APPLICATION OF COLLECTIVE RISK MODEL FOR INDONESIA NATURAL DISASTER LOSS RESERVE

• Main Author: Anderhan Torang, Amaris
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/54941
• Institution: Institut Teknologi Bandung

This final project applied collective risk model for predicting one-year-ahead Indonesia’s natural disaster loss reserve. Statistical test shows that data follows the assumption of independence and identical distribution between frequency and severity component. Modelling process is carried out separately between frequency and severity component to obtain more accurate and flexible model. Frequency component is modelled with two type of model, constant and trend parameter. Frequency modelling shows that Poisson model with trend parameter fits data the best. Severity component is modelled with heavy-tailed distribution: Lognormal, Pareto and Weibull. Severity modelling shows that Lognormal distribution fits data the best. After modelling frequency and severity, reserve for one year ahead is estimated by three risk measures: expected value, Value at Risk and Tail Value at Risk. Expected value is determined by parameters from frequency and severity model. Value at Risk is estimated by Monte Carlo Simulation with large sample, because there is no closed form for value at risk and tail value at risk for aggregate loss distribution. Value at Risk and Tail Value at Risk is estimated for three quantile level, 90% , 95% and 99%.