How Do You Manage Your Model Risk?
An Introduction to Model Risk Management

Our clients have an understanding of the concept of model risk management, but there is uncertainty about the key requirements for an effective and holistic program. As we will explain in future posts, a clearly defined framework significantly enhances business efficiency, quality of output, and regulatory compliance.

So what is model risk management and why is it so important?
The financial services industry has evolved and institutions regularly use models for AML activities such as customer screening, risk rating and transaction monitoring. Depending on the size and complexity of the financial institution, business requirements and purpose of the model a model can be as simple as a spreadsheet or it can be a complex system driven by algorithms.

The Financial Conduct Authority’s Principles for Businesses and its Senior Management Arrangements, Systems and Controls sourcebook require firms to take reasonable care and to establish and maintain systems and controls to manage their risks that are appropriate to their business model. This allows institutions to interpret and determine what is required by their business and tailor their approach accordingly.

The U.S.-based regulators have taken a much more prescriptive approach, outlined in guidance issued by the Office of the Comptroller of the Currency and the Federal Reserve Board. Guidance defines a model as “a quantitative method, system, or approach that applies statistical, economic, financial, or mathematical theories, techniques, and assumptions to process input data into quantitative estimates.”

Simply put, a model has an input, a processing component, and output – and decisions are driven by the output. When operating effectively, models can improve business decisions. If models are incorrect or misused, they can impose costs.

The potential for poor business and strategic decisions, financial loss, or damage to reputation when models play a material role is the essence of model risk.

Global regulations and expectations regarding model risk management differ however the risks and fundamentals of managing and mitigating those risks do not.

What are the components of an effective model risk management program?

For the models to be fit for purpose, an institution must:
– Identify and inventory models
– Assess models for risk
– Document the model development, implementation, and use
– Validate or test the models
– Tune and Optimise (calibrate) the models
– Govern the models

See the white paper “Effective Model Risk Management for Financial Institutions: The Six Critical Components” for an overview of this approach to model risk management.

mrm-intro

Source: Crowe Horwath LLP analysis

As the exhibit shows, each action is a vital component of model risk management. In future posts, Crowe Risk Consulting professionals will outline the regulatory expectations for each component and present a business case for implementing a strong model risk management framework.