THE MINIMIZATION OF ANTI SELECTION SUSPECT USING UNSUPERVISED MODEL BASED HISTOGRAM-BAYESIAN (STUDY CASE: WEST JAVA BPJS KETENAGAKERJAAN DEATH GUARANTEE CLAIM DATA 2019-2021)

• Main Author: Nico Rafikasha, Muhamad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/72881
• Institution: Institut Teknologi Bandung

Participant deaths throughout 2019 to 2021 peaked. BPJS Ketenagakerjaan participants who are guaranteed a Death Benefit submit a large number of death claims for their deaths in the 2019-2021 range. From this condition, BPJS Ketenagakerjaan’s actuaries suspect that there is an opportunity for participants who submit death claims with a low membership period because there is no waiting period in death guarantee program. This behavior is called anti-selection, namely when the participants have already known the level of risk of death but immediately guarantees himself to the BPJS Ketenagakerjaan’s death guarantee. In this study, the Death Claim Waiting Period will be determined and Anti-Selection Suspect Characteristics of Death Claims will be carried out with the help of the Histogram - Bayesian Based Unsupervised Model. As a result, for PU and BPU segment coded data, respectively, two and one year waiting periods are required for anti-selection behavior to be reduced in the submission of Death Benefit claims. Then women who enrolled in the age range of 45 to 55 years became the Non-Specific Participant Characteristics with the highest frequency of suspected anti-selection, followed by the fact that Claims Offices in the Bogor and Bekasi Region became the Office with the most suspected anti-selection claims, and the Office The branch becomes the Register Channel with the most frequency of anti-selection claims. Even so, anti-selection behavior cannot be completely eliminated, but can be reduced by the previous solutive options.