Introduction
Benchmarks are vital either where the client’s own data is sparse or where they wish to compare themselves to the overall marketplace. We have a large quantity of market data that allow us to perform such benchmarking exercises.
Benchmark data is used in two distinct ways in reserving:
• Client data is sparse, meaning that a projection using this data alone will not be credible.
• Client data is credible; the client wishes to know how their account compares to the market as a whole
Benchmark data can take a number of different forms. The two most common types used for reserving are typical loss ratios and typical development patterns.
Coverage in the marketplace may be very heterogeneous, being written on different bases with different exposures and different wordings. As such, care must be taken that the benchmark is appropriate for the data it is being used one. The benchmark data will typically be needed split by class, currency and origin year (if applicable).
Using Benchmarks with Sparse Data
Using Benchmarks with Credible Data
