IFRS 9: moving the credit industry towards account-level

IFRS 9: moving the credit industry
towards account-level provisioning
Fred Crawley
Credit Today
Sponsored
by
Damian Riley
HML
In association
with
Jason Benton
Deloitte
Dan Ray
Deloitte
IFRS 9: moving the credit industry
towards account-level provisioning
Implementing IFRS 9
Overview and
Industry Insight
Jason Benton & Dan Ray
September 2015
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Standard Introduction
IFRS 9 Business-wide Impact
Whilst IFRS 9 can be seen as an accounting policy change, it creates business-wide challenges for organisations. IFRS 9 has a direct, quantifiable
impact on provisions feeding the P&L but it also has an indirect but material impact on a wide range of factors contributing to shareholder value.
IFRS 9 Business-wide Impact
Risk adjusted
pricing
Basel III
COREP
FINREP
Risk and Finance
operating model
efficiency
• People
• Processes
• Data & systems
• Policies
• Models
• MI & Reporting
Pricing
P&L impact of IFRS 9
provisioning, including
on-going volatility
Balance sheet impact of IFRS 9
provisioning, including step
change upon introduction
Deloitte’s 5th Global
IFRS Banking Survey
Sponsored
by
Basel 3 Tier 1 and Tier
2 capital instruments
and leverage ratio
Products
and
volume
Revenue
Growth
Operating
Margin
Capital
Portfolio and
product mix
Reputation
External
audit
IFRS 9
External
rating,
reflecting
increased
P&L
volatility
Liquidity
Pillar 2B capital
planning buffer
for drawn down
Pillar 1 and 2A capital requirements
in a stress
In association
with
Market position
relative to peers
Disclosures and
market
discipline
Cost of
funding
Stakeholder
expectations
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Standard Introduction
IFRS 9 Three Phases
IFRS 9 is the new accounting standard for recognition and measurement of financial instruments that will replace IAS 39. There are 3 main phases
with the biggest change being the move from an incurred loss model to an expected credit loss model; provision levels expected to increase.
Classification and Measurement
Impairment
General Hedge Accounting
Macro Hedge Accounting
Sponsored
by
In association
with
Separate project
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Standard Introduction
IFRS 9 vs. IAS 39
IFRS 9 requires: (i) a 12-month expected credit loss to be recognised on day 1; (ii) an entity to assess whether the ‘credit risk’ on a financial
instrument has increased significantly since initial recognition; and (iii) earlier recognition of lifetime expected credit losses.
IAS 39
At initial
recognition…
Significant
increase in credit
risk …
Credit-impaired …
Sponsored
by


Lifetime expected
credit loss
In association
with
IFRS 9
Stage 1
12-month expected
credit loss
Stage 2
Lifetime expected
credit loss
Stage 3
Lifetime expected
credit loss
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Implementation Considerations
Key Technical Design Decisions
The scale of impact of IFRS 9 will be sensitive to policy and risk modelling decisions as well as timing of transition. The purpose of an impact
assessment is not only to inform key stakeholders about the anticipated financial impact but also to inform design decisions.
Financial Impact
Transition
What will be the
impact of IFRS
9 on impairment
at transition?
Volatility
How will
impairment
volatility be
affected? What
will drive IFRS 9
impairment
volatility?
Key Technical Design Decisions
Capital
How will capital
be affected?
Models
• Will you apply the Expected Credit Loss model on a portfolio
basis?
Data
Controls
Reports
Sponsored
by
• How will you define default in a comparable manner with Basel?
• Do you expect to rebut the presumption (30 and 90)?
IT
What lessons can be learnt from your quantitative
impact assessments – how can you use them to
inform your operating model design choices?
• How are you assessing the significance of an increase in credit
risk?
People
In association
with
• How will you calculate the Expected Credit Loss for undrawn
commitments and guarantees?
• Where Probability of Default (PD), Loss Given Default (LGD) and
Exposure at Default (EAD) are used, how do you generally
expect to approach data gathering?
