Variation of Individual Risks within a Class

tracyguo8tracyguo8
in Bailey.Simon

Hi,

We calculate the ratio of the n+ accident-free year credibility to the annual claim frequency for each class to determine the variation of individual hazards within a class. Would you please explain

  1. What does the ratio actually mean? Why do we divide the credibility by the annual claim frequency for each class?
  2. Why do we use the earned car years to calculate the annual claim frequency for each class, while we use EP for other calculations (frequency, relative frequency etc)?

Thanks!

Comments

  • adminadmin
    1. Bailey and Simon ask whether the credibility measured depends on how well the risks are segmented by the rating plan to begin with. If the rating plan explains all of the differences we see then dividing the credibility of the class by the annual claim frequency for the class should give a constant result across all classes. The ratio itself doesn't have a meaning, it's just used to test whether the underlying rating plan is explaining the credibility. Bailey and Simon conclude since we don't get a constant result across classes, there must be variation within the classes that is being explained/picked up by the credibility. Class 1 is far less homogeneous than classes 2-5 which is why its experience is more credible - the rating plan doesn't do as good of job at segmenting so there's more meaningful signal left over.
    2. Claim frequency = # claims / # exposures. Earned car years are used to measure exposures. The rest of the calculations are either performed using earned premium or earned car years - it depends on whether Hazam's conditions are met. For example, 2012 Q6 should not use EP.
Sign In or Register to comment.