Description: This paper demonstrates that extreme value theory (EVT) can be used as a tool to model policyholder behavior in the extreme tail. This paper applies EVT to the study of variable annuity dynamic lapse behavior in the extreme tail. It illustrates a process whereby a company can fully model policyholder behavior, including the extreme tail, with existing data. No judgment about the extremal behavior is necessary.

Description: The article discusses modeling of variable annuity behavior risk, with a particular focus on how interaction between different behavioral assumptions complicates modeling. The article gives a brief overview of variable annuity behavior risk, discusses the challenges that behavioral interactions pose for VA modeling, and discusses three potential solutions for dealing with this complexity.

Description: We analyze how volatility management techniques in VAs affected policyholders and insurance companies