Utilizing Dual PIT/TTC Ratings to Support
IFRS9 Expected Loss Calculations
CRC – University of Edinburgh Business School
Dr. Scott D. Aguais - Managing Director, Aguais & Associates Ltd.
[email protected]
August 4, 2015 – DRAFT – FINAL
Point-in-Time Methodology & IFRS9 Challenges
Agenda
•
Overview – Key Points – PIT IFRS9 Challenges & Historical Background
•
Overview - PIT-TTC Framework & Methodology Required for Developing PIT
Measures
•

Systematic Credit Cycles, PIT/TTC Measures, Model Calibration & Batch Processing

Utilizing PIT Measures to Project ‘ECL’ for IFRS9 for Retail, Commercial & Wholesale
Summary – Integrated IFRS9 & Stress test Architecture & Key Points in the
Presentation
Aguais & Associates Ltd.
CRC - IFRS9
2
Point-in-Time Methodology & IFRS9 Challenges
Key Points in this PIT/TTC Dual Ratings IFRS9 Challenges
1) Systematic Credit Cycles Exist & Can be Measured – Motivates PIT-TTC
Distinctions
2) Evolving Regulatory Agenda – TTC for Basel II – PIT for IFRS9 & Stress Testing
3) IFRS9 Modelling – Commercial, Corporate & Retail Require Different Approaches
4) Substantial Accuracy Implications Motivate PIT – Wholesale ECLs can vary across
the credit-cycle between starting at a peak or trough by a factor of 3-5 times !!!
5) IFRS9 & Stress Testing Require One Integrated Batch Solution & Architecture
Aguais & Associates Ltd.
CRC - IFRS9
3
Risk & Regulatory Drivers Motivate PIT/TTC Ratings
Dual PIT/TTC Ratings Support Multiple Business Goals
1. Regulatory Compliance:
• Support portfolio-wide PIT/TTC methodology development
• Provide integrated & unified IFRS9 & Stress Testing solutions
• E2E support for internal model governance & regulatory approval
2. Know Your Risk:
•
•
•
•
3. Increase Returns:
Process
Data
Models/
Ratings
E2E PIT/TTC implementation
Dual PIT/TTC ratings framework
Advanced Credit Cycle Measures
Integrated PD/LGD/EAD Solution
• Batch Processing: competitive advantage
• Accurate Asset Valuations/Early Warning
• Optimize Model accuracy & Capital Costs
4. Multiple Business Objectives:
•
•
•
•
Capital (TTC)
IFRS9/Provisioning (PIT)
Stress Testing (Stress PIT)
Risk Appetite (TTC)
•
•
•
•
Aguais & Associates Ltd.
Sanctioning (TTC)
Pricing (PIT)
Early Warning (PIT)
CVA (PIT)
CRC - IFRS9
4
Primary PIT/TTC Drivers - Risk Ratings & Regulatory Evolution
Dual Ratings & Credit Benchmarking – Represent New Paradigms in Risk Management
Basel I
Basel II
(TTC)
Stress Testing
(PIT)
Regulatory
Evolution
Single Internal Rating
90s
Credit
Benchmarking
Across Bank’s
Internal Ratings
Internal AIRB
Models Determine
TTC Rating
2002-07
IFRS 9
(PIT)
Full External &
Internal
Benchmarking
Post B2 – ‘RW
Conundrum’
Leads to
‘AERB’ Model
Calibrations
2015-20
2014
Aguais & Associates Ltd.
CRC - IFRS9
5
Systematic Credit Cycles Require PIT/TTC Distinctions
Dual Ratings (PIT & TTC) - Required for IFRS 9 & Stress Testing Compliance/Accuracy
‘Point‐in‐time risk parameters (PDpit and LGDpit) should be
forward looking projections of default rates and loss rates
and capture current trends in the business cycle. In contrast
to through‐the‐cycle parameters they should not be
business cycle neutral. PDpit and LGDpit should be used for
all credit risk related calculations except RWA under both,
the baseline and the adverse scenario. Contrary to
regulatory parameters, they are required for all portfolios,
including STA and F‐IRB.’
EBA – ‘Methodology EU-wide Stress Test 2014’
Version 1.8, 3 March 2014, P 26
Aguais & Associates Ltd.
CRC - IFRS9
6
Regulatory Evolution Also Motivates Benchmarking
Dual Ratings & Benchmarking – Represent New Paradigms in Risk Management
1990s
Use External
Agency
Ratings &
Default to
Develop Corp
PD Models
‘Agency
Replication’
2000s
MarketBased PD
Models
MKMV
EDFs
2010
Credit
Derivative
Markets
Pricing
Risk
Neutral PD
Basel II
Substantial
focus on
collecting &
using Internal
Credit Data for
‘Internal’ Model
Calibration
2015 +
Benchmarking
Initiatives to
collect &
compare model
calibration
AIRB
Regulatory
Benchmarking
FSA HPE
EBA/FRB
Etc
AERB
PECDC
Loan Loss Data
Collection ‘By
Banks for Banks’
Supports LGD
Benchmarking
Aguais & Associates Ltd.
CRC - IFRS9
* Just Published:
Forest, Chawla &
Aguais, ‘Biased
Benchmarks’,
Spring 2015
Journal of Risk
Model Validation
7
Point-in-Time Methodology & IFRS9 Challenges
Agenda
•
Overview – Key Points – PIT IFRS9 Challenges & Historical Background
•
Overview - PIT-TTC Framework & Methodology Required for Developing PIT
Measures
•

