Simulating Ruin Probabilities for a Class of Semimartingales by Importance Sampling Methods

From MaRDI portal
Publication:4455901





The authors consider the problem of finding the probability of ruin when the risk process is assumed to be a special semi-martingale with absolutely continuous characteristics. Let \(U_{t}=p_{t}+\sigma_{1}W_{1,t}-\sum_{i=1}^{N_{1,t}}S_{1,i}\), where \(W_{1}\) is a Brownian motion, \(N_{1}\) is a Poisson process with intensity \(\lambda_{1}\), and \(S_{1,i}\) is a sequence of i.i.d. positive random variables with finite expectation. The process \(U\) can be viewed as a risk process in a world without economic factors, where \(p\) represents the premium intensity, \(\sum_{i=1}^{N_{1}}S_{1,i}\) claims and \(W_{1}\) fluctuations in premium income and claims. Let \(R\) be another independent integrable Lévy process \(R_{t}=rt+\sigma_{2}W_{2,t}+\sum_{i=1}^{N_{2,t}}S_{2,i}\), where \(W_{2}\) is a Brownian motion, \(N_{2}\) is a Poisson process with intensity \(\lambda_{2}\), and \(S_{2,i}\) is a sequence of i.i.d. random variables with finite expectation. Interpreting \(R\) as a return on the investments generating process, the total asset at time \(t\) can be written as \(Y_{t}=y+U_{t}+\int_{0}^{t}Y_{s-}\,dR_{s}\). It is shown how the generalized Girsanov theorem can be used in connection with Monte Carlo simulation to obtain estimates of the ruin probability in the considered model.











This page was built for publication: Simulating Ruin Probabilities for a Class of Semimartingales by Importance Sampling Methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4455901)