Perbandingan Estimasi Cadangan Klaim Menggunakan Metode Bornhuetter-Ferguson, Booststrapping Bornhuetter-Ferguson dan Generalized Linear Model

Jidan, Ahmad Jundullah Abu (2025) Perbandingan Estimasi Cadangan Klaim Menggunakan Metode Bornhuetter-Ferguson, Booststrapping Bornhuetter-Ferguson dan Generalized Linear Model. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006211078-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Risiko adalah ketidakpastian yang melekat dalam berbagai aspek kehidupan, termasuk sektor asuransi yang berfungsi melindungi dari potensi kerugian finansial. Perusahaan asuransi dan reasuransi wajib memastikan kecukupan cadangan klaim untuk menjaga solvabilitas dan memenuhi kewajiban pembayaran klaim yang belum terselesaikan. Estimasi cadangan klaim, khususnya untuk klaim Incurred But Not Reported (IBNR), menjadi faktor krusial dalam menentukan cadangan teknis. Tantangan ini semakin besar karena pola perkembangan klaim yang bervariasi dan tidak selalu stabil. Penelitian ini membandingkan metode Bornhuetter-Ferguson (BF) yang digunakan perusahaan dengan dua alternatif, yaitu Bootstrapping Bornhuetter-Ferguson (BBF) dan Generalized Linear Model (GLM), dalam estimasi cadangan klaim. Evaluasi metode dilakukan berdasarkan akurasi menggunakan Mean Absolute Percentage Error (MAPE). Metode BBF dipilih karena kemampuannya meningkatkan akurasi estimasi parameter dengan data terbatas, sedangkan GLM menawarkan fleksibilitas pemodelan stokastik untuk distribusi klaim. Data penelitian berupa data sekunder klaim asuransi periode Januari 2023 - Juni 2024 dari PT Reasuransi XYZ dengan periode pengembangan bulanan. Hasil nilai prediction error dengan menggunakan MAPE untuk metode BF dengan dan tanpa menggunakan Bootstrap serta metode GLM adalah masing-masing sebesar 8,4173%, 49,0255% dan 8,4173%. Nilai prediction error metode BF dan GLM bisa dikatakan memiliki tingkat akurasi yang tinggi, sedangkan prediction error metode BBF sudah terbilang layak dalam mengestimasi cadangan klaim. Pada kasus kali ini metode Bornhuetter-Ferguson tanpa pendekatan Bootstrap dan metode GLM lebih baik dalam mengestimasi cadangan klaim. Hasil total estimasi klaim metode Bornhuetter-Ferguson tanpa pendekatan Bootstrap dan GLM adalah Rp2.501.526.489.
=======================================================================================================================================
Risk is an inherent uncertainty in various aspects of life, including the insurance sector that serves to protect against potential financial losses. Insurance and reinsurance companies are required to ensure the adequacy of claims reserves to maintain solvency and fulfill outstanding claim payment obligations. The estimation of claim reserves, especially for Incurred But Not Reported (IBNR) claims, is a crucial factor in determining technical reserves. This challenge is even greater because the pattern of claim development varies and is not always stable. This study compares the Bornhuetter-Ferguson (BF) method used by the company with two alternatives, namely Bootstrapping Bornhuetter-Ferguson (BBF) and Generalized Linear Model (GLM), in estimating claim reserves. The evaluation of the methods is based on accuracy using Mean Absolute Percentage Error (MAPE). The BBF method was chosen for its ability to improve parameter estimation accuracy with limited data, while the GLM offers the flexibility of stochastic modeling for claims distribution. The research data is secondary data of insurance claims for the period January 2023 - June 2024 from PT Reasuransi XYZ with a monthly development period. The results of the prediction error value using MAPE for the BF method with and without using Bootstrap and the GLM method are 8,4173%, 49,0255% and 8,4173% respectively. The prediction error values of the BF and GLM methods can be said to have a high level of accuracy, while the prediction error of the BBF method is considered feasible in estimating claim reserves. In this case, the Bornhuetter-Ferguson method without the Bootstrap approach and the GLM method are better at estimating claim reserves. The total result of the Bornhuetter-Ferguson method claim estimation without the Bootstrap and GLM approach is Rp2.501.526.489.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bootstrapping, Bornhuetter-Ferguson, Cadangan Klaim, Generalized Linear Model, Run-off Triangle, Bootstrapping, Bornhuetter-Ferguson, Claim Reserve, Generalized Linear Model, Run-off Triangle
Subjects:	H Social Sciences > HG Finance > HG4012 Mathematical models H Social Sciences > HG Finance > HG8051 Insurance H Social Sciences > HG Finance > HG8771 Life insurance
Divisions:	Faculty of Mathematics, Computation, and Data Science > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Ahmad Jundullah Abu Jidan
Date Deposited:	30 Jul 2025 05:36
Last Modified:	30 Jul 2025 05:36
URI:	http://repository.its.ac.id/id/eprint/123368

Actions (login required)

View Item