A large reinsurer appointed Re Square in 2024 to design and deliver a digital replacement for its suite of Excel models used to price reinsurance structural features. The client required:
A single application to be built within the hx Renew framework to replace multiple incumbent models across its global offices
Existing loss modelling techniques and reinsurance structural features to be first rationalised, then vastly enhanced such that essentially any treaty structure would be priceable using any loss modelling technique
Application simplicity to be maintained where possible, such that underwriters could continue to carry out their own analyses in simpler cases
In designing the backing python calculation library, it was clear that an overarching loss modelling framework was needed: loss distributions would be represented by YLT simulations, no matter their origin. With a common unit of loss representation, the abstract idea could be introduced that different “components” could be connected together and “absorb” losses, according to their function. Components made available ranged from the very simple and abstract (e.g. a cap, a scalar) to the likes of “Reinstatements”, or “Sliding Scale Commission”.
By appropriately assembling these building blocks, users were empowered to model almost any structure. However, the front-end was designed within Renew to reveal just as much complexity as a user required, but no more: e.g. an underwriter wishing to model standard XoL features (AADs, reinstatements, PCs etc.) would never have to interact with the underlying components, which would be auto-assembled behind the scenes. Helper screens enabled users to create YLTs from e.g. Panjer recursions or frequency / severity simulations, as well to upload their own.
The model was successfully released in October 2024 and is in live use.