Systematic Credit Cycles, PIT/TTC Measures, Model Calibration & Batch Processing

Utilizing PIT Measures to Project ‘ECL’ for IFRS9 for Retail, Commercial & Wholesale
Summary – Integrated IFRS9 & Stress test Architecture & Key Points in the
Presentation
Aguais & Associates Ltd.
CRC - IFRS9
8
Systematic Credit Cycles Require PIT/TTC Distinctions
TTC Ratings for B2 are Conditionally Cycle Neutral – PIT Ratings Reflect ‘True’ Risk
Idiosyncratic Risk
(Specific to the Client)
• Primarily a function of Leverage & Cash Flow Volatility
• Assessed using financials &qualitative assessments
• Changes measured by internal models are typically infrequent
• Credit Conditions deteriorate, client cash flows fall, client PDs rise
• Credit Conditions improve, client cash flows rise, client PDs fall
Systematic Risk
(Specific to the Economy)
• Credit Conditions (assessed using equity data) change frequently
• Monthly changes according to client’s primary region and sector (PIT)
PIT & TTC PD
• Region and Sector values set to their long run averages (TTC)
Aguais & Associates Ltd.
CRC - IFRS9
9
Systematic Wholesale Credit Cycles Can Be Measured
The Existence of Credit Cycles Motivates PIT/TTC Modeling Approaches
• Credit cycles are prominent in corporate defaults, losses & MKMV EDFs
• Using credit cycles in PD estimation significantly improves PIT prediction accuracy
• Empty Glass vs 20% Full Glass
Annualised Quarterly default rates: 3 Quarter
Moving Average – S&P Corporates
Credit Cycles Indices Derived from Various PD,
Rating & Loss Measures: MKMV EDFs, S&P Default
Rates & C&I Loss Rates
Default
Rate
ZGAPs: Smoothed 3 Qtr Moving Average - Corporates
3.00
Annualized Quarterly DR: 3-qtr Moving Average - Corporates
8.00%
Predicting historical defaults
improves when incorporating
credit cycle measures directly
in estimating PIT models
7.00%
2.00
6.00%
1.00
Removing the credit cycle in
implementation generates
correctly calibrated TTC PDs
5.00%
4.00%
0.00
3.00%
-1.00
2.00%
1.00%
-2.00
SP DRs
US Bank LRs
Moody's DRs
S&P Corporate DR
Long run average historical default
rate (TTC) required for capital (25 yrs)
Aguais & Associates Ltd.
CRC - IFRS9
2012-Q1
2011-Q1
2010-Q1
2009-Q1
2008-Q1
2007-Q1
2006-Q1
2005-Q1
2004-Q1
2003-Q1
2002-Q1
2001-Q1
2000-Q1
1999-Q1
1998-Q1
1997-Q1
1996-Q1
1995-Q1
1994-Q1
1993-Q1
1992-Q1
1991-Q1
2012-Q1
2011-Q1
2010-Q1
2009-Q1
2008-Q1
2007-Q1
2006-Q1
2005-Q1
2004-Q1
2003-Q1
2002-Q1
2001-Q1
2000-Q1
1999-Q1
1998-Q1
1997-Q1
1996-Q1
1995-Q1
1994-Q1
1992-Q1
1991-Q1
1993-Q1
KMV EDFs
1990-Q1
0.00%
-3.00
1990-Q1
Credit
Cycle
Index
Moody's Corporate DR
10
Almost All Credit Models Are Blind to Credit Cycles
A Component of Credit Cycles is Predictable - Systematic Cycles in Industries & Regions
Current Credit Models Are Blind to Credit Cycles – 20% Prediction is Therefore Powerful
3.0
NA Corp Z Credit
Index
2.0
Legacy Credit
Models
Z-Gap
1.0
Predicted by CreditCycle Model
0.0
Predicted by CreditCycle model
-1.0
-2.0
-3.0
Legacy Credit
Models
1990
Legacy Credit
Models
1998 1999
2001 2002
2007
