Fundamentals of Model Risk

Catarina Souza, Head of Division, Model Development and Review at Bank of England; Dr Diederick Potgieter, senior technical specialist in capital management at the Prudential Regulation Authority (PRA), Bank of England; Dr Alan Forrest, former lead developing and validating risk models in Virgin Money UK, Royal Bank of Scotland and Halifax Bank of Scotland; and other international specialists come together for this uniquely educational day of Model Risk.

We build and use models in our daily work and research, but how well do we know the risks of using those models? Models can be inaccurate and we routinely test for this, but they may also be built on incorrect assumptions or shifting data; they may be poorly specified, unstable, incorrectly interpreted, unexplainable, or misapplied.

Model Risk is the name given to the probable loss (however measured) caused by an incorrect, inappropriate, or misused model. This risk appears in all models, from simple regressions to complex GenAI systems, and in all contexts, from classic data analysis and inference, to deep learning and data engineering. Model Risk explores and expresses our degree of trust in our models.

But does Model Risk matter? And if it does, how does it appear, what impact does it have, and how is it managed? And if managed well, what benefits does a Model Risk mindset bring to modelling and decision-making? In this introductory session, invited experts from the Bank of England and international banking will address these questions. It will call on the experience of the Financial Services industry where over many years Model Risk has become a deeply embedded discipline.

1. To understand how Model Risk is a feature of all kinds of modelling, and how its scope extends beyond data quality, validation, testing, and monitoring.
2. To learn the Model Risk mindset — systematically considering how a model could fail, and how best to respond to that possibility.
3. To observe how Model Risk is managed in the Financial Services industry, where it is both a professional discipline and a regulatory requirement.
4. To demonstrate the benefits of a strong Model Risk approach for modelling practice, and for business decisions and outcomes.
5. To leave with practical ideas how to identify and size up the key sources of Model Risk in the models you build or use.

Please note that the Beginner sessions are tailored towards early career professionals, end-stage students, and academics new to model risk.

The Evening session is also relevant to the above, but will be of interest to individuals at all stages in their career, from Student to CEO.

This session is aimed at three kinds of people, all ready to look critically at their approaches to building and using models:

· Academics and experienced researchers who build or use data-based models in their analysis and who wish to know more about an industry-professional standard that could help them incorporate data- and model-uncertainty systematically into their conclusions.

· Research students who use structured data and models, and who wish to understand better how to detect and express model assumptions, weaknesses and risks, and how to prioritise and plan for these.

· Analysts and quants in business, data scientists and engineers who build or use data-based models in their work and who are asked to justify their outputs and express the risks involved in a way that can meet independent challenge and audit.

a. Welcome and Introductions:

Breaking the ice with examples and observations about models (20 minutes)

b. The Model Risk mind-set:

A structured approach to Model Management - Model Life Cycle and Model Scope, Model Risk as a kind of Risk Management. (30 minutes)

c. A mature case study: Financial Services

i. How Model Risk appears in banking (30 minutes)

ii. How Model Risk is regulated in UK Financial Services by the Bank of England. (30 minutes)

iii. What are the benefits and the costs? (20 minutes)

d. Final Discussion

How could Model Risk help your modelling and your work? What things could you do differently? (30 minutes)

II. 5:10pm-6:00pm Panel Session and Discussion

A panel discussion featuring regulators, acadmics, and industry leaders will be moderated by Dr Alan Forrest. Chatham house rules are followed (no recording & participants are free to use the information received, but neither the identity nor the affiliation of the speaker(s), nor that of any other participant, may be revealed).

II. 6:00pm-7:30pm Drinks reception

Do not miss it!

The event will be held on the 4th floor of the University of Edinburgh Business School, in the conference room & roof terrace.

For further information on how the University of Edinburgh Business School processes your data, you can view our Privacy Policy here.

---

Follow the Credit Research Centre (CRC) to find out upcoming events.

Find out how to become a member of the Model Risk Managers' International Association (MRMIA) here.

---

We apologise to our remote audience for the in-person nature of this event. However, we strive to provide the panellists' slides, contingent on their permission.

Info: Breaking the ice with examples and observations about models.

Info: A structured approach to Model Management - Model Life Cycle and Model Scope, Model Risk as a kind of Risk Management.

Info: i. How Model Risk appears in banking (30 minutes)
ii. How Model Risk is regulated in UK Financial Services by the Bank of England. (30 minutes)
iii. What are the benefits and the costs? (20 minutes)

Info: How could Model Risk help your modelling and your work? What things could you do differently?

Info: A panel discussion with a selection of speakers from the main sessions and guest panellists. Chatham house rules are followed (no recording & participants are free to use the information received, but neither the identity nor the affiliation of the speaker(s), nor that of any other participant, may be revealed).

Know what’s Happening Next — before everyone else does.

Get App

Host or Publisher Business School