Description: This paper investigates easy-to-implement techniques and tools developed to reduce sample-run errors and sampling bias arising from using small risk scenario samples to replicate a very large population distribution at the tails.

Description: With advancement of computing technology and innovative product design, creating dependent and nested stochastic scenarios to accurately predict insurance business has become more feasible and affordable than it used to be. However, can technology and research be accessible, applicable and beneficial to all the stakeholders? Sharing research findings and its validation tools has been our interest and focus, and even a resource-saving strategy. The presenters will share a case study on a segregated fund with hedging strategy. We investigate the reserve calculation for its GMMB in an extensive but simplified stochastic model so that our method is adoptable, applicable and extendable to real-word segregated fund/separate account model. The segregated fund model theory and implementation will be presented and its cash flows will be simulated and analyzed for reserve calculation. Varying parameter assumptions will be tested and to be modified for real-world segregated fund with hedging risk for guaranteed benefits. At the conclusion of the session, attendees will gain basic understanding of and access to: - The option pricing model and its application to the segregated fund (variable annuity's separate account) guarantee - Stochastic scenario generators for the interest rates and equity returns - R-code to program from the scenarios simulations to cash flows of hedging errors - Reserve calculations and its sensitivity to various parameter values and - Comparison of methods to increasing efficiency level and the accuracy