Errata for the 2018 Exam 8 Cookbook
Bahnemann - Inflation Effects on Excess Layers (10/3/18): In the discussion, the section “Inflation Effects on a Limited Excess Layer” should be changed to the following to be more consistent with the rest of the text:
If the excess layer has an attachment point (a) and limit (l) and we’re interested in the effective trend for limited losses within the layer (a) to (a + l), then we need to modify the claim size formula a bit:
Effective Trend on Excess Layer Claim Size:
The effective trend on the frequency of claims will be the same, but losses greater than a + l will be capped at the layer limit l.
Note: The difference is the interpretation of claims in the layer. The formula I originally showed in the discussion section is relevant for claims that specifically are between a and a + l. The interpretation above is that the claims in the layer are all the claims greater than a, but limiting the total amount at limit l.
Last year the CAS accepted either interpretation, but the formula above makes more intuitive sense and is more consistent with the rest of Bahnemann. Sorry for the confusion on this. Check out the sample solutions for CAS 2017 Q8 to get a better sense of how the CAS is interpreting it.
Robertson - Calinski-Harabasz Statistic (9/17/18): The first bullet point should say: "The trace of the dispersion matrix of class excess loss factors is 0.0470." The paper mentions that trace(S) is the total sample variance, not trace(T). Matrix T is called the dispersion matrix.
Clark - Recursive Aggregate Distribution (9/10/18): On page 144, the key formula in step 2 has an incorrect subscript. The final term should show A_k-i.
NCCI 2 - NCCI Retro Premium (9/5/18): For a problem like this, it's always good practice to show that the calculated retro premium is between the min and max premium. It is in this case, but I could've shown the quick check.
Grossi - Occurrence Exceedance Probability (9/5/18): There is a typo in step 1. OEP(3,000) should be 0.01495.
Clark - Casualty XOL Workers Comp (8/22/18): The exposure factor formula has a typo, but the calculations are correct. The ELF terms should be switched and read: Exposure Factor = ELF(AP) - ELF(AP+L)