MODELLING DEPENDENT INCURRED BUT NOT REPORTED CLAIMS WITH GENERALIZED ESTIMATING EQUATIONS

• Main Author: Yudha Prasietyawan, Adi
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/71061
• Institution: Institut Teknologi Bandung

Generalized linear models (GLMs) are a commonly used approach for estimating insurance claims reserves. These models assume that the claims in different origin and development years are independent variables. However, this assumption may not always hold true, leading to potentially incorrect predictions of the claim reserves or misleading estimates of the prediction error. To address this issue, this project investigates the use of generalized estimating equations (GEEs) for the estimation of claims reserves. GEEs allow for the incorporation of correlated data, such as the claims within the same accident year, which loss development triangle are treated as panel data in the analysis. A range of correlation structures can be considered within the GEE framework, and model selection criteria are proposed. Additionally, a method for estimating the mean square error of prediction for the claims reserves is derived and discussed, along with real data examples to illustrate the utility of the presented approach.