• How are IFRS 9 and Stressed IFRS 9 Impacts going to be
different?
• What do you see as the biggest contributing factors to differences
between the Internal Ratings Based expected loss approach
under Basel and IFRS 9 as proposed?
• How are you planning to bridge this gap from a modelling
perspective?
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Implementation Considerations
Key Implementation Challenges
Implementing your IFRS 9 programme will present many challenges. Insight from industry has highlighted a number of key areas where challenges
are more frequently seen – being aware of these challenges can make a big difference to delivery success.
Key Challenges and Insight
Design
Principles
Data and
System Issues
•
•
Early and clear view is required with regard to the design principles and target state
A prerequisite to deliver a comprehensive solution capturing stakeholder (Finance and Risk) objectives at an early stage
•
Issues arising as part of the data gathering for a Quantitative Impact Study (QIS), highlight design challenges and potential
implementation issues in this space
Identifying data / system weaknesses early in analysis should be used to inform process, database and IT design decisions
•
•
•
Performing QIS is an excellent opportunity to test modelling solutions and understand impact and implementation challenges
Developing various PD, LGD, EAD and behavioural lifetime modelling techniques for each product type helps model developers and
business compare and contrast options, before a final solution is selected
IFRS 9 Change
•
•
‘Do it once and do it right’
IFRS 9 requires a substantial change in functional capabilities, both from a Risk and Finance perspective. A robust project
management structure and change control approach needs to be set up
Complexity
•
•
IFRS 9 accentuates the current complexities of the operating model
The complexity of IFRS 9 requires organisations to challenge their existing operating model in order to avoid adding even more
complexity, operational risk and cost to existing processes
•
•
•
Material cost saving potential in implementing IFRS 9 efficiently
Different implementation approaches have varying impact on project and ongoing costs as well as qualitative benefits
The selection of the TOM design aligned to the bank’s investment principles requires a comprehensive business case
•
•
Leverage streamlining opportunities to maximise benefits
The heart of the IFRS 9 design is in creating a robust but intuitive and manageable model suite. The data, systems and end user
computing solutions need to follow the same principle in order to maximise benefits
Modelling
Techniques
Cost Savings
Nimble
Solutions
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Implementation: focus areas
Modelling Considerations
The introduction of IFRS 9 means that institutions will be required to model a forward looking, lifetime view of expected losses and probability of
default. There are a number of possible approaches to this challenge.
Incorporating Economics into Model Estimates
Experienced Credit Judgement
Extrapolation
IFRS 9 introduces the need to incorporate forward-looking
economic expectations into model estimates – PD, EAD, LGD,
behavioural lifetime. There are various ways this could be
achieved
Model
Considerations
•
External
Metrics e.g.
CDS
Panel
Regression
PD
Modelling
Approach
•
EAD
EMV
Simulation
•
•
•
Migration
Matrices
LGD
Experienced Credit Judgement
Sponsored
by
•
•
In association
with
•
•
Used for loss estimation as well as stage allocation
criteria
Sensitivity to changes in economic factors will be key
Leveraging existing forecast and stress test
methodologies a possibility
Estimation of the utilisation of undrawn commitments
over the lifetime of the asset
Should attrition/prepayment be included in EAD or as a
separate model?
Alignment of balance assumptions between risk and
finance needed
Currently a ‘lifetime’ view of losses from the point of
default to write / charge-off
Stability analysis of LGD over time against macroeconomic factors
HPI probably already included for secured retail products
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS 9 Implementation: focus areas
Data Challenges
IFRS 9 introduces new data quality challenges as well as re-emphasising existing ones. With significantly increased financial disclosure requirements
as well as the introduction of the need to model losses across the lifetime of an asset, data will be at the forefront of any implementation strategy.
Modelling Data
History
• Incomplete data history is going to prove challenging in the development of a robust and forward-looking expected loss
model and in achieving effective portfolio segmentation.
