Modelling a BMS-based automobile liability rating system: the Serbian case

Simeunović, Ivana and Domazet, Ivana and Zubović, Jovan (2016) Modelling a BMS-based automobile liability rating system: the Serbian case. Ekonomika preduzeća, 44 (5/6). pp. 361-370. ISSN 0353-443X

07 Ekonomika preduzeca 5-6-2016.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (434kB) | Preview


In this paper the authors examine the viability of applying a bonusmalus system in establishing automobile liability premium rates. They have developed an optimal tariff system which reflects the equivalence principle between the amounts of the premiums and the reported losses of individual policyholders. Based on past incidents of the policyholders, a homogeneous distribution of the risk classes and an improved insurance rating have been obtained. The developed model was tested with the negative binomial distribution and the expected value principle on a sample of 98,978 policyholders in Serbia. As a result, the set tariff system has been financially balanced, i.e. a balance was struck between the cost amounts of total premiums among policyholders and the cost amount of expected future claims. It has been proven that, in the process of modelling the automobile liability rating system based on the frequency of claims, it is optimal to use sophisticated distribution models such as the negative binomial distribution.

Item Type: Article
Additional Information: COBISS.ID=227238412
Uncontrolled Keywords: bonus–malus system, negative binomial model, Bayesian inference
M Category: H Social Sciences > H Social Sciences (General)
Depositing User: Jelena Banovic
Date Deposited: 13 Feb 2017 07:57
Last Modified: 13 Feb 2017 07:57

Actions (login required)

View Item View Item