Optimal control and dependence modeling of insurance portfolios with Lévy dynamics (Q2276249)

The authors use a multivariate Lévy process to model the risk of an insurance company with several business lines. The risk reserves can be controled by a proportional reinsurance policy as well as by investments in a financial market. The company is interested in maximizing its exponential utility from terminal wealth. Solving the Hamilton-Jacobi-Bellman equation yields that the optimal retention level and investments are constant regardless of time and the company's wealth level. The dependence structure of the multivariate Lévy process is described in terms of its Archimedean Lévy copula. The authors derive a sufficient and necessary condition for an Archimedean Lévy generator to create a multidimensional positive Lévy copula in arbitrary dimension. This condition is related to the notion \(d\)-monotone function. The authors identify stucture conditions with respect to the Archimedean generator and the Lévy measure under which an insurance company reinsures a larger fraction of claims from one business line than from another.