THE ESTIMATION OF THE PURE IBNR CLAIMS USING CLAIMS FREQUENCY MODEL ON INDIVIDUAL DATA WITH THE ASSUMPTION THAT THE DISTRIBUTION OF THE ULTIMATE CLAIMS FREQUENCY IS NEGATIVE BINOMIAL

• Main Author: Beverly, Audrey
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/76261
• Institution: Institut Teknologi Bandung

A Pure IBNR (Incurred but Not Reported) claim is caused by an incident that has occurred but has not been reported to the insurance company by the time of valuation. An IBNR claim is a liability for an insurance company. To minimize this liability, an insurance company needs to set up a reserve for the Pure IBNR claims. In IFRS (International Financial Reporting Standard) 17, which will be implemented in Indonesia in the year 2025, an estimate of the liability for Pure IBNR claims must be made for each policy. As a result, the Mack's Chain Ladder method cannot be used because this method produces an aggregate reserve for the Pure IBNR claims based on the “accident year”. The focus of the research in this Final Project is to estimate the outstanding claims frequency and the reserve for the Pure IBNR claims for each insurance policy, assuming that the ultimate claims frequency follows a Negative Binomial distribution. This Final Project uses two types of data, namely: the claims data and the individually recorded policy data. From the claims data, the probability distribution of the reporting lag time may be determined; and from the policy data, the probability distribution of the frequency of unreported claims may be determined. A Monte Carlo simulation was carried out using the assumptions of both distributions; the assumption that the probability of close-with-pay (CWP) is 0.9; the assumption of the maximum frequency of the unreported claims is 10 claims; and the assumption of the probability distribution of the claims severity is a Lognormal distribution with parameters ???? = 10 and ???? = 2 which are equivalent to the mean and standard deviation of the Lognormal distribution of 162,754,791 and 1.4197?1012. An example of the simulation result of one of the policies, where the duration of the policy since it was in force is 182 days; and the duration of the policy since it was in force up to the valuation date is 1,352 days. The prediction was that: there will be one Pure IBNR claim; and the amount of reserves for the Pure IBNR claims is USD 3,600.00.