Behavioural Lifetime
• High quality historic default, attrition and recoveries data will be necessary for effective modelling over lifetimes. The key to
successful implementation of behavioural lifetime models will be to leverage existing experience across the business whilst
maximising the use of incumbent credit risk modelling expertise.
Forbearance
• The presence of rich and complete forbearance data is going to be highly beneficial to IFRS 9 implementation. Whilst data
quality standards have improved, driven in part by the focus Basel places upon data quality, there is still room for
improvement, supported by a stronger control environment.
Limits
• Risk limit data are typically incomplete, inconsistent and not subject to the same controls as balance data. Historic data may
only be obtainable from paper files or use of proxies. This will be particularly problematic for manually underwritten and
watch list or high risk accounts.
Single Customer
View
Lifetime Probability
of Default
Loss Given Default
Sponsored
by
• Organisations currently recognise that a ‘single customer view’ is essential to effective credit risk management. Data
collation and validation are going to be essential, with data quality under-pinning the successful implementation of a
useable single customer data source.
• The lifetime modelling of credit risk and deteriorations thereof will be dependent on historic risk grades and expectations of
performance across these risk grades. A lack of data quality and history may present challenges in the back-testing of
lifetime models, especially under stressed economic conditions.
• Data concerning historic LGD rates, profiles and recovery curves will be an area with particular focus given that the
collateral modelling requirements for IFRS 9 will be more granular than under IAS 39, requiring institutions to monitor the
use, value and timing of – and proceeds from – disposal of collateral very closely.
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Contact Us
Today’s Presenters
Jason Benton
Senior Manager | Banking and Capital Markets
Tel: +44 (0) 207 007 4562
Mobile: +44 (0) 789 423 1008
Email: [email protected]
Dan Ray
Manager | Risk Advisory
Tel: : +44 (0) 207 303 2115
Mobile: +44 (0) 789 680 0605
Email: [email protected]
Contact Us
IFRS 9 Leadership Team
Sponsored
by
Tom Millar
Partner | Banking and Capital Markets
Tel: +44 (0) 207 007 7241
Mobile: +44 (0) 782 694 3426
Email: [email protected]
Damian Hales
Partner | Risk Advisory
Tel: +44 (0) 207 007 7914
Mobile: +44 (0) 791 720 1661
Email: [email protected]
Mark Rhys
Partner | Banking and Capital Markets
Tel: +44 (0) 207 303 2914
Mobile: +44 (0) 771 005 7625
Email: [email protected]
Tim Thompson
Partner | Risk Advisory
Tel: +44 (0) 207 007 7242
Mobile: +44 (0) 777 554 4047
E-mail: [email protected]
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Important notice
This document has been prepared by Deloitte LLP (as defined below) for the sole purpose of providing a proposal to the parties to whom it is addressed in order that they may
evaluate the capabilities of Deloitte LLP to supply the proposed services.
The information contained in this document has been compiled by Deloitte LLP and includes material which may have been obtained from information provided by various sources and
discussions with management but has not been verified or audited. This document also contains material proprietary to Deloitte LLP. Except in the general context of evaluating our
capabilities, no reliance may be placed for any purposes whatsoever on the contents of this document or on its completeness. No representation or warranty, express or implied, is
given and no responsibility or liability is or will be accepted by or on behalf of Deloitte LLP or by any of its partners, members, employees, agents or any other person as to the
accuracy, completeness or correctness of the information contained in this document or any other oral information made available and any such liability is expressly disclaimed.
Other than as stated below, this document and its contents are confidential and prepared solely for your information, and may not be reproduced, redistributed or passed on, directly
or indirectly, to any other person in whole or in part. Therefore you should not refer to or use our name or this document for any other purpose, disclose them or refer to them in any
prospectus or other document, or make them available or communicate them to any other party. If this document contains details of an arrangement that could result in a tax or
National Insurance saving, no such conditions of confidentiality apply to details of that arrangement (for example, for the purpose of discussion with tax authorities). In any event, no
other party is entitled to rely on our documentation for any purpose whatsoever and thus we accept no liability to any other party who is shown or gains access to this document.
