DETERMINING THE PURE PREMIUM USING GENERALIZED LINEAR MODELS AND MODELS COPULA IN AGRICUTURAL INSURANCE

• Main Author: FITRIANTI (NIM : 20813009); Pembimbing : Sapto Wahyu Indratno, Ph.D, HARIANI
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/20155
• Institution: Institut Teknologi Bandung

Rice farming is one of the important sectors in realizing the availability of food, but it has risk of crop failure. One of the factor that causing crop failure is climate factor, which is a natural phenomenon can not control by human. Agricultural insurance is a program made by government to protect farmers from losses that caused by crop failure. Agricultural insurance discussed in this thesis is one of insurance that cover the extensive loss of crop failure of paddy (ha), where it focused only in the Merauke Regency, especially in Semangga and Tanah Miring. The purpose of this research are (a) to model the extensive loss of crop failure of paddy (ha) which is influenced by climate factors using generalized linear models; (b) to model the relationship of dependency between the extensive loss of crop failure of paddy (ha) for Semangga and Tanah Miring using Copula; and (c) to determine the pure premium based on assuming dependent and independent risk. The extensive loss of crop failure of paddy (ha) and the pure premium obtained by assuming independent risk for both areas are smaller than from the result obtained by assuming dependent one. The extensive loss of crop failure of paddy (ha) and the pure premium obtained from generalized linear models and Copula for Tanah Miring are larger than for Semangga, where climate factor that influence both areas are rainfall. Based on these results, we can conclude that by assuming dependent, the crop failure risk of paddy for both areas are higher then by assuming independent. The crop failure risk of paddy (ha) which influenced by climate factor in the Tanah Miring is higher than in the Semangga. Thus, it can be seen that there exist dependent relation so that by assuming dependent risk is more realistic than by assuming independent one.