ANALYSIS OF DEPENDENCE STRUCTURE OF MOTOR VEHICLE INSURANCE CLAIMS USING COPULA MODEL Baini Sulhi, Yogo Purwono
Univeristy of Indonesia
Abstract
The growth of the insurance industry in Indonesia has growing. Reserves of claims in the insurance business are mandatory and the requirements of an insurance company to be sustainable, but in Indonesia they still apply independent assumptions. Copula is a tool that is able to describe these dependencies so that claims reserves can be calculated correctly. To achieve this goal, a simulation study is given to describe dependencies that have an impact on reserves and Solvabilities Capital Needs. In this thesis, we will discuss the application of copula-based methods to analyze the structure of claim dependencies in a vehicle insurance business line. From the results of the analysis it was found that the Partial Loss TPL and Total Loss Variables had abnormal distribution. To explain the variables or variables Partial Loss, TPl and Total Loss, the type of copula vine chosen was C-vine copula with the smallest AIC and BIC values, namely AIC of 3.169612 and BIC 11.40441 for MLE and AIC of 3.427467 and BIC of 11.66226 for Tau Kendall . Copula from the Archimedean family can be used to explain the structure of the dependencies of the three variables. Copula Archimedean from the Frank family is the best model to explain the dependency structure of Partial Loss variables with TPL and TPL with Total Loss with parameter estimates of -0.1431583 and 0.5113428. Copula Archimedean from the Gumbel family is the best model to explain the dependency structure of Partial Loss variables with Total Loss with parameter estimators of 1.0921127. The best parameter estimation is obtained by using the MLE method with an AIC value of 3.169612 and BIC 11.40441.