This document is not an offer and is not intended to be contractually binding. Should this proposal be acceptable to you, and following the conclusion of our internal acceptance
procedures, we would be pleased to discuss terms and conditions with you prior to our appointment.
In this document references to Deloitte are references to Deloitte LLP. Deloitte LLP is the United Kingdom member firm of Deloitte Touche Tohmatsu Limited (“DTTL”), a UK private
company limited by guarantee, whose member firms are legally separate and independent entities. Please see www.deloitte.co.uk/about for a detailed description of the legal
structure of DTTL and its member firms.
© 2015 Deloitte LLP. All rights reserved.
Deloitte LLP is a limited liability partnership registered in England and Wales with registered number OC303675 and its registered office at 2 New Street Square, London EC4A 3BZ,
United Kingdom.
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Damian
Riley
Damian Riley
– –
director of of
business
director
business
intelligence, HMLHML
intelligence,
[email protected]
07824991857
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
HML Credentials
£35bn AUM / hosted
£50bn standby
•
HML established in 1988
•
Business Intelligence mortgage pool, created in December
2004 under the “Baseline” brand
19 standby clients
•
Collects data to build, run and monitor scorecards for Basel
IRB approach.
FCA Regulated
•
Holds information about more than 1.4 million mortgage
accounts going back to the 1980’s, 46+ million data points, c.
225,000 currently live
Clients include:
The richest source of transactional mortgage data available in
the UK. HML owns the IP.
Building Societies
Developed a range of additional analytical products and
services, principally in the areas of credit management,
benchmarking and provisioning
Investment Banks
•
•
Sponsored
by
30 servicing clients
In association
with
10 analytics clients
Retail Banks
Securitisation Deals
Asset Purchasers
IFRS 9: moving the credit industry
towards account-level provisioning
IFRS9: Terminology
•
•
Probability of Default (PD) has a number of flavours:
−
PD is the risk that the borrower will be unable or unwilling to repay its
debt in full or on time
−
Default can be defined as 180 days or 90 days in arrears
−
Probability must be calculated for a given time period i.e. 12 months, 24
months, remaining term of mortgage, etc.
Exposure At Default (EAD):
−
•
Loss Given Default (LGD):
−
•
In general, EAD is seen as an estimation of the extent to which a bank
may be exposed to a counterparty in the event of, and at the time of,
that counterparty’s default
LGD is the share of an asset that is lost when a borrower defaults.
Expected Loss (EL):
EL = PD x EAD x LGD
At HML the most trusted way of making these estimates is to use historical data to build
statistical models and behavioural scorecards…………
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
How do we build statistical models?
Statistical models utilise Empirical Data with Historic Performance to model Predictive
Data characteristics to predict Future Performance.
Past
Performance
Historic data sample at
observation point with
known performance
outcome.
Analytical
Modelling of
Predictive Data
Future
Performance
Statistical models to link
observation data to known
outcome, e.g. regression
models, decision and
classification trees.
Statistical models enable
decisions to be made
today based on prediction
of future performance.
We use a combination of the lender’s own Behavioural Data and Credit Bureau Data at a
Customer Level. We may also use elements Industry and Macro-Economic Data.
Credit Bureau
Data
Credit Reference Agencies
(CRAs) hold information on
customers behaviour on all
credit products they hold with
a variety of providers and
public information such as
bankruptcies and CCJs.
Sponsored
by
In association
with
Lender’s Own
Behavioural Data
Information from the past
on how customers
perform on the lender’s
products.
IFRS 9: moving the credit industry
towards account-level provisioning
What are scorecards?
