Application of the affine stochastic model on Singapore mortality data.

• Main Authors: Chia, Fu Yuan., Meng, Deshuo., Zou, Deshun.
• Other Authors: Balasooriya, Uditha
• Format: Final Year Project
• Published: 2008
• Subjects: DRNTU::Business::Finance::Actuarial science
• Online Access: http://hdl.handle.net/10356/10336
• Institution: Nanyang Technological University

In recent years, stochastic modeling has become a buzz word in the finance industry and the phenomenal advantages of stochastic models have attracted immense attention from researchers. Here, we fit Singapore's mortality rates to an affine stochastic state-space model. The model is fitted through maximum likelihood estimation process, which in our project relies on the assumption of white and normal innovations in the state-space model. Kalman Filter technique is adopted to facilitate the derivation of the maximum-likelihood estimators (MLE) and EM-algorithm with direct search technique is applied in finding the numeric estimates for the MLE. The affine stochastic state-space model allows us to capture two important features of mortality rates: time dependency and uncertainty of the future development. This model gives us more realistic premiums and reserves, and it quantifies the risk of the underlying mortality rates. We also examine the effects on profitability of insurance products that are linked to the development of the mortality rates.