722%
CCC+ 12.722% ≤ PD ≤ 21.799%
CCC or below 21.799% ≤ PD ≤ 100%
Date: 26/01/2012, Slide 8
Overview of Credit Rating Models
Empirical Model Practitioner Model
Scoring • Altman Z-Score
Structural • Merton
Date: 26/01/2012, Slide 9
Credit Scoring (Accounting) Models
– Models use financial statement data to estimate a PD or credit rating.
– Altman Z-score is the most well-known credit scoring model.
– Other accounting based models have been developed to derive “shadow” credit
–S&P and Fitch developed models to estimate a credit rating using only historical accounting
– Problems with Credit Scoring/Accounting models include:
–Financial statements are reported at historical cost and don’t incorporate current market values
–Exclude volatility as an explanatory factor of default
Date: 26/01/2012, Slide 10
Market Models – Merton (Moody’s KMV)
– Merton Model is the most common type of Structural/Market model.
– Kealhofer, McQuown and Vasicek developed the “KMV” model, the most widely used
Merton based model today.
– Moody’s purchased the KMV model in 2002.
– Merton Model uses option prices theory to capture the relationship between financial
statement information and market values to derive an ex-ante PD (for KMV a
Expected Default Frequency [EDF]).
– Equity prices, volatility and leverage are the primary drivers of the Merton Model.
– Empirical studies have demonstrated that KMV EDF’s are more accurate and timely
predictors of default when compared to credit ratings of S&P, Moody’s and Fitch.
– There are numerous examples of KMV “predicting” default well before CRA’s.
Date: 26/01/2012, Slide 11
KMV – Enron Example
Enron - KMV EDF v S&P Issuer Rating
Date: 26/01/2012, Slide 12
KMV – Kodak Example
Kodak - KMV EDF v Moody’s Issuer Rating
Date: 26/01/2012, Slide 13
KMV – SLM Corp LBO Announcement
[slide shown in class]
Date: 26/01/2012, Slide 14
KMV – Portfolio Example
[slide shown in class]
Date: 26/01/2012, Slide 15
Market Models – CDS & Bond Implied Rating
– Ratings model have been developed to map CDS and bond spreads to credit rating.
– Models allow a direct comparison with issuer credit ratings.
– Models can isolate issuer specific changes in spreads.
Date: 26/01/2012, Slide 16
Example CDS & Bond Implied Ratings – Ford
Date: 26/01/2012, Slide 17
Example CDS & Bond Implied Ratings – SLM Corp
– KMV implied rating improved from
‘BBB’ to ‘A’ following SLM LBO
– CDS implied rating dropped from
‘BBB’ to ‘B+’
– Bond implied rating also dropped
from ‘A-’ to ‘BB+’
– Over a 6 notch rating difference
between KMV and CDS/Bond
implied ratings follow market
reaction to LBO announcement
Date: 26/01/2012, Slide 18
– Scorecards are rating models which incorporate both quantitative (financial risk) and
qualitative (business risk) factors to derive a credit rating or PD.
– Scorecards are developed using a combination of empirical analysis and “expert
– Scorecards are useful for rating industries where default data is limited:
– Mutual or Pension funds almost never default
– Very few Hedge Funds have defaulted since LTCM due to collateral or orderly liquidations
– Large exchanges (e.g. NYSE, CME) have not defaulted
– Scorecards are more transparent than models and aligned with fundamental analysis.
– Analysts consider models such as KMV a “black box” and it can be difficult to interpret model
Date: 26/01/2012, Slide 19
Scorecard – Corporate Example
Company XYZ Ltd
Qualitative Factor Weight Score
Industry Risk 15.0% 1
Geographic Diversity 10.0% 3
Market Position 10.0% 3
Cost Efficiency 10.0% 4
Quality of Management 5.0% 3
Quantitative Factor Weight Score
Revenues 10.0% $500M 2 Debt / EBITDA
Profit Margin 2.5% 10% 3 < 1x 1
ROE 2.5% 15% 2 1x to 2x 2
Interest Coverage 10.0% 5.0x 1 2x to 3x 3
Total Debt / Assets 10.0% 67% 4 3x to 4x 4
Debt / EBITDA 15.0% 3.5x 4 > 4x 5
Total Score 2.7
Final Rating Total Score
BBB+ 1.75 to 2.00
BBB 2.00 to 2.25
BBB- 2.25 to 2.50
BB+ 2.50 to 2.75
BB 2.75 to 3.00
Date: 26/01/2012, Slide 20
Alternatives Approaches to Credit Ratings
Pure Expert Combined Expert Judgment / Empirical Pure Empirical
Fundamental Scorecards Models
– Uses traditional analysis taking – Relies on statistical modeling and – Calibrated to derive a rating or ex-
into account business and qualitative inputs ante PD
financial risk – Based on the ideal that “numbers – Constructed using large amounts
– Methodologies based on don’t tell the whole story” of data
experience – More transparent than other – Only incorporate financial data
– Rank ordering of credits models
– Ratings are more subjective
– S&P, Moody’s and Fitch ratings
are primarily expert judgment
Date: 26/01/2012, Slide 21
Final Credit Ratings and Credit Decisions
– The approach used by banks to determine the “final” credit rating varies bank to bank.
– Most banks use a combination of expert judgment and models to determine the final
rating used for RWA, pricing loans, risk appetite etc.
– Banks incorporate external ratings of S&P, Moody’s Fitch into the final rating decision.
– Key Takeaway: Investors and lenders should not use models or external ratings
as a substitute for internal analysis; models (accounting, market or scorecards)
and external ratings should be used as compliments to internal analysis.
Final Credit Rating
Date: 26/01/2012, Slide 22
Suggestions for Job Search
– Treat the job search like a class – allocate minimum 10 hours a week.
– Prepare a comprehensive list of target firms.
– Internships are the best way to get your foot in the door.
– Network as much as possible:
– career fairs
– company presentations
– alumni networks
– Must be strong with Excel
Date: 26/01/2012, Slide 23
Date: 26/01/2012, Slide 24