Developments in Modelling Risk Aggregation

The report suggests improvements to the current modelling techniques used by complex firms to aggregate risks. It also examines supervisory approaches to firms' use of risk aggregation models, particularly in light of the global financial crisis.

Mr Tony D'Aloisio, Chairman of the Joint Forum and Chairman of the Australian Securities and Investments Commission, said "This report is essential reading for firms considering ways to make more effective use of risk aggregation methods, and for supervisors wanting to understand firms' use of risk aggregation models to help identify shortcomings in a firm's approach."

Key Findings

• Despite recent advances, models currently in use have not adapted to support all the functions and decisions for which they are now used. Firms using these models may not fully understand the risks they face, including tail events.
• While some firms are addressing these issues - particularly the treatment of tail events - others are not.
• Firms face a range of practical challenges when modelling risk aggregation. These include managing the volume and quality of data and communicating results in a meaningful way. Despite these challenges, the Joint Forum found that firms have little or no appetite for fundamentally reassessing or reviewing how risk aggregation processes are managed.
• In carrying out their responsibilities, supervisors generally do not rely on aggregation models currently used by firms as they are generally considered a "work in progress" with best practices yet to be established. Substantial improvements and refinements in methods - particularly in aggregating across risk classes - are needed before supervisors are likely to be comfortable in placing reliance on these models for supervisory purposes.

Key Recommendations

• Firms should improve their risk aggregation techniques, for example by reassessing and reorienting models according to their purpose and function. Such improvements will assist firms to better comprehend the risks they face.
• Firms using models for risk identification and monitoring purposes should ensure they are sufficiently sensitive, granular, flexible and clear. Models used for capital adequacy and solvency purposes should be improved to better reflect tail events.
• Supervisors should recognise the risks posed by continued use of current aggregation processes and methods. Supervisors are urged to communicate their concerns to firms while highlighting the benefits of appropriately calibrated and well-functioning aggregation models for improved decision making and risk management. Supervisors should work with firms to implement these improvements.