Aon: Guiding insurers on natural catastrophe risk
Steve Bowen, head of catastrophe insight at Aon, and Sami Pant, senior scientist at the insurance solutions and advisory company, explain why natural catastrophe risk is so important for insurers
Do you see cat risk becoming more important than ever for insurers given the rise in nat cats around the world?
Steve Bowen (SB): As the topic of climate change has grown in its importance and influence across the industry in recent years, there is a renewed sense of urgency to better account for anticipated changes in hazard behaviour, and the resulting direct and indirect impacts that future events may cause. We need to be focusing on the risks of tomorrow and working across various public and private sector groups to encourage collective best practice mitigation and adaptation measures to help limit the growing loss costs being incurred by re/insurers.
In the last year, which particular areas of cat modelling has Aon helped clients with?
SB: There have been several model development updates, in addition to real-time delivery solutions that are aiding clients in real-time to make important decisions and understand the potential risk to their portfolios. In the past 12 months, Impact Forecasting has expanded its Automated Event Response (AER) tool to provide real time loss projections for a re/insurer portfolio in advance of a pending event. This is for the US hurricane, European windstorm, and Japan typhoon perils. The US hurricane AER solution is used by more than 70 companies in the US alone. Additionally, another big advancement was the approval of the Florida hurricane model by the Florida Commission on Hurricane Loss Projection Methodology.
This accreditation allows clients to file rates in the states based on Impact Forecasting model output, and marks a major milestone for the new model and for future business in the US and beyond. The approved model has already been recognised and used by a number of Florida primary insurers, Lloyd's syndicates, and global re/insurers.
Where do you see the major challenges for insurers looking at cat risk in the future?
SB: The biggest challenge the insurance industry faces is how to properly account for uncertainty. While climate science continues to advance, and we are gaining confidence in how near- and long-term events may behave, there remains large inherent uncertainties of what annual or seasonal cat risk will look like and how we can more granularly quantify anticipated changes.
Further challenges include greater connected or compounded impacts of large-scale concurrent disasters; accounting for growing costs tied to liability or litigation risk; reasonably anticipating what future exposure and structural vulnerabilities will be; and meeting the regulatory or rating agency demands around climate-related financial disclosures.
How important is data in this field?
SB: Data serves as the fundamental foundation of catastrophe modelling and any other type of disaster risk analysis work. The better the data, the better the output. While we live in a golden era for data availability and access, including robust historical event or archived atmospheric / oceanic environmental conditions, there remains significant challenges in gaining enough access to data to better calibrate models.
This is particularly true when accounting for updates to vulnerability assumptions in a model. As data access grows and computational power improves, we should expect better modelled loss performance to better guide insurers in planning for future risk.
How important is data in this field?
Sami Pant (SP): Data is crucial in catastrophe modelling. Compared to the past, we have an abundance of better quality data now. We fully leverage the latest data to update our models and obtain a comprehensive view of risk.
For example, the data from recent storms and recent changes in land cover have been used to update our hurricane model. We are also fortunate to have access to claims data from multiple insurers for the recent storms, which has been crucial in obtaining up-to-date vulnerability functions.
With each passing year, we expect to obtain more and better data that can be used to further improve our models. In addition, we access a breadth of data from different academic collaborations – notably around climate change – which we then incorporate directly into our models.
Did you make any upgrades or improvements in your models or products in the last 12 months?
SP: We regularly update our models to incorporate the latest data and the advances from the scientific community. Recently, we made some major upgrades to our hurricane model to consider climate change impacts based on the latest climate science and to incorporate the newly-released CMIP6 data.
Our Florida hurricane model, which was approved last year by the Florida Commission on Hurricane Loss Projection Methodology includes significant updates to model components and databases, and it was well-received during the review process. Impact Forecasting's Automated Event Response has also been recently updated to provide a more robust view of risk with inclusion of uncertainties.