Home
>
Explore Content
>
All Issues
>
Article View
The Bayesian Estimation on Korean Male Mortality Rates Using Poisson Log-Bilinear Model and its Application
Journal of Insurance and Finance
2016, 27(1),
pp.23-49
DOI :10.23842/jif.2016.27.1.002
Publisher : Korea Insurance Research Institute
Research Area : Business Management
DOI :10.23842/jif.2016.27.1.002
Publisher : Korea Insurance Research Institute
Research Area : Business Management
-
In this paper, we intend to project the future Korean mortality rates, and analyze the effect of mortality improvement on the life expectancies and the actuarial present values of life insurances and life annuities. Under the Poisson log-bilinear model, we estimate the parameters for the Korean males by using the hierarchical Bayesian methodology proposed by Czado et al.(2005). For comparison purpose, we provide the estimation results based on the singular value decomposition and the maximum likelihood estimation. Our results show that the parameter estimates based on three different estimation methods do not exhibit a large difference between each other, but the traditional estimation methods underestimate the prediction interval. We expect that the Bayesian estimation method not only can provide a solution to this underestimation problem, but could be very useful to evaluate insurers’ liabilities in the future.
279
Viewed
Print this page
Download PDF