- A scorecard is a statistically-derived tool using a technique called Multivariate Example Scorecard
Regression Analysis
Constant
- It consists of a series of multiple questions with points awarded to each
answer, which are added together to make a score
- It is designed to predict the probability of a specific event or occurrence
- It rank orders along a spectrum, high likelihood of outcome through to low
likelihood of outcome
- The scorecard can be calibrated in a variety of ways to suit the user
Sponsored
by
In association
with
Age
18-25
26-35
>=36
Time on Book
<= 2 years
2-5 years
>=5 years
LTV
<=50%
50%-80%
>=80%
300
Points
-10
5
15
-20
5
25
30
0
-25
IFRS 9: moving the credit industry
towards account-level provisioning
How are scorecards built?
We use an step-by-step analytical and statistical approach on account and customer-level
data to develop and validate the resulting scorecard.
1. Determine
development data
sample
8. Calibrate
regression
model
9. Validate and test
performance and
stability of model
2. Determine Good Bad
Definition and Outcome
Period
7. Use an iterative process to
obtain the optimised final
regression model
10. Document and
present final model
3. Carry out
exploratory data
analysis
6. Run a multi-variate
regression analysis on all
the characteristics
11. Implement
scorecard
12. Monitor
scorecard
4. Assess predictivity of
characteristics and
remove the weakest
5. Class up the characteristics
to create discrete groups with
similar relationships
13. Assess need
for re-build
For scorecard performance to be optimised and accepted by the FCA, it is essential
that the best quality data is used in the process
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
What data is needed?
•
For PD - circa 1500 “default” events from multiple time points together with outcome data
•
For EAD – sufficient account level information to track exposures including roll rates
•
For LGD:
•
•
Cure rates (proportion of accounts recovering from default)
•
Forced sale discounts (reduction from indexed value generally experienced on repossessions)
•
Time-to-sale haircuts (how long it takes to sell a property)
Forbearance performance
All the above by assets class, vintage, property type, geography, LTV bands, etc. would be ideal.
More data of the right quality will generate more accurate predictions
This could present a problem for:
•
Prime lenders where low level of defaults are experienced
•
Organisations holding low volumes of assets
•
Start-ups who have no data history
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Outsourcing: How it could work
Data provision
Clients provide account details each reporting period (month /quarter/year).
Credit report
Credit Reference data is attached to provide a fully rounded picture of each customer’s financial status.
Data provision
Provision calculation
Service provider “crunches the numbers” to calculate expected loss for each account.
Quality
Outputs quality checked and models performance monitored
Data return
Account and portfolio level provision figures returned to client.
Reporting
Interactive visual tools allow clients to review their portfolios by key segments e.g. LTV, region, cohort (vintage), and so on.
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Outsourcing?: Considerations
Do you have sufficient data and modelling skills to build and run suitable models that can withstand
regulatory scrutiny?
Should look for a service, but its also a partnership.
Expert support and advice should be provided as part of the service – its about more than just the
numbers.
Look for a highly customisable solution to meets the needs of your individual business.
Full details of all models and calculations should available for clients to review and validate.
Contribute your views and “get involved”, it’s your business.
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
The boarding process
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Exclusive: IAS39 to IFRS9 Impact
Calculated provisioning on mortgages (details downloadable) under both approaches
using accounts in the HML Business Intelligence pool
•
Increase in provisioning across the pool = 29%
•
Increase in residential property = 27%
•
Increase in buy-to-let = 53%
•
Increase in prime = 38%
•
Increase in subprime light = 18%
•
Increase in subprime heavy = 1%
•
Increase at:
• <70% LTV high but negligible (circa from 0.0001% to 0.0005%)
• 70% - 89.99%
=
54%
• 90% - 99.99%
=
31%
• 100% +
=
1%
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Question time
Please submit your questions
via the questions tab at the
top of your screen
Sponsored
by
In association
with
IFRS 9: moving the credit industry
towards account-level provisioning
Contact us
Email: [email protected]
Call: 020 7940 4835
Websites: www.credittoday.co.uk
Sponsored
by
In association
with