DETERMINATION OF THE OPTIMAL RETENTION ON REINSURANCE

• Main Author: Yunus, Akmal
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39161
• Institution: Institut Teknologi Bandung

In covering the claim severity of the insurance policies, occasionally not all of the claim severity are insured by the insurance company, especially for the large claim. Consequently, the insurance company must have maximum amo- unt to be paid out for every single claim, which equals to retention. The risk measures have been exploited for determining the retention in the context of insurance such as SDPP, Value-at-Risk (VaR), Tail Value-at-Risk (TVaR), and Conditional Value-at-Risk (CVaR). Some methods for determining reten- tion are Filtered Historical Simulation, Monte Carlo Simulation, Estimative, and Improved. Furthermore, the optimal retention is chosen by checking the accuracy correct VaR or coverage probability proportion close to a given con- dence level. We analyze two data claim severity from two dierent general insurance companies during the contract period. The Lognormal distribution with parameters ^ = 2:465699 and ^2 = 2:681603 returns the best results as model of the rst data claim severity. The best model for the second data is the Skew-t distribution with parameters ^ = 0:9999994, ^ = 0:8240933, ^ = 71916290, and ^ = 1:100162. VaR using improved methods is an optimal retentions of these data because coverage probability proportion close to the condence level. Finally, we can compute insurance and reinsurance premium on reinsurance models using the retention.