The present study lines up with the current focus of the auto insurance literature. The aim
of this study can be highlighted at two levels. The first is theoretical and methodological and
aims to present synthetically the econometric modelling methodology of auto claim frequency.
The second objective is related to the empirical part of this research. Working with a French
auto insurance portfolio, we estimate an econometric model for claim frequency. On this level,
the main contribution of the study is represented by a specific set of explanatory variables that
take into account a number of updates concerning the data registered by insurance companies.
For example, in this study, we introduce as risk factors the variables occupation of insured,
GPS and value of vehicle. Also, in comparison with similar studies, we use a different
classification of the insured based on age intervals on the assumption that more homogenous
groups will be obtained and the calculation of premiums will better correspond to the reality of
studied phenomenon. Although the results cover a portfolio of a French insurance company,
the methodology of data count models can be applied to other insurance portfolios of
companies from other European countries such as Romania.