Retrocession for Portfolio Optimization in Reinsurance

Sammanfattning: Reinsurance is the insurance protection of an insurance company. Retrocession is reinsurance for a portfolio of reinsurance contracts. Reinsurance portfolios can comprise several thousand contracts that may be contingent on the same events, which makes retrocession a complex decision. This thesis develops an optimization model for retrocession, where the aim is to maximize the expected result and satisfy contraints on risk. A review and development of risk measures that can be included in the model is performed. The optimization model is implemented and applied to a large portfolio of reinsurance contracts using mathematical programming algorithms. Results suggest that a benefit amounting to several percent of the annual expected result may be obtained by applying optimal retrocession to the reinsurance portfolio. The results depend on several assumptions that, if not fulfilled, may diminish the benefit.