ENDOWMENT INSURANCE PREMIUM CALCULATION USING WEIBULL COX PROPORTIONAL HAZARDS SPATIAL SURVIVAL MODEL A CASE STUDY OF COVID-19 DATA IN MASSACHUSETTS

• Main Author: Christy, Gabrielle
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/65371
• Institution: Institut Teknologi Bandung

Survival model is a commonly used model to observe the survival probability of an individual/group or any other objects that becomes our center of attention. The survival probability is also driven by another risk factors. Hence, this survival model can be developed into a multiple regression model. One of the most known regression models to represent the survival probability of an individual is Cox Proportional Hazards Model. Furthermore, this model can be developed to a more complex model that consider the frailty effect. The random frailty that is based on individual’s spatial geostatistic location will be taken into account. In particular, this model will be applied to COVID-19 data in Massachusetts, United States of America. It is obtained that the frailty model with factor risks, such as age, gender/sex, respiratory rate, oxygen saturation, systolic blood pressure, and indicator variable whether the individual uses ventilator or not is the best model to represent the survival probability with an AIC of 1712,764. Moreover, the survival probability calculation will be used a base to calculate the endowment insurance premium. There are several scenarios used in the premium calculation in order to represent the real condition in our everday life. The survival probability of an individual may increase or decrease due to the effect of the frailty. Therefore, the annual premium rate that needs to be paid to the insurance company will change accordingly to the risk covered. Using the equivalence principle, the premiums for the sample policyholders are 0,5% - 10% less compared to the premiums derived from Massachussets’ mortality table if only age and gender/sex considered as the factor risks.