Model risk management in banking incl. Regulatory requirements

Model risk management in banking. How to manage model risk including regulatory requirements

1. W W W . P E A K S 2 T A I L S . C O M
Model Risk
Management (MRM)

2. W W W . P E A K S 2 T A I L S . C O M
Discussion Points •What is Model ?
•What is Model Risk & Sources ?
•Types of Models within Bank
•Why MRM ?
•What are Associated Regulations with MRM ?
•What Next ?

3. W W W . P E A K S 2 T A I L S . C O M
What Is A Model Risk?
A model is a method, system or approach that applies statistical, economic, financial or mathematical theories, technique,
expert judgment and assumptions to process inputs to quantitative outputs or reports.
A model is comprised of inputs, calculations and outputs and is considered for the purposes of Model Risk Policy as an
end-to-end concept.
A model may have several internal components, in which case they are all considered as parts of the model (e.g. the
market and payoff component of certain valuation model, the NDAO, cure and charge off component of certain LGD
model).

4. W W W . P E A K S 2 T A I L S . C O M
•It is the risk of potential loss that an institution may incur as a consequence of decisions that are principally based on
output of internal models (as discussed above).
Model Risk may arise principally because of two main reasons:
•The model might have fundamental inaccuracies that results in erroneous output for its intended use
•The incorrect or inappropriate use of models
Model Risk
Sources
Data
Model
Implementation
Methodology
Parameters &
Assumptions
Misuse Interpretation Inventory
What Is A Model Risk?
Sources Of Model Risk

5. W W W . P E A K S 2 T A I L S . C O M
Insurance Models
•Actuarial Models
•Loss Forecasting
•Reserving Models
Valuation & Pricing Models
•Derivatives
•Structured Products
•Risk Based Pricing
models
Portfolio Model
•Capital Forecasting
•Stress Testing
•Economic Capital
Credit Risk Models
•Risk Rating models
•PD, LGD, EAD models
•Impairment / IFRS9
Marketing Models
•Marketing Models
•Client targeting models
Finance and Liquidity Models
•P&L Attribution
•Cashflow/NPV
•FVA and PVA
•ALM and Liquidity Risk
Models
Decision Support Models
•Credit Sanctioning
Models (Scorecards)
•Customer Targeting
Model
•Collection Models
Compliance Models
•AML Model
•Fraud Models
•Trade surveillance
Investment Management
•Trading strategies
•Security/Asset Pricing
Types of Model with Bank

6. W W W . P E A K S 2 T A I L S . C O M
•Models are now used in every department within bank for decisioning
•Market observation shows that models are now used more widely than ever
Increasing Dependence on Models
•Complexity of the models increases as demands are increasing
•Suite of AI/ML models are now been used for decision making e.g. Fraud Models, Acquisition
scorecards etc.
Growing Model Complexity
•Inaccurate model design, data ,governance and usage may lead to suboptimal lending and
pricing decisions which may lead to financial/reputational losses
Poor Model Risk management leads to
Financial/Reputational Losses
•Regulators have increased the frequency of exams and reviews on banks MRM function
•Regulators are publishing new guidelines
Increasing Regulator Focus on MRM
•Bord of directors, senior management, regulators are expecting sophisticated risk
management for the suite of model within a bank
Stakeholder/ Senior Management Demand
•Because of the above reasons there has been a contestant increase in demand in SMEs with MRM expertise.
•There has been an increasing pressure on banks to demonstrate that institution’s MRM is deemed fit for purpose in
comparison to peer institutions
Why Model Risk Management ?

7. W W W . P E A K S 2 T A I L S . C O M
KEY REGULATORY REQUIRMENTS
Regulatory
Requirements
2011- Federal
Reserve SR11/7
2013- Federal
Reserve “Capital
Planning
Requirements ”
2014 – EBA SERP
Guidelines
2016/17 - EBA TRIM
Guidelines
2017 – PRA Stress
Test Model
Management
Principles
2018/19 – ECB
Guidelines to
Internal Models
2019/20 – AI/ML
Guidelines
2023 – PRA SS1/23
Model Validation
•Regular validation of
models(Initial/Periodic/
Change) ensuring Fit for
Purpose
•Independent Validation team
should be established
Model Governance
•Clear governance structure of
model development,
implementation and Usage
•Documentation of model
development process and
decisions
Model Inventory
•Setup a inventory of in use
models
•Regular updates and
maintenance of inventory
Risk Assessment
•Conduct risk assessment of
all in use models to identify
weakness
•Stress testing and scenario
analysis performed as a part
of assessment
Reporting
•Reporting of model
performance and risk to
senior management
•Clear identification of model
limitation and assumptions
Non-Compliance Leads to
Regulatory Fines Increased Capital Charges Reputational Losses
Guidance/Regulation/Inception

8. W W W . P E A K S 2 T A I L S . C O M
What Next ?
Next
Understanding
MRM Structure
Tiering &
Quantification
Workshop on
Model Tiering
Workshop on
Model Risk
Assessment
Management &
Controls
Model Inventory
(Key Items )
Handson session
on Model
Validation
Demystifying MRM
Policies/ Standard &
frameworks
Governance/
responsibilities
Understanding
MRM Governance
MRM Roles and
Responsibilities
Walkthrough of
Key Regulatory
requirements
•MRM validates all in use models within an institution both at inception and ongoing (not just the model with bad
performance)
•Model Validation is critical MRM exercise, but there are additional responsibilities of MRM and is not confined to validation
Key Takeaways

9. W W W . P E A K S 2 T A I L S . C O M
Thank you!
•Call - 86-97-86-97-86
•Website - www.peaks2tails.com