Guojun Gan - Colloquium Speaker
Variable annuities are life insurance policies that contain complex guarantees. Due to the complexity of the guarantees, there is no closed-form formula to calculate the fair market value of the guarantees except for some special cases. Insurance companies rely heavily on the Monte Carlo simulation to value the guarantees. However, it is very time-consuming to the Monte Carlo simulation to value a large portfolio that contains hundreds of thousands of variable annuity policies. As a result, valuation and risk management of such large portfolios pose a big challenge to insurance companies.
Metamodeling has been proposed to address the computational challenges. The metamodeling approach consists of two components: an experimental design method and a metamodel. The experimental design method is used to select a small number of representative variable annuity policies from the portfolio. The metamodel is constructed to replace Monte Carlo simulation to approximate the fair market values (or other quantities of interesting) of policies in the portfolio. In this presentation, Dr. Gan will talk about the computational issues arising from variable annuity valuation and present some metamodeling approaches. In addition, he will talk about some open problems related to this topic and some resources that can be used by researchers and practitioners.