AGGREGATE IBNR CLAIM RESERVE ESTIMATION USING GENERALIZED ADDITIVE MODEL (GAM)

• Main Author: Alfarazh, Matthew
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81879
• Institution: Institut Teknologi Bandung

The management of insurance reserves is crucial for maintaining the financial stability of insurance companies, particularly in addressing IBNR (Incurred But Not Reported) claims, which are claims that have occurred but have not yet been reported. This study utilizes data from the products of a leading insurance company in Indonesia and applies the Generalized Additive Model (GAM) to estimate IBNR claim reserves. GAM is chosen for its flexibility in modeling non-linear relationships between independent and dependent variables, enabling more effective use of complex historical data. Eight modeling variations were performed, combining Gamma and Inverse Gaussian distributions, the number of knots, and smoothing functions to identify the best model. The results indicate that for product A, the best model uses Gamma distribution, 4 knots, and cubic spline basis. For product B, the best models consist of two variations: Gamma distribution, 4 knots, with cubic spline basis and penalized regression splines. Based on these optimal models, the reserves required until 2027 are IDR 591,378,610,000 for product A and an average of IDR 364,758,024,000 for product B, resulting in a total reserve requirement of IDR 956,136,634,000. Accurate reserve estimation helps insurance companies ensure their ability to address unreported claims, maintain financial balance, and minimize financial risk in the future.