MIXING DISTRIBUTION ANALYSIS OF MIXTURE POISSON DISTRIBUTION FOR THIRD PARTY LIABILITY INSURANCE CLAIM FREQUENCY DATA IN INDONESIA

Abstract

The Indonesian government plans to mandate Third Party Liability (TPL) insurance for all vehicle owners in 2025. However, statistical modeling of TPL claim frequency data in Indonesia has received limited attention in academic research. The mixture Poisson distribution can be considered as a distribution for Third Party Liability claim frequency data in Indonesia. This is because claim frequency data often experiences overdispersion. In this study, the mixing distribution of the mixture Poisson distribution for TPL claim frequency data in Indonesia will be analyzed using a bootstrap approach. The data used in this study is policyholder claim frequency data for comprehensive coverage of TPL for underwriting years 2015-2019 for vehicle categories 1, 2, 3 and 6 of PT. X in Indonesia. The results generally show that most distributions with more parameters have a larger p-value (more suitable for use as a mixing distribution for mixture Poisson distribution) than distributions with fewer parameters.