Skip to content

Univariate Risk Classification

Pricing classes within the book.

Goal: Estimate the relative costs between classes
  • Univariate: Pricing one variable at a time.
  • Assume: Rating algorithm is multiplicative. Goal is to get factors.
    • Base level is the one with relativity = 1.
    • When calculating New indicated relativities, try keeping the base relativity as 1.

Approaches

  • Pure Premium

    • Gets Indicated Relativities.

    • assumptions No exposure correlation between rating variable.

      • Also known as Distributional Bias.
      • If Territory B has a high percentage of youthful drivers, we might misclassify Territory B as having higher risk, but the real issue is the youthful drivers.
    • Adjusted Pure Premium Method

    • Loss Ratio

    • Gets Adjustment Factors.

Comments