Source: Moody’s KMV, Aguais/Forest research
Aguais & Associates Ltd.
CRC - IFRS9
2009
2011
11
Joint Projections of PIT PDs, LGDs, EADs Determine ECLs
Bond & Loan Portfolios – B2 Models & Agency Ratings Converted to PIT for IFRS9 ECL
Agency
Ratings
PIT (100%)
PIT Basel Models
& Agency
Ratings
TTC (100%)
BASEL PD or
Scorecard
Models
Credit-cycle indexes allow banks to
convert hybrid indicators to PIT for
accurate outlooks and to TTC for Basel
RWA
Credit +
Cycle
Index 0
-
Time
Projected ‘ECL’ – Utilise Jointly Simulated/Correlated PIT EAD/LGD/EAD
Wholesale Basel PD models & Agency Ratings are mostly TTC -– Therefore Credit
Cycle Indices can be Utilised to Convert Agency Ratings or Internal Scorecard Models
to Pure PIT Measures for IFRS9
Aguais & Associates Ltd.
CRC - IFRS9
12
PIT/TTC Approach Models Detailed Industry/Regions
Industry & Region Systematic Factors are Combined to Develop Credit Cycle Indices
Weighted Avg
Industry Sector ZI
Spot Median ZS/R Gap
Regional ZR
(Corp/FI)
LR Median ZS/R Gap
Aerospace & Defence
Banking
Chemicals & Plastic Products
Construction
Consumer Products
Asia
PD
Continental Europe
Oil & Gas
United Kingdom
Finance, Real Estate & Insurance
Hotels & Leisure
Avg PIT PD
Basic Industries
Machinery & Equipment
Latin America
North America
Pacific
Media
TTC
Medical
Steel & Metal Products
Mining
Motor Vehicle & Parts
Retail & Wholesale Trade
t
Business & Consumer Services
Time
Technology
Transportation
Utilities
Commercial Real Estate
Aguais & Associates Ltd.
CRC - IFRS9
13
PIT/TTC Approach Models Detailed Mean Reversion/Momentum
The Core of the Commercial PIT Methodology - The Evolution of Credit Cycles
Credit Cycle Behavior (Z) is Driven by Two Competing Influences – Mean Reversion &
Momentum
Almost All Data Exhibiting Credit Cycles Shows Two Competing Empirical
Influences
Z Value
Momentum
Historical z
Mean Reversion
Long Term Mean
Today
Aguais & Associates Ltd.
CRC - IFRS9
Time
14
Industry Examples of Systematic Credit Cycles & Forecasts
Banking, Finance, Insurance & Real Estate
Forecast >
< Actual
3.0
2.0
Z Gap
1.0
-
-1.0
-2.0
-3.0
-4.0
191
192
193
194
195
196
197
Banking
198
199
100
101
102
103
104
105
106
Finance, Insurance & Real Estate
Aguais & Associates Ltd.
107
108
109
110
111
112
113
114
115
116
117
118
Long Run Average
CRC - IFRS9
15
Industry Examples of Systematic Credit Cycles & Forecasts
Industries – Basic Industries, Steel & Metal Products, Mining, Utilities, Oil & Gas
3.0
Forecast >
< Actual
2.0
Z Gap
1.0
-
-1.0
-2.0
-3.0
191
192
193
194
195
196
197
198
199
100
101
102
103
104
105
106
107
108
109
110
111
112
113
Basic Industries
Mining
Oil & Gas
Steel & Metal Products
Utilities
Long Run Average
Aguais & Associates Ltd.
CRC - IFRS9
114
115
116
117
118
16
Regional Examples of Systematic Credit Cycles & Forecasts
Regional Corporates - Asia, North-America, Europe, Pacific, UK, Latin America
3.0
< Actual
Forecast >
2.0
Z Gap
1.0
0.0
-1.0
-2.0
-3.0
191
192
193
194
195
196
197
198
199
100
101
102
103
104
105
106
107
108
109
110
112
113
114
115
