Modelling Motor Insurance Policy Claims Using Probability Distributions : Evidence from Ghana

Non-life insurance companies most especially, are often faced with the challenge of predicting the number of claims and claim amount to be incurred at any given time. Overcoming such challenges requires examining claim patterns, which assists in premium determination and setting of reserves. The study, therefore, modelled motor insurance claim frequency and amount with probability distributions in Ghana. Findings revealed that claim frequency is best modelled with the zero-inflated Poisson distribution. The gamma, lognormal and Weibull distributions were also identified to have the best-fit claim amount. However, the Akaike’s Information Criterion and Bayesian Information Criterion values indicated that the gamma distribution comes best in modelling claim amounts. In conclusion, the age of the driver, age of the vehicle and gender have significant influences on claim frequency. Moreover, claim amount was also influenced significantly by the driver’s age, vehicle age and gender whiles the usage or purpose of the vehicle has no significant influence on both claim frequency and amount. It is essential therefore that these factors be well considered by the company in motor insurance pricing and premium determination