Hierarchical Ensemble Actuarial Method for Motor Claim Reserving under Indonesia’s PSAKBI
Abstract
This study develops and evaluates a hierarchical ensemble actuarial approach for motor claim reserving under Indonesia’s PSAKBI framework. The method integrates traditional actuarial techniques with modern machine learning models in a structured ensemble design to enhance predictive accuracy, reliability, and transparency. Using motor insurance claim data, the ensemble was compared against conventional single-model reserving practices. Results show that the proposed approach achieves lower prediction error (MSE = 220.3), accurate calibration (94.7% coverage), and more stable reserve estimates across accident years. Beyond statistical performance, the design emphasizes interpretability by tracing predictions to weighted contributions of base models, thereby avoiding black-box behavior. These findings highlight the practical relevance of hybrid ensemble reserving in regulated environments, offering a transparent and robust solution aligned with PSAKBI requirements. The study contributes to the literature by demonstrating how hybrid actuarial ensembles can balance methodological rigor, machine learning flexibility, and regulatory compliance in insurance reserving.
Keywords
Full Text:
PDFReferences
C. Blier-Wong, H. Cossette, L. Lamontagne, and E. Marceau, “Machine Learning in P&C Insurance: A Review for Pricing and Reserving,” Risks, vol. 9, no. 1, pp. 1–29, 2021, doi: 10.3390/risks9010004.
M. Krùpovà, N. Rachdi, and Q. Guibert, “Explainable Boosting Machine for Predicting Claim Severity and Frequency in Car Insurance,” arXiv, pp. 1–37, 2025.
G. Taylor, “Loss reserving models: Granular and Machine Learning Forms,” Risks, vol. 7, no. 3, 2019, doi: 10.3390/risks7030082.
F. Duval and M. Pigeon, “Individual Loss Reserving Using a Gradient Boosting-Based Approach,” Risks, vol. 7, no. 3, 2019, doi: 10.3390/risks7030079.
J. Pesantez-Narvaez, M. Guillen, and M. Alcañiz, “Predicting Motor Insurance Claims Using Telematics Data—XGBoost versus Logistic Regression,” Risks, vol. 7, no. 2, 2019, doi: 10.3390/risks7020070.
F. F. Jin, “Using decision tree ensemble methods for the estimation of individual claims reserving,” 2021.
M. A. Fauzan and H. Murfi, “The Accuracy of XGBoost for Insurance Claim Prediction,” International Journal of Advances in Soft Computing and its Applications, vol. 10, no. 2, pp. 159–171, 2018.
N. Kotsalo, “Machine learning methods vs. traditional methods in forecasting loss reserves,” 2021.
B. So, “Enhanced gradient boosting for zero-inflated insurance claims and comparative analysis of CatBoost, XGBoost, and LightGBM,” Scandinavian Actuarial Journal, vol. 2024, no. 10, 2024, doi: 10.1080/03461238.2024.2365390.
I. Chaoubi, C. Besse, H. Cossette, and M. P. Côté, “Micro-level reserving for general insurance claims using a long short-term memory network,” Applied Stochastic Models in Business and Industry, vol. 39, no. 3, 2023, doi: 10.1002/asmb.2750.
J. Crevecoeur, J. Robben, and K. Antonio, “A hierarchical reserving model for reported non-life insurance claims,” Insurance: Mathematics and Economics, vol. 104, 2022, doi: 10.1016/j.insmatheco.2022.02.005.
S. C. Vanegas, A. L. Badescu, and X. S. Lin, “Claim reserving via inverse probability weighting: a micro-level Chain-Ladder method,” European Actuarial Journal, vol. 15, no. 1, 2025, doi: 10.1007/s13385-024-00395-3.
P. Cai, A. Abdallah, and P. Jeganathan, “Recurrent Neural Networks for Multivariate Loss Reserving and Risk Capital Analysis,” North American Actuarial Journal, 2025, doi: 10.1080/10920277.2025.2517149.
R. R. Fauzi and J. J. Stevanlim, “Bayesian Chain Ladder For Cumulative Run-Off Triangle Under Half-Normal Distribution Assumption,” arXiv, pp. 1–9, 2024.
E. Ramos-Pérez, P. J. Alonso-González, and J. J. Núñez-Velázquez, “Stochastic reserving with a stacked model based on a hybridized Artificial Neural Network,” 2021. doi: 10.1016/j.eswa.2020.113782.
S. D. Permai and K. Herdianto, “Prediction of Health Insurance Claims Using Logistic Regression and XGBoost Methods,” Procedia Computer Science, vol. 227, pp. 1012–1019, 2023, doi: 10.1016/j.procs.2023.10.610.
L. L. Cheong, T. Meharizghi, W. Black, Y. Guang, and W. Meng, “Explainability of Traditional and Deep Learning Models on Longitudinal Healthcare Records,” arXiv, pp. 1–21, 2022.
D. Chevalier and M. P. Côté, “From point to probabilistic gradient boosting for claim frequency and severity prediction,” European Actuarial Journal, vol. 15, no. 3, 2025, doi: 10.1007/s13385-025-00428-5.
Y. Feng and S. Li, “Advancing the Use of Deep Learning in Loss Reserving: A Generalized DeepTriangle Approach,” Risks, vol. 12, no. 1, 2024, doi: 10.3390/risks12010004.
N. Kollongei and F. Onyango, “Motor Insurance Claim Frequency Prediction Using XGBoost,” Asian Journal of Probability and Statistics, vol. 26, no. 10, pp. 155–170, 2024, doi: 10.9734/ajpas/2024/v26i10665.
AAUI, “Polis Standar Asuransi Kendaraan Bermotor Indonesia (PSAKBI) revisi 2021 – ketentuan penyelesaian sengketa,” 2021.
DOI: http://dx.doi.org/10.30829/zero.v10i1.28134
Refbacks
- There are currently no refbacks.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Publisher : Department of Mathematics Faculty of Science and Technology Universitas Islam Negeri Sumatera Utara Medan | ||||
| ||||
 |  |  |  |  |