Gillam & Snader I - Experience rating

Question 1

Describe how either of the two changes in credibility for small risks support one primary goal of experience rating (by Venter)

Answer 1

Predictive accuracy measures the difference between predicted and actual losses, and increasing excess credibility improves predictive accuracy. On a standard premium basis, the loss ratios should be less dispersed and ideally all equal.

Question 2

Three criteria for credibility

Answer 2

1. Credibility is not less than zero, nor greater than one
   0=0
2. As the size of risk increases, the % charge for any loss of a given size decreases.
   d/dE (Z/E)

Question 3

Give verbal explanation of formula M = 1 + Z* (A-E)/E

Answer 3

standard rate plus a credit or debit depending on whether A>E, stabilized by a value of Z/E

Question 4

Why a no-split plan requiring the experience mod factor to be linear function of loss does not apply well to WC.
Describe a change that can be made.

Answer 4

WC losses have heavy tail. This causes the linear constraint to produce greater squared error in estimations.
Transform losses into primary and excess losses. Then the linear constraint will produce better estimates, since primary and excess losses are both less heavy tailed.

Question 5

Discuss how using full credibility conflicts with one goal of experience rating.

Answer 5

Conflicts with individual risk equity. Goal for individual risk equity is to charge insured for the true potential loss cost. Giving full credibility to one’s experience would penalize them for having random losses that do not reflect their true expected losses.

Question 6

Off-balance is typically a slight credit. Why?

Answer 6

Largest rated risks, those with more credibility, have relatively better experience than smaller rated risks.