Critical factors in order to have precise claim reserves estimated:
- Optimal prediction model / models are used;
- Prediction in risk based segments (not overall portfolio);
- Fast preparation of data and execution of models ensured for both: official reserving and testing models to check improvement ideas;
- Regularly performed back-testing to check the validity of the assumptions.
Insurance Risk Reserves module offers the following main functions:
- Preparation of data sets for claim triangulation (claims paid, claims incurred, etc.; big claims, cases handled to zero may be included or excluded; flexible definition of risk segments is available).
- Pre-calculation, manual correction and saving in system of RBNS and IBNR reserves using classic Chain-ladder method modifications like Bernhuetter-Ferguson, grossing up.
- Upload of estimated reserves into Insurance Risk system - function is useful if: a) other reserving methods are used (e.g. prediction error is performed according T. Mack method, using bootstrapping methodology or other methods); b) estimation is done using market data / personal experience).
- Split of RBNS and IBNR figures into separate policy and object insured level, in order to access them together with real claim flow for flexible chosen segment.
- Check accuracy of previous calculated reserves for improvement of reserving methodology.
