Claims data often forms that basis of our understanding of risk. But what if your risk model becomes so granular that you simply don’t have enough historical claims data to parameterise it with? Or what if the risks you’re insuring are changing too quickly for claims data to even be relevant at all?
Flock argues for a modular, scenario-based approach to modelling risk. One which models individual claims events separately, where every claim event is parameterised using only external datasets.