this research focused on the modeling of claim sizes using various statistical distributions. Four distributions were used,namely, gamma, pareto, exponential and lognormal to model the claims from a motor portfolio in Zimbabwe. Only four
distributions were applied but there are many distributions that can be used for this purpose. Forty four data points were
obtained from a local insurer. Two methods, Classical and Bayesian were used to estimate the parameters from the data and a
comparison was also done to ascertain if one method was better than the other. Of the four distributions, pareto and lognormal
fitted properly to the data. Interestingly it was noted that the lognormal distribution produced the best fit for lower claims while
the pareto produced the best fit for huge claims. Therefore, the researchers, recommends that both distributions be applied at the
same time at those parts which they fit best. On comparing, Bayesian and Classical methods, similar results were obtained for
the pareto distribution and a poor result using Bayesian method on lognormal was obtained. Since each method has some merits
and demerits, it is wise to use both methods at the same so that the one with better results can be used for further analysis.