116
117
118
Latin America
Asia
Europe
Great Britain
North America
Pacific
Long Run Average
Aguais & Associates Ltd.
111
CRC - IFRS9
17
Agency Ratings Can Also Be Modelled with PIT Adjustments
BBB Rated Companies Have Substantial PIT Movements When Converted
MKMV EDFs for S&P BBB Rated Companies (Non-FIs)
-- NA, EU&UK and APAC
Asia&Pacific Corps
EU&UK Corps
N.America Corps
Global Average
0.8%
0.7%
0.6%
0.5%
0.4%
0.3%
0.2%
0.1%
Ja
n11
Ja
n10
Ja
n09
Ja
n08
Ja
n07
Ja
n06
Ja
n05
Ja
n04
Ja
n03
Ja
n02
Ja
n01
Ja
n00
Ja
n99
Ja
n98
Ja
n97
Ja
n96
Ja
n95
0.0%
Source: S& P & Moody’s KMV,
Aguais & Associates Ltd.
CRC - IFRS9
18
Developing Multi-Year Cycle-Adjusted IFRS9 Credit Estimates
Core PD PIT Methodology Needs to Be Adapted for Different Borrowers & Models
Retail/Consumer &
Small Bus up to £1 mil
‘Commercial’ - £1 mil to
£25-50 mil Turnover
Dominated by Behavioral
Scorecards
- ‘Backward Looking’ PIT
Dominated by Commercial
Scorecards
- Financials & Qualitatives
‘Close to ‘Fully TTC’
Need Forward
Looking
PD/LGD/EAD Term
Structures
Need PIT Conversion &
Forward Looking
PD/LGD/EAD Term
Structures
Aguais & Associates Ltd.
‘Corporates & Banks’
Dominated by ‘Agency
Replication’
- ‘Mostly TTC’
Need PIT Conversion &
Forward Looking
PD/LGD/EAD Term
Structures
CRC - IFRS9
19
PIT/TTC Framework for Wholesale Can Be Applied to Retail
PIT & Stress Test Modelling – Model Calibration Utilizes Macro & Credit Factors
Regional Credit Cycle Indices
Industry Credit Cycle Indices
4
4
3
3
2
2
1
0
Z Gap
Z Gap
1
0
-1
-1
-2
-2
-3
-3
-4
-4
-5
Jan90
Jan91
Jan92
Jan93
Jan94
Asia
Jan95
Jan96
Jan97
Europe
Jan98
Jan99
Great Britain
Jan00
Jan01
Jan02
Jan03
Latin America
Jan04
Jan05
Jan06
North America
Jan07
Jan08
Pacif ic
Jan09
Jan10
Jan11
Jan12
Jan13
Jan14
Jan15
-5
Jan16
Jan-90 Jan-91 Jan-92 Jan-93 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10
South Af rica
Hot els and Leisure
M edia
Ret ail & Wholesale Trade
Jan-11 Jan-12
Technology
Macro Factors
GDP, Personal Income,
Unemployment, House Prices
GDP, Equity Indexes & Credit
Spreads
Credit Factors
Personal Income/debt, Unemployment, House
Prices, Consumer Loss Rates, Benchmark DRs
Benchmark PIT PDFs – MKMV,
Kamakura, Bloomberg
Portfolios/
Obligors
Basel TTC
Models
Ret Morts
C Cards
UNSEC
Asset Fin
Soc Hou
PD
SME
CRE
NBFIs
Jan-15
Jan-16
Industry/
Region
Banks
LG Corps
EAD
LGD
Aguais & Associates Ltd.
Sovs
Jan-13 Jan-14
Transport at ion
CRC - IFRS9
20
IFRS 9 – Requires Point-in-Time Measures
PIT/TTC Rating – Accurate (1) Assessment of Significant Deterioration & (2) Lifetime EL
Origination
PD
PD
Bins
S&P
Mapping
S&P
Mid-Point
Mid
AAA
0.01%
AA
0.02%
A+/A
0.04%
A/A0.07%
BBB+
0.11%
0.14%
BBB
0.22%
0.
0.32%
BBB/BBB- 0.
BBB0.47%
BB+
0.71%
BB
1.07%
1.58%
BB2.27%
B+
3.21%
4.45%
B
6.05%
8.07%
B10.56%
CCC+
15.26%
CCC
23.41%
CCC37.07%
D
100.00%
Internal
Ratings
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
D
1 Yr Later
+ 1 Year
Balance of 5-Year Facility/Borrower Term
Credit Risk
Improvement
IFRS 1 Year
Credit EL
Stable Credit
Risk
IFRS
‘Significant
Deterioration’
