Few products in re/insurance are as challenging to model as liability due to its long-tail nature and its susceptibility to changing technological, economic, legal and societal conditions. In addition, we do not have a good understanding of the exposure of a liability portfolio towards risk accumulation.
What is the next asbestos? When will it happen and how big will it be? How exposed are portfolios? These questions are becoming more and more relevant in an increasingly globalized world and where the development of new technologies is speeding up. When I first started working in Casualty years ago, I thought that we'd need something analogue to what is already in place for natural catastrophe risk management. An approach aiming at the better understanding of the cause-effect chain of a liability event, from the loss generating process via all the economic, societal, legal, and judicial drivers, to the final insurance coverage.
A team at Swiss Re developed a unique forward-looking modeling (FLM) approach to address these needs. Swiss Re's Liability Risk Drivers™ (LRD) model starts from loss scenarios and risk drivers, and models the cause-effect chain, using historic payouts to challenge model predictions, i.e. to check if the model was able to predict the past without directly using it.
Forwarding-looking modeling allows us to anticipate the impact of a change or of a trend (such as new class action laws or medical cost inflation) without having to wait for claims to emerge. Re/insurers can use it to confidently grow their business in new markets and segments where we may lack sufficient loss experience such as high-growth markets. FLM gives re/insurers a better understanding of how a portfolio is exposed to big liability events.
Do you also have FLM approaches in place? Do you see other areas of application of FLM? It would be great if you could share them as comments below.