MUNICH CHAIN LADDER METHOD: PREDICTIVE DISTRIBUTION OF THE CLAIMS RESERVE USING BOOTSTRAPPING

• Main Author: TALIB (NIM: 20814016) , TAUFIK
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/24484
• Institution: Institut Teknologi Bandung

An insurance company need to correctly determine the prediction of the outstanding claims reserve in order to pay any claims occurred. One method often used to predict the outstanding claims reserve of a long tail business is the Standard Chain Ladder (SCL) method (Mack, 1993). This method used the runoff triangle of the paid losses data and/or the runoff triangle of the incurred losses data. One problem that may occur is that the prediction based on the paid losses data is significantly different than that obtained using the incurred losses data. Quarg dan Mack (2004) suggested the Munich Chain Ladder (MCL) method to reduce the gap between the two predictions by utilizing the correlations between past paid and incurred losses. Following Liu and Verrall (2010), bootstrapping method is used to determine the predictive distribution of the outstanding claims reserve. As a case study, the data in Dahms (2008) is used. The analysis in this thesis showed that the prediction obtained by the MCL method is better than that obtained by the SCL method. This conclusion is drawn based on the prediction error of the estimator and the predictive distribution of the outstanding claims reserve obtained by the MCL method.