PIT Cycle Adjusted PD Term Structures
IFRS Lifetime
Credit EL
Time
Aguais & Associates Ltd.
CRC - IFRS9
21
IFRS 9 – Requires Point-in-Time Measures
Accurately Assessing IFRS ‘Significant Deterioration’ Requires PIT Risk Measures
But PIT Risk for a Given TTC Grade Can Move Up to 5X
‘Collective IFRS Assessment’
Holistic ‘Forward
Looking’/Qualitative
Assessments
Systematic
Credit
Conditions
AIRB TTC
Models
Agency Ratings
– ‘External
Assessment’
IFRS ‘Low Risk
vs High Risk’
PD
PD
Bins
S&P
Mapping
S&P
Mid-Point
Mid
AAA
0.01%
AA
0.02%
A+/A
0.04%
A/A0.07%
BBB+
0.11%
0.14%
BBB
0.22%
0.
0.32%
BBB/BBB- 0.
BBB0.47%
BB+
0.71%
BB
1.07%
1.58%
BB2.27%
B+
3.21%
4.45%
B
6.05%
8.07%
B10.56%
CCC+
15.26%
CCC
23.41%
CCC37.07%
D
100.00%
Internal
Ratings
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
D
PIT Ratings Changes Driven Only by
Systematic Industry/Region Credit Cycles
‘BBB’ Mapping for ‘Good Times’ PIT Credit Conditions
‘BBB’ Mapping for LR Average Credit Conditions
‘BBB’ Mapping for ‘Bad Times’ PIT Credit Conditions
Aguais & Associates Ltd.
AIRB TTC Models Designed
for Capital – IFRS
‘Significant Deterioration’
Assessment Requires PIT
Measures to be Accurate
CRC - IFRS9
22
PIT PD/LGD/EAD Conversions - Critical to IFRS9 Accuracy
Depending on the Start Point – BB 5-Yr Loss Predictions Can Vary by a Factor of 5 Times
Weighted Avg
IFRS9 Accuracy Requires Both PIT Grades & PIT Prospective Credit Conditions
•
•
•
See below PD outlook for a TTC BB counterparty starting respectively with the credit cycle at TTC, a trough (Z=-2.5),
and a peak (Z=2.5)
TTC model produces the TTC line under all cyclical circumstances, whereas PIT model produces the more accurate
estimates sensitive to initial & prospective credit conditions
Top-to-Bottom
PD term structures are prospective by using a mean reversion momentum model
Accuracy Swing is
Substantial
Portfolio PD
BB Rated Cumulative 5-Year PDs
16.00%
14.00%
12.00%
10.00%
8.00%
AVG
BB PD
6.00%
4.00%
2.00%
0.00%
Year 1
IFRS9 Life Time
Year 2
TTC
Year 3
Cycle Trough
Horizon
Aguais & Associates Ltd.
CRC - IFRS9
Year 4
Year 5
Cycle Peak
23
IFRS 9 Loss Inaccuracy Can Be Significant by Not Using PIT
IFRS9/ECL Accuracy Improvements Using PIT PDs Are Substantial
• ECL estimates that arise from TTC grades & a fixed, TTC matrix will be
accurate only under the special circumstances that the credit cycle is at
TTC and expected to remain there
• In other cases, the ECLs will be inaccurate, especially so at credit-cycle
troughs and peaks (see below)
• Bottom line: need both PIT grades & a PIT forward-looking outlook
Estimates for a five-year, corporate, revolving facility with a TTC default
grade of BB equivalent, expected utilization of 60%, and DT LGD of 40%.
Case
Spot Status
Grade
Cycle
Outlook
TTC
Grade
Spot Credit Cycle Status
Grade Z level Z change
Prospective PDs
Year 1
Year 2
Year 3
Year 4
Year 5
PV
ECL/Limit
TTC
TTC
TTC
BB
BB
0.00
0.00
0.83%
2.15%
3.68%
5.47%
7.45%
2.05%
PIT
Trough
PIT
BB
B+
-2.50
-0.50
2.37%
4.95%
7.74%
10.78%
13.95%
4.48%
PIT
Peak
PIT
