Introduction — Definition of a Bonus-Malus System

Most developed countries use several classification variables to differentiate premiums among automobile third-party liability policyholders. Typical variables include age, sex, and occupation of the main driver, the town where he resides, and the type and use of his car. More exotic variables, such as the driver’s marital status and smoking behavior, or even the color of his car, have been introduced in some countries. Such variables are often called a priori rating variables, as their values can be determined before the policyholder starts to drive. The main purpose for their use is to subdivide policyholders into homogeneous classes. If, for instance, females are proved to cause significantly fewer accidents than males, equity arguments suggest that they should be charged a lower premium. Moreover, if a company was to disregard this variable and charge an average premium to everyone, regardless of gender, most of its female policyholders would switch to another carrier, leaving the company with a disproportionate population of males and insufficient premium income to pay for claims.