Although Poisson distribution is often used to estimate the frequency of claims, the empirical evidences from literature show that it is usually too restrictive for this type of data. The fundamental problem of the Poisson distribution is that for count data the variance usually exceeds the mean, a feature known as overdispersion.