ABSTRACT
Existing approaches for addressing the computational problems associated with variable annuity valuation can be divided into two categories: hardware approaches and software approaches. Hardware approaches speed up the valuation process by using hardware. Risk-neutral scenarios, which reflect unrealistic assumptions of investors, are often used in Monte Carlo simulation. When Monte Carlo is used to value a portfolio of variable annuity policies, the payoffs of the guarantees embedded in each policy are calculated at each time step of each risk-neutral scenario. Software approaches speed up the valuation by using mathematical models/algorithms. The chapter provides an overview of some software approaches that have been proposed to address the computational problems. Representative scenarios are scenario reduction techniques that select a subset of representative scenarios from the full set of scenarios based on certain characteristics of the scenarios. Representative scenarios are often used to reduce the runtime of cash flow testing models.
