Statistics for monitoring the healthiness of a portfolio

When a company offers a new insurance product, there are a set of assumptions to be made in order to start the pricing process. For example, the size or frequency of the claims could be unknown, as could be the variability around their expected value. There are several ways in which actuaries estimate these parameters, but the risk of being too far off the true distribution should be constantly monitored. This project aims to study three categories of this risk/uncertainty:

  1. Assumed Expectation: deviation of the realized losses from their assumed expected value.
  2. Volatility: even if the expected value of the incoming claims looks adequate a higher-than-expected variability could be exposing the company to unwanted risk.
  3. Extreme events: a possible accumulation of high losses or catastrophic events should also be taken into account when monitoring risk.

The goal of the project is to explore the technical risk of an insurance portfolio into the three categories described above and use statistical techniques to derive possible metrics to measure the adequacy of the assumptions.

Students: Linshan Jiang

Supervisor: DaniĆ«l Linders

Graduate Supervisor: Gabriel Casabianca