MODIFIED MULTIPLE DECREMENT TABLE AND ITS CREDIBILITY BASED ON FACTOR CHARACTERISTICS
Randi Deautama (1), Rr Kurnia Novita Sari (2*)

1) Faculty of Mathematics and Natural Science, Actuary, Bandung Institute of Technology, Bandung 40132, Indonesia
2*) Statistics Research Group, Faculty of Mathematics and Natural Science, Bandung Institute of Technology
Jalan Ganesha 10, Bandung 40132, Indonesia
*kurnia[at]math.itb.ac.id


Abstract

The preparation of the Indonesian Mortality Table IV (TMI IV) in 2019
involved 52 life insurance companies in Indonesia in the study period of 2013-2017. From this data, there may be differences in the distribution of insurance agents in each company, namely the different characteristics of the company^s customers. It is feared that this data does not represent every company but only certain companies that have the same agent distribution. In life insurance companies, there is a common type of coverage, namely Natural Death or Personal Accident (NDPA),
which means death due to illness or accident- PAD means reimbursement of medical expenses and cancellation of the policy so that coverage ends. Therefore, companies can construct a multi-cause modified life table, the Multiple Decrement (MD) Table. This table is modified into a Modified Multiple Decrement Table (MDT) by adding factors to the causes in the form of regions or others. Factor grouping is necessary to reduce the complexity of the calculation. MDT is obtained from the relationship between MD and the Associated Single Decrement (ASD). The annual exposure method is used to calculate the probability of the cause. Furthermore, extrapolation to the empty cause level and graduation to the coarse cause level Then,
credibility theory is used to obtain the credible cause level for the industry. The clustering results of the R1-R9 region and the causes of PAD (3 clusters) and SRD (2 clusters) are different regional groupings. The probability of ASD is greater than the probability of MD.

Keywords: associated single decrement, clustering, extrapolation and graduation, crediblity theory, modified multiple decrement

Topic: MATHEMATICS AND STATISTICS

ICMNS 2023 Conference | Conference Management System