Perbandingan Metode Bornhuetter–Ferguson Dan State Space Model Scalar Dalam Estimasi Cadangan Klaim Berbasis Run-Off Triangle

Lesmana, Vanessa Emanuela (2026) Perbandingan Metode Bornhuetter–Ferguson Dan State Space Model Scalar Dalam Estimasi Cadangan Klaim Berbasis Run-Off Triangle. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006221042-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Penelitian ini bertujuan untuk mengestimasi cadangan klaim pada lini bisnis asuransi kendaraan bermotor non-coinsurance dengan menggunakan metode Bornhuetter-Ferguson dan State Space Model. Data yang digunakan merupakan data historis klaim dan premi perusahaan asuransi berdasarkan accident year 2015–2024 dan development period 0–9. Metode Bornhuetter-Ferguson digunakan dengan menggabungkan pola perkembangan klaim historis dan estimasi awal ultimate loss berdasarkan Net Earned Premium. Sementara itu, State Space Model digunakan dengan memodelkan cumulative paid claims sebagai latent state yang diestimasi melalui tahapan Kalman Filter dan Kalman Smoother dengan algoritma Maximum Likelihood Estimation-Expectation Maximization. Hasil penelitian menunjukkan bahwa data klaim pada lini bisnis kendaraan bermotor memiliki karakteristik short-tail, karena sebagian besar pembayaran klaim terjadi pada development period awal. Total estimasi cadangan klaim yang diperoleh dengan metode Bornhuetter-Ferguson adalah sebesar Rp78.803 juta, sedangkan metode State Space Model menghasilkan estimasi cadangan sebesar Rp63.205 juta. Selisih estimasi cadangan antara kedua metode adalah sebesar Rp15.598 juta, dengan metode Bornhuetter-Ferguson menghasilkan nilai yang lebih besar. Perbedaan tersebut menunjukkan bahwa Bornhuetter-Ferguson cenderung menghasilkan estimasi yang lebih konservatif karena mempertimbangkan prior ultimate loss, sedangkan State Space Model menghasilkan estimasi yang lebih dinamis karena mempertimbangkan ketidakpastian pada proses perkembangan klaim. Dengan demikian, kedua metode dapat digunakan sebagai pendekatan dalam estimasi cadangan klaim dengan karakteristik hasil yang berbeda.
======================================================================================================================================
This research aims to estimate claims reserves for the non-coinsurance motor vehicle insurance line of business using the Bornhuetter-Ferguson method and the State Space Model. The data used consists of historical claims and premium data from an insurance company based on accident years 2015–2024 and development periods 0–9. The Bornhuetter-Ferguson method is applied by combining historical claims development patterns with an initial estimate of ultimate loss based on Net Earned Premium. Meanwhile, the State Space Model is applied by modeling cumulative paid claims as a latent state estimated through the Kalman Filter and Kalman Smoother stages using the Maximum Likelihood Estimation-Expectation Maximization algorithm. The results show that the claims data for the motor vehicle line of business has short-tail characteristics, as most claim payments occur in the early development periods. The total claims reserve estimate obtained using the Bornhuetter-Ferguson method is Rp78,803 million, while the State Space Model method produces a reserve estimate of Rp63,205 million. The difference in reserve estimates between the two methods is Rp15,598 million, with the Bornhuetter-Ferguson method yielding the larger value. This difference indicates that Bornhuetter-Ferguson tends to produce more conservative estimates because it takes the prior ultimate loss into account, whereas the State Space Model produces more dynamic estimates because it accounts for uncertainty in the claims development process. Thus, both methods can be used as approaches for estimating claims reserves, each with different result characteristics.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Cadangan klaim, Bornhuetter-Ferguson, State Space Model, Kalman Filter, Segitiga run-off, Claim reserves, Bornhuetter-Ferguson, State Space Model, Kalman Filter, Run-off triangle
Subjects:	Q Science > QA Mathematics > QA274.2 Stochastic analysis
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Vanessa Emanuela Lesmana
Date Deposited:	21 Jul 2026 04:35
Last Modified:	21 Jul 2026 04:35
URI:	http://repository.its.ac.id/id/eprint/136051

Actions (login required)

View Item