Modelling Indonesian Mortality Table Using Heavy-Tailed Distribution on Two Different Age Groups
Marcellius K. Chandra, Udjianna S. Pasaribu*, Adilan W. Mahdiyasa

Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology
Jalan Ganesha 10, Bandung 40132, Indonesia
*udjianna[at]math.itb.ac.id

Abstract

In the field of life insurance, actuaries often have to determine a survival model that matches actual data before deciding a premium price based on a person age. One of the possible methods to determine insurance premiums is to use mortality tables. The mortality table is built based on a predetermined model or available data on the number of deaths. However, through these approaches, the mortality table does not model a person survival chances continuously, which produces considerable limitations in calculating the probability of survival. This study presents the possibility of modelling the Indonesian Mortality Table using heavy-tailed distribution for both genders. The results show that fitting two different distributions based on age groups produces a model that is more in line with the mortality table data than the model that relies only on one distribution. This study also provides some considerations in fitting a distribution model into mortality tables.

Keywords: Survival model, mortality table, heavy-tailed distribution, life insurance

Topic: MATHEMATICS AND STATISTICS