PODRA Model: being simple and useful
When developing a model for partial internal risk assessments of an insurance company, we tried to focus on the major risks and the major risk interactions. Furthermore, considering risk mitigation instruments in a very comfortable was a major aim.
The model should lead to meaningful and consistent results. Therefore sufficient result variables were intended to be observable, including premium volume, loss and risk measures on gross, ceded and net bases.
We set-up a simple and useful model, the PODRA model, which can be dynamically expanded (all submodules customizable):
- Underwriting Information (e.g. per line of business and/or regions) including risk bands per segment.
- Loss generators (perils per line of business, large / cat / attritional loss splits.
- Structural organisation of the company (eg. line of business, regions and/or profit center)
- Risk Mitigation instruments (treaty reinsurance, pool memberships, facultative cession policy)
We select all necessary result variables per submodule in the result descriptor. All the selected output variables, including risk measures on them, are stored in the SQL database for future use.
Most favorable is the generation of outputs in standard presentation format straight from the database. We tried out several reinsurance structures for optimization purposes, and had the results in presentable format ready for all cases.
An example, including VBA programming, will be shortly available on request.
My conclusion is that the PODRA model is a very strong and simple model for partial internal modelling.
Further improvements are on their way: Extensions to include reserve and asset liability mismatch risk are currently within the implementation phase.
-- Jörg