BB
BBB+
2.50
0.10
0.24%
0.61%
1.10%
1.72%
2.50%
0.62%
PIT outlook accounts for Z projections and Z sensitivities of PIT PDs, LGDs,
and EADs.
Aguais & Associates Ltd.
CRC - IFRS9
Huge
Top to
Bottom
Accuracy Swing
24
Developing ‘Best Estimates’ of ‘Expected Credit Losses’- ‘ECL’
Calculate ECL Using Counterparty, Facility, Model & Z Credit Factors
• Extract from source systems the relevant, categorical data (region,
industry, asset class) & PDs, LGDs, and EADs for all facilities,
• Apply the Z credit factors in converting, where necessary, the PD, LGD, and
EAD measures from source systems to spot, PIT ones,
• Run Monte Carlo simulations of the credit factors and apply those creditfactor scenarios together with the spot, PIT measures and the transition,
LGD, and EAD models in creating joint, PD, LGD, and EAD scenarios for
each facility; and,
• Combine & average those scenarios & thereby produce ECLs over the life
of each facility.
Note that to get the completely unbiased estimates anticipated under IFRS 9, will need to
eliminate upward biases from regulatory, credit models; 2013 study of Pillar 3 filings found
that the regulatory models of large banks over-estimate losses by about 50% on average
Aguais & Associates Ltd.
CRC - IFRS9
25
Developing ‘Best Estimates’ of ‘Expected Credit Losses’- ‘ECL’
Calculate ECL Using Counterparty, Facility, Model & Z Credit Factors Using Simulations
•
Draw random effects & run Z scenarios:
4.00
3.00
2.00
1.00
0.00
-1.00
-2.00
-3.00
-4.00
1
2
3
4
5
6
7
8
9
10
11
12
Quarter Number
•
Enter z (=Z) into conditional, transition models and Z into LGD and EAD models
and produce joint, PD, LGD, EAD, and thereby ECL scenarios; bold = average =
unconditional ECL:
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
1
2
3
4
5
6
7
8
9
10
11
12
Quarter Number
Aguais & Associates Ltd.
CRC - IFRS9
26
Extending Credit Cycle Modelling to Retail Portfolios
Preliminary Retail Credit Cycle Modelling: Recurring Cycles Seem Less Evident
•
•
•
Strong pattern of cycles in C&I DDGAPs inferred from US bank loss rates
Less evident in retail, with Great Recession unprecedented; thus, one may apply
the legacy, random-walk model or work harder to remove non-stationary features
before identifying cycles in the history
Correlation between retail & wholesale loss rates moderate = diversification
1990-2014: DDGAPs Imputed from Corp & Retail Indicators
C&I
Cards
Cons Other
RRE
1.00
0.80
0.60
Source: US
FRB/OCC
Quarterly Net
Charge-off
Survey of Banks
0.40
0.20
0.00
-0.20
-0.40
-0.60
-0.80
Aguais & Associates Ltd.
2014Q1
2013Q1
2012Q1
2011Q1
2010Q1
2009Q1
2008Q1
2007Q1
2006Q1
2005Q1
2004Q1
2003Q1
2002Q1
2001Q1
2000Q1
1999Q1
1998Q1
1997Q1
1996Q1
1995Q1
1994Q1
1993Q1
1992Q1
1991Q1
-1.00
1990Q1
C&I Losses
Credit Card Losses
‘Other’ Consumer
Losses
Mortgages (‘RRE’)
CRC - IFRS9
27
Point-in-Time Methodology & IFRS9 Challenges
Agenda
•
Overview – Key Points – PIT IFRS9 Challenges & Historical Background
•
Overview - PIT-TTC Framework & Methodology Required for Developing PIT
Measures
•

