Models are pivotal for governments, businesses such as insurance companies and even our everyday lives. They help us understand complex problems, anticipate the impacts of future decisions and monitor the progress of our plans. However, with their growing importance, it’s crucial to embed a robust actuarial model risk management within your organisation. This guide will provide you with the main factors you need to consider when building your model risk management process.
The Unique Challenges of Model Risk for Insurers
Model risk is critical in the insurance industry because of the complex quantitative and statistical models that are used to underpin strategic decision-making and have substantial financial implications. That’s why insurance companies must deploy robust systems and implement rigorous checks to manage the limitations of actuarial models and related risks effectively.
Model risk arises from two primary sources: fundamental errors in the model relative the model design objectives, and incorrect or inappropriate usage of the model. Long-standing problems, such as incorrect coding, usage of wrong data fields or outdated assumptions can remain undetected due to inadequate or flawed testing and validation processes.
Strategies for Managing Model Risk
Effective model risk management necessitates a robust governance framework, commonly known as a model risk management framework. Despite the increasing reliance on models, regulatory guidance on model risk remains limited for insurers. Existing principles for stress testing in banks are more developed as exemplified by the Prudential Regulatory Authority’s (PRA) policy statement PS6/23 which outlines the model risk management principles for banks.
Insurers cannot afford to restrict their model risk management; a comprehensive framework covering all relevant models across all business units is essential. Here are the key elements of our proven model risk management framework that we implement when designing and redeveloping models for clients:
Elements of an Effective Model Risk Management Framework
Model Risk Governance: The board and senior management must establish a clear policy outlining the roles and responsibilities of stakeholders and define the standards and controls for model development and operation. A company-wide approach to model risk management must fit into the organisation’s broader risk management strategy and cover model development, implementation, use and validation standards. A dedicated model risk function should be responsible for the model risk management framework and governance. Additionally, engaging an independent expert can help further validate the models to ensure objectivity.
Model Risk Tolerance: Define the board or management’s level of acceptance regarding model results based on the model type and adherence to company governance standards. It also includes the acceptance of models that may not entirely align with the insurer’s governance standards.
Model Risk Identification: Create and maintain a comprehensive inventory of models, recording risk ratings such as materiality, complexity and validation status, alongside practical details like model ownership and user information.
Monitoring and Reporting: Monitor regularly and consistently to verify the appropriate usability and functionality of models. After validation and approval, models should be monitored continuously to detect any deviations from targets and to take remedial or preventative action. Monitoring includes process verification and benchmarking to ensure that models remain aligned with their intended purpose and performance limits. Discrepancies between model outputs and benchmarks should trigger further investigation and necessary adjustments.
Model Risk Mitigation: Take action when a model’s risk profile falls outside your organisation’s risk tolerance. Mitigation strategies may include model redevelopment, expert judgment application or additional validation in response to emerging or changing risks.
Model Lifecycle Management: Manage a model’s lifecycle by paying attention to various factors to maintain its quality, including development, documentation, validation, inventory and follow-up. Keep a complete record of all actuarial and financial models, documenting every use, change and approval status. Detailed documentation of model development should include data sources, methodology, calibration, testing plans, user manuals, technical environment and operational risk assessments. Independent validation by external experts can ensure models perform as expected, meet business needs and comply with regulatory requirements.
Model Infrastructure: Optimised actuarial processes and strong technological platforms should support the model risk management framework. There should be clear documentation of the interconnectivity between source data systems, analytical tools and reporting platforms, with a rationale for platform choices. Ensuring the completeness, accuracy and timeliness of data feeding the models should be a top priority. A robust software vendor selection process is also necessary to ensure systems fit within the current infrastructure. Even open-source modelling platforms like Python, Julia and R need to be regularly monitored and audited.
By adhering to these strategies and best practices, you will be able to effectively manage model risk within your organisation, ensuring actuarial models are reliable, appropriately used and continuously monitored to mitigate potential risks and enhance strategic outcomes.
If you need support with putting in place a sustainable model risk management process our team can help you.
