5 Key Facts About Model Risk Management
The unique challenges and demands facing the global economy since early 2020 have demonstrated the value of models. They’ve also highlighted the risks involved in using them.
Deloitte recently provided a valuable piece of research that gives into the reality of modelling and model risk management today. Using fieldwork completed in late 2020 and early 2021, it shines a light on the realities of how models are used and how their users and managers manage the risks they expose.
It’s far more nuanced than you may think.
Model Risk Management (MRM) at a glance
The US SR 11 7 guidelines define 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.
Model Risk Management has been a significant beneficiary of technology advances over the last 30 years. Let’s take a look at some of the highlights (as well as some of the negatives) of this evolution.
The research highlights many types of technology used in the MRM process, including vendor-based platforms, in-house-developed platforms, and Excel spreadsheets.
Spreadsheets were initially used to create models to
help understand credit risks more effectively
This has changed greatly over the years. Now, enterprise modeling applications provide more computing capabilities, extending into regulatory capital management, liquidity capital, and more sophisticated forms of credit risk. Across the board, their use is designed to create value and insight from large, complex, and evolving datasets.
Modeling applications have created operational, regulatory, and reputational risks that can attract scrutiny
Consider this: what if a model used in a recruitment campaign scores applications from men preferentially over applications from women? Beyond opening up you to risk under employment law, this is a nightmare for the P.R. department!
The size of the institution determines the type of model
Institutions with large model estates need scalable platforms and utilize automation to make the best use of limited skills. Smaller institutions, and niche model functions in larger institutions, will likely make use of existing capabilities that reduce the cost and overhead of implementing a Model Risk Management program while still addressing the needs of a best practice approach.
What’s the optimal Model Risk Management environment?
The optimal Model Risk Management environment is one where all the requirements of best practice MRM are aligned across the entire MRM lifecycle. In this environment, the lifecycle process is highly automated, seamless, and streamlined. Unfortunately, the research shows us that companies large and small have some way to go before their MRM processes are optimized this way.
A range of toolsets are available today
With a premium on flexibility, cost-effectiveness, and ease of use, a range of toolsets allow modeling teams to implement effective MRM regimes without the need for complex, expensive, and time-consuming enterprise I.T. applications.
A best-practice-based Model Risk Management environment will provide inventory management, document management, workflow automation, automated attestation, and end-user model support (excel management capabilities). As institutions and enterprise explore new flexible ways of developing their models, they will also need new, more flexible ways of managing ever more diverse models.