Systematic Credit Cycles, PIT/TTC Measures, Model Calibration & Batch Processing

Utilizing PIT Measures to Project ‘ECL’ for IFRS9 for Retail, Commercial & Wholesale
Summary – Integrated IFRS9 & Stress test Architecture & Key Points in the
Presentation
Aguais & Associates Ltd.
CRC - IFRS9
28
Point-in-Time Methodology & IFRS9 Challenges
Stress Test Approach Utilizes Macro Factor & ‘Bridge’ Model on Top of PIT Framework
‘Macro-Merton’ Credit Cycle Stress Test Framework

View GDP & equity measures as asset-value proxies

Project macro debt on the basis of trends in asset-value proxies

Treat Debt/GDP & Debt/Equity as leverage measures

Derive macro DDs (‘Default-Distance’) as ratios of leverage to historical, leverage volatility

Convert macro DDs to macro Zs (by normalising mean & variance)

Use bridging relationship to derive industry-region Zs from macro Zs

Enter industry-region Zs into the PD, LGD, and EAD models and derive stress losses
Macro Scenarios
(GDP, Equity)
Macro DD= L/V =f1 (GDP, Equity)
Macro Z = f2 (Macro DD)
Industry/Region Z = f3 (Macro Z)
Aguais & Associates Ltd.
Conditional Stressed PIT PD =
f4(Obligor’s internal assessment,
Stress Industry/Region Z )
CRC - IFRS9
29
Integrated PIT/TTC & Stress Test Solution
IFRS9 & Stress Testing – Integrated Batch Architecture – Unconditional vs Conditional
Methodology - A Single Unified Framework Across IFRS9 & Stress Testing
Integrated Stress Test ‘Stress PIT’
Batch Analytics
Retail – Commercial - Wholesale
Integrated IFRS9 ‘PIT’ Batch Analytics
Retail – Commercial - Wholesale
Basel AIRB PD-LGD-EAD Models
Enabling Batch Automation Data/Architecture/Financial
Data/CP –Static Data
Aguais & Associates Ltd.
‘CONDITIONAL
SCENARIO BASED’
-
Batch Analytics
E2E Implementation
E2E Governance &
Regulatory Sign-off
Custom Model Calibration
‘UNCONDITIONAL SIMULATION
BASED’ – BUT ‘CONDITIONAL ON
CORRECT CYCLE STARTING POINT
Batch Processing
Leverages All Basel II
PD/LGD/EAD Models
CRC - IFRS9
30
Point-in-Time Methodology & IFRS9 Workshop
Summary - Key Points in this PIT/TTC Dual Ratings IFRS9 Workshop
1) Systematic Credit Cycles Exist & Can be Measured – Motivates PIT-TTC
Distinctions
2) Evolving Regulatory Agenda – TTC for Basel II – PIT for IFRS9 & Stress Testing
3) IFRS9 Modelling – Commercial, Corporate & Retail Require Different Approaches
4) Substantial Accuracy Implications Motivate PIT – ECLs can vary across the creditcycle between starting at a peak or trough by a factor of 3-5 times !!!
5) IFRS9 & Stress Testing Require One Integrated Batch Solution & Architecture
Aguais & Associates Ltd.
CRC - IFRS9
31