Temporal compound events: Are they represented in catastrophe models?

Temporal compound events are defined in recent literature as successive events which impact the same geographical region. These kinds of events have the ability to cause catastrophic impacts. If we treat them as single events in a catastrophe model, the overall event magnitude, impact, and subsequent losses would be underestimated. The United Kingdom is vulnerable to temporally-compounding events due to low-pressure systems from the north Atlantic Ocean: the storms Desmond, Eva, Frank that occurred in December 2015 and Ciara, Dennis, Jorge that occurred in February 2020 are some recent, notable temporally compounding events that caused large economic losses.

 

For insurers and reinsurers to appropriately manage their exposure, it is imperative the tools they use truthfully reflect the risk of an insured asset being inundated several times due to temporal compound events. It has been recognised in previous research that catastrophe models are limited in their ability to handle connected, multi-hazard events. In addition, the risk of loss from temporal compound events should be demonstrated accordingly as the loss from a second event may not be as severe as the initial impact. Therefore, the definition of an event within a catastrophe model’s event set is extremely important. This provided the motivation to review temporal compound event representation in JBA Risk Management’s stochastic event set.

 

We manipulated various versions of stochastic event sets for known historical temporal compound events, and we explored how these different event sets alter the losses from catastrophe models. This research allowed us to interpret the impact various modelling strategies would have on (re)insurance companies should similar events occur in the future and provided further questions on how is best to model natural catastrophes.