Model Risk Governance: A Best Practice Approach
UK Finance webinar covering the challenges & solutions to model risk management.
With the advancements in technology and data analytics, there has been a rise in the number of predictive models used in industries such as finance, and for a variety of different purposes. These models are increasingly influencing key strategic decisions, yet they come with an associated model risk which has created the need for more sophisticated and effective model risk management (MRM) systems.
A model is a statistical method of using data to determine the quantitative estimate of behavior. Quantitative models come in a wide variety of complexity and uses, and in finance they are integral to meet regulations such as CCAR and Basel III. Models are used when decision making can not rely on human input alone, however they must be constantly monitored due to inevitable associated risk. This risk, called model risk, is the potential risk of loss that occurs from these predictive models.
Predictive models may contain technical errors, a model may be used for an incorrect purpose or a user may not fully understand the models capabilities. This means that model risk comes from either inaccuracies of the model or from poorly made decisions based upon the model. Notably model risk can only come from a model in use, and an unused model does not cause a risk.
Models are primarily used to meet regulations in the financial sector, with model risk determining the risk of financial loss. However, models and model risk have been diverging into other industries over time, from predicting fraudulent credit card transactions to assessing the likelihood of terrorists.
With the nature of risk constantly evolving, combined with the increasing complexity of models, the consequences from making decisions based on models and model risk has never been higher. Whoever developed or uses a model will be affected by model risk and so it is integral to use highly accurate models that have undergone thorough model validation and governance. Effective governance and validity testing should lead to the introduction of a model risk management framework.
Predictive models are a necessity for many industries, meaning that model risk management should form a crucial part of any business plan that relies on these models. Effective model risk management has a number of benefits from preventing losses to improving strategic decision making.
MRM helps users to understand models better in terms of their capabilities and limitations, and a greater understanding of predictive models allows them to be used more effectively. With model risk management, problems can be spotted at an earlier stage and inefficiencies can also be identified more easily, helping improve models over time.
By implementing MRM more confidence can be placed in operating models, allowing for better decision making and reducing the risk of losses.
Model risk can be reduced with effective model risk management that includes stress testing, model governance and reporting. However, traditional GRC or risk management systems lack flexibility. Instead, sophisticated risk management models not only satisfy regulations, but they are also designed to simplify and improve workflow. Advanced automated risk management models are more cost effective through decreased financial losses and time saved overall, as well as allowing for better judgements to be made through increased model transparency.
This is where we can help with our model risk management solution.