Asset-liability management for with-profit life insurance policies: A novel multi-stage stochastic programming model

Asset-liability management (ALM) is a critical issue for insurance companies because the premiums received from policyholders should be invested according to regulatory frameworks while providing suitable profitability, and simultaneously, the insurer should fulfill its obligations to policyholders on time. Our focus is on participating (with-profit) life insurance policies, where policyholders not only receive a guaranteed profit but also participate in the return of the insurer's investment-portfolio. Due to the risks of death and surrender, uncertainty in asset returns, the broad range of insurance products and regulations, it is difficult to make optimal decisions. In this paper, we aim to present a new multi-stage stochastic programming ALM model for with-profit life insurance policies. Compared to existing models that involve some simplifications, our model incorporates more details and is closer to reality. Specifically, our model is multi-stage and updates the amount of policies investment reserves based on the realized return of the investment-portfolio. Evaluation of the model across a variety of datasets confirms the effectiveness of the proposed model.