Dependent Risks and Ruin Probabilities in Insurance

Classical risk process models in insurance rely on independency. However, especially when modeling natural events, this assumption is very restrictive. This paper proposes a new approach to introducing dependency structures between events into the model and investigates its effects on a crucial parameter for insurance companies, the probability of ruin. Explicit formulas, numerical simulations and sensitivity results for dependence are established for different dependency models of first-order markovian type indicating that for various scenarios dependency considerably increases the probability of ruin.