From validation to optimisation, tackling a growing model landscape has been saved
From validation to optimisation, tackling a growing model landscape
EMEA Model Risk Management Survey
Model risk management continues to increase in importance, as banks rely on models more than ever. Recent global events such as the COVID‑19 pandemic have also revealed the weaknesses of our models and model risk management practices when the environment around us changes quickly. A mature model risk management framework creates insights into the entire model landscape of a bank and across all steps of the model lifecycle, and it enables awareness and mitigation of model risk within banks. What do these frameworks look like across the different banks and what are the main challenges and next steps?
Model landscape and inventory: The foundation for efficient model risk management
The model inventory is the central repository for all models and the foundation for efficient model risk management. It contains the scope for model risk management, but is also the source for all information about model risk.
- According to the survey most banks have a documented model definition, and most of the banks use a range of decision trees, scorecards and questionnaires to assess model candidates.
- As the reliance of banks on models increases, the models in the model inventory and the scope of the model risk management framework are also expanded. The average number of models in the inventory is around 90 models for small, 170 for medium, and 650 for large banks.
Views on governance and model lifecycle
Stronger model governance across the entire model lifecycle is a key requirement for the model risk management framework.
- The role of model owner is key in model governance, and 86% of the banks indicate that they have clearly defined and documented the role of the model owner. However, banks are facing various challenges concerning the adoption of that role. The core challenge is how to get people to act according to the responsibilities of the model owner role.
- The reporting structure that is used by the majority of the banks – which also evolves as a best practice for banks – is that of the head of model risk management reporting directly to the CRO.
- As the model landscape is expanding, the responsibilities of the committee also increase. More than half of the large banks indicate they make a distinction between operational and strategic model risk management committees.
- A large number of banks indicate that the model development and model validation teams do not have sufficient resources. Banks in the survey indicate that the key reasons are growing numbers of models in scope, increasing regulatory requirements, finding the right quantitative resources and budget constraints.
Technology and tooling: Potential for improvement
Successful model risk management framework implementations are often supported by model risk management tooling.
- Although Excel is the most commonly used tool for all sizes of banks, large banks use a vendor solution or an in‑house developed solution more often than medium or small banks.
- By far the most widely used functionality of the model risk management tooling is model inventory. Storing findings and analytics and risk reporting are also quite common.
- There is great potential for improvement for the model risk management tooling when it comes to using the information that is available for all models more effectively. Especially reporting and analytics components are often lacking at most of the banks.
Mitigating risk: Model monitoring and reporting
Model monitoring can help to alleviate resource pressures in both model development and validation. For instance, it offers more frequent and up to date information on the quality and materiality of models, without performing periodical manual model validations or first line reviews.
- Banks indicate that model performance, model outcomes and portfolio characteristics/stability are most often monitored. 87% of the banks indicate that model monitoring for credit risk models is performed, while this is only 64% for market risk models.
- Almost two third of the banks indicate that the model monitoring process is not automated. If it was, this would result in smarter ways of working across the model lifecycle and more efficient use of scarce resources.
- Of the banks, 77% indicate that they provide periodical reports on model risks to their senior management or management board. Also, 59% of the banks indicate that they have a risk appetite. However, current market insights teach us that these risk appetite statements are often simple, with a focus on model validation results.
The future of model risk management
Most banks have indicated that they intend to enhance their framework in a number of areas within the next two years. At least more than half of the banks have such intentions within the areas of analytics and reporting, scope extension (including model in scope of the model risk management framework), model risk governance, model risk policies and standards, and standardisation of processes. Of these areas, model risk policies and standards and standardisation of processes are considered as the most challenging areas to enhance.
EMEA Model Risk Management Survey
The answers to the questions in this article are included in the first edition of the Deloitte EMEA Model Risk Management Survey. This survey presents our insights into the current model risk management practices and challenges of banks across Europe, Middle East and South Africa. It was conducted between November 2020 and February 2021. A total of 80 banks across EMEA participated. The survey covers all the key building blocks of model risk management across four themes: governance, model landscape and inventory, technology and tooling, and monitoring and reporting. It renders valuable insights into model risk management to help banks to be responsible businesses.