Abstract
In practical situations, we observe the number of claims to an insurance portfolio but not the claim intensity. It is therefore of interest to try to solve the 'filtering problem'; that is, to obtain the best estimate of the claim intensity on the basis of reported claims. In order to use the Kalman-Bucy filter, based on the Cox process incorporating a shot noise process as claim intensity, we need to approximate it by a Gaussian process. We demonstrate that, if the primary-event arrival rate of the shot noise process is reasonably large, we can then approximate the intensity, claim arrival, and aggregate loss processes by a three-dimensional Gaussian process. We establish weak-convergence results. We then use the Kalman-Bucy filter and we obtain the price of reinsurance contracts involving high-frequency events.
Original language | English |
---|---|
Pages (from-to) | 93-107 |
Number of pages | 15 |
Journal | Journal of Applied Probability |
Volume | 42 |
Issue number | 1 |
DOIs | |
Publication status | Published - Mar 2005 |
Externally published | Yes |
Keywords
- Cox process
- Gaussian process
- Kalman-Bucy filter
- Piecewise-deterministic Markov process theory
- Shot noise process
- Stop-loss reinsurance contract